article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
it is well - known that the nonrelativistic symmetries are described by the galilei group , with the galilean boosts relating dynamically equivalent frames which move with relative constant velocity @xmath3 @xmath4 the velocity values @xmath5 are not bounded and three boost generators @xmath6 ( @xmath7 ) commute @xmath8 = 0 \ , , \ ] ] einstein s equivalence of relativistic frames is described by the following modification of ( we choose for simplicity @xmath9 , @xmath10 the lorentz boosts @xmath11 generating pseudo - orthogonal rotations in the lobachevsky planes @xmath12 are described by the deformation of the abelian algebra @xmath13 = \frac{1}{c^2}\ , m_{ij } = \frac{1}{c^2 } \ \epsilon_{ijk } \ , m_k \ , , \ ] ] where @xmath14 generate the space rotations in ( @xmath15 ) plane ( @xmath16 ) .
the relativistic transformation should be applied to moving frames if the relative velocity ration @xmath17 is not negligible .
further , from the invariance of maxwell equations under poincar symmetries follows that @xmath18 should be interpreted as the velocity of electromagnetic waves ( light velocity ) .
the poincar symmetries of special relativity theory are described by 10 generators @xmath19 which satisfy the poincar lie algebra ( to compare with we should put @xmath20 ) @xmath21 = \eta_{\mu\rho } \ , m_{\nu\tau } - \eta_{\nu\rho } \ , m_{\mu\tau } + \eta_{\nu\tau}\ , m_{\mu\rho } - \eta_{\mu\tau}\ , m_{\nu\rho } \ , , \cr\cr & & [ m_{\mu\nu } , p_{\rho } ] = \eta_{\mu\rho } \ , p_\nu - \eta_{\nu\rho}\ , p_\mu \ , , \cr\cr & & [ p_{\mu } , p_{\nu } ] = 0\ , .
\end{aligned}\ ] ] the poincar algebra was considered for almost a century after einstein s discovery in 1905 as quite uncontested way of describing the equivalent space - time frames in relativistic elementary particle physics , however under the assumption that the gravitational effects are negligible .
further in consistency with the classical poincar group structure ( abelian translation group ! ) it was assumed that the relativistic space - time is described by classical commuting minkowski space - time coordinates @xmath22 , i.e. @xmath23 = 0\ , .\ ] ] in last period however , the status of the relations and as describing ultimate geometric description of microworld physics was challenged .
let us recall here some arguments leading to the modification of einstein s relativistic symmetries and classical nature of space - time .
\i ) due to einstein s general relativity theory the space - time manifold with its geometrical structure is a dynamical entity , with its metric ( in quantum theory ) undergoing the quantum fluctuations . in particular any measurement of position due to heisenberg uncertainty relations
leads to disturbed values of the energy density , what affects through einstein s gr equations the values of gravitational field .
if we calculate the fluctuations of gravitational field generated by the energy needed to measure the distance with accuracy @xmath24 one arrives at the conclusion , that it is not possible to measure distances which are smaller than the planck length @xmath25 ( @xmath26 cm ) .
we see therefore that due to quantum gravity effects the space - time from operational point of view ceases to be a continuous manifold , but becomes a discrete set of small cells with the planck length size @xmath25 .
algebraically the noncontinuous structure of `` quantum '' minkowski space can be expressed by the dopplicher - haag - roberts ( dhr ) relation @xcite .
@xmath27 = l_p^2 \ , \theta_{\mu\nu } \ , , \qquad \theta_{\mu\nu}= - \theta_{\nu\mu } \ , , \
] ] where in simplified version of the model the tensor @xmath28 is central , i.e. one can postulate that it takes numerical values .
\ii ) other argument comes from the brane world scenario , in which we assume that the space - time manifold is described by the @xmath29-brane coordinates @xcite .
the @xmath29-branes describe the location of the end points of strings ( in general of fundamental @xmath30-branes ) .
if we accept the view that the fundamental super - strings describing the most elementary objects in universe are ten - dimensional , the @xmath29-branes are coupled not only to the gravitational background @xmath31 but as well to the nonvanishing antisymmetric tensor field @xmath32 ( here @xmath33 ) .
it was shown using canonical quantization techniques ( see e.g. @xcite ) that the quantized end point @xmath34 of the open string after first quantization do satisfy the following noncommutativity relation @xmath35 = \theta_{\mu\nu } \ , ( b(\hat{x } ) ) \ , , \ ] ] i.e. @xmath0 is a local function of @xmath32 .
if we live on @xmath29-brane , the relation describes the noncommutative @xmath36 space - time algebra , which after dimensional reduction @xmath37 implies the quantum nature of our four - dimensional minkowski space .
\iii ) third source of the space - time noncommutativity is provided by the spin degrees of freedom . already in seventies
it was shown @xcite that the space - time coordinates of superparticles satisfy the following equal time ( et ) relation @xmath38 = \frac{i}{m^2 } \ , s_{\mu\nu}\ , , \ ] ] where the relativistic spin operator @xmath39 satisfies lorentz algebra relations and can be expressed in terms of pauli - lubanski vector @xmath40 .
independently , the formula was confirmed in twistor theory , with space - time variables as composites of primary twistor coordinates @xcite .
it should be added that old snyder idea of noncommutative space - time ( see e.g. @xcite ) also relates the noncommutativity with nonvanishing angular momentum . in this
short talk we would like to recall the quantum deformation of einstein s symmetry scheme which is consistent with simple models of noncommutative space - time . because the space - time coordinates can be also described by translation sector of poincar group , in order to describe noncommutative space - time one should consider new algebraic relativistic symmetries with noncommutative symmetry parameters .
we shall consider below in sect . 2 and sect .
3 two types of such quantum symmetries : the canonical @xmath0-deformation poincar symmetries , implying the relation with the constant values of tensor @xmath0 , and the so - called @xmath1-deformed relativistic symmetries .
these two types of deformations introduce new fundamental mass parameter @xmath1 which enters the respective noncommutativity relations of minkowski space coordinates . in sect .
4 we shall mention possible prospects of deformed relativistic symmetries . it should be pointed out that the validity of quantum relativistic symmetries can be also interpreted as the particular violation of classical poincar symmetries @xcite .
for example if we consider the modification of the classical poincar mass casimir , which follows from the @xmath1-deformation of poincar algebra @xcite , we obtain @xmath41 the formula ( [ konlu11 ] ) for @xmath42 can be interpreted in two ways \i ) as a result of a particular violation of classical lorentz invariance by @xmath43 terms \ii ) the indication that the classical lorentz invariance and the lorentz transformations should be modified in a way which leads to the modified mass casimir @xmath44 as a new invariant quantity .
the advantage of second point of view lies in quite strong limitations on the ways in which the classical einstein symmetries can be violated . in this talk
we assume that the modification of classical symmetries is represented by a hopf - algebraic structure .
our hope is that in similar way as hundred years ago the einstein symmetries replaced galilei ones , in future considerations e.g. due to the quantum gravity effects the einstein symmetries will be modified and replaced by quantum relativistic symmetries .
let us expand the rhs of the general noncommutativity relation @xmath46 = \frac{1}{\kappa^2}\theta_{\mu\nu}\left(\kappa\hat{x}\right)\ , , \label{konlu11a}\ ] ] where @xmath1 is a masslike parameter , as follows : @xmath47 if in the power serie only first term is nonvanishing ( @xmath48 ) we shall call the modified poincar symmetries preserving the relation the canonically deformed or @xmath0 -deformed poincar symmetries .
such quantum symmetries were discovered quite recently @xcite@xcite and are obtained by the modification of classical poincar - hopf algebra only in the lorentz coalgebra sector by so - called twist function @xcite @xmath49 the generators @xmath50 satisfy the classical poincar algebra and the modified poincar coalgebra looks as follows @xmath51\ , , \nonumber\label{konlu13}\end{aligned}\ ] ] where ( @xmath52 ) @xmath53 and @xmath54 .
the dual canonical @xmath0-deformed poincar group constructed some years ago @xcite and reconstructed recently @xcite looks as follows @xmath55 = -\frac{i}{\kappa^2}\,\theta ^{\rho \sigma } _ { ( 0 ) } ( \hat{\lambda } _ { \ \rho } ^{\mu } \,\hat{\lambda } _ { \ \sigma } ^{\nu } -\delta _ { \ \rho } ^{\mu } \,\delta _ { \ \sigma } ^{\nu } ) \ , , \cr\cr & & [ \hat{\lambda } _ { \ \tau } ^{\mu } , \hat{\lambda } _ { \ \rho } ^{\nu } ] = [ \hat{a}^{\mu } , \hat{\lambda } _ { \ \rho } ^{\nu } ] = 0\ , , \label{konlu15}\end{aligned}\ ] ] with the coproducts remaining classical .
the twist provides an example of `` soft '' abelian deformation , with the classical @xmath56-matrix @xmath57 having the abelian carrier algebra ( @xmath58=0 $ ] ) .
the twist determines the noncommutative structure of @xmath0-deformed minkowski space @xmath59 . in accordance with general framework for twisted quantum symmetries
@xcite the coproduct formulae enter into the definition of noncommutative associative * -product for the functions @xmath60 @xmath61 in such a description the quantum minkowski space @xmath59 is a quantum representation ( a hopf module ) of canonical @xmath0-deformed poincar algebra .
the relation applied to @xmath62 and @xmath63 provides the formula @xmath64 and leads to the relation .
interestingly enough , the relation @xmath65_\ast \equiv { x}_\mu \ast { x}_\nu - { x}_\nu \ast { x}_\mu = i \theta_{\mu\nu}^{(0 ) } \ , , \label{konlu19}\ ] ] is covariant under the hopf - algebraic action of twisted poincar algebra generators . using the general formula for the action of canonical quantum symmetry generators @xmath66 and choosing @xmath67 ,
one obtains @xmath68_\ast ) - i \theta_{\mu\nu})= 0 \ , .
\label{konlu21}\ ] ] the algebraic multiplication formula after the use of classical realizations @xmath69 can be represented on commutative minkowski space as the moyal - weyl star product @xmath70 the product has been used recently quite often in order to describe the effects of the noncommutativity of space - time coordinates in classical and quantum field theory ( see e.g. @xcite ) .
the next class of quantum deformations is obtained by assuming that the linear term in is nonvanishing . in order to obtain the classical nonrelativistic physics we postulate @xmath71 = 0 \ , .
\label{konlu24}\ ] ] the only remaining @xmath72-covariant deformed relation is the one proposed firstly in the @xmath1-deformed framework of einstein s symmetries @xcite @xmath73 = \frac{i}{\kappa } \hat{x}_i \ , .
\label{konlu25}\ ] ] we see from the relations that the space coordinates remain classical , but the time variable becomes quantum .
the @xmath1-deformed relativistic symmetries preserving the relations are generated by the following @xmath1-deformed poincar - hopf algebra \a ) algebraic sector the only modified classical relation is between boosts ( we put @xmath74 i.e. @xmath75 ) and three momenta @xmath76 : @xmath77 = i \delta_{ij } [ \frac{\kappa}{2}(1-e^{- \frac{2p_0}{\kappa } } ) + \frac{1}{2\kappa } \ , \overrightarrow{p}^2 ] + \frac{1}{\kappa } \
, p_i p_j \ , , \label{konlu26}\ ] ] \b ) coalgebra sector we obtain the following set of coproducts for @xmath11 and @xmath78 , satisfying the relation @xmath79 remaining coproducts for @xmath80 and @xmath81 are primitive , i.e. @xmath82 \c ) antipodes the classical value @xmath83 is modified for the generators @xmath84 @xmath85 using the formulae
one can calculate the @xmath1-deformation of finite lorentz transformations of the fourmomenta .
the classical lorentz boost transformations @xmath86 in deformed case are generalized as follows @xmath87 where the quantum adjoint action is defined by the formula @xmath88 where @xmath89 .
choosing in @xmath90 and @xmath91 one obtains after using and ( see e.g. @xcite ) @xmath92\ , .\ ] ] if we use in we see that the formula takes the standard form .
choosing @xmath93 one obtains the differential equation for @xmath1-deformed lorentz transformations of @xmath94 in ( @xmath95 ) plane @xmath96 \ , .\ ] ] substituting in the rhs of one gets the nonlinear equation for @xmath94 , which has been solved in bicrossproduct basis firstly in @xcite ( the solution for arbitrary vector @xmath97 ( @xmath98 ) has been given in @xcite ) .
the @xmath1-deformed minkowski space as the hopf - algebra module of the @xmath1-deformed poincar algebra has been firstly considered in @xcite . using @xmath99 and the classical form of the actions @xmath100 one gets @xmath101 and subsequently @xmath102- \frac{1}{\kappa^2 } x_i ) = 0\ , .\ ] ]
similarly one can show that @xmath103=0 $ ] .
the multiplication leading to the formulae can be represented by a particular cbh * -product ( cbh@xmath104cambell - baker - haussdorf ) on commutative minkowski space @xmath22 . from
follows that ( see @xcite ) @xmath105 where @xmath106 and @xmath107 .
the cbs star product representing the multiplication rule can be described by the formula @xmath108 in particular we obtain from @xmath109 i.e. we reproduce the relation . after the discovery of @xmath1-deformation of the poincar symmetries in 1991 there was found in 1995 the light - cone deformation of poincar algebra @xcite , with one of the light - cone directions quantized , and the relations replaced by the following ones ( @xmath110 ) @xmath111 = \frac{1}{\kappa}\ , x_r \ , , \qquad [ \hat{x}_- , \hat{x}_+ ] = \frac{1}{\kappa } \hat{x}_-\ , , \cr\cr & [ \hat{x}_r , \hat{x}_s ] = [ \hat{x}_+ , \hat{x}_r ] = 0 \ , , \end{aligned}\ ] ] where @xmath112 . in 1996
both relations and were obtained as special cases of generalized @xmath1-deformation of poincar symmetries @xcite with the quantized noncommutative direction @xmath113 in deformed minkowski space .
we get the following @xmath114-dependent noncommutativity structure of quantum space - time @xmath115 = \frac{i}{\kappa } ( a_\mu \ , \hat{x}_\nu - a_\nu \ , \hat{x}_\mu)\ , .\ ] ] it follows from @xcite@xcite that only if @xmath116 , i.e. when the light - cone direction in minkowski space is quantized , the corresponding deformation of poincar symmetries can be represented by the twist factor , which modifies by the similarity transformation the classical coproduct ( see ) . in such a case only the coalgebraic sector is modified and we deal with twisted quantum relativistic symmetries .
it should be added that @xmath1-deformations were further employed from 2000 as a mathematical structure describing so - called doubly special relativistic ( dsr ) theories ( see e.g. @xcite@xcite ) .
basic aspects of the relation between the dsr framework and the @xmath1-deformation of poincar symmetries were discussed by the present author in @xcite .
the idea of noncommutative space - time can be traced back to heisenberg who first indicated that a space - time uncertainty relation might be useful for the removal of infinities in renormalization procedure .
unfortunately till present time the idea of compensating the divergent terms ( see e.g. @xcite ) in noncommutative field theory has not been successful .
also other ways of justifying the modifications of classical relativistic symmetries , based on the observation of modified kinematics in cosmic ray physics ( see e.g. @xcite ) after a period of hopeful uncertainties in the interpretation of experimental data at present are rather not confirmed .
important theoretical arguments for noncommutativity of space - time are based on quantum gravity and quantized string theory .
they indicate that the implementation of quantum symmetries and quantum modification of relativistic symmetries should be a necessary step for the description of the distances comparable with the planck length @xmath25 ( @xmath117 cm ) . also
the role of noncommutative geometry in the ultimate unification of all interactions - the @xmath118-theory - appears to be important and should be further explored .
now however the main challenge is to find among experimental data in high energy physics and astrophysics the ones which indicate the violation of the postulates of einstein special relativity theory ( e.g. constant velocity of light , classical conservation law of fourmomenta , classical energy momentum dispersion relation , isotropy of `` empty '' space - time etc . ) .
one can conjecture that these possible classical symmetry breaking effects can be recast into a new quantum relativistic symmetry invariance .
j. lukierski , h. ruegg , v.n.tolstoy , in _ quantum groups . formalism and applications _ , proc . of xxx - th karpacz winter school of theoretical physics ,
february 1994 , eds .
j. lukierski , z. popowicz and j. sobczyk , polish scientific publishers , warsaw 1995 , p. 359 | we shall outline two ways of introducing the modification of einstein s relativistic symmetries of special relativity theory - the poincar symmetries . the most complete way of introducing the modifications is via the noncocommutative hopf - algebraic structure describing quantum symmetries .
two types of quantum relativistic symmetries are described , one with constant commutator of quantum minkowski space coordinates ( @xmath0-deformation ) and second with lie - algebraic structure of quantum space - time , introducing so - called @xmath1-deformation .
the third fundamental constant of nature - fundamental mass @xmath1 or length @xmath2 - appears naturally in proposed quantum relativistic symmetry scheme .
the deformed minkowski space is described as the representation space ( hopf - module ) of deformed poincar algebra .
some possible perspectives of quantum - deformed relativistic symmetries will be outlined .
address = institute for theoretical physics + university of wrocaw , pl .
m. borna 9 , 50 - 204 wrocaw , poland |
in this paper , we present the main results from an analysis of the kinematic data obtained in the context of a massive _ absolute _ proper - motion survey , the southern proper - motion program ( spm hereafter ) .
our analysis is based on the spm catalog as described by platais et al .
this catalog provides positions , absolute proper - motions , and @xmath7 photometry for about 30,000 randomly selected stars , among other objects , near the south galactic pole .
the sky coverage of the spm catalog is about 720 deg@xmath8 in the magnitude range @xmath9 .
the accuracy of individual absolute proper - motions is 3 - 8 mas yr@xmath1 depending on the star s magnitude . for a complete description of the catalogue structure , contents , plate measurement and other astrometric / photometric details ,
the reader is referred to girard et al .
( 1998 , paper i ) and platais et al .
( 1998 , paper ii ) .
a detailed analysis of the velocity distribution function , and star- and color - counts for stars from the disk , thick - disk , and halo derived from the spm photometry and astrometry is addressed by mndez et al .
( 1999 , paper iii ) .
the arrangement of this letter is the following : in section 2 we present an overview of the spm photometric and proper - motion data . in section 3
the model used to compare the observed proper - motion distributions with those predicted is described , while section 4 presents the model predictions as compared to the proper - motion data .
section 5 presents the implications of our results for the global mass of the galaxy .
finally , section 6 presents the main conclusions of the paper .
the random sgp - spm sample has been chosen such that , at a given @xmath10 magnitude interval , a fixed number of stars are randomly extracted from all stars available in that magnitude interval . as a result , one - magnitude intervals are chosen in the range @xmath11 for the subsequent analysis . at a given magnitude interval ,
the complete sample and the randomly selected sample differ only by a scale factor
. therefore , color and kinematic properties binned in the proper magnitude intervals should be very similar to those derived from a complete sample , except for the increased uncertainties on the derived values because of the smaller sample - an effect that is fully taken into account in the analysis below .
we model the starcounts concurrently with the kinematics by using the galactic structure and kinematic model presented by mndez and van altena ( 1996 ) .
the starcount model employed here has been tested under many different circumstances , and has proved to be able to predict starcounts that match the observed magnitude and color counts ( in both shape _ and _ number ) to better than 10% , and in many cases to better than 1% .
the kinematic model presented by mndez and van altena ( 1996 ) has also been shown to be able to reproduce the kinematics of disk stars toward different galactic fields . in mndez
et al . ( 1999 ) , a full analysis of the spm - sgp data in terms of star and color - counts , as well as the kinematics is presented .
a basic assumption in the mndez and van altena ( 1996 ) model is that of the gaussian shape of the velocity ellipsoid . while the velocity distribution observed for late - type stars ( e.g. , dehnen 1998 ) is clearly skewed and can not be properly modeled by a gaussian , the error in the predicted median proper motion made by using a shifted gaussian is likely to be small
indeed , mndez et al .
( 1999 ) , show that shifted - gaussian velocity distribution functions do provide a good fit to the full spm proper - motion distributions as a function of apparent magnitude , with perhaps a slight predicted excess of stars with positive proper - motions in the direction of galactic rotation . because of projection effects near the sgp , our model comparisons are most sensitive to the motions in the direction of galactic rotation ( the so - called v - component ) , and along the galactic center - anticenter direction ( the u - component ) , and _ not _ to the motions perpendicular to the galactic plane ( the w - component ) .
there have been a number of determinations for the solar peculiar motion .
the classical result , quoted in mihalas and binney ( 1981 ) , gives @xmath12 km s@xmath1 , essentially based upon delhaye s ( 1965 ) compilation .
more recent values for the solar peculiar motion do not seem to have converged to a single value , especially in the v - component .
dehnen and binney ( 1998 ) find @xmath13 km s@xmath1 from a carefully selected unbiased sample of hipparcos stars , while miyamoto and zhu ( 1998 ) find @xmath14 km s@xmath1 from 159 cepheids , _ also _ from the hipparcos catalogue .
as suggested earlier by robin and oblak ( 1987 ) , such variation in the @xmath15 component is probably due to difficulties in separating the asymmetric drift from the intrinsic solar motion , and also because of the peculiar motions exhibited by the very young ob stars , which are still moving under the influence of the spiral arm kinematics and/or of their parent molecular cloud . in 1985 ,
the iau adopted a value for the motion of the lsr of 220 km s@xmath1 .
kerr and lynden - bell ( 1986 ) have discussed extensively the determinations of @xmath16 , as well as @xmath17 , and the oort constants a and b available until then . from a straight mean of different determinations they obtained @xmath18 km s@xmath1 ( their table 4 ) , while from independent determinations of @xmath17 ,
a , and b ( their tables 3 and 5 ) , we find @xmath19 km s@xmath1 .
although the uncertainties involved in @xmath20 are larger than those of @xmath15 , we shall see that the relative motion of disk stars is more affected by uncertainties in the solar peculiar motion than uncertainties in the motion of the lsr , since the whole nearby disk population is approximately co - moving with the lsr .
the basic standard kinematic parameters employed in the model are described in mndez and van altena ( 1996 ) and mndez et al .
the standard model adopts @xmath21 km s@xmath1 for the peculiar solar motion , and a flat rotation curve with @xmath22 km s@xmath1 . because we are analyzing data near the sgp
, our model predictions do not span a large range in galactocentric distance , and therefore the model is insensitive to the value adopted for the _ slope _ of the rotation curve near @xmath17 ( although see mndez et al .
1999 ) . as for the disk kinematics
, we have adopted the velocity dispersions given in mndez and van altena ( 1996 ) , with the scale - heights adopted in mndez et al .
( 1999 ) which reproduce the observed spm color- and magnitude counts , a scale - length of 3.5 kpc ( again , the model is not sensitive to this last parameter , as it enters as a function of the galactocentric distance , which for the spm - sgp data is quasi - constant ) , and a value of @xmath23 for the velocity ellipsoid ( see eq . 4 in mndez and van altena 1996 ) ; @xmath23 implies a velocity ellipsoid parallel to the galactic plane at all heights from the galactic plane , while @xmath24 is for a velocity ellipsoid that points towards the galactic center ) . as shown by mndez et al .
( 1999 ) the exact choice of kinematical parameters for the thick - disk and halo are not critical to the conclusions of this letter .
the mean velocity of stars with different velocity dispersion is computed via the proper moments of the collisionless boltzmann equation ( cbe ) .
the cbe allow us to relate our adopted velocity dispersions to the density laws , and to derive an expression for the velocity lag as a function of the net rotation of the disk ( the local lsr ) and the pressure support provided by the velocity dispersions ( see , e.g. , kuijken and gilmore 1989 , gilmore et al .
we must emphasize however that , if the data contain information that is inconsistent with the assumptions made ( in particular , the validity of an axisymmetric cbe , and the neglect of stellar streams ) , then these assumptions inevitably could lead to biased results
the classical problem of all parametric methods .
figures ( [ fig1lett ] ) and ( [ fig2lett ] ) show the model - predicted median absolute proper - motions for the standard model parameters compared to the observations .
these figures clearly show that , while the standard model produces a very good fit to the u - component of the proper - motion ( along the galactocentric direction ) , the v - component ( along galactic rotation ) is grossly underestimated , especially for @xmath25 , where the disk component dominates the overall kinematics .
a radical change in the model predictions occurs when we change the peculiar solar motion from the standard value of @xmath26 km s@xmath1 to @xmath27 km s@xmath1 .
the largest effect occurs at the brightest bins , i.e. , for nearby disk stars where the solar peculiar motion dominates the reflex motion . on the other hand
, we may keep the conventional solar peculiar velocity , but change the overall rotation speed for the lsr from the iau adopted value of + 220 km s@xmath1 to + 270 km s@xmath1 , as suggested by recent hipparcos results ( miyamoto and zhu 1998 ) . in this case , the change in the predicted motion becomes more important at fainter magnitudes where one is sampling objects from the other galactic components located at larger distances , and where the dominant effect is that of the overall rotation of the disk with respect to galaxies , and the relative state of rotation between the different galactic components . at the brightest bins , the effect of changing @xmath28 becomes also noticeable because of the larger fraction of bright giants , which can be seen to large distances and , hence , the differential rotation effects become amplified . a model in which we simultaneously change the solar peculiar motion _ and _ the lsr rotational speed to + 5 km s@xmath1 and + 270 km s@xmath1 ,
respectively , produces a much better fit to the observed motions .
we note that , in this model , the velocity lags for the thick - disk and halo are increased in proportion to the increase of the disk s rotational speed , as the net rotation of those two components is kept constant at 180 km s@xmath1 and 0 km s@xmath1 , respectively ( for details see mndez et al .
we conclude that a model with @xmath27 km s@xmath1 and @xmath29 km s@xmath1 clearly provides a much better fit to the median motion in v than does the standard model with @xmath30 km s@xmath1 and @xmath31 km s@xmath1 .
the predicted median proper - motion in the u - component shows a good fit to the observed values , and the changes in the v - component described above do not affect this parameter substantially because of the orthogonality of the projection effects toward the galactic poles .
these results do show us , though , that the adopted value for the solar peculiar motion in this direction is the correct one .
figure ( [ fig1lett ] ) shows the effect of changing the standard @xmath32 km s@xmath1 by @xmath333 km s@xmath1 .
an eye - ball fit from figure ( [ fig1lett ] ) suggests for the u component of the solar motion a value of @xmath34 km s@xmath1 . also , there is no indication of a local expansion or contraction of the galactic disk , as is also found from the kinematics of local molecular clouds ( belfort and crovisier 1984 ) .
a change of the orientation of the velocity ellipsoids from a cylindrical to a spherical projection tends to produce a slightly larger value for the velocity lag , especially at fainter magnitudes . the upper solid line in fig .
( [ fig2lett ] ) shows that , while the fit of the model to the observed data for @xmath35 km s@xmath1 , @xmath36 km s@xmath1 is good at bright magnitudes ( @xmath25 ) , the model underestimates the lag at fainter magnitudes .
however , by changing the orientation of the velocity ellipsoid we can actually increase the lag . this is shown in figure ( [ fig2lett ] ) where a model with @xmath27 km s@xmath1 , @xmath36 km s@xmath1 , and @xmath24 gives a better overall fit to the observed median motion in the v - component than a model with @xmath23 .
the large lsr speed suggested by the spm data poses an interesting problem to dynamical models of the galaxy .
since the lsr is , by definition , in centrifugal equilibrium with the galactic potential , a larger lsr velocity implies a larger mass for the galaxy interior to the solar circle . as emphasized by rohlfs and kreitschmann ( 1988 )
, practically all galactic - mass models follow , to within 10% , the relation : @xmath37 where @xmath38 is the mass interior to the radius @xmath39 , and @xmath40 is the circular speed at a distance @xmath39 from the galactic center .
we see that an increase in the lsr speed from 220 km s@xmath1 to 270 km s@xmath1 would imply as much as 50% more mass interior to the solar circle .
is there any other evidence that this might be plausible ?
an indication that this is perhaps not only possible but , indeed , dynamically required , comes from a recent measurement of the absolute proper - motions of the large magellanic clouds ( lmc ) by anguita et al .
they have used qsos as an inertial reference frame to obtain absolute proper - motions of stars in three fields toward the lmc .
the proper - motions are based on ccd frames taken at 7 to 12 epochs spanning eight years , with a mean accuracy for the proper - motions from these three fields of @xmath41 mas yr@xmath1 .
their derived motion for the lmc is larger than that inferred by previous photographic studies that have used either bright stars or extended galaxies as a reference frame .
anguita et al .
derive a mass for the milky way out to 50 kpc of @xmath42 , in contrast to previous determinations ( e.g. , lin and lynden - bell 1982 ) that implied a value of @xmath43 .
anguita s result is fully consistent with our findings : galactic mass models predict that the mass interior to the solar circle is only about 4% of the total galactic mass ( rohlfs and kreitschmann 1988 ) . the range of mass values found by anguita et al .
implies , through eq .
( 1 ) for @xmath44 kpc , a rotational velocity at the solar galactocentric distance in the range @xmath45 km s@xmath1 , which indeed brackets our results .
interestingly enough , a mass of @xmath46 implies a very improbable rotational speed of 120 km s@xmath1 . on the other hand
, our derived value for @xmath28 seems to be in conflict with a recent determination of the proper motion of sgr a@xmath47 by reid et al .
( 1999 ) , and backer & sramek ( 1999 ) who find @xmath48 ( employing the good assumption that sgr a@xmath47 is at rest ) , which would require @xmath17 larger than 9 kpc , inconsistent with almost all previous determinations .
could this inconsistency point , perhaps , to evidence against our basic model assumptions which would then lead to a biased @xmath28 ?
recently , zaritsky ( 1999 ) has reviewed most of the evidence regarding the mass of the galactic halo and , by extension , the total mass of the galaxy .
he points out that a squared ratio of rotational circular velocities of 1.5 ( such as that found from the 220 to 270 km s@xmath1 ) is not considered a serious discrepancy in the modeling of the total galactic mass .
he furthermore points out that timing arguments involving the relative motion of m 31 and the milky way imply a mass for the latter of @xmath49 .
a similar argument , based on the leo i - milky - way system , leads to a mass of @xmath50 . in both cases ,
the analysis excludes angular momentum , overlapping mass distributions at earlier times , and the evolution ( growth ) of the galactic halo with age , all of which would lead to an increased total mass .
zaritsky also reviews an extension of the two - body calculations into the larger environment , which leads to an even larger mass , in the range @xmath51 , concluding that all this evidence implies a mass out to @xmath52 kpc of @xmath53 .
thus , anguita s results based on the lmc proper motions , and its correspondence to our findings , are in line with zaritsky s arguments .
we have used the distribution of absolute proper motions of @xmath54 randomly selected stars , part of the southern proper motion survey , to constrain a structural & kinematic model of the galaxy .
the absolute proper - motions in the u - component indicate a solar peculiar motion of @xmath2 km s@xmath1 , with no need for a local expansion or contraction term . in the v - component
, the absolute proper - motions can only be satisfactorily reproduced by the model if we adopt a solar peculiar motion of + 5 km s@xmath1 , a large lsr speed of 270 km s@xmath1 , and a ( disk ) velocity ellipsoid that always points towards the galactic center .
the larger than expected lsr speed leads to a larger mass for the galaxy interior to the solar circle by a factor of about 1.5 .
this is compared with other recent evidence indicating that the milky - way might indeed be more massive than previously thought .
this evidence comes mainly from a new measurement of the proper motion of the lmc ( anguita et al . 1999 ) and from the binary motion ( `` timing argument '' ) of the m31 - milky way and leo i - milky way pairs ( zaritsky 1999 ) .
our larger lsr speed is also coincident with a similar finding by miyamoto and zhu ( 1998 ) using o - b5 stars and cepheids from the hipparcos catalogue .
we stress that our result of a large lsr speed rests crucially on the assumptions of both axisymmetry and equilibrium dynamics .
failure to meet these conditions would lead to inconsistencies , such as those related to the solar - galactocentric distance indicated at the end of section 4.2 .
high accuracy proper motion measurements for the smc , and proper motions for some of the nearby dwarf spheroidals will help in confirming a larger value for the mass of the galaxy , while future space interferometric space missions like fame ( http://www.usno.navy.mil/fame ) , gaia ( http://astro.estec.esa.nl/gaia/ ) , and sim ( http://sim.jpl.nasa.gov/ ) , will directly measure the disk rotational speed across the entire galaxy . the spm is a joint project of the universidad nacional de san juan , argentina and the yale southern observatory . financial support for the spm has been provided by the us nsf and the unsj through its observatorio astronmico `` flix aguilar '' .
we would like to also acknowledge the invaluable assistance of lic .
carlos e. lpez , who participated in , and supervised , all of the spm second - epoch observations .
we are indebted to an anonymous referee who provided many useful comments on the limitations and potentialities of the galactic model employed here .
these , and other comments by the referee , have greatly helped to clarify many points of the original manuscript . | predictions from a galactic structure and kinematic model are compared to the absolute proper - motions of about 30,000 randomly selected stars with @xmath0 derived from the southern proper - motion program ( spm ) toward the south galactic pole .
the absolute nature of the spm proper - motions allow us to measure not only the relative motion of the sun with respect to the local disk , but also , and most importantly , the overall state of rotation of the local disk with respect to galaxies .
the spm data are best fit by models having a solar peculiar motion of + 5 km s@xmath1 in the v - component ( pointing in the direction of galactic rotation ) , a large lsr speed of 270 km s@xmath1 , and a disk velocity ellipsoid that points towards the galactic center .
we stress , however , that these results rest crucially on the assumptions of both axisymmetry and equilibrium dynamics .
the absolute proper - motions in the u - component indicate a solar peculiar motion of @xmath2 km s@xmath1 , with no need for a local expansion or contraction term .
the implications of the large lsr speed are discussed in terms of gravitational mass of the galaxy inferred from the most recent and accurate determination for the proper - motion of the lmc .
we find that our derived value for the lsr is consistent both with the mass of the galaxy inferred from the motion of the clouds ( @xmath3 to @xmath4 kpc ) , as well as the timing argument , based on the binary motion of m31 and the milky way , and leo i and the milky way ( @xmath5 to @xmath6 kpc ) . |
the question of whether a given smooth manifold admits a riemannian metric with positive sectional curvature is nearly as old as the subject of riemannian geometry itself .
the classical gauss bonnet theorem @xmath0 relating the sectional curvature @xmath1 of a compact surface @xmath2 to its euler characteristic , @xmath3 , can be considered a first classification theorem of that kind : the only compact @xmath4manifold of positive curvature is the @xmath4sphere or the real projective plane .
moreover it poses one of the motivational questions for this article : to which degree does positive curvature determine the euler characteristic ? we shall approach an answer to this under mild symmetry assumptions . given the historic dimension of the topic , it is somewhat surprising that there exists few known examples of smooth manifolds admitting such metrics .
in fact , besides the rank one symmetric spaces @xmath5 , @xmath6 , @xmath7 and @xmath8 , there are no known simply connected manifolds in dimensions above 24 that admit metrics with positive sectional curvature .
see ziller @xcite for a survey of known examples and @xcite for recent examples in dimension seven .
the complementary task of proving that certain manifolds do not admit a positively curved metric has proven just as difficult . indeed , for simply connected manifolds that admit a nonnegatively curved metric , there is not a single example that is known not to admit positive sectional curvature . the famous conjecture of hopf that @xmath9 does not admit a positively curved metric , and the fact that it remains open , illustrates how little is known about this question .
the only known obstructions to simply connected manifolds admitting positive sectional curvature are already obstructions for weaker curvature properties .
for example , due to lichnerowicz
hitchin , a non - vanishing @xmath10genus ( respectively @xmath11invariant ) obstructs the existence of a metric of positive scalar curvature given on a compact spin manifold .
note that a metric of nonnegative scalar curvature can be deformed to a metric of positive scalar curvature .
another example is the _
gromov betti number bound _
stating that a compact manifold @xmath12 that admits a nonnegatively curved metric satisfies @xmath13 for all coefficient fields @xmath14 , where @xmath15 is a constant depending only on @xmath16 .
in particular , this implies an upper bound , albeit a large one , for the euler characteristic of @xmath2 . a conjecture corresponding nicely to this bound is due to bott .
it states that a compact @xmath16manifold admitting ( almost ) nonnegative sectional curvature is ( rationally ) elliptic .
a consequence of this would be that the betti numbers are bounded by those of the @xmath16torus , that is , the conjecture would imply that the gromov constant @xmath15 could be taken to be @xmath17 , as conjectured by gromov . for simply connected @xmath16manifolds , the bott conjecture would imply @xmath18 , as is the case for a product of two - dimensional spheres .
another crucial conjecture in the field is the conjecture by hopf claiming that the euler characteristic of a positively curved manifold of even dimension is positive .
however , general confirmations of all of these conjectures remain elusive .
fortunately , a research program initiated by karsten grove in the 1990s has been breathing new life into these problems for two decades .
the idea is to restrict one s attention to positively curved metrics that admit large isometry groups .
one important notion of large " might be that the isometry group acts transitively or with codimension one principal orbits ( see verdiani @xcite and grove wilking
ziller @xcite ) , or simply that the dimension of the isometry group is large ( see wilking @xcite ) .
for an application to the hopf conjecture , see pttmann and searle @xcite .
the measure of symmetry we will consider in this paper is the symmetry rank , which is the rank of the isometry group .
many topological classification results of varying strengths ( diffeomorphism , homeomorphism , homotopy etc . ) have been proven under the assumption that the symmetry rank is sufficiently large .
see grove @xcite or wilking @xcite for surveys , as well as theorems [ thm : grovesearle ] and [ thm : wilking03 ] for prototypical examples of such results .
our first result provides improved upper bounds for the euler characteristic in the presence of symmetry .
[ logassumption ] let @xmath12 be a closed , one - connected riemannian manifold with even dimension and positive sectional curvature .
if a torus @xmath19 of dimension at least @xmath20 acts effectively and isometrically on @xmath2 , then @xmath21 here @xmath22 denotes the @xmath23th betti number of @xmath24 , the fixed - point set of the @xmath19action on @xmath2 .
we would like to remark that the upper bound in theorem [ logassumption ] is less than @xmath25 ( for @xmath26 ) , which is asymptotically less than @xmath27 .
hence even a confirmation of the bott conjecture would not imply the theorem .
this result is consistent with the belief that a stronger upper bound for @xmath3 or @xmath28 holds for positively curved manifolds .
based on the known examples , it is possible , if a bit optimistic , to conjecture that @xmath29 for any connected , positively curved @xmath16manifold @xmath2 . since @xmath30 by conner s theorem ( see theorem [ thmconner ] ) , it would follow from this conjecture that * @xmath31 and * the number of components of @xmath24 is also at most @xmath32 .
our next result provides estimates for these two quantities under the assumption of an arbitrary linear symmetry rank bound .
[ thm : linearassumption ] let @xmath33 be an integer , and let @xmath12 be an even - dimensional , one - connected , closed riemannian manifold with positive sectional curvature . if a torus @xmath19 acts effectively by isometries with @xmath34 , then @xmath24 has at most @xmath35 components and @xmath36 where @xmath37 and @xmath38 .
we make the following remarks in order to put theorem [ thm : linearassumption ] in context : * had we allowed @xmath39 in the statement , the assumption on @xmath19 would imply @xmath40 ( for @xmath41 ) .
this contradicts the maximal possible symmetry rank ( see theorem [ thm : grovesearle ] ) , so we excluded this case . * for @xmath42
, the assumption implies @xmath43 ( for @xmath41 ) . in this case
, we already have wilking s homotopy classification ( see theorem [ thm : wilking03 ] ) .
in particular , sharp bounds hold for @xmath44 and @xmath45 .
unfortunately , the formula for the constant @xmath46 shows that @xmath47 , so we can not include the @xmath42 statement in the theorem . * for @xmath48
, the assumption implies @xmath49 . in this case
, we extend wilking s cohomology classification ( see theorem [ thm : wilking03 ] ) to hold for all @xmath16 .
all of the substantial work is contained in the original proof of theorem [ thm : wilking03 ] .
we simply adapt it to cover the cases @xmath50 using the stronger symmetry assumption ( see lemma [ linearassumptionbasecase ] ) .
+ in particular , we obtain a sharp upper bound for @xmath44 , which is reflected in our result by the fact that @xmath51 .
similarly , the sharp bound for @xmath45 holds .
this too is included in our result if we allow ourselves to set @xmath52 in this case , which is reasonable since @xmath53 as @xmath54 from above .
* the @xmath55 case is the first which provides a truly new bound . for clarity
, we restate the result in this case : if @xmath56 is a closed , one connected , positively curved manifold , and if a torus @xmath19 acts isometrically on @xmath2 with @xmath57 then @xmath24 has at most @xmath58 components and @xmath59 * for all @xmath11 , the result implies @xmath60 and @xmath61 under the assumption that @xmath62 grows linearly in @xmath63 . before entering into a discussion of the proof techniques
, we would like to state a few corollaries that follow easily from the above results , their proofs , and the tools we use in these proofs ( see section [ preliminaries ] ) . as mentioned above
, hopf conjectured that there is no metric of positive curvature on @xmath64 .
in that spirit we can provide the following restriction on the existence of positively curved metrics on cartesian products .
[ cor : stablehopf ] let @xmath12 be a closed , one connected riemannian manifold with even dimension , positive sectional curvature , and symmetry rank at least @xmath65 . * if @xmath66 , then @xmath67 . *
if @xmath68 with @xmath69 , then @xmath70 .
we remark on the name of the corollary . in the first conclusion
, we may start with and @xmath71manifold @xmath72 and consider positively curved metrics on @xmath73 with symmetry rank @xmath74 .
the conclusion in this case is that @xmath75 , which again can not be true for arbitrarily large @xmath76 .
we remark that gromov s established betti number bound , his conjectured bound , and the bound @xmath77 coming from the combined bott
hopf conjecture do not yield a statement similar to the first case .
however , they provide a bound in the second case , which we improve in the presence of symmetry .
we also remark that the key tool is the two - sided bound on the euler characteristic we get by combining theorem [ logassumption ] and corollary [ oddbettis ] below .
there is therefore a corresponding statement that uses theorem [ thm : linearassumption ] . in the second statement
, we make the assumption @xmath78 as a means of ensuring that the connected sum is nontrivial . due to the multiplicativity of the euler characteristic in fibrations , the ideas behind the corollary
allow for an application to iterated fibrations .
let us give one example of that kind : suppose given simply connected compact homogeneous spaces @xmath79 ( with @xmath80 , @xmath81 ) with @xmath82 and such that we have a family of fibrations @xmath83 for @xmath84 .
set @xmath85 .
then for @xmath76 large , @xmath86 can not carry a metric of positive curvature and symmetry rank at least @xmath87 .
see section [ corollaries ] for proofs , as well as other related statements .
the subsequent corollary is an improvement of ( * ? ? ?
* theorem a ) : [ corsymmetricspaces - short ] suppose @xmath12 has the rational cohomology of a one - connected , compact symmetric space @xmath72 . if @xmath2 admits a positively curved metric with symmetry rank at least @xmath88 , then @xmath72 is a product of @xmath89 even - dimensional spheres times either a rank one symmetric space , a rank two grassmannian @xmath90 , or a rank three grassmannian @xmath91 .
in particular , @xmath72 has rank less than @xmath92 .
the main new content of this corollary is that theorem [ logassumption ] implies an upper bound on the number of spherical factors .
we prove this corollary in section [ corollaries ] .
one generalization of hopf s conjecture on @xmath93 is that a symmetric space that admits positive curvature has rank one .
this corollary can be seen as further evidence for this conjecture .
the next corollary copes with the question under which assumptions we obtain a general confirmation of the hopf conjecture , i.e. @xmath94 .
it also provides a window to lifting the betti number information about @xmath24 to betti number information about @xmath2 : [ oddbettis ] let @xmath12 be an even - dimensional , one connected , closed riemannian manifold with positive sectional curvature .
assume @xmath19 is a torus acting effectively by isometries on @xmath2 such that 1 .
[ oddbettiass2 ] @xmath95 and @xmath96 , 2 .
[ oddbettiass3 ] @xmath97 and @xmath98 , @xmath99 , or @xmath100 vanishes , 3 . [ oddbettiass4 ] @xmath97 and the bott conjecture holds , or 4 .
[ oddbettiass5 ] @xmath24 has an isolated fixed point and @xmath2 is rationally elliptic , then @xmath101 and @xmath102 . if moreover the action of @xmath19 is equivariantly formal ,
then @xmath103 under the assumption @xmath104 .
we remark that @xmath101 already follows from the first assumption by theorem a in @xcite .
we include it here for completeness .
for the best known general result if @xmath105 , see rong and su @xcite .
we remark that the formality of the action follows immediately , if @xmath2 is rationally elliptic with positive euler characteristic , since in this case the odd betti numbers vanish .
we also remark that theorem [ thm : linearassumption ] provides another upper bound for @xmath45 under the assumption of equivariant formality . for a proof of the corollary
we refer the reader to section [ sec : oddbettis ] . as a spin - off of our considerations concerning the euler characteristic let us state a direct corollary of the proof of theorem [ logassumption ] dealing with the vanishing of the elliptic genus @xmath106 of a positively curved manifold @xmath2 with symmetry .
recall that the _ elliptic genus _
@xmath107 of an oriented connected compact manifold @xmath108 with @xmath109 is a laurent series in the variable @xmath110 with coefficients @xmath111 , @xmath112 , @xmath113 , etc .
we refer the reader to @xcite for an introduction to index theory .
[ indth ] let @xmath12 be a closed , simply connected manifold with @xmath109 .
assume @xmath2 admits a metric with positive sectional curvature and symmetry rank at least @xmath65 . 1 .
if @xmath12 is spin , then the first @xmath114 coefficients of its elliptic genus vanish unless @xmath2 has @xmath115periodic cohomology .
if @xmath2 has @xmath116 , then the first @xmath114 coefficients of its elliptic genus vanish unless @xmath2 is a rational sphere . for the proof see section [ secind ]
a statement similar to theorem [ indth ] was recently independently announced to us by nicolas weisskopf .
the techniques we use are summarized in section [ preliminaries ] .
they build upon methods originally developed in the first named author s ph.d .
thesis @xcite and the second author s ph.d .
thesis ( see @xcite ) . in particular , wilking s connectedness lemma and the resulting periodic cohomology play a role . to sharpen these tools ,
we prove a number of results that provide sufficient conditions for a manifold to have 4periodic rational cohomology ( see proposition [ pro : new4per ] , lemma [ lem : case1lemma ] , and lemma [ linearassumptionlemma2 ] ) .
a key aspect of the proofs involves finding bounds on the topology of the fixed - point set @xmath24 of the torus action . a priori
, @xmath24 might have many components of varying complexity .
we want to draw the reader s attention to the fact that the methods by which we cope with this problem in theorem [ logassumption ] are distinct from the ones we develop in theorem [ thm : linearassumption ] , where a suggestion by burkhard wilking becomes of importance .
* structure of the article . * in section [ preliminaries ]
, we review the required tools and prove a couple of general propositions using them . in section [ proofoflogassumption ] , we state theorem [ logassumptionplus ] and show how it implies theorem [ logassumption ] . in section [ prfplus ] , we prove theorem [ logassumptionplus ] .
section [ proofoflinearassumption ] is devoted to the proof of theorem [ thm : linearassumption ] , section [ sec : oddbettis ] to corollary [ oddbettis ] , section [ corollaries ] to corollaries [ cor : stablehopf ] and [ corsymmetricspaces - short ] , and section [ secind ] to theorem [ indth ] .
* acknowledgements . *
we thank burkhard wilking for a motivating discussion and nicolas weisskopf for email communication .
we begin by stating two important results referenced in the introduction .
both results will serve as the foundation for our inductive proofs over dimension .
we remark that we have restated the results in a way that will best serve us in our proofs .
[ thm : grovesearle ] if @xmath117 acts effectively by isometries on a closed , one connected , positively curved riemannian manifold , then @xmath118 with equality only if @xmath2 is diffeomorphic to @xmath5 or @xmath119 .
[ thm : wilking03 ] let @xmath12 be a closed , one - connected riemannian manifold with positive sectional curvature and symmetry rank @xmath120 . *
( homotopy classification ) if @xmath121 and @xmath122 , then @xmath2 is homotopy equivalent to @xmath5 , @xmath119 , @xmath123 , or @xmath124 . *
( cohomology classification ) if @xmath125 and @xmath126 , then 1 .
@xmath2 is homotopy equivalent to @xmath5 or @xmath119 , 2 .
@xmath2 has the integral cohomology of @xmath123 , 3 .
@xmath127 and , for all fields @xmath14 , the cohomology algebra of @xmath2 with coefficients in @xmath14 is isomorphic to either @xmath128 or @xmath129 , or 4 .
@xmath130 and , for all fields @xmath14 , the cohomology algebra of @xmath2 with coefficients in @xmath14 is isomorphic to either @xmath131 or @xmath132 .
we remark that wilking s homotopy classification was aready known for @xmath133 by theorem [ thm : grovesearle ] and for @xmath134 by the homeomorphism classification of fang and rong under the assumption of almost maximal symmetry rank , i.e. @xmath135 ( see @xcite ) . for @xmath136 ,
see lemma [ linearassumptionbasecase ] next , we state a number of results whose importance to the grove research program is already well established . [ thmconner ] if @xmath19 acts smoothly on a closed manifold @xmath2 , then 1 .
@xmath137 , 2 .
@xmath30 , and 3 .
@xmath138 .
the first statement will allow us to pull some information about @xmath24 up to @xmath2 , while the second and third results will allow us to push information about @xmath2 or its submanifolds down to @xmath24 . [
berger ] suppose @xmath19 is a torus acting by isometries on a closed , positively curved manifold @xmath12 .
if @xmath16 is even , then the fixed - point set @xmath24 is nonempty , and if @xmath16 is odd , then a codimension one subtorus has nonempty fixed - point set .
let @xmath12 be a closed , compact riemannian manifold with positive sectional curvature .
if @xmath139 and @xmath140 are totally geodesic submanifolds of @xmath2 with @xmath141 , then @xmath142 is nonempty .
the bound on the dimensions of @xmath143 and @xmath144 is equivalent to the bound @xmath145 , where @xmath146 is the codimension of @xmath147 .
this result was greatly improved by wilking @xcite : [ connectednesstheorem ] suppose @xmath12 is a closed riemannian manifold with positive sectional curvature . 1 .
if @xmath148 is a closed , embedded , totally geodesic submanifold of @xmath2 , then @xmath149 is @xmath150connected .
if @xmath151 and @xmath152 are closed , embedded , totally geodesic submanifolds of @xmath2 with @xmath153 , then @xmath154 is @xmath155connected .
recall that an inclusion @xmath156 is called @xmath157connected if @xmath158 for @xmath159 . as a consequence ,
the map @xmath160 induced by inclusion is an isomorphism for @xmath161 and a surjection for @xmath162 , and the map @xmath163 induced by inclusion is an isomorphisms for @xmath161 and an injection for @xmath164 . given a highly connected inclusion of closed , orientable manifolds , applying poincar duality to each manifold produces a certain periodicity in cohomology ( again see @xcite ) : [ wilkingpd ] let @xmath12 and @xmath148 be closed , connected , oriented manifolds .
if the inclusion @xmath149 is @xmath165connected , then there exists @xmath166 such that the map @xmath167 induced by multiplication by @xmath168 is a surjection for @xmath169 and an injection for @xmath170 .
this periodicity is especially strong when @xmath171 .
for example , because @xmath2 is connected , @xmath172 generates @xmath173 for @xmath174 .
we call this property periodicity ( see @xcite )
: for an integer @xmath175 , a ring @xmath176 , and a connected space @xmath2 , we say that @xmath177 is _ @xmath76periodic _ if there exists @xmath178 such that the map @xmath179 induced by multiplication by @xmath168 is surjective for @xmath180 and injective for @xmath181 . examples of manifolds with periodic cohomology include spheres , projective spaces , and some products such as @xmath182 or @xmath183 for any manifold @xmath72 of dimension less than eight . in practice , we conclude periodicity from the connectedness lemma and theorem [ wilkingpd ] .
specifically , if @xmath143 and @xmath144 intersect transversely as in part ( 2 ) of the connectedness lemma , then @xmath184 is @xmath185periodic .
it follows that the rational cohomology of @xmath144 is also @xmath185periodic .
the following refinement of this rational periodicity was proven in @xcite : [ per ] let @xmath12 be a closed , one connected riemannian manifold with positive sectional curvature .
let @xmath151 and @xmath152 be connected , closed , embedded , totally geodesic submanifolds that intersect transversely . if @xmath186 , the rational cohomology rings of @xmath187 , @xmath188 , @xmath189 , and @xmath190 are @xmath115periodic .
while the connectedness and periodicity theorems have no symmetry assumptions , their main use has been to the study of positively curved manifolds with large symmetry . indeed ,
fixed point sets of isometries are totally geodesic , and with enough isometries , one can guarantee the existence of totally geodesic submanifolds with small codimension and pairs of submanifolds that intersect transversely .
we will return to how one does this in a moment . for now , we state an application of these ideas proven in ( * ? ? ?
* theorem 2.2 ) : [ thmp ] let @xmath191 be even integers , and let @xmath12 be a closed , one connected manifold with positive sectional curvature .
assume @xmath19 is a torus acting effectively by isometries on @xmath2 such that @xmath192 for all @xmath193 , there exists @xmath194 such that @xmath195 satisfies 1 .
@xmath196 is @xmath115periodic , 2 .
@xmath197 is @xmath198connected , 3 .
@xmath199 , and 4 .
@xmath200 . in particular ,
if @xmath201 , there exists @xmath202 such that the maps @xmath203 induced by multiplication by @xmath168 are surjective for @xmath204 and injective for @xmath205 .
we will use this theorem in the proof of corollary [ oddbettis ] . for
now , we combine it with theorem [ wilkingpd ] to prove the following observation , which will serve as a convenient tool in our proofs of theorem [ thm : linearassumption ] : [ pro : new4per ] let @xmath12 be a closed , one connected manifold with positive sectional curvature . if @xmath2 both contains a totally geodesic submanifold of codimension @xmath206 and has symmetry rank at least @xmath207 , then @xmath208 is @xmath115periodic . by theorem [ wilkingpd ] , the existence of a codimension @xmath76 totally geodesic submanifold implies that there exists @xmath209 such that the maps @xmath210 given by multiplication by @xmath211 are injective for @xmath212 and surjective for @xmath213 .
next by theorem [ thmp ] with @xmath214 , the symmetry rank assumption implies the existence of @xmath202 such that the maps @xmath203 are surjective for @xmath215 and injective for @xmath216 . in particular , @xmath217 for some @xmath218 .
we claim that @xmath168 induces periodicity in @xmath208 .
we first show that it suffices to show injectivity of the multiplication maps @xmath203 for @xmath219 . indeed
, poincar duality would imply that the betti numbers satisfy @xmath220 for all @xmath219 , hence the injective maps @xmath203 must be isomorphisms for dimensional reasons . since the map @xmath221
is already known to be surjective by our choice of @xmath168 , the proof would be complete . we now prove that multiplication maps @xmath222 are injective for @xmath223 . since @xmath224 ,
the four cases that follow constitute a proof : * if @xmath216 , then the map @xmath203 is injective by our choice of @xmath168 . *
if @xmath225 , the injectivity of @xmath226 follows from two facts .
the first is that @xmath227 is injective for these values of @xmath23 .
the second is that multiplication by @xmath211 factors as multiplication by @xmath168 followed by multiplication by @xmath228 .
* if @xmath229 , the injectivity of the map @xmath203 follows from the fact that @xmath230 is a composition of injective maps that equals the composition @xmath231 the first map of which is an isomorphism .
* if @xmath232 , we prove injectivity of the map @xmath203 directy .
let @xmath233 be a nonzero element .
by poincar duality , there exists @xmath234 such that @xmath235 .
the estimate @xmath236 implies that @xmath237 for some @xmath238 by the surjectivity property of @xmath168 .
hence @xmath239 which implies @xmath240 , completing the proof .
the final tool we need to introduce , used for the first time in the subject in @xcite , is the theory of error - correcting codes .
the specific result we use from this theory is the griesmer bound . if @xmath241 is a linear map such that the image of every nonzero element has hamming weight at least @xmath242 , then @xmath243 lemma 1.8 in @xcite uses this bound to prove the existence of totally geodesic submanifolds @xmath72 such that the inclusions @xmath244 are @xmath198connected .
we adapt this proof , at the suggestion of b. wilking , to prove the existence of such a submanifold which has the additional property that it contains a controlled number of fixed points in @xmath24 .
[ lem : griesmer ] let @xmath245 be even .
suppose a torus @xmath117 acts effectively and isometrically on a closed , connected , positively curved manifold @xmath12 .
if @xmath246 are fixed points of the @xmath117action on @xmath2 , and if @xmath247 then there exists a nontrivial involution @xmath248 and a component @xmath249 such that @xmath72 contains @xmath250 of the @xmath251 .
the @xmath252 statement is essentially lemma 1.8 in @xcite , and it implies that we can cover each fixed - point component of @xmath24 by a component of a fixed - point set of an involution with dimension at least @xmath253 . an argument using the fact @xmath117 has only @xmath254 involutions , together with frankel s theorem , will be the key step in the proof of theorem [ logassumption ] . the conclusion for general @xmath255 will be the key step in proving the first conclusion of theorem [ thm : linearassumption ]
( see lemma [ lem : linearassumptionpart1 ] ) . by iterated
used of this proposition , one could find involutions @xmath256 and components @xmath257 such that * @xmath258 , * @xmath259 for all @xmath23 , and * @xmath260 contains all @xmath261 .
a variation of this idea will be used in the proof of the second assertion of theorem [ thm : linearassumption ] ( see lemma [ lem : linearassumptionpart2 ] and its proof ) . choose a basis for each @xmath262 so that the isotropy representation @xmath263 factors through @xmath264 .
it follows that the @xmath265 maps via the isotropy representation to a copy of @xmath266 .
moreover , the map @xmath267 has the property that the hamming weight of the image of @xmath268 is half of the codimension of @xmath269 . consider the direct sum @xmath270 of the isotropy representations at @xmath271 .
restricting to the involutions in @xmath117 , we obtain a homomorphism @xmath272 let @xmath273 denote the projection of the direct sum onto the @xmath23th summand of @xmath274 .
the hamming weight of the image of @xmath268 under the composition of the isotropy representation and @xmath273 is half the codimension @xmath146 of @xmath269 .
set @xmath275 .
we claim that a nontrivial @xmath276 exists with @xmath277 . assuming this for a moment
, we conclude the proof .
relabel the @xmath261 so that the codimensions @xmath278 are increasing in @xmath279 . using frankel s theorem together with the estimate @xmath280 for all @xmath281 ,
we conclude that the components @xmath282 and @xmath283 intersect .
hence they are , in fact , the same component .
since we also have @xmath284 we conclude the component @xmath285 of @xmath286 satisfies @xmath287 and contains at least @xmath250 of the @xmath261 .
this concludes the proof given the existence of @xmath276 as above .
we proceed by contradiction to show that some nontrivial @xmath276 exists with @xmath277 .
suppose therefore that @xmath288 for all nontrivial @xmath268 .
applying griesmer s bound , we obtain @xmath289 we estimate the sum on the right - hand side as follows : @xmath290 rearranging , we conclude that @xmath291 but this contradicts the assumed bound on @xmath120 , so the result follows .
in this section , we state theorem [ logassumptionplus ] , which is the most general version of theorem [ logassumption ] we can prove .
we then state lemma [ logassumptionlemma ] and show how to derive theorem [ logassumption ] from these results . finally , we comment on the quality of the approximation provided by theorem [ logassumption ] , illustrate the result graphically , and prove lemma [ logassumptionlemma ] . the proof of theorem [ logassumptionplus ]
is postponed until section [ prfplus ] . in order to state theorem [ logassumptionplus ]
, we require two definitions : set @xmath292 for @xmath41 , set @xmath293 for @xmath294 , and define @xmath295 for @xmath296 by the recursive formula @xmath297 [ logassumptionplus ] let @xmath12 be a connected , closed , even - dimensional riemannian manifold with positive sectional curvature . if a torus @xmath19 of dimension at least @xmath20 acts effectively and isometrically on @xmath2 , then @xmath298 we will combine this theorem with lemma [ logassumptionlemma ] below , the purpose of which is to estimate @xmath295 . in order to state the lemma ,
we require two more definitions : * set @xmath299 and , for @xmath300 , let @xmath301 be the minimal integer satisfying @xmath302 * set @xmath303 and , for @xmath304 , set @xmath305 see table [ table01 ] for some ( approximate ) values of @xmath301 and @xmath306 .
[ logassumptionlemma ] for all @xmath307 and @xmath308 , @xmath309 combining theorem [ logassumptionplus ] with the @xmath310 statement of lemma [ logassumptionlemma ] , we easily obtain a proof of theorem [ logassumption ] . .some approximated values [ cols="<,^,<",options="header " , ] before proceeding to the proof of lemma [ logassumptionlemma ] , we wish to remark on the shape of our upper bound on @xmath45 .
we shall compare the function @xmath311 with the approximation given in theorem [ logassumption ] against the background of an exponential bound .
for this note that it is easily possible to modify the arguments we provide in order to obtain the exponential function we draw in the pictures below as an alternative upper bound . from the pictures already it should be obvious why we did not spend more effort on this .
already from the graphs in figures [ fig1 ] and [ fig2 ] , one sees that the functions grow strictly subexponentially . as for the asymptotic behavior , the exponential reference function
may easily be chosen such that the power of two is a genuine fraction of @xmath16 .
( blue ) and @xmath312 ( red ) under a logarithmically scaled @xmath211-axis respectively both axes being logarithmically scaled . as a reference
the function @xmath313 , which also bounds @xmath311 , ( black ) is added.,title="fig:",scaledwidth=47.0% ] ( blue ) and @xmath312 ( red ) under a logarithmically scaled @xmath211-axis respectively both axes being logarithmically scaled . as a reference
the function @xmath313 , which also bounds @xmath311 , ( black ) is added.,title="fig:",scaledwidth=47.0% ] ( blue ) and @xmath314 ( red ) under a logarithmically scaled @xmath211-axis respectively both axes being logarithmically scaled . as a reference
the function @xmath313 , which also bounds @xmath311 , ( black ) is added.,title="fig:",scaledwidth=47.0% ] ( blue ) and @xmath314 ( red ) under a logarithmically scaled @xmath211-axis respectively both axes being logarithmically scaled . as a reference
the function @xmath313 , which also bounds @xmath311 , ( black ) is added.,title="fig:",scaledwidth=47.0% ] we give the proof for @xmath315 , and we omit it for @xmath316 . first , the @xmath310 statement follows for @xmath317 by definition of @xmath295 . for @xmath318
, the estimate follows from the @xmath319 statement , the proof of which we give now .
computing the finitely many even values for @xmath320 $ ] shows that @xmath321 for all @xmath16 in this interval .
consequently , we may proceed by induction using the fact that the recursion formula from @xmath322 onwards will only use the values @xmath295 with @xmath296 .
thus , we may compute @xmath323 for @xmath324 as claimed .
this proves the estimate for @xmath325 and hence completes the proof in cases @xmath319 and @xmath310 .
we begin by proving the following sufficient condition for @xmath208 to be 4periodic . for this
we need to set up the following notation : let @xmath19 be a torus acting on a manifold @xmath2 . for a submanifold @xmath244 on which @xmath19 acts ,
we let @xmath326 denote the dimension of the kernel of the induced @xmath19action on @xmath72 .
[ lem : case1lemma ] let @xmath12 be a closed , one - connected riemannian manifold with positive sectional curvature , and let @xmath19 be a torus acting isometrically on @xmath2 .
assume @xmath24 is nonempty , and let @xmath193 .
if there exists a subgroup @xmath327 with @xmath328 such that every @xmath329 satisfies @xmath330 and @xmath331 , then @xmath208 is @xmath115periodic .
inductively choose a @xmath279 algebraically independent involutions @xmath332 such that @xmath333 for all @xmath334 , where @xmath335 for all @xmath336 and @xmath337 .
by maximality of @xmath146 , we have both of the estimates @xmath338 where @xmath339 is the dimension of the intersection of the @xmath340eigenspaces of the actions of @xmath341 and @xmath342 on the normal space of @xmath343 .
geometrically , @xmath339 is the codimension of @xmath344 inside @xmath345 .
together , these two estimates imply that @xmath346 for all @xmath307 .
we draw two conclusions from this .
first , @xmath347 since @xmath348 is even , we conclude that @xmath349 .
second , for all @xmath350 , @xmath351 so we can conclude both of the estimates @xmath352 and @xmath353 we will use the first estimate in a moment .
the second estimate implies that the inclusion @xmath354 is @xmath198connected with @xmath355 . as a consequence ,
if @xmath356 is @xmath115periodic for some @xmath157 , then @xmath208 would be @xmath115periodic as well .
we will use this fact later .
let @xmath357 be the the minimal index such that @xmath358 .
for @xmath359 , let @xmath360 denote the dimension of the intersection of the @xmath340eigenspaces of the action of @xmath361 and @xmath362 on @xmath363 .
geometrically @xmath360 is the codimension of @xmath364 inside @xmath365 . by replacing @xmath362 by @xmath366 for some @xmath367
, we may assume that @xmath368 for all @xmath369 .
let @xmath359 be the maximal index such that @xmath370 .
observe that some @xmath371 since the subgroup @xmath372 acts effectively on @xmath2 .
it follows that @xmath373 is the transverse intersection of @xmath374 and @xmath375 inside @xmath376 . since @xmath377 is equal to @xmath378
, we conclude that @xmath379 is @xmath115periodic by the periodicity theorem .
as already established , this implies that @xmath208 is @xmath115periodic .
we proceed to the proof of theorem [ logassumptionplus ] , keeping the definitions of @xmath380 and @xmath295 from section [ proofoflogassumption ] .
this will finally establish the proof of theorem [ logassumption ] .
first , the assumption @xmath381 implies that @xmath40 for @xmath382 . as this contradicts the maximal possible symmetry rank ( theorem [ thm : grovesearle ] ) , the theorem is vacuously true in these dimensions .
moreover , the theorem continues to hold for dimensions @xmath294 by wilking s homotopy classification theorem ( theorem [ thm : wilking03 ] ) and the definition of @xmath311 .
we proceed by induction to prove the result for @xmath296 .
set @xmath383 , and choose any @xmath384 .
we split the proof of the induction step into two cases .
case 1 : : there exists @xmath385 such that every @xmath386 with @xmath387 and @xmath388 actually has @xmath389 .
case 2 : : for all @xmath385 , there exists @xmath390 with @xmath387 and @xmath391 . fix some @xmath193 as in the statement of case 1 .
let @xmath392 be maximal such that there exists a @xmath393 with the property that every @xmath394 has @xmath395 and @xmath396 .
we claim that @xmath397 .
assuming this for a moment , lemma [ lem : case1lemma ] would imply that @xmath208 is 4periodic .
in particular , poincar duality would imply that every even betti number is at most one , so conner s theorem and the definition of @xmath311 would let us conclude the proof by estimating as follows : @xmath398 we prove now that @xmath397 . to do this ,
suppose instead that @xmath399 .
choose a @xmath400 so that @xmath401 .
observe that @xmath402 so lemma [ lem : griesmer ] applies with @xmath252 and @xmath403 .
choose a nontrivial @xmath404 with @xmath405 .
suppose for a moment that @xmath406 .
then there exists an an involution @xmath407 such that @xmath408 with both inclusions strict .
it follows that @xmath409 is the transverse intersection of @xmath410 and @xmath411 .
since @xmath412 the periodicity theorem implies @xmath208 is @xmath115periodic .
we may assume therefore that @xmath413 and hence , by the assumption in case 1 , that @xmath414 .
fix for a moment any @xmath394 .
observe that @xmath415 , so @xmath416 as in the proof that @xmath417 , we can conclude without loss of generality that @xmath418 .
but now the assumption in case 1 implies that @xmath419 .
hence every @xmath420 in the @xmath421 generated by @xmath372 and @xmath422 satisfies @xmath387 and @xmath423 .
this contradicts the maximality of @xmath279 and hence concludes the proof that @xmath424 .
if case 1 does not occur , it is clear that case 2 does , so we conclude the proof of the bound @xmath425 by proving it under the assumptions of case 2 .
we can not calculate @xmath208 in this case 2 .
instead , we cover @xmath24 by at most @xmath426 submanifolds , each of which satisfies the induction hypothesis .
adding together the estimates for each of these submanifolds and applying the recursive definition of @xmath311 , we conclude the desired bound . throughout the proof , we will use the notation @xmath427 for the sum of the even betti numbers of a space @xmath428 .
suppose @xmath24 has @xmath255 components , and choose @xmath429 , one in each component .
the assumption in case 2 implies that there exist nontrivial involutions @xmath430 such that @xmath431 and @xmath432 .
set @xmath433 for all @xmath279 .
the @xmath19action on @xmath2 restricts to a @xmath19 action on each @xmath434 . after dividing by the kernel of the induced action
, we obtain an effective action by a torus @xmath435 on @xmath434 with @xmath436 .
since @xmath381 and @xmath437 , we see that @xmath438 .
it follows from the induction hypothesis that @xmath439 .
but the components of @xmath440 are precisely the components of @xmath24 that also lie in @xmath434 .
moreover , @xmath441 , so @xmath442 , since both dimensions are even .
hence , we obtain @xmath443 we will use this estimate in a moment .
the key observation is that the set @xmath444 has at most @xmath445 elements .
moreover since each @xmath446 , we can conclude from frankel s theorem that the set @xmath447 also has at most @xmath448 elements .
let @xmath449 denote an index set in one - to - one correspondence with the set of @xmath434 .
since every component @xmath450 is contained in some @xmath451 with @xmath452 , we have the estimate @xmath453 applying the estimate above and the recursive definition of @xmath311 , we see that @xmath454
recall that we are given an even - dimensional , closed , connected manifold @xmath12 .
we have a fixed integer @xmath455 , and we have an effective , isometric action by a torus @xmath19 on @xmath2 such that @xmath456 where @xmath457 our task is threefold .
first we prove in lemma [ linearassumptionbasecase ] that @xmath48 implies that @xmath458 is @xmath115periodic for all fields @xmath14 .
recall that this automatically implies * @xmath459 and * @xmath460 where we interpret @xmath461 as 0 since @xmath462 as @xmath54 from above .
second , for @xmath463 , we claim that the number of components @xmath44 of @xmath24 is at most @xmath35 where @xmath37 .
this is the first part of theorem [ thm : linearassumption ] and is proven in lemma [ lem : linearassumptionpart1 ] .
finally the second part of theorem [ thm : linearassumption ] , namely , that @xmath464 where @xmath465 , is proven in lemma [ lem : linearassumptionpart2 ] .
we first prove the @xmath48 statement .
the proofs of both parts of theorem [ thm : linearassumption ] proceed by induction over @xmath11 .
this lemma therefore serves as a common base case .
[ linearassumptionbasecase ] assume @xmath12 is a closed , one - connected riemannian manifold with positive sectional curvature . if a torus @xmath19 acts effectively by isometries on @xmath2 with @xmath466 then 1 .
@xmath2 is homotopy equivalent to @xmath5 or @xmath119 , 2 .
@xmath2 has the integral cohomology of @xmath123 , 3 .
@xmath127 and , for all fields @xmath14 , the cohomology of @xmath2 with coefficients in @xmath14 is isomorphic to either @xmath128 or @xmath129 , or 4 .
@xmath130 and , for all fields @xmath14 , the cohomology of @xmath2 with coefficients in @xmath14 is isomorphic to either @xmath131 or @xmath132 .
in particular , @xmath2 has @xmath115periodic rational cohomology .
we induct over dimension . for @xmath467 ,
the theorem is trivial . for @xmath468 , @xmath469
, hence the result follows from the grove and searle s diffeomorphism classification ( theorem [ thm : grovesearle ] ) . for @xmath470 ,
the symmetry rank is at least @xmath43 , so the theorem follows from wilking s homotopy classification ( theorem [ thm : wilking03 ] ) . in each of these cases ,
the conclusion of the lemma follows from the universal coefficients theorem .
suppose now that @xmath471 .
the assumption on @xmath19 implies @xmath472 by proposition 8.1 in @xcite , if there exists an involution @xmath407 with @xmath473 , then the conclusion of the lemma holds .
it therefore suffices to prove the existence of such an involution . to see that such an involution exists , we adapt the proof of corollary 3.2.a in @xcite .
let @xmath19 denote the torus acting isometrically on @xmath2 . by berger s theorem
, there exists @xmath474 such that @xmath117 has a fixed point @xmath475 , where @xmath476 and where @xmath477 if @xmath16 is even and @xmath478 if @xmath16 is odd .
we consider the isotropy representation @xmath479 at @xmath168 .
the dimension of the @xmath340eigenspace of the image of an involution @xmath480 in @xmath481 is equal to the codimension of @xmath410 .
the proof of corollary 3.2.a in @xcite states that such an nontrivial involution @xmath248 exists with @xmath482 if @xmath483 where @xmath484 .
the original proof concludes that this is the case for @xmath485 and @xmath486 , and in particular for @xmath487 and @xmath126 . in our case ,
the assumption on @xmath120 in terms of @xmath71 is that @xmath488 except for @xmath489 and @xmath490 , this estimate holds for all @xmath471 . for @xmath489 and @xmath491
, one can instead apply the griesmer bound to find the desired involution . in all cases therefore , if @xmath471 , the involution @xmath420 with @xmath482 exists .
this concludes the proof of the lemma .
with the base case ( @xmath48 ) complete , we pause for a moment to prove a couple of lemmas that will facilitate our induction over @xmath11 . the third estimate of lemma [ salphaestimates ] together with lemma [ linearassumptionlemma2 ] already suggest how this will work .
[ salphaestimates ] for @xmath463 , @xmath492 satisfies the following : 1 . if @xmath493 , then @xmath494 .
2 . if @xmath495 , then @xmath496
. 3 . if @xmath497 , then @xmath498 . the estimate @xmath499 is equivalent to @xmath500 the logarithm term is decreasing in @xmath11 , so it is at most @xmath501 .
in particular , @xmath494 holds if @xmath502 , which is to say if @xmath493 . for
the second estimate , suppose that @xmath495 .
observe that @xmath463 implies @xmath503 hence @xmath504 on the other hand , @xmath505 and @xmath495 imply that @xmath506 combining this estimate with the one above for @xmath492 , we conclude the proof of the second estimate as follows : @xmath507 finally , if @xmath508 , then @xmath509 .
hence we conclude the third estimate as follows : @xmath510 [ linearassumptionlemma2 ] assume @xmath12 is a closed , one - connected riemannian manifold with positive sectional curvature , and assume @xmath19 is a torus acting isometrically on @xmath2 with @xmath511 and @xmath505 . if there exists a nontrivial involution @xmath407 and a point @xmath512 such that @xmath513 or such that @xmath514 and @xmath515 , then @xmath2 has @xmath115periodic rational cohomology . here , we are keeping the notation from section [ prfplus ] : @xmath516 denotes the dimension of the kernel of the induced @xmath19action on @xmath410 .
choose @xmath420 and @xmath168 as in the assumption of the lemma .
if @xmath517 is at most @xmath518 , then proposition [ pro : new4per ] implies that @xmath2 has @xmath115periodic rational cohomology .
indeed , the proposition applies since @xmath463 implies that @xmath519 and hence that @xmath520 on the other hand , if @xmath521 and @xmath515 , then we can write @xmath410 as a transverse intersection and apply the periodicity theorem . to see this ,
observe that @xmath410 is fixed by a @xmath4dimensional torus inside @xmath19 .
choose any nontrivial involution @xmath422 in this 2torus not equal to @xmath420 .
it follows that @xmath410 is the transverse intersection of @xmath409 and @xmath522 .
moreover , effectivity of the @xmath19action implies that the codimensions of these manifolds are positive , and the assumption @xmath523 implies @xmath524 we therefore conclude the lemma from the periodicity theorem .
we are ready to prove the first part of the proof of theorem [ thm : linearassumption ] : [ lem : linearassumptionpart1 ] let @xmath12 be a closed , one connected riemannian manifold with positive sectional curvature , and assume @xmath19 is a torus acting by isometries on @xmath2 . if @xmath33 and @xmath456 , then @xmath525 where @xmath37 . by lemma [ linearassumptionbasecase ]
, the theorem follows in the base case of @xmath48 .
we proceed by induction the prove the theorem for @xmath505 . by lemma [ salphaestimates ] and the induction hypothesis , we may we may assume that @xmath495 . in particular , by lemma [ salphaestimates ]
again , we may assume that @xmath526 next observe that if @xmath2 has @xmath115periodic rational cohomology , then the even betti numbers of @xmath2 are at most one
. hence we can conclude the theorem using conner s theorem : @xmath527 in particular , by lemma [ linearassumptionlemma2 ] , we may assume that , for all nontrivial involutions @xmath407 and for all @xmath512 , if @xmath528 , then @xmath529 and @xmath530 .
we proceed by contradiction .
assume there exist @xmath531 distinct components of @xmath24 .
the estimate on @xmath492 established at the beginning of the proof implies @xmath532 by proposition [ lem : griesmer ] ( with @xmath533 ) , there exists a nontrivial involution @xmath407 and a component @xmath249 with @xmath534 such that @xmath72 contains at least @xmath535 of the @xmath536 components of @xmath24 . as established at the beginning of the proof
, we may assume that @xmath537 and @xmath530 .
let @xmath538 , and observe that @xmath539 acts effectively and isometrically on @xmath72 .
moreover the components of @xmath540 are the components of @xmath24 that are contained in @xmath72 . in particular , the number of components of @xmath540 is at least @xmath535 . on the other hand , since @xmath541 by lemma [ salphaestimates ] , the induction hypothesis implies that the number of components of @xmath540 is at most @xmath542 . putting these estimates together ,
we conclude that @xmath543 and hence that @xmath544 .
this contradicts our definition of @xmath536 , so the proof is complete . with
the proof of the first part of theorem [ thm : linearassumption ] complete , we spend the remainder of the section on the proof of the second part : [ lem : linearassumptionpart2 ] let @xmath12 be a closed , one - connected riemannian manifold with positive sectional curvature , and assume @xmath19 is a torus acting by isometries on @xmath2 . if @xmath33 and @xmath511 , then @xmath545 where @xmath546 .
the first step of the proof is to find submanifolds that have @xmath115periodic cohomology and that cover multiple components of @xmath24 .
the second step is to use these submanifolds together with conner s theorem to conclude the upper bound on @xmath547 . to quantify the first step , we prove the following : [ lem : linearassumptionpart2chains ] let @xmath12 be a closed , one - connected riemannian manifold with positive sectional curvature , and assume @xmath19 is a torus acting isometrically on @xmath2 . if @xmath455 and @xmath456 , then there exist @xmath548 and a chain @xmath549 of submanifolds satisfying 1 .
every @xmath550 contains @xmath551 components of @xmath24 for some @xmath552 , 2 .
@xmath19 acts on every @xmath550 and @xmath553 , and 3 .
@xmath554 has @xmath115periodic rational cohomology .
clearly the chain @xmath555 of length 0 satisfies properties ( 1 ) and ( 2 ) .
let @xmath556 denote a maximal chain satisfying properties ( 1 ) and ( 2 ) .
first note that the maximality of @xmath279 implies @xmath557 .
indeed if this were not the case , then we could define @xmath558 and conclude from lemma [ salphaestimates ] that the chain @xmath559 satisfies properties ( 1 ) and ( 2 ) .
next note that , if @xmath560 , then the subchain @xmath561 has properties ( 1)(3 ) since @xmath562 which implies that @xmath563 has @xmath115periodic rational cohomology by lemma [ linearassumptionbasecase ] .
we may assume therefore that @xmath564 .
we show in this case that @xmath554 has @xmath115periodic rational cohomology .
we will do this by choosing an involution @xmath407 and a component @xmath565 of @xmath566 , using the maximality of @xmath279 to conclude that @xmath565 does not satisfy properties ( 1 ) and ( 2 ) , then using this information to conclude that @xmath554 has @xmath115periodic rational cohomology .
our first task is to choose the involution @xmath420 .
set @xmath567 .
note that @xmath539 is a torus acting effectively on @xmath554 with dimension at least @xmath568 by property ( 2 ) . by property ( 1 ) together with the first part of theorem [ thm : linearassumption ] ( or lemma [ lem : linearassumptionpart1 ] )
, we have @xmath569 in particular , @xmath570 where the last inequality holds by lemma [ salphaestimates ] and our assumption that @xmath571 .
this estimate allows us to apply proposition [ lem : griesmer ] ( with @xmath572 and @xmath403 ) .
choose a nontrivial involution @xmath573 and a component @xmath565 of @xmath566 with @xmath574 such that @xmath565 covers at least @xmath575 of the @xmath576 components of @xmath577 .
in particular , property ( 1 ) and our choice of @xmath565 imply that @xmath565 covers at least @xmath578 of the components of @xmath24 , hence the chain @xmath579 satisfies property ( 1 ) . by maximality of @xmath279
, this chain must fail to satisfy property ( 2 ) .
it follows that @xmath580 or @xmath581 if @xmath582 , then @xmath554 has @xmath115periodic rational cohomology by proposition [ pro : new4per ] , and if @xmath583 , then it follows as in the proof of lemma [ linearassumptionlemma2 ] that @xmath554 is 4periodic .
indeed there exists an involution @xmath422 such that @xmath565 is the transverse intersection in @xmath554 of @xmath584 and @xmath585 .
moreover @xmath422 can be chosen such that the codimensions @xmath185 and @xmath586 of these submanifolds in @xmath554 are positive and satisfy @xmath587 the periodicity theorem implies that @xmath554 has @xmath115periodic rational cohomology .
this concludes the proof of the existence of a chain @xmath588 satisfying properties ( 1 ) , ( 2 ) , and ( 3 ) . with the first step complete ,
the second step of the proof of lemma [ lem : linearassumptionpart2 ] is to cover @xmath24 by chains @xmath589 of varying lengths @xmath590 satisfying properties ( 1 ) , ( 2 ) , and ( 3 ) .
we will see that at most @xmath591 chains are required .
since the @xmath554 in each such chain has 4periodic rational cohomology , it satisfies @xmath592 combining these facts , lemma [ lem : linearassumptionpart2 ] and hence theorem [ thm : linearassumption ] follows .
we proceed with the details .
first , by the claim , there exists a submanifold @xmath593 with @xmath115periodic rational cohomology that covers at least @xmath594 of the @xmath255 components of @xmath24 .
this leaves @xmath595 components of @xmath24 uncovered by @xmath143 , where @xmath596 where @xmath597 .
next , we apply the claim to the remaining @xmath595 components of @xmath24 .
this yields a submanifold @xmath598 with @xmath115periodic rational cohomology such that @xmath599 covers all but @xmath600 components of @xmath24 , where @xmath601 continuing in this way , we claim that we obtain a cover of @xmath24 by rationally @xmath115periodic submanifolds @xmath602 where @xmath603 indeed , after @xmath157 steps of this process , there are @xmath604 components of @xmath24 not yet covered , and @xmath605 if the right - hand side is less than 1 , then we have covered @xmath24 by the submanifolds @xmath606 .
moreover , the right - hand side is less than one if and only if @xmath607 which in turn holds if and only if @xmath608 taking @xmath609 , we see that we can cover @xmath24 by rationally @xmath115periodic submanifolds @xmath602 .
applying this fact together with conner s theorem , we have @xmath610 recall that we have @xmath611 for all @xmath23 , hence , by using the first part of the theorem , we can conclude the theorem by estimating @xmath157 as follows : @xmath612
recall that we are given an even - dimensional , one - connected , closed @xmath16manifold with positive curvature and an isometric action by a torus @xmath19 .
we will require the following theorem ( see ( * ? ? ?
* corollary iv.2.3 , p. 178 ) ) : [ exactlyone ] if @xmath428 is a compact , closed , orientable manifold , then no torus action on @xmath428 can have exactly one fixed point .
assume for a moment that each of the assumptions ( [ oddbettiass2])([oddbettiass5 ] ) imply that @xmath101 .
since @xmath613 by conner s theorem , and since @xmath24 is nonempty by berger s theorem , it follows that @xmath94 .
moreover since @xmath2 is orientable , theorem [ exactlyone ] implies @xmath24 is not a single point .
hence there are at least two components or one of positive , even dimension .
since the components of @xmath24 are orientable , we have in either case that @xmath614 . keeping the assumption that @xmath101 , and adding the assumption that the @xmath19action is equivariantly formal , we obtain that the leray serre spectral sequence of the borel construction degenerates at the @xmath615term .
this implies that @xmath616 applying conner s theorem , this implies that @xmath617 and @xmath618 the upper bounds on @xmath28 under the assumptions of theorems [ logassumption ] or [ thm : linearassumption ] now follow from the corresponding bounds on @xmath45 .
this leaves us with the task of proving @xmath101 under each of the assumptions ( [ oddbettiass2])([oddbettiass5 ] ) .
we first consider assumption ( [ oddbettiass5 ] ) .
that is , we assume that @xmath24 has an isolated fixed point and that @xmath2 is rationally elliptic . according to ( * ? ?
* corollary 7.49 , p. 302 ) , since @xmath2 is a compact , one - connected ( and therefore nilpotent ) , rationally elliptic manifold , if the homotopy euler characteristic @xmath619 , then @xmath24 has no contractible component .
hence the assumed isolated fixed point of @xmath24 implies that @xmath620 . since the rational ellipticity of @xmath2 implies @xmath621 with equality if and only if @xmath619 ( see ( * ? ? ?
* proposition 32.10 , p. 444 ) ) , we conclude @xmath622 .
this together with the assumption of rational ellipticity ( see ( * ? ? ?
* proposition 32.10 ) ) implies @xmath623 , hence @xmath101 by conner s theorem .
next we prove that @xmath101 under each of the assumptions ( [ oddbettiass2])([oddbettiass4 ] ) .
we remark that the case of assumption ( [ oddbettiass2 ] ) is contained in ( * ? ? ?
* theorem 1 , p. 1 ) , but we provide an alternative proof here based on theorem [ thmp ] . fix any component @xmath624 of @xmath24 .
taking @xmath625 in theorem [ thmp ] , we conclude the existence of a submanifold @xmath244 such that a. [ prop1 ] @xmath72 is rationally @xmath115periodic , b. [ prop2 ] @xmath244 is @xmath626connected , c. [ prop3 ] @xmath627 , d. [ prop4 ] @xmath628 , e. [ prop5 ] @xmath72 is totally geodesic in @xmath2 , and f. [ prop6 ] @xmath19 acts on @xmath72 and @xmath624 is a component of @xmath629 .
the last two points follow from the choice of @xmath72 as a component of @xmath630 for some @xmath194 .
indeed , such an @xmath72 is totally geodesic , and @xmath19 acts on @xmath72 since @xmath19 is abelian and connected . by ( [ prop6 ] ) and conner s theorem , @xmath631
hence it suffices to show that @xmath72 has vanishing odd betti numbers .
moreover , by ( [ prop1 ] ) and ( [ prop2 ] ) , @xmath72 is simply connected and has 4periodic betti numbers , so it suffices to prove that @xmath632 .
we prove this in three cases , which together cover assumptions ( [ oddbettiass2])([oddbettiass4 ] ) : * assume @xmath633 . by ( [ prop3 ] ) , @xmath634 is also divisible by four , hence four periodicity and poincar duality imply @xmath635 * assume @xmath127 , and assume @xmath98 , @xmath99 , or @xmath100 is zero . by ( [ prop2 ] ) , this is equivalent to @xmath636 , @xmath637 , or @xmath638 vanishing .
we prove that @xmath639 the first implication holds because four periodicity , poincar duality , and the assumption that @xmath640 imply @xmath641 .
the second implication follows from the definition of four periodicity .
indeed , the element @xmath642 inducing periodicity is zero , hence ( [ prop4 ] ) and four periodicity implies that the map @xmath643 induced by multiplication by @xmath644 is an isomorphism .
this can only be if @xmath632 . *
assume @xmath127 , and assume that the bott conjecture holds .
note that , by the previous case , we may assume that @xmath645 . by ( [ prop5 ] ) , @xmath72 has positive curvature , so the bott conjecture implies that @xmath72 is rationally elliptic .
in particular , @xmath646 . using four periodicity and the condition that @xmath640
, we can calculate @xmath647 in terms of the @xmath637 : @xmath648 by ( [ prop4 ] ) , this inequality implies @xmath649 .
but @xmath637 must be even by poincar duality and the graded commutativity of the cup product , hence @xmath650 . returning to the expression above for @xmath647
, we conclude that @xmath651 , and returning to the assumption of rational ellipticity , we conclude that @xmath632 , as required .
in this section we prove corollary [ cor : stablehopf ] and comment on some consequences of it .
we then state a more elaborate version of corollary [ corsymmetricspaces - short ] .
recall that we have a closed , one - connected manifold @xmath12 with positive sectional curvature and symmetry rank @xmath65 .
we assume that @xmath2 is a @xmath76fold cartesian product or a @xmath76fold connected sum of a manifold @xmath72 .
we claim that @xmath67 or @xmath70 , respectively .
we prove the two cases simultaneously .
observe that @xmath652 , hence @xmath653 by theorem [ thmp ] .
since @xmath654 in both cases , and since we may assume without loss of generality that @xmath655 , we conclude that @xmath656 .
this implies @xmath102 by theorem [ oddbettis ] .
theorem [ logassumption ] implies that @xmath657 . using the multiplicativity and additivity formulas for the euler characteristic ,
we obtain either @xmath658 or @xmath659 in the respective cases .
these inequalities imply either @xmath660 or @xmath661 the first is a contradiction for @xmath662 , while the second implies @xmath663 for @xmath664 , another contradiction .
* note that if there exists a nonnegatively curved metric on @xmath2 , also the product metric is nonnegatively curved .
no similar result on the existence of nonnegatively curved metrics on connected sums seems to be known unless it is the sum of two rank one symmetric spaces ( see @xcite ) .
thus the result on the cartesian power , i.e. the first assertion in the stable hopf conjecture with symmetry ( see corollary [ cor : stablehopf ] ) , appears to be stronger than the second one . * as for the assertion on the iterated fibrations , the proof is similar . since the @xmath81 are compact homogeneous spaces of equal rank lie groups , they have positive euler characteristic and their betti numbers are concentrated in even degrees .
hence @xmath665 for all @xmath23 by poincar duality , and the arguments in the proof of corollary [ cor : stablehopf ] apply equally .
let us make the result provided by corollary [ cor : stablehopf ] more precise in the classical case of a product with factors @xmath666 .
for example , the estimates in theorem [ logassumption ] imply that there is no metric of positive curvature on any of @xmath667 , @xmath668 ,
, @xmath669 compatible with the isometric action of a @xmath670torus .
we now proceed to corollary [ corsymmetricspaces - short ] .
we have @xmath671 for @xmath672 . by theorem 3.3 in @xcite
, there exists a product of spheres @xmath673 with @xmath674 such that 1 .
@xmath675 , 2 .
@xmath676 where @xmath677 , 3 .
@xmath676 where @xmath678 , 4 .
@xmath679 where @xmath678 , or 5 .
@xmath680 where @xmath678 .
note that the number of spherical factors @xmath681 ( as in the statement of the corollary ) is @xmath255 in the first three cases and is @xmath682 in the last two cases .
it follows that @xmath656 or @xmath683 . in particular
, @xmath622 by corollary [ oddbettis ] .
this implies that each @xmath301 is even and that case ( 5 ) can not occur . in cases ( 1 ) , ( 2 ) , and ( 3 ) , we estimate the euler characteristic by @xmath684 and , in case ( 4 ) , we estimate @xmath685 since @xmath686 , we conclude that the number of spherical factors @xmath681 is less than @xmath687 . for this
we make use of @xmath688 and @xmath689 where @xmath690 denotes the respective dimensions of the manifolds .
( we recall that the euler characteristics of the homogeneous spaces can easily be computed using the classical formula @xmath691 , i.e. as the quotient of the cardinalities of the respective weyl groups . )
we leave it to the interested reader to vary this result using the factors @xmath306 from lemma [ logassumptionlemma ] and theorem [ logassumptionplus ] .
besides , a variation of this result using theorem [ thm : linearassumption ] is obvious .
also , the presented result is certainly only interesting for large @xmath16 , since in small dimensions dimension estimates will be better .
in this section we shall provide a proof of theorem [ indth ] . before giving the necessary arguments let us cite the classical reference ( * ? ? ? * corollary , p. 317 ) .
[ ellgen ] let @xmath692 be an involution contained in the compact connected lie group @xmath693 acting differentiably on the @xmath694dimensional spin manifold @xmath428 . if the action is odd , then @xmath695 .
suppose the action is even and @xmath696 .
then in the laurent series the first @xmath120 coefficients beginning with @xmath111 vanish .
if @xmath697 , then @xmath698 does not depend on @xmath110 , it equals the signature of @xmath428 . if @xmath699 , then @xmath700 .
we remark that a similar statement holds if @xmath2 is not necessarily spin , but has finite second and fourth homotopy groups .
for this see ( * ? ? ?
* theorem 2.1 , p. 254 ) in its corrected form in @xcite or @xcite .
let us first deal with the case of a spin manifold @xmath2 .
due to theorem [ ellgen ] we may assume that the action of an involution @xmath701 , the effectively acting torus , is even .
the proof of theorem [ logassumption ] , i.e. the two cases into which we separated it , now gives us the following alternative : either we are in case one and the positively curved manifold @xmath2 has four - periodic cohomology , or case two applies .
if the latter is the case , then , in particular , there exists an involution @xmath692 with a fixed - point component @xmath624 satisfying @xmath702 . due to frankel every component @xmath703 in the fixed - point set @xmath704 has the property @xmath705 .
we derive @xmath706 due to theorem [ ellgen ] the first @xmath707 many coefficients of the elliptic genus @xmath106 vanish .
this proves the first assertion .
as for the proof of the second assertion we proceed in an analogous way using the above cited version of theorem [ ellgen ] . for this we observe that a manifold with four - periodic cohomology and @xmath708 is necessarily a rational cohomology sphere . obviously , it is a rational homotopy sphere in this case then .
h. herrera and r. herrera .
erratum to `` the signature and the elliptic genus of @xmath710-finite manifolds with circle actions '' [ topology appl .
136 ( 1 - 3 ) ( 2004 ) 251259 ] [ mr2023422 ] .
, 157(13):2157 , 2010 . | manifolds admitting positive sectional curvature are conjectured to have a very rigid homotopical structure and , in particular , comparatively small euler charateristics .
however , this structure is still highly speculative and best results in this direction are known under the assumption of large isometric torus actions . in this article , we obtain upper bounds for euler characteristics of closed manifolds that admit metrics with positive curvature and isometric torus actions .
we apply our results to prove obstructions to symmetric spaces and manifold products and connected sums admitting positively curved metrics with symmetry , providing evidence for a conjecture of hopf .
we also derive vanishing properties of the elliptic genus of positively curved manifolds . |
entanglement @xcite has long played a pivotal role in many quantum information processing tasks , such as secure communication using quantum - key distribution @xcite , teleportation of quantum states @xcite , quantum computation @xcite , reduction in communication complexity @xcite , and more recently , expansion and certification of randomness @xcite . as a result ,
the verification and quantification of this resource present in quantum systems is often an indispensable part of quantum - information processing ( qip ) protocols .
traditionally , for low - dimensional composite systems that are made up of only a few subsystems , the process of entanglement certification and/or quantification is carried out using complete quantum state tomography followed by the application of certain separability criteria @xcite .
this approach , however , suffers from the drawbacks that it is unnecessarily resource intensive and that the procedure of tomography may already be intractable for a physical system that is made up of only a handful of qubits ( see , for example , ref .
@xcite and references therein ) .
more recently , there has been considerable interest in verifying the presence of entanglement by measuring directly the expectation value of so - called _ entanglement witnesses _ @xcite .
while this latter approach is much more economical in terms of both the number of measurement settings and the resource required for classical post - processing @xcite , it still requires detailed knowledge of the actual measurements performed on the physical system in question .
this last feature is , of course , undesirable as there are examples of entanglement witnesses that are extremely fragile to the perturbations of the constituent measurement operators @xcite .
also , in some qip tasks such as quantum key distribution , it is highly desirable to make minimal assumptions on the devices used in the protocol .
in contrast , violation of bell - type inequalities @xcite is only a statement about the measurement statistics derived from an experiment .
since entanglement is a prerequisite for bell - inequality violation @xcite , one can deduce the presence of entanglement via a bell - inequality violation without resorting to any knowledge of the actual physical implementation of the measurements .
in modern terminology , bell - type inequalities therefore serve as _ device - independent entanglement witnesses _ and can be used for device - independent quantum - key distribution @xcite , state estimation @xcite , randomness generation @xcite , as well as self - testing quantum circuits @xcite . for a long time
, it was thought that the strength of bell inequality violation is monotonously related to the degree of entanglement ( see , e.g. , refs .
@xcite for a review on this topic ) ; that is , loosely speaking , more entanglement would lead to more nonlocality .
it turns out that the relation between entanglement and nonlocality is much more intricate @xcite .
however , for the clauser - horne - shimony - holt ( chsh ) bell inequality @xcite , it was shown by verstraete and wolf @xcite that the _ maximal _ possible chsh violation of a two - qubit state is monotonously related to the entanglement of the latter . from this result , it is then easy to lower bound the entanglement of a two - qubit state given the observed chsh violation ( see also ref .
@xcite ) . on the other hand , bell inequality violation
as we demonstrate in this paper can also be used to upper bound the entanglement of the underlying state . to achieve this , we present and make use of a family of clauser - horne - like @xcite bell inequalities ( i.e. , involving two binary measurements per party ) that are violated maximally by non - maximally entangled states . moreover , it can be shown that some of these inequalities _ can not _ be violated maximally by maximally entangled states of any dimensions .
this complements the recent result of junge and palazuelos @xcite , who proved the existence of such inequalities .
naturally , our inequalities suggest another _ device - independent _ application of bell inequalities , namely , to bound from above and below
the entanglement present in some underlying quantum system directly from the measurement statistics .
however , in contrast with the scenario considered in _ device - independent quantum - key distributions _
@xcite , here we assume that the dimension of the underlying hilbert space is known . to distinguish the present approach ( where the dimension of the hilbert space is assumed ) from the conventional device - independent approach , hereafter , we will refer to our scenario as _ semi - device - independent_. we focus on a scenario where the experimentalists know on which degree of freedom of the system the measurements are performed thus the relevant hilbert space dimension can be determined but nothing is assumed about the alignment of the measuring devices .
we believe this represents a significant advantage compared to previous techniques , such as entanglement witnesses and tomography , for which small alignment errors can lead to important errors in the estimation of entanglement @xcite .
we note also that it seems unlikely that one can find any useful bounds without such assumption on the dimensionality of the systems , as we discuss later .
the paper is organized as follows . in sec .
[ sec : prelim ] , we introduce a class of bell inequalities that will serve as our main tools for bounding entanglement .
then , in sec . [ sec : maxviolation ] , we investigate the quantum violations of these inequalities by maximally entangled states ( the relevant proof is relegated to appendix [ app : proofs ] ) .
after that , in sec .
[ sec : bounds : qubits ] , we illustrate how these bell inequalities can be used to provide nontrivial lower and upper bounds on the entanglement of a two - qubit state . in sec .
[ sec : qudits ] , we briefly comment on the applicability of these tools to the scenario where one makes no assumption of the dimension of the underlying quantum systems .
we conclude in sec .
[ sec : conclusion ] by comparing our approach with the one presented of ref .
@xcite in the context of device - independent state estimation , and highlighting some possible avenues for future research . in appendix
[ app : projective ] , we also provide another method of upper bounding entanglement assuming that the measurements are projective .
in this section , we first introduce the notations and briefly review the tools that we are going to use in the subsequent analysis . to this end , we consider the simplest scenario where two experimenters alice and bob have access to many copies of a shared quantum state @xmath0 and after many rounds of experiment , the bipartite conditional probability distributions @xmath1 are estimated from the experimental data . here , @xmath2 refers to the joint probability that the outcomes @xmath3 and @xmath4 are observed at alice s and bob s site respectively , conditioned on her performing the @xmath5 and him performing the @xmath6 measurement . clearly , from @xmath7 , the respective marginal probabilities @xmath8 and @xmath9
can also be determined .
our goal here is to obtain some non - trivial bounds on the entanglement of the underlying state from these measurement statistics .
our starting point is the following one - parameter family of modified clauser - horne ( ch ) polynomials : [ eq : bellfunction : tiltedch ] @xmath10 where @xmath11 for @xmath12 .
it is a well - known fact @xcite that with local measurements , separable states can only give rise to probability distributions that can be reproduced by _ shared randomness _ , and hence any measurement statistics that originate from such states must satisfy the bell inequality @xmath13 note that this particular bell inequality arises naturally in the study of the detection loophole in bell experiments . here
the parameter @xmath14 can be interpreted as the ( symmetric ) detection efficiency @xcite . for @xmath15 ,
the bell inequality @xmath16 reduces to the ch inequality @xcite which , in turn , is equivalent to the well - known chsh inequality @xcite .
thus , we can also rewrite eq .
as @xmath17 , \label{eq : bellfunction : tiltedch : viach}\ ] ] where @xmath18 is the ch polynomial . from eq . , it is clear that for any @xmath19 we can see @xmath16 as an inequality that consists of a conic combination of the ch inequality and some positivity constraints on joint probabilities . as a result , any measurement statistics that violate inequality with @xmath20 must also violate the ch inequality .
moreover , it is easily checked that any legitimate probability distribution always satisfies inequality for @xmath21 , since @xmath16 corresponds to a conic combination of positivity constraints in these cases .
henceforth , we restrict our attention to the scenarios where @xmath22 . in the following discussion , we denote by @xmath23 the positive - operator - valued measure ( povm ) element associated with the @xmath3-th outcome of alice s @xmath24-th measurement ( likewise @xmath25 for bob s ) . for any quantum state @xmath0 , we then write the corresponding quantum expectation value of @xmath26 under these measurement operators as : @xmath27 the corresponding maximal quantum violation of inequality for @xmath0 is denoted by : @xmath28 whereas the maximal violation of inequality allowed by quantum mechanics is denoted by @xmath29
as mentioned in sec . [ sec : introduction ] , an important feature of @xmath16 that allows us to establish _ upper _ bound on the underlying entanglement is that , except for @xmath15 , it is a bell inequality that is violated maximally by non - maximally entangled two - qubit states ( see fig . [
fig : concurrencevstau ] ) . a natural question that one may ask
is whether it is possible to recover the maximal quantum violation of @xmath16 if we also consider maximally entangled states of higher dimension .
we show in appendix [ app : mes : nmaxbiv ] that this is not possible for a large range of @xmath30 , as summarized in the theorem below @xcite .
[ conjecture : mescannotproduceq ] no finite - dimensional maximally entangled state can violate maximally the bell inequality @xmath16 with @xmath31 .
( color online ) plot of the concurrence @xmath32 of the two - qubit state @xmath33 which violates maximally the inequality @xmath16 as a function of @xmath30 ( blue solid line ) . also plotted is the concurrence @xmath34 of the critical two - qubit state @xmath35 ( red dashed line ) .
all states with @xmath36 do not violate @xmath16 ( see text at sec .
[ sec : ub : biv ] for details ) ; hence , the violation of @xmath16 allows one to derive upper bounds on the entanglement.,width=340 ] in relation to this , it is worth noting that for two - qubit states , one can show that the maximally entangled state can not even violate @xmath16 for the same interval of @xmath30 ( let alone maximally ) . for higher dimensional
maximally entangled states , despite intensive numerical search , we have also not found any violation of @xmath16 with @xmath31 for local hilbert space dimension up to @xmath37 .
we believe that this feature holds for maximally entangled states of arbitrary dimensions @xcite .
no finite - dimensional maximally entangled state can violate @xmath16 with @xmath31 .
to illustrate the idea of bounding entanglement directly from measurement statistics , we now consider the specific case where the underlying quantum system can be described as a two - qubit state .
it is worth noting that in this case , the entanglement of a quantum state @xmath0 is fully characterized in terms of its _ entanglement of formation _ , which is monotonically related to the _ concurrence _ , defined as @xcite @xmath38,\ ] ] where @xmath39 are the decreasingly ordered eigenvalues of @xmath40 , @xmath41 is the pauli-@xmath42 matrix , and the transposition @xmath43 is to be carried out in any product basis .
now , let us imagine that in some experiments , the probability distributions @xmath7 are found . from here , it is straightforward to compute the quantum value of @xmath26 [ cf .
eq . , ] which we denote by @xmath44 .
clearly , if @xmath44 corresponds to the maximal violation of @xmath16 for some @xmath30 , we can identify the concurrence of the underlying two - qubit state immediately with the help of fig .
[ fig : concurrencevstau ] . in general
, there is of course no reason to expect that the measurement statistics will lead to such extremal correlations , in which case the bounds established below will be useful . to see how a
lower bound on the concurrence of @xmath0 can be obtained from @xmath44 in particular @xmath45 it suffices to recall from ref .
@xcite that for any two - qubit state @xmath0 @xmath46 it then follows from simple calculation that @xcite @xmath47 where the second inequality follows from the fact that in general , @xmath48 , since the measurements operators @xmath49 may not be optimal .
also , it is worth noting that the first inequality in eq . can be saturated by pure states as well as some rank 2 mixed two - qubit states @xcite . as a result of eq .
, for any measurement statistics that violate inequality , that is , giving @xmath50 , we know that its concurrence is bounded from below according to eq . . to obtain a semi - device - independent upper bound on the concurrence of @xmath0 , we make use of @xmath51 for a general two - qubit pure state @xmath0 , cf .
eq . , which can be determined easily using the tools from ref .
@xcite and ref . @xcite .
specifically , if we denote a general two - qubit pure state ( in the schmidt basis ) by @xmath52 where @xmath53 , it can be shown that for any given @xmath54 , the maximal violation of @xmath55 of @xmath16 is upper bounded as follows ( see fig .
[ fig : maximalbiv ] ) @xcite : @xmath56 a direct consequence of this is that for any given @xmath54 , there exists a critical value of @xmath30 , denoted by @xmath57 beyond which no violation of @xmath16 by @xmath55 is possible .
equivalently , for any given value of @xmath58 , there exists a critical state @xmath59 such that any state that is more entangled than @xmath35 ( i.e , one with a higher value of @xmath54 ) can not violate @xmath16 .
see fig .
[ fig : concurrencevstau ] . from the above observation
, we see that for any given @xmath7 that violates @xmath16 for @xmath60 , we can immediately place an upper bound on @xmath54 , and hence the corresponding concurrence @xmath61 by first scanning through @xmath16 for @xmath60 and determining the largest value of @xmath30 for which @xmath16 is violated , that is , @xmath62
\,| \,{{\mathcal{s}}^{(\tau)}_{\mbox{\tiny obs}}}>0\right\}.\ ] ] the concurrence of the two - qubit state that gives rise to @xmath7 is then upper bounded by @xmath63 . while the explicit value of @xmath32 corresponding to @xmath35 for a given @xmath30 which we denote by @xmath64 can be easily determined using the tools from ref .
@xcite , an analytic expression for @xmath64 has remained elusive . nonetheless , from some simple calculation using eq .
, one finds that @xmath64 must satisfy the following upper bound : @xmath65 ( color online ) maximal quantum violation of @xmath66 as a function of @xmath30 ( blue solid line ) .
also shown is the analytic upper bound on @xmath67 computed from eq .
( red dashed line ) , which provides an upper bound on @xmath68 .
, width=302 ] an important point to stress now is that although the upper bound obtained using the above procedures was derived from two - qubit pure state , the same bound is also applicable to any two - qubit mixed state @xmath0 .
this follows from the fact that any @xmath0 having concurrence @xmath32 can always be written as a convex decomposition of pure states , each having concurrence @xmath32 @xcite and that @xmath51 is linear in the convex decomposition of @xmath0 .
note also that stronger bounds can be obtained under the additional assumption that the measurements are projective ( see appendix [ app : projective ] ) . to see how these bounds work in practice ,
let us consider the following measurement statistics that were gathered @xcite using a commercially available setup for producing a maximally entangled two - qubit state @xcite .
@xmath69 where @xmath70 . a simple calculation using eq
. shows that the underlying two - qubit state that generates these probability distributions must have its concurrence lower bounded by 0.9297 .
incidentally , these probability distributions also violate @xmath16 up until @xmath71 . using eq .
, this gives an upper bound on concurrence of 0.9999 @xcite .
if we further make use of that fact the measurements that give rise to eq .
are projective [ cf .
appendix [ app : projective ] ] , then this upper bound can be improved further via eq . to 0.9806 .
let us now relax the assumption that the given quantum system ( or more precisely , its degrees of freedom where the measurements were performed ) are qubits and see what we can learn from the measurement statistics . in general ,
as one might expect , it is very difficult to derive any useful bounds on the entanglement of the underlying state if we are only given @xmath7 .
there are , nonetheless , exceptional cases where concrete statements can be made about the underlying state even in the scenario where we are only given a small set of independent numbers , such as @xmath7 .
the first example is when the measurement statistics give @xmath72 . in this case , the correlations between the measurement outcomes gives rise to the so - called tsirelon bound @xcite the maximal possible quantum violation of the ch ( or chsh ) inequality .
it was known since the early 1990s that in this case the underlying state must be even - dimensional , and is either a coherent superposition or an incoherent mixture of two - qubit maximally entangled states that reside in orthogonal two - qubit subspaces @xcite . likewise , if we denote by @xmath73 the unique two - qubit pure state ( in the form of @xmath55 ) that maximally violates @xmath16 ( see fig .
[ fig : concurrencevstau ] ) , then from the work of masanes @xcite ( see also ref .
@xcite ) , one can show that the underlying quantum state must also be even - dimensional , and can either be a coherent superposition or an incoherent mixture of @xmath73 that reside in orthogonal two - qubit subspaces . clearly , these examples indicate that when the dimension of the hilbert space is not known , then even in these extremal cases , it is possible to have a wide spectrum of states , featuring different amounts of entanglement , that give rise to the same correlations .
hence , as long as we restrict ourself to the framework where only @xmath7 are given , it seems extremely challenging to construct any useful bounds on the underlying entanglement in a fully device - independent manner .
in this paper , we have investigated the possibility of bounding the entanglement of some underlying quantum state directly from measurement statistics , specifically from a given set of probability distributions @xmath7 .
our main tool is a family of bell - type inequalities whose maximal violation can not be obtained from maximally entangled states of any dimension . using these inequalities ,
we have illustrated how useful lower and upper bounds on the concurrence of some underlying two - qubit state may be constructed in a semi - device - independent manner . in the case where the measurements are known to be projective , we also demonstrated how another upper bound can be obtained by analyzing the marginal probabilities given .
an interesting issue is to compare our approach with the one of ref .
@xcite , in which the authors investigate device - independent state estimation based on the chsh inequality violation .
the main differences between these two approaches are as follows . on the one hand , the approach of ref .
@xcite does not require any assumption about the hilbert space dimension of the systems , in contrast to our approach .
on the other hand , the observed chsh violation @xmath74 considered in ref .
@xcite is actually @xmath75 , where @xmath76 refers to some local ( nonunitary ) , completely positive map acting on the original state @xmath0 , whereas in the present paper , we use directly @xmath77 to establish the bounds .
moreover , in the approach of ref .
@xcite , the measurements that give rise to @xmath75 are assumed to be optimal for @xmath78 , whereas we make no such assumption . in general
, it would be interesting to derive useful ( semi-)device - independent bounds on entanglement for higher dimensional systems .
admittedly , the bell inequalities that we have considered here do not seem to provide useful bounds , as one may already anticipate from earlier work on dimension witnesses @xcite .
however , given that there are other bell inequalities @xcite that are more suitable for higher dimensional quantum systems , it will be interesting to investigate if analogous bounds on , say , the generalized concurrence can be derived using these other inequalities . as mentioned in sec .
[ sec : prelim ] , the bell inequalities that we have considered here are closely related to those being used in the analysis of detection loopholes @xcite .
another possible avenue for future research is to exploit results established in these earlier studies @xcite to establish or improve bounds on entanglement beyond what we have derived in this paper .
finally , let us remark that for some quantum resources such as the @xmath79-partite greenberger - horne - zeilinger state , it was shown that there is a high probability of finding bell - inequality violations by performing local measurements in randomly chosen bases @xcite .
if we can also establish non - trivial bounds on multipartite entanglement directly from these measurement statistics , then together , these approaches may offer a much more resource - economical paradigm for characterizing quantum states produced in the laboratory . from a quantum information processing point of view , this is the ultimate goal that we hope to achieve with the approach taken in this paper .
note , however , that in the multipartite scenario , the possibility of bounding entanglement directly from measurement statistics may also become highly non - trivial , as there is no natural choice of entanglement measure , and the whole problem also becomes considerably more complicated .
_ note added . while finishing this manuscript , we became aware of a related result by vidick and wehner @xcite , who showed that the bell inequality @xmath80 of ref .
@xcite shares a similar feature with the ones presented here , namely that its maximal quantum violation can not be reached by maximally entangled states in any dimension .
it is interesting to note that while our inequalities use only two measurements per party ( compared to three measurements for @xmath80 ) , the inequality @xmath80 is tight , in the sense of being a facet of the local polytope .
_ the authors acknowledge helpful comments from an anonymous referee as well as useful discussions with antonio acn , jonathan allcock , joonwoo bae , jean - daniel bancal , cyril branciard , nicolas gisin , charles lim , enrico pomarico , and valerio scarani .
this work is supported by the australian research council , the swiss nccr quantum photonics " , the european erc - ag qore , the jnos bolyai grant of the hungarian academy of sciences and the uk epsrc .
here , we provide a proof of theorem [ conjecture : mescannotproduceq ] which shows that there are bipartite probability distributions @xmath7 that can not be written in the form of @xmath81 where @xmath82 is the @xmath83-dimensional maximally entangled state .
note that similar results for more complicated experimental scenario were also obtained in refs .
@xcite .
the proof is by contradiction .
let us start by assuming the converse , namely , that the maximal violation of @xmath16 for @xmath31 can also be achieved using a @xmath83-dimensional maximally entangled state @xmath84 in conjunction with some choice of ( non - trivial ) projectors @xmath85 , @xmath86 .
we deduce the necessary conditions that follow from this assumption and show that when these necessary conditions are in place , @xmath82 can not even violate @xmath16 . the assumption that @xmath87 can violate
@xmath16 maximally means that there must exist some choice of @xmath85 and @xmath86 such that [ cf .
eq . ]
@xmath88 without loss of generality , we now make use of the main lemma from ref . @xcite and assume that the state @xmath0 and all the povm elements are already written in the basis whereby all the four @xmath85 are simultaneously block diagonal , likewise for @xmath86 , that is , @xmath89 where @xmath90 , @xmath91 are ( local ) block - diagonal projectors consisting of only @xmath92 and/or @xmath93 blocks .
let us now denote by @xmath94 the ( unnormalized ) non - vanishing @xmath95 , @xmath92 , or @xmath93 blocks of @xmath96 and @xmath97 the corresponding probability of projecting into this block , that is , @xmath98 .
the normalized density matrix corresponding to @xmath94 can then be written as @xmath99 . in these notations ,
eq . can be rewritten as @xmath100 where the restricted sum @xmath101 only runs over @xmath102 and @xmath103 where @xmath104 , or equivalently where @xmath105 is well defined .
this last expression indicates that @xmath68 can also be seen as a convex combination of the quantum values of @xmath105 under the povm elements @xmath106 , @xmath107 .
since @xmath68 is the maximal quantum violation of @xmath16 , so we must also have @xmath108 for all @xmath109 where @xmath105 is non - vanishing . clearly , in order for eq . to hold true , for any values of @xmath109 where @xmath104 , the corresponding @xmath105 can not be a block of size @xmath92 or @xmath93 , since such blocks correspond to classical states that can not violate @xmath16 , and hence do not satisfy eq . .
this implies that ( i ) @xmath82 can not have odd schmidt rank and ( ii ) the projectors @xmath90 , @xmath110 can not contain any nonvanishing @xmath93 block , or otherwise there is always a nonzero probability of projecting onto a @xmath105 with size @xmath92 or smaller . combining these observations with the main lemma of ref .
@xcite gives @xmath111 for all @xmath112 , and hence @xmath113 in other words , the marginal probabilities become @xmath114 and likewise for @xmath9 . putting these back into eq .
, we see that for an even - dimensional maximally entangled state and where eq .
is enforced , its quantum value becomes @xmath115 the first part of this expression is simply the ordinary ch polynomial , which admits the maximum quantum value @xmath116 for any even - dimensional maximally entangled state @xcite .
is is then easy to see that the left - hand side of eq .
is always upper bounded by 0 for @xmath117 which contradicts our initial assumption that @xmath82 can violate @xmath16 maximally for this interval of @xmath30 .
is it still possible to upper bound the entanglement of the underlying state when the probability distributions @xmath7 do not violate @xmath16 for @xmath118 ? in this regard , it turns out that a simple upper bound on @xmath32 can always be established if we are also equipped with the knowledge that the measurements that give rise to @xmath7 are projective measurements , that is , are described by @xmath119 , \quad b^{{b}}_{{y}}={\frac{1}{2}}\left[{\mathbf{1}}_2 + ( -1)^{{b}}\hat{b}_{{y } } \cdot\vec{\sigma}\right],\ ] ] where @xmath120 is the vector of pauli matrices , @xmath121 are unit vectors , and @xmath122 is the @xmath92 identity matrix . the key idea here is to realize that as the concurrence @xmath32 increases from 0 to 1 ( i.e. , @xmath54 increases from 0 to @xmath123 ) , the interval of admissible marginal probabilities @xmath8 , @xmath9 shrinks from @xmath124 $ ] to a single point @xmath125 .
this can be seen , for example , by noting that for @xmath55 and with eq . , we have @xmath126 where @xmath127 ( @xmath128 ) is the @xmath129-component of the vector @xmath130 ( @xmath131 ) . clearly , these expressions are bounded as @xmath132 together with eq . , it is then easy to see that @xmath133 ^ 2},\sqrt{1-\left[1 - 2{p_{\mbox{\tiny obs}}({b}| { y})}\right]^2}\right\},\ ] ] where @xmath134 and @xmath135 are , respectively , alice s and bob s marginal probabilities estimated from the experiment .
it is important to note that although eq .
was derived for @xmath55 , it also holds for an arbitrary pure two - qubit state having the same concurrence .
consequently , as with the upper bound derived in the last section , eq .
is also applicable to arbitrary two - qubit mixed states .
naively , one might expect that by first grouping the local subspaces of some higher - dimensional maximally entangled state , it is possible to deterministically produce a two - qubit non - maximally entangled state which can then be used to give maximal violation of @xmath16 .
however , it turns out that such operation , which consists in diluting entanglement , can not be achieved without classical communication @xcite .
note that if this conjecture is valid , then no finite - dimensional maximally entangled states with a symmetric detection efficiency of @xmath136 or less can generate @xmath7 that could be used to close the detection loophole .
note that inequality is also valid with @xmath137 and @xmath74 replaced , respectively , by @xmath68 and @xmath138 .
however , it should be clear from eq .
that the best lower bound on @xmath32 that can be obtained from the measurement statistics is due to the one where @xmath15 , in which case @xmath68 becomes @xmath137 and @xmath138 becomes @xmath74 .
m. navascus , s. pironio , and a. acn , , 010401 ( 2007 ) ; new j. phys . , * 10 * , 073013 ( 2008 ) ; a. c. doherty , y .- c .
liang , b. toner , and s. wehner in _ proceedings of the 23rd ieee conference on computational complexity _
( ieee computer society , college park , md , 2008 ) , pp .
199 - 210 ; eprint arxiv:0803.4373 ( 2008 ) .
s. massar , , 032121 ( 2002 ) ; s. massar , s. pironio , j. roland , and b. gisin , , 052112 ( 2002 ) ; s. massar and s. pironio , , 062109 ( 2003 ) ; n. brunner and n. gisin , a * 372 * , 3162 ( 2008 ) ; k. f. pl and t. vrtesi , , 022120 ( 2009 ) . | detection and quantification of entanglement in quantum resources are two key steps in the implementation of various quantum - information processing tasks . here
, we show that bell - type inequalities are not only useful in verifying the presence of entanglement but can also be used to bound the entanglement of the underlying physical system .
our main tool consists of a family of clauser - horne - like bell inequalities that can not be violated maximally by any finite - dimensional maximally entangled state . using these inequalities , we demonstrate the explicit construction of both lower and upper bounds on the concurrence for two - qubit states .
the fact that these bounds arise from bell - type inequalities also allows them to be obtained in a semi - device - independent manner , that is , with assumption of the dimension of the hilbert space but without resorting to any knowledge of the actual measurements being performed on the individual subsystems . |
in a previous work it was argued@xcite that with maximal vacuum mixing there is agreement , with minor modifications , between extant observations of solar neutrinos and predictions by the standard solar model ( ssm)@xcite . the maximal vacuum mixing case considered was that in which the phase of neutrino oscillations coming from the sun is averaged , leading to @xmath4 of the neutrinos arriving at the earth as electron neutrinos . as a result of this averaging , while @xmath5 was assumed to be maximal ( equal to one ) , @xmath1 was not determined and taken to lie in the approximate range @xmath6 with an exclusion of the approximate range @xmath7 for maximal mixing@xcite due to the lack of an observed day - night effect in the superkamiokande data@xcite . on the other hand , the recent first results of the sno measurement of charged current interactions produced by @xmath8b neutrinos@xcite , taken in combination with the elastic scattering result of the super - kamiokande collaboration@xcite , indicate that only about one third of the neutrinos arriving at the earth from the sun are electron neutrinos , with the other two thirds being @xmath9 or @xmath10 neutrinos .
oscillation into sterile neutrinos now seems relatively unlikely from the sno result .
while at first glance this comparison seems to make maximal vacuum mixing less likely , a global analysis of the sno result with the other solar neutrino experiments , chlorine@xcite , super - kamiokande@xcite , and gallium@xcite has led to the conclusion that `` the cc measurement by sno has not changed qualitatively the globally allowed solution space for solar neutrinos , although the cc measurement has provided dramatic and convincing evidence for neutrino oscillations and has strenghened the ths case for active oscillations with large mixing angles@xcite . ''
furthermore , global analyses@xcite@xcite do not completely exclude solutions to the solar neutrino problem in the mass region @xmath11 for maximal ( or near maximal ) mixing . in the following
, the time varying phase of oscillating @xmath0be neutrinos is investigated as a possible method to discover ( or exclude ) a solution of the solar neutrino problem in that mass region . in the mass region @xmath12
there are so - called `` just - so '' vacuum solutions of the solar neutrino problem , where the phase of the oscillation of @xmath8b neutrinos coming from the sun is not completely averaged@xcite .
recall also@xcite , that there is a large change in the @xmath0be electron neutrino flux over the year in the @xmath8b `` just - so '' region due to the change in phase of order @xmath13 in a year brought about by the @xmath14 yearly orbital variation from the mean distance of the sun to the earth .
as will be shown in the following , when phase averaging due to the temperature of the sun and phase damping due to the msw effect are considered , it turns out that phase variation in @xmath0be neutrinos should be observable for @xmath1 in the range from about @xmath15 up to about @xmath16 ( ev)@xmath3 . this observable range of @xmath1 via @xmath0be neutrinos turns out to be in approximate agreement with a previous analysis by de gouva , friedland , and murayama@xcite .
there is a low energy region of solar neutrinos dominated by the nearly monenergetic 862 kev line from electron capture by @xmath0be in the sun .
it is the purpose of the borexino@xcite experiment , soon to come on line , to measure these neutrinos in real time .
feasability studies are also being carried out for other experiments to measure @xmath0be neutrinos in real time such as lens@xcite and hellaz@xcite .
if the @xmath0be neutrinos were truly monoenergetic then the number of neutrinos detected via electron scattering in an experiment like borexino ( normalized to unity for no oscillations ) would take the following form for vacuum oscillations @xmath17 where @xmath18 is the vacuum mixing angle , @xmath1 is expressed in ( ev)@xmath3 , the @xmath9 or @xmath10 neutrino scattering relative to electron neutrino scattering at 0.862 mev is 0.21@xcite , and @xmath19 the distance from the earth to the center of the sun ( in km . ) , which varies through the year due to the eccentricity @xmath20 of the earth s orbit about the sun . note
that since we take the number of neutrinos detected as a function of @xmath21 , the phase of the earth in its orbit about the sun , rather than of the time of the year there is no @xmath22 seasonal variation in @xmath23 ; it is canceled by the jacobian in going from time as an independent variable to @xmath21 as an independent variable ( kepler s second law ) .
it has been pointed out by pakvasa and pantaleone@xcite that the @xmath0be energy line is thermally broadened not only by the spread in nuclear velocities ( doppler broadening ) but also by the solar temperature of the approximately 80% of the capture electrons that come from the continuum .
we makes use of the published table of bahcall @xcite which was obtained by convoluting both these sources of thermal spreading to obtain the energy profile of the 862 kev @xmath0be solar neutrino shown in figure 1 .
note that the distribution @xmath24 is asymmetric in shape and plotted as a function of @xmath25 in mev .
expressing @xmath1 in @xmath26 one obtains @xmath27 or @xmath28.\ ] ] for clarity and convenience we will retain these constants explicitly and express @xmath1 in units of @xmath26 unless otherwise specified for the rest of this paper .
since @xmath29 is constrained to contribute to eq.(9 ) only when it is much smaller than unity , one may set @xmath30 equal to @xmath31 and carry out the integration to obtain @xmath32,\end{aligned}\ ] ] with @xmath33 in additon to temperature damping of the the oscillations there is a second factor that we may call `` msw damping '' . with
maximal mixing , the phase averaged rate of electron neutrinos does not depend on whether an msw transition has taken place .
however , if an msw transition has taken place in the sun , then the maximally mixed neutrino emerges in the form of a pure mass eigenstate ( i.e. @xmath34 in eq.(a5 ) of appendix a ) . although the pure mass eigenstate @xmath35 is half electron neutrino and half other flavor neutrino , there is no interference from eq.(a5 ) and thus no oscillation .
the probability of remaining an electron neutrino remains one half without variation in the vacuum from the sun to the earth .
in contrast , pure vacuum oscillations with maximal mixing ( no msw transition ) leads to equal parts of each mass eigenstate ; the neutrinos oscillate from pure electron neutrino to pure other flavor neutrino on the path from sun to earth .
however the phase averaged probability of an electon neutrino reaching the earth is still one half .
guth , randall , and serna@xcite have pointed out the relevance of this difference for matter oscillations in the earth : there can be a day - night effect , even for the case of maximal mixing if there has been an msw transition in the sun . appendix a comprises a short digression on this point .
the treatment in appendix a assumes phase averaging over the distances involved , due to the larger @xmath1 values that would come into play in a possible day night effect .
here we are interested in the phase of the vacuum oscillation , since that is our signal .
the onset of msw conversion in the sun with larger @xmath1 can be investigated numerically by utilizing a piece of computer code adapted from a previous investigation@xcite .
the rate of @xmath0be electron neutrinos emerging from the sun ( again normalized to unity for no oscillations ) takes the form @xmath36 with @xmath37 the distance from the surface of the sun plus some constant . for maximal
mixing : @xmath38 , and @xmath39 is 0.5 for vacuum oscillations but @xmath39 vanishes for complete adiabatic conversion .
the top panel in figure 2 shows how the magnitude of the oscillation for maximum mixing is reduced with increasing @xmath1 even though the phase averaged mixing remains a constant .
the filled circles are the values of 2@xmath39 calculated numerically for @xmath40 .
the solid line through these circles approximates 2@xmath39 by @xmath41 $ ] . except for the small oscillations beyond @xmath42
, this `` msw damping '' is well represented by the gaussian factor .
the solid line at .5 represents @xmath43 for maximal mixing .
the dotted lines are the corresponding quantities 2@xmath39 and @xmath43 for @xmath44 , just for comparison .
incorporating `` msw damping '' eq.(6 ) then becomes @xmath45.\end{aligned}\ ] ] this is the expression that we use in the calculations to follow .
( 8) may also be written in the form @xmath46,\end{aligned}\ ] ] where @xmath47 and @xmath48 for the purpose of illustration we ignore the @xmath49 dependence in the temperature damping factor and consider @xmath50 .
the bottom panel of figure 2 shows @xmath50 as the short - dashed line and repeats the gaussian msw damping factor from the above panel as the long - dashed line .
the solid line is the product of the two , the overall damping factor including temperature spreading and msw damping .
it is clear that there is a complete damping out of the oscillations at @xmath51 ( ev)@xmath3 , and that this broadening averages out the phase of the oscillations at higher values of @xmath1 .
finally one should note that the source broadening is insignificant because of the following .
it turns out that the ssm density@xcite of @xmath0be neutrinos produced as a function of the solar radius is very close to a gaussian function of the sun s radius , @xmath52 where @xmath53 , and @xmath54 is the distance from the center of the sun in units of the distance from the earth to the sun .
however , as has been pointed out@xcite , the oscillations effectively start not at the source but at the level crossing point . for the present maximal mixing case
the level crossing point is at surface of the sun .
the neutrinos originating off the sun s axis in the direction of the earth will have a slightly larger distance to travel due to the curvature of the sun s surface . the gaussian density eq.(12 )
leads , in a very good approximation , to a source spreading density of the form @xmath55 where @xmath56 or twice @xmath57 times the ratio of the sun s radius to the mean earth - sun distance , and @xmath58 is the distance from the point on the sun s surface closest to the earth toward its center .
if this small source broadening were the only cause of damping , then by an analytical treatment paralleling that leading to eq.(9 ) , one would find a source damping factor @xmath59 the long and short dashed line in the bottom panel of fig.(2 ) represents @xmath60 .
obviously @xmath60 only starts to deviate from unity at the rightmost part of the plot ( at @xmath61 ) .
thus , source broadening is insignificant for our region of interest .
the region that we will investigate spans the range from @xmath62 , the `` just so '' region for @xmath8b , up to @xmath51 , where the broadening averages the phase .
as noted above and in appendix a , one might in principle begin to see a day - night effect@xcite with the onset of msw damping .
in fact there would be a sizable day - night effect for maximal mixing at @xmath63@xcite ( the so called `` low '' msw solution ) .
figure 3 shows @xmath64 for maximal mixing @xmath40 beginning at the low end with @xmath1 = 0.01 ( again in units of @xmath26 ) .
note from eq.(8 ) that the overall phase of the cosine factor depends on @xmath1 and that this phase changes by @xmath65 when @xmath1 changes by @xmath66 .
this phase sensitivity is illustrated figure 3 where for each panel in addition to the curve for the labeled value of @xmath1 there are also curves for that value plus the appropriate increments to shift the overall phase by @xmath67 , @xmath65 , @xmath68 , and @xmath69 .
figure 4 shows the increasing frequency of the oscillations of @xmath64 as a function of @xmath21 for the @xmath1 region of @xmath70 to @xmath71 .
note also the decreasing amplitude of the oscillations as they come close to being damped out by the temperature plus source broadening and msw damping at @xmath72 .
with such a rapid oscillation period throughout the year seen especially in figure 4 ( on the order of several months to several days ) , one anticipates that for such values of @xmath1 there would be insufficient statistics at an experiment like borexino for a pattern to be obvious .
however , in what follows we will investigate how a fourier type analysis of data from such experiments could give evidence of a phased oscillation and thereby determine the value of @xmath1 if it lies in this range .
fourier analyses of @xmath0be solar neutrino data have been previously proposed@xcite , but what follows here is a somewhat different approach .
since borexino is a detector rather than radiochemical experiment , it records the information on when each count was recorded and thereby the distance of the detector to the sun @xmath73 incorporated as @xmath21 in eq.(8 ) .
we suggest analyzing data from such experiments by effectively integrating data with a factor @xmath74}$ ] and varying @xmath75 over the range @xmath76 to look for a signal .
to test whether the @xmath1 can be determined by such a method , monte carlo data sets have been simulated in the following way .
random numbers are generated uniformly for @xmath21 from 0 to @xmath77 in order to cover the year and make use of eq.(8 ) . in order to weight events according to what would be expected from eq.(8 ) with a specific @xmath1 , a second random number between 0 and 1
is then generated for each @xmath21 and a count is generated if the random number is less than @xmath64 .
this is the data set : the collection of specific angles , @xmath78 , at which single events are recorded durin a year .
figure 5 shows a sample analysis of a data set generated from 15000 monte carlo attempts for @xmath79 in our units .
the top panel shows the expected oscillation pattern , @xmath80 . from eq.(8 ) one would expect about 9075 data points to lie below the curve from 15000 random attempts , and in fact a set of 8993 data points @xmath78 were generated in this sample .
the number of data points in this sample corresponds roughly to a year s running time at borexino .
for analysis one might first consider a fourier type transformation on the set @xmath78 @xmath81 the solid curve in the middle panel displays @xmath82 and it shows a discontinuity in pattern near @xmath83 .
if there were no phase oscillation then one would expect @xmath84 to approach a function that we will call @xmath85 @xmath86 the dotted line in the middle panel displays @xmath87 .
this suggests that we subtract off the bessel function like behavior @xmath85 contained in @xmath88 ( which tends to obscure the signal ) , create a new function @xmath89 and use @xmath90 to analyze our data set @xmath78 .
@xmath91 is displayed in the bottom panel .
the signal of @xmath79 is unambiguous .
figure 6 shows that an unabiguous signal would be obtained with about one year s borexino statistics for @xmath1 in the range 0.10.5 . for @xmath92
there is a signal at @xmath93 that might be a little ambiguous with only one year s statistics , but it retains its shape with increasing statistics ; a secondary peak seen at about @xmath94 with one year s statistics goes away with the higher statistics .
there is no signal apparent for the @xmath95 case , as one would expect from looking at the corresponding curve in figure 4 .
figure 7 shows the extraction of the signal for @xmath1 an order of magnitude lower .
curves correspond to values in figure 3 .
below @xmath96 this particular analysis starts to become ambiguous .
however the slow rate of variation with @xmath21 makes direct comparison with the patterns seen in figure 3 practical . at @xmath97
the analysis is complicated by the large change in magnitude of the rate throughout the year with a small increment in the value of @xmath1 .
based on the sno result it now seem likely that the solar neutrino puzzle has been solved .
it is not a deficiency in the standard solar model that is being observed but new physics .
electron neutrinos are oscillating into some combination of @xmath9 and @xmath10 neutrinos .
exactly how this happens is perhaps not yet clear , whether by one of the msw solutions or some vacuum mixing solution . in the previous sections of this paper it has been shown that if the solution to the solar neutrino puzzle happens to be maximal mixing in the mass range @xmath98(ev)@xmath3 , then a proper analysis of a successful @xmath0be neutrino experiment should be able to unambiguously determine @xmath1 .
not seeing a @xmath1 signal in this mass range would elimate a region of @xmath1 for large mixing angle .
i would like to thank chellis chasman for reading and commenting on the manuscript .
this manuscript has been authored under contract no .
de - ac02 - 98ch10886 with the u. s. department of energy .
guth , randall , and serna@xcite have pointed out that there can be a day - night effect , even for the case of maximal mixing .
what follows is a compact explication of this point with emphasis on the limits of no msw and maximal msw effect in the sun . the general form for two mass eigenstates in two neutrino mixing is @xmath99 and @xmath100 where @xmath101 is presumed to be some linear combination of @xmath102 and @xmath103 . conversely @xmath104 and @xmath105 in free space mass eigenstates propagate independently .
a mixed mass state @xmath106 then has the form @xmath107 the probability @xmath108 that a neutrino born in the sun is an electron neutrino when it reaches the earth is then @xmath109 with @xmath110 the average probability of a mass one or mass two eigenstate arriving at the earth where the phase has been averaged by the distances involved . since @xmath111 this
may also be written equivalently @xmath112 the probability @xmath113 that an electron neutrino born in the sun will be an electron neutino after passing through the sun , traveling to the earth , and then passing through the earth is simply @xmath114 where @xmath115 is the probability that a mass 2 neutrino entering the earth emerges at the detector as an electron neutrino . making use of eq.(a7 )
this becomes @xmath116 this expression is the mikeyev - smirnov expression@xcite for the day night effect , trivially transformed@xcite to be most transparent in various limits .
anthony j. baltz , alfred scharff goldhaber , and maurice goldhaber , phys .
lett . * 81 * , 5730 ( 1998 ) . j. n. bahcall and r. n. ulrich , rev
. mod . phys . * 60 * , 297 ( 1988 ) .
j. n. bahcall and m. h. pinsonneault , rev .
64 * , 885 ( 1992 ) .
john n. bahcall , m. pinsonneault , and g. j. wasserburg , rev .
phys . * 67 * , 781 ( 1995 ) .
j. n. bahcall , s. basu , and m. h. pinsonneault , phys .
* b 433 * , 1 ( 1998 ) j. n. bahcall , p. i. krastev , and a. yu .
smirnov , phys .
d 58 * , 096016 ( 1998 ) .
y. fukuda et al .
( super - kamiokande collaboration ) , phys .
lett . * 81 * , 1158 ( 1998 ) .
q. r. ahmed , et al .
( sno collaboration ) , phys .
lett . * 87 * , 071301 ( 2001 ) b. t. cleveland et al .
, astrophys .
j. * 496 * , 505 ( 1998 ) .
w. hampel et al .
( gallex collaboration ) , phys .
* b 447 * , 127 ( 1999 ) .
j. n. abdurashitov et al .
( sage collaboration ) , phys . rev . * c 60 * , 055801 ( 1999 ) .
john n. bahcall , m. c. gonzalez - garcia , and carlos pea - garay , arxiv : hep - ph/0106258 .
g. l. fogli , e. lisi , d. mantanino , and a. palazzo , arxiv : hep - ph/0106247 . n. hata , and p. g. langacker , phys .
* d 56 * , 6107 ( 1997 ) .
james m. gelb and s. p. rosen , phys . rev . *
d 60 * , 011301 ( 1999 ) .
p. i. krastev and s. t. petcov , nucl .
* b 449 * , 605 ( 1995 ) .
james m. gelb and s. p. rosen , arxiv : hep - ph/9908325 .
andr de gouva , alexander friedland , and hitoshi murayama , phys .
d 60 * , 093011 ( 1999 ) .
borexino collaboration , g. alimonti et al .
, arxiv : hep - ex/0012030 . m. cribier for the lens collaboration , nucl . phys .
* b * ( proc .
suppl.)*87 * , 195 , ( 2000 ) .
p. gorodetzky , a. de bellefon , j. dolbeau , t. patzak , p. salin , a. sarrat , j. c. vanel , nucl . phys .
* b * ( proc .
suppl.)*87 * , 506 , ( 2000 ) . j. n. bahcall , _ neutrino astophysics _ , cambridge university press ( 1989 ) .
sandip pakvasa and james pantaleone , phys .
lett . * 65 * , 2479 ( 1990 ) .
john n. bahcall , phys .
d 49 * , 3923 ( 1994 ) .
alan h. guth , lisa randall , and mario serna , jhep * 9908 * , 018 ( 1999 ) .
a. j. baltz and j. weneser , phys . rev . *
d 50 * 5971 ( 1994 ) .
j. pantaleone , phys . lett .
* b 251 * , 618 ( 1990 ) .
a. j. baltz and j. weneser , phys . rev .
* d 35 * 528 ( 1987 ) ; * d 37 * 3364 ( 1988 ) ; e. d carlson , phys . rev .
* d 34 * , 1454 ( 1986 ) ; j. bouchez , m. cribier , w. hampel , j. rich , m. spiro , and d. vignaud , z. phys .
c * 32 * , 499 ( 1986 ) .
r. s. raghavan , a. b. balantekin , f. loreti , a. j. baltz , s. pakvasa , and j. pantaleone , phys . rev . *
d 44 * , 3786 ( 1991 ) . a. de rujula and s. glashow , cern - th 6655/92 ( 1992 )
. g. l. fogli , e. lisi and d. mantanino , phys .
d 56 * 4374 ( 1997 ) .
s. p. mikheyev and a. yu .
smirnov , in _ new and exotic phenomena , _ proceedings of the moriond workshop , les arc , savoie , france , 1987 , edited by o. fackler and j. tran thanh van ( editions frontieres , gif - sur - yvette , france , 1987 ) , p. 405 . | if @xmath0be solar neutrinos can be observed in real time experiments , then an extended region of @xmath1 can be probed by a proper analysis of the rapidly changing phase of vacuum oscillations due to the eccentricity of the earth s orbit about the sun . for the case of maximal vacuum mixing , a kind of fourier analysis of expected data for one year s time could uniquely pick out @xmath1 if it lies in the region @xmath2(ev)@xmath3 .
+ * pacs : 14.60.pq , 13.10.+q , 25.30.pt * = 100000 specialpagefalse @page=1 oddheaddraftto be submitted to phys .
rev .
d oddhead evenheadoddhead oddfoot@font evenfootoddfoot specialpagefalse oddhead evenheadoddhead |
the total cross section @xmath3 is a fundamental quantity in collisions of strongly interacting particles . at present
one of the main theoretical approaches for the description of @xmath4 is the qcd - inspired formalism @xcite . in this approach
the energy dependence of the total cross section is obtained from qcd via an eikonal formulation .
the high - energy dependence of @xmath4 is driven by the rapid increase in gluon density at small-@xmath5 . in this work
we explore the non - perturbative dynamics of qcd in order to describe , in both @xmath0 and @xmath1 channels , the total cross sections @xmath2 and the ratios of the real to imaginary part of the forward scattering amplitude , @xmath6 . in our calculations
we introduce an energy - dependent dipole form factor which represents the overlap density for the partons at impact parameter @xmath7 .
the behavior of these forward quantities is derived from the qcd parton model using standard parton - parton elementary processes and an updated set of gluon distribution , namely the cteq6l1 set . in our analysis
we have included recent measurements of @xmath0 total cross sections at the lhc by the totem collaboration @xcite .
this work is organized as follows : in the next section we introduce our qcd - based model where the onset of the dominance of semihard partons is managed by the dynamical gluon mass .
we investigate a form factor which the spatial distribution of gluons changes with the energy and introduce integral dispersion relations to connect the real and imaginary parts of eikonals . in section
iii we present our results and in section iv we draw our conclusions .
in our model the increase of @xmath2 is associated with elementary semihard processes in the hadrons . in the eikonal representation @xmath3 and @xmath8
are given by @xmath9 \label{eq01}\end{aligned}\ ] ] and @xmath10 } , \label{eq03}\end{aligned}\ ] ] respectively , where @xmath11 is the square of the total center - of - mass energy , @xmath7 is the impact parameter , and @xmath12 is the ( complex ) eikonal function .
the eikonal functions for @xmath0 and @xmath1 scatterings are the sum of the soft and semihard ( sh ) parton interactions in the hadron - hadron collision , @xmath13 it follows from the qcd parton model that the real part of the eikonal , @xmath14 , can be factored as @xmath15 where @xmath16 is an overlap density for the partons at impact parameter space @xmath7 , @xmath17 and @xmath18 is the usual qcd cross section @xmath19 with @xmath20 , @xmath21 and @xmath22 . the eikonal functions for @xmath0 and @xmath1 scatterings are written as @xmath23 , with @xmath24 and @xmath25 .
since in the parton model @xmath26 decreases rapidly with increasing @xmath11 , the difference between @xmath0 and @xmath1 cross sections is due only to the different weighting of the quark - antiquark ( valence ) annihilation cross sections in the two channels .
thus the crossing - odd eikonal @xmath27 receives no contribution from semihard processes at high energies , and it is sufficient to take @xmath28 and , consequently , @xmath29 .
the connection between the real and imaginary parts of the crossing - even eikonal @xmath30 is obtained by means of dispersion relation @xmath31 valid at @xmath32 gev@xmath33 .
the soft term is given by @xmath34 \right ] , \label{soft01}\end{aligned}\ ] ] where @xmath35 gev and @xmath36 , @xmath37 , @xmath38 , @xmath39 and @xmath40 are fitting parameters .
the odd eikonal @xmath27 is given by @xmath41 where @xmath42 gev and @xmath43 is also a fitting parameter .
in the expression ( [ eq08 ] ) the term @xmath44 is the differential cross section for @xmath45 scattering ( @xmath22 ) .
we select parton - parton processes containing at least one gluon in the initial state , i.e. , we consider the processes @xmath46 @xmath47 @xmath48 where @xmath49 } , \label{eq26}\end{aligned}\ ] ] @xmath50^{- 12/11 } ; \label{mdyna } \end{aligned}\ ] ] here @xmath51 is a non - perturbative qcd effective charge , obtained by cornwall through the use of the pinch technique in order to derive a gauge invariant schwinger - dyson equation for the gluon propagator @xcite .
in the case of semihard gluons we consider the possibility of a `` broadening '' of the spatial distribution .
our assumption suggests an increase of the average gluon radius when @xmath52 increases , and can be properly implemented using the energy - dependent dipole form factor @xmath53 where @xmath54 , with @xmath35 gev . here
@xmath55 and @xmath56 are constants to be fitted . from equation ( [ ref009 ] ) we have @xmath57^{2 } \nonumber \\ & = & \frac{\nu^{2}_{_{sh}}}{96\pi } ( \nu_{_{sh } } b)^{3 } k_{3}(\nu_{_{sh } } b).\end{aligned}\ ] ] since semihard interactions dominate at high energies , we consider an energy - dependence only in the case of @xmath58 . in this way
the soft overlap densities @xmath59 and @xmath60 are static , i.e. , @xmath40 and @xmath61 are not energy dependent .
first , in order to determine the model parameters , we set the value of the gluon mass scale to @xmath62 mev .
this choice for the mass scale is not only consistent to our lo procedure , but is also the one usually obtained in other calculations of strongly interacting processes @xcite .
next , we carry out a global fit to high - energy forward @xmath0 and @xmath1 scattering data above @xmath63 gev , namely the total cross section @xmath64 and the ratio of the real to imaginary part of the forward scattering amplitude @xmath65 .
we use data sets compiled and analyzed by the particle data group @xcite as well as the recent data at lhc from the totem collaboration , with the statistic and systematic errors added in quadrature .
we include in the dataset the first estimate for the @xmath66 parameter made by the totem collaboration in their @xmath66-independent measurement at @xmath67 tev , namely @xmath68 @xcite .
the values of the fitted parameters are given in the legend of figure 1 .
the result @xmath69 for the fit was obtained for 154 degrees of freedom .
the results of the fit to @xmath4 and @xmath66 for both @xmath0 and @xmath1 channels are displayed in figure 1 , together with the experimental data .
the curves depicted in figure 1 were all calculated using the cutoff @xmath70 gev , the value of the cteq6 fixed initial scale @xmath71 .
( @xmath72 ) and @xmath1 ( @xmath73 ) .
the values of the fitted parameters are : @xmath74 gev , @xmath75 gev , @xmath76 gev@xmath77 , @xmath78 gev@xmath77 , @xmath79 gev@xmath77 , @xmath80 , @xmath81 gev and @xmath82 gev@xmath77.,title="fig : " ] [ fig1 ]
in this work we have investigated infrared effects on semihard parton - parton interactions .
we have studied @xmath0 and @xmath1 scattering in the lhc energy region with the assumption that the increase of their total cross sections arises exclusively from semihard interactions . in the calculation of @xmath64 and @xmath65
we have considered the phenomenological implication of an energy - dependent form factor for semihard partons .
we introduce integral dispersion relations to connect the real and imaginary parts of eikonals with energy - dependent form factors . in our analysis
we have included the recent data at lhc from the totem collaboration .
our results show that very good descriptions of @xmath64 and @xmath65 data are obtained by constraining the value of the cutoff at @xmath83 gev .
the @xmath84 for the global fit was 1.062 for 154 degrees of freedom .
this good statistical result shows that our eikonal model , where non - perturbative effects are naturally included via a qcd effective charge , is well suited for detailed predictions of the forward quantities to be measured at higher energies . in the semihard sector
we have considered a form factor in which the average gluon radius increases with @xmath52 . with this assumption
we have obtained another form in which the eikonal can be factored into the qcd parton model , namely @xmath85 . the imaginary part of this _ semi - factorizable _ eikonal was obtained by means of appropriate integral dispersion relations which take into account eikonals with energy - dependent form factors . at the moment we are analyzing the effects of different updated sets of parton distributions on the forward quantities , namely cteq6l and mstw , investigating the uncertainty on these forward observables coming from the uncertainties associated with the dynamical mass scale and the parton distribution functions , and exploring a new ansatz for the energy - dependent form factor .
luna , a.f .
martini , m.j .
menon , a. mihara , and a.a .
natale , phys .
d * 72 * , 034019 ( 2005 ) ; e.g.s .
luna , phys .
b * 641 * , 171 ( 2006 ) ; e.g.s .
luna and a.a .
natale , phys .
d * 73 * , 074019 ( 2006 ) ; e.g.s .
luna , braz .
* 37 * , 84 ( 2007 ) ; d.a .
fagundes , e.g.s .
luna , m.j .
menon , and a.a .
natale , nucl .
a * 886 * , 48 ( 2012 ) ; e.g.s .
luna and p.c .
beggio , nucl .
a * 929 * , 230 ( 2014 ) .
a. corsetti , a. grau , g. pancheri , and y.n .
srivastava , phys .
b * 382 * , 282 ( 1996 ) ; a. grau , g. pancheri , and y.n .
srivastava , phys .
d * 60 * , 114020 ( 1999 ) ; r.m .
godbole , a. grau , g. pancheri , and y.n .
srivastava , phys .
d * 72 * , 076001 ( 2005 ) ; a. achilli _
et al . _ ,
b * 659 * , 137(2008 ) ; a. grau , r.m .
godbole , g. pancheri , and y.n .
srivastava , phys .
b * 682 * , 55 ( 2009 ) ; g. pancheri , d.a .
fagundes , a. grau , s. pacetti , and y.n .
srivastava , arxiv:1301.2925 [ hep - ph ] . | we consider the effect of an energy - dependent dipole form factor in the high - energy behavior of the forward amplitude .
the connection between the semihard parton - level dynamics and the hadron - hadron scattering is established by an eikonal qcd - based model .
our results for the proton - proton ( @xmath0 ) and antiproton - proton ( @xmath1 ) total cross sections , @xmath2 , obtained using the cteq6l1 parton distribution function , are consistent with the recent data from the totem experiment . |
the age of neutron star is an essential parameter , which is relevant to the physics of supernova explosion and thereafter the evolution of stars .
however , it is still a big problem now to determine generally an exact value of age ( except the crab pulsar ) .
it is a conventional and convenient way to obtain the age for rotation - powered neutron stars by equalizing the energy lose rate of spindown to that of magnetodipole radiation , assuming that the inclination angle between magnetic and rotational axes is @xmath0 ( e.g. , manchester & taylor 1977 ) .
the conclusion keeps quantitatively for any @xmath1 , as long as the braking torques due to magnetodipole radiation and the unipolar generator are combined ( xu & qiao 2001 ) .
the resultant age , the so - called characteristic age , is @xmath2 if the initial period @xmath3 is much smaller than the present period @xmath4 .
the age , @xmath5 , is generally considered as the true one for a neutron star with @xmath6 ms since most newborn neutron stars could rotate initially at @xmath7 ms ( e.g. , xu et al . 2002 ) .
it challenges the opinion above that the ages of a few supernova remnants are inconsistent with @xmath5 of their related isolated stars ( table 1 ) , which implies that some additional torque mechanisms do contribute to star braking . among the five stars , three of them have @xmath8 ( the age of supernova remnant ) , and other two @xmath9 .
in addition , electron cyclotron resonant lines are detected in two of the five neutron stars ( 1e 1207.4 - 5209 : bignami et al .
2003 ; 1e 2259 + 568 : iwasawa et al .
1992 ) , but the inferred magnetic fields , @xmath10 , from which are significantly smaller than that in the magnetic dipole radiation model , @xmath11 . the most prominent one in this age discrepancy issue is 1e 1207.4 - 5209 , which has @xmath12 and @xmath13 . one probable and popularly - discussed way to solve the age problem is through an additional accretion torque ( marsden et al .
2001 , alpar et al .
2001 , menou et al .
2001 ) . in this paper ,
whether an additional accretion torque can possibly solve the age discrepancy is investigated , including discussions about possible astrophysical implications .
the key point in the age discrepancy of 1e 1207.4 - 5209 is how to spin down from @xmath14 ms to 424 ms in a short time of @xmath15 kyrs if its true age is @xmath16 .
certainly the problem disappears if one assumes a long initial period @xmath17 ms or a large braking index @xmath18 ( pavlov et al .
2002 ) ; but this is not of occam s razor since it is generally believed that rotation - powered radio pulsars born with @xmath19 ms , and brake with index @xmath20 .
may an additional accretion torque help the spindown ? actually , in an effort to reconcile @xmath11 with @xmath21 , an accretion model for 1e 1207.4 - 5209 was proposed ( xu et al . 2003 ) .
however , a very difficulty in the model is how to choose a time - dependent accretion rate @xmath22 , and to determine the propeller torque with the rate @xmath23 .
nevertheless , the propeller phase works in the centrifugal inhibition regime when @xmath24 ; for a star with mass @xmath25 and magnetic moment @xmath26 , the corotation radius @xmath27^{1/3}p^{2/3}$ ] and the magnetospheric radius @xmath28^{2/7}$ ] . to avoid the complex calculations of magnetohydrodynamics , the rotation energy loss due to propeller torque could be simply introduced as @xmath29 , based on the energy conservation law .
this is unphysical , but should be an limit for accretion braking .
as @xmath30 decreases ( @xmath31 ) , @xmath32 increases , and @xmath33 increases too .
therefore the most efficient spindown ( mesd ) takes place when @xmath34 . in a model where the propeller and electromagnetic torques are combined , in the mesd case
, one can derive the period evolution @xmath35 where @xmath36 g ) and @xmath37 cm ) .
the right hand of eq.([p ] ) is an upper limit of @xmath4 because 1 , a realistic accretion rate may not be as high as that of mesd ; and 2 , the corresponding braking torque is not so effective .
if 1e 1207.4 - 5209 is a conventional neutron star with mass @xmath38 and radius @xmath39 cm , and the line features are related to cyclotron absorptions near the surface ( xu et al .
2003 ; the polar magnetic field is thus @xmath40 g ) , one has @xmath41 ( ms ) .
therefore , assuming 1e 1207.4 - 5209 has a true age @xmath42 kyrs and an initial period @xmath14 ms , the upper limit of the present period is @xmath43 ms ( @xmath44 ms ) , and the age discrepancy can then not be solved in the conventional neutron star model . however ,
if 1e 1207.4 - 5209 is a strange star with low mass , for instance @xmath45 km ( and the mass is thus @xmath46 , since low - mass strange stars have almost a homogenous density @xmath47g/@xmath48 ; alcock et al . 1986 ) , the upper limit is then @xmath49 ( ms ) . in this case , 1e 1207.4 - 5209 could spin down to @xmath50 s during @xmath51 kyrs if its polar magnetic field @xmath40 g. in fact , the fitted radius of 1e 1207.4 - 5209 with a blackbody model is only @xmath52 km ( mereghetti et al .
1996 ; vasisht et al .
1997 ) although a lager radius is possible if a light - element atmosphere is applied ( zavlin , pavlov & trmper 1998 ) . the best - fit tow blackbody model of _ xmm - newton _ data indicates an emitting radius @xmath53 km for the soft component with temperature @xmath54 ev ( bignami et al . 2003 ) . combined with its non - atomic feature spectrum
, we may suggest that 1e 1207.4 - 5209 is a low - mass strange star with bare quark surface ( xu 2002 , xu et al .
2003 ) .
an alternative possibility is that 1e 1207.4 - 5209 is a conventional neutron star , but the cyclotron resonant absorption forms far away from the surface .
the polar magnetic field is @xmath55 ^ 3$]g if the resonant lines form at a height @xmath56 . from eq.([p ] ) , @xmath57 , we estimate a low limit of the polar magnetic field to be @xmath58 g. this implies that the resonant absorption region should be at a level of @xmath59 km height from the surface . certainly , in case of no propeller torque ( i.e. , the polar magnetic field is @xmath60 g ) , the height of resonant absorption region is @xmath61 km .
it is worth noting that 1e 1207.4 - 5209 could be a low - mass neutron star with polar magnetic field @xmath62 . from eq.([p ] ) and the conditions of mesd , one has @xmath63 for @xmath14 ms .
this implies a neutron star with radius @xmath64 km but mass @xmath65 ( e.g. , shapiro & teukolsky 1983 ) .
this result may have difficulties in explaining 1 , a non - atomic spectrum ( xu et al . 2003 ) , and 2 , a possible small radius observed ( mereghetti et al . 1996 ; vasisht et al . 1997 ; bignami et al . 2003 ) and even the fitting result of a neutron star with 10 km and @xmath66 ( zavlin et al .
if other sources list in table 1 are neutron stars with @xmath67 and @xmath68 , the low limits of polar magnetic fields are @xmath69 g for 1e 2259 + 586 , @xmath70 g for psr b1757 - 24 , @xmath71 g for psr j1811 - 1925 , and @xmath72 g for psr j1846 - 0258 . among these sources , only possible cyclotron absorption is found in 1e 2259 + 586 , and the limit field is within the range of that inferred from cyclotron line .
this suggests that the cyclotron resonant may take place just above the stellar surface .
an interesting question is : how much mass could accrete during the propeller phase in case of mesd ? certainly , only a very small part of this matter can accrete onto the stellar surface . when mesd works , one obtains the accretion rate @xmath73 from @xmath74 .
if the quantities are re - scaled , @xmath75 , @xmath76 , and @xmath77 , one has @xmath78 . combining with eq.([p ] ) , one comes to @xmath79 where @xmath80 .
note that the upper limit of accretion mass , @xmath81 , in the right hand of eq.([m ] ) does not depend on the magnetic momentum .
typically for @xmath82 , the upper limit of accretion mass is @xmath83 , which is reasonable since the amount of the fall - back material after supernova explosion could be as high as @xmath84 ( lin , woosley & bodenheimer 1991 ; chevalier 1989 ) . due to @xmath85-mode instability , a nascent neutron star may loss rapidly its angular momentum though gravitational radiation if the initial period is less than @xmath86 ms ( andersson & kokkotas 2001 )
. the upper limit of accretion mass in case of mesd is @xmath87 for @xmath88 .
these results indicate that the fall - back matter is enough to brake the center stars by propeller torque in mesd case .
though the study about mesd torque provides some useful information on the accretion model , including the appropriate magnetic field and the mass of fall - back disk around a neutron star , the accretion rate of mesd torque is questionable in realistic cases . after a dynamical time a fossil disk may form . for a viscosity - driven disk
, the accretion could be in a self - similar way , with an accretion rate of ( cannizzo , lee & goodman 1990 ) @xmath89 where @xmath90 is of order the dynamical time , @xmath91 . assuming @xmath92ms , an initial disk mass @xmath93 , and an opacity dominated by electron scattering ( @xmath94 ) , chatterjee , hernquist and narayan ( 2000 )
develop a first detailed model of fossil disk accretion for anomalous x - ray pulsars ( axps ) .
however it is noted by francischelli and wijers ( 2002 ) that kramers opacity may prevail in the fossil disk ( i.e. , @xmath95 ) . in the regime of conventional neutron stars , we will calculate the accretion torque through the realistic accretion rate of eq.([dotm ] ) , assuming @xmath95 and @xmath96ms , with the inclusion of magnetic dipole radiation . the spinup / spindown torque proposed by menou et al .
( 1999 ) , @xmath97,\ ] ] is applied for the action of fossil disk in the model , where @xmath98 , @xmath99 is the keplerian angular velocity at the magnetospheric boundary .
one may compute the accretion rate , @xmath100 , as well as the spin evolution @xmath101 .
the total disk mass could be @xmath102 .
it is worth noting that the disk mass obtained in this way could be much larger than that in mesd case because of the inclusion of the high accretion in the initial period .
we think that the accretion rate characterized by eq.([dotm ] ) is on average in a sense .
the period derivative , @xmath103 , may be affected by dynamical instabilities or some stochastic processes , whereas the period , @xmath4 , is of the integration over a very long time .
we therefore calculate @xmath104 for any disk mass , @xmath105 , and polar magnetic field , @xmath106 , of neutron stars . for 1e 1207.4 - 5209 ,
the calculated contour of period relative error , @xmath107 ( @xmath108s for 1e 1207.4 - 5209 ) , is shown in fig.1 . a reasonable parameter set (
disk mass @xmath105 and polar magnetic field @xmath106 ) is chosen if the following criteria are met : 1 , the period relative error is smaller ; 2 , @xmath109 ; 3 , @xmath110 ( table 1 ) .
we then have @xmath111 and @xmath112 .
the parameter sets for other sources can also be obtained in this way , which are listed on fig.2 ( except for psr j1811 - 1925 ) .
the period evolution curves , with these parameters , are drawn in fig.2 . note these curves do not change significantly if the parameter sets shift reasonably . the heights , @xmath56 , of cyclotron resonant scattering regions can be obtained , based on the differences of parametric magnetic field , @xmath106 , and the field inferred from absorption features , @xmath113 .
it is found that @xmath114 km and @xmath115 km for 1e 1207.4 - 5209 and 1e 2259 + 586 , respectively , in the model . whether or not the disk will influence the spindown of the neutron star or suppress the radio emission
will depend on the location of @xmath30 relative to the light cylinder radius , @xmath116 , and the corotation radius @xmath117 ( chatterjee et al . 2000 ) .
magnetic dipole radiation dominates , and the disk and the star will effectively evolve independently if @xmath118 ; but in other cases , accretion onto the star will lead to accretion - induced x - ray emission with radio quiet .
we see from fig.2 that the condition of @xmath118 is satisfied only for psr b1757 - 24 when it is older than @xmath119 years . we are therefore not surprise that psr b1757 - 24 is now radio loud whereas the others ( 1e 1207.4 - 5209 , 1e
2259 + 586 , and psr j1846 - 0258 ) are radio quiet .
the axp 1e 2259 + 586 is in a tracking phase , and we expect other two ( 1e 1207.4 - 5209 and psr j1846 - 0258 ) will evolve to be axps when they are in tracking phases too .
psr j1811 - 1925 is an interesting exception among the five sources , whose age is certain if it has physical association with the remnant of a supernova recorded in a.d .
386 . in its calculated contour , we can only choose \{@xmath120 } or \{@xmath121 } ; both the parameter sets are not reasonable ( i.e. , can not meet those 3 criteria ) .
this may imply that the accretion of psr j1811 - 1925 is not self - similar . recalling that the low limit of polar field is only @xmath71 g if in mesd case
, we could suggest that psr j1811 - 1925 has a field within @xmath122 g , with an accretion stronger than that of eq.([dotm ] ) but weaker than that in mesd case .
this result hints that psr j1811 - 1925 is radio quiet ( crawford et al .
in addition , the parametric field , @xmath112 , chosen for 1e 1207.4 - 5209 , which is close to @xmath123 , would also indicate that the real accretion is not described by eq.([dotm ] ) .
in fact the accretion of eq.([dotm ] ) is for the capture of material by black holes where magnetic field is not important , which could differ from that for neutron stars with strong fields .
the possibility of solving the age discrepancy by an accretion - assisted torque is discussed .
we find that : 1 , 1e 1207.4 - 5209 can not be a neutron star , but a low - mass bare strange star , if the cyclotron resonant region is near the polar cap with a magnetic field of @xmath40 g ; whereas it could be a conventional neutron star if the cyclotron lines form at a height of @xmath124 km .
an identification of smaller radius or low altitude cyclotron formation favors a low - mass bare strange star model for 1e 1207.4 - 5209 . 2 , among the five sources with age problems , the magnetic dipole radiation dominates the evolution of psr b1757 - 24 at present , and the others are in propeller ( or tracking ) phases . 3
, the real accretion around these sources may differ from a self - similar one ( eq.([dotm ] ) ) , at least for psr j1811 - 1925 . 4 , by a calculation with self - similar accretion
, it is suggested that psr j1846 - 0258 and 1e 1207.4 - 5209 ( and probably psr j1811 - 1925 ) would evolve to be anomalous x - ray pulsars in the future .
the debris disks formed following supernovae are currently conjectured also for interpreting other astrophysical phenomena , e.g , anomalous x - ray pulsars and soft @xmath125ray repeaters . factually , these disks around the sources could be bright in a wide spectral range .
recent discoveries of possible optical and near - infrared emission from a few axps may be a hint of such kind fallback accretion disks ( 1e 2259 + 586 : hulleman et al .
2001 ; 1rxs j170849 - 400910 : israel et al . 2002 ; 1e 1048.1 - 5937 : wang & chakrabarty 2002 ) .
although a comparison of optical and near - infrared observations with theoretical predictions of spectra of disks around neutron stars ( perna et al . 2000 ) have helped rule out the presence of disks in some cases , more detailed studies in this aspect is still necessary , which may be an effective way to test the fossil - disk model for young neutron stars . _
acknowledgments _ : the helpful suggestions and discussions from an anonymous referee are sincerely acknowledged .
this work is supported by national nature sciences foundation of china ( 10273001 ) and the special funds for major state basic research projects of china ( g2000077602 ) .
llccccccl 1e 1207.4 - 5209 & pks 1209 - 51/52 & 0.424 & 0.07 - 0.3 & 200 - 900 & @xmath51 & 1.7 - 3.6 & 0.06 & 1,2 + 1e 2259 + 586@xmath126 & ctb 109 & 6.98 & 4.84 & 228 & 17 & 59 & 0.4 - 0.9 & 3,4,5 + psr b1757 - 24 & g5.4 - 1.2 & 0.125 & 1.28 & 16 & @xmath127 & 4.0 & & 6 + psr j1811 - 1925 & g11.2 - 0.3 & 0.065 & 0.44 & 24 & 1.6 & 1.7 & & 7 + psr j1846 - 0258 & kes 75 & 0.325 & 71 & 0.72 & 0.9 - 4.3 & 49 & & 8 + | it is found that 1e 1207.4 - 5209 could be a low - mass bare strange star if its small radius or low altitude cyclotron formation can be identified .
the age problems of five sources could be solved by a fossil - disk - assisted torque .
the magnetic dipole radiation dominates the evolution of psr b1757 - 24 at present , and the others are in propeller ( or tracking ) phases . |
[ varying ] in this article we derive an explicit formula for the average @xmath18 where @xmath11 is one of the classical compact lie groups @xmath19 , @xmath1 , or @xmath20 equipped with haar measure @xmath21 of unit mass @xmath22 and @xmath23 depends on a set of complex parameters @xmath24 , which satisfy @xmath25 for all @xmath26 .
the case of @xmath27 is handled in @xcite .
note that @xmath28 with @xmath29 .
this means that @xmath30 is a product of ratios of characteristic polynomials , which explains the title of the article .
the haar average @xmath31 can be regarded as the ( numerical part of the ) character of an irreducible representation of a lie supergroup @xmath32 restricted to a suitable subset of a maximal torus of @xmath33 .
the lie superalgebra @xmath13 is the howe dual partner of the compact group @xmath11 in an orthosymplectic lie superalgebra @xmath34 .
it is naturally represented on a certain infinite - dimensional spinor - oscillator module @xmath35 more concretely , the complex of holomorphic differential forms on the vector space @xmath36 and the irreducible representation is that on the subspace @xmath37 of @xmath11-equivariant forms . to even define the character , we must exponentiate the representation of the lie algebra part of @xmath6 on @xmath38 .
this requires going to a completion @xmath3 of @xmath39 , and can only be done partially .
nevertheless , the represented semigroup contains enough structure to derive laplace - casimir differential equations for its character .
our explicit formula for @xmath31 looks exactly like a classical weyl formula and is derived in terms of the roots of the lie superalgebra @xmath13 and the weyl group @xmath40 .
let us state this formula for @xmath0 , @xmath20 without going into the details of the @xmath41-positive even and odd roots @xmath42 and @xmath43 and the weyl group @xmath40 ( see @xmath44[sect : weylgroup ] for precise formulas ) .
if @xmath45 is the isotropy subgroup of @xmath40 fixing the highest weight @xmath46 , then @xmath47 \in \mathrm{w}/\mathrm{w}_\lambda } \mathrm{e}^ { w(\lambda_n ) } \frac{\prod_{\beta \in \delta_{\lambda,1}^+}(1 - \mathrm{e}^{-w(\beta ) } ) } { \prod_{\alpha \in \delta_{\lambda,0}^+ } ( 1 - \mathrm{e}^{-w(\alpha ) } ) } ( \ln t ) \;.\ ] ] to prove this formula we establish certain properties of @xmath31 which uniquely characterize it and are satisfied by the right - hand side .
these are a weight expansion of @xmath31 ( see corollary [ cor : weightexpansion ] ) , restrictions on the set of weights ( see corollary [ cor : weights ] ) , and the fact that @xmath31 is annihilated by certain invariant differential operators ( see corollary [ cor : dljchi ] ) .
as was stated above , the case @xmath27 is treated in @xcite .
here , we restrict to the compact groups @xmath48 , @xmath49 , and @xmath50 . the cases @xmath51 and @xmath52 can be treated simultaneously .
having established formula ( [ eq : weylcharacter ] ) for @xmath0 , the following argument gives a similar result for @xmath50 .
let @xmath53 and @xmath54 be the unit mass haar measures for @xmath55 and @xmath1 , respectively .
the @xmath55-measure @xmath56 has unit mass on @xmath57 and zero mass on @xmath58 .
it is @xmath57-invariant .
now , @xmath59 with @xmath60 , since @xmath61 . to facilitate the comparison with related work , we now present our final results in the following explicit form .
let @xmath62 and @xmath63 .
consider first the case of the unitary symplectic group @xmath64 ( where @xmath65 ) . then for any pair of non - negative integers @xmath66 in the range @xmath67 one directly infers from ( [ eq : weylcharacter ] ) the formula @xmath68 the sum on the right - hand side is over sign configurations @xmath69 .
the proof proceeds by induction in @xmath70 , starting from the result ( [ eq : weylcharacter ] ) for @xmath71 and sending @xmath72 to pass from @xmath73 to @xmath74 .
in published recent work @xcite the same formula was derived under the more restrictive condition @xmath75 . in @xcite
this unwanted restriction on the parameter range came about because the numerator and denominator on the left - hand side were expanded _ separately _ , ignoring the supersymmetric howe duality ( see @xmath44[sect : osp - howe ] of the present paper ) of the problem at hand . for @xmath76 the same induction process starting from ( [ eq : weylcharacter ] )
yields the result @xmath77 as long as @xmath78 .
please note that this includes even the case of the trivial group @xmath79 with any @xmath80 . for
@xmath81 one has an analogous result where the sum on the right - hand side is over @xmath82 with an _ even _ number of sign reversals .
the very same formulas for @xmath57 and @xmath55 were derived in the recent literature @xcite but , again , only in the much narrower range @xmath83 $ ] . to find an explicit expression for the integral @xmath31
, we first of all observe that the integrand @xmath30 is the supertrace of a representation @xmath84 of a semigroup @xmath85 on the spinor - oscillator module @xmath38 ( cf .
lemma [ lem : strav ] ) .
more precisely , we start with the standard @xmath11-representation space @xmath86 , the @xmath87-graded vector space @xmath88 with @xmath89 , and the abelian semigroup @xmath90 of diagonal transformations in @xmath91 .
we then consider the vector space @xmath92 which is @xmath87-graded by @xmath93 , the infinite - dimensional spinor - oscillator module @xmath94 , and a representation @xmath84 of @xmath85 on @xmath38 .
we also let @xmath95 ( not the weyl group ) .
averaging the product of ratios @xmath30 with respect to the compact group @xmath11 corresponds to the projection from @xmath38 onto the vector space @xmath37 of @xmath11-invariants ( corollary [ cor : weightexpansion ] ) .
now , howe duality ( proposition [ prop : howeduality ] ) implies that @xmath37 is the representation space for an irreducible highest - weight representation @xmath96 of the howe dual partner @xmath13 of @xmath11 in the orthosymplectic lie superalgebra @xmath97 .
this representation @xmath96 is constructed by realizing @xmath98 as a subalgebra in the space of degree - two elements of the clifford - weyl algebra @xmath99 .
precise definitions of these objects , their relationships , and the howe duality statement can be found in @xmath44[sect : osp - howe ] . using the decomposition @xmath100 into weight spaces
, howe duality leads to the weight expansion @xmath101 for @xmath102 . here
@xmath103 denotes the supertrace .
there are strong restrictions on the set of weights @xmath104 .
namely , if @xmath105 , then @xmath106 and @xmath107 for all @xmath108 .
the coefficients @xmath109 are the dimensions of the weight spaces ( multiplied by parity ) .
note that the set of weights of the representation @xmath110 of @xmath13 on @xmath37 is infinite . before outlining the strategy for computing our character in the infinite - dimensional setting of representations of lie superalgebras and groups , we recall the classical situation where @xmath110 is an irreducible finite - dimensional representation of a reductive lie algebra @xmath13 and @xmath84
is the corresponding lie group representation of the complex reductive group @xmath33 . in that case
the character @xmath111 of @xmath84 , which automatically exists , is the trace @xmath112 , which is a radial eigenfunction of every laplace - casimir operator .
these differential equations can be completely understood by their behavior on a maximal torus of @xmath33 . in our case
we must consider the infinite - dimensional irreducible representation @xmath110 of the lie superalgebra @xmath113 on @xmath37 .
casimir elements , laplace - casimir operators of @xmath6 , and their radial parts have been described by berezin @xcite . in the situation @xmath114 at hand
we have the additional feature that every @xmath6-casimir element @xmath115 can be expressed as a bracket @xmath116 $ ] where @xmath117 is the holomorphic exterior derivative when we view @xmath118 as the complex of @xmath11-equivariant holomorphic differential forms on @xmath5 . to benefit from berezin s theory of radial parts , we construct a radial superfunction @xmath119 which is defined on an open set containing the torus @xmath120 such that its numerical part satisfies @xmath121 for all @xmath122 .
if we had a representation @xmath123 of a lie supergroup @xmath124 at our disposal , we could define @xmath119 to be its character , i.e. @xmath125 ( see @xmath44[supergroup characters ] ) . since we do nt have such a representation ,
our idea is to define @xmath119 as a character on a totally real submanifold @xmath126 of maximal dimension which contains a real form of @xmath120 and is invariant with respect to conjugation by a real form @xmath127 of @xmath128 , and then to extend @xmath119 by analytic continuation .
thinking classically we consider the even part of the lie superalgebra @xmath129 , which is the lie algebra @xmath130 .
the real structures at the lie supergroup level come from a real form @xmath131 of @xmath132 .
the associated real forms of @xmath133 and @xmath134 are the real orthogonal lie algebra @xmath135 and the real symplectic lie algebra @xmath136 .
these are defined in such a way that the elements in @xmath137 and @xmath138 are mapped as elements of the clifford - weyl algebra via the spinor - oscillator representation to anti - hermitian operators on @xmath38 with respect to a compatibly defined unitary structure . in this context
we frequently use the unitary representation of the real heisenberg group @xmath139 on the completion @xmath3 of the module @xmath38 .
since @xmath140 has finite dimension , exponentiating the spinor representation of @xmath135 causes no difficulties .
this results in the spinor representation of @xmath141 , a 2:1 covering of the compact group @xmath142 .
so in this case one easily constructs a representation @xmath143 which is compatible with @xmath144 .
exponentiating the oscillator representation of @xmath145 on the infinite - dimensional vector space @xmath146 requires more effort . in @xmath44[metaplectic rep
] , following howe @xcite , we construct the shale - weil - segal representation @xmath147 of the metaplectic group @xmath148 which is the 2:1 covering group of the real symplectic group @xmath149 .
this is compatible with @xmath150 .
altogether we see that the even part of the lie superalgebra representation integrates to @xmath151 .
the construction of @xmath152 uses a limiting process coming from the oscillator semigroup @xmath153 , which is the double covering of the contraction semigroup @xmath154 and has @xmath155 in its boundary .
furthermore , we have @xmath156 where @xmath126 is an analytic totally real submanifold of maximal dimension which contains a real form of the torus @xmath157 ( see @xmath44[oscillator semigroup ] ) .
the representation @xmath158 constructed in @xmath44[sec : semigrouprep ] facilitates the definition of the representation @xmath152 and of the character @xmath119 in @xmath44[subsec : existence ] .
it should be underlined that proposition [ holomorphicity ] ensures convergence of the superfunction @xmath159 , which is defined as a supertrace and exists for all @xmath160 .
on that basis , the key idea of our approach is to exploit the fact that every casimir invariant @xmath161 is exact in the sense that @xmath162 $ ] . by a standard argument
, this exactness property implies that every such invariant @xmath115 vanishes in the spinor - oscillator representation .
this result in turn implies for our character @xmath119 the differential equations @xmath163 where @xmath164 is the laplace - casimir operator representing @xmath115 . by drawing on berezin s theory of radial parts
, we derive a system of differential equations which in combination with certain other properties ultimately determines @xmath119 . in the case of @xmath81 the lie group associated to the even part of the real form of the howe partner @xmath13 is embedded in a simple way in the full group @xmath127 described above .
it is itself just a lower - dimensional group of the same form . in the case of @xmath52
a sort of reversing procedure takes places and the analogous real form is @xmath165 .
nevertheless , the precise data which are used as input into the series developments , the uniqueness theorem and the final calculations of @xmath119 are essentially the same in the two cases .
therefore there is no difficulty handling them simultaneously .
in this chapter we collect some foundational information from representation theory .
basic to our work is the orthosymplectic lie superalgebra , @xmath166 , in its realization as the space spanned by supersymmetrized terms of degree two in the clifford - weyl algebra . representing the latter by its fundamental representation on the spinor - oscillator module ,
one gets a representation of @xmath6 and of all howe dual pairs inside of @xmath167 .
roots and weights of the relevant representations are described in detail .
a _ @xmath87-grading _ of a vector space @xmath168 over @xmath169 or @xmath4 is a decomposition @xmath170 of @xmath168 into the direct sum of two @xmath171-vector spaces @xmath172 and @xmath173 .
the elements in @xmath174 are called _
homogeneous_. the _ parity function _
@xmath175 , @xmath176 , assigns to a homogeneous element its parity .
we write @xmath177 if @xmath178 and @xmath179 .
a _ lie superalgebra _ over @xmath171 is a @xmath87-graded @xmath171-vector space @xmath180 equipped with a bilinear map @xmath181 : \ , \mathfrak{g } \times \mathfrak{g } \to \mathfrak{g}$ ] satisfying 1 .
@xmath182 \subset \mathfrak{g}_{s + s^\prime}\,$ ] , i.e. , @xmath183|= |x| + |y|$ ] ( mod 2 ) for homogeneous elements @xmath184 .
2 . skew symmetry : @xmath185 = -(-1)^{|x||y|}[y , x]$ ] for homogeneous @xmath184 .
jacobi identity , which means that @xmath186 : \ , \mathfrak{g } \to \mathfrak{g}$ ] is a ( super-)derivation : @xmath187 = [ \mathrm{ad}(x)y , z ] + ( -1)^{|x||y|}[y,\mathrm{ad}(x)z ] \;.\ ] ] [ @xmath188 let @xmath189 be a @xmath87-graded @xmath171-vector space .
there is a canonical @xmath87-grading @xmath190 induced by the grading of @xmath168 : @xmath191 the bilinear extension of @xmath185 : = xy - ( -1)^{|x||y|}yx$ ] for homogeneous elements @xmath192 to a bilinear map @xmath181 : \ , \mathrm{end}(v ) \times \mathrm{end}(v ) \to \mathrm{end}(v)$ ] gives @xmath193 the structure of a lie superalgebra , namely @xmath194 .
the jacobi identity in this case is a direct consequence of the associativity @xmath195 and the definition of @xmath185 $ ] .
in fact , for every @xmath87-graded associative algebra @xmath196 the bracket @xmath181 : \ , \mathcal{a } \times \mathcal{a } \to \mathcal{a}$ ] defined by @xmath185 = xy - ( -1)^{|x| |y| } yx$ ] satisfies the jacobi identity .
given a complex @xmath87-graded vector space @xmath197 , consider the complex lie superalgebra @xmath198 .
equip @xmath168 with a unitary structure so that @xmath199 .
for @xmath200 we have the unique decomposition @xmath201 with @xmath202 .
let @xmath203 be the complex anti - linear involution defined by @xmath204 , where @xmath205 ( complex conjugation and transpose ) is defined by the unitary structure .
@xmath206 is an automorphism of the lie superalgebra @xmath194 .
indeed , @xmath207)&=&\sum\nolimits_{s,\ , s^\prime } \sigma(x_s y_{s^\prime})-\sum\nolimits_{s,\ , s^\prime}(-1)^{s s^\prime } \sigma(y_{s^\prime}x_s ) \\&=&-\sum\nolimits_{s,\ , s^\prime } \mathrm{i}^{(s - s^\prime)^2 } \big(\overline{y}_{s^\prime}^\mathrm{\,t } \overline{x}_s^\mathrm{\,t } - ( -1)^{s s^\prime } \overline{x}_s^ \mathrm{\,t } \overline{y}_{s^\prime}^\mathrm{\,t } \big ) \;,\end{aligned}\ ] ] which is the same as @xmath208 \;.\end{aligned}\ ] ] the set of fixed points @xmath209 is a real lie superalgebra , a real form of @xmath210 called the unitary lie superalgebra @xmath211 .
note that @xmath212 is the compact real form @xmath213 of the complex lie algebra @xmath214 .
[ exa : osp ] let @xmath189 be a @xmath87-graded @xmath171-vector space and put @xmath215 .
the @xmath87-grading of @xmath168 induces a @xmath87-grading @xmath216 in the obvious manner : @xmath217 and @xmath218 .
then consider the canonical alternating bilinear form @xmath219 on @xmath220 , @xmath221 and the canonical symmetric bilinear form @xmath222 on @xmath223 , @xmath224 the _ orthosymplectic form _ of @xmath132 is the non - degenerate bilinear form @xmath225 defined as the orthogonal sum @xmath226 @xmath227 note the exchange symmetry @xmath228 for @xmath229
. given @xmath230 , define a complex linear bijection @xmath231 by the equation @xmath232 for all @xmath229 .
it is easy to check that @xmath233 has the property @xmath234 which implies that @xmath233 is an involutory automorphism of the lie superalgebra @xmath235 with bracket @xmath185 = xy - ( -1)^{|x| \mathrm{end}(w)$ ] of @xmath233-fixed points is closed w.r.t .
that bracket ; it is called the ( complex ) _ orthosymplectic lie superalgebra _ of @xmath132 .
[ exa : jh ] using the notation of example [ exa : osp ] , consider the vector space @xmath236 and take it to be @xmath87-graded by @xmath237 and @xmath238 . define a bilinear mapping @xmath239 : \ , \widetilde{w } \times \widetilde{w } \to \widetilde{w}$ ] by @xmath240 = [ \widetilde{w},\mathbb{k } ] = 0 \ ; , \quad [ w , w ] \subset \mathbb{k } \ ; , \quad [ w , w^\prime ] = q(w , w^\prime ) \quad ( w , w^\prime \in w ) \;.\ ] ] by the basic properties of the orthosymplectic form @xmath230 , the vector space @xmath241 equipped with this bracket is a lie superalgebra the so - called jordan - heisenberg algebra .
note that @xmath241 is two - step nilpotent , i.e. , @xmath242 ] = 0\,$ ] .
let @xmath170 be a @xmath87-graded @xmath243-vector space , and recall the decomposition @xmath244 . for @xmath200 ,
we denote by @xmath245 the corresponding decomposition of an operator .
the _ supertrace _ on @xmath168 is the linear function @xmath246 ( if @xmath247 , then usually the domain of definition of @xmath103 must be restricted . )
an @xmath248-_invariant bilinear form _ on a lie superalgebra @xmath249 over @xmath171 is a bilinear mapping @xmath250 with the properties 1 .
@xmath251 and @xmath252 are @xmath253-orthogonal to each other ; 2 .
@xmath253 is symmetric on @xmath251 and skew on @xmath254 ; 3 .
@xmath255,z ) = b(x,[y , z])$ ] for all @xmath256 . we will repeatedly use the following direct consequences of the definition of @xmath103 .
[ lem : str ] if @xmath13 is a lie superalgebra in @xmath193 , the trace form @xmath257 is an @xmath258-invariant bilinear form .
one has @xmath259
= 0\,$ ] . recalling the setting of example [ exa : osp ] , note that the supertrace for @xmath260 is odd under the @xmath261-automorphism @xmath233 fixing @xmath97 , i.e. , @xmath262 .
it follows that @xmath263 for any @xmath264 .
moreover , @xmath265 for any product of an odd number of @xmath6-elements .
let @xmath13 be a lie superalgebra with bracket @xmath266 $ ] .
the universal enveloping algebra @xmath267 is defined as the quotient of the tensor algebra @xmath268 by the two - sided ideal @xmath269 generated by all combinations @xmath270\ ] ] for homogeneous @xmath271 . if @xmath272 is the image of @xmath273 under the projection @xmath274 , the algebra @xmath267 is filtered by @xmath275 .
the @xmath87-grading @xmath276 gives rise to a @xmath87-grading of @xmath277 by @xmath278 and this in turn induces a canonical @xmath87-grading of @xmath279 .
one might imagine introducing various bracket operations on @xmath280 and/or @xmath267 .
however , in view of the canonical @xmath87-grading , the natural bracket operation to use is the _ supercommutator _ , which is the bilinear map @xmath281 defined by @xmath282 for homogeneous elements @xmath283 .
( for the time being , we use a different symbol @xmath284 for better distinction from the bracket @xmath285 $ ] on @xmath286 . ) since by the definition of @xmath269 one has @xmath287 the supercommutator descends to a well - defined map @xmath288 .
[ lem:2.2 ] if @xmath13 is a lie superalgebra , the supercommutator @xmath289 gives @xmath267 the structure of another lie superalgebra in which @xmath290 .
compatibility with the @xmath87-grading , skew symmetry , and jacobi identity are properties of @xmath284 that are immediate at the level of the tensor algebra @xmath277 . they descend to the corresponding properties at the level of @xmath291 by the definition of the two - sided ideal @xmath269 .
thus @xmath267 with the bracket @xmath292 is a lie superalgebra . to see that @xmath293 is contained in @xmath294 , notice that this property holds true for @xmath295 by the defining relations @xmath296 of @xmath279
then use the associative law for @xmath267 to verify the formula @xmath297 for homogeneous @xmath298 .
the claim now follows by induction on the degree of the filtration @xmath299 . by definition ,
the supercommutator of @xmath267 and the bracket of @xmath13 agree at the linear level : @xmath300 $ ] for @xmath301 .
it is therefore reasonable to drop the distinction in notation and simply write @xmath302 $ ] for both of these product operations .
this we now do . for future use , note the following variant of the preceding formula : if @xmath303 are any homogeneous elements of @xmath286 , then @xmath304 = \sum_{i=1}^k ( -1)^{|x|\sum_{j = i+1}^k |y_j|}\ , y_1 \cdots y_{i-1}\ , [ y_i\ , , x ] \ , y_{i+1 } \cdots y_k \;,\ ] ] which expresses the supercommutator in @xmath267 by the bracket in @xmath286 . for a @xmath87-graded @xmath171-vector space @xmath305
let @xmath306 as in example [ exa : osp ] . the _ orthogonal lie algebra _
@xmath307 is the lie algebra of the lie group @xmath308 of @xmath171-linear transformations of @xmath309 that leave the non - degenerate symmetric bilinear form @xmath222 invariant .
this means that @xmath310 is in @xmath307 if and only if @xmath311 similarly , the _ symplectic lie algebra _
@xmath134 is the lie algebra of the automorphism group @xmath312 of @xmath313 equipped with the non - degenerate alternating bilinear form @xmath314 @xmath315 for the next statement , recall the definition of the orthosymplectic lie superalgebra @xmath97 and the decomposition @xmath316 .
[ lem : iso - osp1 ] as lie algebras resp .
vector spaces , @xmath317 the first isomorphism follows directly from the definitions . for the second isomorphism , decompose @xmath318 as @xmath319 where @xmath320 and @xmath321 .
then @xmath322 since both @xmath222 and @xmath219 are non - degenerate , the component @xmath323 is determined by the component @xmath324 , and one therefore has @xmath325 .
we now review how @xmath134 and @xmath307 decompose for our case @xmath326 . for that purpose ,
if @xmath327 is a vector space with dual vector space @xmath328 , let @xmath329 and @xmath330 denote the symmetric resp .
alternating linear maps from @xmath327 to @xmath328 .
[ lem : iso - osp2 ] as vector spaces , @xmath331 there is a canonical decomposition @xmath332 for @xmath333 .
let @xmath334 and write the corresponding decomposition of @xmath310 as @xmath335 substituting @xmath336 and @xmath337 , the defining condition @xmath338 for @xmath339 then transcribes to @xmath340 for all @xmath341 and @xmath342 .
thus @xmath343 , and the maps @xmath344 are alternating .
this already proves the statement for the case of @xmath307 .
the situation for @xmath134 is identical but for a sign change : the symmetric form @xmath222 is replaced by the alternating form @xmath345 , and this causes the parity of @xmath344 to be reversed . by adding up dimensions , lemmas [ lem : iso - osp1 ] and [ lem : iso - osp2 ]
entail the following consequence .
[ cor : dimosp ] as a @xmath87-graded vector space , @xmath346 is isomorphic to @xmath347 where @xmath348 , @xmath349 , and @xmath350
. there exists another way of thinking about @xmath97 , which will play a key role in the sequel . to define it and keep the sign factors consistent and transparent
, we need to be meticulous about our ordering conventions . hence ,
if @xmath351 is a vector and @xmath352 is a linear function , we write the value of @xmath353 on @xmath354 as @xmath355 based on this notational convention , if @xmath168 is a @xmath356-graded vector space and @xmath200 is a homogeneous operator , we define the _ supertranspose _ @xmath357 of @xmath358 by @xmath359 this definition differs from the usual transpose by a change of sign in the case when @xmath358 has a component in @xmath360 . from it , it follows directly that the negative supertranspose @xmath361 , @xmath362 is an isomorphism of lie superalgebras : @xmath363^\mathrm{st } = [ -x^\mathrm{st } , -y^\mathrm{st } ] \;.\ ] ] the modified notion of transpose goes hand in hand with a modified notion of what it means for an operator in @xmath364 or @xmath365 to be symmetric .
thus , define the subspace @xmath366 to consist of the elements , say @xmath367 , which are symmetric in the @xmath87-graded sense : @xmath368 by the same principle , define @xmath369 as the set of solutions @xmath370 of @xmath371 to make the connection with the decomposition of lemma [ lem : iso - osp1 ] and [ lem : iso - osp2 ] , notice that @xmath372 and similar for the corresponding intersections involving @xmath373 .
next , expressing the orthosymplectic form @xmath374 of @xmath375 as @xmath376 and writing out the conditions resulting from @xmath377 for the case of @xmath378 and @xmath379 , one sees that @xmath380 this situation is summarized in the next statement .
[ lem : osp - iso3 ] the orthosymplectic lie superalgebra of @xmath381 decomposes as @xmath382 where @xmath383 , and @xmath384 , and @xmath385 the decomposition of lemma [ lem : osp - iso3 ] can be regarded as a @xmath386-grading of @xmath97 . by the ` block ' structure inherited from @xmath260 ,
this decomposition is compatible with the bracket @xmath181:$ ] @xmath387 \subset \mathfrak{g}^{(m + m^\prime ) } \;,\ ] ] where @xmath388 if @xmath389 .
it follows that each of the three subspaces @xmath390 , @xmath391 , and @xmath392 is a lie superalgebra , the first two with vanishing bracket .
[ lem : iso - osp4 ] the embedding @xmath393 by @xmath394 projected to @xmath97 is an isomorphism of lie superalgebras @xmath395 .
since the negative supertranspose @xmath396 is a homomorphism of lie superalgebras , so is our embedding @xmath397 .
this map is clearly injective . to see that it is surjective , consider any homogeneous @xmath398 viewed as an operator in @xmath399 .
the condition for @xmath358 to be in @xmath97 is ( [ eq : def - osp ] ) . to get a non - trivial condition ,
choose @xmath400 or @xmath401 .
the first choice gives @xmath402 valid for all @xmath403 and @xmath352 , this implies that @xmath404 .
the second choice leads to the same conclusion .
thus @xmath405 is in @xmath97 if and only if @xmath406 .
in the following subsections we will often write @xmath407 for short
. a cartan subalgebra of a lie algebra @xmath251 is a maximal commutative subalgebra @xmath408 such that @xmath251 ( or its complexification if @xmath251 is a real lie algebra ) has a basis consisting of eigenvectors of @xmath409 for all @xmath410 .
recall that @xmath183|=|x|+|y|$ ] for homogeneous elements @xmath184 of a lie superalgebra @xmath286 . from the vantage point of decomposing @xmath13 by eigenvectors or root spaces ,
it is therefore reasonable to call a cartan subalgebra of @xmath251 a cartan subalgebra of @xmath286 .
we will see that @xmath411 and @xmath412 = 0 $ ] for all @xmath413 imply @xmath414 , i.e. , there exists no commutative subalgebra of @xmath6 that properly contains a cartan subalgebra .
lie superalgebras with this property are called of type i in @xcite .
let us determine a cartan subalgebra and the corresponding root space decomposition of @xmath166 .
for @xmath415 choose bases @xmath416 of @xmath417 and associated dual bases @xmath418 of @xmath419 .
then for @xmath420 and @xmath421 define rank - one operators @xmath422 by the equation @xmath423 for all @xmath424 and @xmath425 . these form
a basis of @xmath193 , and by lemma [ lem : iso - osp4 ] the operators @xmath426 form a basis of @xmath427 .
similarly , let bases of @xmath428 and @xmath364 be defined by @xmath429 for index pairs in the appropriate range . then by lemma
[ lem : osp - iso3 ] and equations ( [ eq : sym - b ] , [ eq : sym - c ] ) the subalgebras @xmath391 and @xmath430 are generated by the sets of operators @xmath431 since @xmath432 , a cartan subalgebra of @xmath6 is the direct sum of a cartan subalgebra of @xmath307 and a cartan subalgebra of @xmath134 .
letting @xmath433 be the span of the diagonal operators @xmath434 one has that @xmath433 is a cartan subalgebra of @xmath167 .
indeed , if @xmath435 denotes the basis of @xmath436 dual to @xmath437 , inspection of the adjoint action of @xmath433 on @xmath6 gives the following result .
[ lem : rootsosp ] the operators @xmath438 are eigenvectors of @xmath409 for all @xmath439 @xmath440 = \begin{cases } ( \vartheta_{s j } - \vartheta_{t k})(h ) \ , x^{(m)}_{s j\ , , \ , t k } & m = 0 \ ; , \\ ( \vartheta_{s j } + \vartheta_{t k})(h ) \ , x^{(m)}_{s j\ , , \ , t k } & m = -2 \ ; , \\
( - \vartheta_{s j } - \vartheta_{t k})(h)\ , x^{(m)}_{s j\ , , \,t k } & m = 2\;.\end{cases}\ ] ] a root of a lie superalgebra @xmath13 is called _ even _ if its root space is in @xmath441 , it is called _ odd _ if its root space is in @xmath254 .
we denote by @xmath442 and @xmath443 the set of even roots and the set of odd roots , respectively .
for @xmath444 we have @xmath445 as before , let @xmath249 be a lie superalgebra , and let @xmath446 be its universal enveloping algebra .
denote the symmetric algebra of @xmath251 by @xmath447 and the exterior algebra of @xmath252 by @xmath448 . the poincar - birkhoff - witt theorem for lie superalgebras states that for each @xmath8 there is a bijective correspondence @xmath449 the collection of inverse maps lift to a vector - space isomorphism
, @xmath450 called the super - symmetrization mapping .
in other words , given a homogeneous basis @xmath451 of @xmath452 , each element @xmath453 can be uniquely represented in the form @xmath454 with super - symmetrized coefficients , i.e. , @xmath455 the isomorphism @xmath456 gives @xmath291 a @xmath386-grading ( by the degree @xmath8 ) .
now recall that @xmath267 comes with a canonical bracket operation , the supercommutator @xmath181 : \ , \mathsf{u } ( \mathfrak{g } ) \times \mathsf{u}(\mathfrak{g } ) \to \mathsf{u } ( \mathfrak{g})$ ] .
an element @xmath457 is said to lie in the center of @xmath279 , and is called a _ casimir element _ , iff @xmath185= 0 $ ] for all @xmath458 .
by the formula ( [ eq : brackets ] ) , a necessary and sufficient condition for that is @xmath185 = 0 $ ] for all @xmath459 . in the case of @xmath460 , for every @xmath461
there is a casimir element @xmath462 of degree @xmath463 , which is constructed as follows .
consider the bilinear form @xmath464 given by the supertrace ( in some representation ) , @xmath465 . recall that this form is @xmath248-invariant , which is to say that @xmath255,z ) = b(x,[y , z])$ ] for all @xmath466 .
taking the supertrace in the fundamental representation of @xmath34 , the form @xmath253 is non - degenerate , and therefore , if @xmath467 is a homogeneous basis of @xmath166 , there is another homogeneous basis @xmath468 of @xmath6 so that @xmath469 .
note @xmath470 and put @xmath471 ( notice that , in view of the remark following lemma [ lem : str ] , there is no point in making the same construction with an odd number of factors . )
[ lem : ilcasimir ] for all @xmath472 the element @xmath462 is casimir , and @xmath473 . by specializing the formula ( [ eq : brackets ] ) to the present case , @xmath474 = \sum \sum_{k=1}^{2\ell } ( -1)^{|x| ( |\widetilde{e}_{i_{k+1}}| + \ldots + |\widetilde{e}_{i_{2\ell}}| ) } \widetilde{e}_{i_1 } \cdots \ , \widetilde{e}_{i_{k-1 } } \ , [ \widetilde{e}_{i_k } , x ] \ , \widetilde{e}_{i_{k+1 } } \cdots \ , \widetilde{e}_{i_{2\ell}}\,\mathrm{str}\,(e_{i_{2\ell}}\cdots\ , e_{i_1})\;.\ ] ] now if @xmath475 = \sum_j x_{ij}\ , \widetilde{e}_j$ ] then from @xmath248-invariance , @xmath476 , e_j ) = x_{ij } = b(\widetilde{e}_i \ , , [ x , e_j])$ ] , one has @xmath477 = \sum_i e_i\ , x_{ij}\,$ ] .
using this relation to transfer the @xmath478-action from @xmath479 to @xmath480 , and reading the formula ( [ eq : brackets ] ) backwards , one obtains @xmath481 = \sum \widetilde{e}_{i_1}\cdots \,\widetilde{e } _ { i_{2\ell}}\ , \mathrm{str}\,([x , e_{i_{2\ell}}\cdots\ , e_{i_1 } ] ) \;.\ ] ] since the supertrace of any bracket vanishes , one concludes that @xmath482 = 0 $ ] . the other statement , @xmath483 , follows from @xmath484 , the additivity of the @xmath87-degree and the fact that @xmath485 for any odd element @xmath486 .
now define @xmath496 to be the following odd element of @xmath497 : @xmath498 [ lem : cas - exact ] let @xmath346 be the orthosymplectic lie superalgebra for a @xmath87-graded vector space @xmath168 with isomorphic components @xmath487 .
then for all @xmath461 the casimir element @xmath462 is expressible as a bracket : @xmath499\,$ ] . by the same argument as in the proof of lemma [ lem : ilcasimir ] , @xmath500 = - \sum \widetilde{e}_{i_1}\cdots \ , \widetilde{e}_{i_{2\ell}}\,\mathrm{str}\,([\partial , e_{i_{2\ell } } \cdots
e_{i_1 } ] \ , \widetilde{\partial } \lambda ) \;.\ ] ] using the relations @xmath501 = 0 $ ] and @xmath502 , one has for any @xmath503 that @xmath504\ , \widetilde{\partial } \lambda ) = \mathrm{str}\ , ( \widetilde{\partial } \lambda \ , [ \partial , \ , a ] ) = \mathrm{str}\ , ( [ \widetilde{\partial } , \partial ] \lambda \ , a ) = \mathrm{str}\,(\lambda^2 a ) = \mathrm{str}\ , ( a ) \;,\ ] ] where the second equality sign is from @xmath505 ) = \mathrm{str}\ , ( [ c , b]\ , a)$ ] .
the statement of the lemma now follows on setting @xmath506 .
as we shall see , lemma [ lem : cas - exact ] has the drastic consequence that all @xmath6-casimir elements @xmath462 are zero in the spinor - oscillator representation of @xmath346 for @xmath487 . in the present context , a pair @xmath507 of subalgebras @xmath508 of a lie superalgebra @xmath13
is called a _ dual pair _ whenever @xmath509 is the centralizer of @xmath433 in @xmath13 and vice versa . in this subsection ,
let @xmath510 . given a @xmath87-graded complex vector space @xmath511 we let @xmath512 , where @xmath86 is equipped with the standard representation of @xmath513 , @xmath514 , or @xmath515 , as the case may be .
as a result , the lie algebra @xmath516 of whichever group is represented on @xmath86 is embedded in @xmath346 .
we will now describe the dual pairs @xmath517 in @xmath518 for @xmath519 .
these are known as dual pairs in the sense of r. howe .
let us begin by recalling that for any representation @xmath520 of a group @xmath11 on a vector space @xmath521 , the dual representation @xmath522 on the linear forms on @xmath523 is given by @xmath524 . by this token
, every representation @xmath525 induces a representation @xmath526 of @xmath11 on @xmath260 .
let @xmath527 be any representation of a lie group @xmath528 . if @xmath529 for a @xmath87-graded complex vector space @xmath530 , the induced representation @xmath531 of the lie algebra @xmath516 of @xmath11 on @xmath260 is a subalgebra of @xmath532 .
the @xmath11-action on @xmath533 by @xmath534 preserves the canonical pairing @xmath535 between @xmath86 and @xmath536 .
consequently , the @xmath11-action on @xmath537 by @xmath538 preserves the canonical pairing @xmath539 .
since the orthosymplectic form @xmath540 uses nothing but that pairing , it follows that @xmath541 passing to the lie algebra level one obtains @xmath542 .
the operator @xmath543 preserves the @xmath87-grading of @xmath132 ; therefore one actually has @xmath544 .
let us now assume that the complex lie group @xmath11 is defined by a non - degenerate bilinear form @xmath545 in the sense that @xmath546 we then have a canonical isomorphism @xmath547 by @xmath548 , and an isomorphism @xmath549 by @xmath550 .
[ lem:2.11 ] @xmath551 for all @xmath552 .
if @xmath553 , @xmath554 , and @xmath555 , then by the definition of @xmath556 and @xmath543 , @xmath557 since @xmath253 is @xmath11-invariant , one has @xmath558 , and therefore @xmath559 thus @xmath560 .
let us now examine what happens to the orthosymplectic form @xmath374 on @xmath132 when it is pulled back by the isomorphism @xmath556 to a bilinear form @xmath561 on @xmath562 @xmath563 by definition , @xmath564 , and writing @xmath565 where @xmath566 if @xmath253 is symmetric and @xmath567 if @xmath253 is alternating , we obtain @xmath568 in view of this , let @xmath569 denote the vector space @xmath570 with the twisted @xmath87-grading , i.e.@xmath571 ( @xmath572 ) .
moreover , notice that @xmath556 determines an embedding @xmath573 whose restriction to @xmath574 is an injective homomorphism . in the following we often write @xmath575 and @xmath576 for short .
[ cor:2.2 ] for @xmath81 and @xmath577 , the map @xmath578 defines a lie superalgebra embedding into @xmath97 of @xmath579 resp .
@xmath580 .
for @xmath81 the bilinear form @xmath253 of @xmath86 is symmetric and the bilinear form @xmath374 of @xmath132 pulls back see ( [ eq : pull - back ] ) to the standard orthosymplectic form of @xmath581 . for @xmath577 ,
the form @xmath253 is alternating .
its pullback , the orthosymplectic form of @xmath582 twisted by the sign factor @xmath583 , is restored to standard form by switching to the @xmath87-graded vector space @xmath584 with the twisted @xmath87-grading .
to go further , we need a statement concerning @xmath585 , the space of @xmath33-equivariant homomorphisms between two modules @xmath586 and @xmath587 for a group @xmath128 .
[ lem : homg ] let @xmath588 be finite - dimensional vector spaces all of which are representation spaces for a group @xmath128 .
if the @xmath33-action on @xmath589 and @xmath590 is trivial , then @xmath591 @xmath592 is canonically isomorphic to @xmath593 as a @xmath33-representation space , with @xmath33-equivariant maps corresponding to @xmath33-invariant tensors .
since the @xmath33-action on @xmath594 is trivial , one sees that @xmath595 is isomorphic to the tensor product of @xmath596 with the space of @xmath33-invariants in @xmath597 .
the latter in turn is isomorphic to @xmath598 .
[ prop : dualpairs ] writing @xmath599 for @xmath600 , @xmath601 , @xmath602 , the following pairs are dual pairs in @xmath603 @xmath604 , @xmath605 , @xmath606 . here
we calculate the centralizer of @xmath516 in @xmath97 for each of the three cases @xmath607 , @xmath608 , @xmath609 and refer the reader to @xcite for the remaining details .
since both @xmath610 and @xmath611 are @xmath11-invariant subspaces , @xmath612 decomposes as @xmath613 by schur s lemma , @xmath614 , and therefore lemma [ lem : homg ] implies @xmath615 by the same reasoning , @xmath616 . applying lemma [ lem : homg ] to the two remaining summands , we obtain @xmath617 plus the same statement where each vector space is replaced by its dual . if @xmath618 , then @xmath619 .
hence , @xmath620 for @xmath621 is an isomorphism .
this means that the centralizer of @xmath622 in @xmath97 is the intersection @xmath623 , which can be identified with @xmath624 .
thus we have the first dual pair , @xmath625 . in the case of @xmath0 ,
the discussion is shortened by recalling lemma [ lem:2.11 ] and the @xmath11-equivariant isomorphism @xmath549 . by schur
s lemma , these imply @xmath626 . from corollary [ cor:2.2 ]
it then follows that the intersection @xmath627 is isomorphic as a lie superalgebra to @xmath628 . passing to the lie algebra level for @xmath11
, we get the second dual pair , @xmath629 .
finally , if @xmath577 , the situation is identical except that corollary [ cor:2.2 ] compels us to switch to the @xmath87-twisted structure of orthosymplectic lie superalgebra in @xmath630 .
this gives us the third dual pair , @xmath631 .
let @xmath632 or @xmath633 ( in this subsection the choice of number field again is immaterial ) and recall from example [ exa : jh ] the definition of the jordan - heisenberg lie superalgebra @xmath634 , where @xmath635 is a @xmath87-graded vector space with components @xmath636 and @xmath637 .
the universal enveloping algebra of the jordan - heisenberg lie superalgebra is called the _ clifford - weyl algebra _ ( or quantum algebra ) .
we denote it by @xmath638 .
equivalently , one defines the clifford - weyl algebra @xmath99 as the associative algebra generated by @xmath639 subject to the following relations for all @xmath640 @xmath641 in particular , @xmath642 for all @xmath643 and @xmath644 .
reordering by this commutation relation defines an isomorphism of associative algebras @xmath645 , where the clifford algebra @xmath646 is generated by @xmath647 with the relations @xmath648 for @xmath649 , and the weyl algebra @xmath650 is generated by @xmath651 with the relations @xmath652 for @xmath653 . as a universal enveloping algebra
the clifford - weyl algebra @xmath654 is filtered , @xmath655 and it inherits from the jordan - heisenberg algebra @xmath241 a canonical @xmath87-grading and a canonical structure of lie superalgebra by the supercommutator
see @xmath44[sect : uea ] for the definitions .
the next statement is a sharpened version of lemma [ lem:2.2 ] .
[ lem : n+n-2 ] @xmath656 \subset \mathfrak{q}_{n + n^\prime - 2}(w)$ ] .
lemma [ lem:2.2 ] asserts the commutation relation @xmath657 \subset \mathsf{u}_{n + n^\prime-1 } ( \mathfrak{g})$ ] for the general case of a lie superalgebra @xmath13 with bracket @xmath658 \subset \mathfrak{g}\,$ ] . for the specific case at hand , where the fundamental bracket @xmath659 \subset \mathbb{k}$ ] has zero component in @xmath132 , the degree @xmath660
is lowered to @xmath661 by the very argument proving that lemma .
it now follows that each of the subspaces @xmath662 for @xmath663 is a lie superalgebra .
since @xmath664 \subset \mathfrak{q}_1(w)$ ] , the quotient space @xmath665 is also a lie superalgebra . by the poincar - birkhoff - witt theorem
, there exists a vector - space isomorphism @xmath666 sending @xmath667 to @xmath668 , the space of super - symmetrized degree - two elements in @xmath669 .
hence @xmath670 has a direct - sum decomposition @xmath671 . if @xmath672 is a homogeneous basis of @xmath132 , every @xmath673 is uniquely expressed as @xmath674 by adding and subtracting terms , @xmath675 one sees that the product @xmath676 for @xmath677 has scalar part @xmath678 = \frac{1}{2}q(w , w^\prime)$ ] with respect to the decomposition @xmath679 .
@xmath680 \subset \mathfrak{s}\,$ ] . from the definition of @xmath681 and @xmath682 \subset \mathbb{k}$ ]
it is clear that @xmath683 \subset \mathbb{k }
\oplus \mathfrak{s}\,$ ] .
the statement to be proved , then , is that @xmath684 $ ] for @xmath685 has zero scalar part . by the linearity of the supercommutator
, it suffices to consider a single term of the sum ( [ eq:2.8 ] ) .
thus we put @xmath686 , and have @xmath687 = [ w w^\prime , b ] = w\ , [ w^\prime , b ] + [ w , b ] \ , w^\prime ( -1)^{|w^\prime| |b| } \;.\ ] ] now we compute the scalar part of the right - hand side . using the jacobi identity for the lie superalgebra @xmath99 we obtain @xmath688_\mathbb{k } = [ w,[w^\prime ,
b ] ] + [ [ w , b],w^\prime ] ( -1)^{|w^\prime| |b| } = [ [ w , w^\prime],b ] \;.\ ] ] the last expression vanishes because @xmath689 \subset \mathbb{k}$ ] lies in the center of @xmath99 . as a subspace of @xmath99 which closes w.r.t .
the supercommutator @xmath690 $ ] , @xmath681 is a lie superalgebra .
now from lemma [ lem : n+n-2 ] and the jacobi identity for @xmath99 , one sees that @xmath691 acts on each of the quotient spaces @xmath692 for @xmath693 by @xmath694\,$ ] .
in particular , @xmath681 acts on @xmath695 by @xmath696 $ ] , which defines a homomorphism of lie superalgebras @xmath697 \;.\ ] ] the mapping @xmath233 is actually into @xmath698 .
indeed , for @xmath677 one has @xmath699,w']+(-1)^{|a||w|}[w,[a , w ' ] ] \;,\end{aligned}\ ] ] and since @xmath700 = 0\,$ ] , this vanishes by the jacobi identity .
[ lem : tau - iso ] the map @xmath701 is an isomorphism of lie superalgebras .
being a homomorphism of lie superalgebras , the linear mapping @xmath233 is an isomorphism of such algebras if it is bijective .
we first show that @xmath233 is injective .
so , let @xmath702 be any element of the kernel of @xmath703 .
the equation @xmath704 means that @xmath705,w^\prime ] = [ \tau(a)w , w^\prime]$ ] vanishes for all @xmath706 . to fathom the consequences of this , let @xmath707 and @xmath708 be two homogeneous bases of @xmath132 so that @xmath709 . using that @xmath673 has a uniquely determined expansion @xmath710 with supersymmetric coefficients @xmath711 , \widetilde{e}_i ] = a_{ij } + ( -1)^{|e_i| |e_j| } a_{j\,i } = 2 a_{ij } \;.\ ] ] thus the condition @xmath705,w^\prime ] = 0 $ ] for all @xmath712 implies @xmath713 .
hence @xmath233 is injective . by the poincar - birkhoff - witt isomorphism @xmath714 the dimensions of the @xmath87-graded vector space @xmath715 are @xmath716 these agree with those of @xmath97 as recorded in corollary [ cor : dimosp ] .
hence our injective linear map @xmath717 is in fact a bijection . by the isomorphism
@xmath233 every representation @xmath84 of @xmath718 induces a representation @xmath719 of @xmath97 .
let us conclude this subsection by writing down an explicit formula for @xmath720 .
to do so , let @xmath672 and @xmath721 be homogeneous bases of @xmath132 with @xmath722 as before . for @xmath264 notice that the coefficients @xmath723 are supersymmetric : @xmath724 where the last equality sign uses @xmath725 .
the inverse map @xmath726 is now expressed as @xmath727 to verify this formula , one calculates the double supercommutator @xmath728 ] $ ] and shows that the result is equal to @xmath729 = q(e_i\ , , x e_j)$ ] , which is precisely what is required from the definition of @xmath233 by @xmath730 = x e_j\,$ ] .
as before , starting from a @xmath87-graded @xmath171-vector space @xmath731 , let the direct sum @xmath732 be equipped with the orthosymplectic form @xmath374 and denote by @xmath99 the clifford - weyl algebra of @xmath132 .
consider now the following tensor product of exterior and symmetric algebras : @xmath733 following r. howe we call it the _ spinor - oscillator module _ of @xmath99 .
notice that @xmath38 can be identified with the graded - commutative subalgebra in @xmath99 which is generated by @xmath734 . as such
, @xmath38 comes with a canonical @xmath87-grading and its space of endomorphisms carries a structure of lie superalgebra , @xmath735 . the algebra @xmath38 now is to become a representation space for @xmath99 .
four operations are needed for this : the operator @xmath736 of exterior multiplication by a linear form @xmath737 ; the operator @xmath738 of alternating contraction with a vector @xmath739 ; the operator @xmath740 of multiplication with a linear function @xmath741 ; and the operator @xmath742 of taking the directional derivative by a vector @xmath743 .
the operators @xmath744 and @xmath745 obey the _
canonical anti - commutation relations _ ( car ) , which is to say that @xmath746 and @xmath747 anti - commute , @xmath748 and @xmath749 do as well , and one has @xmath750 the operators @xmath751 and @xmath752 obey the _ canonical commutation relations _ ( ccr ) , i.e. , @xmath753 and @xmath754 commute , so do @xmath755 and @xmath756 , and one has @xmath757 given all these operations , one defines a linear mapping @xmath758 by @xmath759 with @xmath760 , @xmath761 operating on the first factor of the tensor product @xmath762 , and @xmath763 , @xmath764 on the second factor .
of course the two sets @xmath765 and @xmath766 commute with each other . in terms of @xmath767 ,
the relations car and ccr are succinctly summarized as @xmath768 = q(w , w^\prime)\ , \mathrm{id}_{\mathfrak{a}(v)}\;,\ ] ] where @xmath181 $ ] denotes the usual supercommutator of the lie superalgebra @xmath769 . by the relation ( [ eq:2.9 ] )
the linear map @xmath770 extends to a representation of the jordan - heisenberg lie superalgebra @xmath771 , with the constants of @xmath241 acting as multiples of @xmath772 .
moreover , being a representation of @xmath241 , the map @xmath770 yields a representation of the universal enveloping algebra @xmath773 .
this representation is referred to as the _ spinor - oscillator representation _ of @xmath99 . in the sequel
we will be interested in the @xmath518-representation induced from it by the isomorphism @xmath720 .
there is a natural @xmath386-grading @xmath774 , @xmath775 note that the operators @xmath761 and @xmath764 increase the @xmath386-degree of @xmath38 by one , while the operators @xmath760 and @xmath763 decrease it by one .
note also if @xmath776 is the @xmath6-element introduced in @xmath44[sect : osp - cas ] , then a direct computation using the formula ( [ eq : tau - inv ] ) shows that @xmath777 is an eigenspace of the operator @xmath778 with eigenvalue @xmath779 .
thus @xmath780 is represented on the spinor - oscillator module @xmath38 by the _
degree_. we now specialize to the situation of @xmath529 with @xmath88 a @xmath87-graded vector space as in @xmath44[sect : howe - pairs ] , and we require @xmath781 and @xmath782 to be isomorphic with dimension @xmath783 . recall that @xmath784 are howe dual pairs in @xmath97 which we denote by @xmath785 .
there is a decomposition @xmath786 in both cases .
the notation highlights the fact that the operators in @xmath787 change the degree of elements in @xmath38 by @xmath779 .
note that the cartan subalgebra @xmath433 of diagonal operators in @xmath13 is contained in @xmath427 but @xmath788 .
since the lie algebra @xmath516 is defined on @xmath86 , the @xmath516-action on @xmath38 preserves the degree .
this action exponentiates to an action of the complex lie group @xmath11 on @xmath38 .
[ prop : howeduality ] the subalgebra @xmath37 of @xmath11-invariants in @xmath38 is an irreducible module for @xmath286 . the vacuum @xmath789 is contained in it as a cyclic vector such that @xmath790 this is a restatement of theorems 8 and 9 of @xcite . in the case of @xmath791
it matters that @xmath792 , as the connected lie group @xmath76 has invariants in @xmath38 not contained in @xmath793 .
proposition [ prop : howeduality ] has immediate consequences for the weights of the @xmath13-representation on @xmath37 . using the notation of @xmath44[sect : osp - roots ] ,
let @xmath794 be a standard basis of @xmath433 and @xmath795 the corresponding dual basis .
we now write @xmath796 and @xmath797 ( @xmath798 ) .
[ cor : weights ] the representations of @xmath799 on @xmath800 and @xmath801 on @xmath802 each have highest weight @xmath803 .
every weight of these representations is of the form @xmath804 with @xmath805 .
recall from @xmath44[sect : howe - pairs ] the embedding of @xmath806 and @xmath807 in @xmath97 , and from lemma [ lem : tau - iso ] the isomorphism @xmath808 where @xmath681 is the lie superalgebra of supersymmetrized degree - two elements in @xmath99 . specializing formula ( [ eq : tau - inv ] ) to the case of a cartan algebra generator @xmath809 one gets @xmath810 where @xmath811 is a basis of @xmath86 and @xmath812 the dual basis of @xmath536 .
now let @xmath813 act by the corresponding operator , say @xmath814 , in the spinor - oscillator representation @xmath770 of @xmath815 .
application of that operator to the highest - weight vector @xmath816 yields @xmath817 altogether this means that @xmath818 where @xmath819 . from lemma [ lem : rootsosp ] the roots
@xmath820 corresponding to root spaces @xmath821 are of the form @xmath822 where the indices @xmath823 are subject to restrictions that depend on @xmath13 being @xmath806 or @xmath824 . from @xmath825 one then has @xmath826 for every weight @xmath827 of the @xmath13-representation on @xmath37 .
the restriction @xmath828 results from @xmath829 being isomorphic to @xmath830 and the vanishing of @xmath831 for @xmath832 .
[ cor : degree ] for each of our two cases @xmath833 and @xmath834 the element @xmath835 is represented on @xmath37 by the degree operator . since the @xmath11-action on @xmath38 preserves the degree , the subalgebra @xmath37 is still @xmath386-graded by the same degree . summing the above expressions for @xmath836 over @xmath837 and using car and
ccr to combine terms , we obtain @xmath838 which is in fact the operator for the degree of the @xmath386-graded module @xmath37 .
we here record the systems of simple positive roots that we will use later ( in @xmath44[unicity theorem ] ) . in the case of @xmath628
this will be @xmath839 the corresponding system of positive roots for @xmath840 is @xmath841 in the case of @xmath842 we choose the system of simple positive roots @xmath843 the corresponding positive root system then is @xmath844 in both cases the roots @xmath845 form a system of positive roots for @xmath846 .
we now equip the spinor - oscillator module @xmath38 for @xmath847 with a unitary structure .
the idea is to think of the algebra @xmath38 as a subset of @xmath848 , the holomorphic functions @xmath849 . for such functions
a hermitian scalar product is defined via berezin s notion of superintegration as follows . for present purposes ,
it is imperative that @xmath168 be defined over @xmath850 , i.e. , @xmath851 , and that @xmath168 be re - interpreted as a _ real _ vector space @xmath852 with complex structure @xmath853 .
needless to say , this is done in a manner consistent with the @xmath87-grading , so that @xmath854 and @xmath855 . from @xmath856 and @xmath857
we are given an isomorphism @xmath858 .
this induces a canonical isomorphism @xmath859 , which gives rise to a bundle isomorphism @xmath860 sending @xmath861 , the algebra of real - analytic functions on @xmath862 with values in @xmath863 , to @xmath864 , the complex of real - analytic differential forms on @xmath865 .
fixing some orientation of @xmath865 , the berezin ( super-)integral for the @xmath87-graded vector space @xmath866 is then defined as the composite map @xmath867\mapsto \int_{v_0^\prime}\omega[\phi]\;,\ ] ] whenever the integral over @xmath862 exists .
thus the berezin integral is a two - step process : first the section @xmath868 is converted into a differential form , then the form @xmath869 $ ] is integrated in the usual sense to produce a complex number . of course , by the rules of integration of differential forms
only the top - degree component of @xmath869 $ ] contributes to the integral .
the subspace @xmath870 has played no role so far , but now we use it to decompose the complexification @xmath871 into holomorphic and anti - holomorphic parts : @xmath872 and determine an operation of complex conjugation @xmath873 .
we also fix on @xmath189 a hermitian scalar product ( a.k.a .
unitary structure ) @xmath874 so that @xmath199 .
this scalar product determines a parity - preserving complex anti - linear bijection @xmath875 by @xmath876 . composing @xmath877 with complex conjugation
@xmath873 we get a @xmath4-linear isomorphism @xmath878 , @xmath879 . in this setting
there is a distinguished gaussian section @xmath880 singled out by the conditions @xmath881 to get a close - up view of @xmath882 , let @xmath883 and @xmath884 be orthonormal bases of @xmath172 resp .
@xmath173 , and let @xmath885 and @xmath886 be the corresponding coordinate functions , with complex conjugates @xmath887 and @xmath888 . viewing @xmath889 , @xmath890 as generators of @xmath891 , our section @xmath892 is the standard gaussian @xmath893 we fix the normalization of @xmath894 by the condition @xmath895 = 1 $ ] . a unitary structure on the spinor - oscillator module @xmath38 is now defined as follows .
let complex conjugation @xmath896 be extended to an algebra anti - homomorphism @xmath897 by the convention @xmath898 ( without any sign factors ) .
then , if @xmath899 are any two elements of @xmath38 , we view them as holomorphic maps @xmath900 , multiply @xmath901 with @xmath902 to form @xmath903 , and define their hermitian scalar product by @xmath904\;.\ ] ] let us mention in passing that ( [ eq : hsp ] ) coincides with the unitary structure of @xmath38 used in the hamiltonian formulation of quantum field theories and in the fock space description of many - particle systems composed of fermions and bosons .
the elements @xmath905 for @xmath906 and @xmath907 form an orthonormal set in @xmath38 , which in physics is called the occupation number basis of @xmath38 .
[ lem : h.c .
] for all @xmath908 and @xmath909 the pairs of operators @xmath910 , @xmath911 and @xmath760 , @xmath912 in @xmath913 obey the relations @xmath914 i.e. , they are mutual adjoints with respect to the unitary structure of @xmath915 .
let @xmath916 .
since @xmath917 is anti - holomorphic , we have @xmath918 . by the first
defining property of @xmath894 in ( [ eq : def - gamma ] ) and the fact that @xmath755 is a derivation , @xmath919 and passing to the hermitian scalar product by the berezin integral we obtain @xmath920 = \left\langle \mu(cv)\phi_1,\phi_2\right\rangle_{\mathfrak{a}(v)}\;.\ ] ] by the definition of the @xmath921-operation this means that @xmath922 . in the case of @xmath923
the argument is similar but for a few sign changes .
our starting relation changes to @xmath924 since the operator @xmath748 is an _ anti_-derivation .
if @xmath925 denotes the isomorphism @xmath926 , then @xmath927 and the first term on the right - hand side berezin - integrates to zero because @xmath928 lowers the degree in @xmath929 .
therefore , @xmath930 = \left\langle \varepsilon(cv ) \phi_1 , \phi_2 \right\rangle_{\mathfrak{a}(v)}\;,\ ] ] which is the statement @xmath931 . by the hermitian scalar product ( [ eq : hsp ] ) and the corresponding @xmath932-norm , the spinor - oscillator module @xmath38 is completed to a hilbert space , @xmath933 .
a nice feature here is that , as an immediate consequence of the factors @xmath934 in the orthonormal basis ( [ eq : occ - no ] ) , the @xmath932-condition @xmath935 implies absolute convergence of the power series for @xmath936 .
hence @xmath3 can be viewed as a subspace of @xmath937 : @xmath938 in the important case of isomorphic components @xmath487 , we may regard @xmath3 as the hilbert space of square - integrable holomorphic differential forms on @xmath5 .
note that although @xmath755 and @xmath753 do not exist as operators on the hilbert space @xmath933 , they do extend to linear operators on @xmath939 for all @xmath940 and @xmath941 . in this subsection
we define a real structure for the complex vector space @xmath942 and describe , in particular , the resulting real forms of the ( @xmath87-even components of the ) howe dual partners introduced above . recalling the map @xmath943 ,
@xmath944 , let @xmath945 be the vector space @xmath946 note that @xmath131 can be viewed as the fixed point set @xmath947 of the involution @xmath948 by the orthogonality assumption , @xmath949 where @xmath950 .
the symmetric bilinear form @xmath222 on @xmath951 restricts to a euclidean structure @xmath952 whereas the alternating form @xmath219 on @xmath953 induces a real - valued symplectic form @xmath954 please be warned that , since @xmath955 fails to be real - valued on @xmath956 , the intersection @xmath957 is _ not _ a real form of the complex lie superalgebra @xmath97 .
the connected classical real lie groups associated to the bilinear forms @xmath222 and @xmath958 are @xmath959 they have lie algebras denoted by @xmath135 and @xmath960 . by construction
we have @xmath961 , and this in fact is a real form of the complex lie algebra @xmath962 .
the elements of @xmath963 are mapped via @xmath720 and the spinor - oscillator representation to anti - hermitian operators in @xmath964 .
let @xmath965 .
we know from lemma [ lem : tau - iso ] that @xmath966 is a super - symmetrized element of degree two in the clifford - weyl algebra @xmath99 . to see the explicit form of such an element , recall the definition @xmath967\,$ ] . since @xmath968 , and @xmath219 restricts to @xmath969 , the fundamental bracket @xmath181 : \
, w_\mathbb{r } \times w_\mathbb{r } \to \mathbb{c}$ ] given by @xmath970 = q(w , w^\prime)$ ] is real - valued on @xmath971 but imaginary - valued on @xmath972 .
therefore , @xmath973 the proposed statement @xmath974 now follows under the assumption that the spinor - oscillator representation maps every @xmath975 to a self - adjoint operator in @xmath976 .
but every element @xmath977 is of the form @xmath978 and this maps to the operator @xmath979 , which is self - adjoint by lemma [ lem : h.c . ] . given the real structure @xmath131 of @xmath132 , we now ask how @xmath980 intersects with the howe pairs @xmath629 and @xmath981 embedded in @xmath97 . by the observation that @xmath374 restricted to @xmath131 is not real - valued , @xmath982 fails to be a real form of the complex lie superalgebra @xmath628 , and the same goes for @xmath983 . nevertheless , it is still true that the even components of these intersections are real forms of the complex lie algebras @xmath984 and @xmath985 .
the real forms of interest are best understood by expressing them via blocks with respect to the decomposition @xmath260 . since @xmath947 ,
the complex linear endomorphisms of @xmath132 stabilizing @xmath131 are given by @xmath986 writing @xmath358 in block - decomposed form @xmath987 where @xmath988 , @xmath989 , @xmath990 , and @xmath991 , the condition @xmath992 becomes @xmath993 the bar here is a short - hand notation for the complex anti - linear maps @xmath994 when expressed with respect to compatible bases of @xmath168 and @xmath995 , these maps are just the standard operation of taking the complex conjugate of the matrices of @xmath996 and @xmath367 . now , to get an understanding of the intersections @xmath997 and @xmath998 , recall the relation @xmath999 for @xmath1000 and the fact that the action of the complex lie algebras @xmath1001 and @xmath1002 on @xmath132 stabilizes the decomposition @xmath375 , with the implication that @xmath1003 in both cases . combining @xmath1004 with @xmath1005
one gets the anti - hermitian property @xmath1006 , which means that @xmath1007 and @xmath1008 are compact real forms of @xmath1009 and @xmath1010 . turning to the howe dual partners of @xmath1009 and @xmath1010 , recall from @xmath44[sect : howe - pairs ] the isomorphism @xmath1011 and arrange for it to be an isometry , @xmath1012 , of the unitary structures of @xmath86 and @xmath536 .
recall also the embedding of the two lie superalgebras @xmath628 and @xmath1013 into @xmath1014 by @xmath1015 here the notation still means the same , i.e. , @xmath1016 , @xmath1017 , and so on .
let a real structure @xmath1018 of @xmath1019 be defined in the same way as the real structure @xmath1020 of @xmath375 .
[ prop:2.4 ] @xmath1021 .
the intersection is computed by transferring the conditions @xmath1005 and @xmath1022 to the level of @xmath1023 .
of course @xmath1005 just reduces to the corresponding condition @xmath1024 . because the isometry @xmath1011 in the present case is symmetric one has @xmath1025 , so the condition @xmath1026 transfers to @xmath1027 . for the same reason , the parity of the maps @xmath1028 is identical to that of @xmath1029 , i.e. , @xmath1030 is symmetric , @xmath1031 is skew , and similar for @xmath1032 .
hence , computing the intersection @xmath1033 amounts to the same as computing @xmath1034 , and the statement follows from our previous discussion of the latter case . in the case of the howe pair @xmath631
the isometry @xmath1035 is skew , so that @xmath1036 . at the same time , the parity of @xmath1028 is reversed as compared to @xmath344 : now the map @xmath1030 is skew and @xmath1031 is symmetric ( and similar for @xmath1037 .
therefore , @xmath1038 which is a compact real form @xmath1039 of @xmath1040 ; and @xmath1041 which is a non - compact real form of @xmath1042 known as @xmath1043 .
let us summarize this result .
[ prop:2.5 ] @xmath1044 .
as before , we identify the complex lie superalgebra @xmath1045 with the space of super - symmetrized degree - two elements in @xmath670 , so that @xmath1046 the adjoint representation of @xmath13 on @xmath99 restricts to the lie algebra representation of @xmath1047 on @xmath1048 which is just the direct sum of the fundamental representations of @xmath307 and @xmath134 .
these are integrated by the fundamental representations of the complex lie groups @xmath1049 and @xmath312 respectively . since the clifford - weyl algebra @xmath99 is an associative algebra , one can ask if , given @xmath1050 , the exponential series @xmath1051 makes sense .
the existence of a one - parameter group @xmath1052 for @xmath1053 would of course imply that @xmath1054 now @xmath1055 .
since the clifford algebra @xmath646 is finite - dimensional , the series @xmath1051 for @xmath1056 does make immediate sense . in this way one
is able to exponentiate the lie algebra @xmath307 in @xmath646 .
the associated complex lie group , which is then embedded in @xmath646 , is @xmath1057 .
this is a 2:1 cover of the complex orthogonal group @xmath1049 .
its conjugation representation on @xmath309 as in ( [ conjugation ] ) realizes the covering map as a homomorphism @xmath1058 .
viewing the other summand @xmath134 of @xmath251 as being in the infinite - dimensional weyl algebra @xmath650 , it is definitely not possibly to exponentiate it in such a naive way .
this is in particular due to the fact that for most @xmath1059 the formal series @xmath1051 is not contained in any space @xmath1060 of the filtration of @xmath650 . as a first step toward remedying this situation , we consider @xmath99 as a space of densely defined operators on the completion @xmath3 ( cf .
@xmath44[sect:2.6.3 ] ) of the spinor - oscillator module @xmath38 . since all difficulties are on the @xmath313 side , for the remainder of this chapter we simplify the notation by letting @xmath1061 and discussing only the oscillator representation of @xmath1062 .
recall that this representation on @xmath38 is defined by multiplication @xmath1063 for @xmath1064 and the directional derivative @xmath755 for @xmath1065 .
for @xmath1066 there is at least no formal obstruction to the exponential series of @xmath1067 existing in @xmath1068 .
however , direct inspection shows that convergence can not be expected unless some restrictions are imposed on @xmath1069 .
this is done by introducing a notion of unitarity and an associated semigroup of contraction operators .
here we introduce the basic semigroup in the complex symplectic group .
various structures are lifted to its canonical 2:1 covering .
actions of the real symplectic and metaplectic groups are discussed along with the role played by the cone of elliptic elements .
letting @xmath874 be the unitary structure on @xmath168 which was fixed in the previous chapter , we recall the complex anti - linear bijection @xmath875 , @xmath1070 .
there is an induced map @xmath1071 on @xmath1072 by @xmath1073 .
as before , we put @xmath1074 . since we have restricted our attention to the symplectic side ,
the vector spaces @xmath132 and @xmath131 are now equipped with the standard complex symplectic structure @xmath219 and real symplectic form @xmath1075 respectively . from here
on in this chapter we abbreviate the notation by writing @xmath1076 and @xmath1077 .
let an anti - unitary involution @xmath1078 be defined by @xmath1079 .
its fixed point group @xmath1080 is the real form @xmath1081 ) of main interest .
we here denote it by @xmath1082 and let @xmath1083 stand for its lie algebra . given @xmath219 and @xmath1084 ,
consider the mixed - signature hermitian structure @xmath1085 which we denote by @xmath1086 , with subscript @xmath1087 to distinguish it from the canonical hermitian structure of @xmath132 given by @xmath1088 .
the relation between the two is @xmath1089 note also the relation @xmath1090 now observe that the real form @xmath1082 is the subgroup of @xmath1091-isometries in @xmath1092 @xmath1093 then define a semigroup @xmath1094 in @xmath1095 by @xmath1096 note that the operation @xmath1097 of hermitian conjugation with respect to @xmath1098 stabilizes @xmath1095 and that @xmath1082 is defined by the condition @xmath1099 . the semigroup @xmath1094 is defined by @xmath1100 , or equivalently , @xmath1101 is positive definite .
we will see later that @xmath1094 ( or , rather , a @xmath1102 cover thereof ) acts by contraction operators on the hilbert space @xmath3 .
it is immediate that @xmath1094 is an open semigroup in @xmath1095 with @xmath1082 on its boundary .
furthermore , @xmath1094 is stabilized by the action of @xmath1103 by @xmath1104 .
the map @xmath1105 , @xmath1106 , will play an important role in our considerations .
it is invariant under the @xmath1082-action by right multiplication , @xmath1107 , and is equivariant with respect to the action defined by left multiplication on its domain of definition and conjugation on its image space , @xmath1108 .
direct calculation shows that in fact the @xmath1109-fibers are exactly the orbits of the @xmath1082-action by right multiplication .
observe that if @xmath1110 for @xmath1111 , then @xmath1112 and @xmath1113 .
in particular , if @xmath1114 is a cartan subalgebra of @xmath1115 which is defined over @xmath1116 , then @xmath1117 is just the squaring map @xmath1118 .
we now fix a cartan subalgebra @xmath1114 having the property that @xmath1120 is contained in the unitary maximal compact subgroup defined by @xmath1121 of @xmath1122 .
this means that @xmath1123 acts diagonally on the decomposition @xmath519 and there is a ( unique up to order ) orthogonal decomposition @xmath1124 into one - dimensional subspaces so that if @xmath1125 , then @xmath1123 acts via characters @xmath1126 on the vector spaces @xmath1127 by @xmath1128 .
in other words , we may choose @xmath1129 to be an orthonormal basis of @xmath168 and equip @xmath1130 with the dual basis so that the elements @xmath1131 are in diagonal form : @xmath1132 observe that , conversely , the elements of @xmath1095 that stabilize the decomposition @xmath1133 and act diagonally in the above sense , are exactly the elements of @xmath1123 .
moreover , @xmath1134 is the subgroup of elements @xmath1131 with @xmath1135 for all @xmath108 .
note that the complex symplectic planes @xmath1136 are @xmath219-orthogonal and defined over @xmath850 .
we now wish to analyze @xmath1094 via the map @xmath1137
. however , for a technical reason related to the proof of proposition [ m - slice ] below , we must begin with the opposite map , @xmath1138 .
thus let @xmath1139 and write @xmath1140 .
the toral semigroup @xmath1141 consists of those elements @xmath1142 that act as contractions on @xmath995 , i.e. , @xmath1143 for all @xmath108 .
the restriction @xmath1144 is , as indicated above , the squaring map @xmath1118 ; in particular we have @xmath1145 and the set @xmath1146 is likewise contained in @xmath1147 . in the sequel
, we will often encounter the action of @xmath1082 on @xmath157 and @xmath126 by conjugation .
we therefore denote this action by a special name , @xmath1148 .
[ m - slice ] @xmath1149 . for @xmath1150 one
has @xmath1151 . hence if @xmath1152 with @xmath1153 , then @xmath1154
consequently , @xmath1155 for all @xmath1156
. in particular , @xmath1157 and if @xmath1158 is an @xmath1159-eigenvector with eigenvalue @xmath41 , it follows that @xmath1160 and @xmath1161 .
now we have @xmath1162 and hence @xmath1163 . as a result , if @xmath1158 is an @xmath1159-eigenvector with eigenvalue @xmath41 , then so is @xmath1164 with eigenvalue @xmath1165 . since @xmath1166 ,
the product of @xmath1167 with @xmath1168 is always negative .
if @xmath1169 it follows that @xmath1170 and @xmath1171 ; if @xmath1172 then @xmath1173 and @xmath1174 . in both cases
@xmath1175 .
since @xmath1159 does indeed have at least one eigenvector , we have constructed a complex 2-plane @xmath1176 as the span of the linearly independent vectors @xmath1177 and @xmath1178 . the plane @xmath1176 is defined over @xmath850 and , because @xmath1179 , it is @xmath219-nondegenerate .
its @xmath219-orthogonal complement @xmath1180 is therefore also nondegenerate and defined over @xmath850 .
the transformation @xmath1181 stabilizes the decomposition @xmath1182 . hence , proceeding by induction we obtain an @xmath219-orthogonal decomposition @xmath1183 .
since the @xmath1184 are @xmath1159-invariant symplectic planes defined over @xmath850 , there exists @xmath1185 so that @xmath1186 stabilizes the above @xmath1123-invariant decomposition @xmath1187 . exchanging @xmath1177 with @xmath1164 if necessary
, we may assume that @xmath1188 acts diagonally on @xmath1189 by @xmath1190 with @xmath1191 .
in other words , @xmath1192 . if we let @xmath1193 then we now have the following analog of the @xmath1194-decomposition .
[ left / right transitive ] the semigroup @xmath1094 decomposes as @xmath1195 . in particular , @xmath1094 is connected . by definition , @xmath1196 , and from proposition [ m - slice ] one has @xmath1197 .
now the map @xmath1198 , @xmath1199 is surjective . therefore @xmath1200 , which is to say that each point in the fiber of @xmath1201 over @xmath1202 lies in the orbit of @xmath1203 generated by left multiplication with @xmath1119 .
on the other hand , by the property @xmath1204 of @xmath1082-equivariance we have @xmath1205 and hence @xmath1206 . because @xmath1082 and @xmath157 are connected , so is @xmath1207 .
it is clear that @xmath1208 .
furthermore , since both @xmath157 and @xmath1082 are invariant under the operation of hermitian conjugation @xmath1209 and under the involution @xmath1210 , we have the following consequences .
[ cor:3.2 ] @xmath1094 is invariant under @xmath1211 and also under @xmath1210 . in particular , @xmath1094 is stabilized by the map @xmath1212 . letting @xmath1213 one has @xmath1214 and hence @xmath1215 .
the stability of @xmath1094 under @xmath1216 was not immediate from our definition of @xmath1094 , which is why we have been working from the viewpoint of @xmath1217 so far .
now that we have it , we may regard @xmath1094 as the total space of an @xmath1082-principal bundle @xmath1218 .
we are going to see in corollary [ bijective ] that this principal bundle is trivial .
next observe that , since @xmath1219 for @xmath1220 , the maps @xmath1221 and @xmath1222 coincide and are just the square @xmath1223 . thus the claim that the elements of @xmath126 have a unique square root in @xmath126 can be formulated as follows . [ squaring ]
the squaring map @xmath1224 is bijective .
recall from proposition [ m - slice ] that every @xmath1225 is diagonalizable in the sense that @xmath1226 .
since @xmath1227 is surjective , the surjectivity of @xmath1228 is immediate .
for the injectivity of @xmath1109 we note that the @xmath1159-eigenspace with eigenvalue @xmath41 is contained in the @xmath1229-eigenspace with eigenvalue @xmath1230 .
the result then follows from the fact that all eigenvalues of @xmath1159 are positive real numbers .
[ bijective ] each of the two @xmath1082-equivariant maps @xmath1231 defined by @xmath1232 and by @xmath1233 , is a bijection .
consider the map @xmath1234 .
surjectivity is evident from corollary [ left / right transitive ] and @xmath1235 . for the injectivity it suffices to prove that if @xmath1236 and @xmath1237 with @xmath1238 , then @xmath1239 . but
this follows directly from @xmath1240 and the fact that @xmath1241 is the bijective squaring map .
the proof for the map @xmath1233 is similar , with @xmath1109 replacing @xmath1201 .
let us look more carefully at @xmath126 as a geometric object .
first , as we have seen , the elements @xmath1159 of @xmath126 satisfy the condition @xmath1242 .
we regard @xmath1243 , @xmath1244 , as an anti - holomorphic involution and reformulate this condition as @xmath1245 . in the present section
we are going to show that @xmath126 is a closed , connected , real - analytic submanifold of @xmath1094 which locally agrees with @xmath1246 .
this implies in particular that @xmath126 is totally real in @xmath1094 with @xmath1247 .
we will also show that the exponential map identifies @xmath126 with a precisely defined open cone in @xmath1248 .
we begin with the following statement .
[ lem : m - closed ] the image @xmath126 of @xmath1109 is closed as a subset of @xmath1094 .
let @xmath1249 . by the definition of @xmath126
, we still have @xmath1250 . if @xmath1251 is any sequence from @xmath126 with @xmath1252 , then @xmath1253 . but
@xmath1159 has a unique square root @xmath1254 and @xmath1255 .
hence @xmath1256 .
@xmath126 of course fails to be closed as a subset of @xmath1257 .
for example , @xmath1258 is in the closure of @xmath1259 but is not in @xmath1147 .
[ lem : max - rank ] the exponential map @xmath1260 has maximal rank along @xmath1261 .
we are going to use the fact that the squaring map @xmath1262 , @xmath1263 , is a local diffeomorphism of @xmath1095 at any point @xmath1264 . to show this
, we compute the differential of @xmath222 at @xmath1188 and obtain @xmath1265 where @xmath1266 denotes the differential of the left translation @xmath1267 , @xmath1268 .
the middle map @xmath1269 is regular because all of the eigenvalues of @xmath1202 are positive real numbers .
since @xmath1270 and @xmath1271 are isomorphisms , it follows that @xmath1272 is an isomorphism .
turning to the proof of the lemma , given @xmath1273 we now choose @xmath1274 so that @xmath1275 is in a neighborhood of @xmath1276 where @xmath1277 is a diffeomorphism .
it follows that for @xmath327 a sufficiently small neighborhood of @xmath1278 , the exponential map expressed as @xmath1279 is a diffeomorphism of @xmath327 onto its image .
now recall @xmath1280 and consider the cone @xmath1281 it follows from the equivariance of @xmath1277 , i.e. , @xmath1282 , that @xmath1283 is surjective .
furthermore , @xmath1284 is injective and for every @xmath1285 the isotropy groups of the @xmath1082-actions at @xmath1286 and @xmath1287 are the same .
therefore @xmath1288 is also injective .
in fact , much stronger regularity holds . for the statement of this result we recall the anti - holomorphic involution @xmath1289 defined by @xmath1244 and let @xmath1290 denote the connected component of @xmath1246 that contains @xmath126 .
the image @xmath1291 of @xmath1292 is the closed , connected , totally real submanifold @xmath1290 , which is half - dimensional in the sense that @xmath1293 .
the set @xmath1294 , which is an open positive cone in @xmath1295 , is in bijection with @xmath126 by the real - analytic diffeomorphism @xmath1288 .
lemma [ lem : max - rank ] implies that @xmath1296 possesses an open neighborhood @xmath327 in @xmath1297 so that @xmath1298 is everywhere of maximal rank .
because @xmath157 lies in @xmath1299 and @xmath1094 is open in @xmath1257 , by choosing @xmath327 small enough we may assume that @xmath1300 and therefore that @xmath1301 .
since @xmath1277 is a local diffeomorphism on @xmath327 , we may also assume that @xmath1302 , and it then follows that @xmath1303 .
in particular , this shows that @xmath1304 is open in @xmath1305 . in summary , @xmath1306 by the equivariance of @xmath1307 , we also know that it is everywhere of maximal rank on @xmath1308 .
now @xmath1309 is an anti - holomorphic involution .
therefore , @xmath1310 is a totally real , half - dimensional closed submanifold of @xmath1094 , and since @xmath1304 is open in @xmath1295 , we also know that @xmath1311 .
the maximal rank property of @xmath1277 then implies that @xmath1312 is open in @xmath1313 . in lemma
[ lem : m - closed ] it was shown that @xmath126 is closed in @xmath1314 .
thus it is open and closed in the connected manifold @xmath1290 , and consequently @xmath1315 is a local diffeomorphism of manifolds . since we already know that @xmath1316 is bijective , the desired result follows .
[ semigroup product structure ] the two identifications @xmath1317 defined by @xmath1232 and @xmath1318 are real - analytic diffeomorphisms .
the fundamental group of @xmath1094 is isomorphic to @xmath1319 .
the first statement is proved by explicitly constructing a smooth inverse to each of the two maps . for this
let @xmath1320 and note that @xmath1321 and @xmath1322 .
since the square root is a smooth map on @xmath1147 , a smooth inverse in the two cases is defined by @xmath1323 the second statement follows from @xmath1324 by @xmath1277 and the fact that @xmath1082 is a product of a cell and a maximal compact subgroup @xmath528 .
we choose @xmath11 to be the unitary group @xmath1325 acting diagonally on @xmath519 and recall that @xmath1326 .
recall that we are concerned with the lie algebra representation of @xmath1327 which is defined by the identification of @xmath1115 with the set of symmetrized elements of degree two in the weyl algebra @xmath1062 and the representation of @xmath1062 on @xmath38 . in @xmath44[metaplectic rep ]
we construct the oscillator representation of the metaplectic group @xmath1328 , a 2:1 cover of @xmath1119 , which integrates this lie algebra representation . observe that since @xmath1329 and @xmath386 has only one subgroup of index two , there is a unique such covering @xmath1330 .
the method of construction @xcite we use first yields a representation of the 2:1 covering space @xmath1331 of @xmath1094 and then realizes the oscillator representation of @xmath1332 by taking limits that correspond to going to @xmath1082 in the boundary of @xmath1094 .
this representation of the _ oscillator semigroup _
@xmath1333 is for our purposes at least as important as the representation of @xmath1328 .
the goal of the present section is to lift all essential structures on @xmath1094 to @xmath1333 .
we begin by recalling a few basic facts about covering spaces .
if @xmath33 is a connected lie group , its universal covering space @xmath327 carries a canonical group structure : an element @xmath553 in the fiber over @xmath1334 is a homotopy class @xmath1335 $ ] of paths @xmath1336 \to g$ ] connecting @xmath1337 with the neutral element @xmath1338 ; and an associative product @xmath1339 , @xmath1340 , is defined by taking @xmath1341 to be the unique homotopy class which is given by pointwise multiplication of any two paths representing the homotopy classes @xmath1342 .
the fundamental group @xmath1343 acts on @xmath327 by monodromy , i.e. , if @xmath1344 = u \in u$ ] and @xmath1345 = \gamma \in \pi_1(g)$ ] , then one sets @xmath1346 \in u$ ] where @xmath1347 is the path from @xmath1337 to @xmath1348 which is obtained by composing the path @xmath1349 with the loop @xmath877 based at @xmath1350 .
this @xmath1351-action satisfies the compatibility condition @xmath1352 and in that sense is central .
the situation for our semigroup @xmath1314 is analogous except for the minor complication that the distinguished point @xmath1353 does not lie in @xmath1094 but lies in the closure of @xmath1094 .
hence , by the same principles , the universal cover @xmath327 of @xmath1094 comes with a product operation and there is a central action of @xmath1354 on @xmath327 .
moreover , the product @xmath1339 still is associative . to see this , first notice
that the subsemigroup @xmath1355 is simply connected and as such is canonically embedded in @xmath327 .
then for @xmath1356 observe that @xmath1357 where @xmath1358 could theoretically depend on the @xmath1359 .
however , any such dependence has to be continuous and the fundamental group is discrete , so in fact @xmath894 is independent of the @xmath1360 and , since @xmath894 is the identity when the @xmath1360 are in @xmath157 ( lifted to @xmath327 ) , the associativity follows .
let now @xmath1361 denote the subgroup of index two in @xmath1362 and consider @xmath1363 , which is our object of interest .
since the @xmath104-action on @xmath327 is central , i.e. , @xmath1364 for all @xmath1365 and @xmath1366 , the product @xmath1367 descends to a product @xmath1368 .
thus @xmath1369 is a semigroup , and the situation at hand is summarized by the following statement .
the 2:1 covering @xmath1370 , @xmath1371\ , \gamma \mapsto h\,$ ] , is a homomorphism of semigroups .
recall that we have two 2:1 coverings : a homomorphism of groups @xmath1372 , along with a homomorphism of semigroups @xmath1373 .
now , a pair of elements @xmath1374 determines a transformation @xmath1375 of @xmath1314 , and by the homotopy lifting property of covering maps a corresponding action of @xmath1376 on @xmath1377 is obtained as follows .
consider the canonical mapping @xmath1378 given by @xmath703 . by the real - analytic diffeomorphism @xmath1379 , @xmath1380
, this map can be regarded as a 2:1 covering of @xmath1094 , and since any two 2:1 coverings are isomorphic we get an identification of the covering space @xmath1333 with @xmath1381 .
moreover , the action of the group @xmath1332 on itself by left translations induces on @xmath1382 an @xmath1332-action which , by construction , satisfies the relation @xmath1383 this can be viewed as a statement of @xmath1332-equivariance of the covering map @xmath1384 .
now , we have another real - analytic diffeomorphism @xmath1385 by @xmath1386 , which transfers left translation in @xmath1082 to right multiplication on @xmath1094 , and by using it we can repeat the above construction .
the result is another identification @xmath1387 and another @xmath1332-action on @xmath1333 .
altogether we then have two actions of @xmath1332 on @xmath1388 .
the essence of the next statement is that they commute .
[ lifting the product action ] there is a real - analytic action @xmath1389 of @xmath1390 on @xmath1388 such that the covering @xmath1391 is @xmath1392-equivariant : @xmath1393 by construction , the stated equivariance property of @xmath1394 holds for each of the two actions of @xmath1332 separately .
it then follows that it holds for all @xmath1395 if the two actions commute .
but by @xmath1396 and @xmath1397 the commutator @xmath1398 acts trivially on @xmath1094 by @xmath1384 , i.e. , @xmath1399 .
therefore @xmath1337 can be regarded as being in the covering group @xmath1400 of the covering @xmath1401 .
since we can connect both @xmath1402 and @xmath1403 to the identity @xmath1404 by a continuous curve , it follows from the discreteness of @xmath87 that @xmath1405 acts trivially on @xmath1333 .
notice that since the submanifold @xmath1406 is simply connected , there exists a canonical lifting of @xmath126 ( which we still denote by @xmath126 ) to the cover @xmath1333 ; this is the unique lifting by which @xmath1145 is embedded as a subsemigroup in @xmath1333 .
proposition [ lifting the product action ] then allows us to write @xmath1407 .
let us now turn to the issue of lifting the various involutions at hand .
as a first remark , we observe that any lie group automorphism @xmath1408 uniquely lifts to a lie group automorphism @xmath1409 of the universal covering group @xmath1410 , and the latter induces an automorphism of the fundamental group @xmath1411 viewed as a subgroup of the center of @xmath1410 . now
@xmath1412 and both elements of this automorphism group stabilize the subgroup @xmath1413 in @xmath1414 .
therefore @xmath1415 induces an automorphism of @xmath1416 . since the operation @xmath1417 canonically lifts from @xmath1082 to @xmath1332 and @xmath1418
is a lie group automorphism of @xmath1119 , it follows that hermitian conjugation @xmath1209 has a natural lift to @xmath1328 .
the same goes for the lie group automorphism @xmath1210 of @xmath1119 .
hermitian conjugation @xmath1211 and the involution @xmath1419 lift to unique maps with the property that they stabilize the lifted manifold @xmath126 . in particular , the basic anti - holomorphic map @xmath1420 , @xmath1421 lifts to an anti - holomorphic map @xmath1422 which is the identity on @xmath126 and @xmath1376-equivariant in that @xmath1423 for all @xmath1424 and @xmath1425 .
recall that the simply connected space @xmath1426 has a canonical lifting ( still denoted by @xmath126 ) to @xmath1377 .
since all of our involutions stabilize @xmath126 as a submanifold of @xmath1094 , they are canonically defined on the lifted manifold @xmath1147 . in particular , the involution @xmath1309 on @xmath126 is the identity map , and therefore so is the lifted involution @xmath1427 .
note furthermore that the involution defined by @xmath1210 is holomorphic on @xmath1094 and that the other two are anti - holomorphic .
now @xmath1333 is connected and the lifted version of @xmath126 is a totally real submanifold of @xmath1333 with @xmath1428 . in such a situation
the identity principle of complex analysis implies that there exists at most one extension ( holomorphic or anti - holomorphic ) of an involution from @xmath126 to @xmath1429 .
therefore , it is enough to prove the existence of extensions .
since @xmath1430 is uniquely representable as @xmath1431 with @xmath1432 and @xmath1225 , the involution @xmath1211 is extended by @xmath1433 .
similarly , @xmath1210 extends by @xmath1434 , and @xmath1435 does so by the composition of the other two .
the equivariance property of @xmath1427 follows from the fact that @xmath1436 on @xmath1332 coincides with the operation of taking the inverse , @xmath1437 .
here we construct the fundamental representation of the semigroup @xmath1333 on the hilbert space @xmath3 , which in the present context we call fock space .
our approach is parallel to that of howe @xcite : the fock space we use is related to the @xmath932-space of howe s work by the bargmann transform @xcite .
( using the language of physics , one would say that howe works with the position wave function while our treatment relies on the phase space wave function . ) in particular , following howe we take advantage of a realization of @xmath1314 as the complement of a certain determinantal variety in the siegel upper half plane .
let us begin with some background information on the cayley transformation , which is defined to be the meromorphic mapping @xmath1438 if @xmath1439 , then from @xmath1440 we have @xmath1441 for all @xmath1442 . by assuming that @xmath1443 is invertible and then replacing @xmath1177 and @xmath1444 by @xmath1445 resp .
@xmath1446 , we see that @xmath1084 maps the complement of the determinantal variety @xmath1447 into @xmath602 .
the inverse of the cayley transformation is given by @xmath1448 reversing the above argument , one shows that if @xmath1449 is invertible and @xmath1450 , then @xmath1451 .
moreover , by the relation @xmath1452 for @xmath1453 , if @xmath1454 is regular , then so is @xmath1455 , and vice versa .
thus if we introduce the sets @xmath1456 the following is immediate .
[ cayley biholo ] the cayley transformation defines a bi - holomorphic map @xmath1457 now we consider the restriction of @xmath1084 to the semigroup @xmath1458 .
letting @xmath921 be the hermitian conjugation of the previous section , denote by @xmath1459 the real part of an operator @xmath1460 and define the associated siegel upper half space @xmath1461 to be the subset of elements @xmath1460 which are symmetric with respect to the canonical symmetric bilinear form @xmath222 on @xmath1462 with @xmath1463 .
notice the relations @xmath1464 and @xmath1465 , from which it is seen that @xmath358 is symmetric if and only if @xmath1466 .
define @xmath1467 , let @xmath1468 and define a slightly modified cayley transformation by @xmath1469 translating proposition [ cayley biholo ] , it follows that @xmath1470 defines a bi - holomorphic map from @xmath1471 onto the set of @xmath222-symmetric operators with @xmath1472 removed .
the modified cayley transformation @xmath1473 given by @xmath1474 restricts to a bi - holomorphic map @xmath1475 this result is an immediate consequence of the following identity . for @xmath1476 let @xmath1477 and define @xmath1478 .
then @xmath1479 in particular , one has the following equivalence : @xmath1480 it is convenient to rewrite @xmath1481 as @xmath1482 using @xmath1477 one directly computes that @xmath1483 the desired identity follows by inserting these relations in the previous equation .
[ rem : twine - inv ] the modified cayley transformation intertwines the anti - holomorphic involution @xmath1484 with the operation of taking the hermitian conjugate @xmath1485 : @xmath1486 since @xmath1487 is obviously stable under hermitian conjugation , this is another proof of the stability of @xmath1094 under the involution @xmath1488 ; cf .
corollary [ cor:3.2 ] .
let us now turn to the main goal of this section .
recall that we have a lie algebra representation of @xmath1115 on @xmath1489 which is defined by its canonical embedding in @xmath1490 .
we shall now construct the corresponding representation of the semigroup @xmath1333 on the fock space @xmath3 .
it will be seen later that the character of this representation on the lifted toral semi - group @xmath157 is @xmath1491 .
this extends to @xmath1492 by the invariance of the character with respect to the conjugation action of @xmath1328 .
since @xmath1333 is connected and @xmath126 is totally real of maximal dimension in @xmath1333 , the identity principle then implies that if a semigroup representation of @xmath1333 can be constructed with a holomorphic character , this character must be given by the square root function @xmath1493 .
there is no difficulty discussing the square root on the simply connected submanifold @xmath126 .
however , in order to make sense of the square root function on the full semigroup , we must lift all considerations to @xmath1333 . for convenience of notation ,
given @xmath1494 we let @xmath1495 , and for @xmath1496 we simply write @xmath1497 where @xmath1498 is the canonical covering map
. then we put @xmath1499 and wish to define @xmath1500 to be the square root of @xmath1501 which agrees with the positive square root on @xmath1502 .
( here we regard @xmath157 as being in @xmath1331 by its canonical lifting as a subsemigroup of @xmath1314 as in the previous section ) .
this is possible because @xmath1503 is naturally defined on @xmath1504 which is itself a 2:1 cover of @xmath1314 .
since up to equivariant equivalence there is only one such covering , namely @xmath1505 , it follows that we may define @xmath1506 on @xmath1331 as desired . for the construction of the oscillator representation
it is useful to observe that @xmath1503 can be extended to a slightly larger space .
this extension is constructed as follows .
regard the complex symplectic group @xmath1095 as the total space of an @xmath1082-principal bundle @xmath1507 , @xmath1508 . recall that the restricted map @xmath1228 is a diffeomorphism , and that @xmath126 contains the neutral element @xmath1509 in its boundary .
we choose a small ball @xmath253 centered at @xmath1510 in the base @xmath1511 , and using the fact that @xmath126 can be identified with a cone in @xmath1512 we observe that @xmath1513 is contractible .
now @xmath1514 is diffeomorphic to a product @xmath1515 and thus comes with a 2:1 covering @xmath1516 defined by @xmath1517 .
the covering space @xmath569 contains @xmath1429 , and is invariant under the @xmath1332-action by right multiplication . by construction
it also contains the metaplectic group @xmath1328 , which covers the group @xmath1082 in @xmath1257 .
recall the definition of the determinant variety @xmath1518 .
let @xmath1519 denote the set of points in @xmath569 which lie over @xmath1520 by the covering @xmath1516 .
[ mp in the boundary ] there is a unique continuous extension of @xmath1506 from @xmath1333 to its closure in @xmath569 so that @xmath1521 agrees with the lift of @xmath1501 from @xmath327 .
the intersection of @xmath1519 with any @xmath1332-orbit in @xmath569 is nowhere dense in that orbit and the restriction of the extended function @xmath1503 to the complement of that intersection is real - analytic . before beginning the proof
, it should be clarified that at the points of the lifted determinant variety , i.e. , where the lifted square root @xmath1503 of the function @xmath1522 has a pole , _ continuity of the extension _ means that the reciprocal of @xmath1503 extends to a continuous function which vanishes on that set .
the intersection of @xmath1523 with any @xmath1082-orbit in @xmath327 is nowhere dense in that orbit ; therefore the same holds for the intersection of @xmath1519 with any @xmath1332-orbit in @xmath569 .
let @xmath1524 be a point of the boundary of @xmath1333 .
choose a local contractible section @xmath1525 of @xmath1526 with @xmath1527 and a neighborhood @xmath1528 of the identity in @xmath1332 so that the map @xmath1529 , @xmath1530 , realizes @xmath1531 as a neighborhood @xmath1532 of @xmath1067 which has empty intersection with @xmath1533 . by construction @xmath1534
is connected and is itself simply connected .
thus the desired unique extension of @xmath1503 exists on @xmath1532 . at @xmath1067 this extension
is simply defined by taking limits of @xmath1503 along arbitrary sequences @xmath1535 from @xmath1333 .
thus the extended function ( still called @xmath1503 ) is well - defined on the closure of @xmath1536 and is real - analytic on the complement of @xmath1537 in every @xmath1332-orbit in that closure .
it extends as a continuous function on the full closure of @xmath1538 by defining it to be identically @xmath1539 on @xmath1540 , i.e. , its reciprocal vanishes identically at these points .
now let us proceed with our main objective of defining the semigroup representation on @xmath1333 .
recall the involution @xmath1541 , @xmath1542 , and its lift @xmath1427 to @xmath1543 .
the following will be of use at several points in the sequel .
[ psi equivariance ] @xmath1544 . by direct calculation , @xmath1545 .
thus , since @xmath1546 , we have either @xmath1547 or @xmath1548 .
the latter is not the case , as @xmath1503 is not purely imaginary on the nonempty set @xmath1549 .
the semigroup representation @xmath1550 will be given by a certain averaging process involving the standard representation of the heisenberg group .
the latter representation is defined as follows . for elements @xmath1551 of the real vector space @xmath956 ,
the operator @xmath1552 is self - adjoint and its exponential @xmath1553 converges and is unitary ( see , e.g. , @xcite ) .
these operators satisfy the relation @xmath1554 where @xmath1555 is the induced real symplectic structure .
if @xmath1556 denotes the adjoint operation in @xmath1557 , it follows from @xmath922 that @xmath1558 thus if @xmath1559 is equipped with the heisenberg group multiplication law , @xmath1560 then @xmath1561 is an irreducible unitary representation of @xmath1562 on @xmath1563 .
it is well known that up to equivariant isomorphisms there is only one such representation . the oscillator representation @xmath1564 of @xmath1377 is now defined by @xmath1565 here @xmath1566 is the euclidean volume element on @xmath131 which we normalize so that @xmath1567 it should be stressed that we often parameterize @xmath1568 by the map @xmath1569 .
notice that by the positivity of @xmath1570 the gaussian function @xmath1571 decreases rapidly , so that all integrals involved in the discussion above and below are easily seen to converge . in particular , since the unitary operator @xmath1572 ( for @xmath975 ) has @xmath932-norm @xmath1573 , it follows for any @xmath1574 that @xmath1575 where the bound @xmath1576 by direct computation of the integral is a finite number : @xmath1577 thus @xmath1578 is a bounded linear operator on @xmath3 . in proposition [ operator bound ]
we will establish the uniform bound @xmath1579 for all @xmath1580 .
it is also clear that the operator @xmath1578 depends continuously on @xmath1581 .
the main point now is to prove the semigroup multiplication rule @xmath1582 . for this
we apply the heisenberg multiplication formula ( [ heisenberg multiplication ] ) to the inner integral of @xmath1583 to see that @xmath1582 is equivalent to the multiplication rule @xmath1584 where the right - hand side means the twisted convolution @xmath1585 for the proof of the formula @xmath1586 , we will need to know that @xmath1503 transforms as @xmath1587 this transformation behavior , in turn , follows directly from the expression for the semigroup multiplication rule transferred to the upper half - space @xmath1588 ; we record this expression in the following statement and refer to @xcite , p. 78
, for the calculation .
[ prop : semigroup - law ] identifying @xmath1094 with @xmath1487 by the modified cayley transformation and denoting by @xmath1589 the semigroup multiplication on @xmath1590 , one has @xmath1591 given proposition [ prop : semigroup - law ] , to prove the transformation rule ( [ transformation rule ] ) just set @xmath1592 and @xmath1593 and note that , since the semigroup multiplication law for @xmath1333 by definition yields @xmath1594 , the first expression in ( [ transferred multiplication ] ) implies @xmath1595 where @xmath1596 as above .
the transformation rule for @xmath1503 follows by taking the square root of ( [ squared transformation law ] ) . as usual , the sign of the square root is determined by taking the positive sign at points of the lift of @xmath157 in @xmath1333 .
now we come to the main point . the twisted convolution for @xmath1597 satisfies @xmath1598 .
observe that the phase factor for @xmath1599 can be reorganized as @xmath1600 inserting the definitions of @xmath1601 and @xmath1602 in @xmath1603 we then have @xmath1604 since @xmath1605 , the integrand is a rapidly decreasing function of @xmath1606 and convergence of the integral over the domain @xmath131 is guaranteed .
we now wish to explicitly compute the integral by completing the square and shifting variables . for this
it is a useful preparation to write @xmath1607 and similarly for the other terms .
we then holomorphically extend the right - hand side to @xmath1608 in @xmath132 , and by making a shift of integration variables @xmath1609 we bring the convolution integral into the form @xmath1610 where @xmath1611 is the semigroup multiplication on @xmath1487 . using @xmath1612 for @xmath977 and the defining relation @xmath1613 , we see that the first factor on the right - hand side is @xmath1614 . by the transformation rule ( [ transformation rule ] ) the integral
is evaluated as @xmath1615 and multiplying factors it follows that @xmath1616 which is the desired semigroup property .
the mapping @xmath1617 defined by @xmath1618 is a representation of the semigroup @xmath1333 .
we conclude this section by deriving a formula for the adjoint . the adjoint of @xmath1578 is computed as @xmath1619 . in particular , @xmath1620 .
we recall the relations @xmath1621 from proposition [ psi equivariance ] and @xmath1622 from remark [ rem : twine - inv ] .
since @xmath1623 , it follows that @xmath1624 the desired formula , @xmath1625 , now results from this equation and the identities @xmath1626 and @xmath1627 . with this in hand ,
the second statement @xmath1628 is a consequence of the semigroup property .
here we compute the effect of conjugating ( in the semigroup sense ) operators of the form @xmath1629 , @xmath1630 , with operators @xmath1578 coming from the semigroup .
this is an immediate consequence of an analogous result for the operators @xmath1631 . for this
we first allow @xmath1572 to be defined for @xmath1632 by @xmath1633 these operators are no longer defined on fock space , but are defined on @xmath1634 .
they satisfy @xmath1635 note that for @xmath1496 and @xmath1630 the operators @xmath1636 and @xmath1637 are bounded on @xmath3 .
thus we interpret the following as a statement about operators on that space .
[ heisenberg conjugation ] for @xmath1630 and @xmath1496 one has the relation @xmath1638 for convenience of notation we write @xmath1639 thus @xmath1640 now let @xmath1641 and change variables by the translation @xmath1642 .
using the definition @xmath1643 and the relation @xmath1644 for all @xmath1645 , one simplifies the exponent to obtain @xmath1646 reading ( [ heisenberg transformation rule ] ) backwards one sees that this expression equals @xmath1647 .
the basic conjugation rule now follows immediately .
[ basic conjugation formula ] for every @xmath1496 and @xmath1630 it follows that @xmath1648 apply proposition [ heisenberg conjugation ] for @xmath1177 replaced by @xmath1649 and differentiate both sides of the resulting formula at @xmath1650
. numerous properties of @xmath1651 are derived from a precise description of the spectral decomposition of @xmath1578 for @xmath1652 . since every orbit of @xmath1082 acting by conjugation on @xmath126
has nonempty intersection with @xmath1502 , it is important to understand this decomposition when @xmath1653 . for this
we begin with the case where @xmath168 is one - dimensional .
[ 1-d eigenvalues ] suppose that @xmath168 is one - dimensional and that the @xmath157-action on @xmath1654 is given by @xmath1655 where @xmath1656 .
if @xmath1501 is a basis vector of @xmath995 then the monomials @xmath1657 form a basis of @xmath1563 and one has
@xmath1658 first note that if @xmath1659 , then @xmath1660 thus the gaussian function @xmath1661 in the present case is @xmath1662 to apply the operator @xmath1663 to @xmath1664 we use the description @xmath1665 decomposing @xmath1663 in this way is not allowed on the fock space , but is allowed if we regard @xmath1663 as an operator on the full space @xmath1666 of holomorphic functions .
the calculations are now carried out on this larger space . recall that @xmath1667 . from this
we obtain the explicit expression @xmath1668 our goal is to compute @xmath1669 where @xmath1670 corresponds to @xmath1671 by the isomorphism @xmath1672 . expanding the exponential @xmath1673 and using @xmath1674 , the integral
@xmath115 is a sum of gaussian expected values of terms of the form @xmath1675 .
the only terms which survive are those with @xmath1676 .
thus @xmath1677 now @xmath1678 , and @xmath1679 , since we are to take the positive square root at points @xmath1653 .
hence , @xmath1680 as claimed . note that as @xmath1681 goes to the unit element ( or , equivalently , @xmath1682 ) , the expression @xmath1683 converges to @xmath1664 in the strong sense for all @xmath1684 .
now let @xmath168 be of arbitrary dimension and assume that @xmath1681 is diagonalized on @xmath519 in a basis @xmath1685 with eigenvalues @xmath1686 respectively .
since @xmath1687 , we have @xmath1688 for all @xmath1689 . for @xmath1690 and @xmath1691
we employ the standard multi - index notation @xmath1692 and @xmath1693 . in this case
the multi - dimensional integrals split up into products of one - dimensional integrals .
thus , the following is an immediate consequence of the above .
[ eigenvalues in t_+ ] let @xmath1681 be diagonal in a basis @xmath672 of @xmath168 with eigenvalues @xmath1694 ( @xmath1695 ) . if @xmath1664 is a monomial @xmath1696 , then @xmath1697 .
one would expect the same result for the spectrum to hold for every conjugate @xmath1698 , and this expectation is indeed borne out .
however , in the approach we are going to take here , we first need the existence and basic properties of the oscillator representation of the metaplectic group .
the following is a first step in this direction .
[ operator bound ] the operator norm function @xmath1699 , @xmath1700 is bounded by a continuous @xmath1332-independent function @xmath1701 . for any @xmath1496 we already have the bound @xmath1702 where @xmath1576 was computed in ( [ eq : schranke ] ) .
that function @xmath1576 clearly is invariant under conjugation @xmath1703 by @xmath1704 . evaluating it for the case of an element @xmath1705 with eigenvalues
@xmath1694 one obtains @xmath1706 the inequality @xmath1707 now follows from the fact that @xmath1708 for all @xmath1709 . since @xmath1625 and @xmath1710 , we infer the following estimates . [
general operator bound ] for all @xmath1202 and @xmath1432 one has @xmath1711 and @xmath1712 .
recall that we have realized the metaplectic group @xmath1332 in the boundary of the oscillator semigroup @xmath1333 and that @xmath1333 contains the lifted manifold @xmath157 in such a way that the neutral element @xmath1713 is in its boundary .
here we show that for @xmath1681 and @xmath1432 the limit @xmath1714 is a well - defined unitary operator @xmath1715 on fock space and @xmath1716 is a unitary representation .
the basic properties of this _ oscillator representation _ are then used to derive important facts about the semigroup representation @xmath1717 .
convergence will eventually be discussed in the so - called _ bounded strong * topology _ ( see @xcite , p. 71 ) . for the moment , however , we shall work with the slightly weaker notion of bounded strong topology where one only requires uniform boundedness and pointwise convergence of the operators themselves ( with no mention made of their adjoints ) .
note that since @xmath1718 by corollary [ general operator bound ] , we need only prove the convergence of @xmath1719 on a dense set of functions @xmath1720 .
let us begin with @xmath1721 .
if a sequence @xmath1722 converges to @xmath1723 , then the sequence @xmath1724 converges in the bounded strong topology to the identity operator on fock space .
if @xmath1501 is any @xmath157-eigenfunction , the sequence @xmath1725 converges to @xmath1501 by the explicit description of the spectrum given in corollary [ eigenvalues in t_+ ] .
the statement then follows because the subspace generated by these functions is dense . using this lemma along with the semigroup property
, we now show that the limiting operators exist and are well - defined .
[ well defined ] if @xmath1722 converges to @xmath1726 , then for every @xmath1432 the sequence of operators @xmath1727 converges pointwise , i.e. , @xmath1728 for all @xmath1501 in @xmath3 .
the limiting operator @xmath1715 is independent of the sequence @xmath1729 . for any @xmath1730 there
exists some @xmath1731 sufficiently near the identity so that @xmath1732 and @xmath1733 are still in @xmath1502 . by the semigroup property @xmath1734 we then have @xmath1735 letting @xmath1736 it follows from corollary [ general operator bound ] that @xmath1737 thus the cauchy property of @xmath1738 is passed on to @xmath1739 and
therefore the sequence @xmath1739 converges in the hilbert space @xmath3 .
let @xmath1740 . to show that the limit is well - defined , pick from @xmath157 another sequence @xmath1741 ,
let @xmath1742 , and notice that @xmath1743 is no bigger than @xmath1744 using the same reasoning as above , the middle term is estimated as @xmath1745 in the limit @xmath1746 this yields the desired result @xmath1747 .
since @xmath1748 the sequence @xmath1727 converges to @xmath1749 in the bounded strong topology .
such convergence preserves the product of operators , which is to say that if @xmath1750 and @xmath1751 , then @xmath1752 .
indeed , @xmath1753 and convergence follows from @xmath1754 and @xmath1755 , @xmath1751 .
note in particular that if @xmath1756 and @xmath1757 then @xmath1758 .
the bounded strong@xmath1759 topology also requires pointwise convergence of the sequence of adjoint operators .
therefore we must also consider sequences of the form @xmath1760 .
for this ( see the proof of theorem [ thm : weil - repn ] below ) we will use the following fact .
let @xmath1761 and @xmath1762 be sequences of bounded operators and let @xmath1763 . if @xmath1764 and @xmath1765 converge pointwise with @xmath1751 and the sequence @xmath1761 is uniformly bounded , then @xmath1766 converges pointwise on the image of @xmath1767 .
if @xmath1768 then @xmath1769 where @xmath1770 for some hilbert vector @xmath1771 .
write @xmath1772 and use the uniform boundedness of @xmath1766 to show that @xmath1773 converges . applying this with @xmath1774 , @xmath1775 and @xmath1776
, we have the following statement about convergence along the conjugate @xmath1698 .
for @xmath1432 and @xmath1535 any sequence in @xmath157 with @xmath1777 , it follows that @xmath1778 converges pointwise to @xmath1779 .
next , if we take three sequences in @xmath157 and write @xmath1780 then it follows that the sequence @xmath1781 converges to an operator @xmath1782 on the image of @xmath1783 with @xmath1784 .
in particular , the operator @xmath1715 is injective for all @xmath1785 .
finally , we define @xmath1786 by @xmath1787 and write @xmath1788 . taking the limit of both sides of this equation entails that @xmath1789 and since @xmath1715 is injective , this now allows us to reach the main goal of this section .
[ thm : weil - repn ] for every @xmath1432 and every sequence @xmath1790 with @xmath1791 the sequence @xmath1792 converges in the bounded strong@xmath1759 topology .
the limit @xmath1715 is independent of the sequence and defines a unitary representation @xmath1793 . from ( [ last formal step ] )
we have @xmath1794 and , since @xmath1715 is injective , @xmath1795 .
hence @xmath1796 is surjective , and thus @xmath1797 by exchanging @xmath1798 . for the homomorphism property
we write @xmath1799 and take limits to obtain @xmath1800 . convergence in the bounded strong@xmath1759 topology also requires convergence of the adjoint .
this property follows from @xmath1801 and the discussion after ( [ other side convergence ] ) , since @xmath1715 is now known to be an isomorphism .
unitarity of the representation is then immediate from @xmath1802 and @xmath1803 .
finally , we must show that @xmath1804 is continuous .
this amounts to showing that if @xmath1805 is a sequence in @xmath1332 which converges to @xmath1806 , then @xmath1807 for any @xmath1808 .
hence , we let @xmath1535 be a sequence in @xmath157 with @xmath1791 and choose @xmath1809 as in the proof of proposition [ well defined ] so that @xmath1810 and then let @xmath1811 . using the uniform boundedness of @xmath1812 as @xmath1811
, this shows that the convergence @xmath1813 is uniform in @xmath1814 .
since we have @xmath1815 for every fixed @xmath1816 , the continuity of @xmath1817 then implies that @xmath1818 .
let us underline two important consequences . for @xmath1819 and
@xmath1820 it follows that @xmath1821
if @xmath1822 and @xmath1823 are sequences in @xmath157 which converge to @xmath1824 , then , since @xmath1825 is continuous for all @xmath1501 in fock space , @xmath1826 converges pointwise to @xmath1827 . on the other hand
, we have @xmath1828 by the semigroup property , and the right - hand side converges pointwise to @xmath1829 .
we refer to @xmath1830 as the _ oscillator _ representation of the metaplectic group .
it has the following fundamental conjugation property .
let @xmath1831 , @xmath1832 denote the representation of @xmath1332 on @xmath132 defined by first applying the covering map @xmath1833 and then the standard representation of @xmath1257 .
if we let @xmath132 act on @xmath38 by the weyl representation @xmath770 then @xmath1834 since the inverse operator @xmath1835 is now available , this follows from the conjugation property at the semigroup level ( see proposition [ basic conjugation formula ] ) .
note that analogously we have the classical conjugation formula for the representation of the heisenberg group on the fock space @xmath3 , i.e. , @xmath1836 for all @xmath1432 and @xmath977 ( see proposition [ heisenberg conjugation ] ) .
the concrete formula for the eigenvalues of @xmath1578 which is given in proposition [ eigenvalues in t_+ ] shows that if @xmath1681 , then @xmath1578 is of trace class . using the conjugation property proved above ,
we now show that this holds for all @xmath1837 . for every @xmath1496 the operator
@xmath1578 is of trace class .
we must show that the operator @xmath1838 has finite trace . to verify this property observe that @xmath1839 with @xmath1840 .
since @xmath1841 for some @xmath1202 and @xmath1842 , the desired result follows from the explicit formula in proposition [ eigenvalues in t_+ ] for the eigenvalues of @xmath1843 .
[ prop : osc - char ] for every @xmath1496 one has @xmath1844 since the representation of the heisenberg group @xmath1562 on @xmath3 is irreducible , the space of bounded linear operators @xmath1845 that commute with the @xmath1562-action is just @xmath1846 .
now the dual of the space of trace - class operators is the space of bounded linear operators itself , where a bounded linear operator @xmath219 is realized as a functional on the trace - class operators by @xmath1847 ( see @xcite ) .
invariance of @xmath1848 by the @xmath1562-representation is equivalent to the invariance of @xmath219 as an operator .
thus , an @xmath1562-invariant bounded linear functional on the space of trace - class operators is just a multiple of @xmath1849 .
we construct such a functional , @xmath1850 , by letting @xmath1851 be the vacuum in @xmath1563 and averaging the expectation in @xmath1852 over all @xmath1853 : @xmath1854 indeed , the representation of @xmath1562 is unitary and by using @xmath1855 along with the basic property @xmath1856 , one shows that @xmath868 is invariant , i.e. , @xmath1857 for all @xmath1858 .
if @xmath1859 is the projector on the vacuum , then it is a simple matter to show that @xmath1860 thus , this formula holds for any bounded linear operator @xmath1861 of trace class .
now @xmath1862 is computed as @xmath1863 due to the fact that @xmath1864 is rapidly decreasing , we may take the expectation inside the inner integral : @xmath1865 again using the basic properties of @xmath1631 , we see that @xmath1866 thus the inner integral over @xmath1608 is a gaussian integral , and evaluating it we obtain @xmath1867 integrating this quantity with respect to @xmath1671 we obtain @xmath1868 and the desired result for the trace follows from ( [ normalization ] ) and @xmath1869 .
[ holomorphicity ] for every @xmath1870 in the weyl algebra and every @xmath1871 the operator @xmath1872 is of trace class on the fock space @xmath1563 .
furthermore , the function @xmath1873 , @xmath1874 , is holomorphic .
( sketch ) it is enough to show that @xmath1875 is of trace class for the case that @xmath1876 is a monomial operator , i.e. , @xmath1877 . we must then compute the trace of the square root of @xmath1878 for @xmath1879 .
for this , recall that the @xmath1880-operation interchanges multiplication and differentiation .
direct computation of @xmath1881 shows that this number is of the order of magnitude of @xmath1882 where @xmath41 is the torus element corresponding to @xmath1883 .
thus the conjugated operator has effectively the same trace as @xmath1884 itself and in particular @xmath1885 . turning to holomorphicity ,
if we insert the equation @xmath1886 in the formula ( [ general trace ] ) for the trace and use the definitions to compute the various integrals , then the function @xmath1887 is of the form @xmath1888 where @xmath1889 is a rapidly decreasing function on @xmath131 and @xmath1890 is holomorphic on the semigroup @xmath1429 .
since the @xmath1891operator can be exchanged with the integral , it is immediate that @xmath1892 is holomorphic and therefore so is @xmath1893 .
we now show that the semigroup representation @xmath1894 is compatible with the @xmath1115-representation @xmath1895 let @xmath1494 and @xmath1896 .
then , since the semigroup @xmath1094 is open in @xmath1257 , there exists some @xmath1897 so that the curve @xmath1898 \ni t \mapsto \mathrm{e}^{ty } h$ ] lies in @xmath1458 .
fix some point @xmath1899 and let @xmath1900 denote the lifted curve in @xmath1333 . for
all @xmath1496 and all @xmath1901 it follows that @xmath1902 recall that the operator @xmath1578 is the result of integrating the heisenberg translations @xmath1572 against the gaussian density @xmath1903 .
thus @xmath1904 for @xmath1645 the linear transformation @xmath1905 is in @xmath602 , and @xmath1115 is spanned by such transformations .
it is therefore sufficient to prove the statement of the lemma for @xmath1906 of this form .
hence let @xmath1907 and observe that the corresponding element in the weyl algebra is @xmath1908 now , defining @xmath1572 for @xmath1630 by @xmath1909 as before , we have @xmath1910 therefore , for @xmath1911 consider the expression @xmath1912 using @xmath1913 and shifting integration variables @xmath1914 we obtain @xmath1915 comparing eqs .
( [ eq:3.13],[eq:3.14 ] ) with ( [ eq:3.12 ] ) we see that the formula of the lemma is true if @xmath1916 but checking this equation is just a simple matter of computing derivatives .
recall that @xmath1917 and @xmath1918 .
writing @xmath1919 and using @xmath1920 one computes the left - hand side to be @xmath1921 on substituting @xmath1922 , this expression immediately agrees with the result of taking the two derivatives on the right - hand side .
having prepared the algebraic foundations ( @xmath44[sect : osp - howe ] ) and the necessary representation - theoretic tools ( @xmath1923 , we now turn to the main part of our work , which is to prove the formula ( [ eq : weylcharacter ] ) for the @xmath11-haar expectation value @xmath31 of a product of ratios of characteristic polynomials .
we will achieve this goal by exploiting the fact that @xmath31 is the same as the character @xmath1924 of the irreducible @xmath13-representation on @xmath12 .
the key property determining @xmath119 is a system of differential equations coming from the casimir invariants of the lie superalgebra @xmath286 : acting as differential operators on a certain supermanifold @xmath1925 of lie supergroup type , these invariants annihilate @xmath119 as a section of @xmath1925 . for a @xmath87-graded complex vector space @xmath189 let @xmath1926 be the lie superalgebra with bracket @xmath1927 = x x^\prime - ( -1)^{|x| |x^\prime| } x^\prime x$ ] .
then , given any finite - dimensional complex vector space @xmath1928 , define the _ grassmann envelope _ of @xmath1929 by the grassmann algebra @xmath1930 as @xmath1931 the grassmann envelope @xmath1932 is given the structure of an associative algebra by @xmath1933 and it also has the structure of a lie algebra by the usual commutator : @xmath1934 = \omega^\prime \omega \otimes [ x , x^\prime ] \;.\ ] ] more generally , if @xmath1935 is a complex lie ( sub)superalgebra , the grassmann envelope @xmath1936 of @xmath13 by @xmath860 is still defined in the same way : @xmath1937 and it still carries the same structure of associative algebra and lie algebra .
the supertrace @xmath1938 is defined by @xmath1939 .
note that this satisfies @xmath1940 = 0 $ ] for all @xmath1941 .
the grassmann envelope @xmath1942 is @xmath87-graded by @xmath1943 where @xmath1944 .
there also exists a @xmath386-grading , which is induced by that of @xmath860 ; this is @xmath1945 with @xmath1946 .
note that @xmath1947 \subset \tilde{\mathfrak{g } } ^{(k + l)}$ ] .
hence , @xmath1948 is a nilpotent ideal .
the subspace @xmath1949 is a lie subalgebra and we have @xmath1950 .
the following is an elementary statement found , e.g. , in @xcite .
let @xmath1951 and consider the decompositions @xmath1952 and @xmath1953 with respect to the @xmath87- and @xmath386-gradings of @xmath1954 .
the element @xmath1067 is invertible in @xmath1932 if and only if @xmath1955 is invertible , and the latter is the case if and only if @xmath1956 .
let @xmath1957 be any lie group such that @xmath1958 .
following @xcite we write @xmath1959 the objects @xmath128 , @xmath1960 , @xmath1961 , and @xmath1962 are subgroups of @xmath1957 .
the group @xmath1961 is a normal subgroup of @xmath1957 .
here is another elementary result from @xcite : each element @xmath1963 has a unique factorization of the form @xmath1964 with @xmath1965 and @xmath1966 .
each element @xmath1967 can be uniquely represented in the form @xmath1968 with @xmath1969 and @xmath1970 .
a complex _ lie supergroup _ is a pair @xmath1971 consisting of a complex lie superalgebra @xmath276 and a complex lie group @xmath33 that exponentiates @xmath441 , i.e. , @xmath1972 .
given a lie superalgebra @xmath13 there exist , in general , several choices of @xmath33 .
if @xmath1973 is a faithful finite - dimensional representation of @xmath286 , i.e. , @xmath1974 , then one choice is @xmath1975 . a _ superfunction _ on a complex lie supergroup @xmath1976 is a section in the sheaf @xmath1977 of germs of holomorphic functions on @xmath33 with values in @xmath1978 .
this sheaf @xmath1925 is a locally free sheaf of @xmath1979-modules where @xmath1979 is the sheaf of germs of holomorphic functions on @xmath33 .
the gradings on @xmath1978 give a @xmath386-grading @xmath1980 and a @xmath87-grading @xmath1981 .
if @xmath1982 is the decomposition of a superfunction @xmath1501 with respect to the former , then @xmath1983 is called the _ numerical part _ of @xmath1984 .
a _ superderivation _ @xmath1985 is a linear sheaf morphism @xmath1986 satisfying the graded leibniz rule @xmath1987 for @xmath1988 .
the @xmath87-grading of @xmath1925 defines a @xmath87-grading on the vector space @xmath1989 by @xmath1990 if @xmath1991 . in the theory of ordinary lie groups
one has a realization of the lie algebra by left- or right - invariant vector fields acting as derivations of the algebra of differentiable functions on the lie group . in the same spirit ,
one wants to construct a representation of the lie superalgebra @xmath13 by superderivations on the sheaf of algebras @xmath1925 .
the basic tool for this construction is a correspondence between superfunctions @xmath1992 and functions on @xmath1957 with values in @xmath1993 .
there exists no canonical correspondence of such kind ; one possible choice is as follows .
let @xmath1994 be a basis of @xmath441 , and let @xmath1995 be a basis of @xmath252 with dual basis @xmath1996 of @xmath1997 .
then if @xmath1998 and @xmath1999 denotes the corresponding left - invariant vector field on functions @xmath353 on @xmath33 , @xmath2000 we assign to a section @xmath2001 a function @xmath2002 by @xmath2003 where @xmath2004 and @xmath2005 .
the assignment @xmath2006 is injective if @xmath2007 and @xmath2008 ; and is so in particular if @xmath2009 . in that case , the inverse mapping @xmath2010 is given by restriction to elements @xmath2011 with @xmath2012 .
the advantage of passing to the right - hand side of the correspondence @xmath2013 is that functions on @xmath1957 can be shifted by elements of @xmath1957 by multiplication from the left or right .
this possibility now puts us in a position to construct the desired representation of the lie superalgebra @xmath13 by superderivations on the sheaf @xmath2014 .
consider first the case @xmath2015 .
making the choice @xmath2016 and linearizing the function @xmath2017 , @xmath2018 with respect to the parameter of the one - parameter group @xmath2019 , we obtain a first - order differential operator @xmath2020 : @xmath2021 similarly , @xmath2022 is defined by @xmath2023 .
derivations @xmath2024 on @xmath1925 are then given by @xmath2025 , i.e. , by the equation @xmath2026 ( @xmath2027 ) . to handle the case @xmath2028
, we need one extra anti - commuting parameter for the purpose of differentiation . denoting this parameter by @xmath233
we choose @xmath2029 and define @xmath2030 by @xmath2031 and @xmath2032 .
again , the corresponding derivations on @xmath1925 are @xmath2033 .
the assignments @xmath2034 and @xmath2035 are representations of the lie superalgebra @xmath13 on the sheaf @xmath2014 of superfunctions .
the two representations commute in the graded - commutative sense , i.e. , @xmath2036 .
a _ representation _ of a lie supergroup @xmath2037 on a @xmath87-graded vector space @xmath189 is a homomorphism of lie superalgebras @xmath2038 and a homomorphism of lie groups @xmath2039 which are compatible in the sense that @xmath2040 given a @xmath87-graded vector space @xmath530 , consider the lie superalgebra @xmath2041 and the lie supergroup @xmath2042 . for any integer @xmath1159 the two maps @xmath2043 , @xmath2044 , and @xmath2045 , @xmath2046 ,
define a representation of @xmath2047 on the @xmath87-graded one - dimensional vector space @xmath2048 with @xmath2049 and @xmath2050 .
the _ character _ @xmath119 associated with the representation @xmath2051 of a lie supergroup @xmath2037 is the superfunction @xmath2052 provided that the supertrace exists .
note that the numerical part @xmath2053 of the character is just the supertrace of the lie group representation @xmath2054 .
let @xmath119 be the character associated with the representation ( @xmath2061 of a lie supergroup @xmath2062 .
then by the definition of the correspondence @xmath2063 one has @xmath2064 where the notation above ( @xmath1969 , @xmath2004 , @xmath2065 ) is being employed .
this satisfies @xmath2066 , since @xmath2067 for @xmath2068 and the representations @xmath96 and @xmath84 are compatible .
for @xmath2069 the infinitesimal ( or linearized ) version of the same argument gives @xmath2070 and hence @xmath2071 . in chapter [ integrating ] we constructed a semigroup representation @xmath2072 and a group representation @xmath2073 which exponentiate the restriction of the spinor - oscillator representation @xmath110 to @xmath2074 .
these are compatible in the sense that @xmath2075 , @xmath1332 acts on @xmath126 by conjugation , and @xmath2076 for all @xmath1819 and @xmath2077 .
the representation @xmath2078 exponentiates to a lie group representation @xmath2079 .
note that @xmath2080 and @xmath2081 commute , as they act on the two different factors of the tensor product @xmath2082 .
note also that the group @xmath87 acts on @xmath2083 and @xmath1057 by deck transformations of the @xmath1102 coverings @xmath2084 and @xmath2085 .
the non - trivial element of @xmath87 is represented by a sign change , @xmath2086 and @xmath2087 .
given the representations @xmath2080 and @xmath2088 , we form the semigroup representation @xmath2089 we already know the representation @xmath1651 to be compatible with @xmath2090 .
now we define a superfunction @xmath894 on @xmath2091 with values in @xmath2092 by @xmath2093 where @xmath2094 is a basis of @xmath2095 and @xmath2096 is its dual basis .
we refer to @xmath894 as the spinor - oscillator character . by the circumstance that @xmath1651 and @xmath110 are representations on the infinite - dimensional vector space @xmath3 we are obliged to discuss the domain of definition of @xmath882 . for this ,
notice that @xmath1057 acts non - trivially only on the finite - dimensional ( or spinor ) part of @xmath933 .
expanding @xmath2097 we obtain a finite sum @xmath2098 with @xmath2099 .
( here we recall that @xmath2095 can be viewed as part of the clifford - weyl algebra @xmath99 by the isomorphism @xmath2100 . ) by proposition [ holomorphicity ] the operator @xmath2101 is of trace class .
thus the character @xmath894 is defined on the full domain @xmath2102 .
moreover , @xmath2103 depends analytically on @xmath1806 .
does there exist any good sense in which to think of the spinor - oscillator character @xmath894 as a radial superfunction ? there is no question that for every @xmath2104 we do have superderivations @xmath2105 and @xmath1999 on @xmath2106 , and the function @xmath894 by its definition as a character does satisfy @xmath2107 . however , our semigroup elements @xmath2108 do not possess an inverse in the spinor - oscillator representation and we therefore should nt expect such a relation as @xmath2109 . a substitute is this .
let @xmath2110 be the compact real subgroup which is obtained by exponentiating the skew - symmetric degree - two elements of the clifford algebra generated by a euclidean vector space @xmath2111 .
let @xmath2112 .
then on the real submanifold @xmath2113 ( with @xmath2114 formerly denoted by @xmath126 ) the superfunction @xmath894 is @xmath127-radial , since the representations @xmath2115 of @xmath1332 and @xmath2081 on @xmath2116 are compatible . as a result
, we will later be able to apply berezin s theory of radial parts of laplace - casimir operators . here
we show that the restriction of the numerical part of @xmath894 to a toral set @xmath2117 in @xmath2118 gives the autocorrelation function described in @xmath44[introduction ] .
that is , we show that the ratios of characteristic polynomials @xmath2119 and their averages with respect to a compact group @xmath11 ( with @xmath2120 and @xmath1503 being a parameter ) , can be expressed as supertraces of operators on the spinor - oscillator module @xmath38 and the submodule of @xmath11-invariants @xmath37 .
we begin with a summary of the relevant facts . from proposition [ basic conjugation formula ]
we recall the basic conjugation rule for the oscillator representation : @xmath2121 where @xmath2122 , @xmath2123 , and @xmath2124 . on the side of the spinor representation ,
the corresponding conjugation formula is @xmath2125 this defines the 2:1 covering homomorphism @xmath1058 , @xmath2126 , which exponentiates the isomorphism of lie algebras @xmath2127 .
recall also ( from proposition [ prop : osc - char ] ) the formula for the oscillator character : @xmath2128 an analogous result is known for the case of the spinor representation ; see , e.g. , the textbook @xcite . defining the spinor character as the supertrace with respect to the canonical @xmath87-grading of the spinor module , one has @xmath2129 thus the spinor character , just like the oscillator character , is a square root . by taking the supertrace over the total fock representation space
, we obtain the formula @xmath2130 for @xmath2131 , the case of our interest , @xmath2132 intersects with @xmath2133 and the square root @xmath2134 is defined in such a way that @xmath2135 for @xmath2136 .
let now @xmath88 be a @xmath87-graded vector space , and let @xmath529 .
the lie group @xmath2137 acts on @xmath2138 by @xmath2139 this action serves to realize the group @xmath2140 as a subgroup of @xmath2141 . for the purpose of letting @xmath33
act in the spinor - oscillator module @xmath38 , let this representation be lifted to that of a double covering @xmath2142 .
the following statement gives the value of the spinor - oscillator character on @xmath2143 where @xmath2144 and @xmath2145 are diagonal matrices in @xmath2146 ( for @xmath333 ) . in that case @xmath2151 and @xmath2152 are single numbers and @xmath2154 acts on @xmath2155 as @xmath2156 for @xmath2157 . from equation ( [ eq : spin - osc - char ] ) we then have @xmath2158 which turns into the stated formula on pulling out a factor of @xmath2159 from under the square root .
( of course , the double covering of @xmath2160 is to be used in order to define this square root globally . ) in the formula of lemma [ lem : strav ] we now set @xmath2161 and @xmath2162 .
we then put @xmath2163 and integrate against haar measure @xmath21 of unit mass on @xmath11 . this integral and
the summation that defines the supertrace can be interchanged , as @xmath2164 is a finite sum of power series and the conditions @xmath2165 ensure uniform and absolute convergence . the representation of @xmath11 on @xmath38 is induced by the representation of @xmath11 on @xmath168 . therefore , averaging over @xmath11 with respect to haar measure has the effect of projecting from @xmath2166 to the @xmath11-trivial isotypic component @xmath37 , and we arrive at @xmath2167 in the case of an even dimension @xmath2168 , the domain of definition of this formula is a complex torus @xmath2169 where @xmath2170 and @xmath2171 is the open subset determined by the conditions @xmath2172 for all @xmath108 . for odd @xmath2168
we must continue to work with a double cover ( also denoted by @xmath1123 ) to take the square root @xmath2173 .
let @xmath13 be the howe dual partner of @xmath14 in @xmath97 .
we know from proposition [ prop : dualpairs ] that @xmath2174 for @xmath2175 and @xmath2176 for @xmath49 .
recall also from @xmath44[subsubsec : wohdp ] that the @xmath13-representation on @xmath37 is irreducible and of highest weight @xmath2177 .
denote by @xmath2178 the set of weights of this representation .
let @xmath2179 multiplied with the correct sign to form the supertrace . on the right - hand side we recognize the correlation function ( see @xmath44[introduction ] ) which is the object of our study and , as we have explained , is related to the character of the irreducible @xmath13-representation on @xmath37 .
the left - hand side gives this character ( restricted to the toral set @xmath1123 ) in the form of a weight expansion , some information about which has already been provided by corollary [ cor : weights ] of @xmath44[subsubsec : wohdp ] .
being an eigenfunction of ( the radial parts of ) the laplace - casimir operators that represent the center of the universal enveloping algebra @xmath267 , our character @xmath2181 satisfies a certain system of differential equations .
here we first describe the origin and explicit form of these differential equations .
we then prove that @xmath119 is determined uniquely by these in combination with the weight constraints for @xmath2182 .
finally , we provide the explicit function with these properties .
formula ( [ eq:4.3mrz ] ) and corollary [ cor : weightexpansion ] express the character @xmath119 as a function on the toral set @xmath1123 .
our next step is to describe a supermanifold with a @xmath13-action where @xmath119 exists as a @xmath127-radial superfunction and laplace - casimir operators can be applied . here
we give only a rough sketch , leaving the details to the reader .
first of all , the symmetry group @xmath2183 for @xmath119 has to be identified .
recall from the end of @xmath44[sect:4.2.1 ] that the good real group acting in the spinor - oscillator representation @xmath3 is @xmath2184 which contains @xmath81 and @xmath52 as subgroups .
since we are studying the character @xmath119 of the @xmath13-representation on the subspace @xmath12 of @xmath11-invariants , we now seek the subgroup @xmath2185 which centralizes @xmath11 ; this means that we are asking the exponentiated version of a question which was answered at the lie algebra level in @xmath44[sect:2.7 ] . here
, restricting the group @xmath2186 to the centralizer of @xmath11 we find @xmath2187 we observe that @xmath127 for the case of @xmath81 is just the lower - dimensional copy of @xmath2186 which corresponds to @xmath2188 taking the role of @xmath2189 .
we also see immediately that the lie algebras @xmath2190 coincide with the real forms described in propositions [ prop:2.4 ] and [ prop:2.5 ] .
the second object to construct is a real domain @xmath2191 with a @xmath127-action by conjugation such that @xmath2192 where @xmath2193 is a real sub - torus .
the choice we make for @xmath2194 is the one singled out by the parametrization @xmath2195 and @xmath2162 with real - valued coordinates @xmath2196 and @xmath2197 .
we also want the elements of @xmath2198 to commute with those of @xmath11 as endomorphisms of @xmath132 . by these requirements ,
the good real domain @xmath2198 to consider in the case of @xmath81 is the lower - dimensional copy of @xmath2199 which , again , corresponds to @xmath2188 replacing @xmath2189 . by the detailed analysis of @xmath44[integrating ] ( where @xmath2114 was simply denoted by @xmath126 ) and
the fact that the diagonal elements constitute a maximal torus in @xmath2200 , we infer the desired property @xmath2201 . in the case of @xmath52 the same requirements lead to the choice @xmath2202 where the construction of @xmath2203 is fully parallel to that of @xmath2204 : if @xmath2205 denotes the complex orthogonal group of the vector space @xmath2206 , we introduce the semigroup @xmath2207 which is defined by the inequality @xmath2208 for @xmath2209 and then take @xmath2210 to be the totally real submanifold @xmath2210 of pseudo - hermitian elements @xmath2211 . for this choice we easily check that @xmath2201 . having constructed a manifold @xmath2198 with an action of @xmath127 by conjugation , we now consider the supermanifold @xmath2212 which is the sheaf of algebras of analytic functions on @xmath2198 with values in @xmath2213 where , once again , we remind the reader that @xmath2174 for @xmath2175 and @xmath2176 for @xmath49 . by construction , for every point @xmath2214 we have @xmath2215 , the even part of the lie superalgebra @xmath286 .
hence by the basic principles reviewed at the beginning of this chapter , the supermanifold @xmath2216 carries two representations of @xmath13 by superderivations ( @xmath2217 and @xmath2218 ) . the benefit from all this is that for every casimir invariant @xmath161 we now get a laplace - casimir operator @xmath164 on @xmath1925 by replacing each @xmath2219 in the polynomial expression for @xmath115 by the corresponding differential operator @xmath2105 ( or @xmath2220 . utilizing the present setting the character @xmath119 , which was given in ( [ eq:4.3mrz ] ) as a function on @xmath1123 ,
will now be extended to a section of @xmath2221 . by construction ,
the direct product @xmath2222 is contained as a subspace in @xmath2223 , and by restriction of the representation @xmath1651 we get a mapping @xmath2224 whose image still lies in the subspace of trace - class operators .
a good definition of @xmath2225 therefore is @xmath2226 where @xmath2227 now is a basis of @xmath252 and @xmath2228 is its dual basis .
the symbol @xmath96 here denotes the @xmath13-representation on @xmath12 .
it is clear that by restricting the numerical part of @xmath2229 to the toral set @xmath2230 we recover the function described in ( [ eq:4.3mrz ] ) .
consider now the numerical part of the section @xmath2231 .
being a @xmath127-radial function , @xmath2053 is invariant under the action of the weyl group @xmath132 which normalizes @xmath2194 with respect to the @xmath127-action on @xmath2198 by conjugation .
since the latter group action decomposes as a direct product of two factors , so does @xmath132 . for both cases ( @xmath81 , @xmath20 ) the second factor of the weyl group
@xmath132 is just the permutation group @xmath2232 . as a matter of fact , conjugation of a diagonal element @xmath2233 or @xmath2234 by @xmath2235 or @xmath2236 can return another diagonal element only by permutation of the eigenvalues @xmath2237 of @xmath2238 .
( no inversion @xmath2239 is possible , as this would mean transgressing the oscillator semigroup . )
this factor @xmath2240 of @xmath132 will play no important role in the following , as the expressions we will encounter are automatically invariant under such permutations .
the first factors of @xmath132 are of greater significance .
for the two cases of @xmath81 and @xmath52 these are the weyl groups @xmath2241 and @xmath2242 respectively .
an explicit description of these groups is as follows .
let @xmath2243 be an orthonormal basis of @xmath327 and decompose @xmath2244 into a direct sum of 2-planes , @xmath2245 where @xmath1136 is spanned by the vector @xmath2246 and the linear form @xmath2247 ( @xmath2248 ) . in both cases at hand , i.e. , for the symmetric form @xmath222 as well the alternating form @xmath219 ,
this is an orthogonal decomposition .
the real torus under consideration is parameterized by @xmath2249 acting by @xmath2250 and @xmath2251 .
the weyl group @xmath2252 is generated by the permutations of these planes and the involutions which are defined by conjugation by the mapping that sends @xmath2253 and @xmath2254 .
the weyl group @xmath2255 is generated by the permutations together with the involutions which are induced by the mappings that simply exchange @xmath2246 with @xmath2256 .
since we are in the special orthogonal group and the determinant for a single exchange @xmath2257 is @xmath2258 , the number of involutions in any word in @xmath2255 has to be even . in summary ,
the @xmath132-action on our standard bases of linear functions , @xmath2259 and @xmath2260 , is given by the respective permutations together with the action of the involutions defined by sign change , @xmath2261 .
in the sequel , the weyl group action will be understood to be either this standard action or alternatively , depending on the context , the corresponding action on the exponentiated functions @xmath2262 and @xmath2263 . as a final remark on the subject ,
let us note that the weyl group symmetries of the function @xmath1924 can also be read off directly from the explicit expression ( [ eq:4.3mrz ] ) .
in particular , the absence of reflections @xmath2264 is clear from the conditions @xmath2265 .
we are now ready to start deriving a system of differential equations for @xmath119 .
recall that every casimir element @xmath2266 determines an invariant differential operator @xmath164 , called a laplace - casimir operator , by replacing each element @xmath2267 in the polynomial expression of @xmath115 by @xmath2105 or @xmath1999 .
let @xmath119 be the character of an irreducible representation @xmath2268 of a lie supergroup @xmath32 on a complex vector space .
then for any casimir element @xmath2266 the character @xmath119 is an eigenfunction of the differential operator @xmath164 .
we have @xmath2269 .
since @xmath115 is a casimir element , @xmath2270 commutes with @xmath2271 for all @xmath2272 . by schur s lemma for trace - class operators , @xmath2273 with @xmath2274 .
hence , @xmath2275 .
recall now from @xmath44[sect : osp - cas ] that for every @xmath461 we have a casimir element @xmath2276 of degree @xmath2277 .
recall also that under the assumption @xmath2278 we introduced @xmath2279 , @xmath2280 \in \mathfrak{osp}_0 \,$ ] , and @xmath2281 such that @xmath2282 $ ] and @xmath2283 = 0\,$ ] .
let us insert here the following comment .
while @xmath2284 were abstractly defined as @xmath6-generators , they acquire a transparent meaning when represented as operators on the spinor - oscillator module @xmath38 .
in fact , using formula ( [ eq : tau - inv ] ) one finds @xmath2285 here , for notational brevity , we make no distinction between @xmath6-elements @xmath117 , @xmath2286 , etc.and the operators representing them on @xmath38 .
it is now natural to identify the spinor - oscillator module @xmath38 ( or rather , a suitable completion thereof ) with the complex of holomorphic differential forms on @xmath5 .
writing @xmath2287 and @xmath2288 we then see that @xmath117 is the holomorphic exterior derivative : @xmath2289 @xmath2290 becomes the operator of contraction with the vector field @xmath2291 generating scale transformations @xmath2292 , while @xmath2293 = \partial \circ \iota(v ) + \iota(v ) \circ \partial$ ] is interpreted as the lie derivative with respect to that vector field .
consider now any irreducible @xmath6-representation on a @xmath87-graded vector space @xmath2294 with the property that the @xmath2294-supertrace of @xmath2295 ( @xmath2296 ) exists .
let @xmath2297 be the scalar value of the casimir invariant @xmath462 in the representation @xmath2294 .
then a short computation using @xmath2298 $ ] and @xmath2299 = 0 $ ] shows that @xmath2297 multiplied by @xmath2300 vanishes : @xmath2301 = \mathrm{str}_\mathcal{v}\ , [ \partial , \mathrm{e}^{-t \lambda } f_\ell ] = 0 \;,\ ] ] since the supertrace of any bracket is zero .
thus we are facing a dichotomy : either we have @xmath2302 , or else @xmath2303 for all @xmath2304 . now our representation @xmath118 realizes the latter alternative , which leads to the following consequence . [
lem : d(il)chi=0 ] assuming that @xmath88 is a @xmath87-graded vector space with @xmath2305 , let @xmath2306 be the character of the irreducible @xmath15-representation on @xmath37 for @xmath529 .
then @xmath2307 for all @xmath2308 .
for any real parameter @xmath2311 the supertrace of the operator @xmath2312 on @xmath37 certainly exists and is non - zero .
in fact , using formula ( [ eq:4.3mrz ] ) one computes the value as @xmath2313 the dichotomy of @xmath2314 therefore gives @xmath2315 .
since the casimir elements commute with all elements of the lie superalgebra , the laplace - casimir operators leave the set of radial superfunctions invariant .
we denote by @xmath2316 the radial parts of the laplace - casimir operators @xmath2317 for @xmath2304 .
these operators , which arise by restricting the laplace - casimir operators to the space of radial functions , are given by differential operators on the torus @xmath2318 .
they have been described by berezin @xcite , and his results will now be stated in a form well adapted to our purposes .
recall from @xmath44[sect : osp - roots ] that @xmath6-roots are of the form @xmath2319 .
recall also the relations @xmath2320 and @xmath2321 ( @xmath798 ) . in the following
we regard the variables @xmath2322 and @xmath2196 as real ( local ) coordinates for the real torus @xmath2318 . for @xmath2323 let @xmath2324 be the degree-@xmath463 differential operator @xmath2325 let @xmath2326 be the function defined on a dense open subset of @xmath2194 by @xmath2327 where @xmath2328 is a system of even and odd positive roots ( cf .
@xmath44[sect : simple - roots ] ) . up to a constant multiple
the radial part of the differential operator @xmath2317 is of the form @xmath2329 where @xmath2330 is some polynomial ( with constant coefficients ) in the operators @xmath2331 of total degree less than that of @xmath2332 .
the first statement is a direct consequence of theorem 3.2 of @xcite using the explicit form of the casimir elements @xmath462 defined in equation ( [ eq : casimirelement ] ) and the relations @xmath2333 and @xmath2334 .
the statement about the lower - order terms @xmath2330 is from theorem 4.5 of @xcite .
the methods of this section can be used to derive differential equations for the character of a certain class of irreducible representations of @xmath2342 .
define @xmath2343 here , @xmath2344 and @xmath2345 are the sets of even and odd positive roots of @xmath427 .
the following statement is corollary 4.12 of @xcite adapted to the present context and notation .
the idea of the proof is the same as that of lemma [ lem : d(il)chi=0 ] .
[ cor:4.4 ] let @xmath119 be the character of an irreducible representation of the lie supergroup @xmath2346 on a finite - dimensional @xmath87-graded vector space @xmath731 .
if @xmath2309 but @xmath2347 , then @xmath2348 for all @xmath2338 .
recall that the main goal of this paper can be stated as that of explicitly computing the character @xmath119 of a local representation defined on the space of invariants @xmath37 in the spinor - oscillator module .
we have restricted ourselves to the cases where @xmath11 is either @xmath55 or @xmath2 .
the representation on @xmath37 is defined at the infinitesimal level on the full complex lie superalgebra @xmath13 which is the howe partner of @xmath11 in the canonical realization of @xmath6 in the clifford - weyl algebra of @xmath2349 .
recall that @xmath2178 denotes the set of weights of the @xmath13-representation on @xmath37 .
recall also from corollary [ cor : weights ] that the weights @xmath2350 satisfy the weight constraints @xmath2351 .
the highest weight is @xmath2352 .
[ unicity ] the character @xmath2355 is annihilated by the differential operators @xmath2356 for all @xmath2304 , and it has a convergent expansion @xmath2357 where the sum runs over weights @xmath2358 satisfying the constraints @xmath2351 . for the case of @xmath52
it is the unique function on @xmath1123 with these two properties and @xmath2359 .
for @xmath51 it is the unique @xmath132-invariant function on @xmath1123 with these two properties and @xmath2360 , @xmath2361 . to verify the property @xmath2362 which holds for the case of @xmath0 , look at the right - hand side of the formula of corollary [ cor : weightexpansion ] : in order to generate a term @xmath2363 in the weight expansion , you must pick the term @xmath2364 in the expansion of the determinant for @xmath2365 in the numerator ; but the latter term depends on @xmath2366 as @xmath2367 which vanishes upon taking the haar average for @xmath2368 . by @xmath132-invariance
the property @xmath2362 is equivalent to @xmath2369 for all @xmath108 . in view of this remark and corollaries [ cor : weights ] and [ cor : dljchi ] ,
it is only the uniqueness statement of theorem [ unicity ] that remains to be proved here .
this requires a bit of preparation , in particular to appropriately formulate the condition @xmath2370 . for
that we develop @xmath2371 in a series @xmath2372 where the @xmath2373 are @xmath2324-eigenfunctions for _ every _ @xmath2338 . the first step is to determine an appropriate expansion for @xmath2326 . recall that @xmath2374 given a factor in the denominator of this representation , we wish to factor out , e.g. , @xmath2375 to obtain a term @xmath2376 which we will attempt to develop in a geometric series . in order for this to converge uniformly on compact subsets of @xmath1123
it is necessary and sufficient for @xmath2377 to be positive on @xmath2378 .
this of course depends on the root @xmath2379 .
fortunately , the sets of odd positive roots for our two cases of @xmath2380 and @xmath2381 are the same ( see @xmath44[sect : simple - roots ] ) : @xmath2382 so indeed , if we factor out @xmath2383 from each term in the denominator and do the same in the numerator , we obtain the expression @xmath2384 and it is possible to expand each term of the denominator in a geometric series . here
@xmath2385 is half the _ supersum _ of the positive roots .
now let @xmath2386 be a basis of simple positive roots ( cf .
@xmath44[sect : simple - roots ] ) and expand the terms @xmath2387 in geometric series to obtain @xmath2388 which converges uniformly on compact subsets of @xmath1123 . in this expression @xmath2389 and @xmath206 denote the vectors @xmath2390 and @xmath2391 , respectively , and @xmath2392 . following the usual multi - index notation
, @xmath2393 means @xmath2394 for all @xmath1709 .
note @xmath2395 .
now we know that the character has a convergent series representation @xmath2396 thus we may write @xmath2397 for convenience of the discussion we let @xmath2398 and reorganize the sums as @xmath2399 where the inner sum is a finite sum which runs over all @xmath2393 such that @xmath2400 .
we are now in a position to explain the recursion procedure which shows that @xmath119 is unique .
start by applying @xmath2401 to @xmath2402 as represented in the expression ( [ ready for recursion ] ) .
since @xmath2403 is of the form @xmath2404 , we immediately see that it is an eigenfunction with eigenvalue @xmath2405 .
the functions @xmath2406 in the sum are independent eigenfunctions .
hence it follows that @xmath2407 for all @xmath2408 fixed and then for all @xmath2409 . from now on
we consider the equations ( [ recursion equations ] ) only in those cases where @xmath2408 is itself a weight of our representation .
( we have license to do so as only the uniqueness part of theorem [ unicity ] remains to be proved . ) in this case we have the following key fact . our first job is to compute @xmath752 . for the list of even and odd positive roots we refer the reader to @xmath2412 .
direct computation shows that if @xmath0 , then @xmath2413 the same computation for the case of @xmath52 shows that @xmath2414 now we write @xmath2415 with the weight constraints @xmath2416 . the assumption that @xmath2417 vanishes for all
@xmath2418 means that @xmath2419 in the case of @xmath0 . in the case of @xmath64
it means that @xmath2420 in the second case the only solution for @xmath2421 and @xmath2422 satisfying the weight constraints is the highest weight @xmath41 itself . in the first case there is one other solution , namely that which is obtained from the highest weight by replacing @xmath2423 by @xmath2424 .
however , one directly checks that in the @xmath55 case , where @xmath2425 is not a root , it is not possible to obtain such a @xmath894 by adding some combination of roots from @xmath430 to @xmath2426
. we will determine @xmath2427 recursively , starting from @xmath2428 .
let @xmath2429 be a weight that satisfies the weight constraints .
then if @xmath52 we know that @xmath2430 is non - zero for some @xmath2431 .
it therefore follows from equation ( [ recursion equations ] ) and @xmath2395 that @xmath2432 where the sum runs over all @xmath2433 ( recall that @xmath2434 is always the case ) such that @xmath2435 .
since the weights @xmath2436 involved in the sum are smaller than @xmath894 in the natural partial order defined by the basis of simple roots , equation ( [ basic equation ] ) defines a recursion procedure for determining all coefficients @xmath2437 . in the case of @xmath81 we are confronted with the fact that the weight @xmath2438 satisfies the weight constraints and yet gives @xmath2439 for all @xmath2418 .
however , in this exceptional situation the conditions of theorem [ unicity ] provide that @xmath2440 .
thus the expansion coefficients @xmath2427 are still uniquely determined by our recursion procedure .
as before let @xmath2441 be a set of positive roots of @xmath2442 or @xmath2443 .
we now decompose these sets as @xmath2444 which means that @xmath2445 is a set of positive roots of @xmath427 .
let @xmath2446 be further decomposed as @xmath2447 where @xmath2448 and @xmath2449 are the subsets of even and odd @xmath41-positive roots .
then the function @xmath2326 has a factorization as @xmath2450 with @xmath2451 and @xmath2452 is half the supersum of @xmath41-positive roots . for the case of @xmath2175 one
finds @xmath2453 , while for @xmath64 one has @xmath2454 .
the weyl group @xmath132 acts on @xmath1123 and therefore on functions on @xmath1123 .
let @xmath2455 be the subgroup which stabilizes the highest weight @xmath2456 and thus the corresponding function @xmath2457 on @xmath1123 .
note that @xmath2458 is the direct product of the permutations of the set @xmath2263 and the permutations of the set @xmath2459 .
the symmetrizing operator @xmath2460 from @xmath2458-invariant analytic functions to @xmath132-invariant analytic functions on @xmath1123 is given by @xmath2461 \in w / w_\lambda } w(f ) \;.\ ] ] notice that the function @xmath2462 is @xmath2463-invariant ; the symmetrized function @xmath2464 then is @xmath132-invariant .
we now wish to show that this function coincides with our character @xmath119 . in this endeavor , an obstacle appears to be that @xmath2465 by corollary [ cor : weightexpansion ] is a polynomial in the variables @xmath2466 , whereas the function @xmath2467 has poles at @xmath2468 .
hence our next step is to show that these poles are actually cancelled by the process of @xmath132-symmetrization .
an even root @xmath2471 is some linear combination of either the functions @xmath2322 or the functions @xmath2472 . denoting the latter subset of even roots by @xmath2473 ,
let @xmath2474 for @xmath2475 be the complex submanifold @xmath2476 by definition , the function @xmath2477 is holomorphic on @xmath2478 now it is a theorem of complex analysis that if a function is holomorphic outside an analytic set of complex codimension at least two , then this function is everywhere holomorphic .
therefore , since the intersection of two or more of the submanifolds @xmath2479 is of codimension at least two in @xmath1123 , it suffices to show that for any @xmath2480 our function @xmath2481 extends holomorphically to @xmath2482 hence let @xmath820 be some fixed root in @xmath2483 .
there exists a weyl group element @xmath1630 and a @xmath1177-invariant neighborhood @xmath327 of @xmath2484 such that @xmath2485 is a reflection fixing the points of @xmath2486 .
let @xmath2487 be a complex coordinate which is transverse to @xmath2484 in the sense that @xmath2488 . because the root @xmath820 occurs at most once in the product @xmath2467 , the function @xmath2489 has at most a simple pole in @xmath2490 .
we may choose @xmath327 in such a way that @xmath2491 is holomorphic on @xmath2492 .
doing so we have a unique decomposition @xmath2493 where @xmath219 and @xmath253 are holomorphic in @xmath327 . since @xmath2491 is @xmath132-invariant
, we conclude that @xmath2494 and hence @xmath2495 along @xmath2486 . using @xmath2500 we write @xmath2501 .
then , lifting the sum over cosets @xmath2502 \in w / w_\lambda$ ] to a sum over weyl group elements @xmath1630 we obtain @xmath2503 where @xmath2504 is the determinant of @xmath2505 .
the factor @xmath2506 is the character of the representation @xmath2507 of ( a double cover of ) the lie supergroup @xmath2508 .
this representation is one - dimensional , and from corollary [ cor:4.4 ] we have @xmath2509 for all @xmath2510 .
here we carry out the final step in proving the explicit formula for the character @xmath119 of our representation .
since the formula in the case of @xmath76 follows directly from that for @xmath2175 ( see @xmath2512 ) and the case of @xmath2513 has been handled in @xcite , we need only discuss the cases of @xmath81 and @xmath49 . in order to show that the character is indeed given by @xmath2514 with @xmath2515 , it remains to prove that in the series development @xmath2516 of the function defined by @xmath2517 \in w / w_\lambda } \varphi_{[w ] } \ ; , \quad \varphi_{[w ] } : = \mathrm{e}^{w(\lambda _ n ) } \frac{\prod_{\beta \in \delta _ { \lambda,1}^+ } ( 1 - \mathrm{e}^ { -w(\beta ) } ) } { \prod_{\alpha \in \delta_{\lambda,0}^+ } ( 1 - \mathrm{e}^{-w(\alpha ) } ) } \;,\ ] ] the only non - zero coefficients @xmath2427 are those where the linear functions @xmath894 are of the form @xmath2518 with @xmath2519
. we also have to show that @xmath2520 in the case of the exceptional weight @xmath2521 occurring for @xmath0 .
we have shown above that @xmath2515 is holomorphic on the product of the full complex torus of the variables @xmath2522 with the domain defined by @xmath2523 .
although the individual terms @xmath2524}$ ] in the representation of @xmath353 have poles ( which cancel in the weyl group averaging process ) we may still develop each term of @xmath353 in a series expansion ; this will in fact yield the desired weight constraints .
we begin with the situation where @xmath0 . in this case @xmath2525
consists of the roots @xmath2526 ( @xmath2527 ) and @xmath2528 ( @xmath2529 , and @xmath2530 is the set of roots of the form @xmath2531 ( @xmath2532 ) . let us first consider the term of @xmath353 where @xmath2502 = w_\lambda\,$ ] .
its denominator can be developed in a geometric series on the region corresponding to @xmath2533 for all @xmath2534 .
there we may write this term as @xmath2535}=\mathrm{e}^{\lambda_n}\prod_{\beta } ( 1-\mathrm{e}^{-\beta } ) \prod_\alpha \sum_{n \ge 0 } \mathrm{e}^{-n\alpha } \;.\ ] ] here and for the remainder of this paragraph @xmath820 runs through the @xmath41-positive even roots and @xmath2536 through the @xmath41-positive odd roots .
recall that @xmath2537 , and note that all powers of @xmath2522 and @xmath2538 occurring in the series expansion of @xmath2539 are non - positive .
thus , if @xmath2540 is a weight which arises in @xmath2541}\,$ ] , then @xmath2542 and @xmath2543 . in the case of the @xmath2421 this is a statement only about the term @xmath2544}\,$ ] , but , since the action of the weyl group on the variables @xmath2545 is just by permutation of the indices , it follows that @xmath2542 holds always , independent of the term @xmath2546}$ ] under consideration .
hence , we neglect the @xmath2545 in our further discussion and only analyze the powers of the exponentials @xmath2522 which arise in the other terms @xmath2546}\,$ ] . given a fixed index
@xmath2547 we will develop every term @xmath2546}$ ] on a region @xmath2548 defined by certain inequalities which in the case of @xmath2549 are @xmath2550 we now discuss this case in detail .
recall that @xmath2551 occurs in the denominator in factors of the form @xmath2552 for @xmath2553 .
if @xmath2554 , then we expand these factors just as in the case of @xmath2544}\,$ ] .
convergence of the resulting series is guaranteed no matter what @xmath1177 does to @xmath2472 .
in the situation where @xmath2555 we rewrite the factors in the denominator as @xmath2556 and expand , and convergence in @xmath1651 is again guaranteed . adding these series we obtain a series representation @xmath2557 } \varphi_{[w ] } = \sum_\gamma b_\gamma \ , \mathrm{e}^\gamma \;,\ ] ] which is convergent on @xmath1717 . if @xmath2554 , then by the same argument as in the case of @xmath2502 = [ \mathrm{id}]$ ] we see that @xmath2561 occurs in the series development of @xmath2546}$ ] with a power @xmath2562 of at most @xmath2563 . now suppose that @xmath2564 .
then , following the prescription above we rewrite the @xmath2565-dependent factors in @xmath2546}$ ] as @xmath2566 and expand the r.h.s . in powers of @xmath2567 .
it follows that in this case @xmath2568 , which for any positive integer @xmath2168 implies that @xmath2569 .
[ constraints fulfilled ] suppose that @xmath81 and let @xmath2570 be the globally convergent series expansion of the proposed character @xmath2515 .
then for every weight @xmath2558 with @xmath2571 it follows that @xmath2572 . above we showed that on the region @xmath1651 the proposed character @xmath353 has a series development where in every @xmath894 the coefficient @xmath2562 of @xmath2551 is at most @xmath2563 .
recalling the fact that the function @xmath353 is holomorphic on @xmath1123 , we infer that @xmath2573 also holds true for the globally convergent series development @xmath2574 . to get the same statement for @xmath2575 with @xmath2576 we just change the definition of @xmath1651 to @xmath2577 defined by the inequalities @xmath2578 .
arguing for general @xmath2366 as we did for @xmath2549 in the above lemma , we show that the coefficient @xmath2421 of @xmath2575 in every @xmath894 in the series expansion of every @xmath2546}$ ] on @xmath2577 is at most @xmath2563 . by the holomorphic property ,
the same is true for the global series expansion of the proposed character @xmath353 . hence , to complete the proof we need
only show the inequality @xmath2579 .
but for this it suffices to note that for every @xmath2366 there is an element @xmath1177 of the weyl group with @xmath2580 .
indeed , using the weyl invariance of @xmath353 , if there was some @xmath894 where @xmath2581 , then the coefficient of @xmath2575 in @xmath2582 would be larger than @xmath2563 . to complete our work
, we must prove lemma [ constraints fulfilled ] for the case @xmath49 .
for this we use the same notation as above for the basic linear functions , namely @xmath2583 and @xmath2584 . here , compared to the @xmath55 case , there are only slight differences in the @xmath41-positive roots and the weyl group .
the only difference in the roots is in @xmath2585 where @xmath2586 occurs in the larger range @xmath2587 and @xmath2528 in the smaller range @xmath2588 .
the weyl group acts by permutation of indices on both the @xmath2589 and @xmath2322 and by sign reversal on the @xmath2590 . in this case , as opposed to the case above where only an even number of sign reversals were allowed , every sign reversal transformation is in the weyl group . in order to prove lemma [ constraints fulfilled ] in this case
, we need only go through the argument in the @xmath55 case and make minor adjustments .
in fact , the main step is to prove lemma [ m_1 estimate ] and , there , the only change is that the range of @xmath2591 for the factor @xmath2592 is larger .
this is only relevant in the case @xmath2593 , where we rewrite the additional denominator term @xmath2594 as @xmath2595 .
hence the factor in front of the ratio of products on the r.h.s . of equation ( [ eq:4.7mrz ] ) gets an additional factor of @xmath2596 and now is @xmath2597 .
thus @xmath2598 which certainly implies @xmath2599 .
moreover , using the fact that the weyl group transformations for @xmath2175 always involve an even number of sign changes , one sees that @xmath2362 in that case . as a consequence of the uniqueness theorem ( theorem [ unicity ] ) we therefore have @xmath2602 \in w / w_\lambda } \mathrm{e}^{w(\lambda_n)}\frac{\prod_{\beta\in\delta_{\lambda,1}^+ } ( 1 - \mathrm{e}^{-w(\beta)})}{\prod_{\alpha\in\delta_{\lambda,0}^+ } ( 1 - \mathrm{e}^{-w(\alpha)})}\ ] ] in both the @xmath55 and @xmath20 cases . since the @xmath57 case has been handled as a consequence of the result for @xmath2603 , our work is now complete .
1 berezin , f.a . : _ introduction to superanalysis _ , reidel publishing comp . ,
dordrecht , 1987 berline , n. , getzler , e. , vergne , m. : _ heat kernels and dirac operators _ , berlin , springer verlag , 1992 bump , d. , gamburd , a. : on the averages of characteristic polynomials from classical groups , commun .
math.phys .
* 265 * ( 2006 ) 227 - 274 conrey , j.b . ,
farmer , d.w . ,
zirnbauer , m.r . : howe pairs , supersymmetry and ratios of random characteristic polynomials for the unitary group ( math - ph/0511024 ) conrey , j.b . ,
forrester , p. , snaith , n.c . : averages of ratios of characteristic polynomials for the compact classical groups , int .
notices * 7 * ( 2005 ) 397 - 431 folland , g.b . : _ harmonic analysis on phase space _ ,
annals of mathematics studies 122 , princeton , princeton university press , 1989 howe , r. : remarks on classical invariant theory , trans .
ams * 313 * ( 1989 ) 539 - 570 howe , r. : the oscillator semigroup , proc . of symposia in pure mathematics
* 48 * , ams ( 1987 ) reed , m. , simon , b. : _ methods of modern mathematical physics .
i : functional analysis _ , new york , academic press , 1972 | we compute haar ensemble averages of ratios of random characteristic polynomials for the classical lie groups @xmath0 , @xmath1 , and @xmath2 . to that end , we start from the clifford - weyl algebra in its canonical realization on the complex @xmath3 of holomorphic differential forms for a @xmath4-vector space @xmath5 . from it
we construct the fock representation of an orthosymplectic lie superalgebra @xmath6 associated to @xmath5 .
particular attention is paid to defining howe s oscillator semigroup and the representation that partially exponentiates the lie algebra representation of @xmath7 . in the process , by pushing the semigroup representation to its boundary and arguing by continuity , we provide a construction of the shale - weil - segal representation of the metaplectic group . to deal with a product of @xmath8 ratios of characteristic polynomials , we let @xmath9 where @xmath10 is equipped with the standard @xmath11-representation , and focus on the subspace @xmath12 of @xmath11-equivariant forms .
by howe duality , this is a highest - weight irreducible representation of the centralizer @xmath13 of @xmath14 in @xmath6 .
we identify the @xmath11-haar expectation of @xmath8 ratios with the character of this @xmath15-representation , which we show to be uniquely determined by analyticity , weyl group invariance , certain weight constraints and a system of differential equations coming from the laplace - casimir invariants of @xmath13 .
we find an explicit solution to the problem posed by all these conditions . in this way
, we prove that the said haar expectations are expressed by a weyl - type character formula for _ all _ integers @xmath16 .
this completes earlier work of conrey , farmer , and zirnbauer for the case of @xmath17 . |
the main goal of this note is to provide a gentle introduction to the theory of feynman - jackson integrals , and to compute an example of such integrals as explicitly as possible . roughly speaking feynman - jackson integrals are the analogue in @xmath0-calculus of feynman integrals .
the name of jackson is included since integration in @xmath0-calculus was introduced by f.h .
jackson , see @xcite and @xcite .
our computations are done in the one dimensional setting .
extending our concepts to higher dimensions , in particular to infinite dimensions , is the main problem in this subject . in principle
such extension is possible but making the notation easy to handle is hard , as the reader will learn by looking at the simple case of one dimensional integrals .
+ the other fundamental task in this subject matter is to find applications of feynman - jackson integrals in physics .
although in this note we focus on the mathematical properties of such integrals , we believe that our formalism will find applications thanks to the following facts : * in this paper our emphasis is on feynman - jackson integrals which rely on the construction of a @xmath0-analogue of the gaussian measure .
given the widespread range of applications in mathematics and physics alike of the gaussian measure , we expect its @xmath0-analogue to gradually find its natural set of applications .
the subject of @xmath0-probability theory and @xmath0-random processes is still in a developing phase , but already a solid step forward has been taken by kupershmidt in @xcite .
* @xmath0-calculus is adapted to work with arbitrary functions while the usual rules of calculus demand certain kind of regularity .
the @xmath0-gaussian measure is likely to find applications in the context of highly non - regular phenomena . *
jackson integrals are given by infinite sums .
cutting off the number of terms appearing in such sums provides a natural regularization method for diverging jackson integrals . *
classically , we think of the continuous as a limit of the discrete .
indeed , in the limit @xmath1 we recover from feynman - jackson integrals the usual feynman integrals .
thus our feynman - jackson integrals provide a new method for computing feynman integrals : first compute the corresponding feynman - jackson integral and then take the limit as @xmath0 goes to @xmath2 . * from a quantum perspective , it is the continuous that should be regarded as an approximation to the fundamental discrete quantum theory . in particular it has been argued , see @xcite , @xcite , that in the quantization of gravity discrete structures will emerge naturally . if that is so , then integration over discrete structures may become a fundamental issue .
our formalism may shed some light as the form that such theory of discrete integration may take .
we emphasize that the calculus itself in @xmath0-calculus is discrete while the variables involved remain continuous .
let us recall some notions of @xmath0-calculus , see @xcite , @xcite,@xcite and @xcite for more information .
fix a real number @xmath3 .
let @xmath4 be a function and @xmath5 be a real number .
the @xmath0-derivative of @xmath6 at @xmath7 is given by @xmath8 for example if @xmath9 then @xmath10_{q}x^{t-1}$ ] where @xmath11_{q}=\frac{q^{t}-1}{q-1}$ ] . + the @xmath0-integral , better known as the jackson integral , of @xmath6 between @xmath12 and @xmath13 of @xmath6 is given by @xmath14 we also define @xmath15 and @xmath16 notice that in the limit @xmath1 the @xmath0-derivative and the @xmath0-integral approach the usual derivative and the riemann integral , respectively .
the @xmath0-analogues of the rules of derivation and integration by parts are @xmath17 @xmath18 the first goal of this note is to describe the @xmath0-analogue of the gaussian measure on @xmath19 .
the moments of the gaussian measure are given by the integrals @xmath20{2\pi}}\int_{-\infty}^{\infty}e^{-\frac{x^2}{2}}x^ndx.\ ] ] a remarkable property of the gaussian measure is that it provides a bridge between measure theory and combinatorics .
indeed , the moments of the gaussian measure are @xmath21{2\pi}}\int_{-\infty}^{\infty}e^{-\frac{x^2}{2}}x^{2n}dx= ( 2n-1)(2n-3) ... 7.5.3.1,\ ] ] and @xmath22{2\pi}}\int_{-\infty}^{\infty}e^{-\frac{x^2}{2}}x^{2n+1}dx= 0.\ ] ] the number @xmath23 is often denoted by @xmath24 and is called the double factorial .
the reader may consult @xcite for a natural generalization of such numbers .
it can be shown that @xmath23 counts the number of pairings on the set @xmath25=\{1,2, ... ,2n\}.$ ] a pairing on @xmath25 $ ] is a partition of @xmath25 $ ] into @xmath26 blocks each of cardinality two .
so for example we have that @xmath27{2\pi}}\int_{-\infty}^{\infty}e^{-\frac{x^2}{2}}x^{4}dx= 3,\ ] ] since as shown in figure [ fig : pochha graphs ] there are @xmath28 pairings on a set with @xmath29 elements . thus we see that the gaussian measure has a clear cut combinatorial meaning .
this simple fact explains the source of graphs in the computation of feynman integrals . in order to define the @xmath0-analogue of the gaussian measure we must find @xmath0-analogues for the objects appearing in the gaussian measure , namely @xmath30{2\pi},\infty , e^{-\frac{x^2}{2 } } , x^n$ ] and @xmath31 .
the lebesgue measure @xmath31 agrees with riemann integration for good functions .
thus it is only natural to replace @xmath31 by jackson integration @xmath32 while the factor @xmath33 remains unchanged , finding the @xmath0-analogue of @xmath34 is actually quite a subtle matter .
first , we must find a @xmath0-analogue for the exponential function @xmath35 which is characterized by the properties @xmath36 and @xmath37 .
so we look for a function @xmath38 such that @xmath39 and @xmath40 a solution to this couple of equations is @xmath41_{q}!},\ ] ] where @xmath42_{q}!=[n]_{q}[n-1]_{q}[n-2]_{q} ...
[2]_q \mbox { \ \ and \ \ } [ n]_{q}= 1 + q + q^2 + q^3 + ... + q^{n-1}.\ ] ] the @xmath0-analogue of the identity @xmath43 is @xmath44 , where @xmath45_{q}!}.\ ] ] once we have obtained @xmath0-analogues for the exponential map and its inverse one might think that it is straightforward to generalized the term @xmath34 of the gaussian integrals .
however , this is not the case and while our first impulse is to try @xmath46_q}}$ ] , the right answer @xcite is to replace @xmath34 by @xmath47_q}}= \sum_{n=0}^{\infty}\frac{(-1)^{n}q^{n(n+1)}x^{2n}}{(1+q)^{n}[n]_{q^{2}}!}.\ ] ] now we consider the integration limits .
it is amusing that whereas the classical gaussian measure is given by an improper integral , its @xmath0-analogue turns out to be a definite integral whose limits depend on @xmath0 and go to ( plus or minus ) infinity as @xmath0 approaches to @xmath2 .
a similar situation occurs with the integral representations of the @xmath0-analogue of the gamma function , see @xcite and @xcite .
indeed , without further motivation we shall take the boundary limits in the gaussian integrals to be @xmath48 and @xmath49 where @xmath50 to find the @xmath0-analogue @xmath51 of the @xmath52 appearing in gaussian integrals we must demand that @xmath51 be such that @xmath53_q}}d_qx = 1.\ ] ] thus @xmath51 is given by @xmath54_q}}d_qx = 2\int_{0}^{\nu } e_{q^{2}}^{\frac{-q^2x^2}{[2]_q}}d_qx = 2(1-q)\nu\sum_{n=0}^{\infty}q^n e_{q^{2}}^{\frac{-q^2(q^n\nu)^2}{[2]_q}},\ ] ] so @xmath55_{q^{2}}!}\ ] ] and interchanging the order of summation we get the identity @xmath56_{q^{2}}!}.\ ] ] since @xmath57 we obtain the amusing identity @xmath58_{q^{2}}!}.\ ] ] we now look for the @xmath0-analogue of identities @xmath59 and @xmath60 for the moments of gaussian integrals .
a key result is that one can show the identities @xmath61_q}}x^{2n}d_{q}x= [ 2n-1]_{q}[2n-3]_{q} ... [7]_{q}[5]_{q}[3]_{q}[1]_q=[2n-1]_q!!,\ ] ] and @xmath62_q}}x^{2n+1}d_{q}x= 0.\ ] ] identity @xmath63 follows from the fact that @xmath64 is an odd function and @xmath65_q}}$ ] is an even function .
identity @xmath66 is proved recursively .
using formula @xmath67 , one shows that @xmath68_q}}x^{2n+2}d_{q}x= \frac{[2n+1]_{q}}{c(q)}\int_{-\nu}^{\nu}e_{q^{2}}^{\frac{-q^2x^2}{[2]_q}}x^{2n}d_{q}x.\ ] ] identities ( [ qg ] ) and ( [ qgc3 ] ) imply identity ( [ qgc1 ] ) .
next we describe a combinatorial interpretation of the number @xmath69_q!!=[2n-1]_{q}[2n-3]_{q} ... [7]_{q}[5]_{q}[3]_{q}[1]_q.\ ] ] an ordered pairing @xmath70 on @xmath25=\{1,2, ... ,2n\}$ ] is a sequence @xmath71 ^ 2)^n$ ] such that * @xmath72 . * @xmath73 * @xmath74=\bigsqcup_{i=1}^{n } \{a_i , b_i\}.}$ ] we denote by @xmath75 $ ] the set of ordered pairings on @xmath25 $ ] .
we are going to define a weight @xmath76 for each @xmath77 $ ] .
let us introduce the following notation * @xmath78 : a_i < j < b_i\}$ ] for all @xmath79 .
* @xmath80 .
the weight of @xmath70 is defined by the rule @xmath81 figure [ fig:1 ] and figure [ fig:2 ] below show a couple of examples of pairings together with the corresponding weights .[fig:1],width=288 ] .[fig:2],width=288 ] using this language we can state the following result which is proved by induction @xmath82_q!!= \sum_{p \in p[2n]}w(p).\ ] ] notice that as @xmath83 we recover from @xmath84 the well - known identity @xmath85\}|.\ ] ] for example @xmath86_q = 1+q+q^{2}$ ] which agrees with the sum of the weights of the three pairings on @xmath87 $ ] shown in figure @xmath88 . _
q.$ ] [ fig : pochha graphs],width=528 ] in conclusion we have proved that @xmath89_q}}x^{2n}d_{q}x= \sum_{p \in p[2n]}w(p).\ ] ]
we want to study the computation of feynman - jackson integrals of the form @xmath90_q}+g \frac{x^3}{[3]_q ! } } } d_qx.\ ] ] as usual in the theory of feynman integrals we regard @xmath91 as a formal power series in the formal variable @xmath92 , i.e. , @xmath93 $ ] .
the first step in the computation of a feynman - jackson integrals is to reduce it to the computation of a countable number of gaussian - jackson integrals .
this step is carry out with the help of following formula proved in @xcite @xmath94 where @xmath95_{q^{2}}![c - k]_{q^{2}}!}}$ ] . making the substitutions @xmath96_q}}$ ] and + @xmath97_q!}}$ ]
we obtain @xmath98_q}+g\frac{x^{3}}{[3]_q!}}= e_{q^2}^{-\frac{q^{2}x^{2}}{[2]_q}}\sum_{c , d\geq 0 } \lambda_{c , d } \frac{(-1)^{c}q^{2c } x^{2c } x^{3d } } { [ 2]_q^{c}([3]_q!)^{d}}g^{d}=e_{q^2}^{-\frac{q^{2}x^{2}}{[2]_q}}\sum_{c , d\geq 0 } \lambda_{c , d } \frac{(-1)^{c}q^{2c } x^{2c+3d } } { [ 2]_q^{c}([3]_q!)^{d}}g^{d } } , \ ] ] so we get @xmath99_q}+g\frac{x^{3}}{[3]_q!}}=\sum_{c , d , k } \frac{(-1)^{2c - k } { d+k\choose k } q^{(d+k)(d+k-1)+2c}}{[2]_q^c([3]_q!)^{d}[d+k]_{q^{2}}![c - k]_{q^{2 } } ! } x^{2c+3d } g^d } .\ ] ] if we @xmath0-integrate both sides of equation ( [ 26 ] ) we get @xmath100_q}+g\frac{x^{3}}{[3]_q ! } } } d_qx&=&{\displaystyle\sum_{c , d , k } \frac{(-1)^{2c - k } { 2d+k\choose k } q^{(2d+k)(2d+k-1)+2c } [ 2c+6d-1]_q!!}{[2]_q^c([3]_q!)^{2d}[d+k]_{q^2 } ! [ c - k]_{q^2}!}g^{2d } } \\
\mbox { } & \mbox { } & \mbox { } \nonumber \\ & = & { \displaystyle \sum_{c , d , k } \frac{(-1)^{2c - k } { 2d+k\choose k } q^{(2d+k)(2d+k-1)+2c } [ 2c+6d-1]_q!!}{[2]_q^c([3]_q!)^{2d}[2d+k]_{q^2 } ! [ c - k]_{q^2}!}g^{2d}}\label{e2}\end{aligned}\ ] ] the second step in the computation of a feynman integral is to write such as integral as a sum of a countable number of contributions , where each summand is naturally associated to certain kind of graph , see @xcite .
so we want to understand the right hand side of the equation ( [ e2 ] ) in terms of a summation of the weights of an appropriated set of isomorphism classes of graphs .
consider the category @xmath101 whose objects are planar graphs @xmath102 such that 1 .
@xmath103 , where @xmath104 and @xmath105 .
2 . @xmath106 .
@xmath107 this data must satisfy the following axioms : 1 .
@xmath108 and @xmath109 .
2 . @xmath110 and if @xmath111 then @xmath112 for @xmath113 .
3 . if @xmath114 , then @xmath115 or @xmath116 for some @xmath117 or@xmath118 .
4 . if @xmath119 then @xmath120 . 5 . @xmath121 .
+ to each graph @xmath122 as above we associate two polynomials in @xmath0 , @xmath123 and @xmath124 , given respectively , by 1 . @xmath125 .
above @xmath70 is the natural pairing induced by @xmath126 on the flags of @xmath127 , i.e.,the set @xmath128 .
@xmath129 is the weight of @xmath70 as given by formula ( [ peso ] ) .
2 . @xmath130_q^{|v_1|}([3]_q!)^{|v_2|}[|v_2|+|e_2|]_{q^2}![|v_1|-|e_2|]_{q^2}!$ ] .
figure [ fig : wq ] and figure [ fig : wq1 ] show examples of graphs in @xmath101 together with the corresponding polynomials @xmath131 and @xmath132 . with @xmath133 and @xmath134_q^4[3]_q^3$].[fig :
wq],width=192 ] with @xmath135 and @xmath134_q^2([3]_q!)^2[4]_q!$ ] .
[ fig : wq1],width=192 ] now we are ready to state the main result of this paper , namely , we write the integral ( [ int24 ] ) as a sum over the weights of graphs as follows @xmath136_q}+g\frac{x^{3}}{[3]_q ! } } } d_qx=\sum_{(v , e , b)\in{\mathbf{graph}_q^3}/\sim } \frac{\omega_q(\gamma)}{a_q(\gamma)}g^{|v_2| } } \ ] ] where the sum runs over all isomorphisms classes of graphs in @xmath101 .
identity ( [ ult ] ) follows directly from identity ( [ e2 ] ) and the definitions above .
we thank bernardo uribe and the math department at uniandes where this work was finished .
rafael daz and eddy pariguan , _ feynman - jackson integrals _ , journal of nonlinear mathematical physics 13 ( 2006 ) , no . 3 , 365376 .
rafael daz and carolina teruel , _ q , k - generalized gamma and beta functions _ ,
journal of nonlinear mathematical physics 12 ( 2005 ) , no . 1 , 118134 . h.t .
koelink and koornwinder , _ q - special functions , in deformation theory and quantum groups with applications to mathematical physics _ , edited by murray gerstenhaber and jim stasheff , amer .
math soc 134 ( 1992 ) , 141142 .
rovelli carlo and smolin lee , _ discreteness of area and volume in quantum gravity _ , nuclear phys , b442 , 3 , ( 1995 ) , 593 - 619 .
smolin , lee , _ the physics of spin networks _ , the geometric universe , oxford univ . press ,
oxford , ( 1998 ) , 299 - 304 .
a. de sole and v. kac , _ on integral representations of q - gamma and q - beta functions_,atti .
lincei cl .
fis . mat .
lincei ( 9 ) . mat .
16 , 41 , ( 2005 ) , 11 - 29 . | we review the construction of a @xmath0-analogue of the gaussian measure .
we apply that construction to obtain a @xmath0-analogue of feynman integrals and to compute explicitly an example of such integrals . |
in a series of earlier communications @xcite we have proposed the augmented space recursion ( asr ) as an efficient computational technique for the calculation of configuration averaged electronic properties of random binary alloys .
the method is a combination of the augmented space formalism @xcite and the recursion method of haydock @xcite .
when coupled with the local density functional approximation based tight - binding linearized muffin - tin orbital @xcite it provides a first - principles determination of electronic structure and total energy calculations for disordered alloys .
effects such as short - ranged ordering and local lattice distortions due to size mismatch of the constituent atoms can very easily be incorporated in the methodology @xcite . recently
the method has been applied to study dispersion and line widths for phonons in disordered alloys . since recursion can take into account disorder in beyond nearest - neighbour force constants as well as large environmental effects , this application was a step forward in the theory of phonons in disordered alloys @xcite .
it has been pointed out repeatedly in the above publications that the asr with its terminator approximation @xcite goes beyond the standard single - site mean field theories and accurately takes off - diagonal disorder into account .
a number of attempts have been made for practical implementation of mean - field theories beyond the single site approximation .
one of the most sophisticated developments among the mean - field theories , is the travelling cluster approximation of gray and kaplan @xcite and its latest version : the itinerant coherent potential approximation ( icpa ) @xcite , where a small number of classes of selected excitations in augmented space are preserved . in spite of the simplification
so introduced , the resulting equations are still formidable enough to discourage further generalization to larger clusters beyond the nearest neighbour . in the cluster cpa proposed by mookerjee @xcite , in which one retains disorder fluctuations only from a chosen cluster and replaces the rest of the system by an effective medium , one is restricted by the limitation of solving number of coupled self - consistent equations .
the number of equations increases with the size of chosen cluster , which again discouraged extension to larger clusters beyond the nearest neighbour .
this brings us back to the attempt at the implementation of the asr for large environment fluctuations , implying accurate evaluation of a large sequence of recursion coefficients .
we shall discuss this in some detail in the following section .
the asr , in its practical implementation to real alloy systems , has several practical problems
. the formalism maps a disordered hamiltonian , described in a hilbert space spanned by a countable basis in which preferably its representation is sparse , onto an ordered hamiltonian in an enlarged space .
this enlarged or _ augmented _ space is constructed by taking the outer product of the configuration space of the random variables of the disordered hamiltonian together with original hilbert space in which the disordered hamiltonian was described .
the main difficulty in the implementation of recursion in this augmented space is its enormous rank .
a system of @xmath1 sites with binary distribution has an augmented space of rank @xmath2 .
this enormous rank of the augmented space and its description on the computer is what had discouraged earlier workers in implementing the asr , although the augmented space formalism was first proposed about thirty years ago @xcite .
the second problem was the recursive propagation of errors .
very small errors during recursion ( inevitable in numerical implementations ) tends to add up leading to loss of orthogonality in the recursion basis .
this leads to appearance of ghost states .
haydock and te @xcite has overcome this difficulty in proposing the _
dynamical recursion_. we refer the reader to their paper for details .
in this communication we shall describe the reduction of the effective rank of the augmented space on which the recursion is carried out , using the local point - group symmetries in both the hilbert space and the configuration space .
such reduction is well known in various manipulations in reciprocal space for ordered systems .
however , such a reduction has not been implemented in real space recursion , although it had been proposed by gallagher @xcite quite some time ago .
it had been shown by gallagher that if we start recursion with a state belonging to an irreducible subspace of a hilbert space , subsequent recursion always stays within this subspace .
this allows us significant reduction in rank of the required subspace and makes this method practically feasible .
further , this method with its terminator approximation , retains the herglotz properties of the configuration averaged green function .
the recursion method addresses inversions of infinite matrices @xcite .
once a sparse representation of an operator in hilbert space , @xmath3 , is known in a countable basis , the recursion method obtains an alternative basis in which the operator becomes tri - diagonal .
this basis and the representations of the operator in it are found recursively starting from @xmath4 : @xmath5 where , @xmath6 for n@xmath7 2 . in this basis
the representation of @xmath3 is tridiagonal : @xmath8^{1/2}}\ = \ \beta_n \enskip ; \enskip \langle u_n\vert \tilde{h}\vert u_{m}\rangle \ = \ 0 \quad m\geq n+2\end{aligned}\ ] ] and the averaged green function can be written as a continued fraction : @xmath9 the asymptotic part of the continued fraction is obtained from the initial set of coefficients @xmath10 for @xmath11 , using the idea of _ terminators _ @xcite . for small values of @xmath1
we have large inaccuracies and the more the structure in the spectral density of @xmath3 the larger is the @xmath1 needed to maintain the window of accuracy required by us . in earlier communications we have described how to deal with random binary alloys @xcite within the framework of the tight - binding linearized muffin - tin orbital method ( tb - lmto ) @xcite and using the augmented space formalism ( asf ) .
we refer the readers to the seminal papers mentioned earlier for details . here
we shall introduce only the salient features which will be required by us subsequently in this communication .
we begin by setting up a muffin - tin potential with centres at the atomic sites @xmath12 on a lattice .
next , we inflate the muffin - tins into atomic spheres and start from a most - localized tb - lmto representation of the hamiltonian : @xmath13 the _ potential parameters _ , @xmath14 and @xmath15 , are diagonal matrices in the angular momentum indeces , and @xmath16 here @xmath17 is a composite label @xmath18 for the angular momentum quantum numbers .
@xmath19 is the random site - occupation variable which takes values 0 or 1 with probability @xmath20 or @xmath21 @xmath22 respectively , depending upon whether the muffin - tin labelled by @xmath12 is occupied by @xmath23 or @xmath24-type of atom .
the @xmath25 and @xmath26 are projection and transfer operators in the hilbert space @xmath27 spanned by the tight - binding basis @xmath28 .
we can associate the random variable @xmath19 with an operator @xmath29 whose eigenvalues correspond to the observed values of @xmath19 , and whose corresponding eigenvectors , @xmath30 are the _ state vectors _ of the variable .
these state vectors of the set of @xmath1 random variables @xmath19 of rank @xmath31 span a space called configuration space @xmath32 with configurations of the kind @xmath33 where @xmath34 if we define the configuration @xmath35 as the _ reference _ configuration , then any other configuration may be uniquely labelled by the _
cardinality sequence _ : @xmath36 , which is the sequence of positions where we have a @xmath37 configuration .
the _ cardinality sequence _ of the _ reference _ configuration is the null sequence @xmath38 the augmented space theorem states that a(\{n_r } ) where , @xmath39 ( @xmath40 ) is the spectral density of the self - adjoint operator @xmath29 , which is such that the probability density of @xmath19 is given by : @xmath41 @xmath29 is an operator in the space of configurations @xmath42 of the variable @xmath19 .
this is of rank 2 and is spanned by the _ states _ @xmath43 .
@xmath44 the expended hamiltonian @xmath45 in the augmented space is constructed by replacing all the random variables @xmath19 by the corresponding operators @xmath29 .
it is an operator in the augmented space @xmath46 = @xmath47 .
so the asf maps a disordered hamiltonian described in a hilbert space @xmath27 onto an ordered hamiltonian in an enlarged space @xmath46 , where the space @xmath46 is constructed by augmenting the configuration space @xmath48 of the random variables of the disordered hamiltonian together with the hilbert space @xmath27 of the disordered hamiltonian .
the configuration space @xmath32 is of rank 2@xmath49 if there are n muffin - tin spheres in the system .
we may rewrite the expression for the configuration average of the green operator as : @xmath50 where , @xmath51\enskip { \cal p}_r \otimes { \cal p}_{l } \otimes { \cal i } \\ \bhat & = & \sum_{rl } b\left[\rule{0mm}{4 mm } ( e - c_l)/\delta_l \right ] \enskip { \cal p}_{r } \otimes { \cal p}_{l } \otimes { \cal p}_{\downarrow}^r \\ \chat & = & \sum_{rl } c\left[\rule{0mm}{4 mm } ( e - c_l)/\delta_l \right ] \enskip { \cal p}_{r } \otimes { \cal p}_{l}\otimes \left\ { { \cal t}^r_{\uparrow\downarrow } + { \cal t}^r_{\downarrow\uparrow } \right\ } \\ \shat & = & \sum_{rl } \sum_{r'l ' } s_{rl , r'l ' } \enskip { \cal t}_{rr ' } \otimes { \cal t}_{ll ' } \otimes { \cal i } \\ \phantom{x } & & \\ \dhat & = & \sum_{rl } a\left ( \delta^{1/2}_l \right)\enskip { \cal p}_r \otimes { \cal p}_{l } \otimes { \cal i } + \sum_{rl } b\left ( \delta^{1/2}_l \right ) \enskip { \cal p}_{r}\otimes { \cal p}_{l}\otimes { \cal p}_{\downarrow}^r \ldots \\ & & + \sum_{rl } c\left ( \delta^{1/2}_l \right ) \quad { \cal p}_{r } \otimes { \cal p}_{l}\otimes \left\ { { \cal t}^r_{\uparrow\downarrow } + { \cal t}^r_{\downarrow\uparrow } \right\ } \end{aligned}\ ] ] and @xmath52 since , @xmath53 the ket @xmath54 is not normalized .
we first write the above in terms of a normalized ket @xmath55 = @xmath56^{-1/2 } \vert 1 \ } $ ] .
we now have : @xmath57 where , @xmath58 / a\left ( 1/\delta_{l } \right)\right\}\enskip { \cal p}_{r } \otimes { \cal p}_l\otimes { \cal p}_{\downarrow}^{r } \nonumber\\ \fl \chat^{\prime } = \sum_{rl } \left\{c\left [ ( c_l - e)/\delta_l\right ] / a\left ( 1/\delta_l \right)\right\}\enskip { \cal p}_{r } \otimes { \cal p}_{l } \otimes \left\ { { \cal t}^{r}_{\uparrow\downarrow } + { \cal t}^{r}_{\downarrow\uparrow } \right\}\nonumber \\
\fl \shat^{\prime } = \sum_{rl } \sum_{r'l'}\left[a\left ( 1/\delta_l \right)\right]^{-1/2 } \
s_{rl , r'l ' } \ \left [ a\left ( 1/\delta_{l ' } \right)\right]^{-1/2 } { \cal t}_{rr'}\otimes{\cal t}_{ll ' } \otimes { \mathaccent " 7e { \cal i}}\nonumber\\ \label{kasr2 } \end{aligned}\ ] ] this equation is now exactly in the form in which the recursion method may now be applied .
the computational burden is considerably reduced due to this diagonal formulation , the recursion now becomes energy dependent as is clear from the form of the effective hamiltonian as shown in and discussed in @xcite .
the recursion formalism of the ordered hamiltonian was free of this constraint .
this energy dependence makes the recursion technique computationally unsuitable because to obtain the green functions we have to carry out recursion per energy point of interest .
this problem has been tackled using _ seed recursion technique _ @xcite .
the idea is to choose a few seed points across the energy spectrum uniformly , carry out recursion over those points and then interpolate the values of coefficients across the band . in this way
one may reduce computation time .
for example , if one is interested in an energy spectrum of 200 points , in the bare diagonal formulation recursion has to be carried out at all the 200 points but in the seed recursion technique one needs to perform recursions only at 10 - 15 points .
the whole idea stems from the fact that in most of the cases of binary alloys , it is seen that the recursion coefficients @xmath59 and @xmath60 vary quite weakly across the energy spectrum . at this point
we note that the above expression for the averaged @xmath61 is _ exact_.
we mentioned earlier that recursion on the augmented space is not computationally feasible because of its large rank . for a binary alloy with @xmath1 sites and with only @xmath62-orbitals on them ,
the rank of the augmented space is @xmath2 .
implementing recursion on this huge space for a sufficient number of steps to ensure accuracy is often not feasible on available computers .
however , if we exploit the symmetry both of the underlying lattice in real space and of the configuration space ( which arises due to the homogeneity of the disorder and arrangements of atoms on the underlying lattice ) , the rank of the irreducible part of the augmented space in which the recursion is effectively carried out becomes tractable .
the conceptual advantages in asf include apart from analyticity , translational and underlying lattice symmetries automatically built into the augmented space hamiltonian .
this allows us to involve the idea of utilizing symmetry operations present in both the real and configuration spaces , in the context of recursion method , reducing the rank of hamiltonian drastically and making the implementation of asf feasible .
since the augmented space recursion essentially retains all the properties of real space recursion but is described in a much enlarged space , it will be useful to consider symmetry operations in real space recursion first and then to consider those in the full augmented space . during the recursion ,
the basis member @xmath63 is generated from the starting state @xmath64 by repeated application of the hamiltonian .
if the starting state belongs to an irreducible sub - space of the hilbert space then all subsequent states generated from recursion will belong to the same irreducible sub - space .
physically , we may understand this as follows : the recursion states @xmath63 carry information of distant environment of the starting state .
if the hamiltonian is nearest - neighbour _ only _ , then the state @xmath65 , which arises by the application of the hamiltonian on @xmath66 is a combination of states in the nearest shell with which @xmath64 couple via the hamiltonian .
similarly , @xmath63 is a combination of n - th neighbour shell with which @xmath64 is coupled via the hamiltonian .
if @xmath67 is a point group symmetry of the hamiltonian , then all @xmath68-th neighbour - shell states which are related to one another through the symmetry operator must have equal coupling to @xmath64 .
hence , it is useful to consider among the @xmath68-th neighbour - shell states of which @xmath63 is a linear combination , only those belonging to the irreducible subspace and redefine the hamiltonian operation .
= 3.0 in as an example , take a nearest - neighbour @xmath62-state anderson model on a square lattice , with a binary distribution of its diagonal elements .
we shall label the tight - binding basis by the position of the lattice points in units of the lattice constant : e.g. @xmath69 where @xmath70 are integers @xmath71 .
the starting state @xmath72 belongs to the one dimensional representation of the point group of a square lattice .
this state then couples with linear combinations of states on neighbour shells which are symmetric under square rotations : @xmath73 the shows the grouping together of sites with the local symmetry of the square lattice on the first three nearest neighbour shells of the central site .
the first and second groups in are coloured gray in and the last group coloured black .
the labels on the groups ( shown on the left sides of the equations ) are confined _ only _ to the upper right quadrant of the lattice .
if we go up to @xmath1 shells ( for large @xmath1 ) there are about @xmath74 states in the diamond shaped nearest - neighbour cluster
. however , there are only @xmath75 states with square symmetry .
so within this reduction we can work only in 1/8-th of the lattice , provided we incorporate proper weights to the states to reproduce the correct matrix elements .
if @xmath76 and @xmath77 are two states coupled to each other via the hamiltonian , and both belong to the same irreducible subspace , and if @xmath78 are states obtained by operating on @xmath76 by the symmetry group operations of the real space lattice , then @xmath79 is called the _ weight _ associated with the state labelled by @xmath12 .
if we wish to retain only the states in the irreducible subspace and throw out the others and yet obtain the same results , we have to redefine the hamiltonian matrix elements as follows : @xmath80 = 5.4 in = 6.5 in = 6.in in the above prescription , the new irreducible basis introduced by us , reflects only the symmetry of the underlying lattice and holds good for a model system which has @xmath62-like orbitals only . but for a real system , the tb - lmto minimal basis contains members with other symmetries as well .
for example , in a cubic lattice with a @xmath81 minimal basis , we have basis members with @xmath62 , @xmath82 , @xmath83 and @xmath84 symmetries .
the symmetry of the orbitals is reflected in the two - centered slater - koster integrals and this prohibits overlap integrals at certain positions , called _ symmetric positions _ with respect to the overlapping orbitals .
a few of these symmetric positions for a simple cubic lattice are shown in figure [ fig2 ] .
for the sake of clarity we have shown only the projections on @xmath85 plane .
we have indicated the positive and negative parts of the orbital lobes by different shades . from ( b )
it is easy to argue that since the hamiltonian is spherically symmetric and the product @xmath86 is positive in the upper right quadrant and negative in the upper left quadrant : @xmath87 .
the same is true for @xmath88 .
similarly , @xmath89 . on the other hand , @xmath90 .
element & expression & factor & to be zero + @xmath91 & @xmath92 & @xmath93 & + @xmath94 & @xmath95 @xmath96 & @xmath97 & = 0 + @xmath98 & ' '' '' @xmath99 @xmath100 & @xmath101 & =0 or = 0 + @xmath102 & @xmath103(@xmath104-@xmath104 ) @xmath100 & @xmath105 & = + @xmath106 & ' '' '' [ @xmath107 - 1/2(@xmath104+@xmath104 ) ] @xmath100 & @xmath108 & @xmath109 + @xmath110 & @xmath111@xmath104 @xmath112+(1-@xmath104 ) @xmath113 & @xmath114 & + @xmath115 & @xmath111 ( @xmath116 ) & @xmath117 & =0 or = 0 + @xmath118 & @xmath111 [ @xmath119@xmath104 @xmath120 + ( 1 - 2@xmath104 ) @xmath121 & @xmath122 & = 0 + @xmath123 & @xmath124 ( @xmath120 -2 @xmath125 ) & @xmath126 & =0 or = 0 @xmath127 + & & & @xmath127 or = 0 + @xmath128 & ' '' '' [ @xmath129 ( @xmath104-@xmath104 ) @xmath120 + ( 1-@xmath104+@xmath104 ) @xmath121 & @xmath130 & =0 + @xmath131 & ' '' '' ( @xmath104-@xmath104)[@xmath129 @xmath120 - @xmath121 & @xmath132 & = 0 ot = + @xmath133 & ' '' '' @xmath134[@xmath104-(@xmath104+@xmath104)/2 ] @xmath120-@xmath119 @xmath104 @xmath135 & @xmath136 & =0 + @xmath137 & ' '' '' 3@xmath104@xmath104 @xmath138+(@xmath104+@xmath104 - 4@xmath104@xmath104 ) @xmath139 & @xmath140 & + & @xmath127 + ( @xmath104+@xmath104@xmath104 ) @xmath141 & & + @xmath142 & ' '' '' @xmath134 3@xmath104 @xmath138+(1 - 4@xmath104 ) @xmath143 + ( @xmath104 - 1 ) @xmath141 & @xmath144 & =0 or = 0 + @xmath145 & ' '' '' ( @xmath104-@xmath104 ) @xmath134 ( 3/2)@xmath138 - 2 @xmath146 + ( 1/2 ) @xmath147 & @xmath148 & =0 or = 0 @xmath127 + & & & @xmath127 or = + @xmath149 & ' '' '' @xmath134(3/2)(@xmath104-@xmath104 ) @xmath138-[1 + 2(@xmath104-@xmath104 ) ] @xmath146 @xmath127 & @xmath150 & = 0 or = 0 + & @xmath127 + [ 1 + 1/2(@xmath104-@xmath104 ) ] @xmath147 & & + @xmath151 & ' '' '' @xmath119 @xmath134 [ @xmath104 - 1/2(@xmath104+@xmath104 ) ] @xmath138 - 2 @xmath104 @xmath146 @xmath127 & @xmath152 & =0 or = 0 + & @xmath127 + 1/2(1+@xmath104 ) @xmath147 & & + @xmath153 & ' '' '' 3/4(@xmath104-@xmath104)@xmath104 @xmath138 + [ @xmath104+@xmath104-(@xmath104-@xmath104)@xmath104]@xmath146@xmath127 & @xmath155 & + & @xmath127 + [ @xmath104 + 1/4(@xmath104-@xmath104)@xmath104 ] @xmath141 & & + @xmath156 & ' '' '' @xmath119 ( @xmath104-@xmath104)@xmath134 [ @xmath104/2 - 1/4(@xmath104+@xmath104 ) ] @xmath138 -@xmath104
@xmath146@xmath127 & @xmath157 & = + & @xmath127 + 1/4(1+@xmath104 ) @xmath141 & & + @xmath158 & ' '' '' [ @xmath104 - 1/2(@xmath104+@xmath104)]@xmath104 @xmath138 + 3@xmath104(@xmath104+@xmath104)@xmath146@xmath127 & @xmath159 & + & @xmath127 + 3/4(@xmath104+@xmath104)@xmath104 @xmath141 & & + the above is an illustration , detailed arguments for the orbital symmetry has been discussed by harrison @xcite .
the provides the conditions for obtaining the orbital based symmetry factor @xmath160 for a lattice with cubic symmetry .
we shall describe the direction of @xmath161 by the direction cosines ( @xmath162 .
the table gives the details for the calculation of the symmetry factor @xmath163 on a cubic lattice .
a look at the figures [ fig2 ] and [ fig3 ] shows the reason for introduction of this factor . for
the @xmath62-orbital overlaps the symmetry factor is always 1 ( figure 1(a ) ) .
all six neighbours in the cubic lattice in the directions ( 100),(@xmath16400),(010),(0@xmath1640 ) , ( 001 ) and ( 00@xmath164 ) are equivalent with the same overlap integrals @xmath92 ( only four are shown in the @xmath165 plane )
. we may then reduce the lattice and retain only the neighbour in the ( 100 ) direction and scale the overlap @xmath92 by the appropriate weight ( i.e. 2 in this case ) as discussed earlier . for the @xmath166
, however , the overlaps in the ( 010 ) , ( 0@xmath1640 ) , ( 001 ) and ( 00@xmath164 ) are zero .
this can be deduced immediately by noticing that the overlap products in these directions are positive for @xmath167 and negative for @xmath168 . on reduction , although they are related to one another by the symmetry operations of the cubic lattice , we can not club the neighbours in the ( 100 ) and ( @xmath16900 ) direction with those in the ( 010),(0@xmath1640),(001 ) and ( 00 @xmath164 ) and will have to put these overlaps to be zero by introducing the symmetry factor @xmath163 .
@xmath170 where @xmath163 is 0 if @xmath171 is a symmetric position of @xmath12 with respect to @xmath17 and @xmath172 , otherwise it is 1 .
illustrates some of the overlaps involving the @xmath173 states with @xmath83 and @xmath84 symmetries . with this simple reduction procedure the real space part
can be reduced to 1/8-th , keeping only the sites in the ( @xmath174 ) octant and suitably renormalizing the hamiltonian matrix elements as described earlier .
= 5 in -1 cm = 5 in -1 cm = 4 in we still have not exhausted all the symmetries in the full augmented space .
as discussed earlier , this space is a direct product of the real ( lattice - orbital ) space and the configuration space which are disjoint . as a consequence
the symmetry operations apply independently to each of them .
since the disorder is homogeneous , the cardinality sequence in configuration space itself has the symmetry of the underlying lattice . to see this ,
let us look at the where we show a part of a cubic lattice where the central site is occupied by a configuration labelled by @xmath175 , while two of the six nearest neighbours are occupied by configurations labelled by @xmath37 , and four of them by @xmath175s .
we note that the twelve configurations in the first three rows of the figure , where the two @xmath37s sit at distances @xmath176 times the lattice constant , are related to one another by the symmetry operations of the cubic lattice .
for example , the second to the fourth configurations on the top row can be obtained from the first by the rotations @xmath177 , @xmath178 and @xmath179 respectively .
the configurations are described by _
cardinality sequences _ ( as described earlier ) .
the cardinality sequences for the four configurations on the top row of are : @xmath180 , @xmath181 , @xmath182 and @xmath183 . from the figure it is easy to see that :
@xmath184 = 3.5in=2.5 in this equivalence of the configurations on the lattice is quite independent of the symmetries of the hamiltonian in real space discussed earlier .
thus , in augmented space , equivalent states are @xmath185 and the set @xmath186 for all different symmetry operators @xmath187 of the underlying lattice .
again the symmetry of the orbitals also rules out the operation of the hamiltonian at certain symmetric positions discussed earlier .
= 4.0 in = 2.5 in .comparison between system size and cpu time ( in seconds ) taken for recursion on a p4/256 machine for a full fcc lattice and the reduced lattice in real space . [ cols="^,^,^,^,^,^,^ " , ] [ tab3 ] for the case of the ordered fcc lattice , the shows how the size of the map increases as we increase the number of nearest neighbour shells from a starting site , both with and without reduction .
table 2 shows the details of the cpu time and storage space reduction for recursion after applying the symmetry reduction .
we have carried out calculations both on a simple @xmath62-state system on a fcc lattice , as well as for @xmath188 ( with @xmath81 minimal tb - lmto basis ) also on a fcc lattice . for the calculation of a disordered binary alloy on a fcc lattice ,
the shows the enormous decrease in the size of the augmented space map after application of symmetry reduction for a seven nearest neighbour map on a fcc lattice .
tabulates the reduction in storage and cpu time for a 7 shell , 11 step recursion in augmented space carried over 15 seed points using tb - lmto potential parameters and structure matrix to built the hamiltonian .
the power of symmetry reduction of storage space is more evident in this example . such reduction will allow us to stretch our nearest neighbour map up to 9 - 10 shells , stepping up our accuracy .
the indicates the reduction in cpu time as we increase the number of recursion steps .
further , since the number of sites in the map decrease , the number of individual applications of the @xmath189 hamiltonian also decrease significantly , as do the number of operations involved in taking various inner products .
this will reduce the inherent cumulative error of the recursion technique and lessen the probability of the appearance of _ ghost bands _ which often plague recursion calculations .
we have shown that recursion calculations can be carried out much faster and for many more recursive steps exactly , if we perform the recursion on a subspace of the original augmented space reduced by using the symmetries of both the underlying lattice and random configurations on the lattice .
this will allow us to obtain results for disordered binary alloys with enhanced accuracy required for first - principles , self - consistent , density functional based calculations . in this communication
we have described the details of the implementation of this symmetry reduction and the modifications required in the standard recursion method .
we propose using the symmetry reduced version of the augmented space recursion in our future work on disordered alloys .
99 saha t. , dasgupta i and mookerjee a. , 1994 , l245 saha t. , dasgupta i. and mookerjee a. , 1996 1979 mookerjee a. , 1973 l205 ; 1973 1340 haydock r. , heine v. and kelly m.j .
, 1972 2845 andersen o. k. , 1975 b * 12 * 3060 jepsen o. and andersen o. k. , 1971 1763 mookerjee a. and prasad r. , 1993 17724 saha t. , dasgupta i. and mookerjee a. , 1994 13267 saha t. and mookerjee a.,1995 2915 alam a. and mookerjee a. , 2003 ( to be published ) ; cond - mat/0307697 beer n. and pettifor d. g. ( 1982 ) , in _ the electronic structure of complex systems _ , ed .
p. phariseau and w. m. temmerman , nato asi series b , v.113 , p 769 lucini m.u . and nex c.m.m . , 1987 , 3125 kaplan t. and gray l.j . , 1976 l303 , l483 kaplan t. and gray l.j . , 1977 3260 ghosh s. , leath p.l . , and cohen m. h. , 2002 , 214206 kumar v. , mookerjee a. and srivastava v.k . , 1982
, 1939 srivastava v.k . , choudhry v. and mookerjee a. , 1983 4555 haydock r. and te r.l . ,
1995 _ computer phys . communication _ * 90 * 81 gallagher j. ( 1978 ) , _ ph .
d. thesis , _ university of cambridge , u. k. harrison w.a .
, 1999 _ elementary electronic structure _ ( world scientific , singapore ) p 546 . biswas p. sanyal b. , mookerjee a. , huda a. , halder a. and ahmed m. , 1997 _ int
. phys . _ * b 11 * 3715 ghosh s. , das n. and mookerjee a. , 1999 _ int . j. mod .
_ * b 21 * 723 haydock r. , _ solid state physics _ ed .
h. ehrenreich , f. seitz and d. turnbull ( academic press , new york ) ( 1980 ) | we present here an efficient method which systematically reduces the rank of the augmented space and thereby helps to implement augmented space recursion for any real calculation .
our method is based on the symmetry of the hamiltonian in the augmented space and keeping recursion basis vectors in the irreducible subspace of the hilbert space .
@xmath0 # 1 2\ { r } # 1 # 1 |
the problem of biological aging has attracted much attention in recent years . based on the data of human demography and experiments of other living organisms , many important phenomena of longevity have been found @xcite .
for instance , the gompertz law was observed for intermediate ages , that is , the mortality function increases exponentially with age , while at old ages the mortality was found to decelerate or level off , and even decline for some organisms like flies , worms , and yeast @xcite . to reproduce and explain these phenomena , various models of senescence
have been proposed , with genetic or nongenetic mechanisms @xcite . among them , the one widely used by physicists is the penna model @xcite , where one computer word is used to represent the inherited genome of one individual and each bit of the word corresponds to one age of the individual lifetime .
a bit set to one represents a deleterious mutation and the suffering from an inherited disease from this age on , and the individual will die if the accumulation of these set bits exceeds a threshold .
although the penna model has been well applied to many problems related to biological aging @xcite , there exists an important flaw in this model as pointed out by pletcher and neuhauser very recently @xcite .
that is , the model predicts that for a genetically identical population all individuals have their genetic death at the same age , but this is inconsistent with the experimental results @xcite which also exhibited the exponential gompertz law and the deceleration of the old age mortality for the genetically homogeneous case .
thus , a more complicated model has been proposed @xcite . in this paper
we develop a simpler stochastic model bridging the gap between the standard ( deterministic ) penna model and the pletcher - neuhauser approach .
the simulations and analytic results of this model are shown to agree with some features of the biological aging , e.g. , the exponential increase of the mortality function and the deceleration at advanced ages , and the flaw of the penna model mentioned above can be avoided .
as in the standard penna model , here the genome of each individual is characterized by a string ( computer word ) of 32 bits , and each bit is expressed as a particular age in the life of the individual .
a bit @xmath0 is set to @xmath1 if it represents a deleterious mutation , and from this age @xmath0 on this bit will continuously affect the survival probability of the individual .
that is , at age @xmath2 ( @xmath3 ) the death probability contributed by the mutated bit @xmath0 is @xmath4 , with the corresponding survival probability @xmath5 . otherwise , this bit is set to zero and has no effect on death .
thus our assumptions are very different from an earlier `` fermi '' function in another stochastic penna model @xcite .
the individual s survival probability @xmath6 up to age @xmath2 is the product of the contributions from all the bits before @xmath2 : @xmath7 where @xmath8 or @xmath9 ( @xmath10 ) represents the @xmath0th bit . with the form of @xmath4
, one can obtain the mortality function by simulation or analytical work . in this work
we simply assume that @xmath11 with the constant @xmath12 and the limit @xmath13 , which means that the contribution of death probability from bit @xmath0 ( if set to @xmath1 ) is assumed to increase linearly with the age .
the other forms of @xmath4 , such as the exponential and the square root forms , have been tried , and we have also simulated the other probabilistic penna model with fermi function @xcite .
although some phenomena for the genetically heterogeneous steady - state population can be reproduced , they can not give a good result for the genetically homogeneous populations .
the alive individual will generate @xmath14 offsprings from the minimum reproduction age @xmath15 to the maximum one @xmath16 , and the genome of each offspring is the same as the parent one , except for @xmath17 mutations randomly occurring at birth . at each time
step @xmath18 , a verhulst factor @xmath19 denoting the survival probability of the individual due to the space and food restrictions is introduced , where @xmath20 is the current population size and @xmath21 is the carrying capacity of the environment , usually set to @xmath22 . in the next section [ sec3 ] the simulations based on these rules
are presented , while for genetically identical individuals , which have the same genotype randomly sampled from the simulated steady - state population , the analytic results can be derived , as shown in section [ sec4 ] .
moreover , in this paper the mortality function @xmath23 at age @xmath2 is defined as @xmath24 where @xmath25 denotes the number of alive individuals with age @xmath2 , and @xmath26 is the survival rate . to eliminate the effect of the verhulst factor ,
the normalized mortality function is preferred @xcite , i.e. , @xmath27 . \label{mortfunc}\ ] ]
in our simulations , initially the population size @xmath28 is @xmath29 and all bits of all the strings are set to zero , i.e. , free of mutations .
one time step @xmath18 corresponds to one aging interval of the individuals , or reading one bit of all strings . the reproduction range is set from @xmath30 to @xmath31 with the birth rate @xmath32 , and the results are similar if using the maximum value of @xmath33 .
@xmath34 mutation for each offspring genome is introduced at birth , and here only the bad mutations are taken into account , that is , the bit randomly selected for mutation is always set to @xmath1 .
( the good mutations have also been considered , e.g. , 10% good mutations and 90% bad ones , and similar results are found . )
[ fig - pop - mut ] shows the evolution of the whole population size @xmath20 until @xmath35 .
similar to the standard penna model , the steady - state population is obtained at late timesteps , and as a result of evolution and selection , the frequency of deleterious bits ( set as @xmath1 ) for the individual of the steady - state population is low at early ages ( especially before the reproduction age ) and very high at old ones .
this behavior of the frequency ( or the bad mutation rate ) is shown in the inset of fig .
[ fig - pop - mut ] .
the mortality function is calculated using eq .
( [ mortfunc ] ) and averaged over the steady - state population from timesteps 5000 to 10000 , as shown in fig .
[ fig - mortality ] .
the result is consistent with the experimental and empirical observations @xcite , that is , at intermediate ages the mortality function increases exponentially , exhibiting the gompertz law , and deceleration occurs for old ages . for comparison , the mortality simulated by the standard ( deterministic ) penna model is also shown in fig .
[ fig - mortality ] , with the threshold of the accumulated bad mutations @xmath36 and the other parameters unchanged .
the exponential gompertz law can also be obtained for the standard penna model @xcite , however , no deceleration is observed except for suitable modifications summarized in @xcite ; see also @xcite .
to study the genetically homogeneous population , one can randomly sample an individual ( genotype ) from the simulated steady - state population , and then `` clone '' it to create the whole genetically identical population . according to the form of these bit - strings ,
the mortality function can be derived and calculated analytically . as in some experiments of fruit
flies @xcite , reproduction is prevented during the aging of genetically homogeneous individuals .
thus , for this population of single genotype , we have @xmath37 where @xmath25 , @xmath38 , is the number of individuals with age @xmath2 in the population , and the function @xmath39 is defined by eq .
( [ liveprob ] ) .
then the survival rate is easily obtained : @xmath40 for the mortality function , the normalized formula ( [ mortfunc ] ) is used to be consistent with the simulations in sec .
[ sec3 ] , and then we have @xmath41 . \label{hommort}\ ] ] different genotypes have been selected randomly from the stable population of sec .
[ sec3 ] , and the corresponding mortality function of each type is calculated using eq .
( [ hommort ] ) .
some examples are shown in fig .
[ fig - clonemort ] for linear - log plots , where part of them obey the exponential gompertz law at the intermediate ages , similar to that of the above simulation ( sec .
[ sec3 ] ) and experiments @xcite .
moreover , all of these curves exhibit the deceleration for old ages . moreover , the analytic calculation is also available if the reproduction is allowed as in other experiments of genetically identical population , but for the case of no mutation .
the details are shown in the appendix , and the mortality function derived is the same as eq .
( [ hommort ] ) .
in this paper a stochastic genetic model of aging is developed based on the bit - string asexual penna model , and the results of the exponentially increasing mortality at intermediate ages and its deceleration at old ages are obtained for both the genetically heterogeneous steady - state population and the homogeneous individuals .
however , the decrease of mortality for the oldest ages , observed in some experiments @xcite , can not be described by the mechanism of this model .
although the properties for intermediate and old ages have been well simulated in this model , the behavior at early ages can not be well reproduced , which is also an artifact of the penna - type genetic models . from fig .
[ fig - clonemort ] for genetically identical populations , it can be found that some populations have unrealistic zero mortality at some early ages .
thus , the effects for the early ages studied in the experiments , such as the investigations of genetic variation for ln - mortality contributed by steady - state population or by new mutations @xcite , can not be produced in this model .
more efforts should be made to avoid this difficulty , e.g. , by considering different kinds of genes before and after the reproduction age @xcite .
we thank scott d. pletcher and naeem jan for very helpful discussions and comments .
this work was supported by sfb 341 .
here an example of the analytic solution for this stochastic model is presented , for the case where the reproduction is allowed in the aging process of genetically identical population , but no mutation occurs when generating the genomes of offsprings .
thus , the individuals keep homogeneous , characterized by the same bit - string @xmath42 with @xmath43 the length of genome ( @xmath44 in above studies ) .
when the system evolves to the steady state , the population size at timestep @xmath18 of this state @xmath45 as well as the verhulst factor @xmath46 can be considered as constant .
thus , the numbers of individuals with ages from @xmath1 to @xmath43 at this step @xmath18 are @xmath47 where @xmath39 is the living probability of individual at age @xmath2 , as defined in eq .
( [ liveprob ] ) , and the individuals of age zero ( newly born ) are generated by the ones with reproducible age ( from age @xmath15 to @xmath16 ) , that is , @xmath48 \nonumber\\ & = & bv[n_{r_{\rm min}-1}(t-1)g(1,2, ...
,r_{\rm min } ) + n_{r_{\rm min}}(t-1)g(1,2, ...
,r_{\rm min}+1 ) \nonumber\\ & + & \cdots+n_{r_{\rm max}-1}(t-1)g(1,2, ... ,r_{\rm max } ) ] \label{a3}\end{aligned}\ ] ] with the birth rate @xmath14 .
consequently , the number of individuals with certain age @xmath2 ( @xmath49 ) can be expressed as @xmath50 therefore , if @xmath51 is unchanged for the steady state , all the @xmath52 , @xmath53 , will also keep unchanged , i.e. , independent of timestep @xmath18 , and then the survival rate @xmath54 can be obtained from eq .
( [ a2 ] ) , that is , @xmath55 and @xmath56 the verhulst factor can be eliminated when calculating the normalized rate : @xmath57 and from the definition of eq .
( [ mortfunc ] ) one can obtain the normalized mortality function , which is the same as eq .
( [ hommort ] ) .
the constant property of the population size @xmath20 and the number @xmath51 for age @xmath9 , as well as the above analytic result of the mortality function have been confirmed by the simulation .
moreover , the steady state condition can be derived from eqs .
( [ a3 ] ) and ( [ a4 ] ) , which is @xmath58=1 , \label{a5}\end{aligned}\ ] ] depending on the parameters @xmath14 , @xmath15 , and @xmath16 .
j. w. vaupel , j. r. carey , k. christensen , t. e. johnson , a. i. yashin , n. v. holm , i. a. iachine , v. kannisto , a. a. khazaeli , p. liedo , v. d. longo , y. zeng , k. g. manton , and j. w. curtsinger , _ science _ * 280 * , 855 ( 1998 ) . | a stochastic genetic model for biological aging is introduced bridging the gap between the bit - string penna model and the pletcher - neuhauser approach .
the phenomenon of exponentially increasing mortality function at intermediate ages and its deceleration at advanced ages is reproduced for both the evolutionary steady - state population and the genetically homogeneous individuals .
: biological aging ; mortality ; penna model . |
in this paper we try to establish the intrinsic characteristics of the stellar sources for the heavy @xmath0-process nuclei ( @xmath0-nuclei with mass numbers @xmath7 corresponding to the elements ba and above ) using aspects of a phenomenological model presented earlier ( wasserburg & qian 2000 ; qian & wasserburg 2001b , 2002 ) . the bulk of the light @xmath0-nuclei ( with @xmath6 ) appear to have different sources from those for the heavy @xmath0-nuclei ( wasserburg , busso , & gallino 1996 ) and are not extensively discussed here . from the observational data on metal - poor ( i.e. , low [ fe / h ] ) stars
, we will show that the heavy @xmath0-nuclei are produced with only some coproduction of the light @xmath0-elements such as sr , y , and zr , but without any coproduction of the elements from na to zn ( including fe ) .
we argue that supernovae with extended shell structure can not be the source for the heavy @xmath0-nuclei . instead ,
these nuclei must be produced during the formation of an essentially bare neutron star .
such a state may result from a low - mass ( @xmath8@xmath9 ) type ii supernova ( sn ii ) or from the accretion - induced collapse ( aic ) of a white dwarf into a neutron star in a binary system .
the observational basis that establishes the distinctive characteristics of the stellar sources for the heavy @xmath0-nuclei is discussed in 2 . in
3 we discuss the requirement of an essentially bare neutron star for the production of the heavy @xmath0-nuclei and the proposed role of aic events . the special nature of both @xmath0-process and @xmath4-process enrichments in binary systems with aic is discussed in 4 . in
5 we explore some of the consequences for our earlier model and other implications that follow if aic events are a source for the heavy @xmath0-nuclei .
elemental abundances in halo stars of low metallicities have provided fundamental data for understanding nucleosynthesis and chemical evolution of the galaxy .
the observations ( mcwilliams et al . 1995 ; ryan , norris , & beers 1996 ; burris et al .
2000 ; westin et al .
2000 ; sneden et al .
2000 ; norris , ryan , & beers 2001 ) led us to propose earlier that there was a major transition at [ fe / h ] @xmath3 in the production of heavy elements .
the fe production at [ fe / h ] @xmath10 was dominated by very massive ( @xmath11 ) stars ( vmss ) .
it was proposed that the vms activities ceased and major formation of normal stars ( with masses @xmath12@xmath13 ) took over when a critical metallicity corresponding to [ fe / h ] @xmath3 was reached in the interstellar medium ( ism ) .
this is in accord with the recent study by bromm et al .
( 2001 ) , which indicates that substantial formation of a normal stellar population could only occur when a metallicity substantially above @xmath14 times the solar value was achieved in the ism to provide sufficient cooling for the collapse and fragmentation of gas clouds .
the subsequent fe production at @xmath15}<-1 $ ] was then considered to be dominated by sne ii .
the production of the heavy @xmath0-elements relative to fe was very low prior to the achievement of [ fe / h ] @xmath3 in the ism , but increased sharply at [ fe / h ] @xmath3 and subsequently approached a well - defined trend for normal stages of galactic evolution ( wasserburg & qian 2000 ; qian & wasserburg 2001b , 2002 ) .
this behavior is shown for eu ( a predominantly @xmath0-process element in the solar system ) in figure 1a where the available data on @xmath16 are plotted versus [ fe / h ] . , where ( e / h ) is the abundance ratio of element e to hydrogen in a star , and @xmath17}\equiv\log({\rm fe / h})-\log({\rm fe / h})_\odot$ ] . ]
it can be seen that for [ fe / h ] @xmath18 there is a very wide scatter in @xmath16 while for higher values of [ fe / h ] @xmath16 follows [ fe / h ] in a more regular trend .
this trend is indicated by the solid line segment of unit slope in figure 1a and reflects a constant eu / fe ratio for the net production by all stellar sources .
the results for eu are similar to those for ba as shown in figure 1b .
the transition at [ fe / h ] @xmath3 is prominently demonstrated for ba by the contrast between the data at [ fe / h ] @xmath10 and those at [ fe / h ] @xmath19 .
note that this transition reflects the changes in the relative production rates of the heavy @xmath0-elements with respect to fe , and therefore , can be used to identify the nucleosynthetic contributions to the ism from the sources for the heavy @xmath0-elements relative to those from the sources for fe .
for example , the transition for ba corresponds to an increase by a factor of 10 ( solid curve in fig .
1b ) to 20 ( dot - dashed curve in fig .
1b ) in the relative production of ba with respect to fe from [ fe / h ] @xmath10 to [ fe / h ] @xmath19 ( qian & wasserburg 2002 ) .
however , the transition at [ fe / h ] @xmath3 by itself can not identify the relative populations of the different stellar types . to elucidate the nature of the sources for the heavy @xmath0-elements , we exhibit in figure 2 the available data on hd 115444 , hd 122563 ( westin et al .
2000 ) , and cs 31082001 ( hill et al .
2002 ) with [ fe / h ] @xmath20 , @xmath21 , and @xmath22 , respectively , which show clear evidence for varying degrees of @xmath0-process enrichment within a narrow range of [ fe / h ] .
we note that the data ( not shown ) on cs 22892052 with [ fe / h ]
@xmath23 ( mcwillam et al . 1995 ; sneden et al . 2000 ) are very close to those on cs 31082001 . we also note that high th abundances were found in cs 22892052 and hd 115444 while high th and u abundances were found in cs 31082001 .
it can be seen from figure 2a that the abundances of the elements from na to zn ( including fe ) in the three stars shown there are all rather constant with a typical range of @xmath24 dex in @xmath25 .
this is in sharp contrast to the wide range of @xmath26 dex in @xmath25 for the heavy @xmath0-elements ( ba and above ) as shown in figure 2b for the same stars .
figure 2b also shows that the relative abundances of the heavy @xmath0-elements in each star can be described remarkably well by the solar @xmath0-process abundance pattern ( solar @xmath0-pattern ) . indeed , extensive studies ( e.g. , sneden et al .
1996 ; johnson & bolte 2001 ) found that the heavy @xmath0-elements in a number of metal - poor stars often closely follow the solar @xmath0-pattern while having widely varying absolute abundances .
it follows from figure 2 that the sources for the heavy @xmath0-nuclei are not connected with the production of the elements from na to zn , which include not only the fe group but also the so - called `` @xmath27-elements '' mg , si , ca , and ti .
this observation was discussed in our earlier study ( qian & wassurburg 2002 ) of the nucleosynthetic yields which we attributed to vmss and two hypothesized types of sne ii [ sne ii(@xmath28 ) and sne ii(@xmath29 ) ] . in particular ,
the abundances of the elements from na to zn at [ fe / h ] @xmath3 were considered to represent a prompt inventory of the elements dominantly produced by vmss . the prompt inventory of the light @xmath0-elements sr , y , and zr is also significant ( qian & wasserburg 2001b ) .
this can be seen from the data on hd 115444 and hd 122563 shown in figure 2b .
these two stars have very different abundances of the heavy @xmath0-elements but essentially the same abundances of sr , y , and zr , which are dominated by the prompt inventory of these elements .
compared with hd 115444 and hd 122563 , cs 31082001 have much higher abundances of the heavy @xmath0-elements and significantly higher abundances of sr , y , and zr ( see fig .
this indicates that the sources for the heavy @xmath0-nuclei also produce some light @xmath0-nuclei and such contributions overwhelm the prompt inventory of sr , y , and zr in cs 31082001 . the data on cs 31082001 ( see fig .
2b ) and cs 22892052 ( sneden et al .
2000 ) show that the light @xmath0-elements produced by the sources for the heavy @xmath0-nuclei in these two stars also include nb , ru , rh , pd , ag , and cd . however , the sources for the bulk of the light @xmath0-nuclei appear to be different from those for the heavy @xmath0-nuclei as first argued by wasserburg et al .
( 1996 ) based on the inventory of @xmath30hf and @xmath31i in the early solar system .
this was supported by the data on cs 22892052 ( sneden et al .
2000 ) , cs 31082001 ( hill et al . 2002 ) , bd + 17@xmath323248 ( cowan et al .
2002 ) , and several other metal - poor stars ( johnson & bolte 2002b ) , which showed that the light @xmath0-elements such as ag in these stars are clearly deficient relative to the solar @xmath0-pattern translated to pass through the eu data .
the production of the light @xmath0-elements appears to be associated with that of fe .
this was studied in some detail in qian & wasserburg ( 2001b ) but will not be extensively discussed here .
the range in @xmath25 for eu and ba in many halo stars with [ fe / h ] @xmath3 shown in figure 1 and the range in @xmath25 for the pure heavy @xmath0-elements in three such stars shown in figure 2 require special attention . if we consider that the solar @xmath0-process abundances represent the galactic average at the time of solar system formation ( ssf ) , and that the absolute yield @xmath33 of element ( or nuclide ) e per h atom in a standard reference mass of ism which will mix with the nucleosynthetic products is constant for every event producing the heavy @xmath0-nuclei , it follows ( see appendix a ) that @xmath34 where @xmath35 is the solar @xmath0-process abundance ratio of e to hydrogen and @xmath36 is the frequency for replenishing the standard reference mass of ism with newly - synthesized material by the events producing the heavy @xmath0-nuclei .
the term @xmath37 is the number of such events per standard reference mass of ism over the period @xmath38 yr prior to ssf . based on the abundance of @xmath30hf ( with a mean lifetime of @xmath39 yr ) in the early solar system ,
@xmath36 was estimated to be @xmath40 ( wasserburg et al .
1996 ; qian & wasserburg 2000 ) .
this gives @xmath41 .
for a metal - poor star formed from the ism in the early galaxy , the number @xmath42 of the events contributing to the abundance of e in the star is @xmath43 where ( e / h ) is the number abundance ratio of e to hydrogen in the star .
for the three stars with [ fe / h ] @xmath3 shown in figure 2 , their enrichments of heavy @xmath0-elements correspond to @xmath4450 .
while the abundances corresponding to @xmath4410 for e.g. , eu , can be reasonably attributed to the enrichment of an average ism at [ fe / h ] @xmath3 by the events producing the heavy @xmath0-nuclei , it is rather problematic to consider that @xmath45 such events could have occurred in an average ism without any other events also occurring to change the fe abundance . as argued earlier by us ( qian & wasserburg 2001a ) , the high @xmath0-process enrichments in stars with very low [ fe / h ] such as cs 22892052 and cs 31082001 ( eu and ba data shown as filled diamonds and filled circles , respectively , in fig .
1 ) most plausibly resulted from surface contamination by being a low - mass binary companion to the stellar source for the heavy @xmath0-nuclei .
there is some indication that cs 22892052 might be in a binary ( preston & sneden 2001 ) .
a special case of high eu enrichment at low [ fe / h ] was discovered recently by j. cohen and n. christlieb , who found @xmath46 corresponding to @xmath47 for he 21481247 with [ fe / h ] @xmath48 ( cohen et al .
2003 ) . the data point ( open diamond in fig .
1a ) for this star lies far above the trend for the evolution of eu relative to fe at [ fe / h ] @xmath49 .
the result for ba is similar ( see fig .
1b ) . upon being informed of the initial discovery and prior to the detection of pb in this star , we attempted to predict the abundances of all the other elements from only the preliminary data on eu and fe using our earlier model ( qian & wasserburg 2001b , 2002 ) , which assumes no @xmath4-process contributions .
these predictions are shown along with the final data in figure 3a .
it can be seen that in consideration of the observational errors , there is good agreement between the predictions and the data for most of the elements .
this is the case especially for the elements from mg to ni , the predicted abundances of which have significant contributions from the prompt inventory .
however , there is a serious discrepancy for the elements ba , la , ce , pr , and nd as highlighted in figure 3b .
this discrepancy and the detection of high pb abundance in he 21481247 unambiguously demonstrate the presence of high @xmath4-process enrichments in this star [ the pb in this star can not be plausibly attributed to the decay of the actinides produced by the @xmath0-process as the observed value of @xmath50 far exceeds @xmath51 . on the other hand , the ba / eu ratio in this star is @xmath52 times smaller than that for the main @xmath4-component of the solar abundances ( arlandini et al .
we can not construct an @xmath4-process scenario that could decrease the ba / eu ratio greatly below the value for the solar main @xmath4-component .
this and the tentatively inferred th abundance would imply a mixture of high @xmath4-process and high @xmath0-process enrichments in he 21481247 , the explanation of which requires a new approach and is discussed here and in cohen et al .
( 2003 ) .
it was well recognized that the high @xmath4-process enrichments in metal - poor stars are associated with mass transfer from their previous asymptotic giant branch ( agb ) companions in binaries .
four such stars with confirmed binary membership are lp 62544 with [ fe / h ] @xmath53 ( aoki et al . 2000 ) , cs 22948027 with [ fe / h ] @xmath54 ( hill et al . 2000 ; preston & sneden 2001 ) , cs 22942019 with [ fe / h ] @xmath55 ( preston & sneden 2001 ) , and he 00242523 with [ fe / h ] @xmath55 ( lucatello et al . 2003 ) .
there is also evidence for he 21481247 being in a binary ( cohen et al .
2003 ) . from the above discussion
, it is apparent that : ( 1 ) the production of the heavy @xmath0-nuclei is decoupled from the production of the elements from na to zn ( including the `` @xmath27-elements '' and the fe group ) ; ( 2 ) surface contamination of a low - mass binary companion by the stellar source for the heavy @xmath0-nuclei is a relatively common occurrence ; ( 3 ) surface contamination with @xmath4-process products is a fairly common occurrence ; and ( 4 ) there seems to be a coupling of high @xmath4-process and high @xmath0-process enrichments in at least one metal - poor star .
the fundamental criterion for an astrophysical environment to be the site for @xmath0-process nucleosynthesis is the capability of providing a large neutron abundance on a short timescale .
two major candidate environments are the neutrino - driven wind from new - born neutron stars ( e.g. , woosley & baron 1992 ) and the ejecta from neutron star mergers ( e.g. , freiburghaus , rosswog , & thielemann 1999 ) . in the neutron star merger model , the material undergoing nucleosynthesis
is ejected from a bare old neutron star that is disrupted during the merging with another neutron star or a black hole .
regardless of the possibility of an @xmath0-process , production of the elements from na to zn would not be expected to occur in this material .
thus , if neutron star mergers are a source for the heavy @xmath0-nuclei , the observational requirement that the production of these nuclei not be related to the production of the elements from na to zn is satisfied .
however , due to the rarity of neutron star mergers , very large yields per event are required to account for the total @xmath0-process abundances in the present galaxy if these events are the major source for the @xmath0-nuclei .
the required large yields appear to be in conflict with the level of @xmath0-process enrichment observed in metal - poor stars ( qian 2000 ) .
while the possibility of neutron star mergers being the @xmath0-process site deserves further investigation , here we focus on the neutrino - driven wind from new - born neutron stars as the likely source for the heavy @xmath0-nuclei .
a neutron star can be formed from the collapse of ( 1 ) an fe core of stars with @xmath56 , ( 2 ) an o - ne - mg core of stars with @xmath57@xmath58 , and ( 3 ) a white dwarf which has been accreting material from its low - mass companion in a binary system .
the surface layers of a new - born neutron star are heated by an enormous flux of neutrinos through the reactions @xmath59 and @xmath60 , which results in a neutrino - driven wind ( duncan , shapiro , & wasserman 1986 ) .
this wind was studied as a possible @xmath0-process site but with very uncertain results ( e.g. , woosley & baron 1992 ; meyer et al .
1992 ; takahashi , witti , & janka 1994 ; woosley et al .
independent of these theoretical uncertainties , the observational requirement that the production of the heavy @xmath0-nuclei be decoupled from the production of the elements from na to zn ( including the `` @xmath27-elements '' and the fe group ) strongly constrains the possible association of the above neutron star formation scenarios with the sources for the heavy @xmath0-nuclei .
stars with @xmath56 possess extensive shells surrounding the fe core .
these shells contain the elements from na to the fe group when they are ejected in the sn ii explosion of the star subsequent to the collapse of the fe core ( e.g. , woosley & weaver 1995 ; thielemann , nomoto , & hashimoto 1996 ) .
therefore , sne ii from stars with @xmath56 can not be associated with the source for the heavy @xmath0-nuclei that were observed in the stars shown in figure 2 .
stars with @xmath61@xmath58 have insignificant shells surrounding the o - ne - mg core ( nomoto 1984 ) .
the sn ii explosion resulting from the core collapse of such stars produces essentially no elements from na to zn ( nomoto 1987 ) , and therefore , can be associated with the source for the heavy @xmath0-nuclei as argued earlier by us ( qian & wasserburg 2002 ; see also wheeler , cowan , & hillebrandt 1998 ) .
the accretion - induced collapse ( aic ) of a white dwarf produces a bare neutron star which ejects material only through the neutrino - driven wind .
thus , the aic events can also be a source for the heavy @xmath0-nuclei .
the aic of a white dwarf was initially proposed to explain the origin of neutron stars with low - mass binary companions ( see canal , isern , & labay 1990 for a review ) .
the white dwarf in an aic event may be of c - o composition for a progenitor with @xmath62 or of o - ne - mg composition for a progenitor with @xmath57@xmath63 .
[ due to tidal mass loss , the progenitor mass range for the formation of an o - ne - mg core in binary systems is wider than the range @xmath57@xmath58 for single stars ( nomoto & kondo 1991 ) . ] nomoto & kondo ( 1991 ) showed that in both cases a broad range of mass accretion rates would lead to aic .
studies of nucleosynthesis in the neutrino - driven wind associated with aic events ( woosley & baron 1992 ; fryer et al .
1999 ) found no @xmath0-processing .
instead , these studies showed that elements such as sr , y , and zr are produced profusely by aic events .
this was used to argue that these events should be very rare .
consequently , aic events have not been considered as a serious candidate for the @xmath0-process site .
however , the conditions in the neutrino - driven wind are sensitive to many uncertainties in the neutrino and neutron star physics ( e.g. , qian & woosley 1996 ) . in the arguments presented below , we will assume that the conditions in the wind associated with aic events are suitable for producing the heavy @xmath0-nuclei .
we hope that the observational evidence discussed here would stimulate and possibly guide further theoretical studies of these events in connection with @xmath0-process nucleosynthesis .
as argued above , if the aic events are a source for the heavy @xmath0-nuclei , then the production of these nuclei is automatically decoupled from the production of the elements from na to zn .
in addition , as these events occur in binary systems , the ejecta from such events would naturally lead to contamination of the low - mass companion s surface with heavy @xmath0-elements . as material is ejected through the neutrino - driven wind with only a small total mass loss in the aic events , ablation of the companion
may be insignificant compared with the case of sne ii .
further , the aic model for the production of the heavy @xmath0-nuclei requires that there be a coupling of @xmath0-process and @xmath4-process enrichments in binary systems .
in general , a star evolves through the agb phase prior to becoming a white dwarf .
this evolution results in production of neutrons , the capture of which by the fe seeds gives rise to the @xmath4-process . depending on the distribution of neutron fluence or the ratio of neutrons to the fe seeds , low - metallicity agb stars may predominantly produce pb ( e.g. , van eck et al .
2001 ) as predicted by gallino et al .
( 1998 ) or produce an @xmath4-process pattern similar to the main @xmath4-component of the solar abundances ( e.g. , aoki et al . 2000 ) , or produce some superposition of high pb enrichment and the main solar @xmath4-component ( e.g. , johnson & bolte 2002a ; cohen et al . 2003 ) . in any case
, the @xmath4-process products along with the other elements in the envelope of the agb star may be dumped onto the surface of a low - mass companion via mass transfer in a binary .
if the white dwarf left behind by the agb star subsequently accretes mass from the companion and undergoes aic , the surface of the companion will be enriched again , but this time with heavy @xmath0-elements .
it is also plausible that the ratio of neutrons to the fe seeds in some agb stars would be too low for any significant @xmath4-processing to occur - processing would occur in the star .
however , this star still produces a white dwarf , the aic of which can then provide pure @xmath0-process enrichments to the low - mass companion . ]
( the problem of @xmath4-processing at low metallicities is now attracting considerable attention ) .
thus , the low - mass star in a binary may have a surface composition ranging from predominantly @xmath0-process [ e.g. , possibly cs 22892052 ( sneden et al .
2000 ; preston & sneden 2001 ) ] to predominantly @xmath4-process [ e.g. , lp 62544 ( aoki et al . 2000 ) ] in origin .
the elements c , n , and o will then reflect the normal stellar processing in the low - mass star ( e.g. , fujimoto , ikeda , & iben 2000 ) and additions from the standard evolution of the originally more massive agb companion
. it may be rather common to find metal - poor stars with high surface enrichments of both @xmath4-process and @xmath0-process elements .
we note that a number of metal - poor stars were observed with high @xmath4-process enrichments ( e.g. , hill et al .
2000 ; aoki et al .
2000 ; johnson & bolte 2002a ; lucatello et al . 2003 ) .
unfortunately , the element th , which is unique to the @xmath0-process , is either not targeted for observations or is difficult to detect for these stars .
we know of only one star , he 21481247 , the surface composition of which reflects a superposition of @xmath4-process and @xmath0-process products possibly including th ( cohen et al .
2003 ) . without th detection , significant @xmath0-process enrichments may still be identified based on the observed abundance pattern of the elements ba and above .
this is because for these elements , the patterns produced by the @xmath4-process are drastically different from the pattern produced by the @xmath0-process . a convenient measure to characterize this difference is the ba / eu ratio . as mentioned earlier , observations showed that the abundances of the heavy @xmath0-elements in a number of metal - poor stars closely follow the solar @xmath0-pattern .
this suggests that the ba / eu ratio for the events producing the heavy @xmath0-nuclei is close to the solar @xmath0-process value ( ba / eu)@xmath64 ( arlandini et al .
1999 ) . by contrast , the ba / eu ratio for the solar main @xmath4-component is ( ba / eu)@xmath65 ( arlandini et al .
1999 ) . due to the variations in the ratio of neutrons to the fe seeds ,
the @xmath4-process in low - metallicity agb stars will in general produce different ba / eu ratios from ( ba / eu)@xmath66 .
the metallicity dependence of the @xmath4-process production pattern was demonstrated by the data ( van eck et al .
2001 ) on three metal - poor stars , the @xmath4-process abundances of which do not follow the solar main @xmath4-component but are characterized by predominant pb enrichments as predicted by gallino et al .
models that were in good agreement with the data on zr , la , ce , pr , nd , sm , and pb in these stars gave the @xmath4-process ratios ( ba / eu)@xmath67 for [ fe / h ] @xmath68 and ( ba / eu)@xmath69 for [ fe / h ] @xmath70 ( van eck et al .
thus , it appears that the @xmath4-process always produces ( ba / eu)@xmath71 .
the ba / eu ratio in a star is given by @xmath72 \beta_r({\rm ba } ) } , \label{rba}\ ] ] where @xmath73 is the fraction of ba contributed by the @xmath0-process .
the ba / eu ratio in the star can also be given by @xmath74 , \label{seu}\ ] ] where @xmath75 is the fraction of eu contributed by the @xmath4-process .
a general question is when the @xmath0-process and the @xmath4-process contributions can be identified for a star with a mixture of such contributions . from equation ( [ rba ] ) , the @xmath0-process contributions to the star can be well recognized by the abundance pattern for @xmath76 . from equation ( [ seu ] ) , the @xmath4-process contributions to the star can be identified for @xmath77 .
this is illustrated in figure 4 by considering mixtures of the solar @xmath0-pattern and the solar main @xmath4-component . for such mixtures ,
the reference value to be compared with @xmath73 and @xmath75 is @xmath78 .
the dot - dashed curve in figure 4a represents a mixture in which the ba receives equal contributions from the @xmath0-process and the @xmath4-process [ @xmath79 $ ] .
it can be seen that this curve closely follows the solar @xmath0-pattern ( thick solid curve ) and the @xmath4-process contributions are essentially obscured .
this can be understood as only @xmath80 of the eu [ @xmath81 $ ] is contributed by the @xmath4-process .
it follows that if @xmath82 of the ba in a star is from the @xmath0-process , then the abundances of all the elements above ba will exhibit the solar @xmath0-pattern .
the case of @xmath83 corresponding to @xmath84 is also shown in figure 4b .
for comparison , the thin solid curve in figure 4b represents a mixture in which the eu receives equal contributions from the @xmath0-process and the @xmath4-process [ @xmath85 .
it can be seen that this curve closely follows the solar main @xmath4-component ( dashed curve ) .
we emphasize that if @xmath82 of the eu in a star is from the @xmath4-process , the overall abundance pattern will not be easily distinguishable from an @xmath4-pattern .
the identification of the @xmath0-process and the @xmath4-process contributions is also complicated by the observational errors ( @xmath86 dex in @xmath25 ) and by the possible variations of ( ba / eu)@xmath87 . in very uncertain cases ,
the @xmath0-process contributions can only be identified by the data on ir , th , and u , which receive much less or no contribution from the @xmath4-process compared with eu . the observed abundance pattern of the elements from ba to dy except for gd in he 21481247 can be accounted for by a mixture of the solar @xmath0-pattern and the solar main @xmath4-component with @xmath88 corresponding to @xmath89 .
this mixture is shown as the dot - dashed curve in figure 3b .
we can not offer an explanation for the offset of the gd data from the dot - dashed curve .
for the mixture represented by this curve , both @xmath73 and @xmath75 greatly exceed the reference value of @xmath78 .
consequently , the presence of both the @xmath0-process and the @xmath4-process contributions in he 21481247 can be established based on the observed ba / eu ratio alone .
as the dominant @xmath4-process sources in the galaxy are long - lived low - mass stars , the @xmath4-process contributions to the ism are insignificant at low metallicities corresponding to very early times .
as mentioned earlier , it was well recognized that the high enrichments of @xmath4-process elements in metal - poor stars are associated with mass transfer from their previous agb companions in binaries .
it was thought that all the stars with high @xmath4-process enrichments should have a white dwarf companion now .
some such stars indeed show shifts in their radial velocity ( aoki et al .
2000 ; preston & sneden 2001 ; lucatello et al . 2003 ) .
however , preston & sneden ( 2001 ) showed that these cases are not common and a number of highly @xmath4-process enriched stars appear to be single stars .
according to the above discussion of the aic events , the highly @xmath4-process enriched metal - poor stars , if in binaries , may have neutron stars instead of white dwarfs as companions .
further , the neutron star formed from the aic of a white dwarf may receive a large kick from , e.g. , asymmetric neutrino emission ( e.g. , lai & qian 1998 ) .
this would disrupt the binary and may explain the observed single stars with high @xmath4-process enrichments .
the binary disruption scenario would also imply the existence of highly @xmath0-process enriched single stars as significant @xmath4-processing might not occur in some agb companions ( see footnote 4 ) .
the data on metal - poor stars require that the production of the heavy @xmath0-nuclei be decoupled from the production of the elements from na to zn ( including the `` @xmath27-elements '' and the fe group ) .
this appears to exclude sne ii with progenitors of @xmath56 , or more generally any sne with extended shell structure , from being the source for the heavy @xmath0-nuclei . considering that the heavy @xmath0-nuclei are produced in the neutrino - driven wind from a new - born neutron star
, we have argued that only sne ii from stars with @xmath57@xmath58 and aic events in binary systems , which produce essentially bare neutron stars , satisfy the observational requirement .
while neutron star mergers may also satisfy this requirement , they are not considered here as a major source for the heavy @xmath0-nuclei .
this is because the rarity of neutron star mergers requires very large yields per event , which appear to be in conflict with the observed level of @xmath0-process enrichments in metal - poor stars .
further , neutron star mergers can not produce an intrinsic coupling between @xmath0-process and @xmath4-process enrichments in binary systems . as argued above
, the aic events may naturally lead to contamination of the low - mass binary companions surface with both @xmath4-process and @xmath0-process products .
the @xmath4-process contamination is associated with the agb progenitor for the white dwarf which subsequently undergoes aic into a neutron star .
depending on the distribution of the ratio of neutrons to the fe seeds that results from the evolution of the agb progenitor , the surface composition of the low - mass binary companion can range from predominantly @xmath0-process to predominantly @xmath4-process in origin .
the discovery of a highly - enriched mixture of @xmath4-process and @xmath0-process elements including possibly th in he 21481247 provides strong support for the aic model for the production of the heavy @xmath0-nuclei . in general , as the binary might be disrupted during an aic event , this might explain the existence of single stars with high @xmath4-process enrichments .
the binary disruption scenario would also imply the existence of highly @xmath0-process enriched single stars as significant @xmath4-processing might not occur in some agb companions ( see footnote 4 ) .
subsequent to the disruption of the binary , the highly @xmath4-process or @xmath0-process enriched stars may acquire rather large proper motion .
the aic model can be tested by discovering ( 1 ) more metal - poor stars with high enrichments of both @xmath4-process and @xmath0-process elements including ir , th , and u , ( 2 ) neutron star companions of such stars , and ( 3 ) large proper motion of such stars in single configuration . as some light @xmath0-elements are produced along with the heavy ones in lesser quantities ( qian & wasserburg 2001b ) , the detection of the light @xmath0-elements such as ag may help the identification of @xmath0-process enrichments in a star .
the aic events have no optical display due to the lack of an envelope .
direct observations of these events must rely on the detection of neutrinos from the neutron star formed in such events .
gamma rays from decay of the progenitors of the heavy @xmath0-nuclei , as well as radio signals from the shocked ejecta and the neutron star , may also be observable from an aic event .
there are several questions regarding the generality of the sources for the heavy @xmath0-nuclei that are proposed here based on observations of metal - poor stars .
it is not evident that all the heavy @xmath0-nuclei in the galaxy were produced by low - mass ( @xmath8@xmath58 ) sne ii and aic events .
the mass range of @xmath8@xmath58 corresponds to only @xmath90 of all the stars with @xmath57@xmath13 for a salpeter initial mass function ( imf ) of the form @xmath91 . for the low - mass sne ii to constitute @xmath92 of all sne
ii would require a drastic change from @xmath93 to @xmath94 in the exponent of the imf .
thus , low - mass sne ii are perhaps not the major source for the heavy @xmath0-nuclei due to the narrow mass range .
if the proposed aic model is not responsible for the bulk of the heavy @xmath0-nuclei , either , then the question is what are the major sources for these nuclei ?
if the aic events are the dominant or sole source for the heavy @xmath0-nuclei , then it is necessary to provide a detailed description of the @xmath0-process production and the chemical enrichment associated with this scenario .
first consider the issue of chemical enrichment . in our earlier model ( wasserburg & qian 2000 ; qian & wasserburg 2001b , 2002 ) , we attributed the heavy @xmath0-nuclei to @xmath95 of sne ii [ sne ii(@xmath28 ) ] and the bulk of the light @xmath0-nuclei to the rest [ sne ii(@xmath29 ) ] .
in addition , it was required that sne ii(@xmath29 ) but not sne ii(@xmath28 ) produce the elements from na to zn ( including fe ) . for the model proposed here with aic events being the major source for the heavy @xmath0-nuclei , there is no need to require that most sne ii produce no elements from na to zn .
thus , all sne ii from stars with @xmath56 could be sne ii(@xmath29 ) .
this removes the conflict between our earlier model and observations of sn ii light curves , which showed that most sne ii produce @xmath96 of fe ( e.g. , table 1 in sollerman 2002 ) . with this fe yield and a galactic sn ii rate of @xmath97 , @xmath98 sne ii over the galactic history of @xmath99
yr would enrich the total baryonic mass of @xmath100 in the galaxy with @xmath101 of the solar fe mass fraction ( @xmath102 ) .
this is in accord with the expected additional fe contributions from sne ia . as in the case of sne ii
, the amount of ism which will dilute the nucleosynthetic products is determined by the total energy injected into the ism .
the energy injected by an aic event is perhaps comparable to that by an sn ii ( woosley & baron 1992 ) , which corresponds to a standard dilution mass of @xmath103 ( e.g. , thornton et al .
if @xmath41 aic events provided this dilution mass with the solar abundances of the heavy @xmath0-nuclei ( see 2 ) , then a total of @xmath104 aic events must have occurred to provide the same enrichment to the total baryonic mass of @xmath105 in the galaxy .
this is @xmath106 times more than the total number of sne ii over the galactic history , and therefore , suggests that the galactic rate of aic events must be @xmath107 if these events are the major source for the heavy @xmath0-nuclei .
the value of @xmath41 used to obtain this rate corresponds to a frequency of @xmath40 for replenishment of the heavy @xmath0-nuclei in the standard dilution mass .
this frequency was inferred from the original measurements of the @xmath30w/@xmath108w deficiencies in iron meteorites relative to the earth s crust , which imply ( @xmath30hf/@xmath109hf@xmath110 in the early solar system .
recent measurements by yin et al .
( 2002 ) and kleine et al .
( 2002 ) give ( @xmath30hf/@xmath109hf@xmath111 . repeating the argument of wasserburg et al .
( 1996 ) for this new value indicates that the frequency for replenishment of the heavy @xmath0-nuclei should be reduced by a factor of @xmath26 , and therefore , the required galactic rate for aic events to be the major source for the heavy @xmath0-nuclei is @xmath112 .
if the galactic rate of aic events is indeed @xmath112 , then the neutrino signals from such an event will be observed in the near future by detectors such as super kamiokande ( sk ) .
the neutrino signals from an aic event should be similar to those from an sn ii .
the dominant signals in a water erenkov detector such as sk are caused by @xmath60 .
eleven @xmath113 events were observed from sn 1987a at a distance of 50 kpc by the kamiokande ii water detector with a fiducial mass of 2.14 kton ( hirata et al .
thus , if an aic event occurs at a distance of 10 kpc , @xmath114 @xmath113 events will be observed by sk with a fiducial mass of 32 kton .
the aic events were originally proposed to explain the formation of low - mass x - ray binaries ( lmxbs ) , which consist of a low - mass star and a neutron star ( see canal et al .
1990 for a review ) .
however , we can not connect the aic events with lmxbs based on the required galactic rate for these events to be the major source for the heavy @xmath0-nuclei .
there are @xmath115 lmxbs observed in the galaxy ( van paradijs 1995 ) .
if the lifetime of an lmxb is @xmath116@xmath117 yr , the birth rate of lmxbs is @xmath118@xmath119 ( verbunt & van den heuvel 1995 ) .
this is far below the galactic rate of sne ii and that of aic events discussed above .
therefore , if the aic events are the major source for the heavy @xmath0-nuclei , these events may have little to do with the formation of lmxbs .
we note that the observed lmxbs have very short orbital periods of @xmath120400 hr ( white , nagase , & parmar 1995 ) while the binary systems containing low - mass stars with high @xmath4-process enrichments tend to have much longer orbital periods of @xmath121 yr ( aoki et al .
2000 ; preston & sneden 2001 ; cohen et al .
it is plausible that the aic events frequently occur in wide binary systems while lmxbs are the rare outcome of the evolution of close binary systems . the galactic birth rate of pulsars , which are rotating neutron stars with high magnetic fields ,
was estimated to be ( 60330 yr)@xmath122 ( lyne et al .
it is conventional to consider all pulsars as the products of sne ii . as only @xmath123 of the catalogued sn
ii remnants in the galaxy are observed to contain pulsars ( kaspi & helfand 2002 ) , the galactic sn ii rate of @xmath97 is only consistent with the lower end of the estimated pulsar birth rate . if aic events are the major source for the heavy @xmath0-nuclei , the required galactic rate of @xmath112 for these events could easily accommodate the higher end of the estimated pulsar birth rate . in addition , accretion onto the white dwarf in an aic event would provide a natural mechanism to produce a rapidly rotating neutron star , which is required for the pulsar mechanism . according to the above discussion , if aic events are the major source for the heavy @xmath0-nuclei , their occurrences in the galaxy must have produced @xmath124 neutron stars , which are @xmath125 times more than those produced by sne ii .
this implies that aic events are a frequent outcome of binary evolution , far more common than sne ia . on the other hand ,
progenitor systems of aic events only constitute a fraction of all binary systems , most of which would evolve into white dwarf binaries .
thus , the total number of white dwarf binaries must greatly exceed @xmath5 , which is the number of neutron stars produced by aic events .
these white dwarf binaries could contribute significantly to the inventory of machos . as discussed in 4
, aic events would lead to high @xmath4-process and high @xmath0-process enrichments in binary systems .
these enrichments are not diagnostic of the composition of the ism from which the binary systems were formed
. there are three distinct times involved here , which correspond to the formation of the binary , the development of the agb phase , and the aic event .
we note that there may be significant delays between the first and the other two times depending on the stellar masses . due to the time
( @xmath126 yr ) required for the evolution of the most abundant stars ( @xmath127 ) to reach the agb phase , it is plausible that there were no significant @xmath4-process contributions to the ism at low [ fe / h ] .
prior to the onset of sn ia contributions at [ fe / h ] @xmath128 , the elements from na to zn can only be produced by stars with @xmath56 .
we consider these elements instead of the heavy @xmath0-elements or the @xmath4-process elements as more reliable monitors of the ism because surface contamination is not likely an issue in interpreting the stellar data on these elements .
consequently , the abundances of the elements from na to zn can be plausibly associated with ongoing `` average '' evolution of the ism at @xmath15}<-1 $ ] . as we have emphasized here the role of binary systems at low [ fe / h ]
, we must consider the effects of sne ia , which produce dominantly the fe group elements but not the `` @xmath27-elements . ''
the observed transition in the trend for evolution of mg , si , ca , and ti relative to fe at [ fe / h ] @xmath128 indicates that major fe contributions from sne ia only occur at [ fe / h ] @xmath129 .
the role of the aic events as the major source for the heavy @xmath0-nuclei especially at [ fe / h ] @xmath130 then implies severe restrictions on binary evolution .
it is possible that sne ia may only occur through the merging of two white dwarfs , which are produced in binary systems with two low - mass stars ( e.g. , iben 1985 ) . in this case , the onset of sn ia contributions would require a delay of @xmath131 yr similar to that for the onset of major @xmath4-process contributions to the ism .
the parameter space for the mass and the accretion rate of a c - o or o - ne - mg white dwarf leading to aic or an sn ia was studied by nomoto & kondo ( 1991 ) .
more studies of this kind and of binary evolution in general are much needed to assess the possibility of the predominance of aic events over sne ia at [ fe / h ] @xmath130 .
another crucial issue to be addressed is the @xmath0-process production by aic events .
the diverse nature of the @xmath0-process was inferred from the meteoritic data on @xmath30hf and @xmath31i by wasserburg et al .
this was supported by observations of metal - poor stars , which showed that the light @xmath0-elements such as rh and ag in cs 22892052 ( sneden et al .
2000 ) and cs 31082001 ( hill et al .
2002 ) are deficient relative to the solar @xmath0-pattern translated to pass through the eu data ( see also cowan et al . 2002 ; johnson & bolte 2002b ) .
so far ab initio approaches are unable to define the @xmath0-process yield patterns .
observations ( sneden et al . 1996 ; westin et al .
2000 ; johnson & bolte 2001 ) showed that the heavy @xmath0-elements in a number of metal - poor stars closely follow the solar @xmath0-pattern .
this comparison between the pattern resulting from a very small number of @xmath0-process events ( only a single event in some cases ) at low [ fe / h ] and the solar @xmath0-pattern resulting from many events indicates that the production of the heavy @xmath0-nuclei is well behaved and not very sensitive to the individual production site .
one possible means of obtaining uniform yield patterns for the heavy @xmath0-nuclei would be fission cycling , which tends to produce a robust pattern with two peaks at @xmath132 and 195 , respectively ( e.g. , freiburghaus et al . 1999 ) .
the observational data do not require an invariant yield pattern of the heavy @xmath0-nuclei .
for example , the th / eu ratio in cs 22892052 differs from that in cs 31082 - 001 by an amount that is too large to be explained by the possible age difference between the two stars ( qian 2002 ) .
however , if th is produced but if fission cycling does not occur , the initial ratio of neutrons to the seed nuclei for the @xmath0-process is still rather restricted .
this limits but does not fix the resulting yield pattern . it may be possible to obtain a regular but not fixed pattern for the heavy @xmath0-nuclei including th without fission cycling ( e.g. , qian 2002 ) . the basic physics of neutron star formation and neutrino emission is essentially the same for both sne ii and aic events . neither process is sufficiently understood to provide a reliable prediction for @xmath0-process nucleosynthesis .
it is hoped that a focused effort on aic models might be fruitful in exploring the problem of @xmath0-process nucleosynthesis without consideration of the complexity of sn ii dynamics .
the neutrino - driven wind associated with aic events requires further studies . among the important physics issues ,
effects of neutrino flavor mixing ( e.g. , caldwell , fuller , & qian 2000 ) and neutron star magnetic field on the conditions in the wind deserve special attention . in conclusion
, we consider that a strong and coherent case can be made for the production of the heavy @xmath0-nuclei by aic events in binary systems based on observations of metal - poor stars with [ fe / h ] @xmath3 . more specifically , the production of the heavy @xmath0-nuclei occurs in the neutrino - driven wind from the neutron star produced by the aic of a white dwarf in a binary system .
this scenario provides a coupling between the @xmath0-process and the @xmath4-process as the progenitor for the white dwarf may produce @xmath4-process elements during the agb phase .
whether this scenario with aic may also be the dominant source for the heavy @xmath0-nuclei in the galaxy remains to be investigated .
with regard to our earlier model , the assignment of heavy @xmath0-element production to sne ii(@xmath28 ) must be changed to the aic events .
the requirement of a major change in the mode of star formation at [ fe / h ] @xmath3 and the decoupling of the production of the heavy @xmath0-nuclei from that of fe appear to be firm .
prior to [ fe / h ] @xmath3 being obtained in the ism , vmss with very large dilution masses played a dominant role in chemical evolution .
the vms activities ceased and major formation of normal stars took over at [ fe / h ] @xmath3 .
the evolution of neutron - capture elements at @xmath133}<-1 $ ] must be reinvestigated now that some coupling between the @xmath0-process and the @xmath4-process has been established . in particular
, our previous assumption of no significant @xmath4-process contributions at [ fe / h ] @xmath130 and our inferred modifications of the @xmath4-process and the @xmath0-process components of the solar abundances must be reevaluated .
we thank gerry brown for calling our attention to the important issue of lmxbs , petr vogel for discussion of supernova neutrino detection , and an anonymous referee for helpful comments .
this work was supported in part by doe grants de - fg02 - 87er40328 and de - fg02 - 00er41149 ( y.z.q . ) and by nasa grant nag5 - 11725 ( g.j.w . ) , caltech division contribution 8899(1101 ) .
equation ( [ eh ] ) can be derived from a simple model that treats the evolution of the abundance of element e in the ism .
consider a parcel of matter initially consisting of gas only .
we assume that the gas of the parcel is well mixed and its amount can only decrease at subsequent times due to astration or escape from the parcel . at time @xmath134 , the total numbers of e and hydrogen atoms in the gas of the parcel are ( e ) and ( h ) , respectively .
the evolution of ( e ) is governed by @xmath135 where @xmath136 is the number of e atoms injected into the gas per unit time from prototypical sources of type @xmath137 , and @xmath138 represents the total rate of loss of gas due to astration and escape from the parcel .
equation ( [ pe ] ) can be rewritten as @xmath139 where the yield @xmath140 is the number of e atoms produced by a single event of type @xmath137 per h atom in a standard reference mass of gas which will mix with the nucleosynthetic products from such an event .
this yield is assumed to be the same for all such events . in equation ( [ ye ] ) , @xmath141 is the number of the events of type @xmath137 that have occurred in the standard reference mass of gas up to time @xmath134 .
equation ( [ ye ] ) can be solved to give @xmath142 where ( e / h)@xmath143 is the initial value and corresponds to the prompt inventory of e that exists in the gas before normal astration takes place .
equation ( [ eh ] ) is obtained for a single type of source when the prompt inventory is overwhelmed by the subsequent production .
if two parcels of matter with different evolutionary histories are mixed , then the equation for ( e / h ) is of the same form as in equation ( [ ehi ] ) but @xmath141 becomes the weighted value @xmath144 for the mixture .
if matter with no enrichment is mixed with an ism that has been enriched in e , then both the resultant @xmath144 and ( e / h ) are decreased .
it is not possible to decrease @xmath144 and keep ( e / h ) fixed .
the abundance patterns resulting from mixed or unmixed parcels is unchanged as long as the yield pattern of each type and the ratios @xmath145 for all contributing types are fixed .
see qian & wasserburg ( 2001b ) for a full treatment .
if we consider infall of unenriched gas during the evolution of a parcel , then the term @xmath146 in equation ( [ pe ] ) is replaced by @xmath147 with @xmath148 being the rate of infall .
this gives @xmath149 the solution to equation ( [ inf ] ) depends in detail on the rate of infall and can be formally written as @xmath150 compared with the case without infall , equation ( [ ehinf ] ) gives a lower value of ( e / h ) for a given @xmath141 or requires larger values of @xmath141 for a fixed ( e / h ) .
however , the ratio of the @xmath141 values corresponding to two different values of ( e / h ) may not be significantly affected by infall . the argument presented in the text below equation ( [ eh ] )
is based on the ratio @xmath151 , and therefore , is not substantially altered .
the basic problem discussed there has to do with the observations that different metal - poor stars with essentially the same [ fe / h ] have widely varying inventories of heavy @xmath0-nuclei . in some cases ,
the observed abundances of these nuclei are so high that they can not be attributed to enrichments of the ism from which the stars were formed but must have resulted from surface contamination in binary systems . | the stellar sites and the complete mechanism of @xmath0-process nucleosynthesis are still unresolved issues . from consideration of the observed abundances in metal - poor stars
, it is proposed that the production of the heavy @xmath0-process nuclei ( @xmath0-nuclei with mass numbers @xmath1 ) is not related to the production of the fe group elements or of the elements with lower atomic numbers : na , mg , al , si , ca , sc , and ti . this requires that the production of the heavy @xmath0-nuclei not occur in supernovae with extended shell structure , but be associated with either bare neutron stars or type ii supernovae ( sne ii ) in the mass range @xmath2 .
from the observations of stars with [ fe / h ] @xmath3 but with high abundances of @xmath0-elements , it is clear that these @xmath0-process enrichments can not represent the composition of the interstellar medium from which the stars were formed , but must represent very local contamination from binary companions .
further evidence for very high enrichments of @xmath4-process elements in metal - poor stars also requires binary systems for explanation .
we propose that the accretion - induced collapse ( aic ) of a white dwarf into a neutron star in a binary system may be associated with the production of the heavy @xmath0-nuclei and may provide occasional coupling of high @xmath0-process and high @xmath4-process enrichments in the envelopes of low - mass stars with low [ fe / h ] .
if we assume that the bulk of the heavy @xmath0-nuclei are produced in aic events , then these events would have produced @xmath5 neutron stars in the galaxy .
a much larger number of white dwarf binaries would have resulted from the evolution of other binary systems .
the aic scenario removes the assignment in our earlier model that sne ii provide the bulk of the heavy @xmath0-nuclei and relegates @xmath0-process production in sne ii to the light @xmath0-nuclei with @xmath6 .
this new assignment removes the requirement in our earlier model that most sne ii produce no fe and gives fe yields that are in accord with the observed values for most sne ii . |
comets are thought to be largely unaltered reservoirs of leftover material from the primordial mixture of dust in the solar nebula orbiting the newly - formed sun ( e.g. , * ? ? ?
spectral observations of comet hale - bopp by @xcite using the _ infrared space observatory _ ( _ iso _ ; * ? ? ?
* ) allowed for detailed spectral models over a wide range of wavelengths @xcite . the amorphous dust of pyroxene ( [ mg , fe]sio@xmath1 ) or olivine ( [ mg , fe]@xmath2sio@xmath3 ) stoichiometry ( henceforth , `` amorphous pyroxene '' or `` amorphous olivine '' , respectively ) was found to require a stoichiometric ratio of magnesium to iron of about unity , being relatively iron - rich @xcite ; the crystalline pyroxene must be iron - poor , as it is cooler @xcite ; and the larger grains must be porous , fluffy aggregates of smaller grains @xcite .
interplanetary dust particles ( idps ; * ? ? ?
* ) , believed to originate from comets , were often found to be highly porous aggregates of mg - rich crystalline silicate dust , having qualities very similar to those inferred from comet spectra and spectra of young stellar objects ( ysos ; systems in the process of forming stars and planets ) . using increasingly powerful telescopes and sensitive detectors ,
astrophysicists have obtained increasingly higher - quality spectra for decreasingly fainter objects . @xcite
distinguished 10 @xmath0 silicate complexes in absorption versus emission , while @xcite identified amorphous silicates , forsterite , enstatite , and silica using their spectrum of the 10 @xmath0 complex of hen 3 - 600a .
approaches toward analysis of the dust composition of protoplanetary disks in herbig ae / be and t tauri ( sed class ii yso ) systems have taken two forms : computation of indices from flux ratios over the 10 @xmath0 ( and sometimes longer wavelength ) complex , or detailed modeling using opacities of minerals and amorphous dust measured on earth .
studies taking the former approach include @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite .
studies taking the latter approach include @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite .
the infrared spectrograph ( irs ; * ? ? ? * ) on board the _ spitzer space telescope _
@xcite allows good s / n spectra to be obtained at 5 - 37 @xmath0 wavelength for objects with flux densities of a few tens of mjy .
dust at a considerable range of temperatures in protoplanetary disks gives rise to the emission seen from such systems in irs spectra .
our approach has been to model this using two temperatures ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , though others have used more temperatures ( see * ? ? ?
here we model the dust emission over the 7.7 - 37 @xmath0 wavelength range for 65 class ii ysos in the 1 - 2 myr old taurus - auriga star - forming region @xcite using dust at two temperatures , in order to analyze the dust composition in the inner and outer disk regions .
`` crystalline indices '' computed from flux ratios that characterize crystalline silicate emission have the advantage of being quickly computed and measured repeatably for a wide range of objects .
however , they do not make full use of the information on dust emissivity available in a high - quality mid - infrared spectrum ; for instance , they measure the combined effect of grain growth and crystallinity , but such are difficult to disentangle @xcite without detailed modeling using dust opacities , as is done here .
much of the process of data reduction of the 65 spectra analyzed in this study has been described by @xcite and @xcite . here
we give a brief overview and provide details specific to this study .
we note that the spectra of zz tau and roxs 42c used here are exactly the same as used and described by @xcite ; the following discussion of data reduction applies to all spectra except those of zz tau and roxs 42c .
sixty - three of the 65 objects in our sample come from the 85 class ii young stellar objects ( ysos ) analyzed by @xcite .
the other 22 objects whose spectra were analyzed by @xcite were rejected from our study due to high ( here , defined as having a@xmath4@xmath56 ) or unknown extinction , the presence of polycyclic aromatic hydrocarbon bands , saturation of the detectors , and confusion of class ii ysos with class i ysos in the same spectrograph slit .
to the 63 from @xcite we add gg tau b and haro 6 - 28 . for opacity calibration ( see section 3 ) , we reduced data for roxs 42c ( see @xcite ) , hbc 656 , and tw cha .
all of these objects were observed using the combinations of irs modules short - low and long - low or short - low , short - high , and long - high in either staring mode or mapping mode ; for more information , see @xcite .
we obtained basic calibrated data products from the _ spitzer _ science center ( ssc ) for each of the targets in our sample .
bcd data are corrected for the dark current , stray light , and flatfield variations .
the spectrum of hbc 656 was reduced using data from the s16.1.0 data calibration pipeline .
the spectrum of zz tau was reduced using data from the s14.0.0 irs data calibration pipeline , using the methods described by @xcite ; the present discussion of extraction and calibration of spectra excludes the spectrum of zz tau . for the spectra of tw cha and the 64 t tauri stars in taurus - auriga excluding zz tau , the data are from the s13.2.0 data calibration pipeline .
next we identified and fixed by interpolation in the spectral direction bad pixels , as described by @xcite and @xcite .
we then extracted spectra using the spectral modeling , analysis , and reduction tool ( smart ; * ? ? ?
we sky - subtracted data from the low spectral resolution modules sl and ll using the off - order observation .
we did not subtract sky from high spectral resolution ( sh and lh ) observations , as explained by @xcite .
multiple data collection events ( dces ; detector samplings ) at the same telecope position , same order , and same modules were averaged , and the uncertainties were calculated from the standard deviation of the mean . tapered column extraction was used for sl and ll data , and full - column extraction was used for sh and lh data @xcite .
we used relative spectral response functions ( rsrfs ) to calibrate the flux @xcite .
for both orders of sl and for the both orders of ll ( for all observations except for hbc 656 ) , a spectral template of @xmath6 lacertae ( a1v ; m. cohen 2004 , private communication ) of higher spectral resolution than the templates described by @xcite was used . for the ll flux calibration of hbc 656 over second , bonus , and first orders out to 36 @xmath7 m wavelength , rsrfs generated from data of @xmath8 dra and the template for @xmath8 dra from @xcite were used .
past 36 @xmath7 m , a ll first - order rsrf generated from data of markarian 231 and the template for markarian 231 ( j. marshall , private communication ; * ? ? ?
* ; * ? ? ?
* ) were used . for sh and lh , the spectral template for @xmath8 dra ( k2iii ) by @xcite and its irs data
were used to generate rsrfs . except for gg tau a
, gg tau b , zz tau , roxs 42c , hbc 656 , and tw cha , all rsrf - calibrated spectra are the same as presented by @xcite .
we rebin our high - resolution spectra ( see table a1 for whether the spectra were obtained with high or low spectral resolution ) to low spectral resolution .
the width of each bin is determined by requiring spectral resolution r=@xmath9/@xmath10 @xmath11 90 , and rebinned spectrum is sampled such that each bin is @xmath10 from the next bin .
we corrected our spectra for extinction using the same method described by @xcite , except we assume a ratio of v - band extinction to total optical depth at 9.7 @xmath0 , a@xmath4/@xmath12 of 25 .
this value is suggested by data plotted by @xcite .
we use values of a@xmath4 given by @xcite .
we do not correct for extinction if a@xmath4 @xmath13 1.4 , as we want to minimize any artifacts due to overcorrection of extinction ( see * ? ? ?
coku tau/4 , gg tau a , gg tau b , and haro 6 - 28 are not corrected , but v410 anon 13 is , assuming a@xmath4 = 5.8 ( see * ? ? ? * ) .
the extinction correction for roxs 42c is the same as made by @xcite .
the average ( v - r ) colors as measured by @xcite of hbc 656 ( also known as as 216 ) are @xmath11 0.64 , which is slightly bluer than the intrinsic ( v - r ) color of a k2 star ( the spectral type assigned to hbc 656 by * ? ? ?
* ) of 0.74 given by @xcite .
this is consistent with hbc 656 being of slightly earlier spectral type ( e.g. , k0 ) and having a@xmath4 @xmath11 0 , so we assign a@xmath4 = 0 to hbc 656 . finally , we compute a@xmath4 = 1.78 for tw cha using i and j magnitudes from denis @xcite and intrinsic colors for a k4 main - sequence star from @xcite , assigning a spectral type of k4 to tw cha despite @xcite adopting k8 for tw cha after having found various classifications for this star in the literature of k0 , k8 , and m0 . next , we corrected for slight mispointing of the telescope from the standard nod positions by scaling the mispointed nods , as described by @xcite .
the scalars used are given in table a1 and further described in the appendix .
after applying these scalars , the spectral uncertainties were computed . for all objects except zz tau ,
the flux density uncertainties were half the difference between the spectra obtained at the two nod positions ( see * ? ? ? * ) .
the uncertainties were derived from only two independent measurements of flux ( except for zz tau ) , so sometimes the relative uncertainty was less than 1% . because of this and because of uncontrolled systematic errors , when the relative uncertainty was less than 1% , it was set to 1% as described by @xcite .
the spectra plotted by @xcite for coku tau/4 , dm tau , and gm aur showed very little excess emission above the stellar photosphere at @xmath9 @xmath13 8 @xmath0 ; for these , we subtracted planck functions using solid angles and effective stellar temperatures given by @xcite for the same purpose .
photospheres were subtracted for zz tau and roxs 42c in exactly the manner described by @xcite .
for all other spectra , the stellar photosphere contribution to the spectra appears minor , and stellar photospheres are not subtracted before modeling them .
to characterize the silicate dust composition of the upper levels of the t tauri disks giving rise to the emission features in our spectra , we construct detailed models @xcite .
briefly summarized , each model is a sum of emission from blackbodies at temperatures t@xmath14 ( warm ) and t@xmath15 ( cool ) and emission from optically thin dust grains proportional to their opacities multiplied by planck functions at t@xmath14 and t@xmath15 : @xmath16 + \nonumber \\ & & b_{\nu}(\lambda , t_{w } ) \left[{\omega}_{w}+\sum_j a_{w , j}\kappa_{j}({\lambda})\right ] ~~,\end{aligned}\ ] ] where b@xmath17(@xmath9,t ) is the planck function evaluated at wavelength @xmath9 and temperature t , @xmath18 ( @xmath19 ) is the solid angle of the blackbody at temperature t@xmath15 ( t@xmath14 ) , a@xmath20 ( a@xmath21 ) is the mass weight of component j at temperature t@xmath15 ( t@xmath14 ) , and @xmath22(@xmath9 ) is the opacity ( effective cross - section per unit mass ) at wavelength @xmath9 of dust species j. the mass weights , a@xmath20 ( a@xmath21 ) , are equal to @xmath23 ( @xmath24 ) , where @xmath25 ( @xmath26 ) is the mass of dust species @xmath27 at @xmath28 ( @xmath29 ) and @xmath30 is the distance to the t tauri star , which is assumed to be 140 parsecs for t tauri stars in the taurus - auriga star - forming region @xcite .
the blackbody components of each model , @xmath18b@xmath17(@xmath9,t@xmath15 ) and @xmath19b@xmath17(@xmath9,t@xmath14 ) , represent optically thick continuum emission , unsubtracted stellar photosphere emission ( f@xmath17 proportional to 1/@xmath31 , and it should be small for spectra for which we have not already subtracted a planck function at the stellar effective temperature to represent it ) , and emission from grains in the optically thin disk regions whose opacities are featureless continua , like amorphous carbon and very large silicate grains .
the model components proportional to opacity multiplied by a planck function represent emission from submicron- to few - micron - sized dust grains , located in the optically thin disk regions , with strong resonances giving rise to detectable features in their opacities
. we will assume in the following that all grain species in the same location in the disk have the same temperature , the thermal coupling provided by gas .
@xcite find that the layer with a@xmath4 @xmath32 0.1 ( where a@xmath4 is the visual extinction of star into the disk ) , the layer above the layer where dust emission features arise ( what they call the `` superheated surface layer '' ) , has gas and dust temperatures within 10% of each other , suggesting the gas would bring dust grains of differing optical properties to roughly the same temperature . in general , settling of dust towards the disk midplane @xcite should enhance the gas particle density for the optically thin regions in the disk atmosphere that gives rise to the 10 @xmath0 silicate feature because the gas density is higher towards the disk midplane ( see figure 1 of * ? ? ?
* ) , so dust and gas temperatures should be better coupled in more settled disks .
however , @xcite found that , for a fiducial disk at 1 au , the dust grains should remain at their radiative equilibrium temperatures , undisturbed by gas - grain collisions , for the vast majority of plausible gas particle densities , 10@xmath33 to 3@xmath34 @xmath35 , at this distance from the star .
they found that only between about 10@xmath36 and 3@xmath34 @xmath35 would the dust grain temperatures be affected by collisions with gas molecules . at 8 au ,
@xcite find that gas and dust temperatures are only equal for a@xmath4 slightly greater than 1 , suggesting greater difference between dust and gas temperature in the `` superheated surface layer '' than @xcite predict .
for the purposes of dust modeling of t tauri stars , planck functions and opacity - weighted planck functions at two temperatures ( at most ) will be used , while keeping in mind the results of these sophisticated studies of dust and gas coupling .
we allow for the two temperatures in order to study compositional differences of disks between inner disk regions and outer disk regions . if , however , there is not sufficient coupling of dust and gas heating in the optically thin surface layers giving rise to emission features in our tts spectra , this means dust that absorbs starlight poorly at optical and near - infrared wavelengths with respect to mid - infrared wavelengths , like enstatite and forsterite , will be located closer to the star than dust like iron - rich amorphous pyroxene and olivine at the same temperature .
such iron - rich amorphous silicates absorb starlight more efficiently at visible and near - infrared wavelengths with respect to mid - infrared wavelengths @xcite . to determine the best fit of a model to a spectrum , we minimize @xmath37 between the model and the 7.7 - 37 @xmath0 region of the spectrum , with respect to the two temperatures , the solid angles , and the mass weights .
we choose ranges of t@xmath15 and t@xmath14 to explore , and , picking a single pair of [ t@xmath15,t@xmath14 ] , we minimize @xmath37 , @xmath39^{2 } ~~,\end{aligned}\ ] ] with respect to all solid angles and all mass weights .
when this results in mass weights or solid angles that are negative , we eliminate from the model the component whose flux integrated from 7.7 - 37 @xmath0 is most negative , and minimize @xmath37 of the model without this component .
this is repeated at a given temperature pair [ t@xmath15,t@xmath14 ] until there are no negative mass weights .
this entire process is carried out for each temperature pair . the temperature pair for which @xmath37 per degree of freedom ( where the number of degrees of freedom is the number of data points ( @xmath11200 ) minus the number of temperatures used , 2 , minus the number of nonzero solid angles minus the number of nonzero mass weights ) is lowest is the best - fit model for a spectrum .
we also determine uncertainties for our model parameters .
we define the uncertainty in a parameter to be the increment of the parameter from its best - fit model value that changes by 1.0 the value of @xmath37 per degree of freedom , or @xmath40 , of the model with the incremented value of the parameter .
we use a taylor series expansion to approximate this .
we note that our definition of uncertainty in a parameter is similar to what @xcite call `` independent point uncertainty '' in that all model parameters are kept at their best - fit values except for the one whose uncertainty is being determined .
however , our uncertainties are determined differently from how others have been determined in the literature .
@xcite suggest taking the uncertainty to be that increment of the parameter that increases @xmath37 by a number computed from the incomplete gamma function for a given level of significance .
@xcite found that uncertainties computed using the increment in @xmath41 determined in this manner are quite consistent with uncertainties computed from monte carlo analysis .
in such an analysis , the standard deviation in the set of parameter values obtained by fitting multiple versions of the same spectrum , each manipulated by adding noise according to poisson statistics , is taken to be the uncertainty in that parameter .
for example , the model for dg tau has only eight constraints .
the incomplete gamma function implies that an increment of @xmath41 of 9.31 would determine the 68.3% ( one sigma ) uncertainty level on parameters .
this would result in a one sigma uncertainty on the mass weight of cool small amorphous pyroxene of @xmath42 @xmath43 .
this is nearly identical to the uncertainty on this component obtained _ via _ the monte carlo analysis described previously of @xmath44 @xmath43 . using the method adopted in this study , an uncertainty of @xmath45 @xmath43 is obtained , which is almost 5 times higher than the one sigma uncertainties obtained using the incomplete gamma function to determine the increment of @xmath41 or the monte carlo approach .
we note that , since the number of data points ( about 200 ) in one of our spectra is much greater than the number of model constraints ( by a factor of @xmath520 ) , our uncertainty level involves increasing @xmath37 by much more than suggested by @xcite . by so doing , we conservatively bias the uncertainties in our parameter values to many times one sigma .
our models include the opacity of annealed silica from @xcite .
this was shown to provide the best fit to emission features at 9 , 12 , 16 , and 20 @xmath0 in astronomical spectra of tts by @xcite .
we now justify our choice of opacities for submicron amorphous and crystalline silicates and large ( few - micron - sized ) silicate dust grains . in order to choose the best opacity curves to use to model emission from tts
, we assume that all disks in our sample have the same major kinds of dust , but in different relative abundances . assuming all disks have basically the same major kinds of dust - amorphous silicates of olivine and pyroxene composition , crystalline olivines , crystalline pyroxenes , and crystalline silica - our goal is to find the best realistic opacities to represent these major kinds of dust , not necessarily to constrain the specific properties of the grains themselves ( e.g. , sizes , porosities , composition ) .
the first opacity for our models with which we concern ourselves is that of forsterite .
olivines have been used before in modeling dust emission to fit emission ( and sometimes absorption ) complexes at 10.0 , 11.2 , 16 , 19 , 23 , 28 , and 33 @xmath0 ( e.g. , * ? ? ?
one spectrum for which these features appear at very high signal - to - noise ratio ( s / n ) is that of hbc 656 , which we include as our first and most important forsterite exemplar .
others with these features in their spectra at fairly high s / n ratios that we include as forsterite exemplars are dk tau , gn tau , is tau , roxs 42c ( which has already been identified as a silica exemplar ; see * ? ? ?
* ) , v836 tau , and v955 tau . to fit the forsterite features in these exemplars spectra , we have compiled opacities of crystalline olivine from a wide range of sources . from @xcite ,
we include opacities from transmission measurements of the following olivine powders pressed into potassium bromide ( kbr ) pellets : forsterite ( mg@xmath2sio@xmath3 ) , magnesium - rich olivine ( mg@xmath46fe@xmath47sio@xmath3 ) , hortonolite ( mgfesio@xmath3 ) , and fayalite ( fe@xmath2sio@xmath3 ) .
to explore further the effects on opacity of the full range of iron content in the olivine solid series ( olivines of all mg / fe ratios ) , we include 8 opacities measured by @xcite for the range of 0.77@[email protected] for olivines of composition mg@xmath48fe@xmath49sio@xmath3 , and also x=0.218 .
in addition , we created a hybrid forsterite opacity using the x=1 ( `` fo100 '' ) opacity from @xcite for @xmath9 @xmath13 15 @xmath0 and room - temperature forsterite opacity measured by @xcite for @xmath9 @xmath5 15 @xmath0 .
we also include the opacities of forsterite powder measured by @xcite and annealed forsterite by @xcite .
finally , we compute opacities for the continuous distribution of ellipsoids ( see * ? ? ?
* ) and cde2 ( see * ? ? ?
cde2 and cde are distributions of ellipsoids whose size can be described as being in the `` rayleigh - limit '' , or characteristic size ( of the semimajor axes of the ellipsoid ) much less than @xmath50 , with ellipsoids described by every possible set of axial ratios weighted more toward near spherical shapes ( cde2 ) and equally ( cde ) , respectively . for both shape distributions of forsterite ,
we compute opacities using optical constants for the three crystallographic axes of forsterite provided by @xcite . to account for this anisotropy of forsterite
, we compute opacities for forsterite grains in a given shape distribution ( cde2 or cde ) three times , once for each of the 3 crystallographic axes optical constants ; we average these three resulting opacities to account for forsterite grains oriented randomly ( see * ? ? ?
* ) . as a starting mixture ( not the same
as the standard mixture determined and used later ) of opacities for other components to use in modeling our forsterite exemplars , we use the non - forsterite opacities from @xcite ; that is , cde2 opacities for amorphous olivine and amorphous pyroxene using optical constants from @xcite for mgfesio@xmath3 and mg@xmath51fe@xmath52sio@xmath1 , respectively , `` en90 '' enstatite opacity ( mg@xmath53fe@xmath54sio@xmath1 ) from @xcite , and annealed silica opacity from @xcite . for large grains ,
we use the aforementioned amorphous silicate optical properties and mie theory @xcite to compute the opacity of 5 @xmath0 radius spheres that are 60% vacuum by volume , computing effective complex dielectric functions for this porous amorphous silicate material using bruggeman effective medium theory ( emt ; * ? ? ?
the scattering efficiencies of small amorphous and crystalline silicate grains are negligible because they are in the rayleigh limit for _ spitzer _ irs wavelengths and are not included in their opacities .
the scattering component of opacity is not negligible for 5 @xmath0 radius grains .
the scattering cross section is comparable to the absorption cross section in the 10 @xmath0 region for both amorphous pyroxene and olivine dust grains .
since we model the emission from an optically thin region ( the upper disk layers ) , only the absorption component contributes to our model .
full radiative transfer treatments of the emergent intensity from these optically thick disks must include the effects of scattering on the 5 - 37 @xmath0 irs spectra .
we note the detailed radiative transfer models of @xcite , who cite earlier work by @xcite , show that scattering by large grains does not much change the flux from optically thin regions and makes the flux from optically thick regions slightly smaller . for this reason ,
we deem the lack of inclusion of scattering in opacities not of great concern .
we give in table 1 the mass weights in units of our conservative uncertainties , @xmath55 ( a measure of significance ) , of the olivine components for each of the seven forsterite exemplars .
those olivine opacities whose @xmath55 are highest contribute most significantly to the fit of a given forsterite exemplar .
a given entry in a column in table 1 is the mass weight of the component named in the first entry on the same row divided by its uncertainty , giving the significance of a particular opacity used in a given model .
a prefix of `` c '' ( or `` cool '' ) means the opacity was used at the lower temperature , and a prefix of `` w '' ( or `` warm '' ) means the opacity was used at the higher temperature . of the entries
whose names include `` j98 '' ( meaning those opacities were presented by @xcite ) , `` fo '' denotes forsterite , `` ol '' denotes mg - rich olivine , `` ho '' denotes hortonolite , and `` fa '' denotes fayalite .
the subsequent 9 opacities come from @xcite , with the number after `` fo '' indicating the stoichiometric abundance , expressed as a percentage , of mg with respect to fe in the ratio mg/(fe+mg ) for olivine [ mg , fe]@xmath2sio@xmath3 .
entries for fo100 , fo90.7 , and fo77 , and the opacity of forsterite powder presented by @xcite are not included at either temperature because no mass of any of them was ever used in any of the models of the forsterite exemplars .
`` annfor '' denotes the opacity of forsterite presented by @xcite annealed for the longest duration .
`` fohyb '' is the hybrid opacity ( explained in the text ) .
s6 means the optical constants come from @xcite , and cde and cde2 are the shape distributions for which opacities are computed using these optical constants .
we find that a few opacities recur from model to model , and some of those at both model dust temperatures .
arrows in the final column indicate recurring opacities .
forsterite powder measured by @xcite shows up in 4 models as cool dust , but not as warm dust .
our forsterite `` hybrid '' opacity shows up in 3 models as cool dust , but only once as warm dust .
annealed forsterite occurs in 4 models as a warm dust component , but not as a cool component .
fayalite from @xcite occurs as warm dust with frequently high @xmath55 for the 6 exemplars , but as cool dust only once .
the opacity that most repeatedly occurs as both warm and cool dust components in the models of the exemplars is the cde2 opacity using optical constants for forsterite given by @xcite .
frequently associated with this is also the cde opacity for forsterite from the same optical constants .
the best fits seem to come from a combination of cde2 and cde shape distributions of forsterite using optical constants from @xcite , suggesting that a shape distribution intermediate in weighting between cde2 and cde is indicated .
further , we note that the cde shape distribution is somewhat unphysical , requiring ellipsoids either infinitely long or infinitesimally thin .
we therefore computed an opacity curve for a custom - designed shape distribution of ellipsoids intermediate in weighting between that of cde2 and of cde .
we note that a similar attempt was made by @xcite resulting in what those authors called `` modified continuous distribution of ellipsoids '' . as the features for forsterite grains in the cde shape distribution peak at wavelengths slightly longward of the peaks in the data and those of the cde2 shape distribution of forsterite grains peak at wavelengths shortward of the peaks in the data , we expect the intermediate shape distribution to give opacity peaks at more optimal wavelengths than do the cde or cde2 shape distributions .
not knowing exactly how to weight one ellipsoidal shape relative to another , like cde we weight all allowed shapes equally , but unlike cde we do not allow ellipsoidal shapes with l@xmath56 parameter ( see * ? ? ?
* ) equal to zero in the shape distribution parameters are inversely proportional to the semimajor axis of one of the three ellipsoidal axes , there is one l@xmath56 parameter for each of the three axes , and the sum of the l@xmath56 parameters equals 1 .
therefore , two l@xmath56 parameters , l@xmath57 and l@xmath2 , suffice to describe completely the shape of the ellipsoid .
this can be represented graphically with l@xmath57 as the horizontal axis and l@xmath2 as the vertical axis .
any conceivable ellipsoidal shape is then represented on this graph as a point in the first quadrant satisfying l@xmath2 @xmath13 1-l@xmath57 ] .
our restricted shape distribution , which we call the `` truncated continuous distribution of ellipsoids '' , or tcde , weighs all points equally within the bounds of a right triangle on the graph bounded by the three vertices ( l@xmath57 , l@xmath2 ) of ( 0.005 , t ) , ( 0.005 , 0.99 ) , and ( 0.995-t , t ) .
we chose not to search for the best forsterite opacity using the tcde shape distribution of forsterite and an untested enstatite opacity ( en90 ) , but , instead , to search jointly for the best value of the bound , t , on the tcde shape distribution and the best enstatite opacity .
this is because enstatite opacities have features of similar widths and at many of the same wavelengths ( 10 , 19 , 23 , 28 , 33 ) as the opacities of forsterite .
now we focus on choosing the best opacity of enstatite , also searching for the best bound in the tcde shape distribution for forsterite .
@xcite found fn tau to require enstatite to fit narrow features in its 10 @xmath0 complex at 9.3 , 10.6 , 11.2 , and 11.6 @xmath0 . in this study , we are interested in modeling all silicate dust features found in irs spectra , so we look for guidance longward of the 10 @xmath0 complex to the spectrum of fn tau .
fn tau has a very weak 16 @xmath0 feature ( unlike the forsterite exemplars ) , a strong 28 @xmath0 feature relative to both the 23 and 33 @xmath0 complexes , and a double - peaked 33 @xmath0 complex .
we look for these patterns elsewhere in our taurus - auriga sample to find other enstatite exemplars .
dh tau also has a narrow 9.3 @xmath0 feature , strong 28 @xmath0 complex , and double - peaked 33 @xmath0 complex .
we also include haro 6 - 37 and hk tau as enstatite exemplars because both have narrow features at 9.3 , 10.6 , and also between 11 and 12 @xmath0 , though they lack prominent crystalline silicate features at wavelengths longward of the 10 @xmath0 complex . as with the olivines ,
numerous lab measurements of pyroxenes exist .
we include opacities computed from transmission spectra of submicron pyroxene powders in potassium bromide ( kbr ) pellets for orthoenstatite ( mgsio@xmath1 crystals in orthorhombic crystal structure ) , clinoenstatite ( mgsio@xmath1 in monoclinic crystal structure ) , the aforementioned en90 , `` en80 '' ( mg@xmath59fe@xmath47sio@xmath1 ) , `` en70 '' ( mg@xmath51fe@xmath52sio@xmath1 ) , `` en60 '' ( mg@xmath60fe@xmath61sio@xmath1 ) , `` en50 '' ( mg@xmath62fe@xmath62sio@xmath1 ) , and `` en00 '' ( fesio@xmath1 , or ferrosilite ) by @xcite .
we also consider similar measurements of synthetic orthoenstatite and clinoenstatite @xcite ; natural clinoenstatite from akita , japan , orthoenstatite from norway , and synthetic clinoenstatite and orthoenstatite @xcite ; and hypersthene ( mg@xmath62fe@xmath62sio@xmath1 ) , bronzite ( mg@xmath59fe@xmath47sio@xmath1 ) , and clinoenstatite @xcite .
we also compute opacities of orthoenstatite grains in the cde2 shape distribution using optical constants for the three crystallographic axes provided by @xcite . at the same time , for the rest of the dust types in the models , we use the same non - forsterite and non - enstatite opacities as were used in the models of the forsterite exemplars in the previous section , and we explore values of the bound , t , on the tcde shape distribution of forsterite ranging from 0.04 to 0.1 .
we also include the opacity of the cde2 shape distribution of forsterite grains .
we use all of these opacities to model the spectra of the enstatite exemplars fn tau , dh tau , haro 6 - 37 , and hk tau , in addition to the forsterite exemplars .
table 2 gives the significance of the mass weights of the opacities of crystalline pyroxene and crystalline olivine in the models of the enstatite and forsterite exemplars .
the opacities named `` en90 '' , `` en80 '' , `` en70 '' , `` en50 '' , and `` en00 '' were presented by @xcite .
the ones named `` j98cde2 '' use the optical constants for the three crystallographic axes of orthoenstatite presented by @xcite to compute opacity for grains in a cde2 shape distribution .
`` k0so '' is the synthetic orthoenstatite , whose opacity is presented by @xcite .
`` j98b '' and `` j98c '' are bronzite and clinoenstatite opacities , respectively , presented by @xcite obtained via transmission measurements of crushed grains pressed into pellets .
entries for orthoenstatite , clinoenstatite , and `` en60 '' opacities from @xcite , synthetic clinoenstatite from @xcite , and hypersthene from @xcite are not included at either temperature because no mass of any of them was ever used in any of the models of the exemplars of both forsterite and enstatite modeled jointly .
`` s6 '' in the last five rows refers to optical constants presented by @xcite used to compute opacities for the cde2 shape distribution ( the first of the five ) and for the tcde shape distributions with bounding parameter ( see discussion in section 3.4 ) , t , of 0.10 , 0.08 , 0.06 , and 0.04 for the second through fifth of the five opacities . among the pyroxene opacities , cde2 orthoenstatite from optical constants provided by @xcite shows up at significant levels as a cool dust component but only weakly as a warm dust component .
ferrosilite ( en00 ) by @xcite often contributes significantly as a warm dust component but not as a cool component .
the one opacity that contributes significantly most frequently as both a warm and a cool dust component is the clinoenstatite ( enstatite having monoclinic crystalline structure ) en90 from @xcite .
though we use the opacity of enstatite of composition mg@xmath53fe@xmath54sio@xmath1 having slight iron content , the fit at long wavelengths is not ideal and suggests caution in the interpretation of finding crystalline silicates with any iron content .
this opacity was used by @xcite in modeling the 10 @xmath0 complexes of 12 tts and by @xcite in modeling silica exemplars . among the various opacities of forsterite
, the tcde opacity with bound t=0.1 gives the best fit to both the enstatite and forsterite exemplars .
we note the improvement of fit of the spectrum of roxs 42c reported by @xcite when using tcde instead of cde forsterite ( optical constants by * ? ? ?
* ) , with @xmath37 per degree of freedom of 4.3 and 5.1 , respectively .
now we turn to amorphous silicate opacities .
@xcite showed that the transitional disks ( disks whose outer parts are optically thick to mid - infrared radiation but whose inner regions are very optically thin ) coku tau/4 , dm tau , and gm aur were fit well using opacities of mostly submicron amorphous silicates having smooth 10 and 20 @xmath0 features .
lkca 15 was shown by @xcite to be a pre - transitional disk , which are like the aforementioned transitional disks but with more dust in the inner regions .
as with the transitional disks , lkca 15 shows smooth 10 and 20 @xmath0 features .
fm tau , tw cha , and uy aur also show similarly smooth 10 and 20 @xmath0 features .
we consider these seven objects to be amorphous silicate exemplars .
@xcite gives complex indices of refraction for amorphous dust of pyroxene composition ( `` amorphous pyroxene '' ) mgsio@xmath1 , mg@xmath51fe@xmath52sio@xmath1 , and mg@xmath62fe@xmath62sio@xmath1 and also for amorphous dust of olivine composition ( `` amorphous olivine '' ) mgfesio@xmath3 .
we use complex indices of refraction for samples of amorphous forsterite ( mg@xmath2sio@xmath3 ) and amorphous enstatite ( mgsio@xmath1 ) obtained by laser ablation of crystals of forsterite and enstatite , respectively @xcite .
we also use complex indices of refraction for samples of amorphous forsterite and amorphous enstatite prepared by sputtering of samples of mgsi and mg@xmath63si in an atmosphere of 50:50 argon - oxygen @xcite .
additionally , we use optical constants for a sample of amorphous bronzite obtained by quenching a melt of a natural sample of bronzite ( mg@xmath53fe@xmath54sio@xmath64 ) from paterlestein , germany analyzed by @xcite and for amorphous pyroxene of cosmic composition ( ca@xmath65mg@xmath66fe@xmath67sio@xmath64 ; sample `` 1s '' ) by @xcite . for the same reasons provided by @xcite , we assume the amorphous silicate dust grains are in the cde2 shape distribution .
we also compute opacities of 5 @xmath0 radius , 60% vacuum porous grains .
the narrow features in the opacity curves of crystalline grains are very sensitive to the exact details of these grains properties such as shape , composition , porosity , size , etc , as we have already indicated in our exploration of the opacities of submicron crystalline grains used to fit silica , forsterite , and enstatite exemplars .
for this reason , we , as @xcite , do not include opacities of large crystalline grains .
the opacities for large amorphous silicate grains are less sensitive to these details .
large grains ( greater than 1 @xmath0 in characteristic size ) in protoplanetary disks should not be homogeneous .
instead , they should be like interplanetary dust particles , which are heterogeneous aggregates whose components are @xmath11 0.1 @xmath0 in size ( see * ? ? ? * ) and composed chiefly of either amorphous silicate , forsterite , or enstatite ( see the review by * ? ? ? * ) .
how should the opacities of such large heterogeneous grains appear ?
the recent study by @xcite concluded that the opacity of a large heterogeneous aggregate is equal to the sum of opacities of its constituents , the opacities resembling those of homogeneous grains of characteristic size corresponding to the abundance of the particular constituent in the large aggregate .
as an example in their study , they compute the opacity of a large porous aggregate of amorphous silicate , forsterite , and enstatite , the amorphous silicate being the most abundant component of the aggregate .
the opacity of the amorphous component of the aggregate resembled that of homogeneous grains of amorphous silicate of size similar to that of the heterogeneous aggregate , while for the lowest abundance of crystalline grains in the aggregate , the forsterite contribution to the aggregate opacity resembled the opacity of much smaller grains of homogeneous forsterite .
very similar conclusions were also reached by @xcite , finding that the typical grain size of crystalline silicate grains used in their models of seven spectra were submicron , but typical amorphous silicate grain sizes were up to 6 @xmath0 radius ( solid grains ) . because the grains in protoplanetary disks grow primarily by sticking together , it is likely large grains will be heterogeneous , composed of various sub - micron components , similar to the idps .
the small amorphous components are believed to come directly from the ism , with sizes @xmath13 0.25 @xmath0 in radius .
since nearly all our objects show evidence for this amorphous component , one expects a significant amorphous component in the large , porous grains .
furthermore , the amorphous components were present at the initial formative stages of the protoplanetary disks .
on the other hand , the crystalline components were produced later , by as yet unknown processes in the protoplanetary disks ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
thus one would expect amorphous components to dominate the heterogeneous fluffy grains . in this case
, the large grain spectra will resemble the sum of completely amorphous large grains with an admixture of sub - micron crystalline grains .
the models presented here will correctly detect the large grains , through the broadening of the silicate features to longer wavelengths .
any crystalline components of large grains will be modeled as small crystalline grains .
the opacity of a large porous heterogeneous grain with a significant abundance of crystalline components , however , will not be modeled well by our standard set of opacities , especially at longer wavelengths ( @xmath9 @xmath5 20 @xmath0 ) in the irs spectra ( see appendix b ) . because of optical depth effects , this will somewhat underestimate the crystalline mass fraction contributed by large grains . to model our amorphous silicate exemplars
, we use annealed silica , enstatite , and forsterite in the tcde shape distribution ( with bound t=0.1 ) as our crystalline silicate opacities .
we use the aforementioned amorphous silicate optical properties for grains in the cde2 shape distribution to compute opacities of submicron grains of amorphous pyroxene and amorphous olivine and also of 5 @xmath0 radius , 60% vacuum grains of amorphous pyroxene and amorphous olivine . in testing the amorphous pyroxene and amorphous olivine opacities , we chose the material whose opacities of both submicron and 5 @xmath0 grains showed up most frequently as a significant contributor as both warm and cool dust . in table 3
, we list the significance of the mass weights of the amorphous pyroxenes and amorphous olivines used in the models of our seven amorphous silicate exemplars .
`` d95 '' refers to optical constants presented by @xcite , `` d79 '' refers to optical constants presented by @xcite , `` sd96 '' refers to optical constants presented by @xcite , and `` j94 '' refers to optical constants presented by @xcite .
the amorphous olivine `` ol '' from @xcite is mgfesio@xmath3 , and the amorphous olivine from @xcite is amorphous forsterite .
the amorphous pyroxenes with `` py5 '' and `` py10 '' in their names refer , respectively , to mg@xmath62fe@xmath62sio@xmath1 and mgsio@xmath1 presented by @xcite .
the amorphous pyroxene from @xcite is amorphous pyroxene of cosmic composition , and the amorphous pyroxenes from @xcite and @xcite are amorphous enstatite . `` sm '' denotes small rayleigh - limit ( here , submicron ) size grains in the cde2 shape distribution . ``
lg '' denotes large 5 @xmath0 radius 60% vacuum porous spheres .
entries for opacities computed from optical constants for amorphous forsterite from @xcite , amorphous pyroxene of composition mg@xmath51fe@xmath52sio@xmath1 from @xcite , and amorphous bronzite from @xcite are not included at either temperature because no mass of any of them was ever used in any of the models of the amorphous silicate exemplars .
although it is never used as a cool dust component , the amorphous olivine mgfesio@xmath3 from @xcite is the best amorphous olivine because it is a very significant warm dust component in the models of all seven amorphous exemplars .
lkca 15 provides one of the most stringent tests of the amorphous silicate features , as its silicate features are very prominent and have high s / n ratios .
amorphous olivine @xcite almost exclusively fits this spectrum s 10 @xmath0 feature .
amorphous pyroxene can dominate as a cool dust component over amorphous olivine ; for instance , the amorphous pyroxene of cosmic composition @xcite exclusively fits the 20 @xmath0 feature of lkca 15 . although the rest of the amorphous pyroxenes are fairly equal contenders for best amorphous pyroxene opacity , the amorphous pyroxene of cosmic composition by @xcite shows up as the most significant amorphous pyroxene opacity in the most exemplars , so we choose this as our amorphous pyroxene opacity .
the minor modification to the amorphous pyroxene opacity usually improved ( but sometimes made worse ) the @xmath68 by @xmath11 0.2 over the amorphous pyroxene opacity used by @xcite .
the amorphous pyroxene of cosmic composition used in our models has a magnesium - to - iron ratio of 52:45 , while the amorphous olivine used has a mg - to - fe ratio of 50:50 .
these are not far from the mg - to - fe ratio for the `` cosmic '' abundances of these two elements ( mg : fe @xmath11 4:5 , measured by * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we also note the imaginary parts of the complex dielectric functions in the visible and near - infrared regions ( 0.2 - 8 @xmath0 wavelengths ) of the best amorphous olivine @xcite and the best amorphous pyroxene @xcite are very similar to that of `` astronomical silicate '' @xcite , whose near - ultraviolet , visible , and near - infrared imaginary part of the dielectric function was specifically constructed so that astronomical silicate grains would heat to the correct temperatures in the presence of stars .
this was important in attempting to obtain a self - consistent model of the _ spitzer _ irs spectrum of ip tau by @xcite .
we use the best opacities that fit our various forsterite , enstatite , and amorphous silicate exemplars to fit our sample of 65 spectra of tts from the taurus - auriga star - forming region .
the first real test of this ensemble of dust opacities to be used as a standard mixture for dust emission models was to fit the mid - infrared spectrum of the trapezium nebulosity .
the 8 - 13 @xmath0 spectrum of the trapezium @xcite was used by @xcite to derive the emissivity of `` astronomical silicate '' grains , which represent the dust in the interstellar medium .
the 8 - 13 @xmath0 spectrum of the trapezium presented by @xcite was combined with the 16 - 38 @xmath0 spectrum of the trapezium obtained by subtracting 3% of the spectrum of the kleinman - low nebula from the spectrum obtained pointed at the trapezium ( see * ? ? ?
this accounted for contamination by emission from the kl nebula of the wide beam of the detector while pointed toward the trapezium @xcite .
the best - fit model for the trapezium spectrum is shown in figure 1 .
as can be seen , the dominant components used to fit the spectrum are submicron amorphous pyroxene and submicron amorphous olivine .
this is consistent with more recent analyses of dust composition of the ism along the line - of - sight to the galactic center by @xcite , who found their dust models required negligible large grains and very small amounts of crystalline silicates ( less than 2.2% by mass ) .
the parameters for the model shown in figure 1 are given in table 4 .
we describe in the appendix our tests of bruggeman effective medium theory used to compute effective complex dielectric functions and mie theory used to compute opacities of large grains .
we desire to measure the degeneracy between model components .
this is the extent to which one component could be replaced by other components and achieve a similarly good model fit . after computing the best - fit models ( to be described in the next chapter ) for the spectra of fn tau ( an enstatite exemplar ) , is tau ( a forsterite exemplar ) , zz tau ( a silica exemplar ) , and dm tau ( an amorphous silicate exemplar ) , @xmath37 was minimized with respect to the blackbody solid angles and dust mass weights over 7.7 - 37 @xmath0 wavelengths at the two dust temperatures ( found by the best - fit models of these exemplars ) , not eliminating components with negative solid angle or mass weight . in the process of minimizing @xmath37 , a 16@xmath6916 element covariance matrix was computed , each of the rows and columns of which belong to one of the 7 dust species or the blackbody at one of two temperatures .
each of the elements of this matrix is a sum over all concerned wavelengths ( 7.7 - 37 @xmath0 ) of a product .
the product is opacity times the planck function ( or just the planck function in the case of a blackbody component ) specific for a given row times the opacity times the planck function ( or , again , just the planck function in the case of a blackbody component ) specific for a given column , all divided by the square of the flux density uncertainty .
the inverse of this matrix gives the covariance matrix ( see discussion in chapter 14 of * ? ? ?
the diagonal elements of this matrix are the variances of each of the dust components in the model , and the off - diagonal elements are the covariances of the dust component of a given row with the dust component of a given column .
the correlation coefficient , r , is computed for each off - diagonal element by dividing the covariance of that element by the square root of the product of the variances of the two dust components in question .
one variance is the diagonal element of the same column , the other is that of the same row .
two highly degenerate components will be highly anticorrelated , with a correlation coefficient very near -1 .
a correlation coefficient near zero means the two components are not correlated .
in addition , for each off - diagonal element , the probability , p , of a coefficient of equal or greater magnitude being found for a non - correlated data set ( which is the probability of the correlation coefficient having been drawn from a random distribution ; see * ? ? ?
* ) is computed .
p near 0% indicate significant correlation coefficients , and p near 100% indicate insignificant correlation coefficients . here
we discuss the most significantly degenerate component pairs , which are shown in table 5 .
the most negative r s are for the pair of cool large amorphous olivine and cool large amorphous pyroxene . for this pair ,
r is between -0.87 and -0.89 ( and p=0.0% ) for the four representative exemplars fn tau ( enstatite exemplar ) , is tau ( forsterite exemplar ) , zz tau ( silica exemplar ) , and dm tau ( amorphous silicate exemplar ) , indicating the most degeneracy between components .
warm large amorphous olivine and warm large amorphous pyroxene are also fairly highly anticorrelated ( degenerate ) ; for this pair , r is usually around -0.72 over the sample of 65 spectra .
similarly , cool small amorphous pyroxene and cool small amorphous olivine are fairly significantly degenerate with each other , with -0.82@xmath5r@xmath5 - 0.89 over the four representative exemplars ( p=0.0% ) .
the warm small amorphous pyroxene and amorphous olivine are also highly degenerate , with r usually near -0.73 over the sample of 65 spectra .
cool enstatite is typically fairly significantly degenerate with cool forsterite , with r usually near r=-0.46 over the sample of 65 spectra , as enstatite and forsterite share very similar features in their opacities for wavelengths longward of the 10 @xmath0 complex .
silica also shares a 20 @xmath0 feature close to 19 @xmath0 complex of enstatite , explaining the spread of -0.29@xmath70r@xmath70 - 0.50 ( [email protected]% ) in this pair of components for the four representative exemplars . at warm temperatures
, enstatite is only significantly degenerate with silica , with r for the 65 spectra in the sample usually being about -0.50 , as warm silica and warm enstatite have strong , narrow features at 9.3 and 9.1 @xmath0 , respectively , while warm forsterite is not degenerate with either , not sharing a strong feature with silica or enstatite in the 10 @xmath0 region . for the highly degenerate pairs ,
the dust emission models accurately determine the sum of the masses of the two components , but not so accurately the individual masses .
this indicates the model finding that the inner disk ( warm component ) tends to be dominated by amorphous olivine while the outer disk ( cool component ) shows more amorphous pyroxene is probably not physically real .
this , at least partially , explains the negative masses often found to give the very best fit . by zeroing the negative mass ,
the complementary component s mass is increased to give nearly as good a fit .
figure 2 demonstrates this effect by showing the degeneracy between cool small amorphous pyroxene and cool small amorphous olivine for the model of the trapezium shown in figure 1 . for the best - fit model temperature pair ,
the model is recomputed but without setting any of the components in the model to zero , allowing mass weights to be negative if that results in the lowest @xmath37 per degree of freedom ( reduced @xmath37 ) .
all resulting model parameters are held at their new best - fit values except for the mass weights of cool small amorphous pyroxene and cool small amorphous olivine , which are varied over the ranges indicated on the plot .
the levels of reduced @xmath37 resulting from exploring this range of mass weights of these two dust components are shown as contours on the plot , with the levels of the first three contours indicated .
this contour plot suggests that replacing some amount of one of the cool small amorphous silicates with the same mass of the other gives a very similarly good fit , changing reduced @xmath37 only very slightly .
histograms of the correlation coefficient between four representative pairs of dust components for all 65 spectra in the sample are given in figure 3 .
the histograms all show single , well - defined peaks .
this indicates the correlation coefficients are measuring real degeneracy between model components .
dust emission models are fit to the 65 taurus - auriga spectra , which we show in figures 4 - 14 .
table 6 gives the parameters of these models ( temperatures , blackbody solid angles , mass weights , reduced @xmath37 ) .
@xmath37 is minimized over 7.7 - 37 @xmath0 , so both dust temperatures used in the models are well constrained by relative uncertainties of 10% .
the median high and low model dust temperatures are 545k and 127k , respectively . in the dr tau model , which has dust at precisely these temperatures , most of the 10.0 @xmath0 wavelength flux above that from the blackbodies originates from 545k dust , while most of the 20.0 @xmath0 flux above that from the blackbodies originates from 127k dust . assuming a distance of 140pc to dr tau , assuming all of the optically thin dust flux at 10.0 @xmath0 comes from 545k dust in a disk of total optical depth at 10.0 @xmath0 of 0.1 , and assuming all of the optically thin dust flux at 20.0 @xmath0 comes from 127k dust in a disk of total optical depth at 20.0 @xmath0 of 0.1 , the radii of these two disks in the case of dr tau are * 0.75 au * and * 11.5 au * for the * warm * and * cool * optically thin dust , respectively .
we list at the end of table 6 the best opacities that fit our various forsterite , enstatite , and amorphous silicate exemplars to fit our sample of 65 spectra of tts from the taurus - auriga star - forming region .
the spectra of dm tau , do tau , uz tau / e , xz tau , and zz tau irs in our sample show evidence of high abundances of large grains . figure 4 shows our models of these five spectra , breaking into components the model of uz tau / e as an example of the models .
warm large amorphous olivine dominates the 10 @xmath0 complexes of these five objects spectral models .
the 20 @xmath0 features also are partly fit by this same component , but usually more of the flux required to fit the 20 @xmath0 features comes from cool submicron amorphous pyroxene .
the overall shape of the spectrum of dm tau differs from that of the other spectra in figure 4 because it is a transitional disk @xcite , lacking significant dust closer to the central star than @xmath11 3 au .
we note that these five spectra support our use of opacities of large grains of amorphous silicates and do not require those of large crystalline grains . that our models fit our spectra well is consistent with amorphous silicates being more abundant both as small grains and as constituents in larger grains ( see section 3.4 ) .
we note that significant abundances of large grains have been found for sr20 , a tts in the ophiuchus star - forming region @xcite , cs cha @xcite , and cvso 224 @xcite .
@xcite found a large abundance of large grains around hd 98800 b , a somewhat older ( @xmath11 10 myr old ) yso in the tw hydrae association . large abundances of large grains around other somewhat older stars were found by @xcite and @xcite . in figure 5
we show the spectra of six spectra with prominent forsterite features , five of which are our forsterite exemplars , with the model of is tau broken down into components .
we note in support of the forsterite opacities used in our models that the fits at all wavelengths to the spectra of these stars , dk tau , f04147 + 2822 , gn tau , is tau , v836 tau , and v955 tau are quite good .
all major features and complexes of forsterite at 10.0 , 11.2 , 16 , 19 , 23 , and 33 @xmath0 are fairly well fit .
the spectrum of f04147 + 2822 ( figure 5 ) is especially interesting .
its 10 @xmath0 complex lacks any of the distinctive narrow emission features characteristic of the forsterite , enstatite , or silica and is fit well by a combination of emission from submicron amorphous silicate grains and large amorphous pyroxene . at longer wavelengths ,
however , emission features at 19 , 23 , and 33 @xmath0 characteristic of forsterite dominate and , correspondingly , are fit well by our forsterite profile for grains in a tcde shape distribution .
this suggests a greater abundance of forsterite in the outer disk than the inner disk .
one possible explanation is that this spectrum is a sum of emission from two disks , one with very little crystallinity throughout and blue in continuum color and another disk with a lot of crystallinity throughout the disk and red in continuum color , but this system is not known to be multiple .
if this spectrum is from only one protoplanetary disk , this would contrast with the finding by @xcite of greater crystallinity within 2 au than outside of 2 au in disks around 3 herbig ae / be stars .
however , the two are not inconsistent , as t tauri stars are less luminous than herbig ae / be stars .
two au in haebe stars should correspond to a much lesser radius in disks around t tauri stars .
it suggests that , on average , the regions to which we refer as `` inner '' and `` outer '' disk regions in the population of 65 t tauri stars in taurus - auriga are both outside of the regions in t tauri star disks analogous to the region inside 2 au of herbig ae / be disks .
in some spectra , however , the long - wavelength side of the 10 @xmath0 feature is not fit well .
dd tau , de tau , and v710 tau all share this problem , as is seen in figure 6 , with the model of fx tau broken into its components .
there are many ways by which forsterite opacity features can be centered at longer wavelengths than those in our tcde opacity curve .
grain shape distributions more heavily weighted toward extreme shapes ( extremely flat , extremely elongated ) , larger grains ( or large aggregates of small grains , with a large fraction of these small grains being forsterite ) , olivine grains with nonzero fe / mg ratios , and forsterite grains with greater porosity would all result in forsterite emission features at longer wavelengths than resulted from the forsterite tcde curve ( solid spheres of forsterite give rise to opacity features at shorter wavelengths than those of forsterite in the cde shape distribution , and the opacity curve of 60% porous spheres of forsterite is almost identical to that of forsterite in the cde shape distribution * ? ? ?
* ) . also ,
both coku tau/3 and fx tau ( fig .
4 ) have 23 and 33 @xmath0 emission complexes with emission that extends to longer wavelengths than provided by the submicron forsterite grains in the tcde shape distribution in our models .
that both spectra lack significant 28 @xmath0 complexes indicates the dominant form of crystals giving rise to 20 - 37 @xmath0 emission for these two systems is forsterite and not enstatite .
it should be noted , though , that the 10 @xmath0 complexes in the spectra of both coku tau/3 and fx tau are fit fairly well .
we therefore conclude there is variation in the exact details of the olivine grain populations in the protoplanetary disks in our sample , in terms of average grain shape , grain size , grain composition , and grain porosity . for dd tau , de tau , and v710 tau
, this variation occurs for the warmer forsterite grains , while for coku tau/3 and fx tau , this variation occurs for the cooler forsterite grains .
we note that most of these five also seem to have emission features at 14 @xmath0 wavelength ( see below , section 4.8 ) . whatever gives rise to these 14 @xmath0 features
is not necessarily related to the poor fits to the forsterite features because the poor fitting happens over very limited wavelength ranges of 11 - 11.5 @xmath0 and 23 - 25 @xmath0 .
in addition to the olivine component of our models , the silica component may also vary in details of grain properties among all disks in our sample .
while the silica features at 9 , 12.6 , and 20 @xmath0 in the spectrum of zz tau ( and the other silica exemplars ) are fairly fit well with the annealed silica opacity used , others like the 12.6 @xmath0 features in the spectra of dn tau and fz tau are not quite so well fit ( see figure 7 ) .
fz tau has a single - peaked feature in this wavelength range , but it peaks at 12.45 @xmath0 , shortward of the wavelength at which annealed silica peaks .
this could suggest the polymorph of silica giving rise to this feature is @xmath6-quartz at somewhat elevated temperatures of 500 - 625k ( see discussion by * ? ? ?
dn tau appears to have a double - peaked feature , which would suggest @xmath6-quartz at @xmath11 300k @xcite , but the s / n at these wavelengths is low for this spectrum so its reality is more doubtful ( note the discussion on the effect of unresolved emission lines in this section ) .
according to @xcite , the presence of silica argues for high - temperature processing , as it is not present in the interstellar medium and therefore should not be present as the starting dust mixture as the protoplanetary disk formed from its protostellar envelope .
if @xmath6-quartz can be confirmed , the silica , once formed , must cool slowly enough to allow phase transition from a higher - temperature polymorph , like tridymite or cristobalite , to @xmath71-quartz , then @xmath6-quartz .
we show the model fits to our enstatite exemplars in figure 8 , with the model of fn tau broken into components .
other than a mismatch between model and spectrum for fn tau at 11.6 @xmath0 , the fits to the 10 @xmath0 complexes are generally good .
however , we note the poorness of fit of our models to the longest wavelength complexes ( 28 and 33 @xmath0 ) of the spectra of fn tau and dh tau , our two best enstatite exemplars . the height - to - continuum ratios in our models of these two exemplars for the 28 and 33 @xmath0 complexes are unacceptably low compared to those in the spectra .
we note the progressively increasing height - to - continuum ratio of the 19 , 23 , 28 , and 33 @xmath0 complexes in the spectra of both dh tau and fn tau .
the rest of our sample are sufficiently fit using crystalline dust at two temperatures .
this suggests that a population of submicron enstatite or forsterite dust at a third ( very low ) temperature would be required to give acceptable fits to these spectra .
we note this lack of a third model temperature is not a problem for spectra requiring abundant forsterite ( figures 5 and 6 ) .
note in figure 6 the problem is not insufficient peak - to - continuum ratio for the model forsterite features ; rather , the problem is insufficient width of the features , which is a problem with the forsterite opacity . the lh part of the spectrum of haro 6 - 37 ( 19.3 - 37 @xmath0 wavelength ) suffers from an artifact , in which each of the 10 spectral orders are tilted such that the flux density of their short - wavelength end is higher than it should be and the flux density of their long - wavelength end is lower than it should be . these tilted spectral orders , which could be interpreted to resemble narrow crystalline silicate features longward of 19.3 @xmath0 wavelength , result from an artifact of the reduction of the lh spectrum of this object .
this artifact could be due either to unsubtracted sky emission or to mispointing of the telescope with respect to haro 6 - 37 along or perpendicular to the length of the lh slit .
spectra requiring abundant submicron amorphous silicate grains and model fits to these spectra are shown in figures 9 and 10 .
figure 9 shows the spectra and corresponding models of five of our amorphous silicate exemplars ( see section 3.4 ) and the model breakdown into components for fm tau . in general , from 7.7
to slightly longward of 30 @xmath0 wavelength , the fits of the models to the data are excellent . for gm aur and
especially lkca 15 , the each model fits the continuum at wavelengths greater than 30 @xmath0 with emission from a weak 33 @xmath0 complex of forsterite .
this is an artifact , as these two spectra require a third model temperature to fit the data using emission from a blackbody or amorphous silicates at a very low temperature .
figure 10 shows spectra whose models also require relatively high abundances of submicron amorphous silicates , though they require more crystalline silicates and large grains than those shown in figure 9 ; the model of hq tau is broken into its components as an example .
again , the model fits to the data are very good , with the exception of a slight insufficiency in the peak - to - continuum ratio of the forsterite 33 @xmath0 complex in the model of dp tau .
the spectra of cy tau , df tau , dg tau , dl tau , dq tau , it tau , and v807 tau show 10 @xmath0 complexes of small equivalent width ( see figure 11 ; the model of df tau is broken into components ) .
cy tau , df tau , dl tau , and it tau also have essentially flat spectra over all irs wavelengths , being consistent with the models presented by @xcite and their interpretation of settling of dust grains . settling of dust in a disk results in the disk being flatter and less flared , which gives rise to a bluer spectral continuum color in the mid - infrared and also less equivalent width in the distinctive silicate features .
@xcite noted cy tau for its flat spectral colors over infrared wavelengths .
these spectra are well fit mostly by emission from submicron dust grains and the two blackbody components in the models .
cy tau requires a small amount of silica to fit in the 10 @xmath0 complex and a moderate amount of cool forsterite to fit longer wavelength complexes , especially the 33 @xmath0 feature .
the emission required by large grains in the model for df tau to fit the longer wavelengths of its spectrum is the largest of the five low 10 @xmath0 equivalent width spectra , in addition to small amounts of emission from submicron dust grains .
little emission from forsterite is required to give rise to the 10 @xmath0 feature of dg tau , while a fair amount of submicron amorphous pyroxene is required to fit its mild 20 @xmath0 feature .
the 10 @xmath0 feature of dl tau requires modest amounts of emission from forsterite and silica and lesser amounts from other kinds of dust .
the lh spectrum ( 19.3 - 37 @xmath0 ) of dl tau suffers from an order - tilt artifact as does haro 6 - 37 ( see above ) , limiting the precision with which we can determine the crystallinity and large grain content of the cooler outer disk , but it appears to be optimally fit using a mixture of both amorphous and crystalline grains .
dq tau has subtle features at 9.3 , 9.8 , 10.6 , 11.1 , and 11.6 @xmath0 that require a moderate amount of enstatite .
it tau requires small amounts of many of the submicron amorphous and crystalline grains to fit its miniscule 10 @xmath0 complex , but its broad 20 @xmath0 feature is fit well by submicron amorphous pyroxene .
the same applies to the fit of the 20 @xmath0 feature of v807 tau , though it requires a modest amount of large amorphous pyroxene to fit its 10 @xmath0 feature .
the abundance of crystalline grains with respect to amorphous grains in the models of these five spectra is consistent with the finding by @xcite using all ( and more ) of the spectra analyzed in this study that increased crystallinity accompanies more advanced settling of dust in disks as measured by continuum indices n@xmath72 and n@xmath73 .
it is also noted that these seven spectra with small equivalent width 10 @xmath0 features do not indicate large abundances of large grains ( see table 6 ) .
we explore these issues more in section 5 . in figures 12 - 14
, we show the spectra and corresponding models to the rest of our sample , with eight spectra in each figure arranged top - to - bottom alphabetically by their names . for all of these spectra , the abundances of the various dust species required by the models are mixed .
typically , no one dust type has an abundance required by its model much more than the abundances of any of the other dust types in that model .
the spectrum of aa tau ( and others ) have what appear to be a broad emission feature at 14 @xmath0 , but this probably originates from gas emission @xcite ; also , sh and lh spectra of aa tau and a few other objects show unresolved emission lines from h@xmath2o and oh ( see next subsection ) .
the models of fp tau , gi tau , and vy tau are broken into components in figures 12 , 13 , and 14 , respectively .
note that fs tau and fv tau in figure 13 and gk tau , hn tau , and haro 6 - 28 in figure 14 suffer from the same lh order - tilt artifact as did haro 6 - 37 ( see section 4.4 ) .
also note the high point - to - point noise in the 20 - 25 @xmath0 parts of the ll spectra of ft tau , gg tau b , and gh tau in figure 13 and iq tau in figure 14 .
this `` fringing '' artifact is due to the delamination of the ll order - sorting filter and shows up when the telescope has been mispointed with respect to a source in the direction perpendicular to the slit for a ll observation , with increasing severity of the artifact for greater mispointing of the telescope .
there are some very narrow gaseous emission lines in our spectra .
the spectra of dg tau ( figure 11 ) , dm tau ( figure 4 ) , and fs tau ( figure 13 ) all show 12.8 @xmath0 features arising from [ neii ]
. this can affect identification of the silica component , as noted by @xcite with respect to the modeling by @xcite of tw hya .
the presence of this feature in addition to a narrow feature at @xmath11 12.4 @xmath0 in the sl spectrum of tw hya led @xcite to conclude the presence of @xmath6-quartz , which has a double - peaked feature matching these two wavelengths .
investigation of high - resolution spectra of tw hya showed that these two features belonged not to @xmath6-quartz but to the humphreys-@xmath6 hydrogen line ( hi n=7 - 6 ) at 12.37 @xmath0 and [ neii ] at 12.81 @xmath0 .
c@xmath2h@xmath2 appears in the spectra of coku tau/3 ( figure 6 ) , df tau ( figure 11 ) , and dl tau ( figure 11 ) .
this emission line is located at the more innocuous wavelength of 13.7 @xmath0 @xcite , and does not appear to have significantly affected the dust models of the spectra of these three objects . however , emission from hcn centered just longward of 14.0 @xmath0 @xcite is more problematic .
examples of spectra with this feature are aa tau ( figure 12 ) , bp tau ( figure 12 ) , and it tau ( figure 11 ) .
first , the feature occurs often in spectra taken with the sl and ll modules , which we splice together at precisely 14.00 @xmath0 .
often there is a flux level mismatch between sl and ll at this wavelength , making establishing the reality of this feature difficult .
second , the width of this feature varies .
we have no dust component with an emission feature centered around 14 @xmath0 , so when the feature is wider , more data deviate from our model , worsening the model fit to the data .
thirdly , observations of evolved stars in the large magellanic cloud @xcite sometimes show a feature at @xmath11 14 @xmath0 probably belonging to dust related to silicates .
lastly , there are emission features near these wavelengths in spectra of calcium - aluminum inclusions in meteorites @xcite and in laboratory spectra of the melilite solid series @xcite , making identification of all of the 14 @xmath0 features with hcn emission more ambiguous .
finally , emission lines of h@xmath2o have been discovered in high resolution _
irs spectra of classical ttss in taurus , such as aa tau @xcite .
these lines are found in its sh and lh spectra , especially past 25 um .
many of these lines were found in the _ spitzer _ irs spectrum of the class 0 yso protostar ngc 1333 iras 4b @xcite .
the lines are generally spread out over wavelength , but there are clumps of lines around 24.5 - 25.5 , 27.5 - 28.5 , and 30 - 31 @xmath0 wavelength .
these water lines are unresolved , and our sh and lh spectra ( r @xmath11 600 ) have been rebinned to sl / ll resolution ( r @xmath11 90 ) , so the effect of the lines generally being distributed over all wavelengths in sh and lh is to raise slightly the level of the continuum above the continuum underneath the water lines in the high resolution spectra . the clumps at 25 , 28 , and 30.5 @xmath0 in the original high - resolution spectra , when rebinned , will show up as very slight @xmath11 0.5 @xmath0 wide bumps in the rebinned spectra .
we typically see no prominent crystalline silicate features centered at either 25 or 30.5 @xmath0 in our spectra ( though dh tau has a small feature centered around 30.5 @xmath0 wavelength that has been attributed to enstatite before ; see * ? ? ?
* ) , but we do see many 28 @xmath0 features in our data .
both the opacities of forsterite and enstatite have 28 @xmath0 complexes , the 28 @xmath0 complex of enstatite being more prominent with respect to its 23 and 33 @xmath0 complexes , so the net effect of not accounting for the water emission might be obtaining a enstatite abundance slightly too large .
though we attempted to correct for extinction as explained in section 2 , two spectra proved to be difficult in this regard , those of fv tau ( figure 13 ) and coku tau/3 ( figure 6 ) . for fv tau , a@xmath4=5.33 , and for coku
tau/3 , a@xmath4=5 . both were just under the upper limit of a@xmath4=6 above which we would not correct for extinction and therefore not model , and both spectra show absorption at 15.2 @xmath0 from co@xmath2 ice ; additionally , coku tau/3 has ice absorption features at 6.0 and 6.8 @xmath0 wavelengths as do heavily embedded class i ysos in taurus - auriga @xcite .
though the model fit to the 10 @xmath0 complex of coku tau/3 is pretty good , the same is not true for fv tau .
we suspect this to result from imperfect extinction correction . at longer wavelengths , however , the fit to fv tau ( excluding the region around the 15.2 @xmath0 co@xmath2 ice absorption feature ) is quite good , while the fit to the crystalline silicate complexes at 23 and 33 @xmath0 is only adequate . as any extinction correction at longer wavelengths is less than that for the 10 @xmath0 silicate complex , we attribute this not to a problem with extinction correction , but rather to the opacity used to fit forsterite , which is discussed in the next subsection .
from the models for all 65 spectra , we computed the percentage of mass in a given dust grain species at one temperature out of all mass in dust at that temperature . with these mass percentages ,
we compute histograms displaying both warm and cool dust mass percentages . the histograms for total warm and cool large grain mass fraction , total warm and cool crystalline grain mass fraction , warm and cool forsterite mass fraction , warm and cool enstatite mass fraction , and warm and cool silica mass fraction
are given in fig .
the test described and shown in appendix b of fitting a 5 @xmath0 radius heterogeneous grain profile with our standard model suggests that the abundance of crystalline grains could be underestimated by our standard models by up to 50% .
in addition , it also suggests the large grain abundance could be slightly underestimated by 23% .
the net effect of this bias is to underestimate the true crystalline and large grain abundances .
figure 15a shows that the inner disk regions typically have much larger mass fraction in large grains than the outer disk regions .
the mean and median in the histogram for warm large grains are around 50% ; for cool large grains , they are between 0 and 10 percent , though with a considerable `` tail '' up to 90% . noting that our 5 @xmath0 radius porous grains are very similar in shape of opacity to 2 @xmath0 solid grains @xcite , our average and median warm large grain mass fractions of 44% and 45% , respectively , compare favorably to the finding by @xcite of the average grain size for 1 - 2 myr old systems being between 1.5 and 2 @xmath0 radius ( as probed by the 10 @xmath0 feature ) .
our average and median cool large grain mass fraction were 17% and 0% , respectively .
figure 15b shows histograms of warm and cool crystalline dust mass fraction .
both histograms peak at low percentages of about 5% for warm crystalline dust and about 12% for cool crystalline dust ( each of these percentages being the _ mode _ of the distribution ) , and there is considerable overlap .
the cool crystalline dust histogram extends slightly to higher mass fraction bins , but we note the greater large grain mass fractions for the warmer inner disk regions . the average and median mass fractions for warm crystalline dust were 17% and 11% , respectively , and the average and median mass fractions for cool crystalline dust were 23% and 15% , respectively .
greater large grain mass fractions can lead to slight underestimation of the crystalline abundance , so the amount of crystallinity in the inner disk regions could be higher and closer to that of outer disk regions .
our findings of the mean , median , and mode of the warm crystalline dust mass fractions of 17% , 11% , and @xmath11 5% are consistent with those by @xcite , but they are slightly less consistent with those by @xcite .
@xcite find typical warm crystalline dust mass fractions ( as probed by the 10 @xmath0 feature ) typically to be between 5 and 20% , regardless of system age , while @xcite find typical warm crystalline dust mass fractions to be between 5 and 7% .
though crystalline mass fractions may be greater in the inner disk regions than outer disk regions ( see table 6 ) , the total mass of the cool dust in our models is between @xmath11 10 and 1000 times greater than that of the warm dust in the models .
the total mass in crystalline silicates is often greater for the cool crystalline dust than the warm crystalline dust , as table 6 shows for v955 tau .
the crystalline mass fractions for warm and cool crystals are 24.7% and 20.6% , respectively , being very similar .
however , there is @xmath11 250 times more mass is cool dust than warm , so there is much more mass in 115k crystalline dust than in 488k crystalline dust according to the model .
figures 15c , 15d , and 15e show the enstatite , forsterite , and silica mass fraction histograms . in figure 15c ,
the warm enstatite mass fractions are very slightly displaced to larger values than the cool enstatite mass fractions .
the previously discussed potential for slight underestimation of crystalline grain mass fraction for the inner disks due to greater grain growth reinforces our conclusion that the abundance of enstatite is greater in the inner disk regions .
figure 15d shows that both the warm and cool forsterite mass fraction histograms show a peak in the lowest bin ; however , a large fraction of the cool forsterite distribution extends to greater mass fractions .
first , we note the larger uncertainties in cool forsterite mass fractions than warm forsterite mass fractions ( table 6 ) .
second , we once again note the potential to underestimate inner disk crystallinity .
together , these two considerations suggest the warm forsterite abundance is similar to that of cool forsterite for our sample .
figure 15e shows the same is true for silica as for forsterite ; for the same reasons as forsterite , the warm and cool silica abundances in our sample are similar .
the higher relative abundance of enstatite over forsterite in inner disk regions than for outer disk regions was also found by @xcite . in order to search for correlations between pairs of model parameters , stellar properties , and disk properties ,
we have computed the linear correlation coefficient , r @xcite , and its corresponding probability , p , of finding a linear correlation coefficient of magnitude greater than or equal to these coefficients on a null data set , as above ( section 3.5 ) .
the correlation coefficient is weighted by uncertainties which are , for one data point , the square root of the sum of the squares of the uncertainties of the two quantities for which the correlation coefficient is being computed ( see discussion on weighted least - squares fitting with uncertainties in both x and y by * ? ? ?
note that scaling all dust mass weights by the same scalar less than unity would not change the correlation coefficient between two dust mass fractions .
the uncertainty weighting of the correlation coefficient shows up in both numerator and denominator of the correlation coefficient @xcite to the same power , so the scalar applied in the denominator would divide the scalar applied in the numerator to give 1 times the original correlation coefficient . note that if r and the number , n , of data points that are being tested for correlation do not change , p will not change either because p depends only upon r and n. values of r that significantly deviate from zero and values of p close to zero suggest significant correlation or anticorrelation .
we deem as significant [email protected]% , resulting in @xmath74r@xmath74 @xmath70 0.29 .
stellar masses , disk - to - star mass ratios , n@xmath72 and n@xmath73 , and their uncertainties , plus multiplicity counts , stellar luminosities , and mass accretion rates used in this correlation analysis come from @xcite .
submillimeter slopes were obtained from the study by @xcite .
first we explore trends between dust mass fractions .
all trends between crystalline components are positive .
figure 16 shows a positive correlation between warm enstatite and warm forsterite with correlation coefficient r=0.62 and probability p of having been drawn from a random distribution of 0.0% .
this trend is fairly clear , with a heavy concentration of points at low mass fractions and a collection of points with higher enstatite and forsterite mass fractions .
warm enstatite also has a correlation with warm silica , with r=0.31 and p=1.2% .
figure 17 shows a positive correlation between cool forsterite and cool silica , with r=0.29 and p=2.0% .
cool forsterite also correlates with warm large grain mass fraction ( r=0.32 and p=0.9% ) and warm crystalline grain mass fraction ( r=0.32 and p=1.0% ) .
the correlation between warm silica and cool crystalline ( cool enstatite plus cool forsterite plus cool silica ) mass fractions is shown in figure 18 , with r=0.32 and p=0.9% .
this correlation is likely related to the correlation between warm silica and cool forsterite ( r=0.36 and p=0.4% ) .
warm enstatite also correlates with cool forsterite ( r=0.29 and p=1.9% ) .
cool forsterite mass fraction and submillimeter slope ( @xmath6 reported by * ? ? ?
* ) correlate , with r=-0.40 and p=2.0% .
lower @xmath6 means flatter submillimeter slope ( flux density at submillimeter wavelengths , f@xmath75 , is proportional to @xmath76 ) , implying that more growth of grains to millimeter sizes ( instead of sizes of a few microns ) accompanies a greater abundance of cool forsterite .
the general sense of these trends is that the crystalline dust abundances all correlate positively with each other .
if the abundance of one type of crystal is high , it is likely that the abundances of the other two types will also be high .
the general finding that crystal abundances track other crystal abundances has been noted before .
@xcite , _ via _ detailed modeling of the 10 @xmath0 complexes of herbig ae / be stars , noted correlation between forsterite and silica abundance in herbig ae / be stars , and @xcite ( using a similar analysis ) noted correlation between enstatite abundance and total crystalline abundance in a similar population of herbig ae / be stars .
@xcite measured crystalline abundances using indices computed from ratios of flux integrated over small wavelength bands characteristic of emission from pyroxene , olivine , and silica at 10 and 33 @xmath0 ( they call these bands p@xmath77 , o@xmath77 , s@xmath77 , and o@xmath78 , respectively ) ; in their study , they found warm pyroxene correlated with warm silica and warm olivine , and that warm olivine correlated with warm silica and cool olivine . from this
we conclude , as did @xcite , that whatever produces the different kinds of crystals ( forsterite , enstatite , and silica ) produces them at a rate faster than the crystals can transform from one to another . as a specific example : if some silica forms by incongruent melting of enstatite , that amount of silica formed must be much less abundant than any silica formed by another means ( e.g. , _ via _ incongruent melting of amorphous pyroxene ; * ? ? ?
there is no significant correlation of crystalline silicate masses with disk - to - star mass ratio .
a positive correlation of disk - to - star mass ratio might be expected for crystalline dust produced by shock annealing @xcite .
the lack of this correlation does not necessarily mean shock annealing is not connected to dust processing , however , as disk mass estimates are based on assumed opacities of dust at submillimeter wavelengths .
the amount by which these assumed opacities differ from actual opacities varies according to the extent of grain growth to millimeter sizes ( see discussion by * ? ? ?
@xcite computed and interpreted the sed continuum indices n@xmath72 and n@xmath79 as primarily measuring the degree of flaring of disks .
indices that are increasingly positive indicate increasingly flared disks , and more negative indices indicate increasingly flatter disks .
the flattening of these disks was interpreted to mean that dust had settled from high in the disk atmosphere towards the disk midplane .
@xcite used n@xmath73 instead of n@xmath79 to reduce contamination of the indices by the 20 @xmath0 amorphous silicate feature .
we use the n@xmath72 and n@xmath73 indices , interpreting them as primarily measuring the degree of flaring of disks @xcite .
we omit from our searches of correlation of parameters with disk continuum indices the points for the transitional and pre - transitional disks coku tau/4 @xcite , dm tau and gm aur @xcite , and lkca 15 @xcite .
these four objects have spectra whose continuum colors are very red due not to highly flared disks but , rather , to the clearing of almost all small dust grains in the innermost regions of the disks .
the disk indices for these four systems would measure how clear they are of such dust in their inner regions , so they are not included in the search for trends of parameters with disk indices .
figure 19 illustrates a tight anticorrelation between warm crystalline grain abundance and n@xmath72 , with r=-0.50 and p=0.0% .
this can be explained by correlations between each of the three warm crystalline types of grains and n@xmath72 .
warm forsterite and n@xmath72 are anticorrelated with r=-0.42 and p=0.1% , warm enstatite and n@xmath72 are correlated with r=-0.41 and p=0.1% , and warm silica and n@xmath80 are correlated with r=-0.39 and p=0.2% .
cool enstatite correlates with n@xmath73 ( r=-0.31 and p=1.9% ) .
these findings are similar to those of @xcite that greater warm olivine and warm silica , measured by o@xmath77 and s@xmath77 , respectively , anticorrelate with n@xmath72 and that greater warm enstatite , warm olivine , warm silica , and cool olivine ( warm enstatite and cool olivine being measured by p@xmath77 and o@xmath78 , respectively ) anticorrelate with n@xmath73 .
the spectra described in section 1.6 ( low 10 @xmath0 complex equivalent width ) all indicate , with the exceptions of dg tau and v807 tau , highly settled systems with abundant crystalline silicates , though with relatively large uncertainties on these crystalline abundances .
the settling of disks is interpreted here , as did @xcite and @xcite , as part of the evolution of protoplanetary disks .
the processing of dust into crystalline silicates is also interpreted as evolution of such disks .
further , the models by @xcite predict crystalline mass fraction should be correlated with dust sedimentation both in inner and outer disk regions .
therefore , the correlation of crystalline abundance with disk settling is expected as a general result of protoplanetary disk evolution .
warm and cool large grain abundances both anticorrelate with n@xmath73 .
figure 20 shows the correlation between warm large grain abundance and n@xmath73 ( r=-0.36 and p=0.5% ) .
this trend is interpreted as an increase in large grain abundance expected as a result of evolution of protoplanetary disks .
figure 21 shows that cool large grain abundance also anticorrelates with n@xmath73 ( r=-0.32 and p=1.5% ) .
@xcite found other tentative evidence for dust settling with grain growth .
grain growth is , in fact , expected to be the cause of settling of dust in protoplanetary disks . from first - principles modeling , @xcite find in his simulations of grain growth and settling towards disk midplane of protoplanetary disks that at 60,000 years of evolution in the disk ( the initial diameter of their dust grains was 1 @xmath0 ) , at 30 au from the central star they find particles with diameters from one @xmath0 to near one centimeter .
the range of ages assumed for class ii ysos in the taurus - auriga star - forming region is 1 - 2 myr , so substantial grain growth should have occurred for the 65 objects whose spectra are analyzed in this study .
also , at a disk age of 0.1 myr , @xcite predicts that at a disk radius of 30 au , the dust - to - gas mass ratio at 6 au above the midplane should be 10@xmath81 what it was at time zero , when the disk was initially well - mixed , at the same location in the disk .
this ratio of the dust - to - gas mass ratio to its initial well - mixed value is what @xcite call the settling parameter , @xmath82 . as shown by @xcite ,
the disk indices of some of the class ii ysos in the taurus - auriga star - forming region are consistent with such low values of @xmath82 ( 0.01 - 0.001 ) .
@xcite also find that the differing morphology of irs spectra of ysos are consistent with dust growth and settling .
large grain abundance anticorrelating with n@xmath73 is consistent with the expectation by @xcite of increasing grain growth over time and also greater settling of larger dust grains towards the disk midplane . also significant
is the result shown in figure 15a that more large grains are found in inner disk regions than outer disk regions .
the observations and models by @xcite required more large grains for disk regions inside 2 au than for disk regions outside 2au for spectra of two of the three herbig ae / be stars they studied .
@xcite find that the surface density of protoplanetary disks decreases with radius in the disks as r@xmath83 or r@xmath84 , so inner disk regions should be denser .
this could encourage more rapid grain growth . however , as noted previously , the correlations between large grain abundances and n@xmath73 are not very strong .
one possible explanation for the weakness of correlation we find between warm large grain abundance and n@xmath73 is that we are not measuring the true large grain abundance from each spectrum because we do not account for differing amorphous - to - crystalline component ratios in the large inhomogeneous aggregate grains from system to system .
the more crystalline components the average aggregate large grain in a given disk has , the less reliable a measure our large amorphous silicate grains are of the large grain extent in such disks @xcite .
another possibility to explain the weakness of trend of warm large grain abundance with n@xmath73 is that sufficiently high turbulence prevents rapid grain growth and settling of few - micron - sized grains to the disk midplane @xcite .
perhaps different disks can have the same degree of flattening but differing abundances of large grains in the optically thin surface layer .
@xcite note that collisions replenish small grains .
however , @xcite finds that turbulence can also encourage collisions that facilitate grain growth .
these large grains , once formed , can be stored in a dead zone underneath the turbulent live zone , and can be protected even if the dead zone is disturbed by the live zone above it @xcite .
warm large grain abundance also correlates with the known number of stars in the star system , with r=0.33 and p=0.7% , so perhaps multiplicity contributes to the dispersion seen in the trend between warm large grain abundance and n@xmath73 .
figure 22 shows a comparison of the histogram of warm large grain abundance for single star systems and the same for multiple star systems - one immediately notices the higher large grain mass fractions for multiple systems than single systems .
finally , figure 23 shows a fairly weak anticorrelation between warm large grain mass fraction and stellar mass , with r=-0.36 and p=0.5% .
this may be related to the correlation between warm small amorphous grain mass fraction and stellar mass ( r=0.40 and p=0.1% ) . in our sample ,
the spectra of low mass stars indicate higher levels of warm large grain abundance than for higher - mass stars .
@xcite found in their study of six brown dwarf spectra higher mass fractions of crystalline grains for lower mass stars .
they concluded that the region giving rise to the 10 @xmath0 feature was at smaller disk radii for less luminous , lower mass stars , and that in these innermost disk radii the crystalline dust abundance was higher .
we interpret the trend of increasing warm ( inner disk ) large grain mass fraction with decreasing stellar mass similarly , in view of our finding of greater grain growth in inner disk regions than outer disk regions ( figure 15 ) .
we have analyzed the dust composition of 65 t tauri stars using spectra from the irs on the _ spitzer space telescope_. these spectra show very prominent emission features and complexes at 10 , 16 , 19 , 23 , 28 , and 33 @xmath0 wavelengths , which are characteristic of silica and silicate grains ( both amorphous and crystalline ) with sizes from submicron to a few microns .
we have constructed spectral models for each of the spectra that include , firstly , blackbodies at two temperatures and , secondly , planck functions at those two temperatures multiplied by scaled dust emissivities .
these represent inner and outer disk emission from , firstly , the optically thick disk midplane and blackbody grains and , secondly , dust in the optically thin disk atmosphere with strong infrared resonances . the best fit for a given set of dust opacities
is given by the pair of temperatures for which the global @xmath37 per degree of freedom is a minimum .
we have tested the opacities on high - quality irs spectra to find the best opacities to represent emission from submicron grains of enstatite , forsterite , silica , and amorphous silicates and large ( few micron size ) grains in the protoplanetary disks of the taurus - auriga association .
we conclude the following : * high s / n spectra suggest the best crystalline silicate opacities to use in modeling are those of iron - poor pyroxene ( enstatite ) and iron - poor olivine ( forsterite ) .
the best amorphous silicate opacities to use are those whose iron - to - magnesium ratios are nearly 1 ( cosmic ) .
this is consistent with the finding by @xcite that cometary amorphous silicates require significant iron in their compositions in order to heat to inferred dust temperatures .
the adopted amorphous silicate closely matches the heating of @xcite `` astronomical silicates '' . *
though we use opacities of grains with 50% porosity and 5 @xmath0 radius of both amorphous pyroxene and amorphous olivine , calculations of the covariance between these components in models of various high s / n spectra show they are highly degenerate .
this means our models can not readily distinguish between the two species as large grains , but our models are more sensitive to the sum of their masses .
submicron amorphous pyroxene and submicron amorphous olivine are similarly degenerate but slightly less so .
cool silica is degenerate with cool submicron amorphous silicates as they share 20 @xmath0 features .
cool enstatite and cool forsterite are somewhat degenerate because they share prominent 23 , 28 , and 33 @xmath0 features ( only somewhat because the 28 @xmath0 feature of enstatite is relatively stronger than that of forsterite ) .
* a few spectra of t tauri stars show evidence for extensive grain growth in that protoplanetary disks as indicated especially by their 10 @xmath0 complexes . * though most spectra are fit satisfactorily using our truncated cde forsterite opacity , a few are not .
the deviations of the 10 , 23 , and 33 @xmath0 features between our models and the spectra suggest olivine grains of greater size , greater porosity , greater iron content ( namely , a nonzero iron content ) , or greater weighting towards extremely elongated or flattened grain shapes in the shape distribution are required , or some combination of these . *
most spectra that require silica are fit well by the annealed silica opacity in the standard dust model mixture .
the spectra of fz tau and dn tau suggest a silica polymorph other than annealed silica ( cristobalite and tridymite ; see * ? ? ?
* ) would provide a better fit to their spectra .
this would imply different cooling rates for silica , once formed , than implied by the presence of the higher temperature polymorphs of silica @xcite .
however , the 12.6 @xmath0 features of these two exceptional spectra are only mildly inconsistent with the annealed silica in the dust models , so the suggestion of an alternative silica polymorph is weak . *
the models of the best enstatite exemplars , fn tau and dh tau , do not fit very well the strong crystalline silicate emission peaks at 28 and especially 33 @xmath0 , suggesting enstatite or forsterite dust at temperatures significantly lower than those used by our models is required by the data .
it is unknown why such cold material is required to fit spectra indicating high enstatite abundance . *
transitional and pre - transitional disks in this sample require negligible crystalline silicates and modest amounts of large grains ( dm tau being the exception , requiring significant large grains ) to fit their spectra relative to the rest of the disks whose spectra were analyzed in this study .
a similar ( small ) number of spectra of systems not known to be transitional or pre - transitional also require high abundances of submicron amorphous silicates . *
a few spectra have 10 @xmath0 complexes that have small equivalent widths .
most of these are fit well by crystalline silicates as opposed to large grains , which is inconsistent with the idea that decreasing peak - to - continuum ratio of the 10 @xmath0 complex necessarily implies grain growth .
* the goodness of fit of our models to the majority of our spectra means the opacities of large crystalline silicate grains are not typically required .
this , in turn , implies that , if grains grow to larger sizes by agglomeration of grains of differing composition , then on average in the taurus - auriga yso population the more abundant subcomponents of these heterogeneous aggregate large grains are amorphous rather than crystalline . *
higher abundances of large grains are found in inner disk regions than outer disk regions , suggesting grain growth occurs more rapidly in inner disk regions . *
crystalline silicate abundances are very similar in the inner and outer disk regions .
this contrasts interestingly with the finding by @xcite that regions inside 2 au in disks around herbig ae / be stars have much more processed dust than regions outside of 2 au .
f04147 + 2822 exemplifies the contrast , having prominent 19 , 23 , 28 , and 33 @xmath0 complexes but very little indication of crystallinity based on its 10 @xmath0 complex .
this suggests the inner region of this disk giving rise to the 10 @xmath0 complex has a lower abundance of crystals than the outer region giving rise to emission past 19 @xmath0 wavelength .
* cool forsterite also correlates with flatter submillimeter seds , suggesting grain growth to millimeter sizes is more extensive for disks with more cool forsterite .
* each crystalline silicate abundance always correlates positively and significantly with abundances of other crystalline silicate species .
this suggests that whatever produces crystalline silicates does so at a faster rate than that at which any of the crystalline silicate species transform into any of the others .
this , in turn , suggests that amorphous dust is processed into crystalline dust at a rate greater than one kind of crystalline silicate species can transform into another .
* crystalline silicates in inner disk regions are more abundant for bluer , flatter disks , those with more advanced settling of dust grains towards disk midplane . * large grains in the inner disk regions
are , on average , more abundant for more settled ( bluer ) disks .
there is a fair degree of dispersion in the inner disk large grain abundance for a given degree of settling , suggesting different rates of settling of large grains in different disks .
there is also an indication that the abundance of large grains in the inner disk correlates with known multiplicity of stellar system . *
the spectra of disks around less massive stars indicate higher abundances of large grains in their inner disk regions .
this work is based on observations made with the _
spitzer space telescope _ , which is operated by the jet propulsion laboratory , california institute of technology under nasa contract 1407 .
support for this work was provided by nasa through contract number 1257184 issued by jpl / caltech and through the spitzer fellowship program , under award 011 808 - 001 , and jpl contract 960803 to cornell university , and cornell subcontracts 31419 - 5714 to the university of rochester .
the authors wish to thank harald mutschke for sharing the opacity in tabular form of the annealed silica presented by @xcite .
a.l . acknowledges support from the chandra theory program , the hubble theory programs , and the spitzer theory programs .
smart was developed by the irs team at cornell university and is available through the spitzer science center at caltech .
this publication makes use of the jena - st .
petersburg database of optical constants @xcite .
the authors made use of the simbad astronomical database and would like to thank those responsible for its upkeep .
we corrected for slight mispointing of the telescope from the standard nod positions by scaling the mispointed nods , as described by @xcite . for observations taken by the low spectral resolution modules sl and ll ,
only mispointing in the dispersion direction ( perpendicular to the length of the slit ) could give rise to loss of flux .
mispointing of the high resolution slits along the length of the slit ( spatial direction ) towards the center of the slit can actually result in obtaining more flux than is present if one does not consider the rsrfs used to calibrate flux were constructed from observations of flux standards at nominal nod positions .
more flux enters through the slit if the target is positioned closer to the center of the slit than a nominal nod position ( like the flux calibrator is ) ; in this case , one must apply a scalar factor less than 1 to the affected nod spectrum to correct for the extra fraction of total signal from the target not measured for the calibrator .
sh has a smaller slit and is therefore more sensitive to mispointing in either spectral or spatial directions ; therefore , the relative amount of scalar correction is typically higher for sh than lh .
the scalars used are given in table a1 .
we conducted tests of the combination of bruggeman emt and mie theory used to compute effective complex dielectric constants and opacities for our large grains . note that @xcite found that opacity curves of sub - micron spherical 60% vacuum ( porous ) grains of forsterite and @xmath6-quartz were identical to those computed for solid grains of the same materials in the cde shape distribution .
we have computed the effective complex dielectric function for material that is , by volume , 60% vacuum ( complex dielectric function equal to unity ) , 20% forsterite ( complex dielectric function by * ? ? ?
* ) , and 20% amorphous olivine mgfesio@xmath3 ( complex dielectric function by * ? ? ?
* ) as follows .
first we computed using bruggeman emt the effective complex dielectric function of 50% amorphous olivine and 50% forsterite a - axis ; we also computed the same effective complex dielectric functions for 50/50 mixtures of amorphous olivine and each of the other two forsterite crystallographic axes , b and c. then we computed using bruggeman emt the effective complex dielectric constants for material that is 60% by volume vacuum and 40% the 50/50 mixture of amorphous olivine and forsterite ; again , the same was done for the effective complex dielectric functions involving forsterite b and c axes .
this resulted in 3 sets of complex dielectric functions , one for each of the 3 crystallographic axes of forsterite , of material that is , by volume , 60% vacuum , 20% forsterite , and 20% amorphous olivine .
next we computed opacities for each of the three sets of effective complex dielectric functions for rayleigh - limit - size spheres of these materials .
finally , we averaged the three opacities together to account for anisotropy of optical constants ( see * ? ? ?
* ) for forsterite .
we then calculated the flux density from 1 lunar mass of such grains 140 parsecs away at 300k .
1% relative uncertainties were assumed , which are about the best we can expect from our sample of 65 spectra . the 7.7 - 37 @xmath0 part of this spectrum was then modeled using our standard opacities , which are listed in the notes for table 6 , with the exception of using the opacities of forsterite and amorphous olivine grains in the cde shape distribution ( instead of tcde and cde2 ) .
we used the cde for these two dust types in anticipation of the porous - cde correspondence reported by @xcite . in figure a1
, we show the fit of the model to this test spectrum .
although none of the dust opacities were precluded at the outset , the only opacities used were the cde opacities for forsterite and amorphous olivine , with masses of 0.45 and 0.43 m@xmath85 , respectively .
the fit is very good to all wavelengths , though we note deviations for the strongest features at 11.4 , 19 , and 23 @xmath0 ; we also note that the correspondence between porous grain and cde opacities reported by @xcite was not perfect .
the model used grains at 308k , only slightly different from the assumed 300k .
this slightly higher than assumed temperature explains the slightly lower ( 0.88 m@xmath85 ) than assumed ( 1 m@xmath85 ) total mass of dust . we take this as support for the conclusion by @xcite that opacities of aggregates can be expressed as a sum of the opacities of homogeneous grains composed of the same materials as the aggregate s components , as applied to heterogeneous aggregates in the rayleigh size limit .
next we conducted a similar test , but for large aggregates .
we used the same effective complex dielectric function of vacuum , forsterite , and amorphous olivine in a 60/20/20 volume ratio as before but this time to compute opacities for 5 @xmath0 radius grains of the material .
we computed the test spectrum assuming 1 lunar mass of grains at 300k located 140pc away .
here , we used our standard opacities exactly as listed in the notes for table 6 , to gauge the effect of large grains in protoplanetary disks being aggregates of both amorphous and crystalline grains .
dust at only one temperature was allowed , but no dust species at this temperature was precluded from the model at the outset .
this time , the best - fit model dust temperature is lower , 264k , and enstatite is used in addition to forsterite and large ( 5 @xmath0 radius porous ) grains of amorphous olivine , though at a lower significance level than either the amorphous olivine or forsterite .
we show this model in figure a2 .
the total mass of the dust used by the model , 1.25 lunar masses , is a sum of 0.96 m@xmath85 of large amorphous olivine plus 0.20 m@xmath85 of forsterite , 0.09 m@xmath85 of enstatite , and 0.001 m@xmath85 of submicron amorphous pyroxene , which is somewhat greater than the mass assumed for the aggregates .
this increase in mass was needed to compensate for the lower temperature .
the model becomes an increasingly poorer fit at increasing wavelengths .
part of this failure may be due to the use of the restricted shape distribution ( tcde ) for the forsterite instead of the more extreme shape distribution cde ( used previously to test the rayleigh - limit aggregate ) .
an additional population of forsterite grains at a lower temperature than used may improve the fit .
this suggests any serious failures of our standard model to fit spectra with prominent crystalline silicate features may imply a significant amount of grain growth of aggregate grains with significant crystalline silicate abundances ( here , 50% of the solid mass of the aggregate ) .
this test also implies that the crystalline abundance could be underestimated by 50% ( 0.29/1.25 instead of 0.625/1.25 fractional abundance of crystalline silicates in heterogeneous aggregates ) of the true value , and it implies that the large grain abundance could be slightly underestimated ( 0.96 instead of 1.25 lunar masses of large ( 5 @xmath0 radius ) grains ) by 100@xmath69(0.29/1.25)% = 23% .
we tested the necessity of using opacities of large grains in our models .
we used our standard dust model on is tau , one of the objects whose spectrum was modeled using a significant amount of large grains , but we eliminated the opacities of large amorphous pyroxene and large amorphous olivine in the fit .
the resulting model , shown in figure a3 , is a poorer fit to the 10 @xmath0 complex , not `` filling out '' the full width of the 10 @xmath0 complex .
reduced @xmath37 for this large - grains - precluded model was 4.9 , which is 1.6 higher than the reduced @xmath37 of 3.3 for the model with large grains .
we conclude the opacities of large amorphous grains are required in our models .
we tested how sensitive our models are to the exact combination of porosity and grain size when we compute and use opacities of large grains in our models to detect grain growth . instead of 5 @xmath0 radius 60% porous grains , we computed opacities for 20 @xmath0 radius grains that are 88% porous .
we chose this size to preserve the optical depth through the center of the grain .
as noted by @xcite , 5 @xmath0 radius 60% porous grains were chosen to preserve the optical depth through the center of the 2 @xmath0 solid grains of amorphous olivine used by @xcite to represent grain growth in modeling spectra of herbig ae / be stars .
this preservation of the optical depth through the center of the grain resulted in a close correspondence of the opacities of 5 @xmath0 radius porous grains and of 2 @xmath0 radius solid grains .
although 90% vacuum ( by volume ) grains would be the logical extension 88% vacuum 20 @xmath0 radius grains proved to result in an opacity most like that of 5 @xmath0 radius 60% vacuum grains . as a test of this opacity , we modeled the spectrum of uz tau / e , noted in section 4 to require a large amount of large grains ; here , we replace the opacity of 5 @xmath0 60% porous grains of amorphous olivine with that of 20 @xmath0 88% porous grains of amorphous olivine .
we show the best - fit model in figure a4 , compared to the model shown in figure 4 and described in section 4 .
the fit is not much different , and the reduced @xmath37 is 5.6 instead of the 5.4 obtained with our standard opacities .
therefore , we can not make precise statements on grain size and porosity individually for a given protoplanetary disk based only on the irs spectra .
andrews , s. m. , & williams , j. p. 2005
, , 631 , 1134 andrews , s. m. , & williams , j. p. 2007
, , 659 , 705 apai , d. , pascucci , i. , bouwman , j. , natta , a. , henning , t. , & dullemond , c. p. 2005 , science , 310 , 834 armus , l. , et al . 2007 , , 656 , 148 avni , y. 1976 , , 210 , 642 batalha , c. c. , et al . 1998 , , 128 , 561 bevington , p. r. 1969 , new york : mcgraw - hill , 1969 , bohren , c. f. , & huffman , d. r. 1983 , absorption and scattering of light by small particles , new york : wiley boogert , a. c. a. , et al . 2008 , , 678 , 985 bouwman , j. , meeus , g. , de koter , a. , hony , s. , dominik , c. , & waters , l. b. f. m. 2001 , , 375 , 950 bouwman , j. , de koter , a. , dominik , c. , & waters , l. b. f. m. 2003 , , 401 , 577 bouwman , j. , et al .
2008 , arxiv e - prints , 802 , arxiv:0802.3033 bowey , j. e. , & adamson , a. j. 2002 , , 334 , 94 bradley , j. 2003 , lnp vol . 609 : astromineralogy , 609 , 217 calvet , n. , et al . 2005 , , 630 , l185 cambresy , l. , copet , e. , epchtein , n. , de batz , b. , borsenberger , j. , fouque , p. , kimeswenger , s. , & tiphene , d. 1998 , , 338 , 977 carr , j. s. , & najita , j. r. 2008 , science , 319 , 1504 chen , c. h. , et al .
2006 , , 166 , 351 chiang , e. i. , & goldreich , p. 1997
, , 490 , 368 chiar , j. e. , et al .
2007 , , 666 , l73 chihara , h. , koike , c. , & tsuchiyama , a. 2001 , , 53 , 243 chihara , h. , koike , c. , tsuchiyama , a. , tachibana , s. , & sakamoto , d. 2002 , , 391 , 267 chihara , h. , koike , c. , & tsuchiyama , a. 2007 , , 464 , 229 ciesla , f. j. 2007 , , 654 , l159 cohen , m. , & witteborn , f. c. 1985 , , 294 , 345 cohen , m. , megeath , s. t. , hammersley , p. l. , martn - luis , f. , & stauffer , j. 2003 , , 125 , 2645 cox , a. n. 2000 , allen s astrophysical quantities , crovisier , j. , leech , k. , bockelee - morvan , d. , brooke , t. y. , hanner , m. s. , altieri , b. , keller , h. u. , & lellouch , e. 1997 , science , 275 , 1904 dalessio , p. , calvet , n. , & hartmann , l. 2001 , , 553 , 321 dalessio , p. , et al .
2005 , , 621 , 461 dalessio , p. , et al .
2006 , , 638 , 314 day , k. l. 1979 , , 234 , 158 dorschner , j. , friedemann , c. , guertler , j. , & henning , t. 1988 , , 198 , 223 dorschner , j. , begemann , b. , henning , t. , jaeger , c. , & mutschke , h. 1995 , , 300 , 503 draine , b. t. , & lee , h. m. 1984 , , 285 , 89 draine , b. t. 2003 , , 41 , 241 dullemond , c. p. , & dominik , c. 2004 , , 421 , 1075 dullemond , c. p. , & dominik , c. 2005 , , 434 , 971 espaillat , c. , et al . 2007 , , 664 , l111 espaillat , c. , calvet , n. , dalessio , p. , hernndez , j. , qi , c. , hartmann , l. , furlan , e. , & watson , d. m. 2007 , , 670 , l135 espaillat , c. , et al .
2008 , submitted fabian , d. , jger , c. , henning , t. , dorschner , j. , & mutschke , h. 2000 , , 364 , 282 fabian , d. , henning , t. , j " ager , c. , mutschke , h. , dorschner , j. , & wehrhan , o. 2001 , , 378 , 228 forrest , w. j. , gillett , f. c. , & stein , w. a. 1975 , , 195 , 423 forrest , w. j. , houck , j. r. , & reed , r. a. 1976 , , 208 , l133 furlan , e. , et al .
2006 , , 165 , 568 furlan , e. , et al .
2007 , , 664 , 1176 glassgold , a. e. , najita , j. , & igea , j.2004 , , 615 , 972 gorti , u. , & hollenbach , d. 2008 , arxiv e - prints , 804 , arxiv:0804.3381 harker , d. e. , & desch , s. j. 2002 , , 565 , l109 harker , d. e. , wooden , d. h. , woodward , c. e. , & lisse , c. m. 2002 , , 580 , 579 henning , t. , ilin , v. b. , krivova , n. a. , michel , b. , & voshchinnikov , n. v. 1999 , , 136 , 405 higdon , s. j. u. , et al . 2004 , , 116 , 975 holweger , h. , heise , c. , & kock , m. 1990 , , 232 , 510 honda , m. , kataza , h. , okamoto , y. k. , miyata , t. , yamashita , t. , sako , s. , takubo , s. , & onaka , t. 2003 , , 585 , l59 honda , m. , et al .
2006 , , 646 , 1024 houck , j. r. , et al . 2004 , , 154 , 18 hubbard , a. , & blackman , e. g. 2006 , new astronomy , 12 , 246 jaeger , c. , mutschke , h. , begemann , b. , dorschner , j. , & henning , t. 1994 , , 292 , 641 jaeger , c. , molster , f. j. , dorschner , j. , henning , t. , mutschke , h. , & waters , l. b. f. m. 1998 , , 339 , 904 kamp , i. , & dullemond , c. p. 2004
, , 615 , 991 kastner , j. h. , buchanan , c. l. , sargent , b. , & forrest , w. j. 2006 , , 638 , l29 kemper , f. , vriend , w. j. , & tielens , a. g. g. m. 2004 , , 609 , 826 ( erratum : 633 , 534 ) kenyon , s. j. , dobrzycka , d. , & hartmann , l. 1994 , , 108 , 1872 kenyon , s. j. , & hartmann , l. 1995 , , 101 , 117 kessler , m. f. , et al .
1996 , , 315 , l27 kessler - silacci , j. e. , hillenbrand , l. a. , blake , g. a. , & meyer , m. r. 2005 , , 622 , 404 kessler - silacci , j. , et al .
2006 , , 639 , 275 koike , c. , et al .
2000 , , 363 , 1115 koike , c. , chihara , h. , tsuchiyama , a. , suto , h. , sogawa , h. , & okuda , h. 2003 , , 399 , 1101 li , a. , & greenberg , j. m. 1998 , , 498 , l83 li , a. , & lunine , j. i. 2003 , , 590 , 368 li , a. , & lunine , j. i. 2003 , , 594 , 987 li , a. , lunine , j. i. , & bendo , g. j. 2003 , , 598 , l51 luhman , k. l. 2004 , , 602 , 816 marshall , j. a. , herter , t. l. , armus , l. , charmandaris , v. , spoon , h. w. w. , bernard - salas , j. , & houck , j. r. 2007 , , 670 , 129 mcclure , m. k. , et al .
2008 , accepted min , m. , hovenier , j. w. , waters , l. b. f. m. , & de koter , a. 2008 , arxiv e - prints , 806 , arxiv:0806.4038 miyake , k. , & nakagawa , y. 1993 , icarus , 106 , 20 molster , f. j. , waters , l. b. f. m. , & tielens , a. g. g. m.2002 , , 382 , 222 pilipp , w. , hartquist , t. w. , morfill , g. e. , & levy , e. h. 1998 , , 331 , 121 posch , t. , mutschke , h. , trieloff , m. , & henning , t. 2007 , , 656 , 615 press , w. h. , teukolsky , s. a. , vetterling , w. t. , & flannery , b. p. 1992 , cambridge : university press , |c1992 , 2nd ed . , przygodda , f. , van boekel , r. , brahm , p. , melznikov , s. y. , waters , l. b. f. m. , & leinert , c. 2003 , , 412 , l43 rettig , t. , brittain , s. , simon , t. , gibb , e. , balsara , d. s. , tilley , d. a. , & kulesa , c. 2006 , , 646 , 342 sargent , b. , et al .
2006 , , 645 , 395 sargent , b. , et al . 2008 , accepted for publication by the astrophysical journal schegerer , a. , wolf , s. , voshchinnikov , n. v. , przygodda , f. , & kessler - silacci , j. e. 2006 , , 456 , 535 schtz , o. , meeus , g. , & sterzik , m. f. 2005 , , 431 , 165 scott , a. , & duley , w. w. 1996 , , 105 , 401 sicilia - aguilar , a. , hartmann , l. w. , watson , d. , bohac , c. , henning , t. , & bouwman , j. 2007 , , 659 , 1637 sloan , g. c. , devost , d. , bernard - salas , j. , wood , p. r. , & houck , j. r. 2006 , , 638 , 472 sloan , g. c. , kraemer , k. e. , wood , p. r. , zijlstra , a. a. , bernard - salas , j. , devost , d. , & houck , j. r. 2008 , arxiv:0807.2998 snow , t. p. , & witt , a. n. 1995 , science , 270 , 1455 sogawa , h. , koike , c. , chihara , h. , suto , h. , tachibana , s. , tsuchiyama , a. , & kozasa , t. 2006 , , 451 , 357 taylor , j. r. 1982 , a series of books in physics , oxford : university press , and mill valley : university science books , 1982 , uchida , k. i. , et al .
2004 , , 154 , 439 van boekel , r. , waters , l. b. f. m. , dominik , c. , bouwman , j. , de koter , a. , dullemond , c. p. , & paresce , f. 2003 , , 400 , l21 van boekel , r. , et al .
2004 , , 432 , 479 van boekel , r. , min , m. , waters , l. b. f. m. , de koter , a. , dominik , c. , van den ancker , m. e. , & bouwman , j. 2005 , , 437 , 189 watson , d. m. , et al .
2004 , , 154 , 391 watson , d. m. , et al .
2007 , , 448 , 1026 watson , d. m. , et al .
2008 , in preparation weidenschilling , s. j. 1997 , icarus , 127 , 290 werner , m. w. , et al .
2004 , , 154 , 1 wooden , d. h. , harker , d. e. , woodward , c. e. , butner , h. m. , koike , c. , witteborn , f. c. , & mcmurtry , c. w. 1999 , , 517 , 1034 wooden , d. h. , harker , d. e. , & brearley , a. j. 2005 , chondrites and the protoplanetary disk , 341 , 774 zasowski , g. , markwick - kemper , f. , watson , d. m. , furlan , e. , bohac , c. j. , hull , c. , & green , j. d. 2007 , arxiv:0712.2458 zubko , v. g. , mennella , v. , colangeli , l. , & bussoletti , e. 1996 , , 282 , 1321 | mid - infrared spectra of 65 t tauri stars ( tts ) taken with the infrared spectrograph ( irs ) on board the _ spitzer space telescope _ are modeled using populations of optically thin dust at two temperatures to probe the radial variation in dust composition in the uppermost layers of protoplanetary disks .
most spectra with narrow emission features associated with crystalline silicates require mg - rich minerals and silica , but a very small number suggest other components .
spectra indicating large amounts of enstatite at higher temperatures ( 400 - 500k ) also require crystalline silicates , either enstatite or forsterite , at temperatures lower ( 100 - 200k ) than those required for spectra showing high abundance of other crystalline silicates .
a few spectra show 10 @xmath0 complexes of very small equivalent width .
they are fit well using abundant crystalline silicates but very few large grains , inconsistent with the expectation that low peak - to - continuum ratio of the 10 @xmath0 complex always indicates grain growth .
most spectra in our sample are fit well without using the opacities of large crystalline silicate grains .
if large grains grow by agglomeration of submicron grains of all dust types , the amorphous silicate components of these aggregates must typically be more abundant than the crystalline silicate components .
we also find that the more there is of one crystalline dust species , the more there is of the others .
this suggests that crystalline silicates are processed directly from amorphous silicates , whether by evaporation of the amorphous grains and condensation in chemical equilibrium or by annealing of the amorphous precursor grains .
this also suggests that neither forsterite , enstatite , nor silica are intermediate steps along the way to producing either of the other two for the majority of the crystalline dust produced .
crystalline silicate abundance is correlated tightly with disk geometry , in the sense of higher crystallinity accompanying more settled disks , which are commonly associated with growth and settling of grains .
large - grain abundance is also correlated with disks that are more highly settled , but with a wide range of large grain abundance for a given degree of settling .
we interpret this range to mean that the settling of large grains is sensitive to individual disk properties .
we also find that lower - mass stars have higher abundances of large grains in their inner regions . 0 _ |
topological excitations are the hallmark of symmetry broken systems , such as liquid crystals @xcite , superfluid helium systems @xcite , and ultracold atomic gases @xcite .
they appear as spatial variations of order parameters with a topological charge .
examples include a vortex ( line defect ) , a point defect , and a skyrmion .
the topological excitations arise as a result of the spontaneous symmetry breaking of the system and their existence is determined by the broken symmetry . to characterize the broken symmetry of a given order parameter @xmath0
, we introduce a symmetry group @xmath1 whose action @xmath2 @xmath3 does not change the energy of system .
also , its subgroup @xmath4 keeps the order parameter invariant , where @xmath5 is called an isotropy group @xmath6 . in terms of @xmath1 and @xmath5 ,
the broken symmetry is described by a quotient space @xmath7 , which represents the degeneracy of the order parameter and defines an order parameter manifold ( opm ) .
topological excitations are classified by the homotopy group @xcite .
the invariance of a topological charge implies that a topological excitation can not continuously transform to an other topological excitation if their topological charges are different . for the case of a vortex
, the topological charge is characterized by the fundamental group @xmath8 . for general cases , in @xmath9-dimensional ( @xmath9d ) space , a @xmath10-dimensional topological excitation
is classified by @xmath11 .
in particular , for @xmath12 , topological excitations with the dimension @xmath13 correspond to point defects , line defects , and domain walls , respectively .
for example , in the case of a scalar bose - einstein condensate ( bec ) or a superfluid @xmath14he , the opm is @xmath15 , which represents the global gauge degrees of freedom .
the fundamental group @xmath16 gives a topological charge of a point defect in 2d space ( @xmath17 ) or that of a line defect in 3d space ( @xmath18 ) , where @xmath19 is an additive group of integers .
topological excitations are usually classified by the homotopy group @xmath20 , but we can not , in general , directly apply this scheme if more than one type of topological excitations coexist , for there may be a topological influence between them .
the topological influence was first pointed out by mineev and volovik @xcite , and mermin @xcite in condensed matter physics , and schwarz @xcite in cosmology .
for example , a point - defect charge changes its sign as it makes a complete circuit of a half - quantum vortex .
the topological influence depends only on the topology of the order parameter manifold and is mathematically described by the @xmath21 action on @xmath20 .
that is , the homotopy groups no longer uniquely determine the charge of the topological excitation under the topological influence because the charge is not invariant under the relative motion of one topological object against another . in a previous work @xcite
, we have demonstrated that the abe homotopy group @xmath22 can be used to deal with the topological influence .
the abe homotopy group classifies maps from a pinched torus ( see fig . 6 in ref .
@xcite ) to the order parameter manifold , and this map enables us to classify a combined object of line defects and point defects . in this framework , the @xmath21 action on @xmath23 can naturally be described by the group action of @xmath22 .
so far , these researches focus on the topological influence in multiple topological excitation systems , but do not address the question of the charge conservation . in the present paper
, we address the consistency between the topological influence and the charge conservation .
the charge conservation is important for an isolated system like ultracold atomic gases because a topological charge can not escape from the system .
the consistency between topological influence and the charge conservation was first discussed by bucher _
et al_. @xcite in connection with alice cosmic rings .
they discussed a topological charge transfer from the magnetic monopole to the alice cosmic ring .
the purpose of this paper is to revisit the relation between the charge conservation and the charge transfer in isolated systems and to show that the back - action from a point defect can be interpreted as a twist of a vortex " .
we formulate the charge transfer using the commutator of the underlying homotopy groups @xmath24 $ ] , which gives a general classification of the charge transfer . here
, an element of @xmath24 $ ] lies within @xmath23 even for @xmath25 because the @xmath8 action on @xmath23 is thought of as a homomorphism from @xmath26 to @xmath27 which is the automorphism map on @xmath28 @xcite .
we also make concrete classifications of the charge transfer for liquid crystals , ultracold atomic gases , and superfluid @xmath29he and the non - abelian charge transfer in the spin-2 bec . this paper is organized as follows . in sec .
[ sec : influence ] , we review a topological influence using liquid crystals , which exhibit the topological influence on non - abelian vortices for a biaxial nematic phase and on a point defect in a uniaxial nematic phase . in general , the topological influence is described by the automorphism map on @xmath30 induced by @xmath31 . in sec .
[ sec : twist ] , we describe the connection between the back - action and the twist of vortex . in secs .
[ sec : transfer - vortex]-[sec : transfer - monopole ] , we discuss a charge transfer between non - abelian vortices and that between a vortex and a point defect . in sec .
[ sec : examples ] , we gives examples in liquid crystals , spinor becs , and superfluid @xmath29he . in sec .
[ sec : testing ] , we suggest the experimental detection of the total charge and each individual topological charge under the charge conservation . finally , in sec .
[ sec : summary ] , we summarize the main results of this paper . in appendix , we give the mathematical definitions of the semidirect product and the @xmath32-product to make this paper self - contained .
in this section , we review the topological influence , which is defined as a nontrivial effect on a topological charge of a topological excitation due to the presence of another topological excitation .
it is well - known that the topological influence exists between non - abelian vortices @xcite and between a point defect and a vortex @xcite .
this effect is mathematically described as the @xmath8 action on @xmath28 , which makes up an automorphism map on @xmath23 .
especially , for @xmath33 , the topological influence is described by a conjugation in group theory .
recently , we have proposed that the abe homotopy group @xmath34 @xcite provides a mathematical framework for simultaneously classifying a vortex and a topological excitation with dimensionality @xmath35 .
this group naturally includes the topological influence in its group structure . in condensed matter systems ,
the topological influence is predicted for a number of ordered phases and for several topological excitations such as a vortex with a non - abelian charge @xcite , a point defect @xcite , and a @xmath36-texture @xcite in 4d space - time .
first , we discuss the topological influence on a non - abelian vortex . as a concrete example , let us consider a biaxial nematic phase in a liquid crystal .
an order parameter of the liquid crystal is described by a real traceless symmetric tensor @xmath37 .
the symmetry of this system is given by @xmath38 .
we define the group action on @xmath39 as @xmath40 where @xmath41 and @xmath42 means the transpose . in the biaxial nematic phase ,
the matrix order parameter is described by @xmath43 where @xmath44 and they satisfy @xmath45 . in addition , we assume that @xmath46 , @xmath47 and @xmath48 . hence , the isotropy group is the second dihedral group @xmath49 and the opm turns out to be @xcite @xmath50 where @xmath51 is the eight - element quaternion group , which is represented in terms of the pauli matrices @xmath52 as @xmath53 . here
, @xmath54 is the 2-by-2 identity matrix .
since @xmath55 is simply connected , from the exact homotopy sequence , we obtain @xmath56 therefore , vortices of the biaxial nematic phase are labeled by @xmath51 .
furthermore , @xmath51 is a non - abelian group , so that vortices have non - abelian charges .
let us consider a situation in which two vortices labeled by @xmath57 and @xmath58 coexist and a vortex with charge @xmath57 rotates around a vortex with charge @xmath58 . to see the topological nature of this situation
, we carry out this operation adiabatically so as to guarantee that the order parameters remain in the ground - state manifold during the rotation . from fig .
[ fig : loop - deformation ] , we find that the situation remains the same if we replace @xmath59 and @xmath60 with @xmath58 and @xmath57 , respectively , and identify the topological charge of a vortex v@xmath61 after rotating around a different vortex v@xmath62 with the topological charge of a vortex enclosed by @xmath63 . as a result ,
the topological charge @xmath57 changes as @xmath64 according to the definition of the eight - element quaternion group , @xmath57 and @xmath65 are different topological charges .
thus , a topological charge of a vortex can transform to an other topological charge by making a complete circuit around a different vortex .
in other words , the non - abelian charge is not uniquely determined up to conjugation .
this ambiguity is a well - known property in the classification of vortices @xcite .
in fact , it is known that non - abelian vortices are classified by conjugacy classes of the fundamental group .
the conjugacy class of @xmath51 is given by @xmath66 = \ { \ { \bm{1}_2\ } , \{-\bm{1}_2\ } , \
{ \pm i \sigma_x\ } , \
{ \pm i \sigma_y\ } , \ { \pm
i \sigma_z\ } \ } , \end{aligned}\ ] ] where @xmath67 $ ] means conjugacy classes .
therefore , we have five different types of vortices in the biaxial nematic phase .
however , the classification based on the conjugacy classes presupposes that the topological charge need not be conserved @xcite . a classification that respects the topological charge conservation is the primary subject of this paper , which will be discussed in sec .
[ sec : back - action ] in detail . represents a loop enclosing a vortex v@xmath62 , and @xmath68 encloses a vortex v@xmath61 from the left(right)-hand side of v@xmath62 , but does not enclose v@xmath62 .
( b ) loop deformation from @xmath69 to @xmath63 , where @xmath70 is a loop product defined by eq .
( [ eq : loop - prod ] ) and @xmath71 represents a loop in the reverse direction .
this deformation shows that the loops @xmath69 and @xmath63 are homotopy equivalent.,width=302 ] in general , the topological influence on a non - abelian vortex is described by the inner automorphism map @xmath72 where @xmath73 and @xmath74 , and @xmath70 represents the loop product defined for a parameter @xmath75 $ ] by @xmath76 where @xmath77 satisfies @xmath78 .
the loop product satisfies the group structures through the homotopy equivalence @xcite .
thus , @xmath79 is the identity map if @xmath59 is an element of the centralizer @xmath80 .
next , let us consider the topological influence on a topological excitation with dimension higher than one ( @xmath25 ) .
as an example , we treat the topological influence on a point defect in the uniaxial nematic phase in a liquid crystal .
an order parameter of the uniaxial nematic phase is given by @xmath81 where @xmath82 .
the corresponding isotropy group is @xmath83 , where @xmath84 " denotes the semidirect product defined in appendix [ app : a ] .
thus , the opm becomes @xmath85 where @xmath86 is a 2d real projective space , whose first and second homotopy groups are given by @xmath87 equations ( [ eq : un_pi1 ] ) and ( [ eq : un_pi2 ] ) predict the existence of a vortex and a point defect in the uniaxial nematic phase .
let us denote their elements as @xmath88 and @xmath89 , and consider a situation in which a vortex with charge @xmath90 and a point defect with charge @xmath91 coexist as illustrated in fig .
[ fig : surface - deformation ] ( a ) .
when the point defect makes a complete circuit around the vortex , the charge of the point defect changes to @xmath92 . through continuous transformation as illustrated in fig .
[ fig : surface - deformation ] ( b ) , we obtain @xmath92 as @xmath93 where the loop product @xmath70 is defined as @xmath94 equation ( [ eq : inf_alpha ] ) represents the topological influence of the vortex on the point defect . by definition , @xmath95 and @xmath96
is bijective , so @xmath96 acts on the @xmath97-element as an automorphism map .
thus , similar to the non - abelian vortex , the topological influence of the vortex is represented by the automorphism map .
on the other hand , since @xmath98 , the set of automorphism maps @xmath99 is isomorphic to @xmath100 . as a result
, the possible topological influence is given by a @xmath100-action on the point - defect charge .
in the uniaxial nematic phase , a disclination gives the nontrivial @xmath100-action . to show this
, we stat with the standard order parameter of the uniaxial nematic phase @xmath101 and describe the order parameters along a loop around the disclination by @xmath102 , where @xmath103 represents a @xmath104-rotation about the @xmath105 axis in the cartesian coordinates and @xmath75 $ ] is a parameter which specifies the position of the loop .
then , the vector @xmath106 , which describes the direction of the major axis of the nematic tensor at @xmath107 , changes to @xmath108 by the action of @xmath109 at @xmath110 . on the other hand , the topological charge of the point - defect is defined by @xmath111 where @xmath112 are the polar and azimuthal coordinates of a unit sphere enclosing the point defect . the charge @xmath113 transforms into its inverse @xmath114 if the vector @xmath115 changes to @xmath108 .
thus , when the point defect goes around the disclination , the sign of @xmath113 changes due to the influence of the disclination . in general , the topological charge involving the topological influence can be described by the abe homotopy group @xmath116 @xcite . by analogy with the classification of non - abelian vortices ,
the conjugacy class of the abe homotopy group gives the corresponding topological charge if the charge conservation is not required .
the conjugacy class of the second abe homotopy group in the uniaxial nematic phase is given by @xcite @xmath117 \cong \mathbb{z}_2 \times \mathbb{z}_2.\end{aligned}\ ] ] the first and second @xmath100 on the right - hand side represent the charge of a vortex and that of a point defect , respectively . here the topological influence of the vortex changes the charge of the point defect from @xmath19 to @xmath100 in which even and odd integers form two equivalent classes .
in this section , we shift focus to the topological phenomena under the charge conservation . under the topological influence ,
the total topological charge is _ prima facie _ not conserved . however , in an isolated system such as ultracold atomic gases , the topological charge can not escape from the system .
thus , the topological influence needs to be revisited so that it is consistent with the charge conservation .
when imposing the charge conservation , we naturally derive the back - action on a vortex as a counteraction of the topological influence .
we call it a _ topological back - action_. the topological back - action is physically defined as a twist of the vortex " , around which a vortex ( a point defect ) rotates . in the case of the influence of a vortex , the back - action is equivalent to a change in the topological charge of the vortex at the rotating center . on the other hand ,
the back - action of a point defect causes two vortex loops on the vortex line at the rotating center . in
what follows , we show the relationship between the topological influence and the twist of the vortex , which acquires a nontrivial topological charge given by @xmath118 when the vortex and the point defect coexist . furthermore , to characterize the topological influence and the back - action in a unified way
, we introduce a _
topological charge transfer _ , which relates the topological influence with the back - action .
we formulate the charge transfer based on the scheme of bucher _ at el . _
@xcite who discussed the topology of an alice cosmic ring , which has the same topological nature as that of a half - quantum vortex ( hqv ) ring in the polar phase of a spin-1 bec and the superfluid @xmath29he - a dipole - free state . also , we show that the charge transfer can be classified by the commutator @xmath119 $ ] ( @xmath120 ) .
first , we show the connection between the topological back - action on a vortex and a twist of a vortex . for the case of non - abelian vortices ,
the topological back - action is nothing but a change in the topological charge of the vortex since the topological charges are described by the fundamental group @xmath8 . on the other hand , for the topological influence on a point defect ,
the topological back - action on a vortex is no longer described by the fundamental group because these charges belong to the higher homotopy groups .
thus , we call for a new understanding in higher - dimensional ( @xmath25 ) cases . to resolve this problem , we show that a twist of a vortex gives the consistent charge . to understand the relation between the topological influence and the twist of a vortex intuitively , let us consider a hqv and specify a field configuration @xmath121 along a line which connects an order parameter of a point defect with that of the hqv as illustrated in fig .
[ fig : back - ation ] ( a ) . as shown in fig .
[ fig : back - ation ] ( b ) , if the point defect makes a complete circuit around the hqv , @xmath121 also winds itself around the hqv once .
after that , we twist the hqv so as to unwind @xmath121 . as a result ,
the winding of @xmath121 gives rise to the twist of the vortex , which is a direct manifestation of topological back - action .
furthermore , this twist can continuously transform to a pair of hqv rings , each of which carries not only the charge of a vortex , but also that of a point defect @xcite .
each hqv ring thus carries the topological @xmath97 charge of + 1 due to the nontrivial twist , so that the total topological charge is invariant under the topological influence such that @xmath122 . in what follows
, we prove that the twist of a vortex is caused by the topological back - action , which is associated with the topological charge given by @xmath118 . to characterize a topological charge ,
we first define a loop space as @xmath123,\end{aligned}\ ] ] where maps[@xmath124 is a set of maps from @xmath125 to @xmath7 , which is called a loop space .
each element of @xmath126 corresponds to a loop @xmath127 , where @xmath128 $ ] .
when a loop encloses a vortex core , its homotopy equivalence classes define the topological @xmath31 charge of the vortex .
when we consider a vortex line , we have a set of loops along the vortex line . without loss of generality
, we can assume that a loop enclosing the top edge of the vortex line @xmath129 has the same configuration as that enclosing the bottom edge of the vortex line @xmath130 as long as there exist a continuous transformation between them . to characterize a vortex line , we redefine each loop as @xmath131 , where @xmath132 $ ] parametrizes each loop and @xmath133 $ ] parametrizes the location of the loop along a vortex line .
for a given @xmath133 $ ] , @xmath134 corresponds to a loop @xmath135 enclosing a vortex core at @xmath136 . by assumption , @xmath134 satisfies the boundary condition as @xmath137 .
thus , a whole set of loops described by a configuration of a vortex line is characterized by @xmath134 .
this vortex line can transform to another vortex line characterized by @xmath138 if @xmath134 is homotopic to @xmath138 within @xmath7 .
this statement is equivalent to the classification of loops in the loop space @xmath126 so that @xmath134 represents a loop in @xmath126 .
thus , a topological charge of the vortex line is given by the fundamental group of the loop space : @xmath139 .
for a loop space , the following isomorphism holds ( e.g. see ref . @xcite chapter 1 ) : @xmath140 thus , a nontrivial twist of a vortex line is equivalent to a topological charge classified by @xmath141 .
next , we relate the topological influence with the back - action via a charge transfer .
let us consider a situation in which two non - abelian vortices with topological charges @xmath59 and @xmath142 coexist , as illustrated in the leftmost figure in fig .
[ fig : conservation - vortex ] .
( see fig .
[ fig : loop - deformation ] for the definitions of @xmath59 and @xmath60 . ) here we assume @xmath59 does not commute with @xmath60 and the total charge of the system is given by @xmath143 .
as demonstrated in the previous section , the topological influence of @xmath59 on @xmath60 is given by @xmath144 where @xmath63 is the topological charge after the topological influence .
to make @xmath145 invariant under the topological influence , we require the topological back - action to satisfy @xmath146 so that the topological charge conservation is met : @xmath147 . here
@xmath148 is depicted in the rightmost figure of fig .
[ fig : conservation - vortex ] . as a concrete example , let us consider the biaxial nematic phase , where @xmath59 and @xmath60 are given by @xmath58 and @xmath57 , respectively . from eqs .
( [ eq : inf - navortex ] ) and ( [ eq : back - navortex ] ) , @xmath148 and @xmath63 are calculated to be @xmath149 and @xmath65 .
thus , we can readily confirm the charge conservation .
in addition , eqs .
( [ eq : inf - navortex ] ) and ( [ eq : back - navortex ] ) are rewritten as @xmath150^{-1 } , \\ & \gamma_2 ' = [ \gamma_1^{-1 } , \gamma_2 ] \ast \gamma_2,\end{aligned}\ ] ] where the commutator @xmath151 $ ] is defined by @xmath152 : = \gamma_1^{-1 } \ast \gamma_2 \ast \gamma_1 \ast \gamma_2^{-1}.\end{aligned}\ ] ] thus , the topological influence is related to the back - action via the commutator of the fundamental group @xmath151 $ ] , which plays a central role in the charge transfer . and @xmath153 . in the rightmost figure , we show that @xmath148 is equivalent to a loop which encircles v1 , but does not encircle v2.,width=302 ] similarly , we can define the charge transfer for the higher - dimensional case .
we employ the method by bucher _ at el . _
@xcite for an alice cosmic ring .
we consider a system in which an hqv ring and a point defect coexist in 3d space as shown in the left figure of fig .
[ fig : consevation - point ] , where the topological charge of the hqv and that of the point defect are denoted by @xmath154 and @xmath91 @xmath155 , respectively .
in this situation , the total topological charge is given by @xmath156 , where @xmath157 " means the additive operation since @xmath23 ( @xmath25 ) is abelian . according to sec .
[ sec : influence ] , the topological influence is given by @xmath158 in a manner similar to the case of non - abelian vortices , we define the topological back - action as @xmath159 the image of @xmath160 is illustrated in the right figure of fig .
[ fig : consevation - point ] . using a commutator of the homotopy groups ,
we can rewrite eqs .
( [ eq : inf_point ] ) and ( [ eq : back_point ] ) as @xmath161 , \\ & \alpha ' = \alpha + [ \gamma^{-1},\alpha],\end{aligned}\ ] ] where the commutator is defined by @xmath162 : = \gamma^{-1 } \ast \alpha \ast \gamma - \alpha . \end{aligned}\ ] ] thus , in the case of the point defect , the commutator @xmath163 $ ] describes the charge transfer due to the topological influence .
for example , in the uniaxial nematic phase , the charge of a point defect changes from @xmath91 to @xmath164 under the topological influence .
thus , the topological back - action is given by @xmath165 .
especially , for @xmath166 and @xmath167 , we have @xmath168 , which corresponds to the twist of a vortex discussed in the previous subsection . using the same approach
, we can show that the charge transfer is generally classified by the commutator of homotopy groups @xmath169 $ ] , where @xmath170 and @xmath171 . and @xmath172 , where @xmath154 and @xmath91 denote the topological charge of a half quantum vortex ( hqv ) and that of a point defect , respectively .
in the right figure , we show a closed surface @xmath172 in which @xmath92 passes through a hqv and then encloses a point defect.,width=302 ] finally , we make a general remark on the charge transfer for a multiple topological excitation system . as we discussed in previous subsections ,
the charge transfer indeed occurs between two topological excitations . in this subsection , consider a general situation in which there are @xmath173 topological excitations , and investigate the effect of the charge transfer between two of them on the remaining @xmath174 topological excitations .
let us define the total charge of @xmath173 non - abelian vortices @xmath175 ( @xmath176 ) by @xmath177 where a set of labels @xmath178 specifies the order of non - abelian vortices which are placed from the left to the right with an order of increasing @xmath179 .
also , the total charge of @xmath180 vortex rings @xmath181 ( @xmath182 ) and @xmath183 point defects @xmath184 ( @xmath185 )
is defined by @xmath186 where @xmath187 . under the topological charge conservation
, we consider that a vortex with charge @xmath188 rotates around a vortex with charge @xmath189 ( we assume @xmath190 ) . by definition ,
the topological back - action and influence are given by @xmath191^{-1 } , \label{eq : ex_back } \\ & \gamma_{n_j } \mapsto \gamma_{n_j } ' = [ \gamma_{n_i}^{-1 } , \gamma_{n_j } ] \ast \gamma_{n_j } , \label{eq : ex_inf}\end{aligned}\ ] ] where the vortices labeled by @xmath192 stay at their original positions .
although we do not perform any operations on the remaining vortices , the charge of vortices @xmath193 changes so as to respect the conservation law : @xmath194 \ast \gamma_{n_{i+1 } } \ast \cdots \ast \gamma_{n_{j-1 } } \ast [ \gamma_{n_i}^{-1 } , \gamma_{n_j}]^{-1}. \label{eq : ex_remain}\end{aligned}\ ] ] it follows from eqs .
( [ eq : ex_back ] ) , ( [ eq : ex_inf ] ) , and ( [ eq : ex_remain ] ) that @xmath145 is invariant .
thus , the charge transfer from @xmath189 to @xmath188 gives an effect on the charge of vortices in the intermediate region .
in general , the effect of a charge transfer depends on the initial condition of @xmath145 . in other words ,
the topological charges are sensitive to a relative position of each vortex due to the noncommutativity of topological charges . in a similar fashion ,
when a point defect with charge @xmath195 passes through a center of a vortex ring with charge @xmath196 , the topological back - action and influence are given by @xmath197 , \label{eq : ex_back_ndim } \\
& \alpha_j \mapsto \ ; \alpha_j ' = \alpha_j + [ \gamma_i^{-1 } , \alpha_j ] .
\label{eq : ex_inf_ndim}\end{aligned}\ ] ] since topological charges are abelian , we can verify the charge conservation by substituting eqs .
( [ eq : ex_back_ndim ] ) and ( [ eq : ex_inf_ndim ] ) into eq .
( [ eq : tot_ndim ] ) .
thus , in the case of point defects , the charge transfer has no effect on other topological charges due to the commutativity .
in what follows , we show some concrete examples of charge transfers in liquid crystals , spinor becs and superfluid @xmath29he . as described in sec . [
sec : influence ] , the topological influence on a non - abelian vortex and a point defect exists for the uniaxial and biaxial nematic phase , respectively .
thus , these phases show a nontrivial charge transfer under the charge conservation .
let us label a disclination and a point defect by @xmath90 and @xmath91 , respectively .
then , the topological influence on the point defect is described by @xmath198 thus , by definition , the commutator is given by @xmath199 = -2\alpha.\end{aligned}\ ] ] in general , a point defect with charge @xmath200 is labeled by @xmath201 , and thus @xmath202 = -2n\alpha$ ] .
therefore , the charge transfer for the point defect is classified by @xmath203 \cong 2 \mathbb{z}. \label{eq : ct - un}\end{aligned}\ ] ] where @xmath204 implies even integers . in the biaxial nematic phase , disclinations are classified by the eight - element quaternion group @xmath51 in eq .
( [ eq : vortex_bn ] ) .
the commutator of @xmath58 and @xmath57 gives @xmath205 & = ( i \sigma_x ) ( i \sigma_y ) ( i \sigma_x)^{-1 } ( i \sigma_y)^{-1 } \notag \\ & = ( i \sigma_z ) ^2 = - \bm{1}_2.\end{aligned}\ ] ] this result holds for all commutators @xmath206 $ ] with @xmath207 . as @xmath208 is its own inverse in the quaternion group ,
the action and back - action are identical .
the charge transfer is classified by @xmath209 \cong \mathbb{z}_2.\end{aligned}\ ] ] thus , there is only one nontrivial charge transfer .
[ cols="^,^,^,^,^,^,^",options="header " , ] next , we see the charge transfer for spin-1 and spin-2 becs .
the spin-1 becs have been realized for @xmath210na and @xmath211rb and the spin-2 bec for @xmath211rb .
according to the homotopy classification in refs .
@xcite , non - abelian vortices exist for spin-2 biaxial nematic and spin-2 cyclic phases .
also , the topological influence on a point defect arises for the spin-1 polar phase and the spin-2 uniaxial nematic phase .
since the charge transfer of point defects is calculated in a manner similar to the case of the uniaxial nematic phase , we consider here only the non - abelian vortices .
the opm of the biaxial nematic phase of a spin-2 bec is given by @xcite @xmath212 where the subscripts @xmath213 and @xmath214 indicate the global gauge symmetry and the spin rotation symmetry , respectively , @xmath215 is a spin - gauge - coupled fourth dihedral group , and @xmath216 is its binary group due to the lift of @xmath38 .
the fundamental group is given by @xcite @xmath217 where we introduce an @xmath32-product @xmath218 since @xmath219 involves the global gauge degrees of freedom .
the definition of the @xmath32-product is given in appendix [ app : a ] .
since @xmath19 is abelian , we can ignore it and we discuss only the binary fourth dihedral group @xmath219 .
the fourth dihedral group consists of the identity element , three four - fold rotations , and a two - fold rotation about an axis perpendicular to the rotation axis .
since the binary fourth dihedral group is the double representation of @xmath215 , it can simply be described by @xmath220 where @xmath221 and @xmath222 are the generators of the group satisfying the three relations on their right . then , an element of @xmath219 is described by @xmath223 and @xmath224 ( @xmath225 ) .
the nontrivial commutators arise from @xmath226 $ ] and @xmath227 $ ] ( @xmath228 ) , which are calculated to be @xmath229 & = r^{2n } , \label{eq : calc_rrs}\\ [ r^n s , r^m s ] & = r^{2(n - m)}. \label{eq : calc_rsrs}\end{aligned}\ ] ] in eq .
( [ eq : calc_rrs ] ) , @xmath230 has no effect on the commutator because we have @xmath226 = r^{n+m}s r^{-n}s^{-1 } r^{-m}$ ] @xmath231 .
therefore , the commutator subgroup is given by @xmath232 \cong \mathbb{z}_4.\end{aligned}\ ] ] this result implies that we have only three nontrivial charge transfers .
furthermore , in this case , we can immediately see that some pairs of vortices behave as @xmath100 .
in fact , considering winding an @xmath233 vortex around an @xmath223 vortex , we find that the commutator in this case is given in eq.([eq : calc_rrs ] ) as @xmath234 .
thus , the topological back - action becomes @xmath235^{-1 } = r^{-n},\end{aligned}\ ] ] which implies that a second winding will reverse this action .
we therefore have a @xmath100-action .
however , if we consider winding an @xmath233 vortex around an @xmath236 vortex , the commutator is @xmath237 , and therefore @xmath238r^m s & = r^{m+2(n - m)}s \label{eq : calc_rm - rn_a},\\ r^n s \mapsto r^n s [ r^ms , r^ns]^{-1}&= r^n s r^{-2(n - m ) } \notag \\ & = r^n s r^{-2(n - m)}s^{-1 } s \notag \\ & = r^{n+2(n - m)}s .
\label{eq : calc_rm - rn_b}\end{aligned}\ ] ] equations ( [ eq : calc_rm - rn_a ] ) and ( [ eq : calc_rm - rn_b ] ) are simply expressed by @xmath239 thus , we find @xmath240 .
since @xmath241 , we see immediately that this will be a @xmath242-action if @xmath243 is odd , and a @xmath100-action if @xmath243 is even . next , we consider the cyclic phase of a spin-2 bec , whose opm is given by @xcite @xmath244 where @xmath245 is a spin - gauge - coupled tetrahedral group and @xmath246 is its binary form .
the fundamental group becomes @xcite @xmath247 where @xmath19 results from the @xmath15 degrees of freedom . since @xmath246 is the spin - gauge coupled symmetry , @xmath19 and @xmath246 are described by the @xmath32-product defined in appendix [ app : a ] .
similar to the biaxial nematic phase , we can ignore @xmath19 , since @xmath19 is abelian .
we consider the commutator of @xmath246 .
we can show the equivalent form of @xmath246 as @xmath248 where the semidirect product `` @xmath249 '' is defined as follows : element @xmath250 in @xmath251 does nothing ; element @xmath252 performs a permutation , i.e. , @xmath253 , @xmath254 , and @xmath255 ; and element @xmath256 performs the cyclic permutation , i.e. , @xmath257 , @xmath258 , and @xmath259 .
for example , we consider the product of @xmath260 and @xmath261 , which is given by @xmath262 to derive the commutator subgroup , we consider the following two commutators : @xmath263 & = ( \bm{0},i \sigma_x ) ( \bm{0 } , i \sigma_y)(\bm{0},-i \sigma_x ) ( \bm{0 } , -i \sigma_y ) \notag \\ & = ( \bm{0 } , -\bm{1}_2 ) , \\
[ ( \bm{1},i \sigma_x ) , ( \bm{1 } , i \sigma_y ) ] & = ( \bm{1},i \sigma_x ) ( \bm{1 } , i \sigma_y)(\bm{2},-i \sigma_z ) ( \bm{2 } , -i \sigma_x ) \notag \\ & = ( \bm{0 } , i \sigma_z ) , \end{aligned}\ ] ] where the first expression indicates that we have the @xmath264 element of @xmath51 , and the second one shows that @xmath58 , @xmath265 , and @xmath266 belong to the commutator subgroup .
the closure of the commutator subgroup implies that it is at least the quaternions .
however , as the @xmath267 is abelian , the commutator subgroup is isomorphic to the quaternion ; namely , @xmath268 \cong q_8.\end{aligned}\ ] ] thus , the charge transfer in the cyclic phase is described by the eight - element quaternion group @xmath51 , and furthermore since @xmath51 is non - abelian , the cyclic phase accommodates the non - abelian charge transfer .
finally , we show the existence of the topological back - action in the superfluid @xmath29he , in which the order parameter possesses the orbital and spin degrees of freedom , and each internal degree of freedom behaves like a vector in the orbital and spin space due to the spin triplet @xmath269-wave state . in general
, the order parameter is described by @xmath270 where @xmath271 refers to @xmath272 in the spin space and @xmath273 refers to @xmath274 in the orbital space , and @xmath275 is a complex 3-by-3 matrix .
the possible phases and topological excitations are discussed in refs .
@xcite . according to them , the topological influence between the vortex and the point defect exists for the dipole - free state in the superfluid @xmath29he - a phase .
thus , we expect the topological back - action through the topological charge conservation .
the order parameter is given by @xmath276 where @xmath277 is an amplitude of the order parameter , and @xmath278 , @xmath279 , and @xmath280 are unit vectors which satisfy @xmath281 .
also , the opm is given by @xcite @xmath282 where the subscripts @xmath283 and @xmath284 represent the orbital and spin symmetries .
the corresponding first and second homotopy groups are given by @xmath285 the topological influence occurs for @xmath286 and @xmath287 , which give the @xmath100-action on @xmath97 ; i.e. @xmath288 changes into its inverse @xmath289 . therefore
, the calculation of the charge transfer is the same as in the case of the uniaxial nematic phase .
we summarize the classification of charge transfers in table .
[ tab : core ] .
in this section , we discuss the issues concerning the definition of the total charge and the measurement of topological charges .
if the homotopy group is abelian , we can observe topological excitations individually . in this case , the total charge is given by the sum of individual topological charges .
the representative examples are superfluid @xmath14he and scalar becs ( fully polarized spinor becs ) , in which the order parameter manifold is given by @xmath15 . since @xmath16 , it is possible to observe the winding number of the vortices individually .
however , if the homotopy group is non - abelian , the counting of individual defects breaks down due to the influence of the coexisting defects . in this case , the topological charge is determined up to the conjugacy class of the homotopy group .
thus , the sum of each individual topological charge is insufficient to define the total charge .
nevertheless , in the closed system , the total charge would be defined by encoding the values of the order parameter on the boundary of the system , which remains unchanged under the adiabatic manipulation .
thus , the total charge is measurable by probing the system boundary without regard to the topological influence .
next , we state the detection of each individual topological charge in the topological charge conserving framework .
we suggest that we could determine individual topological charges through the testing of pair annihilation .
for example , let us consider the system with a vortex v@xmath62 .
assuming that we can create a vortex - antivortex pair and manipulate vortices adiabatically , we create the vortex - antivortex pair : v@xmath61 and v@xmath61 near the vortex v@xmath62 , and then we rotate the vortex v@xmath61 about the vortex v@xmath62 adiabatically in the clockwise direction . finally , we fuse the vortex v@xmath61 into the remaining antivortex v@xmath61. the combined charge of v@xmath61 and v@xmath61 turns out to be given by a commutator . as a result
, this test is categorized into three cases : ( 1 ) the vortex pair annihilates each other .
hence , each topological charge is abelian and we can observe the topological charge individually . ( 2 ) the vortex pair survives and its charge belongs to the abelian group .
then , the vortices v@xmath62 and v@xmath61 are not observable individually , but each individual commutator is observable .
thus , a relative charge of v@xmath62 and v@xmath61 is possible to detect through the commutator .
( 3 ) the vortex pair survives and its charge belongs to the non - abelian group . in this case
, it is also possible to determine the topological charge uniquely if we have only two types of the vortex .
however , if there are more than two , we can not determine the topological charge uniquely due to the influence between commutators .
namely , we can detect the charge of vortex - antivortex pair up to the conjugacy class .
fortunately , for the real system tabulated at table [ tab : core ] , the commutator subgroup is abelian except for the cyclic phase in the spin-2 becs .
thus , almost all ordered phases are categorized into the case ( 1 ) or ( 2 ) . in the cyclic phase ,
the commutator subgroup is isomorphic to the eight - element quaternion group .
namely , this phase belongs to the case ( 3 ) . as a concrete example , if we consider non - abelian vortices in the spinor becs , the total charge would be estimated by analyzing the spatial variations of the order parameter at the boundary through a stern - gerlach separation , in which the populations in each magnetic sublevel can be measured @xcite .
on the other hand , each individual topological defect is determined up to the conjugacy class due to the topological influence . to observe the topological charge in the charge conserving scheme
, it is necessary to apply the pair annihilation testing to individual topological defects .
however , such manipulation of the topological defects has yet to be achieved in the experiment of spinor becs .
in the present paper , we have discussed the relation between the topological charge conservation and the topological influence . to ensure the consistency with the charge conservation , we have introduced the topological back - action , and shown that the topological back - action causes twisting of a vortex line for the point - defect case .
the twist of a vortex corresponds to two hqv rings ( alice rings ) attached on two edges of a vortex line , whose topological charge is characterized by @xmath141 .
this idea might be generalized to higher - dimensional cases ( @xmath290 ) , in which a dimension of a vortex is @xmath291 , where @xmath9 is the space dimension .
thus , a surface enclosing a vortex is given by @xmath292 , where @xmath293 .
here , @xmath294 includes the degrees of freedom to encircle the vortex core and we assume that each edge has the same configuration .
the topological back - action from a point defect in a @xmath9-dimensional space is given by the homotopy equivalence classes of @xmath294 ; i.e. @xmath295 where @xmath296 describes a topological charge of a point defect in the @xmath9-dimensional space .
using the method by bucher _
et al . _
@xcite , we have related the topological influence with its back - action via the commutators of the homotopy groups @xmath119 $ ] ( @xmath297 ) , which plays a key role in the charge transfer in a multiple topological excitation system . for liquid crystals , spinor becs , and superfluid @xmath29he ,
we have calculated commutators and determined the classifications of the charge transfer for the non - abelian vortices : one type for the bn phase , three types for the spin-2 biaxial nematic phase , and seven types for the cyclic phase .
in particular , we have found non - abelian charge transfer in the cyclic phase , which may lead to new topological phenomena .
it would be worthwhile to mention that the commutator between @xmath31 and @xmath30 represents the whitehead bracket @xcite , which suggests a new direction of this study . in higher - dimensional cases , the un phase , the spin-1 polar phase , the spin-2 uniaxial phase , and the superfluid @xmath29he - a phase
show the charge transfer classified by an even integer due to the topological influence of the hqv ( disclination ) .
this charge transfer depends mainly on a set of the automorphism maps @xmath27 .
also , if we consider the action of @xmath31 on @xmath30 to be a homomorphism from @xmath31 to @xmath298 , then we can rule out topological influence in a number of different cases on group - theoretic grounds .
in particular ,
when @xmath299 , @xmath300 is always @xmath100 .
thus , we can have only the @xmath100-action on the point defect .
for example , in the case of the dipole - free @xmath29he - a phase , @xmath301 and @xmath302 , and the @xmath31 action on @xmath97 is nontrivial @xcite . for @xmath303
, if we consider the homomorphism map from @xmath242 into @xmath304 , the only such homomorphism takes @xmath305 to an identity map .
thus , @xmath306 and @xmath305 will not cause any effect , while @xmath307 and @xmath308 will .
also , we have suggested that the pair annihilation testing enables us to detect the topological charge in the charge conservation framework .
this testing is applicable to the higher dimensional cases and is useful for detecting evidence of the nature of non - abelian charge . in this paper , we have focused on the role of the charge conservation in a multiple topological excitation system .
however , our calculation is also applicable to statistics of topological excitations for situations in which two topological excitations exchange their charges .
it has been shown in refs .
@xcite that non - abelian vortices obey the partially colored braid statistics " because a topological charge of a vortex is indistinguishable within the conjugacy class . the biaxial nematic phase in the liquid crystal exhibits the non - abelian vortex ( line defect ) classified by @xmath51 , but its commutator is @xmath309
thus , the exchange of vortices has a simple algebraic structure . on the other hand ,
the spin-2 biaxial nematic and spin-2 cyclic phases exhibit several different types of the charge transfer , offering a piece of evidence on the richness of statistics .
this work was supported by kakenhi 22340114 , a grant - in - aid for scientific research on innovation areas topological quantum phenomena " ( kakenhi 22103005 ) and the photon frontier network program , from mext of japan . s.k .
acknowledges support from jsps ( grant no .
in this appendix , we give the definitions of @xmath84 and @xmath310 , which are used in eqs .
( [ eq : gh_un ] ) , ( [ eq : gh_sbn ] ) , and ( [ eq : gh_c ] ) to make this paper self - contained .
the semidirect product of two groups @xmath311 and @xmath312 , which is denoted as @xmath313 , is defined as follows ; for @xmath314 and @xmath315 , the semidirect product between @xmath316 and @xmath317 is given by @xmath318 where @xmath319 .
for instance , in eq .
( [ eq : gh_un ] ) , @xmath311 and @xmath312 are given by @xmath320 and @xmath321 , respectively .
for @xmath322 and @xmath323 , the semidirect product is given by @xmath324 where @xmath325 is given by @xmath326 to satisfy the group structure between the spin - gauge - coupled discrete group @xmath327 and @xmath19 , we define an @xmath32-product @xmath310 as follows : given a discrete group @xmath328 involving both the global gauge and the spin rotation , the group structure of @xmath329 is given by @xmath330 where @xmath331 and @xmath332 , @xmath333 , and @xmath334 . here
@xmath32 is a map from @xmath335 to @xmath19 , which is given by @xmath336 examples for the spin-2 biaxial nematic and spin-2 cyclic phases are described in ref .
@xcite .
999 p. d. gennes and j. prost , _ the physics of liquid crystals _
, oxford university press , oxford ( 1993 ) . m. kleman , _ points , lines and walls in liquid crystals , magnetic systems and various ordered media _ wiley , new york ( 1983 ) . r. j. donnelly , _ quantized vortices in helium ii _ , cambridge university press , cambridge ( 1991 ) .
d. vollhardt , and p. wolfle , _ the superfluid phases of helium 3 _ , taylor and francis , london ( 1990 ) .
g. e. volovik , _ the universe in a helium droplet _ , oxford university press , new york ( 2003 ) .
d. m. stamper - kurn and m. ueda , rev .
phys . * 85 * , 1191 ( 2013 ) .
y. kawaguchi and m. ueda , phys . rep . *
520 * , 253 ( 2012 ) .
n. d. mermin , rev .
phys . * 51 * , 591 ( 1979 ) .
l. michel , rev .
phys . * 52 * , 617 ( 1980 ) .
trebin , adv .
phys . * 31 * , 195 ( 1982 ) .
v. p. mineev , _ topological stable defects and solitons in ordered media _ , classical reviews in physics , vol.1 , harwood accademic , amsterdam ( 1998 ) . g. e. volovik and v. p. mineev , zh .
teor . fiz . *
45 * , 1186 ( 1977 ) [ sov .
jetp * 45 * , 1186 ( 1977 ) ] .
n. d. mermin , j. math . phys . * 19 * 1457 ( 1978 ) .
a. s. schwarz , nucl .
b * 208 * , 141 ( 1982 ) .
s. kobayashi , m. kobayashi , y. kawaguchi , m. nitta , and m. ueda , nucl .
b * 856 * , 577 ( 2012 ) .
h. mkl , y. zhang , and k .- a .
suominen , j. phys .
a : math . gen . * 36 * , 8555 ( 2003 ) .
h. mkl , j. phys .
a : math . gen . * 39 * , 7423 ( 2006 ) .
g. w. semenoff and f. zhou , phys .
* 98 * , 100401 ( 2007 ) .
r. m. switzer , _ algebraic topology - homology and homotopy _ , springer - verlag , berlin - heidelberg - new york ( 1975 ) .
lo and j. preskill , phys .
d , * 48 * , 4821 ( 1993 ) .
m. nakahara , _ geometry , topology and physics second edition _ ,
taylor and francis , london ( 2003 ) .
m. bucher , h .- k .
lo , and j. preskill , nucl .
b , * 386 * , 3 ( 1992 ) .
h. mori and h. nakanishi , j. phys .
jpn , * 57 * , 1281 ( 1988 ) .
h. nakanishi , k. hayashi , and h. mori , commun .
. phys . * 117 * , 203 ( 1988 ) .
f. a. bais and p. john , int .
a , * 10 * , 3241 ( 1995 ) . j. l. song , g. w. semenoff , and f. zhou , phys .
lett . * 98 * , 160408 ( 2007 ) .
s. uchino , m. kobayashi , m. nitta , and m. ueda , phys .
lett . * 105 * , 230406 ( 2010 ) .
ho , phys . rev .
* 81 * , 742 , ( 1998 ) .
f. zhou , phys .
lett . * 87 * , 080401 ( 2001 ) .
y choi , w. j. kwon , and y .-
i shin , phys .
lett . * 108 * , 035301 ( 2012 ) .
j. h. c. whitehead , ann . of math .
* 42 * , 409 ( 1941 ) .
l. brekke , a. f. falk , s. j. hughes and t. d. imbo , phys .
b , * 271 * , 73 ( 1991 ) .
l. brekke , h. dykstra a. f. falk , and t. d. imbo , phys .
b , * 304 * , 127 ( 1993 ) . | topological objects can influence each other if the underlying homotopy groups are non - abelian . under such circumstances ,
the topological charge of each individual object is no longer a conserved quantity and can be transformed to each other . yet
, we can identify the conservation law by considering the back - action of topological influence .
we develop a general theory of topological influence and back - action based on the commutators of the underlying homotopy groups . in the case of the topological influence of a half - quantum vortex on a point defect , we point out that the topological back - action from the point defect is a twisting of the vortex .
the total twist of the vortex line compensates for the change in the point - defect charge to conserve the total charge .
we use this theory to classify charge transfers in condensed matter systems and show that a non - abelian charge transfer can be realized in a spin-2 bose - einstein condensate . |
thermal transport at the nanoscale is a problem of great fundamental and practical interest @xcite .
figure [ fig1 ] shows the temperature profiles across silicon films of varying length ( @xmath0 ) , as computed by the phonon boltzmann transport equation ( bte ) @xcite
. the temperature jumps at the interfaces with the two ideal , reflectionless contacts are characteristic features of quasi - ballistic phonon transport , and are commonly observed @xcite in physically - detailed modeling ( such as monte carlo simulations@xcite , molecular dynamics @xcite , or solutions of the phonon bte @xcite ) . in practical situations ,
the routine analysis of nanoscale heat transport phenomena , including ballistic effects , has been limited by the high computational demand of rigorous simulations .
simple , accurate and physically transparent models that provide physical insight could help in understanding the results of detailed simulations , as well as the analysis of experiments .
quasi - ballistic phonon transport can impact the thermal response of materials on the nanoscale and even on the microscale , given that phonon mean - free - paths ( mfp ) can span from @xmath1 to @xmath2 @xcite .
ballistic phonon effects reduce the heat carrying capability of thin - films from the value expected from a simple application of fourier s law @xcite .
this also affects the analysis of experiments probing short time and length scales ( e.g. , time / frequency - domain thermoreflectance ) @xcite , influences heating in small electronic devices @xcite , and can provide a route to extract the mfp distribution of materials @xcite .
a simple , physically sound , accurate , and computationally efficient technique to analyze such problems is presented in this paper .
the lines in fig .
[ fig1 ] are the results of our calculations described below . versus normalized position @xmath3 for a si film of @xmath4 .
symbols are results of the phonon bte ( taken from @xcite ) .
lines are solutions to the approach described in this paper , which are obtained by solving a very simple phonon bte or , equivalently , by solving @xmath5 with physically correct boundary conditions.,width=245 ] in this paper , we begin with an approach originally introduced by mckelvey @xcite and shockley @xcite to describe particle transport , and extend it to treat heat transport .
this technique uses a particularly simple , but accurate , discretization of the bte into forward and reverse fluxes that provide solutions from the ballistic to diffusive limits .
additionally , we show that the mckelvey - shockley phonon bte can be recast exactly as fourier s law and the heat equation with no additional assumptions ( such as restrictions on the size of the structure or assumption of local thermal equilibrium ) . when solved with physically meaningful boundary conditions ,
the solutions are identical to those of the mckelvey - shockley phonon bte .
the lines in fig .
[ fig1 ] are solutions of the heat equation , @xmath6 , using the correct physical boundary conditions , as described in this paper .
the outline of this paper is as follows .
section [ sec : theory ] describes our approach to heat transport .
a simple example calculation is presented in section [ sec : example ] , where we show that the magnitude of the temperature jump is proportional to the phonon transmission across the structure .
section [ sec : fourier ] provides a derivation of familiar heat transport equations from the mckelvey - shockley phonon bte .
although familiar in form , the heat equation includes ballistic effects _ if proper boundary conditions are used_. section [ sec : discussion ] discusses the treatment of a full phonon dispersion with energy - dependent mean - free - path , and explains how the calculations shown in fig .
[ fig1 ] are performed .
finally , we summarize our findings in section [ sec : summary ] .
for this work , we borrow an approach originally developed for electronic transport that is applicable on all length scales , the mckelvey - shockley flux method @xcite , and adapt it for phonon / heat transport .
we assume steady - state 1d transport along @xmath7 with an infinite @xmath8-@xmath9 plane .
the essence of this approach is to _ (
i ) _ describe phonons in terms of fluxes ( i.e. phonon density times average velocity along the transport direction ) and _ ( ii ) _ categorize all phonons into two components : forward and backward fluxes .
the governing equations of the mckelvey - shockley flux method are @xcite : @xmath10 where @xmath11/@xmath12 are the forward / backward phonon fluxes [ units : # phonons m@xmath13s@xmath14ev@xmath14 ] , @xmath16 is the mean - free - path for backscattering and @xmath17 is the phonon energy .
the above coupled equations describe the evolution of each flux type , which can scatter to / from the opposite flux component .
@xmath16 governs the scattering , and is defined as the average distance travelled along @xmath7 by a phonon with energy @xmath17 before scattering into an opposite moving state ( in the isotropic case , @xmath18 is ( 4/3 ) times the regular mfp @xcite ) . note that the mckelvey - shockley flux method can be derived from the bte @xcite .
the boundary conditions are : @xmath19 where @xmath0 is the length of the thermal conductor .
thus one needs to specify the injected phonon fluxes at both ends : @xmath11 on the left side ( @xmath20 ) and @xmath12 on the right side ( @xmath21 ) . the mckelvey - shockley flux method described by eqns.([mk_flux1]-[mk_flux2 ] ) , with the boundary conditions given by eqns.([flux_bound1]-[flux_bound2 ] ) , forms the basis for our approach to heat transport ( see fig .
[ fig2 ] ) . with a given mean - free - path for backscattering @xmath18 .
by specifying the injected phonon fluxes ( solid arrows ) , the mckelvey - shockley flux equations describe the evolution of the flux components inside the material.,width=321 ] the total heat current ( @xmath22 ) and heat density ( @xmath23 ) are written as : @xmath24\,{\rm d}\epsilon,\;\;\;\text{[w\,m$^{-2}$ ] } \label{def : heat_curr } \\
q^{\rm tot}(x ) & = \int_0^{\infty}\epsilon\,\left[\frac{f^+(x,\epsilon)+f^-(x,\epsilon)}{v_x^+(\epsilon)}\right]\,{\rm d}\epsilon,\;\;\;\text{[j\,m$^{-3}$ ] } \label{def : heat_den}\end{aligned}\ ] ] where @xmath25 is the net phonon flux , @xmath26 is the phonon density and @xmath27 is the average @xmath7-projected velocity ( defined as @xmath28 ) .
the heat current and heat density correspond to multiplying the net phonon flux and the phonon density , respectively , by the energy @xmath17 carried by each phonon ( and integrating over all phonon energies ) . from the above definitions
, we can directly replace @xmath29 in the mckelvey - shockley flux equations ( eqns.([mk_flux1]-[mk_flux2 ] ) ) by @xmath30 , as we will assume from this point on unless otherwise stated .
in addition , to ease the notation we will drop the explicit dependence on @xmath17 in @xmath31 , @xmath18 and @xmath27 , although keep in mind that a final integration over energy is required ( eqns.([def : heat_curr]-[def : heat_den ] ) ) .
equations ( [ mk_flux1 ] ) and ( [ mk_flux2 ] ) comprise a simple bte in which the forward and reverse fluxes have been integrated over angle .
this particular discretization is especially effective in handling the correct physical boundary conditions , where a flux is injected from each side .
inside the device , the carrier distribution can be very far from equilibrium , but each half of the distribution is at equilibrium with its originating contact as it is injected and scattering gradually mixes both flux components . in the limiting case of purely ballistic transport ,
each half of the distribution is in equilibrium with its originating contact .
more complicated discretizations are possible and sometimes necessary , but we will demonstrate the effectiveness of these simple equations in the remainder of the paper
. joined by two contacts in equilibium , each at a specified temperature @xmath32 and @xmath33 .
the contact interfaces are reflectionless and do not scatter phonons .
( b ) forward and backward heat currents versus position @xmath7 .
( c ) temperature profile versus position @xmath7 , which shows the temperature jumps ( @xmath34 ) at the contacts .
the temperature profiles of the forward and backward phonons are shown as dashed lines.,width=283 ]
having presented the mckelvey - shockley flux method adapted for heat transport , we will now demonstrate this approach with an example .
we will consider steady - state thermal transport across a dielectric film of length @xmath0 ( the electronic contribution to thermal transport can be neglected ) , contacted by two _ ideal _ thermalizing contacts each at their respective temperatures @xmath32 ( left contact ) and @xmath33 ( right contact ) , as shown in fig .
[ fig3](a ) .
ideal contacts " in this context assumes that _ ( i ) _ each contact is in thermal equilibrium , with phonon statistics given by the bose - einstein distribution , and _
( ii ) _ the interfaces are reflectionless , thus phonons are not scattered at the contacts . while perfect contacts are assumed in this paper , this is not a fundamental limitation of our approach , as we will discuss later .
we begin by subtracting the flux equations ( [ mk_flux1 ] ) and ( [ mk_flux2 ] ) : @xmath35 this is the energy balance equation , or equivalently the first law of thermodynamics , which states that under steady - state conditions the heat current @xmath36 is a constant along @xmath7 . using this relation with eqns.([mk_flux1]-[mk_flux2 ] ) , we have @xmath37 it is straightforward to show that @xmath38 and @xmath39 have the following solutions : @xmath40 where we used the boundary conditions eqns.([flux_bound1]-[flux_bound2 ] ) .
the forward and backward heat currents @xmath31 vary linearly along @xmath7 , as shown in fig .
[ fig3](b ) .
the difference between @xmath41 and @xmath42 is the net heat current @xmath36 , and the slope of @xmath41 and @xmath42 is @xmath43 .
the net heat current @xmath36 can be extracted by subtracting eq .
( [ fminus ] ) from eq .
( [ fplus ] ) , and isolating @xmath36 : @xmath44 , \label{fnet}\end{aligned}\ ] ] where @xmath45 is the phonon transmission coefficient @xmath46 , corresponding to the probability of a phonon traveling from one contact to the other .
thus , the net heat current is simply the difference in injected heat currents times the phonon transmission coefficient .
if the contacts are in equilibrium , then the injected heat currents from the contacts can be written as @xmath47 where @xmath48 is the distribution of modes of the thermal conductor ( depends only on the phonon dispersion ) @xcite , @xmath49 is the bose - einstein occupation function and @xmath50 is planck s constant . inserting these expressions into eq .
( [ fnet ] ) , we obtain : @xmath51 \label{fnet2 } \\ & \approx \underbrace{\epsilon\,\frac{m}{h}\left(\frac{\lambda}{\lambda+l}\right ) \frac{\partial f_{\rm be}}{\partial t}}_{k } \ , \delta t , \label{fnet3}\end{aligned}\ ] ] where @xmath52 is the thermal conductance , @xmath53 is the ballistic thermal conductance @xcite .
note that eq .
( [ fnet3 ] ) applies in the case of a small @xmath54 , while eq .
( [ fnet2 ] ) does not have this limitation .
the total heat current is calculated by integrating over energy ( eq .
( [ def : heat_curr ] ) ) .
the above expressions for @xmath36 are applicable on _ all _ length scales , and span from ballistic to diffusive transport regimes .
the transmission coefficient controls the length dependence of the heat current . in the ballistic limit @xmath55 , @xmath56 and @xmath36
is independent of length .
note that in this limit @xmath36 is equal to the known expression in the casimir limit ( see appendix [ sec : casimir ] ) . in the diffusive limit @xmath57 , @xmath58 and @xmath59 , as expected from classical scaling .
( [ fnet2 ] ) is identical to @xmath36 obtained with the landauer approach @xcite .
one advantage of the mckelvey - shockley flux method , versus landauer , is that it provides spatial information on the heat transport properties .
( [ def : heat_den ] ) shows that the heat density is the sum @xmath60 divided by the average forward projected velocity @xmath27 .
often it is desirable to replace heat density by temperature .
it is important to note that in small structures ( compared to @xmath18 ) the phonon distribution may be highly non - equilibrium and the definition of temperature , an equilibrium quantity , is ambigious ( no such problem arises with @xmath36 and @xmath61 ) . assuming a small applied @xmath62 ensures the thermal conductor remains near equilibrium , where temperature is well defined . in this case , we can rewrite heat density in terms of temperature using @xmath63 , where @xmath64 is the volumetric heat capacity . and , we can expand the heat currents @xmath65 due to a small difference in contact temperatures @xmath54 , where @xmath66 is simply the equilibrium heat current arising from a reference background temperature , chosen as @xmath33 in this case . rewriting @xmath61 in terms of @xmath67 , we find : @xmath68 + t_r .
\label{def : temp}\end{aligned}\ ] ] the temperature profile inside the thermal conductor is proportional to the sum of @xmath69 and @xmath70 or equivalently the average of @xmath71 and @xmath72 , and is presented in fig .
[ fig3](c ) .
note that @xmath73 , since @xmath64 takes into account both the forward and backward phonon states . using this with the property @xmath74
, we obtain the above expression for @xmath75 , which has the appearance of an average over @xmath76 .
we see that @xmath75 varies linearly inside the thermal conductor , and that there are discrete temperature drops at the boundaries .
the separation of temperature for phonons traveling in the forward and backward directions is completely analogous to the way electrochemical potentials are treated with electron transport in nanostructures @xcite .
allowing the forward- and backward - moving phonon populations , and associated temperatures ( @xmath77 and @xmath78 ) , to be different is key to capturing the non - equilibrium nature of ballistic transport . in the diffusive limit
the difference between @xmath77 and @xmath78 , as well as @xmath41 and @xmath42 , becomes vanishingly small , and corresponds to the case of near local thermal equilibrium ( i.e. both halves of the phonon distribution are nearly identical )
. we can use eqns.([def : temp_pm]-[def : temp ] ) with eqns.([fplus]-[fminus ] ) to determine the temperatures at the boundaries : @xmath79 where the relation @xmath80 was used ( the factor of two appears since the forward and backward heat currents are equal in the contacts ) .
the boundary temperatures @xmath81/@xmath82 are weighted averages of the contact temperatures , that depend on @xmath46 .
when @xmath83 ( diffusive limit ) , the interior " boundary temperatures tend to the contact temperatures ( classical result ) , and when @xmath56 ( ballistic limit ) the boundary temperatures tend to the average of the contact temperatures ( constant @xmath75 inside the material ) @xcite . from eqns.([temp_left]-[temp_right ] ) , we can extract the value of the temperature jumps at the contacts ( @xmath34 ) .
@xmath34 is found to be identical at both left and right contacts : @xmath84 where @xmath85 is the ballistic thermal resistance with @xmath86 .
we find the temperature jumps are simply proportional to the transmission , which depends on @xmath0 and @xmath18 , and can be interpreted as an intrinsic contact resistance .
the temperature jumps do not occur because the contact interfaces scatter phonons , since we assume reflectionless contacts .
rather , it is because we specify the injected heat currents @xmath87/@xmath88 at one end of the thermal conductor while the opposite ends are floating " boundaries that depend on @xmath18 and @xmath0 ( as shown in fig . [ fig3](b ) ) .
once the forward and backward heat currents are added to obtain temperature , discrete drops are observed at the boundaries . in this work
we have assumed perfectly thermalizing contacts that are in equilibrium , which allows us to extract a simple analytical expression for the temperature jumps . in general one
must specify the injected heat currents at the boundaries , which in principle can originate from an adjacent material that need not be in equilibrium . versus normalized position @xmath3 of diamond films of length @xmath0 = @xmath89 ( a ) , @xmath90 ( b ) , @xmath91 ( c ) .
lines correspond to our analytical solution , and symbols are obtained from the phonon bte ( taken from @xcite ) .
the values @xmath92 ( @xmath93 ) and @xmath94 ( @xmath95 ) reported in @xcite were adopted.,width=226 ] in fig .
[ fig4 ] we compare our simple model ( lines ) to the numerical solutions of the bte ( symbols ) , in the case of diamond films of length @xmath0 = @xmath89 , @xmath90 , @xmath91 . using the effective @xmath18 values reported in @xcite for @xmath96
, we find our simple analytical model adequately reproduces the results of the bte , including the linear temperature profiles and the temperature jumps at the contacts .
we note that phonons in a material typically have a broad distribution of @xmath18 , and in general using a constant @xmath18 may lead to significant errors ( in this case @xmath34 should be obtained by evaluating @xmath18 and @xmath46 at each energy and integrating over energy following eqns.([def : heat_curr]-[def : heat_den ] ) , as discussed in section [ sec : discussion ] ) . to summarize , writing the phonon bte in the mckelvey - shockley form ( eqns.([mk_flux1 ] ) and ( [ mk_flux2 ] ) ) leads to very accurate solutions for the class of problems considered in this section .
in the next section , we will show that exactly the same solutions can be obtained by solving the traditional diffusion equation if the appropriate boundary conditions are used .
in this section , we demonstrate how the mckelvey - shockley flux equations can be rewritten into familiar expressions for heat transport , that yield exactly the same results . by adding eq .
( [ mk_flux1 ] ) and eq .
( [ mk_flux2 ] ) ( replacing @xmath29 by @xmath31 ) and using the relation @xmath97 , we obtain @xmath98 where @xmath99 @xmath100 is the phonon diffusion coefficient and @xmath101 is the bulk thermal conductivity ( see appendix [ sec : trad_kappa ] ) . eq .
( [ fourier2 ] ) is fourier s law , and comes out directly from the mckelvey - shockley flux method , _ without making any assumption on @xmath0 relative to @xmath18_. this indicates that the ballistic transport physics contained in the mckelvey - shockley flux method are also included in fourier s law . by combining fourier s law with the energy balance equation ( eq .
( [ cont_eq ] ) ) , we find the steady - state heat equation : @xmath102 where we assumed the material parameters are position - independent .
fourier s law has been derived from the bte @xcite , by assuming that the phonons at each @xmath7 were locally at thermodynamic equilibrium .
we find no such assumption is necessary .
the key to capturing ballistic transport effects with fourier s law is to use the correct physical boundary conditions .
traditionally , the contact temperatures @xmath32/@xmath33 are used as the boundary conditions , however the mckelvey - shockley flux method shows it is the injected heat currents that are the physical boundary conditions ( eqns.([flux_bound1]-[flux_bound2 ] ) ) . stated differently , the bte is a first order equation in space , and requires one boundary condition for the phonon distribution .
this is typically accomplished by defining a boundary condition on half of the distribution ( the incoming flux ) at each of the two contacts .
specifying the contact temperatures as the boundary conditions at both ends overspecifies the problem .
the temperature at each of the two ends is determined by the temperatures of the injected fluxes and by the scattering within the film ; it is an outcome of the calculation .
although , in the diffusive limit , the boundary temperatures tend asymptotically to the contact temperatures . in general , using the definitions of heat current and heat density , the boundary conditions for temperature are determined to be ( details in appendix [ sec : bc_temp ] ) : @xmath103 where @xmath104 , @xmath105/2 $ ] and @xmath33 was chosen as our reference background temperature .
the above boundary conditions are mixed , meaning they depend on both temperature and its derivative , and are applicable even when the contacts are not in equilibrium .
solving the heat equation ( eq .
( [ heat_eq ] ) ) is straightforward , and simply gives a linear temperature profile @xmath106 + t(l^-)\,\left[\frac{x}{l}\right ] .
\label{temp_profile}\end{aligned}\ ] ] inserting this solution into eqns.([bc_temp1]-[bc_temp2 ] ) , we can determine @xmath81 and @xmath82 , which in this case ( equilibrium contacts ) are equal to eqns.([temp_left]-[temp_right ] )
. with equilibrium contacts , applying the above boundary conditions is equivalent to replacing the contact temperatures @xmath32 and @xmath33 by the interior " boundary temperatures @xmath81 and @xmath82 .
we note that using eqns.([temp_left]-[temp_right ] ) with eq .
( [ temp_jump2 ] ) , @xmath81 and @xmath82 can also be rewritten as @xmath107 with a given @xmath75 , we can calculate the heat current using fourier s law ( eq .
( [ fourier2 ] ) ) : @xmath108 , \label{heat_curr1}\\ & = \kappa \ , \left[\frac{t_l - t_r}{l+\lambda } \right ] .
\label{heat_curr2}\end{aligned}\ ] ] note the correct usage of fourier s law implies evaluating the gradient of @xmath67 using the interior " boundary temperatures and not the contact temperatures .
this expression for heat current is applicable on _ all _ length scales , and is equal to eq .
( [ fnet3 ] ) . in the diffusive limit ( @xmath57 )
, we have @xmath109 , the classical result ; in the ballistic limit ( @xmath55 ) , we have @xmath110 and the heat current is independent of @xmath0 .
( [ heat_curr2 ] ) shows that @xmath111 is reduced , due to the temperature jumps at the boundaries , and is equivalent to replacing @xmath0 by @xmath112 . as
transport becomes more ballistic , the temperature jumps increase and the absolute temperature gradient inside the thermal conductor decreases . if one assumed @xmath113 for all @xmath0 , then a reduction in the expected @xmath36 could be interpreted as a reduction in the thermal conductivity .
this introduces the concept of apparent thermal conductivity , which is mathematically defined as @xmath114 @xcite .
our analysis shows that , physically , ballistic transport reduces @xmath115 and does not change the bulk thermal conductivity of the material at any length scale as long as @xmath18 is independent of energy .
in section [ sec : example ] , we validated our approach by comparing our solutions to the numerical results of the bte where a constant @xmath18 was used .
in general , the mfp of phonons in a given material can span orders of magnitude @xcite , and may not be well represented by a single mfp .
next , we investigate the impact of an energy - dependent @xmath16 , and demonstrate how to perform detailed modeling using the simple " approach presented in this paper .
specifically , we calculate the temperature profile inside a silicon film , using a first principles - calculated phonon dispersion and a semi - empirical scattering model calibrated to experimental data . from the full phonon dispersion , shown in fig .
[ fig5](a ) ( details in @xcite ) , we obtain @xmath48 and @xmath116 . including umklapp , defect and boundary scatterings , we extract @xmath16 , presented in fig .
[ fig5](b ) ( details in @xcite ) . with this model
, we can reproduce the lattice thermal conductivity of bulk silicon within 15% error from @xmath117 to @xmath95 . versus energy at room temperature .
@xmath18 includes boundary , defect and umklapp scattering , and was calibrated to experimental data .
dashed horizontal line corresponds to the average bulk @xmath118.,width=226 ] fig .
[ fig1 ] shows the normalized temperature profile versus normalized position of si films for @xmath4 .
our solutions ( lines ) are compared to results of the phonon bte ( symbols ) @xcite .
given that small differences would be expected since we do not use the exact same dispersion and mfp distribution as in ref .
@xcite , the agreement is very good . if the average bulk @xmath18 is used ( @xmath118 ) , instead of the energy - dependent @xmath16 , significant differences in the temperature profiles are observed .
for example , with @xmath16 the temperature jumps are @xmath119 ( @xmath120 ) , @xmath121 ( @xmath122 ) and @xmath123 ( @xmath124 ) , however with @xmath125 we find @xmath126 , @xmath127 and @xmath128 , respectively .
thus , using @xmath125 overestimates the phonon ballisticity , because it fails to capture the fact that a large fraction of the phonons have a mfp less than the average ( see fig . [ fig5](b ) ) . how does the temperature profile change when using an energy - dependent @xmath16 ? according to eq .
( [ def : heat_den ] ) and the definition of heat capacity @xmath129 , the energy - averaged temperature at any point @xmath7 is given by : @xmath130 where @xmath131 \frac{\partial f_{\rm be}(\epsilon)}{\partial t}. \label{cv_energy}\end{aligned}\ ] ] note that @xmath132 , and that the term in brackets is simply the phonon density of states @xcite .
from the temperature profile given by eq .
( [ temp_profile ] ) , we only need to compute the energy - averaged temperatures @xmath133 and @xmath134 with eq .
( [ temp_total ] ) .
a straight line connecting both points corresponds to the correct energy - averaged temperature profile .
@xmath135 and @xmath136 are both related to the temperature jump @xmath137 at the contacts ( eq .
( [ temp_jump ] ) ) and thus @xmath16 . hence
, if @xmath48 , @xmath116 and @xmath16 are known ( using full phonon dispersions and energy - dependent scattering rates ) , performing detailed modeling within the simple approach described in this paper boils down to evaluating one integral in energy to extract the energy - averaged @xmath138 or @xmath46 . from this
the temperature profile corresponds to a straight line joining @xmath81 and @xmath82 . note that if @xmath18 is a constant , then @xmath138 does not depend on energy , evaluating eq .
( [ temp_total ] ) is trivial , and @xmath48 and @xmath116 do not need to be specified .
in summary , using the mckelvey - shockley flux method , we have shown that a simple phonon boltzmann equation can be written . the solutions to this equation agree well with the results of the full phonon bte , including temperature jumps ( @xmath34 ) at the boundaries with ideal contacts .
examples with a simple phonon dispersion and energy - independent phonon mean - free - path and with a full band phonon dispersion and energy - dependent mean - free - path distribution were both considered . for the simple case ,
analytical solutions for @xmath75 and @xmath36 , that describe phonon transport from the ballistic to diffusive limits were derived . for the complicated case
, the results of full numerical solutions of the phonon boltzmann equation can be reproduced with a fraction of the computational burden .
in addition to faster solutions , the method introduced here also provides new insights into quasi - ballistic phonon transport .
for example , the nature of @xmath34 was discussed , and we showed that it is simply related to the phonon transmission across the film .
in addition , we showed that @xmath34 can be described in terms of an ideal contact resistance equal to one - half of the ballistic thermal resistance at each contact , which is analogous to the so - called quantum contact resistance for electron transport @xcite .
we also showed that for a constant mean - free - path , the thermal conductivity of a small structure ( @xmath139 ) is the same as the bulk thermal conductivity .
the reduction in heat flux occurs because the temperature difference across the film is less than the difference in temperatures of the two contacts and not , as commonly modeled , because of reduced thermal conductivity .
finally , we showed that our simple phonon boltzmann equation can be rewritten exactly as fourier s law and the heat equation .
when solved with correct boundary conditions , we showed that fourier s law and the heat equation capture ballistic effects and are thus applicable on all length scales .
this work addressed one - dimensional transport where the approach and effects can be most clearly discussed . if extensions to higher dimensions are similarly accurate , this approach may prove useful in extending finite - element heat transfer tools to capture ballistic effects and analyzing realistic structures and experiments probing short length and time scales , such as time / frequency - domain thermoreflectance . for the latter ,
a time - dependent mckelvey - shockley flux method of the type previously used to described electron transport @xcite will be needed .
this work was supported in part by darpa meso ( grant n66001 - 11 - 1 - 4107 ) and through the ncn - needs program , which is funded by the national science foundation , contract 1227020-eec , and the semiconductor research corporation . j.m .
acknowledges financial support from nserc of canada .
in this appendix we show how our derived expression for heat current ( eq . ( [ fnet2 ] ) ) , in the ballistic phonon limit , reduces to the known result in the casimir limit .
starting from eq .
( [ fnet2 ] ) , we assume purely ballistic transport , meaning @xmath140 and @xmath56 , which gives @xmath141 \,{\rm d}\epsilon , \label{app : iq_ball}\end{aligned}\ ] ] where we have integrated over energy to obtain the total heat current ( using eq .
( [ def : heat_curr ] ) ) . with a linear phonon dispersion , as is commonly assumed in the casimir limit , the distribution of modes is @xcite : @xmath142 where @xmath143 is the group velocity and the factor of three is for the three acoustic branches .
inserting eq .
( [ app : dom_ph ] ) and the expression for @xmath49 into eq .
( [ app : iq_ball ] ) , we find @xmath144 \,{\rm d}\epsilon , \label{app : iq_ball2}\end{aligned}\ ] ] where @xmath145 is boltzmann s constant . defining a new variable @xmath146 , eq .
( [ app : iq_ball2 ] ) becomes : @xmath147 one can show that the integral is equal to @xmath148 , which gives the known result for heat current in the casimir limit @xcite : @xmath149 where @xmath150 is the stefan - boltzmann constant for phonons .
note the above expression is valid at temperatures much less than the debye temperature of the material .
the expression for bulk thermal conductivity derived from the approach presented in this work is ( see eq . ( [ kappa_bulk ] ) ) : @xmath151 where @xmath64 is the heat capacity , @xmath18 the mean - free - path _ for backscattering _ and @xmath27 is the average @xmath7-projected velocity of the forward moving carriers .
the commonly encountered relation for thermal conductivity is @xmath152 where @xmath153 is the mean - free - path and @xmath143 the group velocity .
we can show that both eq .
( [ app : kappa_bulk ] ) and eq .
( [ app : trad_kappa_bulk ] ) are identical , in the case of an isotropic phonon dispersion ( note our approach applies in the case of any full phonon dispersion ) . in @xcite
, it is shown that @xmath154 and @xmath155 . inserting these relations into eq .
( [ app : kappa_bulk ] ) , one directly finds eq .
( [ app : trad_kappa_bulk ] ) .
in this note we demonstrate how to derive eqns .
( [ bc_temp1]-[bc_temp2 ] ) .
our starting point is fourier s law , which we showed is applicable on all length scales ( eq .
( [ fourier2 ] ) ) : @xmath156 where @xmath157 is simply the net heat current @xmath36 . using the expression for heat density ( eq .
( [ def : heat_den ] ) ) , we have @xmath158 by adding eqns .
( [ app : net_heat_curr ] ) and ( [ app : heat_den ] ) and evaluating @xmath7 at the left boundary ( @xmath20 ) , we obtain @xmath159 by subtracting eqns .
( [ app : net_heat_curr ] ) from ( [ app : heat_den ] ) and evaluating @xmath7 at the right boundary ( @xmath21 ) , we obtain @xmath160 since we assume the applied temperature difference across the contacts ( @xmath161 ) is small , the heat currents can be expanded as @xmath162 , where @xmath163 is the equilibrium heat current associated with a reference background temperature ( chosen as @xmath33 in this case ) and @xmath164 is a correction due to @xmath54 . rewriting eqns .
( [ app : left_bc_temp ] ) and ( [ app : right_bc_temp ] ) in terms of the injected heat fluxes due to @xmath54 and using the definition for heat capacity @xmath64 ( @xmath165 ) , we find : @xmath166 where @xmath104 .
the above equations relate the injected heat currents at the boundaries to the temperature ( and its gradient ) at the boundaries , and represent the correct physical boundary conditions for the heat equation and fourier s law .
note that eqns .
( [ app : left_bc_temp2]-[app : right_bc_temp2 ] ) are applicable even when the contacts are not in equilibrium .
if the contacts are in equilibrium , then @xmath167 and @xmath168 can be obtained by expanding eqns .
( [ fleft]-[fright ] ) : @xmath169 \label{app : left_inj_flux } \\ & = i_{q,{\rm eq}}^+ + \delta i_{q,0}^+ , \nonumber \\
i_{q , l}^- & = \epsilon\,\frac{m}{h}\,f_{\rm be}(t_r ) \label{app : right_inj_flux } \\ & = i_{q,{\rm eq}}^- .
\nonumber\end{aligned}\ ] ] by choosing @xmath33 as the reference temperature , we have @xmath170 .
00 d. g. cahill , p. v. braun , g. chen , d. r. clarke , s. fan , k. e. goodson , p. keblinski , w. p. king , g. d. mahan , a. majumdar , h. j. maris , s. r. phillpot , e. pop and l. shi , appl . phys . rev . * 1 * , 011305 ( 2014 ) .
r. a. escobar and c. h. amon , j. heat transfer * 130 * , 092402 ( 2008 ) .
a. majumdar , j. heat transfer * 115 * , 7 ( 1993 ) .
g. chen , j. heat transfer * 124 * , 320 ( 2002 ) .
s. mazumder and a. majumdar , j. heat transfer * 123 * , 749 ( 2001 ) .
d. lacroix , k. joulain and d. lemonnier , phys .
b * 72 * , 064305 ( 2005 ) . c. j. gomes , m. madrid , j. v. goicochea and c. h. amon , j. heat transfer * 128 * , 1114 ( 2006 ) .
j. hu , x. ruan and y. p. chen , nano lett .
* 9 * , 2730 ( 2009 ) .
s. e. sullivan , k .- h .
lin , s. avdoshenko and a. strachan , appl .
. lett . * 103 * , 243107 ( 2013 ) .
g. d. mahan and f. claro , phys .
b * 38 * , 1963 ( 1988 ) .
f. claro and g. d. mahan , j. appl . phys .
* 66 * , 4213 ( 1989 ) .
a. a. joshi and a. majumdar , j. appl . phys . *
74 * , 31 ( 1993 ) . c. hua and a. j. minnich , phys
b * 89 * , 094302 ( 2014 ) .
b. vermeersch , j. carrete , n. mingo and a. shakouri , arxiv:1406.7341 ( 2014 ) .
b. vermeersch , a. m. s. mohammed , g. pernot , y. r. koh and a. shakouri , arxiv:1406.7342 ( 2014 ) .
a. s. henry and g. chen , j. comput .
. nanosci . * 5 * , 141 ( 2008 ) .
d. p. sellan , j. e. turney , a. j. h. mcgaughey and c. h. amon , j. appl . phys . * 108 * , 113524 ( 2010 ) . m .- h .
bae , z. li , z. aksamija , p. n. martin , f. xiong , z .- y .
ong , i. knezevic and e. pop , nature comm .
* 4 * , 1734 ( 2013 ) .
y. k. koh and d. g. cahill , phys .
b * 76 * , 075207 ( 2007 ) .
k. t. regner , d. p. sellan , z. su , c. h. amon , a. j. h. mcgaughey and j. a. malen , nature comm .
* 4 * , 1640 ( 2013 ) .
a. j. minnich , phys .
lett . * 109 * , 205901 ( 2012 ) .
f. yang and c. dames , phys .
b * 87 * , 035437 ( 2013 ) .
j. p. mckelvey , r. l. longini and t. p. brody , phys .
rev . * 123 * , 51 ( 1961 ) .
w. shockley , phys .
rev . * 125 * , 1570 ( 1962 ) .
h . rhew and m. s. lundstrom , j. appl .
phys . * 92 * , 5196 ( 2002 ) . c. jeong , r. kim , m. luisier , s. datta and m. lundstrom , j. appl
. phys . * 107 * , 023707 ( 2010 ) .
j. maassen , c. jeong , a. baraskar , m. rodwell and m. lundstrom , appl .
. lett . * 102 * , 111605 ( 2013 ) .
j. maassen and m. lundstrom , appl .
. lett . * 102 * , 093103 ( 2013 ) . c. jeong , s. datta and m. lundstrom , j. appl .
* 109 * , 073718 ( 2011 ) .
r. kim , s. datta and m. lundstrom , j. appl .
* 105 * , 034506 ( 2009 ) .
m. j. mclennan , y. lee and s. datta , phys .
b * 43 * , 13846 ( 1991 ) .
g. chen , j. nanopart .
res . * 2 * , 199 ( 2000 ) . the si phonon dispersion was computed within the harmonic approximation , with the force constants obtained from dfpt as implemented in vasp @xcite .
a @xmath171@xmath172@xmath171@xmath172@xmath171 cubic supercell containing si 216 atoms was used ( lattice constant of @xmath173 ) .
a plane - wave energy cutoff of @xmath174 , a @xmath171@xmath172@xmath171@xmath172@xmath171 @xmath175-mesh ( monkhorst - pack scheme ) were tested to provide good accuracy . the paw method and gga - pbe
were adopted for the core and exchange - correlation potentials , respectively .
the phonon dispersion was obtained , from the force constants , using the phonopy simulation package @xcite .
the mean - free - path for backscattering is calculated using @xcite @xmath176 \,\tau(\epsilon)$ ] , where @xmath177 is the scattering time and the term in square brackets depends only on phonon dispersion .
the average is defined as @xmath178 .
boundary , defect and umklapp scattering are included using phenomenological expressions @xcite : @xmath179 , @xmath180 and @xmath181 .
using @xmath182 , @xmath183 , @xmath184 , @xmath185 and @xmath186 , we reproduce the measured lattice thermal conductivity of bulk silicon ( a. v. inyushkin , a. n. taldenkov , a. m. gibin , a. v. gusev and h .- j .
pohl , phys .
( c ) * 11 * , 2995 ( 2004 ) ) within 15% error from 5 k to 300 k. g. nilsson and g. nelin , phys .
b * 6 * , 3777 ( 1972 ) .
s. datta , _ electronic transport in mesoscopic systems _ , cambridge university press ( 1997 ) .
m. a. alam , s .-
tanaka and m. s. lundstrom , solid - state electron . * 38 * , 177 ( 1995 ) .
g. kresse and j. furthmuller , phys .
b * 54 * , 11169 ( 1996 ) ; g. kresse and j. furthmuller , comput .
sci . * 6 * , 15 ( 1996 ) .
a. togo , f. oba and i. tanaka , phys .
b * 78 * , 134106 ( 2008 ) . | it is generally understood that fourier s law does not describe ballistic phonon transport , which is important when the length of a material is similar to the phonon mean - free - path . using an approach adapted from electron transport ,
we demonstrate that fourier s law and the heat equation _ do _ capture ballistic effects , including temperature jumps at ideal contacts , and are thus applicable on all length scales .
local thermal equilibrium is not assumed , because allowing the phonon distribution to be out - of - equilibrium is important for ballistic and quasi - ballistic transport .
the key to including the non - equilibrium nature of the phonon population is to apply the proper boundary conditions to the heat equation .
simple analytical solutions are derived , showing that _
i ) _ the magnitude of the temperature jumps is simply related to the material properties and _ ii ) _ the observation of reduced apparent thermal conductivity physically stems from a reduction in the temperature gradient and not from a reduction in actual thermal conductivity .
we demonstrate how our approach , equivalent to fourier s law , easily reproduces results of the boltzmann transport equation , in all transport regimes , even when using a full phonon dispersion and mean - free - path distribution . |
for many integrable by quadratures dynamical systems in holonomic and nonholonomic mechanics the generic level set of integrals of motion can be related to the jacobian of some hyperelliptic curve @xcite .
it allows us to study symmetries of integrals of motion using group operations in this jacobian [ add - jac ] d+ d=dd = d , where @xmath0 and @xmath1 are reduced divisors , @xmath2 and @xmath3 $ ] denote addition and scalar multiplication by an integer , respectively . in 1826 , abel made a great advance : he discovered an addition theorem of sweeping generality @xcite .
the main legacy of the abel theorem is the string of developments leading to our understanding of the picard variety , functor of points and divisor class group of an algebraic variety .
the history of these developments is rich and fascinating , and it is a significant part of the history of algebraic geometry , see article @xcite , which gives many precise references to the original sources and to the secondary literature .
in @xcite cantor proposed an algorithm for performing computations in jacobian groups of hyperelliptic curves which consists of two stages : composition stage which generally outputs an unreduced divisor , and reduction stage which transforms the unreduced divisor into the unique reduced divisor .
now we have a lot of algorithms and their professional computer implementations for the divisor arithmetics on low , mid and high - genus hyperelliptic curves , for generic divisor doubling , tripling , halving , for degenerate divisors operations and other special cases , in particular see @xcite . in hamiltonian mechanics
authors usually consider only full degree divisors @xmath4 and weight one divisor @xmath5 in ( [ add - jac ] ) when reduction coincides with inversion .
this partial group operation and the corresponding one - point auto bcklund transformations ( bts ) have been studied from different points of view in many publications , in particular see @xcite and references therein .
in fact , this construction of auto bts was specially developed for one - parametric discretization of continuous hamiltonian flows .
our main our is to explicitly show two examples of auto bts associated with the generic divisor doubling at @xmath6 in ( [ add - jac ] ) .
such auto bts represent hidden symmetries of the level manifold which yield new canonical variables on original phase space and , therefore , we can use these new variables for construction of new integrable systems , i.e. for construction of hetero bcklund transformations @xcite .
in this section we give some brief background on hyperelliptic curves and divisor arithmetics @xcite .
let @xmath7 be a positive integer , hyperelliptic curve @xmath8 of genus @xmath7 is defined by equation @xmath9 where @xmath10 is a monic polynomial of degree @xmath11 with distinct roots , @xmath12 is a polynomial with deg @xmath13 .
a rational point on curve @xmath8 is denoted @xmath14 , and @xmath15 is a point at infinity .
the inverse of @xmath16 is point @xmath17 . if @xmath18 then @xmath19 is so - called ramification point .
divisor @xmath20 , @xmath21 is a formal sum of points on the curve , and degree of divisor @xmath22 is the sum @xmath23 of the multiplicities of points in support of the divisor . set @xmath24 of the divisors form an abelian group under the formal addition rule @xmath25 set @xmath26 of the degree zero divisors is a subgroup of @xmath24 .
two divisors are equivalent @xmath27 if their difference is equal to the divisor of zeroes and poles of rational function .
i.e. a principal divisor @xmath28
. quotient group of @xmath26 by the group of principal divisors @xmath29 is called the divisor class group of @xmath8 or picard group of @xmath8 .
the divisor class group , where the elements are equivalence classes of degree zero divisors on @xmath8 , is isomorphic to the jacobian of @xmath8 .
semi - reduced divisor is a divisor of the form @xmath30 where @xmath31 , @xmath32 for @xmath33 , no @xmath34 satisfying @xmath35 appears more than once . for each divisor
@xmath36 there is a semi - reduced divisor @xmath22 so that @xmath37 .
however semi - reduced divisors are not unique in their equivalence class . in @xcite
we used this fact in order to get auto bts associated with an equivalence relation .
a semi - reduced divisor @xmath22 is called reduced if @xmath38 , i.e. if the sum of multiplicities is no more that genus of curve @xmath8 .
this is a consequence of riemann - roch theorem for hyperelliptic curves that for each divisor @xmath36 there is a unique reduced divisor @xmath22 so that @xmath37 .
below we assume that equations ( [ add - jac ] ) involve only reduced divisors unless stated otherwise .
the reduced degree or weight of reduced divisor @xmath22 is defined as @xmath39 .
the jacobian of hyperelliptic curve is a group of degree zero divisors modulo principal divisors , and the group operation is formal addition modulo the equivalence relations . to compute the additive group law ,
cantor gave a concrete algorithm which is applicable to a hyperelliptic curve of any genus @xcite .
the composition part of this algorithm computes the semi - reduced divisor @xmath40 that is equivalent to @xmath1 .
the reduction part computes the reduced divisor @xmath1 . in the examples considered below reduction coincides with inversion @xmath41 . in cantor
s algorithm we use of the polynomial representation of group elements @xmath42 proposed by mumford for the reduced and semi - reduced divisors @xmath43 here monic polynomial @xmath44 may have multiple roots and @xmath45 is tangent to the curve according to multiplicity roots . in classical mechanics jacobi polynomials @xmath44 and @xmath45 form so - called lax matrix @xcite @xmath46 and group operations are similar transformations @xmath47 where matrix @xmath48 depends on a group operation .
some matrices @xmath48 associated with weight one divisor @xmath5 in ( [ add - jac ] ) are discussed in @xcite . instead of the lax matrices we prefer to use divisors @xmath42 , computer implementations of cantor s type algorithms and well - known explicit formulae for arithmetic on low - genus hyperelliptic curves from textbooks on algebraic geometry and handbooks on hyperelliptic cryptography .
in cryptography we have to distinguish odd and even characteristic fields @xcite , but in classical mechanics we can always apply the substitution @xmath49 and study the genus two hyperelliptic curve which is defined by equation [ eq - c ] c : y^2=f(x),f(x)= a_5x^5 + a_4x^4 + a_3x^3 + a_2x^2+a_1x+a_0 . let us consider the addition of full degree reduced divisor @xmath50 with other reduced divisor @xmath5 in the following cases [ cases ] llll 1.&d+d=d,&w(d)=2,&(d)=\ { p_1,p_2}\{p _ } , + + 2.&d = d , & d=d,&(d)=\ { p_1,p_2}\{p _ } , + + 3.&d+d=d , & w(d)=1 , & ( d)=\ { p_1}\{p_}. the result is full degree reduced divisor @xmath1 with supp@xmath51 and @xmath52 . in the second case operation [ doub - gen ]
[ 2]d = d+d = d is called doubling of divisor @xmath22 , i.e. scalar multiplication on integer @xmath6 .
its inverse is called halving of @xmath1 and for a given @xmath1 equation ( [ doub - gen ] ) has @xmath53 solutions , any two of which differ by a 2-torsion divisor @xcite , for efficient implementation see @xcite and references within .
below we do not consider halving , tripling and other operations because the resulting equations for the corresponding auto bcklund transformations are quite bulky and unreadable .
the third case in ( [ cases ] ) corresponds to one - point auto bcklund transformations from @xcite .
in fact , there are two main methods for deriving group law in the jacobian of hyperelliptic curve : the algebraic method based on harley s implementation @xcite of cantor s algorithm @xcite , and the geometric method using interpolation of points @xcite , which is based on clebsch s geometric interpretation of the abel results @xcite .
in the second case we add two divisors @xmath22 and @xmath5 by finding a function which divisor contains the support of both @xmath22 and @xmath5 , and then the sum is equivalent to the negative of the compliment of that support . following abel s main idea from @xcite
such a function @xmath54 can be obtained by interpolating the points in the support of the two divisors .
compliment of the support @xmath22 and @xmath5 in the support of @xmath55 consists of other points of intersection of @xmath56 with the curve @xmath8 .
let us consider intersection of @xmath8 with the second plane curve defined by equation [ eq - p ] y - p(x)=0,p(x)=b_3x^3+b_2x^2+b_1x+b_0 .
substituting @xmath56 into the equation ( [ eq - c ] ) , we obtain the so - called abel polynomial @xcite @xmath57 which has no multiple roots in the first and third cases and has double roots in the second case : @xmath58 the cases 1 and 2 are presented in figure 1 .
equating coefficients of @xmath59 in the first case gives & & x_1+x_2= -x_1-x_2-x_1-x_2 + , [ g2-x12 ] + + & & x_1x_2= -(x_1+x_2+x_1+x_2)(x_1+x_2 ) -x_1(x_2+x_1+x_2 ) -x_2(x_1+x_2)-x_1x_2 . in the second case
we have to put @xmath60 in these equations , whereas in the third case we have [ g2-x3 ] l x_1+x_2=-x_1-x_2-x_1 + , + + x_1x_2=-(x_1+x_2+x_1)(x_1+x_2)-x_1x_2-x_1x_1-x_2x_1 .
+ four coefficients @xmath61 and @xmath62 of polynomial @xmath54 ( [ eq - p ] ) are calculated by solving four algebraic equations : [ eq - b ] cl 1.&y_1,2=p(x_1,2),y_1,2=p(x_1,2 ) ; + + 2 . &
y_1,2=p(x_1,2),.|_x = x_1,2=.|_x = x_1,2 f(x_1,2 ) , + + 3.&y_1,2=p(x_1,2),y_1=p(x_1),b_3=0 , + where @xmath63 is derivative of @xmath10 ( [ eq - c ] ) by @xmath64 .
substituting coefficients @xmath65 into ( [ g2-x12]-[g2-x3 ] ) one gets abscissas @xmath66 , whereas the corresponding ordinates @xmath67 are equal to [ g2-y ] y_3,4=-p(x_1,2 ) , where polynomial @xmath54 is given by [ ab - p ] lcl 1.p(x ) & = & + + + & + & + , + + 2.p(x)&= & + + + & + & + , + + 3.p(x)&= & + + .
+ in ( [ g2-y ] ) we also made a reduction , which coincides with inversion in our cases .
if @xmath22 is a full degree divisor on the hyperelliptic genus three , then at the composition stage we compute coefficients of the quintic polynomial @xmath68 using six equations @xmath69 and @xmath70
. then we compute abscissas @xmath71 of the remaining four points of intersection using abel polynomial @xmath72 and find the corresponding ordinates @xmath73 . as a result , we obtain unreduced divisor @xmath74 . at the reduction stage
we are looking for the cubic polynomial @xmath75 which interpolates the points in support of @xmath76 and intersects @xmath8 in three more places to form @xmath77 . as a result , we obtain reduced divisor @xmath1 which is equivalent to @xmath76 @xcite .
to perform such computations in generic cases we can take any professional implementation of cantor s algorithm or implementations of more effective improvements and extensions to cantor s algorithm for mid- or high - genus curves with general divisor doubling , tripling etc @xcite .
let us consider rotation of a rigid body around a fixed point in a homogeneous gravity field in the lagrange case . in terms of the euler angles
@xmath78 , and momenta @xmath79 the hamilton function looks like @xmath80 it is invariant with respect to canonical transformations of valence @xmath81 [ lag - can1 ] = c+a,= c+b,_=p_,_=p_. for canonical transformation @xmath82 of valence @xmath81 the jacobi matrix of transformation @xmath83 is a generalized symplectic matrix of valence @xmath81 @xmath84 see details in @xcite .
canonical transformation ( [ lag - can1 ] ) is a symmetry and , according to noether s theorem , this symmetry is related to integrals of motion @xmath85 we can also use this canonical transformation for construction of integrable discretization @xmath86 for dynamics of spin @xmath87 and precession @xmath88 . in order to describe similar canonical transformation , symmetry and discretization for nutation we can use the standard arithmetic on elliptic curves .
indeed , in @xcite lagrange noted that nutation @xmath89 is the weierstrass elliptic function @xmath90 because hamilton - jacobi equation @xmath91 [ lagr - curve ] + u = e , v= p _ , defines the elliptic curve . for elliptic curve
@xmath92 the jacobian of @xmath8 is isomorphic to the curve itself .
so , let us consider points @xmath93 , @xmath94 and @xmath95 on @xmath8 so that @xmath96p = p''\,,\qquad \mbox{case 3:}\quad [ 3]p = p''\,.\ ] ] according to @xcite coordinates of @xmath97 are equal to [ ell - xy ] lll 1.&x=-x - x-+()^2,&y=-y+y , + + 2.&x=-2x-+()^2,&y=-y- , + + 3.&x=-3x--,&y=b_2x^2+b_1x+b_0 , where @xmath98 are coefficients of the polynomial @xmath99 in order to get auto bcklund transformations @xmath100 for the lagrange top we have to combine these equations with equation ( [ lagr - curve ] ) : [ pq - lag ] 1.&=-u-+h+()^2,&=-v+ , + + 2.&=-2u+h-,&=-v- , where we put @xmath101 and @xmath102 following @xcite . in the third case we present polynomial @xmath103 only : @xmath104 the corresponding variables @xmath105 and @xmath106 are obtained from ( [ ell - xy ] ) equations ( [ pq - lag ] ) determine canonical transformations @xmath107 of valencies one , two and three , respectively .
these canonical transformations preserve the form of hamilton - jacobi equation @xmath91 , i.e. they are auto bcklund transformations for the lagrange top .
the proof is a straightforward calculation in which we have taken into account that @xmath108 is a function on @xmath109 , @xmath110 and parameter @xmath111 .
first transformation is the so - called one - point bt similar to bt from @xcite , second and third auto bts of valencies two and three are new . for inverse transformations associated with division , we have valencies @xmath112 , @xmath113 etc . in the framework of the elliptic curve cryptography various combinations of addition , multiplication and division @xmath114p''_{k-1}+p'_k\,,\qquad p''_0=p\,,\ ] ] determine the so - called cryptographic protocols @xcite . in classical mechanics we can identify these protocols with various integrable discretizations of continuous flows , for instance with discretizations for dynamics of nutation angle @xmath109 in the lagrange case of rigid body motion .
let us take hnon - heiles system with hamiltonians @xmath115 separable in parabolic coordinates on the plane @xmath116 standard momenta associated with parabolic coordinates @xmath117 are equal to @xmath118 to describe evolution of @xmath119 with respect to @xmath120 we use the canonical poisson bracket [ poi1 ] \{q_i , p_j}=_ij,\{q_1,q_2}=\{p_1,p_2}=0,\{u_i , p_u_j}=_ij,\{u_1,u_2}=\{p_u_1,p_u_2}=0 . and expressions for @xmath121 [ hh - ham - u ] h_1=-a(u_1+u_2)(u_1 ^ 2+u_2 ^ 2),h_2=+au_1u_2(u_1 ^ 2+u_1u_2+u_2 ^ 2 ) to obtain @xmath122 and @xmath123 using hamilton - jacobi equations
@xmath124 we can prove that these variables satisfy to the following separated relations @xmath125 expressions ( [ hh - eq1]-[hh - eq2 ] ) and ( [ hh - sep ] ) yield standard abel quadratures [ hh - ab - q1 ] + = 2dt_2,+=2dt_1 , on hyperelliptic curve @xmath126 of genus two defined by equation [ hh - eq - c ] c : y^2=f(x),f(x)=x(ax^4+_1x+_2 ) .
suppose that transformation of variables [ st - b - hh ] ( q_1,q_2,p_1,p_2)(_1,_2,_1,_2 ) preserves hamilton equations ( [ hh - eq1]-[hh - eq2 ] ) and the form of hamiltonians ( [ hh - ham - u ] ) .
it means that new parabolic coordinates @xmath127 satisfy to the same equations [ hh - ab - q2 ] + = 2dt_2,+=2dt_1 . subtracting ( [ hh - ab - q2 ] ) from ( [ hh - ab - q1 ] ) one gets abel differential equations @xmath128 where @xmath129 and @xmath130 form a base of holomorphic differentials on hyperelliptic curve @xmath8 of genus @xmath131 @xmath132
these abel equations are related to the picard group which is isomorphic to jacobian of @xmath8 @xcite and we can consider canonical transformation ( [ st - b - hh ] ) of variables in the phase space as some group operation in the jacobian . let us suppose that generic points @xmath133 and @xmath134 form divisor @xmath22 in ( [ cases ] ) , whereas points @xmath135 belong to support of resulting divisor @xmath1 .
canonical transformations ( [ st - b - hh ] ) associated with arithmetic operations ( [ cases ] ) are completely defined by coefficients @xmath98 of polynomial @xmath136 ( [ ab - p ] ) . in the first case ,
these coefficients @xmath98 are defined by ab_0&=&q_1x_1x_2(x_1-x_2)(p_1(q_1 ^ 2 + 4q_2 ^ 2 - 2q_2(x_1+x_2)+x_1x_2)+ q_1p_2(x_1+x_2 + 2q_2 ) ) + + & -&2q_1 ^ 2x_2(q_1 ^ 2 + 2q_2x_2-x_2 ^ 2)y_1 + 2q_1 ^ 2x_1(q_1
^ 2 + 2q_2x_1-x_2 ^ 2)y_2 ; + + ab_1&=&(x_2-x_1 ) ( q_1p_1((q_1 ^ 2 + 4 q_2 ^ 2)(x_1+x_2)-2q_2(x_1 ^ 2+x_1x_2 + x_2 ^ 2 ) ) . + & & -.p_2(2 q_1 ^ 2 q_2(x_1+x_2)-q_1 ^ 2(x_1 ^ 2+x_1x_2 + x_2 ^ 2)+x_1 ^ 2 x_2 ^ 2 ) ) + & + & 2 ( q_1 ^ 2 - 2 q_2 x_2 ) ( q_1 ^ 2 + 2 q_2x_2-x_2 ^ 2 ) y_1 - 2 ( q_1 ^ 2 - 2 q_2 x_1 ) ( q_1 ^ 2 + 2 q_2x_1-x_1 ^ 2 ) y_2 , + + ab_2&= & ( x_1-x_2)(q_1p_1(q_1 ^ 2 + 4 q_2 ^ 2-x_1 ^ 2-x_1x_2-x_2 ^ 2 ) - ( 2q_1 ^ 2q_2+(x_1+x_2)x_1x_2 ) p_2 ) + + & + & 2(2 q_2+x_2 ) ( q_1 ^ 2 + 2 q_2 x_2-x_2 ^ 2)y_1 - 2 ( 2 q_2+x_1 ) ( q_1 ^ 2 + 2 q_2 x_1-x_1 ^ 2 ) y_2 , + + ab_3&=&(x_2-x_1)(q_1p_1(2q_2-x_1-x_2)-(q_1 ^ 2+x_1x_2)p_2)-2(q_1 ^ 2 + 2q_2x_2-x_2 ^ 2)y_1 + + & -&2(q_1 ^ 2 + 2q_2x_1-x_1 ^ 2)y_2 , where @xmath137 substituting these coefficients into the following expressions one gets explicit formulae for the new variables [ pq - case1 ] l
_ 2= -q_2 - -+ , + + _ 1 ^ 2 = -q_1 ^ 2 + 2_2(2q_2+x_1+x_2)+2q_2(x_1+x_2)+x_1x_2 - - , + + _ 1= --4_1b_3 - 2_1b_2,_2=-2(_1 ^ 2 + 4_2 ^ 2)b_3 - 4_2b_2 - 2b_1 . in case 2 we have bb_0&= & -8aq_1 ^ 4(2p_1q_1q_2-p2q_1 ^ 2 - 2p_2q_2 ^ 2)+q_1 ^ 2p_1(p_1 ^ 2q_1 + 2p_1p_2q_2 - 2p2 ^ 2q_1 ) , + + bb_1&=&-8aq_1 ^ 2(p_1q_1 ^ 3 + 10p_1q_1q_2 ^ 2 - 4p_2q_1 ^ 2q_2 - 4p_2q_2 ^ 3)-2p_1 ^ 3q_1q_2 + 3p_1 ^ 2q_1 ^ 2p_2 - 4p_1 ^ 2p_2q_2 ^ 2 + + & + & 4p_1p_2 ^ 2q_1q_2 - 2p_2 ^ 3q_1 ^ 2 , + + bb_2&=&-8aq_1(12p_1q_2 ^ 3+p_2q_1 ^ 3 - 2p_2q_1q_2 ^ 2)+p_1 ^ 2(p_1q_1 + 4p_2q_2 ) , + + bb_3&=&8aq_1(p_1q_1 ^ 2 + 6p_1q_2 ^ 2 - 2p_2q_1q_2)-p_2p_1 ^ 2,b = 4q_1(p_1 ^ 2q_1 + 2p_1p_2q_2-p_2 ^ 2q_1)so that _
2&=&- 2q_2 + + + & & ( 64 q_1 ^ 2 ( 12 p_1 q_2 ^ 3+p_2 q_1 ^ 3 - 2 p_2 q_1 q_2
^ 2 ) ( p_1 q_1 ^ 2 + 6 p_1 q_2 ^ 2 - 2 p_2 q_1 q_2 ) a^2+p_1 ^ 4 p_2 ( p_1 q_1 + 4 p_2 q_2 ) . + + & & .-8a q_1 ( 6 p_1 ^ 4 q_1 q_2 ^ 2 - 2 p_1 ^ 3 p_2 q_1 ^ 2 q_2 + 36 p_1 ^ 3 p_2 q_2 ^ 3 + 3 p_1 ^ 2 p_2 ^ 2 q_1 ^ 3 - 14 p_1 ^ 2 p_2 ^ 2 q_1 q_2 ^ 2 + 4 p_1 p_2 ^ 3 q_1 ^ 2 q_2-p_2 ^ 4 q_1 ^ 3 ) ) , + + _ 1 ^ 2&=&-()^2,[pq - case2 ] + + _ 1_1&= & -2b_0 - 2 ^ 2_1(b_2 + 2b_3),_2=-2(_1 ^ 2 + 4_2 ^ 2)b_3 - 4_2b_2 - 2b_1.in case 3 coefficients are equal to [ b - case3 ] l b_0= + b_1= - , b_2= , so that _ 2&=&-q_2-+=-q_2-+ , + + _ 1 ^ 2&=&-q_1 ^ 2 + 2_2(2q_2+x_1)+2q_2x_1+[pq - case3 ] + + & = & -q_1 ^ 2 + 2_2(2q_2+x_1)+2q_2x_1 - , + + _ 2&= & -4_2b_2 - 2b_1=-p_2 - , + + _ 1 _ 1&=&-2b_0 - 2q^2_1b_2= -q_2p_2 - .
these explicit formulae for @xmath138 and @xmath139 can be easily obtained using any modern computer algebra system .
we present these bulky expressions here only so that any reader can verify that these transformations @xmath82 are different , i.e. can not be obtained from each other for special values of parameters , and that these transformations have the following properties .
equations ( [ pq - case1],[pq - case2 ] ) and ( [ pq - case3 ] ) determine canonical transformations ( [ st - b - hh ] ) on @xmath140 of valencies one and two for which original poisson bracket ( [ poi1 ] ) has the following form in new variables @xmath141 respectively .
these canonical transformations preserve the form of integrals of motion ( [ hh-1 ] ) , i.e. they are auto bcklund transformations of the hnon - heiles system . the proof is a straightforward calculation . in the same manner ,
i.e. using a computer implementation of cantor s algorithm , we can prove that auto bt associated with tripling @xmath142d$ ] of divisor has valence three .
summing up , using jacobian arithmetic for hyperelliptic curves , we can identify various cryptographic algorithms and protocols @xcite with various schemes of the discretization of continuous hamiltonian flows in classical mechanics .
the fact most interesting to us is that the corresponding auto bts also yield new canonical variables on the original phase space , which can be used for construction of new integrable systems in the framework of the jacobi method .
for instance , let us suppose that divisor @xmath5 in case 3 consists of ramification point @xmath143 on @xmath8 .
in this case variables @xmath144 are defined by equations ( [ g2-x3 ] ) and ( [ g2-y ] ) for @xmath145 and @xmath146 : @xmath147 where coefficients @xmath98 are given by ( [ b - case3 ] ) . substituting @xmath148 and @xmath149 into the separated relation @xmath150 which defines genus three hyperelliptic curve @xmath151 , and solving the resulting equations with respect to @xmath152 , one gets hamiltonian @xmath153 we can identify this hamiltonian with well - known second integrable hnon - heiles system with quartic additional integral @xmath154 @xcite . in this case divisor
@xmath1 in ( [ cases ] ) belongs simultaneously to jacobians of genus two and genus three curves @xmath8 and @xmath151 .
it is an open question how to construct such second curve @xmath151 .
we prove that addition and scalar multiplication of divisors yield two different types of auto bcklund transformations of hamilton - jacobi equation . in hyperelliptic curve cryptography addition and scalar multiplication
are associated with coding / decoding and , therefore , we can use well - studied cryptographic algorithms and protocols to construct various discretizations of hamiltonian flows .
the main purely technical problem is how to find a compact , readable representation of bcklund transformation in terms of original variables .
it is well - known that using variables of separation @xmath155 we can construct infinite dimensional pencil of poisson bivectors @xmath156 compatible with canonical poisson bivector @xmath157 .
auto bts map this pencil to another infinite dimensional pencil of poisson bivectors compatible with @xmath157 .
these new poisson bivectors produce new families of bi - hamiltonian systems on the original phase space .
some of these integrable systems prove to be interesting in various applications .
in @xcite we also consider auto bts associated with the equivalence relations between unreduced divisors .
one more type of auto bt may be associated with cryptographic attacks .
some of these attacks are quite efficient if the underlying hyperelliptic curve has special form , see @xcite . consequently , using well known cryptographic attacks we could try to determine auto bts and discretizations of the corresponding dynamical system at least numerically .
it may be interesting to perform some computer experiments for dynamical systems associated with the hyperelliptic curve of a special form .
galbraith s. g. , harrison m. , mireles morales d.j .
, , in algorithmic number theory 8th international symposium ( ants - viii ) , a.j .
van der poorten and a. stein ( eds . ) , lecture notes in computer science , 5011 , springer , ( 2008 ) , pp.342 - 356 | the hyperelliptic curve cryptography is based on the arithmetic in the jacobian of a curve . in classical mechanics
well - known cryptographic algorithms and protocols can be very useful for construct auto bcklund transformations , discretization of continuous flows and study of integrable systems with higher order integrals of motion .
we consider application of a standard generic divisor doubling for construction of new auto bcklund transformations for the lagrange top and hnon - heiles system . |
a few of years ago , a novel hadron beam cooling technique capable to deal with the accelerators operating in the range of few tevs , the coherent electron cooling ( cec ) , was proposed @xcite .
currently , a test facility is under construction at brookhaven national laboratory . for the present status of the developments of the machine
, we refer to @xcite .
the cec is the modern realization of the stochastic electron cooling , wherein the electron beam serves as a pick - up and a kicker .
it records the information about the hadron beam via the electron density perturbations resulting from the shielding of the hadrons .
then , these perturbations are amplified in the free electron laser ( fel ) section , and then , in the kicker , every hadron experience the electric field produced by its own amplified perturbation receiving kicks . before the kicker , in the dispersion section , each hadron is emplaced such that these kicks accelerate or decelerate it depending on its energy deviation , thereby reducing the energy spread of the hadron beam . to analyze the performance of the cec ,
all the sections of the device must be studied in detail ; in particular , the shielding of a hadron in an electron beam should be computed with high precision .
+ in this article , we offer a theoretical description of the modulator section of the coherent electron cooling , i.e. , we address the problem of dynamical shielding of a charged particle in an electron beam .
the simplest interpretation of this problem is a screening of a stationary particle in an infinite plasma ( by _ plasma _ we mean a collisionless single - species electron plasma ) , i.e. , the well - known debye screening .
the next step is to consider a moving ion in an infinite plasma .
this problem was studied recently and the density perturbation for the lorentz distribution was expressed as a one - dimensional integral @xcite . the most advanced approach to resolving
this problem is a general method for a shielding in a finite electron beam that takes into account the focusing field and the space - charge effects ; this methodology was proposed last year by the authors of the present paper @xcite and is under development now .
there are also simulations of this effect using the pic ( `` particle - in - cell '' ) method @xcite . in this article , we present a solution of this problem via the fourier and laplace transforms for the 1d , 2d , and 3d infinite plasmas and various equilibrium distributions .
we derived the expressions for the solution in the fourier - laplace space , then inverted them numerically .
for the 1d plasma with the cauchy equilibrium distribution , we found a fully analytical solution in the space - time domain , which gives an opportunity to test the semi - analytical ones involving numerical evaluations of the inverse integral transforms and to perform many - particle computations much faster than with non - exact solutions .
this method can also work with empirical equilibrium distributions . as a fast and robust solution
, it has its own practical value and it will also serve as a testing ground for the pic simulations and the general solution mentioned for a realistic case of a finite beam .
generally , the shielding of a charged particle in a plasma is described by the vlasov - maxwell system of equations @xcite , i.e. , the dynamics of the electron density is governed by the vlasov equation and the electro - magnetic field by the maxwell equations .
we first describe the system in a co - moving frame of reference , then derive a formal solution via the integral transforms , then introduce convenient dimensionless variables , and finally write a solution for a particle moving along a straight line .
we consider the vlasov - maxwell system for the 1d , 2d and 3d plasmas simultaneously , which means that @xmath0 is a one- , two- or three - dimensional vector depending on the dimensionality of the plasma we are considering and by @xmath1 we denote its absolute value , even for the 1d case ; the same conventions are applied for the dimensionless vectors that we will introduce in subsection [ dimless ] . for the electron phase - space density @xmath2 , the hamiltonian @xmath3 , and the electric potential @xmath4 , we have : @xmath5 and the charge density is @xmath6 . for this system
, we consider the test charge problem with an external time - dependent density @xmath7 .
we assume that @xmath8 , where @xmath9 is an equilibrium electron density , and @xmath10 is an unknown perturbation resulting from the interaction with the test charge .
the linearized maxwell - vlasov system looks as follows : @xmath11 the poisson equation ( [ pois ] ) can be solved via the fourier transform : @xmath12 where @xmath13 , @xmath14 , and @xmath15 are the fourier images of the corresponding functions . using this solution , we transform the equation ( [ vlas ] ) to @xcite : @xmath16 where , @xmath17 and @xmath18 are , respectively , the laplace - fourier images of @xmath19 and @xmath7 : @xmath20 @xmath21 we introduce one more notation : @xmath22 denoting the inverse fourier and laplace transforms , respectively , by @xmath23 and @xmath24 , we obtain the following expression : @xmath25 for the details on definitions of the integral transforms and our notations , see appendix [ apena ] .
generally , the expression ( [ general1 ] ) can be complex .
looking back to our initial equations and assuming complex @xmath26 , we note that the equation with @xmath27 corresponds to the equation without an external charge , while the equation with @xmath28 is the one with it , consequently , @xmath29 and it is confirmed by further computations .
hence , the expression ( [ general1 ] ) is real , as it should be . even though the poisson equations and their green s functions differ for the 1d , 2d , and 3d cases , their solutions in the fourier domain ( [ furim ] ) and the expression ( [ general1 ] ) for @xmath30 have the same form . in proceeding further ,
we need to specify the dimension of the problem , the external charge density @xmath7 , and the equilibrium distribution , but first we introduce dimensionless variables .
we define the dimensionless variables , denoting them using the sans - serif font , as follows : @xmath31 where @xmath32 are , respectively , the root - mean - square velocity , the plasma frequency , and the debye radius .
the equilibrium density is normalized via : @xmath33 we introduce the dimensionless equilibrium density @xmath34 by the relation : @xmath35 wherein all the dimensional constants are gathered into @xmath36 and @xmath37 stands for the dimensionality of the space , and can be @xmath38 , @xmath39 or @xmath40 .
we have the following dimensionalities for other quantities : @xmath41=\frac{c^2t^2}{l^dm},\:\:\:\:[n(\vec{x},t)]=[\rho]=l^{-d},\:\:\:\:[{f_d}]=[{v_{\mathrm{rms}}}]^{-d}.\end{aligned}\ ] ] we note that @xmath36 and @xmath42 are not the same for the different equilibrium densities and must be computed via ( [ rms ] ) and ( [ norm ] ) ; for the non - integrable densities , the values have to be chosen voluntarily ; among the densities we consider , only the cauchy one is of that type .
using the dimensionless units , we rewrite formula ( [ general1 ] ) as follows : @xmath43,\end{aligned}\ ] ] where @xmath44 and @xmath45 are the dimensionless analogs of ( [ lf1 ] ) and ( [ lf2 ] ) , respectively , @xmath46 is a dimensionless factor , and @xmath47 , @xmath48 are the inverse laplace and fourier transforms for the dimensionless variables .
we assume that the charge s trajectory is unaffected by the space charge fields and consider the charge moving along a straight line @xmath49 , we have : @xmath50 this assumption is reasonable for a hadron moving in an electron beam , as the hadron s mass is much larger than the electron s . for simplicity , we assume @xmath51 and the final density for the non - unitary charge can be recovered just by multiplying it by @xmath52 . using the dimensionless units introduced , for any number of the spatial dimensions
, we have : @xmath53 finally , we can write the expression for the electron density perturbation resulting from the interaction with the external charge moving along a straight line @xmath54 , valid in 1d , 2d , and 3d spaces : @xmath55.\end{aligned}\ ] ] in the next section , we consider this solution for some particular equilibrium densities @xmath56 ; for each case , we just need to compute @xmath45 and @xmath57 and insert them into ( [ genfin ] ) .
generally , the equilibrium distribution @xmath56 has to be a solution of the unperturbed vlasov equation , i.e. , it has to be a function of the unperturbed hamiltonian , in our dimensionless units it is @xmath58 , thus we consider the following functions : @xmath59 they correspond to the kapchinskij - vladimirskij ( kv ) , water - bag ( wb ) , normal ( or maxwell ) , and cauchy ( or lorentz ) equilibrium distributions , @xmath60 stands for the heaviside step function , @xmath37 is the dimensionality of the space , and @xmath61 is a one- , two- , or three - dimensional vector .
however , for the case of an infinite plasma that we are considering here , any function of velocity is a solution of the unperturbed vlasov equation ; thus , all our formulae can be easily generalized for the equilibrium distributions of the form : @xmath62 corresponding to an anisotropic plasma , where @xmath63 , @xmath64 are dimensionless constants characterizing the plasma s temperatures .
the changes should be applied only to the expression for @xmath45 , i.e. , @xmath65 should be substituted with @xmath66 for @xmath64 , and the whole expression should be divided by @xmath67 . + for the 1d cauchy distribution , the inverse laplace and fourier transforms in ( [ genfin ] )
can be evaluated analytically , while for the other distributions , the numerical techniques should be applied .
we describe the 1d cauchy case in detail and just quote the results for the other distributions starting with the kv and wb , which have different expressions for @xmath45 in spaces of different dimensionalities .
we conclude this section with the cauchy and maxwell distributions that have the same expressions for this quantity in all cases .
although , below we present @xmath42 computed via ( [ rms ] ) , we considered dimensionless equilibrium distributions , @xmath56 , corresponding to @xmath68 with this @xmath42 , in all cases , @xmath69 and @xmath56 has a simpler form . here , we assume @xmath70 , where @xmath71 and @xmath72 are dimensional constants that can be used for fitting the experimental distributions .
computing @xmath73 via ( [ lf2 ] ) , we obtain the following expression : @xmath74 or , using the dimensionless variables : @xmath75 then , we insert the expression for @xmath76 into the formula ( [ genfin ] ) and obtain : @xmath77.\end{aligned}\ ] ] for all distributions we are considering , excepting the 1d cauchy , the inverse integral transforms in the corresponding expressions for @xmath78 have to be inverted numerically , while , for the 1d cauchy , they can be computed analytically giving the following expression : @xmath79 where @xmath80 and @xmath81 and @xmath82 are the exponential integral functions @xcite that can be computed via the series expansions , for details , see appendix [ eiig ] . the whole expression ( [ huge3 ] ) is real even though it contains complex numbers .
[ comp ] , we show the densities obtained via the exact formula ( [ huge3 ] ) and the ones obtained by the discussed in section [ num ] numerical inversion of the integral transforms for @xmath83 and @xmath84 .
we note the perfect agreement of the exact solution and the numerical one .
the solution has several interesting features , i.e. , starting from some time , the left tail of the density has negative values , meaning that there is an accumulation of the charge of the same sign as that of the external perturbation , its maximum is oscillating , and the shape of the peak depends on the charge s velocity , being spiky for small velocities , widening as it increases , and , for large velocities , a discontinuity of the density s shape derivative appears in the right tail .
all these features are equally well captured by the numerical computations and the analytical formula .
we will comment further on the parameters values in subsection [ res ] .
the exact values and the ones obtained numerically via the fft for the 1d cauchy distribution , for the various velocities of the external charge .
solid lines represent the exact values and the fft values are shown by marks of different shapes regarding the times.,title="fig : " ] the exact values and the ones obtained numerically via the fft for the 1d cauchy distribution , for the various velocities of the external charge .
solid lines represent the exact values and the fft values are shown by marks of different shapes regarding the times.,title="fig : " ] the exact values and the ones obtained numerically via the fft for the 1d cauchy distribution , for the various velocities of the external charge .
solid lines represent the exact values and the fft values are shown by marks of different shapes regarding the times.,title="fig : " ] for the 1d kv distribution , we have @xmath85 . using the dimensionless variables ,
we obtain : @xmath86 @xmath87 for the 1d wb , we have : @xmath88 @xmath89 @xmath90 in the 2d case , we have for the kv distribution : @xmath91 @xmath92 @xmath93 and , for the 2d wb : @xmath94 @xmath95 @xmath96 the expressions are slightly bulkier in the 3d case ; we obtain for the 3d kv : @xmath97 where @xmath98 and @xmath99 is a third component of @xmath100 ; for the 3d wb , we obtain : @xmath101 the integral in ( [ 3dwb ] ) has to be computed numerically . for the cauchy distribution
, we obtained the expressions valid in 1d , 2d , and 3d cases : @xmath102 for the maxwell distribution , we also found universal formulas valid in 1d , 2d , and 3d cases : @xmath103,\\ & { v_{\mathrm{rms}}}=\sqrt{\frac{dh_c}{2\beta}},\ , { f_d}=\left(\frac{\beta}{h_c}\right)^{\frac{d}{2}},\,{f_d}^{-1 } { v_{\mathrm{rms}}}^{-d}=\left(2/d\right)^{\frac{d}{2}},\end{aligned}\ ] ] where @xmath104 is the complementary error function @xcite , for its definition and some computational details , see appendix [ erf ] .
the density @xmath105 for the 1d kv for @xmath106 . ]
in this section , we briefly discuss the numerical methods we employed and present our results . to evaluate the expression ( [ genfin ] ) for a particular distribution , we first need to compute @xmath76 via the formulas presented in the previous section ; the expressions are either elementary functions or include special functions or a one - dimensional integral , all these things can be computed straightforwardly .
next step is an evaluation of the inverse fourier and laplace transforms .
it is well - known that the inverse fourier transform can be approximated by the discreet fourier transform and then computed using the fft algorithm , for details , see appendix [ apena ] . in this algorithm ,
the domain of interest of the resulting function is divided into @xmath107 segments .
the inverse laplace transform can be expressed via the inverse fourier transform : @xmath108 which can be evaluated in a way that we described above , @xmath109 is a real constant greater than the real parts of all singularities of @xmath110 . in the upcoming section
, we graphically present our results using the dimensionless units .
we note that the dimensionless values for the different distributions are not always comparable to each other , since the values for @xmath42 can be different ; the corresponding conversion factors should be applied . in this section ,
we discuss the results obtained numerically and shown in fig .
[ 1d ] , [ 1dkv ] , [ 2d ] , and [ 3d ] ; the velocity of the external charge , @xmath111 , is measured in units of @xmath42 corresponding to the electron s equilibrium distribution and the initial position of the charge , @xmath112 , is measured in units of @xmath113 .
the possible space - time ranges differ for the different distributions and are limited by the required precision and the number of points @xmath107 in the fft algorithm . for each plot
, we increased @xmath114 until the values stabilized ; the values used are shown in the legends in each plot .
the most well - behaving case corresponds to the cauchy distribution , the kv and normal distributions require greater values of @xmath114 . for the 1d cauchy distribution , the numerical results were already shown in fig .
[ 1d ] , we show the densities computed numerically for all 1d distributions for @xmath115 . for the 1d kv distribution , we see that beam s response is a delta function - like peak , for the wb , the density is very spiked and asymmetric . for all distributions , excepting the kv , an accumulation of the charge of the same sign as the external charge occurs . for the normal equilibrium distribution ,
the perturbation is skewed and spiked resembling the shape of the @xmath116-stable distribution . for the cauchy case ,
the perturbation is also skewed and spiked for @xmath117 , as illustrated in fig .
[ comp ] . while all other distributions
exhibit a symmetric peak around the external charge for @xmath106 , the kv distribution has two peaks that spread out with time , as shown in fig .
[ 1dkv ] ; with the increase of the velocity , the relative sizes of the peaks change , and , for @xmath115 , the left peak almost disappears and the right one looks almost like delta - function , as evidenced in fig .
[ 1d ] . in fig .
[ 2d ] , the lines of equal densities , for a certain set of times , for all 2d distributions considered , are shown for @xmath118 . for the 2d kv distribution , we see the spreading out delta function - like `` fronts '' , similar to the 1d case for @xmath106 . for the 2d wb distribution ,
the lines are triangular with a peak following the charge . for the 2d cauchy distribution ,
outer lines are almost circular ; for the normal distribution , they have a bit more complicated shape . for the smaller velocities ,
the profiles are less directed toward the charge . in fig .
[ 3d ] , the lines of equal densities in certain planes , i.e. , in three planes , each of which is parallel to the two out of the three coordinate axes , are illustrated .
the shape of the lines for every distribution has the same features as the ones for 1d and 2d cases .
we emphasize that our 1d and 2d cases are not simple reductions of the 3d case , but correspond to the 1d and 2d theories of electrodynamics with the different from the 3d case green s functions ; the expressions for @xmath45 also have different forms for the same distribution , but for the different dimensionalities .
although , in the method we consider in the present article , the green s functions are not used and the solutions of the poisson equation in the fourier space were employed , which have the same form ( [ furim ] ) in all cases , the general method for a finite beam @xcite uses green s function explicitly and its singularities is the main difficulty there . for the 1d case , the green s function is not singular and the general method can be implemented without accounting for the singularities and the results can be compared with the exact solution ( [ huge3 ] ) for the infinite beam ; after a certain change of variables , the general method is also able to deal with the infinite beam case .
this will be a reliable test of the method itself ; and the 2d and 3d results of the present paper will help to develop and test good ways to handle the singularities .
the general method is currently under development and we will elaborate on this in our future publications .
the similarity of the 1d , 2d and 3d cases is also a sign of the applicability of the exact 1d results for estimating the shielding in the real device .
the method we discussed herein was implemented as an object - oriented program in c++ .
the solution is stored as a multidimensional array over some grid in space - time ; for further usage , it can be evaluated for any point using interpolation .
the program is easily expendable for other external charge densities and equilibrium distributions and , in particular , can deal with the empirical ones .
the visualization is also very flexible : it is possible to adjust time values , the number of equal density lines , set the particular values of interest , and look at different projections and cross - sections of the 2d and 3d densities .
as it was mentioned in the introduction , the proof - of - principle ( pop ) experiment is planned at brookhaven national laboratory and the corresponding facility is currently under construction @xcite . in this subsection
, we describe how the results can be applied to the modulator of the real device . to recover the dimensional quantities ,
we need the debye radius and the plasma frequency ; for the pop , we have @xmath119 and @xmath120 .
we obtain for the dimensional density perturbation : @xmath121 where @xmath37 is a spatial dimension of the problem , for the real 3d case , @xmath122 . for the pop experiment ,
the modulator is constructed such that the interaction time is about one half of the plasma period , it depends on the hadron s velocity , as the modulator length is constant .
the velocity is measured in units of @xmath42 , in the pop experiment , we have @xmath123 . our computations , shown in fig .
[ 1d ] , demonstrate that extending the modulator up to a few plasma oscillations can significantly increase the density perturbation , i.e. , its maximum will be up to four times greater .
further increasing of the modulation time does nt increase the perturbation , as the modulator saturates , as shown in fig .
the amplification of the perturbation in the fel section is limited by the fel saturation . for the model - independent description of the fel saturation and its application to the theory of cec ,
see @xcite and @xcite , respectively .
these considerations provide limitations on a possible amplified perturbation that we can get , which , in their turn , determine the performance of the cec device .
in the present article , we considered a possible way to model the modulator section of the coherent electron cooling , i.e. , we developed a method for evaluating the dynamical shielding of an external charge in an infinite electron plasma ; for the certain case , we found analytical solution .
the software package we developed gives reliable results for a variety of equilibrium distributions and initial conditions .
we plan to use it in the analysis of the next section of the cec , the fel section , wherein the electron density perturbation from the modulator evolves in a free electron laser @xcite .
the results obtained can also be used as a testing ground for the more general method for a finite beam and for the pic simulations .
+ recently , we proposed a full theoretical model of the cec @xcite .
all sections , i.e. , the modulator , the fel amplifier , and the kicker , were described using the inverse integral transforms . the kicker can be described in a very similar way to the modulator . for the details on the fel section
, we refer to @xcite .
the methods for the inverse integral transforms inversion that we developed and tested in the present article open an opportunity to implement this model and get a reliable and fast complete numerical model of the cec .
various communications with g. wang , s. webb , i. pogorelov , and a. fedotov are gratefully acknowledged .
we thank a. woodhead for proofreading .
we use the following definition of the fourier transform : @xmath124 and the inverse fourier transform : @xmath125 when we write the fourier transform or its inverse for the dimensionless variables , we use @xmath126 and @xmath48 , respectively . the discrete fourier transform ( dft )
assigns to the set of points @xmath127 the set of points @xmath128 : @xmath129 we use the following notation : @xmath130.\end{aligned}\ ] ] the dft can be computed numerically using the effective fast fourier transform ( fft ) algorithm , there are parallel algorithms allowing to increase the speed of computations .
the laplace transform is defined as follows : @xmath131 and the inverse : @xmath132 where @xmath109 is a real constant greater than the real parts of all singularities of @xmath133 , we use @xmath134 and @xmath47 for the dimensionless variables .
the inverse fourier transform can be approximated by the discrete one .
approximating the integral by @xmath135 and , introducing the following notations @xmath136 we obtain : @xmath137.\label{invfur}\end{aligned}\ ] ] similar expressions can be written for the multidimensional fourier transform .
the inverse laplace transform can be expressed via the fourier transform : @xmath138 where @xmath109 is the same as in ( [ invlap ] ) and subscript @xmath139 stands for the fact that the transform is over @xmath139 . evaluating the inverse fourier transform via ( [ invfur ] ) , we obtain : @xmath140.\end{aligned}\ ] ] for the 1d problems , the inverse laplace - fourier transform can be computed via the dft as follows : @xmath141\left(-1\right)^{n^{\left(1\right)}}\}_{0\leq n^{\left(1\right)}\leq n^{\left(1\right)}-1}\big],\end{aligned}\ ] ] where superscript @xmath142 corresponds to the laplace transform , and @xmath143 to the fourier one . for the 2d and 3d cases , the similar expressions can be easily written .
the exponential integral functions are defined by @xcite : @xmath144 these functions can be computed via the convergent series for a small argument : @xmath145 where @xmath146 is the euler-mascheroni constant , and via the asymptotic series for a large one : @xmath147 @xmath148 where it is assumed that @xmath149 . the error function and the complementary error function
are defined by @xcite : @xmath150 respectively , these functions can be computed via the convergent series for a small argument and via the asymptotic series for a large one : @xmath151 to compute the density perturbation for the normal equilibrium distribution , we need to compute the following expression : @xmath152,\end{aligned}\ ] ] where @xmath153 . for certain values of @xmath154 , @xmath155 diverges , while @xmath156 is finite , thus we use the series expansions for the whole expression : @xmath157 @xmath158 to perform the fft , one needs to compute the expression ( [ expr ] ) exactly at @xmath159 , for this case , we write the following series : @xmath160 thus @xmath161 99 v. n. litvinenko , y. s. derbenev , phys . rev . lett . * 102 * , 114801 ( 2009 ) .
i. pinyaev _
et al_. , in _ proceedings of the international particle accelerator conference 2012 _ , ( new orleans , usa , 2012 ) , p. 400 , moppd016 , url : http://accelconf.web.cern.ch/accelconf/ipac2012/papers/moppd016.pdf ; in _ proceedings of the international particle accelerator conference 2013 _ , ( shanghai , china , 2013 ) , p. 1535
, tupfi081 , url : http://accelconf.web.cern.ch/accelconf/ipac2013/papers/tupfi081.pdf . g. wang , m. blaskiewicz , phys .
e * 78 * , 026413 ( 2008 ) .
a. elizarov , v. litvinenko , g. wang , in _ proceedings of the international particle accelerator conference 2012 _ , ( new orleans , usa , 2012 ) , p. 3171
, weppr099 , url : http://accelconf.web.cern.ch/accelconf/ipac2012/papers/weppr099.pdf . g. bell _
et al_. , in _ proceedings of the 2011 particle accelerator conference _ , ( new york , usa , 2011 ) , p. 235
, mop067 , url : http://accelconf.web.cern.ch/accelconf/pac2011/papers/mop067.pdf ; in _ proceedings of the international particle accelerator conference 2012 _ , ( new orleans , usa , 2012 ) , p. 3231
, theppb002 , url : http://accelconf.web.cern.ch/accelconf/ipac2012/papers/theppb002.pdf .
a. a. vlasov , sov .
usp . * 10 * , 721 ( 1968 ) .
a. j. turski , ann .
( n.y . ) * 35 * , 240 , ( 1965 ) .
m. abramowitz and i. stegun , handbook on mathematical functions ( dover publications , inc .
, new york , 1965 ) .
s. webb , g. wang , and v. litvinenko , phys . rev .
beams * 14 * , 051003 ( 2011 ) .
a. elizarov and v. litvinenko , in _ proceedings of the free electron laser conference 2013 _ , ( new york , usa , 2013 ) , mopso17 , url : http://accelconf.web.cern.ch/accelconf/fel2013/papers/mopso17.pdf .
v. n. litvinenko , report no .
c - a / ap/480 , tech - note , bnl , 2013 , http://public.bnl.gov/docs/cad/documents/on%20the%20fel%20gain%20limit.pdf . v. n. litvinenko , y. s. derbenev , phys . rev . lett . *
110 * , 269504 ( 2013 ) . | in the coherent electron cooling , the modern hadron beam cooling technique , each hadron receives an individual kick from the electric field of the amplified electron density perturbation created in the modulator by this hadron in a co - propagating electron beam .
we developed a method for computing the dynamics of these density perturbations in an infinite electron plasma with any equilibrium velocity distribution a possible model for the modulator .
we derived analytical expressions for the dynamics of the density perturbations in the fourier - laplace domain for a variety of 1d , 2d , and 3d equilibrium distributions of the electron beam . to obtain the space - time dynamics , we employed the fast fourier transform ( fft ) algorithm .
we also found an analytical solution in the space - time domain for the 1d cauchy equilibrium distribution , which serves as a benchmark for our general approach based on numerical evaluation of the integral transforms and as a fast alternative to the numerical computations .
we tested the method for various distributions and initial conditions . |
the search for general classes of solutions of the einstein equations that either have pure vacuum for their source or a non - zero cosmological constant , that are of petrov type n , and have principal null rays with non - zero twist has been continuing for a very long time now .
we will use the ( relatively common ) nomenclature `` einstein spaces '' for solutions with this sort of a source , i.e. , either the pure vacuum or that vacuum with a non - zero cosmological constant , @xmath0 , appended to it .
with this idea firmly in mind , there are only two solutions known so far for twisting type n einstein spaces : the one described by i. hauser @xcite , which has @xmath3 , and the one by j. leroy @xcite , with a non - zero value for @xmath0 , but which in the limit as @xmath4 degenerates to a flat solution , rather than one of type n. because of the apparent difficulty of the problem , many different approaches have been used to attempt the finding of such a solution . with a requirement of one or more killing vectors
, the problem can be reduced to the solution of a single , nonlinear ode , which has been produced in several forms by different authors @xcite ; nonetheless this approach has produced no new solutions .
looking at the problem as a reduction from complex - valued manifolds via plebaski s hyperheavenly equation @xcite has produced no new solutions @xcite .
therefore we were quite interested when we became aware of a different approach via a recent paper by pawe nurowski @xcite , looking for exact solutions of this type with non - zero cosmological constant .
many of nurowski s research articles use the fact that one can productively study ( 4-dimensional ) lorentz geometries which admit a shearfree , geodesic null congruence of curves by viewing the 3-parameter space that picks out any particular curve in the congruence , as a ( 3-dimensional ) cr structure @xcite . in @xcite , he and his collaborators use the cr function in such a structure to create a very appropriate choice of coordinates for a twisting type n einstein space , and reduce the einstein equations to a set of nonlinear pde s for a couple of functions of three variables . then in @xcite , he makes a clever ansatz depending only on a single variable and discovers a particular twisting solution ; unfortunately that solution turns out to be the same as the one mentioned above as first found by leroy , as he notes in a more recent paper @xcite .
however , we were quite intrigued by the approach and have made some efforts to follow it through with the hopes of obtaining more general solutions of the equations in nurowski s article .
a spacetime of type n allows one , and only one , congruence of twisting , shearfree , null geodesics , referred to as a principal null direction ( pnd ) ; this ( 3-parameter ) family of null geodesics allows the option to choose a single coordinate @xmath5 along any such geodesic , and to associate the other three degrees of freedom in the parameter space as a model for a cr manifold .
we first insist that our manifold admit a killing vector in the real direction in this ( 3-dimensional ) cr manifold , so that the remaining unknown functions depend only on the complex coordinates there , and then calculate the ( infinite - dimensional ) classical symmetries for the system .
this allows us to derive a quite simple nonlinear third - order ode which the invariant solutions of the classical symmetries must satisfy .
because this equation does not contain the independent variable explicitly , it can be immediately reduced to the following second - order ode , for @xmath6 , with two slightly different forms that differ by a constant : @xmath7 we are then able to show that the leroy - nurowski solution is indeed a special solution for this equation . at this point
it is worthwhile to introduce a question as to how one knows that the new solution of nurowski does indeed describe locally the same manifold as the solution found by leroy .
the method was originally created by e. cartan @xcite to prove equivalencies of cr structures , without the need of actually determining an explicit coordinate transformation between the two sets of coordinates on two manifolds .
instead , one determines the values of a set of invariant quantities for a cr structure , the same for all equivalent such structures .
therefore it is necessary to calculate the invariants for leroy s solution and compare them with the ones already known to nurowski for his solution , noting that they are the same constants .
we have therefore also calculated these invariants for our class of solutions , which we find to be quite different . for the case
when the constant @xmath8 in ( [ geqn ] ) is zero , an integral transformation may be performed to reduce that equation further , to a first - order ode of abel type , for @xmath9 : @xmath10 which , quite unfortunately , we have not been able to identify as any of the known solvable types of abel equations @xcite .
nonetheless , we believe that these two equations , at the time of publication , are the simplest odes available that determine nontrivial twisting type n einstein spaces . returning to the case when @xmath11 , our current examinations suggest the possibility that the solutions of this equation might define a new class of transcendental functions , which constitutes the major result of this paper , establishing a new set of solutions to the type n problem with @xmath12 . the remainder of the paper will describe the process involved in this and our reasons for stating that these are indeed new solutions . in particular , we will present solutions to in the forms of power series and puiseux series , both shown to be locally convergent .
although prior to this time there were indeed only two known twisting type n einstein spaces , it is always worth remembering that the solution space for the problem is in fact quite large .
it has been shown by sommers @xcite that the full set of solutions for type n is given by two complex functions of two real variables .
surely the requirement of non - zero twist puts a very strong constraint on this , but it is expected that there should be a large number of new analytic functions involved in the full solution of the twisting type n problem .
as already mentioned those ( 4-dimensional ) lorentz geometries which admit a shearfree , geodesic null congruence of curves have some history of people using cr structures to study them .
they were first introduced into mathematics by poincar and extensively studied by e. cartan @xcite .
good sources of background on the relation of these two geometric concepts may be found , for instance in the thesis of nurowski @xcite , and also in the very detailed discussion of their use for einstein spaces in his article with hill and lewandowski @xcite , which generalized earlier work to include the possibility of a non - zero cosmological constant .
the approach to a lorentz - signature manifold begins with the usual form for the metric @xcite in terms of a basis of 4 null 1-forms , modified for our choice of signature , @xmath13 : @xmath14 where @xmath15 is the complex conjugate of @xmath16 while @xmath17 and @xmath18 are real , and the product is the usual symmetric product of 1-forms to create the metric . for einstein spaces of type n , which have only one pnd
, we call its tangent vector @xmath19 , which is a well - defined vector field in the neighborhood of each point , and take @xmath20 as its dual 1-form .
the requirement that the twist of @xmath19 should be non - zero is equivalent to the restriction that @xmath21 .
we also choose a real coordinate @xmath5 as a parameter along this congruence , and a real function @xmath22 on the manifold so that @xmath23 ( @xmath24 , @xmath25 ) .
this congruence , still locally , is a 3-parameter family of null shearfree geodesics , so that at each value of @xmath5 , the manifold still needs three additional coordinates . following the work described in detail in @xcite ,
one is able to describe those degrees of freedom as coming from a 3-dimensional , strictly pseudoconvex cr manifold , @xmath26 , equipped with ( at least ) one non - constant cr function @xmath27 , and classes of pairs of 1-forms @xmath28 ( real ) and @xmath29 ( complex ) such that @xmath30 with @xmath31 a complex - valued function on @xmath26 .
a cr structure is a 3-dimensional real manifold @xmath26 equipped with an equivalence class of pairs of 1-forms , @xmath32 , as above .
another pair @xmath33 is considered equivalent to @xmath32 , and therefore simply another representative of the same equivalence class , iff there are functions @xmath34 ( real ) and @xmath35 , @xmath36 ( complex ) on @xmath26 such that @xmath37 this cr manifold can be lifted into a spacetime , so that the tetrad may be displayed in the following way @xcite : @xmath38 of course this is exactly the form that a spacetime with such a distinguished pnd is supposed to have @xcite .
referring back to , we can see that the twist is just proportional to @xmath39 . using to determine this , we see that these forms have been chosen to put the non - zero character of the twist very clearly in evidence as non - zero . however , still looking at those equations , the function @xmath31 that appears there is quite an important new function on the manifold . using the closure of the equations in
, one determines an important reality condition on the derivatives of @xmath31 : @xmath40 once a dual basis of vector fields is introduced , which , however , is not a commutative basis : @xmath41 = -\rmi \partial_0 , \qquad \left[\partial_0 , \partial \right ] = c \partial_0 , \qquad \left[\partial_0 , \bar\partial\ , \right ] = \bar c \partial_0.\end{aligned}\ ] ] at this point one has sufficient information to write out the explicit forms of the einstein equations , which , in generality , say that @xmath42 with all the other components of the ricci tensor vanishing .
we quote from @xcite and @xcite which show that the results are the following : @xmath43 where the functions @xmath44 , @xmath45 ( complex ) and @xmath46 , @xmath47 ( real ) , all independent of @xmath5 , satisfy @xmath48 given all the above , the functions @xmath49 and @xmath45 define a twisting type n einstein space , of the form given in and , iff the unknown functions @xmath31 and @xmath47 satisfy the following system of pdes on @xmath26 @xmath50 p = \case{2}{3 } \lambda p^3 , \label{nurowskieqn } \\ \psi_3 = 0 , \qquad r_{33 } = 0 .
\label{tn3}\end{aligned}\ ] ] as well as one inequality @xmath51 in order that the spacetime should not be conformally flat . in terms of those variables
already defined , the ricci tensor component @xmath52 and the weyl scalars @xmath53 and @xmath54 take on quite nasty - looking expressions @xmath55
\cos^4 \left(\frac{r}{2 } \right),\end{aligned}\ ] ] @xmath56 \mathrm{e}^{\rmi r/2 } \cos^3 \left(\frac{r}{2 } \right),\end{aligned}\ ] ] @xmath57 \right\ } \mathrm{e}^{-\rmi r/2 } \cos^3 \left(\frac{r}{2 } \right),\end{aligned}\ ] ] where the function @xmath58 is defined by @xmath59 following the procedure of @xcite ( with the use of maple ) which calculated the curvature tensor using cartan s structure equations , we present the calculated @xmath54 above ( _ simplified _ with the use of @xmath60 ) for @xmath12 , which we believe has not been previously published . despite the frightening appearance of @xmath52
, the equations and @xmath60 together do imply the requirement @xmath61 .
this tells us that within the established formalism the twisting type n solutions to the einstein equations automatically satisfy the condition for an einstein space , i.e. , vacuum with or without a cosmological constant . for @xmath3 ,
the statement is obviously true ( see also @xcite p.451 ) and was used in @xcite to prove the cr embeddability of twisting type n vacuums , without cosmological constant . for @xmath62 , one uses to substitute the term @xmath63 in @xmath52 and notices that the resulting expression is a linear combination of @xmath64 and @xmath53 .
the equation @xmath61 is therefore superfluous for the type n problem , which facilitates our calculation greatly .
it is useful to understand the meaning of the operator @xmath66 by introducing a real coordinate system @xmath67 on @xmath26 such that we have @xmath68 with @xmath69 a complex - valued function @xcite .
in addition , the function @xmath70 relates to the function @xmath31 in the following way @xmath71 one major difficulty of fully solving the system ( [ tn1]-[tn3 ] ) is that unlike ordinary coordinate differentiations , the selected ( dual ) basis for the tangent space is not commutative , and , even worse , the operator @xmath66 itself involves the unknown function @xmath31 .
when this dependence on the real coordinates is written out explicitly , the original pdes become formidably lengthy . instead of facing this entire conundrum ,
we have decided to circumvent it , at least in this paper , by looking at the special case where the unknowns @xmath47 and @xmath31 have no @xmath65-dependence , i.e. , @xmath72 .
geometrically speaking , we insist that the spacetime admits a killing vector in the @xmath65-direction .
such an assumption simplifies the problem greatly in that one can treat the operator @xmath66 the same as @xmath73 , when acting on either @xmath47 or @xmath31 .
this is a generalization of the assumption made by nurowski @xcite , where it was simplified to just dependence on @xmath74 , i.e , two killing vectors assumed .
* theorem 1 * ( _ cr embeddability _
@xcite ) a cr structure with @xmath75 is cr embeddable . from this particular form of @xmath31 , a @xmath65-independent form of the function @xmath70
can be constructed as @xmath76 with a real - valued function @xmath77 satisfying @xmath78 associated to this @xmath70 , the tangential cr equation @xmath79 yields a second cr function : @xmath80 _ proof : _ because of the restraint @xmath81 , the system is compatible and has a real solution @xmath82 .
henceforward one can directly check that satisfies and that satisfies the equation @xmath83 . clearly , the cr functions @xmath84 and @xmath27 are functionally independent , i.e. , @xmath85
. therefore we acquire a second cr function . for a given function @xmath86 , the equation , viewed as a pde for @xmath70 , may give rise to multiple choices of the function @xmath70 , hence various @xmath28 s .
however , such an ambiguity only constitutes different representatives of the same cr structure . to see this
, one may look into the six cartan invariants ( details in section [ crsymmetry ] ) and notice that they are all uniquely determined by the function @xmath86 ( see , e.g. , and appendix b ) , given that the function @xmath5 defined in does not vanish . for cr structures with @xmath87 ,
they are all locally equivalent to a 3-dimensional hyperquadric inside @xmath88 @xcite .
an alternative proof would be to show that there always exists a coordinate transformation @xmath89 ( @xmath90 ) that takes a function @xmath91 satisfying to the @xmath65-independent form .
this can be confirmed by checking the compatibility of pdes regarding such an existence . in conclusion ,
the cr structure on @xmath26 is _ uniquely _ determined once a function @xmath86 is given .
the converse of the last statement above is , however , not true .
in fact various choices of the function @xmath31 may correspond to the same cr structure .
we will see examples of this in later sections .
related to this matter , our assumption of the function @xmath31 being @xmath65-independent is thus not a cr invariant property .
a function @xmath86 may acquire @xmath65-dependence through the transformation that takes one representative @xmath32 of the cr structure to another .
now we apply the assumption and the following notations @xmath92 and then rewrite the system ( [ tn1]-[tn3 ] ) as @xmath93 where the last equation arises from @xmath60 .
these are the pdes we aim to solve .
moreover , the weyl scalar @xmath54 reads @xmath94 \frac{\mathrm{e}^{-\rmi r/2}}{p_0 ^ 2 } \cos^3\left(\frac{r}{2 } \right ) .
\label{psi4}\ ] ]
we follow the standard procedure ( e.g. , @xcite ) to calculate the classical symmetries of the system ( [ cc]-[psi3 ] ) . since is generally complex
, we have to include its complex conjugate as well in the calculation .
moreover , we treat as a constraint and encode it and its differential consequences directly into the choice of intrinsic coordinates so that this equation no longer needs further attention .
this gives us three pdes for three dependent variables @xmath47 , @xmath31 and @xmath95 which depend on two independent variables @xmath27 and @xmath96 .
the intrinsic coordinates within the first four jets that are relevant to the calculation are chosen as follows @xmath97 the rest of the jet variables , such as @xmath98 , @xmath99 , @xmath100 etc .
, can be expressed in terms of the intrinsic coordinates through the pdes and their differential consequences . with a considerable amount of manual work on the algebraic computer program maple , we have managed to find the classical symmetries with the generating section given by @xmath101 where @xmath102 is an arbitrary function of @xmath27 that is sufficiently differentiable .
the lie bracket of two symmetries with , respectively , @xmath103 and @xmath104 yields a third symmetry with a new @xmath105 given by @xmath106 : = a_1 \partial_\zeta a_2 - a_2 \partial_\zeta a_1.\ ] ] therefore , we indeed obtain an infinite - dimensional set of classical symmetries for the system ( [ cc]-[psi3 ] ) .
in particular , they reduce to translational symmetries for nonzero constant @xmath107 , and scaling symmetries for @xmath108 .
setting the generating section @xmath109 to zero and solving for @xmath47 , @xmath31 and @xmath95 , we are able to obtain a remarkable ansatz for the einstein equations : @xmath110 with a new real argument @xmath111 where the constant @xmath112 and the undetermined functions @xmath113 are all real - valued .
one may easily verify these expressions by direct calculation . substituting the ansatz into and and noticing that all dependence on @xmath114 , except those in the argument @xmath115 , can be factored out
, we have a neat reduction from the pdes to a system of two odes for @xmath116 and @xmath117 only : @xmath118 the satisfaction of the second ode above is given by the introduction of a single new , real - valued function @xmath119 such that @xmath120 then the first ode simply becomes @xmath121 since this ode does not have the argument @xmath115 appearing explicitly , we can lower the order of the ode through the standard transformation @xmath122 and obtain an even simpler equation of the second - order @xmath123 a solution @xmath124 to can give rise to a solution @xmath119 to at least locally by inverting @xmath125 with @xmath126 constant
. this solution will be physical if it also makes @xmath113 real - valued via , which requires that locally @xmath127 therefore we are only interested in solutions for @xmath128 that are monotonically increasing , or equivalently , positive @xmath129 .
we can also consider the special case of with @xmath130 and @xmath12 , i.e. , @xmath131 by introducing the following integral transformation @xmath132 of which the inverse has the form @xmath133 we can further reduce to an abel ode of the first kind @xcite , as already noted in the introduction : @xmath134 once the general solution @xmath135 is acquired with a constant @xmath136 , we can find the general solution @xmath129 of by solving the following ode @xmath137 of which the solution is given by @xmath138 simple as both and may appear , so far we have had no luck finding their explicit general solutions . for more comments on and abel odes in general ,
see appendix a.
to identify new twisting type n einstein spaces obtained from , we refer to the following theorem as a natural way to classify metrics equipped with cr structures . * theorem 2 * ( @xcite , see _ theorem 1.2 _ and references therein ) let @xmath139 be a 4-dimensional manifold equipped with a lorentzian metric and foliated by a 3-parameter family of shearfree null geodesics .
then @xmath140 is locally a cartesian product @xmath141 .
the cr structure @xmath142 on @xmath26 is uniquely determined by @xmath143 and the shearfree null congruence on @xmath140 . by definition ,
a type n spacetime at each point has a unique pnd . in the case of vacuums ( with or without @xmath0 )
, a pnd must be geodesic and shearfree @xcite .
thus for every twisting type n einstein space , the shearfree null congruence is unique .
hence to confirm a new twisting type n vacuum metric , it is sufficient to show that its cr structure is distinct from the one of known metrics .
this can be routinely done by computing the six cartan invariants @xcite , which are denoted respectively by @xmath144 cartan showed that two local cr structures are equivalent iff their six cr invariants ( defined when @xmath145 in ) are identical , except possibly for a sign difference in both @xmath146 and @xmath147 @xcite . with the assumption of @xmath65-independence , we can write down , for instance , the simplest invariant computed from the 1-forms defined in : @xmath148{(r\bar{r})^{7 } } } , \cr r = \case{1}{6}\left(\partial_{\bar\zeta } \bar{l } + 2 \bar{c } \bar{l}\right ) , \qquad l = -\partial_\zeta \partial_{\bar\zeta } c - c \partial_{\bar\zeta } c. } \label{alphai}\end{aligned}\ ] ] here the function @xmath145 , following the notation of cartan , is not to be confused with the coordinate @xmath5 along the null congruence . for our calculated @xmath149 and @xmath150 ,
see appendix b. note that this @xmath146 only relies on @xmath151 , @xmath152 and their derivatives , which is also the case for all the other cartan invariants .
we first point out a remarkable feature of these invariants computed from the ansatz ( [ pccb],[zdef ] ) .
* proposition * given the ansatz ( [ pccb],[zdef ] ) , all the following quantities are independent of @xmath153 and @xmath154 except those in the argument @xmath115 : @xmath155 in another word , they are all functions of @xmath115 only , e.g. , @xmath156 .
_ proof _ : except for a lengthy but straightforward symbolic computation with maple , we are , at the moment , still not aware of any other more insightful way of proving this result . here
we only emphasize that the law @xmath157 is in general _ not _ true in the complex domain ; failing to notice this may cause an erroneous conclusion .
* remark * for a fixed @xmath115 , the presence of the functions @xmath107 and @xmath158 in @xmath146 and @xmath147 themselves only affects their signs .
more specifically , the only dependence on @xmath107 and @xmath158 takes the following forms : @xmath159 hence the product @xmath160 is a function of @xmath115 only . according to cartan @xcite , this sign situation
is accounted for by a local cr diffeomorphism and therefore does not generate a new cr structure .
hence we have proved the following theorem .
* theorem 3 * locally , the cr structure ( as an equivalence class ) determined by the function @xmath31 given in ( [ pccb],[zdef ] ) is independent of the choice of the function @xmath161 , once the form of @xmath162 is fixed . altogether , the freedom of choosing various @xmath161 does not affect the cr structure of a type n metric which we are considering .
hence , for the simplicity of representing new metrics distinguished by cr structure , we can just set @xmath163 ( see the conclusions ) .
in hindsight , the classical symmetries we have obtained are nothing more than a particular manifestation of the underlying cr equivalency .
we believe this connection between the two may as well suggest a more general concern if one aims to find , through the ( classical or higher ) symmetries , additional exact solutions to the einstein equations formulated with cr structures .
we will see later examples of solutions that have _ constant _ cr invariants , and remarkably , one of them is the solution of leroy - nurowski .
nonetheless , this feature is generally not true for other solutions .
before we try to solve for type n solutions , it is important to find out in advance those conformally flat solutions satisfying @xmath164 which are automatically contained in the general solution of .
we insert the ansatz ( [ pccb],[zdef ] ) into the expression for @xmath54 given by , and re - normalize @xmath54 to pull out just a simple complex - valued function of @xmath115 : @xmath165 we now apply and use to substitute for @xmath166 , which gives us @xmath167 + \rmi \left [ -2 c_1 \left(j '' + 3 \lambda j j ' \right ) \right],\ ] ] or in terms of @xmath129 , @xmath168 + \rmi \left [ -2 c_1 p \left(p ' + 3\lambda j \right ) \right],\ ] ] where we have put the real and imaginary parts in separate brackets . replacing @xmath169 with the help of , we can rewrite the equation as @xmath170 which clearly has @xmath171 , i.e , all conformally flat solutions , as some of its solutions .
if @xmath129 is not restricted to the real domain , then solving the first - order ode @xmath171 for @xmath129 leads to the following general solution @xmath172 with a complex constant @xmath136 .
if , instead , we restrict @xmath129 to be real , a simultaneous vanishing of the real and imaginary parts of @xmath173 respectively yields the following set of two equations , provided @xmath22 , @xmath174 both of which are consistent with .
there are now two cases for solutions .
the case @xmath175 requires that both odes be satisfied , so that we have a unique solution @xmath176 which , by solving @xmath177 , gives rise to @xmath178 in the limit @xmath179 , the above solution becomes even simpler : @xmath180 from , we have @xmath181 with negative - valued @xmath182 .
note that it is only at this stage that the reality condition on @xmath183 , i.e. , @xmath184 , requires @xmath185 , i.e , a negative cosmological constant .
an important remark that can be made is that the extended form of the leroy - nurowski solution ( see the next section ) resembles this solution greatly with simply differences in the coefficients . for the other case
when @xmath186 , we have @xmath187 with a real constant @xmath136 . from , the solution @xmath128 is determined by @xmath188 since @xmath184 , we can not have both @xmath189 and @xmath190 . hence , we can discuss three other sign possibilities , the details of which are put in appendix c. we note that the special solution corresponding to @xmath191 serves as the single `` point '' where these two families of conformally flat solutions are joined up .
modulo possible sign differences in @xmath146 and @xmath147 caused by square roots as already discussed , the cartan invariants for both and , as calculated via and the equations for the other invariants , as presented in appendix b , are given by @xmath192{\frac{2}{5 } } , \qquad \beta_i = \frac{41}{2\sqrt{10 } } , \qquad \gamma_i = \frac{29}{2 \sqrt{10 } } , \cr \theta_i = 3\rmi \sqrt{\frac{2}{5 } } , \qquad \eta_i = -\frac{\rmi}{\varepsilon } \cdot \frac{2^{19/4}}{5^{3/4 } } , \qquad \zeta_i = -\frac{327}{40 } , \qquad \varepsilon=\pm 1 . }
\label{cartaninvarpflat}\end{aligned}\ ] ] remarkably , they are all constant and do not depend on @xmath112 . nonetheless , this is not the case for the other conformally flat solutions obtained from with @xmath193 of which the cartan invariants are generally functions of @xmath115 and @xmath136 .
for instance , simplified by and @xmath177 , the first cartan invariant satisfies @xmath194 where @xmath119 belongs to one of the three cases described in appendix c. two conformally flat einstein spaces may have non - equivalent cr structures .
this does not conflict with the previous theorem because in a conformally flat spacetime , one is free to make different choices from among the multiple shearfree null congruences and therefore may have non - equivalent cr structures attached to them . for a further integration of
, we have three separate cases .
_ case 1 : _ @xmath291 .
we always have @xmath292 .
then the solution is determined by @xmath293 in the real domain , the inverse function @xmath119 is well defined over @xmath294 instead of the entire real line , and has singularities at @xmath295 .
_ case 2 : _ @xmath296 .
we need @xmath297 for @xmath292 .
the solution is determined by @xmath298 in the real domain , the inverse function @xmath119 is well defined over @xmath299 , and has singularities at @xmath300 .
_ case 3 : _ @xmath301 .
we need @xmath302 for @xmath292 .
the solution is determined by @xmath303 in the real domain , the inverse function @xmath119 from above is well defined over the entire real line .
now we can reveal a fuller extent of the exact twisting type n solution first discovered by leroy , and re - derived by nurowski within the framework of cr geometry , upon the latter of which our current work is mainly based .
we hope that our derivation of this solution will make the process behind the previous discoveries appear clearer . given nurowski s form of the solution ( see @xcite or ) and recasting it into the form of the ansatz ( [ pccb],[zdef ] ) and
, we find the following special solution to @xmath195 which gives rise to a solution to : @xmath196 in the limit @xmath179 , the above expression becomes even simpler : @xmath197 which is quite similar to that of .
back to the case with @xmath198 , using , we have @xmath199 with a negative @xmath182 .
note that it is only at this stage that the reality condition on @xmath116 requires @xmath185 . in the end
, our extended version of the leroy - nurowski solution takes the form @xmath200 , \label{nurwoskisolnc } \\ z = -\rmi \left(\int \frac{1}{a } \mathrm{d } \zeta - \int \frac{1}{\bar a } \mathrm{d } \bar{\zeta } \right).\end{aligned}\ ] ] the flexibility of choosing the function @xmath153 and real constant @xmath201 may perhaps facilitate a possible future application of the solution . from this extended version ,
one can obtain the original form of nurowski @xcite by setting @xmath202 and consequently , @xmath203 note that all @xmath112 s are canceled out in the above expressions .
hence another way of obtaining is by taking the limit @xmath204 in and ( cf . ) and setting @xmath205 and @xmath206 .
modulo possible sign differences in @xmath146 and @xmath147 caused by square roots as already discussed , the cartan invariants calculated from and are both given by @xmath207 like , they are all constant and do not depend on @xmath112 .
for simplicity , assume that @xmath130 in the ode . now consider the power series solution of satisfying the regular initial conditions @xmath208 , @xmath209 and @xmath210 . a simple calculation gives us the first few terms of this series @xmath211 moreover , this series solution , convergent in a neighborhood of @xmath212 according to the cauchy existence and uniqueness theorem , is of type n with a non - vanishing weyl scalar @xmath213 .
particularly , @xmath214 to see that the solution is not equivalent to the leroy - nurowski solution , we calculate the first cartan invariant @xmath215 via which , in this case , is no longer a constant . in particular , this series solution has @xmath216 with @xmath215 and also @xmath217 continuous at @xmath212 , while the values given in are always non - zero constants .
this is sufficient to assert that the ode as well as its reductions and indeed contain new twisting type n solutions .
one feature that makes preferable to the other two odes and is that the conformally flat solution and the extended leroy - nurowski solution are just simple quadratic functions , without poles in the complex plane , compared to their counterparts and .
also note that these quadratic solutions with @xmath130 do not correspond to any solution of the abel equation since the form with the non - constant function @xmath218 , excludes all quadratic functions as solutions .
these well - behaved quadratic solutions facilitate a study of the power series solutions near them .
additionally in appendix e , we also present a puiseux series solution to , the existence of which is suggested by the weak painlev tests performed in appendix d. to simplify the notation , we apply the scaling transformation , one may choose a sign for @xmath0 in order to have @xmath219 . ]
@xmath220 , @xmath221 with @xmath62 , @xmath175 such that takes on the form already noted as with @xmath11 : @xmath222 we look for power series solutions for this equation corresponding to the regular initial conditions @xmath223 , @xmath224 .
the first few terms of this series read @xmath225 where all odd order terms vanish .
the remainder of the coefficients in the series can be determined by a recursion relation which is valid beginning with @xmath226 : @xmath227 while @xmath228 and @xmath229 can be easily read off from .
it is clear that this relation allows one to calculate the coefficients to whatever order desired .
one can easily see that the coefficient of @xmath230 , namely @xmath231 , is a @xmath19th - order polynomial , @xmath232 , divided by @xmath233 .
remarkably , this infinite series reduces to simple quadratic functions in two special cases .
the reason for this is that for every value of @xmath234 , the polynomial @xmath232 has the factors @xmath235 , as can be seen in the few terms demonstrated in above and can easily be shown by induction .
hence for @xmath236 , we retrieve the leroy - nurowski solution , which in this notation is simply @xmath237 as well , for @xmath238 , we retrieve a conformally flat solution , which has the form @xmath239 for all other values of @xmath240 , the formal series solution may then be viewed as a generalization of these two known solutions , in terms of a power series with infinitely many terms .
it is interesting that in every one of these polynomials , @xmath232 , all coefficients are negative , so that the only possible real roots would be negative .
our numerical calculations suggest that none are smaller than @xmath241 , and that there are no other roots common to all these different polynomials .
the series does define , in the complex domain , a function holomorphic in some neighborhood of the origin as is shown by the following method of determining a non - zero radius of convergence for it .
we present the proof in appendix f. * theorem 4 * given the series with the recursion relation and a fixed @xmath240 , one has the following bound : @xmath242 provided that one can pick two constants @xmath243 and @xmath244 such that they satisfy @xmath245 the existence of such an upper bound on @xmath246 guarantees a lower bound @xmath247 on the radius of convergence . for instance , if we take @xmath248 , which lies nicely in the interval between @xmath249 and @xmath241 , we can at least pick @xmath250 satisfying both and .
the bound is by no means optimal at every @xmath240 .
in fact , our numerical integrations of with @xmath251 sampled between @xmath241 and @xmath249 all indicate that in the _ real _ domain , the series solutions with @xmath252 are all well sandwiched between the parabolic curves of and , and therefore suggest an infinite radius of convergence on the real line .
moreover , by applying the transformation @xmath253 to and studying the formal ( puiseux ) series expansion of the transformed ode at the origin , we find the following asymptotic expansion in a complicated way , hence not presented here . ] of as @xmath254 ( cf . ):
@xmath255 where @xmath256 is an arbitrary constant .
this asymptotic behaviour at infinity , consistent with our numerical calculations , again suggests that we may significantly extend the radius of convergence for at least in the real domain .
an additional comment is that the cartan invariant @xmath146 , computed from with @xmath257 , is not constant and has a dependence on @xmath112 , contrary to the special cases for those values of @xmath251 at the two endpoints of the interval of values for @xmath251 being considered .
we have begun with the advantage of prior work done on the use of ( 3-dimensional ) cr manifolds to look for solutions of the einstein field equations that correspond to algebraically - degenerate einstein spaces , with twisting principal null directions . a general solution of those reduced field equations for the two functions of three variables would generate all twisting solutions of petrov type n. of course we did not achieve this ; however ,
after the assumption of a single killing vector in a particular direction , our ansatz for group - invariant solutions obtained from the infinite - dimensional classical symmetries of the field equations , allowed us to obtain a single ode , the solutions of which would generate a family of solutions of type n with twisting principal null directions . that ode is either a rather simple , third - order nonlinear equation for @xmath119 in which the independent variable @xmath115 does not appear or , equivalently , an even simpler , second - order nonlinear equation for @xmath6 , where @xmath2 is a dimensionless re - scaling of @xmath258 and @xmath36 is a re - scaling of @xmath259 , which includes a non - zero value for @xmath0 , the cosmological constant . within the same ansatz
, we have also investigated all the cases of solutions corresponding to conformally flat spacetimes to which type n solutions may degenerate , which helps us look for non - trivial cases .
we have studied this second - order equation at some length .
in particular it contains one parameter , @xmath112 , which may always be re - scaled to the value @xmath260 unless it happens to be zero . in the case
that it is zero , the equation can be reduced still further , to a first - order equation of abel type . following standard approaches to abel equations we were unable to determine any method that we thought would generate reasonable type n solutions , although this is still an ongoing project of considerable interest .
however , when @xmath112 is not zero we have considered various sorts of solutions which it might have . we have shown that it does have solutions which are holomorphic , in the complex plane in a neighborhood of the origin , and have found an asymptotic behavior near the ( real ) infinity .
in particular we have picked out especially those solutions which are even functions of @xmath2 and looked at power - series solutions about the origin , both analytically and numerically via a maple computer program .
we have determined a moderately - simple recursion relation for the coefficients of the powers of @xmath261 in the series solutions , which determines the coefficient @xmath262 , of @xmath263 ( @xmath234 ) in terms of all the previous coefficients , looking at all of them as determined by the value of @xmath264 .
this series terminates quickly for just two particular values of @xmath251 , in the form @xmath265 , with @xmath44 constant , different for the two values of @xmath251 .
the value @xmath236 generates the previously - known leroy - nurowski solution , while the other one @xmath238 is unfortunately simply a conformally flat solution . to ensure that these series solutions are distinct from the leroy - nurowski solution , we have used the work of cartan on the question of the equivalence of two cr manifolds , which requires the equality of the set of six cartan invariants
we have found that any value of @xmath251 between these two special values generates cartan invariants that are quite different from those at the endpoints of this interval , and therefore distinct from those of the leroy - nurowski solution .
the solutions characterized by values of @xmath251 between @xmath241 and @xmath249 have an asymptotic behavior , via a puiseux series around the ( real ) infinity , that has the same form @xmath266 as the conformally flat solution aforementioned , but also lower - order terms involving third - roots of @xmath2 , which doubtless generate algebraic singularities there .
numerical integrations via maple agree with this behavior , showing negative values of @xmath1 as needed and very simple structure , for all real values of @xmath2 .
the same numerical integrations do show singularities in the solutions for @xmath267 . as well , numerical calculations of the coefficients @xmath262 , for several values of @xmath268 ( e.g. , @xmath269 ) show that starting at a large enough @xmath19 , they alternate in sign while their absolute values are monotonically decreasing at rapid rates .
we therefore postulate that these solutions are everywhere non - singular and well - behaved on the real axis , and believe that they might define new well - behaved , transcendental functions with algebraic singularities off the real @xmath2-axis .
the proof of such a conjecture is still being pursued ; nonetheless , we feel that the numerical calculations justify the belief that this is a sufficiently interesting result as to merit the attention of a wider audience . to conclude the discussion , we present here our new class of metrics , which
, without loss of generality , may be considered by setting @xmath206 in the ansatz .
although we will present it here with the new real coordinate @xmath115 introduced in , with this choice of @xmath153 it is the same as the usual coordinate @xmath74 used in . as well , our studies with the equation for @xmath129 , equivalently @xmath1 , allow us to replace @xmath115 by its form in terms of @xmath258 as determining the imaginary part of @xmath270 , via @xmath271 , namely , @xmath272 for simplicity of presentation , we show both forms below , with coordinates @xmath273 or @xmath274 : @xmath275\ ] ] with real - valued @xmath119 , @xmath276 and @xmath277 such that @xmath278 where @xmath112 is an arbitrary real parameter .
the function @xmath70 as in @xmath279 can be chosen so as to be real - valued : @xmath280 such that from , @xmath281 \rmd x } { \exp\left(\int f_2 \mathrm{d}z \right ) } = \frac{\mathrm{e}^{c_1 x } \rmd u - 2 \left[\int p^{-1 } \exp\left(\int f_2 p^{-1 } \rmd j \right ) \rmd j \right ] \rmd x } { \exp\left(\int f_2 p^{-1 } \rmd j \right)},\ ] ] where @xmath117 is given by @xmath282 meanwhile , the functions @xmath128 and @xmath129 respectively satisfy @xmath283 in particular , the original metric by nurowski @xcite corresponds to the case @xmath130 , @xmath284 and a proper choice of the integration constants in @xmath28 .
we can here note the philosophy that certain odes themselves may serve the purpose of defining _ new _ transcendental functions ; for instance , we recall the painlev functions and the associated odes .
hence our situation with new type n solutions being determined by a second - order _ nonlinear _ ode is presumably not too different from that of hauser s solution ( in terms of hypergeometric functions @xcite ) which is determined by a second - order _ linear _ ode , although it is true that there has already been much more extensive studies made on the properties of hypergeometric functions than have been made for newer functions defined by solutions of nonlinear odes that may not even have the painlev property .
the equation actually does have the following special solution @xmath285 however , it can be shown to correspond to a conformally flat solution , and hence is not interesting .
unfortunately , we have had no luck so far finding the general solution to or any other special solution other than . since constructing
the general solution to the generic abel ode has remained an open problem for decades , the general strategy of integration nowadays mainly lies in recognizing , within a suitable class of transformations , the ode in question as equivalent to a previously solved equation .
such a procedure has been programmed into the current state - of - the - art maple code ` dsolve ` ( or ` abelsol ` ) @xcite , which presumably covers all / most of the integrable classes presented in kamke s book @xcite and various other references ( e.g. , @xcite ) . however , this code , as tested by us , does not recognize as a known solved type , e.g. , the air class .
other attempts by us , such as the symmetry method , on finding special solutions all have failed or just led to .
so far we have not been able to find a similar reduction for the ode with @xmath198 , nor can we negate the possibility that with @xmath198 may contain different type n solutions other than the case with @xmath130 .
in fact , the cartan invariants calculated with generally do have a dependence on the constant @xmath112 even though this is not the case for all the conformally flat solutions and the leroy - nurowski solution ( see and ) .
a concise description of cartan invariants can be found in section 2.1.3 of @xcite . given @xmath86 and @xmath286 taken from where @xmath146 is presented , the next three cartan invariants , when @xmath145 , read @xmath287,\end{aligned}\ ] ] @xmath288,\end{aligned}\ ] ] @xmath289.\end{aligned}\ ] ] due to the length of @xmath290 and @xmath147 as calculated with maple for our studies , we will not put them here .
all cartan invariants are uniquely determined by the function @xmath86 .
we have three odes , , and at hand that may be explored for new twisting type n solutions .
a particular , probably useful way to decide which one of these equations has a better chance for one to find a solution is given by the ( weak ) painlev test @xcite .
this test reveals the nature of the movable singularities of the general solution of a nonlinear ode .
failing the test means the occurrence of certain undesirable movable singularities , e.g. , infinitely branched singularities , that relate to non - integrability @xcite , although it may still be possible to find special solutions .
associated to the ( weak ) painlev test is the global property called the ( weak ) painlev property .
an ode possesses the painlev property if the general solution can be made single - valued . for the weak painlev property
, it requires that the general solution be at most finitely branched around any movable singularity .
the tests themselves are by design sets of necessary conditions respectively for these properties . in this appendix
, we will show that none of the three odes pass the painlev test , and that also fails the weak painlev test while the other two pass . to begin
, we detail the test procedures on .
then we briefly comment on and simply point out where the tests fail for without dwelling on details .
the equation surely does not have the painlev property for the coefficient of the @xmath304 term clearly violates the necessary conditions for the painlev property @xcite ( see p.127 ) .
this is also confirmed by the test conclusion that has movable algebraic singularities
. _ step 1 _ ( _ dominant behaviours _ ) . assume the leading behaviour of a solution @xmath129 to be @xmath305 with @xmath306 not a positive integer .
substitute this form into and select out all possible lowest order terms as listed below @xmath307 since @xmath308 , we only have two possibilities . for @xmath309 , @xmath310 is the lowest order term and the vanishing of its coefficient requires @xmath311 given @xmath312 . for @xmath313 ,
the constant @xmath314 is the lowest order term , which does not vanish in general , hence not interesting for the purpose . to summarize ,
we obtain @xmath315 with arbitrary @xmath240 , i.e , that @xmath316 is the only detected dominant behaviour .
_ step 2 _ ( _ resonance conditions _ @xcite ( see p.87 ) ) . having found the dominant behaviour , now we consider the possibility to extend it to a puiseux series expansion @xmath317 this requires the determination of the locations @xmath318 , called _
fuchs indices _ or _ resonances _ , where arbitrary coefficients may enter the puiseux series .
consider the dominant terms @xmath319 of that contribute to the leading behaviour @xmath320 . then compute the derivative @xmath321 the fuchs indices satisfy the so - called _ indicial equation _
hence we obtain a fractional resonance at @xmath323 with @xmath324 .
_ step 3 ( compatibility conditions ) .
_ at @xmath324 , we know , from the the first step , that @xmath325 is indeed an arbitrary coefficient .
this completes the test . in conclusion , passes the weak painlev test .
* remark * note that no _ pole _ is detected from the test above .
the ode for @xmath326 still involves a puisuex series instead of a laurent series since the cubing does not eliminate all third roots of @xmath327 . according to @xcite ,
the presence of movable algebraic singularities is not incompatible with integrability . the very design of the weak painlev test limits its usage only as necessary conditions for the weak painlev property .
the test can neither detect movable ( branched ) essential singularities themselves nor exclude an accumulation of algebraic singularities forming a movable essential one that may be severely branched .
these possibilities make a rigorous proof of the weak painlev property not at all a trivial one , which by itself may deserve a specialized article to discuss .
see examples in @xcite . according to painlev @xcite ,
the only movable singularities of solutions to the first - order ode @xmath328 where @xmath329 is rational in @xmath74 with coefficients that are algebraic functions of @xmath330 , are poles and/or algebraic branch points .
in addition , the only nonlinear ode in this class that has the painlev property is the riccati equation which is certainly not .
hence the equation automatically has the weak painlev property , but not the painlev property , and it is free from movable essential singularities . the equation admits two families of dominant behaviours ( cf . and ):
@xmath331 it fails the weak painlev test for having _ irrational _ resonances .
this means that has an infinitely branched movable singularity , which is a strong indicator for non - integrability @xcite . since our attempt of solving
has not been successful , we decided to focus on and explore some of its features that may facilitate constructing new solutions .
as indicated by the weak painlev test , the ode for @xmath129 possesses a formal puiseux series solution @xmath332 ( j - j_0)^2 + \cdots \nonumber \\\end{aligned}\ ] ] with two arbitrary complex constants @xmath240 and @xmath333 .
in particular , this puiseux series solution contains a special case for @xmath334 ( @xmath12 ) such that @xmath335 this finite expression coincides with the known solution ( setting @xmath336 ) .
* theorem 5 * given that @xmath240 and @xmath251 , @xmath337 , the ode admits a formal puiseux series solution such that it converges in a neighborhood of @xmath333 . _
proof : _ the idea of the proof , following many standard proofs of painleve property , is to convert the puiseux series into a power series solution of a regular initial value problem ( e.g. , @xcite ) . first we define @xmath338 then differentiate it once with respect to @xmath258 and substitute @xmath339 using .
hence we obtain @xmath340 the system ( [ zeq1],[zeq2 ] ) is equivalent to the ode . now by introducing a new variable @xmath341 , we can transform the system into @xmath342 which has a unique power series solution about @xmath343 @xmath344 by the cauchy existence and uniqueness theorem , both series have non - vanishing radii of convergence . from the series , the corresponding solutions to ( [ zeq1],[zeq2 ] )
then take the form @xmath345^{2/3 } \!\!\!\ !
+ \sum_{k=1}^\infty u_k ( j - j_0)^{(k+2)/3 } , \nonumber \\ z
= z_0 + \sum_{k=0}^\infty v_k ( j - j_0)^{(k+2)/3}.\end{aligned}\ ] ] with @xmath346 .
this completes the proof .
the series clearly contains type n solutions that are not equivalent to leroy - nurowski s since they all continuously deform to the conformally flat solution in the limit @xmath347 , @xmath204 .
we already know that the latter has a non - constant cartan invariant @xmath146 given by .
the induction begins with @xmath348 which holds by the assumption .
now assume that for @xmath234 and @xmath349 , the bound is true . then for @xmath350 and @xmath351
, we can bound the product @xmath352 by @xmath353 \frac{c^2 m^{2k+2}}{(2l+2)^2 ( 2k-2l)^2 } \nonumber \\ & = & 2\left[\frac{1}{(2l+2)^2}+\frac{1}{(2k-2l)^2}\right ] \frac{c^2 m^{2k+2}}{(2k+2)^2}. \label{ind2}\end{aligned}\ ] ] the second inequality above is due to @xmath354 . rearranging and using the triangular inequality together with
, we obtain an upper bound for @xmath355 : @xmath356 where we define @xmath357 we can evaluate the first summation above in terms of the digamma function @xmath358 where @xmath359 is euler s constant , which is approximately @xmath360 .
the second summation has the following bound @xmath361 note that the trigamma function @xmath362 for all integers @xmath350 and that @xmath363 for @xmath364 .
combining these two bounds , for @xmath350 , we obtain @xmath365 where we use the fact that @xmath366 for all integers @xmath350 .
in addition , the first term in is bounded by @xmath367 where the last inequality becomes an equality for @xmath368 . altogether
, we obtain for @xmath234 @xmath369 \frac{c m^{2k+2}}{(2k+2)^2 } \leq \frac{c m^{2k+2}}{(2k+2)^2},\ ] ] given the assumption . here
@xmath370 is the kronecker delta .
this completes the induction .
100 herlt e 1986 reduktion der einsteinschen feldgleichungen fr type - n - vakuum - lsungen mit killingvektor und homothetischer gruppe auf eine gewhnliche differentialgleichung 3 .
ordnung fr eine reelle funktion _ wiss .
friedrich - schiller - univ .
jena , naturwiss .
r. _ * 35 * 73540 .
plebaski j f and robinson i 1976 left - degenerate vacuum metrics _ _ * 37 * 4935 + a good review with references is boyer c p , finley j d and plebaski j f 1980 complex general relativity , and spaces a survey of one approach _ general relativity and gravitation _
vol 2 ed held a ( new york : plenum ) pp 24181 krasilshchik i s and vinogradov a m ( eds ) 1999 _ symmetries and conservation laws for differential equations of mathematical physics _
( _ translations of mathematical monographs _
vol 182 ) ( providence : ams ) | in the search for vacuum solutions , with or without a cosmological constant , @xmath0 , of the einstein field equations of petrov type n with twisting principal null directions , the cr structures to describe the parameter space for a congruence of such null vectors provide a very useful tool .
a work of hill , lewandowski and nurowski has given a good foundation for this , reducing the field equations to a set of differential equations for two functions , one real , one complex , of three variables .
under the assumption of the existence of one killing vector , the ( infinite - dimensional ) classical symmetries of those equations are determined and group - invariant solutions are considered .
this results in a single ode of the third order which may easily be reduced to one of the second order .
a one - parameter class of power series solutions , @xmath1 , of this second - order equation is realized , holomorphic in a neighborhood of the origin and behaving asymptotically as a simple quadratic function plus lower - order terms for large values of @xmath2 , which constitutes new solutions of the twisting type n problem .
the solution found by leroy , and also by nurowski , is shown to be a special case in this class .
cartan s method for determining equivalence of cr manifolds is used to show that this class is indeed much more general .
in addition , for a special choice of a parameter , this ode may be integrated once , to provide a first - order abel equation .
it can also determine new solutions to the field equations although no general solution has yet been found for it . |
the canadian hydrogen intensity mapping experiment ( chime ) is a radio telescope currently being constructed at the dominion radio astrophysical observatory ( drao ) in penticton , bc , canada @xcite .
chime is a transit telescope with no moving parts , composed of four cylindrical reflecting surfaces , each 100 m in length north - south ( n - s ) and 20 m wide east - west ( e - w ) . on each of the four focal lines
is a linear array of 256 dual polarization antennas with a spacing of 0.3048 m .
this geometry provides an extremely wide effective fov of @xmath2120@xmath3 in n - s and [email protected]@xmath3 ( frequency dependent ) in e - w , that is , a primary beam size of @xmath0 square degrees .
chime is expected to be operational in 2017 .
apart from the original science goal of mapping redshifted 21-cm hydrogen emission to study the baryon acoustic oscillation signal at redshifts [email protected] @xcite , a custom backend is being deployed to conduct a commensal blind survey of frbs .
a recently discovered phenomenon @xcite , frbs appear to be a new class of radio transient with unknown astrophysical origin .
the current challenges in the frb field are mainly twofold : first , traditional radio telescopes employed for the study of point sources tend to have a limited fov .
this has meant low survey efficiency in the frb search effort thus far , and the small sample of only twenty or so known frbs has limited the precision with which frb rates can be estimated . secondly , only one frb has been observed to repeat @xcite .
it remains the only frb that has been localized by subsequent interferometric observations , which allowed for the identification of a host galaxy @xcite .
for all other frbs , the localization is only approximately as good as the full width at half - maximum ( fwhm ) of the beam of the initial single dish detection , typically of the order of a few square arcminutes .
this has made it impossible to identify multi - wavelength counterparts , which is crucial for determining the origin of frbs .
interferometers that are utilized to study time - domain astronomy are typically employed in a phased array mode , where targeted beams are formed by coherently summing antenna signals .
@xcite have shown that for an array of regularly spaced antennas as in chime , the interferometric phasing required for a complete basis of formed beams resembles the functional form of a fourier transform .
fft beamforming is computationally interesting for the context when the number of antennas @xmath5 is large , as it reduces the computational cost from @xmath6 to @xmath7 .
commercially - available graphics processing units ( gpus ) have now provided an affordable solution to the powerful computing necessary for this task .
we have implemented a real - time fft - beamforming pipeline ( see section [ sec : pipeline ] ) , which will generate an array of discrete but closely spaced , high - spectral resolution synthetic beams , allowing us to efficiently search the entire primary beam for short - duration transient signals . by searching these beams
, chime has the potential to discover tens of frbs per day @xcite , and is well positioned to revolutionize the field of frb research .
the chime correlator follows a hybrid fx design .
input voltage data are collected from 256 dual - polarization receivers on each of the 4 cylinders , i.e. , a total of 2048 distinct inputs .
custom fpga boards ( f - engine ) @xcite digitize and channelize these data to 1024 frequency channels via a 4-tap polyphase filter bank ( pfb ) , then scale this voltage data to 4 + 4-bit complex numbers , at a 2.56@xmath8s cadence ( see table [ tab : table ] ) .
spatial correlation ( x - engine ) is performed in a gpu cluster that consists of 256 processing nodes each with 4 amd fiji gpus , similar to that presented in @xcite .
this gpu cluster is also where the frb fan - beam pipeline takes place .
.the specifications of the input and output data of the chime / frb beamformer .
[ cols= " < , < , < " , ]
the frb fan - beam pipeline has been implemented as three opencl kernels .
the first kernel carries out a hybrid beamforming task . in the n - s direction , fft beamforming
is used to form 256 beams . in the e - w direction , exact phasing
is employed to form four beams .
the second kernel transposes the data from time
polarization
beams@xmath9beams@xmath10 , where time is the slowest varying index , to the order of polarization
beams@xmath9beams@xmath10time , where time becomes the fastest varying index as required by the next step .
a third kernel applies a fine channelization , increasing spectral resolution by 128 times to 3khz channels . as a final step ,
stokes i beams are formed by squaring the real and imaginary parts .
the output data are integrated to 16k frequency channels ( i.e. , at spectral resolution @xmath11khz ) at 0.983ms cadence and rescaled to 8-bit integers ( see table [ tab : table ] ) . a total output data rate of 131.072gbps , or 0.512gbps per node ,
is transmitted via 1gbe output data links in a custom udp packet format to the subsequent frb search , where dedispersion and triggering take place .
the current implementation of the end - to - end fan - beam pipeline utilizes 15% of the available gpu computational budget , within the planned allocation among other tasks of the correlator gpu cluster . in the following
we discuss some of the key elements of the frb fan - beam pipeline in further detail : in the n - s direction , chime has @xmath12 antennas in a linear and regular array , which makes spatial fft beamforming a sensible choice .
figure [ fig : chromatic ] shows that the nominal position of peak sensitivity of the fft - formed beams is a function of the observing frequency @xmath13 .
the formed beams at higher frequencies are closer together than those at lower frequencies .
hence , for a wide bandwidth instrument like chime , a straightforward implementation of fft beamforming will lead to chromatic smearing , where the formed beams across the 400mhz band line up poorly . instead , we employ an approximate method using a zero - padded fft . by extending the input data with a large number of zeros before taking the fft , additional redundant beams can be formed .
the output samples are then sufficiently dense that there will be a sample close to a given desired steering angle .
the more we increase the padding , the closer one can get to any arbitrary steering angle .
however , the computational cost also increases to roughly @xmath14 , where @xmath15 is the padding factor . taking both the computational cost and the beam sensitivity into account ,
a padding factor of @xmath162 has been chosen to create 512 beams , which will be subsampled by selecting the subset of beams at each frequency closest to the 256 desired pointings . in figure
[ fig : chromatic ] , the coloured dots enclosed by the gray areas are the selected nearest - neighbour clamped beams , whereas the fainter coloured dots are the additional and discarded beams .
as shown in @xcite , the clamping can be efficiently done by calculating a set of reference beam indices ( @xmath17 ) for a reference frequency @xmath13 using equation ( 1 ) , and then finding the adjustment to those indices ( @xmath18 ) at each observing frequencies ( @xmath19 ) using equation ( 2 ) . in both equations ,
the speed of light is denoted by @xmath20 and @xmath21 is the antenna separation .
these indices can be pre - computed and stored as look - up tables for the beamforming pipeline .
@xmath22 @xmath23 the frequency - summed beam pattern becomes progressively more elongated towards the horizon , with fwhm of @xmath24 at zenith and up to @xmath25 at @xmath26 elevation .
the extent of the n - s horizon coverage is tunable , and we have the flexibility to form denser or sparser beam tilings .
also note that the e - w formed beams can be configured as static or tracking .
the interstellar medium acts as a cold , ionized plasma , dispersing the broadband frb emission as it propagates towards the observer . dispersion across frequency channels can be corrected for by delaying the time series at higher frequencies by the appropriate amount to align with the signal arrival time at the lower frequencies .
intra - channel dispersion , however , can only be corrected for through coherent de - dispersion , which is computationally too expensive in a search operation .
if not accounted for , intra - channel dispersion will smear the observed pulse - like frb signal , limiting our sensitivity to frbs .
the beamformer input has only 1024 frequency channels across the 400-mhz band , i.e. , a spectral resolution of 390khz .
as shown in figure [ fig : smearing ] , this is far from ideal compare to the pulse widths of the known frbs .
the frb fan - beam processing , therefore , includes a second stage channelization , transforming every 128 time samples with an fft , to achieve a 3khz spectral resolution .
the improved frequency resolution is crucial for detecting narrow - width transient signals , hence maximizing our sensitivity towards frb events .
we are also investigating the feasibility of including a fourth kernel which will incoherently dedisperse the post - upchannelized data to a high nominal dm in order to boost our fiducial sensitivity to high - dm frb events .
spatial aliasing occurs when the physical spacing of the detectors equals half of the observing wavelength or shorter .
chime has a n - s antenna spacing of 0.3048 m , which means above an observing frequency of @xmath2490mhz we will inevitably have aliased signals .
the strength of an aliased signal could be of comparable magnitude to the actual source in a single frequency channel .
however , the location of the aliased beams are highly frequency dependent and hence in the frequency - summed response would appear ` smeared ' across a range of sky location relative to the main beam .
the higher the zenith angle , the wider the extent of smearing in the corresponding aliased beams .
figure [ fig : sim ] is the frequency - summed sky response of simulated chime data .
an injected signal is detected at the correct location near @xmath27 from zenith , whereas its aliased signals are smeared across roughly @xmath28 to @xmath29 .
this plot also confirms that the aliased artifact appears much weaker compared to the unaliased beam in the frequency - summed response data , which is what downstream frb search pipeline will be manipulating .
we have implemented a custom beamforming algorithm to the chime / frb backend in order to generate 1024 intensity beams at high spectral and time resolution .
this will enable 24/7 searching in chime s entire fov and potentially bring about the discovery of a large population of fast transients .
work is on - going to compare simulated data with response power from the fan - beam pipeline .
figure [ fig : sim ] shows one example of an injected signal re - detected from the gpu pipeline output .
further proposed extensions of the chime / frb beamforming project include beam interpolation , which would enable the fft - formed beams to be perfectly aligned across observing frequencies ( masui et al .
, in prep ) .
we are very grateful for the warm reception and skillful help we have received from the staff of the dominion radio astrophysical observatory , which is operated by the national research council of canada .
we acknowledge support from the canada foundation for innovation , the natural sciences and engineering research council of canada , the b.c . knowledge development fund , le cofinancement gouvernement du quebec - fci , the ontario research fund , the cifar cosmology and gravity program , the canada research chairs program , and the national research council of canada .
we thank xilinx university programs for their generous support of the chime project , and amd for donation of test units .
cn acknowledges support from the dunlap institute for astronomy & astrophysics at the university of toronto .
we thank mandana amiri , nolan denman , mark halpern , tom landecker , jeff peterson , ziggy pleunis , masoud ravandi and kris sigurdson for useful comments .
99 p. chawla , et al .
, `` a search for fast radio bursts with the gbncc pulsar survey , '' submitted to _ apj _ , 2017 , arxiv:1701.07457 .
the chime collaboration ( k. bandura , et al . ) , `` canadian hydrogen intensity mapping experiment ( chime ) pathfinder , '' _ spie astronomical telescopes+ instrumentation . international society for optics and photonics _ , 2014 , pp .
914522 - 914522 .
the chime collaboration ( k. bandura , et al . ) , `` ice - based custom full - mesh network for the chime high bandwidth radio astronomy correlator , '' _ j. astron .
_ , * 05 * , 2016 , 1641004 . the chime collaboration ( k. bandura , et al . ) ,
`` ice : a scalable , low - cost fpga - based telescope signal processing and networking system , '' _ j. astron .
_ , * 05 * , 2016 , 1641005 . l. connor , et al .
, `` constraints on the frb rate at 700 - 900 mhz , '' _ mnras _ , * 460 * , 2016 , pp . 1054 - 1058 .
n. denman , et al .
, `` a gpu - based correlator x - engine implemented on the chime pathfinder , '' _ press .
ieee asap _ , 2015 ,
p. klages , et al .
, `` data packing for high - speed 4-bit gpu correlators , '' _ press .
ieee asap _ , 2015 .
d. r. lorimer , et al .
, `` a bright millisecond radio burst of extragalactic origin , '' _ science _ , * 318 * , 2007 , 777 - 780 . b. maranda , `` efficient digital beamforming in the frequency domain , '' _ j. acoust
_ , * 86 * , 1989 , pp .
1813 - 1819 .
a. recnik , et al . , `` an efficient real - time data pipeline for the chime pathfinder radio telescope x - engine , '' _ press .
ieee asap _ , 2015 ,
l. g. spitler , et al .
, `` a repeating fast radio burst , '' _ nature _ , * 531 * , 2016 , pp . 202 - 205
. m. tegmark & m. zaldarriaga , `` the fast fourier transform telescope , '' _ phys .
d _ , * 79 * , 8 , 2008 . m. tegmark & m. zaldarriaga , `` omniscopes : large area telescope arrays with only @xmath5log@xmath5 computational cost , '' _ phys .
d _ , * 82 * , 10 , 2010 , p.103501 .
s. p. tendulkar , et al . , `` the host galaxy and redshift of the repeating fast radio burst frb 121102 , '' _ apjl _ , * 834 * , 2017 , l7 .
d. thornton , et al .
, `` a population of fast radio bursts at cosmological distances , '' _ science _ , * 341 * , 2013 , pp . | we have developed fft beamforming techniques for the chime radio telescope , to search for and localize the astrophysical signals from fast radio bursts ( frbs ) over a large instantaneous field - of - view ( fov ) while maintaining the full angular resolution of chime .
we implement a hybrid beamforming pipeline in a gpu correlator , synthesizing 256 fft - formed beams in the north - south direction by four formed beams along east - west via exact phasing , tiling a sky area of @xmath0 square degrees . a zero - padding approximation is employed to improve chromatic beam alignment across the wide bandwidth of 400 to 800mhz .
we up - channelize the data in order to achieve fine spectral resolution of @xmath124khz and time cadence of 0.983ms , desirable for detecting transient and dispersed signals such as those from frbs . |
the group - iii sesquioxides ga@xmath2o@xmath3 and in@xmath2o@xmath3 are currently popular in basic materials science and technology for being , respectively , deep - uv large - breakdown and near - uv transparent - conducting materials . a natural development that can be envisaged
is the growth of a solid solution ( ga@xmath0in@xmath1)@xmath2o@xmath3 , typically ( but not necessarily ) epitaxially on the parent compounds .
this would enable one to combine and tune the functionalities of the two parent compounds , and to export the band - engineering and nanostructuration concepts well known in other semiconductor systems ( such as arsenides and nitrides ) to much higher absorption energies and breakdown voltages .
progress in this directions requires a knowledge of the miscibility , as well as the behavior of relevant properties ( gap , specific volume , band offsets , etc . ) , of a solid - solution substitutional alloy composed , in a given proportion , of the two parent materials . in this paper
we report the modeling of ( ga@xmath0in@xmath1)@xmath2o@xmath3 over the entire range of @xmath5 by ( mostly ) ab initio density - functional - theory techniques .
previous results @xcite for the low-@xmath5 end of the composition range are integrated in a comprehensive picture of the miscibility and attendant properties .
the main result is that the alloy will phase - separate in a large composition range . on the low @xmath5 side ,
the favored phase is isostructural with @xmath4-ga@xmath2o@xmath3 ; at large @xmath5 , it has a bixbyite structure .
we also find that as function of @xmath5 the gap , volume , and band offsets to the parent compound exhibit discontinuities typical of a first - order phase transition as function of @xmath5 .
geometry and volume optimizations as well as electronic structure calculations are performed using density - functional theory ( dft ) in the generalized gradient approximation ( gga ) , and the projector augmented - wave ( paw ) method as implemented in the vasp code @xcite . for mixtures of insulators ,
it is almost invariably the case that the structure of the mixture is that ( or one of those ) of the parent compounds .
( this certainly does not generally apply to metal alloys . )
the stable phase for ga@xmath2o@xmath3 is monoclinic @xmath4 , for in@xmath2o@xmath3 it is bixbyite .
these are discussed at length in previous works ( @xcite and @xcite , respectively ) . while there are no competing structures for in@xmath2o@xmath3 , for ga@xmath2o@xmath3 the @xmath4 phase can be transformed under pressure into the @xmath6 phase
; thus we checked these latter phases , and the bixbyite structures .
we find that the @xmath4 phase is lowest in energy ; the bixbyite is 80 mev / formula unit higher , and the @xmath6 is 250 mev / formula unit higher .
we therefore neglect the latter and consider the @xmath4 and the bixbyite structures . for the @xmath4 structure we use , as in @xcite , supercells containing 1@xmath72@xmath72 unit cells ( 80 atoms ) and a 4@xmath74@xmath72 k - point grid . for the bixbyite structure , we use the cubic primitive cell ( 80 atoms ) , and a 4@xmath74@xmath74 k - point grid .
the plane wave cutoff is 300 ev , exceeding by 10% or more for all paw sets the maximum suggested value .
the structure of the @xmath4 phase , reported in @xcite , is in good agreement with experiment ; for bixbyite the lattice constant is 10.29 , or 1.5% larger than experiment as usual for gga calculations .
the bixbyite structure ( group t@xmath8 ) has six - fold coordinated cations ( large spheres ) occupying 8@xmath9 high - symmetry and 24@xmath10 wyckoff sites . in the left panel
we highlight the 8@xmath9 cation sites by surrounding them with the local octahedra with oxygen ( small spheres ) at their vertexes ; in the right panel we do the same with the 24@xmath10 lower - symmetry cation sites.,width=340 ] we simulate the compositional variation explicitly mixing in and ga cations , as dictated by the mole fraction @xmath5 of in . for low @xmath5 , we consider the monoclinic @xmath4-ga@xmath2o@xmath3 phase doped with in .
this alloy is free - standing , i.e. its energy is calculated at zero stress ; we have checked that epitaxial constraints on this phase do not change any of our conclusions , so we neglect them here for clarity .
as found previously @xcite , this phase is only relevant up to about @xmath5=0.25 .
we then study the bixbyite phase over the whole range of @xmath5 ; this is obtained naturally substituting ga for in in in@xmath2o@xmath3 , which is indeed a bixbyite as many other sesquioxides @xcite .
specifically , we study the bixbyite alloy in two settings : free - standing and in@xmath2o@xmath3-epitaxial . in the latter ,
the in - plane lattice parameters are fixed to that of in@xmath2o@xmath3 and the vertical lattice parameter and all internal coordinates are optimized .
both the @xmath4 and bixbyite supercells contains 80 atoms , i.e. 32 cations .
the choice of configurations in the @xmath4 phase has been discussed in ref.@xcite . for the bixbyite phase , we find that ga substitution is slightly favored at the high - symmetry cation site ( see fig.[fig1 ] ) .
we substitute ga for in on those sites first , and then on the lower symmetry ones .
for each @xmath5 , we sample a few configurations ; as ga s generally try avoid each other , the geometric constraints on the configurations are stringent , so there are not many possible inequivalent configurations to begin with .
our discussion is based on the lowest energy found at each @xmath5 ; of course , these may well not be the absolute energy minima for that @xmath5 .
also , we neglect the possible occurrence of higher - energy configurations in small proportions at finite temperature .
to address the occurrence of phase separation , we calculate the specific ( i.e. referred to one cation ) helmholtz free energy ( the enthalpy vanishes because the pressure is alway zero to numerical accuracy ) of the mixture as a function of @xmath5 . the internal energy is calculated directly from first principles using the supercells described above .
the entropy is modeled as the sum of mixing and vibrational terms .
the mixing entropy has the standard form @xmath11 since growth temperatures are comparable to or higher than debye temperatures ( @xmath12=420 k , @xmath13=870 k ) , the vibrational entropy per cation can be approximated as that of a single - frequency oscillator at the debye frequency .
thus @xmath14,\ ] ] where @xmath15 is the planck - bose distribution @xmath16 and the mixture s debye temperature @xmath17(@xmath5 ) is assumed to be an interpolation @xmath18 of that of the parent compounds .
we now recall that phase separation in a mixture occurs when the specific free energy is a negative - curvature function of an extensive parameter such as @xmath5 .
the values , say , @xmath19@xmath20@xmath21 , at which the curvature becomes negative and goes back to positive , respectively , delimit the phase separation region ; in general these bounds depend on temperature t , and the @xmath5 range they identify is the miscibility gap .
if the negative curvature region vanishes as t increases , i.e. @xmath19 and @xmath21 get to coincide , there is complete miscibility .
our results , as mentioned earlier and discussed below , suggest a large miscibility gap surviving up to above the melting temperatures of the parent compounds .
we discuss our results in terms of the _ mixing _ specific free energy , i.e we subtract out the bulk free energy @xmath22 which interpolates the values for two equilibrium bulk phases ( bixbyite and @xmath4 , respectively ) .
fig.[fig2 ] , left panel , compares the mixing free energies of the free - standing @xmath4 phase ( circles ) with that of the epitaxial bixbyite phase ( squares ) .
fig.[fig2 ] , right panel , compares the same quantities for the same @xmath4 phase ( circles ) with that of the free - standing bixbyite phase ( diamonds ) .
the temperatures considered are 800 k , a typical growth temperature , and 2000 k , near the melting temperature of the parent compounds .
the free energy is evidently upward - convex in a wide region of intermediate @xmath5 , indicating that a phase separation occurs . the borders of that region define the miscibility gap . mixing free energy vs @xmath5 at 810 k and 2000 k for @xmath4-phase vs epitaxial bixbyite ( left panel ) , and @xmath4-phase vs free - standing bixbyite ( right panel ) .
the phase separation region extends between the vertical dashed and dash - dotted lines.,title="fig:",width=226 ] mixing free energy vs @xmath5 at 810 k and 2000 k for @xmath4-phase vs epitaxial bixbyite ( left panel ) , and @xmath4-phase vs free - standing bixbyite ( right panel ) .
the phase separation region extends between the vertical dashed and dash - dotted lines.,title="fig:",width=226 ] on the low-@xmath5 end , the @xmath4 phase prevails in all cases , and the change in curvature occurs ( hence the phase separation region starts ) at about @[email protected] .
this confirms largely our previous estimate @xcite of 10% maximum in solubility , and experiments @xcite giving similar results . at high @xmath5
, the end of the miscibility gap region is estimated at @[email protected]@xmath240.55 for the epitaxial case ( left panel ) , subject to large uncertainties in locating the free - energy downturn from the epitaxial phase .
therefore the miscibility gap is approximately @xmath5@xmath25(0.15,0.55 ) for the epitaxial bixbyite and @xmath4 phase .
comparison with the growth and x - ray diffraction study by zhang _
@xcite suggest that this prediction is quite plausible , even accounting for their epitaxial conditions being different from those simulated .
most importantly , at 2000 k the borders of the phase separation region are about the same as at 800 k , i.e. the miscibility gap hardly changes ( it actually may widen slightly ) .
since the melting temperatures of the parent compounds are around 2200 k , we conclude that in the practical range of t the miscibility gap between the epitaxial and @xmath4 phases is @xmath5@xmath25(0.15,0.55 ) independent of t. we now come to the @xmath4 phase vs free - standing bixbyite competition ( fig.[fig2 ] , right panel ) .
a phase separation region exists here too , involving the structure change to the @xmath4 phase at low @xmath5 : the lower limit is again @xmath5=0.15 . the free - standing bixbyite phase is favored over the @xmath4 phase ( as well as over the epitaxial ) over the rest of the @xmath5 range , from @xmath5=0.2 or so onward .
however , its own free energy is upward - convex for most of the range ; this indicates a phase separation between ga@xmath2o@xmath3 and in@xmath2o@xmath3 _ within _ the bixbyite phase ; the change in curvature on the high @xmath5 side is approximately at @[email protected] .
therefore the overall miscibility gap is @xmath5@xmath25(0.15,0.85 ) in the free - standing case .
this is quite clearly the case at both 800 k and 2000 k. thus , as in the epitaxial case , we conclude that in the practical range of t the miscibility gap for free - standing bixbyite is @xmath5@xmath25(0.15,0.85 ) independently of t. an experimental determination of the ga@xmath2o@xmath3-in@xmath2o@xmath3 phase diagram @xcite suggests that indeed at the in end of the range the single - crystal stability region is quite marginal , being limited to @[email protected] or so .
not unexpectedly , the properties of the alloy exhibit anomalies as function of concentration due to the changes in phase and structure . in fig.[fig4 ] , left panel , we show the calculated optimized volume in the two free - standing bixbyite and @xmath4 phases , showing a clear volume discontinuity at any given @xmath5 .
the fundamental gap , shown in fig.[fig4 ] , right panel , also exhibits analogous interesting features .
the @xmath4 phase has a linear decrease in good agreement with pressure experiments @xcite .
the bixbyite gap is also linear at low @xmath5 , but picks up a significant bowing near @xmath5=1 . to correct for the semilocal density - functional error ,
we supplement the gga calculated gap with an ad hoc scissor "- like correction @xmath27 which brings the gga gap to the experimental value in ga@xmath2o@xmath3 and in@xmath2o@xmath3 @xcite ( incidentally , the correction reduces the bowing as obtained from gga eigenvalues ) . since the lowest gap is dipole - forbidden , to compare with the experimental optical onsets @xcite we estimate the position of the optical onset at all @xmath5 as the gga gap value ( corrected by eq.[delta ] ) plus the difference of optical onset and minimum gap in in@xmath2o@xmath3 ( 0.55 ev ) .
left panel : volume vs mole fraction for the free - standing @xmath4 and bixbyite phases .
right panel : fundamental gap in the same phases and interpolations vs @xmath5 ( quadratic for bixbyite ; linear at low @xmath5 for @xmath4 ) .
a correction for the gap error has been introduced ( see text ) .
the gap show a sizable bowing in bixbyite at large @xmath5.,title="fig:",width=238 ] left panel : volume vs mole fraction for the free - standing @xmath4 and bixbyite phases .
right panel : fundamental gap in the same phases and interpolations vs @xmath5 ( quadratic for bixbyite ; linear at low @xmath5 for @xmath4 ) .
a correction for the gap error has been introduced ( see text ) .
the gap show a sizable bowing in bixbyite at large @xmath5.,title="fig:",width=221 ] the agreement is decent , but should be taken with more than a grain of salt : on the experiment side , the data are for films grown on sapphire , the in content is generally lower than the nominal one especially at intermediate @xmath5 , and the gap in the @xmath5=0 and @xmath5=1 limits is larger than in most reports ( including our own @xcite ) ; on the theory side , we have applied a simple correction that offers no guarantee of being equally appropriate for all transitions and all @xmath5 . as we now discuss , an interesting crossover behavior is more easily observable in the band offsets at the interface with the parent compounds .
band offsets at interfaces are key quantities in the design and engineering of heterostructures . ab initio theory has been predicting reliable offsets all along ; a review of the theory concepts and of the techniques in common use is in ref.@xcite .
recently we predicted a staggered b - type offset at the ( 100 ) interface of @xmath4-phase ga@xmath2o@xmath3 with low-@xmath5 ( ga@xmath0in@xmath1)@xmath2o@xmath3 . here we extend the work to a much larger range in the bixbyite structure .
we simulate the ( 001)-like interface with in@xmath2o@xmath3/(ga@xmath0in@xmath1)@xmath2o@xmath3 superlattices epitaxially constrained to in@xmath2o@xmath3 , containing 160 atoms in the primitive cell , and with explicit atomic substitutions .
we also calculate the same quantities relaxing all lattice parameters of the superlattice ; this mimics a substrate that is locally compliant , i.e. deforms along with the film . for @xmath5 inside a phase separation region , the calculated values refer to the mixed phase , and not to the potentially compositionally ( or structurally ) segregated one ( see sec.[phsep ] ) .
valence ( vbo , top ) and conduction ( cbo , bottom ) interface band offsets between in@xmath2o@xmath3 and ( ga@xmath0in@xmath1)@xmath2o@xmath3 in the bixbyite phase , both epitaxially constrained on in@xmath3o@xmath3 , or with compliant substrate .
the offset between ga@xmath2o@xmath3 and low-@xmath5 ( ga@xmath0in@xmath1)@xmath2o@xmath3 at @[email protected] in the @xmath4 phase for the same conditions are also reported.,width=340 ] on the bixbyite side of the phase separation region , the offsets are again type - b ( a relatively uncommon occurrence in itself ) , but most interestingly they are staggered in the opposite direction , i.e. both the conduction and valence offsets encountered in going from the parent compound into the alloy are generally positive , whereas they were negative in the low-@xmath5 limit ( see the sketches in fig.[fig5 ] ) .
this suggests interesting perpectives for interface offset tuning in this alloy system .
the offset values are also rather large , and hence interesting in terms of potential charge confinement .
recalling the phase separation region discussed above , the most promising @xmath5 is around 0.7 in the epitaxial bixbyite .
we note that the gap error of density functional theory is immaterial here , as any gap corrections will largely cancel out of the offsets themselves .
we thus purposely refer to offsets only , starting from zero at @xmath5=0 and @xmath5=1 .
using ( mostly ) density - functional ab initio theoretical techniques , we have established that ( ga@xmath0in@xmath1)@xmath2o@xmath3 will exist in the @xmath4 phase at low @xmath5 and in the bixbyite phase at high @xmath5 .
for the epitaxial bixbyite case , the compound will phase - separate above 15% , and the two phases will coexist up to about 55% .
while the free - standing bixbyite s coexistence with the @xmath4 phase is limited to about @xmath5=0.25 , but bixbyite will itself separate into ga - rich and in - rich regions for @xmath5 up to about 0.85 .
thus the effective miscibility gap extends all the way from @xmath5=0.15 to @xmath5=0.85 .
importantly , both miscibility gaps are practically independent of temperature , and survive up to the melting temperature . the behavior of the calculated volume , gap ( in decent agreement with experiments for films ) , and band offsets also confirm the picture .
interestingly , we find that the interface band offsets are largely type - b staggered and positive at large @xmath5 , whereas they are staggered and negative in the low-@xmath5 limit .
work supported in part by miur - prin 2010 project _ oxide _ , car of university of cagliari , fondazione banco di sardegna grants , cineca computing grants .
mbm acknowledges the financial support of her phd scholarship by sardinia regional government under p.o.r .
sardegna f.s.e .
operational programme of the autonomous region of sardinia , european social fund 2007 - 2013 - axis iv human resources , objective l.3 , line of activity l.3.1 .
9 maccioni m b , ricci f , and fiorentini v 2015 _ _ a__ppl . phys .
express * 8 * 021102 .
maccioni m b , ricci f , and fiorentini v 2014 _ j. phys .
ser . _ * 566 * 012016 .
kresse g and furthmuller j 1996 _ phys .
b _ * 54 * 11169 .
marsella l and fiorentini v 2004 _ phys .
b _ * 69 * 172103 .
delugas p , fiorentini v , and filippetti a 2009 _ phys . rev .
b _ * 80 * 104301 .
baldini m , gogova d , irmscher k , schmidbauer m , wagner g , and fornari r 2014 _ cryst .
technol . _ * 49 * , 552 .
zhang f , saito k , tanaka t , nishio m , guo q 2014 _ solid state comm . _ * 186 * 28 .
edwards d d , mason t o , goutenoire f and poeppelmeier k r 1997 _ appl .
lett . _ * 70 * 1706 .
ricci f , boschi f , baraldi a , filippetti a , higashiwaki m , kuramata a , fiorentini v , and fornari r 2015 , submitted for publication . king p d c , veal t d , fuchs f , wang c y , payne d j , bourlange a , zhang h , bell g r , cimalla v , ambacher o , egdell r g , bechstedt f , and mcconville c f , 2009 b * 79 * 205211 .
peressi m , binggeli n , and baldereschi a , 1998 1273 . | using density - functional ab initio theoretical techniques , we study ( ga@xmath0in@xmath1)@xmath2o@xmath3 in both its equilibrium structures ( monoclinic @xmath4 and bixbyite ) and over the whole range of composition .
we establish that the alloy exhibits a large and temperature - independent miscibility gap . on the low-@xmath5 side ,
the favored phase is isostructural with @xmath4-ga@xmath2o@xmath3 ; on the high-@xmath5 side , it is isostructural with bixbyite in@xmath2o@xmath3 .
the miscibility gap opens between approximately 15% and 55% in content for the bixbyite alloy grown epitaxially on in@xmath2o@xmath3 , and 15% and 85% in content for the free - standing bixbyite alloy .
the gap , volume and band offsets to the parent compound also exhibit anomalies as function of @xmath5 .
specifically , the offsets in epitaxial conditions are predominantly type - b staggered , but have opposite signs in the two end - of - range phases . |
young star clusters are potentially bright , long lived , tracers of star formation within their host galaxies .
it is therefore of paramount importance to understand what governs the cluster formation process ( e.g. , * ? ? ?
* ) and how the host galactic environment may affect their formation and evolution ( e.g. , * ? ? ?
statistically we can describe the cluster formation as a stochastic process .
numerous works report of an initial cluster mass and luminosity function ( icmf and iclf , respectively ) being described with good approximation by a single power - law slope of index close to @xmath3 .
the stochastic nature of cluster formation combined with the power - law distributions of the cluster mass generate the so called size - of - sample effects .
these effects are well described by the positive relations between the number of clusters @xcite or the youngest brightest cluster in v band , m@xmath8 , versus the galactic star formation rate ( sfr ; * ? ? ?
* ; * ? ? ?
in other words , galaxies with higher sfr have more numerous cluster populations , therefore they are more likely to sample the icmf at higher mass ( luminosity ) ranges
. the size - of - sample effect also applies to the time - binning of the cluster population .
massive clusters are statistically more likely to be older because the galaxy has had a longer period of time to sample the high mass bins of the icmf @xcite .
increasing evidence , however , suggests that this is not the full picture .
a steepening at the bright luminosity bins of the iclf has widely been observed and reported .
this steepening could be explained by the presence of a truncation at the high mass end of the icmf .
indeed , the presence of a truncation mass may explain why we do not observe very massive young star clusters close to @xmath9 mass in the milky way @xcite .
however , it is difficult to establish the presence of a truncation due to low number statistics @xcite .
several works report an icmf compatible with a schechter function of slope @xmath3 , and an exponential cutoff at masses above a certain characteristic mass , .
interestingly it has been observed by these authors that may change as function of the host galactic environment .
such an environmental dependence would be expected theoretically due to the variation of the maximum mass scale for gravitational instability in galaxy discs ( the `` toomre mass '' , toomre 1964 ) , which should lead to an increase of the maximum cluster mass with the gas pressure ( kruijssen 2014 ) .
more evidence of the role played by the galactic environment on the properties of their cluster population has indirectly been suggested by the so called t@xmath10(u ) versus sfr relation @xcite .
t@xmath10(u ) is the fraction of u band light locked in star clusters with respect to the total u band light of the galaxy .
this fraction is observed to increase as function of the sfr of the galaxy , suggesting that the fraction of star formation happening in clusters ( hereafter or cluster formation efficiency , cfe ) is increasing @xcite .
indeed direct evidence of a varying cfe as function of the average galactic star formation rate density ( @xmath11 ) has been observed in numerous galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* among many others ) . in the model proposed by @xcite a variation in the cfe
is directly linked to the star and cluster formation process itself .
this model describes star and cluster formation continuously across the density spectrum of the hierarchically structured ism , in which the highest star formation efficiencies ( and thus bound stellar fractions ) are reached in the highest density peaks .
integration of these local physics over the gas density spectrum then results in a cfe that increases with the gas surface density or pressure , as well as the sfr density ( kruijssen 2012 ) .
given the density gradient observed in disc galaxies , the model thus predicts that the cfe decreases with the galactocentric radius .
indeed , using a small fraction of the whole cluster population of m83 , ( * ? ? ?
* hereafter sv13 ) report evidence of such a variation of the cfe as function of galactocentric distances .
the aim of this article is to probe the imprints that the galactic environment leaves on the properties of its cluster population .
the results will shed light on how the star cluster population is a natural outcome of the physics driving galactic star formation , and will also help to use star cluster populations as tracers of the sfh of their host galaxies . to achieve this goal we study the cluster population of the nearby ( @xmath12 mpc , corresponding to a distance modulus of 28.28 mag ) spiral galaxy m83 .
this galaxy hosts a rich cluster population which has been widely analysed in recent years .
several studies suggest that the current ongoing starburst at the centre of this galaxy ( e.g. , * ? ? ?
* ) was mostly likely produced by a recent minor merger @xcite . outside this region
the most active star - forming regions in the galaxy are at the end of the bar ( see figure [ fig : m83 ] ) .
however , we notice that star formation is occurring over the entire spiral arm system and in spurs extended inside the inter - arm regions , as in m51 @xcite . only recently
m83 has been fully observed with the hubble space telescope ( hst ) with broad and narrow imaging bands covering from the @xmath13 to the @xmath14 and several optical emission lines @xcite .
previous studies of the cluster population have therefore focused on the two pointings f1 and f2 ( see figure [ fig : m83 ] ) .
@xcite report of a similar icmf and cluster disruption rate in both inner ( f1 ) and outer ( f2 ) fields .
they notice a steepening of the icmf slope and a milder disruption rate in f2 , but they conclude that within the uncertainties the results support a cluster formation and evolution scenario independent of the galactic environment . with a similar cluster catalogue ( comparison presented in both @xcite ) , ( * ? ? ?
* ; * ? ? ?
* hereafter b12 ) point out the differences in the photometric properties , icmf and iclf , disruption rate of the cluster populations in the two pointings .
the differences can not be accounted for by a recent , significant change in the sfh , so are therefore most likely linked to the cluster formation and disruption properties of the local galactic environment .
such a change in cluster properties is also supported by the change in the cfe as function of distance from the centre of the galaxy observed by sv13 .
the access to a full cluster catalogue for m83 now allows us to verify and put stronger constraints on previous findings . using the whole m83 cluster catalogue , ( *
* hereafter sv14 ) find a clear environmental dependency on the strength of the cluster disruption with higher cluster disruption in the inner regions of the galaxy , as already suggested by b12 . in a recent publication ,
@xcite has studied the h@xmath15 morphology of the very young star clusters in m83 , reporting that clusters are already partially or fully exposed within a few myr after their formation .
ryon et al ( 2015 ) has focused on the analysis of m83 cluster sizes , important for the dynamical evolution of these systems . in the present work
we will expand the icmf analysis published by b12 and the analysis of the cfe by sv13 to the whole cluster population within the m83 disk .
all together , these works are presenting an unprecedentedly detailed and complete study of the cluster population of a nearby galaxy .
the paper is organised as follows .
section 2 describes the photometric analysis and the properties of the final cluster catalogue .
we also include a description of the ancillary data used in the analysis . in section 3
, we present our results .
section 4 contains a discussion of the results within the wider context of cluster and star formation in galaxies .
a summary of the paper is presented in the conclusions ( section 5 ) .
this analysis is based on the complete m83 cluster catalogue recently published in sv14 .
the final cluster catalogue has been obtained from two multi - band hst imaging datasets , ( go 11360 , pi : oconnell , and go 12513 , pi : blair ) .
the two combined datasets cover within 7 pointings the face - on m83 disk up to a distance of 4.5 kpc ( 3.4 ) from the centre of the galaxy ( see figure [ fig : m83 ] and table [ tab1 ] ) .
our analysis is based on wfc3 imaging in 5 optical bands , f336w ( hereafter @xmath16 band ) , f438w ( @xmath17 band ) , f555w or f547 m ( @xmath18 band ) , f657n ( narrow filter centred on h@xmath15 ) , and f814w ( hereafter @xmath19 band ) .
cluster detection has been performed on the v band with optimised parameter inputs of the source extraction algorithm sextractor . for each extracted source
, aperture photometry has been performed in all the bands with an aperture radius of 5 px ( 0.2@xmath20 4.4 pc ) and a sky annulus located at 8 px ( 0.32 ) and 2 px ( 0.08 ) wide .
isolated compact clusters , visually selected , were used to estimate and add the amount of missed flux lost because of the small aperture size used . moreover
small differences in calibration between different pointings were corrected as explained in sv14 .
the final catalogue retains only sources with a concentration index larger than 1.25 mag and positive detections in 4 filters ( @xmath21 ) .
we consider detection in h@xmath15 as additional but not required constraint .
the vega zero points used to flux calibrate the photometry have been retrieved from the wfc3 instrument webpage . milky way foreground extinction correction has been applied @xcite .
as described in sv14 , the initial catalogue has been visually inspected and each source has been assigned a flag describing the morphology of the cluster .
class 1 objects are clusters with symmetric and compact light distribution .
class 1 objects are most likely bound clusters as tested in b12 .
class 2 objects show asymmetries in the light distribution , multiple peaks , potentially contaminant neighbours .
these objects closely resemble ob associations , therefore there is a fair chance of them being unbound @xcite .
class 3 objects are spurious detections ( bad pixels mostly located at the end of the chip , background and foreground sources ) and have been removed from the final catalogue .
any cluster candidate located within a radius of 0.45 kpc has also been excluded from the catalogue .
we excluded this region because of the high luminosity gradient which causes a change in the completeness of the cluster detection and because of the different sfh that this region has experienced ( b12 ) .
the number of cluster candidates detected in each field is listed in table [ tab1 ] . + galaxy & m@xmath8 & sfr & @xmath11 & + & [ mag ] & m@xmath22yr@xmath23&m@xmath22yr@xmath23kpc@xmath24 & % + m82@xmath25 & -14.80 & 7.00 & & + ngc3597@xmath25 & -13.30 & 10.80 & & + iras19115@xmath25 & -16.80 & 192.00 & & + ngc1533@xmath25 ( a1 ) & -7.17 & 0.37@xmath26 & & + ngc1533@xmath25 ( a2 ) & -5.71 & 0.25@xmath26 & & + ngc1533@xmath25 ( a3 ) & -6.16 & 0.18@xmath26 & & + ngc2623@xmath25 & -14.50 & 51.00 & & + ngc3256@xmath25 & -15.70 & 46.00 & 0.62 & 22.9@xmath27@xmath28 + ngc7673@xmath25 & -14.70 & 4.90 & & + ngc6745@xmath25 & -15.00 & 12.20 & & + ngc1140@xmath25 & -14.80 & 0.80 & & + milky way@xmath28 & & 0.1508 & 0.012 & 7.0@xmath29 + ngc45@xmath30 & -10.83 & 0.12 & & + ngc45@xmath31 & & 0.05 & 1.02@xmath26 & [email protected] + ngc406@xmath30 & -11.75 & 0.29 & & + ngc628@xmath30 & -11.84 & 0.23 & & + ngc1300@xmath30 ( f1 ) & -11.00 & 0.30 & & + ngc1300@xmath30 ( f2 ) & -11.52 & 0.32 & & + ngc1309@xmath30 & -13.80 & 1.70 & & + ngc1313@xmath30 & -10.98 & 0.22 & & + ngc1313@xmath31 & & 0.68 & 0.011 & [email protected] + ngc1483@xmath30 & -10.01 & 0.11 & & + ngc3627@xmath30 & -11.97 & 0.37 & & + ngc4038@xmath30 & -15.25 & 2.43 & & + ngc4394@xmath30 & -10.25 & 0.14 & & + ngc4395@xmath30 & -9.79 & 0.07 & & + ngc4395@xmath31 & & 0.17 & 4.66@xmath26 & [email protected] + ngc4736@xmath30 & -10.44 & 0.04 & & + ngc5055@xmath30 & -9.61 & 0.02 & & + m101@xmath30 ( f1 ) & -11.38 & 0.25 & & + m101@xmath30 ( f2 ) & -11.57 & 0.20 & & + ngc6503@xmath30 & -10.51 & 0.07 & &
+ ngc7793@xmath30 & -9.65 & 0.07 & &
+ ngc7793@xmath31 & & 0.15 & 6.51@xmath26 & [email protected] + ngc4449@xmath33 & -12.60 & 1.0 & 0.04 & 9.0 + ngc1569 & -13.9@xmath34 & 0.36@xmath35 & 0.03@xmath35 & 13.9@xmath36 + ngc1705@xmath37 & -13.8 & 0.31 & 0.046 & + smc@xmath38 & -9.94 & 0.043 & 0.001 & 4.2@xmath39@xmath28 + he2 - 10@xmath40 & -12.5 & 0.20 & 0.20 & + ngc2366@xmath41 & -8.52 & 0.094 & 2.4@xmath26 & + ugc4305@xmath41 & -8.88 & 0.114 & 2.2@xmath26 & + ugc4459@xmath41 & -7.9 & 0.003 & 1.0@xmath26 & + ugc5336@xmath41 & -8.59 & 0.004 & 0.00037 & + ic2574@xmath41 & -9.12 & 0.11 & 1.2@xmath26 & + ugc5692@xmath41 & -9.44 & 7.34@xmath26 & 0.5@xmath26 & + ugca292@xmath41 & -7.46 & 5.15@xmath26 & 0.01 & + ugc8201@xmath41 & -9.16 & 31.6@xmath26 & 2.1@xmath26 & + ugc9128@xmath41 & -5.45 & 0.3@xmath26 & 0.3@xmath26 & + ugc9240@xmath41 & -5.90 & 5.24@xmath26 & 1.0@xmath26 & + dwarf sample@xmath42 & & & [ 0.01 , 10.0]@xmath26 & 5.0 + ic10@xmath43 & -10.4 & 0.07 & 0.03 & 4.2 + | we present a study of the m83 cluster population , covering the disc of the galaxy between radii of 0.45 and 4.5 kpc .
we aim to probe the properties of the cluster population as a function of distance from the galactic centre .
we observe a net decline in cluster formation efficiency ( , i.e. the amount of star formation happening in bound clusters ) from about 26% in the inner region to 8% in the outer part of the galaxy .
the recovered values within different regions of m 83 follow the same versus star formation rate density relation observed for entire galaxies .
we also probe the initial cluster mass function ( icmf ) as a function of galactocentric distance .
we observe a significant steepening of the icmf in the outer regions ( from @xmath0 to @xmath1 ) and for the whole galactic cluster population ( slope of @xmath2 ) of m83 .
we show that this change of slope reflects a more fundamental change of the truncation mass at the high - mass end of the distribution .
this can be modelled as a schechter function of slope @xmath3 with an exponential cut - off mass ( @xmath4 ) that decreases significantly from the inner to the outer regions ( from 4.00 to @xmath5 ) while the galactic is @xmath6 .
the trends in and icmf are consistent with the observed radial decrease of the @xmath7 , hence in gas pressure .
as gas pressure declines cluster formation becomes less efficient .
we conclude that the host galaxy environment appears to regulate 1 ) the fraction of stars locked in clusters ; 2 ) the upper mass limit of the icmf , consistently described by a near - universal slope @xmath3 truncated at the high - mass end . [ firstpage ] galaxies : spiral galaxies : star clusters galaxies : star formation star : formation |
quasiorders , also known as preorders , on a set @xmath3 form a complete lattice @xmath2 .
so do the transitive relations on @xmath3 ; their complete lattice is denoted by @xmath5 .
similarly , @xmath0 will stand for the lattice of all equivalences on @xmath3 .
the natural involution , which maps a relation @xmath6 to its inverse , @xmath7 , is an automorhpism of each of the three lattices mentioned above .
if , besides arbitrary joins and meets , the involution is an operation of the structure , then we speak of the _ complete involution lattices _ @xmath2 and @xmath5 .
however , it would not be worth considering the involution on @xmath0 , because it is the identity map . unless otherwise stated , _ generation _ is understood in complete sense .
that is , for a subset @xmath8 of @xmath0 , @xmath2 , or @xmath0 , we say that @xmath8 generates the complete ( involution ) lattice in question if the only complete sub lattice ( closed with respect to involution ) including @xmath8 is the whole lattice itself . for @xmath9
, we say that a complete lattice @xmath10 is _ @xmath11-generated _ if it can be generated by a @xmath11-element subset @xmath8 ; @xmath11-generated complete involution lattices are understood similarly .
since the involution commutes with infinitary lattice terms , we obtain easily that @xmath12 note that when dealing with _ finite
_ sets @xmath3 or finite lattices , then the adjective `` complete '' is superfluous ; this trivial fact will not be repeated all the time later .
if a complete lattice is generated by a four - element subset @xmath13 such that @xmath14 but both @xmath15 and @xmath16 are antichains , the we say that this lattice is _
@xmath17-generated_. we need also the concept of _ accessible cardinals_. a cardinal @xmath18 is _ accessible _ if it is finite , or it is infinite and for every @xmath19 , * either @xmath20 for some cardinal @xmath21 , * or there is a set @xmath22 of cardinals such that @xmath23 , @xmath24 , and @xmath21 for all @xmath25 .
since all sets in this paper will be assumed to be of accessible cardinalities , two remarks are appropriate here .
first , zfc has a model in which all cardinals are accessible ; see kuratowski @xcite .
second , we do not have any idea how approach the problem if @xmath1 is an inaccessible cardinal .
since a detailed historical survey has just been given in czdli @xcite , here we mention only few known facts . by strietz
@xcite and @xcite , zdori @xcite , and czdli @xcite , the complete lattice @xmath0 of all equivalences is four - generated , provided the size @xmath1 of @xmath3 is an accessible cardinal and @xmath26 .
also , we know from these papers that @xmath0 can not be generated by less than four elements if @xmath27 .
we know from chajda and czdli @xcite and takch @xcite that the complete _ involution _ lattice @xmath2 is three - generated for @xmath26 accessible , whereby we conclude from that @xmath2 is six - generated as a complete lattice .
actually , we know from dolgos @xcite for @xmath28 and from kulin @xcite for the rest of accessible cardinals that the complete lattice @xmath2 is five - generated .
furthermore , it was proved in czdli @xcite that the complete lattice @xmath2 is four - generated for @xmath29 .
it is also shown in @xcite that the complete lattice @xmath2 can not be generated by less than four elements , provided @xmath30 .
special variants of the above results , without considering the lattices @xmath0 and @xmath2 complete , were given in czdli @xcite and @xcite .
dolgos @xcite has recently shown that the complete lattice @xmath5 is eight - generated for @xmath28 . finally , we know from zdori @xcite , which improves strietz @xcite by reducing 10 to 7 , and from czdli @xcite that the complete lattice @xmath0 is @xmath17-generated provided @xmath31 and @xmath1 is an accessible cardinal .
although the proofs given here are short , sometimes very short , these proofs rely on nontrivial earlier constructions .
if someone wanted to replace the proofs of the theorems that are included in the rest of the present paper , then he would need to add several additional pages to each of these proofs ; typically , about 15 - 20 pages to the proofs dealing with all accessible cardinals .
as it is usual in lattice theory , we use `` @xmath32 '' to denote proper inclusion , which excludes equality .
[ lemmakulin ] if @xmath33 and @xmath34 is a complete sublattice of @xmath2 such that @xmath35 , then @xmath36 . for @xmath37 ,
let @xmath38^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}&:={\{{(a , a ) } , { ( a , b ) } , { ( b , b ) } , { ( b , a)}\}}\cr { \langle a , b \rangle^{\kern-1pt{\scriptscriptstyle\textup{q } } } } & : = { \{{(a , a ) } , { ( a , b ) } , { ( b , b)}\}},\text { and}\cr { \{\kern-3pt(a , b)\kern-3pt\}^{\kern-0.5pt{\scriptscriptstyle\textup{tr}}}}&:={\{{(a , b)}\ } } ; \end{aligned}\ ] ] they are the least equivalence , the least quasiorder , and the least transitive relation , respectively , containing the pair @xmath39 .
while @xmath40 is always an atom of @xmath5 and all atoms of @xmath5 are of this form , @xmath41^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}$ ] is an atom of @xmath0 iff @xmath42 is an atom of @xmath2 iff @xmath43 , and all atoms of @xmath0 and @xmath2 are of this form . by a _ zdori configuration _ of rank @xmath44 , we mean an edge - colored graph @xmath45 with @xmath46-colored _ horizontal edges _ @xmath47 and @xmath48 for @xmath49 and @xmath50 , @xmath51-colored _ vertical edges _
@xmath52 for @xmath53 , and @xmath54-colored _ slanted edges @xmath55of slope @xmath56 _ @xmath57 for @xmath58 ; these edges are _ solid _ edges in our figures . for example
, @xmath59 is given in figure [ fig1 ] but we have to disregard the dotted edges .
we do not make a notational distinction between the graph and its vertex set , @xmath60 .
the colors @xmath46 , @xmath51 , and @xmath54 are also members of @xmath61 ; we let @xmath62 if there is an @xmath46-colored path from @xmath63 to @xmath64 in the graph , and we define the equivalences @xmath65 analogously . the following lemma is due to zdori @xcite .
note that it is implicit in @xcite , and it was used , implicitly , in czdli @xcite , @xcite , @xcite , and @xcite .
the lattice operations join and meet are also denoted by @xmath66 and @xmath67 ( or concatenation ) , respectively .
[ lemmazadori ] if @xmath68 and @xmath3 is the base set of the zdori configuration @xmath60 , then @xmath0 is generated by @xmath69^{\kern-0.5pt{\scriptscriptstyle\textup{e } } } } , { [ a_n , b_{n-1}]^{\kern-0.5pt{\scriptscriptstyle\textup{e } } } } \}}$ ] .
the following straightforward lemma was also used , explicitly or implicitly , in several earlier papers ; see chajda and czdli ( * ? ? ?
* second display in page 423 ) , czdli ( * ? ? ?
* circle principle in page 12 ) , ( * ? ? ?
* last display in page 55 ) , ( * ? ? ?
* first display in page 451 ) , and ( * ? ? ?
* lemma 2.1 ) , kulin ( * ? ? ?
* lemma 2.2 ) , takch @xcite , and zdori ( * ? ? ?
* second display in page 583 ) .
[ circlemma ] for an arbitrary set @xmath3 and @xmath70 , if @xmath71 , @xmath72 and @xmath73 are pairwise disjoint subsets of @xmath3 , @xmath74 , and @xmath75 , then @xmath76^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}&= \bigl(\sum_{i=1}^j { [ x_{i-1},x_i]^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}\bigr ) \cdot \bigl(\sum_{i=1}^k { [ y_{i-1},y_i]^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}\bigr ) \text.\end{aligned}\ ] ] [ lemmatransa ] assume that @xmath77 are _ antisymmetric _ @xmath55 in other words , they are orderings@xmath78 and @xmath79 , , @xmath80 generates the complete involution lattice @xmath2 . then @xmath81 is a generating set of the complete involution lattice @xmath5 .
the same holds if we consider @xmath2 and @xmath5 complete lattices @xmath55without involution@xmath78 .
let @xmath82 stand for the complete involution lattice of all binary relations over @xmath3 .
the meet in this lattice is the usual intersection , the involution is the map @xmath83 , but the join is defined in the following way : for @xmath84 and @xmath85 , we have @xmath86 iff there is an @xmath44 , there exists a finite sequence @xmath87 of elements of @xmath88 and there are @xmath89 such that @xmath90 for all @xmath91 .
note that @xmath5 and @xmath2 are complete involution sublattices of @xmath82 . for a relation @xmath6 , denote @xmath92 by @xmath93 . instead of @xmath94 and @xmath95 , we write @xmath96 and @xmath97 , respectively .
we need @xmath11-ary @xmath1-complete involution lattice terms , which are defined in the usual way by transfinite induction , see , for example , @xcite ; these terms are built from at most @xmath1-ary joins and meets and the involution operation @xmath98 .
for such a term @xmath99 , @xmath100 and @xmath101 will stand for @xmath102 and @xmath103 .
then , for every @xmath11-ary @xmath1-complete involution lattice term @xmath99 , we have that @xmath104 if the rank of @xmath99 is 0 , then @xmath99 is a variable and holds obviously .
if holds for a term @xmath99 , then it also holds for @xmath105 , because @xmath98 is a lattice automorphism .
next , assume that @xmath106 and holds for all the @xmath107 .
then @xmath108 whereby holds for @xmath99 .
next , assume that @xmath109 . in order to show the validity of for @xmath99 ,
assume first that @xmath110 .
then @xmath111 and @xmath112 .
so there is _ shortest _ a finite sequence @xmath87 of elements of @xmath3 and there are @xmath89 such that @xmath113 for all @xmath91 . since @xmath111 and
we use a shortest sequence , @xmath44 is at least 1 and @xmath114 for @xmath91 .
thus , @xmath115 , whereby the induction hypothesis gives that @xmath116 .
therefore , @xmath117 . but
@xmath111 , whence @xmath118 .
this proves that @xmath119 .
conversely , since the lattice operations and the involution are monotone , @xmath120 . subtracting @xmath121 , we obtain that @xmath122 .
this proves .
armed with , let @xmath37 . since @xmath123 generates the complete involution lattice @xmath2 ,
there is a @xmath11-ary @xmath1-complete involution lattice term @xmath99 such that @xmath124 .
subtracting @xmath121 from both sides , we obtain that @xmath125 .
thus , by , @xmath126 .
this means that for all @xmath37 , the complete involution sublattice @xmath10 generated by @xmath127 in @xmath82 contains @xmath40 . but
@xmath10 is also what @xmath127 generates in @xmath5 .
thus , what we need to prove is that @xmath128 . for @xmath43 , @xmath129 .
based on this containment , for each @xmath130 , we can pick @xmath131 such that @xmath132 ; then @xmath133 finally , for an arbitrary @xmath134 , we obtain from @xmath135 that @xmath136 . consequently , @xmath128 is generated by @xmath127 as required .
the following lemmas will lead a theorem . for a set @xmath3
such that @xmath137 and @xmath1 is odd , @xmath2 is @xmath17-generated . with dotted @xmath138-edges , twice ]
take @xmath60 for @xmath139 from lemma [ lemmazadori ] , see figure [ fig1 ] .
define @xmath140^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}+{[b_0,b_{n-1}]^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}+{\langle a_2,a_4 \rangle^{\kern-1pt{\scriptscriptstyle\textup{q } } } } \in{\textup{quo}}(a ) ; \label{eqdhfhbezv}\ ] ] the join above is denoted by plus and it is taken in @xmath2 . note that makes sense since , say , @xmath141^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}\in{\textup{equ}}(a)\subseteq{\textup{quo}}(a)$ ] . in the figure , @xmath138 is visualized by the dotted lines .
let @xmath142\leq { \textup{quo}}(a)$ ] .
the @xmath143-block of @xmath144 is @xmath145 , see the black - filled elements on the left , whereby it follows easily that @xmath146^{\kern-0.5pt{\scriptscriptstyle\textup{e } } } } = \beta(\gamma+\delta)$ ] .
similarly , the @xmath147-block of @xmath144 consists of the black - filled elements on the right , and we conclude that @xmath148^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}=\gamma(\beta+\delta)$ ] . by lemma [ lemmazadori ] , @xmath149 . actually , @xmath150 , since @xmath151 .
thus , the statement follows from lemma [ lemmakulin ] .
let us agree that every infinite cardinal is even .
[ lemmatvnylc ] for @xmath152 , if @xmath1 is even , then the complete lattice @xmath2 is @xmath17-generated . ] for @xmath153 , define the graph @xmath154 in the same way ( but with a new notation ) as in czdli @xcite ; see figure [ fig2 ] for @xmath155 .
note that , for example , @xmath156 is a @xmath54-colored edge , no matter how large @xmath99 is .
let @xmath157 .
the dotted lines stand for @xmath138 again ; note that because of @xmath158 but @xmath159 , @xmath160 .
let @xmath142\leq { \textup{quo}}(a)$ ] . clearly , @xmath161 ranges in @xmath162 . for @xmath163
, we let @xmath164 as in @xcite . since the @xmath138-edge @xmath165 does not disturb anything in the proof given in @xcite , @xmath149 .
this inclusion , @xmath166 , and lemma [ lemmakulin ] yields the lemma . also , by the argument of @xcite , @xmath167 [ eqtxtghbuebg ] next , we formulate the `` large accessible '' counterpart of lemma [ lemmatvnylc ]
. if @xmath168 is accessible , then @xmath2 is @xmath17-generated . instead of @xmath169 in czdli (
* figure 1 ) , start with @xmath170 . instead of the four switches of @xmath169 ,
designate five switches in @xmath170 but use only four of them exactly in the same way as in @xcite . follow the construction of @xcite with @xmath170 instead of @xmath169 and , of course , not using the fifth switch .
this change does not disturb the argument , and we obtain a @xmath17-generating set of the complete lattice @xmath0 ; the only difference is that very many unused switches remains by the end of the construction .
now , we pick one of the unused switches and turn it to the , say , upper half of ( * ? ? ?
* figure 4 ) but in a slightly modified form : instead of the non - oriented dotted arc ( for @xmath138 ) , now we use an oriented arc . since this arc changes neither @xmath171 , nor @xmath172 , @xmath173 , we still have that @xmath174 $ ] .
this fact , @xmath160 and lemma [ lemmakulin ] complete the proof .
the following lemma adds @xmath175 , @xmath176 , and @xmath177 to the scope of the main result of czdli @xcite ; unfortunately , the case @xmath178 remains unsettled .
furthermore , it simplifies the approach of @xcite for finite sets @xmath3 with @xmath1 even .
[ htolkd ] for @xmath179 even , the @xmath55complete@xmath78 lattice @xmath2 is four - generated . for @xmath180 for @xmath181 , in accordance with our previous constructs and notation , take the _
one - point extension _ @xmath182 of @xmath60 ; see figure [ fig3 ] for @xmath183 .
let @xmath142 $ ] . also , let @xmath184 , and let @xmath185 .
for @xmath186 , let @xmath187 .
by czdli @xcite and @xmath188 , @xmath189 .
clearly , we have that @xmath190^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}=\beta({[a_0,a_n]^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}+\gamma)$ ] and @xmath191^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}=\gamma({[a_0,a_n]^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}+\beta)$ ] belong to @xmath10 .
hence , lemma [ circlemma ] gives that @xmath149 .
thus , lemma [ lemmakulin ] applies .
now , the conclusion of this section is summarized in the following theorem .
let @xmath3 be a non - singleton set with accessible cardinality .
then the following statements hold . * if @xmath192 , then the complete lattice @xmath2 is four - generated . * if @xmath193 and either @xmath1 is an odd number , or @xmath194 is even , then the complete lattice @xmath2 is @xmath17-generated . *
if @xmath195 and either @xmath1 is an odd number , or @xmath194 is even , then lattice @xmath2 @xmath55not a complete one now@xmath78 contains a @xmath17-generated sublattice that includes all atoms of @xmath2 .
[ lemmakjonway ] if @xmath196 and @xmath1 is an accessible cardinal , then the complete lattice @xmath5 is six - generated . by czdli
@xcite , there are @xmath197 such that @xmath198 .
let @xmath6 be a strict linear order on @xmath3 ; for example , it can be a well - ordering . in order to see that the complete sublattice @xmath199 $ ] is actually @xmath5 ; it suffices to show that @xmath10 contains all the atoms of @xmath5 .
take an atom ; it is of the form @xmath40 .
first , assume that @xmath43 .
then either @xmath6 , or @xmath200 contains the pair @xmath201 .
hence , @xmath40 is either @xmath41^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}\wedge\rho$ ] , or @xmath41^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}\wedge\rho^{-1}$ ] . in both cases , since @xmath41^{\kern-0.5pt{\scriptscriptstyle\textup{e}}}}\in { \textup{equ}}(a)=[\alpha_1,\dots,\alpha_4]\subseteq l$ ] , we obtain that @xmath129 .
second , assume that @xmath202 ; that is , we need to deal with @xmath203 .
the assumption @xmath33 allows us to pick @xmath131 such that @xmath204 . using with @xmath63 in place of @xmath205
, we obtain that @xmath206 , as required . note that this proof is more complicated than the proof of lemma [ lemmakjonway ] , because this proof uses lemma [ lemmatransa ] .
note also that and lemma [ lemmardhgt ] imply lemma [ lemmakjonway ] .
now , based on lemmas [ lemmakjonway ] and [ lemmardhgt ] , we are in the position to conclude this section and the paper with the following theorem .
zdori , l. ( 1986 ) generation of finite partition lattices . in : _ lectures in universal algebra .
szeged , 1983 ) _ colloq .
jnos bolyai , vol .
amsterdam : north - holland , pp . | by h. strietz , 1975 , and g. czdli , 1996 , the complete lattice @xmath0 of all equivalences is four - generated , provided the size @xmath1 is an accessible cardinal .
results of i. chajda and g. czdli , 1996 , g. takch , 1996 , t. dolgos , 2015 , and j. kulin , 2016 , show that both the lattice @xmath2 of all quasiorders on @xmath3 and , for @xmath4 , the lattice @xmath5 of all transitive relations on @xmath3 have small generating sets . based on complicated earlier constructions ,
we derive some new results in a concise but not self - contained way . |
photoacoustics has demonstrated optical contrast imaging in biological tissues at depth beyond 1 mm @xcite .
this non - invasive hybrid modality uses the conversion of transient illumination to ultrasound wave though thermoelastic expansion to detect optical absorption . besides the detection of endogenous photo - absorbers such as hemoglobin and melanin ,
the use of exogenous contrast agents has shown to provide functional and molecular information @xcite .
optical reporter agents in photoacoustics range from molecular agents to nanoparticles .
organic dyes like icg @xcite and methylene blue @xcite are of molecular agents and have already been used to enhance visualization of the circulatory system and its dynamic .
gold nanoparticles have been used in particular to image the enhanced permeability and retention effect in tumors @xcite or their long term biodistribution in small animals @xcite .
although relative variations of the amplitude of the photoacoustic signal could be sufficient if images ar taken at different time points before and after the injection , multispectral approaches are usually employed to improve the specificity of the detection .
the retrieval of spectral signature of contrast agents can be challenging . indeed ,
most of reconstruction algorithms @xcite and analytical derivation of the photoacoustic signal @xcite assume that the reconstructed initial pressure in the medium is directly proportional to the optical absorption and the fluence .
however , because of the wavelength dependent absorption of the excitation before reaching the region of interest , the relationship between the fluence on the sample surface and at the imaged position can be non - trivial .
specific detection by suppressing background signal has also been demonstrated using magnetic contrast agents for magnetomotive photoacoustic imaging @xcite . in this work ,
we consider photoacoustic nonlinearity as a candidate mechanism to discriminate between various type of optical absorbers .
several phenomena may induce non - linear relationships between the photoacoustic signal amplitude and the energy of the incident light , such as optical saturation @xcite or temperature - dependence of thermodynamic parameters @xcite .
nonlinear phenomena could be on their own a mean of selectively detect contrast agents , similarly as what is done in the field of ultrasound imaging . in the context of biomedical imaging ,
one recent experimental study reported a nonlinear signal increase with the laser fluence , presumably caused by thermal coupling within aggregated nanoparticles in cells @xcite . in the present study
, we compare experimental results to theoretical predictions regarding the occurence of thermal - based nonlinear photoacoustic phenomena with two different types of absorbers in aqueous solution , an organic dye and gold nanospheres .
the main objective of this study is to demonstrate the ability to discriminate between different types of absorbers based on thermal nonlinearity .
we first give an introduction to the so called thermal nonlinearity in photoacoustics , and discuss some predictions from numerical simulations .
experimental results from _ in vitro _
experiments are then qualitatively compared to theoretical predictions .
in a theoretical work published in 2001 @xcite , calasso _
et al_. analytically calculated a nonlinear contribution to the photoacoustic signal emitted by a point absorber . in this simplified model , the absorber immersed in a surrounding fluid was of vanishingly small size but with a finite optical absorption cross - section .
as a consequence , the photoacoustic generation was dictated by the thermodynamics properties ( speed of sound @xmath4 , density @xmath5 , specific heat capacity @xmath6 and coefficient of thermal expansion @xmath7 ) of the mere fluid , in particular via its grneisen coefficent @xmath8 . amongst the relevant thermodynamics properties , @xmath7 shows the most significant temperature dependency , illustrated on fig.[fig : fig1 ] .
the nonlinearity predicted in @xcite arised from the temperature - dependence of @xmath7 in the photoacoustic wave equation : @xmath9 p(\mathbf{r},t)= \rho
\beta(t ) \frac{\partial^2 t}{\partial t^2}(\mathbf{r},t)\ ] ] where @xmath10 is the temperature field in water . by linearizing the temperature - dependence of @xmath11 around the equilibrium temperature in the medium , according to the following equation @xmath12 calasso
et al_. separately calculated two terms in the expression of the photoacoustic signal : a linear term corresponding to the photoacoustic emission assuming that @xmath7 had not changed in response to the temperature rise induced by the optical absorption , and a nonlinear term attributed to the local temperature rise .
this so called thermal nonlinearity , earlier discussed in @xcite , results from the dynamic transient change of the coefficient of thermal expansion during the pulsed illumination . the nonlinear contribution may become significant only if temperature rises are high enough to affect the value of @xmath11 during the illumination .
for water , the linear contribution vanishes around @xmath13 , and therefore only the nonlinear contribution should remain , as experimentally observed from protons absorption experiments in water @xcite . as a function of temperature @xcite .
note that @xmath7 vanishes for @xmath14 c. ] in summary , thermal - based photoacoustic nonlinearity is expected to be significant in comparison to the linear contribution , and then accessible to experiments , either when the illumination fluence is large enough to induce a significant temperature rise , or when the equilibrium temperature approaches @xmath1 so that the linear contribution becomes negligible as compared to the nonlinear one .
the objective of the current work is to confront these qualitative predictions to experimental results performed with two types of absorbers , a solution of molecular organic dye and a solution of gold nanospheres .
the temperature rise around gold nanospheres in aqueous solution may become significant , due to the large absorption cross - section of gold nanoparticles and the partial heat confinement caused by the nanometric size . on the other hand
, negligible local temperature rise will be encountered with a dye solution of equivalent optical density : in this case , light is absorbed by a much larger number of dye molecules with as much lower absorption cross - sections , resulting in a very weak temperature rise at the scale of individual absorbers .
therefore , one expects thermal nonlinearity to possibly manifest itself with solution of gold nanoparticles , while dye solutions of equivalent optical density are expected to behave linearly .
the following section discusses orders of magnitude involved with nanosecond illumination and heating of gold nanospheres .
gold nanospheres with a diameter of a few tens of nanometers can not be reduced to point absorbers @xcite .
on the other hand , it has recently been demonstrated experimentally that for nanosecond illumination , it is mostly the fluid surrounding the nanoparticles that emits photoacoustic waves @xcite . from a theoretical point of view
, the prediction of the photoacoustic wave emitted by a gold nanosphere requires to first solve the diffusion equation for the temperature field in the sphere and its environment , and then to use this temperature field as a source term in the photoacoustic wave equation . because the heat diffusion caracteristic times for gold nanospheres are of the same order of magnitude as the illumination time , the problem is analytically intractable , even in the linear regime where @xmath7 is assumed to be constant .
moreover , taking into account the temperature - dependence of @xmath11 makes the problem even more difficult to solve by means of analytical methods .
et al_. could provide a solution only under the assumption of a point absorber and after linearization of the temperature dependence of @xmath11 @xcite . taking into account
the size of the absorber has been done in @xcite , but results could only be found for either thermally small or large absorbers , assumptions that do not hold for gold nanospheres .
numerical approaches are required to obtain accurate predictions of temperature rise and subsequent thermal - based photoacoustic nonlinearity .
it is out of the scope in this experimental report to describe the numerical methods that can be used to solve both the thermal and the photoacoustic problems .
suffice it to say that theoretical predictions from a numerical resolution , detailed elsewhere in a publication by our group @xcite , are in qualitative agreement with the main prediction that can be drawn from the work by calasso _
et al _ , namely that significant temperature rise around efficient optical absorbers yields a nonlinear relationship between the amplitude of photoacoustic signals and the light fluence .
figure [ fig : fig2 ] summarizes additional quantitative results from @xcite , which justify the experimental approach presented in this work .
this figure shows results obtained for a 40-nm diameter gold nanosphere immersed in water and illuminated with a 4 ns ( fwhm ) laser pulse .
first , fig.[fig : fig2](a ) predicts that a relatively low fluence ( typically a few mj/@xmath0 ) are needed to obtain a significant nonlinear dependence of the photoacoustic peak - to - peak amplitude with the incident fluence .
second , fig.[fig : fig2](b ) shows that the nonlinearity is predicted mostly around the centrer frequency corresponding to that of the temporal derivative of the laser pulse , i.e. approximately 100 mhz here .
this feature has apparently never been mentioned nor received any attention to the best of our knowledge .
it is however very important for experimental implementation as it suggests that a high frequency ultrasound detection should be used to measure thermal nonlinear effects in photoacoustics .
third , when one considers the evolution of the photoacoustic amplitude as a function of the equilibrium temperature , two different situations may be predicted , as shown in fig.[fig : fig2](c ) : in the linear regime ( small temperature rises , @xmath15 ) , the temperature dependence of the photoacoustic amplitude simply reflects the temperature dependence of the coefficient of thermal expansion , as is well known @xcite .
however , when large transient temperature rises are involved ( @xmath16 for instance in fig.[fig : fig2](c ) ) , i.e. when nonlinear contributions are significant , the temperature - dependence of the photoacoustic amplitude now reflects some effective value of the coefficient of thermal expansion .
in particular , a significant photoacoustic amplitude is predicted around @xmath17 .
an acoustic - resolution photoacoustic microscope was used to detect signals from a tube filled with absorbing aqueous solutions .
the experimental setup is illustrated in fig.[fig : fig3 ] .
nanosecond pulses from a q - switched laser ( brilliant , quantel , france ) -wavelength of 532 nm , 4 ns pulse duration , 10 hz repetition rate - , were directed towards the sample .
the beam was shaped by the combination of a lens ( f=75 mm , la1608 , thorlabs , usa ) and an engineered diffuser ( edc-10 , rpc photonics , usa ) so that the laser illumination was homogeneous on the tube in the focal region of the transducer .
the photoacoustic signals were detected with a 20 mhz spherically focused transducer ( 12.7 mm focal distance , 6.3 mm diameter , p120 - 2-r2.00 , olympus , japan ) , amplified ( dpr500 , remote pulser rp - l2 , jsr ultrasonics , usa ) and digitized with an oscilloscope ( dlm 2024 , yokogawa ) . a high frequency transducer was chosen to favor the observation of photoacoustic nonlinearities , as suggested by the theoretical predictions ( sec .
[ subsec : generationgns ] ) .
polycarbonate tubes with a 100 @xmath18 m inner diameter ( 200 @xmath18 m outer diameter , ctpc100 - 200 , paradigm optics , usa ) were used to approximately match the frequency of the emitted photoacoustic waves to that of the transducer .
the tubes were filled with aqueous solutions of either 40-nm diameter gold nanospheres ( hd.gc40.od10 , bbi solutions , uk ) or an organic blue dye ( colorant e133 , acidifier e330 , preservative e202 , scrapcooking , france ) .
the dilutions for each type of absorbers were set to obtain similar optical densities of about 10 ( @xmath19 0.1 ) at 532 nm .
532 nm as chosen as the excitation wavelength as it is close to the absorption peak of gold nanospheres with diameters on the order of a few tens of nanometers . for both samples
, the peak - to - peak amplitude of the photoacoustic signal was studied as a function of either the incident fluence or the equilibrium temperature .
the tubes were immersed in a large water tank providing a slow varying equilibrium temperature @xmath20 when the water temperature was different from the room temperature .
temperature was monitored using a thermocouple ( type k , rs components , uk ) placed in the vicinity of the tube , and connected to a computer via a thermocouple data loger ( tc-08 , pico technology , uk ) .
a magnetic stirrer was used to maintain a homogeneous temperature distribution in the water tank .
the tube was fixed on a beaker to prevent motion induced by water convection .
a hot air flow was used to prevent condensation to form and perturb the laser beam on the entrance wall of the water tank .
the incident fluence was controlled using a laser - integrated variable attenuator . for each position of the attenuator
, the fluence values on the sample were deduced from a prior calibration and direct measurements of the average laser output powers using a thermal power meter . from a sample to the other ,
only the position of the tube was adjusted so as to be in the focus of the transducer and within the area where the calibration for the fluence remained valid . to improve the signal - to - noise ratio ,
the signals were coherently averaged on the oscilloscope ( @xmath21 ) , before the peak - to - peak value computed on the scope was transferred to the computer .
the system noise was estimated from measurements with no illumination on the samples . for two equilibrium temperature @xmath22 and @xmath23 ,
the fluence was varied from 0 mj/@xmath0 up to approximately 7 mj/@xmath0 . for @xmath22 ,
water cooled in a freezer was used , the temperature of the tank remained around @xmath24 during the few minutes of the experiment over which the incident fluence was varied .
the experiments at @xmath23 were done at room temperature . to study the evolution of the photoacoustic signal as a function of @xmath20 from @xmath2 to @xmath3 ,
the incident fluence was fixed at 5 mj/@xmath0 , and the water tank initially filled with water from a freezer ( at @xmath25 ) .
as the temperature slowly raised towards room temperature , both the water temperature and the photoacoustic signals were simultaneously measured and stored in the computer every 30 seconds .
the same protocol was used for the two types of absorbing solutions ( dye and gold nanospheres ) . for both samples ,
each type of experiment was repeated at least 3 times , and turned out to be reproducible .
: both absorbers behave linearly . (
b ) results obtained for @xmath26 : no signal is detected from the dye sample , while a nonlinear signal is detected from the sample with gold nanospheres . ]
figure [ fig : fig4](a ) shows typical results obtained for @xmath23 .
no nonlinearity is observed for both types of absorbers as the peak - to - peak amplitude of the photoacoustic increased linearly with the fluence .
however , when the temperature was approximately @xmath22 , no photoacoustic signal could be detected from the dye solution , while a photoacoustic signal well above the noise floor could be observed from the solution of gold nanospheres .
moreover , the photoacoustic amplitude from the gold nanospheres shows a clear nonlinear dependence with the incident fluence . at @xmath23 , the fact that a linear behavior was observed as a function of fluence , as opposed to what is predicted from fig .
[ fig : fig2](a ) , is likely due to the fact that the detection bandwidth centered around 20 mhz is not high enough to observe a significant nonlinear contribution .
[ fig : fig2](a ) predicts the peak - to - peak evolution of the unfiltered signal , whereas in the experiment the photoacoustic signals are filtered by both the detection bandwidth and the spatial extension of the excited absorbing solution . on the other hand ,
the fact that a photoacoustic signal of significant amplitude is observed with gold nanosphere at @xmath26 , while under the same condition none is observed from the dye solution , can be explained by the dynamic change of @xmath11 around the nanosphere caused by the significant temperature rise . in this case , although the nonlinear generation is predominantly centered at much higher frequency than the detection frequency , it is the vanishing linear contribution at @xmath26 that makes the nonlinear contribution detectable . for @xmath23 ,
the nonlinear contribution is probably too small compared to the linear one .
, for an incident fluence of 5 mj/@xmath0 . ]
when the photoacoustic signal was measured as a function of the equilibrium temperature , as illustrated on fig.[fig : fig5 ] , for a fluence of 5 mj/@xmath0 ( at which a nonlinear contribution was seen at @xmath13 from the gold nanospheres ) , the experimental results agreed with the predictions from fig.[fig : fig2](c ) : the temperature dependence of the photoacoustic signal is significantly affected by the nanoscale local heating of the water surrounding the nanoparticles .
several conclusions can be drawn from these experimental results .
first , two absorbing media with the same optical density can be discriminated either from the dependence of photoacoustic amplitude with the incident fluence or from its dependence with the equilibrium temperature , if significant change of the coefficient of thermal expansion takes places during the laser illumination for one of them .
thermal based photoacoustic nonlinearity could be easily observed with gold nanospheres in water , with a detection frequency bandwidth centered around 20 mhz , at relatively low fluence provided that the temperature is close enough to @xmath1 so that the linear contribution is negligible .
the relevance of these results to biomedical applications requires further work but could be envisioned for imaging of biopsies . to date , only one experimental study reported the observation of presumably thermal based photoacoustic nonlinearity @xcite .
however , the required temperature rise was suggested to arise from the aggregation of nanoparticles within cells which led to some heat confinement , rather than from temperature rise around individual nanoparticules as considered in our work .
it may be hypothesized that one reason for the lack of studies reporting thermal - based photoacoustic nonlinearity in tissue with nanoparticles may be that most experiments used rather low frequency ( in the mhz range ) , in combination with the limitation on fluence required by safety considerations ( typically 20 mj/@xmath0 ) .
further work needs to be done to assess whether it can be possible to detect nonlinear contribution in tissue at physiological temperature ( @xmath27 ) by using a higher frequency ultrasound detection .
photoacoustic systems operating at frequencies up to 125 mhz have recently been reported @xcite .
for _ in vitro _ application , our results suggest that if biological tissue behaves as water , high frequency photoacoustic microscopy realized at a temperature close to @xmath28 may allow discriminating nanoparticles from endogenous absorption , a feature that may be interesting for experimental studies of particle uptake in tissue for instance .
our study also suggests that the temperature dependence of photoacoustic amplitude , a feature that has been proposed as a mean to monitor temperature changes in tissue during thermal therapies @xcite , may not only depend on the type of tissue , but if thermal nonlinearities are involved , may also depend on both the incident fluence and the type of absorbers used to photoacoustically probe the medium .
further works is also required to check the relevance of this to _ in vivo _ conditions .
in this work , we demonstrated that thermal - based photoacoustic nonlinearities can be measured using a 20 mhz detection frequency with 40-nm gold nanospheres in water at relatively low fluence .
the nonlinearity is observed either as a nonlinear dependence of the photoacoustic amplitude against the incident fluence , or as a fluence- and absorber- dependent relationship between the photoacoustic amplitude and the equilibrium temperature . as a consequence ,
two solutions of identical optical density were discriminated based on the effect of the nonlinearity .
the experimental observations agree qualitatively with theoretical predictions based on the temperature - dependence of the coefficient of thermal expansion in water .
the authors thank mickal mounaix for his assistance with the reproducibility experiments .
olivier simandoux gratefully acknowledges funding from the direction gnrale de larmement ( dga ) .
this work was funded by grants from the anr ( golden eye ) , the inca ( gold fever ) , and cnrs - psl*.
25 natexlab#1#1[2]#2 , , ( ) . , , , ( ) . , , , , , ( ) . , , , , , ( ) . , , , , , , ( ) . , , , , , , , ( ) . , , , , , ( ) . , , , ( )
, , , , , ( ) . , , , ( ) .
, , , , ( ) . , , , , , ( ) . , , ( ) .
, , , , , ( ) . , , , , ( ) .
, , , , ( ) . , , , , , ( ) . , , , , edition , .
, , , , , , , , , , , , , ( ) . , , , , , ( ) . , , , ( ) .
, , , , ( ) . , , , , , , , , , , ( ) . , , , , , ( ) . , , , , ( ) . | in the thermoelastic regime , photoacoustic sensing of optical absorption relies on conversion from light to acoustic energy via the coefficient of thermal expansion . in this work ,
we confront confront experimental measurements to theoretical predictions of nonlinear photoacoustic generation based on the dynamic variation of the coefficient of thermal expansion during the optical excitation of absorbers in aqueous solution .
the photoacoustic generation from solutions of organic dye and gold nanospheres ( with same optical densities ) , illuminated with 532 nm nanosecond pulses , was detected using a high frequency ultrasound transducer ( center frequency 20 mhz ) .
photoacoustic emission was observed with gold nanospheres at low fluence ( a few mj/@xmath0 ) for an equilibrium temperature around @xmath1 , where the linear photoacoustic effect in water vanishes , highlighting the nonlinear emission from the solution of nanospheres . under the same condition ,
no emission was observed with the absorbing organic dye . at a fixed fluence of 5 mj/@xmath0
, the photoacoustic amplitude was studied as a function of the equilibrium temperature from @xmath2 to @xmath3 . while the photoacoustic amplitude from the organic dye followed the coefficient of thermal expansion in water at the equilibrium temperature and vanished around @xmath1 , the photoacoustic amplitude from the gold nanospheres remained significant over the whole temperature range .
these experimental results are shown to be in qualitative agreement with theoretical calculations based on the nanoscale and nanosecond transient changes of the coefficient of thermal expansion of water around the gold nanospheres .
our results suggest that in the context of high frequency photoacoustic imaging , nanoparticles may be discriminated from molecular absorbers based on either the nonlinear fluence - dependence or the equilibrium temperature - dependence of the photoacoustic amplitude . |
the basic paradigm of many body physics is to use analytical and numerical tools to investigate the low - lying states and in particular the ground state properties of a system governed by a given hamiltonian . at very low temperatures , when thermal fluctuations are dominated by quantum fluctuations , quantum phase transitions can occur due to the change of character of the ground state @xcite .
what is exactly meant by `` character '' has been investigated in numerous works @xcite , @xcite , @xcite , especially in recent years after the discovery of exact measures @xcite of entanglement(purely quantum mechanical correlations ) .
for example , it has been a matter of debate whether a quantum phase transition is always accompanied by a divergence of some property in entanglement of the ground state wave function .
+ unfortunately , except for a few exactly solvable examples , the task of finding the exact ground state of a given hamiltonian is notoriously difficult . as always in dealing with difficult problems ,
one way round the difficulty is to investigate the inverse problem , that is to start from states with pre - determined properties and investigate quantum phase transitions which occur by smoothly changing some continuous parameters of these states .
the suitable formalism for following this path is the matrix product formalism @xcite , @xcite , @xcite which in recent years has been followed in constructing various models of interacting spins @xcite . in this paper
we want to take one step in this direction and in particular , we want to construct as concrete models , ladders of spin one - half particles and see what happens to the entanglement between various spins when the system undergoes a phase transition .
we construct general class of models having a number of discrete and continuous symmetries .
these are symmetry under the exchange of legs of the ladder , symmetry under spin flip , and symmetry under parity ( left - right reflection of the ladder ) in addition to a continuous symmetry , namely rotation of spins around the @xmath1 axis or all three axes ( full rotational symmetry ) .
+ in these models we calculate the entanglement of one rung with the rest of the ladder as measured by the von neumann entropy of the state of the rung and the entanglement of the two spins of a rung with each other as be measured by the concurrence of the same state .
+ we will see that in all these models quantum phase transitions occur at a critical point of the coupling constant , and this point is where a rung of the ladder becomes completely disentangled from the rest of the ladder and the two spins of the rung become fully entangled with each other .
the derivatives of these two types of entanglement are also divergent .
+ we should stress here that we are using the term phase transition in a wider sense than usual @xcite , that is we call any discontinuity in an observable quantity ( i.e. a two point correlation function ) , a phase transition .
+ the structure of this paper is as follows . in section [ mps ] ,
we review the formalism of matrix product states ( mps ) with emphasis on the symmetry properties of such states . in section [ ladder ] , we specify the mps construction to ladders of spin @xmath0 particles and set the general ground for construction of concrete models . in this same section
we construct multi - parameter families of models which have specific symmetries . in section [ stateproperties ]
, we study in detail the properties of the constructed states and in particular calculate the exact correlation functions of spins on the rungs .
we also investigate the connection between quantum phase transitions and divergence of entanglement properties in these models .
finally we derive in the appendix , the hamiltonian which governs the interaction of these models for which the state we have constructed is the exact ground state .
first let us review the basics of matrix product states @xcite . consider a homogeneous ring of @xmath2 sites , where each site describes a @xmath3level state .
the hilbert space of each site is spanned by the basis vectors @xmath4 .
a state @xmath5 is called a matrix product state if there exists @xmath6 dimensional complex matrices @xmath7 such that @xmath8 where @xmath9 is a normalization constant given by @xmath10 and @xmath11 here we are restricting ourselves to translationally invariant states , by taking the matrices to be site - independent . for open boundary conditions , any state can be mps - represented , provided that we allow site - dependent matrices @xmath12 , where @xmath13 denotes the position of the site @xcite .
the mps representation ( [ mat ] ) is not unique and a transformation such as @xmath14 leaves the state invariant . in view of this
we can find the conditions on the matrices which impose discrete symmetries on the state .
the state @xmath15 is reflection symmetric if there exists a matrix @xmath16 such that @xmath17 where @xmath18 is the transpose of @xmath19 and time - reversal invariant if there exist a matrix @xmath20 such that @xmath21 .
+ let @xmath22 be any local operator acting on a single site . then we can obtain the one - point function on site @xmath13 of the chain @xmath23 as follows : @xmath24 where @xmath25 in the thermodynamic limit @xmath26 , equation ( [ 1point ] ) gives @xmath27 where we have used the translation invariance of the model and @xmath28 is the eigenvalue of @xmath29 with the largest absolute value and @xmath30 and @xmath31 are the right and left eigenvectors corresponding to this eigenvalue , normalized such that @xmath32 . here
we are assuming that the largest eigenvalue of @xmath29 is non - degenerate .
+ the n - point functions can be obtained in a similar way .
for example , the two - point function @xmath33 can be obtained as @xmath34 where @xmath35 .
note that this is a formal notation which allows us to write the n - point functions in a uniform way , it does not require that @xmath29 is an invertible matrix . in the thermodynamic limit
the two point function turns out to be @xmath36 for large distances @xmath37 , this formula reduces to @xmath38 where @xmath39 is the second largest eigenvalue of @xmath29 for which the matrix element @xmath40 is non - vanishing and we have assumed that the eigenvectors of @xmath29 have been normalized , i.e. @xmath41 . thus the correlation length is given by @xmath42 any level crossing in the largest eigenvalue of the matrix @xmath29 signals a possible quantum phase transition . also , due to ( [ corr ] ) , any level crossing in the second largest eigenvalue of @xmath29 implies the correlation length of the system has undergone a discontinuous change .
here we are using a broader definition of quantum phase transition , that is we call any non - analytical behavior of a macroscopic property , a quantum phase transition @xcite .
of course in the models that we construct we observe a more direct change of observable physical properties , namely in one regime we have correlations between spin operators on different sites and in the other we have no such correlation .
consider now a local symmetry operator @xmath43 acting on a site as @xmath44 where summation convention is being used .
@xmath43 is a @xmath45 dimensional unitary representation of the symmetry . a global symmetry operator @xmath46
will then change this state to another matrix product state @xmath47 where @xmath48 a sufficient but not necessary condition for the state @xmath15 to be invariant under this symmetry is that there exist an operator @xmath49 such that @xmath50 thus @xmath43 and @xmath49 are two unitary representations of the symmetry , respectively of dimensions @xmath45 and @xmath6 . in case that @xmath43 is a continuous symmetry with generators
@xmath51 , equation ( [ symm ] ) , leads to @xmath52,\ ] ] where @xmath51 and @xmath53 are the @xmath3 and @xmath54dimensional representations of the lie algebra of the symmetry
. equations ( [ symm ] ) and ( [ symmalg ] ) will be our guiding lines in defining states with prescribed symmetries . given a matrix product state , the reduced density matrix of @xmath13 consecutive sites is given by @xmath55 the null space of this reduced density matrix includes the solutions of the following system of equations @xmath56 given that the matrices @xmath57 are of size @xmath58 , there are @xmath59 equations with @xmath60 unknowns .
since there can be at most @xmath59 independent equations , there are at least @xmath61 solutions for this system of equations .
thus for the density matrix of @xmath13 sites to have a null space it is sufficient that the following inequality holds @xmath62 let the null space of the reduced density matrix be spanned by the orthogonal vectors @xmath63 .
then we can construct the local hamiltonian acting on @xmath13 consecutive sites as @xmath64 where @xmath65 s are positive constants .
these parameters together with the parameters of the vectors @xmath66 inherited from those of the original matrices @xmath57 , determine the total number of coupling constants of the hamiltonian . if we call the embedding of this local hamiltonian into the sites @xmath67 to @xmath68 by @xmath69 then the full hamiltonian on the chain is written as @xmath70 the state @xmath15 is then a ground state of this hamiltonian with vanishing energy .
the reason is as follows : @xmath71 where @xmath72 is the reduced density matrix of sites @xmath67 to @xmath68 and in the last line we have used the fact that @xmath73 is constructed from the null eigenvectors of @xmath74 for @xmath13 consecutive sites . given that @xmath75 is a positive operator , this proves the assertion .
we now specify the above generalities to a ladder of spin @xmath0 particles .
the ladder consists of @xmath2 rungs and obeys periodic boundary conditions . to each rung of the ladder
we associate four matrices @xmath76 and @xmath77 respectively pertaining to the local states @xmath78 and @xmath79 . here
we are using the qubit notation which corresponds to the spin notation as @xmath80 and @xmath81 the first and the second indices refer respectively to the states of sites on the legs 1 and 2 as shown in figure ( 1 ) .
the rungs are numbered from 1 to n with a periodic boundary condition .
the operator @xmath82 refers to the pauli operator @xmath83 on leg number @xmath84 and the rung number @xmath13 .
the total spin operator on a rung at site @xmath13 is denoted by @xmath85 .
let us label the legs of the ladder by @xmath86 and @xmath87 as in figure ( 1 ) . in each @xmath88 the first and
the second indices refer respectively to legs number 1 and 2
. refers to the pauli operator @xmath83 on rung @xmath13 and leg number @xmath89 .
, width=302,height=94 ] [ ladder ] any operator corresponding to leg number @xmath84 is designated with a superscript @xmath89 . with these conventions ,
one can easily use ( [ mpsop ] ) and write down the mps operator corresponding to an observable .
for example , for the magnetization in the @xmath90 and @xmath1 directions in rung 1 , we have respectively @xmath91 and @xmath92 for the total magnetization in the @xmath1 direction in a rung we have @xmath93 other mps operators can be obtained in a similar way . in this paper
we restrict the dimensions of our matrices to @xmath94 , and demand that our models have time reversal symmetry so that the matrices @xmath88 are chosen to be real . we will be looking for models which have rotational symmetry in the @xmath95 plane of spin space .
thus we require that there be a matrix @xmath96 such that @xmath97=a_{00 } , \ \ \ \ [ { \cal t}_z , a_{11}]=-a_{11 } , \ \ \ \ [ { \cal t}_z , a_{01}]=[{\cal t}_z , a_{10}]=0.\ ] ] it is an easy exercise to show that the solution of these equations is @xmath98 where @xmath99 is found to be @xmath100 and we have excluded the solutions with @xmath101 or @xmath102 , which lead to trivial uncorrelated states .
hereafter we use the freedom in re - scaling the matrices ( without changing the matrix product state ) to set the parameter @xmath103 .
let us now consider extra discrete symmetries in addition to the above continuous symmetry .
these are as follows : + * a : spin flip symmetry * represented by a matrix @xmath104 , such that @xmath105 where @xmath106 and @xmath107 means @xmath108 .
this symmetry imposes the following condition on the solution ( [ tzsolution ] ) @xmath109 where @xmath104 is found to be @xmath110 . *
b : symmetry under the exchange of legs of the ladder * represented by a matrix @xmath111 such that @xmath112 where @xmath113 .
this symmetry imposes the following condition on ( [ tzsolution ] ) @xmath114 where @xmath111 is found to be @xmath115 .
* c : parity * or a left - right symmetry , represented by a matrix @xmath116 , such that @xmath117 with @xmath118 . + this symmetry imposes the condition @xmath119 on ( [ tzsolution ] ) where @xmath116 is found to be @xmath120 .
+ thus when we impose any of these discrete symmetries we are left with two three - parameter families of models , each family being distinguished by the discrete parameter @xmath121 , @xmath122 or @xmath123 . +
it is now readily seen that imposing any two of these symmetries makes the model symmetric under the third one too . a model which has all three symmetries
is defined by the following set of matrices : @xmath124 such matrices satisfy ( [ y ] ) with @xmath125 .
thus equation [ z23solution ] defines four two - parameter families of models on spin ladders which have @xmath126 symmetry in addition to three types of @xmath127 symmetries .
the families are distinguished by the pair of discrete parameters @xmath128 . + * d : full rotational symmetry * let us now see if we can construct models which have full @xmath129 symmetry , symmetry under rotations in the spin space . to this order we note that to have full rotational symmetry , the matrices defined by @xmath130 and @xmath131 should respectively transform like the spin @xmath86 and spin @xmath132 representations of the @xmath133 algebra , that is , we should have @xmath134=0 , \h a = x,\ y,\ z,\ ] ] and @xmath135=m \ b_{l , m } , \h [ { \cal t}_{\pm } , b_{l , m}]= \sqrt{2-m(m\pm 1)}b_{l , m\pm 1 } , \h l=1 , \ m=-1,0,1,\ ] ] where @xmath136 form the two dimensional representation of the @xmath133 algebra .
it is well known that the matrices @xmath137 transform like a vector under the adjoint action of @xmath129 .
the matrix @xmath138 should also be a multiple of identity .
thus we should set @xmath139 , and satisfy the following equations @xmath140 this puts the constraints @xmath141 with the unique solution @xmath142 where @xmath143 is an arbitrary real parameter .
this will give us a one - parameter family of models with full rotational symmetry and spin - flip symmetry @xmath144 * remark : * note that comparison of these parameters with the constraints ( [ xsolution ] , [ ysolution ] ) and ( [ pisolution ] ) shows that full rotational symmetry is compatible with spin - flip symmetry for arbitrary values of the parameter @xmath143 and compatible with parity or leg - exchange symmetries only for @xmath145 .
+ the model ( [ rotsolution ] ) has already been studied in @xcite . to see the correspondence with that work , we can collect the above matrices in a vector - valued matrix @xmath146 defined as @xmath147 . in view of our notation for spins @xmath148 and the notation of @xcite in which the single and the triplet states are denoted respectively by @xmath149 and @xmath150 )
becomes @xmath151 which modulo an overall constant is identical to the matrix given in @xcite .
we now study the properties of the states constructed above . for simplicity
we consider in detail only two general classes .
the first class is defined by ( [ z23solution ] ) and has @xmath126 symmetry in addition to three @xmath127 symmetries , and the second class is defined by ( [ rotsolution ] ) which has full rotational symmetry in addition to one @xmath127 symmetry , the spin flip symmetry . as mentioned in the previous section , full rotational symmetry
is compatible only with spin flip symmetry for generic values of the parameter @xmath143 and is compatible with the other two symmetries only when @xmath145 . in order to derive results which can be specialized to the two classes mentioned above we study in detail the properties of the state , defined by the equation ( [ xsolution ] ) .
the matrices are now given by @xmath152 the matrix @xmath29 for this class has the following form @xmath153 with eigenvalues @xmath154 for @xmath155 , the largest eigenvalue is @xmath156 and for @xmath157 it is @xmath158 .
hence the point @xmath101 is a point of phase transition .
the right and left eigenvectors of @xmath29 are simply obtained and one can determine all the relevant quantities of the ground state in closed form , in straightforward way after some rather lengthy calculations .
the reduced one - rung density matrix is obtained from ( [ rho ] ) : @xmath159 where @xmath160 .
this matrix can be rewritten as @xmath161,\ ] ] or @xmath162,\ ] ] where we have used the notation introduced in equation ( [ polisha ] ) .
+ from this density matrix one can obtain a lot of information about the observables pertaining to a single rung .
for example it is readily seen that the average magnetization at each single site and hence each single rung is zero , i.e. @xmath163 where @xmath164 is the spin of a rung , implying an anti - ferromagnetic state in which every single site is in a completely mixed states .
it is also seen that @xmath165 where @xmath166 is any unit vector in the @xmath95 plane . + defining the total spin of a single rung as @xmath167
, we find from the above result that @xmath168 thus for @xmath169 , each rung will be in a mixture of spin one states , but for @xmath101 , the spin one and spin zero multiplets can mix , depending on the value of @xmath170 and @xmath171 and @xmath121 .
the entanglement of this rung with the rest of the lattice is measured by the von - neumann entropy of this state , defined as @xmath172 . from the eigenvalues of the one - rung density matrix ,
@xmath173 we readily find @xmath174.\ ] ] as @xmath169 @xmath175 .
therefore in this limit each rung is still entangled with the rest of the ladder .
this means that a rung is not in a pure state and the state of the ladder is not a product of single rung states .
+ the entanglement of the two spins of a single rung with each other is measured by the concurrence @xcite of this state which for the density matrix ( [ rhosolution ] ) is given by @xmath176 where @xmath177 is the largest eigenvalue of the matrix @xmath74 .
a careful analysis of the eigenvalues shows that @xmath178 as @xmath169 , @xmath179 which means that although each single rung is not a pure state , it is a separable state .
in fact these can also be verified directly by looking at the one - rung density matrix in this limit . from ( [ rhosolutionketbra ] )
we have @xmath180 finally we calculate the correlation functions of different components of spins of the rung as a function of the distance between the rungs .
we find from ( [ 2pointthermodynamiclimit ] ) that @xmath181 with a longitudinal correlation length @xmath182 and @xmath183 with a transverse correlation length @xmath184 it is seen that the longitudinal correlation length depends on a single parameter @xmath185 as @xmath186 and the transverse correlation length depends on @xmath90 and another parameter @xmath187 as @xmath188 .
figure ( 2 ) shows the behavior of these correlation functions for different values of the parameters @xmath90 and @xmath189 .
[ correlationlengths ] , as a function of the parameter @xmath90 for different values of the parameter @xmath190 . at @xmath191
both the transverse and longitudinal correlation lengths diverge at @xmath192.,title="fig:",width=302,height=226 ] an interesting point about transverse correlation functions is that depending on the sign of @xmath121 , it becomes identically zero , on one side of the @xmath193 axis and different from zero on the other side .
thus if one insists on a definition of an order parameter to be zero in one phase and non - zero in the other , then we can safely say that in these models , the transverse correlation function is an order parameter which signals a quantum phase transition .
+ let us see study some limiting cases in these general models .
at the point of phase transition @xmath101 , as seen from ( [ rhosolutionketbra ] ) , the state of a rung becomes a mixture of spin - zero states , and the two spins of a rung become fully anti - correlated in the @xmath1 direction , a fact which is reflected in ( [ correlationsolution ] ) . near this point
the correlation length @xmath194 becomes very large as seen from ( [ longitudinallength ] ) , although the amplitude becomes small since it is proportional to @xmath195 .
thus at @xmath101 there is no long - range order in the model . also from ( [ concurrence ] ) and ( [ eigenvalues of rho ] )
it is readily seen that at @xmath101 , the concurrence ( or entanglement ) of the two spins of a rung becomes maximum and equal to @xmath196 on the other hand from ( [ entropysolution ] ) we see that as we approach the point @xmath101 from both sides , the von - neumann entropy decreases .
in addition , the derivative of both types of entanglement become singular at @xmath101 .
these facts show that the point of phase transition in these systems , is a point where the spins of a rung become highly entangled with each other and each rung becomes only slightly entangled to the rest of the lattice .
+ we can also obtain the explicit form of the state in this limit . from ( [ properties ] ) , and ( [ mat ] ) we see that in this limit , no two spins in a rung can be in a @xmath197 state .
a little reflection shows that they can not be in the state @xmath198 either , since in the absence of @xmath199 , the only string of matrices with non - zero trace is a string of matrices @xmath200 and @xmath201 in arbitrary order .
since these matrices commute with each other , the resulting state has a simple description .
define two local states of a rung as @xmath202 where for convenience we have re - introduced the spin notations @xmath203 and @xmath204 instead of qubit notation @xmath132 and @xmath86 .
define a global un - normalized state @xmath205 to be the equally weighted linear combination of all states with @xmath13 local @xmath143 states and @xmath206 local @xmath45 state . then from ( [ properties ] ) and ( [ mat ] ) we find that the ground state of the chain in the limit @xmath101 is given by @xmath207|u^k , d^{n - k}\ra,\ ] ] where @xmath9 is the normalization constant given by @xmath208 ^ 2 = 2\left[(a^2+b^2)^n+(2ab)^n\right].\ ] ] at the other extreme when @xmath209 , the state of each rung becomes a mixture of fully aligned spins either in the positive or negative @xmath1 direction .
this is also reflected in ( [ correlationsolution ] ) . in this limit
a rung becomes entangled with the rest of the lattice , sine @xmath210 and the two spins of a rung become disentangled from each other since @xmath211 .
+ the explicit form of the state can also be obtained in this limit . in this case
the only strings of matrices with non - vanishing traces are strings of @xmath199 and @xmath77 in alternating order .
thus if we define two local states @xmath212 then the ground state in this limit will be a ghz state @xmath213 in the next subsection we specialize these results to the two classes we discussed in the beginning of this subsection .
this is the class which has the so(2 ) symmetry ( rotation around the @xmath1 axis in spin space ) and three @xmath127 symmetries ( spin flip , parity , leg exchange ) . for this class
we have from ( [ pisolution ] ) that @xmath214 .
the parameter @xmath90 defined after equation ( [ transverselength ] ) becomes equal to @xmath215 insertion of @xmath214 in various quantities of the previous subsection shows that all the quantities can be expressed as a function of @xmath90 , namely we find @xmath216 @xmath217 and @xmath218 we also find @xmath219 and @xmath220 figure ( 2 ) show the entropy and the concurrence of the state of a single rung for this model .
[ finalladdercurvesabnew ] , which is taken to be dimensionless.,title="fig:",width=302,height=226 ] this is the class which has full so(3 ) symmetry ( rotation in spin space ) and one discrete symmetry ( spin flip ) for generic values of @xmath143 and all the three @xmath127 symmetries for @xmath145 . for this class
we have from ( [ rotsolution ] ) @xmath221 , @xmath222 and @xmath223 . inserting these values in the equations of the previous section we find the following : @xmath224 and @xmath225 where @xmath166 is any direction in the spin space .
the eigenvalues of the one - rung density matrix ( [ eigenvalues of rho ] ) in this case will be @xmath226 thus we find @xmath227 and @xmath228 figure ( 3 ) show the entropy and the concurrence of the state of a single rung for this model .
[ ucurves ] , which is taken to be dimensionless.,title="fig:",width=302,height=226 ]
although ladders of spin 3/2 models has already been studied in the context of vertex state models in @xcite , in which the ground state is constructed by a suitable concatenation of vertices assigned to single sites , the method developed in this work , in which each single rung is considered as a single site in a hyper - chain and the ground state is constructed as a matrix product state seems to us as more powerful and very easy to generalize to other spin models . in this work
we have applied the matrix product formalism for construction of models on spin @xmath0 ladders .
these models have been constructed to have special discrete and or continuous symmetries and to display a quantum phase transition in a broad sense , that is displaying non - analytical behavior in their correlation functions .
naturally these non - analytical behavior can also be observed in the entanglement properties of pairs of spins in these models .
this route can be followed to develop other models , e.g models with higher spins or alternating spins on the rungs , with alternating coupling constants , or next - nearest neighbor interactions . by exploiting higher dimensional matrices and having more free parameters at our disposal
, we may be able to construct continuous families of models on spin ladders which have full rotational symmetry . in this article
we have restricted ourselves to frustration - free hamiltonians and mps states with fixed - size . relaxing this latter condition
allows one to represent any state @xmath229 as a matrix product state @xcite and then one may be able to study more diverse kinds of phase transitions on spin ladders .
these matters will be taken up in separate publications .
we would like to thank a. langari for very valuable discussions and the members of the quantum information group of sharif university , specially s. alipour and l. memarzadeh for instructive comments .
s. sachdev , _ quantum phase transitions _
( cambridge university press , cambridge , 1999 ) .
a. osterloh , l. amico , g. falci , and r. fazio , nature 416 , 608 ( 2002 ) .
t. osborne , m. nielsen , phys .
rev . a , * 66 * , 032110 ( 2002 ) .
m. c. arnesen , s. bose , and v. vedral , phys .
lett . * 87 * , 277901 ( 2001 ) ; d. gunlycke et al , phys .
rev . a * 64 * , 042302 ( 2001 ) .
s. hill and w. k. wootters , phys .
lett . * 78 * , 5022 ( 1997 ) ; w. k. wootters , phys . rev
. letts . * 80 * , 2245 ( 1998 ) .
a. klumper , a. schadschneider and j. zittartz , j. phys .
a ( 1991 ) l293 ; z. phys .
b , * 87 * ( 1992 ) 281 ; europhys .
* 24 * ( 1993 ) 293 .
m. fannes , b. nachtergaele and r. f. werner , commun .
phys . * 144 * , 443 ( 1992 ) . m. m. wolf , g. ortiz , f. verstraete and i. cirac , phys .
97 , 110403 ( 2006 ) .
d. peres garcia , et al , quant - ph/0608197 .
m. a. ahrens , a. schadschneider , and j. zittartz , europhys .
lett . * 59 * 6 , 889 ( 2002 ) .
e. bartel , a. schadschneider and j. zittartz , eur . phys . jour .
b , * 31 * , 2 , 209 - 216 ( 2003 ) .
h. niggemann , and j. zittartz , j. phys .
a : math . gen .
31 , 9819 - 9828 ( 1998 ) .
f. verstraete , d. porras , j. i. cirac , phys .
93 * , 227205 ( 2004 ) .
f. verstraete , j. j. garcia - ripoll , and j. i. cirac , phys .
lett . * 93 * , 207204 ( 2004 ) .
f. verstraete , j. i. cirac , phys . rev .
b 73 , 094423 ( 2006 ) ; t. j. osborne , phys .
97 , 157202 ( 2006 ) ; m. b. hastings , phys . rev .
b 73 , 085115 ( 2006 ) .
g. vidal , phys .
lett . * 91 * , 147902 ( 2003 ) .
a. k. kolezhuk and h. j. mikeska , phys .
* 80 * , 2709 ( 1998 ) ; int .
b , vol.12 , 2325 - 2348 ( 1998 ) .
in this appendix we briefly discuss the derivation of the explicit form of the hamiltonian in terms of local spin operators . for simplicity
we consider models in class a. the hamiltonian for models in class b , can be constructed along similar lines . as explained in the main text , the starting point is to solve the system of equations @xmath230 this solution space is 12 dimensional since we have 4 equations for 16 unknowns .
thus we should find a set of 12 orthogonal vectors which span this solution space and then form a non - negative linear combination of the corresponding one dimensional projectors .
the hamiltonian constructed in this way does not necessarily have the symmetries imposed on the state , unless we choose new linear combinations of these vectors which transform suitably under the symmetry operators .
these new basis vectors are found to be : @xmath233 \cr |2,0\ra & : = & \frac{1}{\sqrt{6}}\left[-{4a^2}(|0011\ra+|1100\ra ) + g\left(|0101\ra + \s\e |0110\ra+ \s\e|1001\ra + |1010\ra\right)\right]\cr & & \cr |1,1\ra & : = & \frac{1}{2}\left[\e|0001\ra + |0010\ra - \e|0100\ra - |1000\ra \right ] \cr |1,0\ra & : = & \frac{1}{\sqrt{2}}\left[|0110\ra-|1001\ra\right ] \cr & & \cr |1',1\ra & : = & \frac{1}{2}\left[-\s\e|0001\ra + |0010\ra + \e |0100\ra -\s |1000\ra\right ] \cr |1',0\ra & : = & \frac{1}{\sqrt{2}}\left[|0101\ra-|1010\ra\right ] \cr & & \cr |0,0\ra & : = & \frac{1}{2}\left[|0101\ra -\s\e |0110\ra - \s\e |1001\ra + |1010\ra\right ] , \end{aligned}\ ] ] with @xmath234 + here we have organized the states in multiplets , with a labeling reminiscent of the one used in labeling the states of @xmath235 representations .
the reason is that in a certain limit ( @xmath236 ) these states actually comprise the irreducible representations of @xmath235 which emerge from the decomposition of the product of four spin one - half representations living on the sites of two rungs of the ladder .
in fact for the product of these spin 1/2 representations we have @xmath237 however only the representations @xmath238 and @xmath132 span the solution space of ( [ equations ] ) in this limit .
this labeling is useful because we can see under what conditions , the hamiltonian become fully rotational invariant . note also that we have used the labeling @xmath232 in accordance with the labeling of ( [ equations ] ) for @xmath231 ( see figure ( 5 ) .
the reader can easily check that each of the above states ( or more precisely the corresponding one dimensional projector ) is invariant under the parity operation @xmath239 and leg - exchange of the ladder @xmath240 .
they also transform to each other under the spin - flip operation @xmath241 . the local hamiltonian which has the three discrete symmetries mentioned in the text and the symmetry around rotation along the @xmath1 axis in spin space ,
is constructed as follows : @xmath242 where the coefficients @xmath243 are non - negative parameters and together with the parameters @xmath170 and @xmath244 form the 10 free parameters of the hamiltonian .
of course the total number of coupling constants ( interaction strength ) is 8 , since we can always shift the ground state energy and also set the scale of energy by redefintion of these parameters .
re - expressing the local operators in terms of pauli spin operators and rearranging terms , we find , after a rather lengthy calculation , the hamiltonian acting on the ladder where we have multiplied @xmath73 by a factor of @xmath245 for convenience , ( see figure ( 6 ) for labeling of sites of the ladder ) . note that we use @xmath246 as an abbreviation to denote the operator @xmath247 in the following .
the total hamiltonian is then given by @xmath253 this is a complicated looking hamiltonian , with many types of interactions , but in view of the large number of parameters , it is possible to look at specific subsets of the parameter space , where some of the interactions are absent .
indeed the parameter space consists of four disconnected parts , each of which corresponds to one choice of the pair @xmath254 .
let us for example consider the subset on which the hamiltonian has full rotational symmetry .
it is well known that any operator of the form @xmath255 is a scalar .
thus in the limit @xmath256 if we set @xmath257 and @xmath258 , then we will have a hamiltonian which is fully rotational invariant .
on can see that in this limit , all the coupling constant corresponding to non - scalar terms in the hamiltonian vanish and we are left with @xmath259 | we investigate quantum phase transitions in ladders of spin @xmath0 particles by engineering suitable matrix product states for these ladders .
we take into account both discrete and continuous symmetries and provide general classes of such models .
we also study the behavior of entanglement between different neighboring sites near the transition point and show that quantum phase transitions in these systems are accompanied by divergences in derivatives of entanglement . |
we consider the * grassmannian * @xmath0 . for a partition @xmath1 contained in a @xmath2 box ,
consider the path from the northeast corner to the southwest corner of the box that traces the partition .
for the standard flag ( @xmath3 ) , we define the * schubert variety * as @xmath4 we denote the * schubert class * in cohomology as @xmath5_h \in h^\star(gr_k { \mathbb{c}}^n)$ ] . the set @xmath6 forms a basis for @xmath7 , where @xmath8 for @xmath9 , and @xmath10 the littlewood - richardson coefficients .
this follows from the surjective homomorphism @xmath11 @xmath12 @xmath13 for schur functions @xmath14 .
the * mbius function @xmath15 * is defined recursively on a poset @xmath16 as the unique function satisfying @xmath17 the connection of this definition to k - classes is shown in @xcite .
since we are primarily interested in working in k - theory , we will use @xmath18 $ ] to denote the k - class of a subscheme of @xmath19 , and @xmath18_h$ ] to denote its homology class .
any subvariety @xmath20 of a flag manifold is rationally equivalent to a linear combination of schubert cycles with uniquely determined non - negative integer coefficients @xcite .
we say @xmath20 is * multiplicity - free * if these coefficients are 0 or 1 .
@xcite [ thm : mobiusbrion ] let @xmath20 be a multiplicity - free irreducible subvariety of @xmath21 , in the sense of @xcite , with @xmath22_h = \sum_{d\in d } [ x_d]_h$ ] .
let @xmath23 be the set of schubert varieties contained in @xmath24 ( an order ideal in the bruhat order on @xmath25 ) .
then as an element of @xmath26 , @xmath27 = \sum_{x_e \subseteq \bigcup_{d\in d } x_d } \mu_{\mathcal p}(x_e)\ [ x_e].\ ] ] we will give an independent combinatorial proof of this fact in the case that @xmath20 is a multiplicity - free * richardson variety * in a grassmannian , the intersection of a schubert variety @xmath28 with an opposite schubert variety @xmath29 , for @xmath30 the longest word . for any @xmath31 , let @xmath32 $ ] .
we have that @xmath33 form a basis for @xmath34 . for certain symmetric polynomials
@xmath35 which we will define in the next section , we have a surjective homomorphism @xcite : @xmath36 @xmath37 @xmath38 our main theorem will show that , for a poset @xmath16 that we will define , @xmath39 where the sum is over @xmath40 such that @xmath41 box and @xmath42 . our proof will proceed with sign - reversing involutions on this poset , and many reductions in the sizes of the partitions in the product . for finite non - empty sets in @xmath43 , @xmath44 and @xmath45
, we say @xmath46 if @xmath47 , and @xmath48 if @xmath49 . for a partition @xmath1 ,
buch defined a * set - valued tableau ( svt ) * as a filling of a young diagram with nonempty sets in @xmath43 @xcite .
if each box has a single entry , it is a * young tableau*. a tableau is a * semistandard tableau ( ss ) * if it is weakly increasing across rows and strictly increasing down columns .
the * superstandard * filling of a tableau is the one in which each box @xmath50 has a single entry , @xmath51 ( its row ) . in all of our examples
, we will use numbers smaller than 10 , so we can avoid the use of set notation : we use 45 to denote the set @xmath52 . + recall the combinatorial definition for the schur polynomials , @xmath53 we consider the * grothendieck polynomials * of lascoux and schtzenberger @xcite . for @xmath1 a partition ,
buch @xcite gives the formula @xmath54 where @xmath55
he proves that this is a special case of the lascoux - schtzenberger formula ( which we will not need ) for @xmath56 in the case when @xmath57 is a grassmannian permutation .
these are the @xmath58 representing the @xmath59 in the last section . as with the schur polynomials
, it is not obvious from the combinatorial definition that these polynomials are in fact symmetric and a basis for the symmetric polynomials @xcite .
linear independence follows from the fact that the lowest homogeneous component of @xmath35 is @xmath14 .
+ we define the * word of a tableau @xmath60 * to be the entries read right to left , top to bottom .
note that entries in a set are listed in increasing order , so that they occur in decreasing order in the word .
a word is called a * reverse lattice word ( rlw ) * if for any initial string , @xmath61 a word that satisfies this condition is sometimes called an _
election word_. for tableaux of shape @xmath1 and @xmath62 , we define the shape @xmath63 as the skew tableau formed by placing @xmath62 directly southwest of @xmath1 . when we refer to a filling of the shape @xmath63 , we will call @xmath1 the `` northeast '' partition , and @xmath62 the `` southwest '' partition .
buch @xcite gives a combinatorial rule for the product of two grothendieck polynomials : @xmath64 where the coefficients are given by @xmath65 for ss - svt @xmath66 of shape @xmath63 , content @xmath40 , with @xmath60 a rlw .
we call these the k - theoretic littlewood - richardson numbers , since if @xmath67 , then @xmath68 , the usual littlewood - richardson number .
first , we note that the reverse lattice word condition requires that the filling of the northeast tableau @xmath1 always be superstandard .
we will construct a poset out of all of the possible fillings of the southwest tableau @xmath62 , where each vertex is labeled with all tableaux of a given content , and for vertices @xmath69 , @xmath70 if @xmath71 .
note that for each tableau , the row in the poset corresponds to the number of extra " elements in the filling ( e.g. the top row has only semistandard young tableaux ) .
for example , consider the product @xmath72 and its poset in figure [ g21g22 ] .
note that the product is h - multiplicity - free , but not k - multiplicity - free .
the latter cases are extremely rare , occurring only when both partitions @xmath1 and @xmath62 are rectangles or one of them is a single box or empty ( ( * ? ? ?
* proposition 7.2 ) ) .
we will consider our products as being inside an ambient box of size @xmath2 .
that is , we limit the terms in the expansion to those indexed by partitions that fit inside this box .
we note that this restriction gives us a sub - poset of the full poset .
the mbius function on the remaining terms is unaffected by the removal of vertices with content exceeding the box size , since all terms above a vertex @xmath40 have content contained in the content of @xmath40 .
that is , for a given vertex @xmath40 with content in the ambient box , no vertex in its upwards order ideal will have content exceeding the ambient box .
we are interested in cases in which the terms in the grothendieck expansion which correspond to the schur expansion are multiplicity - free , i.e. that their coefficients are 0 or 1 .
then theorem [ thm : mobiusbrion ] implies the following : [ mainthm ] consider partitions @xmath73 and @xmath74 such that @xmath75 in a @xmath2 box is a schur - multiplicity - free product . in the corresponding poset , for each vertex @xmath76 , @xmath77 gives the coefficient of @xmath78 in the buch expansion of the product , where @xmath79 .
these schur - multiplicity - free cases have been classified by stembridge @xcite , and our proof explicitly uses his analysis .
we now mention some speculative geometry that motivated our _ combinatorial _ proof of theorem [ mainthm ] .
buch shows that the expansion of @xmath80 into schubert classes has signs that alternate with dimension ( @xcite ) .
this suggests that there exists an exact sequence on sheaves @xmath81 where the @xmath82 nonzero term sums over buch littlewood - richardson tableaux with @xmath83 extra entries .
this leads to a sequence for the point in the grassmannian corresponding to each @xmath1 , @xmath84 one can hope that this sequence is in fact exact .
our main result is [ sequencethm ] there exists such an exact sequence of vector spaces , and it can be explicitly constructed as a direct sum of exact sequences with exactly two non - zero terms .
the proof requires an involution which pairs terms differing in size by one . in some cases
, we provide a single rule that matches all terms required .
in other cases however , we must resort to a multistage divide and conquer approach , where the involution is defined differently on several disjoint subsets .
we will come back to this theorem in section [ last ] .
assuming theorem [ sequencethm ] , we can prove theorem [ mainthm ] as a corollary .
the exactness of the sequence ( [ eq : pointseq ] ) gives us that the alternating sum of dimensions is 0 .
thus the sum of the coefficients of the pairs of buch tableaux , with signs alternating with number of extra numbers , is also 0 .
together with the extra 1 from the single fixed point tableau , this is equivalent to the statement that the coefficient of @xmath76 is given by the mbius function .
we begin by recalling stembridge s definitions and classification of schur - multiplicity - free cases .
@xcite a partition @xmath62 with at most one part size ( i.e. , empty , or of the form @xmath85 for suitable @xmath86 ) is said to be a * rectangle*. if it has @xmath87 rows or @xmath87 columns ( i.e. , @xmath88 or @xmath89 ) , then we say @xmath40 is a * k - line rectangle*. a partition @xmath62 with exactly two part sizes ( i.e. , @xmath90 for suitable @xmath91 and @xmath92 ) is said to be a * fat hook*. if it is possible to obtain a rectangle by deleting a single row or column from the fat hook @xmath62 , then we say that @xmath62 is a * near rectangle*. we will call these * top , bottom , left , * or * right * near rectangles , to denote the location of the extra row or column .
we say that a product of schur functions is * multiplicity - free * if all of the littlewood - richardson coefficients of the expansion are 0 or 1 .
[ stembridge ] @xcite the product of schur functions @xmath93 is multiplicity - free if and only if 1 .
@xmath1 and @xmath62 are rectangles , or 2 .
( pieri rule ) @xmath1 is arbitrary , and @xmath62 is a 1 . one - row rectangle , or 2 .
one - column rectangle , or 3 .
@xmath1 is a rectangle and @xmath62 is a 1
. left near - rectangle , or 2 .
bottom near - rectangle , or 3 .
top near - rectangle , or 4 .
right near - rectangle , or 4 .
@xmath62 is a fat hook and @xmath1 is 1
. two - row rectangle , or 2 .
two - column rectangle , ( or vice - versa ) .
we note that stembridge classifies these as 4 cases , but in our analysis they naturally split further and thus have listed them as such .
we will start with case ( 1 ) and use it to show all remaining cases except for ( 3d ) , which requires its own proof technique .
we also note that stembridge s products are not restricted inside a box . putting our products inside an ambient box
gives us a larger class of cases that may not be inherently multiplicity - free , but which lose the terms with multiplicity when considered as being inside a small enough box .
these have been classified by thomas and yong ( @xcite ) , and we discuss how to generalize to these in the last section of this paper . henceforth , @xmath62 will be used to denote the southwest term , and @xmath1 the northeast partition . in this way , it will be clear of which term we are considering the fillings .
first , let us consider a graphical interpretation of the reverse lattice word condition .
we create a diagram by putting @xmath1 in the northwest corner of the ambient box . as we read the word of @xmath62 , for each @xmath51 , we place a box in our diagram in row @xmath51 adjacent to the rightmost box .
note that the reverse lattice word condition is equivalent to this diagram always being a partition . for values @xmath44 and @xmath94 in the word of a tableau ,
we say that @xmath94 * depends on * @xmath44 if the corresponding @xmath94 box occurs directly below the corresponding @xmath44 box .
that is , if that particular @xmath44 were nt there , we would not be allowed to add that particular @xmath94 .
for example , in figure [ rectcols ] , we say that 3 depends on 2 , but 5 does not depend on anything . consider the product @xmath95 where @xmath96 and @xmath97 in an ambient box of size @xmath2 .
a filling of a column in tableau @xmath66 of shape @xmath62 is built out of a basic set of blocks , given by @xmath98^{tr } \text { for } p \leq \alpha
\text { and } [ \beta+1,\beta+2,\ldots , q]^{tr } \text { for } q \leq \beta + \alpha\ ] ] where one overlap between the blocks is allowed , and @xmath99 indicates transpose . we claim that in terms of the requirements of the reverse lattice word , each column is independent of the others . that is , for each entry
, any entries on which it depends occur in the same column .
if not , we show that there would be a gap in the column that is not of the form above .
say we have a column of the form @xmath100^{tr}$ ] .
then in order to satisfy the reverse lattice word condition , there must be a ( @xmath101 in @xmath66 that comes before @xmath102 in @xmath60 , which implies that the @xmath103 would be either to the right in the same row ( which would violate the semi - standardness condition ) or in one of the rows above .
then @xmath103 must also come after an @xmath51 in the word , and either * @xmath103 depends on the @xmath51 in @xmath104 , or * @xmath103 depends on an @xmath51 in a different column . in the first case
, @xmath103 would have to be to the left of @xmath51 in the same row ( also violates semi - standardness ) or below ( same row as @xmath102 ) .
in the second case @xmath103 must have a corresponding @xmath105 preceding it in @xmath60 , but this creates the same problem .
a similar argument holds for a gap in @xmath106^{tr}$ ] .
finally , we note that more than one overlap would violate semi - standardness .
this lemma allows us to take the complicated reverse lattice word condition , and turn it into a simple description of the valid column fillings .
let @xmath107 , @xmath108 .
then the only possible column fillings for @xmath62 are : @xmath109 , \left[\begin{array}{c } 4\\ 5 \end{array}\right ] , \left[\begin{array}{c } 1\\ 4 \end{array}\right ] , \left[\begin{array}{c } 14\\ 5 \end{array}\right ] , \left[\begin{array}{c } 1\\ 24 \end{array}\right].\ ] ] in figure [ g22g333 ] , we can see how these columns form tableaux .
we define the * snake * of a filling of @xmath62 in the 2-rectangle case as the union of the lines below all single entries @xmath51 occurring in box @xmath50 , and the lines above all entries @xmath110 .
the lines are co - linear except for where they enclose entry @xmath51 in boxes with entry @xmath111 .
note that a snake uniquely characterizes a filling of a tableau , as the values both above and below the snake are fixed .
+ for a given poset , let @xmath112 denote the first tableau in lexicographic order , and @xmath113 the upper block defined by the snake . in @xmath113 ,
each box @xmath50 has the single entry @xmath51 .
note that @xmath112 is a semi - standard young tableau . in order to prove theorem [ sequencethm ]
, we will define sign - reversing involutions on the poset to show that we can match all terms except one , @xmath112 . more specifically
, our involutions will match terms differing in content size by one and thus terms in adjacent rows .
this will show that the k - theoretic littlewood - richardson coefficients are given by the mbius function .
we define the function @xmath114 as follows : compare each box in @xmath113 to the corresponding box in @xmath66 , top to bottom down a column , and left to right across the tableau . if they all match , let @xmath115 .
otherwise , call the first box that does nt match @xmath50 . *
if @xmath116 is not in @xmath117 , let @xmath118 .
* if @xmath116 is in @xmath117 , let @xmath119 .
then @xmath120 matches @xmath66 with @xmath121 .
graphically , applying @xmath120 is equivalent to narrowing or widening the snake by the one box @xmath50 .
[ 2rect ] @xmath120 is a sign - reversing involution on the poset corresponding to the product @xmath95 in the case where both partitions are rectangles , @xmath96 and @xmath122 , whose only fixed point is @xmath112 .
the choice of box @xmath50 is well - defined . in a tableau @xmath66 ,
if @xmath117 does nt match @xmath123 , then either @xmath124 $ ] or @xmath125 $ ] .
since @xmath120 is only dependent on those terms above it in the @xmath126 column , adding or removing it will not affect the rest of the filling , and will give another valid tableau . since @xmath120 matches a tableau @xmath66 with another tableau @xmath121 that has either one more or one fewer element , it is clearly sign - reversing .
consider the product @xmath75 in a @xmath2 box for a rectangle of shape @xmath96 and an arbitrary tableau of shape @xmath74 .
define the * upper rectangle * of @xmath62 , * @xmath127*. let @xmath128 denote the filling of the upper rectangle of @xmath112 .
[ redlem ] consider the poset of fillings of the two tableaux representing the product @xmath75 in the ambient box @xmath2 , where @xmath96 and @xmath74 .
let @xmath129 . under @xmath120 ,
the poset of fillings of @xmath62 reduces to the product of @xmath130 and @xmath131 , where @xmath132 and @xmath133 , with the product in ambient box @xmath134 .
apply @xmath120 to the upper rectangle of @xmath62 .
then by theorem [ 2rect ] , all terms are canceled except those whose upper rectangle filling matches that of @xmath112 . consider @xmath135 obtained from the standard filling of ( @xmath1 ) and this filling of the upper rectangle .
then the remaining poset is equivalent to the poset of the product @xmath136 , where @xmath130 is the shape determined by a standard filling of the partition of shape corresponding to @xmath137 with the top @xmath138 rows truncated , with the added restriction on the number of 1 s to the number of @xmath138 s in @xmath137 .
note that by our construction of the reduction lemma , @xmath139 and @xmath140 .
thus we have reduced to a product of two smaller tableaux .
now , we will use our reduction lemma to prove theorem [ sequencethm ] in each of the stembridge cases . as in the statement of the lemma
, we use the notation @xmath96 and @xmath74 , and the poset of fillings of the two tableaux representing the product @xmath75 in ambient box @xmath2 .
we denote the filling of the minimal element @xmath112 ( of shape @xmath62 ) as @xmath141 .
we will successfully apply the reduction lemma to most of the stembridge cases .
we omit the discussion of the ambient box in the following proofs as it is not particularly enlightening , but the size of the ambient box after the lemma is applied is easily determined in each case .
[ 2a ] [ pieri - stembridge case ( 2a ) ] theorem [ sequencethm ] holds in the case of a single row and an arbitrary tableau @xmath142 : let @xmath143 and @xmath74 .
we proceed by induction .
consider the base case @xmath144 : the only possible fillings for @xmath145 are @xmath146 \text { or } [ 1,\ldots , 1 , [ 12],2 , \ldots , 2]$ ] .
the involution matches terms with 2 or @xmath147 $ ] in the @xmath148 spot .
+ now , assume the claim holds for the products @xmath149 for some @xmath150 .
then consider @xmath151 . applying the reduction lemma
, we get the product @xmath152 .
note that @xmath153 .
[ 2b ] [ pieri - stembridge case ( 2b ) ] theorem [ sequencethm ] holds in the case of a single column and an arbitrary tableau @xmath154 : let @xmath155 and @xmath156 .
we proceed by induction .
consider the base case @xmath157 : the only possible fillings for @xmath145 are @xmath158 \text { or } [ 1,\ldots , 1 , 2 ] \text { or } [ 1,\ldots , 1 , [ 12]].\ ] ] the involution matches the latter two cases with each other , leaving the first .
+ assume the claim holds for the products @xmath159 for some @xmath150 .
then consider @xmath160 . applying the reduction lemma
, we get the product @xmath161 where @xmath162 .
[ 3a ] [ stembridge case ( 3a ) ] theorem [ sequencethm ] holds in the case of a rectangle and a left near rectangle @xmath163 : let @xmath164 and @xmath165 .
apply the reduction lemma to the upper rectangle @xmath166 . then @xmath167 and @xmath168
this is pieri - stembridge case ( 2b ) .
[ 3b][stembridge case ( 3b ) ] theorem [ sequencethm ] holds in the case of a rectangle and a bottom near rectangle @xmath169 : let @xmath164 and @xmath170 .
apply the reduction lemma to the upper rectangle @xmath166
. then @xmath171 and @xmath168 .
this is pieri - stembridge case ( 2a ) .
[ 3c][stembridge case ( 3c ) ] theorem [ sequencethm ] holds in the case of a rectangle and a top near rectangle @xmath172 : let @xmath164 and @xmath173 .
apply the reduction lemma to the upper rectangle @xmath174 . then @xmath175 and @xmath176 for @xmath177 .
this is pieri - stembridge case ( 2b ) .
[ 4a ] [ stembridge case ( 4a ) ] theorem [ sequencethm ] holds in the case of a two row rectangle @xmath178 and a fat hook : let @xmath179 and @xmath180 .
apply the reduction lemma to the upper rectangle @xmath166 . then @xmath175 and @xmath181 .
this is stembridge case ( 3b ) or ( 3c ) .
[ 4b ] [ stembridge case ( 4b ) ] theorem [ sequencethm ] holds in the case of a two column rectangle @xmath182 and a fat hook : let @xmath183 and @xmath180 .
apply the reduction lemma to the upper rectangle @xmath166 . then @xmath175 and @xmath184 for some @xmath185 .
this is stembridge case ( 3a ) .
finally , consider stembridge case ( 3d ) , of a rectangle @xmath186 , and a right near rectangle @xmath187 .
we consider this as almost " the rectangle times rectangle case , and we consider the fillings of the rectangle . as in the 2 rectangles case , we can characterize the fillings of the columns of the rectangle in terms of column blocks : consider the product of @xmath186 and @xmath187 in an ambient box of size @xmath2 . a filling of a column in @xmath62 is built out of a a basic set of blocks , given by * @xmath188^{tr}$ ] for @xmath189 . *
@xmath190^{tr}$ ] for @xmath191 . plus one initial string of @xmath192 $ ] for @xmath193 which can occur of any length from 0 to @xmath194 running southwest through the filling , occurring in order in @xmath60 .
additionally , two overlaps are allowed in each column ( either as two blocks with two entries each , or one block with 3 entries ) .
+ a parallel argument to that in the 2-rectangles column characterization holds as to why the blocks occur and why there can be no other gaps besides a jump between blocks .
the extra string must occur in order in the word , which means each entry must occur weakly southwest of the previous .
( by semi - standardness , it must occur strictly south , weakly west . )
overlaps are allowed between the column blocks , as well as with entries of the snake .
+ first we apply @xmath120 , to the upper rectangles of the entries in the poset .
all terms where @xmath195 differs from @xmath128 will cancel under this involution , leaving terms where @xmath195 is identical to @xmath128 .
now , we define two more involutions , which when applied in succession will cancel the rest of the terms
. we call the * intersnake * those entries in the filling that correspond to the missing column piece in the fat hook ; more explicitly , the entries that correspond to * the first occurrence of entries @xmath196 * the @xmath197 occurrence of entries @xmath198 in the word of the tableau .
we extend our previous definition of a * snake * for this case to be the union of four lines : the line below all single entries @xmath51 occurring in box @xmath50 , the line above all entries @xmath199 , and the lines defining the intersnake region .
note that a snake uniquely defines a filling of a tableau , as the values in each region are determined .
[ snakelem ] the unique tableau in which the intersnake is maximally northeast is @xmath112 , the first tableau in lexicographic order .
we will show that any weakly southwest shift of a snake box in an arbitrary tableau @xmath66 will yield a tableau @xmath121 that is lexicographically after @xmath66 .
consider a column in @xmath66 : @xmath200\ ] ] for @xmath201 in the intersnake of @xmath66 and @xmath202 , and a column to the left @xmath203\ ] ] for @xmath204 .
if we move @xmath201 to the left column and shift appropriately to get another tableau @xmath121 where the snake box is southwest of the box in @xmath66 , we get columns in @xmath121 of the form @xmath205\ ] ] and @xmath206\ ] ] respectively .
we get @xmath207 by taking @xmath60 and replacing @xmath208 $ ] with the larger @xmath209 $ ] .
that is , @xmath121 comes after @xmath66 lexicographically .
we will use * @xmath210 * to denote those semi - standard set valued tableaux left after involution @xmath87 .
we define the function @xmath211 as follows : compare @xmath112 and @xmath66 along the intersnake as defined by @xmath112 , and find the first box where they differ .
let this be the @xmath82 position in the intersnake of @xmath112 .
let @xmath212 be the length of the intersnake .
* if @xmath116 in @xmath117 , remove it as follows : * * if @xmath213 , remove @xmath116 . * * if @xmath214 , remove @xmath116 and replace @xmath215 with @xmath216 for @xmath217 $ ] . *
if @xmath116 is not in @xmath117 , add it as follows : * * if @xmath213 , add @xmath116 . * * if @xmath214 , add @xmath116 to @xmath117 and replace @xmath216 with @xmath215 for @xmath218 $ ] . [ inv2 ] @xmath219 is a sign - reversing involution .
this function changes the content of a tableau @xmath66 by either adding or removing the last value in the intersnake of @xmath66 .
this is clearly sign - reversing , as we are changing the size of the content by one .
note that since we are applying this to terms in which the above - snake region matches @xmath112 , now the only overlaps in the columns are of the form @xmath220 $ ] for @xmath221 .
then , @xmath222 $ ] or @xmath223 $ ] .
the shifting is necessary to get a tableau that satisfies the reverse lattice word condition .
note that we apply @xmath219 , we can always remove an entry from the intersnake , but we can not always add one : tableaux which are not matched by this involution are those in which * @xmath219 wants to add an entry * we can not pull the intersnake through , because the full intersnake already appears in @xmath66 .
then the only terms left in the poset are those with @xmath224 , and the full intersnake appears in @xmath66 but does not match @xmath112 s intersnake .
we will define one last involution on the remaining terms .
we define the function @xmath225 as follows : compare @xmath112 and @xmath66 along the intersnake as defined by @xmath112 , and find the first box where they differ .
let this be the @xmath82 position in the snake of @xmath112 , and call the box in @xmath66 , @xmath50 , and in @xmath112,@xmath226 . in column @xmath227 ,
* if every entry in the column is single , let @xmath228=box above entry @xmath229 ( @xmath230 if @xmath229 is not there ) . then @xmath231 $ ] and replace entry @xmath232 with @xmath233 for @xmath234 $ ] . *
if box @xmath228 has two entries , @xmath235 and replace @xmath232 with @xmath236 for @xmath234 $ ] .
then @xmath237 matches @xmath66 with @xmath121 .
this function changes the content of a tableau @xmath66 by shifting one column in the below - snake region .
[ inv3 ] @xmath237 is a sign - reversing involution .
the choices of @xmath238 and @xmath239 are well - defined .
it is clearly sign - reversing , as we are changing the size of the content by one .
+ note that while we can always shift down ( remove the last entry in ) a column to get another valid filling , we can not always shift up ( add the last entry ) .
let us consider those cases in which we shift up .
if @xmath240 , there is a @xmath229 in column @xmath51 , so we can clearly add the last term to the shifted column .
however , if @xmath241 , adding @xmath229 might violate the reverse lattice word condition . since each snake box occurs strictly south of the previous , @xmath241 implies that @xmath87 is the last snake entry . note that by our choice of box @xmath228 , @xmath66 matches @xmath112 in columns east of column @xmath242 and in any columns in between @xmath227 and @xmath242 , so we need only consider those two columns . by lemma [ snakelem ] , @xmath228 is strictly southwest of @xmath226 .
this implies that @xmath243 , so both columns @xmath242 and @xmath227 have @xmath229 in @xmath112 , which means that the reverse lattice word condition is not violated by two occurrences and we can thus add it to @xmath244 .
[ stembridge case ( 3d ) ] theorem [ sequencethm ] holds in the case of a rectangle and a right near rectangle .
as shown , by applying @xmath120 , @xmath219 , and @xmath237 in that order , all terms are canceled except for @xmath112 .
let us now go back to theorem [ sequencethm ] .
recall that we want to construct the exact sequence @xmath245 we claim that the functions are given by @xmath246 from our involutions , we have the following two - term exact sequences , where each tableau filling @xmath66 is represented by @xmath247 . for each pair of tableaux of sizes @xmath248 and @xmath249 and matched by one of the above involutions , we define the functions in our sequence by : @xmath250 for @xmath251 the identity function .
that is , we obtain a sequence @xmath252 for each pair of tableaux .
each sequence is clearly exact . for a given @xmath1
, we can construct the desired sequence as a direct sum of sequences of this form , one corresponding to each matched pair in the order ideal of @xmath1 .
consider partitions @xmath1 and @xmath62 in a @xmath253 box .
we will review the notation introduced in @xcite .
we call @xmath254 a * richardson quadruple * , and use the notation @xmath255 to denote the associated poset of fillings of @xmath62
. place @xmath1 in the upper left corner of the box , then rotate @xmath62 by @xmath256 ( call this * rotate(@xmath62 * ) ) and place it in the lower right corner .
this quadruple @xmath257 is called * basic * if @xmath258 does not contain any full rows or columns .
if it is not basic , we can remove all full rows and columns to get a * basic demolition * @xmath259 .
we call each row ( column ) removal a * row ( column ) demolition*. notice that if @xmath260 , then @xmath261 .
@xcite a richardson quadruple is multiplicity - free if and only if its basic richardson quadruple is multiplicity - free .
if a basic richardson quadruple @xmath262 is multiplicity - free , then it must be in the cases classified by @xcite .
for example , consider the case @xmath263 .
this product is not multiplicity - free , but has a basic demolition of @xmath264 , which is multiplicity - free .
+ we will show that our analysis of the @xcite multiplicity - free cases extends to this larger class of products by showing that the posets of a richardson quadruple and its basic demolition are isomorphic .
let us define the * accessible word @xmath265 * as the independent values of @xmath1 read in increasing order , or equivalently @xmath266 [ coldem ] for @xmath254 , if column @xmath267 is full , then @xmath268 is isomorphic to @xmath255 .
we will show that the function given by removal of column @xmath267 from every tableau in the poset gives the isomorphism .
+ let @xmath269 be the number of rows in column @xmath267 in @xmath1 , so @xmath270 is the number of rows in column @xmath267 in @xmath62 .
we claim that a full column implies that in every filling of @xmath62 , column @xmath267 is filled as @xmath271 $ ] .
consider box @xmath272 in @xmath1 : filling it with a value greater than @xmath273 would mean that the last box in the column @xmath274 must contain an entry that is at least @xmath275 , which is larger than the ambient box .
filling it with a value less than @xmath273 would lead to an invalid filling of boxes @xmath276 to @xmath277 .
then @xmath272 contains the single entry @xmath273 , and in order to fit in the ambient box , the proposed column filling is the only possibility .
we define a function @xmath278 as follows : for each tableau @xmath66 in @xmath255 , we remove column @xmath267 and shift columns @xmath279 to @xmath280 left by one , where entries are shifted with their associated boxes . this operation does not change semistandardness .
column @xmath267 is independent in the sense of the reverse - lattice word , and since the ambient box has @xmath87 rows , there can not be a @xmath275 that depends on it .
thus for @xmath66 of shape @xmath281 we get a valid tableau filling @xmath121 of shape @xmath282 in @xmath283 .
the inverse function @xmath284 shifts columns @xmath267 to @xmath285 to the right by one and adds the column back in .
this is clearly an isomorphism .
let @xmath286 and @xmath287 .
there is a full row in the diagram if and only if @xmath288 .
[ fulltoprow ] for @xmath254 , if @xmath289 , then @xmath255 is isomorphic to @xmath283 for @xmath290 .
first we observe that @xmath289 implies that @xmath62 can not have any 1 s in any filling .
then we define the function @xmath291 as the function that takes each entry @xmath51 to entry @xmath292 , with obvious inverse @xmath293 .
@xmath294 is clearly an isomorphism .
[ rowdem ] for @xmath254 , if row @xmath295 is full , then the @xmath283 is isomorphic to @xmath255 .
we will show that the function given by removal of the first @xmath296 boxes in row 1 of @xmath62 , followed by a shifting up of the first @xmath296 columns , gives the isomorphism .
let @xmath297 be the number of columns in row @xmath295 in @xmath1 , so @xmath298 is the number of columns in row @xmath295 in @xmath62 .
first we claim that the entry in @xmath299 is weakly less than @xmath295 .
there are only @xmath296 accessible entries that satisfy this condition , so these must be the filling of this block in every @xmath62 , for @xmath227 from 1 to @xmath300 ( i.e. box @xmath301 contains the single entry @xmath302 ) . then we want to remove this extraneous information from the fillings of @xmath62 as follows : take out boxes @xmath301 for @xmath227 from 1 to @xmath296 , then shift up columns 1 to @xmath296 by one box
. this will give a new tableau of shape @xmath303 .
we instead record this information in @xmath304 .
that is , @xmath305 is now the product @xmath306 .
first we show that semistandardness is preserved .
the only place problems may occur are along the vertical line between columns @xmath307 and @xmath308 .
we start at the top of the column .
consider box @xmath309 with entry @xmath310 .
since @xmath311 , box @xmath312 must be filled with entries @xmath313 .
next we consider box @xmath314 : we claim that the entries are weakly less than @xmath315 .
if it contained entry @xmath316 , then by semistandardness , the last box in the column @xmath317 would have to have entries that were at least @xmath275 , which would exceed the ambient box .
when we shift box @xmath314 to @xmath309 , semistandardness in row 1 is preserved .
next , assume for contradiction that the shifting violates semistandardess in a block @xmath318 $ ] ; that is , that @xmath319 .
but , semistandardness stipulates that @xmath320 and @xmath321 , which together give the contradiction .
then semistandardness will be preserved throughout the column .
next , we show that the reverse lattice condition still holds after the shift .
assume that entry @xmath50 depends on entry @xmath322 .
the only place where the shifting will affect the order is if @xmath323 , @xmath324 and @xmath325 .
however , if this condition existed , semistandardness would be violated by shifting , which we have already shown can not happen .
thus , this shifting gives a valid tableau filling .
finally , we consider how this shifting affects the partial order .
since we are removing an identical block from every tableau , the relative content of any pair of tableaux is the same , and thus the relations in the poset are unchanged . by construction , @xmath326 , so we can apply lemma [ fulltoprow ] to reduce the ambient box .
each step of the function is easily reversed , so it is an isomorphism .
we note that the row and column demolition lemmata define commutative operations on the poset of a richardson quadruple .
figure [ tydemo ] is an example of a case with both a full row and column .
( this product is stembridge multiplicity - free in any ambient box , but is a good example of the row and column demolition commutativity . )
theorem [ sequencethm ] holds for any richardson quadruple whose basic demolition is a stembridge case .
this follows clearly from the previous two lemmata .
i would like to extend special thanks to allen knutson for both the statement of the question and continuing guidance throughout the process .
a. lascoux and m .-
schtzenberger , structure de hopf de lanneau de cohomologie et de lanneau de grothendieck dune varit de drapeaux , c. r. acad .
paris sr .
i math . 295
( 1982 ) , no . 11 , 629633 . | we consider buch s rule for k - theory of the grassmannian , in the schur multiplicity - free cases classified by stembridge . using a result of knutson
, one sees that buch s coefficients are related to mbius inversion .
we give a direct combinatorial proof of this by considering the product expansion for grassmannian grothendieck polynomials .
we end with an extension to the multiplicity - free cases of thomas and yong . |
dissipation and dephasing from two - level systems ( tls ) are a serious problem in many superconducting qubits .
the aggregate effect of many weakly coupled fluctuators causes @xmath5 charge noise , broadband dielectric loss , and magnetic flux noise , as well as inhomogeneous broadening and decreased measurement fidelity in qubits.@xcite an individual tls quantum - coherently coupled to a qubit can typically be identified when it leads to a resolvable avoided crossing in the qubit spectrum .
such avoided level crossings have been observed in phase,@xcite flux,@xcite charge,@xcite quantronium,@xcite and transmon@xcite qubits . while qubit performance is typically severely degraded near such an avoided crossing,@xcite strong qubit - tls interactions
allow the microscopic details of the tls to be determined.@xcite coherent coupling to a long - lived tls also makes it possible to observe coherent oscillations between a qubit and a tls@xcite or use the tls as a quantum memory.@xcite two - level fluctuators in superconducting devices can be classified into three types charge , flux , or critical current depending on the nature of the interaction with the qubit .
the microscopic origin of charge and critical current fluctuators is believed to be impurity ions such as h@xcite or low coordination bonds in the amorphous dielectric used to build the devices . in phase and flux qubits
, it appears to be possible in principle but difficult in practice to identify the exact nature of the qubit - tls interaction .
in contrast , detailed spectroscopy on charge qubits or cooper - pair boxes ( cpb ) has enabled the identification of discrete charge fluctuators.@xcite for example , kim _
et al_. found tls s that behaved as pure charge fluctuators.@xcite a moving charge could also modulate the critical current if it was located in the tunnel barrier.@xcite critical current fluctuations have been frequently seen in josephson junction devices,@xcite but apparently not in josephson based qubits .
this may be due to the difficulty of conclusively distinguishing a critical current fluctuator from a charge fluctuator .
alternatively , the relatively small area of qubit junctions compared to that of conventional junctions leads to far fewer total fluctuators and a corresponding decrease in the probability of observing one .
also , qubit measurements are typically made at less than @xmath6{mk}$ ] , where critical current fluctuators appear to be frozen out .
josephson junctions are a fundamental building block of all superconducting qubits and an understanding of the origin of critical current fluctuations is important for continued improvement of qubit performance . in this paper
we report on a cpb with an unusual spectrum that has multiple spectroscopic features displaced in both frequency and in gate charge instead of an avoided level crossing .
we find that the spectrum , including the curvature of the spectral features , can be modeled well with a critical current fluctuator coupled to a cpb with an excitation energy for the fluctuator much less than the qubit energy . by fitting our model to the spectrum
we extract microscopic parameters for the fluctuators .
our cooper - pair box ( cpb ) consists of a superconducting island connected to a superconducting reservoir ( ground ) through two ultrasmall josephson tunnel junctions ( critical current @xmath7 and junction capacitance @xmath8 ) [ see fig . [
fig : images+setup](d ) ] .
we can apply gate voltage @xmath9 to a capacitively coupled gate ( capacitance @xmath10 to the island ) to control the system s electrostatic energy . applying flux @xmath11 to the loop formed by the two junctions tunes the effective total critical current @xmath12 via the relation @xmath13 where @xmath14 is the magnetic flux quantum . neglecting quasiparticle states , the hamiltonian describing a cpb in the charge basis
is given by@xcite @xmath15 where @xmath16 is the charging energy , @xmath17 is the josephson energy , @xmath18 is the total island capacitance to ground , @xmath19 is the reduced gate voltage and @xmath20 is the excess number of cooper - pairs on the island . for @xmath21
the system is highly anharmonic and only a few charge states are needed to accurately describe the lowest energy states . for charge qubits with @xmath22 and @xmath23
, the hamiltonian can be reduced to@xcite @xmath24 which yields the excited state transition energy @xmath25 . near the charge degeneracy point
@xmath26 the transition energy varies parabolically as @xmath27 . to measure the state of the qubit
, we coupled our qubit to a thin - film quasi - lumped element lc resonator [ see fig .
[ fig : images+setup](d ) ] that was in turn weakly coupled to a microwave transmission line patterned on the sample chip . to read out the state of the qubit
, we apply microwave power at the resonance frequency of the resonator and record the transmitted microwave signal @xmath28.@xcite this is a dispersive readout in which the qubit state modulates the resonance frequency of the resonator . in the case of weak qubit - resonator coupling @xmath29 and large detuning @xmath30
the combined cpb - resonator system hamiltonian is approximately@xcite @xmath31 where @xmath32 is the strength of the qubit - resonator coupling energy , @xmath33 is the coupling capacitance between the resonator and the island of the cpb , @xmath34 is the capacitance of the lc resonator , @xmath35 is the resonance frequency , @xmath36 is the number operator for excitations in the resonator , and @xmath37 is the pauli spin operator .
this jaynes - cummings hamiltonian yields transitions in which the bare resonator frequency @xmath35 is dispersively shifted by @xmath38 depending on the state of the qubit . if @xmath39 , where @xmath40 is the resonator linewidth , the average phase of the transmitted signal at @xmath41 is linearly dependent on the excited state occupation probability . on the other hand
if @xmath42 , then the in - phase or quadrature transmitted voltage is proportional to the excited state occupation probability.@xcite additional complications can arise when the qubit and resonator are coupled to another quantum system , such as a tls .
if multiple energy levels in the combined system have similar detunings from the resonator , the effective dispersive shift @xmath43 will have a contribution from each level.@xcite qubit state readout can still be performed as described for the two level case , but the sensitivity to a particular state depends on the choice of resonator probe frequency . as we see below , this is our situation .
or @xmath44 with corresponding energies @xmath45 or @xmath46 and can tunnel between them with energy @xmath47 .
additionally the josephson energy @xmath48 of the cpb depends on whether the fluctuator is at @xmath49 or @xmath44.[fig : tls - diagram ] ] to include the effects on a cpb produced by a combined charge and critical current fluctuator , we expand on the charge defect model previously reported by z. kim , _
et al_.@xcite we assume the fluctuator acts as a two - level system in which a point charge in the tunnel barrier can tunnel between two potential well minima . in the tls position basis the fluctuator hamiltonian is given by @xmath50 where @xmath46 and @xmath45 are energies of the charge in the left and right position states and @xmath47 is the tunneling matrix element [ see fig .
[ fig : tls - diagram ] ] . for an isolated fluctuator
the excited state transition energy is given by @xmath51 .
the charge coupling between the cpb and tls originates from changes in the electrostatic potential when the defect tunnels between its two sites .
using green s reciprocation theorem@xcite the change in the induced polarization charge on the island of the cpb when the fluctuator tunnels from the left to the right well is @xmath52 where @xmath53 is the tls charge , @xmath54 is the angle the tls displacement vector makes relative to the electric field in the junction and @xmath55{nm}$ ] is the thickness of the tunnel junction . for fixed net charge on the island this in turn results in a change in the electrostatic potential of the island given by @xmath56 accounting for the electrostatic
charging energy and the work done by the gate voltage source when the point charge moves , the coupling hamiltonian is given by @xmath57 where @xmath58 is the cpb charge operator that counts the number of excess cooper - pairs on the island and @xmath59 is the tls position operator.@xcite combining eqs .
( [ eq : hcpbmatrx ] ) , ( [ eq : htlsmatrx ] ) and ( [ eq : hcoupl ] ) we can write the total hamiltonian for a cpb coupled to a single charge fluctuator as @xmath60 . in block matrix form this becomes @xmath61 where @xmath62 , @xmath63 is the @xmath64 identity matrix , and @xmath65 and @xmath66 are the cpb hamiltonian with the tls in either the left or right well . if we assume @xmath67 then @xmath68 as given by eq .
( [ eq : hcpbmatrx ] ) and @xmath69 where @xmath70 sets the energy scale for the charge coupled interaction between the fluctuator and the cpb .
if the tls is in the junction tunnel barrier , it can also modulate the critical current depending on its position.@xcite this coupling can be accounted for by making the substitution @xmath71 in @xmath65 and @xmath72 in @xmath66 . numerically diagonalizing the resulting @xmath73 hamiltonian @xmath74 , we find the energy levels and the transition frequencies from the ground state to the excited states of the system .
an avoided crossing occurs if the excited state of the tls is resonant with the first excited state of the cpb at some value of the gate voltage @xmath75.@xcite however if the tls excited state energy lies below the cpb transition minimum the cpb spectrum is twinned , with one parabola corresponding primarily to the excited state of the cpb and the other to a joint excitation of the cpb and the tls .
considered individually , each parabola bears a strong resemblance to the spectrum of a tls - free cpb .
when the tunneling energy @xmath47 is small we can identify the qualitative effects of each parameter on the twinned parabolas .
@xmath76 creates an offset along the frequency axis and a change in the effective curvature while @xmath77 creates an offset along the @xmath75 axis and `` tilts '' the parabolas .
@xmath45 also creates an offset along the frequency axis that adds to or subtracts from the effect of @xmath76 .
finally @xmath47 determines the size of any avoided crossings that are present in the spectrum and determines the transition rate induced by a gate perturbation between the ground state and excited states involving the tls .
we can further extend the model by considering the effect of two critical current fluctuators .
this is motivated by the observation of quadrupling of the spectral lines in our data which ca nt be explained by the presence of a single tls .
the total hamiltonian for a cpb coupled to two fluctuators in block matrix form is @xmath78 where @xmath79 , @xmath80 , and @xmath81 where @xmath82 accounts for any possible tls - tls coupling and the indices refer to the first or second tls .
@xmath83 with @xmath84 is the cpb hamiltonian with the respective tls in either the left or right well .
for example , @xmath85 is given by @xmath86 and @xmath87 has the respective indices swapped .
@xmath88 includes the contribution of both tls and in addition present on the diagonal is a cpb mediated tls - tls interaction term@xcite of the form @xmath89 .
we fabricated a thin - film lumped - element superconducting microwave resonator using standard photolithography and lift - off techniques .
it was made from a @xmath6{nm}$ ] thick film of thermally evaporated al on a c - plane sapphire wafer that was patterned into a meander inductor ( @xmath90{nh}$ ] ) and interdigital capacitor ( @xmath91{ff}$ ] ) coupled to a coplanar waveguide transmission line [ see fig . [
fig : images+setup](a , b ) ] .
the resonance frequency was @xmath92{ghz}$ ] with loaded quality factor @xmath93 , external quality factor @xmath94 , and internal quality factor @xmath95 .
the cpb was subsequently defined by e - beam lithography and deposited using double - angle evaporation and thermal oxidation of aluminum to create the @xmath96{nm}\times\unit[150]{nm}$ ] josephson tunnel junctions [ see fig . [
fig : images+setup](c)].@xcite for the e - beam lithography we used a bilayer stack of mma(8.5)maa copolymer and zep520a e - beam resist to facilitate lift - off and reduce proximity exposure during writing .
a @xmath97{nm}$ ] thick al island and @xmath98{nm}$ ] thick al leads were deposited in an e - beam evaporator . as discussed below , measurements of the cpb yielded @xmath99 in the @xmath100{ghz}$ ] range and we tuned @xmath101 from @xmath102{ghz}$ ] to @xmath103{ghz}$ ] .
the chip was enclosed in a rf - tight cu box that was anchored to the mixing chamber of an oxford instruments model 100 dilution refrigerator at @xmath2{mk}$ ] .
connections to the chip were made with al wirebonds .
we used cold attenuators on the input microwave line and isolators on the output line to filter thermal noise from higher temperatures [ see fig .
[ fig : images+setup](d ) ] .
a filtered dc bias voltage line was coupled to the input line using a bias tee before the device and a dc block was placed after the sample box . for spectroscopic measurements
the resonator was probed with a weak continuous microwave signal while a second pump tone was applied to excite the qubit .
the transmitted microwave signal at the probe frequency was amplified with a hemt amplifier sitting in the he bath [ see fig . [
fig : images+setup](d ) ] .
we implemented a coherent heterodyne setup to record the phase and amplitude of the transmitted probe signal at @xmath104{ns}$ ] time steps .
after the hemt , the signal was further amplified at room temperature , mixed with a local oscillator tone to an intermediate frequency of @xmath105{mhz}$ ] and then digitally sampled at a typical sampling rate of @xmath106{msa / s}$ ] .
a reference tone split off from the probe signal was directly mixed and digitally sampled .
both signals passed though a second stage of digital demodulation on a computer to extract the amplitude and phase .
all components were locked to a @xmath107{mhz}$ ] rb atomic clock .
both the probe and pump tone powers were optimized for ease of data acquisition while also minimally disturbing the qubit .
the probe tone power was calibrated via the ac stark shift.@xcite during measurement of the qubit state , the probe tone power was set to populate the resonator with an average @xmath108 photons while the concurrent pump tone power was slightly above that needed to saturate the cpb transitions .
we measured the transition spectrum of the qubit by recording the transmitted microwave probe signal while sweeping the dc gate voltage and stepping the frequency of the second microwave pump signal .
[ fig : cpb+tls - fits](d ) shows a plot of the transmitted probe signal amplitude as functions of @xmath75 and pump frequency @xmath109 with @xmath48 tuned near @xmath110 .
several unexpected and anomalous features are evident . rather than a single parabola , we observed two parabolas with varying curvatures offset by @xmath111{ghz}$ ] in frequency and @xmath112 in charge .
this spectral structure was stable over the course of four months and persisted as we tuned the transition frequency from @xmath1{ghz}$ ] .
close examination of the figure reveals sections of two more quite weak parabolas . a notable change in the spectrum occurred when we tuned @xmath48 to bring the transition frequency below that of the resonator .
as seen in fig .
[ fig : cpb+tls - fits](a ) , four parabolas are clearly visible with the stronger new pair displaced @xmath113{ghz}$ ] below the original two .
we note two additional anomalies we observed .
first , a `` dead zone '' was present between @xmath114{ghz}$ ] where no spectrum was visible .
second , only half of the spectral parabolas one from each pair were visible when measured with a pulsed probe readout .
for instance , in fig .
[ fig : cpb+tls - fits](d ) both parabolas were present when we used a continuous measurement but only the bottom parabola was visible when we used a pulsed measurement at a fixed gate voltage @xmath26 .
some clues about the nature of the fluctuator are evident from an examination of the spectrum .
the frequency offset between the two parabolas in fig .
[ fig : cpb+tls - fits](d ) could be caused by a flux fluctuator that modulates @xmath48 .
however such a fluctuator s effect on @xmath115 would be minimal when the applied flux is near zero and increase as @xmath48 is reduced by an external flux bias . as discussed below , this is the opposite of the behavior we observed . another argument against a simple flux fluctuator ( such as a vortex ) or a simple charge fluctuator
is that there are correlated shifts in @xmath75 and frequency between the parabolas . in contrast , the observed offsets and curvature changes are consistent with a two - level system that is coupled to the cpb via both charge and critical current . ) versus frequency for two different preparations of the qubit state .
the filled black squares and black curve show the transmission with the qubit biased at @xmath26 and driven to a mixed ground and excited state .
the two dips are resonances at @xmath116 .
the filled red circles and red curve show the transmission with the qubit in the ground state and far detuned from the resonator at @xmath117 .
the single dip is the bare resonator frequency at @xmath35 .[fig : ng - cuts ] ] several factors indicate that the fluctuator is coherently coupled to the cpb .
an incoherently coupled low frequency critical current fluctuator would be expected to produce twinning in the resonator dispersive shift @xmath118 in addition to twinning of the spectral lines .
this twinning of the dispersive shift would be manifest either as twinning of the ground state resonator frequency or broadening of the resonator linewidth .
we did nt observe either of these effects .
instead we observed an effective dispersive shift @xmath119 consistent with contributions from multiple levels [ see fig .
[ fig : ng - cuts ] ] .
we determined the effective dispersive shift @xmath119 and effective resonator frequency @xmath120 by recording the resonator response with the qubit in the ground and excited states .
we also measured the bare resonator frequency @xmath35 by far detuning the qubit from the resonator by biasing at @xmath117 [ see fig .
[ fig : ng - cuts ] ] .
as expected @xmath121 and the effective dispersive shift @xmath119 differed between the excited states corresponding to the various parabolas .
finally , in previous cases of incoherent fluctuator coupling we found that the qubit was rendered inoperable.@xcite yet in this case we were able to measure qubit excited state lifetimes @xmath122 in the @xmath123{{\ifmmode\begingroup\def\b@ld{bold } \text{\ifx\math@version\b@ld\bfseries\fi\textgreek{m}}\endgroup\else\textgreek{m}\fi}s}$ ] range and record rabi oscillations for all of the parabolas .
the strength of the qubit - tls coupling indicates that the tls was located close to the cpb josephson junctions , either in the tunnel barrier itself or on the surface of the cpb island .
furthermore , we note that the spectra were @xmath124 periodic in @xmath75 .
this is the expected periodicity for a charge fluctuator that is in the tunnel barrier@xcite , and such a fluctuator would need to be in the @xmath125 tunnel barrier to produce a critical current change .
.fit parameters for the model of a single two - level fluctuator coupled to a cpb .
the corresponding spectra are plotted in fig .
[ fig : cpb+tls - fits ] .
@xmath126 and @xmath48 are the cpb charging and josephson energies .
@xmath45 is the tls potential energy well asymmetry , @xmath77 is the charge coupling between the tls and the cpb , @xmath76 is the change in the cpb josephson energy when the tls tunnels between wells and @xmath47 is the tls tunneling rate.[tab : cpb+1tls - fits ] [ cols="<,<,<,<,<,<,<,<,<",options="header " , ] we also fit the entire spectrum of four parabolas to the two tls model [ see eqs .
[ eq : hcpb2tls ] and [ eq : q(rl ) ] ] .
the solid red curves in fig .
[ fig : cpb+2tls - fits ] show the best fit spectrum superposed on the data .
the optimal fit parameters are summarized in table [ tab : cpb+2tls - fits ] .
the vertical lines at @xmath127 and @xmath128 are due to the resonant crossing between the qubit parabolas and the resonator line at @xmath35 .
although the fits are reasonable and capture all of the major features , the fit parameters contain one surprise .
if we assume two independent fluctuators , the simplest assumption in light of the strong shielding of electric fields in the dielectric of the josephson junction by the superconducting electrodes , then we would expect @xmath129 and @xmath130 while @xmath131 .
however our fit yields @xmath132 while @xmath130 and @xmath133 which suggests coupled tls s or more complicated microscopic behavior .
furthermore , we note that several of the tls parameters , such as the charge coupling @xmath134 , change values when switching from the single to the double tls model .
this indicates that the two tls model is needed to explain the full quadrupled spectrum and suggests that there is significant interaction between the tlss .
there are some noteworthy implications from the magnitude of the fit parameters .
first , we note that @xmath135
. the large relative size of @xmath76 to @xmath48 suggests that the junction tunnel barrier is non - uniform with a few dominant conduction channels and that the tls is located near and modulates one of these channels.@xcite second , the tls tunneling matrix element @xmath136{ghz}$ ] is small compared to the other energies in the system , indicating that the tls is tunneling between fairly well isolated sites .
we can also place a lower bound on @xmath47 by noting that for @xmath137{ghz}$ ] the spectra would be too faint to observe .
if the excited state of such a tls were resonant with the first excited state of the cpb , the resulting avoided crossing would be very small and difficult to resolve .
our extracted tunneling matrix element values are also significantly smaller than those reported by z. kim , _
et al_.,@xcite which were in the @xmath138{ghz}$ ] range .
there is a similar relation between the range of well asymmetry values extracted by us , @xmath139{ghz}$ ] , and those reported by z. kim , _
et al_. , @xmath140{ghz}$ ] . assuming a tls charge of @xmath141 and a tunnel barrier thickness of @xmath142{nm}$ ] , we estimate the maximum hopping distance of the defect at @xmath143 .
this is in agreement with the bounds of @xmath144 found by z. kim , _
et al_. discrete critical current fluctuators have been reported in current biased josephson junctions , identified via either a random telegraph signal in the voltage time trace or a signature lorentzian bump in the noise spectrum.@xcite one way we can compare our tls s to others is to calculate the effective defect area @xmath145 given by @xmath146 . for our device find @xmath147{nm^{2}}$ ] where @xmath148{nm^{2}}$ ] is the junction area .
this value is much larger than the @xmath149{nm^{2}}$ ] reported in similar junctions,@xcite the @xmath150{nm^{2}}$ ] seen in larger area junctions,@xcite or the @xmath151{nm^{2}}$ ] found in similar area @xmath152 superconductor grain boundary junctions.@xcite on the other hand , the absolute value of the critical current fluctuation @xmath153{na}$ ] we observed is close to that reported in both similar area ( @xmath154{na}$])@xcite and larger junctions ( @xmath155{na}$]).@xcite one notable difference that might account for some of these discrepancies is that the critical current density of our sample ( @xmath156{a / cm^{2}}$ ] ) is smaller by an order of magnitude or more than the referenced samples .
if we assume that the conductance of a tunneling channel is similar between the various devices , this is consistent with a small number of tunneling hot spots in our junction .
the longitudinal relaxation rate of a tls in an amorphous solid is expected to be limited by @xmath157 where @xmath158 is temperature and @xmath159 is a material dependent constant.@xcite from the results of z. kim , _ et al_.@xcite we estimate @xmath160{\text{\textgreek{m}}s\cdot ghz^{3}}\cdot h^{3}$ ] for the dielectric @xmath125 in the tunnel junction barrier .
our fit values then place an upper bound on the tls excited state lifetime of @xmath161{ms}$ ] .
this bound is consistent with a relatively long tls lifetime and with our qubit @xmath162{\text{\textgreek{m}}s}$ ] .
the excited states of the system are mixtures of pure cpb and tls excited states , so the decay rate is a weighted average of the pure cpb and tls decay rates . for example , according to our fits to the model at @xmath26 the lower parabola in fig .
[ fig : cpb+tls - fits](d ) is composed of a @xmath163 cpb excitation and an @xmath164 joint cpb plus tls excitation while the upper parabola is an @xmath164 cpb excitation and a @xmath163 joint cpb plus tls excitation .
only when both the qubit and tls decay rates are small , as is our case , will the system decay time be long in both parabolas .
finally it s worthwhile to speculate why this behavior was observed in our sample . in order to observe spectral twinning rather than an avoided crossing
, the tls needs to be coupled to the qubit but have a transition frequency less than @xmath101 . that this occurred is a statistical coincidence .
observing two such defects in the same sample is less likely , and the tls fit parameters suggest they are correlated . furthermore , we are biased in selecting samples for detailed study that have especially conspicuous features , such as large avoided crossings or anomalous spectra , and the parameter values of such samples are likely to be somewhat unusual .
while our simple model provides a good fit to the recorded spectrum , it leaves other questions unanswered .
the resonator was nt included in the model but some of our observations suggest that it may produce significant effects on the spectrum .
inclusion of the resonator in the model would allow a theoretical calculation of the expected dispersive shift and a comparison with the data .
a more complete model may also elucidate the role , if any , the resonator played in the the large difference in the visibility of the different parabolas when the qubit was tuned from below to above the resonator @xmath35 or the `` dead zone '' we observed between @xmath114{ghz}$ ] where no spectrum was visible . perhaps the most puzzling feature was that half of the spectral parabolas were nt visible when measured with a pulsed probe .
unfortunately additional data on this issue was nt obtained . in conclusion
we have examined the transition spectrum of a cpb that had an anomalous quadrupling of the spectral lines .
a microscopic model of one or two charged critical current fluctuators coupled to a cpb was used to fit the spectrum .
the fits were in good agreement with the data , reproduced the key features in the spectrum , and allowed us to extract microscopic parameters for the tls s .
our tunneling terms were much smaller than those reported by z. kim , _
et al_.@xcite in their measurements of avoided crossings .
finally , the large fractional change @xmath165 of @xmath3 suggests that the tunnel barrier is non - uniform in thickness with the tls hopping blocking a dominant conduction channel .
fcw would like to acknowledge support from the joint quantum institute and the state of maryland through the center for nanophysics and advanced materials .
the authors would like to thank m. khalil , z. kim , p. nagornykh , k. osborn , and n. siwak for many useful discussions .
43ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/physrevb.72.134519 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.93.077003 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.210503 [ * * , ( ) ] link:\doibase 10.1103/physrevb.72.060506 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.137002 [ * * , ( ) ] link:\doibase 10.1103/physrevb.76.214503 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.93.180401 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.144503 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.050502 [ * * , ( ) ] link:\doibase 10.1103/physrevb.78.144506 [ * * , ( ) ] link:\doibase 10.1103/physrevb.77.180502 [ * * , ( ) ] link:\doibase 10.1038/nature06124 [ * * , ( ) ] link:\doibase 10.1103/physrevb.74.100502 [ * * , ( ) ] link:\doibase 10.1088/0953 - 2048/22/1/015004 [ * * , ( ) ] http://arxiv.org/abs/1007.2577 [ ( ) ] link:\doibase 10.1103/physrevb.80.172506 [ * * , ( ) ] link:\doibase 10.1038/nphys972 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.077001 [ * * , ( ) ] \doibase doi:10.1063/1.3309703 [ * * , ( ) ] link:\doibase 10.1007/bf00660072 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.207001 [ * * , ( ) ] in @noop _ _ , ( , ) \doibase doi:10.1063/1.97738 [ * * , ( ) ] \doibase doi:10.1063/1.97051 [ * * , ( ) ] http://adsabs.harvard.edu/abs/1998phst...76..165b [ * * , ( ) ] link:\doibase 10.1103/physreva.76.012325 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.060501 [ * * , ( ) ] link:\doibase 10.1103/physreva.69.062320 [ * * , ( ) ] link:\doibase 10.1103/physreva.76.042319 [ * * , ( ) ] http://arxiv.org/abs/0805.4429 [ ( ) ] @noop _ _ ( , ) \doibase doi:10.1063/1.89690 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.94.123602 [ * * , ( ) ] _ _ , @noop , ( ) _ _ , @noop , ( ) \doibase doi:10.1063/1.1569986 [ * * , ( ) ] \doibase doi:10.1063/1.2357010 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.184526 [ * * , ( ) ] @noop _ _ ( , ) | we report on the quadrupling of the transition spectrum of an @xmath0 cooper - pair box ( cpb ) charge qubit in the @xmath1{ghz}$ ] frequency range .
the qubit was coupled to a quasi - lumped element al superconducting resonator and measured at a temperature of @xmath2{mk}$ ] .
we obtained good matches between the observed spectrum and the spectra calculated from a model hamiltonian containing two distinct low excitation energy two - level systems ( tls ) coupled to the cpb . in our model , each tls has a charge that tunnels between two sites in a local potential and induces a change in the cpb critical current . by fitting the model to the spectrum
, we have extracted microscopic parameters of the fluctuators including the well asymmetry , tunneling rate , and a surprisingly large fractional change ( @xmath3 ) in the critical current ( @xmath4{na}$ ] ) .
this large change is consistent with a josephson junction with a non - uniform tunnel barrier containing a few dominant conduction channels and a tls that modulates one of them . |
the experiment is conducted in a tow - tank of length 60 m , width 1.8 m and depth 2 m. the plate is towed by a fully automated carriage capable of speeds up to @xmath1 ms@xmath2 .
the plate is 5.0 m long and 1.2 m wide with an elliptical leading edge .
the boundary layer formed on the bottom surface of the plate is tripped using a 1 mm diameter trip wire .
fluorescein is injected through a spanwise oriented slot located immediately downstream of the trip .
the developing boundary layer is then illuminated using a continuous 4 w argon - ion laser which has been fanned into a streamwise / wall - normal sheet .
images are acquired using a redlake y-3 classic high - speed camera , with @xmath3 pixel resolution .
the approximate field of view for all visualisations is @xmath4 mm ( streamwise @xmath5 wall - normal ) .
two flow visualisations are shown : one towed at a slower speed of 0.22 ms@xmath2 and the other at @xmath6 ms@xmath2 . by the trailing edge of the plate
, the boundary layer develops to a friction reynolds number @xmath7 and @xmath8 respectively for the low and high speed experiments .
here @xmath9 , where @xmath10 is boundary layer thickness , @xmath11 is wall - shear velocity and @xmath12 is the kinematic viscosity .
the camera frame - rate is 25 hz for the low speed , and 90 hz for the high speed case . by viewing the visualisations in slow motion
, one can distinguish the origins of the interfacial bulging that forms at the edge of the boundary layer , leading to the well - known highly intermittent fluctuation statistics in this region .
one can also observe eruptions of strongly vortical motions , originating close to the wall , and growing beyond this interface . connected with these events
is the entrainment of irrotational fluid from the freestream into the boundary layer .
these entrainment motions occasionally penetrate deep into the boundary layer , transporting high momentum fluid from the freestream to the near - wall region . by comparing the high and low speed visualisations ,
the impact of reynolds number on the turbulent structure is clearly demonstrated . at the high reynolds number ,
the increased mixing reduces the contrast in the flow visualisation .
one can also observe smaller - scale motions and a more complicated multi - scale interface separating the freestream from the turbulence .
ultimately this unique facility will be developed for high - speed time - resolved particle image velocimetry ( piv ) experiments to enable statistical analysis of the evolving structure of turbulence . in the meantime
, these visualisations illustrate the complexity and beauty of the fluid motion within wall - bounded turbulence . | this fluid dynamics video submitted to the gallery of fluid motion shows a turbulent boundary layer developing under a 5 metre - long flat plate towed through water .
a stationary imaging system provides a unique view of the developing boundary layer as it would form over the hull of a ship or fuselage of an aircraft .
the towed plate permits visualisation of the zero - pressure - gradient turbulent boundary layer as it develops from the trip to a high reynolds number state ( @xmath0 ) .
an evolving large - scale coherent structure will appear almost stationary in this frame of reference .
the visualisations provide an unique view of the evolution of fundamental processes in the boundary layer ( such as interfacial bulging , entrainment , vortical motions , etc . ) . in the more traditional laboratory frame of reference , in which fluid
passes over a stationary body , it is difficult to observe the full evolution and lifetime of turbulent coherent structures .
an equivalent experiment in a wind / water - tunnel would require a camera and laser that moves with the flow , effectively ` chasing ' eddies as they advect downstream @xcite . |
since the discovery of jets in ss433 in 1978 , there has been discussion about their uniqueness . by now the zoo of various jets has enlarged .
there have already been detected hundreds relativistic jets of agns , low velocity jets of young stars and a dozen jets of compact stars .
jets of ss433 stand out among the others for they harbour cool clouds within the relativistic flow .
it is not clear , however , how the clouds could survive in the relativistic jets and be observed .
it is also worth mentioning , that the thermal fraction is likely to be dominant in the jets of ss433 . on the contrary
, the lack of observational evidence of thermal gas in agn jets strongly restricts its volume - filling factor ( @xmath2 ) and implies that it can not be important in the energy budget of these jets ( @xcite ) .
these peculiarities make the jets of ss433 unique .
jets of ss433 are highly collimated , @xmath3.4 , mildly relativistic , @xmath4 cms@xmath5 , and powerful , with kinetic luminosity @xmath6 ergs@xmath5 .
they appear from funnels of the supercritical accretion disk and follow its precession rotation with a half opening angle of about @xmath7 and a period of @xmath8 ( see fig .
[ wall ] ) . at the distance of a few
@xmath9 cm the jets rapidly cool and radiate in the x - ray domain .
the x - ray emission does not show any structure of the jets .
therefore , the x - ray jets are thought to be continuous ( @xcite ) . by contrast
, the optical jets are observed as fragmented ( @xcite ) .
their clumpness may be roughly deduced from energetic considerations .
the gas must be dense enough to radiate the bulk of the emission : @xmath10 and mass loading of the jet by dense gas must approximately satisfy : @xmath11 where @xmath12 is the volume of the jet , @xmath13 is its length , @xmath14 is electron density , @xmath15 is the @xmath16 emission rate .
thus , one finds that @xmath17 @xmath18 and a volume - filling factor of @xmath19 for appropriate parameters : luminosity in @xmath16 line @xmath20 ergs@xmath5 , total jet kinetic luminosity @xmath21 ergs@xmath5 , @xmath22 erg@xmath23s@xmath5 , @xmath24 cm length of _ e_-folding @xmath16 line brightness .
it seems that the jets become lumpy in the gap between the x - ray part and the optical jet beginning at about @xmath25 cm ( @xcite ) .
the gas could condense there due to a thermal instability as brinkmann et al .
( @xcite ) have shown in a simulation of the thermal evolution of the x - ray jets .
the instability takes place only in the interior jet , until a distance of @xmath26 cm .
the parameters of the clouds found in their work are close to those obtained by panferov & fabrika ( @xcite ) from the balmer decrements of the jets : hydrogen density @xmath27 @xmath18 and size of the clouds @xmath28 cm .
it should be noted that these parameters as well as the filling factor @xmath29 found there are more appropriate for the distance of maximum optical brightness of the jets @xmath30 cm .
these parameters will be representative ones in what follows .
the clouds are grouped in clusters with sizes of about @xmath9 cm ( @xcite ) .
still bigger structures are quite possible in the jets as may be deduced from the structure of the jet emission line .
brown et al .
( @xcite ) consider cloudy structure of the jets as a result of a temporal switch on and off of conditions for thermal instability .
they found that the parameters of ss433 are just suitable for the instability to occur .
however , the instability can work with a wide range of the parameters and in ss433 jets the conditions for cloud generation exist rather continuously in time . indeed
, the data on optical emission of the jet ( e.g. @xcite ) demonstrate that the generation of the clouds in the jets is continuous and the disappearance of the optical jets happens very rarely .
the intermittent structure of the jets is natural if the radiating clouds form in such _ in situ _
variable process as the thermal instability .
the optical jets terminate at a distance of about @xmath31 cm ( @xcite ) and then are observed to be quasi - continuous in radio ( @xcite ) . it is observed as separated emission peaks grow and decay at a unchanged wavelength reflecting an evolution of `` bullet '' along the jets ( @xcite ; @xcite ) .
thus , the clouds form in the interior jets and further exist a long time giving a signature of the optical jets .
the actual problem is how do the clouds evolve and survive in the relativistic jets ? as long as the clouds are involved in a motion through a medium , that is , the dense wind of the accretion disk , the applied ram pressure leads to gas - dynamical instabilities that try to disrupt the clouds on a crushing time scale ( e.g. , @xcite ) : @xmath32 where @xmath33 is a cloud density , @xmath34 is a wind density . for the clouds of the jets of ss433 this time is on the order of @xmath35 s. besides this
, the clouds undergo an evaporation in the hotter intercloud medium
. the evaporation characteristic time may be rather short , too ( see eq.([tev ] ) ) .
namely the capability of the clouds to survive permits us opportunity to observe the jets of ss433 in the optical . in this paper
we show that the collisional interaction that seems to be the destructive one is a most necessary condition of the clouds existence . in sect.2
we show that the collisional interaction may be comparable or even more essential than radiation for the thermal state of the clouds . although this has been concluded by another authors ( e.g. , @xcite ) , we use here the last data on the jets parameters ( @xcite ) and on parameters of uv radiation of the system
a possible source of heat of the clouds ( @xcite ) . in sect.3
we discuss evolution and confinement of the clouds .
there are two main ways to heat the clouds : by interaction of jets with the gaseous wind of the supercritical accretion disk and by radiation of the system ( @xcite ) . here
we consider competition of these two mechanisms .
as long as the cloud sound - crossing time is shorter than the dynamical time - scale @xmath36 of the jet , the clouds are in pressure equilibrium with surroundings
. the influence of ram pressure of the wind on the equilibrium is most essential ( see sect.3 ) .
density in the wind scales as @xmath37 , where @xmath38 is the radial distance from the source of the jets . under plausible conditions of approximate constancy of temperature and mass of a cloud
this implies the following dynamical dependences of cloud parameters : @xmath39 where @xmath40 and we have used the values of the cloud parameters at @xmath41 , given in sect.1 . in what follows , we assume these to be valid everywhere along the optical jets .
let us compare the rates of cloud heating by the impinging gas and by radiation .
the jets sweep out the gas of the wind due to the change of the jet direction in the course of the precession and nodding motions and a small amplitude jitters of unclear nature ( @xcite ) .
the energy deposition into a cloud by crossing protons with energy @xmath42 depends on the collisional thickness , measured by the ratio @xmath43 , where @xmath44 @xmath45 is the column density of the cloud , and @xmath46 is the column density required to stop a proton .
the rate of energy loss of a high energy proton in an ionized gas is approximately ( @xcite ) : @xmath47 for temperature @xmath48 , @xmath49 .
the factor @xmath50 , where @xmath51 is the coulomb logarithm . in an ionized plasma
almost all of the lost energy goes into heating the electron gas .
the stopping column density of a proton with energy @xmath52 mev is : @xmath53 when @xmath54 , which is satisfied for the entire jet ( see eq.([colden ] ) ) , the cloud heating rate by the fast protons is : @xmath55 this is lowest estimate .
the magnetic field may be amplified to an equipartition value at the surface of a cloud moving through a medium with a weak magnetic field ( @xcite ) . for assumed in eq.([mf ] ) magnetic field the gyro - radius of a 32 mev proton is @xmath56 cm .
this is smaller than the clouds radius , therefore the trajectories of protons may be tangled and the heat input boosted .
the clouds could also be heated by the collimated radiation from funnels of the accretion disk ( @xcite ; @xcite ) and by the disk uv radiation .
the heating is realized by @xmath57-electrons high energy electrons broken away from an atom due to ionization . already for ionization degree @xmath58 , almost all the energy of @xmath57-electrons is utilized in the heating of the electron gas of a cloud .
we take into account radiation only with frequency beyond the lyman edge .
a cloud intercepts radiation flux with density @xmath59 , where @xmath60 is radius of the radiation source , @xmath61 is the planck function , and relativistic factor @xmath62 is omitted .
the part of the radiation which is absorbed is @xmath63 , where the optical thickness is @xmath64 , the ionized fraction is @xmath65 .
then the radiative heating rate of one cloud is : @xmath66 where @xmath67 is frequency of lyman edge .
we adopt here @xmath68 @xmath69 , for photon with energy @xmath70 in the range 13.6 ev 12.4 kev , for cosmic abundances ( @xcite ) .
parameters of the collimated radiation are not well known since its orientation to the earth is unfavourable . here
we adopt rather rough estimates for its luminosity @xmath71 ergs@xmath5 and temperature @xmath72 kev ( @xcite ; @xcite )
. then the radius of a source of the collimated radiation , in blackbody approach , is @xmath73 cm .
the supercritical accretion disk is a source of near isotropic uv radiation , its parameters in blackbody approach are : @xmath74 k , @xmath75 cm ( @xcite )
. for given parameters and @xmath76 eq.([hr ] ) gives @xmath77 , i.e. heating by uv radiation of the disk could dominate heating by the x - ray collimated radiation .
this ratio might be bigger because the clouds have rather @xmath78 for the uv radiation , however situation is complicated by the screening of the clouds in a jet . the cross - over path length of some particle moving in the direction of the jet velocity across the jet is determined by the curvature of the jet .
the motion of the jet caused by nodding is more rapid than that caused by precession and its rate is @xmath79 rads@xmath5 ( @xcite ) . using this rate we obtain the cross - over length for a jet with fixed pattern @xmath80 cm , with accuracy of factor 2 .
column density along this path is : @xmath81 consequently , we have @xmath82 everywhere in the optical jets and the jets are transparent for the rapid protons .
so , the screening effect is not important for the collisional heating .
meanwhile the screening factor of clouds in the jet @xmath83 may be essential for the uv radiation .
we now determine @xmath84 as the column density of a cloud layer where @xmath85 for the uv radiation .
then the screening factor for the uv radiation is @xmath86 and essentially depends on the ionization degree .
therefore we will consider heating @xmath87 of the cloud by uv radiation as an upper limit of radiative heating ( [ hr ] ) .
really , this heating is bigger than mean heating rate of one cloud , on average over the whole jet , in @xmath88 times ; the opacity of the jet to uv radiation causes a dominance of the collimated radiation in the heating of screened clouds .
the ratio of the heating rate from fast protons ( eq.([hc ] ) ) to the radiation heating rate ( eq.([hr ] ) ) is : @xmath89 where we have used the reduced form of eq.([hr ] ) for @xmath90 , substituting 1 instead of @xmath91 , and used density of wind of the accretion disk : @xmath92 the mass loss rate and wind velocity are taken to be @xmath93y ( @xcite ) and @xmath94 km / s ( from the width of the emission lines ) respectively .
thus the heating of the jet by fast protons is comparable at least with the radiation heating .
their rates become equal at a distance @xmath95 cm , until which the main bulk of the heat input into the jet takes place and equals to 64% of the whole input given by eq.([hj ] )
. the collisional heating may be rather much bigger than radiative one , because the value of @xmath96 is overestimated approximately in @xmath97 times due to the screening effect .
this idea concerning the collisional heating of the jet clouds is observationally supported : @xmath16 emission of the jets is anisotropic , the maximum of its directional pattern is directed to side of jet movement ( @xcite ) .
this implies that some screening effect exists for the impinging protons .
it is possible if a magnetic field tangles trajectories of the protons in a cloud and the stopping column density @xmath46 becomes smaller than that given by eq.([ns ] ) .
however , the directional pattern can not be explained straightforwardly as a result of a head - on collision of the clouds with the wind , because the maximum is inclined to the jet axis .
the inclination of the head maximum is towards the direction of the precessional motion and is about @xmath98 to the jet axis .
this rather shows that a maximum outcome of the emission line radiation is from jet sides moving in the wind . as a result of collision with the wind a cloud decelerates at a rate
: @xmath99 here we used @xmath41 as the lower limit of the integration because evolution of the jet at distances @xmath100 is not well established .
the deceleration ( [ dec ] ) agrees well enough with the observed value of the deceleration @xmath101 ( @xcite ) .
jets sweep out wind gas with a pattern speed of @xmath102
. then the rate of the heating of the whole of optical jet provided by fast protons is : @xmath103 this integral is calculated numerically and dependence of @xmath104 on a geometry of the precessing conical jet is taken into account .
the calculated value of the whole heat @xmath105 agrees with the observable radiation output of the jets in h@xmath106 line @xmath107 ergs@xmath5 ( @xcite ) : a part @xmath108 of absorbed energy is emitted in h@xmath106 .
the small clouds in question would be overheated and their h@xmath106 radiation would have ceased unless the clouds are dense enough to establish a balance between radiative losses and the heating by the fast protons . in the temperature region near @xmath109 k , which is the threshold ionization temperature of hydrogen ,
the efficiency of the radiative cooling @xmath110 strongly depends on temperature and changes by 3 orders of magnitude : @xmath111 @xmath112 ( @xcite ) . from this and eq.([dyn ] ) one finds limits of possible values of the radiative energy loss of the clouds , corresponding to the end and the beginning of the jet : @xmath113 since limits on the heating rate @xmath114 from eq.([hc ] ) correspond to : @xmath115 for the jet end and beginning respectively , the heating rate never can exceed the cooling capability , given by inequality ( [ clim ] ) , and the thermal balance of the clouds can be maintained near a temperature of @xmath109 k throughout the whole of the optical jets .
the lifetime of the clouds in the jets , about 4 days , is much longer than the cloud sound - crossing time .
therefore , the clouds must be in pressure equilibrium with the ambient medium and should not be exposed to destructive processes .
the pressure of the ionizing radiation @xmath116 dyn is much smaller than the gas pressure in a cloud @xmath117 dyn , where we use a temperature of @xmath118 k ( @xcite ) .
therefore , radiation is not important for dynamical balance of a cloud .
the pressure equilibrium of clouds with surroundings imposes constraints on density @xmath119 and on temperature @xmath120 k of the surrounding gas at a distance @xmath41 . in this hot medium
the clouds will evaporate with a timescale of : @xmath121 where @xmath122 is the mass of the cloud , @xmath123 is sound velocity in the hot medium , @xmath124 is some function ( given by equation ( 62 ) in @xcite ) , which is @xmath125 for our case .
the evaporation of the clouds with so short timescale contradicts to the lifetime of the clouds .
therefore the clouds evaporation must be suppressed . on the other hand ,
the ram pressure of the wind is important in pressure balance of the clouds : @xmath126 .
ram pressure of the wind gas seems to be strongest and must govern the clouds pressure . but again , gas - dynamical instabilities are capable of disrupting the clouds on a timescale given by eq.([tcr ] ) .
this is the puzzle of ss433 jets : how can thermal gas survive there ?
the optical jets end at a distance of about @xmath127 cm .
further , at a distance @xmath128 cm , the jets brighten in radio which is a consequence of jet expansion ( @xcite ) .
it seems that the expansion is connected with the termination of the optical jet .
as long as the jets precess and move through the dense wind of the accretion disk they sweep out the wind gas to the surface of the precession cone . after the passage of the jets
the wind fills the jet channels and during the precessional period a region of length @xmath129 cm can be refilled .
consequently , at distances larger than the sweep - out distance the lobe along the precession cone surface is empty ( see fig . [ wall ] ) .
thermal diffusion of wind gas is not capable of filling the lobe : @xmath130 at @xmath131 .
the dependences ( [ dyn ] ) of the clouds parameters on a distance and the clouds heating balance have not peculiarities in point @xmath132 . but external to the clouds conditions abruptly change there .
possibly , this forces the confinement of the clouds to switch off and results in the expansion of the jet . from laws in eqs.([dyn ] ) we derive the approximate size of the cloud @xmath133 cm and the filling factor @xmath134 at the end of the optical jets . with these parameters
the clouds start to expand freely and fill the whole volume of the jets in time : @xmath135 where @xmath136 is sound velocity in the cloud .
the estimated time of expansion corresponds well to the time interval required for the jet to move between the end of the optical part and the radio brightening zone .
the last could develop due to the increase of the rate of generation of relativistic particles , when the clouds expand and turbulence is enhanced .
the _ coincidence _ of lengths of the optical jets and the sweep - out zone indicates that ram pressure of the wind gas causes confinement of the clouds .
the ram pressure switch off is equivalent to a `` hard wall '' which crushes the clouds .
however , the ram pressure by itself is not able to prevent clouds from stripping off the gas and thermal evaporation . on the contrary , it seems from eq.([tcr ] ) that the higher the ram pressure the shorter the lifetime of a cloud . in reality , ionized clouds and their surroundings harbour magnetic fields , which should be important for stability of the clouds .
the magnetic field inferred from radio observations is @xmath137 g at the distance of radio brightening ( @xcite ) .
magnetic flux constancy implies that @xmath138 scales as @xmath139 and we assume magnetic field in the jet is : @xmath140 this field could not confine the clouds having internal pressure @xmath141 dyn .
jones et al . ( @xcite ) showed in their numerical simulations that a `` magnetic shield '' forms around a cloud moving through a medium .
the field lines of the medium are caught by a cloud , stretched and the magnetic field is amplified .
the `` shield '' has magnetic pressure comparable to the ram pressure and is able to quench gas - dynamical instabilities and prevent evaporation of a cloud .
energy of the `` magnetic shield '' is converted from the kinetic energy of the cloud .
so , the magnetic field needed for the confinement of the clouds in the jets of ss433 can be generated by the motion of the clouds through the wind of the accretion disk .
this mechanism of the confinement naturally explains the termination of the optical jet beyond the `` crushing wall '' at a distance @xmath142 , where collisional interaction of the clouds with the wind ceases .
we have considered consequences of collisional interaction of clouds in the jets of ss433 with the wind of the accretion disk .
this interaction appears to be a determining factor of the clouds state and evolution .
this is emphasized by the fact that the clouds exist only in boundaries of the sweep - out zone over which the jet sweeps out the wind .
we propose that in the sweep - out zone the clouds are prevented from destruction essentially by a magnetic field which is amplified due to a collisional interaction with the wind .
the rapid expansion of the clouds after their exit from the sweep - out zone naturally explains the disappearance of the jet hydrogen line emission and the brightening of the jets at radio wavelengths .
the collisional interaction of the jets of ss433 with its surroundings is possible for two reasons : the powerful wind from the supercritical accretion disk and the precession with cone opening being much larger than the jet opening angle .
these factors are necessary for the existence of the cold clouds radiating the hydrogen lines
. their combination may be the clue to the uniqueness of the optical relativistic jets of ss433 .
vermeulen r.c . ,
murdin p.g .
, van den heuvel e.p.j . ,
fabrika s.n . ,
wagner r.m . , margon b. , hutchings j.b . , schilizzi r.t . ,
van kerkwijk m. , van den hoek l.b . , ott e. , angebault l.p .
, miley g.k . , dodorico s. , borisov n. , 1993b , a&a 270 , 204 | the relativistic jets of ss433 are outstanding for their optical thermal radiation .
the radiation is produced by small clouds ( @xmath0 cm ) whose lifetime is about @xmath1 times larger than the gas - dynamical crushing time .
we show that the clouds reside in thermal and dynamical balance as long as they collisionally interact with the wind of the supercritical accretion disk .
the interaction is caused by the precessional movement of the jets and takes place only in the sweep - out zone . beyond the sweep - out zone the interaction ceases and optical jets just terminate
the cloud magnetic field amplified in course of movement through a medium could play a role in containing a cloud .
thus , the clue to the uniqueness of the optical jets of ss433 is thought to be their precessional movement , which provides an opportunity for collisional interaction of the clouds with the wind . |
blazars are active galactic nuclei that have their jet axis oriented at a small angle with respect to the observer @xcite .
they are observationally classified as either flat - spectrum radio quasars ( fsrq ) or bl lacertae ( bl lac ) objects according to the broad line emission in their optical spectra .
recent studies interpreting the differing spectral high - energy ( he ) @xmath2-ray properties of the fsrqs and bl lacs based on physical mechanisms can be found in e.g. , @xcite .
blazars are known to emit non - thermal radiation characterized by a double - peaked spectral energy distribution ( sed ) .
the low - energy component , generally covering radio to uv / x - ray bands , is usually explained as due to synchrotron emission from relativistic electrons in the blazar jet .
the origin of the he component , occurring in the x - ray to @xmath2-ray regime , is still not completely resolved and could be due to emission from a relativistic particle beam consisting of leptons and / or hadrons . in leptonic models , very high energy ( vhe ) photons
are produced by inverse - compton ( ic ) scattering of low - energy photons off the synchrotron - emitting electrons .
the soft seed photons for the ic process can be the synchrotron photons ( synchrotron self - compton , ssc , e.g. , @xcite ) , or they may originate from ambient radiation ( external - compton , ec , e.g. , @xcite ) .
hadronic models include synchrotron emission from protons ( e.g. , @xcite ) and @xmath18-decay from hadronic interactions with subsequent electromagnetic cascades ( e.g. , @xcite ) .
rbs0413 was discovered in the x - ray band ( 1e0317.0 + 1834 ) during the einstein medium sensitivity survey and was optically identified as a bl lac @xcite .
the object was also detected as a radio emitter with the very large array of the national radio astronomy observatory @xcite .
it exhibits significant and variable optical polarization @xcite . having a `` featureless '' optical spectrum @xcite and an estimated synchrotron peak frequency log@xmath19 @xcite , rbs0413 is classified as a high - frequency - peaked bl lac object ( hbl , @xcite ) .
it is located at a redshift of 0.190 @xcite .
the magic collaboration observed rbs0413 in 2004 december2005 february for a livetime of 6.9 hr and reported a vhe flux upper limit of @xmath20 , at 200gev , assuming a power - law spectrum with a photon index of 3.0 @xcite .
veritas observed the source in the 20082009 season and obtained a marginal significance of @xmath21 . in 2009 , _
fermi_-large area telescope ( lat ) detected he emission from the direction of rbs0413 @xcite , triggering new veritas observations .
these new observations , combined with the previous data , resulted in the detection of rbs0413 as a vhe @xmath2-ray emitter in 2009 october @xcite .
veritas is a ground - based @xmath2-ray observatory sensitive to @xmath2 rays with energy between 100gev and 30tev . located at the fred lawrence whipple observatory ( flwo ) near amado , southern arizona , usa ( 1.3 km above sea level , n @xmath22 @xmath23 ,
w @xmath24 @xmath25 ) , the array consists of four imaging atmospheric cherenkov telescopes , each having a diameter of 12 m and a field of view of @xmath26 @xcite . during the 2009 annual shutdown ( july - august ) one of the telescopes was relocated to give a more symmetrical array layout .
in addition , a new mirror - alignment system applied in 2009 spring contributed to an improvement in the point - spread function ( psf ) , with a decrease of @xmath27 in the 80% containment radius @xcite . as a consequence ,
the sensitivity showed a significant improvement , and the observation time required for a @xmath28 crab nebula detection dropped from @xmath29 hr to less than 30 hr . for observations at @xmath30 elevation , the energy resolution is @xmath31 , and the angular resolution , defined as the @xmath32 containment radius , is less than @xmath33 @xcite .
veritas observed rbs0413 for @xmath34 hours in total , using _
mode @xcite , with north , south , east and west wobble positions . after discarding observing runs compromised by bad weather , and a small number affected by hardware problems
, 26 hours remained for analysis .
one third of these data were obtained with the old array configuration ( sep 2008 - feb 2009 , mjd 5473254883 ) and the rest with the new array ( sep 2009 - jan 2010 , mjd 5509255485 ) .
approximately @xmath35 hours of data with the old array were taken under weak moonlight , which leads to a higher energy threshold for those observations . the source elevation in the data set ranges from @xmath36 to @xmath37 , with an average of @xmath38 .
data analysis steps consist of calibration , image parameterization @xcite , event reconstruction , background rejection and signal extraction as described in @xcite . for signal extraction ,
a @xmath39 cut @xcite of 0.0169 , optimized for a point source of 1% strength of the crab nebula , was used .
rbs0413 is a weak source in the vhe regime . using a `` reflected - region '' background estimation @xcite ,
an excess of 180 events and a significance of @xmath40 are obtained , for the source location at @xmath41 and @xmath42 ( j2000 coordinates ) .
the veritas signal is consistent with a point source , and we name the object verj0319 + 187 .
the energy distribution of @xmath2-ray events extends from @xmath43gev to @xmath44tev ( see table [ tab : verspectrum ] for a list of spectral data points ) and is well described by a power - law function , @xmath45 .
the best fit is obtained with photon index @xmath46 and flux normalization @xmath47 @xmath48 @xmath6 @xmath7 at 0.3 tev , with a value of @xmath49 per degree of freedom ( @xmath49/dof ) of 0.14/2 ( see figure [ fig : verspectrum ] ) .
the integral flux above 250gev is @xmath4 @xmath6 @xmath7 , corresponding to a flux level of approximately @xmath28 the flux of the crab nebula .
no significant flux variability is detected ( see figure [ fig : lightcurves ] top panel and caption for details of the light - curve analysis ) . an upper limit ( 99% confidence level ) on the fractional variability amplitude ( @xmath50 , @xcite ) yields @xmath51 .
based on our calculations , all three models are good at describing the observed data .
it appears that if the criterion of equipartition is taken as a reasonable measure of successful blazar emission models , ssc+ec is preferred over ssc for this hbl , which seems to be in contrast with some previous blazar studies .
see @xcite for arguments relating the presence of an ec component with the blazar sequence and @xcite for a discussion of issues encountered in explaining blazar seds with a simple one - zone homogeneous ssc model . on the other hand
, we can not discriminate between leptonic and lepto - hadronic mechanisms , since the ssc+ec and lepto - hadronic models provide equally reasonable descriptions for the observed non - thermal continuum , and we did not detect variability in the he and vhe regimes given the limited statistics . since the synchrotron cooling timescales for electrons and protons are different , the detection of intraday variability would be harder to explain with a lepto - hadronic scenario and would accordingly favor a purely leptonic scenario .
therefore , any future observation of rapid variability would be helpful in distinguishing between the ssc+ec and lepto - hadronic models .
the veritas research is supported by grants from the us department of energy office of science , the us national science foundation , the smithsonian institution , and the nasa _ swift _ guest investigator program , by nserc in canada , by science foundation ireland ( sfi 10/rfp / ast2748 ) , and by stfc in the uk .
we acknowledge the excellent work of the technical support staff at the flwo and at the collaborating institutions in the construction and operation of the instrument . the _ fermi_-lat collaboration acknowledges generous ongoing support from a number of agencies and institutes that have supported both the development and the operation of the lat as well as scientific data analysis .
these include the national aeronautics and space administration and the department of energy in the united states , the commissariat lenergie atomique and the centre national de la recherche scientifique / institut national de physique nuclaire et de physique des particules in france , the agenzia spaziale italiana and the istituto nazionale di fisica nucleare in italy , the ministry of education , culture , sports , science and technology ( mext ) , high energy accelerator research organization ( kek ) and japan aerospace exploration agency ( jaxa ) in japan , and the k. a. wallenberg foundation , the swedish research council and the swedish national space board in sweden .
additional support for science analysis during the operations phase is gratefully acknowledged from the istituto nazionale di astrofisica in italy and the centre national detudes spatiales in france .
abdo , a. a. , ackermann , m. , agudo , i. , et al .
2010a , apj , 716 , 30 abdo , a. a. , ackermann , m. , ajello , m. , et al .
2010b , apj , 715 , 429 aharonian , f. a. 2000 , new astron . , 5 , 377 aharonian , f. a. 2001 , a&a , 370 , 112 albert , j. , aliu , e. , anderhub , h. , et al .
2008 , apj , 681 , 944 atwood , w. b. , abdo , a. a. , ackermann , m. , et al . 2009 , apj , 697 , 1071 barnes , j. , 1993 , a beginner s guide to using iraf ( tucson , az : national optical astronomy observatory ) bttcher , m. 2010 , in fermi meets jansky - agn at radio and gamma - rays , ed .
t. savolainen , e. ros , r. w. porcas , & j. a.zensus ( max - planck - institut fr radioastronomie , bonn , germany ) , 41 burrows , d. , hill , j. e. , nousek , j. a. , et al .
2005 , ssrv , 120 , 165 cash , w. 1979 , apj , 228 , 939 daniel , m. k. , et al .
2008 , in proc .
cosmic ray conf .
r. caballero , j. c. dolivo , g. medina - tanco , l. nellen , f. a. snchez , & j. f. valds - galicia ( universidad nacional autnoma de mxico , mexico city , mexico , 2008 ) , 1325 dermer , c. d. , & schlickeiser , r. 1993 , apj , 416 , 458 finke , j. d. , razzaque , s. , & dermer , c. d. 2010 , apj , 712 , 238 gehrels , n. , chincarini , g. , giommi , p. , et al .
2004 , apj , 611 , 1005 ghisellini , g. , celotti , a. , fossati , g. , maraschi , l. , & comastri , a. 1998 , mnras , 301 , 451 ghisellini , g. , maraschi , l. , & tavecchio , f. 2009 , mnras , 396 , l105 gioia , i. m. , maccacaro , t. , schild , r. e. , et al .
1984 , apj , 283 , 495 hillas , a. m. , 1985 , in proc .
cosmic ray conf . 3 held at la jolla , usa , august 11 - 23 , 1985 , ed . f. c. jones , j. adams & g. m. mason ( nasa conf .
, nasa cp-2376 ) , 445 holder , j. , acciari , v. a. , aliu , e. , et al .
2008 , aip conf .
proc . 1085 , high energy gamma - ray astronomy , ed . f. a. aharonian , w. hofmann , f. m. rieger ( melville , ny : aip ) 657 kalberla , p. m. , burton , w. b. , hartmann , d. , et al .
2005 , a&a , 440 , 775 maraschi , l. , ghisellini , g. , & celotti , a. 1992 , apj , 397 , l5 massey , p. , & davis , l. e. , 1992 , a user s guide to stellar ccd photometry with iraf , ( tucson , az : national optical astronomy observatory ) mattox , j. r. , bertsch , d. l. , chiang , j. , et al .
1996 , apj , 461 , 396 mccann , a. , hanna , d. , kildea , j. , & mccutcheon , m. 2010 , astropart .
phys . , 32 , 325 mcke , a. , protheroe , r. j. , engel , r. , rachen , j. p. , & stanev , t. 2003 , astropart .
, 18 , 593 nieppola , e. , tornikoski , m. , & valtaoja , e. 2006 , a&a , 445 , 441 nolan , p. l. , abdo , a. a. , ackermann , m. , et al .
2012 , apj , in press ong , r. a. , & fortin p. 2009 , atel , 2272 padovani , p. , & giommi , p. 1995 , apj , 444 , 567 perkins , j. s. , et al . 2009 , 1st fermi symposium , washington , d.c .
, november 2 - 5 2009 , ed .
d. j. thompson ( slac national accelerator laboratory , menlo park , ca ) , econf proceedings c091122 poole , t. s. , breeveld , a. a. , page , m. j. , et al .
2008 , mnras , 383 , 627 roming , p. w. a. , koch , t. s. , oates , s. r. , et al .
2009 , apj , 690 , 163 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , apj , 500 , 525 stocke , j. t. , liebert , j. , schmidt , g. , et al .
1985 , apj , 298 , 619 stocke , j. t. , morris , s. l. , gioia , i. , et al . 1990 ,
apj , 348 , 1 stocke , j. t. , morris , s. l. , gioia , i. , et al .
1989 , in bl lac objects , ed .
l. maraschi , t. maccacaro & m .- h .
ulrich ( springer - verlag ) , 242 urry , c. m. , & padovani , p. 1995 , pasp , 107 , 803 vaughan , s. , edelson , r. , warwick , r. s. , & uttley , p. 2003
, mnras , 345 , 1271 xiang , y. , & dai , b .- z .
2007 , pasj , 59 , 1061 zacharias , n. , monet , d. g. , levine , s. e. , et al .
2005 , vizier on - line data catalog : i/297 | we report on the discovery of high - energy ( he ; @xmath0gev ) and very high - energy ( vhe ; @xmath1gev ) @xmath2-ray emission from the high - frequency - peaked bl lac object rbs0413 .
veritas , a ground - based @xmath2-ray observatory , detected vhe @xmath2 rays from rbs0413 with a statistical significance of 5.5 standard deviations ( @xmath3 ) and a @xmath2-ray flux of @xmath4 @xmath5 @xmath6 @xmath7 ( @xmath8 of the crab nebula flux ) above 250gev . the observed spectrum can be described by a power law with a photon index of @xmath9 .
contemporaneous observations with the large area telescope ( lat ) on the _ fermi _ gamma - ray space telescope detected he @xmath2 rays from rbs0413 with a statistical significance of more than @xmath10 , a power - law photon index of @xmath11 and a @xmath2-ray flux between 300mev and 300gev of ( @xmath12 ) @xmath13 10@xmath14 photons m@xmath15 s@xmath16 .
we present the results from _ fermi_-lat and veritas , including a spectral energy distribution modeling of the @xmath2-ray , quasi - simultaneous x - ray ( _ swift_-xrt ) , ultraviolet ( _ swift_-uvot ) and @xmath17-band optical ( mdm ) data .
we find that , if conditions close to equipartition are required , both the combined synchrotron self - compton / external - compton and the lepto - hadronic models are preferred over a pure synchrotron self - compton model . |
granular materials such as sand and powders have generated much interest of late .
such an ensemble of particles with macroscopic size is challenging since it may behave as a solid , a liquid or a gas .
size separation , pattern formation , avalanches , compaction and convection are just a few examples of the wide array of observed phenomena@xcite .
flow underlies most of these phenomena and therefore , theoretical studies so far focus on formulating a hydrodynamical description appropriate to sand@xcite .
these theories , stemming from the boltzmann equation , depend on the assumption of `` molecular chaos '' , _
i.e. _ , the assumption that no interparticle correlations exist .
this assumption is far from obvious . as a dissipative dynamical system , a granular system has attractors in its phase space , which may cause correlations between particles . under certain conditions ,
these attractors lead to a singularity , inelastic collapse , which can not be explained by hydrodynamics . in one dimension , these attractors are so strong that hydrodynamics breaks down for the entire parameter space@xcite in a confined geometry , all particles , save one , form a practically stationary clump against the elastic wall , while the remaining particle moves rapidly back and forth between the clump and the heated wall .
such a state clearly violates partition of energy . in this study , we investigate the corresponding situation in two - dimensions@xcite .
we consider inelastic hard spheres in a box where one wall is kept at a fixed temperature and the other three are reflecting ( see fig .
[ fig : syspict ] ) .
energy input at the heated wall balances the dissipation due to inter - particle collisions and the system can achieve a steady state . unlike the one - dimensional situation , the density and the temperature profiles are smooth functions of the distance from the heated wall . in the steady - state , the momentum balance equations imply that the pressure is constant throughout the system .
particles move faster close to the energy source and more slowly deeper inside the system due to energy loss .
thus the density is greater farther away from the wall to maintain a constant pressure .
density variations may be a consequence of the above mechanism or an effect of the intrinsic attractors in the system which build up correlations among particles .
assuming that the particles are not coherent in the quasi - elastic limit , we derive a differential equation that describes the density variation throughout the system .
heuristic expressions for the mean - free path , the equation of state , and the thermal conductivity are incorporated into the energy flux balance equation to obtain a closed equation describing the steady - state density profiles .
the theory compares well with simulational results over a wide range of densities in the quasi - elastic limit .
= our theory is appropriate for a steady - state which is very close to an equilibrium state .
we expect this to be applicable only in the nearly elastic limit .
we argue that in the complementary situation of stronger inelasticity , hydrodynamics may still be relevant . to analyze the nature of the steady - state ,
we study velocity distributions throughout the system .
we observe that the velocity distributions exhibit scaling .
this observation is then used to obtain a qualitative description of the behavior of the system .
granular materials are different from ordinary fluids or gases in that the diameter of a particle may be comparable to the collisional mean free path . to formulate a hydrodynamic theory ,
it is necessary to describe how the mean free path , the pressure , and the thermal diffusivity depend on the grain diameter , the number density , and the temperature . in this section ,
we obtain heuristic expressions for these quantities .
let us denote the number density of grains by @xmath0 , the mean free path by @xmath1 , and the grain diameter by @xmath2 . without loss of generality
, we set the particle mass and the boltzmann constant to unity : @xmath3 .
the granular temperature can be defined as the average kinetic energy per particle : @xmath4 .
this quantity is well defined for an equilibrium state in which the particles have a symmetric velocity distribution .
in such a case , @xmath5 , can be used as an approximate value for the average grain speed . in our system , quantities such as temperature , density , etc .
, are position dependent , and there is not a global equilibrium .
however , in the quasi - elastic limit , we expect the system to be very close to local thermal equilibrium .
let @xmath6 be the distance from the heated wall .
in the steady state , all quantities vary only in this transverse direction .
the energy balance equation can therefore be written @xmath7 where @xmath8 is the energy flux in the @xmath6 direction . the sink term @xmath9 accounts for the energy lost per unit area per unit time due to inelastic collisions .
the energy flux is induced by a temperature gradient , @xmath10 , where the coefficient of thermal diffusivity , @xmath11 , is positive .
consequently , we find @xmath12 since collisions between grains are inelastic , kinetic energy is continually transferred into heat .
for simplicity we neglect rotation and thus , the degree of inelasticity can be parameterized by @xmath13 .
when two particles collide , their tangential velocities are unchanged , while the relative normal velocity is decreased by a factor of @xmath14 , where the negative sign merely indicates that they move apart after a collision .
using momentum conservation one can write the final velocities ( indicated by primes ) in terms of the initial velocities @xmath15 in the above equation , the subscript @xmath16 denotes the velocity component along the line connecting the centers of particles 1 and 2 .
the energy lost in each collision is therefore @xmath17 in this study , we focus on the quasi - elastic limit , i.e. @xmath18 .
physically , this limit is relevant to hard particles such as glass or steel beads . using the above expression for the energy dissipated in each collision
, one can estimate the sink term , @xmath9 , the mean energy lost per unit area per unit time .
consider a particle moving with speed @xmath19 . during each collision
, it loses , on average , energy proportional to @xmath20 . in unit time , it collides roughly @xmath21 times . in unit area
, there are @xmath0 particles , and consequently @xmath22 suppose there is a temperature gradient in the @xmath6-direction . as a result , there will be an energy flux along this axis . to calculate this flux ,
let us consider the number of particles crossing a line perpendicular to this direction in a time interval @xmath23 .
we define `` crossing the line '' as having any part of the particle over the line during this time interval .
to cross the line within @xmath23 , a particle with speed @xmath19 must have its rightmost point @xmath24 or closer to the line .
thus , only grains in an area ( @xmath25 ) to the left of the line can pass the line from the left in @xmath23 .
in fact , only one half of these particles are moving to the right , so the number of particles crossing the line per unit cross - sectional length in @xmath23 is @xmath26 . in a steady state ,
the number of crossing events from each side must balance .
any energy flux is due to the fact that particles coming from the right are at a different temperature ( @xmath27 ) than those from the left ( @xmath28 ) and thus , @xmath29 .
consider two grains on opposite sides of the line .
their centers are approximately a distance @xmath30 apart , so the temperature difference is roughly @xmath31 .
the coefficient of thermal conductivity is therefore @xmath32 .
a natural choice for @xmath23 is the typical collision time @xmath33 .
this choice is small enough to avoid multicounting and is sufficiently large to ensure that heat transfer does occur .
our heuristic picture therefore estimates the thermal diffusivity by @xmath34 this is a rough approximation ; the actual prefactors depend on the velocity distribution of the grains .
therefore , we generally assume @xmath35 the energy balance equation ( [ eq : m2 ] ) takes the form @xmath36 = ( 1-r^2)\frac{\rho t^\frac{3}{2}}{\gamma l } \label{eq : m3}\ ] ] where @xmath37 is the ratio of prefactors in the expressions for @xmath9 and @xmath11 . in the following subsections ,
we discuss how to obtain the dimensionless coefficients @xmath38 and @xmath37 self - consistently .
to proceed , it is necessary to relate @xmath0 and @xmath39 through the equation of state .
additionally , the mean free path @xmath1 must be expressed in terms of @xmath0 .
for the system to be in a steady state , the pressure must be constant throughout .
the equation of state relates the pressure and the temperature to the density .
for example , in the low density limit , the ideal gas law holds , @xmath40 . using @xmath41 and @xmath42
, one has @xmath43 on the other hand , in the high density limit , the mean free path , which is simply the interparticle spacing , is much less than the particle diameter ( @xmath44 ) . denoting the close - packing density by @xmath45 , one finds @xmath46 in this limit , the center of a grain is confined to an area of the order of @xmath47 , so the entropy per particle , @xmath48 , equals @xmath49 plus a function of temperature . from eq .
( [ eqn : denl ] ) , @xmath48 depends on density only through the term @xmath50 . using maxwell s relation , @xmath51
we obtain the pressure in the limit @xmath52 , @xmath53 we therefore propose the following interpolation formula for the pressure @xmath54 indeed , in the limits of high and low density this expression reduces to ( [ eqn : pressld ] ) and ( [ eqn : presshd ] ) , respectively .
it is useful to compare this expression with the van der waal s equation of state which takes into account long range attraction and hard core repulsion @xcite .
for an inelastic gas of hard spheres , there are no long range forces , and the van der waal s pressure for a two - dimensional gas reads @xmath55 for @xmath56 . in the low density limit the pressure given by eq .
( [ eqn : pressgen ] ) agrees with the van der waal s expression to second order in @xmath57 .
furthermore , one can also compare ( [ eqn : pressgen ] ) to tonks series expansion @xcite for the pressure of a two - dimensional gas of hard spheres , @xmath58/ [ 1 - 1.307(\rho/\rho_c)^3 + 0.307(\rho/\rho_c)^4]$ ] , valid for all densities . over the entire density range ,
the two expressions differ by less than 1.3% .
in contrast , we found that the van der waal s formula is inadequate for describing the high density limit .
hence , we use the interpolation formula ( [ eqn : pressgen ] ) for the pressure .
the mean free path can be expressed in terms of the density and the diameter . in the low density limit one has @xmath59 while in the high density limit
, eq . ( [ eqn : denl ] ) gives @xmath60 again , we use these high and low density limits to interpolate a general expression for the mean free path . using the 2d close - packing value @xmath61
, we find @xmath62 where @xmath63 .
( [ eqn : pressgen ] ) and ( [ eqn : lgen ] ) express the temperature and the mean free path in terms of the density . substituting these expressions into eq .
( [ eq : m3 ] ) yields a second order differential equation for @xmath0 . using for convenience the variable @xmath64
, we have @xmath65 = \nonumber\\ & & \hspace{-6cm}\frac{32}{3d^2 } \frac{1-r^2}{\gamma } \frac{z - a}{(z+1)^{3/2 } } \sqrt{\frac{z-1}{z}}. \label{eqn : zgen}\end{aligned}\ ] ] this equation is complemented by the boundary conditions @xmath66 where @xmath67 is the width of the system and @xmath68 is the total number of particles .
the latter condition merely reflects conservation of particles .
the former condition is a consequence of the fact that the temperature gradient vanishes at the elastic wall . in principle ,
@xmath37 and @xmath38 are two dimensionless factors which can be calculated exactly from the velocity distribution .
however , this distribution is poorly understood . nevertheless , it is still possible to estimate these prefactors by comparing the theoretical predictions of eq .
( [ eqn : bcnorm ] ) with numerical simulations in the limit of high and low densities . in both of these extreme cases
the governing equation ( [ eqn : zgen ] ) can be solved analytically . in the high density limit , @xmath69 , or equivalently , @xmath70 , eq .
( [ eqn : zgen ] ) reduces to @xmath71 .
it is convenient to write this equation in terms of the temperature , @xmath72 . from ( [ eqn : pressgen ] ) , @xmath73 when @xmath70 , and the temperature obeys @xmath74 with @xmath75 . solving this equation subject to the boundary condition @xmath76 at @xmath77 gives the temperature profile @xmath78 far from the elastic wall , @xmath79 , the temperature decays exponentially in agreement with haff s calculation @xcite @xmath80 eq . ( [ eqn : lgen ] ) shows that in the high density limit , @xmath81 . both the temperature and the mean free path decay exponentially with the distance from the heated wall .
the decay length @xmath82 is much larger than the mean free path .
for the continuum description to be valid , @xmath82 must also be much larger than the diameter of a particle , @xmath2 , i.e. , @xmath83 since the prefactor @xmath37 is of order unity , we learn that the continuum theory is valid only in the quasi - elastic limit , @xmath84 .
( note that the restriction @xmath84 is stricter than @xmath18 . ) to test the theoretical predictions , we performed an event - driven simulation of the system@xcite .
the heated wall is implemented in such a way that for @xmath85 , particles have a boltzmann velocity distribution with an average temperature equal to one .
specifically , any particle that collides with the wall at @xmath86 is ejected with a positive @xmath87 drawn from the probability distribution @xmath88 and a @xmath89 drawn from @xmath90@xcite . however , the behavior of particles in the bulk of the system is independent of the details of the boundary condition ( see fig . [
fig : vxscal ] below ) .
this is not surprising , but rather a necessity for a thermodynamic theory to be valid .
numerical simulations confirm the exponential decay in the quasi - elastic limit .
furthermore , the decay length @xmath82 can be measured for various degrees of inelasticity @xmath91 ( see fig . [
fig : denscal ] ) .
we verified that indeed @xmath92 as suggested by eq .
( [ eqn : hdtemp ] ) . the value of the prefactor @xmath93 can be found from the simulations as well .
= we now turn to the low density limit . here , @xmath94 , and eq .
( [ eqn : zgen ] ) reduces to @xmath95 . as a result ,
@xmath96 where @xmath97 , and we have used the boundary condition of eq .
( [ eqn : bcnorm ] ) and the notation @xmath98 .
simulations show that the density near the heated wall is significantly smaller than the density at the elastic wall .
thus , for @xmath99 and @xmath100 , we find the following approximate solution @xmath101 the inverse of the density , and hence the temperature , decays linearly with @xmath6 in the low density regime near the heated wall .
that the low density decay length @xmath102 is similar to the high density decay length @xmath103 , reflects the fact that the underlying differential equation is second order . in the nearly elastic case ( @xmath18 ) , @xmath104 is close to @xmath105 , and eq .
( [ eqn : lddzdx ] ) can be integrated exactly @xmath106_{z_0}^{z(x)}.\ ] ] comparing this prediction with low density numerical simulations allows us to determine @xmath107 at various @xmath91 values ( see fig .
[ fig : dilscal ] ) and verify that indeed @xmath108 .
the constant of proportionality and the previously calculated @xmath37 yields @xmath109 .
= in the low density limit , the equation of state @xmath110 , implies that @xmath111 . for the state to be locally very close to equilibrium , the mean free path must be much less than the length scale over which the temperature is changing , i.e. , @xmath112 .
this condition is @xmath113 when the temperature is of order unity .
since @xmath38 and @xmath37 are also of order unity , the hydrodynamic description is again valid only in the quasi - elastic regime ( @xmath84 ) .
our treatment so far has concentrated on either the high or the low density limit , where analytical expressions were possible . for systems that include both high and low density regions , eq .
( [ eqn : zgen ] ) can be solved numerically using the boundary and normalization conditions ( [ eqn : bcnorm ] ) and the previously calculated values for @xmath37 and @xmath38 . by examining a number of simulations that include a range of densities we determined that the optimal values for the prefactors are @xmath114 and @xmath115 .
while this value of @xmath37 is consistent with the value obtained in the high density calculation , this @xmath38 value is slightly lower than our prediction . a typical system with @xmath116 is shown in fig . [ fig : samp ] , and
it is seen that the predictions of hydrodynamic theory match the numerical data over wide density variations . for this simulation ,
the ratio @xmath117 is approximately @xmath118 , which is not of order unity , so the condition of eq .
( [ eq : valid2 ] ) is satisfied .
the calculation in the previous section assumes that the steady state is very close to equilibrium , and that pressure and temperature can be used to describe the system .
this requires that condition ( [ eq : valid1 ] ) is satisfied in the high density limit , or condition ( [ eq : valid2 ] ) is satisfied in the low density limit . an important question is : when does the behavior of the system changes qualitatively ?
one such transition occurs when @xmath91 becomes low enough for the system to undergo inelastic collapse . here
strong correlations and large density variations develop , and applying hydrodynamics becomes impossible .
even when @xmath91 is slightly higher than the critical value for inelastic collapse , the attractors mentioned in the introduction may still be strong enough to build correlations . where this breakdown occurs is determined by the degree of inelasticity , the density and the total number of particles@xcite .
for each pair of values of density and total number of particles , there is a value of @xmath91 which divides two different kinds of behavior : loose sand and coherent sand .
however , the theory developed in the previous section is strictly for the quasi - elastic limit , @xmath18 , so there may exist systems that , although elastic enough to avoid inelastic collapse , still have @xmath91 far enough from one that the hydrodynamics do not apply .
since correlations between particles are built up through inelastic collisions , high density regions are more liable to inelastic collapse@xcite , while for low density regions , correlations are harder to establish .
therefore , we will investigate the low density limit in order to observe the breakdown of the hydrodynamic description as the degree of inelasticity increases .
when the temperature variation within a mean free path is significant , the system is unable to reach local equilibrium .
therefore , particles carrying energy away from the heated wall can not share this energy with the slow particles returning from the higher density region near the elastic wall
. this inefficient mixing leads to a temperature gap the average energy of particles with @xmath119 is greater than that of the particles with @xmath120 ( see fig . [
fig : tempdiff ] ) .
furthermore , near the heated wall , the temperature drops by approximately @xmath121 over a mean free path , which suggests that such a system will be unable to reach local thermal equilibrium ( see fig .
[ fig : mfp ] ) .
= = the probability distribution functions for @xmath87 ( the velocity perpendicular to the heated wall ) provide an illuminating way to measure this deviation from equilibrium ( see fig . [
fig : vxskew ] ) . note the asymmetry : the @xmath122 tail is longer than the @xmath120 tail .
this is consistent with our understanding that the @xmath122 particles have more energy than the @xmath120 ones .
a rough quantification of the deviation from equilibrium is provided by condition ( [ eq : valid2 ] ) : when @xmath117 is of order unity , the theory breaks down . for @xmath123 ( the value used in the figures ) , this quantity is approximately @xmath124 . note that this condition involves @xmath91 only and is not sensitive to the local density @xmath0 .
this suggests that the behavior of regions of the system with different densities should be similar .
in fact , we observed that the velocity distribution function obeys scaling ( see fig .
[ fig : vxscal ] ) , i.e. , @xmath125 the function @xmath126 is independent of boundary conditions ( see fig . [
fig : vxscal ] ) .
thus @xmath127 , where the constant of proportionality depends only on the shape of the function @xmath126 .
note that this shape will depend on @xmath91 , the degree of inelasticity ; the more inelastic the particles are , the more skewed the velocity distribution is .
the probability distribution for the @xmath128-component of the velocity behaves similarly : @xmath129 but here @xmath130 is a symmetric , nearly gaussian , function .
additionally , the same velocity scale @xmath131 characterizes the transverse and the longitudinal velocity distributions .
therefore , while at each position @xmath6 there is no longer a single hydrodynamic temperature , there is a well defined characteristic velocity scale , @xmath131 , so that the granular temperature , @xmath132 , is proportional to @xmath133 .
this scaling suggests that , although the system may deviate significantly from equilibrium , it can still be treated using some of the tools of the previous sections .
specifically , return to the equation for energy balance ( [ eqn : enerbal ] ) : @xmath134 , where @xmath9 is the energy loss due to collisions per unit time per unit area , and @xmath8 is the energy flux , the heat transfer per unit cross - sectional length per unit time .
as discussed previously , @xmath135 while the heat flux is approximately @xmath136 conservation of momentum flux suggests that @xmath137 is constant , and the energy balance equation gives @xmath138 which indicates that @xmath139 , and hence the temperature , depends linearly on @xmath6 .
this prediction is consistent with our simulational results ( fig .
[ fig : gandg2 ] ) .
in this work , we examined the steady - state behavior of a weakly inelastic two - dimensional driven granular system .
we found that a hydrodynamic formulation provides a satisfactory description of the near - equilibrium behavior of the system in the quasi - elastic limit .
we extended haff s theory to the low density limit and found that the corresponding temperature profile varies linearly in space .
for slightly higher inelasticities , the system is no longer close to equilibrium in the low density limit .
however , the scaling behavior of the velocity distributions suggests that a hydrodynamic treatment can still be useful in describing the system . in this non - equilibrium regime
, we found that the particle velocity distributions were non - gaussian .
indeed , deviations from normal distributions have been observed in theoretical @xcite and experimental studies@xcite .
in addition to this variation in the velocity distributions , systems of this sort can produce highly inhomogeneous spatial distributions , as has been noted elsewhere in one @xcite and two @xcite dimensions .
a recent experimental study examined the spatial distribution of hard particles in two dimensions in the presence of an energy input @xcite .
their data is in qualitative agreement with our theoretical predictions , and inhomogeneous spatial distributions reminiscent of fig .
( 1 ) are observed .
* acknowledgements : * we are indebted to leo kadanoff for stimulating our interest in this problem .
we also thank bill young , tom witten , sergei esipov and thorsten pschel for useful discussions .
this work was supported by the national science foundation under awards # dmr-9415604 and # dmr-9208527 and the mrsec program , award # dmr-9400379 .
99 h. m. jaeger , s. r. nagel , and r. p. behringer , phys . today * 49 * , 32 ( 1996 ) .
k. haff , j. fluid mech . * 134 * , 401 ( 1983 ) .
j. t. jenkins and m. w. richman , j. fluid mech . *
192 * , 313 ( 1988 ) .
y. du , h. li , and l. p. kadanoff , phys .
lett . * 74 * , 1268 ( 1995 ) .
e. l. grossman and b. roman , phys .
fluids * 8 * , 3218 ( 1996 ) .
j. maddox , nature * 374 * , 11 ( 1995 ) .
s. e. esipov and t. pschel , `` boltzmann equation and granular hydrodynamics , '' preprint .
h. hayakawa , s. yue , and d. c. hong , phys .
lett . * 75 * , 2328 ( 1995 ) .
l. tonks , phys .
50 * , 955 ( 1936 ) .
i. goldhirsch and g. zanetti , phys .
, 1619 ( 1993 ) .
s. mcnamara and w. r. young , phys .
e * 53 * , 5089 ( 1996 ) .
y. d. lan and a. d. rosato , phys .
fluids * 7 * , 1818 ( 1995 )
. y. h. taguchi and h. takayasu , europhys .
lett . * 30 * , 499 ( 1995 ) .
s. warr , g. t. h. jaques , and j. m. huntley , powder tech .
* 81 * , 41 ( 1994 ) .
s. mcnamara and j .- l .
barrat , `` the energy flux into a fluidized granular medium at a vibrating wall , '' preprint .
a. kudrolli , m. wolpert , and j. p. gollub , `` cluster formation due to collisions in granular material , '' preprint . | we study steady - state properties of inelastic gases in two - dimensions in the presence of an energy source .
we generalize previous hydrodynamic treatments to situations where high and low density regions coexist .
the theoretical predictions compare well with numerical simulations in the nearly elastic limit .
it is also seen that the system can achieve a nonequilibrium steady - state with asymmetric velocity distributions , and we discuss the conditions under which such situations occur .
pacs numbers : 05.20.dd , 47.50.+d , 81.35.+k 2 |
black holes are perhaps the most tantalizing objects in general relativity . recently
, the study of black holes in a background anti - de sitter spacetime has been well motivated from developments in string / m - theory , which naturally incorporate black holes as solitonic d - branes , or simply branes as the higher - dimensional progenitors of black holes .
an intriguing example of this is the conjectured duality @xcite between string theory on @xmath2 background and @xmath3 super yang - mills theory in four dimensions , and in particular , witten s interpretation @xcite of the hawking - page phase transition between thermal ads and ads black hole @xcite as the confinement - deconfinement phases of the dual gauge theory defined on the asymptotic boundaries of the ads space .
much effort has been put into the weak ads gravity regime , analyzing the implications of ads black holes on dual ( gauge ) theories at non - zero temperature , using the conjectured ads / cft correspondence . in this context ,
the most interesting black hole solutions are presumably the five dimensional kerr - ads solutions for a stationary black hole @xcite .
the thermodynamics of ads quantum gravity has been extensively used to infer the thermodynamics of quantum field theory in the large @xmath4 ( or weak field ) limits , with an ads gravity dual , such as schwarzschild - ads @xcite , kerr - newman - ads @xcite and hyperbolic - ads @xcite black holes .
the kerr metric @xcite is a simple explicit exact solution of the einstein vacuum equations describing a rotating black hole in a four - dimensional flat spacetime .
shortly after kerr s discovery , carter @xcite provided an elegant generalization of the kerr solution in four - dimensional de - sitter and anti - de sitter backgrounds .
a higher dimensional generalization of kerr metric in a flat background was given by myers and perry @xcite .
but its generalization to five and higher dimensions with a non - zero cosmological constant was given , only recently , by hawking _
et al . _
@xcite and gibbons _
et al . _
there has been recent interest in constructing the analogous charged rotating solutions in gauged supergravity in four , five and seven dimensions @xcite , and also on non - uniqueness @xcite of those solutions in five and higher dimensions .
the study of non - charged rotating ( kerr ) black holes is interesting at least for two reasons .
firstly , the thermodynamics of kerr black holes , in a background ads space , can give rise to interesting descriptions in terms of cfts defined on the ( conformal ) boundary of ads space , leading to a better understanding of the ads / cft correspondence @xcite . secondly ,
astronomically relevant black hole spacetimes are , to a very good approximation , may be described by the kerr metric . as not much is known about the stability of kerr - ads black holes in higher dimensions , in this paper we study the thermodynamic stability of these black holes in five and higher dimensions .
we also investigate the gravitational stability of a background kerr-(a)ds spacetime under metric perturbations .
the layout of the paper is as follows .
we begin in sec .
ii by outlining the ( anti)-de sitter background metrics in @xmath5 dimensions and their generalizations to kerr - ads solutions . in sec .
iii we pay special attention to the thermodynamic stability of kerr - ads black holes by studying the behavior of hawking temperature , free energy and specific heat in various dimensions . in section
iv we study the gravitational stability of background kerr-(a)ds metrics under linear tensor perturbations .
the stability of a rotating anti - de sitter background spacetime in dimensions higher than four was not previously studied .
our linearized perturbation equations have other interesting applications .
in particular , they allow us to study the stability of background ads metrics with non - trivial rotation parameters .
separability of hamilton - jacobi and klein - gordon equations in the kerr ( anti)-de sitter backgrounds was discussed in @xcite , especially in the limit when all rotation parameters take the same value , see @xcite for a discussion in five dimensions . an earlier work on separability of the hamilton - jacobi equation and quantum radiation from a five dimensional kerr black hole with two rotation parameters ,
but in an asymptotically flat background , can be found in @xcite
. however , our analysis is different .
it corresponds not to a separability of the wave equations for a particle but rather to a separability of radial and angular wave equations under linear tensor perturbations .
one of the interesting features of the kerr metric in ( anti)-de sitter spaces is that it can be written in the so - called kerr - schild form , where the metric @xmath6 is given exactly by its linear approximation around the ( anti)-de sitter metric @xmath7 as follows @xcite : @xmath8 where @xmath9 is a null geodesic with respect to both the full metric @xmath6 and the ( a)ds metric @xmath7 .
moreover , the ricci tensor of @xmath6 is related to that of @xmath10 by @xmath11 where @xmath12 , with @xmath13 and @xmath14 being the parameters proportional to the mass and gravitational potential of a kerr black hole respectively .
thus , the stability of a kerr metric under metric perturbation is specific to the stability of the background metric , which is given by @xmath15 .
let us begin with a five dimensional ( anti)-de sitter metric in the standard form : @xmath16 which satisfies @xmath17 , with @xmath18 in ads space .
the apparent singularities at @xmath19 are merely coordinate singularities . by defining @xmath20
, one sees that the coordinate @xmath21 has range @xmath22 while @xmath23 have a period @xmath24 , so @xmath25 parameterizes ( topologically ) a @xmath26-sphere , while @xmath27 parameterizes an @xmath28-fiber .
the metric ( [ 5dads ] ) is easily generalized to six and higher dimensions . in six dimensions
, one has @xmath29 where @xmath30 @xmath31 and @xmath32 .
the apparent singularities at @xmath19 are again merely coordinate singularities . by defining @xmath33
one easily sees that @xmath34 parameterize ( topologically ) a @xmath35-sphere .
in fact , the generalized ( anti)-de sitter metric can be written in a more compact form : @xmath36 satisfying @xmath37 where @xmath38 , @xmath39 ( if @xmath5 is odd ) , or @xmath40 , @xmath41 ( if @xmath5 is even ) . both in odd and even dimensions , there are @xmath4 azimuthal coordinates @xmath42 , each with period @xmath24 .
but when @xmath5 is even there is an extra coordinate @xmath43 , which lies in the interval @xmath44 . in @xmath45 spaces the rotation group is @xmath46 and
the number of independent rotation parameters for a localized object is equal to the number of casimir operators , which is the integer part of @xmath47 .
thus in four dimensions the metric of a kerr black hole can have only one casimir invariant of the rotation group @xmath48 , which is uniquely defined by an axis of rotation , while in five dimensions it can have two independent rotation parameters associated with two possible planes of rotation .
one may introduce to ( [ compact1 ] ) @xmath4 rotation parameters , for example , using the following coordinate transformation : @xmath49 where @xmath50 .
the constants @xmath1 which are introduced in ( [ first - co - trans ] ) merely as parameters in a coordinate transformation may be interpreted as genuine rotation parameters after one adds to ( [ compact1 ] ) the square of an appropriate null vector , as in ( [ kerr - schild1 ] ) .
using the following coordinate transformations @xcite : @xmath51 and combining the expressions ( [ kerr - schild1 ] ) , ( [ compact1 ] ) , ( [ first - co - trans ] ) , one would obtain the kerr ( a)ds metrics in boyer - lindquist coordinates .
we are not going into details of this construction but refer to ref .
@xcite for an elegant discussion . in five dimensions ,
the metric of kerr - ads solution is @xmath52 where @xmath53 , @xmath54 , @xmath55 in the limit @xmath56 , one recovers the standard schwarzschild - ads metric . as we see
shortly , black holes with non - zero rotation parameters , or , in general , kerr - ads black holes , enjoy many interesting properties distinct from schwarzschild - ads black holes .
using the standard technique of background subtraction , gibbons @xcite have recently calculated the regularized ( euclidean ) actions for the kerr - ads black holes in arbitrary @xmath57 dimensions .
the results are @xmath58 for odd @xmath59 , and @xmath60 for even @xmath61 , where @xmath62},\ ] ] is the volume of the unit @xmath63-sphere . in the above
we have defined @xmath64 , with @xmath65 being the curvature radius of the ( bulk ) ads space .
the dimensionless parameters are : @xmath66 and @xmath67 . , where , as usual , @xmath68 is the radius of the horizon , which occurs at a root of @xmath69 , and @xmath70 .
the hawking temperature , which is the inverse of euclidean period , @xmath71 , is given by @xmath72,\ ] ] where @xmath39 for odd @xmath5 and @xmath73 for even @xmath5 .
the calculation of total energy in an asymptotically ( a)ds background is somewhat trickier ( see e.g. @xcite ) , mainly because the analogous komar integral for the relevant time - like killing vector diverges , which then requires a regularization , see also ref.@xcite which presents a general analysis for the conserved charges and the first law of thermodynamics for the four dimensional kerr - newman - ads and the five dimensional kerr - ads black holes . in this context , the conserved charges ( energies ) @xmath74 and @xmath75 associated with different killing vectors , respectively , @xmath76 and @xmath77 are different .
however , the calculation of free energy itself is unambiguous .
in fact , one can always identify the free energy of a kerr - ads black hole as @xmath78 , and hence @xmath79 this result is modified from that of a schwarzschild - ads black hole by certain terms in the product which are now functions of @xmath80 and the rotation parameters @xmath81 . in four spacetime dimensions
black holes are stable ( see e.g. @xcite ) , but the issue of stability may be raised in five and higher dimensions .
the five - dimensional kerr - ads solutions are particularly interesting as these could be embedded into iib supergravity in ten dimensions . from ( [ free - energy - gen ] )
, it is readily seen that a phase transition between the background ads space and the black hole is set by the scale @xmath82 , so that @xmath83 corresponds to ads black hole ( @xmath84 ) and @xmath85 to a thermal ads space ( @xmath86 ) .
this behavior may be seen also in terms of the charge or potential if present . in general ,
when the values of the rotation parameters @xmath81 are decreased , the free energy lowers towards zero at low temperature . for @xmath87 , in the small @xmath80 region , @xmath88 nearly approaches but never touches the @xmath89 axis ( see figs . @xmath90 and @xmath26 ) .
that is , the free energy curve crosses the @xmath89 axis only once , namely when @xmath82 , which usually corresponds to the hawking - page phase transition point .
but , in dimensions @xmath91 , this alone does not mean that a first order phase transition of hawking - page type is essentially present . in five dimensions , with @xmath92 ,
there is a minimum @xmath80 below which the temperature appears to be negative and also diverges as @xmath93 ( see fig .
[ figure1 ] ) , which is clearly unphysical .
in fact , there is a minimum in temperature below which the kerr black holes simply do not exit . nevertheless , the plots in fig .
[ figure3 ] show the free energy can be a well defined function of temperature .
we also note that the specific heat is a monotonically increasing function of temperature when @xmath94 .
the hawking temperature of a kerr - ads black hole with a non - vanishing rotation parameter approaches zero as @xmath80 goes to zero .
the free energy is still a smooth function of both the horizon size and the temperature ( see figs .
[ figure2 ] and [ figure4 ] ) .
these all imply a thermodynamic stability of a small kerr black hole in @xmath95 space .
the thermodynamic behavior above is essentially the opposite in six dimensions , where the temperature always diverges at @xmath96 . as the plots in figs .
[ figure5][figure8 ] show the thermodynamics of single parameter solutions are quite different from those with equal rotation parameters .
we also note that the @xmath84 region in figs .
[ figure9][figure12 ] corresponds to @xmath83 , while the region @xmath97 corresponds to @xmath98 .
when @xmath91 , in the case of equal rotation parameters , only small black holes are globally preferred and locally stable , while in the case of single rotation parameter , a thermal ads phase is more preferred .
the behavior in five dimensions is special in that there is no high temperature thermal ads phase regardless of the choice of rotation parameters .
one of the simplest ways of calculating the energy in an asymptotically ads background is to integrate the first law of ( bulk ) thermodynamics : @xmath99 where the entropy @xmath100 and angular momenta ( of a rotating black hole ) @xmath101 are defined via @xmath102 where @xmath103 in ref .
@xcite , the mass ( energy ) of kerr ads black hole was evaluated , by demanding _ as a priori _ that entropy of the black hole is one - quarter the area , @xmath104 , in order to satisfy ( [ firstlaw ] ) .
the results are @xmath105 @xmath106 this result differs from the expression of energy suggested by hawking _
et al . _ in @xcite , both in odd and even dimensions , @xmath107 the reason for this is that the energy ( [ energy - hawking ] ) is measured in a frame rotating at infinity with the angular velocities : @xmath108 instead of ( [ ang - velo ] ) . since the angular velocities differ by @xmath109 , the two results above , ( [ energy - odd - d ] ) or ( [ energy - even - d ] ) , ( [ energy - hawking ] ) , agree only in the limit @xmath110 ( i.e. @xmath111 ) .
a remark is in order .
the energy of background ads spacetime ( i.e. @xmath112 ) is expected to be the same as the casimir energy of a dual field theory in one dimension lower , up to a conformal factor .
but from the above result one finds @xmath113 when @xmath112 . to understand this apparent discrepancy
, it should be noted that the adm mass @xmath13 is only a local definition of black hole energy , while the total energy of a localized object in a curved background normally takes into account the asymptotic value of the background itself ( which is non - zero in the ads space ) . and , in general , one can write @xmath114 , where @xmath115 is an integration constant .
for example , for a schwarzschild - ads black hole with hyperbolic symmetry ( @xmath116 ) , @xmath115 may be given by @xmath117 , where @xmath118 is the black hole mass at the extremal limit , see e.g. @xcite . a black hole as a thermodynamic system is unstable if it has negative specific heat .
as it known , small schwarzschild - ads black holes ( i.e. with @xmath119 ) have negative specific heat but large size black holes have positive specific heat .
while there also exists a discontinuity of the specific heat as a function of temperature at @xmath120 , and so small and large black holes are found to be somewhat disjoint objects .
however , this is essentially not the case when some of @xmath1 are non - trivial , and especially , when @xmath121 , e.g. , the small kerr black holes in @xmath95 space also have positive specific heat . to this end
, we shall study the thermodynamic stability of a kerr - ads black hole by evaluating its specific heat , which is given by @xmath122 figures [ figure13 ] and [ figure14 ] show the plots of energy and temperature differentials as functions of the horizon size @xmath80 . in the @xmath123 case , with equal rotation parameters ,
there is clearly a minimum @xmath80 , below which the temperature diverges .
there is also a minimum value of rotation parameter below which @xmath124 can be negative , which is @xmath125 in five dimensions . above this value ,
both the temperature differential and the specific heat are positive , see figs .
[ figure15 ] and [ figure16 ] .
when plotted as a function of temperature , the minimum of energy corresponds to the minimum in temperature , and hence the specific heat is a monotonically increasing function of hawking temperature .
this is also the case with a single rotation parameter ( see fig .
[ figure17 ] ) , but now the critical value is @xmath126 . it should also be noted that , with @xmath123 and @xmath127 , the thermodynamic behavior of a kerr - ads black hole at high temperature can be very different from that at low temperature ( see fig .
[ figure17 ] ) .
a similar behavior is found in the case of a single non - vanishing rotation parameter , though up to a slightly larger value of @xmath128 ( @xmath129 ) .
put another way , in the @xmath95 background , small rotating black holes are unstable only for rotation parameters of order @xmath130 or less ; the precise limit is dimension dependent and black holes with larger angular velocities are thermodynamically stable .
similarly , in dimensions @xmath91 , the kerr - ads black holes become unstable below some critical values of rotation parameters , for which a new branch would appear . when @xmath131 , with equal rotation parameters , the critical value is @xmath132 ( see figs . [ figure18][figure20 ] ) and it is slightly higher in the case of a single rotation parameter . looking at the behavior of free energy and specific heat as functions of horizon position @xmath80 , we remark that a five dimensional kerr - ads black hole with a single rotation parameter is thermodynamically more stable over two ( equal ) rotation parameter solutions . but
this is essentially not the case in dimensions six or more . in all odd dimensions ,
the specific heat has a single branch at high rotation but two branches at low rotations : the critical value of @xmath128 which distinguishes these two cases increases with the number dimensions , and also with number of non - trivial rotation parameters .
a similar behavior is observed in all even dimensions @xmath91 , but in this case an interesting difference is that the specific heat can never be zero with @xmath133 .
it seems relevant to ask what happens at the critical angular velocity limit , @xmath134 .
apparently , the action as well as the entropy is divergent in this limit .
nevertheless , as discussed in @xcite ( see also @xcite ) , there exists a scaling of the mass parameter @xmath135 which makes the physical charges of the configuration finite . with equal rotation parameters , when @xmath136 , a kerr - ads black hole is more preferred than a thermal ads phase even at low temperature .
in fact , in all dimensions @xmath91 , small kerr - ads black holes with a single non - vanishing rotation parameter are unstable . in our plots
we have used the energy expressions suggested by gibbons _
et al . _
@xcite , which differ from those suggested by hawking _
_ @xcite by some overall constant factors .
this itself does not introduce any significant difference in the behavior of specific heat and hence the thermodynamic stability of kerr - ads solutions . at any rate ,
the energy measured in a non - rotating frame appeared more suggestive to be used because it can be derived using various other methods @xcite ; the energy ( or total mass ) expressions given in @xcite , however , disagree with those in @xcite in odd spacetime dimensions .
it was shown recently in @xcite that at fixed entropy , the temperature of a rotating black hole is bounded above by that of a non - rotating black hole , in four and five dimensions , but not in six or more dimensions .
we verify this claim by plotting temperature as a function of entropy , in various dimensions ; some of the plots are depicted in figs .
[ figure21]-[figure23 ] . in dimensions six or more
, the minimum of entropy is not always the minimum of temperature , it actually depends upon the choice of rotation parameters .
this is precisely the case where the inequality @xmath137 may be realized with a very small entropy .
but in this limit the temperature actually diverges , so the effect like this might be absent in a physical picture . at fixed entropy , but @xmath138 , the hawking temperature of a rotating black hole is always suppressed relative to that of a non - rotating black hole and the inequality @xmath139 holds in all dimensions .
this result , presumably , holds with various charges and classical matter fields ( such as gauge fields , dilaton , etc ) and is in accord with the earlier observation made by visser while studying a static spherically symmetric case in four dimensions with no cosmological term @xcite .
a five dimensional kerr - ads black hole with a single non - vanishing rotation parameter possesses an interesting ( and perhaps desirable ) feature ; in this case the entropy vanishes when the temperature becomes zero .
a similar feature is present in seven dimensions , but with two equal rotation parameters : @xmath140 , @xmath141 . in a recent work @xcite ,
on the evolution of a five dimensional rotating black hole via the scalar field radiation , maeda observed that , in a flat background ( @xmath142 ) , the asymptotic state of a five dimensional rotating black hole with a single non - vanishing parameter is described by @xmath143 .
it would be interesting to know a similar result in an anti - de sitter background , @xmath18 .
following @xcite , one would expect the partition function of a kerr - ads black hole to be related to the partition function of a cft in a rotating einstein universe on the ( conformal ) boundary of the ads space . a curious observation in ref .
@xcite is that the cardy - verlinde entropy formula works more naturally using the bulk thermodynamic variables defined by hawking _
et al . _
this seems to indicate that the energy expression ( [ energy - hawking ] ) is still relevant in a dual cft .
the killing vector is then given by @xmath144 where @xmath145 are the angular coordinates .
this property normally allows the thermal radiation to rotate with black hole s angular velocity all the way to conformal infinity . one could ask whether or not the bulk thermodynamic variables suggested by gibbons
_ et al . _
@xcite , which were measured with respect to a frame that is non - rotating at infinity , can be mapped onto the boundary cft variables by using the usual scaling argument .
this does not seem to be the case as long as the cft is assumed to be on a surface of large @xmath80 in boyer - linquist coordinates .
however , such a mapping might exist when the cft is assumed to be on a large spherical surface , that is one for which the coordinate @xmath146 at large @xmath147 .
that is to say , it is possible that the set of bulk variables for kerr - ads black holes given by gibbons _ et al._@xcite , in some ( modified ) form , match onto the boundary cft variables that satisfy the first law of thermodynamics .
this was indeed shown to be the case in @xcite .
let us briefly discuss the role of non - trivial rotation parameters on the existence of an equilibrium between kerr - ads black hole and rotating thermal radiation around it .
for this the requirement of a positive specific heat is a necessary condition . in five dimensions ,
the specific heat is always positive and also a monotonically increasing function of temperature when one ( or both ) of the rotation parameters takes a value at least one - quarter the ads length scale @xmath65 .
this means , unlike in minkowski ( infinite ) space , the rotating kerr - ads black holes can be in equilibrium with rotating thermal radiation around it , when @xmath148 ; that is , the rotation parameter is not significantly smaller than the ads curvature radius , so as to attain a stable equilibrium .
in this section we study the gravitational stability of kerr - ads background metrics ( with @xmath15 ) in dimensions five and higher . for this purpose , it is sufficient to consider the following @xmath5-dimensional ( time - independent ) metric _ ansatz _ : @xmath149 where the metric @xmath150 is effectively separated into two parts : a diagonal `` bulk '' line element and @xmath151 , which is the metric on an @xmath152-dimensional base manifold whose curvature has not been specified , ( so @xmath153 or @xmath154 ) , and hence can be replaced by any einstein - khler metric with the same scalar curvature .
however , in the present work we study only the @xmath155 case , and hence the base @xmath156 may be viewed as an @xmath28 fiber over @xmath157 ( for odd @xmath152 ) or as @xmath158 ( for even @xmath152 ) . for example , in the @xmath123 ( or @xmath159 ) case , the event horizon is @xmath160 . under a small linear metric perturbation @xmath161 with @xmath162 , the variation in the ricci tensor
is given by @xmath163 where the spin-2 lichnerowicz operator @xmath164 is defined by ( see , for example , @xcite ) @xmath165 the stability of background metrics of the form ( [ metric - gen ] ) with @xmath166 , under certain metric perturbations , is specific to tensor perturbations .
we therefore would like to restrict our analysis here to the tensor mode fluctuations that satisfy @xmath167 unless @xmath168 , where the indices @xmath169 run from @xmath170 and the indices @xmath171 will run from @xmath172 for the @xmath152-dimensional base .
the variation of the ricci tensor on the base @xmath173 must then satisfy @xmath174 where @xmath175 is the @xmath5-dimensional cosmological constant , with @xmath176 h_{i j}\nonumber \\ & + & \sum_{\nu=1}^{d - n}\left[\partial^\sigma g_{\sigma \nu } -\frac{1}{2}g^{\sigma\rho}\partial_\nu g_{\sigma\rho } + ( 4-n)\frac{\partial_\nu\gamma}{\gamma}\right]\partial^\nu h_{ij}\nonumber \\ & - & \frac{4}{\gamma^2}\left[g^{\mu\nu}\partial_\mu \gamma(x ) \partial_\nu \gamma(x)\right ] h_{i j},\end{aligned}\ ] ] where @xmath177 is the spin-2 lichnerowicz operator acting on the base @xmath173 .
the lichnerowicz operator @xmath178 is compatible with the transverse , trace - free ( de donder ) gauge for @xmath179 : @xmath180 , see e.g. @xcite .
let us first consider a background spacetime where @xmath181 , such that we can write the metric as @xmath182 we can write the lichnerowicz operator as @xmath183 h_{i j } \nonumber \\ & + & \left[\frac{\beta_r}{\beta}-\frac{\alpha_r}{\alpha } + ( 4-n)\frac{\gamma_r}{\gamma}\right ] \frac{\partial_r h_{i j}}{\beta^2}\nonumber\\ & - & \frac{4}{\gamma^2 } \frac{\gamma_r^2}{\beta^2 } \,h_{i j},\end{aligned}\ ] ] where the subscripts @xmath184 , @xmath185 denote derivatives w.r.t .
@xmath184 , @xmath185 respectively . in this case
we find it convenient to choose @xmath186 such that @xmath187 where @xmath188 are coordinates on @xmath173 and @xmath189 is the eigenvalue of the lichnerowicz operator on @xmath156 .
we want to write the perturbed equations ( [ perturbed1 ] ) in the form : @xmath190 to facilitate this we introduce two transformations : @xmath191 with @xmath192 we then find ( see appendix for details ) @xmath193 where , @xmath194.\end{aligned}\ ] ] the above potential correctly reproduces the result in @xcite ( cf .
( 41 ) with @xmath195 and @xmath196 ) , see also @xcite . apparently , the case @xmath197 is special .
while we believe the stability analysis of kerr - ads background metrics can be generalized to non - equal rotation parameters ( or angular momenta ) , we shall focus on the case with equal rotation parameters . in the case of an odd number of spacetime dimensions @xmath198 , the zero - mass ( @xmath15 ) kerr ( anti)-de sitter background metric may be given by @xmath199 where the rotation parameters are set equal ( i.e. , @xmath200 ) .
the base space @xmath173 , which is topologically @xmath201 , may be parameterized by the metric @xmath202 where @xmath203 is the canonically normalized fubini - study metric on an @xmath204 dimensional complex projective space @xmath205 , and @xmath206 is a local potential for the khler form @xmath207 on @xmath205 .
for example , in five dimensions , the metric on base @xmath208 is @xmath209 . in the above background
, the linear tensor perturbations satisfy @xmath210 where @xmath211 h_{i j } + \frac{1-ca^2}{r^2+a^2 } ( \tilde{\delta}_l h)_{i j}.\end{aligned}\ ] ] in terms of the regge - wheeler type coordinate @xmath212 , which may be defined by @xmath213 and using eqs .
( [ trans1]),([tranf2 ] ) the differential equation is cast in the standard form : @xmath214 where the potential is @xmath215 this potential is well behaved around @xmath216 unlike for the ads - schwarzschild metric ( i.e. @xmath217 ) .
there exists a criterion for stability ( e.g. the schrdinger equation not possessing a bound state with @xmath218 ) , in terms of the minimum lichnerowicz eigenvalue , @xmath219 , on the base manifold @xmath173 . in the case of a vanishing cosmological constant ( @xmath142 ) , this criterion is the same as that for a schwarzschild - ads background @xcite : @xmath220 a requirement that @xmath221 constrains the spacetime dimensions to @xmath222 ( or @xmath223 ) .
the stability of a potential depends on the eigenvalue @xmath189 , ensuring that the potential is positive everywhere and bounded from below . defining @xmath224 , with @xmath225 ,
we require @xmath226 for @xmath18 and we find @xmath227 the lower bound on @xmath228 required for gravitational stability of the background metric is found to be stronger than that for thermodynamic stability . in the de sitter case ( i.e. @xmath229 ) , there is a mass gap , so @xmath189 starts from a finite value , and @xmath230 is unbounded from below . instead of solving the schrdinger equation directly in terms of @xmath212 , one can solve the radial part of equation ( [ lich - eq ] ) by expressing it as a hypergeometric equation , whose solution is given by a linear combinations of @xmath231 where @xmath232 , and @xmath233 we note that reality of @xmath234 immediately gives the stability condition ( [ instab ] ) .
reality of the solution also requires @xmath235 which implies that there are no exponentially growing ( unstable ) modes . requiring the solution to be bounded as @xmath236 fixes one arbitrary constant which leaves @xmath237 decaying as @xmath238 .
given that @xmath239 we find that @xmath240 decays as @xmath241 . by considering the large @xmath185 limit of potential ( [ potr1 ] )
we also see that @xmath242 so that eq ( [ schro r ] ) remains bounded as required to make the total energy finite .
the radial equation is easily solved to yield @xmath255 however , regularity of the radial solution at @xmath256 requires @xmath257 and hence as @xmath258 the radial solution behaves as @xmath259 equation ( [ angular - eq ] ) , together with boundary conditions of regularity at @xmath260 and @xmath261 , constitute an eigenvalue problem for the separation constant @xmath254 . for @xmath262 ,
the solution is @xmath263,\ ] ] where @xmath264 the criterion for gravitational stability , in terms of the minimum lichnerowicz eigenvalue @xmath219 on the base manifold @xmath156 , namely @xmath265 , now translates into the requirement that @xmath266 .
however we note that @xmath267 in this case . in ads
space , since @xmath18 , for reality of the solution we also require , @xmath268 for real @xmath269 , @xmath270 and hence @xmath271 is imaginary , but this is not allowed by the radial wave equation .
therefore there are no normalisable solutions with @xmath272 . for @xmath273 ,
one requires @xmath274 .
a useful inequality for stability of the background ads metric ( [ background - d - even ] ) is therefore , @xmath275 instead of considering the large @xmath185 limit in ( [ main - eq - r - theta ] ) , let us now consider the special case where the angular velocity approaches the critical limit , @xmath276 ( or @xmath277 ) .
the eigenfunctions are then the associated legendre polynomials @xmath278 , @xmath279 , where , @xmath280 an interesting case is @xmath281 , which allows one to study supergravity solutions in @xmath282
. it would be interesting to know what the limit @xmath283 corresponds to in a dual field theory .
we leave this issue to future work .
in this paper we have studied the thermodynamics and stability of higher - dimensional ( @xmath284 ) rotating black holes in a background ( anti)-de sitter spacetime .
the thermodynamic quantities for kerr - ads black hole solutions suggested by gibbons _
et al . _
@xcite have been used to study the behavior of the free energy and specific heat ( which are defined unambiguously in all spacetime dimensions @xmath285 ) as functions of temperature and horizon positions .
the two apparently different expressions of energy in the kerr - ads background suggested by hawking _ et al . _
@xcite and gibbons _
et al . _
@xcite do not introduce any significant difference in the behavior of bulk thermodynamic quantities ( such as entropy , free energy , specific heat , etc ) and therefore the stability of kerr - ads solutions .
nevertheless , the gibbons _
_ bulk variables are more suggestive to be used as they map onto the boundary variables with the natural definition of conformal boundary metric , that is the one for which the coordinate @xmath146 for large @xmath147 , and satisfy the first law of thermodynamics . as for thermodynamic stability , rotating black holes are found to be stable down to a critical value of the rotation parameter , below which the specific heat becomes negative .
for example , a five dimensional kerr - ads black hole is thermodynamically stable when the rotation parameters take values @xmath286 ; larger angular velocities usually stabilize the black hole .
similarly , a zero - mass kerr - ads background is gravitationally stable down to a critical eigenvalue , below which the schrdinger equation may involve growing tensor mode perturbations .
again larger angular velocities stabilize the background kerr - ads spacetimes , although the bound on the rotation parameters required for the gravitational stability of rotating black holes is not directly related to that of thermodynamic stability . an obvious extension to study , for completeness ,
would be the inclusion of non - zero charges and all of the possible rotation parameters in dimensions five and higher .
a particularly interesting problem would be a study of the gravitational stability of _ massive _ kerr - ads black hole spacetimes in @xmath95 and @xmath287 spaces .
some of the problems will be discussed in a follow - up paper @xcite . * acknowledgements * it is a pleasure to thank roy kerr and david wiltshire for discussions .
i.p.n . also wishes to thank gary gibbons for useful discussions and chris pope for helpful remarks .
this work was supported in part by the mardsen fund of the royal society of new zealand .
consider a second order differential equation of the form @xmath288 where @xmath289 are functions of @xmath185 only .
we find it is convenient to choose @xmath290 , such that @xmath291 .
we then have @xmath292 + for non - zero fluctuations , @xmath293 , this implies that @xmath294 where @xmath295 .
we would like to write this in the form @xmath296 to facilitate this we introduce two transformations : @xmath297 the differential equation then takes the form @xmath298
\varphi^\prime \nonumber \\ & & ~~+\left[\frac{a}{r_x^2}\frac{\chi^{\prime\prime } } { \chi}+\left(\frac{b}{r_x}-\frac{a\,r_{xx } } { r_x^3}\right)\frac{\chi^\prime}{\chi}+ \tilde{c}\right]\varphi=0,\nonumber \\\end{aligned}\ ] ] where @xmath299 .
let us define @xmath300 this implies @xmath301 the differential equation then takes the standard form : @xmath302 where @xmath303 where @xmath304 .
j. m. maldacena , adv .
* 2 * , 231 ( 1998 ) ; s. s.gubser , i. r. klebanov and a. m. polyakov , phys . lett .
b * 428 * , 105 ( 1998 ) ; e. witten , adv .
* 2 * , 253 ( 1998 ) .
e. witten , adv .
theor . math .
* 2 * , 505 ( 1998 ) .
s. w. hawking and d. n. page , commun .
phys . * 87 * , 577 ( 1983 ) .
s. w. hawking , c. j. hunter and m. m. taylor - robinson , phys .
d * 59 * , 064005 ( 1999 ) .
s. w. hawking and h. s. reall , phys .
d * 61 * , 024014 ( 2000 ) .
r. b. mann , phys .
d * 61 * , 084013 ( 2000 ) ; m. m. caldarelli , g. cognola and d. klemm , class .
* 17 * , 399 ( 2000 ) .
a. m. awad and c. v. johnson , phys .
d * 61 * , 084025 ( 2000 ) .
k. landsteiner and e. lopez , jhep * 9912 * , 020 ( 1999 ) .
d. birmingham , class .
* 16 * , 1197 ( 1999 ) .
r. g. cai , phys .
d * 63 * , 124018 ( 2001 ) .
i. p. neupane , phys .
d * 69 * , 084011 ( 2004 ) .
r. p. kerr , phys .
* 11 * , 237 ( 1963 ) .
b. carter , commun .
* 10 * , 280 ( 1968 ) .
r. c. myers and m. j. perry , annals phys . *
172 * , 304 ( 1986 ) .
g. w. gibbons , h. lu , d. n. page and c. n. pope , j. geom .
phys . * 53 * , 49 ( 2005 ) .
g. w. gibbons , h. lu , d. n. page and c. n. pope , phys .
* 93 * , 171102 ( 2004 ) .
m. cvetic , h. lu and c. n. pope , phys .
b * 598 * , 273 ( 2004 ) ; phys . rev .
d * 70 * , 081502 ( 2004 ) ; m. cvetic , g. w. gibbons , h. lu and c. n. pope , arxiv : hep - th/0504080 . o. madden and s. f. ross , class .
* 22 * , 515 ( 2005 ) . m. vasudevan , k. a. stevens and d. n. page , class . quant .
* 22 * , 339 ( 2005 ) .
h. k. kunduri and j. lucietti , phys .
d * 71 * , 104021 ( 2005 ) .
v. p. frolov and d. stojkovic , phys .
d * 67 * , 084004 ( 2003 ) ; v. p. frolov and d. stojkovic , phys .
d * 68 * , 064011 ( 2003 ) .
g. w. gibbons , m. j. perry and c. n. pope , arxiv : hep - th/0408217 .
i. papadimitriou and k. skenderis , arxiv : hep - th/0505190 .
r. gregory and r. laflamme , phys .
lett . * 70 * , 2837 ( 1993 ) .
m. cvetic , p. gao and j. simon , arxiv : hep - th/0504136 .
m. henneaux and c. teitelboim , commun .
phys . * 98 * , 391 ( 1985 ) .
a. ashtekar and s. das , class .
* 17 * , l17 ( 2000 ) .
s. das and r. b. mann , jhep * 0008 * , 033 ( 2000 ) .
n. deruelle and j. katz , class .
* 22 * , 421 ( 2005 ) [ arxiv : gr - qc/0410135 ] .
s. deser , i. kanik and b. tekin , arxiv : gr - qc/0506057 . g. w. gibbons , m. j. perry , and c. n. pope , arxiv : hep - th/0506233 . m. visser , phys . rev .
d * 46 * , 2445 ( 1992 ) .
h. nomura , s. yoshida , m. tanabe and k. i. maeda , arxiv : hep - th/0502179 .
r. g. cai , l. m. cao and d. w. pang , arxiv : hep - th/0505133 . i. p. neupane , class .
quant . grav .
* 19 * , 1167 ( 2002 ) .
g. gibbons and s. a. hartnoll , phys .
d * 66 * , 064024 ( 2002 ) .
h. kodama and a. ishibashi , prog .
phys . * 110 * , 701 ( 2003 ) .
carter and i. p. neupane ( unpublished ) . | we study the thermodynamic and gravitational stability of kerr anti - de sitter black holes in five and higher dimensions .
we show , in the case of equal rotation parameters , @xmath0 , that the kerr - ads background metrics become stable , both thermodynamically and gravitationally , when the rotation parameters @xmath1 take values comparable to the ads curvature radius . in turn
, a kerr - ads black hole can be in thermal equilibrium with the thermal radiation around it only when the rotation parameters become not significantly smaller than the ads curvature radius .
we also find with equal rotation parameters that a kerr - ads black hole is thermodynamically favored against the existence of a thermal ads space , while the opposite behavior is observed in the case of a single non - zero rotation parameter .
the five dimensional case is however different and also special in that there is no high temperature thermal ads phase regardless of the choice of rotation parameters .
we also verify that at fixed entropy , the temperature of a rotating black hole is always bounded above by that of a non - rotating black hole , in four and five dimensions , but not in six and more dimensions ( especially , when the entropy approaches zero or the minimum of entropy does not correspond to the minimum of temperature ) . in this last context ,
the six dimensional case is marginal . |
recently , a number of _ chandra _ and _ xmm_-newton observations of quasars have shown local ( @xmath3 ) x - ray absorption lines ( @xcite ) .
these background quasars are among the brightest extragalactic x - ray sources in the sky and some of them were used as calibration targets .
the typical high ionic column densities of these x - ray absorbers ( @xmath4 @xmath5 ) imply the existence of large amounts of hot gas with temperatures around @xmath6 k. recent ultraviolet observations of the local high velocity absorbers also reveal such hot gas but at lower temperatures ( see .
e.g. , @xcite ) .
given the spectral resolution of _ chandra _ and _ xmm_-newton , it is still unclear where this hot gas is located : in the interstellar medium , in the galactic halo , or in the local group as the intragroup medium . in sharp contrast , so far only four targets were reported showing intervening absorption systems ( @xmath7 ) , all with low ion column densities .
@xcite confirmed their first detection @xcite of an absorption system at @xmath8 , along the sight line towards pks 2155 - 304 .
they found the observed column density is @xmath9 @xmath5 , and set a 3@xmath10 upper limit on the column density of @xmath11 @xmath5at the same redshift .
@xcite reported the detection of two absorption systems along the sight line towards mkn 421 , but both systems showed absorption lines with column densities less than @xmath12 @xmath5 .
@xcite reported a number of x - ray absorption systems towards h 1821 + 643 at 2 - 3 @xmath10 level , and @xcite reported the detection of an absorption line along the sight line towards 3c 120 .
to understand the difference between the local and intervening absorption systems , we conduct a survey of a number of extragalactic targets observed with _
chandra _ and _ xmm_-newton to search for local x - ray absorption lines . in this paper , we report the result of this survey .
we find that a model in which local x - ray absorbers are associated with the milky way can explain our survey results better than a model in which they are associated with the intragroup medium in the local group .
recently , @xcite conducted a systematic survey of 15 nearby agns to investigate the hot x - ray absorbing gas in the vicinity of the galaxy .
while their work mainly focuses on associating individual line of sight with known local structures , our work is different in that we study the generic properties of this hot gas .
all the background sources are selected from the _ chandra _ and _ xmm_-newton data archives .
we focus on instruments that can provide both high spectral resolution and moderate collecting area in the soft x - ray band , and this results in four different instrument combinations : ( 1 ) rgs , the reflection grating spectrometer ; ( 2 ) letg ( the low energy transmission grating ) + hrc ( the high resolution camera ) ; ( 3 ) letg + acis ( the advanced ccd imaging spectrometer ) ; and ( 4 ) hetg ( the high energy transmission grating ) + acis .
the first one ( rgs ) is on board _ xmm_-newton and the last three are instruments on board _ chandra_. the instrumental resolving power is typically characterized by the line response function ( lrf ) , the underlying probability distribution of a monochromatic source .
hetg - acis has the highest resolving power : the lrf of the medium energy grating ( meg ) has a full width at half maximum ( fwhm ) of @xmath13 around 20 .
the other three combinations have a roughly constant resolution of @xmath14 across the relevant wavelength range . in some cases ,
a target has been observed with several different instrument configurations , and we select those that give the highest number of continuum counts .
so far , nearly all the sources with local x - ray absorption lines show the detection of with high confidence , so in this paper we concentrate on the he@xmath1 resonance line with a rest wavelength of 21.6 . to ensure the significance of the absorption line
, we select targets that have a strong continuum between 21 and 23 .
specifically , we bin the spectrum to roughly half of the lrf fwhm and select targets which have at least 10 counts per bin around 21.6 . in this way
, we can ensure a signal - to - noise ratio ( snr ) of at least 4 within the lrf .
we test with other binning size and find essentially similar results . both _
chandra _ and _ xmm_-newton data are analyzed using standard software , i.e. , ciao 3.1 and sas 6.0 .
we refer readers to @xcite for detailed data analysis procedures .
the local he@xmath1 absorption lines from 8 of our 20 targets targets have been reported previously by other papers ; however , to ensure consistency in our data analysis , we re - analyze all the data sets and fit the absorption lines to obtain the equivalent width ( ew ) independently ( see column 5 of table 1 ) , and we refer the readers to the references we list in the column 7 of table 1 for spectra of detected lines . due to the complex nature of the sources in our sample and possible residual calibration effects , instead of fitting the continuum with a physically meaningful model such as an absorbed power law
, we fit the continuum between 21 and 23 with a polynomial of an order between 3 to 5 , to effectively remove residual feature on scales @xmath15 .
the residual is then fitted with a gaussian at @xmath16 , using isis ( interactive spectral interpretation system , see @xcite ) .
while in all cases we confirm the detections ( and non - detections ) that were reported by the original authors , in some cases our measured ews are smaller than those originally reported .
for example , we obtain an ew of @xmath17 m for ngc 4593 , while @xcite report a much larger value , @xmath18 m.a careful reexamination of both data analysis procedures indicates that the main discrepancy comes from the determination of the continuum level ( steenbrugge , private communication . ) nevertheless , our method is more stringent and the measured ews can be taken as conservative lower limits .
some of the sources in our sample are known to have intrinsic warm absorbers .
one may worry that the absorption lines detected at 21.6 could be associated with the intrinsic warm absorbers , particularly in the case of ngc 3783 where may account for part of the absorption @xcite .
however , this does not have a significant impact on our conclusions .
our case - by - case study of all the warm - absorbers in our sample indicates that and are the only possible sources of confusion . is unlikely because except in ngc 4051 , none of the known warm - absorbers show the outflow velocities that can compensate their redshifts . for
, our study indicates that always co - exists with but with at least @xmath19 times higher column density , so a non - detection of or a detection of with column density lower than @xmath20 @xmath5would simply rule out the presence of . by looking at each case
, we find that can have a contribution only for ngc 3783 and mcg6 - 30 - 15 .
while can contribute at most part of the 21.6 line in ngc 3783 , the contribution to mcg6 - 30 - 15 is unknown yet . to be conservative
, we therefore consider a subsample that excludes ngc 3783 and mcg6 - 30 - 15 .
the sight line to 3c 273 extends into the galactic halo through the edges of radio loops i and iv , which have been attributed to supernova remnants ( see , e.g. , @xcite ) .
@xcite estimated that the contribution to the column density from the supernova remnants accounts for at most 50% of the observed equivalent length .
if the absorption line is unsaturated , the equivalent width of the he@xmath1 absorption line can be converted to a column density by @xcite @xmath21 however , saturation could be an important issue , as revealed by high - order transition lines discovered in the mkn 421 spectrum @xcite , the highest quality spectrum in our sample . in that case
, the column density of , determined by high - order transition lines is @xmath22 @xmath5 , more than five times higher than that estimated from eq.([eq : n ] ) . with this consideration , column densities converted from eq .
( [ eq : n ] ) can serve as conservative lower limits , and we adopt @xmath23 as a fiducial value in the following discussion for simplicity . in figure [ f1 ]
we show an all - sky projection of the entire sample in our survey .
the green dots show the positions of those targets with detections , and the red dots show the positions of those without detections .
we wish to estimate the fraction of the sky covered by gas with an equivalent width of at least ew , @xmath24 .
the difficulty we face is that we have a limited sample size , and not all sources in our sample have enough counts to permit detection of an absorption line of equivalent width ew . for each source
, we define the threshold equivalent width , @xmath25 , which is the minimum equivalent width that can be detected at 3@xmath10 : @xmath26 where @xmath27 is wavelength , @xmath28 is the photon flux , @xmath29 is the area of the detector , @xmath30 is the observation time , and @xmath31 is the resolving power where @xmath32 is the width of a resolution element .
@xmath25 therefore depends on both the detector and the source ; as more data are gathered for a given source , its value of @xmath25 will decrease .
table 1 includes values of @xmath25 for each source in our sample . to estimate @xmath24 , we consider all the sources with @xmath33 , since only these sources have enough counts to ensure detection of an absorption system with an equivalent width @xmath34 .
let @xmath35 be the number of these sources , and let @xmath36 be the number of sources in this sample with detectable ovii absorption ( by definition , such sources must have @xmath37 ) .
the detection fraction is then @xmath38 and in the limit of large @xmath36 it will approach @xmath24 .
figure [ f2 ] shows the detection fraction as a function of the minimum detectable equivalent width , and the error bars are the standard 1@xmath10 errors for a binomial distribution .
we show the error bars for the entire sample only for demonstration purposes .
the detection fraction starts at @xmath3910% for the entire sample and @xmath39 5% for the subsample , corresponding to lines with @xmath40 m , and then gradually increases to 100% with increasing continuum level or decreasing ew threshold .
the most striking feature is that the detection fraction approaches 100% as the sensitivity approaches @xmath41 m , which occurs at a continuum level of @xmath42 counts per bin . while it is tempting to infer that the detection fraction would be 100% if our sample had enough sensitivity , we must be cautious : given the detections of absorption lines in the five highest - quality data sets , we can conclude only that we have 90% confidence that the sky detection fraction is at least @xmath43 , or 63%
while both _ chandra _ and _ xmm_-newton have unprecedented resolving power , the highest resolution we can achieve ( hetg - acis , in this case ) is about @xmath44 @xmath45 . based on the hubble flow and adopting a hubble constant of @xmath46 , this corresponds to a distance of @xmath47 mpc at @xmath48 , which makes it impossible to distinguish between a milky way - origin ( with a halo radius of @xmath49 mpc ) and a local group - origin ( with a radius of @xmath50 mpc ) for the x - ray absorbers .
nevertheless , with such a large sample size , we can begin to understand the properties of these absorbers , i.e. , their size , density , temperature , spatial distribution , etc . in the previous section
, we estimate the sky detection fraction @xmath51 of this hot gas .
the detection fraction @xmath51 can be taken as an estimate of sky covering fraction , defined as @xmath52 .
our estimation then indicates that the sky cover fraction is at least higher than 63% , and is likely close to unity . in the following ,
based on this estimation we ( 1 ) calculate the expected number of absorbers along lines - of - sight toward distant agns , ( 2 ) make joint analysis with x - ray emission measurements , and ( 3 ) estimate the total baryonic mass .
our calculation indicates that the observed x - ray absorbers are associated with our galaxy .
given the high quality of spectra in our sample , any absorbers distributed between us and background agns with similar column densities , i.e. , @xmath53 @xmath5 , would be detected with high confidence .
now let us estimate the expected number of _ intervening _ absorbers with a similar column density in our sample .
a key assumption in our model is that the x - ray absorbers detected at @xmath48 are not unique , i.e. , an observer located in another galaxy similar to the milky way should be able to see similar x - ray absorption .
the next step is to convert the covering fraction that we presented in the previous section to the detectability of these absorbers around systems similar to the milky way or the local group , along the sight lines towards background agns .
we assume that the x - ray absorbers are uniformly distributed within the halo ( of a galaxy or a group of galaxies ) that has a radius @xmath54 .
let @xmath55 be the sky covering factor as observed from the center of the distribution of absorbers , which can be obtained from observation .
if we view the absorbers in a different group of galaxies , the observed covering factor would be @xmath56 ( see , e.g. , @xcite ) . given the spatial density of the absorbers within the halo , we can then calculate the expected number of absorbers along the total line of sight towards all the background sources in our sample .
if the spatial density of halos with radius @xmath54 is @xmath57 , the expected number is then @xmath58 where @xmath59 is the cross section and @xmath60 is the cumulative distance along the sight lines to all the targets in the sample . in reality , both the covering fraction and the available pathlength , and so @xmath61 , depend on the detection threshold @xmath25 .
taking the covering fraction from eq .
( 3 ) , we can calculate @xmath61 from eq .
( 4 ) , as a function of @xmath25 ( figure [ f3 ] ) . let us consider two scenarios , in which the absorbers can be associated either with milky way - type halos or with local group - type halos .
for the milky way - type halo , @xmath62 kpc , and the spatial density of halos is just the spatial density of @xmath63 galaxies , i.e. , @xmath64 @xcite . on the other hand ,
if the observed local absorbers are associated with the local group , @xmath54 would be @xmath50 mpc .
an integral over the press - schechter mass function @xcite from @xmath65 to @xmath66 gives a halo spatial density of @xmath67 . in figure
[ f3 ] , local group case and the milky way case are represented by solid and dashed lines , respectively .
dark lines are for the entire sample , and red lines are for the subsample .
given the fact that _ none of the targets in the sample show any intervening absorption with column densities even close to @xmath68 @xmath5 _ , clearly the observations are consistent with the association of the local x - ray absorbers with milky way - type halos , but inconsistent with a local group association .
in the local group case , for the entire sample the maximum expected number of absorbers is @xmath69 if we take a threshold of @xmath70 m .
the poisson probability of detecting zero if the expectation is 9.4 is just @xmath71 , at above gaussian @xmath72 level .
if we take the subsample , the expected number decrease to @xmath73 , and the probability increases to @xmath74 , roughly corresponding to a gaussian @xmath72 level .
we can also estimate the total baryon mass in these x - ray absorbers .
let @xmath75 be the hydrogen density in an absorber and let @xmath76 be the volume filling factor of the absorbers in the halo .
let @xmath77 be the typical hydrogen column density along lines of sight in which x - ray absorption is seen .
averaged over the sky , the average column density is @xmath78 , so that the baryonic mass of the absorbers is @xmath79 , where @xmath80 g is the mass per h. let @xmath81 ( ) be the ionization fraction of and @xmath82 be the abundance of oxygen relative to the solar abundance ( taken as @xmath83 @xcite ) .
we then have @xmath84\ ; , \ ] ] where the factor in brackets equals @xmath77 .
if the absorbers were distributed throughout the local group so that @xmath85 mpc , their total mass would be @xmath86 ( ) ] @xmath87 .
since the dynamical mass of the local group is @xmath88 ( see , e.g. , @xcite ) , the total baryon mass in the local group is @xmath89 for a baryon fraction of @xmath39 15% .
this means the covering fraction at most can be 30% for solar metallicity . since figure [ f2 ] clearly indicates that the covering fraction can be much higher ( and probably be as high as @xmath90 ) given enough instrumental sensitivity ,
we conclude that a local group origin for these x - ray absorbers is unlikely because their estimated total mass would exceed the baryon mass of the local group .
our conclusion is consistent with that of @xcite on the other hand , if we associate these absorbers with our galaxy , then the radius decreases by a factor of 10 - 100 and the total mass is substantially reduced .
having presented evidence associating the @xmath91 x - ray absorbers with the hot gas within the milky way , we can further constrain the properties of the x - ray absorbers by assuming that the same absorbers are also responsible for at least some of the hot halo foreground observed in x - ray emission .
the soft x - ray background is believed to be produced by three components : the local hot bubble ( lhb ) , the extragalactic background ( mainly from point sources ) , and a halo component . by analyzing _
rosat _ all sky survey ( rass ) data , @xcite concluded that the halo emission actually consists of two components , a hard component with @xmath92 and a soft component with @xmath93 . because the hard component is too hot to produce a substantial amount of in collisional ionization equilibrium
, we assume that most of the observed resides in the soft component .
let @xmath94 be the observed emission measure .
since @xmath94 is based on solar abundances and is averaged over the sky , it is related to the actual emission measure , em , along a line of sight through an absorber by @xmath95 , where @xmath96 and @xmath97 and @xmath98 are the electron and proton densities .
@xmath99 is the average pathlength through an absorber in pc .
@xcite found an emission measure of @xmath100 for the soft component . from the absorption measurement , we have @xmath101 @xmath5 . since @xmath102 at @xmath93 , we have @xmath103 ( independent of metallicity ) , and @xmath104 kpc .
this also provides independent evidence that these x - ray absorbers should be associated with our milky way instead of the intragroup medium in the local group .
@xcite argued that based on a combination of emission line and absorption line measurement , the scale length of the o vii absorber should be at least @xmath105 kpc .
however , their argument depended crucially on the temperature of the emitter / absorber
. a shift in temperature by 0.1 to 0.2 in log space will change the scale length by a factor of 10 .
they determined the temperature should be between @xmath106 k , based on the and line ratio . a recent work , based on much higher quality chandra data on mkn
421 @xcite , indicated that just from chandra measurement of the and line ratio , the temperature of the mkn 421 local absorber must be lower than @xmath107 k. such a temperature brings the scale length down to the galactic scale .
we estimate this x - ray absorbing gas has a density of a few @xmath108 with temperature around @xmath6 k. our estimations are consistent with values predicted from other models such as the dynamics of the magellanic stream @xcite .
@xcite suggested , based on observations of absorption in the disk and halo of the milky way and in the intergalactic medium , that these absorption systems belong to radiative cooling flow of hot gas .
their predicted column density is consistent with what we find in this paper .
a key question then is : how to keep this gas hot without cooling ?
for gas with such a high density , the cooling time will be less than the hubble time .
in collisional ionization equilibrium , the typical cooling time scale is @xmath109 with @xmath110 and @xmath111 . here
, @xmath112 is the radiative cooling rate in units of @xmath113 . under the assumption that the cooling rate at @xmath114 scales as the fe abundance , the cooling rate of @xcite is @xmath115 , where the solar abundances are taken from @xcite . for the density and temperature
we have estimated for the absorbers , the cooling time is much less than the hubble time , unless the metallicity is extremely low ( @xmath116 ) . to keep this relatively dense gas ( @xmath117 @xmath118 ) at temperatures around @xmath6 k without cooling down over a hubble time scale ,
some sort of heating mechanism must play an important role .
supernova heating is a potential candidate .
the cooling rate , according to the calculations above , would be @xmath119 for @xmath120 ; note that the dependence on @xmath55 , @xmath81(o vii ) , and @xmath82 has canceled out . assuming that the supernova rate is about 0.02 yr@xmath121 in our galaxy ( see , e.g. , @xcite ) , and that the energy output of each supernova explosion is about @xmath122 , the heating rate would be about @xmath123 . for @xmath124 kpc ,
which is consistent with our estimate for @xmath125 , only a small fraction of the supernova energy is needed to maintain the temperature of the absorbers at @xmath126 k. in this paper , we find that the sky covering fraction of this hot , -absorbing gas is at least 63% , at 90% confidence .
this is based on the detections of absorption lines at @xmath127 level in the spectra of the top five high quality spectra . while we can not obtain the exact location of this hot gas within our galaxy , joint analysis with the x - ray background data
indicates the scale height of this hot gas should be @xmath128 kpc .
finally we conclude that , based on three independent estimations , the x - ray absorbing gas detected locally are part of the hot gas in our galaxy .
future observations will provide robust test of this result .
we thank julia lee for help with mcg 6 - 30 - 15 data and rik williams for help with mkn 421 data .
we also thank the anonymous referee for valuable suggestions .
tf was supported by the nasa through _ chandra _ postdoctoral fellowship award number pf3 - 40030 issued by the _
chandra _ x - ray observatory center , which is operated by the smithsonian astrophysical observatory for and on behalf of the nasa under contract nas 8 - 39073 .
the research of cfm was supported in part by nsf grant ast00 - 98365 and by the support of the miller foundation for basic research .
mgw was supported in part by a nasa long term space astrophysics grant nag 5 - 9271 . | recent _ chandra _ and _ xmm _ observations of distant quasars have shown strong local ( @xmath0 ) x - ray absorption lines from highly ionized gas , primarily he - like oxygen .
the nature of these x - ray absorbers , i.e. , whether they are part of the hot gas associated with the milky way or part of the intragroup medium in the local group , remains a puzzle due to the uncertainties in the distance .
we present in this paper a survey of 20 agns with _
chandra _ and _ xmm _ archival data .
about 40% of the targets show local he@xmath1 absorption with column densities around @xmath2 ; in particular , absorption is present in all the high quality spectra .
we estimate that the sky covering fraction of this -absorbing gas is at least 63% , at 90% confidence , and likely to be unity given enough high - quality spectra . based on ( 1 ) the expected number of absorbers along sight lines toward distant agns , ( 2 ) joint analysis with x - ray emission measurements , and ( 3 ) mass estimation , we argue that the observed x - ray absorbers are part of the hot gas associated with our galaxy .
future observations will significantly improve our understanding of the covering fraction and provide robust test of this result .
nn |
basically , there are two competing explanations of the experimentally observed power - law behaviour of elastic hadronic form factors : hard scattering @xcite and the feynman mechanism @xcite . at sufficiently large momentum transfer ,
the feynman mechanism contribution is dominated by configurations in which one of the quarks carries almost all the momentum of the hadron . in qcd , this results in an extra @xmath5 suppression compared to the hard scattering term generated by the valence configurations with small transverse sizes and finite light - cone fractions of the total hadron momentum carried by each valence quark .
the hard term , which eventually dominates , can be written in a factorized form @xcite,@xcite,@xcite , as a product of the perturbatively calculable hard scattering amplitude and two distribution amplitudes accumulating the necessary nonperturbative information . however , this mechanism involves exchange of virtual gluons , each exchange bringing in a suppression factor @xmath6 .
hence , to describe existing data by the hard contribution alone , one should intentionally increase the magnitude of the hard scattering term by using distribution amplitudes with a peculiar `` humped '' profile @xcite . in this case
, passive quarks carry a small fraction of the hadron momentum and , as pointed out in ref.@xcite , the `` hard '' scattering subprocess , even at rather large momentum transfers @xmath7 , is dominated , in fact , by very small gluon virtualities .
this means that the hard scattering scenario heavily relies on the assumption that the asymptotic pqcd expressions are accurate even for momenta smaller than @xmath8 , @xmath9 in the region strongly affected by finite - size effects , nonperturbative qcd vacuum fluctuations @xmath10 including these effects shifts the hard contributions significantly below the data level even if one uses the humpy distribution amplitudes and other _ ad hoc _ modifications intended to increase the hard term ( see , @xmath11 @xcite ) .
furthermore , as argued in ref.@xcite , the derivation of the humpy distribution amplitudes in @xcite implies a rather singular picture ( infinite correlation length ) of the qcd vacuum fluctuations . under realistic assumptions , it is impossible to get distribution amplitudes strongly differing from the smooth `` asymptotic '' forms .
the latter are known to produce hard contributions which are too small to describe the data on elastic form factors .
thus , there is an increasing evidence in favour of the alternative scenario , @xmath12 that for experimentally accessible momentum transfers the form factors are still dominated by the purely soft contribution corresponding to the feynman mechanism . in the language of the light - cone formalism @xcite , the soft term is given by the overlap of the soft parts of the hadronic wave functions ,
@xmath9 is an essentially nonperturbative object . among the existing approaches to the nonperturbative effects in qcd , that which is closest to pqcd is the qcd sum rule method @xcite .
qcd sum rules were originally used to calculate the soft contribution for the pion form factor in the region of moderately large @xcite,@xcite and then small momentum transfers @xcite .
it should be emphasized that , in the whole region @xmath13[0ex]{$\sim$}}$}}3 \ , gev^2 $ ] , the results are very close to the experimental data : the feynman mechanism alone is sufficient to explain the observed behaviour of the pion form factor . for higher @xmath2 , the direct qcd sum rule method fails due to increasing contributions from higher condensates .
however , a model summation of the higher terms into nonlocal condensates @xcite indicates that the soft term dominates up to @xmath14 @xcite .
this conclusion is also supported by a recent calculation within the framework of the light - cone sum rules @xcite .
an important observation made in ref.@xcite is that the results of the elaborate qcd sum rule analysis are rather accurately reproduced by a simple local quark - hadron duality prescription .
the latter states that one can get an estimate for a hadronic form factor by considering transitions between the free - quark states produced by a local current having the proper quantum numbers , with subsequent averaging of the invariant mass of the quark states over the appropriate duality interval @xmath15 .
the duality interval has a specific value for each hadron , @xmath11 @xmath16 for the pion and @xmath17 for the nucleon .
the local duality ansatz , equivalent to fixing the form of a soft wave function , was used to estimate the soft contribution in the case of the proton magnetic form factor @xcite .
the results agree with available data @xcite , @xcite over a wide region , @xmath18[0ex]{$\sim$}}$}}q^2 \ , { \raisebox{-.2ex}{$\stackrel{\textstyle < } { \raisebox{-.6ex}[0ex][0ex]{$\sim$}}$}}\ , 20\ , gev^2 $ ] .
furthermore , the calculation of ref.@xcite correctly reproduces ( without any adjustable parameter ) , the observed magnitude of the helicity - nonconservation effects @xmath19 with @xmath20 @xcite .
it is difficult to understand the origin of such a large scale within the hard scattering scenario , since possible sources of helicity nonconservation in pqcd include only small scales like quark masses , intrinsic transverse momenta @xmath10 , and one would rather expect that @xmath21 .
thus , the study of spin - related properties provides a promising way for an unambiguous discrimination between soft and hard scenarios . of a particular interest
there is the @xmath0 transition .
a special attention to this process was raised by the results @xcite of the analysis of inclusive slac data which indicated that the relevant form factor drops faster than predicted by the quark counting rules .
the relevant hard scattering contribution was originally considered in ref.@xcite , where it was observed that the hard scattering amplitude in this case has an extra suppression due to cancellation between symmetric and antisymmetric parts of the nucleon distribution amplitude , and it was conjectured that the faster fall - off found in @xcite can be explained by the dominance of some non - asymptotic contribution .
later , it was claimed @xcite that , by appropriately choosing the distribution amplitudes , one can get a leading - twist hard term comparable in magnitude with the data . furthermore , the results of a recent reanalysis @xcite of the inclusive slac data are rather consistent with the @xmath22 behaviour , and this revived the hope that the @xmath0 form factor can be still described by pqcd . however , the important result of the pqcd calculation @xcite is that the lowest - twist hard contribution has the property @xmath23 .
experimentally , the ratio @xmath1 is rather small @xcite , which indicates that the leading - twist pqcd term is irrelevant in the region @xmath24[0ex]{$\sim$}}$}}\ , 3 \ , gev^2 $ ] . in the present paper ,
we use the local quark - hadron duality to estimate the soft contribution for the @xmath25 and @xmath26 form factors of the @xmath0 transition to study whether the soft contribution is large enough to describe the data and whether the relative smallness of the electric form factor persists in the region of moderately large momentum transfers @xmath27[0ex]{$\sim$}}$}}q^2 { \raisebox{-.2ex}{$\stackrel{\textstyle < } { \raisebox{-.6ex}[0ex][0ex]{$\sim$}}$}}\ , 15 \ , gev^2 $ ] .
the starting object for a qcd sum rule analysis of the @xmath0 transition is the 3-point correlator : @xmath28 of the electromagnetic current @xmath29 and two ioffe currents @xcite @xmath30 we use the following parameterization for the projections of @xmath31 and @xmath32 onto the nucleon and @xmath33-isobar states , respectively : @xmath34 here , @xmath35 is the dirac spinor of the nucleon while @xmath36 is the spin-3/2 rarita - schwinger wave function for the @xmath33-isobar , @xmath9 @xmath37 , @xmath38 , @xmath39 , @xmath40 ; with @xmath41 being the nucleon mass and @xmath42 that of @xmath33 .
we use the notation @xmath43 . on the hadronic level
, the @xmath0 transition makes the following contribution to the correlator ( [ 1 ] ) : @xmath44 where @xmath45 is the @xmath0 vertex function @xmath46 ( @xmath47 ) and @xmath48 the projector onto the isobar state @xmath49 the form factors @xmath50 are related to a more convenient set @xmath51 by @xmath52 @xmath53 @xmath54 ( see , @xmath11 @xcite ) .
one should not confuse the magnetic form factor @xmath26 given by eq.([8 ] ) with the effective form factor mentioned in refs.@xcite,@xcite .
in particular , the form factor @xmath55 defined by eq.(6.2 ) of ref.@xcite can be written in terms of @xmath26 and @xmath25 ( defined by eqs.(8),(9 ) above ) as @xmath56 where @xmath57 is the energy of the virtual photon in the proton rest frame .
note that , for large @xmath2 , our @xmath26 and @xmath55 of eq.(6.2 ) of ref.@xcite have the same power behaviour .
multiplying all the factors in eq.([5 ] ) explicitly , one ends up with a rather long sum of different structures @xmath58 accompanied by the relevant invariant amplitudes @xmath59 , each of which is a combination of the three independent transition form factors listed above . to incorporate the local quark - hadron duality
, we write down the dispersion relation for each of the invariant amplitudes : @xmath60 where @xmath61 , @xmath62 .
the perturbative contributions to the amplitudes @xmath63 can also be written in the form of eq.([11 ] ) .
evidently , the physical perturbative spectral densities @xmath64 differ from their perturbative analogues , the difference being most pronounced in the resonance region , @xmath9 for small @xmath65 and @xmath66 values . in particular , @xmath64 contains the double @xmath67-function term corresponding to the @xmath0 transition : @xmath68 while the perturbative spectral densities @xmath69 are smooth functions of @xmath65 and @xmath66 .
the local quark - hadron duality assumes , however , that the two spectral densities are in fact dual to each other : @xmath70 @xmath9 they give the same result when integrated over the appropriate duality intervals @xmath71 . the latter characterize the effective thresholds for higher states with the nucleon or , respectively , @xmath33-isobar quantum numbers . as noted in ref.@xcite ,
the local duality prescription can be treated as a model for the soft wave functions : @xmath72 @xmath9 @xmath15 and @xmath73 also set the scale for the width of the transverse momentum distribution . using this model ,
we can obtain a good estimate for the overlap of the soft wave functions only in the intermediate @xmath2-region where the soft contribution is sensitive mainly to the @xmath74-widths of the quark distributions rather than to their detailed forms . from experience with the proton form factor calculations
, we expect that local duality will work in the region between @xmath75 and @xmath76 . the low-@xmath2 region @xmath24[0ex]{$\sim$}}$}}1 \
, gev^2 $ ] , in which there appear large nonperturbative contributions due to the long - distance propagation in the @xmath2-channel , can be analyzed within a full - framed qcd sum rule approach supplemented by the formalism of induced condensates @xcite or bilocal operators @xcite . applying the local quark - hadron duality to the two - point correlators of @xmath31- or , respectively , @xmath77-currents
, we obtain simple relations between the duality intervals @xmath15 , @xmath73 and the residues @xmath78 , @xmath79 of the ioffe currents : @xmath80 after the duality intervals are fixed ( @xmath11 from the qcd sum rule analysis of the relevant two - point function @xcite ) , the local duality estimates for the form factors do not have any free parameters .
now , choosing a particular lorentz structure @xmath58 , one can get the local duality estimate for the relevant combination of the form factors .
one should remember , however , that not all the invariant amplitudes are equally reliable . to compare the contributions related to different structures
, one should specify a reference frame . in our case ,
the most relevant is the infinite momentum frame where @xmath81 while @xmath82 is fixed .
so , @xmath83 , the structures with the maximal number of the `` large '' factors @xmath84 are more reliable than those in which @xmath84 is substituted by the `` small '' parameter @xmath85 or by @xmath86 . however , dealing with the @xmath77-current in the @xmath33-channel , one should realize that @xmath32 has also a nonzero projection onto the spin-1/2 isospin-@xmath87 states @xmath88 : @xmath89 where @xmath90 is a constant , @xmath91 is the mass of the spin-1/2 state @xmath88 and @xmath92 the relevant dirac spinor satisfying @xmath93 . from eq.([15 ] ) , it follows that any amplitude containing the @xmath84-factor , is `` contaminated '' by the transitions into the spin-1/2 states .
these states lie higher than the @xmath33-isobar and , in principle , one can treat them as a part of the continuum .
then , however , there will be strong reasons to expect that the duality interval @xmath73 for the `` contaminated '' invariant amplitudes deviates from that for the amplitudes containing only the spin-3/2 contributions in the @xmath77-channel .
another possibility is to project out the amplitudes which have contributions due to the transitions into spin-@xmath94 isospin-@xmath87 states . to achieve this , it is convenient to use the basis in which @xmath95 is always placed at the leftmost position
. then , according to eqs.([5]-[7 ] ) , the invariant amplitudes corresponding to the structures with @xmath96 and @xmath97 are free from the contributions due to the spin-1/2 isospin-3/2 states . in this basis , keeping only the terms with @xmath96 and @xmath97 in eq.([5 ] ) , we get @xmath98 \frac{3(m+m)}{8 m } ( g_m^*(q^2)+g_e^*(q^2))\right
. \nonumber \\ \left .
+ \frac{q_\mu}{2 } \left ( m [ \gamma_\nu , \hat{p } ] + m [ \gamma_\nu , ( \hat{p } - \hat{q } ) ] \right ) g_1(q^2 ) \right . \nonumber \\ \left .
- p_\nu [ \hat{p},\hat{q } ] g_2(q^2 ) - q_\nu [ \hat{p},\hat{q } ] \left(g_3(q^2 ) -
\frac12 g_2(q^2 ) \right ) + ... \right ) \ . \label{16}\end{aligned}\ ] ] hence , from the invariant amplitudes related to the structures proportional to @xmath96 , we can get the local duality estimates for the form factors @xmath99 . similarly , extracting the structure @xmath100 $ ] , we get an expression for @xmath101 . counting the powers of @xmath102 , we expect that results for @xmath103 are less reliable than those for @xmath104 and @xmath105 , while results for @xmath106 are less reliable than those for @xmath103
. the number of independent amplitudes can be diminished by taking some explicit projection of the original amplitude @xmath107 .
in particular , if one multiplies @xmath108 by @xmath109 , the invariant amplitude corresponding to the structure @xmath110 $ ] is proportional to the quadrupole form factor @xmath111 : @xmath112 g_c^*(q^2 ) + \ldots \right \ } \ . \label{17}\end{aligned}\ ] ] another possibility is to take the trace of @xmath113 .
the result is proportional to the magnetic form factor @xmath26 : @xmath114 however , one should remember that the trace of @xmath108 is not free from contributions due to spin-1/2 isospin-3/2 states .
though the invariant amplitude related to the trace of @xmath108 is contaminated by the transitions into spin-1/2 isospin-3/2 states , it makes sense to consider this amplitude because it has the simplest perturbative spectral density : @xmath115 where @xmath116 imposing the local duality prescription , we get @xmath117 where @xmath118 is a universal function @xmath119 and @xmath120 . as we will see , the results for other invariant amplitudes can be conveniently expressed through @xmath118 .
the function @xmath118 depends on the duality intervals @xmath15 and @xmath73 .
we fix the nucleon duality interval @xmath15 at the standard value @xmath121 extracted from the analysis of the two - point function and used earlier in the nucleon form factor calculations .
the results of the existing two - point function analysis for the @xmath33-isobar @xcite are compatible with the @xmath33 duality interval @xmath73 in the range @xmath122 to @xmath123 .
to fine - tune the @xmath73 value , we consider two independent sum rules for the @xmath104 form factor @xmath124 and @xmath125 extracted from the invariant amplitudes corresponding to the structures @xmath126 $ ] and @xmath127 $ ] , respectively ( recall that @xmath128 is the proton s momentum and @xmath129 is that of @xmath33 ) . taking the ratio of these two relations
, one can investigate their mutual consistency and test the overall reliability of the quark - hadron duality estimates . indeed , on the `` hadronic '' side
, we have the ratio @xmath130 of the isobar and nucleon masses , while on the `` quark '' side we have the ratio of two explicit and non - trivially related functions .
the consistency requires , first , that the ratio of these functions must be close to a constant and , second , that this constant must be close to the experimental value for the ratio of the isobar and nucleon masses : @xmath131 .
on fig.1 , we plot the @xmath2-dependence for the ratio of the right hand sides of eqs.([23 ] ) and ( [ 24 ] ) for the standard value @xmath121 of the nucleon duality interval and three different values of @xmath73 .
one can see that one should not rely on local duality estimates below @xmath132 .
however , above @xmath132 , the ratio is pretty constant for all three values of @xmath73 , and rather close to 1.3 .
the best agreement is reached for @xmath133 , and we will use this value as the basic isobar duality interval in further calculations . in particular , the @xmath134 parameter will be fixed by @xmath135 ( cf .
eq.([14 ] ) ) . from eqs.([8 ] ) and ( [ 9 ] ) , it follows that @xmath104 is proportional to the difference of the magnetic @xmath136 and electric @xmath137 transition form factors : @xmath138 according to eq.([16 ] ) , the sum @xmath139 of these form factors can be obtained from the invariant amplitude corresponding to the structure @xmath140 $ ] . applying the local duality prescription ,
we obtain @xmath141 \ . \label{26}\end{aligned}\ ] ] now , having expressions both for @xmath142 and @xmath143 , we can calculate @xmath26 and @xmath25 .
the results for the combinations @xmath144 and @xmath145 are shown on figs.2 and 3 , respectively .
it should be noted that , though @xmath118 has the @xmath146 asymptotics for large @xmath2 ( see eq.([22 ] ) , the local duality results are fairly consistent with the @xmath22-behaviour in the wide range @xmath147[0ex]{$\sim$}}$}}q^2 { \raisebox{-.2ex}{$\stackrel{\textstyle < } { \raisebox{-.6ex}[0ex][0ex]{$\sim$}}$}}20 \ , gev^2 $ ] .
an important observation is that @xmath25 is predicted to be much smaller than @xmath26 ( see fig.3 ) .
it should be noted that if the @xmath0 transition form factors are calculated in a purely pqcd approach ( in which only the @xmath148 double - gluon - exchange diagrams are taken into account ) , the sum of electric and magnetic form factors @xmath149 is suppressed for asymptotically large @xmath2 by a power of @xmath5 @xcite .
this is because the matrix element @xmath150 violates the helicity conservation requirement for the hard subprocess amplitude . in other words ,
the pqcd prediction is that should behave asymptotically like @xmath146 , while each of @xmath136 and @xmath137 behaves like @xmath22 . as a result ,
asymptotically @xmath151 .
however , we consider here only the soft contribution generated by the feynman mechanism for which the helicity conservation arguments are not applicable .
thus , for the soft term , there are no _ a priori _ grounds to expect that @xmath151 .
the smallness of @xmath152 dictated by local duality , strongly contrasts with the pqcd - based expectation that @xmath153 , and this allows for an experimental discrimination between the two competing mechanisms .
one should realize , however , that @xmath25 is obtained in our calculation as a small difference between two large combinations @xmath154 and @xmath155 , both dominated by @xmath26 .
hence , even a relatively small uncertainty in either of these combinations ( which is always there , since the local duality gives only approximate estimates ) can produce a rather large relative uncertainty in the values of @xmath137 .
in this situation , we restrict ourselves to a conservative statement that the electric form factor @xmath25 is small compared to @xmath26 in the whole experimentally accessible region without insisting on a specific curve for @xmath25 .
experimental points for @xmath136 shown in fig.2 were taken from the results for the @xmath55 form factor obtained from analysis of inclusive data @xcite , @xcite . since our results give a very small value for the ratio @xmath156 , the @xmath137 term in eq.([gt ] ) can be neglected .
one can see that , in the @xmath157[0ex]{$\sim$}}$}}3 \
, gev^2 $ ] region , the local duality predictions @xmath26 are close to the results of the recent analysis @xcite
. the magnetic form factor @xmath26 can also be obtained from eq.([21 ] ) . if one takes the basic duality interval @xmath133 , the resulting values of @xmath26 ( fig.2 ) are somewhat smaller than those obtained by combining the results for @xmath154 and @xmath155 . as emphasized earlier
, the spin-1/2 states also contribute to the trace of @xmath108 , and the duality interval in this case can be different from the basic value .
in fact , taking @xmath158 in eq.([21 ] ) , we get a curve for @xmath159 ( fig.2 ) essentially coinciding with those obtained from the sum of @xmath154 and @xmath155 .
the quadrupole ( coulomb ) form factor @xmath111 can be calculated from the expression ( [ 17 ] ) for the contracted amplitude @xmath160 : @xmath161 \ .
\label{30}\end{aligned}\ ] ] again , @xmath111 is essentially smaller than @xmath26 ( see fig.4 ) .
furthermore , eq.([30 ] ) predicts that , for large @xmath2 , the quadrupole form factor @xmath111 has an extra @xmath5 suppression compared to @xmath26 .
in fact , if the duality intervals are equal , @xmath162 , the suppression is even stronger , namely , by two powers of @xmath5 .
we applied the local quark - hadron duality prescription to estimate the soft contribution to the @xmath163 transition form factors .
we observed a reasonable agreement between the results obtained from different invariant amplitudes .
we found that the transition is dominated by the magnetic form factor @xmath26 while electric @xmath25 and quadrupole @xmath111 form factors are small compared to @xmath26 for all experimentally accessible momentum transfers .
numerically , our estimates for @xmath55 are close to those obtained from a recent analysis of inclusive data @xcite .
hence , there is no need for a sizable hard - scattering contribution to describe the data .
furthermore , if future exclusive measurements at cebaf would show that the ratio @xmath1 is small above @xmath132 , this would give an unambiguous experimental proof of the dominance of the soft contribution .
we are very grateful to p.stoler , n.isgur , v.burkert and c.e.carlson for discussions which strongly motivated this investigation .
we thank c. keppel for providing us with the results of her analysis and f.gross , r.schiavilla , m.j.musolf , w.j .
van orden , j.l .
goity and i.v.musatov for interest in this work and critical comments .
s.j . brodsky and g.r .
farrar , phys .
( 1973 ) 1153 .
feynman , `` photon - hadron interaction '' ; w.a.benjamin , reading , mass .
v.l . chernyak and a.r .
zhitnitsky , jetp lett .
* 25 * ( 1977 ) 510 ; yad . fiz .
* 31 * ( 1980 ) 1053 .
efremov and a.v .
radyushkin , phys . lett .
* b94 * ( 1980 ) 245 ; theor .
* 42 * ( 1980 ) 147 .
lepage and s.j .
brodsky , phys .
b87 * ( 1979 ) 359 ; phys . rev .
* d22 * ( 1980 ) v.l .
chernyak and a.r .
zhitnitsky , phys . rep .
* 112 * ( 1984 ) 173 .
n. isgur and c.h .
llewellyn - smith , nucl .
b317 * ( 1989 ) 526 .
j. bolz , r. jakob , p. kroll , m. bergmann and n.g .
stefanis , z.phys . *
c66 * ( 1995 ) 267 .
shifman , a.i .
vainshtein and v.i .
zakharov , nucl .
b147 * ( 1979 ) 385 , 448 .
ioffe and a.v .
smilga , phys . lett .
* b114 * ( 1982 ) 353 .
nesterenko and a.v .
radyushkin , phys .
* b115 * ( 1982 ) 410 .
nesterenko and a.v .
radyushkin , jetp .
39 ( 1984 ) 707 .
s.v . mikhailov and a.v .
radyushkin , phys . rev .
* d45 * ( 1992 ) 1754 .
bakulev and a.v .
radyushkin , phys . lett . * b271 * ( 1991 ) 223 .
braun and i. halperin , phys .
b328 * ( 1994 ) 457 .
nesterenko and a.v .
radyushkin , phys . lett .
* 128b * ( 1983 ) 439 .
. phys.rev .
* d48 * ( 1993 ) 29 .
p. bosted et al .
* 68 * ( 1992 ) 3841 ; phys.rev .
* d50 * ( 1994 ) 5491 .
p. stoler , phys .
* 66 * ( 1991 ) 1003 ; phys . rev . * d44 * ( 1991 ) 73 .
carlson , phys.rev . *
d34 * ( 1986 ) 2704 .
n.g.stefanis and m.bergmann , phys . lett .
* b304 * ( 1993 ) 24 . c. keppel , ph.d .
thesis , the american university ( 1995 ) ; also in : workshop on cebaf at higher energies , ed . by n.
isgur and p. stoler , cebaf , newport news ( 1994 ) 237 .
burkert , in : few - body problems in physics , ed . by f.
gross , aip conference proceedings 334 , ( 1995 ) 127 .
burkert and l. elouadrhiri , preprint cebaf - pr-95 - 001 , newport news ( 1995 )
ioffe , nucl .
phys . * 188b * ( 1981 ) 317 .
j. mathews , phys .
* 137 * ( 1965 ) b444 ; + h.f .
jones and m.d .
scadron , annals of physics * 81 * ( 1973 ) 1 .
p. stoler , phys .
reports * 226 * ( 1993 ) 103 .
ioffe and a.v .
smilga , jetp lett * 37 * ( 1983)298 ; + v.m .
belyaev and i.i .
kogan , int.j.mod.phys .
* a8 * ( 1993)153 .
balitskii , v.m .
braun , a.v .
kolesnichenko , nucl.phys .
* b312 * ( 1989 ) 509 .
belyaev and b.l .
ioffe , sov.phys.jetp * 56 * ( 1982 ) 493 . | we use local quark - hadron duality to estimate the purely nonperturbative soft contribution to the @xmath0 form factors .
our results are in agreement with existing experimental data .
we predict that the ratio @xmath1 is small for all accessible @xmath2 , in contrast to the pqcd expectations that @xmath3 .
-48pt -0.5 cm -0.5 cm cebaf - th-95 - 11 * quark - hadron duality and @xmath0 form factors * v.m . belyaev@xmath4 _ continuous electron beam accelerator facility + newport news , va 23606 , usa _ a.v . radyushkin + _ physics department , old dominion university , _ + _ norfolk , va 23529 , usa _ + _ and _ + _ continuous electron beam accelerator facility , _ + _ newport news , va 23606 , usa _ |
pseudo - orbits , or @xmath0-chains , are important tools for investigating properties of discrete dynamical systems .
these @xmath0-chains detect recurrent and mixing behaviors that may not be evident by studying actual orbits .
however , what is missing from the definitions of terms such as chain recurrence , chain transitivity , and chain mixing is any information about the lengths of the recurrence or mixing times .
they also do not give fine detail about the `` dynamics '' of the @xmath0-chains .
let @xmath1 be a compact metric space and @xmath2 be a continuous map .
the map @xmath3 is chain recurrent if for every @xmath4 and every point @xmath5 , there is an @xmath0-chain from @xmath5 to itself ( see section [ sec : defns ] for definitions ) . the chain recurrence time ( that is
, the length of the shortest such chain ) depends on @xmath0 . similarly , if @xmath3 is chain mixing , then the chain mixing time also depends on @xmath0 .
we discuss the properties and relationship of these two quantities , and in particular relate the chain mixing time to topological entropy .
we also study the structure of chain transitive maps , which resembles that of the symbolic dynamics determined by irreducible graphs and is rather different from that of topologically transitive maps .
for example , an irrational rotation of a circle is topologically transitive ( actually minimal ) , but certainly not topologically mixing .
however , as we will see ( cor . [
cor : connected ] ) , a map on a connected space is chain mixing if and only if it is chain transitive ( and , in fact , if and only if it is chain recurrent ) . the paper is organized as follows . in section [ sec : defns ] we give definitions . in section [ sec : structure ]
we discuss the structure of chain transitive maps . in section
[ sec : estimates ] we explore the properties of the chain recurrence and mixing times , and in section [ sec : entropy ] we relate the chain mixing time to topological entropy . in section [ sec : examples ] we study examples .
let @xmath1 be a compact metric space .
( to avoid technical difficulties later , we assume that @xmath1 is not a single point . ) throughout the paper , let @xmath2 be a continuous map
( @xmath6)-chain ( or ( @xmath6)-pseudo - orbit ) _ from @xmath5 to @xmath7 is a sequence @xmath8 such that @xmath9 for @xmath10 . to simplify notation , in most cases
we write simply `` @xmath0-chain '' without referring explicitly to the map @xmath3 .
we define the _ length _ of the @xmath0-chain @xmath11 to be @xmath12 .
there are a variety of ways to use @xmath0-chains to describe recurrence . in this paper
we investigate three such notions chain recurrence , chain transitivity , and chain mixing .
they generalize the well - known topological notions of recurrence of non - wandering , topological transitivity , and topological mixing . for the sake of comparison , let us recall the definitions of these topological notions of recurrence .
a point @xmath13 is _ non - wandering _ if for any open neighborhood @xmath14 of @xmath5 , there exists @xmath15 such that @xmath16 .
the map @xmath3 is non - wandering if every point of @xmath1 is non - wandering .
a map @xmath3 is _ topologically transitive _ if for any nonempty open sets @xmath17 , there is an @xmath18 such that @xmath19 .
a map @xmath3 is _ topologically mixing _ if for any nonempty open sets @xmath20 there is an @xmath21 such that @xmath22 for all @xmath23 .
_ chain recurrence .
_ a point @xmath5 is _ chain recurrent _ if for every @xmath4 , there is an @xmath0-chain from @xmath5 to itself .
the map @xmath3 is _ chain recurrent _ if every point of @xmath1 is chain recurrent .
if @xmath5 is a chain recurrent point , define the _
@xmath0-chain recurrence time _
@xmath24 to be the smallest @xmath12 such that there is an @xmath0-chain of length @xmath12 from @xmath5 to itself .
if @xmath3 is chain recurrent , define @xmath25 to be the maximum over all @xmath5 of @xmath24 .
( to see that this maximum exists , observe that if there is an @xmath26-chain of length @xmath12 from @xmath5 to itself , then there is a neighborhood @xmath27 of @xmath5 such that for all @xmath28 , there is an @xmath0-chain of length @xmath12 from @xmath7 to itself .
then the compactness of @xmath1 gives an upper bound on @xmath29 . ) for the sake of brevity , we usually write `` @xmath30 '' and `` @xmath29 . ''
_ chain transitivity . _
a map @xmath3 is _ chain transitive _ if for every @xmath31 and every @xmath4 , there is an @xmath0-chain from @xmath5 to @xmath7 . following
@xcite we say that a map @xmath3 is _ totally chain transitive _ if @xmath32 is chain transitive for all @xmath33
. _ chain mixing .
_ by compactness , a map @xmath3 is topologically mixing if and only if for any @xmath34 , there is an @xmath35 such that for any @xmath31 and any @xmath23 , there are points @xmath36 and @xmath37 with @xmath38 . to mimic this
, we say that @xmath3 is _
@xmath0-chain mixing _ if there is an @xmath21 such that for any @xmath31 and any @xmath23 , there is an @xmath0-chain from @xmath5 to @xmath7 of length exactly @xmath12 .
the map @xmath3 is _ chain mixing _ if it is @xmath0-chain mixing for every @xmath4 . if @xmath39 and @xmath13 , define the _ chain mixing time _ @xmath40 to be the smallest @xmath35 such that for any @xmath23 and any @xmath41 , there is an @xmath0-chain of length exactly @xmath12 from some point in @xmath42 to @xmath7 .
( in other words , @xmath40 is the smallest @xmath35 such that @xmath43 . )
we define @xmath44 to be the maximum over all @xmath5 of @xmath40 .
( this maximum exists by compactness . )
again , to simplify notation , in most cases we write `` @xmath45 '' and `` @xmath46 . ''
if @xmath3 is a physical system , then @xmath47 represents the uncertainty in the measurement of the initial state of the system , and @xmath0 the possible error in evaluating the map at each step . the chain mixing time
@xmath46 is the time after which all information about the initial state of the system is lost , since any point could be anywhere by then .
we are indebted to the referee for some extremely helpful comments and ideas .
in this section we investigate the structure of chain transitive dynamical systems . to the best of our knowledge , the main result about this structure
, theorem [ thm : structure ] , appears only in the exercises of @xcite .
thus a self - contained discussion of it and its consequences may be useful .
we show that for a fixed @xmath4 , the space @xmath1 decomposes into finitely many subsets ; @xmath3 permutes these @xmath12 subsets cyclically and @xmath48 is @xmath0-chain mixing on each of them . at one extreme , if @xmath0 is greater than the diameter of @xmath1 , @xmath3 is @xmath0-chain mixing on @xmath1 itself . as
@xmath0 goes to zero , the subsets divide repeatedly , either finitely or infinitely many times . for a fixed @xmath4 ,
the dynamical structure of @xmath0-chains for a chain transitive map is very similar to that of the symbolic dynamics determined by a finite irreducible graph .
( see @xcite , from which we borrow some terminology . )
as @xmath0 goes to zero , the size of the graph may grow , possibly without bound .
let us consider two elementary examples .
let @xmath49 ( considered as @xmath50 ) .
the doubling map @xmath51 is chain mixing .
[ ex : doublecircle ] let @xmath1 be the disjoint union of two circles and @xmath3 be the map sending a point @xmath5 to the point @xmath52 in the other circle .
the map @xmath3 is chain transitive but not chain mixing , because it is not @xmath0-chain mixing for any @xmath0 smaller than the distance between the two circles .
however , @xmath53 restricted to one circle _ is _ chain mixing .
we should not be misled by these examples ; they are not prototypes for all chain transitive maps .
as @xmath0 decreases the number of permuted subsets may go to infinity . as we see in the following example , it is possible that no power of @xmath3 has associated invariant subsets on which it is chain mixing .
let @xmath1 be the cantor set of sequences @xmath54 , where each @xmath55 . give @xmath1 the metric @xmath56 , where @xmath57 if @xmath58 and 1 otherwise .
define @xmath2 by adding @xmath59 , where addition is defined componentwise base 2 , with carrying to the right .
( for example , @xmath60 , and @xmath61 . )
the map @xmath3 , which is minimal , is called an _ adding machine _ or _ odometer map_. if we look only at the first digit , then @xmath3 has period two . on the first two digits , @xmath3 has period four ( @xmath62 ) , and so on .
thus @xmath3 is @xmath0-chain mixing for @xmath63 , but not for @xmath64 .
however , when @xmath65 , @xmath53 restricted to @xmath66 or @xmath67 is @xmath0-chain mixing , and in general , for @xmath68 , @xmath1 decomposes into @xmath69 disjoint sets on which @xmath70 is @xmath0-chain mixing , but @xmath71 is not .
in fact , this is essentially the only example of a chain transitive map @xmath3 such that no @xmath72 is a disjoint union of finitely many chain mixing maps , in the sense that any such map factors onto an adding machine map , which we now define more generally .
( adding machines arise in many dynamical contexts ; see , for example , @xcite .
see @xcite for more detailed definitions and background . )
let @xmath73 be a sequence of integers greater than or equal to 2 .
let @xmath74 be the cantor set of all sequences @xmath75 where @xmath76 for all @xmath77 , with @xmath78 .
define the _ adding machine map _ @xmath79 by @xmath80 , where addition is defined componentwise mod @xmath81 , with carrying to the right .
it is easy to see that every adding machine @xmath82 is minimal
. the following alternative definition of an adding machine as an inverse limit of cyclic groups will be useful .
[ lem : eqdefn ] let @xmath83 be an increasing sequence of positive integers such that @xmath84 divides @xmath85 for all @xmath77 .
let @xmath86 be the cyclic permutation given by @xmath87 .
then the inverse limit map @xmath88 is topologically conjugate to the adding machine map @xmath82 , where @xmath89
. the following result is essentially contained in exercises 8.22 and 9.18 of @xcite .
we include it here for completeness and to fill in the gaps in the proofs sketched there .
[ thm : structure ] let @xmath2 be a chain transitive map
. then either 1 .
there is a period @xmath90 such that @xmath3 cyclically permutes @xmath91 closed and open equivalence classes of @xmath1 , and @xmath72 restricted to each equivalence class is chain mixing ; or 2 .
@xmath3 factors onto an adding machine map .
the general idea for the proof is to define an equivalence relation @xmath92 on @xmath1 .
the induced map on the quotient space @xmath93 is a periodic orbit . sending @xmath0 to zero yields an equivalence relation @xmath94 ; thus @xmath3 factors onto the inverse limit of the periodic orbits , which will be either a periodic orbit or an adding machine .
let @xmath2 be a chain transitive map and let @xmath4 .
there exists @xmath95 such that for any @xmath13 , @xmath96 is the greatest common divisor of the lengths of all @xmath0-chains from @xmath5 to itself . for each @xmath13 define @xmath97 to be the greatest common divisor of the lengths of @xmath0-chains from @xmath5 to itself .
consider two points @xmath31 .
let @xmath98 be any @xmath0-chain from @xmath7 to itself .
we show that @xmath97 divides @xmath12 .
let @xmath99 be an @xmath0-chain from @xmath5 to @xmath7 and back to @xmath5 .
the @xmath0-chain @xmath100 from @xmath5 to itself has length @xmath101 .
since @xmath101 is necessarily a multiple of @xmath97 , @xmath12 is as well .
define a relation on @xmath1 by setting @xmath102 if there is an @xmath0-chain from @xmath5 to @xmath7 of length a multiple of @xmath103 . by concatenating chains , it is easy to see that this is an equivalence relation .
in fact , by the definition of @xmath103 , if @xmath102 , then _ any _ @xmath0-chain from @xmath5 to @xmath7 must have length a multiple of @xmath103 .
we may now define another equivalence relation on @xmath1 by saying @xmath104 if @xmath102 for all @xmath4 .
since @xmath3 is chain transitive , it is easy to see that @xmath102 if and only if there is a point @xmath105 such that there are @xmath0-chains from @xmath5 to @xmath105 and from @xmath7 to @xmath105 of the same length . thus another way to define
the equivalence relation @xmath94 is as the set of pairs @xmath106 such that the set of points reachable from @xmath106 with @xmath107-chains ( @xmath108 , in the notation of @xcite ) meets the diagonal .
let @xmath2 be a chain transitive map .
for any @xmath4 the equivalence relation @xmath92 is both open and closed .
the equivalence relation @xmath94 is closed .
it is clear that the relations @xmath92 and @xmath94 are closed . to see that @xmath92 is open ,
observe that an @xmath26-chain is a fortiori an @xmath0-chain . since @xmath3 is chain transitive
, there is an @xmath26-chain from @xmath5 to @xmath7 for any points @xmath5 and @xmath7 , and , if @xmath102 , the length of the chain must be a multiple of @xmath103 .
thus there is an @xmath0-chain of the same length from @xmath5 to any point within @xmath26 of @xmath7 ( use the same @xmath26 chain and change only the last element ) .
now , let @xmath109 be the quotient space and @xmath110 be the map induced by @xmath3 . to begin , let @xmath4 be fixed .
there are @xmath103 equivalence classes for @xmath92 , @xmath3 cycles among the classes periodically , and each class is invariant under @xmath72 .
the quantity @xmath103 is nondecreasing as @xmath111 , and in fact @xmath112 divides @xmath113 if @xmath114 .
either @xmath103 stabilizes at some @xmath91 , or it grows without bound .
we consider the two cases separately .
if @xmath103 stabilizes at @xmath115 , then there is only one @xmath94 equivalence class , and @xmath3 is chain mixing . to see this ,
observe that for any @xmath0 and any @xmath5 , since 1 is the gcd of the lengths of @xmath0-chains from @xmath5 to itself , there exist @xmath0-chains from @xmath5 to itself of lengths @xmath116 and @xmath12 with @xmath116 and @xmath12 relatively prime .
sylvester s formula ( @xcite ) tell us that by concatenating copies of these loops , we can get an @xmath0-chain from @xmath5 to itself of length @xmath35 for any @xmath117 . by compactness , there is an @xmath118 such that between any two points in @xmath1 there is an @xmath0-chain of length less than or equal to @xmath119 . by adding a loop at @xmath5 to a chain from @xmath5 to @xmath7
, we can get a chain from @xmath5 to @xmath7 of any length greater than @xmath120 , so @xmath3 is chain mixing . more generally , if @xmath121 , then the equivalence relation @xmath94 is the same as @xmath122 .
thus there are @xmath91 equivalence classes , @xmath3 cycles among the classes periodically , and each class is invariant under @xmath72 .
an argument similar to that for @xmath115 shows that @xmath72 is chain mixing on each equivalence class .
if , on the other hand , @xmath103 grows without bound , then the period of @xmath3 s cycling goes to infinity as @xmath0 shrinks . in this case , @xmath3 factors onto an adding machine map .
more precisely , let @xmath123 be the quotient space @xmath1 modulo @xmath124 , with the quotient topology , and let @xmath125 be the induced map .
let @xmath110 be the induced map on the quotient space @xmath109 .
the map @xmath125 is conjugate to the cyclic permutation @xmath126 given by @xmath127 .
thus , if we take a sequence @xmath128 decreasing to 0 such that the sequence @xmath129 is strictly increasing , then @xmath130 is conjugate to the map @xmath131 , where @xmath132 is the inverse limit of the sequence @xmath133 and @xmath134 is given by @xmath135 .
then , by lemma [ lem : eqdefn ] , @xmath134 is topologically conjugate to the adding machine map @xmath82 , where @xmath136 .
thus the conjugacy class of @xmath82 is independent of the choice of the sequence @xmath128 .
in fact , it is shown in @xcite that the conjugacy class of an adding machine @xmath137 depends only on the sums of the powers to which prime numbers occur as factors of the @xmath81 s .
we have a number of nice corollaries that follow from theorem [ thm : structure ] .
let @xmath2 be a chain transitive map .
if @xmath3 has a point of period @xmath12 , then there is a @xmath91 dividing @xmath12 such that @xmath1 decomposes into @xmath91 closed and open equivalence classes , @xmath3 permutes the equivalence classes cyclically , and @xmath72 restricted to each equivalence class is chain mixing .
if @xmath13 has period @xmath12 under @xmath3 , then the image of @xmath5 in @xmath138 has period @xmath91 under @xmath130 , for some @xmath91 dividing @xmath12 .
thus @xmath130 can not be an adding machine map .
let @xmath2 be a chain transitive map .
if one of the @xmath94-equivalence classes is open , then there is a @xmath91 such that @xmath1 decomposes into @xmath91 closed and open equivalence classes , @xmath3 permutes the equivalence classes cyclically , and @xmath72 restricted to each equivalence class is chain mixing .
if one of the @xmath94-equivalence classes is open , then its projection in @xmath138 is a single point that is both open and closed .
thus @xmath138 can not be a cantor set . by definition ,
if @xmath3 is chain mixing , then it is totally chain transitive . by theorem [ thm :
structure ] , the converse is also true .
@xmath2 is totally chain transitive if and only if it is chain mixing . in fact ,
as is shown in @xcite , these two conditions are also equivalent to chain transitivity for @xmath139 . to put this result in perspective ,
we contrast it with topologically transitive maps .
( see also @xcite . )
an irrational rotation @xmath140 is totally topologically transitive , but not topologically mixing .
conley s fundamental theorem of dynamical systems ( @xcite ) , says that the interesting dynamics of @xmath3 takes place on the chain recurrent set @xmath141 , with all other orbits simply going from one piece of @xmath141 to another . in some special cases (
if @xmath141 has a hyperbolic structure , for example ( @xcite ) ) , @xmath141 decomposes into finitely many chain transitive pieces ( basic sets ) , so there is not much difference between chain recurrence and chain transitivity . in general , the situation is more complicated , but we do have the following result .
[ cor : connected ] let @xmath1 be connected and @xmath2 be continuous
. then the following are equivalent : 1 .
@xmath3 is chain recurrent .
2 . @xmath3 is chain transitive .
@xmath3 is totally chain transitive .
4 . @xmath3 is chain mixing . clearly ( 4 ) @xmath142 ( 3 ) @xmath142 ( 2 ) @xmath142 ( 1 ) . by theorem [ thm :
structure ] , connectivity and chain transitivity imply that @xmath3 is chain mixing .
so it is enough to show that chain recurrence implies chain transitivity .
this follows from the proof of proposition 1.1 in @xcite , but for completeness we now give a more direct proof .
assume that @xmath3 is chain recurrent and pick @xmath4 .
we say that @xmath5 and @xmath7 are @xmath0-chain equivalent if there are @xmath0-chains from @xmath5 to @xmath7 and from @xmath7 to @xmath5 .
since @xmath3 is chain recurrent , this is an equivalence relation .
so to finish the proof it suffices to show that this is an open condition .
let @xmath5 and @xmath7 be @xmath0-chain equivalent .
choose @xmath143 such that if @xmath144 , then @xmath145 .
it suffices to show that @xmath5 is @xmath0-chain equivalent to an arbitrary @xmath146 .
let @xmath147 be an @xmath0-chain and let @xmath148 be an @xmath149-chain from @xmath7 to itself . then @xmath150 is an @xmath0-chain from @xmath5 to @xmath105 .
similarly , let @xmath151 be an @xmath0-chain from @xmath7 to @xmath5 .
then @xmath152 is an @xmath0-chain from @xmath105 to @xmath5 .
thus @xmath5 is @xmath0-chain equivalent to @xmath105 .
recall ( from @xcite , for example ) the definition @xmath153 for a factor map @xmath154 .
the system @xmath155 is _ weak mixing _
if @xmath139 is topologically transitive ; the factor map @xmath156 is _ weak mixing _ if the restriction of @xmath139 to @xmath157 is topologically transitive .
analogously , we will define the factor map @xmath156 to be _ chain mixing _ if the restriction of @xmath139 to @xmath157 is chain transitive .
the following proposition shows that when @xmath156 is the quotient map associated to @xmath94 , then @xmath156 is chain mixing .
the relation @xmath158 is a maximal closed , chain transitive , invariant subset of @xmath159 .
it is clearly closed and invariant . to see that it is chain transitive , fix @xmath4 and points @xmath160 with @xmath161 and @xmath162 .
we must show that there is an @xmath0-chain for @xmath139 from @xmath163 to @xmath164 .
let @xmath165 be the @xmath124-equivalence class containing @xmath166 and @xmath167 , and @xmath168 be the @xmath124-equivalence class containing @xmath169 and @xmath170 .
there is an @xmath12 such that @xmath32 maps @xmath171 onto @xmath172 .
since @xmath173 restricted to @xmath165 is chain mixing , there is an @xmath0-chain for @xmath139 from @xmath163 to @xmath174 , which we can extend by an actual orbit to @xmath164 . to see that @xmath94 is a maximal chain transitive subset , observe that if @xmath175 and @xmath176 is an @xmath0-chain , then @xmath177 for all @xmath77 .
thus there can not be an @xmath0-chain from @xmath163 to @xmath106 unless @xmath102 .
we conclude this section with a few results from the literature that are similar to theorem [ thm : structure ] .
however , instead of looking at the induced map on the space of equivalence classes @xmath138 , these relate to the induced map on @xmath178 , the set of connected components of @xmath1 .
see @xcite for details and more general results .
the first is a `` folklore '' result similar to theorem [ thm : structure ] for topologically transitive maps .
let @xmath1 be a compact metric space , @xmath2 be topologically transitive map , and @xmath179 be the induced map on the connected components of @xmath1 . then either 1 .
@xmath178 is finite and @xmath180 is a cyclic permutation , or 2 .
@xmath178 is a cantor set and @xmath180 is topologically transitive .
notice that under these hypotheses if @xmath5 and @xmath7 are in the same connected component , then necessarily @xmath104 .
thus @xmath179 factors onto @xmath110 .
if @xmath1 has only finitely many connected components , then the two decompositions and the two maps are the same .
let @xmath181 be a compact invariant set . 1 .
@xmath182 is topologically transitive ( resp .
chain transitive ) if @xmath183 is topologically transitive ( resp
. chain transitive ) .
2 . @xmath182 is _ liapunov stable _ if every neighborhood @xmath27 of @xmath182 contains a neighborhood @xmath184 of @xmath182 such that @xmath185 for all @xmath33 .
@xmath182 is an _ attractor _ if there exists a neighborhood @xmath27 of @xmath182 such that @xmath186 and @xmath187 .
let @xmath1 be a locally compact , locally connected metric space and @xmath2 be a continuous map .
let @xmath181 be a compact , liapunov stable transitive set , and @xmath179 be the induced map on the connected components of @xmath182
. then either 1 .
@xmath178 is finite and @xmath180 is a cyclic permutation , or 2 .
@xmath178 is a cantor set and @xmath180 is topologically conjugate to an adding machine map .
[ thm : hh ] let @xmath1 be a locally compact , locally connected metric space and @xmath2 be a continuous map .
let @xmath181 be a compact , chain transitive attractor . then @xmath179 , the induced map on the connected components of @xmath182 , is a cyclic permutation of a finite set .
theorems [ thm : structure ] and [ thm : hh ] imply the following result .
let @xmath1 be a locally compact , locally connected metric space and @xmath2 be a continuous map .
let @xmath181 be a compact , chain transitive attractor .
then there is a @xmath91 such that @xmath182 decomposes into @xmath91 closed and open equivalence classes , @xmath3 permutes the equivalence classes cyclically , and @xmath72 restricted to each equivalence class is chain mixing .
it is worth remarking here that on many compact metric spaces , including manifolds , homeomorphisms generically have no chain transitive attractors .
see @xcite .
we conclude this section with an example .
construct a denjoy map @xmath188 by starting with an irrational rotation @xmath189 and replacing @xmath190 , the orbit of 0 , with a sequence of intervals @xmath191 such that @xmath192 and @xmath193 is mapped into @xmath194 by an orientation - preserving homeomorphism ( see @xcite for details ) .
so defined , @xmath3 is chain mixing , but not topologically transitive .
in fact , the omega limit set of any @xmath195 is the invariant cantor set @xmath196 .
if we restrict the map to @xmath197 , then @xmath198 is still chain mixing and @xmath198 is topologically transitive .
thus the space @xmath178 of connected components of @xmath197 is @xmath197 itself , while @xmath138 is a single point .
let @xmath2 be a chain transitive map .
we wish to estimate the chain recurrence time @xmath29 and , in the case that @xmath3 is chain mixing , the chain mixing time @xmath46 for a given @xmath0 .
our technique is to model @xmath0-chains with a subshift of finite type , an idea which has been used before in some other contexts ( @xcite , for example ) .
the approach we use here is adapted from that in chapter 5 of @xcite . for calculations of @xmath29 and @xmath46 for a variety of dynamical systems ,
see section [ sec : examples ] .
let @xmath199 be the upper box dimension of @xmath1 ( see @xcite for a definition of upper box dimension ) .
there exists a constant @xmath200 such that for small enough @xmath0 : 1 . if @xmath2 is chain transitive , then @xmath201 .
2 . if @xmath2 is chain mixing , then @xmath202 . let @xmath203 be a minimal cardinality collection of balls of radius @xmath26 whose interiors cover @xmath1 .
observe that for small enough @xmath0 , @xmath204 for some constant @xmath197 , where @xmath199 is the upper box dimension of @xmath1 ( @xcite ) .
construct the @xmath205 transition matrix @xmath206 by setting @xmath207 if @xmath208 and 0 otherwise .
if @xmath209 is an allowable word in the subshift of finite type determined by @xmath206 , then for any @xmath210 and any @xmath211 , there is an @xmath0-chain @xmath212 with @xmath213 for each @xmath214 .
conversely ( see lemma 5.1 of @xcite ) , if @xmath215 is an @xmath216-chain , where @xmath216 is less than the lebesgue number of the cover @xmath217 and @xmath213 , then the word @xmath218 is an allowable word .
thus , because @xmath3 is chain transitive , for any @xmath77 and @xmath214 , there is an allowable word starting at @xmath77 and ending at @xmath214 . by removing any subwords that begin and end at the same symbol
, we can assume that the word has length less than or equal to @xmath219 , and so for any two points @xmath5 and @xmath7 there is an @xmath0-chain from @xmath5 to @xmath7 of length less than or equal to @xmath204 for small enough @xmath0 .
this proves the first part of the result . to prove the second part , observe that since @xmath3 is chain mixing , the matrix @xmath206 is eventually positive .
so if @xmath220 is the smallest integer such that each entry of @xmath221 is strictly positive , then @xmath222 . since @xmath223 ( @xcite ) , we have that @xmath224 for small enough @xmath0 .
we also have a lower bound for @xmath46 in the case that @xmath3 is lipschitz .
we know that the lipschitz constant @xmath225 is at least 1 , since otherwise the banach fixed point theorem says that @xmath3 has an attracting fixed point , contradicting the fact that @xmath3 is chain mixing .
[ prop : lipest ] let @xmath2 be chain mixing and have lipschitz constant @xmath225 .
let @xmath226 be the diameter of @xmath1 .
then for @xmath47 sufficiently small , @xmath227 if @xmath228 , and @xmath229 if @xmath230 .
the diameter of the @xmath12th iterate of a ball of radius @xmath47 under @xmath231 is @xmath232 so , if @xmath228 , then @xmath233 thus if @xmath234 is @xmath226 , then @xmath12 is at least @xmath235 . when @xmath230 , @xmath236
. give the product space @xmath237 the sup metric , i.e. , @xmath238 .
[ prop : recpowers ] let @xmath1 and @xmath239 be compact , @xmath2 and @xmath240 be chain recurrent , and @xmath241
. then @xmath242 and @xmath243 are also chain recurrent , and for all @xmath4 , @xmath13 , and @xmath244 , 1 .
@xmath245 , 2 .
@xmath246 , 3 .
@xmath247 , and 4 .
there exists an @xmath248 such that @xmath249 .
observe that @xmath250 is an @xmath0-chain for @xmath251 if and only if @xmath252 and @xmath253 are @xmath0-chains for @xmath3 and @xmath134 respectively .
statement ( 1 ) and the fact that @xmath251 is chain recurrent if and only if @xmath3 and @xmath134 both are follow immediately . to prove ( 2 ) , let @xmath254 and @xmath255 be @xmath0-chains
. then the @xmath0-chain @xmath256 formed by concatenating @xmath254 with itself @xmath257 times has length @xmath258 , as does the @xmath0-chain @xmath259 formed by concatenating @xmath255 with itself @xmath260 times .
combining the two gives an @xmath0-chain from @xmath106 to itself . to show that @xmath3 is chain recurrent
if @xmath72 is , and to prove ( 3 ) , it suffices to observe that if @xmath261 is an @xmath262-chain , then @xmath263 is an @xmath264-chain of length @xmath265 . to see that @xmath72 is chain recurrent if @xmath3 is , pick @xmath266 small enough that for any @xmath267-chain @xmath268
, @xmath269 is within @xmath0 of @xmath270 .
( such an @xmath266 exists because of the uniform continuity of @xmath3 . )
thus , if @xmath271 is an @xmath272-chain of length @xmath265 , then @xmath273 is an @xmath262-chain of length @xmath12 . to prove ( 4 ) , observe that by taking an @xmath272-chain of length @xmath12 from @xmath5 to @xmath5 and concatenating it with itself @xmath91 times , we get an @xmath272-chain of length @xmath265 , and thus an @xmath262-chain of length @xmath12 .
if we give @xmath237 the euclidean metric , then ( 2 ) becomes `` @xmath274 . ''
let @xmath1 and @xmath239 be compact , @xmath2 and @xmath240 be chain mixing , and @xmath241
. then @xmath242 and @xmath243 are also chain mixing , and for all @xmath4 , 1 .
@xmath275 2 .
@xmath276 3 .
there exists an @xmath248 such that @xmath277 .
statement ( 1 ) and the fact that @xmath251 is chain mixing if and only if both @xmath3 and @xmath134 are follow from the definition of chain mixing .
the proofs of ( 2 ) and ( 3 ) and of the fact that @xmath72 is chain mixing are essentially identical to those of the corresponding statements for chain recurrent maps in proposition [ prop : recpowers ]
. notice that if @xmath2 is chain recurrent and @xmath278 is an invariant subset , then @xmath279 may be greater than or less than @xmath25 ; while clearly @xmath280 for @xmath281 , we take the maximum over a smaller set ( @xmath282 instead of @xmath1 ) to get @xmath279 .
consider the following example .
let @xmath283 and @xmath284 where @xmath285 , @xmath286 , and @xmath287 are the circles shown in figure [ fig : circs ] .
let @xmath2 be a homeomorphism with fixed points at the south poles and motion indicated by the arrows .
assume that away from the fixed points the motion on @xmath285 is much faster than the motion on @xmath286 which is much faster than the motion on @xmath287 .
if @xmath288 stays close enough to @xmath289 near their common fixed point , then a point @xmath5 near the fixed point can return to itself more quickly via an @xmath0-chain passing through @xmath288 than via one that stays in @xmath289 , so @xmath290 , and in fact @xmath291 . on the other hand , points move very slowly on @xmath292 , so @xmath293 .
the following theorem shows that the growth of the chain mixing times can give information about topological entropy .
[ thm : entest ] let @xmath2 be chain mixing .
then the topological entropy , @xmath294 , satisfies @xmath295 where @xmath296 is the lower box dimension of @xmath1 . in @xcite
, misiurewicz shows that we can use the growth rate of @xmath0-chains to calculate topological entropy .
specifically , @xmath297 where @xmath298 is the maximum cardinality of a set of @xmath47-separated @xmath0-chains of length @xmath12 .
( two @xmath0-chains @xmath299 and @xmath300 are @xmath47-separated if @xmath301 for some @xmath77 . ) for @xmath302 , let @xmath303 ( i.e. , the maximum cardinality of an @xmath216-separated collection of points ) .
we claim that for @xmath304 , @xmath305 . to see this ,
let @xmath306 be a @xmath307-separated set of points . by the definition of @xmath46 ,
for each of the @xmath308 sequences @xmath218 ( @xmath309 ) there is a sequence of points @xmath310 such that for each @xmath214 , @xmath311 and there is an @xmath0-chain of length @xmath46 from @xmath312 to @xmath313 .
since the points @xmath314 are @xmath307-separated , the sequences @xmath310 are @xmath47-separated .
thus @xmath305 .
next , let @xmath315 be the minimum number of balls of radius @xmath216 necessary to cover @xmath1 . by the definition of lower box dimension ( @xcite ) , for small enough @xmath216 , @xmath316 for some positive constant @xmath197 .
clearly @xmath317 ( ( * ? ? ?
* lemma 8.1.10 ) ) , so @xmath318 as well .
finally , we have that @xmath319 since @xmath320 ( recall that @xmath1 is not a single point , and it can not be a finite collection of points since @xmath3 is chain mixing ) . in the following section
we see that the formula in theorem [ thm : entest ] can give the precise value of the topological entropy .
we conclude with three examples rigid rotations of the circle , subshifts of finite type , and the doubling map on the circle in which we illustrate the ideas in the previous sections .
consider the dynamical system @xmath321 which is the rigid rotation by @xmath322 given by @xmath323 ( mod 1 ) .
we know that when @xmath216 is rational every point is periodic and when @xmath216 is irrational every point has a dense orbit .
thus , in either case @xmath324 is chain mixing .
the chain mixing time is just the least integer greater than or equal to @xmath325 .
the chain recurrence time for @xmath326 depends on the diophantine properties of @xmath216 .
let @xmath4 be given . clearly , @xmath327 for all @xmath328 and @xmath329 .
so we may assume that @xmath330 and @xmath331 .
notice that for a rotation , the set of points that can be reached with an @xmath0-chain of length @xmath12 beginning at @xmath332 is precisely those points within @xmath333 of @xmath334
. therefore @xmath29 is the smallest integer @xmath15 such that @xmath335 .
equivalently , we want to find the smallest @xmath15 such that @xmath336 for some @xmath337 .
if @xmath216 is rational then @xmath332 is periodic with some period , @xmath338 . in this case , for @xmath0 sufficiently small , the shortest @xmath0 chain from @xmath332 to @xmath332 is the periodic orbit .
thus , @xmath339 .
now assume @xmath216 is irrational . to analyze this case we need to use continued fractions ( for details , see @xcite ) .
let @xmath340=a_{0}+\cfrac{1}{a_{1}+\cfrac{1}{a_{2}+\dotsb}}\ ] ] be the continued fraction expansion for @xmath216 , and @xmath341 $ ] the @xmath91th convergent ( assume that @xmath342 is in lowest terms ) .
we say that the rational number @xmath343 ( @xmath344 ) is a _
best approximation _ of @xmath216 if the inequalities @xmath345 and @xmath346 imply @xmath347 in other words , the only way to get a better rational approximation of @xmath216 than a best approximation is by increasing the denominator . from theorem [ thm : best ] we can conclude that for a given @xmath0 , the values of @xmath12 and @xmath116 that we wish to find are precisely @xmath348 and @xmath349 for some convergent @xmath342 of @xmath216 .
that is , the function @xmath29 is a nonincreasing piecewise constant function taking on the values @xmath350 .
thus , to understand the behavior of @xmath29 as @xmath351 we must investigate the rate of convergence of @xmath342 to @xmath216 .
the following theorem gives a bound on this rate of convergence .
in fact , this worst case can be realized .
for example , when @xmath358 $ ] we have ( @xcite ) @xmath359 where @xmath360 as @xmath361 .
thus , when @xmath362 we have @xmath363 for large @xmath91 , so @xmath364 . on the other hand
, there are @xmath216 for which @xmath365 grows slowly as slowly as desired as @xmath351 .
in particular , if @xmath366 is any decreasing function with @xmath367 as @xmath368 , then there is an @xmath216 such that @xmath369 for some sequence @xmath370 . to see this we need the following theorem .
define @xmath373 .
let @xmath216 be the irrational number guaranteed by theorem [ thm : fastconv ] for this @xmath374 .
then there exist infinitely many @xmath220 and @xmath338 with @xmath372 .
let @xmath375 denote this collection of denominators , listed in increasing order , and let @xmath376 .
then @xmath377 .
let @xmath378 be a subshift of finite type given by an @xmath379 transition matrix @xmath182 , with @xmath380 , and metric @xmath381 . since an @xmath0-jump can take any element of a cylinder set of diameter @xmath0 to any other element in the same cylinder set , it is easy to see that @xmath382 is chain transitive if and only if it is topologically transitive .
( this is because , for a one - sided shift , we can consider only @xmath0-chains that have a jump only at the first step , and afterwards are actual orbits ; for a two - sided shift , they can have a jump at the beginning and at the end . )
the shift space @xmath383 decomposes into finitely many topologically transitive pieces ( @xcite ) , so we assume that @xmath382 is topologically transitive . for notational convenience ,
we consider one - sided shifts ( the proofs for two - sided shifts are essentially the same ) .
given @xmath4 and @xmath384 , we want to find @xmath385 .
let @xmath91 be the smallest nonnegative integer satisfying @xmath386 .
since @xmath382 is topologically transitive , the transition matrix @xmath182 is irreducible ( @xcite ) , so there is an allowable word @xmath387 of length @xmath388 from @xmath389 to @xmath390 .
thus the sequence @xmath391 defined by @xmath392 and @xmath393 for @xmath394 is an @xmath0-chain from @xmath395 to itself of length @xmath396 .
notice that for some @xmath397 we can find an @xmath0-chain from @xmath395 to itself that is shorter than this , but for most points in @xmath383 this is the shortest possible @xmath0-chain .
thus , the worst case is that @xmath398 .
let @xmath403 be the doubling map on the circle @xmath404 . if we consider binary expansions , then @xmath405 , with the identifications @xmath406 , @xmath407 , etc .
thus the argument in example [ ex : finitetype ] shows that @xmath408 , where @xmath409 is the least integer function .
arguing as in the proof of proposition [ prop : lipest ] , we see that @xmath410 .
then theorem [ thm : entest ] says that @xmath411 so in this case the estimate is exact . | we investigate the properties of chain recurrent , chain transitive , and chain mixing maps ( generalizations of the well - known notions of non - wandering , topologically transitive , and topologically mixing maps ) .
we describe the structure of chain transitive maps .
these notions of recurrence are defined using @xmath0-chains , and the minimal lengths of these @xmath0-chains give a way to measure recurrence time ( chain recurrence and chain mixing times ) .
we give upper and lower bounds for these recurrence times and relate the chain mixing time to topological entropy . |
network reciprocity @xcite is a general mechanism responsible for the development of spatial correlations within a viscous population , opening the possibility of persistent cooperation .
several specific models have been proposed showing how these correlations are related to stable groups of cooperating individuals , whose bulk benefits of self - defense and mutual support outcompete the surface exploitation by defectors @xcite .
although actual experiments have been performed @xcite , most of our knowledge comes from these simple models . in particular , a prevailing characteristic in real systems and an important ingredient for cooperation is the heterogeneous contact in systems whose interactions are given by complex @xcite or diluted networks @xcite .
when we consider the prisoner s dilemma ( pd ) dynamics @xcite on a diluted lattice that , albeit heterogeneous , has only short range interactions , intermediate densities present an enhancement of cooperation @xcite and , in the presence of a small amount of noise , the optimal dilution is closely related to the ( random site ) percolation threshold for that lattice @xcite .
whatever the level of heterogeneity , the contact network topology may evolve in time .
although several rewiring mechanisms can be devised ( see ref .
@xcite and references therein ) , this may also be accomplished when the high viscosity restriction is relaxed and the agents become mobile . mobility patterns on different scales of human activity , and their far fetched consequences , have been studied in recent decades .
for example , airplane displacement and its connection with disease spread @xcite , on a global level , can be contrasted with the more local dynamics of pedestrians , crowds or traffic @xcite .
of particular interest is how the observed patterns can affect the outcome of the competition between agents and , in turn , be influenced by it as well . within the evolutionary game theory framework ,
after several sparse , early attempts to include mobility @xcite , it was only recently that the interest in the combined effects of mobility and cooperation in the pd game had a significant increase .
some level of information processing capability is required , for example , when the movement is strategy dependent @xcite or driven by payoff @xcite , success @xcite or the neighborhood composition @xcite .
however , the simplest scenario is when mobility is diffusive @xcite . indeed , as hypothesized in ref .
@xcite , random mobility may have evolved prior to contingent mobility , allowing bacteria to move away from each other while exploring new resources .
our previous results on a lattice @xcite show that even in the framework of random , noncontingent mobility of unconditional agents , diffusion is favorable to cooperation , under rather broad conditions , if velocities are not too high .
analogous conclusions , attesting the robustness of the results , were also found in off - lattice models @xcite . when diffusion occurs on a lattice and the one agent per site constraint holds , this area exclusion couples the diffusivity of the agents with the free area .
this dependence on density , on the other hand , is not immediately present in off - lattice systems with point - like particles @xcite .
moreover , while on a lattice the number of simultaneous interactions is limited by its coordination number , there is no such restriction on the number of point - like particles within the range of interaction in off - lattice systems ( unless it is explicitly included as in ref .
a relevant question concerns the universal effects of such geometrical hindrance on the emergence and persistence of cooperation .
for example , letting the average body size be a coevolving trait , there may be some evolutive pressure for not too small cooperators because , assuming random diffusion , a group of small individuals will more easily evaporate from the cluster surface
. they should not be too large either and , consequently , not be able to evade defectors and avoid exploitation .
analogously , for defectors , they should neither be too small in order to stay closer to their prey for longer periods , nor too large so new , more promising regions will not be explored .
therefore , one intuitively expects that intermediate , optimal sizes may be beneficial to cooperation and thus be selected for .
another possible interpretation for an exclusion zone around each agent is its protected region , and the resources within .
whatever the interpretation , it is important to better understand the relevance of area exclusion in these games . as a first approximation
, we consider an effective radius of exclusion , modeled as a hard disk . here
we study , by explicitly taking into account the excluded area of the agents , the interplay between geometry , density and mobility on the capability of a simple model to sustain cooperation and the question of whether the transitions in this class of model have a geometric interpretation .
although the connection between the threshold of geometric percolation , that is independent of the game dynamics , and cooperation has already been reported in refs .
@xcite , we also explore the geometry of clusters of cooperators and defectors , and the connection between their critical properties and the transition between regions with and without cooperation , thus providing a geometric interpretation of these transitions .
we study an off - lattice model @xcite in which the @xmath3 agents living in a square of side @xmath4 ( with periodic boundary conditions ) are characterized by an unconditional strategy ( cooperate , c , or defect , d ) and two independent geometric parameters : an interaction radius @xmath1 and a hard disk radius @xmath0 to account for excluded area .
the radius @xmath1 determines the neighborhood of each agent and , as a consequence , its instantaneous contact network .
the area fraction occupied by the hard disk particles is @xmath5 .
we use @xmath6 as our length scale .
these geometric parameters are illustrated in fig .
[ fig.desenho ] . the particular case studied by meloni _
@xcite is recovered in the limit of point - like particles , @xmath2 . and the interaction radius @xmath1 .
while the former sets an exclusion area around each agent , the latter defines its contact network .
the characteristic length @xmath7 , as indicated on the right , is defined by dividing the available area , @xmath8 , by the area of the square box around each of the @xmath3 particles of diameter @xmath7 , that is , @xmath9.,width=264 ] initially , @xmath3 individuals with probability 1/2 of being either c or d are randomly placed in such a way that there is no overlap between any two individuals @xmath10 and @xmath11 , i.e. , their center - to - center distance @xmath12 satisfies @xmath13 .
moreover , they are allowed to randomly diffuse while playing the pd game with their neighbors .
two agents are considered neighbors if @xmath14 .
a time step is defined as a sequence of @xmath3 attempts of diffusion and a complete , synchronous round of the pd in which each of the @xmath3 agents plays with all its neighbors . during the diffusive part
, the position @xmath15 of the center of particle @xmath10 at time @xmath16 is updated if there is no overlap between particles in the final position : @xmath17 each step has a constant size , @xmath18 , and a random orientation @xmath19 drawn from a uniform distribution in the interval @xmath20 $ ] .
when @xmath21 there is no mobility and we consider here that @xmath22 is small enough so that jumps over other agents do not occur . under mutual cooperation ( defection ) , both receive payoff @xmath23 ( @xmath24 ) as a reward ( punishment ) ; if one cooperates and the other defects , then the latter receives @xmath25 ( temptation ) and the former , @xmath26 . to characterize the pd
, the following inequalities should hold @xmath27 and @xmath28 . in particular
, we use @xmath29 , @xmath30 and @xmath31 , a common parametrization known as the weak form of the pd game .
the evolution follows the finite population analog of the replicator dynamics @xcite .
each individual @xmath10 , after accumulating the payoff from all combats , randomly chooses a neighbor @xmath11 with whom to compare their respective payoffs @xmath32 and @xmath33 . if @xmath34 , then @xmath10 maintains its strategy . on the other hand ,
if @xmath35 , @xmath10 will adopt the strategy of @xmath11 with probability proportional to the payoff difference @xmath36 where @xmath37 and @xmath38 are the number of neighbors of @xmath10 and @xmath11 , respectively . under this update rule , the total number of individuals is kept constant .
most of our results are for @xmath39 , @xmath40 , @xmath41 and @xmath42 .
we then check the robustness of the model by testing finite size effects with up to @xmath43 particles , as well as the dependence on @xmath25 and @xmath22 .
averages are taken over 100 or more samples .
two macroscopic asymptotic quantities , once averaged , are used to characterize the system : the fraction of cooperators @xmath44 ( those , among the @xmath3 agents , that cooperate ) and the fraction @xmath45 of initial conditions whose evolution ends in the absorbing state @xmath46 .
their difference , @xmath47 , is a measure of the coexistence of both strategies .
four regimes are present in the time evolution , as seen by the behavior of @xmath48 in fig .
[ fig.temporal_evolution ] . as is often the case for this class of model , there is an initial drop in the fraction of cooperators from @xmath49 , since small cooperator clusters are easily preyed on in the beginning of the simulation .
as @xmath48 approaches its minimum value at @xmath50 , fluctuations may lead to extinctions in finite size systems .
away from the minimum , the surviving clusters of cooperators resume growth .
these two initial regimes are quite independent of the occupied area fraction , as indicated in fig .
[ fig.temporal_evolution ] by the close proximity of all curves up to @xmath51 . in the third regime
, @xmath48 attains a plateau where it stays , indicating persistent coexistence of both strategies , if @xmath52 is large enough ( in the case of fig .
[ fig.temporal_evolution ] , the threshold is at @xmath53 ) .
when @xmath52 is below the threshold , after the stasis period on the plateau , the system enters the last regime in which cooperators take over the system , that is , @xmath54 . the time to reach this
asymptotic state seems to diverge as the threshold area fraction is approached from below . , @xmath40 , @xmath55 , and @xmath41 for various area fractions @xmath52 .
below @xmath53 , cooperators eventually invade the whole system .
as @xmath52 approaches this threshold , the time spent close to the critical plateau at @xmath56 also increases.,width=264 ] fig .
[ fig.diagram ] summarizes our most important results , showing , in the stationary regime , the average fraction of cooperators @xmath44 as a function of both @xmath57 and @xmath52 .
the lines are independent measures of percolative properties that will be explained below .
several regimes may be identified : two absorbing phases in which all agents eventually become either defectors ( @xmath58 , labeled ` d ' ) or cooperators ( @xmath59 , ` c ' ) and two coexistence ones ( @xmath60 and @xmath61 , ` c and d ' ) .
finite systems may also present a bi - stable phase , in which all initial conditions lead to one of the absorbing states . in this case , although the average cooperativeness still obeys @xmath60 , it differs from the coexistence state since @xmath62 .
however , by increasing the system size , the probability of becoming dominated by defectors goes to zero inside the c region in fig .
[ fig.diagram ] and , taking this into account ( see fig .
[ fig.finite_size ] ) , we already properly labeled it .
we now concentrate on the results for a finite system with @xmath39 particles , and discuss the finite size effects at the end of this section .
there are two limits of the diagram that have trivial results .
for very large values of @xmath1 , very distant particles interact , increasing the number of contacts and decreasing the effects of spatial correlation .
thus , we recover the mean field result in which all agents become defectors . this trivial region of the phase diagram was not explored . in the other limit ,
when @xmath63 , the hard core prevents any interaction between the agents and the fraction of cooperators remains equal to the initial one , @xmath64 and @xmath61 .
this is the trivial coexistence region , labeled ` c and d ' on the left of fig .
[ fig.diagram ] , above the line @xmath65 , corresponding to @xmath66 . immediately to the right of this line ,
interaction , albeit weak , is possible and clusters are formed .
however , they are small and do not favor cooperation , therefore @xmath67 , and this all defector phase is labeled ` d ' . as we discuss below , the transitions between the other phases have geometric origins and are closely related to the percolating properties of the contact networks . and @xmath52 for @xmath39 , @xmath41 and @xmath40 .
the color code is the average fraction of cooperators , @xmath44 . besides the phases corresponding to the absorbing states , ` c ' ( @xmath59 ) and ` d ' ( @xmath58 ) , there are two coexistence phases , ` c and d'.,width=302,height=207 ] [ [ geometric - percolation . ] ] geometric percolation .
+ + + + + + + + + + + + + + + + + + + + + + for the area fractions considered here , non - trivial cooperation first appears around @xmath68 . for small @xmath52 ,
the transition between phases d to c corresponds to a change in stability of the absorbing state , from the defector to the cooperator dominated phase , while for larger @xmath52 , roughly @xmath69 , the emergent phase is one in which cooperators and defectors coexist , c and d. this onset of cooperation is strongly correlated with the appearance of a _ geometric _ percolating cluster , indicated by the steep line in fig .
[ fig.diagram ] at the point where the probability of finding a percolating cluster is @xmath70 .
this is a purely geometric problem of disks with both an inner hard core in addition to a soft , penetrable region , and thus independent of the game dynamics .
in other words , the network of contacts of the percolating cluster spans the whole length of the system . for @xmath71 ,
our result is consistent both with the percolation threshold obtained numerically in refs .
@xcite , and the exact bounds of ref .
@xcite . for finite @xmath52 ,
the threshold is slightly smaller than for @xmath71 , since as @xmath0 increases , the overlap between the disks decreases and percolation is attained with a smaller @xmath1 . [
[ percolating - clusters - of - defectors . ] ] percolating clusters of defectors .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + besides the clusters of particles irrespective of their strategies , we also consider the geometry of clusters composed only by defectors ( that are , in turn , intimately connected with the geometry of cooperator clusters ) , which depends on the particular strategy evolving dynamics .
[ fig.perc ] shows the probability of percolation of d clusters as a function of time , @xmath72 , for @xmath73 .
notice that in this region there always is a percolating geometric cluster .
initially , as the fraction of cooperators decreases towards the minimum , there is a sea of defectors that obviously percolates and all curves overlap at @xmath74 .
it is only when the curves of @xmath48 , for different values of @xmath52 , start to separate , around @xmath51 , that @xmath72 starts decreasing .
the asymptotic probability of there being a percolating cluster of defectors attains a limiting value , as shown in the inset of fig .
[ fig.perc ] , from which the threshold can be obtained . for the particular value of @xmath1 shown in this figure , when the area fraction is below the threshold at ( roughly ) 0.37 , @xmath75 and , as can be seen in fig .
[ fig.temporal_evolution ] , @xmath54 . the dotted transition line in the phase diagram ( fig . [ fig.diagram ] )
is obtained in the same manner : it is the asymptotic value of @xmath52 at which @xmath75 as a function of @xmath1 .
in the c region , finite size fluctuations sometimes lead to the all - d state @xmath76 , but these configurations are not taken into account for the calculation of the asymptotic value of @xmath52 .
this transition line suggests an important ingredient to understand the coexistence between cooperators and defectors : only under the presence of a percolating sea of defectors is that a stable coexistence between cooperators and defectors is possible . in other words , differently from compact clusters of cooperators , isolated groups of defectors either grow and percolate or eventually become extinct . , for several area fractions @xmath52 and the same parameters of fig .
[ fig.temporal_evolution ] .
inset : asymptotic limit of @xmath72 as a function of @xmath52 .
the linear fit indicates that the percolation probability goes to zero at @xmath53 .
the small deviation seen close to this point is due to the long time of convergence.,width=302 ] [ [ finite - size - effects . ] ] finite size effects .
+ + + + + + + + + + + + + + + + + + + + we now discuss how the system size can affect the phase diagram , fig .
[ fig.diagram ] . for finite sized systems ,
a small region inside the c phase has a bi - stable equilibrium , in which all initial conditions lead to an absorbing state , either @xmath77 or 1 .
the all - d state is due to the fact that the population of cooperators becomes quite small during the initial drop in the first generations and , therefore , sensitive to fluctuations which occasionally cause extinctions .
[ fig.finite_size ] shows @xmath78 for two different values of @xmath52 and several system sizes . for point particles
@xmath79 , the absorbing all - c state region grows as the system size increases . on the other hand , for @xmath80 , the all - c region shrinks . in both cases
, however , the c phase width converges to a finite value and the bi - stable region decreases as the system size increases .
furthermore , the size of the c region in the limit of large systems is consistent with the transition line obtained from the percolating defector cluster analysis for @xmath39 , which explains why this phase is not homogeneously painted in fig .
[ fig.diagram ] .
notice that when the system is bi - stable , the average fraction of cooperators is not a good measure since it represents neither one of the final states @xcite . on the other hand , in the coexistence state , both strategies are present in the asymptotic state and , while @xmath60 , @xmath61 . for the @xmath2 case studied by meloni _
@xcite , the @xmath71 line in the phase diagram , there is no coexistence phase and the system eventually enters an absorbing state . absorbing state as a function of @xmath57 for several system sizes and two area fractions @xmath71 and 0.28 . for large values of @xmath57 , as @xmath78 increases for larger system sizes , the existence of the @xmath77 state for point particles ( @xmath71 ) is due to finite size fluctuations .
for @xmath81 the behavior is the opposite and the transition becomes sharper when the system size increases.,width=302 ] [ [ robustness - against - mobility - and - temptation . ] ] robustness against mobility and temptation .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we finally consider the robustness of our results when the mobility @xmath22 and the temptation @xmath25 are varied , fig .
[ fig.mu-t ] .
the top panel shows several values of @xmath22 , with the temptation fixed at @xmath40 . whatever the velocity , no cooperation is possible for @xmath82 due to the absence of a percolating geometric cluster .
notwithstanding , for small mobilities , existence of cooperators is possible in a wide range of @xmath1 .
when comparing the low mobility case with the one with immobile agents @xmath83 , it can be seen that there is an improvement only for low values of the @xmath57 ratio .
for high values of this ratio , the curves overlap , which is expected , since the agents have a large neighborhood that is minimally perturbed by the small random movements . on the other hand , for low values of the ratio ,
the contact range is small and is more affected by the diffusion .
random movements lead to the evaporation of cooperative clusters and , as the velocity increases , cooperation levels decrease until a threshold above which it is no longer possible .
this effect of cooperation enhancement driven by a low mobility is in accordance with previous simulations on a lattice @xcite .
the bottom panel of fig .
[ fig.mu-t ] shows the effect of the parameter @xmath25 . as expected , as the temptation to defect increases , the fraction of cooperators decreases . for @xmath81 and several values of @xmath22 ( top ) ,
including @xmath21 ( empty symbols ) , and @xmath25 ( bottom ) . in both cases , @xmath39 ; for the top panel , @xmath40 while for the bottom one , @xmath41.,width=302 ]
we presented numerical results for an off - lattice model of mobile agents playing the pd game while randomly moving across a closed region .
this model combines ingredients found in two distinct models for such systems : the excluded area found in lattice simulations and an available continuous space .
we arrive at two important results : first , if the agents are not point - like , a non - trivial coexistence phase with cooperators and defectors becomes possible ( besides the trivial one when the exclusion range is larger than the interacting radius ) ; second , it is possible to geometrically interpret , in terms of percolation , the observed transitions . an important role is played not only by geometric percolation , irrespective of the game dynamics , but also by the percolative properties of defector clusters , whose threshold depends on the details of the game .
when the agents interaction is prevented by the hard core , trivial coexistence between cooperators and defectors is possible . beyond that region of the phase diagram ,
no cooperation is possible in the absence of a percolating geometric cluster .
random mobility provides an evaporating mechanism for groups of cooperators what is detrimental to cooperation as isolated cooperators are easily preyed on .
however , in the presence of a percolating clusters , it is easier for a detaching cooperator to be in contact with another cluster and be protected .
when hard cores are included , movements are hindered and the agents spend some time rattling around the same region , while large displacements become less probable , what increases the correlation among agents and benefits cooperation even further .
this localization is probably the mechanism responsible for the coexistence between cooperators and defectors that is not present for @xmath71 .
interestingly , the presence of defectors is only possible if they form a percolating cluster and no finite cluster of defectors is stable : they either grow , merge with others and span the whole lattice or the isolated cluster becomes extinct .
this is our main result : finite size cooperators and defectors , whose hard core is an effective , averaged interaction restraining their movements , are able to coexist over a broad region of the phase diagram only if defectors are organized in a interconnected cluster , a percolating sea of defectors .
a similar effect was found for the public goods game played on a lattice with empty sites and no mobility @xcite .
it would be interesting to investigate whether this condition for coexistence between cooperators and defectors also occurs in lattice models , where excluded area is inherent to the formulation of the problem . in this manuscript we focused on the particular homogeneous case of equal sizes and equal velocities for both cooperators and defectors
. following refs .
@xcite , it is essential to explore the whole @xmath84 parameter space and the dependence on the chosen dynamic rule in order to check the robustness of cooperation .
furthermore , several extensions are possible .
for example , velocities may not be constant @xcite and depend on the neighborhood @xcite or strategy .
the hard core radius may also correlate with strategy , @xmath85 and @xmath86 for cooperators and defectors , respectively .
in particular , if individuals coevolve with mutations : is there an optimal equilibrium radii ratio to which the system converges to or , instead , a permanent arms race ?
what are the effects of having size dispersion ? if velocity and size coevolve along with strategies , defectors may become small and fast while cooperators become large and slow .
finally , what happens if there is a fraction of fundamentalists ( both cooperators and defectors or just defectors ) whose strategies or positions never change ? in all these cases , it is important to study also the geometric properties of the interfaces between cooperator and defector clusters since these are the places where all strategy flips occur . from a more physical perspective
, it would be interesting to find out , for each transition line , which is the dynamical universality class that it belongs to @xcite . these and other relevant questions are being considered .
research partially supported by the brazilian agencies cnpq , capes and fapergs .
jja also thanks the inct - sistemas complexos ( cnpq ) for partial support .
we thank the supercomputing laboratory at universidade federal da integrao latino - americana ( lcad / unila ) , where the simulations were run , for computer time . | we study the conditions for persistent cooperation in an off - lattice model of mobile agents playing the prisoner s dilemma game with pure , unconditional strategies .
each agent has an exclusion radius @xmath0 that accounts for the population viscosity , and an interaction radius @xmath1 that defines the instantaneous contact network for the game dynamics .
we show that , differently from the @xmath2 case , the model with finite sized agents presents a coexistence phase with both cooperators and defectors , besides the two absorbing phases in which either cooperators or defectors dominate .
we provide , in addition , a geometric interpretation of the transitions between phases . in analogy with lattice models ,
the geometric percolation of the contact network ( i.e. , irrespective of the strategy ) enhances cooperation .
more importantly , we show that the percolation of defectors is an essential condition for their survival . differently from compact clusters of cooperators ,
isolated groups of defectors will eventually become extinct if not percolating , independently of their size . |
the spiral galaxy maffei 2 was first detected by paolo maffei in 1968 on a near - infrared schmidt plate @xcite .
recent @xmath8-band photometry of this galaxy reveals a large highly - inclined barred spiral of hubble type sbc with isophotes visible out to @xmath10 from the nucleus @xcite . if the galaxy lies within 2 - 6 mpc as suggested in previous studies @xcite , it must be among the dominant galaxies in our galactic neighbourhood , which raises the question of its dynamical role in the early evolution of the local group .
unfortunately , a precise distance to maffei 2 has remained elusive owing to the heavy obscuration by dust in the milky way disk associated with its low galactic latitude ( @xmath11 ) .
attempts to determine the galactic extinction have so far produced ambiguous results : @xcite derived a value of @xmath12 mag by comparing the nuclear spectrum with that of m31 , and estimated that the galaxy suffers more extinction than the giant elliptical maffei 1 by about 1 magnitude in @xmath13 .
however , a modern estimate of the reddening of maffei 1 has been measured by fingerhut et al .
( 2003 ; hereafter @xcite ) using a well - defined correlation between the @xmath14 index and @xmath15 colour .
they find @xmath16 mag , which is over 1.5 mag lower than the result reported by @xcite .
there is clearly a need to revisit the problem of the galactic extinction of maffei 2 before an attempt can be made to determine its distance . in this study , we derive the extragalactic extinction of maffei 2 using a relationship among spiral galaxies between the extragalactic extinction of hii regions and the column density of extragalactic hydrogen gas along the line of sight of the hii regions .
the correlation arises because the bulk of the extragalactic extinction of an hii region is due to dust outside the region of emitting gas , where the extinguishing material is widely distributed and well - mixed with hydrogen gas @xcite .
a correlation between @xmath17 and the annular - averaged column density of hi has been observed in both ngc 2403 @xcite and ic 342 @xcite , substantiating the above claims . using a sample of giant extragalactic hii regions in late - type spirals , we construct the dust
gas relation and use it to determine the extragalactic extinction of two hii regions in maffei 2 . by subtracting the extragalactic extinction from the total extinction of the hii regions derived from their balmer decrements ,
we obtain a measurement of the galactic extinction . in deriving the galactic extinction in various bandpasses
, we account for shifts in the effective wavelengths of broadband filters , which can be significant for highly reddened galaxies ( see * ? ? ?
* ) . to check the validity of our result
, we use the method applied to maffei 2 to obtain the galactic extinction of the nearby late - type spiral ic 342 , for which an accurate independent estimate is available owing to its more moderate obscuration .
armed with the extinction to maffei 2 , we determine its distance using the tully - fisher relation constructed in the @xmath8-band . our observations of hii regions in maffei 2 and ic
342 are presented in
[ observations ] and reductions are outlined in [ reductions ] . in [ measurements ] , we describe the measurements of the total extinctions ( galactic plus extragalactic ) and the hydrogen column densities for the observed hii regions . in [ extinction ] , we construct the relation between dust and gas and use it to determine the galactic extinction of maffei 2 and ic 342 . in [ distance ] , we re - determine zero - points for the extragalactic distance scale in a self - consistent fashion .
we then derive the tully - fisher distance to maffei 2 as well as a revised distance to ic 342 using recent observations of cepheids by @xcite .
the revised properties of both galaxies in addition to maffei 1 are presented in [ implications ] and the implications for the local group are discussed .
long slit spectroscopic observations of two hii regions in maffei 2 were obtained on 1997 march 2 and 3 ut with the ritchey
chrtien spectrograph on the mayall 4m telescope at kitt peak national observatory .
observations were optimized to detect both @xmath18 and @xmath19 emission to determine the extinction toward maffei 2 .
the t2 kb 2048@xmath202048 ccd with 24 @xmath0 m pixels was used in combination with the kpc10-a grating for wavelength coverage between 3500 and 7600 at 7 resolution .
the spatial scale at the focal plane was 0 per pixel .
the slit was set to a width of 2 and aligned in the north south direction .
the slit was placed over the southwest portion of maffei 2 at the position of the brightest hii region ( * ? ? ?
two individual clumps were also detected within a second hii region at the position of object number 2 in @xcite .
two 1800second exposures were taken on the night of march 2 , and an additional three 1800second exposures were obtained the following night .
the effective air masses for the first and second night were 1.65 and 1.54 , respectively .
inspection of a single 1800second dark frame showed that the total dark count rate was negligible compared to the sky .
internal and twilight flats were obtained to correct for variations over small and large spatial scales , respectively , on the ccd .
henear arc
lamp spectra were used for wavelength calibration .
flux calibrations were achieved by observing the standard stars feige 67 and g191b2b @xcite .
standard star exposures were interspersed among object exposures over a range of airmasses spanning those of maffei 2 .
in addition to the maffei 2 observations , spectrophotometry of two hii regions in ic 342 was carried out to augment existing data for that galaxy .
the spectra were acquired on 1992 january 28 ut with the 2.3-m f/9 bok telescope and boller and chivens spectrograph at steward observatory .
the detector was a texas instruments @xmath21 ccd with @xmath22 square pixels .
observations were made in first order with a 300 lines @xmath23 grating blazed at @xmath24 .
the grating was rotated to give coverage from 4300 to @xmath25 , over which the dispersion was @xmath26 .
the slit was opened up to @xmath27 to enable absolute calibration of line fluxes and to negate the possibility of errors in line ratios as a result of differential refraction .
precise pointing was guaranteed by rotating the slit to a position angle ( @xmath28 ) which admitted the light of a reference star only when the hii region was centered .
a single 1800-second exposure was acquired , in which @xmath18 and @xmath29\lambda 6584 $ ] were clearly resolved . in order to minimize the effects of flexure on flat fielding , a spectrum of an internal quartz lamp was acquired at the same location immediately following the target exposure .
a helium - argon lamp was observed at a comparable position to calibrate the wavelength scale .
a spectrum of twilight was acquired at the beginning of the night to map out the illumination pattern . to calibrate fluxes ,
the standard stars hiltner 102 , feige 15 , g191b2b , and hd 84937 were observed with the same wide slit as employed for the hii regions .
the maffei 2 spectra were reduced using standard iraf routines .
reductions were carried out separately for each night . the object and flat field frames were overscan corrected and bias corrected .
the internal flat and sky flat frames were combined to produce a single internal flat frame and a single averaged sky flat frame .
the averaged sky flat frame was corrected for pixel - to - pixel variations in response using a normalized internal flat frame .
an illumination image was created from the processed sky flat to correct for the slit function .
then , each object frame was corrected by dividing the product of the illumination image and normalized internal flat . with two - dimensional fits to the arc spectra
, geometric distortions were corrected so that the dispersion axis was made perpendicular to the spatial axis .
spectra of the two standard stars were combined to calibrate fluxes .
air mass corrections were derived using a mean atmospheric extinction curve for kpno .
the five maffei 2 spectral exposures were then aligned , shifted , and combined into a single spectral frame .
the ic 342 spectra were reduced using standard iraf routines in a similar manner as for the maffei 2 spectra .
final one - dimensional flux - calibrated spectra were obtained via unweighted summed extractions .
table [ tablines ] lists the observed flux ratios relative to @xmath30 for the hii regions observed in maffei 2 and ic 342 .
the extinction of a source is best quantified by its optical depth at some wavelength . knowing optical depth , it is possible to evaluate the extinction in any filter by applying an appropriately scaled reddening law to the spectrum and integrating through the response function for the filter .
the best choice of wavelength is 1 @xmath0 m , as it falls in a part of the reddening law which is not very sensitive to environment , and because the optical depth there is comparable numerically to @xmath31 ( see * ? ? ?
the total extinction ( galactic plus extragalactic ) of an hii region can be determined from the degree to which the observed @xmath32 ratio differs from the theoretical intrinsic ratio .
the optical depth at 1 @xmath0 m ( @xmath33 ) can be computed from @xmath34/(r^1_{{\ensuremath{\rm h\alpha } } } - r^1_{{\ensuremath{\rm h\beta}}})\ ] ] where @xmath35 is the intrinsic balmer ratio , @xmath36 is the observed balmer ratio corrected for underlying stellar absorption , and @xmath37 and @xmath38 are the reddening coefficients normalized to @xmath33 ( i.e. , @xmath39 , where @xmath40 is the extinction corresponding to @xmath33 at the wavelength @xmath41 ) . to estimate the reddening coefficients , we require a monochromatic extinction curve appropriate for the diffuse interstellar medium , which is primarily responsible for obscuring extragalactic sources . @xcite has developed an algorithm for determining the monochromatic reddening law associated with any particular value of the ratio of total to selective extinction , @xmath42 . for the diffuse interstellar medium , @xmath43 for a star of zero colour in the limit of zero extinction @xcite .
the reddening law chosen to evaluate the reddening coefficients is that given by the algorithm of @xcite which , when applied to the spectrum of vega , yields @xmath44 after integrating the flux passed by response curves characterizing the @xmath45 and @xmath13 filters .
the resulting balmer coefficients are @xmath46 and @xmath47 .
the intrinsic balmer ratio @xmath35 for each hii region was estimated from the emissivities of @xcite .
an initial approximation for @xmath33 was derived from eq .
[ eqtau ] using @xmath48 , which was computed by assuming a temperature of 8000 k and an electron density of 100 @xmath49 , these being typical values for giant extragalactic hii regions @xcite .
then , the approximation for @xmath33 was used to correct @xmath50 } \lambda3727$]/@xmath19and @xmath51 } \lambda5007$]/@xmath19 for reddening using reddening coefficients @xmath52 and @xmath53 computed in the same manner as for @xmath18 and @xmath19 .
the corrected oxygen line ratios were used to estimate the oxygen abundance , @xmath54 , via the semi - empirical calibration of @xcite .
next , @xmath55 was used to obtain an improved approximation of the hii region temperature using the theoretical correlations between electron temperature and oxygen abundance formulated by @xcite .
the correlations are parameterized by the volume - averaged ionization parameter ( @xmath56 ) and the upper mass limit for the ionizing stars .
however , in the oxygen abundance range inhabited by the reference hii regions ( @xmath57 ) , the relations converge to straight lines of nearly - identical slope with a spread of approximately @xmath58 k for intermediate values of @xmath56 , which are typical of observed hii regions @xcite .
thus , the temperature ( t ) in kelvins is given by @xmath59 eq .
[ eqtemp ] was used to derive the next approximation of @xmath35 , which was then substituted back into eq .
[ eqtau ] to obtain the next estimate of @xmath33 .
this process was iterated until @xmath35 , and thereby @xmath33 , converged .
table [ tabhiiex ] lists the observed and intrinsic fluxes as well as the oxygen abundances , temperatures and total extinctions found for the hii regions in maffei 2 and ic 342 .
the surface brightnesses of hi 21-cm and co 2.6-mm line radiation are traced by atomic and molecular hydrogen gas , respectively , which is distributed all the way from the far side to the near side of a galaxy . in the milky way ,
the scale heights of the hi and @xmath60 gas are @xmath61 pc and @xmath62 pc , respectively @xcite .
the hii regions , being associated with the young stellar population , are confined to a layer with a scale height of only @xmath63 pc @xcite .
the column density of extragalactic gas in the foreground of an hii region can therefore be approximated as half the column density of hydrogen derived from observations of hi and co along the line of sight , with some scatter caused by deviations in position from the exact mid - plane and by concentrations of dust in the immediate vicinity of the hii region .
thus , the hydrogen column density @xmath64 in atoms @xmath65 along the line of sight is given by @xmath66 where @xmath67 and @xmath68 are the column densities of hi and @xmath69 respectively .
the @xmath68 term is doubled to preserve the proportionality between the number of dust particles and hydrogen atoms .
the annular - averaged column densities of hi at the de - projected radii of the two hii regions in maffei 2 were measured from the hi radial profile of @xcite .
the co ( j=1 - 0 ) intensities at the locations of the hii regions were obtained from the radial profile provided by mason & wilson ( private communication ; see also @xcite ) . for ic 342 ,
the hi column densities and co intensities at the locations of the hii regions were measured from the radial profiles of @xcite and @xcite , respectively . the conversion from co intensity ( @xmath70 ) to @xmath60
column density was found to have a metallicity dependence in independent studies by @xcite and @xcite .
the conversion relation adopted here is from the former source , in which measurements of @xmath71 and @xmath72 are determined in a homogeneous manner by applying the virial theorem to individual molecular clouds in nearby spiral and dwarf irregular galaxies . for each hii region ,
the conversion factor ( @xmath73 ) along with the hi and @xmath60 column densities are summarized in table [ tabhiiex ] .
we have constructed the relationship between extragalactic extinction and hydrogen column density for a sample of 74 reference hii regions in 10 spirals of type sbc and scd .
we have restricted our sample to galaxies for which ( 1 ) the galactic extinction is small and thus well - known ( i.e. , @xmath74 mag ) ; ( 2 ) the radial profile of hi has been determined from aperture synthesis maps ; and ( 3 ) there is a published radial profile of the co ( j=1 - 0 ) line .
we have further required that the constituent hii regions be non - nuclear to insure the reliability of abundance calibrations , where `` non - nuclear '' is defined here as having a radius greater than 15% of the host galaxy s disk scale length .
in addition , we have rejected any hii region located beyond the radial boundaries of the hi maps .
lastly , all hii regions were required to have measurements of the flux at @xmath18 and @xmath19as well as the equivalent width of the @xmath19 emission line in order to account for the depression of balmer emission by underlying balmer absorption from stars ( for typical hii regions , the equivalent width of @xmath18 emission is so strong that the correction to the @xmath18 flux is negligible ) .
the galaxies meeting the above criteria are listed in table [ tabspi ] .
the total extinctions of the reference hii regions were derived in a manner consistent with those employed for maffei 2 and ic 342 ( see [ s_extot ] ) .
several of the hii regions were missing measurements for the oxygen line fluxes as a consequence of poor signal - to - noise .
their temperatures were determined from a correlation between temperature and galactocentric radius found for the hii regions for which oxygen line fluxes were available . as can be seen in fig .
[ figtr ] , the slope and intercept of the correlation appear to have a dependence on spiral morphology in the sense that hii regions in late - type spirals have higher temperatures .
we therefore adopt two linear fits ; one for the spirals of type sbc and one for those of type scd .
the fits are described by @xmath75 @xmath76 where t is in kelvins and @xmath77 is the deprojected galactocentric distance of the hii region normalized to the host galaxy s disk scale length .
the root - mean - square ( rms ) deviations for the sbc and scd samples are 562 k and 934 k , respectively .
the extragalactic extinction of each hii region ( quantified by @xmath78 ) was extracted from the total extinction ( @xmath79 ) observed in the balmer decrement ( eq . [ eqtau ] ) by subtracting the galactic extinction ( @xmath80 ) derived from the galactic reddenings of ( * ? ? ?
* hereafter sfd ) . for each host galaxy
, we computed @xmath80 from @xmath31 using the iterative process outlined in @xcite .
briefly , the integrated spectral energy distribution ( sed ) of a typical unreddened elliptical galaxy was extinguished by successive approximations of @xmath80 with the aid of the scaled monochromatic reddening law described in [ s_extot ] .
the application of this process was greatly facilitated by the york extinction solver ( yes ) .
the total gas column density ( @xmath64 ) at the location of each reference hii region was determined according to the procedure outlined in [ s_gas ] . the gas column densities and extinctions of the hii regions
are provided in table [ tabhii ] along with the sources of the hi and co data .
the correlation between @xmath78 and @xmath64 is shown in fig .
[ figdust ] and does not appear to depend on spiral morphology within the range sbc to scd . we have therefore adopted the following linear least - squares fit to the entire reference dataset : @xmath81 where @xmath82 is in units of @xmath83 . the rms deviation in @xmath78 is 0.17 .
[ eqdust ] improves upon the dust
gas relation found previously by @xcite by expanding the hii region sample to include recent observations , by excluding ngc 6946 due to its high galactic extinction , and by including @xmath60 in the gas diagnostic .
this last step has reduced the scatter considerably .
the error in an estimate of @xmath78 for an hii region from eq .
[ eqdust ] can be computed from @xmath84 ^ 2 + \sigma_{{\ensuremath{\tau_1}}}^2/n\ ] ] @xmath85 is the standard error associated with @xmath86 , the slope of the dust
gas correlation ( eq . [ eqdust ] ) .
@xmath87 is the measurement error associated with the estimate of @xmath64 for the hii region .
@xmath88 is the rms deviation of 0.17 in @xmath78 of the @xmath89 reference data points about the linear fit .
@xmath90 is the mean @xmath82 for the hii region sample , found to be @xmath91 .
the measurement uncertainty in @xmath82 for each reference hii region was computed from one half of the quadrature sum of the errors in @xmath67 and 2@xmath68 .
the greatest source of uncertainty in a measurement of @xmath82arises from taking the annular average of intensity measurements at a given radius .
the uncertainty increases with distance from the nucleus ; the larger the circumference of an annulus , the larger the area of galaxy contained within the annulus , and therefore the greater the possibility of intensity fluctuations .
we have derived an expression for this uncertainty by examining the reference galaxies for which measurements of hi were supplied separately for each half of the disk . for a given annulus
, the difference between the column densities within each semi - annulus ( @xmath92 ) relative to the mean ( @xmath93 ) is found to be a linear function of the fractional galactocentric radius @xmath77 , with an rms deviation of 0.16 .
the uncertainty in a measurement of @xmath67 from an annular - averaged radial distribution can therefore be estimated from @xmath94 adopting this approach , the average error in @xmath67 for the reference hii regions is 15% .
the dominant contributions to the uncertainty in @xmath68for each reference hii region are the measurement uncertainty in @xmath95 , for which we find a mean value of 26% , and the uncertainty in the co to@xmath69conversion factor ( @xmath73 ) .
this latter source is dominated by the scatter of @xmath96 in the adopted metallicity relation ( see [ s_gas ] ) and the measurement uncertainty in the oxygen abundance , for which we adopt 0.2 dex as recommended by @xcite . taking the above sources into account
, we find a mean measurement uncertainty in @xmath82 of 19% for the reference hii regions .
we note that the mean measurement uncertainty in @xmath82for the sbc galaxies exceeds that found for the scd galaxies by 6% , as a result of the considerably larger uncertainties in @xmath68 .
this is likely due to the higher @xmath69 content in the sbc galaxies , as nearly 40% of the scd hii regions are found in regions of negligible co. the measurement uncertainty in @xmath78 for each hii region is simply the quadrature sum of the uncertainties associated with @xmath79 and @xmath80 .
the uncertainty in @xmath79is dominated by the measurement error in the observed @xmath97 flux , given the weak dependence of the intrinsic balmer ratio on temperature .
( the mean error in @xmath35 owing to the temperature estimate was found to be under 2% and was therefore considered negligible . ) the mean measurement error in @xmath36 is 11% , which was computed from the quadrature sum of the signal - to - noise error in the flux and a 10% error arising from the uncertainty in the overall response correction @xcite .
the contributions from the uncertainties in the reddening coefficients were found to be negligible . adopting a 0.16 fractional error in @xmath80as recommended by sfd ,
the mean measurement uncertainty in @xmath78 for the reference hii regions is 0.10 .
thus , the scatter of 0.17 in the dust gas relation may be mostly due to the measurement uncertainties .
the scatter above that associated with the measurement errors may be arising from dust concentrated near the hii regions , which is not traced by @xmath64 .
the non - zero intercept proves the existence of such localized dust .
however , the fact that this scatter does not swamp the dust
gas trend implies that the extinction due to this dust is roughly constant among hii regions . a value of @xmath78 obtained from eq .
[ eqdust ] therefore includes both the optical depth due to widespread dust in the extragalactic foreground as well as the average optical depth due to dust associated with the star formation regions .
the derived extragalactic extinctions for the hii regions in maffei 2 and ic 342 are given in table [ tabhiiex ] .
the quoted errors were derived in the same manner as for the reference hii regions ( see [ s_err ] ) .
we find mean total extinctions of @xmath98 for maffei 2 and @xmath99 for ic 342 .
@xcite estimated the total reddening of hot stars in ic 342 from a comparison of the @xmath100 colour of the blue `` plume '' in the colour - magnitude diagram with that of ic 1613 .
they obtained @xmath101 mag , corresponding to @xmath102 based on the sed of a b0 v star .
their result is consistent within errors with the mean total extinction found for hii regions in ic 342 . in fig .
[ figdust ] , the total extinctions for the heavily obscured hii regions are plotted over the dust gas correlation found for the reference hii regions .
the weighted mean of the extinction offsets of the heavily obscured hii regions imply galactic extinctions of @xmath103 for maffei 2 and @xmath104 for ic 342 .
the sfd reddening for ic 342 is @xmath105 mag , which corresponds to a galactic optical depth of @xmath106 according to the treatment outlined in [ s_gasdust ] .
this value is consistent with our estimate of @xmath107 within errors , which validates our extinction analysis in general as well as the extinction estimate for maffei 2 in particular . while sfd caution against the use of their reddening maps for objects within @xmath108 of the galactic plane , the galactic latitude of ic 342 is safely outside this range at @xmath109 .
we therefore have no reason to reject the sfd value of @xmath80 for this galaxy , so we adopt the weighted mean of both values , which yields @xmath110 . the extinction parameters and corrected magnitudes of maffei 2 and ic 342 are given in table [ tabprop ] .
the lower values of the reddening coefficients @xmath111 observed for maffei 2 for each broadband filter @xmath112 are due to shifts in the effective wavelengths of @xmath112 towards the red caused by the higher galactic extinction and inclination .
as explained in [ s_extot ] and [ s_gasdust ] , analyses in this paper are founded upon a modern framework for handling extinction which eliminates the biases that accompany traditional approaches . as a consequence , zero - points for distance indicators employed in the hst key project on the extragalactic distance scale ( hereafter @xcite )
must be updated using the same approach .
specifically , revised distances are required to the ursa major cluster , which defines the tully - fisher ( tf ) relation employed here , and to the coma cluster , which defines the fundamental plane ( fp ) used to determine the distance of maffei 1 in @xcite . behind the tf and fp calibrations
are cepheid distances to nearby galaxies , which in the @xcite are anchored to the large magellanic cloud ( lmc ) . as a result ,
the period - luminosity ( pl ) relations for the lmc must be re - visited .
the @xcite pl relations come from @xcite , and the extinction corrections inherent to the pl relations , which are based on the apparent magnitude of the red clump , are those of @xcite .
@xcite presumed that the red clump has a constant absolute magnitude in cousins @xmath8 ( henceforth @xmath113 ) , and that changes in apparent magnitude are due to fluctuations in extinction with respect to some zero - point .
the zero - point was determined from colour - magnitude diagrams for an eclipsing binary ( hv 2274 ) and for two young open clusters .
it appears that a reddening law for the lmc @xcite was used to define @xmath31 for each of these sources ( see * ? ? ?
* ) , but the determination of corrections to @xmath31 from red - clump magnitude fluctuations and the conversion of total reddenings to extinctions in @xmath13 and @xmath113 were based upon extinction coefficients tabulated by @xcite , which are founded upon the broadband reddening law of @xcite for the milky way . fortunately , the mean reddening of the cepheids behind the @xcite pl relations , @xmath114 mag ( range 0.11 to 0.20 mag ) , is very close to the value measured for the objects defining the zero - point ( 0.13 to 0.15 mag ) .
this means that red - clump adjustments to the cepheid extinctions amount to second - order corrections , and any errors in these adjustments ought to average out .
thus , red clump stars are of less concern than the matter of what sets the zero - point .
the reddening that @xcite adopted for the binary was @xmath115 mag @xcite .
however , @xcite derived @xmath116 mag by simultaneously solving for atmospheric temperatures and the interstellar extinction curve . in making their determination , @xcite incorporated photometry from @xcite .
this removed the reddening degeneracy which afflicted a preliminary determination criticized by @xcite .
the result of @xcite has to be taken very seriously because the sed , reddening law , and colour excess are not only consistent with the lmc environment , but also consistent with each other .
it is concluded that the @xmath31 scale of @xcite is too high by @xmath117 . correcting for the change in the reddening zero - point
, one gets @xmath118 mag for the mean reddening of lmc cepheid fields as tracked by early - type stars .
this is close to the value ( 0.10 mag ) adopted by @xcite . according to @xcite ,
@xmath119 mag is the galactic reddening based upon what is observed in annuli around the lmc . with the rest - frame elliptical sed of @xcite and the reddening law of @xcite tuned to give @xmath120 for vega , the optical depth of galactic dust toward the lmc at @xmath121 is @xmath122 . adopting the sed b12 iii of @xcite for the stars in hv 2274 , a heliocentric velocity of 265 @xmath123@xcite and the reddening law of @xcite for dust inside the lmc ,
the mean value @xmath118 mag for the cepheid fields leads to an optical depth @xmath124 for dust `` localized '' in the lmc . in turn
, applying the redshift and optical depths to seds of cepheids , which here are adopted to be g0 supergiants based upon the periods ( there is hardly any difference in results between f8 i and g2 i ) , one gets the following sums of galactic extinction , localized extinction , and the difference in k - corrections in @xmath13 and @xmath113 : @xmath125 mag ( versus 0.476 mag from @xcite ) , @xmath126 mag ( versus 0.288 mag from @xcite ) , and @xmath127 mag ( versus 0.147 mag from @xcite ) .
this means that the zero - points for the pl relations for the lmc cepheids change as follows . for a distance modulus of 18.50 mag , @xmath128 where p is the period in days . as the starting point for correcting @xcite cepheid distances , apparent distance moduli @xmath129 in @xmath13 and @xmath130 in @xmath113 recorded in the @xcite
were adopted .
first , all were corrected to account for the new pl relations for the lmc ( i.e. , the revised extinction zero - point for lmc cepheids ) .
next , foreground values of @xmath31 from @xcite were converted to @xmath80 by applying the reddening law of @xcite ( tuned to give @xmath131 for vega ) to the rest - frame elliptical sed of @xcite .
values of @xmath78 were derived individually for each galaxy from @xmath132 , the redshift , and the sed of a g0 supergiant ( a cepheid ) , again using the reddening law of @xcite .
then , values of galactic extinction , @xmath133 , localized extinction , @xmath134 , and the k - correction , @xmath135 , were computed .
the corrections were subtracted from @xmath130 to arrive at corrected distance moduli @xmath136 on a scale where the distance modulus of the lmc is 18.50 mag .
note that @xmath137 can reach as high as 0.009 mag for a virgo cluster galaxy , so k - corrections are just starting to become relevant [ @xmath137 , of course , affects the determination of @xmath78 via @xmath138 .
despite all the revisions , cepheid distances increased by only @xmath139 mag on average . as in the @xcite ,
metallicity - corrected distance moduli @xmath140 were computed from oxygen abundances @xmath141 estimated from hii regions via the calibration of @xmath142 } \lambda3727}}+{\ensuremath{{\rm [ o iii ] } \lambda4959}}+{\ensuremath{{\rm [ o iii ] } \lambda5007}})/{\ensuremath{\rm h\beta}}$ ] provided by @xcite .
formally , @xmath143 where @xmath144 from @xcite , @xmath145 and @xmath146 .
the adopted value of @xmath147 is slightly more negative than that employed in the @xcite .
@xcite specify the cepheid calibrators used to establish the zero - point for the tf relation and in turn distances to the ursa major and coma clusters .
besides correcting the cepheid distances , magnitudes for both the calibrators and the cluster galaxies must be re - computed using a self - consistent approach to extinction . to make this possible , s. sakai ( personal communication ) provided the apparent magnitudes and corrections employed in the @xcite . in the process , an error in the @xcite galactic extinction corrections was found ; the @xmath113-band correction for all tf galaxies was accidentally set to 59% of the desired value . for each cepheid calibrator ,
new estimates of the galactic extinction and the k - correction were computed from the value of @xmath31 given by @xcite and the redshift using the sed of @xcite closest to the revised hubble type of the calibrator , adjusted appropriately for tilt .
the internal extinction corrections of @xcite were retained . with weights set by the random uncertainties in the absolute magnitudes ,
the mean offset between the new and old absolute magnitudes amounts to @xmath148 improved galactic extinction and k - corrections for ursa major galaxies were computed from the mean value of @xmath31 ( from @xcite ) and redshift , which were judged from the mean cluster coordinates and velocity recorded by @xcite . to this end , the sbc sed of @xcite was adopted , adjusted to the mean tilt and luminosity of the local calibrators . for consistency with the calibrators ,
the corrections for internal extinction adopted by @xcite were retained . with @xmath149 mag and a heliocentric velocity of 899 @xmath123 , the weighted mean offset between the new and old apparent magnitudes corrected for galactic extinction , redshift , and internal extinction amounts to @xmath150 the revisions to apparent and absolute magnitudes lead to an offset in the distance modulus for ursa major of @xmath151 the distance modulus for ursa major determined by @xcite was 31.58 mag , so the revised distance modulus comes out to be @xmath152 mag for an lmc distance modulus of 18.50 mag .
the quoted error is the random error in the zero - point of the tf relation .
the distance to the coma cluster adopted in this paper is the mean of results from the tf relation and the fp .
the revision to the tf distance for coma was undertaken in exactly the same manner as that for ursa major .
apparent magnitudes shift as follows : @xmath153 combining this with the absolute magnitude shift computed above for the cepheid calibrators , the distance modulus shifts by @xmath154 since @xcite derived a distance modulus of 34.74 mag , the new distance modulus for coma works out to be @xmath155 mag , where the uncertainty is the random error in the zero - point of the tf relation .
the fp analysis for the @xcite was conducted by @xcite . for any particular galaxy , the zero - point of the fp
is defined by @xmath156 where @xmath157 is the effective metric radius ( of an @xmath158 law ) , @xmath159 is the velocity dispersion , and @xmath160 is the mean surface brightness within the effective radius . of course
, one does not measure @xmath157 , but rather @xmath161 , the effective radius in angular units .
if @xmath157 is replaced by @xmath161 , variations in @xmath162 from galaxy to galaxy arise from variations in distance .
the fiducial value of @xmath162 was defined using the leo group and the virgo and fornax clusters , in which cepheid calibrators are located .
revisions to extinction or to k - corrections affect @xmath162 through @xmath160 and through @xmath157 , the latter being determined by the distances assigned to the reference clusters ( which depend upon the treatment of cepheids ) .
photometry in @xmath13 was used to define @xmath162 for leo , virgo and fornax , and photometry in gunn @xmath163 was used to define @xmath162 for the coma cluster .
revising extinction estimates as recommended by @xcite , and computing k - corrections self - consistently , the changes to @xmath160 decrease the weighted mean value of @xmath162 for the calibrating clusters by @xmath164 and decrease the value of @xmath162 for coma by @xmath165 .
the revised cepheid analysis described above moves the distance moduli for leo , virgo and fornax farther away by @xmath166 on average with respect to the distance moduli adopted in the @xcite . in this computation , clusters were weighted on the basis of the uncertainty in @xmath162 @xcite .
the combination of all effects leads to an increase in the distance modulus of the coma cluster by @xmath167 with respect to the @xcite . on a scale where the distance modulus of the lmc is 18.50 mag , the revised distance modulus
to coma becomes @xmath168 mag , where the uncertainty is that due to random errors .
the distance moduli for coma derived from the tf relation and the fp differ by only 0.028 mag .
we adopt the unweighted average , which is @xmath169 mag . in this paper , the authors prefer to adopt the maser distance to ngc 4258 as the zero - point for distances , rather than the lmc .
this galaxy was among the cepheid calibrators in the @xcite . from the re - analysis above , @xmath170 mag with a random error of 0.07 mag .
this is 0.016 mag higher than the @xcite value . from (
* see also @xcite ) , @xmath171 mag , so @xmath172 referencing to the maser zero - point , the distance moduli for ursa major and coma are @xmath173 mag and @xmath174 mag , respectively . a proper treatment of the shift in the @xcite hubble constant ( @xmath175 ) due to the above extinction corrections would necessitate a re - analysis of the zero - points for @xmath176 distance indicators behind the measurement . since our distance measurements to maffei 2 and ic 342
do not depend on @xmath175 , we do not require a value of @xmath175 on the extinction and distance scale adopted in this paper .
however , for the sake of the discussion in
[ implications ] in which we compare the new distance to the ic 342/maffei group with that implied by the hubble flow , we adopt the @xcite value of @xmath177 after ( 1 ) correcting for the mean increase of @xmath178 mag in the distance moduli of the cepheid calibrators due to the extinction corrections described above ; and ( 2 ) correcting for the decrease of 0.236 mag in all @xcite distance moduli due to the shift from the lmc to the maser zero - point . after applying these corrections ,
we obtain @xmath179 .
this value is consistent with the most recent measurement of @xmath180 obtained from the sunyaev - zeldovich effect measured by the chandra x - ray observatory for high - redshift galaxy clusters @xcite , a result which is completely independent of the @xcite .
in addition , @xcite use the planetary nebula luminosity function ( pnlf ) to derive a distance to ngc 4258 which is in close agreement with the maser distance and which increases the @xcite value of @xmath175to @xmath181 .
the agreement between these two studies substantiates the distance scale adopted in this paper . for a very long time
, studies of maffei 2 were encumbered by the lack of an apparent magnitude .
a reliable measurement of the total apparent magnitude of maffei 2 was finally made by @xcite in @xmath8 , which allows application of the tf relation to determine the distance .
we require a tf relation in which the luminosity diagnostic is derived from hi rotation curves as opposed to line widths , owing to the contamination of galactic hi in the hi line profile of maffei 2 . as a tf relation of this form
was not available in @xmath8 , we constructed the relation from the @xmath8-band photometry and hi synthesis observations of spirals in the ursa major ( uma ) cluster conducted by ( * ? ? ?
* hereafter v01 ) . for the distance modulus of uma
, we adopted @xmath173 mag ( see [ s_zp ] ) .
of the rotational velocity parameters measured by v01 , we chose the plateau rotational velocity ( @xmath182 ) , defined as the average amplitude of the flat outer region of the rotation curve .
v01 found the least scatter in the tf relation constructed from this diagnostic , as opposed to the maximum observed rotational velocity .
we restricted our tf analysis to spirals for which @xmath182 could be measured with confidence ( referred to by v01 as the rc / df sample ) .
for consistency with the galaxy sample from which our adopted uma distance was derived , we excluded the 6 galaxies from the rc / df sample with i - band tilt corrections greater than 0.6 mag .
we also excluded ngc 3992 based on the arguments of v01 that the galaxy may be in the background of uma .
our final sample of tf calibrators contains 15 spirals .
great care was taken to correct the tf ingredients in the same manner as for maffei 2 .
galactic extinctions were computed from sfd reddenings as described in [ s_gasdust ] .
the total @xmath8 magnitude of each galaxy was corrected to its face - on value by employing yes to solve for the internal extinction from the rotational velocity and the apparent axis ratio ( see * ? ? ?
k - corrections to the uma photometry were found to be negligible in @xmath8 ( less than 0.01 mag ) .
the tf data are given in table [ tabtf ] and plotted in fig .
[ figtf ] . to construct the tf relation , we followed the treatment of @xcite and determined the slope and zero point using a bivariate linear fit , minimizing errors in both @xmath183 and @xmath9 .
as in v01 , we assumed that all galaxies have equal relative uncertainties of 5% in @xmath182 and equal photometric uncertainties of 0.05 mag in @xmath9 .
the fit yields @xmath184 with an rms scatter of 0.33 mag in @xmath9 .
the scatter is due in large part to the depth of the uma cluster ; the virial radius of 880 kpc @xcite corresponds to a range of 0.21 mag in @xmath9 , which is comparable to the rms deviation over and above the measurement uncertainties .
our fit is in close agreement with v01 , who found a slope , zero - point and rms scatter of @xmath185 , @xmath186 mag , and 0.30 mag , respectively , using the same galaxy sample but anchored to a uma distance of 18.6 mpc and employing different correction methods for the galactic and internal extinction .
we also note that eq .
[ eqtf ] is consistent with the relation found by @xcite for a sample containing both field and cluster galaxies , which demonstrates that the relation does not vary significantly with environment .
the tf parameters for maffei 2 are given in table [ tabtf ] .
using our @xmath8-band tf relation ( eq .
[ eqtf ] ) , we obtain a distance to maffei 2 of @xmath7 mpc ( @xmath187 mag ) .
the random uncertainty is computed from @xmath188 ^ 2
+ [ \delta m ( \log{2{\ensuremath{v_{\rm flat}}}}-\langle\log{2{\ensuremath{v_{\rm flat}}}}\rangle)]^2 + \sigma_{m_i}^2/n\ ] ] where @xmath85 is the standard error associated with the slope @xmath86 of the tf relation ( eq . [ eqtf ] ) , @xmath189 is the uncertainty in the estimate of @xmath190 for maffei 2 , @xmath191 is the rms deviation in @xmath9 of the @xmath192 galaxies in the v01 sample , and @xmath193 is the mean value of @xmath190 for the v01 sample , found to be 2.45 . the systematic uncertainty associated with the calibration of the distance to the uma cluster amounts to 0.13 mag .
v01 also measured total @xmath194 magnitudes and report a @xmath194band tf relation with a slightly reduced scatter of 0.26 mag in @xmath195 as opposed to 0.33 mag in @xmath9 .
maffei 2 has been observed in @xmath196 by the two micron all sky survey ( 2mass ) , and an estimate of its total magnitude , @xmath197 , is available .
however , we find that the values of @xmath197 reported by 2mass for the v01 sample differ considerably from the total @xmath194 magnitudes measured by v01 , with a mean difference of @xmath198 mag and the discrepancy reaching as high as 1.2 mag .
the large standard deviation in the residuals reflects a significant difference in the measurement techniques adopted by the two surveys .
the v01 total magnitude of a galaxy was obtained by integration of the fit to the growth curve out to infinity , while the 2mass extrapolated magnitude was computed by integration of the fit out to a finite radius judged to contain the extent of the galaxy .
hence , the v01 total magnitudes allow for a more meaningful comparison of the integrated properties .
further evidence of this is the significantly larger scatter of 0.52 mag found for the tf relation constructed from the 2mass @xmath197 magnitudes for the v01 sample . in an attempt to derive a transformation from 2mass to v01 magnitudes , we investigated the magnitude residuals as a function of various galaxy parameters .
the only dependency we observed is on the apparent @xmath194-band central disk surface brightness , @xmath199 , from @xcite , in the sense that @xmath200 is larger for the galaxies with low surface brightness ( ie . , @xmath201 ) .
this is illustrated in fig .
[ figresid ] .
unfortunately , the large scatter in @xmath200 does not allow for a reliable estimate of a @xmath194 magnitude for maffei 2 from @xmath197 .
the lack of a v01 measurement of the total @xmath194 magnitude for maffei 2 therefore precludes a reliable tf distance in this bandpass , so we adopt the @xmath8band value of 3.34 mpc .
the small inclination of ic 342 makes it difficult to determine an accurate distance from the tf relation .
however , recent observations of cepheids make possible a reliable determination from the pl relation .
@xcite observed cepheid variables in ic 342 in thuan - gunn @xmath163 and @xmath202 , deriving a distance modulus of @xmath203 mag and extinction @xmath204 mag .
however , the adopted pl relations for @xmath163 and @xmath202 were taken from @xcite , who derived them from the lmc relations for cousins @xmath205 and @xmath8 presented by @xcite , who in fact employed a different @xmath8-band pl relation from that adopted in the @xcite . also , the transformation equations employed to get the pl relations in @xmath163 and @xmath202 @xcite were founded upon johnson @xmath205 and @xmath8 ( hereafter @xmath206 and @xmath207 ) , not cousins @xmath205 and @xmath8 ( hereafter @xmath208 and @xmath113 ) .
thus , a re - examination of the cepheid distance is in order .
pl relations for the lmc in @xmath206 and @xmath207 have been derived by combining the @xcite relations for @xmath13 and @xmath113 as revised in
[ s_zp ] , the period - colour relation for @xmath209 @xcite , and transformation equations connecting @xmath209 to @xmath210 and @xmath211 to @xmath212 @xcite . in turn , revised pl relations for @xmath163 and @xmath202 have been derived from those for @xmath206 and @xmath207 using the transformation equations of @xcite . on a scale where the lmc is at a distance modulus of 18.50 mag , @xmath213 where p is the period in days . with the new pl relations ,
the mean apparent distance moduli in @xmath163 and @xmath202 for the cepheids in ic 342 become @xmath214 mag and @xmath215 mag , respectively , and the mean colour excess becomes @xmath216 mag .
the colour excess permits a solution for the optical depth of dust localized in ic 342 ( @xmath78 ) , since the optical depth of galactic dust at @xmath121 has been shown to be @xmath217 ( [ s_ext ] ) .
knowledge of the optical depth of both components then permits corrections for extinction in @xmath163 and @xmath202 .
calculations have been accomplished with yes using the reddening law of @xcite tuned to give @xmath120 for vega ( see * ? ? ?
* ) . adopting a heliocentric velocity of @xmath218 @xcite , the observed mean colour excess of the cepheids gives @xmath219 for spectral types between f8 i and g2 i , typical of luminous cepheids . solving for the galactic and localized components of the extinction in @xmath202 over the same span of spectral types (
the k - correction is negligible ) , the true distance modulus @xmath136 ranges from 27.504 to 27.492 mag .
the distance modulus can be corrected for metallicity using observations of hii regions in ic 342 made by @xcite . as was done for @xcite cepheids , measurements of @xmath220 across the disk
were converted to oxygen abundances using the calibration of @xcite .
based upon their positions , the appropriate mean metallicity to adopt for the cepheids observed in ic 342 is @xmath221 . with @xmath146 @xcite ,
the corrected distance modulus is @xmath222 mag on the scale where the lmc is at 18.50 mag .
the uncertainty is that due to random errors , specifically the quadrature sum of the errors associated with @xmath223 and with the apparent distance modulus in @xmath202 .
the result is 0.06 mag higher than the value estimated by @xcite .
note also that the revised extinction of the cepheids in @xmath13 is @xmath224 ( galactic plus localized ) , which is @xmath225 more than estimated by @xcite . correcting to the maser zero - point
, the distance modulus for ic 342 becomes @xmath226 mag .
recently , the extinction of maffei 1 was found from the colour-@xmath14 relation to be @xmath227 @xcite .
the corresponding broadband extinctions were derived from extinction coefficients computed by an older version of yes than that employed in this paper for maffei 2 and ic 342 . in the updated version of yes
, the accuracy of extinction computations has increased owing to improvements in the template seds ( see * ? ? ? * ) . for consistency with the treatment of maffei 2 and ic 342 ,
we have re - computed the extinction parameters of maffei 1 using the same version of yes employed in this paper .
the revised extinction coefficients , broadband extinctions and extinction - corrected magnitudes for maffei 1 are given in table [ tabprop ] .
@xcite used the @xmath8-band fp relation constructed from ellipticals in the coma cluster to determine the distance to maffei 1 .
the authors arrived at a distance of @xmath228 mpc , which is the weighted mean of their fp estimate combined with measurements from the @xmath229 relation of @xcite using photometry in @xmath45 and @xmath194 .
the @xmath229 distance estimates depend on the hubble constant . to place @xmath175 on the same extinction scale as that adopted in this paper
would require a re - analysis of @xmath176 @xcite zero - points ( see [ s_zp ] ) .
given that the two @xmath229 distances to maffei 1 are consistent within errors with the fp distance derived for this galaxy , here we adopt the fp distance alone in order to obtain a value that is on the same extinction and distance scale as the distances to maffei 2 and ic 342 derived in this paper .
the fp estimate depends on the @xmath8-band extinction of maffei 1 as well as the adopted distance to the coma cluster .
the modifications to the extinction of maffei 1 and the distance to the coma cluster ( see [ s_zp ] ) result in a revised distance to maffei 1 of @xmath230 mpc ( corrected to the ngc 4258 maser zero - point ) .
the @xmath8-band imaging of @xcite revealed a close grouping of galaxies apparently dominated by the spirals maffei 2 and ic 342 and the giant elliptical maffei 1 .
unfortunately , the gross uncertainties in the distances to these three galaxies , owing to their heavy obscuration , precluded a definitive statement about their relative positions and luminosities .
the results of this paper now allow for a reliable comparison of all three galaxies . the revised properties of maffei 1 , maffei 2 and ic 342 are summarized in table [ tabprop ] .
all values are now anchored to the same distance and extinction scale .
the implications of our results for these three galaxies are as follows : \1 .
the distances found in this study place the three galaxies within 0.49 mpc in depth .
their maximum spread in angle is @xmath231 , which corresponds to a metric distance of 0.53 mpc at the average distance to the galaxies of @xmath232 mpc
maffei 1 , maffei 2 and ic 342 have nearly identical face - on luminosities , spanning a range of only 0.02 mag in @xmath9 . as predicted by @xcite , maffei 2
suffers more galactic extinction than maffei 1 by nearly a full magnitude in @xmath13 .
although a magnitude fainter in @xmath9 than m31 , the three galaxies clearly dominate the ic 342/maffei group .
the next brightest member , dwingeloo 1 , has a rotational velocity comparable to m33 and can be considered to be @xmath233 mag fainter than the giants in @xmath8 .
velocities in the local group reference frame , computed using the solar apex vector of @xcite , are 311 @xmath123 , 220 @xmath123 and 255 @xmath123 for maffei 1 , maffei 2 and ic 342 , respectively .
these values significantly exceed the velocity dispersion of 61 @xmath123 for the local group @xcite , placing all three galaxies well beyond its dynamical range .
previously , the role that maffei 1 , maffei 2 and ic 342 have played in the evolution of the milky way and its neighbours has been a subject of controversy due to their uncertain distances .
early estimates of the radial velocities and distances to maffei 1 and ic 342 suggested that the galaxies could have been in the vicinity of the local group within the last 7 billion years , thereby affecting the early dynamical evolution of the local group @xcite . shortly following this analysis
, @xcite used the angular separations and radial velocities of 12 galaxies identified as members of the ic 342/maffei group to argue that the evolution of the group must have been independent of the local group .
more recently , distances to several dwarf galaxies in the ic 342/maffei group have been derived by @xcite via the tip of the red giant branch ( trgb ) , placing the dwarfs outside the early dynamical influence of the local group . however , past distance estimates to maffei 2 were too uncertain to ascertain its placement within the group or its role as a dominant member .
the modern , homogeneous properties presented in this paper for all three galaxies firmly establish the existence of a compact group of galaxies dominated by three giants at an average distance of 3.03 mpc from the milky way .
the mean velocity of the galaxies relative to the local group is 253 @xmath123 . assuming a smooth hubble flow with @xmath234 ( see [ s_zp ] ) , the hubble distance to the ic 342/maffei group is 3.17 mpc .
this is in close agreement with the mean distance of @xmath232 mpc found in this study .
it is therefore highly unlikely that the dominant members of the ic 342/maffei group have interacted with the local group since the big bang .
this conclusion lends credence to the timing model of @xcite , which is based on the simple evolutionary scenario of a two - body system involving m31 and the milky way with negligible perturbation from other galaxies .
the model provides an estimate for the age of the universe and mass of the local group , in particular offering strong evidence that most of the mass in the local group is dark .
the predictions of the model are therefore strengthened by the results of this study .
we have constructed a correlation between dust and gas in spiral galaxies as a tool for determining the galactic extinction of a heavily obscured spiral .
the correlation can be used to estimate the extragalactic extinction of constituent hii regions from measurements of the column density of hydrogen in the heavily obscured galaxy .
the spiral s galactic extinction can then be determined by taking the average difference between the extragalactic extinction and the total extinction of each hii region found from observations of the balmer decrement .
we have used the dust gas correlation to measure the galactic extinction of the heavily obscured spirals maffei 2 and ic 342 , for which we find values of @xmath235 and @xmath236 , respectively , where @xmath80 is the optical depth of galactic dust at 1 @xmath0 m .
the galactic extinction of maffei 2 is the most accurate measurement to date , while the result for ic 342 is consistent with the optical depth of @xmath4 obtained from the reddening maps of @xcite .
we therefore adopt the weighted mean of the two values for ic 342 , which yields @xmath110 .
thus , @xmath2 mag for maffei 2 and @xmath6 mag for ic 342 .
the reddening coefficients used in this analysis were computed by the york extinction solver ( yes ) , which accounts for the colour - dependent shifts in the effective wavelengths of broadband filters . to determine the distance to maffei 2 ,
we have formulated the tully - fisher ( tf ) relation in @xmath8 using rotation curves instead of hi line widths as the luminosity diagnostic , as line widths for nearby galaxies like maffei 2 can be unreliable due to contamination by galactic hi . the tf distance
to maffei 2 is found to be @xmath7 mpc , where the zero - point is set by the maser distance to ngc 4258 . using the extinction treatment and distance scale adopted in this paper , we re - derive the cepheid distance to ic 342 @xcite and the fundamental plane distance to maffei 1 @xcite , obtaining @xmath237 mpc and @xmath230 mpc , respectively .
the revised distances and magnitudes reveal that the ic 342/maffei group is dominated by three giant galaxies with nearly identical luminosities .
further , the average distance to the three galaxies is in excellent agreement with expectations for the hubble flow , making it highly unlikely that these galaxies interacted with the local group since the big bang .
rlf and mlm gratefully acknowledge the continuing support of the natural sciences and engineering research council of canada and the ontario graduate scholarship program .
mlm is also grateful to s. sakai for very helpful communications and to r. ross for performing initial data reductions .
hl and mgr thank noao for the use of their facilities .
mgr gratefully acknowledges the financial support of grants from conacyt ( conacyt 43121 ) and unam ( dgapa in 112103 , in 108406 - 2 and in 108506 - 2 ) .
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 .
some data were accessed as guest user , canadian astronomy data centre , which is operated by the dominion astrophysical observatory for the national research council of canada s herzberg institute of astrophysics .
in addition , 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 .
cccccc @xmath238 & & & & @xmath239 & + @xmath19 & @xmath240 & @xmath241 & & @xmath242 & @xmath243 + @xmath51 } \lambda4959 $ ] & & & & @xmath244 & @xmath245 + @xmath51 } \lambda5007 $ ] & & & & @xmath246 & @xmath247 + hei @xmath248 & & & & @xmath249 & + @xmath250 \lambda6300 $ ] & & & & @xmath251 & + @xmath252 \lambda6548 $ ] & @xmath253 & @xmath254 & & & + @xmath18 & @xmath255 & @xmath256 & & @xmath257 & @xmath258 + @xmath252 \lambda6583 $ ] & @xmath259 & @xmath260 & & @xmath261 & @xmath262 + hei @xmath263 & & & & @xmath264 & + @xmath265 \lambda6716 $ ] & @xmath266 & @xmath267 & & @xmath268 & @xmath269 + @xmath265 \lambda6731 $ ] & @xmath270 & @xmath271 & & @xmath272 & @xmath273 + @xmath274 \lambda7136 $ ] & & & & @xmath275 & + @xmath276 & @xmath277 & @xmath278 & & @xmath279 & @xmath280 + @xmath281 ( ) & @xmath282 & @xmath283 & & @xmath284 & @xmath285 + @xmath286 $ ] & 168 & 6.6 & @xmath287 & -2.98 & 4,572 & 3.07 & @xmath288 & @xmath289 & 2.24 & @xmath290 & @xmath291 & @xmath292 & + @xmath293 $ ] & 187 & 16.5 & @xmath294 & -3.00 & 4,748 & 3.06 & @xmath295 & @xmath296 & 2.33 & @xmath297 & @xmath298 & @xmath299 & + @xmath300 $ ] & 564 & 197.0 & @xmath301 & -3.76 & 10,330 & 2.86 & @xmath302 & @xmath303 & 7.61 & @xmath304 & @xmath305 & @xmath306 & mr85 + @xmath307 $ ] & 303 & 64.3 & @xmath308 & -3.01 & 4,835 & 3.05 & @xmath309 & @xmath310 & 2.37 & @xmath311 & @xmath312 & @xmath313 & mr85 + @xmath314 $ ] & 916 & 53.9 & @xmath315 & -3.76 & 10,390 & 2.86 & @xmath316 & @xmath317 & 7.71 & @xmath304 & @xmath318 & @xmath319 & mr85 + @xmath320 $ ] & 160 & & @xmath321 & -3.24 & 6,527 & 2.97 & @xmath322 & @xmath323 & 3.40 & @xmath324 & @xmath325 & @xmath326 & bk82 + @xmath327 $ ] & 357 & 442.4 & @xmath328 & -3.33 & 7,147 & 2.94 & @xmath329 & @xmath330 & 3.87 & @xmath331 & @xmath332 & @xmath333 & + @xmath334 $ ] & 321 & 112.0 & @xmath335 & -3.31 & 7,032 & 2.95 & @xmath336 & @xmath337 & 3.78 & @xmath338 & @xmath339 & @xmath340 & + lcccccccccll ngc2903 & 4 & 09:32:09.7 & + 21:30:02 & 560 & 60 & 29 & 58.2 & 0.031 & 0.035 & wv86 & yx95 + ngc4303 & 4 & 12:21:54.7 & + 04:28:20 & 1,568 & 25 & 7 & 38.4 & 0.022 & 0.025 & w88 & ky88 + ngc4321 & 4 & 12:22:55.2 & + 15:49:23 & 1,574 & 30 & 153 & 63.6 & 0.026 & 0.030 & w88 & ky88 + ngc4501 & 3 & 12:31:59.6 & + 14:25:17 & 2,276 & 58 & 140 & 49.7 & 0.038 & 0.043 & w88 & yx95 + ngc5055 & 4 & 13:15:49.3 & + 42:02:06 & 500 & 55 & 99 & 102.6 & 0.018 & 0.020 & wv86 & yx95 + ngc5194 & 4 & 13:29:53.3 & + 47:11:48 & 464 & 20 & 170 & 111.6 & 0.035 & 0.040 & ta91 & yx95 + + maffei2 & 4 & 02:41:54.6 & + 59:36:11 & -23 & 67 & 206 & 118.9 & & & ht96 & mw04 + + ngc598 & 6 & 01:33:50.9 & + 30:39:37 & -181 & 57 & 22 & 494.4 & 0.042 & 0.048 & n80a & yx95 + ngc2403 & 6 & 07:36:54.5 & + 65:35:58 & 125 & 60 & 125 & 163.2 & 0.040 & 0.045 & wv86 & yx95 + ngc4654 & 6 & 12:43:56.6 & + 13:07:33 & 1,030 & 52 & 128 & 39.7 & 0.026 & 0.030 & w88 & yx95 + ngc5457 & 6 & 14:03:12.5 & + 54:20:55 & 243 & 18 & 39 & 126.6 & 0.009 & 0.010 & bg81 & ks91 + + ic342 & 6 & 03:46:49.7 & + 68:05:45 & 25 & 25 & 39 & 209.3 & & & n80b & yx95 @xmath341 $ ] & 53 & 30.8 & 3.99 & -2.88 & 3,877 & 3.13 & 0.22 & 14.26 & 1.94 & 11.92 & 38.10 & 0.19 & mr85 + @xmath342 $ ] & 89 & 27.7 & 5.61 & -2.99 & 4,641 & 3.07 & 0.55 & 15.90 & 2.28 & 10.42 & 36.74 & 0.52 & mr85 + @xmath343 $ ] & 93 & 27.6 & 6.56 & -2.92 & 4,134 & 3.11 & 0.68 & 15.78 & 2.04 & 9.24 & 34.26 & 0.65 & mr85 + @xmath344 $ ] & 152 & 73.0 & 4.36 & -3.21 & 6,272 & 2.98 & 0.35 & 11.14 & 3.22 & 13.94 & 39.02 & 0.31 & mr85 + @xmath345 $ ] & 157 & 31.1 & 3.12 & -3.01 & 4,820 & 3.05 & 0.02 & 10.51 & 2.36 & 10.24 & 30.99 & -0.02 & mr85 + @xmath346 $ ] & 161 & 41.8 & 4.55 & -3.21 & 6,280 & 2.98 & 0.39 & 10.07 & 3.22 & 13.95 & 37.97 & 0.35 & mr85 + @xmath347 $ ] & 45 & 33.0 & 3.84 & -2.93 & 4,255 & 3.10 & 0.20 & 11.67 & 2.10 & 19.97 & 51.61 & 0.16 & ss91 + @xmath348 $ ] & 46 & 36.0 & 3.40 & -3.09 & 5,393 & 3.02 & 0.11 & 11.62 & 2.67 & 24.63 & 60.88 & 0.07 & ss91 + @xmath349 $ ] & 51 & 160.0 & 5.18 & -2.98 & 4,569 & 3.07 & 0.48 & 11.36 & 2.24 & 17.49 & 46.34 & 0.44 & ss91 + @xmath350 $ ] & 51 & 19.0 & 3.65 & -2.93 & 4,185 & 3.10 & 0.15 & 5.76 & 2.07 & 16.09 & 37.94 & 0.11 & ss91 + @xmath351 $ ] & 54 & 64.0 & 3.68 & -3.09 & 5,422 & 3.02 & 0.18 & 6.55 & 2.69 & 18.78 & 44.11 & 0.15 & ss91 + @xmath352 $ ] & 71 & 21.0 & 3.34 & -2.94 & 4,295 & 3.09 & 0.07 & 10.07 & 2.12 & 8.25 & 26.57 & 0.03 & ss91 + @xmath353 $ ] & 73 & 46.0 & 4.12 & -3.01 & 4,824 & 3.05 & 0.27 & 10.56 & 2.37 & 8.42 & 27.40 & 0.24 & ss91 + @xmath354 $ ] & 107 & 207.0 & 3.27 & -3.31 & 7,002 & 2.95 & 0.09 & 9.17 & 3.76 & 4.11 & 17.39 & 0.06 & ss91 + @xmath355 $ ] & 144 & 117.0 & 3.12 & -3.39 & 7,647 & 2.92 & 0.06 & 7.09 & 4.31 & 1.32 & 9.73 & 0.02 & ss91 + @xmath356 $ ] & 57 & 21.2 & 5.87 & -2.93 & 4,196 & 3.10 & 0.58 & 5.87 & 2.07 & 12.81 & 31.49 & 0.55 & mr85 + @xmath357 $ ] & 68 & 39.0 & 4.59 & -2.97 & 4,481 & 3.08 & 0.36 & 6.13 & 2.20 & 10.63 & 27.39 & 0.33 & ss91 + @xmath358 $ ] & 81 & 33.4 & 5.55 & -2.92 & 4,135 & 3.11 & 0.53 & 5.71 & 2.04 & 7.72 & 21.15 & 0.49 & mr85 + @xmath359 $ ] & 108 & 62.0 & 4.06 & -2.98 & 4,625 & 3.07 & 0.25 & 7.06 & 2.27 & 5.08 & 17.22 & 0.22 & ss91 + @xmath360 $ ] & 127 & 43.5 & 4.67 & -3.01 & 4,795 & 3.06 & 0.39 & 6.95 & 2.35 & 3.68 & 14.31 & 0.35 & mr85 + @xmath361 $ ] & 153 & 137.0 & 3.77 & -3.11 & 5,516 & 3.02 & 0.20 & 5.80 & 2.74 & 2.66 & 11.12 & 0.17 & ss91 + @xmath362 $ ] & 152 & 33.0 & 4.70 & -3.23 & 6,404 & 2.97 & 0.42 & 5.87 & 3.31 & 3.28 & 12.43 & 0.38 & ss91 + @xmath363 $ ] & 158 & 68.0 & 4.22 & -3.08 & 5,346 & 3.02 & 0.30 & 9.29 & 2.64 & 2.34 & 13.97 & 0.27 & ss91 + @xmath364 $ ] & 159 & 51.6 & 4.50 & -3.04 & 5,003 & 3.04 & 0.36 & 9.00 & 2.46 & 2.14 & 13.28 & 0.32 & mr85 + @xmath365 $ ] & 41 & 67.0 & 4.79 & -2.89 & 3,911 & 3.13 & 0.39 & 9.83 & 1.95 & 15.46 & 40.75 & 0.35 & sk96 + @xmath366 $ ] & 72 & 39.0 & 5.16 & -2.98 & 4,616 & 3.07 & 0.47 & 6.81 & 2.26 & 8.46 & 23.73 & 0.44 & sk96 + @xmath367 $ ] & 74 & 12.0 & 3.31 & -2.88 & 3,880 & 3.13 & 0.05 & 6.92 & 1.94 & 6.90 & 20.72 & 0.02 & sk96 + @xmath368 $ ] & 124 & 70.0 & 3.31 & -3.14 & 5,776 & 3.00 & 0.09 & 5.55 & 2.90 & 3.14 & 11.83 & 0.05 & sk96 + @xmath369 $ ] & 116 & 26.0 & 3.84 & -3.12 & 5,589 & 3.01 & 0.22 & 6.28 & 2.78 & 3.68 & 13.64 & 0.19 & sk96 + @xmath370 $ ] & 96 & 16.3 & 5.06 & -2.92 & 4,164 & 3.11 & 0.45 & 15.53 & 2.06 & 9.95 & 35.43 & 0.41 & mr85 + @xmath371 $ ] & 169 & 18.7 & 5.20 & -2.99 & 4,650 & 3.07 & 0.48 & 15.78 & 2.28 & 11.03 & 37.84 & 0.45 & mr85 + @xmath372 $ ] & 152 & 52.2 & 6.13 & -3.10 & 5,442 & 3.02 & 0.65 & 16.16 & 2.70 & 13.05 & 42.26 & 0.61 & mr85 + @xmath373 $ ] & 281 & 118.6 & 3.36 & -3.24 & 6,524 & 2.97 & 0.11 & 8.07 & 3.39 & 16.42 & 40.91 & 0.08 & mr85 + @xmath374 $ ] & 61 & 50.4 & 4.79 & -2.89 & 3,913 & 3.13 & 0.39 & 5.07 & 1.95 & 37.59 & 80.25 & 0.35 & mr85 + @xmath375 $ ] & 125 & 51.9 & 3.84 & -2.90 & 3,996 & 3.12 & 0.19 & 4.87 & 1.99 & 9.32 & 23.51 & 0.15 & mr85 + @xmath376 $ ] & 1,816 & 4.3 & 2.67 & -3.20 & 6,184 & 2.98 & -0.10 & 9.42 & 3.16 & 0.00 & 9.42 & -0.15 & mr85 + @xmath377 $ ] & 846 & 20.5 & 3.95 & -3.24 & 6,518 & 2.97 & 0.26 & 18.01 & 3.39 & 0.00 & 18.01 & 0.21 & mr85 + @xmath378 $ ] & 435 & 203.0 & 4.23 & -3.28 & 6,773 & 2.96 & 0.33 & 19.03 & 3.58 & 0.00 & 19.03 & 0.28 & mr85 + @xmath379 $ ] & 86 & 28.7 & 4.05 & -3.14 & 5,739 & 3.01 & 0.27 & 14.93 & 2.87 & 7.56 & 30.05 & 0.22 & mr85 + @xmath380 $ ] & 268 & & 4.84 & -3.27 & 6,752 & 2.96 & 0.45 & 15.61 & 3.56 & 0.00 & 15.61 & 0.40 & ka81 + @xmath381 $ ] & 368 & 117.0 & 3.49 & -3.17 & 5,998 & 2.99 & 0.14 & 16.09 & 3.04 & 0.00 & 16.09 & 0.09 & vp88 + @xmath382 $ ] & 444 & & 4.76 & -3.37 & 7,488 & 2.93 & 0.44 & 19.48 & 4.16 & 0.00 & 19.48 & 0.40 & ka81 + @xmath383 $ ] & 615 & 107.0 & 2.98 & -3.21 & 6,292 & 2.98 & 0.00 & 18.26 & 3.23 & 0.00 & 18.26 & -0.05 & vp88 + @xmath384 $ ] & 782 & & 4.31 & -3.55 & 8,814 & 2.89 & 0.36 & 15.19 & 5.52 & 0.00 & 15.19 & 0.32 & ka81 + @xmath385 $ ] & 871 & 86.0 & 3.78 & -3.33 & 7,190 & 2.94 & 0.23 & 20.21 & 3.91 & 0.00 & 20.21 & 0.18 & vp88 + @xmath386 $ ] & 1,129 & 20.0 & 2.84 & -3.42 & 7,830 & 2.92 & -0.02 & 19.87 & 4.48 & 0.00 & 19.87 & -0.07 & vp88 + @xmath387 $ ] & 1,352 & & 3.07 & -3.39 & 7,642 & 2.92 & 0.04 & 17.83 & 4.30 & 0.00 & 17.83 & 0.00 & ka81 + @xmath388 $ ] & 1,473 & & 4.48 & -3.63 & 9,425 & 2.88 & 0.40 & 19.87 & 6.28 & 0.00 & 19.87 & 0.36 & ka81 + @xmath389 $ ] & 1,476 & 97.0 & 3.20 & -3.46 & 8,163 & 2.91 & 0.09 & 19.93 & 4.80 & 0.00 & 19.93 & 0.04 & vp88 + @xmath390 $ ] & 1,675 & & 4.48 & -3.61 & 9,245 & 2.88 & 0.40 & 15.22 & 6.04 & 0.00 & 15.22 & 0.35 & ka81 + @xmath391 $ ] & 592 & 157.0 & 3.49 & -3.55 & 8,797 & 2.89 & 0.17 & 11.05 & 5.50 & 0.02 & 11.09 & 0.13 & mr85 + @xmath392 $ ] & 393 & 155.9 & 3.40 & -3.43 & 7,906 & 2.92 & 0.14 & 15.67 & 4.55 & 0.09 & 15.85 & 0.09 & mr85 + @xmath393 $ ] & 65 & 83.1 & 3.76 & -3.27 & 6,747 & 2.96 & 0.22 & 20.30 & 3.56 & 1.55 & 23.40 & 0.17 & mr85 + @xmath394 $ ] & 165 & 180.5 & 4.26 & -3.40 & 7,678 & 2.92 & 0.34 & 21.17 & 4.33 & 0.56 & 22.29 & 0.30 & mr85 + @xmath395 $ ] & 80 & 72.0 & 4.06 & -3.16 & 5,938 & 3.00 & 0.28 & 20.24 & 3.00 & 0.96 & 22.16 & 0.23 & mr85 + @xmath396 $ ] & 416 & 213.0 & 3.60 & -3.59 & 9,087 & 2.89 & 0.20 & 15.07 & 5.84 & 0.09 & 15.25 & 0.16 & mr85 + @xmath397 $ ] & 390 & 61.9 & 3.40 & -3.30 & 6,945 & 2.95 & 0.13 & 8.63 & 3.71 & 1.08 & 10.79 & 0.12 & mr85 + @xmath398 $ ] & 308 & 169.3 & 4.52 & -3.40 & 7,686 & 2.92 & 0.40 & 9.96 & 4.34 & 2.55 & 15.06 & 0.39 & mr85 + @xmath399 $ ] & 291 & 68.2 & 3.42 & -3.32 & 7,088 & 2.95 & 0.14 & 10.17 & 3.82 & 2.60 & 15.37 & 0.13 & mr85 + @xmath400 $ ] & 269 & 234.3 & 3.12 & -3.28 & 6,807 & 2.96 & 0.05 & 10.08 & 3.60 & 2.93 & 15.94 & 0.04 & mr85 + @xmath401 $ ] & 287 & 249.4 & 4.51 & -3.56 & 8,879 & 2.89 & 0.40 & 10.22 & 5.59 & 3.93 & 18.08 & 0.39 & mr85 + @xmath402 $ ] & 701 & 213.6 & 3.35 & -3.67 & 9,658 & 2.87 & 0.14 & 5.91 & 6.60 & 0.13 & 6.17 & 0.13 & mr85 + @xmath403 $ ] & 172 & 162.2 & 4.14 & -3.29 & 6,918 & 2.95 & 0.31 & 8.33 & 3.69 & 6.91 & 22.15 & 0.30 & rp82 + @xmath404 $ ] & 113 & & 4.84 & -3.25 & 6,611 & 2.97 & 0.45 & 7.34 & 3.46 & 10.74 & 28.82 & 0.44 & rp82 + @xmath405 $ ] & 212 & & 4.41 & -3.32 & 7,124 & 2.94 & 0.37 & 9.68 & 3.85 & 5.14 & 19.96 & 0.36 & rp82 + @xmath406 $ ] & 704 & 173.8 & 3.84 & -3.67 & 9,676 & 2.87 & 0.26 & 5.86 & 6.62 & 0.13 & 6.12 & 0.25 & rp82 + @xmath407 $ ] & 288 & 195.0 & 5.03 & -3.38 & 7,517 & 2.93 & 0.49 & 10.21 & 4.19 & 2.92 & 16.05 & 0.48 & rp82 + @xmath408 $ ] & 78 & 83.0 & 4.14 & -3.28 & 6,800 & 2.96 & 0.31 & 20.00 & 3.60 & 4.09 & 28.18 & 0.28 & sk96 + @xmath409 $ ] & 106 & 100.0 & 3.08 & -3.31 & 7,054 & 2.95 & 0.04 & 12.68 & 3.80 & 1.47 & 15.62 & 0.01 & sk96 + @xmath410 $ ] & 75 & 127.0 & 3.57 & -3.36 & 7,364 & 2.93 & 0.18 & 19.72 & 4.06 & 5.12 & 29.96 & 0.15 & sk96 + @xmath411 $ ] & 55 & 59.0 & 3.57 & -3.11 & 5,554 & 3.01 & 0.15 & 17.71 & 2.76 & 7.27 & 32.25 & 0.12 & sk96 + @xmath412 $ ] & 37 & 280.0 & 3.54 & -3.07 & 5,222 & 3.03 & 0.14 & 14.19 & 2.58 & 11.47 & 37.13 & 0.11 & sk96 + @xmath413 $ ] & 23 & 15.0 & 4.14 & -3.23 & 6,408 & 2.97 & 0.30 & 13.08 & 3.31 & 19.14 & 51.36 & 0.27 & sk96 + llccccccccc ngc3726 & sc & 53 & 0.62 & 865.6 & @xmath414 & 9.51 & 0.019 & 0.03 & 0.18 & -21.98 + ngc3729 & sab & 49 & 0.68 & 1059.8 & @xmath415 & 10.30 & 0.012 & 0.02 & 0.13 & -21.13 + ngc3917 & scd & 79 & 0.24 & 964.6 & @xmath416 & 10.85 & 0.025 & 0.04 & 0.60 & -21.06 + ngc3949 & sbc & 55 & 0.62 & 800.2 & @xmath417 & 10.28 & 0.024 & 0.04 & 0.18 & -21.22 + ngc3953 & sbc & 62 & 0.5 & 1052.3 & @xmath418 & 9.02 & 0.034 & 0.05 & 0.36 & -22.67 + ngc4085 & sc & 82 & 0.24 & 745.7 & @xmath419 & 11.28 & 0.020 & 0.03 & 0.59 & -20.62 + ngc4088 & sbc & 69 & 0.37 & 756.7 & @xmath420 & 9.37 & 0.023 & 0.04 & 0.46 & -22.41 + ngc4100 & sbc & 73 & 0.29 & 1074.4 & @xmath421 & 10.00 & 0.026 & 0.04 & 0.58 & -21.90 + ngc4102 & sab & 56 & 0.56 & 846.3 & @xmath422 & 9.93 & 0.023 & 0.04 & 0.25 & -21.63 + ngc4138 & sa & 53 & 0.63 & 893.8 & @xmath423 & 10.09 & 0.016 & 0.03 & 0.16 & -21.37 + ugc6399 & sm & 75 & 0.28 & 791.5 & @xmath424 & 12.88 & 0.018 & 0.03 & 0.00 & -18.43 + ugc6446 & sd & 51 & 0.62 & 644.3 & @xmath425 & 12.58 & 0.018 & 0.03 & 0.05 & -18.78 + ugc6667 & scd & 89 & 0.12 & 973.2 & @xmath426 & 12.63 & 0.019 & 0.03 & 0.56 & -19.24 + ugc6917 & sd & 56 & 0.54 & 910.7 & @xmath427 & 11.74 & 0.031 & 0.05 & 0.14 & -19.73 + ugc6983 & scd & 49 & 0.66 & 1081.9 & @xmath428 & 11.91 & 0.031 & 0.05 & 0.09 & -19.51 + + maffei2 & sbc & 67 & 0.421 & -23 & @xmath429 & 9.29 & 2.017 & 3.15 & 0.37 & -21.87 + lccccc apparent magnitudes & @xmath430 & mag & @xmath431 & @xmath432 & @xmath433 + & @xmath434 & mag & @xmath435 & @xmath436 & @xmath437 + & @xmath438 & mag & @xmath439 & @xmath440 & @xmath441 + + galactic optical depth at @xmath442 & @xmath443 & & @xmath444 & @xmath445 & @xmath446 + + extinction normalized to @xmath443 & @xmath447 & & @xmath448 & @xmath449 & @xmath450 + computed by yes & @xmath451 & & @xmath452 & @xmath453 & @xmath454 + & @xmath455 & & @xmath456 & @xmath457 & @xmath458 + + galactic extinction & @xmath459 & mag & @xmath460 & @xmath461 & @xmath462 + & @xmath17 & mag & @xmath463 & @xmath464 & @xmath465 + & @xmath466 & mag & @xmath467 & @xmath468 & @xmath469 + + internal extinction & @xmath470 & mag & & 0.646 & 0.099 + ( excess over face - on ) & @xmath471 & mag & & 0.548 & 0.081 + & @xmath472 & mag & & 0.385 & 0.053 + + apparent face - on magnitudes & @xmath473 & mag & @xmath474 & @xmath475 & @xmath476 + corrected for galactic extinction & @xmath477 & mag & @xmath478 & @xmath479 & @xmath480 + & @xmath481 & mag & @xmath482 & @xmath483 & @xmath484 + + distance modulus & @xmath0 & mag & @xmath485 & @xmath486 & @xmath487 + + distance & @xmath488 & mpc & @xmath489 & @xmath490 & @xmath491 + + absolute face - on magnitudes & @xmath492 & mag & @xmath493 & @xmath494 & @xmath495 + corrected for galactic extinction & @xmath496 & mag & @xmath497 & @xmath498 & @xmath499 + & @xmath9 & mag & @xmath500 & @xmath501 & @xmath502 + | we have obtained spectra of hii regions in the heavily obscured spiral galaxy maffei 2 .
the observations have allowed for a determination of the galactic extinction of this galaxy using a correlation between extinction and hydrogen column density observed among spirals .
the technique reveals that the optical depth of galactic dust at 1 @xmath0 m obscuring maffei 2 is @xmath1 , which implies that @xmath2 mag , significantly higher than observed for the giant elliptical maffei 1 despite its similar latitude . for comparison
, we apply the same technique to ic 342 , a neighbouring spiral to maffei 2 but with more moderate obscuration by galactic dust , owing to its higher galactic latitude . for this galaxy ,
we obtain @xmath3 , which agrees within errors with the value of @xmath4 derived from the reddening estimate of @xcite .
we therefore adopt the weighted mean of @xmath5 for the extinction of ic 342 , which implies that @xmath6 mag .
a new distance estimate for maffei 2 of @xmath7 mpc is obtained from a self - consistent tully - fisher relation in @xmath8 adjusted to the ngc 4258 maser zero - point . with our new measurement of @xmath9 , maffei 2
joins maffei 1 and ic 342 as one of three giant members of the nearby ic 342/maffei group of galaxies .
we present the revised properties of all three galaxies based on the most accurate extinction and distance estimates to date , accounting for shifts in the effective wavelengths of broadband filters as this effect can be significant for highly reddened galaxies .
the revised distances are consistent with what would be suspected for the hubble flow , making it highly unlikely that the galaxies interacted with the local group since the big bang . |
questions of how to compare hadronic observables to the apparent underlying fundamental theory of qcd lie at the heart of understanding the nature of strong interactions .
thirty years after its inception , qcd in @xmath7 spacetime dimensions still stubbornly refuses to admit a global solution .
the asymptotic freedom property of the theory permits the perturbative calculation of ( euclidean ) green functions involving large values of momentum transfer or energy release in terms of quarks and gluons , the fundamental objects of qcd .
but at lower scales one enters the nonperturbative regime , which not only invalidates ( or at least complicates ) the standard perturbative methods of field theory developed in qed , but also leads to a dramatic change in the physical spectrum of the theory . instead of quarks and gluons ,
only colorless hadrons are produced as asymptotic states in any process , even at arbitrarily large energy .
many nontrivial theoretical techniques respecting qcd first principles have been developed to study nonperturbative features of the theory . yet despite numerous advances , no one has been able to compute the masses , wavefunctions , or transition amplitudes of hadrons in terms of quark masses and couplings directly from the qcd lagrangian . moreover , many existing theoretical tools are expressed through various expansions in certain small parameters ; the actual range of each parameter where the expansions are applicable is often not well known . in such a situation , it is clearly advantageous to build a soluble toy field theory that incorporates as many features of the qcd lagrangian as possible .
such a theory does indeed exist , the famous t hooft model @xcite , which is defined by the yang - mills lagrangian in @xmath8 spacetime dimensions in the limit of a large number of colors @xmath9 .
as was shown in the original paper , the quark - antiquark sector of the theory admits an infinite tower of confined , color - singlet solutions that can be obtained , in principle , to an arbitrary degree of numerical accuracy .
the reason for this solubility lies precisely in the defining features of the model .
large @xmath9 eliminates all feynman diagrams with internal @xmath10 loops and nonplanar gluons . on the other hand ,
@xmath8 allows gluon self - couplings to be eliminated by gauging away one component of the gauge potential @xmath11 . since only two components are initially present
, the commutator term @xmath12 $ ] in the covariant derivative , which gives gluon self - coupling , vanishes identically in such gauges .
then the only remaining feynman diagrams to be summed for the quark - antiquark green function are `` rainbow '' and `` ladder '' diagrams , whose schwinger - dyson equations can be solved , giving rise to an integral expression called the t hooft equation ( discussed in sec .
[ thm ] ) .
the t hooft model provides an excellent laboratory for testing various approaches to strong interaction physics .
after all , the t hooft equation provides a means to compute hadronic masses , wavefunctions , and transition amplitudes in terms of the underlying partonic degrees of freedom . in this work
we are specifically interested in questions of local quark - hadron duality in the inclusive decays of heavy quarks . the notion of duality in general terms was first introduced in the early days of qcd in ref .
@xcite but not pursued for quite some time .
a more detailed consideration was given a few years ago by shifman @xcite and later reiterated in a number of papers ( see , e.g. , refs .
@xcite ) , with applications relevant to minkowskian observables amenable to study via an operator product expansion ( ope ) .
this allows the formulation of the concept of local duality in a more quantitative way , including nontrivial nonperturbative effects ; we refer the reader to these recent publications for the theoretical aspects . here the question of duality is studied concretely by comparing the weak decay width of a meson containing a heavy quark computed in two ways . in terms of partonic degrees of freedom
, one has an ope depending upon the free quark diagram ( with perturbative corrections ) and a number of nonperturbative matrix elements suppressed by powers of the heavy quark mass . in terms of the hadronic degrees of freedom ,
one simply computes the weak decay amplitude for each allowed exclusive channel , and adds them up one by one .
this comparison is especially instructive since one may consider the behavior of solution as the mass @xmath13 of the heavy decaying quark is varied .
such a problem was first considered in ref .
@xcite , where the main elements in numerical computations of exclusive decay rates were annunciated .
the hadronic result was compared to the born - level free - partonic diagram as a function of @xmath13 . in terms of the ope ,
the latter is the tree - level piece of the wilson coefficient corresponding to the unit operator .
the numerical agreement was seen to be remarkable , in that the onset of the asymptotic agreement was clearly visible already for relatively small values of @xmath13 . the intrinsically limited numerical accuracy for sufficiently heavy quarks , however , prohibited drawing a definite conclusion about the size of nonperturbative corrections for asymptotically large @xmath13 .
additional numerical studies @xcite considered similar questions for weak decay topologies other than the simple spectator tree diagram , in particular weak annihilation ( wa ) .
the validity of the ope was addressed analytically in refs .
@xcite , which considered on one hand the nature of the ope for heavy quark decays , and on the other an explicit @xmath3 expansion of the decay amplitudes , which allows an analytical summation of the individual decay rates in the asymptotic regime .
the agreement of the two approaches through relative order @xmath14 was obtained by means of a number of sum rules derived directly from the t hooft equation , the archetype of which first appeared in ref .
@xcite . while adequate to illustrate the theoretical validity of the ope for the inclusive decay widths of heavy flavors , the analytic methods per se can not help in answering the practical question relevant to phenomenology of beauty and charm quarks : namely , how accurately do the ope - improved parton computations describe the true weak decay width of a heavy flavor meson with finite mass , only a few times larger than the typical strong interaction scale ?
a purely analytic expansion can hardly be used for this purpose , since it is a priori unknown how small an expansion parameter must be for the expansion to start yielding a reasonable approximation , not to mention achieving the necessary precision . to obtain insights into the size of deviations between the actual decay widths and the expressions obtained from the ope for quarks in the intermediate mass range , one must employ real numerical computations . in this paper
we focus on semileptonic decays of heavy quarks . in the contexts of both real qcd and the t hooft model , they are technically simpler than nonleptonic decays . moreover , the magnitude of local duality violation is phenomenologically most important in semileptonic decays when one extracts @xmath1 and @xmath2 .
we use the techniques developed in ref .
@xcite to evaluate the required decay rates , and confront the total decay width with the expansion in terms of a power series in @xmath3 of ref .
moreover , by making use of a number of relations derived in the large-@xmath13 limit of the t hooft equation @xcite , members of the set of nonperturbative matrix elements involved can be related to each other , providing an economical description of the nonperturbative physics .
these are the tools that allow us to study the onset of quark - hadron duality .
as explained in appendix [ newmult ] , we use a scheme based on the modified multhopp method , by which the t hooft equation is converted into an infinite - dimension eigenvector system that for practical reasons must be truncated at some number @xmath15 of eigenvector modes . the asymptotic convergence of this approach has not been rigorously studied , although it apparently must yield unlimited accuracy when the number of the multhopp modes @xmath15 goes to infinity . yet
the rate of convergence at large @xmath15 is not well known .
additionally , large quark masses turn out to require one to use a larger @xmath15 for sufficient numerical accuracy , as discussed in sec .
it therefore seems mandatory to make an independent cross - check of the numerical accuracy .
we investigate this problem by comparing the numerical values of a number of static properties of heavy mesons at different values of @xmath13 , with the results of their @xmath3 expansions obtained analytically from the t hooft equation ; this is the topic of sec .
[ check ] .
we find that our solutions have sufficient numerical accuracy for masses @xmath13 corresponding to physical values ( in the sense explained in sec . [ check ] ) as large as @xmath16gev . the duality of the inclusive widths of heavy - flavor hadrons to the parton - level widths , including the power corrections from the ope , emerges through a set of sum rules that equate sums of weighted transition probabilities to possible final states and expectation values of the local heavy quark operators . since our main interest lies in @xmath17 transitions , which carry in practice a limited energy release , the most relevant are the so - called small velocity ( sv ) sum rules , which we study here in the heavy quark limit .
the behavior of these sum rules not only shapes the semileptonic @xmath17 decays in actual qcd , but is also important for the determination of the basic parameters of the heavy quark expansion .
an additional advantage of the heavy quark limit for our investigation is that we are able to compute the sv amplitudes semi - analytically , using the exact relations @xcite derived from the t hooft equations and relying for input only on a few static parameters , which can be computed with a high precision .
a discussion of these relations appears in sec . [ dual ] .
we find that the sv sum rules in the t hooft model are saturated to an unexpectedly high degree by the first excitation above the ground state ( which we henceforth call the `` @xmath18-wave '' excitation , despite the fact that in @xmath8 only radial excitations occur ) .
its contributions to even the darwin ( @xmath19 ) and kinetic ( @xmath20 ) expectation values constitute over @xmath21 and @xmath22 of the totals , respectively , while it saturates the `` optical '' sum rule for @xmath23 to a @xmath24 accuracy .
this appears to be an intriguing dynamical feature of the model .
a similar high - saturation effect has been observed in a quark flux - tube model @xcite , for the contribution from the `` valence '' quarkonium states .
we study the size of violations of local duality in the semileptonic decays @xmath25 assuming vectorlike weak currents and massless leptons .
these assumptions are important for comparison with qcd far beyond the obvious parallel of closely resembling the actual world : the strength of the resonance - related duality violation crucially depends on the threshold behavior in the decay probabilities , which is completely different in two and four dimensions .
the two - body phase space , while @xmath26 in @xmath7 , is @xmath27 , that is , infinite at threshold , in @xmath8 . on the other hand , the situation is special for massless leptons : their invariant mass is always zero if they are produced by a vectorlike source , and the weak vertex is then proportional to the momentum . as a result ,
in this case the threshold behavior of the decay rate becomes @xmath26 much in the same way as in real qcd .
this is a crucial detail if one tries to draw practical lessons from the t hooft model .
the need for a vectorlike coupling in @xmath8 is even more stark for the parton - level calculation .
there one finds that the integrated three - body phase space actually diverges for massless leptons , and only the behavior of the weak decay amplitude renders the width finite .
we provide more arguments in favor of such a choice in sec .
[ wid ] , which is dedicated to the inclusive decay widths . in sec .
[ vac ] we briefly illustrate how well the duality works for the vacuum correlator of light quarks in the timelike domain .
in the context of the heavy quark expansion this is relevant for the nonleptonic decay widths , including spectator - dependent effects like wa .
section [ concl ] summarizes our investigation and discusses the conclusions that can be drawn for actual qcd .
appendices describe the computational technique employed and contain a number of relations for the heavy quark limit of the t hooft equation employed in these numerical studies .
we first review some well - known properties of the t hooft model both as a reminder and to establish notation .
confinement is manifest in 1 + 1 spacetime dimensions with large @xmath9 , and the quark(@xmath28)-antiquark(@xmath29 ) two - particle irreducible green function , i.e. , the meson wavefunction @xmath30 , is given by the t hooft equation : @xmath31 where @xmath32 is the momentum fraction in light - cone coordinates carried by the quark , and @xmath33 since @xmath34 is finite in the large-@xmath9 limit , it provides a natural unit of mass .
thus , all masses in this paper are understood as multiples of @xmath34 . indeed , as pointed out in ref .
@xcite , @xmath34 fills the role in 1 + 1 dimensions of served by @xmath35 in 3 + 1 .
we discuss the estimation of @xmath34 as a particular number in sec .
[ check ] . the singularity of the qcd coulomb interaction in eq .
( [ the ] ) is regularized using a principal value prescription , indicated by p in eq .
( [ the ] ) .
solutions @xmath36 of the t hooft equation alternate in parity , with the lowest being a pseudoscalar .
the general analytic solution in closed form is not known .
as the eigenvalue index @xmath37 increases , the eigenvalues @xmath38 asymptotically approach @xmath39 $ ] . the static limit @xmath40 is most easily studied @xcite by employing the `` nonrelativistic '' variables @xmath41 , @xmath42 and @xmath43 , in terms of which eq .
( [ the ] ) assumes the form @xmath44 we solve the finite - mass t hooft equation using a numerical method called the multhopp technique @xcite , a venerable system for solving integral equations with singular kernels .
it was first applied to the t hooft equation in ref .
the idea is to expand the wavefunction in a series of modes , not unlike fourier analysis , and then turn the equations for the mode coefficients into an equivalent infinite - dimension eigenvector problem . in practice ,
one then truncates at some point where the higher modes are deemed to have little effect upon the wavefunction solutions , which is of course strongly dependent on the highest value of @xmath37 used .
the detailed formulas for applying the _ standard _ multhopp technique to mesons with unequal quark masses in the t hooft model appear in appendix a of ref . @xcite .
intrinsic to the original multhopp technique is the evaluation of the wavefunction at a discrete set of points called `` multhopp angles , '' which in the current problem are equivalent to @xmath45 \ ; , \hspace{1em } \;\ ; k = 1 , \ldots , n \ ; , \ ] ] where @xmath15 is the number of modes retained in the numerical solution .
the mode coefficients are then obtained by the use of a discrete inversion formula [ ( a7 ) in @xcite ] .
however , the multhopp solutions can be seen to vanish as @xmath46 and @xmath47 at the endpoints @xmath48 and @xmath49 , respectively [ see ( a10)(a11 ) in @xcite ] , while the exact solutions are known to vanish as @xmath50 and @xmath51 , respectively , where @xmath52 leading to a type of gibbs phenomenon in the multhopp solutions .
since the multhopp angles cease to sample the wavefunction at some finite distance from the endpoints , it may be expected that the wavefunctions thus obtained are numerically inaccurate there .
this shortcoming led brower , spence , and weis @xcite to improve the multhopp technique by eliminating the multhopp angles and using instead a continuous inversion formula .
the algebraic details are presented in appendix [ newmult ] , and it is this improved numerical technique that is used in obtaining our results .
let us first establish a bit of notation .
the mass of a heavy quark of flavor @xmath53 is labeled as @xmath54 ; in the weak transitions considered in subsequent sections , the final - state quark @xmath55 is assigned the mass @xmath56 .
the spectator antiquark mass @xmath57 is labeled by @xmath58 , or @xmath59 if there is any chance of confusion .
as explained in the previous section and appendix [ newmult ] , we use the modified multhopp technique to find numerical solutions of the t hooft eigenstate problem .
since the heavy meson wavefunctions are peaked near the end of the interval , the accuracy deteriorates with increasing @xmath13 . the same , in principle , applies to the high excitations of light hadrons .
a more appropriate strategy for heavy quarks is to start with a solution of the infinite - mass ( static ) equation .
this has been done analytically @xcite , and full consistency created by the weak current have @xmath60 , this agreement was shown up to and including @xmath61 terms in the weak decay width in @xcite , while terms up to and including @xmath62 were shown to coincide with those in the ope in @xcite . in the current work we take @xmath63 . ] with the ope was demonstrated
. however , our practical interest lies in the properties of heavy hadrons with @xmath13 lying in the intermediate domain , specifically for @xmath13 one order of magnitude larger than @xmath34 .
the convergence of the @xmath3 expansion in this case is too difficult to quantify analytically .
this is just the situation where the numerical computations are best employed .
therefore , an important element of the analysis is to check the accuracy of the numerical computations of both the heavy hadron masses and wavefunctions at different values of @xmath13 . to this end , we compute the masses and certain moments for the ground and first excited states , and compare them to the analytic @xmath3 expansion . in general , the terms in the @xmath3 expansion depend on a number of expectation values in the static limit , like the kinetic one @xmath64 , etc .
however , one can show @xcite that the parameters appearing here through high order in @xmath3 can be expressed in terms of just the asymptotic value @xmath65 and the corresponding decay constant .
these quantities are the ones most accessible to numerical evaluation ; in particular , @xmath66 is expected to be the most accurately determined quantity .
our main computations refer to the case of @xmath67 , as chosen in ref .
it corresponds ( see sec . [ dual ] ) to a mass of the strange quark in qcd .
the choice of a noticeable light quark mass may be motivated by an attempt to mimic the effect of the transverse gluons absent in @xmath8 , which in a certain respect supply some effective mass to the light quark .
clearly , this can be only a rather crude approximation , since the bare quark mass breaks chiral invariance .
one can suppose , nevertheless , that this side effect is not too important for our purposes . the chiral symmetry is spontaneously broken anyway , and the presence of a massless versus a massive pion does not seem to be essential for the range of problems we address here . on the other hand , the effect of the transverse degrees of freedom is known to soften the @xmath68 singularity of the heavy quark distribution function @xcite , similar to the impact of the light quark mass in the t hooft model .
the behavior of the distribution function affects the inclusive decays of the heavy quarks in an essential way .
we also present some results for @xmath69 , partly to explore light quark dependences of matrix elements and partly to investigate the beginnings of failure of the numerical solutions as @xmath70 .
the number @xmath15 of multhopp modes used is @xmath71 ; we considered smaller @xmath15 as well to study this dependence , but since the behavior was found to be stable , we do not dwell on it further here .
the masses of heavy hadrons obey @xcite @xmath72 where @xmath73 and these expectation values refer to the infinite mass limit . in the @xmath74 expression
, @xmath55 is the momentum variable conjugate to @xmath32 , and diagonal transitions within the correlator have been removed . since @xmath66 in qcd traditionally denotes the mass difference between a ground - state pseudoscalar meson and its corresponding heavy quark in the large @xmath13 limit [ as it is defined in eq .
( [ 50 ] ) ] , and we need it for a number of the excited states @xmath75
as well , we assign the notation @xmath76 and use @xmath77 and @xmath66 throughout the paper on equal footing .
equations ( [ 50])([58 ] ) hold for each state @xmath75 with @xmath78 , so that an implicit superscript @xmath79 is to be understood in these expressions . according to ref .
@xcite , the following relations hold in the t hooft model : @xmath80 \;. \label{52}\ ] ] here @xmath81 is the scaled decay constant in the heavy quark limit , i.e. , @xmath82 where the superscript is suppressed if there is no ambiguity , @xmath83 and the exact relation between @xmath84 and the decay constant of the @xmath37th excitation is given in eq .
( [ dec ] ) . in the heavy quark limit
one has @xmath85 but there are @xmath86 corrections to these relations . for further applications to the decay widths we also consider the scalar expectation value @xcite @xmath87 then the following expansions hold : [ 58 ] c_n & = & ( 1 - ) f^(n ) + o ( ) , + m_q & = & - + + o ( ) , + m_q^2(-^2 ) & = & _ ^2- ( 8 _ ^2 + ^2 f^2 ) + o ( ) , + & = & 1- - + o ( ) .
we note that values of @xmath20 , @xmath19 , or @xmath74 determined from the expansions eqs .
( [ 58 ] ) suffer degraded numerical accuracy compared to those taken directly from eqs .
( [ 52 ] ) since @xmath66 and @xmath81 are determined from more stable expansions ( in particular , they do not depend upon close numerical cancellations ) .
therefore , we use eqs .
( [ 52 ] ) as primary information and relegate eqs .
( [ 58 ] ) to numerical checks .
our method of determining @xmath66 from the @xmath88 expression , designed to minimize the influence of potentially large uncertainties at large @xmath13 , is described in appendix [ calc ] .
values of @xmath88 and and the averages in eqs .
( [ 58 ] ) as functions of @xmath13 from @xmath89 to @xmath90 are presented in table [ t1 ] for both the ground and first excited states .
similar results for just the ground state with @xmath91 are presented in table [ t2 ] . based upon the 10 data points presented in table [ t1 ] for the ground state
, one may fit to a polynomial in @xmath3 , obtaining @xmath92 the corresponding expressions using the approach of appendix [ calc ] ( neglecting the one for @xmath88 , which is used as input and hence is identical through @xmath93 ) read @xmath94 this agreement between the two approaches is quite excellent and is exhibited in figs .
[ m56][mx2 ] for @xmath88 and the quantities in eqs .
( [ 58 ] ) ; in general , the exact results are presented as points on a solid line , while each fit using eqs .
( [ 1mfits ] ) is presented as a dashed line . in fig .
[ m56p ] the analogous expression @xmath88 for the @xmath95 first excited state is presented , while fig .
[ m26 ] uses the same methods and values from table [ t2 ] to present @xmath88 for the @xmath96 ground state . in fig .
[ m56 ] and especially in fig .
[ m26 ] , the quality of numerical results is seen ( as expected ) to begin breaking down at large @xmath13 and small @xmath58 , since @xmath97 is fixed .
we conclude that the numerical routine we rely upon is sufficiently accurate for @xmath97 up to @xmath98 .
the critical value of @xmath13 also depends , however , on the meson s light quark mass , decreasing for small @xmath58 .
this is expected since at small @xmath58 the sharpness of the wavefunction as @xmath99 becomes stronger , and more multhopp functions are required to approximate it : each multhopp function vanishes like @xmath47 .
likewise , the required @xmath15 increases for the excited states . still , one can check that it is possible to go as high as @xmath100 even for @xmath58 as small as @xmath101 .
it turns out that a numerically significant cancellation occurs in the value of @xmath102 in @xmath103 corrections and , in particular , at the @xmath104 level between @xmath19 and @xmath74 , for the ground state just around our primary value @xmath89 .
such a numerical suppression of the power corrections is accidental and does not occur for the excited states , nor for @xmath105 .
let us note that the expectation value of the _ light - quark _ scalar density in the heavy meson turns out very close to unity for the ground state , which may be seen by taking @xmath106 and @xmath107 in eq .
( [ 56 ] ) and referring to table [ light ] ; this is a characteristic feature of a nonrelativistic ( with respect to the light quark ) bound - state system .
it implies an almost simple additive dependence of @xmath66 on the light quark mass @xmath58 , @xmath108 and indeed one can verify this feature by comparing @xmath88 values between table [ t1 ] ( @xmath95 ) and table [ t2 ] ( @xmath96 ) . while this pattern is expected when the spectator quark is heavy , it a priori needs not hold when it is light
this supports the naive expectation that the chiral symmetry breaking may lead to a description in some aspects resembling the nonrelativistic constituent quark model .
the above expectation value , however , decreases for the excited states , as expected from such a picture . drawing semi - quantitative conclusions for qcd
requires a translation rule between the mass parameters in the two theories , that is , an estimate of the value of @xmath34 in @xmath109 .
different dimensionful quantities can be taken as the yardstick ; since the theories are not identical , this translation rule must be introduced with some care .
as follows from the heavy quark sum rules , the physics of duality in the decay widths of heavy flavors crucially depends on the properties of the lowest excited heavy - quark states , in particular the @xmath18-wave excitations with opposite parity to the ground - state multiplet .
it will become evident from the next section that they are of primary importance for the @xmath3 expansion of static properties as well .
therefore , we choose the mass difference between the lowest parity - even ( @xmath18-wave ) state and the parity - odd ground - state meson to gauge the translation between the mass scales . in the t hooft model
the mass difference @xmath110 for light spectators amounts to about @xmath111 .
real charm spectroscopy suggests that the first @xmath18-wave excitations are between @xmath112 and @xmath113 above the ground state .
taking the larger value for sake of illustration , we arrive at the estimate @xmath114 which is adopted in our analysis .
this falls rather close to the estimate of ref .
@xcite , which relied on a quite different type of effects in the light - quark systems .
assuming a value for the `` bare '' @xmath115 quark mass in qcd ( normalized at the appropriate scale @xmath116 ) of about @xmath117 , we conclude that mesons with quarks of masses @xmath118 represent in the t hooft model the actual beauty and charm mesons .
the value of @xmath119 seems to be in a reasonable correspondence with the size of this difference in qcd when it is normalized at a low hadronic scale , @xmath120 @xcite .
it should be noted , however , that the kinetic expectation value in the t hooft model turns out to be rather small , @xmath121 .
this is not surprising , since the chromomagnetic field is absent in two dimensions , while it was shown @xcite to be crucial in the real case .
indeed , the comparison is better justified for the difference @xmath122 in actual qcd versus the value of @xmath20 in the t hooft model .
these questions were discussed in detail in ref .
@xcite , and can be easily understood using the sum rule representation . due to the absence of spin in two dimensions ,
there is no difference between the would - be spin-@xmath123 and spin-@xmath124 light degrees of freedom . in particular , labeling the `` oscillator strengths '' @xmath125 defined in the next section by spin rather than excitation number , @xmath126 and @xmath127 effectively hold .
then the sum @xmath128 and the latter sum is just the general expression for @xmath129 : @xmath130 accepting such an identification suggested in ref .
@xcite and the estimate @xmath131 , we again observe a reasonable agreement with the findings of the t hooft model .
a useful theoretical limit
the so - called small velocity ( sv ) regime was suggested in the mid 80 s @xcite as a theoretical tool for studying semileptonic heavy quark decays .
this refers to kinematics where both @xmath115 and @xmath132 quarks are heavy , but the energy release is limited , so that the velocity of the final charm hadron is small . at large energy release the ope for the width must converge rapidly to the actual hadronic width .
still , at fixed energy release the deviations , although @xmath3 suppressed , are present regardless of the absolute values of masses . in the sv regime
the semileptonic decays proceed either to the ground - state charm final state , @xmath133 or @xmath134 ( the semi - elastic transitions ) , or to excited `` @xmath18-wave '' states of the opposite parity .
other decays are suppressed by higher powers of velocity , or by heavy quark masses .
the equality of the sum of partial decay widths and its ope expansion is achieved through the sum rules that relate the sums of the @xmath18-wave transition probabilities , weighted with powers of the excitation energies , to the static characteristics of the decaying heavy hadron .
the onset of convergence of the ope expansion for the widths is then directly related to the pattern of saturation of the sum rules by the lowest excitations .
if higher states contribute significantly , they delay the onset of duality , while their absence leads to a tight quark - hadron duality after the first @xmath18-wave channel is open .
knowledge of degree of saturation of the heavy quark sum rules is also important for another reason : it determines the hadronic scale above which one can apply the perturbative treatment to compute corrections or account for evolution of the effective operators .
the lower this scale , the more predictive in turn is the treatment of the nonperturbative effects in the ope
. a recent review of the sv sum rules can be found in ref .
@xcite ( the perturbative aspects are discussed in more detail in ref .
@xcite ) . for most practical purposes addressed here
, one can consider the perturbative effects to be absent in the t hooft model .
in particular , the heavy quark parameters do not depend perturbatively on the normalization point , and there is no need in the explicit ultraviolet cutoff to introduce a normalization point . the sum rules we address are _ k^2- & = & _ n [ oscsum1 ] _
nk^2 , + _ k & = & _ n ( _ n-_k ) _
nk^2 , + ( _ ^2)_k & = & _ n ( _ n-_k)^2 _ nk^2 , + ( _ d^3)_k & = & _ n ( _ n-_k)^3 _ nk^2 .
[ oscsum4 ] here @xmath135 and @xmath37 denote excitation indices for the initial and final states , respectively ( in practice only transitions from the ground state are interesting , so we limit ourselves to @xmath136 ; in this case the index @xmath135 is omitted ) . the so - called `` oscillator strengths '' @xmath125
parameterize the transition amplitudes into the opposite - parity states in the sv limit , = _ nk _
v^+ o(^3 ) , [ 102 ] where @xmath137 is the velocity of the final state hadron . in the diagonal transition
@xmath138 is the slope of the isgur - wise ( iw ) function of state @xmath139 : = 1-_k^2 + o ( ^4 ) .
[ 104 ] the expressions for @xmath140 and @xmath141 in terms of the light - cone wavefunctions are _ nk & = & _ 0^ dt _
k(t ) = -_m_q _ 0
n(x ) ( 1-x ) _
k(x ) , + ^2_k & = & _ 0^ dt |(t + ) _ k(t)|^2 = _ m_q _ 0 ^ 1 dx |_k(x)|^2 .
+ [ 106 ] the finite-@xmath13 corrections to the @xmath32-integral forms turn out to be rather significant , leading to significant problems in precision numerical studies .
to avoid this problem we use the analytic expression for the inelastic amplitudes obtained in ref .
@xcite : @xmath142 where @xmath143 are the asymptotic values of the decay constants @xmath84 scaled up by the factor @xmath144 , as in eq .
( [ 112 ] ) .
the constants @xmath143 are computed as the values of @xmath145 at @xmath146 ( see table [ t1 ] ) augmented by the @xmath3 corrections detailed in the first of eqs .
( [ 58 ] ) , while values of @xmath147 are computed using the procedure described in appendix [ calc ] .
the results of the computations for the case @xmath100 , @xmath148 are presented in table [ osc ] .
our central result is a surprisingly good saturation of the sum rules : the first @xmath149 excitation generates @xmath150 of @xmath151 , @xmath152 of @xmath66 , @xmath22 of @xmath20 , and even @xmath153 of @xmath19 .
the rest is almost completely saturated by the second @xmath18-wave state ( @xmath154 ) , where the cumulative values for the same quantities read @xmath155 , @xmath156 , @xmath157 , and @xmath158 , respectively . in terms of absolute numbers , the sum rules eqs .
( [ oscsum1])([oscsum4 ] ) would give @xmath159 , @xmath160 , @xmath161 , and @xmath162 , the last of which gives @xmath163 , in fine agreement with the values obtained from the values obtained in the previous section via the methods described in appendix [ calc ] .
the few - percent discrepancy corresponds to the accuracy in determinations of squared decay constants .
the level of saturation by the lowest open channels is extraordinary .
the explicit reason for such a perfect saturation of the sum rules involving even rather high , @xmath164 powers of the excitation energy can be read off eq .
( [ 110])@xmath165 s are inversely proportional to the third power of the excitation energy . with the asymptotics @xmath166 , @xmath167 ,
the first excitation energy @xmath110 is notably smaller than the next one @xmath168 including three energy gaps .
the general peculiarity of the t hooft model leading to such a saturation is not understood completely . with this pattern of saturation of the sv sum rules for the ground - state meson ,
one expects an early onset of the accurate duality for the inclusive widths in the @xmath17 transitions , only slightly above the threshold of the first excitation .
demonstrating this result through direct evaluation of the decay widths is one of the purposes of the next section .
the semileptonic widths described in this work were considered in detail in ref .
@xcite . here
we recapitulate a few basic points .
the weak decay lagrangian is @xmath169 in terms of the previous notation , @xmath170 , @xmath171 ( or , later in this section , @xmath172 ) , and @xmath173 .
the key property of all @xmath8 vectorlike currents is that for @xmath174 , the invariant mass @xmath175 of the lepton pair is always zero . for all computational purposes
decays into this massless lepton pair are equivalent to decays into a single massless pseudoscalar particle @xmath176 weakly coupled to quarks according to @xmath177 several arguments favor our choice of a vectorlike weak decay interaction in the t hooft model .
one is of course the simplicity of eq .
( [ 122 ] ) .
another is that for @xmath178 some difficult problems of renormalization are absent , as we now discuss .
the central problem in applying the ope in practice is disentangling perturbative and nonperturbative effects .
more precisely , this refers to the separation of short - distance effects attributed to the coefficient functions from long - distance effects residing in the matrix elements of the effective heavy - quark operators .
the perturbative corrections , for example those that renormalize the weak quark current , are generally rather nontrivial , even in the t hooft model . however , according to the nonrenormalization theorem of ref .
@xcite , such vertex corrections are absent from the decays with @xmath178 .
this allows one to isolate the problem of renormalization of the underlying current from the question of interest in our study : possible deviations of the full decay widths due to the presence of thresholds in the production of the hadronic resonances . in reality , from the ope viewpoint some short - distance corrections still remain even in this special kinematic region due to the high - momentum tails in the meson wavefunctions .
these tails come from the hard gluon exchanges between the constituents . in principle
, these `` hard '' components can also be separated from the `` soft '' bound - state dynamics explicitly . however , in practice this is not necessary : these effects are completely contained in the meson wavefunctions . another advantage of vectorlike currents is apparent when one notes that the @xmath8 three - body `` semileptonic '' phase space diverges logarithmically for massless leptons .
explicitly , for the decay @xmath179 ( equal lepton masses are assumed to render the expressions simpler ) , the three - body phase space turns out to be @xmath180}{(m\!-\!m)^2 \left [ ( m\!+\!m)^2 \!-\!4m_\ell^2 \right ] } \right],\end{aligned}\ ] ] where @xmath181 is the complete elliptic integral of the first kind . as @xmath182 , one regains eqs .
( 4.1)(4.3 ) of @xcite , while as @xmath183 , the argument of the elliptic integral goes to unity , and @xmath184 .
this is a manifestation of the logarithmic infrared divergence of the massless scalar green function at large distance in @xmath8 .
a detailed calculation shows that the vector nature of the weak coupling regularizes the phase space integral , preventing the partonic rate from diverging in the limit of massless leptons .
furthermore , as discussed in the introduction , this also removes the @xmath185 singularity in the threshold behavior for hadronic two - body decays . as a final advantage of vectorlike currents and the special kinematic point @xmath178 ,
note that at @xmath178 the @xmath186 transition amplitudes are directly expressed in terms of the overlap between the initial and the final wavefunctions : @xmath187 where @xmath188 , so that the partial decay width for @xmath189 is given by @xmath190 the threshold suppression mentioned above is manifested in the explicit factor @xmath191 : the reciprocal of this factor in the phase space is removed by @xmath192 from the matrix element .
it is also possible to derive this result directly using the methods of ref .
@xcite ; note , however , that these expressions are much simpler than those of ref .
@xcite , because the vectorlike current with massless leptons restricts @xmath175 to 0 .
the sum of these widths over all open channels is to be compared to the ope prediction .
the remarkable speed of saturation in @xmath37 , anticipated in the last section , is illustrated for one sample case in table [ satur ] .
turning to the ope , we mention one more problem associated with an accurate understanding of local duality violation . apart from the purely theoretical aspect that ope power series are generally only asymptotic and ,
thus have a formally zero radius of convergence in @xmath3 , one normally has additional practical limitations .
only a limited number of the terms , as well as the associated expectation values , are usually known , which places additional theoretical uncertainties that dominate in practice at sufficiently large @xmath13 .
this feature can be naturally incorporated in the analysis of our concrete model .
we account completely only through terms that scale like @xmath14 , the highest order that emerges from the ope free from the four - fermion operators @xcite .
the rest , although calculable in principle term - by - term in the t hooft model , are taken to represent the ope `` tails '' discarded by the unavoidable truncation .
using the sum rules of the t hooft model , ref .
@xcite established the following exact representation for the total decay width : @xmath193 where @xmath194 at @xmath195 are understood as given by eq .
( [ 124 ] ) without the explicit @xmath196-function singling out the open channels ; such @xmath194 are therefore all negative . on the other hand ,
the ope yields the result @xmath197 \ ; , \label{132}\ ] ] with @xmath198 generically denoting the ope expansion parameter ; we do not specify here if it is @xmath199 or @xmath200 , or some other combination .
it was shown in ref .
@xcite that the @xmath194 term in eq .
( [ 130 ] ) is dual to the order term in eq .
( [ 132 ] ) ; however we do not use this here and rather treat the latter as an intrinsic uncertainty in the `` practical '' version of the ope .
thus , our strategy is to compare the exact width @xmath201 to @xmath202 the expectation value @xmath203 above can either be evaluated numerically , or in the spirit of the ope , computed in the form of a @xmath204 expansion , the last of eqs .
( [ 58 ] ) .
it turns out that the expansion converges very rapidly to the exact result , so that this does not significantly affect the observed pattern of local duality at the quantitative level .
the born - term partonic rate is simply given by @xmath205 with this expectation value set to unity ,
@xmath206 the main practical interest of these calculations lies in the @xmath207 width with its limited energy release @xmath208 . in general
, @xmath208 can be small either if @xmath199 is not large enough , or even at large @xmath199 if @xmath209 ( or @xmath210 if @xmath175 is nonzero ) is insufficient due to a significant @xmath132 quark mass .
the latter case falls into the sv category , and the violations of duality are suppressed here even at the maximal @xmath175 by heavy quark symmetry , as was pointed out in the mid-80 s @xcite .
therefore , one a priori expects a different pattern in the two cases .
we try to separate the possible effects by considering different choices for @xmath199 and @xmath211 rather than by only taking them close to their realistic values . with these arguments in mind
, one can expect to find significant effects of duality violation in the cases where @xmath212 or @xmath204 effects are important . as suggested in ref .
@xcite , in this case it is advantageous to fix @xmath199 close to its actual value , and vary @xmath211 from near @xmath199 down to smaller values , changing in this way the energy release . at one end of the interval
the local duality is supported by the heavy quark symmetry with large quark masses and sv kinematics , while at another end it rests on the large energy release .
we start from the sv case when @xmath199 is fixed and large and @xmath211 is large as well , varying the energy release by increasing @xmath211 towards @xmath199 .
since the violation of local duality is expected to be suppressed for all values of @xmath211 , high numerical accuracy is vital .
we fix @xmath213 ( @xmath214 ) , and vary @xmath211 from @xmath215 up to @xmath199 .
the results are given in table [ dual15 ] and fig .
[ dual15fig ] .
we note that the difference between the two widths is so small that one must plot @xmath216 rather than the widths themselves .
this is expected since the sv sum rules are very well saturated , as detailed in the previous section the higher thresholds are then strongly suppressed numerically at finite energy release . but for @xmath211 approaching @xmath199 , where they could be noticeable , the heavy quark symmetry works efficiently since both quarks are very heavy .
in fact , the only prominent features on the plot occur when thresholds to the first few @xmath133 states of opposite parity to the ground - state @xmath217 meson are crossed , for example between @xmath218 and @xmath219 .
the deviation is extremely small also for smaller @xmath211 where the @xmath132 quark velocity is rather large
yet there the energy release is significant , and a large number of excited states ( up to @xmath220 at @xmath221 ) are produced .
table [ dual10 ] and fig
. [ dual10fig ] show analogous results for @xmath222 , @xmath148 .
to render the duality violation more apparent , we consider ( table [ dual5 ] and fig .
[ dual5fig ] ) the same decay widths for a @xmath115 quark with half the mass , @xmath223 .
even here the deviation is below per mill as soon as the first excitation can appear with sufficient phase space .
the duality - violating component at last exhibits the proper oscillating behavior ( note the decrease between @xmath224 and @xmath225 or 1 and @xmath226 ) , but this effect is too small to be extracted reliably at larger energy release where this property becomes an asymptotic rule . as follows from our computations , local duality is violated at a tiny level in the @xmath17 decays in the t hooft model whenever it is a priori meaningful to apply ope .
a possible reason behind this might be that for unidentified reasons the heavy quark symmetry works for the inclusive widths too effectively , down to relatively low masses and velocities of order @xmath227 .
this was conjectured in the early papers on the subject @xcite
. therefore , our final attempt in the quest for the sizeable duality violation in beauty is considering the @xmath228-type transitions , where the heavy quark symmetry per se does not constrain the individual transition form factors .
we fix in our expressions @xmath229 or @xmath230 ( but still keep the two quarks flavor - distinguished ) and vary @xmath199 from @xmath231 to @xmath232 .
the results are shown in table [ dualmc56 ] and fig .
[ dual_mc56 ] , and table [ dualmc26 ] and fig .
[ dual_mc26 ] , respectively .
although the difference between the actual width and its ope approximation is larger , it still is very small and approaches a percent level for @xmath199 as low as @xmath233 .
the total decay width is no longer saturated to such a high degree by transitions to the ground state , especially for larger @xmath199 .
nevertheless , the duality is amazingly well satisfied when just the first few open channels are summed .
again , the only prominent features in the plots appear when crossing kinematic thresholds due to the lightest @xmath133 mesons of opposite parity to the ground - state @xmath217 .
the extraordinary agreement between @xmath234 and @xmath235 may be underscored by instead plotting ( fig .
[ part_mc56 ] , final column of table [ dualmc56 ] ) the difference between @xmath234 and the born - term partonic rate @xmath236 given in eq .
( [ born ] ) . from an algebraic point of view , @xmath234 and @xmath235 differ generically at @xmath237 , while @xmath234 and @xmath236 begin to differ already at @xmath238 .
thus , we find local duality between the actual semileptonic decay width and its ope expansion to be very well satisfied in all cases . before concluding this section ,
let us briefly address duality in the differential distribution @xmath239 . in the heavy quark limit
the shape of the final - state hadronic mass distribution follows the heavy quark distribution function in the decaying meson ; for the @xmath240 decays under consideration , this is the light - cone distribution function @xmath241 . in decays with @xmath178 the recoil energy of the lepton pair @xmath242
is directly related to the final state mass @xmath243 : e = . [ 150 ] since @xmath178 , these decays are analogous to @xmath244 in the standard model . in the large-@xmath199 limit one has = f ( ) .
[ 152 ] at finite @xmath199 in a theory with narrow resonances the actual distribution is given by the comb of @xmath245-functions with spacing in the argument of eq .
( [ 152 ] ) of order @xmath246 . in order to obtain a continuous result
, we adopt the simple ansatz of averaging over the peaks . using eq .
( [ 150 ] ) to define the energy @xmath247 of the @xmath37th state @xmath243 , we integrate the @xmath245-function for the @xmath37th state evenly over the energy range @xmath248 to @xmath249 , i.e. , the midpoints between energy eigenvalues . letting @xmath15
be the maximum number of kinematically allowed @xmath243 values , we establish the endpoint bins by defining @xmath250 and @xmath251 .
we find that our numerical computations yield a distribution resembling the light - cone distribution function @xmath252 ; specifically , f(y)=^2 ( ( 1-y ) ) = _ m_q ^2 ( 1-(1-y ) ) .
[ 154 ] recalling that @xmath253 and combining eqs .
( [ 152 ] ) and ( [ 154 ] ) yields @xmath254 the two sides of this expression are plotted in fig .
[ dgde_25v10 ] , using @xmath255 to represent the limit @xmath256 , while the actual distribution is considered at @xmath257 .
the agreement is quite remarkable .
the continuous distribution appears to pass approximately through the midpoint of each bin ; owing to the near - equal spacing of t hooft model eigenvalues in @xmath38 , eq .
( [ 150 ] ) shows that these bin midpoints are very close to the values @xmath247 themselves .
it is also interesting to consider integration over a range of @xmath242 . in particular , define @xmath258 as the cumulative fractional width from maximum energy @xmath259 down to the given @xmath242 ; then @xmath260 and @xmath261 .
while the exact result for @xmath262 amounts to an integration of the @xmath245-function differential widths renormalized so that the cumulative result approaches unity , the integral of the continuous distribution gives @xmath263 these two curves are presented in fig . [ inc10 ] .
two features particularly stand out in this plot .
first , even for @xmath199 as large as @xmath264 , the overwhelming part of the decay probability falls into the transitions to at most four lowest states .
second , the continuous curve seems to provide a nearly optimal description possible for the step - like exact distribution .
the point - to - point deviation for all plotted values with @xmath265 does not exceed half of the contribution of the nearest threshold .
in this section we briefly illustrate the onset of duality for the absorptive part of the vector current correlator with light quarks , of the type that determines the normalized cross section @xmath266 as a function of energy . in the context of the heavy quark
decays this is relevant in nonleptonic decay widths in two kinds of processes : in spectator - independent decays , where @xmath267 determines the weight with which the semileptonic width at given @xmath175 must be integrated over @xmath175 ( see ref .
@xcite ) , and in the effects of wa decays . in either case , at @xmath268 the cross section appears as a comb - like collection of @xmath245-functions : r(q^2)= _
n c_n^2 ( q^2-m_n^2 ) ; c_n= _ 0 ^ 1 dx _
[ 200 ] the above expression for the residues refers to the case where a vector current is considered .
we suppress here the factor of @xmath269 relating @xmath84 to @xmath270 [ eq .
( [ dec ] ) ] .
we also assume in what follows that the light quark masses are equal , @xmath271 , and are @xmath272 or less , in order to reach asymptotic @xmath175 more quickly . in the extreme situation of infinitely narrow resonances one can not , of course , discuss a point - to - point equality of the cross section @xmath267 with its ope in the form of @xmath273 expansion .
a meaningful comparison is possible if each resonant peak is somehow averaged over an interval no smaller than the distance between adjacent peaks , the latter being approximately given by @xmath274 @xcite .
it is worth recalling that @xmath267 is proportional to @xmath275 , so one must consider nonvanishing masses for the vector current , and address the ope terms formally suppressed by @xmath276 .
this question was first addressed in the context of nonleptonic decays in ref . @xcite using the numerical approach .
duality for the average cross section in the same manner as above , i.e. , using sum rules derived from the t hooft equation and analytically matching terms in the @xmath3 expansion , was obtained in ref .
yet establishing the asymptotics per se can not tell us beforehand how early one can expect the onset of duality . here
we study this question numerically , in the domain of intermediate @xmath175 .
the concrete amount of the deviation between @xmath267 and @xmath277 in the case of direct resonances may depend in an essential way on the chosen smearing procedure .
interested in the qualitative features only , we choose a rather simplified , crude method : we spread the integral of @xmath267 evenly over the interval between the successive resonances .
more precisely , we put @xmath278 , with @xmath279 , the partonic pair production threshold . here
we use the fact that @xmath84 vanish for odd @xmath37 when @xmath280 .
this smearing is very similar to that described for the differential width in the last section , except that averaging is performed in @xmath175 rather than @xmath242 .
the free quark loop @xmath267 , which is of course the leading term of the ope , is given by r_0(q^2)= . [ 212 ] table [ vaccor ] and fig . [ vacfig ] show the results for our reference case @xmath148 .
the agreement of the average hadronic cross section with the parton - computed probability again turns out to be very good .
apparently , this can be related to two facts : the heavy suppression of power corrections to @xmath267 in the ope ( see eqs .
( 34)(35 ) in @xcite ) , and an early onset of the asymptotics in the spectrum , @xmath281 which even at @xmath282 is satisfied to about @xmath283 .
the main motivation behind the present study has been to assess the magnitude of local duality violations in the inclusive semileptonic decays of beauty particles .
we considered this question using the t hooft model as a toy theory in which all relevant decay amplitudes can be evaluated numerically . the t hooft model , while retaining certain key features of full @xmath7 qcd that shape the spectrum of hadrons ( quark confinement , chiral symmetry breaking ) , still differs from @xmath7 in many respects .
yet using it as a lab for exploration carries an important advantage
it allows no `` wiggle room '' for interpretation of the results .
there are no ad hoc parameters to choose or adjust , and as soon as the underlying weak decay lagrangian is fixed , the numerical results are unambiguous and must be accepted at face value .
this positively distinguishes this approach from various models where often the conclusions , even qualitatively , depend on the arbitrary choice of parameters according to one s preferences .
the question of a particular model being compatible with the general dynamical properties of qcd underlying the ope approach , often quite problematic in simplified quark models , does not arise for the t hooft model .
although the simplest illustration of the asymptotic nature of the decay width @xmath3 expansion and related violations of local duality @xcite follows just from the existence of hadronic thresholds ( see , e.g. , @xcite ) , violation of local duality is a more universal phenomenon that is _ not _ directly related to existence of hadronic resonances nor even confinement itself .
this has been illustrated in ref .
@xcite by the example of soft instanton effects that do not lead , at least at small density , to quark confinement but do indeed generate computable oscillating duality - violating contributions to the total decay rates . nevertheless , there is a widespread opinion that decays with manifest resonance structure in the final state are most difficult for if compatible at all with the standard ope .
even the possibility that the ope does not fully apply in the case of `` hard '' confinement has been occasionally voiced in the literature .
the analytic studies performed in refs .
@xcite , which explicitly demonstrate in the t hooft model the applicability of the ope to the total widths , should help to allay such conceptual concerns .
nevertheless , the intuition remains that resonance dominance is not `` favorable '' for the ope , and problems might show up , for instance , through a delayed numerical onset of duality , in that the approximate equality of the ope predictions and the actual decay widths may set in only after a significant number of thresholds has been passed . to address such issues
, the t - hooft model seems to represent the most certain testing ground for local duality in the domain of decays of moderately heavy quarks .
contrary to naive expectations , we found surprisingly accurate duality between the ( truncated ) ope series for @xmath284 and the actual decay widths .
the deviations are suppressed to a very high degree almost immediately after the threshold for the first excited final state hadron is passed . no suspected delay in the onset of duality was found . the key property that governs the onset of the @xmath3 expansion for the semileptonic widths is the pattern of saturation of the heavy quark sum rules .
we examined a particular class , the sv sum rules in the heavy quark limit , that has the most transparent quantum mechanical meaning .
we found them saturated to an amazing degree by the very first excitation .
the contribution of the remaining , higher states to the slope of the iw function , @xmath66 , and @xmath20 does not exceed a few percent . even in the darwin operator sum rule , the first excitation accounts for @xmath21 of the whole expectation value , despite the fast - growing weight , @xmath285 of higher - order contributions .
this peculiarity underlies the early onset of duality for the case when initial- and final - state quarks are both heavy .
some of the duality - violating features observed in these studies have natural explanations . at fixed energy release @xmath286 the magnitude of the deviations is smaller if @xmath13 , @xmath56 are both large ( as in @xmath287 ) than if they are both small .
this is expected , since in the former case the heavy quark symmetry for the elastic amplitude additionally enforces approximate duality even when no expansion in large energy release can be applied .
it is interesting , however , that at fixed @xmath199 the duality violation decreases rapidly as @xmath211 decreases , in full accord with the ope where the higher order terms are generally suppressed by powers of @xmath288 .
this is clearly a _
dynamical _ feature that goes beyond heavy quark symmetry per se , the quality of which deteriorates as @xmath211 decreases .
it is also instructive to note that including the calculated power - suppressed ope terms significantly reduces the difference between the actual decay width and its purely partonic evaluation .
moreover , the seeds of oscillations inherent to duality violation ( as functions of quark masses ) , can be seen . since we adopted the truncated ope expansion to mirror the existing implementation of the ope in qcd
, the deviations do not average to zero but rather oscillate around the ( rapidly dissipating ) contributions attributed to discarded higher - order terms .
the numerical effects of duality violation we study turn out to be typically quite small .
partially this can be attributed to moderate size of the corresponding expectation values multiplying @xmath289 corrections in the ope . yet
certainly not all power corrections in heavy quarks are suppressed in the model .
it is well known from ordinary quantum mechanics that masses ( eigenvalues ) typically are much more robust against perturbations than wavefunctions themselves ( or transition amplitudes ) .
we observe a similar pattern in the t hooft model .
for example , @xmath3 corrections to the meson decay constants turn out very significant even at the scale of the @xmath115 quark mass .
apparently , the inclusive decay rates fall into the class of `` robust '' observables , although , as explained above , this was difficult to anticipate beforehand .
we note here another `` fragile '' observable , the light - cone heavy quark distribution function , which can be measured in decays of the type @xmath290 . in @xmath8 the scaled spread @xmath291 of the @xmath32 distribution approaches @xmath20 at large @xmath13 . yet , as seen in fig .
[ mx2 ] , even at the @xmath115 quark mass one would obtain from this distribution only about @xmath292 of the actual value of @xmath20 , due to significant @xmath3 corrections .
this caveat may be important for existing analyses of the decay distributions in @xmath217 decays , where such effects routinely are not included .
we also briefly addressed the inclusive differential decay distributions in the analogues of @xmath293 or @xmath294 decays .
generally , we find good agreement ( at the scale corresponding to the physical @xmath115 mass ) with the parton - based prediction incorporating effects of the `` fermi motion , '' and in particular for the partially integrated probability ( x ) =
_ 0^xm_b^2 dm_h^2 .
[ c4 ] this distribution , following refs .
@xcite , is examined in real @xmath217 decays in the quest for @xmath2 @xcite . however ,
the point - to - point deviations are clearly still significant , for the decays to only the 4 or 5 lowest final states saturate the overwhelming fraction of the total decay probability .
it is quite conceivable , though , that such deviations are less pronounced in actual qcd owing to the significant resonance widths and to a richer resonance structure .
the vacuum current correlator also turns out to be especially robust ; even neglecting all ope corrections except the leading partonic contribution leads to excellent agreement with the hadronic result . turning to the direct phenomenological conclusions that can be inferred from our studies , we see that , to the extent our findings can be transferred to real qcd , violation of local duality in the total semileptonic widths of @xmath217 mesons is not an issue .
the scale of duality violation lies far below the phenomenologically accessible limits , and can not affect the credibility of @xmath1 or @xmath2 extractions . in reality
there are , of course , essential conceptual differences between the two theories , including those aspects that are expected to be essential for local duality ( for a discussion , see ref .
although many seem to optimistically suggest that duality violation is more pronounced in the t hooft model than for actual heavy flavor hadrons , some differences may still work in the opposite direction .
in @xmath8 there are no dynamical gluons , nor a chromomagnetic field that in @xmath7 provides a significant scale of nonperturbative effects in heavy flavor hadrons .
likewise , there is no spin in @xmath8 , and no corresponding @xmath18-wave excitations of the light degrees of freedom ( the so - called @xmath295 states ) , which seem to play an important role in @xmath7 .
two - dimensional qcd neither has long perturbative `` tails '' of actual strong interactions suppressed weakly ( by only powers of @xmath296s of the energy scale ) . in @xmath8
the perturbative corrections are generally power - suppressed , as follows from the dimension of the gauge coupling . as discussed in ref .
@xcite , it is conceivable that the characteristic mass scale for freezing out the transverse gluonic degrees of freedom is higher than in the `` valence '' quark channels .
this would imply a possibly higher scale for onset of duality in @xmath297 corrections to various observables .
regardless of these differences , we conclude that presence of resonance structure per se is not an obstacle for fine local quark - hadron duality tested in the context of the ope . as we see in the t hooft model
, resonances themselves do not seem to demand a larger duality interval .
as soon as the mass scale of the states saturating the sum rules in a particular channel ( quark or hybrid ) has been passed , the decay width can be well approximated numerically by the expansion stemming from the ope .
the ground states of heavy mesons in the t hooft model exhibit relatively small expectation values of nonperturbative operators ( @xmath20 , @xmath19 , but not @xmath66 ) compared to real qcd , if our identification @xmath298 is adopted .
this may be regarded as a reason for small duality violation for @xmath299 in the model . however , even if we scale @xmath34 up to @xmath300@xmath301 to make up for smallness of the nonperturbative ope effects , the duality violation is still very small , and superficially rather insignificant even in charm .
we note , however , that the specific choice eq .
( [ 120 ] ) of the weak interaction effectively requires decays to occur only at @xmath178 , and therefore the effects of four - fermion operators of the type @xmath302 are totally absent , at least in the lowest orders of perturbation theory ( cf.ref .
@xcite , sec .
iii.b.3 ) . as was suggested in ref .
@xcite , it is conceivable that the apparent excess in @xmath303 is simply related to a noticeable magnitude of the non - valence ( nonfactorizable ) expectation values @xmath304 .
if this conjecture is true , similar effects in @xmath305 are still suppressed but possibly detectable in future precision experiments . in the context of the present study , it suffices to say that this would be a legitimate ope effect rather than a manifestation of a significant local duality violation in the strict sense . +
* acknowledgments : * r.l .
thanks the department of energy for support under contract no.de-ac05-84er40150 ; n.u . acknowledges the support of the nsf under grant number phy96 - 05080 , by nato under the reference pst.clg 974745 , and by rffi under grant no . 99 - 02 - 18355 .
we are grateful to n. isgur for inspiring interest and discussions , and to m. burkardt for invaluable insights .
n.u . also thanks i. bigi , m. shifman and a. vainshtein for encouraging interest and collaboration on related issues , and a. zhitnitsky for useful comments .
n.u . enjoyed the hospitality of physics department of the technion and the support of the lady davis grant during completion of this paper .
the brower - spence - weis @xcite ( bsw ) improvement of the multhopp technique avoids the need for evaluating the wavefunction at a discrete set of points called `` multhopp angles , '' thus improving the behavior of the solutions in the endpoint regions , as described in sec .
[ thm ] . here
we exhibit the expressions used by bsw , correcting along the way some minor typographical errors in their work . starting with the t hooft equation ( [ the ] ) with bare quark masses @xmath28 and
@xmath57 , one converts the kinematic variables @xmath306 to angular variables : @xmath307 in terms of which the t hooft equation reads @xmath308 \varphi_p ( \theta ) \nonumber \\ & & + \int_0^\pi { \rm d}\theta^\prime \ , \varphi_p ( \theta^\prime ) \ , { \rm p } \frac{1}{(\cos { \theta } \!-\ !
\cos { \theta^\prime})^2 } .\end{aligned}\ ] ] expanding @xmath309 and using the continuous inversion identity ( contrast with eq .
( a7 ) of ref .
@xcite ) @xmath310 one obtains the infinite - dimensional eigenvector system @xmath311 where @xmath312\ , \sin { n\theta } \,\sin { m\theta } , \\ v_{nm } & = & -\frac{4}{\pi } \beta^2 \int_0^\pi { \rm d}\theta \ , \sin { n\theta}\ , \int_0^\pi { \rm d}\theta^\prime \,\sin{\theta^\prime } \sin { m\theta^\prime } \ , { \rm p } \frac{1}{(\cos { \theta } \!-\ ! \cos{\theta^\prime})^2 } \;. \label{vmn1}\end{aligned}\ ] ] both of these integrals can be evaluated , with the result @xmath313 , \\
v_{nm } & = & v_{n-1,m-1 } \left ( \frac{m}{m\!-\!1 } \right ) + \frac{8m}{n\!+\!m\!-\!1 } \left [ \frac{1+(-1)^{n+m}}{2 } \right ] , \end{aligned}\ ] ] where @xmath314 for @xmath315 .
this recursive form for @xmath316 is most convenient for numerical calculations ; however , one may also write the closed - form solution , @xmath317 \left [ \psi \left ( \frac{1\!-\!n\!-\!m^{\vphantom{2}}}{2 } \right ) - \psi \left ( \frac{1\!-\ ! @xmath318 in eq .
( [ vmn2 ] ) is real but not symmetric , owing to the extra @xmath319 in eq .
( [ vmn1 ] ) ; therefore , the `` hamiltonian '' @xmath320 is not hermitian , and the eigenvectors @xmath321 are not orthogonal .
this is a direct result of converting the exact wavefunctions , which are eigenfunctions of a hermitian hamiltonian when written in terms of the variable @xmath32 ( and therefore orthogonal in @xmath32 ) , into orthogonal functions of the variable @xmath196 .
this transformation is nonunitary because the number of modes used is not infinite ; therefore , the overlap of different eigenvector solutions should be small when a large number of modes are used .
indeed , this turns out to be empirically true ; nevertheless , we take the further step of orthogonalizing the numerical eigenvector solutions recursively by means of the standard gram - schmidt procedure , i.e. , @xmath322 for @xmath323 modes , this typically changes expectation values by one part in @xmath324 .
the expressions for these overlaps and other matrix elements in terms of the mode coefficients @xmath325 are presented in appendix [ ovlap ] .
a number of useful overlaps and other integrals are straightforward to evaluate in terms of the mode coefficients , using the expressions ( [ exp ] ) . solving them amounts to evaluating a number of trigonometric integrals .
such expressions are especially convenient since they permit a number of integrations that introduce no numerical uncertainties ( except due to machine precision ) beyond those of solving the original multhopp - bsw eigenvector equation ( [ mulbsw ] )
. in particular , denote the @xmath326th eigenstate wavefunction presented in eq .
( [ exp ] ) by @xmath327 and that for some other set of masses in the @xmath55th eigenstate by @xmath328 ; the latter wavefunction then has an expansion like ( [ exp ] ) with mode coefficients @xmath329 .
truncating after @xmath15 modes , one then finds @xmath330 \frac{1}{\left [ 1 \!-\ !
( m\!-\!n)^2 \right ] \left [ 1 \!-\ !
( m\!+\!n)^2 \right ] } .
\nonumber \\ & & \end{aligned}\ ] ] indeed , the normalization integral @xmath331 is just the case @xmath332 and @xmath333 , in agreement with eq .
( a9 ) of ref .
@xcite .
other useful expectation values include @xmath334 ^ 2 \nonumber \\ & = & -\sum_{m=1}^n m a_m^{(p ) } \sum_{n=1}^n n a_n^{(p ) } \left [ \frac{1 \!-\ ! ( -1)^{m+n}}{2 } \right ] \frac{1}{\left [ 4 \!-\ ! ( m\!-\!n)^2 \right ] \left [ 4 \!-\ ! ( m\!+\!n)^2 \right ] } , \nonumber \\ & & \end{aligned}\ ] ] @xmath335 ^ 2 \nonumber } \\ &
= & -\frac 1 2 \sum_{m=1}^n m a_m^{(p ) } \sum_{n=1}^n n a_n^{(p ) } \left [ \frac{1 \!+\ ! ( -1)^{m+n}}{2 } \right ] \left [ 21 \!-\ !
6(m^2\!+\!n^2 ) + ( m^2\!-\!n^2)^2 \right ] \nonumber \\ & & \times \left [ \left ( 1 \!-\ ! ( m\!-\!n)^2 \right ) \left ( 1 \!-\ ! ( m\!+\!n)^2 \right ) \left ( 9 \!-\ ! ( m\!-\!n)^2 \right ) \left ( 9 \!-\ ! ( m\!+\!n)^2 \right ) \right]^{-1 } .\end{aligned}\ ] ] note that the spread of the wavefunction may be computed about any convenient point in @xmath32 , viz . , @xmath336 so that the additive constants of @xmath337 above are irrelevant . also , @xmath338 ^ 2 \:=\ : \sum_{m=1}^n a_m^{(p ) } \,\sum_{n=1}^n a_n^{(p ) } \ , i_{mn } , \ ] ] where @xmath339 one also finds @xmath340 ^ 2 \,=\ , \sum_{m=1}^n \,a_m^{(p ) } \,\sum_{n=1}^n \,a_n^{(p ) } \ , j_{mn } \;,\ ] ] where , using the notation of eq .
( [ imn ] ) , one finds @xmath341 for @xmath342 even , and @xmath343 for @xmath342 odd . finally , the decay constant of the @xmath326th excitation [ cf.eqs .
( [ 112])([cn ] ) ] is given by @xmath344
the numerical calculation of large-@xmath13 matrix elements with acceptable accuracy relies on achieving a balance between competing effects .
on one hand , multhopp solutions to the t hooft equation with @xmath345 tend to suffer degraded numerical accuracy since they are highly concentrated into the small kinematic region @xmath346 . as discussed in sec .
[ thm ] , the endpoint regions @xmath347 and @xmath227 are where the multhopp solutions or more precisely , their derivatives tend to break down .
this effect is compounded when @xmath348 , since lighter quark masses force sharper endpoint behavior in the wavefunction .
although the bsw solution ameliorates this behavior , as @xmath13 is increased one eventually faces the problem of attempting to represent a function with only a very small region of support in @xmath32 by a finite number of modes with support over the full range @xmath349 $ ] . in practice , we gauge the errors committed through such `` lattice spacing '' effects by computing a given quantity with @xmath350 and noting the amount by which its value shifts if one uses instead @xmath351 , and as expected , such errors become substantial ( as much as a few percent ) by the time one reaches @xmath352 or @xmath353 .
we adopt an intermediate strategy of employing certain exact relations that hold for the t hooft solutions .
to determine the relevant static expectation values , we solve the finite-@xmath13 heavy hadron mass expansion for @xmath356 [ eq .
( [ 50 ] ) ] : @xmath357 neglecting the order term and using the relations [ eq . ( [ 52 ] ) ] @xmath358 , \label{c52}\ ] ] we thus arrive at an equation cubic in @xmath66 that depends on @xmath359 .
we solve it at @xmath360 .
the asymptotic value of the scaled decay constant @xmath361 must also be evaluated at a finite value of @xmath13 , thus including @xmath3-suppressed pieces .
we account for them explicitly using the expansion @xcite [ the first of eqs .
( [ 58 ] ) ] @xmath362}{3m_q } \right ) f^{(n ) } + o\left(\frac{\beta^{5/2}}{m_q^2}\right)\;. \label{c58}\ ] ] we likewise solve this equation for @xmath143 at @xmath360 . turning to the analysis of the sv sum rules eqs .
( [ oscsum1])([oscsum4 ] ) in sec . 4
, we note that their rapid saturation demands an exceptionally high precision in evaluating both the oscillation strengths @xmath125 in the r.h.s . and the expectation values in the l.h.s .
reaching such an accuracy through direct computation seems impossible .
therefore , we use a number of identities to get meaningful results .
first , we employ the expression for @xmath140 in terms of @xmath147 , @xmath363 , and the corresponding decay constants : @xmath364 then we make use of the fact that the discussed sum rules , being completeness sums , are exact when summation includes all excitations ( see ref .
therefore , one has & = & _
= n+1^_k^2 , [ csv1 ] + & = & _ = n+1^(_-_k ) _
k^2 , + & = & _ = n+1^(_-_k)^2 _ k^2 , + & = & _ = n+1^(_-_k)^3 _ k^2 .
the sums on the r.h.s . can be accurately evaluated since the higher contributions fall off in magnitude very quickly . in practice , we truncate the sum at @xmath365 .
a similar approach was used to evaluate the duality - violating difference @xmath366 as a function of @xmath13 .
we use the exact relation @xcite [ eqs .
( [ 124 ] ) , ( [ 130 ] ) , ( [ 141 ] ) ] @xmath367 and therefore , @xmath368 the summation runs over all final excited states kinematically _ forbidden _ in the decay .
once again , the sum converges rapidly and is dominated by the lowest couple of states .
n. uraltsev , preprint und - hep-98-big1 , in _ heavy flavour physics : a probe of nature s grand design _ , proceedings of the international school of physics `` enrico fermi , '' course cxxxvii , varenna , july 7 - 18 , 1997 , eds .
i. bigi and l. moroni ( ios press , amsterdam , 1998 ) p. 329
[ hep - ph/9804275 ] .
0.56 & 1.21918 & 0.7300 & 0.280 & 0.017 & @xmath374 + 1.0 & 1.24633 & 0.9534 & 0.432 & 0.048 & @xmath375 + 3.0 & 1.28764 & 1.4210 & 0.791 & 0.211 & @xmath376 + 5.0 & 1.29904 & 1.6061 & 0.944 & 0.333 & @xmath377 + 7.0 & 1.30423 & 1.7503 & 1.029 & 0.417 & @xmath378 + 10.0 & 1.30820 & 1.7901 & 1.102 & 0.500 & @xmath379 + 15.0 & 1.31131 & 1.8633 & 1.166 & 0.582 & @xmath380 + 25.0 & 1.31375 & 1.9271 & 1.222 & 0.661 & @xmath381 + 35.0 & 1.31475 & 1.9560 & 1.248 & 0.700 & @xmath382 + 50.0 & 1.31545 & 1.9783 & 1.268 & 0.732 & @xmath383 + + 0.56 & 2.82831 & 0.0000 & 0.280 & 0.032 & @xmath384 + 1.0 & 2.77888 & 0.0922 & 0.476 & 0.091 & @xmath385 + 3.0 & 2.66569 & 0.4429 & 1.094 & 0.457 & @xmath386 + 5.0 & 2.61977 & 0.6427 & 1.437 & 0.775 & @xmath387 + 7.0 & 2.59522 & 0.7648 & 1.649 & 1.014 & @xmath388 + 10.0 & 2.57436 & 0.8775 & 1.848 & 1.267 & @xmath389 + 15.0 & 2.55644 & 0.9812 & 2.033 & 1.529 & @xmath390 + 25.0 & 2.54088 & 1.0765 & 2.205 & 1.797 & @xmath391 + 35.0 & 2.53382 & 1.1213 & 2.287 & 1.933 & @xmath392 + 50.0 & 2.52833 & 1.1566 & 2.351 & 2.046 & @xmath393 | we address numerical aspects of local quark - hadron duality using the example of the exactly solvable t hooft model , two - dimensional qcd with @xmath0 .
the primary focus of these studies is total semileptonic decay widths relevant for extracting @xmath1 and @xmath2 .
we compare the exact channel - by - channel sum of exclusive modes to the corresponding rates obtained in the standard @xmath3 expansion arising from the operator product expansion .
an impressive agreement sets in unexpectedly early , immediately after the threshold for the first hadronic excitation in the final state . yet even at higher energy release it is possible to discern the seeds of duality - violating oscillations .
we find the `` small velocity '' sum rules to be exceptionally well saturated already by the first excited state .
we also obtain a convincing degree of duality in the differential distributions and in an analogue of @xmath4 . finally , we discuss possible lessons for semileptonic decays of actual heavy quarks in qcd .
pacs numbers : 12.38.aw , 11.10.kk , 13.20.-v 15.2 true cm 22.0 true cm 0 cm 0 cm 0.15 true in 0.4 true in 0.25 true in .7ex
.7ex 255=255 by 60 255 by-60255 by jlab - thy-00 - 21 + und - hep-00-big05 richard f. lebed@xmath5 and nikolai g.
uraltsev@xmath6 + ( june , 2000 ) |
[ sec : intro ] this paper initiates the study of the minimum number of distinct edge - lengths in a drawing of a given graph is denoted by @xmath8 , and its edge set by @xmath9 . a graph with @xmath2 vertices , @xmath10 edges and maximum degree at
most @xmath6 is an @xmath2-vertex , @xmath10-edge , degree-@xmath6 graph .
a graph in which every vertex has degree @xmath6 is _
@xmath6-regular_. for @xmath11 , let @xmath12 $ ] be the subgraph of @xmath0 induced by @xmath13 , and let @xmath14 $ ] . for each vertex @xmath15 ,
let @xmath16 .
standard notation is used for graphs : complete graphs @xmath17 , complete bipartite graphs @xmath18 , paths @xmath19 , and cycles @xmath20 . a graph
@xmath21 is a
_ minor _ of a graph @xmath0 if @xmath21 can be obtained from a subgraph of @xmath0 by contracting edges . throughout the paper
, @xmath22 is a positive constant .
of course , different occurrences of @xmath22 might denote different constants . ] .
a _ degenerate drawing _ of a graph @xmath0 is a function that maps the vertices of @xmath0 to distinct points in the plane , and maps each edge @xmath23 of @xmath0 to the open straight - line segment joining the two points representing @xmath24 and @xmath25 .
a _ drawing _ of @xmath0 is a degenerate drawing of @xmath0 in which the image of every edge of @xmath0 is disjoint from the image of every vertex of @xmath0 .
that is , no vertex intersects the interior of an edge . in what follows
, we often make no distinction between a vertex or edge in a graph and its image in a drawing .
the _ distance - number _ of a graph @xmath0 , denoted by @xmath26 , is the minimum number of distinct edge - lengths in a drawing of @xmath0 .
the _ degenerate distance - number _ of @xmath0 , denoted by @xmath27 , is the minimum number of distinct edge - lengths in a degenerate drawing of @xmath0 . clearly , @xmath28 for every graph @xmath0 . furthermore , if @xmath21 is a subgraph of @xmath0 then @xmath29 and @xmath30 .
[ sec : motivation ] the degenerate distance - number and distance - number of a graph generalise various concepts in combinatorial geometry , which motivates their study .
a famous problem raised by @xcite asks for the minimum number of distinct distances determined by @xmath2 points in the plane@xmath31 of the monograph by @xcite . ] . this problem is equivalent to determining the degenerate distance - number of the complete graph @xmath17 .
we have the following bounds on @xmath32 , where the lower bound is due to @xcite ( building on recent advances by @xcite , @xcite , and @xcite ) , and the upper bound is due to @xcite . [
lem : degencomplete ] the degenerate distance - number of @xmath17 satisfies @xmath33 observe that no three points are collinear in a ( non - degenerate ) drawing of @xmath17 . thus @xmath34 equals the minimum number of distinct distances determined by @xmath2 points in the plane with no three points collinear .
this problem was considered by szemerdi ( see theorem 13.7 in @xcite ) , who proved that every such point set contains a point from which there are at least @xmath35 distinct distances to the other points .
thus we have the next result , where the upper bound follows from the drawing of @xmath17 whose vertices are the points of a regular @xmath2-gon , as illustrated in ( a ) .
( a ) a drawing of @xmath36 with five edge - lengths , and ( b ) a drawing of @xmath37 with three edge - lengths . ]
[ lem : complete ] the distance - number of @xmath17 satisfies @xmath38 note that lemmas [ lem : degencomplete ] and [ lem : complete ] show that for every sufficiently large complete graph , the degenerate distance - number is strictly less than the distance - number .
indeed , @xmath39 .
degenerate distance - number generalises another concept in combinatorial geometry .
the _ unit - distance graph _ of a set @xmath13 of points in the plane has vertex set @xmath13 , where two vertices are adjacent if and only if they are at unit - distance ; see @xcite for example .
the famous hadwiger - nelson problem asks for the maximum chromatic number of a unit - distance graph .
every unit - distance graph @xmath0 has @xmath40 . but the converse is not true , since a degenerate drawing allows non - adjacent vertices to be at unit - distance .
gives an example of a graph @xmath0 with @xmath41 that is not a unit - distance graph .
in general , @xmath40 if and only if @xmath0 is isomorphic to a subgraph of a unit - distance graph . a graph with distance - number @xmath42 that is not a unit - distance graph . in every mapping of the vertices to distinct points in the plane with unit - length edges , @xmath24 and @xmath25 are at unit - distance . ]
the maximum number of edges in a unit - distance graph is an old open problem .
the best construction , due to @xcite , gives an @xmath2-vertex unit - distance graph with @xmath43 edges .
the best upper bound on the number of edges is @xmath44 , due to @xcite .
( @xcite found a simple proof for this upper bound based on the crossing lemma . )
more generally , many recent results in the combinatorial geometry literature provide upper bounds on the number of times the @xmath45 most frequent inter - point distances can occur between a set of @xmath2 points .
such results are equivalent to upper bounds on the number of edges in an @xmath2-vertex graph with degenerate distance number @xmath45 .
this suggests the following extremal function .
let @xmath46 be the maximum number of edges in an @xmath2-vertex graph @xmath0 with @xmath47 .
since every graph @xmath0 is the union of @xmath48 subgraphs of unit - distance graphs , the above result by @xcite implies : [ lem : numedges ] @xmath49 equivalently , the distance - numbers of every @xmath2-vertex @xmath10-edge graph @xmath0 satisfy @xmath50 results by @xcite ( building on recent advances by @xcite , @xcite , and @xcite ) imply : [ lem : extremal ] @xmath51 equivalently , the distance - numbers of every @xmath2-vertex @xmath10-edge graph @xmath0 satisfy @xmath52 note that lemma [ lem : extremal ] improves upon lemma [ lem : numedges ] whenever @xmath53 . also note that lemma [ lem : extremal ] implies the lower bound in lemma [ lem : complete ] .
the above results give properties of various graphs defined with respect to the inter - point distances of a set of points in the plane .
this paper , which is more about graph drawing than combinatorial geometry , reverses this approach , and asks for a drawing of a given graph with few inter - point distances .
our first results provide some general families of graphs , namely trees and graphs with no @xmath1-minor , that are unit - distance graphs ( ) . here
@xmath1 is the graph obtained from @xmath54 by deleting one edge .
then we give bounds on the distance - numbers of complete bipartite graphs ( ) .
our main results concern graphs of bounded degree ( ) .
we prove that for all @xmath55 there are degree-@xmath6 graphs with unbounded distance - number .
moreover , for @xmath56 we prove a polynomial lower bound on the distance - number ( of some degree-@xmath6 graph ) that tends to @xmath7 for large @xmath6 . on the other hand , we prove that graphs with bounded degree and bounded treewidth have distance - number in @xmath3 .
note that bounded treewidth alone does not imply a logarithmic bound on distance - number since @xmath57 has treewidth @xmath4 and degenerate distance - number @xmath58 ( see ) .
then we establish an upper bound on the distance - number in terms of the bandwidth ( ) .
then we consider the distance - number of the cartesian product of graphs ( ) .
we conclude in with a discussion of open problems related to distance - number .
graph invariants related to distances in higher dimensions have also been studied .
@xcite defined the _ dimension _ of a graph @xmath0 , denoted by @xmath59 , to be the minimum integer @xmath45 such that @xmath0 has a degenerate drawing in @xmath60 with straight - line edges of unit - length .
they proved that @xmath61 , the dimension of the @xmath2-cube is @xmath4 for @xmath62 , the dimension of the peterson graph is @xmath4 , and @xmath63 for every graph @xmath0 .
( here @xmath64 is the _ chromatic number _ of @xmath0 . )
the dimension of complete @xmath65-partite graphs and wheels were determined by @xcite .
the _ unit - distance graph _ of a set @xmath66 has vertex set @xmath13 , where two vertices are adjacent if and only if they are at unit - distance .
thus @xmath67 if and only if @xmath0 is isomorphic to a subgraph of a unit - distance graph in @xmath60 .
@xcite proved for all @xmath45 there is a finite bipartite graph ( which thus has dimension at most @xmath68 ) that is not a unit - distance graph in @xmath60 .
this highlights the distinction between dimension and unit - distance graphs .
@xcite also proved that every finite graph with maximum degree @xmath6 is a unit - distance graph in @xmath69 , which was improved to @xmath70 by @xcite .
these results are in contrast to our result that graphs of bounded degree have arbitrarily large distance - number . a graph is @xmath45-_realizable _ if , for every mapping of its vertices to ( not - necessarily distinct ) points in @xmath71 with @xmath72 , there exists such a mapping in @xmath60 that preserves edge - lengths .
for example , @xmath73 is @xmath4-realizable but not @xmath42-realizable . @xcite and @xcite proved that a graph is @xmath4-realizable if and only if it has treewidth at most @xmath4 .
they also characterized the @xmath65-realizable graphs as those with no @xmath74-minor and no @xmath75-minor .
[ sec : unitdistancegraphs ] this section shows that certain families of graphs are unit - distance graphs . the proofs
are based on the fact that two distinct circles intersect in at most two points .
we start with a general lemma .
a graph @xmath0 is obtained by _ pasting _ subgraphs @xmath76 and @xmath77 on a cut - vertex @xmath24 of @xmath0 if @xmath78 and @xmath79 .
[ lem : pasting ] let @xmath0 be the graph obtained by pasting subgraphs @xmath76 and @xmath77 on a vertex @xmath24 . then : + ( a ) if @xmath80 then @xmath40 , and + ( b ) if @xmath81 then @xmath82 .
we prove part ( b ) .
part ( a ) is easier .
let @xmath83 be a drawing of @xmath84 with unit - length edges .
translate @xmath85 so that @xmath24 appears in the same position in @xmath86 and @xmath85 .
a rotation of @xmath85 about @xmath24 is _ bad _ if its union with @xmath86 is not a drawing of @xmath0 .
that is , some vertex in @xmath85 coincides with the closure of some edge of @xmath86 , or vice versa . since @xmath0 is finite , there are only finitely many bad rotations .
since there are infinitely many rotations , there exists a rotation that is not bad .
that is , there exists a drawing of @xmath0 with unit - length edges .
we have a similar result for unit - distance graphs .
[ lem : pastingunitdistance ] let @xmath76 and @xmath77 be unit - distance graphs .
let @xmath0 be the ( abstract ) graph obtained by pasting @xmath76 and @xmath77 on a vertex @xmath24
. then @xmath0 is isomorphic to a unit - distance graph .
the proof is similar to the proof of lemma [ lem : pasting ] , except that we must ensure that the distance between vertices in @xmath87 and vertices in @xmath88 ( which are not adjacent ) is not @xmath42 . again
this will happen for only finitely many rotations .
thus there exists a rotation that works .
since every tree can be obtained by pasting a smaller tree with @xmath89 , lemma [ lem : pastingunitdistance ] implies that every tree is a unit - distance graph .
the following is a stronger result .
[ lem : trees ] every tree @xmath90 has a crossing - free drawing in the plane such that two vertices are adjacent if and only if they are unit - distance apart . for a point @xmath91 in the plane ,
let @xmath92 be the ray from @xmath24 to @xmath93 .
we proceed by induction on @xmath2 with the following hypothesis : every tree @xmath90 with @xmath2 vertices has the desired drawing , such that the vertices have distinct @xmath94-coordinates , and for each vertex @xmath95 , the ray @xmath96 does not intersect @xmath90 .
the statement is trivially true for @xmath97 . for @xmath98 ,
let @xmath24 be a leaf of @xmath90 with parent @xmath99 .
by induction , @xmath100 has the desired drawing .
let @xmath25 be a vertex of @xmath100 , such that no vertex has its @xmath94-coordinate between @xmath101 and @xmath102 .
thus the drawing of @xmath100 does not intersect the open region @xmath103 of the plane bounded by the two rays @xmath104 and @xmath105 , and the segment @xmath106 .
let @xmath107 be the intersection of @xmath103 with the unit - circle centred at @xmath99 .
thus @xmath107 is a circular arc .
place @xmath24 on @xmath107 , so that the distance from @xmath24 to every vertex except @xmath99 is not @xmath42 .
this is possible since @xmath107 is infinite , and there are only finitely many excluded positions on @xmath107 ( since @xmath107 intersects a unit - circle centred at a vertex except @xmath99 in at most two points ) .
since there are no elements of @xmath100 in @xmath103 , there are no crossings in the resulting drawing and the induction invariants are maintained for all vertices of @xmath90 . ]
recall that @xmath1 is the graph obtained from @xmath54 by deleting one edge .
every @xmath1-minor - free graph @xmath0 has a drawing such that vertices are adjacent if and only if they are unit - distance apart . in particular , @xmath0 is isomorphic to a unit - distance graph and @xmath108 .
clearly we can assume that @xmath0 is connected .
if @xmath0 is not @xmath4-connected then @xmath0 can be obtained by pasting smaller graphs with no @xmath1-minor on a single vertex .
thus by lemma [ lem : pastingunitdistance ] , it suffices to prove that every @xmath4-connected @xmath1-minor - free graph is a unit - distance graph .
below we prove that every @xmath4-connected graph @xmath0 with no @xmath1-minor is a cycle .
the result follows since @xmath20 is a unit - distance graph ( draw a regular @xmath2-gon ) .
suppose on the contrary that @xmath0 has a vertex @xmath24 of degree at least @xmath65 .
let @xmath109 be the neighbours of @xmath24 .
so there is an @xmath110-path @xmath111 avoiding @xmath24 ( since @xmath0 is @xmath4-connected ) and avoiding @xmath112 ( since @xmath0 is @xmath1-minor free ) .
similarly , there is an @xmath113-path @xmath114 avoiding @xmath24 . if @xmath115 is the only vertex in both @xmath111 and @xmath114 , then the cycle @xmath116 plus the edge @xmath117 is a subdivision of @xmath1 .
now assume that @xmath111 and @xmath114 intersect at some other vertex .
let @xmath118 be the first vertex on @xmath111 starting at @xmath115 that is also in @xmath114 .
then the cycle @xmath119 plus the sub - path of @xmath111 between @xmath115 and @xmath118 is a subdivision of @xmath1 .
hence @xmath0 has no vertex of degree at least @xmath65 .
therefore @xmath0 is a cycle , as desired .
[ sec : completebipartite ] this section considers the distance - numbers of complete bipartite graphs @xmath18 . since @xmath120 is a tree , @xmath121 by lemma [ lem : trees ] .
the next case , @xmath57 , is also easily handled .
[ lem : k2n ] the distance - numbers of @xmath57 satisfy @xmath122 let @xmath123 with colour classes @xmath124 and @xmath125 , where @xmath126 .
we first prove the lower bound .
consider a degenerate drawing of @xmath0 with @xmath27 edge - lengths .
the vertices in @xmath125 lie on the intersection of @xmath27 concentric circles centered at @xmath24 and @xmath27 concentric circles centered at @xmath25 .
since two distinct circles intersect in at most two points , @xmath127 . thus @xmath128 for the upper bound , position @xmath24 at @xmath129 and @xmath25 at @xmath130
. as illustrated in , draw @xmath131 circles centered at each of @xmath24 and @xmath25 with radii ranging strictly between @xmath42 and @xmath4 , such that the intersections of the circles together with @xmath24 and @xmath25 define a set of points with no three points collinear .
( this can be achieved by choosing the radii iteratively , since for each circle @xmath132 , there are finitely many forbidden values for the radius of @xmath132 . )
each pair of non - concentric circles intersect in two points .
thus the number of intersection points is at least @xmath2 . placing the vertices of @xmath125 at these intersection points results in a drawing of @xmath57 with @xmath131 edge - lengths .
illustration for the proof of lemma [ lem : k2n ] . ]
now we determine @xmath133 to within a constant factor .
[ lem : k3n ] the degenerate distance - number of @xmath134 satisfies @xmath135 the lower bound follows from lemma [ lem : k2n ] since @xmath57 is a subgraph of @xmath134 .
now we prove the upper bound .
let @xmath107 and @xmath125 be the colour classes of @xmath134 , where @xmath136 and @xmath126 .
place the vertices in @xmath107 at @xmath129 , @xmath137 , and @xmath130 .
. for @xmath139 $ ] , let @xmath140 note that @xmath141 .
let @xmath142 be the circle centred at @xmath129 with radius @xmath143 .
for @xmath144 $ ] , let @xmath145 be the circle centred at @xmath130 with radius @xmath146 .
observe that each pair of circles @xmath142 and @xmath145 intersect in exactly two points .
place the vertices in @xmath125 at the intersection points of these circles .
this is possible since @xmath147 .
let @xmath148 and @xmath149 be the two points where @xmath142 and @xmath145 intersect .
thus @xmath150 and @xmath151 .
it follows that @xmath152 thus @xmath153 .
that is , the distance from @xmath148 to @xmath137 equals @xmath154 which is the same distance from @xmath149 to @xmath137 .
thus the distance from each vertex in @xmath125 to @xmath137 is one of @xmath155 values ( determined by @xmath156 ) .
the distance from each vertex in @xmath125 to @xmath129 and to @xmath130 is one of @xmath45 values .
hence the degenerate distance - number of @xmath134 is at most @xmath157 .
illustration for the proof of lemma [ lem : k3n ] . ]
now consider the distance - number of a general complete bipartite graph .
[ lem : kmn ] for all @xmath158 , the distance - numbers of @xmath18 satisfy @xmath159 in particular , @xmath160 the lower bounds follow from lemma [ lem : extremal ] .
for the upper bound on @xmath161 , position the vertices on a regular @xmath162-gon @xmath163 alternating between the colour classes , as illustrated in ( b ) . in the resulting drawing of @xmath164 ,
the number of edge - lengths is @xmath165 . since @xmath166 is an edge if and only if @xmath156 is odd , the number of edge - lengths is @xmath167 .
the upper bound on @xmath168 follows since @xmath169 is a subgraph of @xmath164 .
[ sec : degree ] lemma [ lem : k2n ] implies that if a graph has two vertices with many common neighbours then its distance - number is necessarily large .
thus it is natural to ask whether graphs of bounded degree have bounded distance - number .
this section provides a negative answer to this question .
[ sec:7 ] this section proves that for all @xmath56 there are @xmath6-regular graphs with unbounded distance - number .
moreover , the lower bound on the distance - number is polynomial in the number of vertices . the basic idea of the proof is to show that there are more @xmath6-regular graphs than graphs with bounded distance - number
. it will be convenient to count labelled graphs .
let @xmath171 denote the family of labelled @xmath6-regular @xmath2-vertex graphs .
let @xmath172 denote the family of labelled @xmath2-vertex @xmath10-edge graphs with degenerate distance - number at most @xmath45 .
our results follow by comparing a lower bound on @xmath173 with an upper bound on @xmath174 with @xmath175 , which is the number of edges in a @xmath6-regular @xmath2-vertex graph .
the lower bound in question is known . in particular , the first asymptotic bounds on the number of labelled @xmath6-regular @xmath2-vertex graphs were independently determined by @xcite and @xcite . @xcite further refined these results .
we will use the following simple lower bound derived by @xcite from the result of @xcite .
[ lem : numberregular ] for all integers @xmath176 and @xmath177 , the number of labelled @xmath6-regular @xmath2-vertex graphs satisfies @xmath178 the proof of our upper bound on @xmath174 uses the following special case of the milnor - thom theorem by @xcite .
let @xmath179 be a sequence of polynomials on @xmath99 variables over @xmath180 .
the _ zero - pattern _ of @xmath181 at @xmath182 is the set @xmath183 .
[ lem : zeros ] let @xmath184 be a sequence of polynomials of degree at most @xmath185 on @xmath186 variables over @xmath180
. then the number of zero - patterns of @xmath181 is at most @xmath187 .
recall that @xmath46 is the maximum number of edges in an @xmath2-vertex graph @xmath0 with @xmath47 .
bounds on this function are given in lemmas [ lem : numedges ] and [ lem : extremal ] .
our upper bound on @xmath174 is expressed in terms of @xmath46 .
[ lem : num ] the number of labelled @xmath2-vertex @xmath10-edge graphs with @xmath47 satisfies @xmath188 where @xmath189 is the base of the natural logarithm . let @xmath190 for every @xmath191 .
for every @xmath191 , there is a point set @xmath192 and a set of edge - lengths @xmath193 such that @xmath0 has a degenerate drawing in which each vertex @xmath194 is represented by the point @xmath195 and the length of each edge in @xmath9 is in @xmath196 .
fix one such degenerate drawing of @xmath0 . for all @xmath197 with @xmath198 and @xmath199 , and for every graph @xmath191 , define @xmath200 consider @xmath201 to be a set of @xmath202 degree-@xmath4 polynomials on the set of @xmath203 variables @xmath204 .
observe that @xmath205 if and only if the distance between vertices @xmath194 and @xmath206 in the + degenerate drawing of @xmath0 is @xmath207 .
@xmath208 recall the well - known fact that @xmath209 . by lemma [ lem : zeros ] with @xmath210 , @xmath211 and @xmath212 , the number of zero - patterns determined by @xmath181 is at most @xmath213 fix a zero - pattern @xmath214 of @xmath215 .
let @xmath216 be the set of graphs @xmath0 in @xmath172 such that @xmath214 is the zero - pattern of @xmath181 evaluated at @xmath0 . to bound @xmath174 we now bound @xmath217 .
let @xmath218 be the graph with vertex set @xmath219 and edge set @xmath220 where @xmath221 if and only if @xmath222 for some @xmath223 .
consider a degenerate drawing of an arbitrary graph @xmath223 on the point set @xmath224 .
by @xmath208 , @xmath224 and @xmath196 define a degenerate drawing of @xmath21 with @xmath45 edge - lengths .
thus @xmath225 and by assumption , @xmath226 . since every graph in @xmath216 is a subgraph of @xmath218 , @xmath227 .
therefore , @xmath228 as required . by comparing the lower bound in lemma [ lem : numberregular ] and the upper bound in lemma [ lem : num ]
we obtain the following result . [
lem : polybound ] suppose that for some real numbers @xmath229 and @xmath230 with @xmath231 and @xmath232 , @xmath233 then for every integer @xmath234 , for all @xmath235 , and for all sufficiently large @xmath236 , there exists a @xmath6-regular @xmath2-vertex graph @xmath0 with degenerate distance - number @xmath237 in this proof , @xmath229 , @xmath230 , @xmath6 and @xmath238 are fixed numbers satisfying the assumptions of the lemma .
let @xmath45 be the maximum degenerate distance number of a graph in @xmath239 .
the result will follow by showing that for all sufficiently large @xmath236 , @xmath240 by the definition of @xmath45 , and since every @xmath6-regular @xmath2-vertex graph has @xmath241 edges , every graph in @xmath171 is also in @xmath242 . by lemma [ lem : numberregular ] with @xmath177 , and by lemma [ lem : num ] , @xmath243 since @xmath244 , and since @xmath45 is a function of @xmath2 , there is a constant @xmath22 such that @xmath245 for sufficiently large @xmath2 .
thus ( and since @xmath209 ) , @xmath246 hence @xmath247 by lemma [ lem : complete ] , @xmath248 , implying @xmath249 for all large @xmath250 .
thus @xmath251 hence @xmath252 observe that @xmath253 for all large @xmath254 .
thus @xmath255 hence @xmath256 as required .
we can now state the main results of this section .
by lemma [ lem : numedges ] , the conditions of lemma [ lem : polybound ] are satisfied with @xmath257 and @xmath258 ; thus together they imply : [ thm : degree-7 ] for every integer @xmath170 , for all @xmath235 , and for all sufficiently large @xmath254 , there exists a @xmath6-regular @xmath2-vertex graph @xmath0 with degenerate distance - number @xmath259 by lemma [ lem : extremal ] , the conditions of lemma [ lem : polybound ] are satisfied with @xmath260 and @xmath261 ; thus together they imply : [ thm : degree-8 ] for every integer @xmath262 , for all @xmath235 , and for all sufficiently large @xmath254 , there exists a @xmath6-regular @xmath2-vertex graph @xmath0 with degenerate distance - number @xmath263 note that the bound given in theorem [ thm : degree-8 ] is better than the bound in theorem [ thm : degree-7 ] for @xmath264 .
[ sec:5 ] theorem [ thm : degree-7 ] shows that for @xmath56 and for sufficiently large @xmath2 , there is an @xmath2-vertex degree-@xmath6 graph whose degenerate distance - number is at least polynomial in @xmath2 .
we now prove that the degenerate distance - number of degree-@xmath5 graphs can also be arbitrarily large .
however , the lower bound we obtain in this case is polylogarithmic in @xmath2 . the proof is inspired by an analogous proof about the slope - number of degree-@xmath5 graphs , due to @xcite .
[ thm : degree-5 ] for all @xmath266 , there is a degree-@xmath5 graph @xmath0 with degenerate distance - number @xmath267 .
consider the following degree-@xmath5 graph @xmath0 .
for @xmath268 , let @xmath269 be the graph with vertex set @xmath270 and edge set @xmath271 .
let @xmath272 .
no pair of vertices in @xmath13 are adjacent in @xmath269 , and @xmath273 is even
. let @xmath274 denote the set of all perfect matchings on @xmath13 .
for each perfect matching @xmath275 , let @xmath276 .
as illustrated in , let @xmath0 be the disjoint union of all the @xmath277 .
thus the number of connected components of @xmath0 is @xmath278 , which is at least @xmath279 by lemma [ lem : numberregular ] with @xmath280 .
here we consider perfect matchings to be @xmath42-regular graphs .
( it is remarkable that even with @xmath280 , lemma [ lem : numberregular ] gives such an accurate bound , since the actual number of matchings in @xmath13 is @xmath281 ignoring lower order additive terms , let @xmath282 be the number of perfect matchings of @xmath283 $ ] .
here we determine the asymptotics of @xmath284 . in every such matching , @xmath2 is matched with some number in @xmath285 $ ] , and the remaining matching is isomorphic to a perfect matching of @xmath286 $ ] .
every matching obtained in this way is distinct .
thus @xmath287 , where @xmath288 .
hence @xmath289 , where @xmath290 is the double factorial function .
. thus @xmath292 by stirling s approximation . ] . )
the graph @xmath0 with @xmath293 . ]
suppose , for the sake of contradiction , that for some constant @xmath45 , for all @xmath294 such that @xmath268 , @xmath0 has a degenerate drawing @xmath295 with at most @xmath45 edge - lengths .
label the edges of @xmath0 that are in the copies of @xmath269 by their length in @xmath295 .
let @xmath296 be the label of the edge @xmath166 in the copy of @xmath269 in the component @xmath277 of @xmath0 .
this defines a labelling of the components of @xmath0 .
since @xmath269 has @xmath297 edges and each edge in @xmath269 receives one of @xmath45 labels , there are at most @xmath298 distinct labellings of the components of @xmath0 .
let @xmath299 be the degenerate drawing of @xmath277 obtained from @xmath295 by a translation and rotation so that @xmath300 is at @xmath137 and @xmath301 is at @xmath302 .
we say that two components @xmath303 and @xmath304 of @xmath0 _ determine the same set of points _
if for all @xmath305 $ ] , the vertex @xmath306 in @xmath307 is at the same position as the vertex @xmath306 in @xmath308 .
partition the components of @xmath0 into the minimum number of parts such that all the components in each part have the same labelling and determine the same set of points .
observe that two components of @xmath0 with the same labelling do not necessarily determine the same set of points .
however , the number of point sets determined by the components with a given labelling can be bounded as follows . for each component @xmath277 of @xmath0
, @xmath300 is at @xmath137 and @xmath301 is at @xmath309 in @xmath299 .
thus for a fixed labelling , the positions of @xmath300 and @xmath301 in @xmath299 are determined .
now for @xmath310 , in each component @xmath277 , the vertex @xmath306 is positioned in @xmath299 at the intersection of the circle of radius @xmath311 centered at @xmath312 and the circle of radius @xmath313 centered at @xmath314 .
thus there are at most two possible locations for @xmath306 ( for a fixed labelling ) .
hence the components with the same labelling determine at most @xmath315 distinct points sets .
therefore the number of parts in the partition is at most @xmath316 . finally , we bound the number of components in each part , @xmath103 , of the partition .
let @xmath317 be the graph with vertex set @xmath318 where @xmath319 if and only if @xmath320 for some component @xmath321 .
since the graphs in @xmath103 determine the same set of points , the union of the degenerate drawings @xmath299 , over all @xmath321 , determines a degenerate drawing of @xmath317 with @xmath45 edge - lengths .
thus @xmath322 and by lemma [ lem : numedges ] , @xmath323 for some constant @xmath324 .
every component in @xmath103 is a subgraph of @xmath317 , and any two components in @xmath103 differ only by the choice of a matching on @xmath13 .
each such matching has @xmath325 edges .
thus the number of components of @xmath0 in @xmath103 is at most @xmath326 hence @xmath327 , and by the lower bound on @xmath278 from the start of the proof , @xmath328 the desired contradiction follows for all @xmath329 .
[ sec : tw ] this section proves a logarithmic upper bound on the distance - number of graphs with bounded degree and bounded treewidth .
treewidth is an important parameter in robertson and seymour s theory of graph minors and in algorithmic complexity ( see the surveys @xcite ) .
it can be defined as follows .
a graph @xmath0 is a _
@xmath330-tree _ if either @xmath331 , or @xmath0 has a vertex @xmath24 whose neighbourhood is a clique of order @xmath330 and @xmath332 is a @xmath330-tree .
for example , every @xmath42-tree is a tree and every tree is a @xmath42-tree .
then the _ treewidth _ of a graph @xmath0 is the minimum integer @xmath330 for which @xmath0 is a subgraph of a @xmath330-tree .
the _ pathwidth _ of @xmath0 is the minimum @xmath330 for which @xmath0 is a subgraph of an interval is an _ interval graph _ if each vertex @xmath15 can be assigned an interval @xmath333 such that @xmath334 if and only if @xmath335 . ]
graph with no clique of order @xmath336 .
note that an interval graph with no @xmath337-clique is a special case of a @xmath330-tree , and thus the treewidth of a graph is at most its pathwidth .
lemma [ lem : trees ] shows that ( @xmath42-)trees have bounded distance - number .
however , this is not true for @xmath4-trees since @xmath57 has treewidth ( and pathwidth ) at most @xmath4 . by theorem [ thm : degree-8 ] , there are @xmath2-vertex graphs of bounded degree with distance - number approaching @xmath7 . on the other hand ,
no polynomial lower bound holds for graphs of bounded degree _ and _ bounded treewidth , as shown in the following theorem .
[ thm : tw ] let @xmath0 be a graph with @xmath2 vertices , maximum degree @xmath6 , and treewidth @xmath330
. then the distance - number of @xmath0 satisfies @xmath338 to prove theorem [ thm : tw ] we use the following lemma , the proof of which is readily obtained by inspecting the proof of lemma 8 in @xcite .
an _ @xmath21-partition _ of a graph @xmath0 is a partition of @xmath8 into vertex sets @xmath339 such that @xmath21 is the graph with vertex set @xmath340 where @xmath341 if and only if there exists @xmath342 and @xmath343 such that @xmath344 . the _ width _ of an @xmath21-partition is @xmath345 .
[ lem : helper ] let @xmath21 be a graph admitting a drawing @xmath295 with @xmath346 distinct edge - slopes and @xmath347 distinct edge - lengths .
let @xmath0 be a graph admitting an @xmath21-partition of width @xmath25 .
then the distance - number of @xmath0 satisfies @xmath348 the general approach is to start with @xmath295 and then replace each vertex of @xmath21 by a sufficiently scaled down and appropriate rotated copy of the drawing of @xmath349 on a regular @xmath25-gon .
the only difficulty is choosing the rotation and the amount by which to scale the @xmath25-gons so that we obtain a ( non - degenerate ) drawing of @xmath0 .
refer to @xcite for the full proof .
let @xmath25 be the minimum width of a @xmath90-partition of @xmath0 in which @xmath90 is a tree .
the best known upper bound is @xmath350 , which was obtained by @xcite using a minor improvement to a similar result by an anonymous referee of the paper by @xcite . for each vertex @xmath351 , there are at most @xmath352 edges of @xmath0 incident to vertices mapped to @xmath115 .
hence we can assume that @xmath90 is a forest with maximum degree @xmath352 , as otherwise there is an edge of @xmath90 with no edge of @xmath0 mapped to it , in which case the edge of @xmath90 can be deleted .
similarly , @xmath90 has at most @xmath2 vertices .
@xcite proved that @xmath90 has pathwidth at most @xmath353 ; see @xcite .
@xcite proved that every tree @xmath90 with pathwidth @xmath354 has a drawing with @xmath355 slopes and @xmath356 edge - lengths .
thus @xmath90 has a drawing with at most @xmath357 slopes and at most @xmath358 edge - lengths . by lemma [ lem : helper ] ,
@xmath359 which is in @xmath360 .
every @xmath2-vertex graph with bounded degree and bounded treewidth has distance - number @xmath3 .
since a path has a drawing with one slope and one edge - length , lemma [ lem : helper ] with @xmath361 implies that every graph @xmath0 with a @xmath111-partition of width @xmath330 for some path @xmath111 has distance - number @xmath362 .
[ sec : bandwidth ] this section establishes an upper bound on the distance - number in terms of the bandwidth .
let @xmath0 be a graph .
a _ vertex ordering _ of @xmath0 is a bijection @xmath363 .
the _ width _ of @xmath214 is defined to be @xmath364 .
the _ bandwidth _ of @xmath0 , denoted by , is the minimum width of a vertex ordering of @xmath0 .
the _ cyclic width _ of @xmath214 is defined to be @xmath365 .
cyclic bandwidth _ of @xmath0 , denoted by , is the minimum cyclic width of a vertex ordering of @xmath0 ; see @xcite .
clearly @xmath366 for every graph @xmath0 .
[ lem : bandwidth ] for every graph @xmath0 , @xmath367 given a vertex ordering @xmath214 of an @xmath2-vertex @xmath0 , position the vertices of @xmath0 on a regular @xmath2-gon in the order @xmath214 .
we obtain a drawing of @xmath0 in which the length of each edge @xmath23 is determined by @xmath368 thus the number of edge - lengths equals @xmath369 which is at most the cyclic width of @xmath214 .
the result follows .
the distance - number of every @xmath2-vertex degree-@xmath6 planar graph @xmath0 satisfies @xmath370 @xcite proved that @xmath371 .
the result follows from lemma [ lem : bandwidth ] .
[ sec : cartesianproducts ] this section discusses the distance - number of cartesian products of graphs . for graphs @xmath0 and @xmath21 , the _ cartesian product _ @xmath372 is the graph with vertex set @xmath373 , where @xmath374 is adjacent to @xmath375 if and only if ( 1 ) @xmath376 and @xmath377 is an edge of @xmath21 , or ( 2 ) @xmath378 and @xmath379 is an edge of @xmath0 .
thus @xmath372 is the grid - like graph with a copy of @xmath0 in each row and a copy of @xmath21 in each column .
type ( 1 ) edges form copies of @xmath21 , and type ( 2 ) edges form copies of @xmath0 . for example
, @xmath380 is the planar grid , and @xmath381 is the toroidal grid .
the cartesian product is associative and thus multi - dimensional products are well defined .
for example , the @xmath45-dimensional product @xmath382 is the @xmath45-dimensional hypercube @xmath383 .
it is well known that @xmath383 is a unit - distance graph .
@xcite proved that the cartesian product operation preserves unit - distance graphs .
that is , if @xmath0 and @xmath21 are unit - distance graphs , then so is @xmath384 , as illustrated in .
the following theorem generalises this result . ] [ fig:333 ] [ thm : cartprod ] for all graphs @xmath0 and @xmath21 , the distance - numbers of @xmath384 satisfy @xmath385 the lower bounds follow since @xmath0 and @xmath21 are subgraphs of @xmath384 .
we prove the upper bound for @xmath386 .
the proof for @xmath387 is simpler .
fix a drawing of @xmath0 with @xmath26 edge - lengths .
let @xmath388 be the coordinates of each vertex @xmath24 of @xmath0 in this drawing .
fix a drawing of @xmath21 with @xmath389 edge - lengths , scaled so that one edge - length in the drawing of @xmath0 coincides with one edge - length in the drawing of @xmath21 .
let @xmath229 be a real number in @xmath390 .
let @xmath391 be the coordinates of each vertex @xmath25 of @xmath0 in this drawing of @xmath21 rotated by @xmath229 degrees about the origin .
position vertex @xmath374 in @xmath384 at @xmath392 . this mapping preserves edge - lengths .
in particular , the length of a type-(1 ) edge @xmath393 equals the length of the edge @xmath394 in @xmath21 , and the length of a type-(2 ) edge @xmath395 equals the length of the edge @xmath394 in @xmath0 .
thus for each @xmath229 , the mapping of @xmath384 has @xmath396 edge - lengths .
it remains to prove that for some @xmath229 the mapping of @xmath384 is a drawing .
that is , no vertex intersects the closure of an incident edge .
an angle @xmath229 is _ bad _ for a particular vertex / edge pair of @xmath384 if that vertex intersects the closure of that edge in the mapping with rotation @xmath229 .
observe that the trajectory of a vertex @xmath374 of @xmath384 ( taken over all @xmath229 ) is a circle centred at @xmath388 with radius @xmath397 .
now for distinct points @xmath99 and @xmath398 and a line @xmath347 , there are only two angles @xmath229 such that the rotation of @xmath99 around @xmath398 by an angle of @xmath229 contains @xmath347 ( since the trajectory of @xmath99 is a circle that only intersects @xmath347 in two places ) , and there are only two angles @xmath229 such that the rotation of @xmath347 around @xmath398 by an angle of @xmath229 contains @xmath99 .
it follows that there are finitely many bad values of @xmath229 for a particular vertex / edge pair of @xmath384 .
hence there are finitely many bad values of @xmath229 in total .
hence some value of @xmath229 is not bad for every vertex / edge pair in @xmath384 .
hence @xmath399 is a valid drawing of @xmath384 .
note that @xcite proved a result analogous to theorem [ thm : cartprod ] for dimension .
let @xmath400 be the @xmath45-fold cartesian product of a graph @xmath0 .
the same construction used in theorem [ thm : cartprod ] proves the following : [ thm : cartprodsimple ] for every graph @xmath0 and integer @xmath401 , the distance - numbers of @xmath400 satisfy @xmath402
[ sec : conclusion ] we conclude by mentioning some of the many open problems related to distance - number . * what is @xmath34 ?
we conjecture that @xmath403 .
that is , every set of @xmath2 points in general position determine at least @xmath404 distinct distances .
note that @xcite proved this conjecture for points in convex position .
* what is the relationship between distance - number and degenerate distance - number ?
in particular , is there a function @xmath284 such that @xmath405 for every graph @xmath0 ?
* theorems [ thm : degree-7 ] , [ thm : degree-8 ] and [ thm : degree-5 ] establish a lower bound for the distance - number of bounded degree graphs .
but no non - trivial upper bound is known .
do @xmath2-vertex graphs with bounded degree have distance - number in @xmath406 ?
* outerplanar graphs have distance - number in @xmath407 by theorem [ thm : tw ] .
do outerplanar graphs ( with bounded degree ) have bounded ( degenerate ) distance - number ? * non - trivial lower and
upper bounds on the distance - numbers are not known for many other interesting graph families including : degree-@xmath65 graphs , degree-@xmath68 graphs , @xmath4-degenerate graphs with bounded degree , graphs with bounded degree and bounded pathwidth . * as described in , determining the maximum number of times the unit - distance can appear among @xmath2 points in the plane is a famous open problem .
we are unaware if the following apparently simpler tasks have been attempted : determine the maximum number of times the unit - distance can occur among @xmath2 points in the plane such that no three are collinear . similarly , determine the maximum number of edges in an @xmath2-vertex graph @xmath0 with @xmath82 . * determining the maximum chromatic number of unit - distance graphs in @xmath60 is a well - known open problem .
the best known upper bound of @xmath408 is due to @xcite .
exponential lower bounds are known @xcite .
unit - distance graphs in the plane are @xmath409-colourable @xcite , and thus @xmath410 .
can this bound be improved ? * degenerate distance - number is not bounded by any function of dimension since @xmath164 has dimension @xmath68 and unbounded degenerate distance - number . on the other hand , @xmath411 .
is @xmath59 bounded by a polynomial function of @xmath27 ? *
every planar graph has a crossing - free drawing . a long standing open problem involving edge - lengths , due to harborth et al .
@xcite , asks whether every planar graph has a crossing - free drawing in which the length of every edge is an integer .
@xcite recently answered this question in the affirmative for cubic planar graphs .
@xcite extended this question to nonplanar graphs and asked what is the minimum @xmath45 such that a given graph has a crossing - free drawing in @xmath60 with integer edge - lengths .
note that every @xmath2-vertex graph has such a drawing in @xmath412 . *
the _ slope number _ of a graph @xmath0 , denoted by @xmath413 , is the minimum number of edge - slopes over all drawings of @xmath0 .
@xcite established results concerning the slope - number of planar graphs .
@xcite proved that degree-@xmath65 graphs have slope - number at most @xmath5 . on the other hand ,
@xcite and @xcite independently proved that there are @xmath5-regular graphs with arbitrarily large slope number .
moreover , for @xmath56 , @xcite proved that there are @xmath2-vertex degree-@xmath6 graphs whose slope number is at least @xmath414 .
the proofs of these results are similar to the proofs of theorems [ thm : degree-7 ] , [ thm : degree-8 ] and [ thm : degree-5 ] . given that theorem [ thm : tw ] also depends on slopes , it it tempting to wonder if there is a deeper connection between slope - number and distance - number .
for example , is there a function @xmath284 such that @xmath415 and/or @xmath416 for every graph @xmath0 .
note that some dependence on @xmath417 is necessary since @xmath418 but @xmath419 .
b. keszegh , dmtr plvlgyi , jnos pach , and gza tth .
drawing cubic graphs with at most five slopes . in _
graph drawing 2006 _ , vol .
4372 of _ lecture notes in computer science _ , pp . 114125 .
springer , 2007 . | the _ distance - number _ of a graph @xmath0 is the minimum number of distinct edge - lengths over all straight - line drawings of @xmath0 in the plane .
this definition generalises many well - known concepts in combinatorial geometry .
we consider the distance - number of trees , graphs with no @xmath1-minor , complete bipartite graphs , complete graphs , and cartesian products .
our main results concern the distance - number of graphs with bounded degree .
we prove that @xmath2-vertex graphs with bounded maximum degree and bounded treewidth have distance - number in @xmath3 .
to conclude such a logarithmic upper bound , both the degree and the treewidth need to be bounded .
in particular , we construct graphs with treewidth @xmath4 and polynomial distance - number .
similarly , we prove that there exist graphs with maximum degree @xmath5 and arbitrarily large distance - number
. moreover , as @xmath6 increases the existential lower bound on the distance - number of @xmath6-regular graphs tends to @xmath7 . |
one of the interesting and challenging sub - fields of low - dimensional quantum magnetism which has emerged from the wealth of activity directed at improving the understanding of high - temperature superconductors is that of ladder systems @xcite . these consist of @xmath2 parallel , interacting chains of @xmath3 = ions , which can be considered as a spin ladder with @xmath2 legs , and rungs of @xmath2 - 1 bonds .
the ladders have only weak mutual interactions .
a combination of experimental and theoretical efforts has in the last few years produced significant advances in the realization and understanding of the properties of spin ladders , some of which are not at all intuitive .
ladder cuprates emerged first with the discovery by hiroi _
et al._@xcite that in the series of materials sr@xmath4cu@xmath5o@xmath6 it is possible to create two - dimensional , stoichiometric copper oxide planes of composition cu@xmath5 o@xmath6 by removing from the uniform cuo@xmath7 plane parallel , equally - spaced lines of oxygen atoms .
it was pointed out by rice _
et al._@xcite that because these shear defects give rise to only weak , ferromagnetic interactions between neighboring copper spins , the remaining strips of cuo@xmath7 plane will appear as isolated @xmath2 + 1-leg ladders of antiferromagnetically coupled spins .
these authors proposed that the systems should then illustrate the contrasting , and now well - established properties of even- and odd - leg ladders , that the former show a gap to spin excitations ( spin gap ) with consequent exponential spatial decay of correlations , while the latter are gapless with power - law decays .
subsequent susceptibility @xcite and nuclear magnetic resonance ( nmr ) @xcite experiments have amply borne out this conjecture .
the theoretical understanding of ladders has been in large part based on numerical techniques , which are particularly well suited to systems of such restricted dimensionality , and this is reviewed in ref .
the first indications that the 2-leg ladder should exhibit a spin gap came from numerical work @xcite , and a variety of methods has since been applied to investigate the spin gap and spin correlations , and the properties of multi - leg ladders.@xcite these studies not only confirm the picture of the spin system emerging from analytic approaches @xcite valid in certain limits , but provide the most accurate information available on the properties of this class of strongly - correlated systems .
the second proposal concerning the properties of ladder compounds made in ref .
@xcite , that the doped ladder should become superconducting with a @xmath8-wave order parameter , has proved harder to test for materials reasons .
however , the compound lacuo@xmath0 , recently synthesized by hiroi and takano,@xcite has been found to admit hole doping without altering its structure of coupled , 2-leg copper oxide ladders , and so constitutes the first case in which one may seek doping - dependent behavior analogous to that of the high - temperature superconductors .
the high - pressure phase of la@xmath9sr@xmath10cuo@xmath0 is derived from a cubic , three - dimensional perovskite structure .
the absence of oxygen atoms along lines leaves 2-leg ladders which relax from a relative angle of 90@xmath11 in the primitive structure to 62@xmath11 in the depleted one .
the material is then orthorhombic ( space group p@xmath12 ) , with four copper atoms per unit cell , and may be considered as a set of ladders oriented along the @xmath13-axis , and periodically arrayed in the @xmath14- and @xmath15-directions , as shown in fig .
this structure is not altered with sr doping to @xmath16 .
the interladder coupling in the ( @xmath17 ) plane arises because each oxygen atom at the outside edges of the 2-leg ladders , which is part of the planar cuo@xmath18 unit , is also effectively apical to a copper atom in a neighboring ladder , contributing to a finite transfer integral .
susceptibility measurements @xcite on the parent compound up to temperatures of 500k , interpreted by a formula proposed by troyer _
et al._,@xcite suggest the presence of a spin gap in the excitation spectrum , and therefore a spin liquid state .
in contrast , nmr studies of the same samples @xcite indicate that the system orders antiferromagnetically at low temperatures , below an apparent @xmath19k
. since the spin susceptibility should evolve continuously as one passes through the quantum critical point separating the spin - liquid state and the state with long - range antiferromagnetic order , it is possible that even in the latter , close to the quantum critical point , the spin susceptibility will decrease substantially as the temperature is lowered . such behavior may be difficult to distinguish from that of a true spin liquid with a finite gap .
therefore it is plausible that the seemingly contradictory experimental results may be reconciled if the magnetic state is close to the crossover from an antiferromagnet to the spin liquid .
here we seek evidence , by examining the effects of interladder interactions , that the system is indeed close to this quantum critical point .
the outline of this paper is as follows . in section
ii we present a tight - binding fit to the lda bandstructure to extract the interladder hopping matrix elements , and use these to obtain the superexchange interactions .
we consider in section iii the nature of the spin - liquid ground state for ladders of spins coupled in both orthogonal directions , using a mean - field approach to estimate the location of the quantum critical point . in section
iv we introduce a double - occupancy term to the tight - binding bands , to investigate the degree of three - dimensional character in the electronic structure as a result of interladder interactions .
section v contains our conclusions and a brief discussion .
the bandstructure of lacuo@xmath0 has been computed by mattheiss,@xcite using the local density approximation ( lda ) method . here
we examine the dispersion of only the highest occupied valence bands , and use a tight - binding model based on the single , copper - based orbital in each planar cuo@xmath18 unit within the ladder which lies closest to the chemical potential : in the cu@xmath5o@xmath6 strip this is the antibonding , cu - centered 3@xmath20 orbital . restricting the set of hopping matrix elements to those between nearby copper atoms
, we obtain an effective one - band model , albeit with four mixed levels ( one from each cu atom in the unit cell ) .
these generate a complex of four bands , which lie close to or cut the fermi energy .
the complex is half - filled in the undoped system , and well separated from other bands on the small energy scales of most physical interest , so can be taken to determine the low - energy behavior of the model . with this interpretation
, one may deduce the ratio @xmath21 , of inter- and intraladder atomic orbital overlap , and thus estimate the ratio @xmath22 of the magnetic interactions , by using the superexchange result @xcite @xmath23 .
the tight - binding hamiltonian is @xmath24 in which @xmath25 and @xmath26 each denote a pair ( @xmath27 ) , where @xmath2 labels the unit cell , and @xmath28 the different atoms within each cell .
the hopping matrix elements @xmath29 may be taken to be short - ranged , and the maximal set which we need to achieve a reasonable fit is illustrated in fig .
starting from tight - binding parameter fits in the two - dimensional cuo@xmath7 plane,@xcite we choose the nearest - neighbor intraladder parameters , @xmath30 for hopping along a rung and @xmath31 for hopping along a leg , to have the same value @xmath32ev .
we use as the next - nearest - neighbor parameters @xmath33@xmath34 and @xmath35@xmath36 for hopping across a square plaquette in the ladder , and along two leg bonds , respectively .
finally , we introduce a parameter @xmath37 for hopping from a copper atom on one ladder to its neighbor on the adjacent ladder , and also the next - neighbor analog @xmath38 for transfer between atoms on adjacent ladders with a relative displacement of one leg bond . because of the low symmetry of the lacuo@xmath0 structure , the cu-(apical ) o - cu interladder bond is far from straight , with the cuo@xmath39 pyramids quite irregular , and this distortion may allow a substantial value of @xmath37 . the role of each of the terms in the fitting scheme
will be illustrated below .
the next - neighbor parameters @xmath40 are expected to be significantly smaller than their direct counterparts , due to the short - range nature of the overlaps , and the important parameter to fit will be the ratio of @xmath37 to @xmath30 . the hamiltonian ( [ esh ] ) may be expressed in matrix form in reciprocal space as @xmath41 where @xmath42 is the vector of creation operators @xmath43 for the bloch states formed by separate linear combination of the @xmath20 orbitals of each copper atom in the unit cell , and @xmath44 is a vector in the orthorhombic brillouin zone . setting the lattice constants @xmath45 , @xmath46 and @xmath47 to unity , the hamiltonian matrix is @xmath48 where @xmath49 , @xmath50@xmath51 , in which the factor @xmath52@xmath51 arises because @xmath53 , @xmath54 , and the bond vectors @xmath55(@xmath56 ) and @xmath57(@xmath58 ) are shown in fig .
the eigenvalue problem gives an equation quadratic in the squares of the mode frequencies , whose solutions are the dispersion relations of the four energy bands @xmath59^{1/2 } \label{eqpb } \nonumber \\ & & \;\;\;\ ; - 2 t_z \cos k_z - 2 t_{z}^{\prime } \cos 2 k_z .\end{aligned}\ ] ] this concise , closed form emerges because the exponential factors @xmath60 may be collected as @xmath61 , as @xmath62 .
the general dispersion simplifies further for particular values of @xmath63 and @xmath64 in the brillouin zone : in particular , on the zone faces ( @xmath65 ) , the last term in the square root vanishes and the bands are doubly degenerate , as required by the group - theoretical analysis of structures with non - symmorphic space groups .
the dispersions of the 4 bands are shown in fig .
2(a ) for a series of high - symmetry lines in the orthorhombic brillouin zone , and for the ratio @xmath66 .
the labelling of points is shown in fig .
2(c ) , and their order is chosen to match the results of mattheiss.@xcite also shown in the figure are the energy - spacings dictated by the choice of the tight - binding parameters , as these vary between their maximum and minimum values along the chosen directions . for an isolated ladder along @xmath67 ,
i.e. with no interladder interactions ( @xmath68 ) , the bands would be completely flat around @xmath69xsy@xmath69 and zurtz , with only a cosine dispersion ( mildly perturbed by the @xmath70 term ) along @xmath69z . in this situation
there would be two doubly - degenerate bands around the zone center and zone face , corresponding to the bonding and antibonding bands of each ladder .
these features are reflected in the two branches observed in fig .
2(a ) . from eq .
( [ eqpb ] ) , the separation of the degenerate bands along the zone face xsy and edge urt is governed by the combinations @xmath71 , and the splitting of the degeneracy along @xmath69x , @xmath69y and zu , zt by the combinations @xmath72 .
considering first the intraladder parameters @xmath31 , @xmath70 , @xmath30 and @xmath73 chosen from the two - dimensional cuo@xmath7 plane , we find good agreement of the tight - binding result with that from lda .
particularly notable is that the negative sign of @xmath73 , and the relative magnitude @xmath34 , are required to reproduce the bands separations along xsy and urt .
there is no evidence that a value of @xmath30 different from @xmath31 would improve the fit .
the next - neighbor hopping parameter along the ladder , @xmath70 , appears only as an asymmetry of the cosine dispersion along @xmath69z , and the value chosen is in qualitative accord with the lda result .
that cuo@xmath7 plane parameters remain appropriate for the ladder confirms the predominantly local picture of the interactions between copper sites . turning to the band splitting at the @xmath69 and z points , we find that the relatively large value @xmath74@xmath34 ( fig .
2(b ) ) gives the best qualitative reproduction of the bands crossing the chemical potential in the full lda calculation by mattheiss,@xcite but that the value @xmath75@xmath34 appears closer to the results of a `` 12-parameter fit '' illustrated in the same reference .
the difference of the @xmath69 and z splittings is given rather well by the starting choice of @xmath76@xmath77 .
it is clear from fig .
2 that the primary feature of the dispersion remains that in the @xmath78 direction , i.e. along the ladder .
the fermi surfaces for each band are determined almost exclusively by this part of the dispersion , in that they appear as sheets which are almost flat , perturbed little by the @xmath30 and @xmath37 terms , and have @xmath78 as normal . at half - filling ,
the lowest band is an exception to this situation , because part of it is also filled close to the z point .
on doping with holes , this region is rapidly emptied ( below 5% ) so that all four fermi surfaces are sheets parallel to the ( @xmath79 ) plane .
only when the doping level reaches 20% does the chemical potential drop below the highest band in the @xmath69xsy@xmath69 plane , causing a pocket to open at the @xmath69 point for this band . in summary , we find that a simple model of a single orbital per copper atom provides a good fit to the bandstructure . while there is scope for some variation in the choice of intraladder parameters , with this level of agreement between the tight - binding results and those of lda it is not worthwhile to optimize further .
we choose to work with the above values of the intraladder hopping matrix elements , and with the interladder overlap @xmath74@xmath80 , bearing in mind that this latter will be close to the upper limit of the narrow range of probable values . estimating the superexchange interaction by @xmath81 leads us to conclude that the interaction between spins on neighboring ladders will have a magnitude @xmath82 , where @xmath83 is the intraladder magnetic coupling of both rung and leg spins . because @xmath84 is an appreciable fraction of @xmath83
, it is clear that the spin interactions in lacuo@xmath0 will have significant three - dimensional character .
a mean - field analysis of the spin state for ladder systems was introduced by gopalan _
@xcite , and in this section we follow closely the treatment of these authors .
they employed a bond operator representation of @xmath85 quantum spins , used initially @xcite to investigate dimerized spin phases in two - dimensional systems , exploiting the fact that the topology of the ladder favors dimerization , particularly when the spin interaction , @xmath83 , on a rung exceeds that on a leg , @xmath86 .
the authors proceeded in the mean - field approximation to consider first the properties of an isolated ladder as a function of the interaction ratio @xmath87 , then of periodic arrays of ladders in two dimensions , and finally of an array of frustrated double ladders of the type found in the srcu@xmath7o@xmath88 system.@xcite this approach can be considered to be exact in the limit where @xmath89 and the spins form dimers on every rung , while its accuracy will diminish on extrapolating through finite @xmath87 towards the desired isotropic point , @xmath90 .
we begin by representing the system of spin ladders , each coupled by two identical bonds per spin to separate , neighboring ladders , in a geometrically simpler but topologically identical form , as shown in fig .
3 . note that the spin configuration for local , antiferromagnetic interactions is unfrustrated , so a transition to an ordered antiferromagnet may be expected with increasing interladder coupling .
the magnetic interactions are taken to be @xmath83 for spins on the same rung , @xmath86 for spins separated by a leg bond , and @xmath91 for neighboring spins on different ladders , while no other couplings are considered .
the hamiltonian for the spins is @xmath92 where @xmath26 is a rung bond index and the labels @xmath93 and @xmath94 denote spins on the left and right sides of the ladder . following ref .
@xcite , transformation to the bond - operator representation yields @xmath95 where @xmath96 @xmath97 and @xmath98 in these equations , @xmath99 is the creation operator for a spin singlet on bond @xmath26 , the operators @xmath100 create the three possible triplet states on the same bond , and the lagrange multiplier , @xmath101 , introduced to ensure the constraint @xmath102 on each bond , which restricts the physical spin states to singlets or triplets , appears as an effective chemical potential .
the part @xmath103 in ( [ ebosh ] ) contains terms with three and four @xmath104 operators , and will be neglected in our approximation ; in ref .
@xcite it was shown in addition that the effects of such higher - order terms are small . because the singlet on each bond has the lowest energy , we assume that the system condenses into this state , leading to a finite expectation value of the bosonic @xmath105 operator , @xmath106 . this is the average expectation value , or mean - field value of the operators @xmath105 , and the site - dependent chemical potential @xmath101
is also replaced by a global average value @xmath107 .
working with the physical unit cell ( fig .
3 ) , which contains two rungs , we may transform the operators @xmath104 in @xmath108 , @xmath109 and @xmath110 to those for two types of triplets , @xmath111 and @xmath112 .
the hamiltonian in this approximation is @xcite @xmath113 + \left [ \lambda_{\bf k}^{\prime } t_{{\bf k } \alpha}^{1 \dag } t_{{\bf k } \alpha}^{2 } \right . \\ & & \left .
+ \delta_{\bf k}^{\prime } \left ( t_{{\bf k } \alpha}^{1 \dag } t_{-{\bf k } \alpha}^{2 \dag } + t_{{\bf k } \alpha}^{1 } t_{-{\bf k } \alpha}^{2 } \right ) \right ] + [ 1 \leftrightarrow 2 ] \right\ } , \nonumber\end{aligned}\ ] ] in which @xmath114 @xmath115 @xmath116 and @xmath117 denotes the total number of ladder rungs .
the part of @xmath118 ( [ emfsh ] ) dependent on the triplet operators is diagonalized by the non - unitary , bosonic bogoliubov transformation @xmath119 whose coefficients are given by @xmath120 where in turn @xmath121 yield the dispersion relations of the two magnon branches .
the hamiltonian now takes the form @xmath122 which contains the mean - field part and the zero - point quantum corrections from the triplet magnon excitations .
the mean - field equations to be solved self - consistently for @xmath107 and @xmath123 are given by @xmath124 and @xmath125 where @xmath126 contain the dispersive parts of @xmath127 .
@xmath128 denotes the magnon thermal occupation function , and will be discussed in more detail in a future publication .
the factor of 3 preceding the @xmath44-summations in both equations is the result of the sum over @xmath129 for the three triplet magnon states.@xcite this factor was omitted in ref .
@xcite , and we comment below on the effect of the correction on the results presented there .
the mean - field equations are solved at zero temperature , where the thermal factor becomes unity , and by taking the continuum limit in which the @xmath44-sum becomes an integral over three - dimensional reciprocal space .
as in ref .
@xcite we reduce the two equations to a single one for the variable @xmath130 which has the form @xmath131 as a characteristic parameter of the spin - liquid ground state , we will be most interested in the value of the spin gap , the minimum excitation energy of the lower branch of the triplet magnon excitations , which is given by @xmath132 with @xmath8 determined by the mean - field equation ( [ escd ] ) . from eq .
( [ eqd ] ) we see that the excitation spectrum has a minimum at the wavevector @xmath133 in the reciprocal lattice of the bipartite structure shown in fig .
3 . the value of @xmath134 where @xmath135 is driven to zero will give the transition from the spin liquid state , where the spin orientation fluctuates with a time - averaged value of zero and with short - range correlations primarily at @xmath136 , to a magnetically ordered state characterized by @xmath136 .
this wavevector corresponds to uniform polarization of the spin singlets on ladder rungs in the ( @xmath17 ) plane , with spins oppositely directed between neighboring planes in the @xmath13-direction , i.e. a simple antiferromagnetically aligned spin pattern . in the limit of no interladder coupling
we obtain the spin gap @xmath137 of the isolated , isotropic ( @xmath90 ) 2-leg ladder .
this is only a mean - field result , but is in very good agreement with the result @xmath138 of numerical studies @xcite by the density matrix renormalization group technique .
in fact this agreement is largely serendipitous , and deteriorates on taking into account the higher - order terms @xcite ; in the mean - field approximation , the spin gap of the isolated ladder diverges logarithmically in the limit of large @xmath87,@xcite and the effects of this increase are already manifest at @xmath90 , causing the mean - field result , which initially underestimates the spin gap , to recover towards the exact value determined numerically . in this treatment
we have not had to invoke a self - energy correction term : in ref .
@xcite , the authors investigated the curious qualitative behavior of the solution to their ( erroneous ) mean - field equation by expanding in small @xmath87 about the limit of strong rung coupling where the dimer treatment is accurate . in the corrected mean - field theory ,
one obtains @xmath139 which corresponds reasonably well to the result @xmath140 of a detailed strong - coupling analysis including excitation modes.@xcite previously , the coefficient of the quadratic term had been found to be negative , and so a self - energy term @xmath141 was introduced to correct for short - range interaction effects which appeared to have been missing at the mean - field level ; the chosen value @xmath142 brought the results into good argeement with previous numerical ones , and with the above approximate treatments .
the spin gap obtained from the solution of ( [ escd ] ) for the three - dimensionally coupled ladder system is shown in fig .
4(a ) as a function of the interladder coupling @xmath134 , for fixed @xmath90 .
we see immediately that the spin gap decreases monotonically , with the transition point at @xmath143 .
comparison with the result of section ii indicates that the lacuo@xmath0 system should lie within the ordered antiferromagnetic regime , but that it is indeed located in the vicinity of the quantum critical point marking the phase transition from spin liquid to magnetic order .
it is also instructive to compare the appearance of the spin gap with that in a two - dimensional , unfrustrated , periodic array of coupled ladders.@xcite in this case there is only one magnon branch , and the dispersive factor in the excitation spectrum is @xmath144 where @xmath63 is a wavevector parallel to a ladder rung .
again @xmath145 corresponds to a simple antiferromagnetic spin alignment in the ordered phase , and the factor of appears because there is only one bond per spin between ladders in such a system .
the spin gap for the two - dimensional array is shown as a function of @xmath134 in fig .
4(b ) , where again @xmath90 and the gap at @xmath146 is that of the isolated , isotropic ladder .
the critical value of the interladder coupling , @xmath147 , is seen to be significantly greater than twice that in the three - dimensional case above , as might be expected for the simple reason that there are half as many interladder interactions , which may be taken as an indication that the spin liquid state is more robust in lower dimensions .
the almost linear decrease , in contrast to the downward curvature of the function in fig .
4(a ) , illustrates a further effect of dimensionality , and agrees well with the qualitative result of ref .
@xcite .
we may conclude that to within the accuracy of the mean - field approach for an isotropic ladder system , the above analysis provides good evidence that lacuo@xmath0 lies on the magnetically ordered side of the quantum critical regime of the transition between spin liquid and antiferromagnet .
we have seen in the previous sections that the interladder interactions in lacuo@xmath0 are quite strong , and possibly sufficiently strong to change the magnetic structure from a spin liquid to a three - dimensionally ordered antiferromagnet . to examine further the character of the electronic structure ,
we introduce an on - site coulomb interaction , @xmath1 , to the tight - binding bandstructure of section ii , and perform a hartree - fock calculation of an ordered antiferromagnetic state . in a strictly one - dimensional system , the critical value of @xmath1 necessary to stabilize an insulating antiferromagnet vanishes at half - filling , whereas in a general , three - dimensional electronic structure it is of the order of the bandwidth .
introduction of the on - site coulomb interaction leads to a hubbard hamiltonian @xmath148 examining the stability of an antiferromagnetic state with wavevector @xmath136 , we note first that there will be no increase of the unit cell in the ( @xmath17 ) plane , since there are already two atoms for each spin direction , but that it doubles along @xmath149 when the spins order along the ladder legs . in the notation of @xmath150 and @xmath151 sublattices for the bipartite system , sites @xmath152 of the original unit cell ( fig .
1(c ) ) in every second plane , and @xmath153 in the alternating planes , will belong to the @xmath150 sublattice , while the remaining sites will belong to @xmath151 .
introducing the parameter @xmath154 as the difference between average site occupation by particles of each spin orientation on a site of the @xmath150 sublattice , we require @xmath155 for all sites .
we proceed by solving the problem in the hartree - fock approximation , with the value of @xmath1 where @xmath156 becomes finite marking the antiferromagnetic transition . following the treatment of section ii
, the hamiltonian may be written as in eq .
( [ esmh ] ) , with now @xmath157 , where the superscripts @xmath45 and @xmath46 denote atoms in the two ( @xmath17 ) planes of the doubled unit cell , and @xmath158 in which the @xmath159 matrices @xmath160 @xmath161 @xmath162 ( defined identically using @xmath56 ) and @xmath163 are the generalizations of the previous expressions to the new unit cell , and the new @xmath47-axis dimension is set to unity .
this matrix can not be block - diagonalized , but the structure of the solution for the eigenmodes and eigenvectors is evident from section ii .
schematically , if the ( @xmath164 ) dispersion contained in the square root in eq .
( [ eqpb ] ) is denoted by @xmath165 , the eight band dispersions will have the form @xmath166^{1/2 } - 2 t_{z}^{\prime } \cos 2 k_z , \label{eshmb}\ ] ] where @xmath25 labels the bands .
the @xmath167 term splits the former four bands into two sets , which as @xmath1 becomes large will not overlap , ensuring that the half - filled system becomes insulating .
the corresponding eigenvectors are the states @xmath168 where the index @xmath26 runs over the eight atomic sites , and @xmath169 is the bloch state created by the operator @xmath170 .
the hartree - fock equations for the system , which will determine @xmath156 and the chemical potential @xmath107 at fixed @xmath1 when solved self - consistently , are the equations for the total and site occupancies @xmath171 and @xmath172 these equations have been generalized to arbitrary band filling , which is expressed in terms of the deviation @xmath173 from half - filling , which in turn is normalized to be @xmath174 .
the chemical potential is contained in @xmath175 , and the second expression is valid for the coefficients of any chosen bloch function @xmath26 . the operation of diagonalizing the hamiltonian matrix ( [ eshhm ] ) can be performed numerically at sufficient speed that it is still possible to solve the hartree - fock equations , which involve three - dimensional @xmath44-integration , on a workstation within a reasonable amount of time . in fig .
5(a ) are shown the energy bands for the incipient antiferromagnetic system for half - filling ( @xmath176 ) and @xmath177 .
the splitting of the bands into upper and lower branches as a result of @xmath1 is clearly evident , as is the characteristic @xmath69xsy@xmath69 and zurtz structure of the purely kinetic hamiltonian in both sets of bands , at positive and negative energies , as a result of the folding back of bands due to the change in meaning of the coordinate @xmath78 , which now spans a brillouin zone half the former size . at this value of @xmath1
, the half - filled system would appear to be close to the transition from metallic to insulating behavior , which one expects near the point where there is no longer any overlap between upper and lower band energies in any region of reciprocal space . at smaller values of @xmath1 ,
the two sets of bands are characteristic of the doubled brillouin zone , and have a semi - metallic overlap , while as @xmath1 is raised to large values , the energy gap increases , and the bands become progressively more flat .
in other cuprate compounds , a hubbard model has been found @xcite to give an accurate description of the low - energy behavior with a ratio @xmath178 of order unity , where @xmath179 is the bandwidth . here
the total bandwidth of the uppermost valence bands is approximately the same as in the cuprates with cuo@xmath7 planes , @xmath180ev .
since the local environment of the cu@xmath181 ions is similar , the on - site coulomb repulsion should also be the same , @xmath182ev .
these parameter values place lacuo@xmath0 well within the mott insulating region .
we note that the actual magnetic structure can not be determined in the hartree - fock approximation , as the quantum corrections which act to stabilize the spin liquid phase are not included .
the form of the hartree - fock solutions for smaller values of @xmath1 are sensitive to the effective dimensionality of the electronic structure . in general , values of @xmath183
are required to obtain a mott insulator at half - filling , but the perfect - nesting property of a one - dimensional band gives an insulating state for arbitrarily small values of @xmath1 . in fig .
5(b ) , the average antiferromagnetic order ( parameterized by @xmath156 ) in the half - filled system is shown as a function of the on - site repulsion @xmath1 , which is measured in terms of the bandwidth @xmath184ev . in lacuo@xmath0
we find that the critical value @xmath185 for the antiferromagnetic transition is rather small , @xmath186 , which indicates substantial one - dimensional nature .
the interladder hopping matrix elements , although quite strong , are not sufficient to destroy the nesting character completely , which would make the electronic structure effectively three - dimensional .
this quasi - one dimensional behavior of the bands may also be reflected in the sensitivity of the system to random potential fluctuations .
it is well known that the onset of localization is strongly dependent on dimensionality .
clearly , in ladder compounds there is an inherent conflict between the need to change the valence of the counterions in order to induce hole carriers , and the need to avoid strong , random potential fluctuations .
thus it appears that in la@xmath9sr@xmath10cuo@xmath0 the random potential fluctuations act to cause localization for @xmath187 , in spite of the substantial interladder overlap , suggesting that a more gentle hole - doping technique will be required to retain itinerant character at small doping concentrations .
this could perhaps be achieved in structures where the counterions are further from the ladders than in the present case of lacuo@xmath0 .
we have investigated the basic electronic and magnetic properties of the three - dimensionally coupled 2-leg ladder compound lacuo@xmath0 , a material which is of significant experimental and theoretical interest as it is the first ladder compound to be discovered in which the cu@xmath7o@xmath88 ladders may be doped with holes .
we present a tight - binding fit in which the bands are derived from a single ( @xmath20 ) orbital close to the fermi energy on each copper atom .
the results of lda studies are well reproduced by reasonable values of the most significant transfer integrals : those within each ladder are found to be similar to the cuo@xmath7 planar system , emphasizing the short - range nature of the dominant physical processes , while the interladder hopping term @xmath188@xmath189 is found to be quite large . as a consequence ,
the interladder magnetic coupling @xmath84 is also relatively large , and the compound may be expected to exhibit some three - dimensional characteristics .
the effective spin interactions in this structure are those of unfrustrated antiferromagnetism , and a mean - field treatment of the magnetic state from the basis of dimerized singlets on the rungs of decoupled ladders gives a spin - liquid phase whose spin gap decreases with increasing interladder coupling .
the spin gap is found to vanish , signalling a transition to an ordered phase , at an interladder coupling ratio @xmath190 somewhat smaller than that deduced for lacuo@xmath0 , indicating that the system is located in the antiferromagnetically ordered state , albeit not far from the quantum critical point of the ordering transition .
we may take the very small value of the intrinsic susceptibility measured at low temperatures @xcite as evidence that proximity to the critical point , and the possibility this allows of significant critical fluctuations , plays an important role in determining the physics of the system .
further insight into the electronic properties is provided by the introduction of an on - site repulsion term , @xmath1 , to the band scheme : within the hartree - fock approximation we find that in the half - filled system the transition to an antiferromagnetic insulator occurs for values of @xmath1 quite small compared to the bandwidth .
this is a reflection of the fact that the predominant feature of the bands remains the dispersion in the ladder direction , and this one - dimensionality , also apparent in the fermi surfaces of the partially - filled bands , makes the system inherently susceptible to potential fluctuations . that @xmath185 is finite illustrates an absence of perfect nesting conditions in the bandstructure , and the instability occurs at the antiferromagnetic wavevector .
our results represent the first theoretical consideration of the lacuo@xmath0 system . while by the nature of the approximate techniques used they are somewhat inexact
, we believe that they are important in establishing the parameter space for the magnetic state , and in focussing the direction of further research .
the properties of such `` nearly critical '' magnetic systems have been studied for the two - dimensional , planar case by detailed analytical @xcite and numerical @xcite techniques .
we propose the application of similar methods for the case of ladder systems with variable interladder coupling in two additional dimensions , to investigate the nature of the critical point , and the surrounding `` quantum critical '' regime , in the same formalism , with particular emphasis on the appearance of physical quantities accessible in experiment .
such studies may prove valuable in establishing the framework within which to interpret the data from experiment , and hopefully will serve to reconcile the apparent contradiction in current results .
the observation that the spin liquid is not the appropriate description of the magnetic ground state in lacuo@xmath0 is itself important , particularly with regard to the current interpretation of the susceptibility data.@xcite while it may be possible to deduce the form of the susceptibility suitable for the critical regime from a mean - field picture , we await the results of numerical studies of the same system in order to make a more detailed comparison with the data .
a further direction not strongly emphasized in this communication is the nature of the doped system : the methods used here become less accurate at finite doping , and so were not studied in great detail in this regime , but may nonetheless be used to obtain additional insight into the evolution of spin properties on moving towards the metallic state .
we are grateful to y. kitaoka , m. sigrist , m. takano , m. troyer and especially z. hiroi and m. zhitomirsky for helpful discussions , and to the swiss national fund for financial support .
99 e. dagotto and t. m. rice , science * 271 * , 618 ( 1996 ) . z. hiroi , m. azuma , m. takano and y. bando , j. solid state chem . * 95 * , 230 ( 1991 ) .
t. m. rice , s. gopalan and m. sigrist , europhys . lett . * 23 * , 445 ( 1993 ) .
m. azuma , z. hiroi , m. takano , k. ishida and y. kitaoka , phys .
lett . * 73 * , 3463 ( 1994 ) .
k. ishida , y. kitaoka , k. asayama , m. azuma , z. hiroi and m. takano , j. phys .
63 * , 3222 ( 1994 ) .
e. dagotto , j. riera and d. j. scalapino , phys .
b * 45 * , 5744 ( 1992 ) .
r. hirsch , diplomarbeit , _ uni .
, unpublished ( 1988 ) .
t. barnes , e. dagotto , j. riera and e. swanson , phys .
b * 47 * , 3196 ( 1993 ) . s. white , r. noack and d. j. scalapino , phys . rev .
lett . * 73 * , 886 ( 1994 ) .
b. frischmuth , b. ammon and m. troyer , preprint .
g. sierra , preprint .
z. hiroi and m. takano , nature * 377 * , 41 ( 1995 ) .
m. troyer , h. tsunetsugu and d. wrtz , phys .
b * 50 * , 13515 ( 1994 ) .
s. matsumoto , y. kitaoka , k. ishida , k. asayama , z. hiroi , n. kobayashi and m. takano , phys .
b * 53 * 11942 ( 1996 )
. l. f. mattheiss , solid state commun .
* 97 * , 751 ( 1996 ) .
p. w. anderson , in _ magnetism i _
, edited by g. t. rado and h. suhl ( academic press , new york , 1963 ) .
t. tanamoto , h. kohno and h. fukuyama , j. phys .
jpn . , * 62 * , 717 ( 1993 ) .
b. normand , h. kohno and h. fukuyama , j. phys .
jpn . , * 64 * , 3903 ( 1995 ) .
s. gopalan , t. m. rice and m. sigrist , phys .
b * 49 * , 8901 ( 1994 ) .
s. sachdev and r. bhatt , phys .
b * 41 * , 9323 ( 1990 ) .
we are indebted to m. e. zhitomirsky for pointing out to us an error in the original version of the following calculation .
m. e. zhitomirsky , o. a. starykh and k. ueda , private communication .
m. reigrotzki , h. tsunetsugu and t. m. rice , j. phys .
* 6 * , 9325 ( 1994 ) .
m. hybertsen , e. b. stechel , m. schlter and d. r. jennison , phys .
b * 41 * , 11068 ( 1990 ) .
a. v. chubukov , s. sachdev and j. ye , phys .
b * 49 * , 11919 ( 1994 ) .
m. troyer , private communication . | the recently - discovered `` ladder '' compound lacuo@xmath0 has been found to admit hole doping without altering its structure of coupled copper oxide ladders .
while susceptibility measurements on the parent compound suggest a spin gap and a spin - liquid state , nmr results indicate magnetic order at low temperatures .
these seemingly contradictory results may be reconciled if in fact the magnetic state is near the crossover from spin liquid to antiferromagnet , and we investigate this possibility . from a tight - binding fit to the valence lda bandstructure , we deduce that the strength of the interladder hopping term is approximately half that of intraladder hopping , showing that the material is three - dimensional in character .
a mean - field treatment of the insulating magnetic state gives a spin - liquid phase whose spin gap decreases with increasing interladder coupling , vanishing ( signalling a transition to the ordered phase ) at a value somewhat below that obtained for lacuo@xmath0 .
the introduction of an on - site repulsion term , @xmath1 , to the band scheme causes a transition to an antiferromagnetic insulator for rather small but finite values of @xmath1 , reflecting the predominance of ( one - dimensional ) ladder behavior , and an absence of any special nesting features . |
fission is a physical phenomenon taking place in heavy atomic nuclei that leads to the disintegration of a parent nucleus into two or more emerging fragments .
it involves the evolution of the nucleus from its ground state to scission going through a variety of intrinsic shapes that cover a wide range of different intrinsic deformation parameters @xcite .
fission properties depend upon the competition between the surface energy term coming from the strong nuclear interaction and the coulomb repulsion and therefore they are often used as constraints and/or guidance to refine the parameters of effective nuclear interactions .
a typical example is the d1s parametrizaton of the gogny @xcite force with parameters fine tuned to reproduce the fission barrier of @xmath3pu @xcite .
more recently fission related constraints have been used with skyrme interactions to define the unedf1 parametrization @xcite .
the gross features of fission can be understood from a mean - field perspective using the hartree - fock - bogoliubov ( hfb ) theory @xcite and therefore it is not surprising the large amount of studies devoted to this subject with skyrme interactions @xcite , gogny ones @xcite or based on the relativistic mean field @xcite .
fission observables also depend on the inertia of the system to the relevant collective degrees of freedom and therefore they are sensitive to pairing correlations . as a consequence , fission is a good testing ground to test both the theories and interactions commonly used in nuclear structure . in addition , the theoretical understanding of fission is relevant to other areas outside traditional nuclear physics like safe energy production with nuclear reactors , radioactive waste degradation or the nucleosynthesis of heavy elements in the explosive galactic environments through the the r - process .
last but not least , a better understanding of fission could open the door to a better estimation of magic numbers and hence extra stability of super - heavy nuclei beyond z=114 . in this paper
we explore the ability of a newly proposed energy density functional ( edf ) denoted as barcelona - catania - paris - madrid ( bcpm ) @xcite to describe fission .
the bcpm is a recent parametrization of the bcp edf @xcite devised for nuclear structure calculations .
its free parameters have been adjusted to reproduce the binding energies of even - even nuclei all over the nuclide chart , including deformed ones . instead of the more traditional approaches where some central potential form is guessed ( contact , gaussian , yukawa , etc ) and used afterwards to fit nuclear matter properties and/or the nuclear matter equations of state eos ( both symmetric and neutron ) , in the bcpm functional we start from a microscopic eos that is fitted by means of a low order polynomial in the density .
that polynomial fit is translated to a finite nuclei edf just by replacing the nuclear matter density by the density of the finite nucleus .
this procedure is inspired by the local density approximation ( lda ) and is common practice in practical applications of the kohn - sham theory in condensed matter physics .
the edf is supplemented with a finite range surface term , a contact spin - orbit interaction of the same form as in skyrme or gogny forces , the coulomb interaction and finally , a density dependent zero range pairing interaction @xcite with strengths fitted to reproduce gogny s neutron matter pairing gap .
the parameters of the functional ( essentially those of the finite range surface term plus some freedom in the polynomial fit to fine tune the binding energy per nucleon ) are fitted to reproduce binding energies of the 518 even - even nuclei of the 2003 mass table evaluation of audi and wapstra .
the properties of the interaction concerning quadrupole , octupole and fission dynamics have also been explored @xcite . as shown below
, the bcpm functional gives reasonable results for fission observables including spontaneous fission half lives , fission isomer excitation energies , inner and outer barrier heights and mass distribution of fragments .
we have also shown that those results could be improved by slightly modifying the amount of pairing correlations , either by modifying the pairing strengths or by going beyond the mean field approximation to restore the particle number symmetry broken by the hfb procedure . as a consequence of the satisfactory performance of bcpm in describing fission ,
we have explored fission properties of the uranium isotopic chain from proton drip line to the neutron drip line .
both the bcpm energy density functional @xcite and its predecesor bcp @xcite are made of a bulk part which is determined by fully microscopic and realistic calculations of symmetric and neutron matter equations of state @xcite as in the lda of condensed matter physics .
the two equations of state ( symmetric and neutron matter ) given as a function of the nuclear density are parametrized by low order polynomials of the densities . to account for finite size effects related to the surface energy , a phenomenological finite range gaussian interaction
is included .
in addition , the coulomb interaction and the spin - orbit term are taken exactly as in the skyrme or gogny forces . to deal with open - shell nuclei
we include in the bcpm and bcp functionals a zero range density - dependent pairing interaction fitted to reproduce the nuclear matter gaps obtained with the gogny force @xcite .
the calculations in finite nuclei are carried out with a modification of the hfbaxial computer code developed by one of the authors @xcite . to describe fission
we follow the usual procedure based on the mean field approach with pairing correlations : the hartree- fock- bogoliubov ( hfb ) theory with constraining fields . as constraining operators
we have used mainly the axially symmetric quadrupole moment operator @xmath4 although some exploratory calculations have also been performed with the octupole @xmath5 and hexadecapole @xmath6 moment operators and the necking operator @xmath7 .
axial symmetry is preserved in the calculations because of the high computational cost involved in the release of this restriction .
we are aware of the relevance of triaxiality specially in the height of the inner fission barrier but its effect is merely quantitative and to a much lesser extent , qualitative . on the other hand ,
reflection symmetry is allowed to break at any stage of the calculation permitting octupole deformation and asymmetric fission . as a consequence of the breaking of the parity symmetry
we are forced to constraint the center of mass to the origin as to prevent spurious solutions corresponding to a translation of the nucleus as a whole .
the quasiparticle creation and annihilation operators of the hfb theory are expanded in a harmonic oscillator basis ( ho ) preserving axial symmetry and containing ho states with @xmath8 quantum numbers up to 35/2 and up to 26 quanta in the @xmath9 direction .
the basis contains over 3000 levels but time reversal invariance and the axial block structure reduces the computational complexity to a manageable level .
the two lengths characterizing the ho basis , @xmath10 and @xmath11 , have been optimized in a few nuclei for each value of the quadrupole moment .
for the others the oscillator lengths computed for nearby nuclei are used . as the number of hfb configurations for each nucleus is large , a robust and fast gradient - like algorithm to solve the hfb equations
is used @xcite .
the most evident advantage of this method is the way it handles the constraints , which allows an easy generalization to an arbitrary number of them .
the spontaneous fission lifetime formulas eqs [ eq : tsf ] and [ eq : s ] below , depend crucially on the theory of the collective mass @xmath12 .
we shall use two methods to calculate it and compare in our results .
the first is the well - known cranking approximation to the adiabatic time dependent hfb approximation @xcite .
the resulting mass is expressed in terms of the moments @xmath13 of the generating field @xmath14 @xmath15 as @xmath16 here @xmath17 are distinct 2-quasiparticles excitations and @xmath18 is the excitation energy , neglecting the quasiparticle- quasiparticle interaction ( cranking approximation @xcite ) . an alternative method to calculate the mass
is based on the gaussian overlap approximation to the generator coordinate method ( gcm ) .
it is often simplified to obtain the expression @xmath19 we shall calculate the lifetimes with both eq [ eq : atdhfb ] and [ eq : gcm ] and compare .
we note that ref @xcite compares several forms of the mass , including eq [ eq : atdhfb ] and [ eq : gcm ] , in the context of the skyrme functionals .
it is also important to mention the dependence of the mass with the amount of pairing correlations : it has been shown in refs @xcite that the mass is inversely proportional to some power of the pairing gap .
it means that the stronger the pairing correlations are the smaller the mass is .
zero point energy corrections to the hfb energy @xmath20 are also considered in the atdhfb and gcm approaches .
in addition , the rotational energy correction computed following the phenomenological prescription of ref @xcite is also substracted .
this correction is very important to the shape of the potential energy as its value increases with deformation and can reach several mev for large deformations .
the spontaneous fission half life is computed with the standard wkb formalism of quantum mechanics . in the wkb formalism
the @xmath21 is given ( in seconds ) by the formula @xmath22 the action @xmath23 along the @xmath14 constrained path is given by @xmath24 where the integration limits correspond to the classical turning points for the energy @xmath25 .
for the collective quadrupole inertia @xmath12 we have used both the atdhfb and the gcm expressions .
the results obtained with the two different theories can differ in several orders of magnitude as the atdhfb mass is known to be a factor in between 1.5 and 2 larger than the gcm mass .
the potential energy @xmath26 is given by the hfb mean field energy corrected by zero point energies as described above , @xmath27 .
finally , an additional parameter @xmath28 is added to the ground state energy @xmath29 .
it is meant to represent the true ground state energy obtained after considering quantal fluctuations in the quadrupole degree of freedom .
this quantity could be estimated to be half of the square root of the curvature around the minimum divided by the collective inertia but it is often taken as a free parameter or kept fixed at some reasonable value .
we have followed the later approach with @xmath30 mev and estimated the impact of considering a larger value by repeating the calculations with @xmath31 mev .
the spontaneous fission half life obtained in this way is subject to several uncertainties that can lead to differences of several orders of magnitude .
the uncertainties are : 1 ) the height of the inner fission barrier gets reduced when triaxial shapes are allowed in the mean field calculation .
the amount of reduction is typically of a couple of mev , but it can show some isotopic variation ( see @xcite for a recent account in the actinide region ) ; 2 ) the value of @xmath28 .
it can also make a difference in @xmath0 specially for long lived isotopes where the fission barrier is wide .
the reason is that the value of @xmath28 alters the classical turning points ; 3 ) the values of the correlation energy corrections to the hfb energy included in @xmath26 .
the values are computed under certain assumptions and approximations and a better estimation of their values can lead to some changes to @xmath26 ; 4 ) the approximations involved in the evaluation of the collective masses can lead to differences of the order of 40 or 50 % ; 5 ) the pairing correlation .
it is an important ingredient both in the evaluation of the zero point energy as well as in the evaluation of the collective inertia . as shown below
, changes of a few percent in the pairing strength values can lead to changes in the theoretical estimation of the half lives in the range of five to twelve orders of magnitude . on the other hand , the experimental values of the parameters defining the potential energy of the fission process , namely the inner and outer barrier heights and the excitation energy of the fission isomer are more robust quantities to compare with as they are not as sensitive to pairing correlations as the other parameters .
however , these experimental quantities are obtained by model dependent assumptions that can mask the physical meaning of the parameters .
therefore , although we have compared our values with the experimental ones , we prefer to compare the trends in spontaneous fission life times as a function of n and z rather than the absolute values of the lifetimes themselves . we will also study the impact of changing various quantities entering the wkb formula .
obtained with three different interactions for the @xmath32u nucleus .
the bcpm edf ( black curves , bullet symbol ) as well as the d1s ( red , times symbol ) and d1 m ( blue curves , plus symbol ) parametrizations of the gogny force are included . in panel
a ) the hfb energy is given .
in panel b ) the particle - particle correlation energy @xmath33 is plotted for protons ( dashed lines ) and neutrons ( full lines ) for the three different kinds of calculations . in panel c ) the octupole and hexadecapole moments are given .
finally , panel d ) the atdhfb collective inertia is depicted .
[ fig : comparison ] ]
u for three different quadrupole deformation parameters indicated in each panel .
contour levels correspond to densities between 0.02 @xmath34 and 0.16 @xmath34 in steps of 0.02 @xmath34 .
the calculations correspond to the bcpm edf .
[ fig : contdens ] ] but for different pairing strengths .
the pairing strengths are given in terms of the reference value and a scaling parameter @xmath35 taking the values 1.00 ( the standard calculation ) , 1.05 and 1.10.[fig : pairing ] ] before comparing with the experimental data a comparison with the hfb results obtained with the two gogny interactions , namely , d1s and d1 m , is in order .
the gogny d1s interaction has been used in a thorough study of heavy nuclei properties , including fission , in ref @xcite and it has proved to reproduce quite nicely most of the properties analyzed . on the other hand , the fission properties of d1 m @xcite have not been analyzed in detail yet but its nice behavior regarding other aspects of nuclear structure like binding energies@xcite , radii @xcite quadrupole @xcite and octupole @xcite properties make it a good candidate to compare with . as we have already made comparison with d1s concerning fission properties @xcite of actinides ( @xmath3pu ) and super - heavies ( @xmath36sg )
, we will just explore another actinide : the nucleus @xmath32u . in panel
a ) of fig [ fig : comparison ] we compare the hfb energy as a function of @xmath14 for the three functionals in the nucleus @xmath32u .
we observe how the shape of the three curves starting at @xmath37b look rather similar up @xmath38 b , apart from a constant shift of a few mev . from there on , the d1 m interaction declines more gently than d1s .
the bcpm results are in between d1 m and d1s but closer to d1 m than to d1s .
the decline of energy with increasing quadrupole moment is correlated with the surface energy coeficient in nuclear matter which is larger in d1 m and bcpm than in d1s @xcite .
the other three curves only present at large @xmath14 values correspond to hfb solutions with two well separated fragments .
they intersect the one fragment curves at @xmath14 values around 130 b and they show a fast quasi - linear decrease in energy as the quadrupole moment increases .
the charge and mass of each fragment are the ones that lead to the minimum energy for each quadrupole moment .
for very large @xmath14 values the distance between the fragments is larger than the range of the nuclear strong interaction and therefore only the coulomb repulsion energy between the fragments plays a role . in this two - fragment regime
the quadrupole moment is directly linked to the separation distance between fragments @xcite and therefore increasing the quadrupole moment is equivalent to separating further away the fragments reducing the total energy of the system as a consequence of the reduction of coulomb repulsion .
although the two - fragment curves seems to intersect the one - fragment ones this is just a consequence of projecting out paths in a multidimensional space of collective variables ( quadrupole , octupole , hexadecapole , necking , etc ) into an one dimensional plot ( see below ) .
there is a minimum action path with a ridge connecting both the one and the two fragments curves that goes along the multidimensional space .
this path contributes to the action that enters the wkb formula to compute the @xmath39 half lives .
as the determination of this path is cumbersome and its contribution to the action is small we will neglect its contribution to the action .
this simplification amounts to consider both curves as really intersecting ones .
a measure of pairing correlations , the pairing interaction energy @xmath33 , is shown as a function of the mass quadrupole moment for protons ( dashed lines ) neutrons ( full lines ) in panel b ) of fig [ fig : comparison ] . again
the shape of the curves look rather similar for the three functionals but bcpm yields lower values of @xmath40 than d1s and d1 m .
those lower values correspond to less intense pairing correlations in the bcpm case with a severe quenching of the proton s pairing correlations . in panel c ) , the octupole and hexadecapole moments as a function of @xmath14 are given .
the results for the three edfs are very similar , to the extent that they appear as a single curve for many @xmath14 values .
the full curves correspond to the one - fragment solutions whereas the long dashed ones correspond to the scissioned configurations . as mentioned before , the values of the multipole moments of the two kind of curves are quite different and therefore the paths are quite separated in the multidimensional space of parameters .
finally , in panel d ) the collective inertia in the atdhfb approximation is plotted . as in the case of the hfb and particle - particle energies the shape of the curves for the different interactions
are quite similar but the bcpm inertia is up to a factor of three larger than the inertias obtained with the gogny forces .
the large value of the bcpm inertia is a direct consequence of the quenched pairing correlations : less pairing correlations imply a lower gap and therefore smaller two quasiparticle energies .
as the two quasiparticle energies enter the collective inertia in the denominator , quenched pairing correlations imply enlarged collective inertias .
on the other hand , the d1 m inertia is around 15 % smaller than the d1s one , consistent again with the quenched pairing correlations in d1 m as compared to the d1s ones .
the gcm inertias , not depicted , look rather similar in shape to the atdhfb ones but are a factor of 0.5 - 0.6 smaller .
the inertia for the two fragment solutions corresponds to the reduced mass of the two fragments @xcite and is constant with quadrupole deformation . in the calculation of the @xmath0 half life with the wkb formula
the configuration with the lowest energy is always choosen .
this is an approximation that neglects the path in the multidimensional space that connects the one fragment with the two fragment solution and therefore the @xmath0 obtained are to be considered as lower bounds . on the other hand
, triaxiallity leads to a reduction of the inner barrier height that is somehow compensated by an increase in the corresponding inertia .
this effect has not been included in the present calculation .
the values of the @xmath0 half life obtained for the three edfs computed with the gcm inertias are @xmath41s , @xmath42s and @xmath43s for bcpm , d1 m and d1s , respectively .
the large differences observed of up to 15 orders of magnitude can be attributed partly to the difference in the hfb energy curve but mostly to the very different values of the collective inertias .
the previous @xmath0 values have been obtained without taking into consideration the reduction of the inner barrier height as a consequence of triaxiallity . also , increasing the value of the @xmath28 parameter to 1.5 mev reduces the half lives by 6 , 2 and 4 orders of magnitude respectively . in any case , the values obtained are several orders of magnitude larger than the experimental value of @xmath44 s expect for gogny d1s that is very close to experiment .
if the atdhfb inertias are used instead of the gcm ones a much longer lifetimes are obtained : @xmath45s , @xmath46s and @xmath47s for bcpm , d1 m and d1s , respectively . this tendency to produce longer lifetimes when the atdhfb inertias are used is common to all the isotopes considered in this study .
the atdhfb inertias are typically a factor 1.5 larger than the gcm ones ( see refs @xcite for examples ) implying a 20 % increase in the action and therefore a 20% increase in the exponent of the lifetimes .
this is a source of theoretical uncertainty in the evaluation of @xmath0 that deserves further investigation .
another source of uncertainty comes from the fact that the inertias are computed in the `` cranking approximation '' where the energy of elementary excitations is replaced by the sum of hfb quasiparticle energies .
the approximation @xcite can lead to overestimations in the inertias as large as 40 - 50% for ground state configurations .
given the impact of these effects on the fission observables a better quantitative understanding is highly desirable . in the following , to simplify the presentation , we will consider only the gcm inertias in the evaluation of the lifetimes .
finally in fig [ fig : contdens ] contour plots of the densities for three values of the quadrupole moment are depicted .
they are obtained in calculations with bcpm and differ little from the same quantities computed with d1 m and d1s . for the quadrupole moment
@xmath14=130 b two densities , corresponding to the one fragment and two fragment solutions are presented .
the two fragment solution shows a spherical fragment that corresponds to z=51.60 and n=82.00 and an oblate deformed fragment with z=40.40 and n=60.00 .
the non integer proton and neutron number is due to the existence of low density nuclear matter between the two fragments .
the oblate and slightly octupole deformed fragment ( @xmath48 and @xmath49 acquires this shape to minimize the rather large coulomb repulsion energy ( assuming point like fragments the classical repulsion energy amounts to 196 mev ) .
this is an interesting result because it is commonly assumed @xcite that the fission fragments can only have prolate shapes .
.fission barrier height parameters @xmath50 ( inner ) and @xmath51(outer ) as well as excitation energy of the fission isomer @xmath52 .
the three parameters are given in mev .
the theoretical values have been obtained from the rotational energy corrected hfb potential energy surface .
the experimental values are taken from @xcite for the @xmath52 and from @xcite for the @xmath53s.[tab : fission - barrier - heigth ] [ cols="^,^,^,^,^,^,^",options="header " , ] in the bcpm functional the pairing interaction is taken as a density dependent contact pairing interaction with strength parameters fixed to reproduce the neutron matter pairing gap of the gogny force @xcite .
we have shown in the previous subsection that the particle - particle correlation energy , a quantity related to the amount of pairing correlations , was much smaller for bcpm than for the gogny forces leading to much larger collective inertias .
it is therefore reasonable to investigate the behavior of fission properties as a function of the pairing strength for the same functional . to this end , a parameter
@xmath35 has been introduced as a multiplicative factor in front of the pairing gap field @xmath54 . for the sake of simplicity
we have considered an unique parameter for both protons and neutrons although different parameters will give more flexibility to adjust the experimental data .
the outcome of the calculations with @xmath35 values of 1.05 and 1.10 for the nucleus @xmath32u are presented in fig [ fig : pairing ] .
we observe in panel a ) that increasing the pairing strength by 10 % ( @xmath55 leads to an overall gain of the order of 1 mev in binding energy .
for the ground state the 1 mev energy gains has to be compared to the 1 mev of pairing correlation energy for the standard bcpm .
the gain is even larger ( 1.6 mev ) for the configuration with @xmath56 b and corresponding to the top of the inner barrier .
however , the standard bcpm pairing correlation energy is 2.06 mev for that configuration . the net effect of increasing the pairing strength by 10% is to decrease the inner barrier eight ( @xmath57 by 0.6 mev whereas the other parameters , namely the outer barrier height @xmath51 and the fission isomer excitation energy remain more or less the same .
the particle - particle correlation energies @xmath40 shown in panel b ) for protons and neutrons increase with increasing @xmath35 but the slope is larger for neutrons than for protons .
the multipole moment values depicted in panel c ) do not change at all when the pairing strenght is increased and the different values lie on top of each other for different @xmath35 values .
finally , the impact on the collective inertia is clearly visible in panel d ) and is associated to the inverse dependence of the mass on the square of pairing gap @xcite . increasing the pairing strength by 5 % reduces the collective inertia by roughly 30 % whereas a 10 % increase leads to a reduction of 50 % .
the consequences of these reductions on the @xmath0 are dramatic , decreasing its value by 11 orders of magnitude in going from @xmath58 to @xmath59 ( @xmath60 s ) and six additional orders of magnitude in going from @xmath59 to @xmath61 ( @xmath62 s ) .
this result is a clear indication of the very important role played by pairing correlations in the description of fission .
the result suggests that experimental fission data could be used to fine tune the pairing strength instead of more traditional approaches based on odd - even staggerings . from a theoretical perspective
the result also points to the very important role that the correlations associated to particle number symmetry restoration should have in fission dynamics . restoring particle number symmetry usually leads to larger pairing correlations than the ones present at the mean field level and therefore will have a tremendous impact on fission lifetimes . in this respect
it is worth mentioning that the dependence on density of bcpm is on integer powers of the density allowing the use of the regularization techniques suggested to solve some technical problems associated to the evaluation of the energy kernel overlaps required by symmetry restoration theories ( see refs @xcite and references therein ) . given the large variability of the lifetimes with the different parameters entering the wkb formula , a direct comparison with the experimental data is meaningless and only comparisons with the trends along a series of nuclei or isotopes , all of them computed in the same conditions , can lead to meaningful conclusions regarding fission properties .
half lives ( bullets ) are compared to different theoretical results ( open symbols ) for several isotopic chains where experimental data exists .
the @xmath39 are plotted as a function of the fissibility parameter @xmath63 .
see text for details .
[ fig : spontaneus - fission - half ] ] ) , neutron ( @xmath64 ) and mass number ( @xmath65 ) of the two fragments emitted by the fission of the parent actinides ( perpendicular labels on top ) .
the magic numbers 50 and 82 as well as its sum are highlighted with horizontal lines .
[ fig : fragmassdistacinited ] ] in order to validate bcpm as a functional able to describe fission properties , we have performed calculations for those even - even nuclei where the spontaneous fission half life has been measured .
we will also compare the parameters defining the theoretical potential energy surface , namely the inner and outer barrier heights ( @xmath50 and @xmath51 ) and the excitation energy of the fission isomer @xmath52 with available experimental data @xcite .
it has to be mentioned that the experimental data for @xmath50 and @xmath51 @xcite is model dependent and therefore less reliable that the pure @xmath0 data . in table
[ tab : fission - barrier - heigth ] the experimental and theoretical values for @xmath50 , @xmath51 and @xmath52 are given for all nuclei where experimental data exists @xcite .
the theoretical values have been obtained by considering the hfb energy as a function of @xmath14 with the rotational energy correction ( computed in the way described in the previous section ) subtracted .
the effect of the zero point energy correction @xmath20 has not been included mainly because it is almost constant as a function of @xmath14 .
we notice that the theoretical predictions for @xmath50 are typically one or two mev larger than experiment .
this is not surprising as it is well known that the theoretical inner fission barrier is affected by triaxiallity and its height typically decreases by one or two mev when the effect is included in the calculation @xcite .
triaxiallity is not included at present because we still do not have access to a triaxial code incorporating the bcpm functional but work in this direction is in progress .
the situation is slightly better in the comparison with the @xmath52 and @xmath51 values . for them ,
no significant triaxial effects are expected and the agreement with experiment is better than for the @xmath50 . in ref
@xcite a thorough comparison of these data with various model predictions has been made . in this paper , the rms deviations for the fission isomer energy and second barrier height are given for several mean field models .
the bcpm values @xmath66 mev and @xmath67 mev are similar in magnitude to the ones of unedf1 @xcite a skyrme variant specifically tailored to describe fission .
this is a quite satisfactory result taking into account that bcpm does not use any fission data in its fit . in fig
[ fig : spontaneus - fission - half ] the theoretical @xmath0 results obtained for different choices of the @xmath28 and @xmath35 parameters are compared to the known experimental values .
the experimental @xmath0 values @xcite span a range of 27 orders of magnitude for a mass range a=232 - 286 .
the theoretical predictions , not including triaxial effects and computed with the gcm masses and zero point energies , span an even larger range of values and show a large variability depending upon the choices for the parameters . focusing on the `` standard '' theoretical values
@xmath58 and @xmath30 we observe differences with the experiment of up to 16 orders of magnitude for the lighter nuclei that steadily decrease to differences of just a couple of orders of magnitude for the heavier ones .
the largest differences are observed for nuclei with higher and wider barriers where the impact of parameters like @xmath28 is larger .
the comparison in isotopic chains indicate that the trend with neutron number compares much better with the experiment than the absolute values .
the same conclusion can be extracted from the overall trend with mass number obtained from the table .
therefore , we conclude that the hfb predictions , although subject to large uncertainties due to uncontrolled approximations in the evaluation of the different parameters , can be used to guess with a reasonable precision the trends of @xmath0 with mass number . the second conclusion drawn from this plot is the extreme sensitivity of the half lives to changes in @xmath35 and @xmath28 : increasing the pairing strength by 10% ( @xmath68 ) decreases @xmath0 by several orders of magnitude . in the uranium isotopes
the reduction is of 12 orders of magnitude bringing the theoretical predictions on top of the experimental data . on the contrary , in the fm and
no isotopic chains the reduction represents only 6 orders of magnitude but worsens the agreement with experiment . on the other hand ,
the increase of @xmath28 from 1 mev to 1.5 mev also reduces @xmath0 by several orders of magnitude , but the reduction is not as severe as with the increase of pairing strength . in the uranium case ,
the reduction represents on the average 6 orders of magnitude .
incidentally , the @xmath0 values obtained with @xmath69 and @xmath31 mev are in most of the cases very close to the results ( not shown ) corresponding to @xmath59 and @xmath70 mev .
the sensitivity of the results to the pairing strength demands a theory beyond hfb to describe pairing correlations .
a first candidate would be particle number restoration supplemented with configuration mixing using the pairing gaps as collective coordinates .
also the sensitivity to the @xmath28 parameter justifies an effort to better understand its rationale .
this is obviously a task for the future . for some neutron rich uranium isotopes .
the energies have been shifted upwards for the heavier isotopes in order to fit all the curves in a single plot . along with the one fragment curves starting at @xmath37 the curves corresponding to the energy of the two fragments resulting of fission are given.[fig : hfb - uranium ] ] another important piece of relevant information is the mass distribution of the fission fragments .
the mass of the fragments is determined by the nuclear shape in the neighborhood of the scission point . as the scission point is difficult to characterize in a mean field theory that explores just a few degrees of freedom , we have preferred to take a different approach that involves the evaluation of quasi - fusion configurations .
they are obtained by constraining the number of particles in the neck of the parent nucleus @xmath71 to a small value and then releasing the constraint to do a self - consistent calculation .
most of the time the self - consistent solution ends up in a solution with two well separated fragments . to make sure that the configuration is the lowest in energy
the procedure is repeated with different choices of the neck operator parameters @xmath72 ( position of the neck along the @xmath9 direction ) and @xmath73 ( the width of the neck distribution ) .
those configurations are constrained to larger quadrupole moments in order to separate the fragments ( please remember that for two fragments the quadrupole moment is proportional to the separation of the fragments @xcite ) .
an example of those quasi - fusion curves has already been presented in fig [ fig : comparison ] . in fig
[ fig : fragmassdistacinited ]
the proton ( z ) and neutron ( n ) numbers of the fragments obtained for the actinides considered are given .
the z values of the fragments are mostly determined by the z=50 magic number except in @xmath74fm . also for the heaviest nucleus
considered @xmath75fl larger z values are observed . for neutrons
, the magic n=82 seems also to be dominant neutron number but here the exceptions are more numerous . for the plutonium and heavier isotopes
the heaviest fragment has a mass number between 130 and 132 , ten unit less than the average experimental value of refs @xcite .
the discrepancy can be attributed to the lack of quantum fluctuations in our model that can modify substantially the raw mean field numbers @xcite .
obviously , the numbers given here are meant to represent the peaks of the fragments mass distribution which is a broad distribution as a consequence of exchange of particles during the scission process as well as a consequence of neutron evaporation .
a better dynamical theory is required ( see @xcite as an example of such theory ) in order to reproduce the experimental broad distribution . in the previous section we concluded that the description of fission based on the hfb theory is subject to large uncertainties coming from the poor understanding of the way the different quantities entering the wkb formula should be computed .
however , we also concluded that the hfb theory is reproducing reasonably well the experimental @xmath0 trends with mass number .
encouraged by the result , we have performed calculations in the uranium isotopic chain from the light uranium @xmath1u up to the neutron drip line corresponding to @xmath2u with the aim of understanding and analyzing the trends in spontaneous fission half lives and the mass of the emerging fission fragments . to illustrate the results the hfb potential energies for the @xmath3u , @xmath76u , @xmath77u , @xmath78u and @xmath79u isotopes
are depicted as a function of @xmath14 in fig [ fig : hfb - uranium ] .
the energies of the heavier isotopes have been shifted by different amounts of energy ( 55 , 100 , 135 and 160 mev , respectively ) to fit all the curves in a single plot .
also the curves corresponding to the two fragment solution with the lowest energy are depicted .
half lives ( bullets ) are compared to different theoretical results ( open symbols , fission ; asterisk @xmath80-decay ) for the uranium isotopes up to the drip line nucleus @xmath2u .
[ fig : tsfuranium ] ] we observe the ground state evolution from a quadrupole deformed ground state in @xmath3u with @xmath81 to an spherical one for @xmath78u ( corresponding to n=178 ) and for @xmath79u .
it is also worth mentioning the existence of a second fission isomer in @xmath3u ( excitation energy of 3.9 mev ) and @xmath76u ( excitation energy of 3.8 mev , lower than the excitation energy of the first isomer ) .
the situation in @xmath77u is not as well defined as in the previous cases and there are three very swallow minima .
the second one could be associated to the first fission isomer that is shifted to larger quadrupole moment values and zero octupole moment .
two fission isomers reappear in @xmath78u but both are located at a very high excitation energy ( around 9 mev ) and different quadrupole deformations than the ones in the light uranium isotopes . as a consequence of the increasing height and widening of the fission barriers as the neutron number approaches the neutron drip line we expect increasing @xmath0 values as can be observed in the next figure . in fig
[ fig : tsfuranium ] the spontaneous fission half lives @xmath0 of the uranium isotopes and computed with different choices of the @xmath35 and @xmath28 parameters are plotted as a function of mass number a. as in previous cases , the @xmath0 values have been obtained with the gcm collective mass and not taking into account the effects of triaxiallity in the first barrier .
the usual range of up to 12 orders in magnitude depending on the choice of parameters is observed .
however , the trend with mass number a is the same in all the four sets of parameters considered .
this again gives us confidence on the validity of the conclusions extracted from the trends with a. a decrease in the @xmath0 values is observed up to mass number a=256 where it becomes a steady increase with mass number up to @xmath82u where the @xmath0 values reach a maximum that corresponds to a neutron number of 184 that corresponds to one of the classical magic numbers .
the two neutron separation energy drops by 3 mev in going from a=276 ( 4.99 mev ) to a=278 ( @xmath83 mev what is a clear indication of extra stability . the half lives for this and the other isotopes beyond @xmath77u are very large and the corresponding nuclei can be considered as stable against the spontaneous fission decay channel . as the bcpm functional has been created to give a reasonable description of masses , it is reasonable to use its predictions for the binding energies of uranium and thorium to compute the half lives of @xmath80-decay using the phenomenological viola - seaborg formula @xcite .
the results for the uranium isotopes are plotted as asterisks in the figure .
we observe a steady increase of @xmath84 with mass number that reaches its maximum at a=264 where @xmath80 decay is no longer favorable energetically . from a=244 on fission
is faster than @xmath80 decay . ) , neutrons ( @xmath64 ) and mass number ( @xmath65 ) of the fission fragments is plotted as a function of the mass number a of the parent uranium isotope.[fig : fragmassdistu ] ] in fig [ fig : fragmassdistu ] the proton , neutron and mass numbers of the fragments in the fission of the uranium isotopes are plotted as a function of mass number a. those numbers are obtained by integrating the densities of each of the fragments coming out of the two fragment ( fusion valley ) self - consistent solutions mentioned before and corresponding to the lowest energy . in some cases ,
there are additional fusion valleys with different fragments but they lie higher in energy .
the issue of how to describe dynamically the evolution of the system through those valleys is a very interesting subject of research ( see for instance @xcite ) with many practical applications ( the real fission fragment mass distribution ) but exceeds the scope of the present paper .
the numbers discussed below are to be taken as the ones of the peaks ( most favorable mass ) of the mass distribution of fragments prior to neutron emission . except for @xmath85u and @xmath82u
the number of neutrons in one of the fragments always corresponds to the magic number n=82 .
the neutron number of the other fragment varies linearly accordingly to the mass number of the parent . on the other hand ,
the number of protons , that is close to the magic number z=50 for the light isotopes ( in good agreement with experiment @xcite ) varies linearly with mass number except for the isotopes @xmath86 u where it stabilizes at z=50 and for @xmath85u and @xmath82u ( symmetric fission ) where it is distributed equally between the two fragments .
also for the @xmath87u isotope a symmetric splitting with equal fragments is obtained .
concerning the mass distribution of the fragments , the heavy fragment has a mass number around 136 for the light isotopes that decreases with the mass number of the parent until the @xmath87u isotope is reached . at this point
the mass number of the heavy isotope starts to increase linearly with the exception of @xmath85u and @xmath82u .
this change in tendency is due to the increasing number of neutrons : as the number of neutrons increases , the fragment with n=82 is no longer the heaviest one . in any case , the behavior beyond a=256 is a rather academic issue as fission for those isotopes is a extremely unlikely process as discussed above .
the fission properties of several actinides and super - heavy nuclei have been computed with the recently proposed bcpm edf .
the theoretical results for the spontaneous fission half lives show a large variability consequence of uncertainties in the evaluation of some parameters of the theory and also on the strong dependence of the collective inertia with pairing correlations . as a consequence of the large uncertainties in the theoretical results we are only able to compare with the experimental data trend with mass number ( for instance the reduction by 27 orders of magnitude in the spontaneous fission half lives in going from a=232 to a=286 ) .
the theoretical predictions seem to reproduce such trend giving us confidence in the convenience of the method and edf for the study of fission properties of neutron rich uranium isotopes .
there we find that the spontaneous fission half lives remain more or less constant up to a=260 where they increase enormously as a consequence of the proximity to the magic neutron number n=184 .
therefore , it is confirmed the prevalence of this magic number in a extreme neutron rich case . on the other hand , a comparison of the parameters defining the potential energy surface for fission ( inner and outer barrier heights and fission isomer excitation energies ) with the model dependent
`` experimental data '' show a rather good agreement that gives us additional confidence on the validity of our conclusions .
the results obtained clearly show that more attention has to be paid to a proper description ( including beyond mean field effects ) of pairing correlations in the configurations relevant to fission .
the evaluation of the quasi - fission valley in the hfb model also allows to predict the peaks of the mass distribution of fission fragments .
it is shown that the magic proton number z=50 and the magic neutron number n=82 play an important role in determining the mass of the fragments . to conclude
, our study has shown the applicability of the bcpm edf for fission studies in heavy and super - heavy nuclei .
we have also pointed out the large variability of the theoretical predictions to the models used to evaluate the relevant parameters . however
, this variability seems to respect the trend with mass number of the spontaneous fission half life and therefore we have applied our method to study fission properties of the uranium isotopes up to the neutron drip line .
it is shown that beyond a=264 the uranium isotopes can be considered as stable against fission and @xmath80-decay . | fission dynamics properties of the barcelona - catania - paris - madrid ( bcpm ) energy density functional are explored with mean field techniques .
potential energy surfaces as well as collective inertias relevant in the fission process are computed for several nuclei where experimental data exists .
inner and outer barrier heights as well as fission isomer excitation energies are reproduced quite well in all the cases .
the spontaneous fission half lives @xmath0 are also computed using the standard semiclasical approach and the results are compared with the experimental data . the experimental trend with mass
number is reasonably well reproduced over a range of 27 orders of magnitude .
however , the theoretical predictions suffer from large uncertainties when the quantities that enter the spontaneous fission half life formula are varied .
modifications of a few per cent in the pairing correlation strengths strongly modify the collective inertias with a large impact on the spontaneous fission lifetimes in all the nuclei considered .
encouraged by the quite satisfactory description of the trend of fission properties with mass number we explore the fission properties of the even - even uranium isotope chain from @xmath1u to @xmath2u .
very large lifetimes are found beyond a=256 with a peak at neutron number n=184 . |
consider the singular boundary value problems for the emden
fowler equation @xmath2 & \alpha u(0)-\beta u'(0)=0,\quad \gamma u(1)+\delta u'(1)=0,\end{aligned}\ ] ] where @xmath3 and @xmath4 and may be singular at @xmath5 .
when @xmath6 , see @xcite for the result concerning the above problem .
when @xmath7 , @xcite shows the existence and uniqueness to ( 1 ) and ( 2 ) in the case of @xmath8 by means of the shooting method . for the following problem @xmath9 & \alpha u(0)-\beta u'(0)=0,\quad \gamma u(1)+\delta u'(1)=0,\end{aligned}\ ] ] where @xmath10 , @xmath11 and @xmath12
mao @xcite gave a sufficient and necessary condition when @xmath13 .
in this paper we shall consider the case of @xmath14 for the problems ( 3 ) and ( 4 ) .
a function @xmath15\cap c^2 ( 0,1)$ ] is a positive solution of ( 3 ) and ( 4 ) if @xmath16 satisfies ( 3 ) and ( 4 ) and @xmath17
we state the following hypothesis , which is used throughout this paper .
we now state the main results of this paper as follows : suppose that ( h ) is satisfied
. then suppose that ( h ) is satisfied .
then problems @xmath19 and @xmath20 have a @xmath21 $ ] positive solution if and only if the following inequalities hold .
( h1 ) @xmath22\ , \hbox{\rm d}t < \infty , \ \
\beta=0 , \delta\not=0.\ ] ] ( h2 ) @xmath23\ , \hbox{\rm d}t < \infty,\ \
\beta\not=0 , \delta=0.\ ] ] ( h3 ) @xmath24\ , \hbox{\rm d}t < \infty,\ \beta\!=\!\delta\!=\!0.\ ] ]
first we prove theorem 2.1 .
we will prove all the necessary conditions first then all the sufficient conditions .
_ case i _ : @xmath25 let @xmath26\cap c^2 ( 0,1 ) $ ] is a positive solution of ( 3 ) and ( 4 ) . from ( 4 ) and the nontrivial concave function @xmath27 , we know that @xmath27 must satisfy the following case : @xmath28 then there exists @xmath29 $ ] with @xmath30 @xmath31 yield @xmath32.$ ] let @xmath33 be a constant which satisfies @xmath34,$ ] and @xmath35 then @xmath36 q(t)u^{-m}(t ) & \geq q(t)(4c_0)^m.\end{aligned}\ ] ] by means of ( 12 ) and ( 13 ) , we have @xmath37\,\hbox{\rm d}s\\[.5pc ] & \geq \int_t^{t_0 } [ ( 4c_0)^{\lambda}p(s)+(4c_0)^m q(s)]\,\hbox{\rm d}s\\[.5pc ] & \geq ( 4c_0)^{\min\{\lambda , m\}}\int_t^{t_0 } [ p(s)+q(s)]\,\hbox{\rm d}s , \ \
t\in[0,t_0)\\[.5pc ] -u'(t ) & = \int^t_{t_0 } [ p(s)u^{-\lambda}(s)+q(s)u^{-m}(s)]\,\hbox{\rm d}s \\[.5pc ] & \geq ( 4c_0)^{\min\{\lambda , m\}}\int^t_{t_0 } [ p(s)+q(s)]\,\hbox{\rm d}s,\ \ \ t\in [ t_0,1].\end{aligned}\ ] ] so , @xmath38\,\hbox{\rm d}s \\[.5pc ] & = \int^{t_0}_0 [ p(s)+q(s)]\,\hbox{\rm d}s + \int^1_{t_0 } [ p(s)+q(s)]\,\hbox{\rm d}s\\[.5pc ] & \leq ( 4c_0)^{-\min\{\lambda , m\}}[u'(0)-u'(1)]\\[.5pc ] & < \infty.\end{aligned}\ ] ] therefore , ( 5 ) holds . _ case ii _ : @xmath39 let @xmath40\cap c^1 ( 0,1]\cap c^2 ( 0,1)$ ] be a positive solution of ( 3 ) and ( 4 ) . from ( 4 ) we obtain
@xmath41 then by the concavity of @xmath16 there exists @xmath42 $ ] with @xmath43 let @xmath44 be a constant satisfying @xmath45 by means of ( 12 ) and ( 13 ) , we obtain , @xmath46\,\hbox{\rm d}s = \int_0^{t_0 } \,\hbox{\rm d}\tau\int^{t_0}_{\tau}[p(s)+q(s)]\,\hbox{\rm d}s\\[.5pc ] & \leq ( 4c_1)^{-\lambda}\int^{t_0}_0 \,\hbox{\rm d}\tau\int^{t_0}_{\tau } p(s)u^{-\lambda}(s)\,\hbox{\rm d}s\\[.5pc ] & \quad
\ + ( 4c_1)^{-m}\int^{t_0}_0 \,\hbox{\rm d}\tau \int^{t_0}_{\tau}q(s)u^{-m}(s)\,\hbox{\rm d}s\ ] ] @xmath47\,\hbox{\rm d}s\\[.5pc ] & = ( 4c_1)^{-\min\{\lambda , m\ } } \int^{t_0}_0 ( -u^{\prime}(t_0)+u^{\prime}(\tau))\,\hbox{\rm d}\tau\\[.5pc ] & = ( 4c_1)^{-\min\{\lambda , m\}}u(t_0)\\[.5pc ] & < + \infty.\end{aligned}\ ] ] similarly , @xmath48\,\hbox{\rm d}s < + \infty.\ ] ] hence we conclude @xmath49\,\hbox{\rm d}t < \infty.\ ] ] _ case iii _ : @xmath50 .
the proof for case iii is almost the same as that for case ii .
_ case iv _ : @xmath51 let @xmath52 $ ] be a positive solution of ( 3 ) and ( 4 ) . integrating ( 3 )
twice gives @xmath53\,\hbox{\rm d}s,\nonumber\\[.5pc ] u'\left(\frac{1}{2}\right)\left(t-\frac{1}{2}\right ) - u(t ) + u\left(\frac{1}{2}\right ) & = \int^t_{{1}/{2 } } \,\hbox{\rm d}\eta \int^{\eta}_{{1}/{2 } } [ p(s)u^{-\lambda}(s)\nonumber\\[.5pc ] & \quad\ , + q(s)u^{-m}(s)]\,\hbox{\rm d}s\nonumber\\[.5pc ] & = \int^t_{{1}/{2 } } ( t - s)[p(s)u^{-\lambda}(s)\nonumber\\[.5pc ] & \quad \ , + q(s)u^{-m}(s)]\,\hbox{\rm d}s.\end{aligned}\ ] ] since the limit of ( 23 ) as @xmath54 exists and is finite , by the monotone convergence theorem , @xmath55\,\hbox{\rm d}s < \infty.\ ] ] so , @xmath56\,\hbox{\rm d}s <
\infty.\ ] ] similarly @xmath57\,\hbox{\rm d}s < \infty.\ ] ] hence @xmath58\,\hbox{\rm d}s < \infty.\ ] ] _ case i _ : @xmath25 suppose that ( 5 ) is satisfied .
let @xmath59 \,\hbox{\rm d}s\\[.5pc ] & \quad + \frac{\alpha t+\beta}{\rho}\int^1_t ( \gamma(1-s)+\delta)[p(s)+q(s)]\,\hbox{\rm d}s,\ t \in[0,1].\end{aligned}\ ] ] then @xmath60 \cap c^2 ( 0,1 ) $ ] satisfies ( 4 ) and solve the equation @xmath61 @xmath62 , t\in ( 0,1).$ ] let @xmath63\,\hbox{\rm d}s , l_2= 1/\rho \int^1_0 ( \alpha s+\beta)$ ] @xmath64\,\hbox{\rm d}s.$ ] then it is easy to check that @xmath65 @xmath66.$ ] let @xmath67 @xmath68,$ ] where @xmath69 @xmath70then @xmath71 @xmath72 \cap c^2 ( 0,1 ) , 0 < \alpha(t ) \leq \beta(t),$ ] @xmath73 , $ ] and @xmath74 satisfy the boundary condition ( 4 ) .
furthermore , @xmath75 + p(t)[k_1 q_1 ( t)]^{-\lambda}\\[.5pc ] & \quad \ + q(t)[k_1 q_1 ( t)]^{-m } \\[.5pc ] & \quad \
\geq p(t)[(k_1l_2)^{-\lambda}- k_1]\\[.5pc ] & \quad
\ + q(t)[(k_1l_2)^{-m } - k_1]\\[.5pc ] & \quad \
\geq 0,\ \ t\in(0,1),\\[.5pc ] \beta^{{\prime\prime } } ( t)+ p(t)\beta^{-\lambda}(t ) + q(t)\beta^{-m}(t ) & = - k_2[p(t)+q(t)]+ p(t)[k_2 q_1(t)]^{-\lambda}\\[.5pc ] & \quad \ + q ( t)[k_2q_1(t)]^{-m}\\[.5pc ] & \quad \ \leq p(t)[(k_2l_1)^{-\lambda}-k_2]\\[.5pc ] & \quad \ + q(t)[(k_2l_1)^{-m}-k_2]\\[.5pc ] & \quad \
\leq 0,\ \ t\in(0,1).\end{aligned}\ ] ] thus , @xmath76 and @xmath77 are respectively lower and upper solutions of problems ( 3 ) and ( 4 ) .
we will now prove that problems ( 3 ) and ( 4 ) admit a @xmath21 $ ] positive solution @xmath78 satisfying @xmath79.$ ] first , define an auxiliary function @xmath80 p(t)u^{-\lambda}(t)+ q(t)u^{-m}(t ) , & \alpha\leq u\leq \beta(t),\\[.5pc ] p(t)\beta^{-\lambda}(t)+ q(t)\beta^{-m}(t ) , & u > \beta(t ) .
\end{cases}\ ] ] from ( h ) , @xmath81 is continuous .
consider the boundary value problem @xmath82 & \alpha u(0)- \beta u^{\prime}(0)= 0,\quad \gamma u(1)+ \delta u^{\prime}(1)= 0.\end{aligned}\ ] ] it is clear that the above problem is equivalent to the integral equation @xmath83 where @xmath84 ( \alpha t+\beta)(\gamma(1-s)+\delta ) , \
t\geq s. \end{cases}\ ] ] let @xmath85.$ ] for @xmath86 if for some @xmath73\ \alpha(t)\leq
u(t ) \leq \beta ( t)$ ] , we obtain that @xmath87 & \leq p(t)(k_1l_1)^{-\lambda}+ q(t)(k_1l_1)^{-m}\leq ( k_1l_1)^{-\min\{\lambda , m\ } } [ p(t)+q(t)].\end{aligned}\ ] ] therefore , from ( 5 ) , ( 14 ) and ( 18 ) , we know that @xmath88 is continuous and @xmath89 is a bounded set .
in addition , @xmath90 $ ] is a solution of problems ( 15 ) and ( 16 ) if and only if @xmath91 . since @xmath92,$ ] by the standard application of arzela
ascoli theorem , we obtain that @xmath93 is compact . by means of schauder fixed point theorem , we obtain that @xmath93 has at least one fixed point @xmath94.$ ] we will show @xmath95,\ ] ] which will imply that @xmath96 $ ] is a positive solution of ( 3 ) and ( 4 ) .
suppose ( 19 ) is not satisfied .
then there exists @xmath97 $ ] such that either @xmath98 or @xmath99 let us consider the second case .
let @xmath100 $ ] denote the maximal interval containing @xmath101 such that @xmath102 on i. then , it is clear that either @xmath103 on @xmath104 or both @xmath16 and @xmath105 satisfy the same boundary conditions given by ( 4 ) on @xmath104 .
let @xmath106 then @xmath107 on @xmath108 and either @xmath109 for @xmath110 or @xmath111 satisfies the boundary condition ( 4 ) for @xmath110 . from the maximum principle @xmath112 for @xmath113 for @xmath114
this is a contradiction . in the same way , @xmath115 for @xmath116 so , @xmath117 is a @xmath21 $ ] positive solution of ( 3 ) and ( 4 ) .
_ case ii _ : @xmath39 suppose ( 6 ) is satisfied .
choose @xmath118 so that @xmath119 .
let @xmath120\,\hbox{\rm d}s\right.\\[.5pc ] & \left.\quad \ + \
t \int^1_t \frac{\gamma ( 1-s)+\delta}{\gamma+\delta}[p(s)+q(s)]\,\hbox{\rm d}s\right)^{1/(n\min\{\lambda , m\})}.\end{aligned}\ ] ] then @xmath121 \cap c^2 ( 0,1),$ ] satisfies @xmath122 and the boundary conditions @xmath123 .
we now estimate @xmath124r^{-\min\{\lambda , m\}}(t)\,\hbox{\rm d}t\\[.5pc ] & = \int^1_0 t[p(t)+q(t)]\bigg(\frac{\gamma(1-t)+\delta}{\gamma+\delta}\int^t_0 s[p(s)+q(s)]\,\hbox{\rm d}s\\[.5pc ] & \quad \ + t\int^1_t \frac{\gamma(1-s)+\delta}{\gamma+\delta}[p(s)+q(s ) ] \,\hbox{\rm d}s\bigg)^{-({1}/{n})}\,\hbox{\rm d}t\\[.5pc ] & \leq \int^1_0 t[p(t)+q(t ) ] \left ( \int^1_0 s[p(s)+q(s)]\,\hbox{\rm d}s\right)^{-({1}/{n})}\,\hbox{\rm d}t.\end{aligned}\ ] ] let @xmath125 \,\hbox{\rm d}s\\[.5pc ] & \quad\ + t\int^1_t \frac{\gamma(1-s)+\delta}{\gamma+\delta}[p(s)+q(s)]\,\hbox{\rm d}s,\\[.5pc ] \gamma_2(t ) & = \frac{\gamma(1-t)+\delta}{\gamma+\delta}\int^t_0 s[p(s)+q(s ) ] r^{-\min\{\lambda , m\}}\,\hbox{\rm d}s\\[.5pc ] & \quad \ + t\int^1_t \frac{\gamma(1-s)+\delta}{\gamma+\delta } [ p(s)+q(s)]r^{-\min\{\lambda , m\}}(s)\,\hbox{\rm d}s\\[.5pc ] & \quad
\ + r(t),\ \
t \in [ 0,1].\end{aligned}\ ] ] it is clear that @xmath126 and @xmath127,$ ] where @xmath128\left(\frac{\gamma(1-s)+\delta}{\gamma+\delta}\right)\,\hbox{\rm d}s,\\[.5pc ] l_4 & = \int^1_0 s\left(\frac{\gamma(1-s)+\delta}{\gamma+\delta}\right)[p(s)+q(s)]r^{-\min \{\lambda , m\}}(s)\,\hbox{\rm d}s\\[.5pc ] & \quad \
r_0 = \max_{t\in [ 0,1]}r(t).\end{aligned}\ ] ] we also check by direct computation that @xmath129,\
\gamma_2^{\prime\prime}(t ) \leq -(p(t)+q(t))r^{-\min\{\lambda , m\}}(t),\ \
( 0,1)$ ] .
let @xmath130,$ ] where @xmath131 then we have @xmath132+p(t)(k_1l_3)^{-\lambda}\\ & \quad \ + q(t)(k_1l_3)^{-m } \geq 0,\end{aligned}\ ] ] @xmath133r^{-\min\{\lambda , m\}}(t)\\ & \quad \ + p(t)(k_2 r(t))^{-\lambda}(t)\\ & \quad
\ + q(t ) ( k_2r(t))^{-m } \leq 0.\end{aligned}\ ] ] in addition , @xmath134 \cap c^1 [ 0,1 ] \cap c^2 ( 0,1 ) , \gamma\alpha(1)+\delta\alpha^{\prime}(1 ) = 0 , \gamma\beta(1 ) + \delta\beta^ { \prime}(1 ) \geq 0.$ ] let @xmath135 then @xmath136 also ,
if we assume @xmath137 on an interval @xmath100,$ ] we find @xmath138r^{-\min\{\lambda , m\ } } + k_1[p(t)+q(t ) ] \leq -k_2 [ p(t)+q(t)]l_4^{-\min\{\lambda , m\ } } + [ p(t)+q(t ) ] \leq 0 $ ] in @xmath108 .
as before , for @xmath139 , we have either @xmath140 or if @xmath141 from the maximum principle , @xmath142,$ ] i.e. @xmath143,$ ] which is a contradiction .
hence @xmath74 are respectively the lower and upper solutions of ( 3 ) and ( 4 ) . in the following
we will prove that problems ( 3 ) and ( 4 ) have a @xmath144 $ ] positive solution @xmath27 satisfying @xmath145 let @xmath146 be a sequence satisfying @xmath147 with @xmath148 as @xmath149 let @xmath150 be a sequence satisfying @xmath151 .
for each @xmath152 , consider the following singular boundary problem : @xmath153 & u(a_n ) = r_n,\quad \gamma u(1 ) + \delta u^{\prime}(1 ) = 0.\end{aligned}\ ] ] from ( 6 ) , we see that @xmath154\,\hbox{\rm d}s \geq \int^1_{a_n } s[p(s)+q(s)]\,\hbox{\rm d}s \geq a_n \int^1_{a_n } [ p(s)+q(s)]\,\hbox{\rm d}s.\ ] ] therefore , @xmath155\,\hbox{\rm
d}s \leq \frac{1}{a_n}\int^1_0 s[p(s)+q(s)]\,\hbox{\rm
d}s < \infty.\ ] ] following the proof of case i , with ( 22 ) , we can say that for each @xmath152 , the singular boundary value problems ( 20 ) and ( 21 ) have at least one positive solution @xmath156 $ ] satisfying @xmath157 $ ] .
@xmath158 & |u_n^{\prime}(1)| = |(\gamma/\delta)u_n(1)| \leq \frac{\gamma}{\delta}m,\ \ n = 1 , 2,\ldots.\end{aligned}\ ] ] without loss of generality , we can assume @xmath159 , \ \
n \arrow \infty,\\[.5pc ] u_n^{\prime}(1 ) & \arrow -(\gamma/\delta ) u_0 , \ \
n \arrow \infty.\end{aligned}\ ] ] similar to theorem 3.2 in @xcite , we can prove ( 3 ) has a positive solution @xmath27 with @xmath160 its maximal interval of existence is @xmath161 , and @xmath162 converges to @xmath163 uniformly in any compact subset of @xmath161 ( @xmath164 converges to @xmath165 uniformly ) , @xmath166 . since @xmath167
$ ] and @xmath168 = [ 0,1],$ ] we have @xmath169.$ ] from continuation theorem we obtain that @xmath72 \subset ( \omega^-,\omega^+)$ ] .
since @xmath170 we also obtain that @xmath171 .
in addition @xmath172 . thus @xmath173 \cap c^1 ( 0,1 ) \cap c^2 ( 0,1)$ ]
is a positive solution of problems ( 3 ) and ( 4 ) .
the proof for case ii is complete .
_ case iii _ : @xmath174 the proof for case iii is almost the same as that for case ii .
_ case iv _ : @xmath51 let @xmath175\,\hbox{\rm d}s + t\int^1_t ( 1-s ) [ p(s)$]@xmath176\,\hbox{\rm d}s)^{{1}/{n\min\{\lambda , m\}}},\ \
t\in [ 0,1].$ ] then @xmath177 \cap c^2 ( 0,1)$ ] with @xmath178@xmath179 and @xmath180 .
let @xmath181\,\hbox{\rm d}s + t\int^1_t ( 1-s)[p(s)+q(s)]\,\hbox{\rm d}s,\\[.5pc ] \gamma_2 ( t ) & = ( 1-t)\int^t_0 s[p(s)+q(s)]q^{-\min\{\lambda , m\}}(s)\,\hbox{\rm d}s\\[.5pc ] & \quad \ + t \int^1_t ( 1-s)[p(s)+q(s)]q^{-\min\{\lambda , m\}}(s)\,\hbox{\rm d}s + q(t ) \
t \in [ 0,1].\end{aligned}\ ] ] then @xmath182\,\hbox{\rm d}s , \ l_6 = \int^1_0 s(1-s)[p(s)+q(s ) ] q^{-\min\{\lambda , m\}}(s)\,\hbox{\rm d}s + q_0,\ \ q_0 = \max q(t ) .
$ ] let @xmath183 $ ] . here
@xmath184 then @xmath74 are respectively the lower and upper solutions of ( 3 ) and ( 4).the remaining proof is analogous to that of case ii .
thus we complete the proof of theorem 2.1 .
_ proof of theorem _ 2.2 proof for case ( h1 ) : @xmath185 .
suppose that @xmath16 is a @xmath21 $ ] positive solution of ( 3 ) and ( 4 ) .
then both @xmath186 and @xmath187 exist , and @xmath188 from ( 4 ) and the fact that @xmath16 is a positive concave function , we know that @xmath189 then there exists @xmath42 $ ] such that @xmath190 . since @xmath191 and
@xmath192 $ ] , one easily sees that @xmath193 $ ] and there exist constants @xmath194 and @xmath195 which satisfy @xmath196.\ ] ] hence @xmath197\,\hbox{\rm
d}s \leq i_2^{\max\{\lambda , m\ } } \int^1_0 [ p(s)u^{-\lambda}(s)+q(s)u^{-m}(s)]\,\hbox{\rm d}s\end{aligned}\ ] ] @xmath198 the above inequality shows that ( 9 ) holds .
suppose that ( 9 ) is satisfied .
let @xmath199\,\hbox{\rm d}s\\[1pc ] & \quad \
+ t\int^1_t \frac{\gamma(1-s)+\delta}{\gamma+\delta}[p(s)s^{-\lambda } + q(s)s^{-m}]\,\hbox{\rm d}s , t \in [ 0,1].\end{aligned}\ ] ] then @xmath200 \cap c^2 ( 0,1)$ ] .
replace @xmath27 with @xmath201 in ( 24 ) and let @xmath202\,\hbox{\rm d}s,\\[.8pc ] i_2 & = \int^1_0 [ p(s)s^{-\lambda}+q(s)s^{-m}]\,\hbox{\rm d}s.\end{aligned}\ ] ] then @xmath201 satisfies ( 24 ) .
let @xmath203 then @xmath204.$ ] then @xmath76 and @xmath77 are respectively lower and upper solutions of problems ( 3 ) and ( 4 ) .
it is clear that @xmath205 , \alpha(0 ) = \beta(0 ) = 0 , \gamma \alpha(1)+\delta\alpha'(1 ) = 0 , \gamma\beta(1)+\delta\beta'(1 ) = 0.$ ] on the other hand , when @xmath206 , we have @xmath207 & = p(t)(k_1 i_1 t)^{-\lambda } + q(t)(k_1 i_1 t)^{-m } = f(t).\end{aligned}\ ] ] from ( 9 ) , we have @xmath208 . the same argument that we have given in the sufficiency of theorem 2.1 assures us that problems ( 3 ) and ( 4 ) admit a positive solution @xmath209 \cap c^2 ( 0,1)$ ] such that @xmath210 $ ] .
the proof for case ( h1 ) is complete .
similarly we can prove cases ( h2 ) and ( h3 ) .
thus we complete the proof of theorem 2.2 .
this work is supported in part by the nsf(youth ) of shandong province and nnsf of china . | this paper investigates the existence of positive solutions of a singular boundary value problem with negative exponent similar to standard emden fowler equation . a necessary and sufficient condition for the existence of @xmath0 $ ] positive solutions as well as @xmath1 $ ] positive solutions
is given by means of the method of lower and upper solutions with the schauder fixed point theorem .
= msam5 at 9pt = tibi at 10.4pt [ theo]*theorem * # 1#2 |
the morphological division of extragalactic radio sources into two classes introduced by @xcite has proved to be remarkably robust .
fanaroff & riley had already noted that the edge - brightened ( frii ) sources tend to have higher radio luminosities than the edge - darkened ( fri ) sources , but there is also a dependence on stellar luminosity of the host galaxy which makes the division even cleaner @xcite .
that said , the fri class is not homogeneous : examples of the range of structures are shown in fig .
[ fig : fris ] .
although almost all fri sources show jets on small scales , some appear to be confined to the nuclear regions ( e.g. 3c84 ; fig .
[ fig : fris]a ) while others persist almost to the outer edges of the source structure ( e.g. 3c296 ; fig .
[ fig : fris]f ) . it has long been surmised ( e.g. @xcite ) that the characteristic structures of fri sources result from deceleration and disruption of their jets , perhaps triggered by interaction with the surrounding intergalactic medium . turning this perception from a vague idea into a quantitative description of source dynamics
proved to be a slow process , but significant progress has recently been made using a combination of deep radio imaging , sophisticated jet modelling and an understanding of the external environments of the sources from x - ray observations .
these results bear on five of @xcite s tasks : map jet velocity fields ; understand the changing composition ; measure jet pressures ; deduce jet confinement mechanisms and infer jet powers and thrusts .
deceleration of a symmetric , initially relativistic flow provides a very natural explanation for the decrease in jet / counter - jet sidedness ratio @xmath0 with distance from the nucleus observed in fri jets ( e.g. @xcite ) .
there is now direct evidence for relativistic motions of jet knots on kpc scales in m87 and cena @xcite and many other results are mostly easily explained if fri jet bases are relativistic .
for example , @xmath1 is correlated with core fraction ( * ? ? ? * fig .
[ fig : relclues]a ) and the counter - jet lobe depolarizes more rapidly with increasing wavelength than that associated with the brighter jet ( * ? ? ? * fig .
[ fig : relclues]b ) .
an example of the characteristically asymmetric appearance of fri jet bases is shown in fig .
[ fig : ipdata296](b ) . as well as this asymmetry , the main and counter - jets also show differences in their transverse brightness profiles , in the sense that the main jet is more centrally peaked ( e.g.fig . [ fig:3c296side ] ) .
this is a natural manifestation of a transverse velocity gradient , with higher flow speeds on - axis .
there is also a systematic difference in polarization structure between the main and counter - jets ( e.g.fig .
[ fig : ipdata296]a ) , in the sense that the brighter jet base has a smaller degree of polarization on - axis with more prominent regions of longitudinal apparent magnetic field .
if the jets are intrinsically symmetric and relativistic , this effect must be due to aberration , which causes the two jets to be observed at different angles to the line of sight in the rest frames of their emitting material and therefore to have different observed polarizations .
the combination of asymmetries in total intensity and linear polarization provides enough information to break the degeneracy between flow velocity and angle to the line of sight , as well as to constrain the three - dimensional structure of the magnetic field , allowing us to make detailed models of the jets .
we have now modelled the flows in five fri radio galaxies : 3c31 , b20326 + 39 , b21553 + 24 , ngc315 and 3c296 @xcite .
we assume that the jets are precisely symmetrical , axisymmetric , relativistic flows and derive their geometries , velocity fields , emissivity and magnetic - field component ratios by fitting to deep vla observations .
we find that a decelerating jet model fits the observed brightness and polarization structures and asymmetries well ( e.g. fig .
[ fig : ivec296 ] ) .
fri jets on kpc scales can generally be divided into a flaring region ( where the expansion rate first increases rapidly and then decreases again ) and an outer region in which the expansion is uniform .
we can not model the jets within a kpc or so of the nucleus , where they tend to be faint and poorly resolved by the vla .
where they first brighten , we derive characteristic on - axis velocities @xmath2 .
rapid deceleration then occurs over distances of 1 10kpc ( always before the jets recollimate ) , after which the velocity either stays constant or decreases less rapidly . for all sources except 3c296 ,
the transverse velocity profiles are consistent with edge / on - axis ratios @xmath30.7 everywhere ( although we can not rule out evolution from a top - hat profile at the initial brightening ) . in 3c296 , the only source with a bridged twin - jet structure that we have observed in sufficient detail ( fig .
[ fig : fris]f ) , the velocity falls to a low fractional value @xmath4 at the jet edge .
the jets in 3c296 may be embedded within the lobes rather than propagating in direct contact with the interstellar medium of the host galaxy , as appears to be the case for the other sources .
velocity fields for four of the sources are shown in fig .
[ fig : vels ] . the data hint at further complexity in the jet velocity fields before the initial brightening
. measurements of both sidedness ratio and proper motion on pc scales in ngc315 indicate an apparent _ acceleration _ ( * ? ? ?
[ fig : accel315 ] ) .
our vla observations also show slightly smaller sidedness ratios in the faint , kpc - scale jet bases than at or immediately after the brightening point for the four sources where we can resolve the former .
if we observe a jet stratified in velocity with inadequate transverse resolution , the speed inferred for its _ apparently _ brightest emission will be influenced by how changes in its transverse emissivity profile interact with the ( orientation - dependent ) doppler favouritism , and may give a misleading account of the true velocity variation along the jet .
both the pc - scale and faint kpc - scale regions are barely resolved in width , so we can not yet tell whether the bulk flow really accelerates . on average ,
the largest single magnetic - field component is toroidal .
the longitudinal component is significant close to the nucleus but decreases with distance and the radial component is always the smallest of the three ( e.g.fig .
[ fig : bgrey ] ) .
we also find that the toroidal component is stronger relative to the longitudinal component at the edge of the jet .
the evolution of the field component ratios along the jets before they recollimate is not consistent with flux freezing in a laminar flow , which requires a much more rapid transition from longitudinal to transverse field than we infer .
after recollimation , flux freezing is consistent with our results , however .
before the jets decelerate and recollimate , the quasi - one - dimensional adiabatic approximation ( together with flux freezing ) is grossly inconsistent with the emissivity evolution inferred from our models . after deceleration and recollimation , however , this approximation describes the emissivity evolution along the jets quite well .
details of the radio spectra and high - energy emission are given by @xcite .
it is likely that fri jets decelerate as a result of entraining matter , either by _ injection _ of mass lost from stars within the jet volume or by _ ingestion _ of ambient gas from the surrounding igm via a mixing layer .
the kinematic models of section [ models ] prescribe the variation of velocity along a jet , together with its geometry . given the external pressure and density profiles derived from x - ray observations , we can then apply conservation of energy , momentum and mass in the quasi - one - dimensional approximation , following @xcite and including the effects of buoyancy .
two key assumptions are needed to close the problem : the jet is assumed to be in pressure equilibrium with the surrounding medium after recollimation and @xmath5 , as expected for an initially relativistic jet ( @xmath6 is the energy flux with the rest - mass contribution subtracted and @xmath7 is the momentum flux ) . in addition , a relativistic equation of state is assumed and anisotropic magnetic stresses are neglected . the results of this calculation are estimates of the jet energy flux and the variations of internal pressure , density , mass flux and entrainment rate as functions of distance from the nucleus .
this technique was first applied to 3c31 by @xcite . in fig .
[ fig : cons ] we also show the preliminary profiles of pressure , density , mach number and entrainment rate for two other sources b20326 + 39 and 3c296 ( laing et al .
, in preparation ) .
the top panels show the internal , external and synchrotron minimum pressures . by hypothesis , the first two
are matched in the outer regions , but we also note that the pressure gradients are quite similar ( for 3c296 it is possible that the external pressure should be that of the lobe rather than the surrounding igm ) .
the synchrotron minimum pressure is comparable with or slightly less than the internal pressure in the outer parts , as expected if the latter is dominated by relativistic particles and fields close to equipartition . in all three sources
, there is a significant overpressure at the brightening point ( where the analysis starts ) and this drives the initial expansion in the flaring region .
the pressure becomes less than that of the surroundings at the end of the flaring region before the jet recollimates and reaches pressure equilibrium ( the large under - pressure in b20326 + 39 between 2 and 4kpc is probably due to an inaccurate estimate of the outer velocity , which is poorly constrained ) .
the jet densities are very low ( typically equivalent to 1 proton m@xmath8 ) , and the initial density ratios with the surroundings are @xmath910@xmath10 .
the jets are transonic , with mach numbers @xmath91 .
the entrainment rate for 3c31 continues to rise at large distances whereas the profiles for b20326 + 39 and 3c296 both show peaks in the flaring region before dropping to very low values .
the reason is that the kinematic model for 3c31 requires continued deceleration after recollimation , whereas those for the other two sources indicate an approximately constant asymptotic velocity .
differences in external density and/or shielding by lobe material may be responsible .
the entrainment rate profile for 3c31 also shows a prediction for the mass input from stars in the jet volume @xcite .
this is comparable with the required entrainment rate close to the start of the model , but not at large distances , suggesting that mass injection and ingestion are both important .
the derived energy fluxes are @xmath11 , @xmath12 and @xmath13w , respectively , for 3c31 , 3c296 and b20326 + 39 .
it will be of interest to compare these estimates with values derived from cavity dynamics ( * ? ? ? * and references therein ) .
estimates of the mass density and the number density of radiating electrons can be combined to constrain the composition of the jets ( see @xcite for a detailed review of assumptions ) . at the brightening point ,
limiting cases include : 1 .
a power - law lorentz - factor distribution @xmath14 of relativistic electrons , each associated with a ( cold ) proton . in this case
, there must be a low - energy cut - off at @xmath15 ( 3c31 and b20326 + 39 ) or @xmath16 ( 3c296 ) .
2 . an electron - positron jet with entrained thermal matter providing essentially all of the mass . unfortunately
, it is not yet possible to distinguish between these ( and other ) compositions .
the techniques reviewed here have led to a detailed picture of the evolution of the fri jets as they flare and decelerate .
they work because we can compare both jets in the same source : sizes match the resolution of our most sensitive instruments , surface brightnesses are ( just ) high enough , faraday rotation is easy to correct and jet speeds are modest .
the key project when evla and emerlin become operational will be an attempt to apply the same methods to other types of relativistic jet : closer to the launching scale and in frii sources or microquasars .
there are also more immediate questions , including : 1 . what makes fri jets brighten so abruptly in the radio band at distances of a few kpc from the nucleus ? 2
can we decide what combination of stellar mass - loss ( injection ) and external entrainment ( ingestion ) operates in fri jets ?
can we ( finally ) detect internal depolarization and hence estimate densities of thermal matter now that we understand foreground faraday screens much better ?
4 . do different methods for estimating fri jet powers ( conservation - law , cavity dynamics , bow - shock physics ) give the same answer ? can they be made consistent with radiation - loss timescales ?
do jets really accelerate on pc scales and/or as they flare or is this effect an artefact of inadequate transverse resolution ?
bicknell , g. v. , 1994 , apj , 422 , 542 biretta j.a . , zhou f. , owen f.n .
, 1995 , apj , 447 , 582 blandford , r.d .
, these proceedings canvin , j.r . , laing , r.a .
, 2004 , mnras , 350 , 1342 canvin , j.r . , laing , r.a . ,
bridle , a.h . , cotton , w.d .
, 2005 , mnras , 363 .
1223 cotton w.d . , feretti l. , giovannini g. , lara l. , venturi t. , 1999 , apj , 519 , 108 fanaroff b.l .
, riley j.m . , 1974 , mnras , 167 , 31p fanti , r. , lari , c. , parma , p. , bridle , a.h . , ekers , r.d . , fomalont , e.b . , 1982 , a&a , 110 , 169 hardcastle m.j . , worrall d.m . , kraft r.p . , forman w.r
, jones c. , murray s.s . , 2003 , apj , 593 , 169 harris , d.e . , cheung , c.c . , stawarz , l. , biretta , j.a . , sparks , w. , perlman , e.s . , wilson , a.s . , 2007 , these proceedings laing , r.a . , 1993 , in burgarella , d. , livio , m. , odea , c.p .
, eds , space telescope sci . inst .
6 : astrophysical jets .
cambridge university press , cambridge , p. 95
laing , r.a . ,
bridle , a.h .
, 2002a , mnras , 336 , 328 laing , r.a . ,
bridle , a.h .
, 2002b , mnras , 336 , 1161 laing , r.a . ,
canvin , j.r . ,
bridle , a.h . ,
hardcastle , m.j .
, 2006 , mnras , 372 , 510 laing , r.a . ,
bridle , a.h . , cotton , w.d . , worrall , d.m .
, birkinshaw , m. , 2007 , these proceedings laing , r.a . , parma , p. , de ruiter , h.r . ,
fanti , r. , 1999 , mnras , 306 , 513 leahy , j.p . ,
bridle , a.h . , strom , r.g . , 2000 , http://www.jb.man.ac.uk/atlas ledlow , m.j . ,
owen , f.n .
, 1996 , aj , 112 , 9 mcnamara , b.r . , brzan , l. , rafferty , d.a . , nulsen , p.e.j . ,
carilli , c. , wise , m.w . , these proceedings morganti , r. , parma , p. , capetti , a. , fanti , r. , de ruiter , h.r . , 1997 , a&a , 326 , 919 simon , a.j.b . , 1978 , mnras , 184 , 537 | there is now unequivocal evidence that the jets in fr i radio galaxies are initially relativistic , decelerating flows . on the assumption that they are axisymmetric and intrinsically symmetrical ( a good approximation close to the nucleus )
, we can make models of their geometry , velocity , emissivity and field structure whose parameters can be determined by fitting to deep vla observations .
mass entrainment either from stellar mass loss within the jet volume or via a boundary layer at the jet surface is the most likely cause for deceleration .
this idea is quantitatively consistent with the velocity field and geometry inferred from kinematic modelling and the external gas density and pressure profiles derived from x - ray observations .
the jets must initially be very light , perhaps with an electron - positron composition . |
we are concerned about the numerical solution of the following integrodifferential equation : @xmath0 @xmath1 , \bar{x } \in \omega \subset { \mathbb{r}}^2,\ ] ] where the unknown @xmath2 is a continuous function @xmath3 \rightarrow { \mathbb{r}}$ ] , @xmath4 , @xmath5 and @xmath6 are given functions ; @xmath7 is a constant .
we search for a solution @xmath8 of this equation which satisfies the initial condition @xmath9 along with equation we will also consider @xmath10 @xmath1 , \quad \bar{x } \in \omega \subset { \mathbb{r}}^2,\ ] ] where @xmath11 is a delay , depending on the spatial variables . in the last case , the initial condition has the form @xmath12,\ ] ] where @xmath13 . by integrating both sides of with respect to time on @xmath14 $ ] ,
we obtain the volterra - fredholm integral equation : @xmath15 @xmath1 , \bar{x } \in \omega \subset { \mathbb{r}}^2.\ ] ] the existence and uniqueness of solution of equation was proved in @xcite , both in the case of a smooth and discontinuous function @xmath6 .
an analytical study of equation was carried out in @xcite , where the authors have addressed the problems of existence , uniqueness and stability of solutions .
when solving numerically equations of the forms and , they are often reduced to the form ; therefore we begin by discussing literature on computational methods for volterra - fredholm equations . starting with the one - dimensional case , without delay
, brunner has analysed the convergence of collocation methods @xcite , while kauthen has proposed continuous time collocation methods @xcite . in @xcite an asymptotic error expansion for the nystrm method
was proposed , which enabled the use of extrapolation algorithms to accelerate the convergence of the method .
another approach was developped by z. jackiewicz and co - authors @xcite , @xcite , who have applied gaussian quadrature rules and interpolation the approximate the solution of integrodifferential equations modelling neural networks , which are similar to equation . in the two - dimensional case ,
the required computational effort to solve equations and grows very fast as the discretization step is reduced , and therefore special attention has to be paid to the creation of effective methods .
an important approach are the low - rank methods , as those discussed in @xcite , when the kernel is approximated by polynomial interpolation , which enables a significant reduction of the dimensions of the matrices . in @xcite ,
the authors use an iterative method to solve linear systems of equations which takes into account the special form of the matrix to introduce parallel computation . concerning equation , besides the existence and stability of solution ,
numerical approximations were obtained in @xcite .
the computational method applies a quadrature rule in space to reduce the problem to a system of delay differential equations , which is then solved by a standard algorithm for this kind of equations .
a more efficient approach was recently prpoposed in @xcite @xcite , where the authors introduce a new approach to deal with the convolution kernel of the equation and use fast fourier transforms to reduce signifficantly the computational effort required by numerical integration .
the above mentioned equations are known as neural field equations ( nfe ) and have played an important role in mathematical neuroscience for a long time .
equation was introduced first by wilson and cowan @xcite , and then by amari @xcite , to describe excitatory and inhibitory interactions in populations of neurons .
while in other mathematical models of neuronal interactions the function @xmath8 ( membrane potential ) depends only on time , in the case of nfe it is a function of time and space .
the function @xmath4 represents external sources of excitation and @xmath6 describes the dependence between the firing rate of the neurons and their membrane potential
. it can be either a smooth function ( typically of sigmoidal type ) or a heaviside function .
the kernel function @xmath16 gives the connectivity between neurons in positions @xmath17 and @xmath18 . by writing the arguments of the function in this form we mean that we consider the connectivity homogeneous
, that is , it depends only on the distance between neurons , and not on their specific location . according to many authors , realistic models of neural fields
must take into account that the propagation speed of neuronal interactions is finite , which leads to nfe with delays of the form . here
the delay @xmath19 is space dependent , that is , it depends on the distance between the positions @xmath20 and @xmath21 . in @xcite a new numerical algorithm for the solution of equation was introduced .
this algorithm combines a second order implicit scheme for the space discretization with gaussian quadrature and collocation methods for the space discretization .
we refer to this article for the detailed description of this method and the analysis of its convergence . in @xcite
this algorithm was applied to the solution of 3 problems with physical interest .
the aim of the present paper is to present some more applications of the above mentioned algorithm and to discuss its performance .
* example 1 . *
this algorithm is targeted directly to the application in neuroscience and robotics . as an illustration of such applications , we first consider a neural field described in @xcite , where the firing rate function has the form @xmath22 where @xmath23 , and the connectivity function
is given by @xmath24 where @xmath25 and @xmath26 .
it is known @xcite that when there is no external input ( @xmath27 ) , the stability of the trivial solution of this equation depends on the value of @xmath28on the parameters of the connectivity function and on the propagation speed . in particular , for each set of values of these parameters , it is possible to compute a bifurcation value @xmath29 , such that if @xmath30 the zero solution is asymptotically stable , and otherwise it is nt .
we have considered some test cases , in order to analyse how the behaviour of solutions depends on all these parameters . in each case , we have used our algorithm to approximate the solution with @xmath31 on the time interval @xmath14 $ ] , where @xmath32 depends on the specific example ( usually @xmath33 ) .
the parameters of the space discretization were @xmath34 , @xmath35 .
we have used as initial condition @xmath36 ( in the delay case , we have @xmath37 , if @xmath38 $ ] ) .
it is worth to note that according to @xcite all the solutions of equation ( [ 2 ] ) are bounded , so they can either tend to the trivial solution ( @xmath39 ) , to another steady state or just oscillate . *
case 1*. @xmath40 . in this case , the interaction between neurons is excitatory at short distances and inhibitory , at long ones ( so - called mexican hat connectivity ) . for this case , we consider two different values of @xmath41 : a ) @xmath42 , b ) @xmath43 ; and two different forms of propagation : ( i ) no delay ; ii ) v=1 .
the case @xmath43 , without delay , was studied in @xcite . there
the authors report that as time tends to infinity the solution tends to a certain stationary state , with a biperiodical pattern .
we have observed the same kind of behaviour , as can be seen in fig.1 . in this figure ( as in other figures of the same type below )
the graphs on the left - hand are the surface plots of the solution at equidistant moments in time ( @xmath44 ) , while the graphs on the left - hand side are the corresponding contour plots .
note that the solution takes a long time to attain the stationary point .
this can be observed in the graph of fig.2 .
as other graphs of the same type below , in this graph the plots of the fucntions @xmath45 and @xmath46 are displayed , as functions of time , where @xmath47 and @xmath48 are points located at the boundary and at the middle of the domain @xmath49 , respectively .
note that it is important that we apply an implicit method , which allows us to use a not very small stepsize in time .
the fact that the solution does not tend to the zero stationary state indicates that this state is in this case unstable .
we have also studied the same case with propagation speed @xmath50 . in this case
the solution has a very different behaviour .
it also tends to some non - zero stationary state , but obviously a different one ( see fig.3 ) . when @xmath42 , without delay , the behaviour of the solution is typical to the case of a stable zero state : it tends rapidly to 0 ( see fig .
though during a certain time it exhibits a biperiodical pattern ( see fig.4 ) , we observe that the amplitude decreases very fast with time . in the case
@xmath50 , the solution with @xmath42 has a similar behaviour to the no - delay case ( see fig.6 ) .
as could be expected , the main difference is that in this case the solution decays not so fast ( see fig.7 ) . *
case 2*. let @xmath51 . in this case
, we also have mexican - hat connectivity , but with a stronger excitatory component for this example , we also consider two different values of @xmath41 : a)@xmath52 , b ) @xmath42 ; and two different forms of propagation : ( i)no delay ; ii)@xmath53 . when @xmath42 , without delay , the solution tends to a pattern , which is periodic with respect to @xmath20 and constant with respect to @xmath21 ( see fig .
this suggests that the zero solution is unstable .
the evolution of the solution at a point in the boundary and another at the midlle is displayed in fig .
9 . in this graph
we observe clearly that the solution tends to a nonzero stationary state .
the graphs of the solution for the case @xmath42 , but with @xmath53 , are displayed in fig.10 .
we again observe convergence to a nonzero stationary state , which is confirmed by the graphs in fig .
. however , this is clearly not the same stationary state as in the non - delay case . when @xmath52 we observe a quite different situation .
the graphs for this case , with no delay ( see fig.12 ) , show clearly the convergence of the solution to the zero state .
the evolution is plotted in the graphs of fig.13 . finally , in fig .
14 the graphs of the solution are plotted in the case @xmath52 , @xmath50 .
the corresponding evolution graphs are plotted in fig . 15 . in this case
, the graphics also suggest that the solution tends to the zero state , but not so fast .
@xmath54 the results of the numerical simulations concerning example 1 are summarized in table 1 . both in case 1 and case 2 , when the value of @xmath41 is higher than a certain threshold we observe the convergence of the solution to certain non - trivial stationary states , while for lower values of @xmath41 the solution converges to the zero state , which in this case is stable .
this is in agreement with the results presented in @xcite , where it is mentioned that there is exists a bifurcation value for @xmath41 .
concerning the influence of propagation speed , we observe that in the case of a stable zero state the solution converges to this state also in the case of a finite propagation speed , though the convergence is slower in this case ( as it could be expected ) . in the case of the nontrivial stationary states
, the solution converges to different states , depending on the propagation speed .
the main difference between case 1 and case 2 is that in case 2 the critical value of @xmath41 ( for which the zer state becomes unstable ) is lower .
we recall that both in case 1 and case 2 we have mexican hat connectivity , but in case 2 the parameters @xmath55 and @xmath56 are higher .
* example 2 .
* we are now concerned about a neural field , which is described in @xcite . in this case
we have a connectivity kernel of the form : @xmath57 where @xmath58 , @xmath59 and @xmath60 , @xmath61 and @xmath62 are positive constants .
in our numerical simulations we have used @xmath63 , @xmath64 , @xmath65 .
this kernel reflects spatial hexagonal connections , which have been found in the visual cortex of monkeys @xcite ( see fig .
16 ) . in order to test our program ,
we have reproduced the numerical simulation carried out by hutt and rougier in @xcite .
the firing rate function in this case has the form of the sigmoidal function @xmath66 and the external input is given by @xmath67 with @xmath68 and @xmath69 .
note that this external input is localized at the domain center , so that the function @xmath4 has very high derivatives in the neighborhood of this point .
this property is reflected by the solution , which ( as can be seen in fig .
17 ) has a very thin peek at the domain center .
this creates some difficulties to the application of our program .
namely , we can not apply the rank reduction technique , described in @xcite , to reduce the size of the matrices .
it is not possible in this case to approximate accurately the solution by an interpolating polynomial with a small number of interpolation points .
we have overcome this difficulty by applying our algorithm without rank reduction . actually , since the discretization method converges fast , it is possible to obtain an accurate approximation of the solution with a moderate number of gridpoints and a large stepsize in time ( in our simulation , we have used @xmath35 and @xmath70 ) . in this way , it is possible to carry out numerical simulations with a not too high computational effort .
the resolution of the solution graph can then be improved , obtaining additional solution values by interpolation . with this purpose ,
we have used cubic interpolating polynomials at each interval @xmath71 $ ] . just as reported in @xcite
, we note that the activity of the neural field is first concentrated in the domain center and then propagates to the boundary , with a sequence of peeks , which exhibit the same hexagonal pattern as the connectivity function .
the aim of the numerical experiments concerning this example is to investigate the influence of the delay @xmath72 on the solution of equation ( [ 2 ] ) . as described in sec . 1
, we assume that this delay results from the time spent by electric signals to travel between neurons .
therefore , it has the form @xmath73 where @xmath74 is the velocity of signal propagation ( which is supposed to be constant in time and uniform in space ) .
figures 17 - 20 illustrate the case where @xmath75 . in each figure
, we have a 3d graphic of the solution ( left - hand side ) and the corresponding contour plot ( right - hand side ) .
as we can see from these figures , at each instant of time , it is possible to identify a region of activity , that is , a region @xmath76 , such that if @xmath77 , then @xmath78 , where @xmath79 is some threshold value of the potential , under which we consider that there is no significant activity . in our experiments , we have considered @xmath80 , which was found experimentally and seems to be a good criterium .. as mentioned before , this domain has the approximate form of a hexagon ( reflecting the characteristics of the connectivity function ( [ ker2 ] ) ) .
we define the radius of activity @xmath81 by @xmath82 the figures show clearly that @xmath81 is an increasing function , and our purpose is to use the numerical results to analyse this function and investigate how it depends on the delay and , in particular , on the velocity @xmath74 .
we note also that in the case of * no delay * ( which corresponds to infinite velocity @xmath74 ) the solution of the problem in the initial period ( @xmath83 ) has much higher values in the center than in the periphery of the domain @xmath84 ( as it happens in the delay case ) .
however in this case there is no sharp transition between activity and non - activity zones , and therefore we can not identify clearly a domain of activity @xmath76 .
this case is illustrated by figure 21 , where the contour plot of the solution is displayed in the cases of @xmath85 and @xmath86 . in table 2
, we display the values of @xmath81 for @xmath87 in three different cases : @xmath75 , @xmath88 and no delay .
some interesting conclusions can be deduced from this table .
* in the first moments after the start of stimulation ( @xmath89 ) , the velocity @xmath74 seems not to influence the activity radius @xmath81 .
this is not surprising , because in the first moments the activity of the field is significant only in a small region and the effect of delay within this region very is small ( the time spent by signals to travel within this region does not play an important role ) .
the radius of activity in this case depends mainly on the properties of the connectivity kernel , which does not depend on @xmath74 .
* however , as time passes , the situation changes .
for @xmath90 , we see that in the case of no delay the radius of activity is @xmath91 , which means that from this moment there is activity in the whole domain . in the delay case
, @xmath92 continues to grow with @xmath93 , which physically means that there is a wave front moving from the center to the periphery of the domain .
* this wave front does not move with a constant speed ( the differences between @xmath94 are not proportional to the time interval lengths ( @xmath95 .
* however , when we compare the cases @xmath75 and @xmath88 ( for @xmath96 ) , we see that the average propagation speed of the wave front ( defined as @xmath97 ) is nearly the double in the second case , compared with the first one .
this means that for sufficiently large time , the transmission delay is the main factor for the definition of the activity radius .
again that can be explained by the fact that when we have a large activity domain , the transmission time between the neurons has a strong influence in the dynamics of the neural field . .
the numerical results were obtained from a numerical simulation using the method under discussion , with @xmath75 , with stepsize @xmath98 and @xmath35 . in each figure , the contour plot displayed on the right - hand side corresponds to the 3d - graphic given on the left - hand side .
@xmath99{surf16 } & \includegraphics[clip = true , trim= 2 cm 9 cm 2 cm 9 cm , width=8cm]{image16 } \end{array}\ ] ] @xmath99{surf48 } & \includegraphics[clip = true , trim= 2 cm 9 cm 2 cm 9 cm , width=8cm]{image48 } \end{array}\ ] ] @xmath99{surf72 } & \includegraphics[clip = true , trim= 2 cm 9 cm 2 cm 9 cm , width=8cm]{image72 } \end{array}\ ] ] @xmath99{surf96 } & \includegraphics[clip = true , trim= 2 cm 9 cm 2 cm 9 cm , width=8cm]{image96 } \end{array}\ ] ] @xmath99{image3nodelay } & \includegraphics[clip = true , trim= 2 cm 9 cm 2 cm 9 cm , width=8cm]{image5nodelay } \end{array}\ ] ] @xmath100 * example 3 .
* in this example we consider a kind of neural field with spatially localized oscillations which occur in excitable neural media with short - range excitation and long - range inhibition ( mexican hat connectivity ) , in the case of a spatially localized input .
such neural fields are known as _
breathers _ and their dynamics are analysed in @xcite .
a similar example was described and computed in@xcite . in this case
the connectivity kernel has the form : @xmath101 the firing rate function has the the sigmoidal form @xmath102 and the external input is given by @xmath103 our aim is to investigate how the dynamic behaviour of the solution is influenced by the propagation speed of the impulses , so we consider different values of @xmath74 . as initial condition
we consider @xmath104 .
note that when there is no delay ( infinite propagation speed ) the solution tends to a stable stationary state . in the case of finite propagation speed ,
oscillatory solutions occur .
the frequency of the oscillations decreases as the propagation speed becomes lower . as an example , in fig.22 the graphs of the solution maximum and minimum are depicted , as a function of time , in the case @xmath105 .
we have carried out simulations over the time interval @xmath106 $ ] with stepsize in time @xmath107 .
the parameters of the space discretization were @xmath108 , @xmath35 . in fig .
23 and 24 , the 3d - graphs of the solution , in the case @xmath105 , are displayed for different moments of time . from the pictures it is clear that the solution has oscillatory behaviour and the amplitude of oscillations decreases with time . as referred in @xcite , we also can see a symmetry breaking dynamical instability , when a two - dimensional radially symmetric stationary pulse bifurcates to a nonradially symmetric breather with an integer number of lobes ( in this case 4 )
in this paper we present some numerical simulations obtained by an algorithm which was previously described in @xcite and @xcite . the three numerical examples considered here have been analysed before by other authors and are directly connected with real world applications .
the main purposes are 1 ) to test the performance of the mentioned algorithm , by comparing the numerical results with those obtained by other authors ; 2 ) to analyse with more detail the properties of the solutions and take conclusions about their physical meaning .
the numerical experiments have shown that the computational method applied in these simulations is an efficient tool and provides sufficiently accurate results to analyse the physical phenomena under consideration .
all the numerical simulations were carried out in a personal computer with a intel processor of 2.4 ghz .
the computing time for each time step is about 10 - 20s , depending on the example considered .
the interpolation method used to reduce the matrix size has proved to be very efficient in the cases where the input data are sufficiently smooth . however , in the case of an external input which changes very fast in the neighborhood of a certain point ( like in example 2 ) this method may not be applicable . in such cases the algorithm can still be used but the computing time is much higher .
this suggests that it would be useful in such cases to consider the use of nonuniform meshes , which take into account the specific behaviour of the solution .
such improvement of the algorithm will be the subject of future research .
this research was supported by a marie curie intra european fellowship within the 7th european community framework programme ( pief - ga-2013 - 629496 ) . h. brunner , on the numerical solution of nonlinear volterra - fredholm integral equations by collocation methods , siam j. numer .
27 ( 1990 ) 987 - 1000 .
a. cardone , e. messina , and e.russo , a fast iterative method for discretized volterra - fredholm integral equations , j. comput . applied math .
189 ( 2006 ) 568 - 579 .
g. faye and o. faugeras , some theoretical and numerical results for delayed neural field equations , physica d 239 ( 2010 ) 561578 .
han guoqiang , asymtotic error expansion for a nonlinear volterra - fredholm integral equation , j. comput . applied math .
59 ( 1995 ) 49 - 59 .
a. hutt and n. rougier , numerical simulations of one- and two - dimensional neural fields involving space - dependent delays , in s. coombes et al .
, eds . , neural fields theory and applications , springer , 2014 .
p. m. lima and e. buckwar , numerical solution of the neural field equation in the two - dimensional case , arxiv ref .
http://arxiv.org/abs/1508.07484 , 2015 .
p.m. lima and e. buckwar , numerical investigation of the two - dimensional neural field equation with delay , to appear in proceedings of ieee .
j. lund , a. angelucci , p. bressloff , anatomical substracts for functional columns in macaque monkey primary visual cortex , cerebral cortex , * 13 * ( 2003 ) 15 - 24 .
r. pothast and p. beim graben , existence and properties of solutions for neural field equations , math .
, 33 ( 2010 ) 935 - 949 . | in the present paper we are concerned with a numerical algorithm for the approximation of the two - dimensional neural field equation with delay .
this algorithm was described in @xcite and @xcite .
we consider three numerical examples that have been analysed before by other authors and are directly connected with real world applications .
the main purposes are 1 ) to test the performance of the mentioned algorithm , by comparing the numerical results with those obtained by other authors ; 2 ) to analyse with more detail the properties of the solutions and take conclusions about their physical meaning . |
in this section , we discuss potential measurements at the lhc that can constrain the pomeron structure . the pomeron structure in terms of quarks and gluons has been derived from qcd
fits at hera and at the tevatron and it is possible to probe this structure and the qcd evolution at the lhc in a completely new kinematical domain . the high energy and luminosity at the lhc allow the exploration of a completely new kinematical domain .
one can first probe if the pomeron is universal between @xmath0 and @xmath1 colliders , or in other other words , if we are sensitive to the same object at hera and the lhc .
the different diagrams of the processes that can be studied at the lhc are namely double pomeron exchange ( dpe ) production of dijets , of @xmath2jet , sensitive respectively to the gluon and quark contents of the pomeron , and the jet gap jet events .
all diagrams were included in the fpmc @xcite generator that was used for this analysis .
the dijet production in dpe events at the lhc is sensitive to the gluon density in the pomeron . in order to quantify how well we are sensitive to the pomeron structure in terms of gluon density at the lhc
, we display in fig .
[ d0 ] , left , the dijet cross section as a function of the dijet mass fraction @xcite , assuming for instance the protons to be tagged in the afp @xcite proton detectors at 210 m. the central black line displays the cross section value for the gluon density in the pomeron measured at hera including an additional survival probability of 0.03 .
the yellow band shows the effect of the 20% uncertainty on the gluon density taking into account the normalisation uncertainties .
the dashed curves display how the dijet cross section at the lhc is sensitive to the gluon density distribution especially at high @xmath3 . for this sake
, we multiply the gluon density in the pomeron from hera by @xmath4 where @xmath5 varies between -1 and 1 . when @xmath5 is equal to -1 ( resp .
1 ) , the gluon density is enhanced ( resp , decreased ) at high @xmath3 . from fig .
[ d0 ] , we notice that the dijet cross section is indeed sensitive to the gluon density in the pomeron and we can definitely check if the pomeron model from hera and its structure in terms of gluons is compatible between hera and the lhc
. this will be an important test of the pomeron universality .
this measurement can be performed for a luminosity as low as 10 pb@xmath6 since the cross section is very large ( typically , one day at low luminosity without pile up at the lhc ) .
[ d0 ] , right , displays a possible observable at the lhc that can probe the quark content in the pomeron , namely the @xmath2jet to the dijet cross section ratio @xcite as a function of diffractive mass ( @xmath8 ) for different assumptions on the quark content of the pomeron , @xmath9 varying between 0.25 and 4 in steps of 0.25 .
we notice that the cross section ratio varies by a factor 2.5 for different values of @xmath10 .
the aim of the diffractive mass distribution measurement is twofolds : is the pomeron universal between hera and the lhc and what is the quark content of the pomeron ?
the qcd diffractive fits at hera assumed that @xmath11 , since data were not sensitive to the difference between the different quark component in the pomeron .
the lhc data will allow us to determine for instance which value of @xmath9 is favoured by data .
let us assume that @xmath12 is favoured .
if this is the case , it will be needed to go back to the hera qcd diffractive fits and check if the fit results at hera can be modified to take into account this assumption .
if the fits to hera data lead to a large @xmath13 , it would indicate that the pomeron is not the same object at hera and the lhc . on the other hand ,
if the hera fits work under this new assumption , the quark content in the pomeron will be further constrained .
the advantage of measuring the cross section ratio as a function of diffractive mass is that most of the systematic uncertainties will cancel .
soft color interaction models ( sci ) is another model to explain diffraction at hadronic colliders @xcite . in fig .
[ d0 ] , right , we notice that the distribution of the @xmath14jet to dijet ratio as a function of the total diffractive mass distributions may allow to distinguish between the herwig / dpe and pythia / sci models because the latter leads to a more flat dependence on the total diffractive mass , giving further insight into soft qcd . in this process ,
both protons are intact after the interaction and detected in afp at 210 m , two jets are measured in the atlas central detector and a gap devoid of any energy is present between the two jets @xcite .
this kind of event is important since it is sensitive to qcd resummation dynamics given by the bfkl @xcite evolution equation .
this process has never been measured to date and will be one of the best methods to probe these resummation effects , benefitting from the fact that one can perform the measurement for jets separated by a large angle ( there is no remnants which ` pollute ' the event ) . as an example , the cross section ratio for events with gaps to events with or without gaps as a function of the leading jet @xmath15 is of the order of 20% which is much higher than the expectations for non - diffractive events .
this is due to the fact that the survival probability of 0.03 at the lhc does need to be applied for diffractive events .
in the standard model ( sm ) of particle physics , the couplings of fermions and gauge bosons are constrained by the gauge symmetries of the lagrangian .
the measurement of @xmath18 and @xmath19 boson pair productions via the exchange of two photons allows to provide directly stringent tests of one of the most important and least understood mechanism in particle physics , namely the electroweak symmetry breaking @xcite .
the parameterization of the quartic couplings based on @xcite is adopted .
the cuts to select quartic anomalous gauge coupling @xmath16 events are the following , namely @xmath20 for the tagged protons corresponding to the afp detector at 210 and 420 m , @xmath21 20 gev , @xmath22 between the two leptons .
in addition , a cut on the @xmath15 of the leading lepton @xmath23 gev and on the diffractive mass @xmath24 gev are requested since anomalous coupling events appear at high mass .
after these requirements , we expect about 0.7 background events for an expected signal of 17 events if the anomalous coupling is about four orders of magnitude lower than the present lep limit @xcite ( @xmath25 10@xmath26 ) for a luminosity of 30 fb@xmath6 , and about two orders of magnitude better than the present cms limits @xcite .
the search for quartic anomalous couplings between @xmath27 and @xmath18 bosons was performed again after a full simulation of the atlas detector including pile up @xcite assuming the protons to be tagged in afp at 210 m only .
integrated luminosities of 40 and 300 fb@xmath6 with , respectively , 23 or 46 average pile - up events per beam crossing have been considered . in order to reduce the background , each @xmath18 is assumed to decay leptonically ( note that the semi - leptonic case in under study ) .
the full list of background processes used for the atlas measurement of standard model @xmath16 cross - section was simulated , namely @xmath28 , @xmath16 , @xmath29 , @xmath17 , @xmath30jets , drell - yan and single top events .
in addition , the additional diffractive backgrounds were also simulated , since only leptonic decays of the w bosons are considered , we require in addition less than 3 tracks associated to the primary vertex , which allows us to reject a large fraction of the non - diffractive backgrounds ( e.g. @xmath28 , diboson productions , @xmath30jet , etc . ) since they show much higher track multiplicities .
remaining drell - yan and qed backgrounds are suppressed by requiring the difference in azimuthal angle between the two leptons @xmath31 3.1 . after these requirements , a similar sensitivity with respect to fast simulation without pile up was obtained . of special interest
will be also the search for anomalous quartic @xmath32 anomalous couplings which is now being implemented in the fpmc generator .
let us notice that there is no present existing limit on such coupling and the sensitivity using the forward proton detectors is expected to be similar as the one for @xmath33 or @xmath34 anomalous couplings .
if discovered at the lhc , @xmath32 quartic anomalous couplings might be related to the existence of extra - dimensions in the universe , which might lead to a reinterpretation of some experiments in atomic physics . as an example , the aspect photon correlation experiments @xcite might be interpreted via the existence of extra - dimensions .
photons could communicate through extra - dimensions and the deterministic interpretation of einstein for these experiments might be true if such anomalous couplings exist . from the point of view of atomic physics , the results of the aspect experiments would depend on the distance of the two photon sources .
99 m. boonekamp , a. dechambre , v. juranek , o. kepka , m. rangel , c. royon , r. staszewski , e - print : arxiv:1102.2531 ; m. boonekamp , v. juranek , o. kepka , c. royon `` forward physics monte carlo '' , `` proceedings of the workshop : hera and the lhc workshop series on the implications of hera for lhc physics , '' arxiv:0903.3861 [ hep - ph ] .
c. marquet , c. royon , phys .
rev . d*79 * ( 2009 ) 034028 ; o. kepka , c. royon , c. marquet , r. peschanski , eur .
c*55 * ( 2008 ) 259 - 272 ; phys .
lett . b*655 * ( 2007 ) 236 - 240 ; h. navelet , r. peschanski , c. royon , s. wallon , phys .
lett . b*385 * ( 1996 ) 357;h .
navelet , r. peschanski , c. royon , phys .
b366 ( 1996 ) 329 . | we describe different physics topics which can be performed at the lhc using tagged intact protons , namely a better understanding of the pomeron structure in terms of quarks and gluons , and the serach for quartic anomalous couplings . |
to trace gas and stars in the outskirts of galaxy disks , we typically use hi mapping ( e.g. , ( * ? ? ?
* huchtmeier & bohnenstengel 1981 ) ; ( * ? ? ?
* koribalski & lpez - snchez 2009 ) ; ( * ? ? ?
* heald et al . 2011 ) ; ( * ? ? ?
* serra et al . 2012 ) ; ( * ? ? ?
* lee - waddell et al .
2012 ) ; ( * ? ? ?
* koribalski et al . 2016 ) ) and deep , wide - field optical imaging ( e.g. , ( * ? ? ?
* martnez - delgado et al . 2010 ) ; ( * ? ? ?
* duc et al . 2015 ) ) . the hi disks of galaxies , which often extend a factor two or more beyond the bright stellar disk ( ( * ?
* warren et al .
2004 ) ) , are excellent tracers of their total mass ( visible and dark matter ) .
the gas kinematics allows us to model their 3d shapes and obtain rotation curves , from which the radial distribution and amount of dark matter is derived .
galex ultra - violet ( uv ) imaging of nearby galaxies led to the discovery of extended uv disks ( xuv - disks ) in the galaxies m83 and ngc 4625 ( ( * ? ? ?
* thilker et al . 2005 ) ; ( * ? ? ?
* gil de paz et al . 2005 ) ) , indicating star formation well beyond the radius where hii regions are typically found . here
i suggest that the hi distributions of xuv - disk galaxies extend about twice as far .
the term 2x - hi disk was introduced by @xcite , who found the giant hi disk of ngc 1512 to extend a factor two beyond its xuv - disk .
another prominent spiral galaxy with an 2x - hi disk is m83 ( ( * ? ? ?
* koribalski 2008 ) ; ( * ? ? ?
* jarrett et al . 2013 ) ; ( * ? ? ?
* koribalski 2015 ) ) .
cc & + c +
a search for more xuv disks by ( * ? ? ? * thilker et al .
( 2007 ) ) , revealed many more galaxies , among them the nearby , southern spirals ngc 300 ( ( * ? ? ?
* westmeier , braun & koribalski 2011 ) ) , ngc 1512 ( ( * ? ? ?
* koribalski & lpez - snchez 2009 ) ) , ngc 1672 ( ( * ? ? ?
* houghton 1998 ) ) , and ngc 3621 ( ( * ? ? ?
* walsh 1997 ) ) . atca hi mosaics reveal their 2x - hi disks , some of which are shown in fig.[koribalski.fig1 ] .
the galaxy hi diameters are as expected from their hi masses and lie firmly on the @xmath1 @xmath2 relation ( ( * ? ? ? * wang et al . 2016 ) ; wang , these proceedings ) . an hi atlas of nearby galaxies
is presented as part of the local volume hi survey ( lvhis ; ( * ? ? ?
* koribalski 2008 ) , ( * ? ? ?
* koribalski et al . 2016 ) ) .
lvhis consists of a complete sample of @xmath380 hi - rich galaxies in the southern sky , observed with the australia telescope compact array ( atca ) , @xmath330 hours each , and supplemented by multi - wavelength images
. the optical @xmath4 diameter of m83 s bright stellar disk is @xmath313@xmath5 ( 17 kpc at @xmath6 = 4.5 mpc ) .
deep optical photography by @xcite reveals a much larger stellar disk ( @xmath320@xmath5 ) as well as faint streams and loops further out , indicative of dwarf galaxy accretion .
the most prominent stellar stream lies @xmath320@xmath5 north of m83 s center . for comparison
, the hi disk of m83 , mosaiced by the atca ( see fig.[koribalski.fig1 ] ) , has a diameter of over 60@xmath5 ( 80 kpc ) , reveals a large tidal tail / arm wrapping at least 180 degrees around , from west to east , and shows signs of ram pressure stripping on the compressed , western side . also shown in fig.[koribalski.fig1 ] are ngc 3621 , which resembles m83 but seen at a larger inclination angle , and ngc 1512 , which is known to interact with its blue compact dwarf companion ngc 1510 .
galex provided a beautiful image of m83 s xuv disk ( ( * ? ? ?
* thilker et al . 2005 ) ) , with uv emission detected out to a radius of @xmath320@xmath5 .
the star - forming regions in m83 s outer disk agree well with the high density hi clumps , suggesting that hi is an excellent tracer of star formation in the galaxy outskirts ( see also ( * ? ? ?
* koribalski & lpez - snchez 2009 ) ; ( * ? ? ?
* for , koribalski & jarrett 2012 ) ) . a colorful overlay of m83 s galex xuv - disk with the vla single - pointing hi map by @xcite featured on one of the iau s321 conference posters , greeting participants every morning at our lovely venue in toledo , spain .
we note that the hi emission seen by the vla in m83 and ngc 3621 ( ( * ? ? ?
* walter et al . 2008 ) ) is only 22% and 77% , respectively , of the total hi flux , which is well established from single - dish imaging ( e.g. , ( * ? ? ?
* huchtmeier & bohnenstengel 1981 ) ; ( * ? ? ?
* koribalski et al . 2004 ) ) .
+ the askap hi all sky survey ( wallaby , dec @xmath7 + 30 degr ) is expected to detect over 500000 galaxies and provide well - resolved maps for @xmath35000 of these ( ( * ? ? ?
* koribalski 2012 ) ) . | the outskirts of galaxies especially the very extended hi disks of galaxies are strongly affected by their local environment .
i highlight the giant 2x - hi disks of nearby galaxies ( m83 , ngc 3621 , and ngc 1512 ) , studied as part of the local volume hi survey ( lvhis ) , their kinematics and relation to xuv disks , signatures of tidal interactions and accretion events , the @xmath0 relation as well as the formation of tidal dwarf galaxies . - using multi - wavelength data , i create 3d visualisations of the gas and stars in galaxies , with the shape of their warped disks obtained through kinematic modelling of their hi velocity fields . |
the theory of error - correcting codes for classical information has been extensively studied for almost fifty years . a fundamental question in coding theory concerns
what _ capacity _ of information can be successfully transmitted through a noisy channel . call a code that maps @xmath3-bit inputs into @xmath4-bit codewords an @xmath5 code , and define its capacity to be @xmath6 . for any specific @xmath4-bit channel ,
let its _ capacity _ be the maximum capacity of all @xmath5 codes that successfully transmit information through it .
this capacity is often taken as an asymptotic limit as @xmath4 tends towards infinity .
we consider two basic kinds of @xmath4-bit noisy channels .
the first is one that flips any subset of up to @xmath7 bits of each codeword that passes through it . in this model ,
transmission through the channel is considered successful if the @xmath3 bits of data can always be perfectly recovered .
call a code that achieves this a _ @xmath7-error - correcting code_. asymptotically , it is natural to take @xmath7 as some fixed fraction @xmath0 of @xmath4 , written as @xmath8 and understood to mean @xmath9 . a second model of a noisy channel is one where each bit of each codeword that passes through it is flipped independently with probability @xmath0 . in this model , commonly referred to as the _ binary symmetric channel _ , there is no absolute bound on the number of possible errors that occur .
therefore , a probabilistic definition of successful transmission is required . call a code
for which the probability of successful recovery for any @xmath3 bits of data is at least @xmath10 an _ @xmath11-error - correcting code_. the subject of error - correcting codes for _ quantum _ information is much younger , developing within the past couple of years , though it has received considerable attention during this time @xcite .
much of the above terminology extends naturally to quantum information by considering qubits instead of bits .
call a quantum code that maps @xmath3-qubit data to @xmath4-qubit codewords an @xmath12 code .
we need to specify the behavior of noisy quantum channels .
a natural quantum analogue of the first model is to allow any @xmath7 qubits of each codeword that passes through it to be altered
. we can take this to mean : apply an arbitrary unitary transformation to all the qubits selected for alteration . an apparently stronger definition allows the unitary transformation to also involve another set of qubits , representing an `` external environment '' , thereby simulating the effect of `` decoherence '' .
an apparently weaker definition limits the unitary operations to being among : @xmath13 , @xmath14 , and @xmath15 , the standard pauli spin matrices , and @xmath16 , the unit matrix .
it turns out that , by reasoning similar to that in @xcite , these three definitions of `` alter '' can be shown to be equivalent , in the sense that a code that is _ @xmath7-error - correcting _ with respect to the apparently weaker one will automatically be @xmath7-error - correcting with respect to the apparently stronger one .
a quantum analog of the binary symmetric channel is the _ depolarizing channel _
, where each bit of each codeword that passes through it is independently subjected to : @xmath16 with probability @xmath17 , and @xmath13 , @xmath14 , @xmath15 each with probability @xmath18 .
call a code that achieves a fidelity of at least @xmath10 on such a channel an _ @xmath11-error - correcting code_. it has previously been shown @xcite how to take some special classical linear codes ( binary and over @xmath19 ) with certain properties and transform them into quantum codes with error - correcting capabilities . in the present paper , we show how to transform quantum codes with certain properties into classical codes with error - correcting capabilities .
we do not propose this as a means for constructing new classical codes ; rather , this is a means for translating existing proofs of the _ non_existence of certain classical codes into new proofs of the nonexistence of certain quantum codes .
our specific results , which apply to the class of _ stabilizer _ quantum codes ( defined in the next section ) , are : by these results , we can immediately assert that , when we restrict our attention to stabilizer codes , the classical upper bounds in @xcite apply .
in particular , when a quantum channel is subject to @xmath8 errors , the asymptotic capacity is bounded above by @xmath20 , where @xmath21 is the binary entropy function defined as @xmath22 . in fact , a slightly stronger but more complicated upper bound is proven in @xcite .
this stronger bound is plotted in fig . 1 .
this upper bound is stronger than the previously established @xmath23 bound in @xcite , though the latter bound has the advantage that it applies to nonstabilizer codes as well .
it is noteworthy that all of the quantum codes proposed to date for the channels described above are stabilizer codes .
other upper bounds exist for _ nondegenerate _ quantum codes ( see @xcite for a definition of nondegenerate ) .
one is @xmath24 , and is based on an analogue of the classical `` sphere packing bound '' @xcite , and another asserts that the asymptotic capacity is zero if @xmath25 @xcite .
it remains an open question whether the @xmath26 bound also applies to degenerate codes .
very recently , it has been announced that the @xmath25 threshold bound _ does _ extend to nondegenerate codes , and this will be appear in a forthcoming paper @xcite .
it is interesting to note that there exist some degenerate stabilizer codes that outperform all known nondegenerate codes on the depolarizing channel , for some values of @xmath0 @xcite .
the best lower bound for this channel that we are aware of is @xmath27 @xcite . for the depolarizing channel with error probability @xmath0 ,
our results imply that , for stabilizer codes , the capacity is upper bounded by @xmath2 , the bound for the classical binary symmetric channel @xcite . for some values of @xmath0 , this is stronger than the previously established upper bound of @xmath23 @xcite , though the latter bound applies to nonstabilizer codes as well .
the best lower bound that we are aware of for this channel is @xmath26 @xcite , and a slightly larger value for some values of @xmath0 @xcite .
should any improvements to the upper bounds in @xcite for classical coding occur , they will automatically apply to quantum stabilizer codes .
our results demonstrate interesting connections between quantum stabilizer codes and classical linear codes , and , for some instances of channels , yield stronger upper bounds than those that have appeared to date . in sections
ii and iii , we provide a brief overview of quantum stabilizer codes and classical linear codes . in section iv
, we describe how to construct a binary linear code from a quantum stabilizer code , and , in section v , we show that this construction yields the error - correcting properties required for theorems 1 and 2 .
in @xcite it has been shown that many quantum codes can be described in terms of _
stabilizers_. define a stabilizer as a set of @xmath4-qubit unitary operators such that : each operator is a tensor product of @xmath4 matrices of the form @xmath13 , @xmath14 , @xmath15 , and @xmath16 , with a global phase factor of @xmath28 ; and , the set of operators is an abelian group .
the code that is defined by a stabilizer is the set of all @xmath4-qubit quantum states that are fixed points of each element of the stabilizer .
a stabilizer can be most easily described by a set of operators that generate it .
if one negates the phase factor of some of the generators , the resulting code will change , but will have identical characteristics to the original code .
thus , one can always take the phase of each generator to be @xmath29 without any loss of generality .
it is convenient to denote the generators of a stabilizer in the language of binary vector spaces , as in @xcite .
denote the generator @xmath30 as the @xmath31 bit vector @xmath32 , where , for @xmath33 , @xmath34 and @xmath35 for example , @xmath36 is denoted as latexmath:[$(\ , 1 \ 0 \ 1 \ 0 \ 0 \ 1 \ 0 \ 1 \ , of any two generators @xmath32 and @xmath38 is equivalent ( modulo a phase factor of @xmath39 ) to @xmath40 , where @xmath41 denotes the bit - wise sum in modulo two arithmetic . also , @xmath32 and @xmath38 commute if and only if @xmath42 where @xmath43 denotes the inner product in modulo two arithmetic . a stabilizer
can then be written as an @xmath44 matrix whose rows represent the generators .
for example , @xmath45 represents the eight generators of the stabilizer of a specific @xmath46 code that is 1-error - correcting ( see @xcite for a detailed analysis of this code ) .
in general , if there are @xmath4 qubits and @xmath47 generators , we can encode @xmath48 data qubits ( in the above example , @xmath49 ) .
the error correcting capabilities of the code are related to commutativity relationships between the error operators and the generators @xcite .
an @xmath5 binary linear code is a @xmath3 dimensional subspace of @xmath50 over modulo two arithmetic .
it is sufficient to specify a basis @xmath51 for such a code .
then the codeword for the @xmath3-bit string @xmath52 can be taken as the linear combination @xmath53 ( this mapping is a bijection between @xmath54 and the code ) . a natural way of specifying such a code is by an @xmath55 _ generator matrix _ , whose rows are @xmath51 .
an example of such a code is @xmath56 which is a @xmath57 code that is 1-error - correcting .
consider a quantum stabilizer code specified by a @xmath58 matrix @xmath59 .
we shall show how to construct the generator matrix of a classical binary linear code with similar error - correcting capabilities .
our construction involves a transformation of the generator matrix into a useful standard form along the lines of that in @xcite .
this conversion is accomplished by applying a series of basic transformations of the following two types , each of which leaves the error - correcting characteristics of the code unchanged .
the first is a _ row addition _
, where the @xmath60row is added to the @xmath61row , where @xmath62 .
this corresponds to replacing the @xmath61generator with the product of the @xmath61and @xmath60generator , and setting its phase to @xmath29 .
the second is a _ column transposition _
, where @xmath61column is transposed with the @xmath60column in the submatrices @xmath63 and @xmath64 simultaneously .
this corresponds to transposing the @xmath61qubit position with the @xmath60qubit position .
we begin by applying transformations of the above types to the matrix @xmath59 in order to obtain @xmath65 where @xmath66 is the rank of the submatrix @xmath63 , and @xmath67 .
this is like performing gaussian elimination on the @xmath63 submatrix .
next , by performing row additions among the last @xmath68 generators and column transpositions among the last @xmath69 qubit positions , the matrix can be further converted to the form @xmath70 where @xmath71 is the rank of @xmath72 , @xmath73 , and @xmath74 .
this is like performing gaussian elimination on the submatrix @xmath72 of ( [ stand1 ] ) .
note that if @xmath75 and @xmath76 and then one of the last @xmath77 generators would not commute with one of the first @xmath66 generators .
therefore , we can set @xmath78 , @xmath79 , and @xmath80 .
thus , the form ( [ stand2 ] ) becomes @xmath81 where @xmath82 .
call any set of generators in this form ( [ standard ] ) _ in standard form_. for a generator matrix in standard form , consider the classical binary linear code generated by the @xmath83 matrix @xmath84 we claim that this classical code has similar characteristics to the original quantum code . before stating this precisely ,
consider as an example the aforementioned @xmath46 code , that corrects one error , whose stabilizer was given by ( [ eight ] ) .
converting it to standard form yields @xmath85 which is an equivalent @xmath46 code .
the resulting binary linear code is @xmath86 which is a well - known @xmath87 code that corrects one error .
thus , in this example , we obtain a classical code with slightly better capacity .
the reader may recall that previous constructions of @xmath88 quantum codes have been based on the same classical code ( [ seven ] ) @xcite .
it should be noted that the present connection is quite different , as it involves an @xmath46 quantum code . in the next section
we shall show that , for any @xmath7-error - correcting @xmath12 quantum stabilizer code that is in standard form ( [ standard ] ) , the matrix ( [ classical ] ) generates a @xmath7-error - correcting @xmath89 binary linear code .
also , for the case of the depolarizing channel , we shall show that , for any @xmath11-error - correcting @xmath12 quantum stabilizer code in standard form ( [ standard ] ) , the matrix ( [ classical ] ) generates an @xmath11-error - correcting @xmath89 binary linear code .
furthermore , by applying operations of the types below , which also do not affect the error - correcting capabilities of the code @xcite , we can guarantee the additional property that @xmath90 , which slightly sharpens the result .
the first operation is the _ column switch _ , in which the @xmath61column of submatrix @xmath63 is transposed with the @xmath61column of submatrix @xmath64 .
this corresponds to : in the @xmath61qubit position , changing each instance of a @xmath13 in each generator to a @xmath15 , and each instance of a @xmath15 to a @xmath13 ( while leaving each @xmath16 and @xmath14 intact ) .
the second operation is the _ column addition _ , in which the @xmath61 column of submatrix @xmath64 is added to the @xmath61column of submatrix @xmath63 .
this corresponds to : in the @xmath61qubit position , changing each instance of a @xmath14 in each generator to a @xmath15 , and each instance of a @xmath15 to a @xmath14 ( while leaving each @xmath16 and @xmath13 intact ) .
see @xcite for an explanation of why these two operations do not affect the characteristics of the code .
in this section , we show that the constructions of the previous section satisfy the claimed error - correcting properties .
the classical code whose generator matrix is given by ( [ classical ] ) consists of @xmath91 codewords in the space @xmath92 .
we shall construct an _ isomorphism _ between this code and a restricted version of the quantum code specified by ( [ standard ] ) .
the restricted version of the quantum code consists of @xmath91 codewords that are contained in a special set @xmath93 of @xmath94 distinct @xmath4-qubit states .
this set @xmath93 has the property that it is closed with respect to @xmath15 errors among the first @xmath95 qubit positions .
intuitively , the effect of bit errors on classical codewords within the space @xmath92 is equivalent to the effect of @xmath15 errors in the first @xmath95 qubit positions on quantum codewords within the space @xmath93 .
formally , the isomorphism that we shall construct is a mapping @xmath96 , such that : 1 .
@xmath97 is bijective .
2 . for each @xmath98 that is a codeword of the classical code , @xmath99 is a codeword of the quantum code .
3 . for each codeword @xmath98 of the classical code , and each error vector @xmath100 , @xmath101 the existence of such an isomorphism means that an error in the @xmath61bit of the classical code ( for any @xmath102 ) corresponds to a @xmath15 error in the @xmath61qubit of the restricted version of the quantum code . more precisely ,
if the quantum code can correct any @xmath7 errors then it can correct any @xmath7 @xmath15 errors among the first @xmath95 qubit positions , and then the following procedure for correcting any @xmath7 errors in the classical code exists . given a codeword @xmath103 subjected to an error vector @xmath104 of weight bounded by @xmath7 ,
first apply the mapping @xmath97 to it . by the second and third properties of @xmath97 , the result is @xmath99 subjected to at most @xmath7 @xmath15 errors among the first @xmath95 qubit positions , which can therefore be corrected . by the first property
, @xmath105 can be applied to this corrected quantum codeword , yielding the correction of the original codeword . therefore , if we establish that there exists a @xmath97 that satisfies the above three properties then the classical code specified by ( [ classical ] ) must correct at least as many errors as the quantum code specified by ( [ standard ] ) . for the case of the depolarizing channel with parameter @xmath0 ,
if the quantum code attains fidelity @xmath10 then it attains fidelity @xmath10 for a channel that applies @xmath15 in each of the first @xmath95 qubit positions independently with probability @xmath0 ( in fact we may need to slightly modify the code by applying some column switch and column addition operations defined in section iv along the lines of the `` twirling '' techniques explained in @xcite ) .
therefore , the corresponding classical code is correcting with probability at least @xmath10 on a binary symmetric channel with parameter @xmath0 .
thus , the existence of the above @xmath97 also suffices for this noisy channel model . in order to construct a bijection @xmath97 with the above properties
, we shall construct a useful basis for the quantum code .
we begin with the stabilizer specified by the matrix in standard form ( [ standard ] ) .
call the operators corresponding to the respective rows of this matrix @xmath106 . define the additional operators @xmath107 in terms of the matrix @xmath108 where @xmath109 , and @xmath110 in terms of the matrix @xmath111 by considering ( [ standard ] ) , ( [ phase ] ) , ( [ bit ] ) , and recalling the criterion for commutativity ( [ commute ] ) ,
it is straightforward to verify that : * @xmath112 is a set of @xmath4 independent commuting operators .
* @xmath113 is a set of @xmath4 independent commuting operators . *
each @xmath114 and @xmath115 commute if @xmath116 and anticommute if @xmath117 .
using these properties , we can construct a basis @xmath118 for the code with some useful structural features .
first , set @xmath119 to be the quantum state stabilized by @xmath112 ( this state is unique up to a global phase factor ) .
in fact , @xmath120 is a quantum state with this property .
next , for each @xmath121 , set @xmath122 . since @xmath114 commutes with @xmath115 if and only if @xmath117 , @xmath123 is in the + 1-eigenspace of each @xmath106 and the @xmath124 eigenspaces of @xmath107 , respectively .
therefore , these states are an orthogonal basis for the quantum code .
now , define the function @xmath96 as @xmath125 for each @xmath98 .
we shall show that @xmath97 satisfies the three required properties . by considering ( [ standard ] ) and
( [ phase ] ) , the operator that applies @xmath15 to the @xmath61qubit ( and @xmath16 to all other qubits ) anticommutes with the @xmath61generator in the sequence @xmath126 and commutes with all others .
therefore , @xmath99 is in the @xmath127 eigenspaces of @xmath126 , respectively ( recall that @xmath128 ) .
thus , @xmath99 is orthogonal for each distinct @xmath103 .
this proves the first property , that @xmath97 is a bijection .
also , due to the close similarity between ( [ classical ] ) and ( [ bit ] ) , @xmath129 so the second property for @xmath93 holds .
finally , the third property for @xmath93 holds because , using ( [ phi ] ) and ( [ similar ] ) , @xmath130 thus , @xmath97 satisfies the three required properties .
i am very grateful to hans - benjamin braun for help in analyzing and plotting the functions in @xcite , david divincenzo for several interesting discussions about quantum coding theory and comments about an earlier draft of this paper , emanuel knill for providing references to existing bounds for classical codes , and juan paz for interesting discussions about stabilizer representations of codes .
i am also grateful for the hospitality of the program on quantum computers and quantum coherence at the institute for theoretical physics , university of california at santa barbara , where this work was completed .
this research was supported in part by nserc of canada and the u.s .
national science foundation under grant no .
phy94 - 07194 . | we show that within any quantum stabilizer code there lurks a classical binary linear code with similar error - correcting capabilities , thereby demonstrating new connections between quantum codes and classical codes . using this result which applies to degenerate as well as nondegenerate codes
previously established necessary conditions for classical linear codes can be easily translated into necessary conditions for quantum stabilizer codes .
examples of specific consequences are : for a quantum channel subject to a @xmath0-fraction of errors , the best asymptotic capacity attainable by any stabilizer code can not exceed @xmath1 ; and , for the depolarizing channel with fidelity parameter @xmath0 , the best asymptotic capacity attainable by any stabilizer code can not exceed @xmath2 .
# 1 # 1*#1 * 1\{0,1 } |
due to the lack of analytical solutions , numerical integration is required to solve the equations of motion of the gravitational @xmath0-body problem for almost any initial conditions for @xmath1 .
there are many textbooks with algorithms related to general purpose numerical integration of ordinary differential equations ( odes , see e.g. * ? ? ?
* for an introduction ) .
the method has many advantages : it is one of the most efficient methods if we consider long - term and high precision computations , adaptive forms can be implemented without losing computation time , roundoff errors are smaller than other algorithms , etc .
( see e.g. * ? ? ?
* ; * ? ? ?
* ) . however , the need of derivations of the respective recurrence series for any new problem is a major drawback .
first , @xcite have obtained the recurrence relations for the @xmath0-body problem , taking into account mutual and purely newtonian gravitational forces . soon after , the relations have been derived for the restricted three - body problem @xcite .
many methods for stability analysis require the computation of linearized equations . the relations for the linearized @xmath0-body problem including the equations where one of the bodies is fixed have been presented by @xcite .
the algorithm of lie - integration has widely been applied for stability studies related to known planetary systems ( see e.g. * ? ? ? * ) or special resonant systems ( see e.g. * ? ? ?
in addition , more sophisticated semi - numerical methods can be based on the lie - series ( see e.g. * ? ? ? * about the numerical computation of partial derivatives of coordinates and velocities with respect to the initial conditions and the direct applications for exoplanetary analysis ) .
the aim of this paper is to present the recurrence relations for the osculating orbital elements and the mean longitude in the case of the planar @xmath0-body problem .
here we employ a reference frame where one of the bodies ( i.e. , the central body ) has been fixed . choosing this reference frame
has the advantage that all of the bodies orbiting the center have constant osculating orbital elements _ if _ we neglect mutual interactions .
as we show later on , all of the non - trivial terms depend purely on the mutual terms between the orbiting bodies .
the recurrence equations for constants of motion are derived in sec .
[ sec : lieorbitals ] while the relations for the mean longitude are obtained in sec .
[ sec : liemeanlongitude ] .
our results and conclusions are summarized in sec .
[ sec : summary ] .
throughout this paper we follow the conventions used in @xcite or @xcite .
the newtonian gravitational constant is denoted by @xmath2 , the mass of the central body is @xmath3 while the orbiting ones have a mass of @xmath4 ( @xmath5 , hence we deal with @xmath6 bodies ) .
coordinates and velocities ( with respect to the central body ) are denoted by @xmath7 and @xmath8 ( where @xmath9 or @xmath10 ) if we consider vector notations .
the components of these vectors are denoted by @xmath11 and @xmath12 . for simplicity , specific mass
is denoted by @xmath13 . based on @xcite ,
the relations for the fixed - center problem are the following series of equations .
these are @xmath14 for the coordinates , @xmath15 , \nonumber\end{aligned}\ ] ] for the velocities , @xmath16 for the auxiliary quantities @xmath17 and @xmath18 , and @xmath19 for the distances @xmath20 , the mutual distances @xmath21 and the reciprocal cubic distances @xmath22 , @xmath23 . here
if we evaluate the above relations in the order of equations ( [ eq : lbcoord ] ) ( [ eq : lbphiij ] ) , for all values of @xmath25 and then increase @xmath26 by one in each step ( thus starting over with @xmath27 , etc . ) , we obtain the lie - terms for the coordinates and the velocities .
the solution of the original ode after @xmath28 time can be approximated as @xmath29 note that for the last value of @xmath30 , we need only to evaluate equations ( [ eq : lbcoord ] ) and ( [ eq : lbveloc ] ) . in order to bootstrap these relations
, one could consider the fact that for any quantity @xmath31 , @xmath32 .
hence , the above definitions and relations for @xmath33 and @xmath34 are self - explanatory .
in order to introduce the features of the lie - series for the classical keplerian orbital elements , first , we compute the relations for the specific angular momentum , @xmath35 since the definition of @xmath36 is similar to the relations for @xmath33 ( both are second - order _ and _ bilinear functions of the coordinates and velocities ) , one could expect a similar type of relations like equation ( [ eq : lblambdai ] ) . indeed , the relations for the @xmath37 terms can be written as @xmath38 .
\label{eq : lbci}\ ] ] here , equations for the coordinates and velocities should be computed using equations ( [ eq : lbcoord ] ) ( [ eq : lbphiij ] ) up to some order of @xmath39 . in the case of @xmath40
, @xmath37 must be equal to @xmath41 for any @xmath42 since @xmath43 is an integral of motion .
however , equation ( [ eq : lbci ] ) does not imply this property . in order to obtain the values for @xmath37 ,
first we compute @xmath44 : @xmath45 since @xmath46 and @xmath47 , we get @xmath48 now , equation ( [ eq : lbveloc ] ) is substituted for @xmath49 : @xmath50\right]- \nonumber \\ & & -y_i\left[-\mu_i\phi_ix_i - g\sum_{i\ne j}m_j[\phi_{ij}(x_i - x_j)+\phi_jx_j]\right ] .
\end{aligned}\ ] ] by expanding the above summations and multiplications , the following can easily be seen .
in addition to the keplerian terms ( the first ones , proportional to @xmath51 ) , one part of the terms corresponding to the direct perturbations also cancels . therefore , @xmath52 for higher orders , the set of relations can be written as @xmath53 where we introduce @xmath54 and @xmath55 for simplicity .
@xmath56 due to the properties of the lie - operator ( linearity and leibniz product rule ) , the components of the above equation can easily be expanded once @xmath57 is known .
indeed , similarly to @xmath58 , it can be shown that @xmath59= ( -1/2)(\rho_i^2)^{-3/2}l(\rho_i^2)=-1/2\phi_i2\lambda_i=-\phi_i\lambda_i,\label{eq : lierhoinverse}\ ] ] see also @xcite or @xcite .
@xmath60 then , we have to substitute equations ( [ eq : liecperturb ] ) , ( [ eq : lbveloc ] ) , ( [ eq : lierhoinverse ] ) , @xmath61 and @xmath62 for @xmath63 , @xmath64 , @xmath65 , @xmath66 and @xmath67 , respectively , and then perform a full expansion on equation ( [ eq : mulki ] ) .
the keplerian terms indeed cancel and the remaining parts can be written as @xmath68\ ] ] @xmath69 can be computed in a similar manner , thus the relations for @xmath70 are @xmath71.\ ] ] in order to obtain higher order lie - derivatives , @xmath72 , we should use leibniz product rule for the multilinear expressions appearing in the above relation .
this can either be done directly using the multilinear form @xmath73 or by introducing auxiliary quantities ( e.g. @xmath74 , @xmath75 ) and subsequently apply the bilinear leibniz product rule for these ones .
the specific energy is defined as @xmath76 where @xmath77 .
the semimajor axis can then be computed as @xmath78 . for simplicity ,
in the following we compute relations for the quantity @xmath79 . using the relations for @xmath67 and the velocities ( see equation [ eq : lbveloc ] ) , derivation schemes presented above yields @xmath80,\label{eq : lieenergy}\ ] ] where we introduce @xmath81 . the higher order lie - derivatives
are then obtained as it is described at the end of the previous section .
the previously obtained relations for the orbital elements can applied not only for closed ( circular or elliptic ) orbits but for parabolic and hyperbolic orbits , as well . in the following ,
due to its relevance , we handle only closed orbits .
hence , eccentricity @xmath82 is expected to be smaller than unity for all orbits and the reciprocal semimajor axis @xmath83 is also positive .
the mean longitude @xmath84 can be computed using the analytic equation @xmath85-\frac{\lambda_i}{c_i}j_i.\ ] ] here we introduced @xmath86 , the oblateness of the orbit and @xmath87 .
regarding to the differentiation , the @xmath88 function behaves like the arc tangent function , @xmath89 : @xmath90=\mathrm{d}\left[\mathrm{arc\,tg}\left(\frac{y}{x}\right)\right]= \frac{x\mathrm{d}y - y\mathrm{d}x}{x^2+y^2}.\ ] ] the first - order lie - derivative of @xmath84 is then @xmath91 the denominator of the first ( apparently large ) fraction can significantly be simplified to the form @xmath92 .
now one has to simplify the above equation in order to depend mostly on the mutual interactions .
since @xmath93 even for non - perturbed orbits , this simplification can not @xmath94 the terms appearing above can be expanded as : @xmath95,\\ v_ih_i - w_ik_i & = & c_i\left(\frac{1}{\rho_i}-\frac{u_i^2}{\mu_i}\right ) , \\ v_ilh_i - w_ilk_i+k_ilw_i - h_ilv_i & = & -c_i\phi_i\lambda_i- \sum\limits_{i\ne j}gm_j\hat\phi_{ij}s_{ij}\left(\frac{1}{\rho_i}+\frac{u_i^2}{\mu_i}\right ) , \\ k_ilh_i - h_ilk_i & = & \sum\limits_{i\ne j}\frac{gm_j}{\mu_i}\left[-\frac{c_i^2}{\mu_i}\hat\phi_{ij}\hat c_{ji}+\frac{c_i^3}{\mu_i}\phi_{ij}+\right .
\nonumber \\ & & + \left.\frac{\hat\phi_{ij}}{\rho_i}(\lambda_is_{ij}+c_ir_{ij})-c_i\rho_i\phi_{ij}\right ] , \\
\lambda_il\hat\rho_i-\rho_il\lambda_i & = & ( 1+j_i)(\lambda_i^2\rho_i^{-1}-\rho_il\lambda_i)+\lambda_i\rho_ilj_i\end{aligned}\ ] ] and @xmath96.\end{aligned}\ ] ] where @xmath97 and @xmath98 . using the well - known relations from classical celestial mechanics
, it can be shown that the double - negative specific energy , @xmath99 relates to the oblateness @xmath100 and the specific angular momentum @xmath36 as @xmath101 . from this relation , by taking the lie - derivative of both sides , we got @xmath102 therefore , the last term in equation ( [ eq : llambdai ] ) can be written as @xmath103 here the first and third terms cancel each other , thus @xmath104 from the above series of equations we collect those where terms after the summation @xmath105 do not occur .
this part , denoted as @xmath106 is @xmath107c_i\left(\frac{1}{\rho_i}-\frac{u_i^2}{\mu_i}\right)\bigg\}. \label{eq : llambda0}\end{aligned}\ ] ] by substituting the relations @xmath108 , @xmath101 and @xmath109 , equation ( [ eq : llambda0 ] ) @xmath110 let us write the full lie - derivative of @xmath111 in the form @xmath112_{ij}.\ ] ] this is similar to the forms obtained for the angular momentum , specific energy and lagrangian orbital elements , with the exception of the presence of the term related to kepler s third law .
the goal now is to compute the terms @xmath113_{ij}$ ] as simple as possible .
it can be shown that this term is @xmath114_{ij } & = & + \frac{\hat\phi_{ij}}{1+j_i}\left[\left(-\frac{2j_i(1+j_i)}{c_i}+\frac{2c_i}{\mu_i\rho_i}\right)r_{ij}-\left(\frac{\rho_i}{\mu_i}+\frac{c_i^2}{\mu_i^2}\right)\hat c_{ji}\right ] + \nonumber \\ & & + \frac{\phi_{ij}}{1+j_i}\left[\frac{c_i^3}{\mu_i^2}-\frac{c_i}{\mu_i}\rho_i+\frac{2j_i(1+j_i)}{c_i}\rho_i^2\right ] .
\label{eq : llambdaij}\end{aligned}\ ] ] this equation ( [ eq : llambdaij ] ) can be simplified in terms of computation implementation by introducing the dimensionless quantity @xmath115 : @xmath114_{ij } & = & + \frac{\hat\phi_{ij}}{1+j_i}\left[\frac{2}{c_i}\left(g_i^{-1}-j_i(1+j_i)\right)r_{ij}-\frac{\rho_i}{\mu_i}\left(g_i^{-1}+1\right)\hat c_{ji}\right ] + \nonumber \\ & & + \frac{\phi_{ij}}{1+j_i}\frac{\rho_i^2}{c_i}\left[g_i^{-2}-g_i^{-1}+2j_i(1+j_i)\right ] .
\label{eq : llambdaij2}\end{aligned}\ ] ] therefore , the first lie - derivative of @xmath84 can be written as @xmath116\end{aligned}\ ] ] where @xmath117 higher order derivatives can then be computed using the relation @xmath118\end{aligned}\ ] ] let us suppose that the lie - derivatives of the arbitrary quantity @xmath31 are known up to the order of @xmath119 .
it can be shown by mathematical induction that the @xmath120th lie - derivative of @xmath121 can be computed using the relation @xmath122l^{n - k}\left(q^p\right)l^{k+1}q\label{eq : liepower}\ ] ]
in this paper we presented recurrence formulae of the orbital elements related to the planar @xmath0-body problem .
as we showed , the structure of these formulae depends only on the terms related to the mutual interactions .
therefore , the relations for the two - body problem reduces to a constant motion that can be integrated with arbitrary step size .
following studies could investigate the relations for the spatial problem . in some cases
, this extension could be straightforward for scalar quantities like the specific energy .
another interesting point can be the elimination of the need for computing the recurrence formulae for coordinates and velocities and employ directly the orbital elements .
* acknowledgments . *
the author would like to thank the anonymous referees for their valuable comments .
the author also thanks lszl szabados for the careful proofreading .
this work has been supported by the hungarian academy of sciences via the grant lp2012 - 31 .
bancelin , d. , hestroffer , d. & thuillot , w. : numerical integration of dynamical systems with lie series .
relativistic acceleration and non - gravitational forces .
. astron . * 112 * , 221234 ( 2012 ) | lie - integration is one of the most efficient algorithms for numerical integration of ordinary differential equations if .
the method is based on the computation of the taylor - coefficients of the solution as a set of recurrence relations . in this paper
we present these recurrence formulae for orbital elements and other integrals of motion for the planar @xmath0-body problem .
we show that if the reference frame is fixed to one of the bodies for instance to the sun in the case of the solar system , the higher order coefficients for all orbital elements and integrals of motion depend only on the mutual terms corresponding to the orbiting bodies . |
the wave of colonization by migrating farmers and establishment of farming communities in europe between @xmath0 and @xmath1 bc is currently a topic of great interest in prehistoric archaeology , linguistics and anthropology ch , pr .
ammerman and cavalli - sforza developed a model for the expansion of farming as a demic diffusion which spread into europe in the form of wave of advance @xcite . using the radiocarbon dates
, they found that farmers spread at an average rate of about one kilometer a year .
interest in simulation and spatial modelling of spread of agriculture has been growing rapidly in the last decade , especially in the physics community @xcite .
one of the main reasons for this is that the geographical spread of population can be effectively described by the classical fisher - kpp equation and its various generalizations @xcite .
these models have attracted considerable interest in physics and biology , because of the huge number of potential applications .
fort and mendez have applied a time - delayed theory for the neolithic transition @xcite which involves a hyperbolic correction to the fisher - kpp equation . the transition from hunter - gathering
to farming did not happen in a uniform way , and that is why davison et al .
have taken into account both advection and spatial variation in diffusivity and carrying capacity , in the framework of the fisher - kpp equation @xcite .
aoki and shida @xcite have studied the spread of farmers into an area occupied by hunter - gatherers in terms of a system of reaction - diffusion equations . the main objective of those works was to reproduce the observed rate at which agricultural expansion took place in europe . despite the interest in establishment of farming communities in europe
, there remains no published material on the spatial structure formed behind the wave of advance .
the main challenge of our work presented below is to set up a model which provides an explanation to not only the propagation of farming but also the formation of human settlements .
we develop a model that demonstrates the tendency of the distribution of population to form clusters , and furthermore , that this large - scale pattern in population evolution is a transient phenomenon , which disappears due to degradation of land in the form of an extinction wave .
our specific motivation is the successive migration of settlements along parallel river valleys in the tripolye - cucuteni system , which has been thoroughly investigated and documented , e.g. @xcite .
this system can be considered as being one - dimensional , and so is particularly amenable to numerical investigation .
in our model the population consists of semi - sedentary foragers and sedentary farmers who share the same territories .
the semi - sedentary foragers are the population of individuals randomly moving from place to place along a river valley and searching for food and other resources .
an implicit consequence of this behaviour is the foundation of new settlements ( large localized values of population density ) , and an interchange between farming and foraging populations . on the contrary the sedentary farmers are individuals who do not migrate .
they live in small villages scattered near cultivated land in the major river valleys .
their main activities are the cultivation of soil and crop production . in this paper
we are interested in the total population density @xmath2 at location @xmath3 along the river at time @xmath4 we define this density as @xmath5 , where @xmath6 is the density of semi - sedentary foragers and @xmath7 is the density of sedentary farmers .
we assume that there is not a strict distinction between foraging and sedentary lifestyles .
there are always random transitions from a sedentary to a foraging way of life , and _
vice versa _ ; these transitions depend strongly on the local food supply .
this is one of the main features of our random walk model . regarding the movement of semi - sedentary foragers ,
they do not jump from place to place completely randomly . unlike brownian particles in physics , the migration of people can not be explained by a standard diffusion law , in which the flux is proportional to the gradient of number density of individuals . to describe the random migration of semi - sedentary foragers and random transitions from one lifestyle to another
we adopt a biased random walk whose statistical characteristics depend on the local food supply . in our model ,
the probability of a random migration event making a jump @xmath8 in the time interval @xmath9 to @xmath10 is @xmath11 the probability of transition from foraging lifestyle to the farming is @xmath12 the probability of the conversion of farmers to semi - sedentary foragers is @xmath13 thus we introduce a new variable , the local crop production per individual per year @xmath14 so that the frequency of jumps @xmath15 and transition rates @xmath16 and @xmath17 depend on the crop production : @xmath18it is natural to assume that the frequency @xmath19 is a decreasing function of @xmath20 that is , when sedentary farmers are not able to produce enough food to sustain their population , some of them start to migrate from their neighborhood at rate @xmath21 .
we now set up the balance equation for the density of semi - sedentary foragers , @xmath22 according to our random walk model , the density of foragers @xmath6 at location @xmath3 at time @xmath23 can be written as follows @xmath24this equation is a conservation law for foragers .
the first term on the right hand side represents those foragers who stay at location @xmath3 and do not move during time @xmath25 and do not become sedentary farmers .
the second term gives the number of foragers who arrive at @xmath3 during time @xmath25 from different places @xmath26 , where the jump distance @xmath8 is distributed by dispersal kernel @xmath27 the last term is the number of sedentary farmers who become the semi - sedentary foragers during the time @xmath28 the sedentary farmers do not migrate , and their density @xmath29 obeys the balance equation involving logistic growth and lifestyle transitions : @xmath30where @xmath31 is the growth rate of the sedentary population . the carrying capacity @xmath32 is in general , an increasing function of the local crop production @xmath33 .
the last term gives the number of foragers who become the farmers during the time @xmath28 in the limit @xmath34 , from ( [ n1 ] ) and ( [ n2 ] ) we obtain two differential equations for @xmath35 and @xmath36 @xmath37@xmath38 in this paper we are interested in the evolution of population density @xmath2 on characteristic time scales around @xmath39 years .
therefore we can adopt a local equilibrium for the two populations , describing the balance between the nomadic and sedentary ways of life in proportions @xmath40 and @xmath41 we write @xmath42where the proportion of foragers @xmath40 is the function of the crop production @xmath43 @xmath44 in the asymptotic regime when @xmath45 one can write the dependence of @xmath40 on the transition rates @xmath46 and @xmath47 as @xmath48the evolution equation for biased migration ( [ nnn1 ] ) and population growth ( [ nnn2 ] ) can be rewritten as one balance equation for the total population density@xmath49 by adding equations ( [ nnn1 ] ) and ( [ nnn2 ] ) and using ( [ main ] ) we obtain a single equation for @xmath2 @xmath50now we need to find an approximate form for the function @xmath51 it is natural to assume that for a large crop production @xmath52 the population mostly consists of sedentary farmers , that is , @xmath40 close to zero , while for a low crop production the population mostly is semi - sedentary foragers ( @xmath40 is close to one ) . a simple approximation for @xmath40 might be @xmath53where @xmath54 is the heaviside step function : @xmath55 if @xmath56 @xmath57 if @xmath58 thus farmers do not migrate if the crop supply is greater than a critical value @xmath59 . in our numerical simulations we use a piece - wise approximation : a linearly decreasing function in the interval between the minimum value @xmath59 and the maximum value @xmath60 @xmath61based on @xcite , we take @xmath62 kg per person per year respectively . for the results discussed below we set @xmath63 , @xmath64 ( see fig .
1 ) but the exact values are not crucial to our results . the dispersal kernel @xmath65 is assumed to have two cutoffs : a minimum cutoff ( foragers do not travel very small distances ) and a maximum cutoff ( foragers do not migrate over very large distances in a single jump ) . in numerical simulations we use minimum cutoff @xmath66 of between @xmath67 and @xmath68 km . the maximum cutoff is more difficult to estimate .
however , since the typical length of the river is around @xmath69 km , it seems reasonable to assume that the maximum cutoff is around @xmath70 km ( see fig .
we assume that the human population derives most of their food from the cultivation of land .
thus we suggest the following formula for the local food production @xmath71 @xmath72where @xmath73 is the density of soil nutrients , @xmath74 is the production rate coefficient , and @xmath75 is the parameter that determines how the yield depends on the nutrients .
this equation describes how the rate of food supply @xmath33 increases due to the increase in the population density @xmath76 , and how the degradation of land ( the decrease of soil nutrients @xmath77 leads to a decrease of food production through the factor @xmath78 note that the factor @xmath79 describes a tendency toward group solidarity that increases the efficiency of food production .
we know from existing data that the yield , i.e. the production of food per unit area , can be up to @xmath80 kg / m@xmath81year @xcite.we estimate that the area cultivated by people is approximately @xmath82 m@xmath81 per person .
thus , we have an estimate for @xmath83 of @xmath84 kg per person per year . in general , it is difficult to model how the production of food depends on the nutrients , since this relation is strongly dependent on the environmental conditions .
one of the existing models is given by the mitscherlich - baule yield response @xcite which takes into account productivity losses due to soil erosion .
following this model , in @xcite the factor @xmath85 has been proposed , and this is what we have introduced into our model .
the value @xmath86 m@xmath81kg@xmath87 is found for the corn cultivation .
however , other studies @xcite seem to suggest lower values ( @xmath88 m@xmath89kg@xmath87 ) . [ cols= " < , < " , ] we also explored the dependence of the speed of population advance , and the separation of the clusters , as a function of the parameters .
the speed of advance is rather ill - defined , but for simplicity we monitored the movement of the main peak of the distribution .
( arguably we could have , e.g. , studied the position of the cluster furthest from the origin . )
we found that the speed of advance depends approximately linearly on both @xmath90 and @xmath91 , and to be insensitive to @xmath92 .
the separation of clusters is independent of both @xmath90 and @xmath91 and depends approximately linearly on @xmath92 ( @xmath93 ) . if @xmath94 or @xmath95 the clustering phenomenon does not occur .
while we have concentrated our efforts on studying variation among a manageable subset of parameters while retaining fixed plausible values for the others , we did also make a small study of the effects of varying @xmath96 and @xmath97 .
increasing the value of @xmath96 , up to @xmath98 from our canonical value of @xmath99 , prolongs the timescale of the population evolution , by a factor of up to about @xmath100 .
values of @xmath96 significantly smaller than @xmath99 do not give clustering for typical choices of the other parameters .
it means that for non - fertile soils transition to farming and the emergence of clusters are unlikely to occur . for larger values of @xmath101 ,
the regeneration is so strong that , although clustering occurs initially much as for @xmath102 , the final state is a spatially uniform population . for smaller values of @xmath97 , e.g. @xmath103
, we still see strong clustering , but the overall phenomenon persists for a significantly shorter time , as the episodes of cluster regeneration seen when @xmath102 do not now occur .
we have explored a continuous model of linear ( one - dimensional ) population migration and clustering .
we took our inspiration and guidance from the successive south to north migration of the tripolye - cucuteni cultures along parallel river valleys @xcite .
we attempted only to illustrate the migration along a single valley .
the main result is that our model gives an explanation for the formation of settlements as a dynamical phenomenon .
the individuals have a tendency for aggregation and clustering as a result of non - linear random migration .
moreover our model describes subsequent decay of these clusters ( settlements ) due to land degradation in the form of an extinction wave . clearly our model , being mathematically continuous ,
can not be expected to reproduce in detail the essentially discrete phenomenon of the establishment , growth and decay of major settlements and their satellites , as revealed by the archaeological record our goal is far more modest . for reasonable choices of parameters
we obtain migration over distances of order @xmath69 km in times of @xmath104 yr . without fine tuning ,
our solutions exhibit distinct clustering of population , at intervals of @xmath105 km .
plausibly these clusters are estimated to have @xmath106 individuals .
these estimates are all consistent with what is known about the tripolye - cucuteni cultures @xcite .
overall , we feel that our quite naive modelling captures some of the essence of the clustering seen in population migration , and points the way for more sophisticated modelling .
we can think of a number of significant developments in the future , including taking @xmath107 and allowing for a latency effect , in which population in a cluster ( ` settlement ' ) has a reduced probability of moving , representing an attachment to the investment of building a house and clearing land .
it is easy to allow for additional resources available from e.g. forest or river , which are maybe harder to exhaust than the fertility of the land and so can act as a reservoir of resource
. it would be interesting to apply our model to understand societal collapses to which environmental problems as habitat destruction , soil degradation and overpopulation contribute @xcite .
the research is supported by the european community s sixth framework programme , grant nest-028192-fepre . | we present a stochastic two - population model that describes the migration and growth of semi - sedentary foragers and sedentary farmers along a river valley during the neolithic transition .
the main idea of this paper is that random migration and transition from sedentary to foraging way of life and backward is strongly coupled with the local crop production and the associated degradation of land .
we derive a non - linear integral equation for the population density coupled with the equations for the density of soil nutrients and crop production .
our model provides an explanation for the formation of human settlements along a river valley .
the numerical results show that the individual farmers have a tendency for aggregation and clustering .
we show that the large - scale pattern is a transient phenomenon which eventually disappears due to land degradation . |
the production of quantum degenerate gases has revolutionized the field of atomic physics .
such gases are now routinely used as a means towards understanding complex many - body quantum phenomena from the realms of condensed matter and nuclear physics @xcite . as atom sources with precisely controlled properties , these gases can also significantly advance applications such as atom interferometry @xcite and quantum information processing @xcite . while the production and measurement methods of quantum gas experiments are well established , the measurement rate remains substantially limited by the lack of a general method for rapid sample production .
cycle times for such experiments are dominated by the production time , typically tens of seconds , while the actual experiment on the prepared sample lasts for about a second before destructive measurement .
this separation of timescales is a severe impediment to the employment of quantum degenerate gases towards precision devices such as atomic clocks , inertial sensors and gravimeters @xcite .
bridging these timescales can significantly contribute to all classes of quantum gas explorations and applications , as most measurements rely on the statistics of results from many experimental iterations .
the root of this timescale problem lies in the speed of collisional evaporative cooling . in a standard
degenerate gas production sequence , the initial step of laser cooling produces temperatures in the few tens of @xmath5k , while light - induced processes keep the density below @xmath6@xmath7 . the resulting phase space density ( psd ) of @xmath8
is increased to quantum degeneracy by subsequent evaporative cooling in either magnetic or optical traps .
typical magnetic traps have large volumes and relatively low initial densities , yielding low collision rates and long evaporative cooling timescales of tens of seconds .
the production of becs in small - volume magnetic chip traps has provided one solution to the timescale problem @xcite .
while this method has recently achieved cycle times of one second with @xmath9rb @xcite , it can only be applied to magnetic atoms .
optical dipole traps ( odts)@xcite provide the flexibility to cool all atoms , enabling applications with non - magnetic atoms and liberating the magnetic degree of freedom for interaction control during cooling @xcite .
standard odts have small volumes and high collision rates , leading to smaller atom numbers but shorter evaporation timescales in the range of a few to @xmath10s .
an overall bec production time as low as @xmath11s has been accomplished in an odt @xcite by combining one broad and one very narrow transition for laser cooling in @xmath12sr , to realize an extremely high initial psd of @xmath13 before evaporative cooling @xcite .
a bec production time of @xmath14s was achieved in @xmath9rb by combining sub - doppler laser cooling to a high initial psd of @xmath15 with fast evaporative cooling in an odt in which the volume was dynamically compressed by a moving lens @xcite .
these methods however are not easily adaptable in a general way to other atomic species and experimental setups .
in this paper we present a general technique for rapid quantum degenerate gas production , where evaporative cooling is optimized by dynamically controlling the odt shape with the time - averaged potential of a single rapidly - moving laser beam .
this method straightforwardly allows for large initial and small final trap volumes , combining the advantages of earlier evaporative cooling strategies .
it is applicable to all atoms and requires no hardware beyond a standard optical trapping setup .
applied to bosonic @xmath0yb with modest laser cooling to psds @xmath16 , our experiment produces becs containing @xmath17 atoms with an overall cycle time of @xmath18s . by suitably altering the time dependence of the trap parameters , we also produce large @xmath0yb becs with @xmath19 atoms .
the observed evaporation dynamics are successfully captured over three orders of magnitude in psd by our theoretical model .
( a ) schematic of the optical trap setup
. gravity points along @xmath20 .
the horizontally oriented aom is driven by an fm waveform ( inset ) resulting in parabolic intensity profile .
arrows on the laser beam indicate direction of cpm .
( b ) evolution of time - averaged trap shape @xmath21 from gaussian to parabolic with increasing cpm amplitude for fixed laser power in units of unmodulated beam waist @xmath22 and trap depth @xmath23 .
( c ) fractional reduction of trap depth and frequency in the painting direction @xmath24 from unmodulated values @xmath23 and @xmath25 for fixed power . ] forced evaporative cooling works by removing the high - energy tail of the maxwell - boltzmann distribution and allowing the remaining atomic sample to re - equilibrate by elastic collisions to a lower temperature .
keeping the removal point fixed at @xmath26 relative to the temperature @xmath27 allows the derivation of scaling laws for the evaporation dynamics @xcite .
the rate of elastic collisions @xmath28 and @xmath29 determine the per - particle evaporative loss rate as @xmath30 for large @xmath29 @xcite , where @xmath31 is the peak particle density , @xmath32 is the scattering cross section , @xmath33 , and @xmath34 is the particle mass .
we define the evaporation efficiency as @xmath35 , where @xmath36 and @xmath37 are the final(initial ) psd and particle number , respectively . in the absence of additional loss processes
, @xmath38 can be made arbitrarily high by using a large @xmath29 and thus a long cooling timescale .
the reality of other loss processes tempers this idea and introduces a new timescale which competes with that of evaporative cooling .
this competition outlines an important experimental challenge and is captured by the ratio of good " to bad " collisions @xmath39 .
the prescription for optimum efficiency however crucially depends on the nature of the dominant loss process . for 1-body loss dominated systems characteristic of standard magnetic traps
, @xmath40 is proportional to the elastic collision rate and scales as @xmath41 , where we have assumed a 3d harmonic trap with ( geometric ) mean trap frequency @xmath42 . comparing against the scaling @xmath43
, we find that maintaining or increasing the collision rate at every step and achieving runaway " evaporation is an excellent prescription for efficient cooling @xcite .
importantly , this prescription simultaneously improves the speed of evaporative cooling and final particle number . in an odt , while 3-body processes can be neglected in certain situations @xcite , it is often the dominant loss mechanism .
then @xmath44 , and @xmath38 and @xmath40 can not be simultaneously optimized .
crucially , unlike in 1-body loss dominated systems , the inverted scaling of @xmath40 with density means maintaining a large @xmath45 leads to lowered @xmath42 , reduced collision rates and longer evaporation timescales .
numerical modeling of evaporative cooling in an odt @xcite can help optimize @xmath38 , but the challenges of large number and speed remain .
our solution to these challenges involves the implementation of the time - averaged optical potential of a laser beam moving rapidly , or painting " , in one dimension . the ability to dynamically control the center position modulation ( cpm ) amplitude of the beam in addition to its total power , results in independent , arbitrary control over both the trap depth ( @xmath46 ) and frequency as a function of time , a key advantage over methods with a fixed power - law relationship between @xmath47 and @xmath48 @xcite .
the rest of the paper is organized as follows .
section [ sec : expset ] consists of the experimental setup and our model for the cpm trap shape . in section [ sec : evapmodel ]
we present our numerical model for forced evaporative cooling in this trap along with a procedure for optimizing evaporation efficiency , and apply our model to an experimentally optimized evaporation trajectory . in section [ sec : rapidbecs ] we describe rapid bec production using our method , while in section [ sec : largebecs ] we focus on the production of large becs .
finally , section [ sec : summary ] provides a summary of this technique and an outlook for applications .
we perform our experiment in the apparatus described in @xcite using @xmath0yb bosons . our laser cooling procedure can produce @xmath49 atoms in @xmath50s at @xmath3k in a compressed magneto - optical trap ( mot ) operating on the @xmath51 transition . in the rest of this section
, we describe the setup of our odt and characterization of the parabolic cpm trap .
the odt ( fig.[fig1 ] ) is generated by sending the output of a fiber laser at 1064 nm ( ipg ylr-100-lp ) through an acousto - optic modulator ( aom , 80 mhz , intraaction atm-804da6b ) and focusing the diffracted beam ( 1@xmath52 order ) to a gaussian waist of 35@xmath53 m at the atoms .
this light is then refocused with orthogonal polarization back onto the atoms with a waist of 30@xmath53 m , at an angle of 65@xmath54 with respect to the first pass . to implement the trap center position modulation
, we modulate the center frequency of the voltage controlled oscillator ( vco , minicircuits zos-100 ) supplying rf to the aom at 10 khz with the waveform shown in fig.[fig1](a ) , which results in a nearly perfect parabolic trap shape ( see appendix ) .
the largest cpm amplitude used is 260 @xmath5 m ( 520 @xmath5 m peak - to - peak ) , and corresponds to shifting the center frequency of the aom by 7 mhz ( 14 mhz peak - to - peak ) on top of 80 mhz .
the orientation of the aom is such that the cpm occurs in the horizontal plane .
we control the overall odt power with the rf drive strength to the aom . using a power of @xmath55w and cpm amplitude of @xmath56 m at the atoms , we capture up to @xmath57 atoms from the compressed mot into the odt .
comparison of cpm trap model ( open triangles ) to measurements ( solid circles ) of the trap frequency .
different cpm amplitudes are indicated with different colors .
boxes indicate trap configurations used for measurements in fig .
the black dashed line is a parametric plot of @xmath58 corresponding to the evaporation trajectory used in fig .
[ fig5 ] . ]
crucial to the implementation of our forced evaporative cooling model ( see section [ sec : evapmodel ] ) is an accurate model for the optical trap shape as a function of laser power @xmath59 and cpm amplitude @xmath60 @xcite . without cpm and neglecting gravity
, one can apply the scalings @xmath61 and @xmath62 in the absence of beam imperfections ( e.g. thermal lensing , astigmatism , etc . ) .
however , as depicted in fig .
[ fig1 ] , cpm allows for a large range of depths and frequencies at each power .
we calculate the trap frequencies and trap depth for a single odt beam with cpm in the @xmath63 direction and traveling in the @xmath64 direction as @xmath65 where @xmath66 is the atomic polarizability at the odt wavelength @xmath67 and @xmath68 is the rayleigh range . @xmath69 and @xmath70 are the fractional reduction factors shown in fig .
[ fig1](c ) , conveniently written as a function of the cpm amplitude in units of beam waist .
to arrive at a reliable model for our specific time - averaged potential , we measure the physical amplitude @xmath60 ( in @xmath5 m ) of the center position modulation at the odt focus for a given applied voltage to the vco .
furthermore , we account for the effects of ellipticity and thermal lensing by allowing power dependent waists @xmath71 and @xmath72 , changing the arguments of the fractional reduction functions to @xmath73 and @xmath74 , and appropriately altering the equations in ( [ eqn : cpmfreqsanddepth ] )
. measurements of yb number and temperature evolution for trap depth model calibration .
the 3 different trap configurations are as follows : ( 1 , solid circles ) @xmath59 = 58 w , @xmath60 = 259 @xmath5 m ; ( 2 , solid squares ) @xmath59 = 58 w , @xmath60 = 130 @xmath5 m ; ( 3 , solid triangles ) @xmath59 = 2.2 w , @xmath60 = 113 @xmath5 m .
the lines are fits using the evaporation model discussed in section [ sec : evapmodel ] .
the extracted trap depth for each curve is indicated on the plot above .
our time - averaged trap model predicts @xmath75 @xmath5k , @xmath76 @xmath5k , and @xmath77 @xmath5k . ] finally , we include the second pass of the crossed dipole trap by using a magnification of 5/6 given by the lenses used to reimage the beam back onto the atoms at 65@xmath54 .
we write @xmath78 , where the subscripts 1 and 2 refer to the first and second passes of the beam , and the coordinates @xmath79 are related to @xmath80 by a 65@xmath54 rotation in the @xmath81 plane .
we then compute the roots of @xmath82 for a densely spaced grid of @xmath59 and @xmath60 values , and use these roots to compute @xmath83 and @xmath84 referenced to the equilibrium position @xmath85 .
the comparison of our trap model to trap frequency measurements made using @xmath0yb is shown in fig .
as discussed above , all trap parameters are independently determined by observations of the beam itself , so there are no fit parameters used here .
we find good agreement between the theoretical and experimental values .
to compare our model for the trap depth with the experiment , we perform measurements of yb number and temperature evolution for three different fixed settings of odt power and cpm amplitude ( indicated in fig .
[ fig2 ] with boxes ) .
we then fit the observed dynamics with our evaporation model ( see section [ sec : evapmodel ] ) for fixed power and cpm amplitude , using the trap depth as a fitting parameter .
the results are shown in fig .
the predicted trap depth values from our model are given in the caption to fig .
[ fig3 ] , and are in good agreement with those extracted from number and temperature dynamics .
we model the number and temperature evolution in the high @xmath29 limit , where the equation of state is well approximated by @xmath86 .
the dynamical equations are then @xcite @xmath87 where @xmath88 is the per - particle loss rate for 3-body inelastic loss , @xmath89 is a temperature - independent 3-body inelastic loss rate coefficient , @xmath90 , and @xmath91 and @xmath92 are the spontaneous scattering rate and recoil energy for @xmath0yb in our 1064 nm odt . for the odt intensities used here ,
heating from spontaneous scattering is small .
the background lifetime @xmath93 is independently measured to be 35@xmath94s . by neglecting the effects of 3-body inelastic loss , background gas collisions , and spontaneous scattering
, one can construct analytical solutions to equations ( [ ndot])-([tdot ] ) based on scaling laws by specifying a trajectory @xmath95 for all times @xmath96 and assuming a fixed relationship @xmath97 @xcite . without these simplifying assumptions , however , it is essential to turn to numerical techniques , especially with an additional dynamically controllable parameter such as cpm amplitude .
therefore , to inform our experimental choice of ramp profiles @xmath98 and @xmath99 , we run an optimization algorithm in mathematica@xmath100 using numerical solutions to equations ( [ ndot])-([tdot ] ) . to speed up the numerical integration of @xmath101 and @xmath102 ,
we compute the trap parameters @xmath83 and @xmath84 for a dense grid of power and cpm amplitude values .
we then convert these tables into interpolating functions , making the determination of @xmath103 and @xmath104 throughout the evaporation ramp extremely fast . for the functional form of the power ramp
@xmath98 we try single exponential profiles of the form @xmath105 and bi - exponential profiles of the form @xmath106 , guided by the fact that evaporation occurs on an exponential timescale .
furthermore , a bi - exponential could potentially handle the presence of two dominant timescales ( e.g. evaporation and 3-body loss ) . for the cpm amplitude ramp @xmath99
, we try out many different functional forms , including offset exponential @xmath107 , linear @xmath108 , and concave functions @xmath109 and @xmath110 . for the above functions ,
the parameters @xmath111 and @xmath112 are fixed as they correspond to the trap loading conditions , while the remaining parameters are varied as part of the optimization algorithm .
our optimization algorithm utilizes the gradient ascent method where @xmath113 is the quantity to be maximized , and @xmath114 satisfies @xmath115 .
for the power ramp profile , we find that the optimization procedure pushes the bi - exponential profile towards a single exponential ( i.e. @xmath116 ) .
cpm ramp optimization suggests that the system is very robust to the form of @xmath99 .
in fact , if we use @xmath117 with @xmath118s , and run the optimization algorithm for each of the proposed functions @xmath99 above , the optimized values @xmath119 are all within 1% of each other . for these reasons
we choose to adopt the single exponential power ramp and linear cpm ramp , as these are the simplest options .
cpm amplitude trajectory optimization for various power ramp timescales @xmath120 . for each value
@xmath120 we run an optimization of the cpm reduction slope @xmath121 ( solid blue triangles ) to maximize the evaporation efficiency @xmath38 ( solid red circles ) at @xmath122 . ]
we find the system to be extremely robust to the timescale of evaporation , as indicated in fig .
[ fig4 ] . for these simulations
, we fix the timescale @xmath120 and optimize the slope of the cpm ramp @xmath121 .
we restrict @xmath123 , as we believe that our model can not accurately capture the evaporation dynamics on timescales faster than this due to complications involving thermal lensing and decoupling of horizontal and vertical temperatures . as seen in fig .
[ fig4 ] , the optimized evaporation efficiency @xmath119 varies little over the range @xmath124 .
in fact , the slight slope of @xmath119 versus @xmath120 is caused by the introduction of a new timescale , the background lifetime @xmath93 .
as described in section [ sec : comparemodandexp ] , we observe the same behavior in the experiment as the maximum number of atoms in the yb bec is quite resilient to the evaporation timescale .
we experimentally investigate the evaporation efficiency by maximizing final bec number . in agreement with our simulations
we find that over timescales where 1-body loss is negligible , the evaporation efficiency is robust and an exponential reduction of power and linear reduction of cpm amplitude yield the largest @xmath38 .
a typical optimized evaporation trajectory is shown in fig.[fig5 ] .
our theoretical model successfully captures the dynamics over 3 orders of magnitude in psd . for these measurements ,
we wait @xmath125ms after odt loading before beginning forced evaporation to allow atoms in the wings of the trap to escape . for the model curves in fig.[fig5 ] ,
the only free parameters are the initial number and temperature , and @xmath89 .
we assume @xmath126-wave scattering only as the @xmath127-wave threshold for @xmath0yb is 75 @xmath5k , and use @xmath128 , where @xmath129 nm @xcite . from a least squares fit to the data , we extract @xmath130 @xmath131s@xmath132 . from the behavior of @xmath133 and @xmath134 in fig.[fig5](d ) , we see that the dynamically shaped odt maintains dominance of evaporative over 3-body loss . furthermore , the elastic collision rate @xmath135 falls less than a factor of 2 from 1.7 khz to 1.1 khz over the course of the evaporation sequence .
the evaporation efficiency @xmath38 for this ramp is 3.8 , close to the highest value found using the optimization algorithm discussed above with our initial trap conditions and ramp profiles .
we note that although runaway evaporation where @xmath136 is easily achieved with the dynamically shaped odt , we do not find this to be the optimal evaporation strategy due to enhanced 3-body loss .
example of optimized evaporation trajectory .
( a , b ) measured number ( blue circles ) , temperature ( red circles ) , and phase space density ( black circles , inset ) evolution during forced evaporative cooling show excellent agreement with our theoretical model ( solid black line , see text ) . in ( b )
we also plot @xmath137 ( dashed line ) . from the fit in ( b ) we find @xmath138 .
( c ) trajectories of trap frequency and depth ( inset ) for measurements in ( a , b ) .
dashed lines correspond to the same trajectories without reduction of cpm amplitude .
( d ) evolution of per - particle loss rates for evaporation ( solid line ) and 3-body inelastic loss ( dashed line ) during forced evaporation . ]
in addition to providing a platform for highly efficient evaporation of large atom number clouds , dynamical trap shaping can be applied to the rapid production of becs . for this purpose
we shorten the mot loading and compression time to a total of @xmath139s , and begin forced evaporative cooling immediately following loading of the odt with an initial psd of @xmath16 .
the experimentally optimized fast bec ramp measurements are shown in fig.[fig6 ] .
for fast bec production , we find the use of three distinct evaporation stages to be optimal . in the first stage , we exponentially reduce the power by a factor of 20 in @xmath140ms while linearly reducing the cpm amplitude to zero in @xmath141ms .
second , we exponentially reduce the laser power by a factor of 6 in @xmath142ms . in a third , relatively slow evaporation stage , we linearly decrease the power another 30% in @xmath143ms , and then hold at constant depth for @xmath141ms . in fig.[fig6](b ) we see that the horizontal and vertical temperatures initially decouple due to the rapid decrease of cpm amplitude in the first @xmath141ms .
this can be understood by considering the adiabatic temperature evolution terms , since @xmath144 during this time . since the assumption of thermal equilibrium
is violated at these short timescales , we can not apply our model in this regime .
fig.[fig6](c ) shows absorption images and horizontally integrated optical density ( od ) profiles after the third evaporation stage for a few final laser powers near the condensation transition .
we fit the density profiles to a bimodal distribution consisting of gaussian thermal and thomas - fermi bec profiles .
we detect nearly pure @xmath0yb condensates of @xmath145 atoms for @xmath146w , with a total cycle time of @xmath147s . by shortening the mot loading and compression time to @xmath148s , we produce becs of @xmath149 atoms with a total cycle time of @xmath150s . rapid production of @xmath0yb becs with a total cycle time of @xmath147s .
( a , b ) evolution of number and horizontal ( @xmath151 ) and vertical ( @xmath152 ) temperatures during first 2 phases of rapid evaporation ( see text ) .
( c ) absorption images ( insets ) and horizontally integrated od profiles after the 3rd phase of evaporation and 25 ms time - of - flight ( tof ) to a variable final laser power @xmath153 .
solid lines are fits to a bimodal density distribution ( see text ) , with dashed lines indicating fits to the thermal component . for @xmath154
w , we find a nearly pure condensate of @xmath145 atoms . ]
we now turn our attention to the production of large atom number condensates . as shown in fig.[fig5](d ) , the 3-body loss rate grows noticeably near the point @xmath122 due to a large increase in the density .
therefore , in order to produce the largest condensates we evaporate with the same functional ramps as in fig.[fig5 ] until @xmath155 , and subsequently continue forced evaporation by fixing the power and increasing the cpm amplitude .
fig.[fig7 ] shows an example absorption image and integrated od profile of a pure bec produced from such an evaporation ramp . for this particular measurement
we finish the initial evaporation stage at @xmath155 with @xmath156 m and @xmath157 w , and subsequently increase the cpm amplitude to @xmath158 m .
the resulting trap frequencies are @xmath159 hz . following this method we can reliably create pure @xmath0yb condensates of @xmath160 atoms , a factor of 4 improvement over the largest reported yb bec number @xcite , with a total cycle time of @xmath161s .
while an additional cpm degree of freedom from a second orthogonal aom can allow more control over atom cloud compression @xcite or final bec shape @xcite , it is unlikely that it will provide significant improvements to the speed and efficiency of evaporative cooling that we have demonstrated here . in our current implementation ,
we choose the vertical waist and large initial cpm amplitude to load about 50% of the compressed mot .
once a large atom number has been loaded , we control all the relevant parameters for efficient evaporative cooling by dynamically shaping the trap ( fig.[fig5 ] ) to provide an appropriate @xmath42 to maximize @xmath38 and simultaneously provide large @xmath162 for tight confinement against gravity . the cpm amplitude provides a simple way to control the trap aspect ratio over a large range .
furthermore , a large cpm amplitude with @xmath163 can be used to realize 2d confinement , and square - wave modulation can be used to generate multi - well traps .
production of large , pure @xmath0yb becs using dynamical trap shaping to combat 3-body inelastic loss .
absorption image ( left ) and horizontally integrated od profile ( right ) after 50 ms tof , with fit ( solid line ) to a thomas - fermi bec density distribution , yielding a total of @xmath160 condensed atoms .
the thermal fraction is consistent with zero . ]
our fast and efficient quantum gas production methods can be applied fruitfully to other atomic species , fermions and mixtures , and can significantly impact various applications , such as atom interferometry @xcite .
the demonstrated fast bec production time in this work is at least one order of magnitude shorter than that in typical quantum degenerate gas experiments .
the leading limitation to our cycle time is the mot loading rate , stemming from the relatively narrow linewidth of the @xmath164 transition .
combining broad- and narrow - line laser cooling either spatially @xcite or temporally @xcite can shorten the cycle time further , and is also applicable to other alkaline - earth @xcite as well as lanthanide @xcite atoms .
our scheme should also improve the cycle time of alkali atom experiments which feature the advantage of sub - doppler cooling with relatively broad transitions as well as the combination of broad and narrow transitions @xcite .
the broad applicability of our method to all laser - cooled atomic species gathers additional appeal when one considers that most optical trapping experiments already involve one aom to control the power during evaporative cooling .
the short cycle time can also allow for further technical simplifications including reduced vacuum requirements and the use of lower odt powers at wavelengths closer to atomic resonance . in alkali atoms such as rb and cs , where laser cooling can produce samples at a few @xmath5k with psd @xmath165 @xcite ,
an order of magnitude lower initial trap depth than used here is adequate .
estimating for the commonly used @xmath9rb atom , where @xmath32 and @xmath89 @xcite are similar to @xmath0yb , it should be possible to have @xmath166s bec production times with only @xmath167w of odt power at @xmath168 nm detuning .
we thank chao yi chow for various technical contributions to the experiment .
we gratefully acknowledge financial support from nsf grant no .
phy-1306647 , afosr grant no .
fa 9550 - 15 - 1 - 0220 , and aro muri grant no .
w911nf-12 - 1 - 0476 .
fm waveform for parabolic time - averaged potential . ( a )
the 3 roots of the implicit equation @xmath169 plotted in the domain @xmath170 , corresponding to a half period . the root of interest for cpm is centered on the @xmath171 axis ( root 2 ) .
( b ) time - averaged cpm potentials using root 2 from ( a ) as the fm waveform for delta function ( dashed black line ) and gaussian ( solid red line ) initial beams . ] in order to derive the fm waveform necessary to create the parabolic time - averaged potential shown in fig .
1(b ) , we work with 1d transverse profiles in the painting dimension , @xmath63 .
consider the unmodulated beam shape to be a delta function @xmath172 . in this case , for a cpm amplitude of @xmath112 we want the time - averaged intensity distribution to be of the form @xmath173 where @xmath174 is the heaviside function .
to realize this potential we want to find the function @xmath175 such that the time average of @xmath176 from @xmath177 to @xmath178 equals the expression in equation ( [ eqn : timeaveintensity ] ) , where @xmath179 and @xmath180 .
we reason that the rastered delta function must spend an amount of time at each point @xmath181 that obeys @xmath182 writing @xmath183 as @xmath184 we find @xmath185 treating @xmath186 as a constant determined by @xmath187 and @xmath112 , we arrive at a differential equation for @xmath175 @xmath188 solving equation ( [ forcepsode ] ) gives the implicit equation @xmath189 .
the 3 roots are plotted in fig . [
fig8](a ) .
clearly the solution that is centered on the y axis will be the desired root . from the constraint @xmath190
, we find @xmath191 .
anticipating that we will construct the periodic waveform shown in the inset to fig .
[ fig1](a ) , we define the modulation period @xmath192 .
next we compute the time average with @xmath193 and @xmath194 , @xmath195 where @xmath196 satisfies @xmath197 . from equation ( [ forcepsode ] ) we have @xmath198 .
lastly , the integral over the delta function evaluates to zero unless @xmath199 , which is equivalent to @xmath200 .
therefore , @xmath201 fig .
[ fig8](b ) shows the time - averaged potentials for delta function @xmath172 and gaussian @xmath202 initial beam shapes , where @xmath203 .
as seen in the figure , the cases of the delta function and gaussian differ very little , and become indiscernible when @xmath112 is much greater than the gaussian waist . in order to perform frequency modulation with the waveform
@xmath175 corresponding to the second root in fig . [ fig8](a ) , we utilize the arbitrary waveform functionality of a stanford research systems ds345 function generator , passing it an array of values very closely approximating the periodic continuation of @xmath175 ( see inset to fig . [ fig1 ] ) . | we report on a general method for the rapid production of quantum degenerate gases . using @xmath0yb
, we achieve an experimental cycle time as low as @xmath1s for the production of bose - einstein condensates ( becs ) of @xmath2 atoms .
while laser cooling to @xmath3k proceeds in a standard way , evaporative cooling is highly optimized by performing it in an optical trap that is dynamically shaped by utilizing the time - averaged potential of a single laser beam moving rapidly in one dimension .
we also produce large ( @xmath4 ) atom number becs and successfully model the evaporation dynamics over more than three orders of magnitude in phase space density .
our method provides a simple and general approach to solving the problem of long production times of quantum degenerate gases . |
in the last few years it has become possible to observe details of absorption by singly ionized helium .
the observations combine new information about the history of quasars , intergalactic gas , and structure formation .
early observations of the heii ly@xmath1 absorption spectral region included the quasars q0302 - 003 ( z=3.285 , jackobsen et al .
1994 , hs 1700 + 64 ( z=2.72 , davidsen et al . 1996 ) and pks 1935 - 692 ( z=3.18 tytler & jackobsen 1996 ) .
higher resolution ( ghrs ) observations of q0302 - 003 hogan , anderson & rugers 1997 and he 2347 - 4342 ( z=2.885 , reimers et al .
1997 ) revealed structure in the absorption which could be reliably correlated with hi absorption .
heap et al .
have followed up with stis on q0302 - 003 ( 1999a ) and he 2347 - 4342 ( 1999b ) .
the second observation should be particularly illuminating since he 2347 - 4342 is relatively bright allowing a high resolution grating to be used .
the anderson et al .
( 1998 ) observations of pks 1935 - 692 with stis yield good zero level estimates important for estimating the optical depth @xmath6 .
preliminary reductions of longer stis integrations of pks 1935 - 692 with the 0.1 slit by anderson et al .
( 1999 ) confirm the 1998 results . taken together , these data now appear to be showing the cosmic ionization of helium by quasars around redshift 3 .
although it is possible that the medium is already ionized to heiii by other sources at a lower level ( miralda - escud et al .
1999 ) .
all of the objects show absorption with mean @xmath7 at redshifts lower than the quasar . for the higher redshift qso s q0302 - 033 and pks 1935 - 692 ( shown in fig .
1 ) there is a clear shelf of @xmath8 in a wavelength region of order 20 in observed wavelength blueward of the quasar emission line redshift , dropping to a level consistent with zero flux or @xmath9 beyond that .
the sharp edge led anderson et al . to conclude that gas initially containing helium as mostly heii was being double - ionized in a region around the quasars .
the lack of a strong emission line for heii ly@xmath1 suggests that ionizing flux is escaping so that the 228 flux may be similar to a simple power - law extension of the observed 304 rest frame flux .
hogan et al .
( 1997 ) used the same reasoning to estimate the time required for quasars to create the double ionized helium region to be 20 myr for a 20 shelf ( dependent on the hubble parameter , spectral hardness , cosmology , baryon density and the shelf size ) .
the features present in heii ly@xmath1 spectra are reflected in the hi ly@xmath1 forest for these quasars .
attempts to model the heii absorption with line systems detected in hi suggest that very low column hi absorbers , difficult to differentiate from noise in hi spectra , provide a substantial contribution to the heii absorption .
typically , in the shelf region , the ratio of heii to hi ions is of order 20 or more , rising to at least 100 farther away ( the cross - section for heii ly@xmath1 absorption is 1/4 that of hi ) .
both pks 1935 - 692 and he 2347 - 4342 display conspicuous voids in the heii absorption near the apparent edge of the heiii bubble with corresponding voids in the hi spectra .
reionization and the origin of the heii ly@xmath1 forest can addressed with detailed theoretical treatments ( eg .
zheng & davidsen 1995 , zhang et al .
1998 , fardal et al .
1998 , abel & haehnelt 1999 , gnedin 1999 ) .
wadsley , hogan & anderson ( 1999 ) presented numerical models of the onset of full helium reionization around a single quasar .
the models integrated one - dimensional radiative transfer along lines of sight taken from cosmological hydrodynamical simulations and used flux levels comparable to pks 1935 - 692 . the results reinforced the basic intepretation of pks 1935 - 692 by hogan et al .
( 1997 ) as the growth of an he iii bubble over time in a medium that was mostly heii .
in addition void recoveries in the large ionized bubbles similar to that observed for pks 1935 - 692 occurred often among a random set of simulated lines of sight .
the gas temperature in such bubbles is strongly affected by the ionization of he ii to he iii , particularly because the first photons to reach much of the gas will be quite hard since the softer photons are absorbed close to the quasar until the gas is optically thin .
in particular the underdense medium reaches temperatures of order 15000k .
we make a simple analytical argument for this result in the next section .
given our fairly good guess as to the physical conditions near pks 1935 - 692 we are in position to attempt to convert the observed optical depths into a density in baryons .
comparing optical depths measured for the small sample of voids in the shelf region of pks 1935 - 692 to distributions of void widths and densities from simulations we can build an estimate of @xmath0 , the total cosmic density in baryons .
the are still uncontrollable systematic uncertainties in this estimate but these differ from other techniques and can be addressed with better simulations and a larger sample of quasars .
the main focus of this paper is the technique .
in he ii ly@xmath1 absorption spectra only the voids are sufficiently low density to allow measurements of the optical depth .
the highly ionized void gas is optically thin to the ionizing radiation and cool enough to ignore collisional effects , resulting in a fairly simple relation between optical depth and gas density , where @xmath10 is the optical depth for absorption of he ii ly@xmath1 , @xmath11 is the number density of he ii ions , @xmath12 converts from real space to wavelength ( velocity ) space and @xmath13 is the gas density .
the recombination rate @xmath14 contains the only temperature dependence .
the photo - ionizations per second @xmath15 is determined from the rest frame heii ionizing flux , @xmath16 can be extrapolated from the rest frame flux at @xmath17 which is estimated ( assuming a cosmological model ) from the observed continuum flux at @xmath18 , where @xmath19 is the redshift of the quasar .
the fraction of heii depends on temperature roughly as @xmath20 through the recombination coefficient .
the radiative transfer models of wadsley et al .
( 1999 ) gave temperatures around 15000k for the underdense gas near quasars .
the optical depth to a he ii ionizing photon travelling @xmath21 ( the size of the pks 1935 - 692 shelf region ) at redshift @xmath2 ( scdm , @xmath22 , @xmath23 ) is approximately @xmath24 .
the optical depth falls to 1 for photons with energies , @xmath25 .
thus the heii in the outer half of the shelf region will be ionized preferentially by photons in the energy range 100 - 200 ev . injecting 100 - 54.4
ev per helium atom is equivalent to a temperature increase of 13000 k when the energy is distributed among all the particles .
the cooling time for the underdense gas is very long , dominated by adiabatic expansion .
gas elsewhere will gain @xmath26 4000 k due to heiii reionization .
as voids evolve they approach an attractor solution resembling empty universes and their relative growth slows when they reach around @xmath27 times the mean density , making that value fairly representative .
we used @xmath28-body simulations of 3 cosmologies ( standard cdm , open cdm and @xmath5cdm ) to get probability distributions for the quantity we label normalized opacity @xmath29 as a function of void width .
@xmath30 may be directly related to the void optical depths via ( [ opacity ] ) .
the set of voids allows a statistical maximum likelihood comparison of the simulated distribution of @xmath30 to the distribution of void widths and optical depths observed .
figure [ taunorm ] shows the probability distribution of @xmath30 in voids ( equivalent to void optical depth ) and void widths for the @xmath5cdm cosmology .
the optical depth from the simulation was normalized so that mean density gas expanding with the hubble flow gives an optical depth of 1.0 .
the effect of gas pressure on the dynamics of gas in the voids should be negligible and was ignored in our simulations .
thus the gas follows the dark matter exactly .
the simulations were performed using hydra ( couchman 1991 ) with 64 mpc periodic simulation volumes ( @xmath31 @xmath32 at @xmath2 ) containing @xmath33 particles .
a preliminary convergence study indicated that the statistics of @xmath30 are relatively insensitive to resolution however our resolution is still lower than that of the best current lyman-@xmath1 forest studies which also include gas .
we were constrained by the need to include larger scales so that the large void portion of our sample was reasonable .
higher resolution gives slightly larger estimates for @xmath0 .
the ratio of @xmath10 to @xmath30 that has the maximum likelihood gives an estimate for the total baryon density as follows , @xmath34 in this equation all the quantities are given as ratios with typical values estimated from pks 1935 - 692 , with @xmath35 being the local density divided by the cosmological mean density , @xmath36 the local expansion rate along the line - of - sight compared to the mean , @xmath37 local hubble parameter value , @xmath1 the spectral slope of the quasar ( @xmath38 ) and @xmath39 the helium mass fraction .
the value of @xmath40 is @xmath41 at @xmath2 for an @xmath42 scdm cosmology .
the exact systemic redshift of pks 1935 - 692 is uncertain due to an absence of narrow ir emission lines .
we use @xmath43 .
we used the 5 large voids observed over the range @xmath44 for pks 1935 - 692 .
this range should be far enough from the quasar to avoid `` associated '' absorbers caught in the quasar outflow .
the fits are shown in figure [ taufit ] .
he ii and h i absorption are both plotted with the models for each overplotted as dashed lines .
the he ii spectrum was modeled with the known hi ly@xmath1 absorption lines and 5 parameter values for the optical depth in each void .
the ratio of the hi to heii optical depths was a sixth free parameter .
the models were convolved with the same spectral point spread function as the observations and include same level of photon shot noise .
the best model was determined using maximum likelihood and the fitting errors determined with monte carlo realizations .
the emptiest void has a reasonable probability according to the simulations , however since it has been suggested that it could be due to a local ionizing source we calculated the effect of removing it , which was a @xmath45 increase in the estimate for @xmath0 .
the ionizing flux was inferred from the observed pks 1935 - 692 continuum level at 304 @xmath46 rest frame and extrapolated to @xmath47 and 228 @xmath46 . assuming the quasar redshift is @xmath43 this gives a flux of @xmath48 ( for standard cdm ) at @xmath49 which is in the middle of the voids .
the sample of helium quasars is quite small and though pks 1935 - 692 does not have any known variability all quasars are thought to be intrinsically variable due to the nature of the fueling and emission processes .
selection effects favour the idea that pks 1935 - 692 is currently brighter than its average value over the last few thousand years ( the ionization response timescale ) .
if pks 1935 - 692 has recently brightened then ( [ ombh2eqn ] ) implies that our @xmath0 estimate is biased upward
. if there is additional foreground continuum absorption then our estimates are biased low .
the results are tabulated below with 1-@xmath4 fitting errors indicated ( monte carlo ) . aside from the uncertainties mentioned above , there is still uncertainty in the temperature through the @xmath50 dependence .
if the gas was pre - ionized to heiii the underdense gas could be colder by a factor of order 2 which would lower the @xmath51 estimates by 20 % .
a larger sample of quasars would make the greatest improvement in the robustness of this measurement . | a new method to extract @xmath0 from high redshift intergalactic absorption is described , based on the distribution of heii ly@xmath1 optical depths in the voids in the ionization zone of quasars .
a preliminary estimate from recent hst - stis spectra of pks 1935 - 692 at @xmath2 gives @xmath3 ( 1-@xmath4 statistical errors , for a @xmath5cdm cosmology ) consistent with other estimates . |
the spectral slope optical / uv variations in agns have been quantified by @xcite , by means of the _ spectral variability parameter _ @xmath0 , @xmath1 being the slope of the spectrum and @xmath2 its flux in the given band .
a positive value was found , which means that the spectrum is harder when the flux is higher .
+ however , in the x - ray band , the opposite behaviour has been found for individual sources , and very few systematic studies have been performed , e.g. @xcite , who found this trend for a sample of 10 nearby seyfert galaxies .
+ in our study we investigated the spectral variability of quasars , using the mexsas catalogue @xcite , created cross - matching the multi - epoch observations of the xmm - newton serendipitous source catalogue , data release 5 @xcite with two partially overlapping sloan digital sky survey catalogues , sdss - dr7q @xcite and sdss - dr12q @xcite , obtaining a catalogue of 7,837 x - ray observations of 2,700 quasar sources .
in order to study the spectral variability in the x - ray band , we redefine the spectral variability parameter in terms of the photon index @xmath3 , defined after @xmath4 .
then , the spectral variability parameter becomes @xmath5 , since @xmath6 . in order to better compare sources we computed the linear fit between the variations of @xmath3 and @xmath7 ( @xmath8 in band @xmath9kev ) from the source mean values .
the computed spectral variability parameter is @xmath10 ( see fig . [
fig : ensemble ] ) .
a negative @xmath11 implies that the spectral slope becomes higher for increasing flux , making the spectrum steeper .
this means that the spectrum is softer when brighter .
+ we tried to investigate the dependence of @xmath11 from some source parameters , such as black hole mass , eddington ratio , redshift and x - ray luminosity , finding no evidence of such dependence .
the value of @xmath11 in some bins deviates significantly from the ensemble value , though , suggesting that different sources may have different values of @xmath11 between each other .
therefore we studied single sources , selecting the ones with low probability of finding a @xmath12 correlation by chance ( @xmath13 ) .
we find all negative @xmath11s in a wide range of values , from @xmath14 to @xmath15 .
( band @xmath9kev ) correlation of the whole sample . ]
the softer when brighter trend , found in both the ensemble and single - source analysis , may be caused by the superposition of the primary x - ray emission with an additional reflected component .
according to some models ( e.g. , * ? ? ?
* ) the primary emission could be variable in flux , but not in spectrum , while the constant reflected component is not variable at all , producing the observed spectra .
other models ( e.g. , * ? ? ?
* ) , suggest that the primary component may be variable in spectrum as well .
+ as for the single source analysis , the wide range of @xmath11 values may be dependent on several unconsidered features of the source , such as black hole spin , angle of view or radio - loudness , or it may be influenced by stochastic processes .
liang , e. p. t. , 1979 , _ apj _ , * 231 * , l111 .
pris , i. , petitjean , p. , ross , n. p. et al . , 2016 , ` arxiv:1608.06483 ` rosen , s. r. , webb , n. a. , watson , m. g. et al . , 2016 , _
a&a _ , * 590 * , a1 .
schneider , d. p. , richards , g. t. , hall , p. b. , 2010 , _ aj _ , * 139 * , 2360 .
shih , d. c. , iwasawa , k. , fabian , a. c. , 2002 , _ mnras _ , * 333 * , 687 .
sobolewska , m. a. , papadakis i. e. , 2009 , _ mnras _ , * 399 * , 1597 .
trevese , d. , vagnetti , f. , 2002 , _ apj _ , * 564 * , 624 .
vagnetti , f. , middei , r. , antonucci , m. et al , 2016 , _ a&a _ ,
* 593 * , a55 . | variations of the x - ray spectral slope have been found in many active galactic nuclei ( agn ) at moderate luminosities and redshifts , typically showing a `` softer when brighter '' behaviour .
however , similar studies are not usually performed for high - luminosity agns .
we present an analysis of the spectral variability based on a large sample of quasars in wide intervals of luminosity and redshift , measured at several different epochs , extracted from the fifth release of the xmm newton serendipitous source catalogue .
our analysis confirms a `` softer when brighter '' trend also for our sample , extending to high luminosity and redshift the general behaviour previously found .
these results can be understood in light of current spectral models , such as intrinsic variations of the x - ray primary radiation , or superposition with a constant reflection component . |
in the simplest models undergoing absorbing state phase transitions in the directed percolation ( dp ) universality class , like the contact process ( cp ) , the stationary state of the system in the inactive phase is the state devoid of particles .
other models in the same class , like the pair contact process ( pcp ) , have however a richer structure , associated with the existence of an infinite number of absorbing states @xcite . in the inactive phase ,
the absorbing state the system evolves to depends on the initial conditions and the distance from the critical point , and so does the average density of isolated particles in the stationary state . in the case of pcp ,
the field responsible for the dynamics ( the density of pairs of particles , @xmath8 ) is coupled to another field ( the density of isolated particles , @xmath1 ) .
this background of isolated particles is responsible for the non - universality of some dynamic properties of the system at criticality @xcite .
recently the one - dimensional pcp with particle diffusion ( known as pcpd or annihilation / fission model ) has received a lot of attention and was at the center of some controversy @xcite ; nevertheless its critical behaviour is not yet fully clarified .
@xcite presented a pair approximation study of this model which however is not valid for the non - diffusive case .
the expected behaviour of pcp is qualitatively different from the @xmath9 limit of pcpd : in the non - diffusive case it is crucial to distinguish between _ particles _ and _ isolated particles _ , since all activity stops as soon as the number of pairs vanishes , whereas isolated particles may produce a pair if particle diffusion is allowed . in pcp ,
as the system approaches the critical point from the active phase , the nonordering field approaches a ` natural ' value ( @xmath10 ) , and its behavior is described by the same power laws as those of the order parameter @xcite . in the present work
we investigate the behavior of this model in the inactive phase by looking at properties of its _ natural absorbing states _ , the absorbing states selected by the system s dynamics .
as shown by both mean - field - like approximations and monte carlo ( mc ) simulations , the stationary density of isolated particles develops a power law singularity as it approaches the critical point .
in the inactive phase , the concentration of pairs is known to decay exponentially : @xmath11 , with @xmath12 .
the exponent @xmath13 is associated with the temporal correlation length and mean - field theory predicts @xmath14 .
now , if one is interested in studying the stationary single particle concentration as a function of @xmath15 , one can refer to the coupled langevin equations describing this dynamical process @xcite . at the mean - field level , the time evolution of @xmath1 is given by @xmath16 obviously coupled to the evolution of @xmath8 . with the change of variables
@xmath17 , equation ( 1 ) reads @xmath18 in the following , we will assume @xmath19 .
then , equation ( 2 ) can be solved exactly : @xmath20 with @xmath21 in the limit @xmath22 , one has @xmath23 in the vicinity of the critical point , when @xmath24 , @xmath25 decreases more rapidly than @xmath26 , and one can therefore conclude that @xmath27 at the critical point , @xmath28 , i.e. @xmath29 @xmath3 is then given by @xmath30 , as already found in @xcite .
but now we can see that @xmath31 varies with @xmath2 , in the inactive phase : @xmath32 we must point out that the present argument relies on a mean - field approach and on the assumption that @xmath33 ( which may not be true at early times ) .
a variation of @xmath31 with @xmath2 is then plausible .
however , the prediction that the critical exponent of the quantity @xmath34 should be equal to @xmath13 is only valid within a mean - field approximation and is not expected to apply to the @xmath0 pcp model .
mean - field - like kinetic equations for the pcpd were obtained in @xcite .
whereas the single - site approximation is only appropriate to the high diffusion limit ( @xmath35 ) , the pair approximation gives a good qualitative picture of the model for @xmath36 , but does not show some important characteristics of the pcp without diffusion . indeed , according to the pair approximation , the steady state single particle density vanishes at the critical point ( @xmath37 for the pair approximation with @xmath9 ) ; however , it is well established that , at the critical point , the single particle density approaches a nonzero value , @xmath3 ( @xmath38 , in a sequential dynamical process ) . also , the pair approximation predicts a power - law temporal decay for the pair concentration in the inactive phase , contrary to the exponential decay found in the simulations . in the coarse - grained langevin description ,
the fields that are used to characterize the system configurations are the local pair density and the local density of isolated particles .
if one uses two - site clusters , as in the pair approximation , configurations with isolated particles are not treated appropriately .
indeed , if one aims at building a better approximation , one has to go to three - site clusters and consider @xmath39 , the probability of having an isolated particle at the center of the cluster .
the price to pay is that of increasing complexity : the number of variables and equations increases and these have to be solved numerically ; carlon _ et al . _
dealt with two variables and two equations , so they were able to obtain some results analytically .
we have used a @xmath40 - cluster approximation @xcite whose technical details are given in the appendix . in fig.1
, we have plotted the stationary values of @xmath41 and @xmath42 as functions of @xmath2 . as shown
, the concentration of pairs vanishes for @xmath43 , i.e. @xmath44 within the present approximation .
the estimate obtained by mc simulations is @xmath45 ; so , as expected , the present result is an improvement when compared to what was obtained by the pair approximation .
more interestingly , the concentration of particles at the critical point ( @xmath46 , since @xmath47 ) is nonzero ; indeed @xmath48 , a value not far from @xmath49 , as obtained by the simulations @xcite . on the other hand , a linear decrease of @xmath1 for @xmath5 can also be noticed .
in this work , we concentrate on the critical behaviour of @xmath1 in the inactive phase .
the simulations that we present were done on systems of size @xmath50 ; to ensure a sufficient number of pairs we chose an initial particle concentration of 0.5 .
the numbers of particles and of pairs were recorded versus time up to a number of monte carlo steps , @xmath51 , which ranged from @xmath52 for @xmath2 very close to the transition ( @xmath53 ) down to @xmath54 for @xmath55 .
these times were chosen such that most of the samples ( typically around @xmath56 ) had already entered the absorbing state .
for each @xmath2 we evaluated the average activity time @xmath57 . to obtain the stationary value of @xmath58
, we have averaged the final number of particles in those samples which entered the absorbing state at a time around @xmath57 .
the short lived samples were excluded in order to eliminate finite size effects . on the other hand , in systems of this size , and for @xmath2 close to @xmath59 ,
the concentration of pairs is still considerably high ( typically around @xmath60 ) at very long times - thus the concentration of isolated particles is lower than in samples where the number of pairs is vanishingly small .
for this reason , we have also ignored the contribution from these long - lived samples .
we have chosen to consider samples with @xmath61 , but have checked that the results are stable with respect to other choices in a reasonable range . in fig.@xmath62
we show the stationary single particle concentration as a function of @xmath2 .
fig.@xmath63 shows a log - log plot of @xmath64 vs @xmath65 , with @xmath66 and @xmath67 .
a linear fit of the data is clearly appropriate and leads to an exponent @xmath68 .
this seems to be a new exponent , not simply related to the dp exponents that describe the critical behaviour of other quantities , in the active phase of the pcp model .
in this work , we show that the approach to the critical point in the in inactive phase of the @xmath0 pcp model is signaled by a power - law singularity of a static quantity - the deviation of the particle concentration from its critical point value , @xmath10 .
a cluster - mean - field approximation was developed that predicts a finite value of @xmath3 , in reasonable agreement with mc simulations and a linear decrease of @xmath1 with @xmath69 ; this is confirmed by a phenomenological approach . such a decrease in @xmath1 is also shown in our mc simulations ; the best fit to the data is consistent with an exponent @xmath70 , different from the mean - field prediction .
critical behavior of a static quantity in the inactive phase has been observed by lipowski and droz @xcite in a rather different model with infinite number of absorbing states . in that case , a random walk argument leads to a relationship between the corresponding exponent and the order parameter exponent @xmath71 . in the case of pcp
such an argument can not be applied and whether @xmath72 is related to the ( dp ) critical exponents that characterize the active phase of pcp requires further investigation .
the same applies to other systems with infinitely many absorbing states .
we thank ronald dickman for a stimulating discussion and a critical reading of the manuscript and adam lipowski for communicating his work prior to publication .
partial support from project pocti/1999/fis/33141 is acknowledged .
in the @xmath40-cluster approximation @xcite one uses @xmath73site clusters and allows for an overlap of @xmath62 sites between adjacent clusters .
thus , the probability of a @xmath74site cluster in the state @xmath75 is written as we use @xmath77 , @xmath78 , @xmath79 , @xmath80 as independent variables .
it can be easily shown that @xmath81 , and @xmath82 .
we then study all the processes that can occur when a pair belonging to a @xmath74site cluster is selected .
take for example the configuration @xmath83 . according to the above approximation , the probability of this configuration is @xmath84 , where @xmath85 .
when the central pair is selected , then , with rate @xmath2 , the configuration @xmath86 is generated .
the variation in the number of @xmath73site clusters in the state @xmath87 is then @xmath88 ; analogously , the changes in the number of @xmath73site clusters in configurations @xmath89 and @xmath90 are , respectively , @xmath91 . on the other hand , according to the pcp dynamic rules , a pair can also create a new particle at a randomly chosen nearest neighbour provided this is vacant . if one considers the configuration @xmath92 ( whose probability is given , within the present approximation , by @xmath93 , with @xmath94 ) , then , with rate @xmath95 , the configuration @xmath83 is generated .
this change corresponds to @xmath96 .
the kinetic equations for @xmath97 are obtained by considering all the contributions from all the possible @xmath74site configurations : @xmath98\nonumber\\ & + & ( 1-p ) [ 3a_{11}+a_{12}+3a_{13}+2a_{14}+a_{15}+a_{16}\nonumber\\ & + & 2a_{17}+a_{18 } ] \nonumber\\ \frac { dy}{dt } & = & p [ a_{1}-a_{4}-a_{5}-2a_{7}-2a_{8}-2a_{9}-a_{10}]\nonumber\\ & + & ( 1-p ) [ -a_{11}-a_{13}]\nonumber\end{aligned}\ ] ] | steady state properties in the absorbing phase of the @xmath0 pair contact process ( pcp ) model are investigated .
it is shown that , in typical absorbing states ( reached by the system s dynamic rules ) , the density of isolated particles , @xmath1 , approaches a stationary value which depends on the annihilation probability ( @xmath2 ) ; the deviation from its natural value at criticality , @xmath3 , follows a power law : @xmath4 for @xmath5 .
monte carlo simulations yield @xmath6 .
a cluster approximation is developed for this model , qualitatively confirming the numerical results and predicting @xmath7 .
the singular behavior of the isolated particles density in the inactive phase is explained using a phenomenological approach . 2 |
a deconfined phase of nuclear matter is expected to be formed at the high energy densities created in relativistic heavy ion collisions @xcite .
it is widely believed that important signatures of such a phase are reflected in the space - time extent and shape of particle emission source functions .
recently , 1-dimensional source imaging techniques @xcite have revealed a non - trivial long range structure in the two - pion emission source at rhic @xcite .
the origins of this structure are still unclear .
the presence / absence of such a structure in the pion emission sources in heavy ion collisions at intermediate sps energies could yield important information which could help resolve the structure s origins .
the na49 collaboration has carried out pb+pb collisions over a wide range of bombarding energies at the cern sps during the last decade @xcite .
such a rich data set provides a unique opportunity to search for this long range structure at the sps and study its evolution with beam energy with a view to unraveling its nature . in this paper ,
the 3-dimensional emission source images for pions produced in central pb+pb collisions over the incident energy range 40 and 158 agev are presented .
the results are discussed in the context of a gaussian shape assumption .
the data presented here were taken by the na49 collaboration during the years 1996 - 2002 .
the incident beams of 40 and 158 agev were provided by the cern sps accelerator .
the na49 large acceptance hadron detector @xcite achieves large acceptance precision tracking ( @xmath0 ) and particle identification using time projection chambers .
charged particles are detected by tracks left in the tpc and identified by the energy deposited in the tpc gas .
mid - rapidity particle identification is further enhanced by a time - of - flight wall ( resolution 60ps ) .
event centrality is determined by a forward calorimeter which measures the energy of spectator matter .
3d correlation functions , c(@xmath1 ) , were obtained as the ratio of pair to uncorrelated reference distributions in relative momentum @xmath1 for @xmath2 pairs . here , @xmath3 is half of the relative momentum between the two particles in the pair center - of - mass system ( pcms ) .
the pair distribution was obtained using pairs of particles from the same event and the uncorrelated distribution was obtained by pairing particles from different events .
track merging and splitting effects were removed by appropriate cuts on both the pair and uncorrelated distributions .
momentum resolution effects were negligible . in the cartesian harmonic decomposition technique @xcite , the 3d correlation function
is expressed as @xmath4 where @xmath5 , @xmath6 , @xmath7 are cartesian harmonic basis elements ( @xmath8 is solid angle in @xmath1 space ) and @xmath9 are cartesian correlation moments given by @xmath10 the coordinate axes are oriented so that z is parallel to the beam ( long ) direction , x points in the direction of the total momentum of the pair in the locally co - moving system ( lcms ) ( out ) and y is perpendicular to the other two axes ( side ) .
the correlation moments , for each order @xmath11 , are calculated from the measured 3d correlation function using equation ( [ eqn2 ] ) .
each independent correlation moment is then imaged using the 1d source imaging code of brown and danielewicz @xcite to obtain the corresponding source moment for each order @xmath11 .
bose - einstein symmetrisation and coulomb interaction ( the sources of the observed correlations ) are contained in the source imaging code .
thereafter , the total source function is constructed by combining the source moments for each @xmath11 as in equation ( [ eqn3 ] ) @xmath12
figure [ 40_160gev_fit ] shows the @xmath13 ( @xmath14 ) and @xmath15 ( @xmath16 and @xmath17 ) moments for mid - rapidity ( @xmath18 , where @xmath19 is particle laboratory rapidity and @xmath20 is cm rapidity ) , low @xmath21 ( @xmath22 ) @xmath2 pairs from 158 agev central ( 7@xmath23 ) pb+pb collisions , as a function of the relative momentum q in the pcms frame . here , @xmath16 is shorthand for @xmath24 etc .
moments for other l s are either zero or negligible .
the lorentz transformation of @xmath1 from the laboratory frame to the pcms is done by transforming to the lcms along the beam axis and then transforming to the pcms along the pair transverse momentum .
the data are represented in fig .
[ 40_160gev_fit ] by solid circles , while squares represent the results of fits to the independent moments with model sources , an ellipsoid shape ( 3d gaussian ) in panels ( a)-(c ) and a 2-source shape in panels ( d)-(f ) .
the last function arises by assuming a linear combination of two gaussians for the single particle distribution .
the best - fit parameters for the two shapes are listed in the lower panels .
the 2-source shape yields a much better representation of the data .
-1.cm given that the only significant imaged moments are found for @xmath25 and 2 multipolarities , the net imaged source function in the x , y and z directions is simply the sum of the 1d source @xmath26 and the corresponding @xmath27 moment @xmath28 where i= x , y or z. in panels ( a)-(c ) , figure [ 40_160gev_src ] compares the net imaged source ( squares ) for mid - rapidity , low @xmath21 pion pairs from central pb+pb collisions at 40 agev ( i.e @xmath29agev ) to the best - fit parametrized functions , gaussian ( triangles ) and 2-source model ( circles ) . the source image and the 2-source model
agree very well in all 3 directions and both disagree with the gaussian fit in the z direction . at 158 agev ( i.e @xmath30agev ) , represented in panels ( d)-(f ) of fig .
[ 40_160gev_src ] , the findings are similar , but in addition to the z direction , the source function starts to develop a non - gaussian tail in the x direction
. -1.cm
the ratio of the rms radii of the source functions in the x and y directions is [email protected] at 40 agev and [email protected] at 158 agev .
this deviation from unity , evident visually , points to a finite pion emission time . moreover , the rms pair separation in the z direction , from figure [ 40_160gev_src](c ) and ( f ) , is 11fm at 40 agev and 12fm at 158 agev .
these dimensions are much smaller than the lorentz - contracted nuclear diameters of 3fm at 40 agev and 1.5fm at 158 agev and in fact give the rms pair separation due to the longidutional spread of nuclear matter created by the passage of the two nuclei .
since the latter are moving with almost the speed of light , one can infer the lower bound formation time of the created nuclear matter to be 8fm / c at 40 agev and 10fm / c at 158 agev .
10 qm2002 , nucl . phys . * a 715 * , 1c ( 2003 ) p. chung et al , nucl . phys . * a 749 * , 275c ( 2005 ) s.s .
adler et al.(phenix collaboration ) , nucl - ex/0605032 d.a .
brown and p. danielewicz , phys . lett . * b 398 * , 252 ( 1997 ) d.a .
brown and p. danielewicz , phys . rev .
* c 57 * , 2474 ( 1998 ) c. alt et al.(na49 ) , cern - spsc-2005 - 041 , cern - spsc - p-264-add-12 , nov 2005 . s. afanasiev et al . , nucl .
a430 * 210 ( 1999 ) .
p. danielewicz and s. pratt , phys
* b 618 * 60 ( 2005 ) . | the na49 experiment at cern sps has acquired a huge data set of pb+pb events over a broad range of energy and centrality during the last several years .
this high statistics data set , coupled with a state - of - the - art analysis technique , allows for the first model - independent extraction and energy scan of 3d emission sources for pion pairs at sps energies .
these 3d pion emission sources provide new insights into the nature of a long - range source previously reported by phenix at rhic .
the new results indicate that the pion source displays significant non - gaussian tails in the longitudinal direction at 40 and 158 agev and in the outward direction at 158 agev . |
polarization in spectral lines is the most direct and reliable signature of the presence of magnetic fields on the surfaces of stars .
while the zeeman broadening of unpolarized spectra can be used , with some ambiguity , to diagnose kg - strength fields in the extremely active stars @xcite , circular polarimetry enables detection of global magnetic fields on the order of one gauss and weaker in moderately active stars @xcite .
the diagnostic power of polarimetry comes from the possibility to measure very precisely magnetic signals in spectral lines against an unpolarized continuum and to boost the signal - to - noise ratio by combining weak polarization signals in thousands of spectral lines recorded simultaneously with a wide wavelength coverage chelle spectrometer @xcite .
furthermore , doppler imaging analysis of the rotational modulation of circular polarization in spectral lines has resulted in detailed maps of stellar magnetic fields ( e.g. , * ? ? ?
* ) , leading to many new insights into the magnetism and activity of different groups of late - type stars @xcite .
despite spectacular results of the recent investigations , spectropolarimetry of stellar magnetic fields remains fundamentally limited .
most importantly , high - resolution spectropolarimetric observations of all late - type stars and of the majority of early - type stars are limited to circular polarization ( stokes @xmath5 ) . a significantly weaker linear polarization ( stokes @xmath6 and @xmath7 )
signal due to the transverse zeeman effect has been detected and used for diagnosing stellar magnetic topologies only for a few strongly magnetic ap stars @xcite and for one magnetic white dwarf @xcite .
previous spectropolarimetric observations of late - type stars in the stokes @xmath6 and @xmath7 parameters were limited to the objects where the presence of a relatively strong linear polarization was unrelated to magnetic fields ( e.g. , * ? ? ?
* ; * ? ? ?
incomplete stokes parameter data sets typically employed for magnetic stars are plagued by a reduced sensitivity to the horizontal magnetic field component and , in general , do not allow a unique reconstruction of the stellar magnetic field geometries @xcite . four stokes parameter studies of ap stars also demonstrated that magnetic models based on the interpretation of circular polarization alone miss the smaller - scale magnetic structures @xcite .
high - resolution full four stokes parameter observations of active cool stars represent a crucial step towards more robust and complete analysis of their magnetic field topologies . here
we present the first results of simultaneous circular _ and _ linear polarization observations of active stars with a new polarimetric device : harpspol .
[ cols="^,^,^,^,^,^,^,^,>,^ " , ] the harps spectrometer @xcite , fed by the 3.6-m telescope at the european southern observatory ( eso ) , is one of the most powerful astronomical tools for precision studies of stellar spectra . it is the highest resolution eso spectrograph , providing a coverage of the 380690 nm wavelength region in a single exposure with the resolving power of @xmath8=115,000 and a long - term velocity stability of better than 1 ms@xmath9 .
the new polarimeter , harpspol @xcite , is installed in the cassegrain focus of the 3.6-m telescope , feeding the two existing harps optical fibers .
the polarimeter consists of two separate polarimetric units : one for circular and another for linear polarization measurements .
each unit is equipped with a modulator a superachromatic quarter - wave or half - wave plate .
the light is split into two orthogonally polarized beams by a polarizing beam - splitter ( foster prism ) .
the corresponding spectra are recorded simultaneously on the 4k@xmath104k ccd mosaic .
a slider inserts the polarimeter from a completely retracted position , and allows for selection of either the circular or the linear polarimeter .
the two orthogonally polarized beams injected into the fibers can be exchanged by rotating the half - wave plate with the increment steps of 45@xmath11 and the quarter - wave plate with 90@xmath11 steps .
a combination of exposures obtained with different modulator orientations allows compensation of , to first order , flat - fielding errors and effects of seeing and other artifacts @xcite .
the high spectral resolution of harpspol is especially useful for detecting complex and weak linear polarization signatures in cool stars @xcite .
each of our polarization observations of cool stars consisted of four sub - exposures , which were calibrated and processed with the help of the reduce package @xcite .
the resulting one - dimensional spectra were demodulated using the `` ratio '' method described by @xcite and re - normalized to the continuum intensity , yielding a spectrum in one of the stokes parameters ( @xmath12 , @xmath13 or @xmath14 ) , the corresponding intensity spectrum ( @xmath15 ) , associated error bars , and a diagnostic null spectrum .
the circular to linear polarization cross - talk is estimated to be below 0.5% using symmetry properties of the stokes profiles for the slowly rotating ap star @xmath16 equ .
table [ tbl : obs ] summarizes the four stokes parameter observations of three late - type stars
hr1099 , @xmath0eri , @xmath3cena obtained during our harpspol runs in may 2009 and january 2010 .
the first six columns in that table give the target name , the observation date , the heliocentric julian date , the @xmath5 , @xmath6 , and @xmath7 stokes parameters , the exposure time , and the peak signal - to - noise ( @xmath17 ) ratio per 0.8 kms@xmath9 velocity bin .
polarization signatures are typically not detectable in individual spectral lines of active late - type stars . among the targets observed in our program ,
only hr1099 shows evidence of the circular polarization signal in a few strong , magnetically sensitive spectral lines . to detect and characterize magnetic fields of the target stars we use a multi - line technique known as least - squares deconvolution ( lsd , *
this method approximates the stellar spectrum with a superposition of identical profiles , centered at the position of individual lines and scaled according to the line s strength and magnetic sensitivity . a mean profile extracted by lsd combines common information from thousands of spectral lines , boosting the sensitivity of polarization studies .
we computed mean profiles for all stokes parameters and the corresponding null polarization spectra using the methodology and the code developed by @xcite .
the lsd line mask was based on the atomic line parameters extracted from the vienna atomic line database ( vald , * ? ? ?
depending on the star , we used between 7500 and 9000 metal lines stronger than 10% of the unpolarized continuum , achieving a multiplex gain of 30100 .
table [ tbl : obs ] summarizes results of the lsd analysis of our targets .
columns 710 give information on the noise level ( per 0.8 kms@xmath9 velocity bin ) in the lsd profiles , the maximum amplitude of the lsd polarization signal , the false alarm probability ( fap ) of the signal detection computed according to @xcite , and the polarization detection flag . following the widely used spectropolarimetric convention
, we consider profiles with fap@xmath18 as a definite detection ( dd ) , @xmath19fap@xmath18 as a marginal detection ( md ) , and fap@xmath20 as no detection ( nd ) .
the k1 subgiant primary of the rs cvn binary hr1099 ( v711 tau ) is one of the most magnetically active late - type stars .
it was frequently studied with doppler imaging @xcite and circular spectropolarimetry @xcite .
these investigations found cool spots evolving on a time scale of about one year and longer - lived magnetic field structures , dominated by regions with predominantly azimuthal field orientation .
we observed hr1099 in all four stokes parameters on two different nights in january 2010 .
[ fig : lsd ] shows the lsd stokes @xmath21 profiles obtained for this star .
the contribution of both stars is evident in stokes @xmath22 .
the distortions in the mean intensity profile of the primary indicate significant spot coverage at the time of our observations .
the rapid changes of stokes @xmath22 are due to the orbital motion and rotation of the primary with @xmath23@xmath24@xmath25 d @xcite .
the stokes @xmath5 profile of the primary shows a complex field morphology .
using the first moment of lsd stokes @xmath5 @xcite , we derive a mean longitudinal magnetic field of @xmath26=@xmath27 g and @xmath28 g for the observations on january 4 and 16 , respectively . a weak magnetic signature in the stokes @xmath5 profile of the secondary component is also clearly detected at @xmath29 kms@xmath9 in the latter observation .
the @xmath6 and @xmath7 signatures of the primary are detected in both harpspol observations with a very high confidence ( fap@xmath30 ) .
this is the first detection of the linear polarization in line profiles of a cool active star .
the lsd stokes @xmath31 profiles of hr1099 are significantly more complex than the corresponding stokes @xmath5 spectra and have an amplitude of @xmath1@xmath32 .
no signal is present in the @xmath6 and @xmath7 null profiles ( fap@xmath33 ) , confirming that detected linear polarization is intrinsic to the star and does not arise due to an instrumental or data reduction artifact .
the k2 dwarf @xmath0eri is one of the nearest solar - type stars and the closest exoplanet host @xcite .
this star is surrounded by a debris disk and shows a high level of magnetic activity .
previous studies of the zeeman broadening in optical and infrared magnetically sensitive lines suggested the presence of fields with spatially averaged strength of 100200 g @xcite .
no detection of the magnetically induced line polarization was previously reported for this star .
we observed @xmath0eri in all four stokes parameters during 11 nights , fully covering the 11.3 d rotation period of the star @xcite .
[ fig : lsd ] illustrates the lsd intensity and polarization spectra of @xmath0eri for the three selected nights .
there is no evidence of variability in stokes @xmath22 , but a variable magnetic field is detected in the circular polarization for all nights .
the stokes @xmath5 lsd profiles change systematically from night to night throughout the rotation period of the star .
the mean longitudinal magnetic field of @xmath0eri varies between @xmath34 and @xmath35 g. despite a high @xmath17 ratio of our polarimetric data , the lsd @xmath6 and @xmath7 profiles of @xmath0eri generally do not reveal magnetic signatures .
however , on the night 11 january 2010 the @xmath6 signature is definitely detected , while on the other two nights a marginal signal may be present in @xmath6 and/or @xmath7 profiles
. the amplitude of these linear polarization signatures does not exceed @xmath36 .
rotational modulation of the linear polarization signal and its occurrence within the stellar photospheric lines suggests that polarization is unrelated to the dusty disk around @xmath0eri .
the solar analogue @xmath3cena was observed with in june 2009 , during partly cloudy conditions .
nevertheless , sufficient photons were collected to reach a polarimetric sensitivity of @xmath1@xmath37 with the help of the lsd technique . as demonstrated by fig .
[ fig : lsd ] , no signals were detected in either @xmath5 , @xmath6 or @xmath7 profiles .
the corresponding null spectra are also featureless , confirming the lack of polarization artifacts down to the level of @xmath1@xmath37 .
the non - detection of the stokes @xmath5 signature corresponds to the @xmath38 upper limit of 0.2 g for the disk - averaged longitudinal magnetic field . in comparison ,
the net longitudinal field of the sun as a star reaches 1 g during activity maximum and is below 0.1 g during the minimum @xcite .
we therefore conclude that harpspol can detect the global magnetic fields of active solar twins . in its present low activity state
, @xmath3cena appears to be reminiscent of the sun at cycle minimum , in agreement with recent x - ray observations @xcite .
for both hr1099 and @xmath0eri we detected linear polarization in spectral lines with an amplitude of 0.615@xmath39 of the unpolarized continuum , which is about a factor of 10 lower than the amplitude of the circular polarization signal recorded on the same nights .
for both objects the lsd linear polarization profiles appear to be morphologically complex , requiring resolving power @xmath40 for secure detection .
it is interesting to discuss the cause of linear polarization in cool stars .
solar - like spots are connected with a patchy , magnetic environment called `` plage '' through low - lying magnetic loops @xcite .
these plage regions are usually brighter than the quiet photosphere and much brighter than the cool spots , and therefore dominate the disk - integrated magnetic signal . for an on - disk active region
, linear polarization can be created if the plage is not symmetrically distributed around the spot . near the limb
, the vertical components of plage can create significant linear polarization as long as there are not too many active regions such that their magnetic signals average out over the disk .
the two polar regions that exhibit a weak but quite uniform outward magnetic field over a large area @xcite could contribute to linear polarization as well . on the other hand , for stars with enhanced activity and non - solar dynamo properties , such as rs cvn binaries , linear polarization can arise from the large - scale toroidal fields mapped by circular polarization zeeman doppler imaging studies @xcite . four stokes parameters observations with complete rotational phase coverage are essential to confirm the reality of these azimuthal fields and distinguish different scenarios of their formation . | the application of high - resolution spectropolarimetry has led to major progress in understanding the magnetism and activity of late - type stars . during the last decade , magnetic fields have been discovered and mapped for many types of active cool stars using spectropolarimetric data .
however , these observations and modeling attempts are fundamentally incomplete since they are based on the interpretation of the circular polarization alone . taking advantage of the newly built harps polarimeter
, we have obtained the first systematic observations of several cool active stars in all four stokes parameters .
here we report the detection of magnetically - induced linear polarization for the primary component of the very active rs cvn binary hr1099 and for the moderately active k dwarf @xmath0 eri . for both stars
the amplitude of linear polarization signatures is measured to be @xmath110@xmath2 of the unpolarized continuum , which is approximately a factor of ten lower than for circular polarization .
this is the first detection of the linear polarization in line profiles of cool active stars .
our observations of the inactive solar - like star @xmath3 cen a show neither circular nor linear polarization above the level of @xmath110@xmath4 , indicating the absence of a net longitudinal magnetic field stronger than 0.2 g. |
a long - standing problem in quantum field theory is a nonperturbative formulation of chiral fermions .
our only general nonperturbative formulation of quantum field theory is the renormalization - group limit of functional integrals defined on a lattice .
but when chiral symmetry is an issue , there are notorious problems @xcite .
briefly , one must either sacrifice locality or positivity , tolerate additional states ( doubling or mirror states ) , or break chiral symmetry explicitly . when the coupling of fermions to gauge fields is vector - like , the standard formulations @xcite are adequate , if imperfect . on the other hand ,
when fermions couple to chiral gauge fields , it has been difficult to prove a conceptually clean theory ; see ref .
@xcite for a review .
an encouraging proposal replaces the functional integral over fermions with an auxiliary quantum - mechanical system @xcite , inspired by domain - wall @xcite and lattice pauli - villars @xcite methods .
this paper offers constructions of vector - like and chiral gauge theories , coupling continuum fermions to lattice gauge fields by introducing an interpolation of the latter .
ideas of this type were first discussed by flume and wyler for the schwinger model @xcite , and recently t hooft advocated a similar approach for four - dimensional gauge theories @xcite .
the appeal stems from the nontrivial ( instanton ) topology of continuum gauge fields , because the atiyah - singer index theorem @xcite implies an intimate relation between chiral properties of fermions and the topology of the gauge field . in 19871988
there was some discussion about topology and fermions in lattice gauge theory @xcite . except for a conference
reports @xcite , however , none of the applications to chiral gauge theories have been published .
spurred by ref .
@xcite , i would like to present my variation on the theme .
@xcite regulates the gauge field with the lattice and the fermions with a standard pauli - villars scheme .
the number of fermionic degrees of freedom ( per unit volume ) remains infinite in the words of smit , the method is desperate @xcite .
in particular , a numerical evaluation of the effective action would require infinite computation , even at fixed cutoff .
this paper , on the other hand , examines a sharp cutoff on determinants , which was first studied in ref .
the number of fermionic degrees of freedom is now finite , so the numerical computation of the effective action is finite too .
another difference between this paper s proposal and the one in ref .
@xcite is the strategy for removing the cutoffs .
let @xmath0 denote the lattice spacing and @xmath1 the ultraviolet cutoff on fermions . in the formulation of ref .
@xcite one takes @xmath2 for @xmath0 fixed , and afterwards @xmath3 . as stressed in sect .
[ gth cutoff ] the cutoff in ref .
@xcite maintains the gauge symmetry of the chiral theory only in the @xmath2 limit . with the sharp cutoff formulated below , however , it is permissible and natural to take @xmath3 , @xmath2 with @xmath4 constant .
the latter approach is far superior in a numerical computation .
if @xmath1 must vary at fixed @xmath0 , it will be extremely difficult to generate useful ensembles of lattice gauge fields , because to obtain a renormalized theory the bare gauge coupling must depend on @xmath1 .
[ vector - like ] begins with a discussion of vector - like theories .
the analysis starts with anomalies , because they are such a stumbling block for lattice formulations @xcite .
the functional integral over fermion fields is defined in the manner of fujikawa @xcite , and a specific cutoff procedure is formulated and justified in sect . [ regulator ] .
these sections define an effective action for a continuous background gauge field .
[ link ] summarizes the essential features of an interpolation from the lattice field of parallel transporters to a connection .
[ metric ] derives a heuristic relation between the fermion measure of this proposal and the usual measure of lattice fermions . while inessential to the main line of argument , the derivation suggests a rationale for relating the cutoff on determinants to the lattice spacing .
[ mc ] briefly considers numerical aspects . as usual , the generalization to chiral gauge theories is not immediate , but sect .
[ chiral ] produces a satisfactory definition of the chiral effective action , including fermion nonconservation .
finally , sect .
[ conclusions ] remarks on some of the loose ends , and compares the status of this formulation with ref . @xcite .
the formal expression for the euclidean functional integral for fermions is @xmath5 let us assume a background gauge potential ( or connection ) @xmath6 . in the application to lattice field theory , this connection is a determined from the lattice gauge field , cf .
[ link ] . staying momentarily `` in the continuum , ''
the action is @xmath7 formal integration over the fermion fields yields the boltzmann factor @xmath8 the objective is to give a rigorous meaning to the measure @xmath9 , and/or to the determinant .
consider the eigenfunctions and eigenvalues of the dirac operator @xmath10 @xmath11 since @xmath10 is anti - hermitian , the @xmath12 are real .
the dirac operator transforms covariantly under the gauge group , so the @xmath12 are gauge invariant .
the dirac operator anti - commutes with @xmath13 , i.e. @xmath14 .
hence , if @xmath15 is an eigenfunction with eigenvalue @xmath12 , then @xmath16 is an eigenfunction with eigenvalue @xmath17 . as usual
, one imagines that the theory is defined on a compact space - time , and the infinite volume limit is taken at the end .
then the spectrum of @xmath10 is discrete .
the dirac operator can possess zero modes , @xmath18 . in this subspace
it is convenient to sort the eigenfunctions according to chirality , i.e. the eigenvalue of @xmath13 .
let @xmath19 be the number of modes with @xmath20 and @xmath21 , and let @xmath22 .
the determinant should then be @xmath23 except that the infinite product still requires an ultraviolet regulator ; this is postponed to sect . [ regulator ] .
( the second equality follows because nonzero eigenvalues come in pairs @xmath24 . )
it is easy to see that the eigenvectors are orthonormal and form a complete set : @xmath25 @xmath26 where @xmath27 , @xmath28 denote spinor and color indices .
these properties of the eigenfunctions permit the expansions @xmath29 \bar{\psi}(x ) & { \displaystyle \sum_n } \bar{a}_n \varphi^\dagger_n(x ) , \end{array}\ ] ] where the coefficients @xmath30 and @xmath31 are grassman numbers . to obtain eq .
( [ unregulated ] ) from eqs .
( [ basic integral ] ) and ( [ vlike action ] ) one takes the functional integral over fields given by eqs .
( [ eigenbasis ] ) , i.e. the fermion measure is _ defined _ to be @xmath32 fujikawa @xcite makes a formal argument pursues a similar , yet complementary , line of thought . ] to relate the right - hand side of eq .
( [ measure ] ) to @xmath33 .
since the product over a continuous index is formal , it is logically cleaner to assert eq .
( [ measure ] ) as a definition .
the fujikawa measure is analogous to one based on fourier modes , @xmath34 the momenta correspond to eigenvalues of @xmath35 , and are also discrete ( in a box ) . with a coupling to gauge fields , however , the momenta are not gauge invariant .
hence , the definition based on eigenfunctions of @xmath10 is preferable .
one must check that the formalism reproduces the axial anomaly .
consider space - time dependent chiral transformations @xmath36 \bar{\psi}(x ) & \bar{\psi}'(x ) e^{-\alpha^a(x)t^a\gamma_5 } , \end{array}\ ] ] and @xmath37 is a generator of a global symmetry group .
( for u(1 ) take @xmath38 . ) under such transformations the measure of eq .
( [ measure ] ) is not invariant , but @xmath39 and similarly for @xmath40 , with sums on @xmath41 implied
. the rules of berezin integration imply @xmath42 the jacobian determinants are then responsible for the anomaly . using an eigenfunction expansion of @xmath43 one can write @xmath44 with @xmath45 then @xmath46 where the anomaly @xmath47 collecting the anomaly from @xmath48 and @xmath49 , as well as terms from the chiral transformation of the action , yields the ( anomalous ) ward - takahashi identity @xmath50
the preceding discussion skirts the need to regulate the determinants in the ultraviolet .
let us start with a kind of pauli - villars regulator .
( [ anomaly ] ) becomes @xmath51 let the index @xmath52 run over @xmath53 with the convention that @xmath54 denotes nonzero modes ; let us order the nonzero modes by @xmath55 and take @xmath52 odd ( even ) if @xmath12 is negative ( positive ) .
the regulating function @xmath56 is chosen to look like the sketch in fig .
[ regulator - function ] .
= the cutoff @xmath57 and smearing parameter @xmath58 should satisfy @xmath59 @xmath60 where @xmath61 is even ( so @xmath62 ) , and one takes @xmath63 with @xmath57 fixed .
the fermi function @xmath64 changes rapidly from 1 to 0 as @xmath65 passes through 1 .
since @xmath66 and @xmath67 and all its derivatives vanish at @xmath68 , it supports the standard analysis @xcite that leads to @xmath69\ ] ] in four dimensions , where the two traces are over global and gauge - group indices .
details in appendix [ convergence ] show that all higher dimension terms , e.g. @xmath70 , are proportional to @xmath71 as @xmath63 . to check that the regulator does not spoil the derivation of eq .
( [ awti ] ) , one restricts the chiral transformations @xmath72 to those for which @xmath73 if one of @xmath74 . if both @xmath52 and @xmath75 , then the left - hand side of eq .
( [ block c ] ) need not vanish .
the jacobian matrix @xmath76 takes a block form , and the regulator decouples the block with large eigenvalues . nevertheless , the transformation function @xmath72 is sufficiently arbitrary to derive the ward - takahashi identity .
the @xmath63 limit corresponds to keeping the first @xmath61 modes of @xmath10 completely and eliminating the rest . in a loose analogy with the lattice cutoff , in which the fourier measure of eq .
( [ fourier measure ] ) is truncated , one can carry out the truncation of the modes throughout .
thus , the regulated measure is @xmath77 with eigenvalues ordered as above .
the functional integral is now over fields given by @xmath78 \bar{\psi}(x ) & { \displaystyle \sum_{n=1-n_0}^n } \bar{a}_n\varphi_n^\dagger(x ) .
\end{array}\ ] ] the effective boltzmann factor becomes @xmath79 similarly , the regulated fermion propagator is @xmath80 the finite - mode cutoff has been examined before @xcite , with emphasis on anomalies .
those papers did not notice the disappearance of higher - dimensional terms as the fermi function becomes infinitely sharp , a feature that is especially important in the chiral gauge theory , cf .
[ chiral ] . in perturbation theory
, one must be careful to impose the sharp cutoff with @xmath10 , rather than with loop momenta .
for example , the dependence of the regulator on @xmath10 , and hence on @xmath6 , induces additional terms in the gauge current , to maintain gauge invariance .
how should one choose @xmath61 ?
one ought to take nonzero modes in pairs : it would be silly to take @xmath12 and not @xmath17 .
one also ought to retain all zero modes , because they are the most infrared of all ! but from the atiyah - singer index theorem @xcite the number of zero modes is even ( odd ) if the topological charge @xmath81 is even ( odd ) .
thus , @xmath61 must depend on the gauge field and be even ( odd ) if @xmath81 is even ( odd )
. the order of magnitude of @xmath61 can be specified only vaguely .
an examination of cutoff effects , appendices [ convergence ] and [ cutoff ] , introduces expansions . for a generic cutoff , one would need to maintain @xmath82 .
the absence of non - universal terms with the finite - mode cutoff , however , permits a lower cutoff .
if , as in the next section , @xmath6 is obtained from a lattice gauge field , @xmath83 , where @xmath76 is a gauge - group dependent constant . with the finite - mode cutoff
it is thus natural to take @xmath84 .
indeed , on an @xmath85 lattice one expects @xmath86 fermionic degrees of freedom in the lattice fermion field ( per flavor ) , where @xmath87 is the dimension of the fermion s gauge - group representation .
an obvious choice would be to augment the usual number with the zero modes , i.e. @xmath88 .
to extend the formulation to lattice gauge theory , one must provide an interpolation determining a connection @xmath89 from the lattice gauge field @xmath90 .
@xcite gives a prescription adequate for defining traces of products of @xmath91 , but does not prescribe @xmath6 .
other possibilities , based on definitions of the topological ( instanton ) charge of a lattice gauge field @xcite , do prescribe @xmath6 explicitly @xcite .
the latter have two crucial properties : 1 .
the only singularities in @xmath6 are instanton - like
the interpolation @xmath6 transforms as a connection under _ lattice _ gauge transformations . ref .
@xcite does prove a useful theorem on the spectrum of the dirac operator for a bounded gauge field .
the ( manifestly ) topological interpolations obey the hypothesis of the theorem . the great insight of ref .
@xcite was to recognize that the lattice gauge field can be used to define a fiber bundle .
then the topological charge is the second chern number of the bundle .
lscher originally provided expressions for transition functions , which encode changes of gauge @xmath92 from a patch @xmath93 of space - time to a neighboring one @xmath94 . for consistency , @xmath95 . in the fiber - bundle formalism , the gauge is fixed separately within each patch , such that @xmath96 has no singularities . the winding responsible for topological charge
then resides in the transition functions @xcite .
the more familiar patch - independent connection is @xmath97 where the section @xmath98 is related to the transition functions by @xmath99 the first step of refs .
@xcite is a patch - wise continuous , bounded interpolation for the nonsingular @xmath96 .
( [ connection ] ) produces a continuous , bounded connection @xmath6 , but when the transition functions have nontrivial winding , the sections , and thus @xmath6 , have directional singularities . by construction , therefore , the _ only _ singularities in the globally defined @xmath6 are those induced by the instanton - winding of the section . the other crucial property of the reconstructed gauge potentials of refs .
@xcite is the response to a lattice gauge transformation .
the transformation law of the lattice gauge field is @xmath100 where @xmath101 , @xmath102 denote lattice sites . for @xmath103 in the patch @xmath104 associated with @xmath101 ,
the reconstructed section obeys @xcite @xmath105 the interpolation @xmath106 is independent , eq .
( [ v = ww ] ) requires that the interpolation obey @xmath107 .
] of @xmath104 , but , of course , @xmath108 is independent of @xmath103 .
the interpolated connection transforms as @xmath109 with the _ same _ function @xmath106 .
if @xmath110 at each site , the three interpolated gauge transformation fields obey the composition law @xmath111 note that the interpolation depends on the underlying lattice gauge field , which is emphasized here by the second argument .
consequently , every interpolated @xmath112 can be built up from infinitesimal , site - by - site steps .
for the present discussion the complicated expressions for the interpolated @xmath106 and @xmath6 are not illuminating .
interested readers can consult refs .
@xcite for lscher s bundle .
( for phillips and stone s bundle , analogous results can be obtained @xcite . ) our prescription for coupling fermions to lattice gauge fields is to start with the lattice gauge field @xmath113 , interpolate to obtain the connection @xmath6 , use the associated dirac operator to define the measure , and regulate the determinant using the sharp cutoff . because the dirac operator is constructed using fiber bundles
, it automatically satisfies the atiyah - singer index theorem @xmath114 where @xmath81 is the topological charge ( as defined by ref .
@xcite ) of the lattice gauge field . moreover , from eq .
( [ transform a ] ) the eigenvalues of the dirac operator , and hence the effective action , are invariant under lattice gauge transformations . from eq .
( [ regulated effective action ] ) the boltzmann factor @xmath115 is also positive and finite .
it is intriguing to contrast the regulated measure of eq .
( [ regulated measure ] ) with the usual one @xmath116 where @xmath117 and @xmath118 denote the lattice fermion field and its conjugate at site @xmath101 .
there is an interpolation of the fermion field analogous to @xmath6 . in particular , under a lattice gauge transformation the interpolated fermion field transforms as @xmath119 with the same interpolated @xmath106 mentioned in sect .
[ link ] .
let us expand the interpolated field @xmath120 as in eq .
( [ eigenbasis ] ) , use the orthonormality to solve for the @xmath30 and take differentials . then @xmath121 d\bar{a}_n & { \displaystyle\sum_s\int}d^4x\,d\bar{\psi}(s ) \dfrac{\partial\bar{\psi}(x)}{\partial\bar{\psi}(s ) } \varphi_n(x ) = : { \displaystyle\sum_s } d\bar{\psi}(s)\bar{m}_{sn}. \end{array}\ ] ] if the number of modes @xmath61 equals the number of lattice sites , @xmath41 and @xmath122 are @xmath123 matrices and one can write @xmath124 the determinants here are akin to ones appearing in ref .
formally , one can combine the determinants into a `` metric '' @xmath125 and rewrite the measure as @xmath126 the metric depends covariantly on the gauge field , by construction of the interpolation .
one can develop a geometric picture by imagining a grassman line element @xmath127 that is gauge invariant . in this language
the customary lattice fields are curvilinear coordinates , and the expansion coefficients are rectilinear . according to this picture
, the usual lattice measure mistakenly neglects the curvature .
note that without the gauge interaction , the metric becomes flat : the two bases are then fourier transforms of one another .
the metric also becomes trivial in the naive continuum limit .
then the interpolation is unnecessary and @xmath128 .
the factor @xmath129 is essential , however , for obtaining the correct anomaly , index theorem , etc . in particular
, the anomaly arises because one may remove the regulators only after calculating with the functional integral , not ( as in the naive continuum limit ) before . of course , eq .
( [ metric measure ] ) is merely heuristic . to make the manipulations rigorous
, the interpolated fields must be smooth enough that the dirac eigenmode expansions leading to eq .
( [ a - psi ] ) stop at the @xmath61th term .
here @xmath61 is both the number of modes kept and the number of lattice sites .
this condition presumably puts constraints on the smoothness of the lattice gauge field similar to , but perhaps more stringent than , those imposed by uniqueness considerations in the fiber bundle constructions @xcite .
one should emphasize , however , that such constraints are on the connection between the present measure and the usual lattice one .
they do not detract from eq .
( [ measure ] ) as a definition of the functional integral , or @xmath115 as a definition of the effective action .
since the effective action @xmath130 is real and positive , the functional integral over the gauge bosons is amenable to the monte carlo method with importance sampling .
the dirac eigenvalues , and hence the effective action , depend on the lattice gauge field @xmath131 so they must be recomputed for every change of @xmath131 .
if the effective action is defined with a gaussian @xmath132 @xcite or standard pauli - villars @xcite regulator , one would need to compute all eigenvalues of @xmath10 and weight them appropriately an infinite computation . with eq .
( [ regulated effective action ] ) , however , only the lowest @xmath61 eigenvalues are needed a finite computation . times more costly , where @xmath133 characterizes machine precision . ] to compute the eigenvalues one can introduce an auxiliary lattice much finer than the original one .
parallel transporters for the fine lattice are constructed from the interpolated gauge field .
now consider any discretization of the dirac operator , and denote its eigenvalues by @xmath134 .
the discretization and auxiliary lattice spacing must be chosen such that @xmath135 ( up to tolerable floating - point precision ) for @xmath136 . on a fine enough auxiliary lattice , `` any '' discretization becomes precise enough . to avoid problems sorting the eigenvalues , however , a discretized operator without a doubled spectrum is preferable .
suitable examples would be the wilson discretization @xcite , or one derived by gauging the fixed - point action of a free fermion @xcite .
the discretization may break chiral symmetry , provided the breaking is numerically significant only for modes above the cutoff @xmath61 .
similar remarks apply to the construction for chiral gauge theories , sect .
[ chiral ] .
an important difference is that the boltzmann factor @xmath115 can be complex ( for fermions in `` complex '' representations ) , and in any case not positive definite .
monte carlo integration is then much more difficult , because fluctuations in sign reduce the effectiveness of importance sampling , cf .
[ how to ] .
nevertheless , @xmath115 can be evaluated with finite computation .
the preceding sections provide a definition of the functional integral for vector - like fermions .
it is nonperturbative , has the correct axial anomaly , and by construction provides a natural association between a dirac operator and a topological charge , so that the index theorem is obeyed .
the ideas are now extended to chiral fermions , which appear in the standard model and in grand unified theories .
the essential feature of chiral gauge theories is that positive and negative chirality fermions transform under different representations of the gauge group . in a basis of the dirac matrices with @xmath13 diagonal
it is useful to split the four - component dirac spinor into two two - component weyl spinors . without loss of generality one can charge - conjugate the negative chirality part and
assemble everything into one positive chirality field .
this weyl spinor is henceforth denoted @xmath137 , and the representation of the gauge group under which it transforms is denoted @xmath138 , with generators @xmath139 .
the kinetic term of the action is @xmath140 as before one wants to define the boltzmann factor @xmath141 but @xmath142 maps positive chirality weyl spinors into negative chirality weyl spinors .
the underlying difficulty in constructing a chiral gauge theory is that the right - hand side of eq .
( [ chiral gamma ] ) is _ not _ a ( functional ) determinant .
a well - formulated chiral gauge theory should also exhibit fermion nonconservation @xcite . if the vector - like operator @xmath10 has zero modes , the right - hand side of eq .
( [ chiral gamma ] ) should vanish : recall that the zero modes have definite chirality .
thus the zero modes possess natural projections from dirac spinors @xmath143 onto weyl spinors @xmath144 . for positive chirality
the projection implies @xmath145 , and for negative chirality @xmath146 ( integration by parts implied ) . in eq .
( [ chiral gamma ] ) integration over the @xmath147 component of @xmath137 or over the @xmath148 component of @xmath149 yields zero . on the other hand , if there are @xmath19 zero modes of each chirality ( counting individual species appropriately ) the integral @xmath150 does not vanish if @xmath151 .
( the notation @xmath152 is schematic for a product of @xmath153 suitable components or positions of @xmath137 . )
amplitudes for fermion nonconservation are proportional to integrals like @xmath154 . before discussing how to regulate the integrals in eqs .
( [ chiral gamma ] ) and ( [ b violation ] ) , one should list the properties of eq .
( [ chiral gamma ] ) that the regulator should respect .
in addition to the connection between zero modes and fermion nonconservation , one wants 1 .
[ invariant ] @xmath155 , under all circumstances .
[ anomaly free ? ] @xmath156 , only if @xmath138 is `` anomaly - free . ''
[ complex ] @xmath157 , if @xmath138 is complex ; such that @xmath158 , then @xmath138 is real ; otherwise it is complex .
the representation generated by @xmath159 is denoted @xmath160 .
if @xmath138 is real , then @xmath161 is real . ]
indeed @xmath162 .
here @xmath163 is a gauge transform of @xmath6 , and the notation @xmath164 stresses the fermion representation .
condition [ anomaly free ? ] is imposed so that the nonperturbative definition reproduces perturbation theory .
condition [ complex ] identifies a diagnostic feature of the effective action of chiral fermions .
even though the right - hand side of eq .
( [ chiral gamma ] ) is by nature not a determinant , ultraviolet regulators almost always turn it into one . for example , in the most naive ( and unsuccessful ) lattice formulation , @xmath142 becomes a large , square , numerical matrix , and @xmath165 is the matrix determinant .
so let us provisionally assume @xmath166 in anticipation of a cutoff analogous to the one for vector - like theories ( sect .
[ regulator ] ) , the product runs only over the first @xmath61 eigenvalues . the most transparent way to realize conditions [ invariant ] and [ anomaly free ?
] is as follows : write @xmath167 , with @xmath168 and @xmath169 real , and note that @xmath170 if the moduli @xmath171 of the eigenvalues are gauge invariant and one orders the eigenvalues by @xmath172 , condition [ invariant ] is satisfied . by condition [ anomaly free ? ] , in anomalous ( sub)representations the phases of the eigenvalues would be gauge variant , but the variation from one species could cancel that of another .
unfortunately , it seems that eigenvalue problems with these simple gauge - transformation properties leave the total phase @xmath173 unspecified , flouting condition [ complex ] .
is related to the vector - like theory and @xmath174 to the spectral asymmetry of a certain operator @xcite .
this method , however , requires three regulators : one for the gauge fields , one for the vector - like fermions , and one for @xmath175 . ] a standard way to cast the effective action as a determinant is to introduce a new negative chirality partner @xmath176 with no dynamics .
the action is now @xmath177 where the four - component spinor @xmath178 and @xmath179 the matrix forms of eqs .
( [ dirac spinor ] ) and ( [ d hat ] ) presumes a @xmath180-matrix basis with @xmath13 diagonal .
formally , the functional integral over @xmath181 is trivial , but the combined integral @xmath182 can be expressed as a determinant , as in the vector - like theory .
the operator @xmath183 is not self - adjoint , but it is elliptic @xcite , so on a compact space - time it still has a discrete spectrum .
the eigenvalue problem now has different right and left eigenfunctions .
there are zero modes @xmath184 where @xmath185 are zero - mode eigenfunctions of the vector - like operator @xmath10 , with chirality @xmath186 .
nonzero modes come in pairs .
this operator is again not self - adjoint , though elliptic , so there are left and right eigenfunctions @xmath187 .
most papers either ignore this subtlety , or try to circumvent it .
for example , some tricks turn @xmath183 into a hermitian operator , thus leaving @xmath173 unspecified .
they can not be adopted here . ]
@xmath188 ( i\hat{d})^\dagger\chi_n & \lambda_n^ * \chi_n \end{array}\ ] ] that are mutually orthonormal : @xmath189 in addition @xmath190 .
furthermore , since @xmath191 , if @xmath192 has eigenvalue @xmath12 , then @xmath193 has eigenvalue @xmath17 , and similarly for @xmath194 , @xmath195 .
the functional integral is now defined to be over fields @xmath196 \bar{\psi}(x ) & { \displaystyle \sum_{n=1}^{n_-}}\bar{z}_n \varphi_{-,n}(x ) & { \displaystyle \sum_{n=1}^n } \bar{a}_n
\chi^\dagger_n(x ) , \end{array}\ ] ] i.e. the measure is @xmath197 as in sect .
[ regulator ] the functional integral is cut off by retaining only the lowest @xmath61 nonzero modes .
the principle for ordering the eigenvalues is revealed below , but clearly the two modes with the same @xmath55 should be adjacent .
analogously to sect .
[ regulator ] , it is convenient to take @xmath52 odd ( even ) if @xmath198 is negative ( positive ) . let us first integrate over the @xmath30 and @xmath31 .
( [ chiral fields ] ) and ( [ chiral measure ] ) yield @xmath199 for the functional integral . in a real representation
these eigenvalues come in pairs @xmath200 ; @xmath201 satisfies condition [ complex ] .
but one should not expect @xmath202 to be a suitable definition of the effective action hence the tilde because neither the moduli nor the phases of the nonzero eigenvalues are gauge invariant . under the gauge transformation
@xmath203 the chiral dirac operator transforms as @xmath204 , where @xmath205 .
the zero modes are gauge invariant . to first order in @xmath206
the nonzero eigenvalues vary by @xmath207 which immediately yields the variation of @xmath201 : @xmath208 where @xmath209 the last expression applies if the eigenvalues are ordered by increasing @xmath210 , which is justified because it reproduces the consistent gauge anomaly . from eq .
( [ chiral anomaly ] ) and appendix [ convergence ] , the imaginary part of @xmath211 is @xmath212,\ ] ] the familiar consistent anomaly @xcite .
it vanishes if @xmath213 in representation @xmath138 @xcite ; @xmath201 satisfies condition [ anomaly free ? ] .
even if the anomaly cancels , however , the real part @xmath211 is not gauge invariant . with @xmath214 and @xmath215 from appendix [ convergence ] @xmath216 after taking @xmath63 , all higher - dimension terms drop out because they are proportional to @xmath217 .
following bardeen , the gauge variation of the real part can be compensated by counter - terms @xcite .
let @xmath218 s_4 & { \displaystyle \frac{1}{48\pi^2}\int}d^4x\,\tr_\rho\left [ { \mbox{\small $ \frac{1}{2}$}}a_\mu\partial^2a_\mu + ( \partial\cdot a)^2 + { \mbox{\small $ \frac{1}{4}$}}a_\mu a_\nu a_\mu a_\nu - { \mbox{\small $ \frac{1}{2}$}}(a^2)^2 \right ] , \end{array}\ ] ] and @xmath219 .
if @xmath220 is computed using the _ interpolated _ gauge field , its gauge variation cancels that of @xmath221 exactly .
the choice of @xmath222 is not unique , because one could also add a gauge invariant term proportional to @xmath223 .
but the ambiguity corresponds to a shift in the ( inverse ) bare gauge coupling , so it should make no difference once all cutoffs are removed . the counter - term @xmath224 is supposed to remove a quadratic divergence , but the number of modes @xmath61not the mass @xmath57 that appears in eq .
( [ counter - terms])defines the cutoff .
the number of fermion modes with momentum below a cutoff @xmath1 is @xmath225 approximating the sum by an integral yields @xmath226 .
but @xmath61 is an integer , so the error in this approximation can be eliminated by taking @xmath227 this value of @xmath57 is the one needed to cancel the quadratic divergence . if one chooses @xmath228 , then @xmath229 .
the combination @xmath230 is thus gauge invariant under infinitesimal gauge transformations . by eq .
( [ composition ] ) this is enough to show that @xmath130 is invariant under all lattice gauge transformations .
hence , @xmath130 satisfies all three conditions .
this is the main result .
finally , let us integrate over the zero modes . unless there are enough factors of the fermion field in the amplitude , the integral vanishes . with the minimal number of fields in eq .
( [ b violation ] ) @xmath231 where @xmath232 , @xmath233 and @xmath234 , @xmath235 denote discrete indices and positions of the fields in eq .
( [ b violation ] ) .
functional integration alone would lead to eq .
( [ amplitudes ] ) with @xmath201 instead of @xmath236 .
the zero modes present no substantive changes in the computation of the gauge variation of @xmath201 , so the same counter - terms restore gauge symmetry . although ref .
@xcite focuses primarily on the vector - like theory with pauli - villars cutoff , it does prove its important convergence theorem for theories with vector and axial - vector couplings .
this suggests that the chiral coupling is also intended as an application . with
the cutoff proposed there , the analysis of cutoff effects leads to somewhat different conclusions .
appendix [ pv ] recasts t hooft s cutoff in a way that makes appendix [ convergence ] directly applicable .
one finds that the quadratic divergence drops out , but the universal term @xmath215 still spoils gauge invariance of the real part .
moreover , the coefficients of the higher - dimension terms , @xmath237 and so forth , no longer vanish , though they are suppressed by powers of @xmath238 . to eliminate these violations of gauge symmetry
, one must take @xmath1 to infinity , on each gauge field separately .
this is cumbersome , and perhaps logically inconsistent , because to obtain a renormalized theory , the bare gauge coupling must be adjusted to keep physical masses properties of the ensemble average cutoff independent . to summarize the results of this section
it is worth sketching how to compute correlation functions . because the integrals @xmath239 sometimes vanish trivially ,
let us denote the sector of lattice gauge fields with @xmath19 zero modes @xmath240 .
nonzero integrals with then be over one or so sectors .
consider first fermion - conserving observables .
one would like to compute a ratio of the form @xmath241 where @xmath242 is the lattice - gauge - field action . here
the gauge - invariant observable @xmath243 is constructed from the gauge field and fermion propagators .
( [ observable ] ) one requires an ensemble of fields in @xmath244 , distributed with weight @xmath245 for amplitudes of this type , all other sectors carry weight 0 , because of the zero modes , so in the monte carlo they are simply omitted .
consider next a fermion - violating amplitude .
for simplicity , suppose that @xmath243 does not contain fermion species that are being created or annihilated .
now one would like to compute a ratio of the form @xmath246 the numerator and denominator are averages over different sectors , is typically the sector with topological charge @xmath247 , which would have one zero mode for each species . ] and recall that in sectors with zero modes @xmath248 is defined to be the product of the first @xmath61 nonzero eigenvalues , ordered by @xmath210 .
one can re - write eq .
( [ violation ] ) as @xmath249 where the average @xmath250 is over @xmath251 .
in addition to the weight @xmath252 , one must compute the zero - mode eigenfunctions @xmath253 for created and annihilated species .
the other factor in the fermion - nonconserving amplitude is the ratio of two partition functions . to compute it accurately some special numerical techniques @xcite are available , which keep track of the system s preference for one sector or the other .
various versions of these `` histogram methods '' @xcite may also prove useful in obtaining the nontrivial phase of @xmath252 , inherent to a complex representation .
some theories , the simplest of which is su(2 ) with one weyl doublet , are afflicted by a global anomaly @xcite .
the representations in question are real , and therefore @xmath115 , as defined by eqs .
( [ gamma tilde ] ) and ( [ main result ] ) , is real and positive .
thus @xmath130 is real .
let us focus on su(2 ) . because @xmath254 , there are nontrivial gauge transformations , for which the proof of gauge invariance of @xmath130 breaks down . )
shows that the interpolations inherit this property , the two classes can be separated by lattice gauge transformations for which the interpolation is ill - defined .
] let @xmath255 be in the nontrivial class .
the variation @xmath256 is real , but does it vanish ? following ref .
@xcite one can compute the difference by embedding su(2 ) into su(3 ) and taking a trajectory from @xmath257 to @xmath258 in su(3 ) .
@xmath259 does not change for any infinitesimal step , and since @xmath260 , the trajectory can be constructed from infinitesimal steps .
thus @xmath261 for the embedded field , and hence likewise for the su(2 ) fields themselves . on the other hand
, witten argued that the two configurations @xmath131 and @xmath262 should have boltzmann weights equal in magnitude but _ opposite _ in sign @xcite . indeed the gauge variation of @xmath263 integrated along the su(3 ) trajectory supports his conclusion @xcite .
but , given a lattice gauge field @xmath264 , the algorithm for @xmath265 can not determine whether @xmath266 or @xmath267 . and thanks to the pains taken ensure gauge invariance , the computed weight is the same in either case .
this is not a serious drawback , however .
if the global anomaly applies , one can replace numerator and denominator of eq .
( [ observable ] ) by witten s original result , @xmath268 , eliminating all computation .
the seeming incompatibility of lattice fermions and chiral symmetry has inspired the recurring idea @xcite of treating the fermions in the continuum , even if the underlying gauge field is on the lattice .
smit calls the idea desperate @xcite .
how desperate are the main results presented here ?
the colorful terminology refers to the tacit assumption that a continuum requires an infinite number of degrees of freedom ( per unit volume ) .
then the arithmetic needed to evaluate functional integrals is infinite : we are desperate because a computer can not do the job . but with a finite - mode cutoff there are @xmath61 eigenvalues ; the computation is finite .
alas , even in the vector - like theory the effort needed to obtain higher eigenvalues will be high ( cf .
the construction for chiral fermions is yet more computationally intensive , because , first , the counter - terms must now be computed accurately , and , second , the phase , which stems not from the regulator but from the chiral coupling itself , requires extra care .
there is a potential shortcoming to the finite - mode regulator .
one would like to verify that the spectrum remains chiral , in perturbation theory and beyond .
appendix [ cutoff ] demonstrates perturbative universality for fermion loops .
@xcite and appendix [ convergence ] do the same for anomalous diagrams only . )
but there is ( as yet ) no comparable proof for diagrams with external fermion lines , which are needed to examine the fermion spectrum .
the resolution is not straightforward , because it depends on details of the gauge - boson propagator , i.e. on the lattice gauge action . and
even if the perturbative test is a success , one should be cautious until the nonperturbative spectrum has been checked .
the counter - terms required in the chiral gauge theory may be unsettling at first sight .
but their necessity arises from the unassailable observation @xcite that a _ regulated _ functional integral either respects a symmetry or it does not . to obtain the anomaly ( and hence the physically powerful requirement of anomaly cancellation ) the functional integral @xmath269 can not be gauge symmetric .
consequently , the final result for the chiral boltzmann factor @xmath270 is the product of a functional integral and a symmetry - restoring factor .
a related peculiarity is the fate of witten s global anomaly @xcite .
again , an ansatz for the effective action is either gauge invariant or not . the effective action @xmath130 _ is _ gauge invariant , even when the gauge transformation is in the nontrivial class . without asserting that this paper s formulation succeeds at defining the globally anomalous theories , one might suggest that it could shed light on the dynamical puzzles that originally motivated ref .
an optimistic possibility is that the dynamics of @xmath130 fail in globally anomalous theories , but not otherwise .
a serious complication is that the formulation is in euclidean field theory . at the nonperturbative level
the wick rotation does no good , and instead one constructs the minkowski theory via the imaginary - time evolution operator on the hilbert space of states .
this procedure defines a hamiltonian that can then be used to propagate the states in real time .
the eigenfunctions used to define the functional integral are fundamentally four - dimensional , so the constructive approach @xcite does not seem helpful .
perhaps the axiomatic approach @xcite will prove more promising .
finally , let us compare the present chiral construction with the overlap formalism @xcite .
both define an effective action with a gauge - invariant real part .
both generate an imaginary part in a complex fermion representation , but not in a real representation . with the interpolation and
@xmath271 the gauge variation of the imaginary part is the consistent anomaly and with the finite - mode regulator ( or in the limit @xmath2 @xcite ) nothing else . with the overlap
the gauge variation of the imaginary part contains the anomaly and higher - order gauge breaking terms , analogous to @xmath272 ( notation of appendix [ convergence ] ) , as well .
@xcite argues they should be tolerably small , and tests in two dimensions @xcite indicate that this conclusion may be correct .
both methods provide fermion nonconserving amplitudes : here the gauge - field topology @xcite drives fermion nonconservation , and in ref .
@xcite the fermion nonconserving amplitudes define the gauge - field topology .
( for smooth lattice gauge fields the two topologies coincide . )
whereas appendix [ cutoff ] includes an explicit verification of fermion - loop coupling - constant renormalization , it does not seem that an explicit calculation starting from the lattice overlap is available yet @xcite .
on the other hand , ref .
@xcite includes several numerical cross checks that have not been done here .
an important , dynamical test is whether the fermion spectrum remains chiral after integrating over gauge fields .
neither construction has been subjected to this test yet , because it requires a full - fledged monte carlo calculation .
sects . [ vector - like][metric ] are based on an unpublished draft ( 1988 ) .
that work was aided by discussions with m. gckeler , g. schierholz , and u .- j .
i also thank h. joos for discussions on this topic then and since . while formulating sect .
[ chiral ] , i had helpful conversations with w.a .
bardeen , m.f.l .
golterman , and t. onogi .
fermilab is operated by universities research association , inc .
, under contract de - ac02 - 76ch03000 with the u.s .
department of energy . while this paper was being written , a preprint appeared by s.d.h .
hsu ( yale university report yctp - p5 - 95 , hep - th/9503058 ) , expanding on the idea of gckeler and schierholz @xcite to exploit the @xmath175-invariant formulated in ref .
this appendix fills in the steps from eq .
( [ regulated anomaly ] ) to eq .
( [ fdualf ] ) .
the modifications needed to obtain eq .
( [ consistent anomaly ] ) are provided in sect .
[ chiral convergence ] .
the analysis is standard , but it is provided to demonstrate the form of higher - dimension terms suppressed by powers of @xmath273 .
the sharp limit @xmath63 turns out to be special , because it eliminates these terms _ before _ taking @xmath274 .
first some preliminaries on fourier transforms . for definiteness
the space - time is a box with sides @xmath275 , and volume @xmath276 .
the fourier transform @xmath277 allowed values of @xmath278 depend on the boundary conditions .
this appendix presents details for a wide class of almost periodic boundary conditions , such that @xmath279 , @xmath280 .
strictly periodic directions have @xmath281 ; anti - periodic directions have @xmath282 ; @xmath76-periodic directions have a certain combination of the foregoing and @xmath283 .
the inverse transform ( assuming convergence ) is @xmath284 the application needed here is the fourier transform of eq .
( [ complete ] ) @xmath285 consider any function @xmath286 obeying @xmath287 and @xmath288 @xcite , and let @xmath289_{ij},\ ] ] with the summation convention over the spin - flavor - color multi - indices @xmath290 . in sect .
( [ regulator ] ) the fermi function @xmath291 appears , with mass @xmath57 and the limit @xmath63 . for comparison with refs .
@xcite , however , it is convenient to keep @xmath286 arbitrary .
fourier transforming the eigenfunctions and using eq .
( [ k complete ] ) one obtains @xmath292_{ij}.\ ] ] from eq .
( [ detc ] ) one recalls that here @xmath293 . hence , @xmath294_{ij},\ ] ] expanding in @xmath295 @xmath296 where now @xmath297 , and @xmath298 .
because the functions under consideration are smooth and vanish rapidly at infinity , the sums can be approximated by integrals @xmath299 with little error .
odd powers vanish .
note that the finite - size effects also vanish when the ultraviolet regulator is removed . reorganizing the terms according to the power of @xmath1 @xmath300 + & \left .
m^{d-4 } f^{(2-d/2)}(0)\,\alpha_4^a(x ) + m^{d-6 } f^{(3-d/2)}(0)\,\alpha_6^a(x ) + \cdots\right ] , \end{array}\ ] ] where @xmath301 , @xmath302,\ ] ] @xmath303 + { \mbox{\small $ \frac{1}{6}$}}[d_\mu , d_\nu][d_\mu , d_\nu ] \right)\right\},\ ] ] and @xmath304 contains terms with 6 @xmath305 s , combined to produce a function ( rather than a differential operator ) .
the result depends on the cutoff function via the coefficients @xmath306 , defined by @xmath307 , @xmath308 the @xmath1 independent term has the universal coefficient @xmath287 ; the other coefficients differ for different cutoffs . with the fermi function @xmath309 , @xmath54 , plus terms of order @xmath310 .
hence , the @xmath311 , @xmath312 , are power - law divergences , unless the traces vanish . on the other hand , for @xmath54 then @xmath313 as @xmath314 .
thus , the sharp cutoff has no `` scaling violations . '' for the vector - like theory @xmath315f_{\mu\nu}$ ] . if @xmath316 , the only surviving term @xmath214 yields the well - known result . if @xmath317 , the dirac trace makes @xmath214 vanish as well as everything in @xmath318 except the term @xmath319 , yielding the familiar axial anomaly , eq .
( [ fdualf ] ) . in the chiral gauge theory one wants @xmath320_{ij}.\ ] ] the left and right eigenfunctions of @xmath183 are not complete , but @xmath321 where @xmath322 projects onto zero modes .
this nuisance is easiest to handle with anti - periodic boundary conditions .
if the function @xmath286 drops to zero below the lowest momentum mode , then @xmath323 .
thus , the projector can be dropped after fourier transforming .
the hole in @xmath286 can not affect the momentum sums , so the correct approximation in eq .
( [ integrals ] ) uses a function without the hole on the right - hand side . the algebraic manipulations still hold , but one must replace @xmath324 by @xmath139 , @xmath10 by @xmath271 _ and _ the symbol @xmath325 by @xmath326 . from eq .
( [ d hat ] ) @xmath327 then @xmath328 consequently , @xmath329 where the trace is now only over gauge indices . in @xmath316
this is the universal anomaly , plus a term that can be compensated by a local counter - term ; in @xmath317 it only the latter survives , and it is a power - law divergence .
the function in @xmath318 is too lengthy to present explicitly , but after the trace @xmath330 , where @xmath331\quad(d=4)\ ] ] is the well - known consistent anomaly @xcite , and @xmath332 - 2 [ \partial_\mu \partial{\!\cdot\!}a , a_\mu ] + \{\partial{\!\cdot\!}a , a^2\ } \rule{0.0em}{0.9em}\right.\right .
\\[1.0em ] & \left.\left.\rule{0.0em}{0.9em } - a_\mu(\partial{\!\cdot\!}a)a_\mu - [ \partial_\mu a_\nu , [ a_\mu , a_\nu ] ] + a_\mu(\partial_\mu a_\nu + \partial_\nu a_\mu)a_\nu \right)\right ] \end{array}\ ] ] is the almost as well - known quantity that , like @xmath333 , can be compensated by local counter - terms @xcite .
the analysis of the previous section can be applied directly to the effective action .
it shows that the finite - mode cutoff is in the same ( perturbative ) universality class as pauli - villars regulators . as in appendix
[ convergence ] , the analysis is performed for an arbitrary smooth function , but once again the sharp limit is special , because it has no power corrections . for a good infrared behavior ,
this section considers only anti - periodic boundary conditions .
one can write the regulated effective action as @xmath334 where the effective lagrangian @xmath335 and , for any function @xmath286 obeying @xmath287 and @xmath288 , @xmath336 and @xmath337 .
the notation applies to the vector - like theory , but the substitutions needed for the chiral gauge theory are obvious .
the effective lagrangian for arbitrary @xmath286 will be denoted @xmath338 .
fourier transforming the eigenfunctions , using eq .
( [ k complete ] ) , and using @xmath293 one obtains @xmath339_{ij},\ ] ] where @xmath340 . expanding in @xmath295 @xmath341 where @xmath342 .
the derivatives take the form @xmath343 the @xmath344 term can be called universal , because after summing over @xmath278 the universal @xmath287 remains . in the vector - like theory ,
the fermion mass ( presumed nonzero ) regulates the infrared and one can use eq .
( [ integrals ] ) . in the chiral gauge theory ( or any massless case ) , infrared singularities make the integrals poor estimates of the sums . for the present purposes ,
however , the function @xmath286 can also be used as an infrared regulator .
one simply chooses @xmath286 to drop to zero for momentum less than the smallest allowed by the anti - periodic boundary condition , say for @xmath345 .
the delicate infrared behavior has consequences when reorganizing the effective lagrangian according to the dimension of the interactions .
one finds @xmath346 + & \left .
m^{d-4 } l_f^{(2-d/2)}(\delta^2)\,l_4(x ) + m^{d-6 } l_f^{(3-d/2)}(\delta^2)\,l_6(x ) + \cdots\right]+\delta{{\cal l } } , \end{array}\ ] ] where @xmath347 yields a constant of no dynamical significance , @xmath348 @xmath349 + { \mbox{\small $ \frac{1}{6}$}}[d_\mu , d_\nu][d_\mu , d_\nu]\right\},\ ] ] and @xmath350 contains same combination of 6 @xmath305 s as @xmath304 , which is a function rather than a differential operator .
the interactions in @xmath351 are proportional to @xmath352^l$ ] ; hence they survive only when @xmath353 .
the effective lagrangian depends on the cutoff via the coefficients @xmath354 .
if @xmath355 ( i.e. @xmath356 ) these become @xmath357 . for example , with the fermi function @xmath358 plus terms of order @xmath359 .
the higher - dimension functions have coefficients @xmath360 , @xmath54 . from eq .
( [ lfn ] ) one sees that they consist of a universal , @xmath286- and @xmath1-independent term , plus terms non - universal proportional to @xmath361 .
the remaining dimensions are balanced by infrared scales @xmath41 or @xmath362 .
again , for the fermi function the non - universal terms are suppressed by @xmath363 and drop out in the sharp limit . since interactions of all dimension appear in eq .
( [ lagrangian expansion ] ) , it must be interpreted as a perturbative series . in the vector - like theory , regulated as in sect
. [ regulator ] , @xmath364 vanishes .
indeed , the gauge invariance of the _ regulated _ theory forbids any dimension 2 terms . in @xmath365 the dirac trace eliminates the nested commutator leaving , for @xmath317 , @xmath366 which renormalizes the gauge coupling . in the chiral gauge theory of sect .
[ chiral ] this analysis applies to @xmath201 .
the dimension - two term @xmath364 does not vanish , because now the regulator breaks the gauge symmetry . in four dimensions
one has @xmath367 to restore gauge symmetry , one must add the counter - term @xmath224 in eq .
( [ counter - terms ] ) . on the other hand
, @xmath365 induces coupling constant renormalization .
( the gauge non - invariant pieces cancel . ) with some patience one can accumulate the nonvanishing @xmath180-matrix traces in @xmath365 to obtain @xmath368 up to a total derivative .
notice that , as expected , the renormalization term is half that of the vector - like theory .
the dimension - four interactions in @xmath351 , present because @xmath369 , include gauge - breaking terms ; they are cancelled by @xmath222 .
@xcite defines the fermion functional integral with a time - honored @xcite pauli - villars regulator : @xmath370 in the vector - like case , and similarly @xmath371 with @xmath271 in the chiral case .
the determinants are products over the infinitely numerous eigenvalues .
the masses for vector - like gauge theories ; @xmath372 for chiral gauge theories . ]
@xmath373 and signatures @xmath374 satisfy @xmath375 @xmath376 for @xmath377 , eq .
( [ logs ] ) defines the overall scale @xmath1 .
( [ powers ] ) and ( [ logs ] ) require infinite series only if one requires the identities for all @xmath52 .
the function needed to describe the effective lagrangian , cf .
( [ gamma_f])([lagrangian expansion ] ) , is @xmath378 note that @xmath379 and all its derivatives vanish at infinity , by virtue of eqs .
( [ powers ] ) and ( [ logs ] ) , so the manipulations of appendix [ cutoff ] still hold .
( [ awti ] ) and ( [ gauge variation ] ) also hold as before , but with @xmath380 replaced by @xmath381 . here
@xmath382 again @xmath383 and all its derivatives vanish at infinity , so the manipulations of appendix [ convergence ] hold . to apply eq .
( [ general result ] ) one notes the universal normalization @xmath384 .
power - law divergences drop out , but `` scaling violations '' remain ( @xmath54 ) : @xmath385 and analogously for @xmath386 .
one might remark that the absence of power - law divergences relies on integrating @xmath103 to infinity .
they re - appear if one truncates @xmath387 when , say , @xmath388 becomes small .
karsten and j. smit , nucl .
* b183 * ( 1981 ) 103 .
nielsen and h. ninomiya , nucl . phys . *
b185 * ( 1981 ) 20 , ( e ) * b195 * ( 1982 ) 541 ; * b193 * ( 1981 ) 173 ; + d. friedan , commun . math
* 85 * ( 1982 ) 481 .
wilson , in _ new phenomena in subnuclear physics _ , edited by a. zichichi ( plenum , new york , 1977 ) .
l. susskind , phys .
rev . * d16 * ( 1977 ) 3031 ; + t. banks , j.b .
kogut , and l. susskind , phys . rev . * d13 * ( 1976 ) 1043 .
petcher , nucl .
b ( proc . suppl . ) * 30 * ( 1993 ) 50 .
r. narayanan and h. neuberger , phys .
lett . * 71 * ( 1993 ) 3251 ; nucl .
b412 * ( 1994 ) 574 ; institute for advanced study report iassns - hep-94/99 ( hep - th/9411108 ) .
kaplan , phys .
b288 * ( 1992 ) 342 .
frolov and a.a .
slavnov , phys .
b309 * ( 1993 ) 344 ; nucl .
b411 * ( 1994 ) 647 ; + r. narayanan and h. neuberger , phys . lett .
* b302 * ( 1993 ) 62 .
r. flume and d. wyler , phys .
108b * ( 1982 ) 317 . g. t hooft , utrecht university report thu-94/18 ( hep - th/9411228 ) . m. atiyah and i.m .
singer , ann . math . * 93 * ( 1971 ) 139 .
kronfeld , nucl .
b ( proc . suppl . )
* 4 * ( 1988 ) 329 .
j. smit , nucl .
b ( proc . suppl . ) * 4 * ( 1988 ) 451 . m. gckeler and g. schierholz , nucl .
b ( proc . suppl . )
* 29 * ( 1992 ) 114 ; * 30 * ( 1993 ) 609 .
a. andrianov and l. bonora , nucl . phys . *
b233 * ( 1984 ) 232 , 247 .
k. fujikawa , phys .
lett . * 42 * ( 1979 ) 1195 ; phys . rev . *
d21 * ( 1980 ) 2848 , ( e ) * d22 * ( 1980 ) 1499
. m. lscher , comm .
* 85 * ( 1982 ) 29 .
a. phillips and d. stone , comm .
math . phys .
* 103 * ( 1986 ) 599 .
m. gckeler , a.s .
kronfeld , g. schierholz , and u .- j .
wiese , nucl .
* b404 * ( 1993 ) 839 .
kronfeld , unpublished ( 1988 ) .
van baal , comm .
* 85 * ( 1982 ) 529 .
m. gckeler , a.s .
kronfeld , m.l .
laursen , g. schierholz , and u .- j .
wiese , nucl .
b292 * ( 1987 ) 349 .
wiese , phys .
* b315 * ( 1993 ) 417 . g. t hooft , phys .
* 37 * ( 1976 ) 8 ; phys . rev .
* d14 * ( 1976 ) 3432 , ( e ) * d18 * ( 1978 ) 2199 .
l. alvarez - gaum and s. della pietra , in _ recent developments in field theory and statistical mechanics _ , edited by j. ambjrn , b.j .
durhuus , and j.l petersen ( elsevier , amsterdam , 1985 ) ; + l. alvarez - gaum , s. della pietra , and v. della pietra , phys . lett . *
166b * ( 1986 ) 177 ; + s. della pietra , v. della pietra , and l. alvarez - gaum , commun .
* 109 * ( 1987 ) 691 .
l. alvarez - gaum and p. ginsparg , nucl
b243 * ( 1984 ) 449 .
balachandran , g. marmo , v.p .
nair , and c.g .
trahern , phys .
rev . * d25 * ( 1982 ) 2713 .
m.b . einhorn and d.r.t .
jones , phys .
* d29 * ( 1984 ) 331 . w.a .
bardeen , phys . rev .
* 184 * ( 1969 ) 1848 .
j. wess and b. zumino , phys . lett . * 37b * ( 1971 ) 95 .
gross and r. jackiw , phys . rev . * d6 * ( 1972 ) 477 .
a. borelli , l. maiani , g.c .
rossi , r. sisto , and m. testa , nucl .
* b333 * ( 1990 ) 335 .
alonso , ph .
boucaud , j.l .
corts , and e. rivas , mod .
a5 * ( 1990 ) 275 ; phys .
d44 * ( 1991 ) 3258 .
bodwin and e.v .
kovcs , nucl .
b ( proc . suppl . ) * 20 * ( 1991 ) 546 ; * 30 * ( 1993 ) 617 .
g. bhanot , s. black , p. carter , and r. salvador , phys . lett .
* b183 * ( 1987 ) 331 ; + g. bhanot , k. bitar , and r. salvador , phys . lett .
* b187 * ( 1987 ) 381 ; * b188 * ( 1988 ) 246 ; + g. bhanot , a. gocksch , and p. rossi , phys . lett . *
b199 * ( 1988 ) 101 .
m. karliner , s. sharpe , and y.f .
chang , nucl .
b302 * ( 1988 ) 204 .
a. gocksch , phys . lett .
* b206 * ( 1988 ) 290 ; + u .- j .
wiese , nucl .
b318 * ( 1989 ) 153 .
e. witten , phys .
* 117b * ( 1982 ) 324 .
f. klinkhamer , phys . lett .
* b256 * ( 1991 ) 41 .
wilson and j. kogut , phys .
reports * 12c * ( 1974 ) 75 ; + m. lscher , commun . math . phys .
* 54 * ( 1977 ) 283 . k. osterwalder and r. schrader , commun .
* 31 * ( 1973 ) 83 , * 42 * ( 1975 ) 281 ; + k. osterwalder and e. seiler , ann . phys .
* 110 * ( 1978 ) 440 .
r. narayanan and h. neuberger , rutgers university report ru-94 - 99 ( hep- lat/9412104 ) ; + r. narayanan , h. neuberger , and p. vranas , institute for advanced study report iassns - hep-95 - 19 ( hep - lat/9503013 ) .
s. randjbar - daemi and j. strathdee , ictp ( trieste ) report ic-95 - 6 ( hep - lat/ 9501027 ) .
m.f.l . golterman and y. shamir , institute for theoretical physics report nsf - itp-95 - 60 ( hep - lat/9501035 ) .
g. t hooft , nucl . phys .
* b33 * ( 1971 ) 173 . | lattice fermions have well - known difficulties with chiral symmetry . to evade them
it is possible to couple _ continuum _ fermions to _ lattice _ gauge fields , by introducing an interpolation of the latter . following this line of thinking ,
this paper presents two euclidean formulations of the effective action that appears after functional integration over fermion fields , one for vector - like and the other for chiral couplings . with suitable finite - mode regulators both
effective actions can be evaluated in a finite computation .
the prescriptions provided here contain some details not found in previous work marrying continuum fermions to the lattice via an interpolation .
for example , the counter - terms needed to maintain chiral gauge invariance are explicitly given . by construction
coupling - constant renormalization , anomaly structure , and ( in the chiral gauge theory ) fermion nonconserving amplitudes all satisfy one s expectations from perturbative and semi - classical analyses . addtoresetequationsection # 1#1^ # 1#1 # 1 # 1#1 # 1#2 # 1#2 |
_ near field communication ( nfc ) _
@xcite is a highly - practical emerging technology @xcite . indeed ,
nfc - enabled smartphones are being used in several domains , such as payment @xcite , access control @xcite and ticketing @xcite . in the following ,
we focus on mobile ticketing for public transport .
such a ticketing system is operated by a transport authority representing the transport service provider , and it usually consists of three phases @xcite .
first , the user registers ( _ registration _ ) . secondly , the user obtains the product ( a set of m - tickets ) to be used ( _ provisioning _ ) .
the last phase is the _ validation _ , during which an m - ticket is validated by the transport authority using the nfc connection of a mobile phone . during this phase
, the m - ticketing system must ensure the integrity and authenticity of the m - ticket .
another important aspect is that a network connection @xcite is sometimes required , either for the validator gate or for the smartphone itself .
moreover , the validation is subject to a hard time constraint imposed by the transport operators @xcite : the m - ticket validation must occur in less than @xmath2 . in the literature
, many proposed solutions did not consider the security aspects .
even when security issues are addressed , the concept of user s _ privacy _ , that is users anonymity and users trips unlinkability with respect to the transport service provider , is often overlooked @xcite .
some recent works @xcite have shown a special interest in users privacy .
although they managed to ensure user s anonymity , these solutions suffer from privacy weaknesses . in this paper ,
we propose a new cryptographic protocol for m - ticketing that provides strong authentication , anonymity , and unlinkability properties , whilst remaining efficient when implemented in constrained environments such as sim cards . towards this goal
, we also make various cryptographic optimizations . in particular
, we present a new efficient set - membership proof enabling a prover to prove , in a zero - knowledge way , that his secret belongs to a given public set . unlike previous constructions
, this proof does not require pairing computations , especially on the prover side , which are quite costly , in the order of seconds , e.g. for the following microprocessors : @xmath3 on an atmega @xcite , @xmath4 on a philips hipersmartmips @xcite , @xmath5 on an msp430x @xcite .
we also propose several optimizations of boneh - boyen signature schemes .
these results are of independent interest and increase the performance and efficiency of bb - signatures .
based on these cryptographic primitives , our m - ticketing protocol enables to securely validate an m - ticket without disclosing any personal information to the transport operator , even if the latter is malicious .
we push forward a strict privacy requirement when using one of the numbered m - tickets : for example , if the user uses the m - ticket number @xmath6 from the set @xmath7 $ ] , the transport operator will only be able to check that the used m - ticket is one among the ten m - tickets @xmath7 $ ] .
we implemented our system on a standard nfc sim card . to the best of our knowledge ,
this is the first implementation of _ efficient _ and _ practical _ set - membership proofs and bb - signatures on sim cards .
the entire validation process can be performed without any connection to a back - end server or computations delegation to the smartphone hosting the sim card .
this avoids any tracking of the user by a malware installed in the smartphone .
we also show that the validation process occurs , on average , in @xmath8 when the mobile is switched on and in @xmath9 when the mobile is switched off or its battery is flat ( battery - off ) .
moreover , we show that our protocol supports the post payment approach ( which provides more flexibility ) and give countermeasures to detect any misuse of the service owing to the regular reports of unused m - tickets .
the paper is structured as follows . section [ sec : related ] discusses related work on privacy - preserving m - ticketing schemes . in section
[ sec : framework ] , we detail the framework of our m - ticketing protocol . then , we introduce in section [ sec : req ] our assumptions and the desired security , privacy and functional requirements that an m - ticketing system should fulfil . in section [ sec : cryp ] , we describe the cryptographic assumptions and building blocks required for the sequel . we introduce our new set - membership proof in section [ sec : proof ] and our m - ticketing protocol in section [ sec : sol ] .
section [ sec : analysis ] gives the security proofs . in section
[ sec : poc ] , we present our implementation before concluding in section [ sec : conclusion ] .
heydt - benjamin et al . were the first to propose a cryptographic framework for transport service @xcite .
they discuss the challenges that a privacy - preserving transit system should solve .
using cryptographic transit tickets should not disable basic capabilities like cloning detection , virtual ticket backups or transfers , ticket revocation .
the efficiency of the use of a virtual ticket , especially over the air , is also important .
later , many applied m - ticketing solutions have been proposed @xcite .
however , each proposal has some limitations that this paper tries to solve .
we briefly describe these limitations in this section . first , the rsa based m - ticketing solution proposed by ekberg and tamrakar @xcite , the rfid e - tickets proposed by sadeghi et al .
@xcite and the ticketing solution proposed by blass et al .
@xcite only protect user s privacy with respect to outsiders and not the transport authority .
the transport authority is considered honest which is not anymore a reasonable hypothesis .
second , the most recent protocols that include the expected level of privacy by protecting users privacy even against the transport authority still have some efficiency problems .
for instance , the bbs based protocol proposed by isern - deya et al .
@xcite validates a ticket with a duration of few seconds , even if the implementation has been done on a smartphone .
some recent solutions were close to find the right balance between efficiency and users privacy .
derler et al .
proposal @xcite provides anonymous tickets but users trips are still linkable .
rupp et al .
proposal @xcite provides a privacy preserving pre - payment protocol with refund for public transport with a refund verification procedure with a too high cost for constrained devices .
removing this verification makes possible to use the protocol , but it also enables a malicious transport authority to link users trips , as detailed in appendix [ appendix : flaw ] @xcite . our m - ticketing protocol tries to fill the remaining gap between efficiency and strong privacy .
it can be used in very constrained devices , like sim cards and offers a high level of privacy .
in privacy - preserving m - ticketing system , we consider three different entities . the user ( u ) is the owner of an nfc - enabled smartphone and wants to use the transport service .
the transport authority ( ta ) is the manager of the transport service .
additionally , we consider a third actor which is a revocation authority ( ra )
. it can revoke m - tickets and users anonymity , and is completely independent of the transport authority .
this role may be split between several authorities in such a way that they should all agree and cooperate before recovering the identity of an m - ticket holder .
an m - ticketing system consists of six different phases in our protocol .
( 1 ) the m - ticketing system parameters and keys are initialized during the _ initialization_. ( 2 ) the _ registration _ phase enables a user to register to the transport service .
( 3 ) in the _ permission token request _
phase , a user gets a _ permission token _ allowing him to generate @xmath10 m - tickets .
( 4 ) the _ validation _ phase consists in generating and validating an m - ticket . ( 5 )
the _ revocation _ phase enables to retrieve the identity of a user who generated a given m - ticket and the m - tickets obtained by a given user . finally , ( 6 ) the _ reporting _ phase enables a user to report his usage ( i.e. the m - tickets that he did not use ) to the transport authority that can detect any duplication of m - tickets ( i.e. m - tickets that have been validated several times ) . in the sequel ,
these phases are modeled with various algorithms and protocols executed by the above entities .
* initialization . * @xmath11 : this probabilistic algorithm outputs @xmath12 a description of the system parameters .
we assume that @xmath12 are implicit to the other algorithms , and that they include @xmath13 , the security parameter , and @xmath10 , the number of m - tickets that each book / set of m - tickets contains .
they are also an implicit input to the adversary , we will then omit them .
@xmath14(@xmath12 ) : this probabilistic algorithm outputs the two following secret / public key pairs : @xmath15 for the revocation authority and @xmath16 for the transport authority .
the system public key @xmath17 is eventually set as ( @xmath12 , @xmath18 , @xmath19 )
. * registration . * @xmath20(@xmath17 , @xmath21 ) : this probabilistic algorithm outputs a secret / public key pair @xmath22 for a user identifier @xmath21 .
@xmath23(@xmath24 ) ) : this is an interactive protocol between a new user that takes as input his identity @xmath21 and his public key @xmath25 , and the ta that takes as input the database @xmath26 where the identifiers of the registered users will be stored .
if ta accepts the protocol , the user s identity and public key are stored within @xmath26 .
* permission token request . * @xmath27(u@xmath28 , ta@xmath29 ) : this is an interactive protocol between a user that takes as input @xmath30 @xmath31 , and the ta that takes as input @xmath29 .
if the user accepts the protocol , his output is a permission token @xmath32 that will enable him to generate / validate a set of @xmath10 m - tickets .
if ta accepts the protocol , its output is a transcript _ view _ of the protocol .
* validation .
* @xmath33(@xmath34 ) : this probabilistic algorithm takes as input a user s permission token @xmath32 and outputs an m - ticket @xmath35 with a serial number @xmath36 such that @xmath37 $ ] .
@xmath38(@xmath39 ) : this deterministic algorithm takes as input a ticket @xmath35 .
if @xmath35 is valid , it outputs @xmath40 ( otherwise @xmath41 ) and @xmath35 is stored within the database @xmath42 that will be used to detect the m - tickets that have been used several times .
* revocation . * @xmath43(@xmath44 , @xmath26 , @xmath35 ) : this deterministic algorithm takes as input the private key @xmath44 , the database @xmath26 and a valid m - ticket @xmath35 .
it outputs the identifier @xmath21 of the user who obtained @xmath35 .
if @xmath21 does not belong to @xmath26 , it outputs @xmath45 .
@xmath46(@xmath44 , _ view _ , @xmath21 ) : this deterministic algorithm takes as input the private key @xmath44 , a user s identifier @xmath21 and a transcript _ view _ of an execution of @xmath27 with this user .
it outputs all the m - tickets that can be generated from the token obtained after the execution of the @xmath27 protocol that led to _
* reporting .
* @xmath47 : this algorithm is executed by a user with his permission token @xmath32 .
the user generates all the unused m - tickets and collects them in a _ usage report _ @xmath48 .
@xmath48 is then sent to the transport authority .
@xmath49(@xmath50 ) : this deterministic algorithm takes as input @xmath36 the serial number of a valid m - ticket @xmath35 and @xmath42 , and outputs the number of occurrences of @xmath36 in @xmath42 .
in this section , we first describe our trust assumptions .
then , we detail the requirements of our m - ticketing system .
we consider two types of requirements , _ functional _ requirements , which are `` efficiency '' and `` versatility '' and _ security and privacy _ requirements , which consist in `` correctness '' , `` unforgeability '' , `` unlinkability '' and `` non - frameability '' . at the user s side
, we consider an untrusted mobile platform ( smartphone ) with an incorporated trusted secure element .
we use a sim card as a tamper resistant secure element since it could be certified eal4 + @xcite . using a sim card
will ensure the integrity of the implementation .
indeed , the m - ticketing cardlet manages the cryptographic credentials of the m - ticketing application and executes required cryptographic computations without delegating any of them to the smartphone .
consequently , if the mobile application is compromised , the security will not be impacted because all the sensitive data and credentials are stored and managed in the sim card . if the sim card gets compromised , we planned countermeasures ( cf .
section [ subsub : couter ] ) .
m - ticketing systems must fulfil functional requirements imposed by transport operators @xcite , in particular the validation of an m - ticket must be performed in less than @xmath51 .
we must consider that a sim card has limited computation capabilities .
in particular , pairing apis are not available on current sim cards .
the mobile phone and the validator can not be assumed to be connected to a back - end server during the validation phase .
this enables the user to use an m - ticket in any kind of situation , especially in areas with low connectivity , e.g. , underground , or if the battery of his mobile is flat .
moreover , the m - ticketing system must support the post - payment mode , i.e. , charged later ( after its use ) . besides the correctness property ( which is obvious )
, we formally define three security and privacy properties of our m - ticketing protocol in which the attack capabilities of a probabilistic polynomial time adversary @xmath52 are modeled by providing him access to some oracles . in the sequel , @xmath53 will denote the set of honest users and @xmath54 the set of corrupted users .
we assume that @xmath52 receives all the exchanged messages in our system .
@xmath52 acts as an active adversary as regards to the messages issued by malicious users and as a passive adversary with respect to honest users .
@xmath55 is an oracle that will be used by an adversary in order to register honest users . by calling this oracle with @xmath21 as argument
, the adversary adds a new user .
the oracle runs @xmath56 and adds @xmath21 ( along with @xmath25 ) to the set @xmath53 .
the private key @xmath57 is kept secret and public key @xmath25 is returned to the adversary .
@xmath58 is an oracle that will be used by an adversary in order to register malicious users .
the adversary calls this oracle with argument the identifier @xmath21 of a user and sets his public key to @xmath25 and his private key to @xmath57 . the identity @xmath21 ( along with @xmath25 ) is added to the set @xmath54 .
@xmath59 is a user secret key oracle enabling the adversary to obtain the private key @xmath57 of a user @xmath60 .
the oracle transfers @xmath21 to @xmath54 and returns @xmath57 .
@xmath61 is an oracle that runs the user s side in the @xmath27 protocol .
this oracle will be used by an adversary playing the role of a malicious ta .
the adversary gives to the oracle an identity @xmath21 of an honest user and his public key @xmath25 .
the adversary is then given a transcript view of the protocol .
@xmath62 is an oracle that runs the transport authority side in the @xmath27 protocol .
this oracle will be used to simulate the execution of the protocol between a user ( corrupted or not ) and an honest ta .
@xmath63 is an oracle that takes as input the identifier @xmath21 of an honest user and a transcript _ view _ of an execution of the @xmath27 protocol with this user and outputs an m - ticket @xmath35 using a fresh index @xmath64 that has not been used in a previous query of @xmath65 on @xmath21 and _ view_. the oracle records ( @xmath21 , @xmath35 ) in a list _ set_. @xmath66 is an oracle that takes as input the identifier of a user @xmath21 and a transcript _ view _ of an execution of @xmath27 with this user and outputs all the m - tickets that this user is able to generate .
@xmath67 is an oracle that returns the identifier @xmath21 of the user who generated an m - ticket @xmath35 .
@xmath68 is an oracle that takes as input the identifier @xmath21 of an honest user and a transcript _ view _ of a @xmath27 execution with this user and outputs the set of unused m - tickets .
for each unused m - ticket @xmath69 , the oracle records @xmath70 in _
set_. in the sequel , we denote by @xmath52(_keys _ , _
db _ : _ oracles _ ) an adversary who receives the keys `` _ _ keys _ _ '' .
this adversary has only _ read access _ to the databases `` _ _ db _ _ '' and can query the oracles `` _ _ oracles _ _ '' . informally speaking
, it should be impossible for anyone to falsely accuse an honest user of having spent an m - ticket .
we formally define the non - frameability experiment @xmath71 in figure [ fig : nonfra ] .
the scheme is non - frameable , if for any probabilistic polynomial time adversary @xmath52 , the probability + @xmath72[@xmath71=1 ] is negligible .
l + informally speaking , it should be impossible for anyone ( 1 ) to validate more m - tickets than what he obtained i.e. an adversary who retrieved @xmath73 tokens @xmath32 ( @xmath73 sets of @xmath10 m - tickets ) should not be able to generate more that @xmath74 m - tickets ; ( 2 ) to validate m - tickets such that the algorithm @xmath75 returns @xmath45 , i.e. , an identifier @xmath21 that does nt appear in @xmath26 .
we formally define the unforgeability experiment @xmath76 in figure [ fig : unf ] .
the scheme is unforgeable if for any probabilistic polynomial time adversary @xmath52 , the probability @xmath77 $ ] is negligible .
l + informally speaking , it should be impossible , except for the revocation authorities , to trace the m - tickets obtained by a user , in particular : ( 1 ) to link m - tickets obtained during the permission token request phase to the validated / used ones ; ( 2 ) to link two m - tickets validated by the same user or to decide whether two validated m - tickets have the same number / index or not ; ( 3 ) to link validated m - tickets to non - used m - tickets reported by the user to the transport authority . for this
, an adversary has full control over the transport authority ( in particular it owns the private key @xmath78 ) and all the users except two honest users @xmath79 and @xmath80 .
the adversary can initiate the @xmath75 protocol over any m - ticket and can get the user s identity behind it , except for the m - tickets generated by @xmath79 and @xmath80 .
he can also initiate the @xmath81 protocol for all the users except for @xmath79 and @xmath80 .
we define the unlinkability experiment @xmath82 in figure [ fig : unlink ] .
the scheme is unlinkable if for any probabilistic polynomial time adversary @xmath52 , the advantage @xmath83- 1/2|$ ] is negligible .
l + * remark : * another basic requirement that our ticketing system should fulfill is the _ traceability _ meaning that revocation should succeed in the face of attacks by malicious users . in fact , our formulations of _ unforgeability _ and _ non - frameability
_ capture this requirement . indeed ,
an attacker who could produce a ticket which either ( 1 ) can not be traced to a user or ( 2 ) identify an honest user who did nt obtain this ticket , would either break the _ unforgeability _ requirement ( 1 ) or the _ non - frameability _ requirement ( 2 ) .
in this section , we introduce the notations , definitions and cryptographic tools used in the description of our set - membership proof and m - ticketing protocol .
we consider throughout this document , except when it is explicitly mentioned , bilinear maps @xmath84 where all groups @xmath85 , @xmath86 and @xmath87 are multiplicative and of prime order @xmath88 .
the mapping @xmath89 satisfies the following properties : @xmath90 * decisional diffie - hellman ( ddh ) .
* for any multiplicative group @xmath91 of prime order @xmath88 , the ddh assumption states that given a random generator @xmath92 , two random elements @xmath93 , @xmath94 in @xmath91 , and a candidate @xmath95 , it is hard to decide whether @xmath96 or not . *
external diffie - hellman ( xdh ) .
* given three groups @xmath85 , @xmath86 , @xmath87 , and a bilinear map @xmath84 , the xdh assumption states that ddh assumption holds in @xmath85 .
* q - strong diffie - hellman ( q - sdh ) . *
the q - sdh assumption holds for some group @xmath85 if it is hard , given ( @xmath97 , @xmath98 , @xmath99 ) @xmath100 @xmath101 , to output a pair @xmath102 . * q - decisional diffie - hellman inversion ( q - ddhi ) . * in any multiplicative group @xmath91 of prime order @xmath88 , the q - decisional diffie - hellman inversion assumption states that , given a random generator @xmath92 and the values @xmath103 , for a random @xmath104 and a candidate @xmath95 , it is hard to decide whether @xmath105 or not . * the decisional composite residuosity assumption .
* there is no probabilistic polynomial time distinguisher for @xmath106-th residues modulo @xmath107 . in other words
, there is no probabilistic polynomial time adversary that can distinguish @xmath108 from @xmath109 , where @xmath110 @xmath111 .
roughly speaking , a zero knowledge proof of knowledge , denoted zkpk , is an interactive protocol during which a prover convinces a verifier that he knows a set of secret values ( @xmath112 ) verifying a given relation @xmath113 without revealing anything else .
such a proof will be denoted in the sequel @xmath114 .
a zkpk should be * complete * ( a valid prover is accepted with overwhelming probability ) , * sound * ( a false prover should be rejected with overwhelming probability ) and * zero - knowledge * ( no information about the secret is revealed ) . these zkpk can be made non - interactive using the generic transformation introduced by fiat and shamir @xcite . the resulting non - interactive proofs , sometimes called _ signatures of knowledge _
@xcite , can be proven secure in the random oracle model of @xcite ( see @xcite ) .
such a signature of knowledge , for the relation @xmath113 , will be denoted @xmath115 .
these signatures , proposed by camesnich and lysyanskaya @xcite , are equipped with additional protocols .
one of these protocols allows a signature to be issued on messages that are not known by the signer , but for which the signer only knows a commitment .
informally , in a protocol for signing a committed value , we have a signer with public key @xmath116 , and the corresponding secret key @xmath117 , and a user who queries the signer for a signature .
the common input to the protocol is a commitment @xmath118 , known by both parties , on secret values @xmath119 known only by the user . at the end of this protocol , the user obtains a valid @xmath120 and the signer learns nothing about @xmath119 .
another protocol allows to prove knowledge of a signature on a tuple of messages @xmath119 without releasing any information on the corresponding signature .
each message can either be revealed to the verifier , sent in a committed form , or it may be such that the verifier has no information on it . in particular , it is possible to prove the knowledge of a cl - signature on committed values . a set membership proof allows a prover to prove , in a zero - knowledge way , that his secret lies in a given public set .
such proofs can be used , for instance , in the context of electronic voting , where the voter needs to prove that his secret vote belongs to the set of all possible candidates .
recent propositions of set membership proofs @xcite follow the same idea : a designated authority produces public signatures on each element of the public set @xmath121 ( and only on these elements ) .
the proof of knowledge of a secret @xmath122 consists in proving the knowledge of a ( public ) signature on the secret @xmath123 ( which will be only possible if @xmath123 belongs to the set @xmath121 ) without revealing @xmath123 or the used signature .
solutions in @xcite require the prover to perform pairing computations . however , these cryptographic tools are not often convenient for many constrained platforms .
the signature scheme used by the designated authority ( which could be the verifier ) to sign each element of the set @xmath121 is the one proposed by boneh and boyen @xcite .
based on their scheme , which is secure under the q - sdh assumption , it is possible to prove knowledge of a signature on a message , without revealing the signature nor the message . having a secret key @xmath124 and two random generators @xmath125
, @xmath126 of @xmath85 , the signature of a message @xmath127 is obtained by computing @xmath128 .
given a bilinear pairing @xmath89 , a signature @xmath129 of @xmath130 is valid if @xmath131 , where @xmath132 is the signer s public key .
let @xmath91 be a cyclic group with prime order @xmath88 where the decisional diffie - hellman ( ddh ) problem is assumed to be hard and @xmath125 , @xmath126 two random generators of @xmath91 .
the signer s private key is @xmath133 and its public key is @xmath134 .
similarly to boneh - boyen signatures with pairings , the signature on a message @xmath130 is the value @xmath135 .
this implies that @xmath136 .
since we work in a group @xmath91 not equipped with a bilinear map , the signer must additionally prove that the signature on @xmath130 is valid , which is done by generating a zkpk @xmath137 that the discrete logarithm of ( @xmath138 ) in the base @xmath139 is equal to the discrete logarithm of @xmath140 in the base @xmath126 : @xmath141 .
such a proof of equality of discrete logarithms has been introduced in @xcite .
finally , the signature @xmath139 on @xmath130 is valid if the proof @xmath137 is valid . * theorem 1 . * the boneh - boyen ( bb for short ) signature scheme without pairings is existentially unforgeable under a weak chosen message attack under the q - strong diffie - hellman assumption , in the random oracle model .
* proof ( sketch ) .
* under the q - strong diffie - hellman assumption , it is impossible to find a message @xmath130 ( which was not given to the signing oracle ) and a value @xmath139 such that @xmath142 , as proved in @xcite .
moreover , in the random oracle model , the proof of knowledge @xmath137 is unforgeable @xcite , which concludes the proof .
> p0.4- > p0.1 - 2 > p0.5- & & + * public input : * public parameters , sets @xmath121 and @xmath143 & & * public input : * public parameters , sets @xmath121 and @xmath143 + * private input : * @xmath64 & & * private input : * @xmath124 + + + + + + + & @xmath144 & * choose * @xmath145 ( a random challenge ) + & & + & & * check * that @xmath146 + & & @xmath147 if the verifier and the designated entity are the same entity , it implies that the verifier holds the private signature key @xmath124 * ( first case)*. hence the prover does not send the value @xmath148 .
the verifier can compute it : @xmath149 and goes to ( * ) + & & @xmath147 otherwise , if the verifier does nt know the private signature key @xmath124 * ( second case ) * , then , it checks that @xmath150 and goes to ( * ) + & & * ( * ) compute : * @xmath151 and @xmath152 + & & * check * that : @xmath153 + the el gamal cryptosystem @xcite works as follows .
let @xmath91 be a cyclic group with prime order @xmath88 where the decisional diffie - hellman ( ddh ) problem is assumed to be hard .
the public key consists of the elements @xmath154 , where @xmath155 is a random generator of @xmath91 , and the corresponding private key is formed by @xmath156 .
the el gamal ciphertext of a message @xmath157 is @xmath158 , where @xmath159 is a random number .
the decryption of the ciphertext @xmath160 is obtained through the following computation : @xmath161 .
the el gamal cryptosystem is semantically secure under the ddh assumption .
the paillier cryptosystem @xcite works as follows .
let @xmath162 and @xmath163 be random primes for which @xmath164 and @xmath165 .
let @xmath166 , @xmath167 , @xmath168 , and @xmath169 .
the public key is @xmath170 and the secret key is @xmath171 . to encrypt @xmath172 , the user chooses @xmath173 and @xmath174 .
the decryption algorithm is given by @xmath175 .
the paillier cryptosystem is secure under the decisional composite residuosity assumption . in a threshold version , the el gamal public key and its corresponding private key @xcite ( resp .
the paillier public key and its corresponding private key @xcite ) are cooperatively generated by @xmath106 parties ; however , the private key is shared among the parties . in order to decrypt a ciphertext ,
a minimal number of @xmath176 ( the threshold ) out of @xmath106 parties is necessary .
a useful building block of our m - ticketing protocol is the pseudo - random function introduced by dodis and yampolskiy @xcite .
their construction works as follows .
let @xmath91 be a cyclic group with prime order @xmath88 , @xmath177 a generator of @xmath91 , @xmath178 a private secret .
the pseudo - random function ( prf for short ) @xmath179 takes as input a message @xmath180 and outputs @xmath181 .
this construction is secure under the q - ddhi assumption in @xmath91 .
in this section , we present a new set membership protocol . our set membership proof is a zkpk which allows the prover to convince any verifier that the committed integer @xmath64 lies in the public discrete set @xmath121 .
this zkpk can be made non - interactive using the fiat - shamir heuristic and the random oracle model @xcite .
the commitment scheme that we use is the one proposed by pedersen @xcite .
our set membership proof bears some similarities with the protocol proposed by camenisch et al .
@xcite . yet
, our set membership proof does not ask the prover ( a sim card and/or a smartphone in our setting ) nor the verifier ( in a specific scenario ) to perform pairing computations .
our set membership proof is also more efficient and conceptually simpler than the recent one proposed by canard et al . at europki 2013 @xcite .
their scheme which involves several verifiers , who share a decryption key , seems to be tailored to specific applications such as e - voting where the verifiers ( tallying authorities ) own such a key to open ballots .
let @xmath85 , @xmath86 and @xmath87 be three ( multiplicative ) bilinear groups of prime order @xmath88 and a bilinear map @xmath84 .
let @xmath182 and @xmath183 be three generators of @xmath85 and consider a generator @xmath184 of @xmath86 .
a designated authority ( which may or may not be the verifier ) randomly chooses a value @xmath185 ( its private key ) and publishes the corresponding public key @xmath186 . after generating its key
, this designated authority can issue bb - signatures on each element of the public set @xmath121 .
let @xmath64 denote an element of the set @xmath121 and @xmath187 the bb - signature on the value @xmath64 , i.e. , @xmath188 .
the set of all message - signature pairs @xmath189 , that we denote by @xmath143 , is published by the designated authority .
@xmath190 will denote a hash function , for instance , sha-256 .
in figure [ tab : proof ] , we propose a protocol enabling to prove that the value @xmath64 committed by the prover in the commitment @xmath191 belongs to the set @xmath121 which comes down to proving knowledge of the bb - signature @xmath187 .
* remark 1 : * proving knowledge of a valid bb - signature @xmath188 on a value @xmath64 committed in @xmath192 , without revealing neither @xmath64 nor the signature and without using pairings on the prover s side can be done as follows .
the prover first randomizes @xmath187 by choosing a random value @xmath193 and computes @xmath194 .
since @xmath188 this implies that @xmath195 and then that @xmath196 .
it implies , if we note @xmath197 and @xmath198 that @xmath199 . as a result , in order to prove that the value @xmath64 committed in @xmath191 belongs to the set @xmath121 , the prover just has to send @xmath200 and @xmath148 to the verifier and compute the following zkpk : @xmath201 .
the computation of the zkpk @xmath202 is described in figure [ tab : proof ] . upon receiving @xmath202 , the verifier proceeds to the verification .
we distinguish two cases . in the _ first _ case , the verifier and the designated authority are the same .
thus , the verifier holds the private signature key @xmath124 .
consequently , the verification of @xmath202 occurs without requiring any pairing computations .
this implies that our set membership proof does not require pairing computations for either the prover or the verifier . in the _ second _
case , the verifier does not have the private signature key @xmath124 of the designated authority .
then , the verifier needs to perform some pairing computations in order to verify @xmath202 .
nevertheless , the prover still has no pairing computations to perform .
this is particularly interesting for the m - ticketing use case .
* theorem 2 . *
if the @xmath203sdh assumption holds , then the protocol in figure [ tab : proof ] is a zero - knowledge argument of set membership for a set @xmath121 .
the proof is in appendix [ appendix : proof2 ] .
this section begins with an overview of our protocol which main notations are summarized in table [ tab : notation ] , before moving to our m - ticketing protocol and its post - payment aspects .
.notation used in the m - ticketing protocol [ cols= " < , < " , ] in a post - payment approach , the m - ticketing application must report unused m - tickets to the back - end server , before the pre - defined @xmath148 day .
thus , the regular reports does not question the privacy of the user .
the following example gives further clarification .
suppose that the user retrieved a permission token of @xmath204 m - tickets .
before @xmath148 day , he used 4 m - tickets , i.e. , m - tickets number 1 , 2 , 3 and 4 .
on @xmath148 day , the m - ticketing application will report to the back - end server the m - ticket number 5 using a network connection .
the report contains @xmath205 = @xmath206 and the proof @xmath207)$ ] .
regularly revealing the unused m - tickets enables the transport authority to ( 1 ) charge the user , ( 2 ) check the reliability and accuracy of the reports without questioning the user s privacy .
indeed , the unlinkability of user s m - tickets @xmath208 , both during the use of the service and during the reporting phase , is ensured owing to the q - ddhi and ddh assumptions ( cf .
_ unlinkabililty _ proof ) . a malicious user s goal simply consists on not paying or paying less than what he is supposed to . to do so
, he may try to block the reporting or attack the cardlet of the sim .
if the user blocks network communications in order to block reporting , the sim cardlet will refuse to issue m - tickets for a permission token after reaching the limit of @xmath10 , or the limit of the time window controlled by @xmath209 .
this countermeasure @xmath118 relies on the security of the sim card .
if the user performs a successful physical attack against the sim card , which is extremely difficult @xcite , he may try to ( 1 ) defeat the countermeasure @xmath118 and never do reports , or ( 2 ) use the same m - ticket several times or ( 3 ) report a used m - ticket .
this will be detected in @xmath42 and ta can decide to break the user s anonymity and retrieve all his m - tickets usage history .
thus , the user is forced to report his consumption and renew his permission token lest the service is unusable or ta breaks his anonymity .
note that forging an m - ticket is not possible , even with a compromised sim because of the _ unforgeability _ property .
we prove that our m - ticketing protocol provides the security and privacy properties defined in section [ sub : secmodel ] . * theorem 4 ( non - frameability ) . *
our m - ticketing protocol is non - frameable , in the random oracle model , under the q - ddhi assumption .
the proof is detailed is appendix [ appendix : proofs ]
. * theorem 5 ( unforgeability ) . *
our m - ticketing protocol satisfies the unforgeability requirement , in the random oracle model , under the q - sdh assumption .
the proof is detailed is appendix [ appendix : proofs ]
. * theorem 6 ( unlinkability ) . *
our m - ticketing protocol satisfies the unlinkability requirement , in the random oracle model , under the q - ddhi and ddh assumptions .
the proof is detailed is appendix [ appendix : proofs ] . because of space limitation
, the proof of theorem 6 is omitted from this paper but is detailed in the extended version @xcite .
we implemented the user side of our solution on a sim card and the validator side in a regular pc ( cf .
appendix [ appendix : param ] ) .
we used a 256-bit barreto - naehrig curve @xcite over @xmath210 since this family of curves provides an optimal size for @xmath211 and @xmath212 while preventing the mov attack @xcite due to their embedding degree of 12 .
@xmath213 is the group of @xmath210 - rational points of order p and @xmath214 is the subgroup of trace zero points in @xmath215 $ ] .
our pairing is of type-3 @xcite .
more details are given in appendix [ appendix : param ] .
* pre - computations . *
the pre - computations for preparing one signature occurs in @xmath216 .
it consists in computing elements involved in proving that the user knows a valid signature on his secret @xmath217 .
the total size of the elements that the card has to store for one signature is @xmath218 ( 24 @xmath219 elements , @xmath220 compressed points and one digest output ) . * validation phase .
* table [ tab : timing ] gives timings ( average over 50 trials and standard deviation between parentheses ) for the whole validation protocol which includes the validator authentication , the signature generation and an m - ticket verification .
the timings include as well the nfc exchanges duration .
we denote by `` battery - off '' a powered - off phone either by the user , or because the battery is flat . in this situation ,
as stated by nfc standards , nfc - access to the sim card is still possible , but with degraded performances . regarding the validator authentication , we chose to use rsa with a @xmath221 bits key ( this is the greatest size supported by the sim card ) and a short public verification exponent @xmath222 ( @xmath223 ) .
the validator asks the card for a challenge ( @xmath224 ) .
then , he sends back his response to this challenge , his own challenge ( @xmath225 ) ( @xmath226 bytes ) and the current date ( @xmath227 ) ( @xmath228 bytes ) .
the column `` card signature '' gives the duration of real - time computations of a signature ( computing an m - ticket @xmath36 , @xmath229 and @xmath202 ) and the nfc communication time .
the considered operations for generating the signature are only one hash value and lightweight operations in @xmath219 .
the size of the computed signature is @xmath230 ( sent by 4 apdus ) . regarding the communication between a sim card and a reader , it is slow ( @xmath231 ) , but the whole process ( signature+nfc ) remains very fast , i.e. , @xmath232 on average . for the signature verification by the validator , we distinguish two cases : the validator holds the private signature keys ( @xmath124 , @xmath233 ) , or not .
the extra pairing computations performed in the second case do not add an important delay to the verification because a regular desktop pc can efficiently achieve such computations . in total , the m - ticket validation occurs for the first case ( without pairing at the verifier side ) in @xmath0 on average and in @xmath234 for the second case ( with pairing at the verifier side ) .
when the battery is flat , the validation occurs in at most @xmath235 on average .
in this paper , we proposed various cryptographic enhancements : ( 1 ) optimizations of bb - signature schemes and ( 2 ) a new set - membership proof that does not require pairing computations at the prover side .
these contributions enable to design efficient protocols over nfc . then , based on these cryptographic primitives , we designed a secure m - ticketing protocol that prevents a malicious transport authority from linking users trips .
moreover , as the entire computations are done within the sim card , we enhance the user s privacy with regards to a compromised smartphone . owing to regular reports of unused m - tickets ,
the user is able to securely post - pay for his m - tickets without disclosing any information about the performed trips .
moreover , we planned countermeasures against m - ticket cloning and multiple usage . finally , we prove that our protocol is completely off - line and efficient : the validation occurs in @xmath0 and in @xmath1 when the battery is flat . 10 p. s. barreto and m. naehrig . .
in b. preneel and s. tavares , editors , _ selected areas in cryptography _ , volume 3897 of _ lncs _ , pages 319331 .
springer berlin heidelberg , kingston , on , canada , 2006 .
bellare , namprempre , pointcheval , and semanko .
the one - more - rsa - inversion problems and the security of chaum s blind signature scheme .
, 16(3):185215 , 2003 .
doi : http://dx.doi.org/10.1007/s00145-002-0120-1[10.1007/s00145-002-0120-1 ] .
m. bellare and p. rogaway . .
in _ 1st acm conference on computer and communications security _ , ccs 93 , pages 6273 , fairfax , virginia , usa , 1993 .
tickets , fares and route maps .
http://www.berlin.de/en/public-transportation/1772016-2913840-tickets-fares-and-route-maps.en.html .
blass , a. kurmus , r. molva , and t. strufe . .
, 36(4):468480 , 2013 .
doi : http://dx.doi.org/10.1016/j.comcom.2012.10.012[10.1016/j.comcom.2012.10.012 ] .
d. boneh and x. boyen . .
in c. cachin and j. l. camenisch , editors , _ advances in cryptology - eurocrypt 2004 _ , volume 3027 of _ lncs _ , pages 5673 .
springer berlin heidelberg , interlaken , switzerland , 2004 .
d. boneh and x. boyen . .
, 21(2):149177 , 2008 .
doi : http://dx.doi.org/10.1007/s00145-007-9005-7[10.1007/s00145-007-9005-7 ] .
d. boneh , x. boyen , and h. shacham . .
in m. franklin , editor , _ advances in cryptology - crypto 04 _ , volume 3152 of _ lncs _ , pages 4155 .
springer berlin heidelberg , santa barbara , california , usa , 2004 .
j. camenisch , r. chaabouni , and a. shelat . .
in j. pieprzyk , editor , _ advances in cryptology - asiacrypt 2008 _ , volume 5350 of _ lncs _ , pages 234252 .
springer berlin heidelberg , melbourne , australia , 2008 .
j. camenisch and a. lysyanskaya . .
in m. franklin , editor , _ advances in cryptology - crypto 04 _ , volume 3152 of _ lncs _ , pages 5672 .
springer berlin heidelberg , santa barbara , california , usa , 2004 .
j. camenisch , j .-
m . piveteau , and m. stadler . .
3rd acm conference on computer and communications security _ , ccs 96 , pages 8894 , new delhi , india , 1996 .
s. canard , i. coisel , a. jambert , and j. traor . . in s.
katsikas and i. agudo , editors , _ public key infrastructures , services and applications _ , volume 8341 of _ lncs _ , pages 4764 .
springer berlin heidelberg , egham , uk , 2014 .
r. chaabouni , h. lipmaa , and b. zhang . . in a.
keromytis , editor , _ financial cryptography and data security _ ,
volume 7397 of _ lncs _ , pages 179199 .
springer berlin heidelberg , kralendijk , bonaire , 2012 .
d. chaum and t. p. pedersen . .
in e. f. brickell , editor , _ advances in cryptology - crypto 92 _ , volume 740 of _ lncs _ , pages 89105 , santa barbara , california , usa , 1993 .
springer berlin heidelberg .
s. chaumette , d. dubernet , and j. ouoba . .
in s. balandin , y. koucheryavy , and h. hu , editors , _ the 11th international conference on next generation wired / wireless networking _ , volume 6869 of _ lncs _ , pages 165177 , st .
petersburg , russia , 2011 .
springer berlin heidelberg .
d. derler , k. potzmader , j. winter , and k. dietrich . .
in l. chen , m. yung , and l. zhu , editors , _ trusted systems _
, volume 7222 of _ lncs _ , pages 6683 , beijing , china , 2012 .
springer berlin heidelberg .
y. g. desmedt and y. frankel . threshold cryptosystems . in g.
brassard , editor , _ advances in cryptology - crypto 89 _ , volume 435 of _ lncs _ , pages 307315 , santa barbara , california , usa , 1989 .
springer berlin heidelberg .
a. devegili , m. scott , and r. dahab . .
in t. takagi , t. okamoto , e. okamoto , and t. okamoto , editors , _ pairing - based cryptography - pairing 2007 _ , volume 4575 of _ lncs _ , pages 197207 .
springer berlin heidelberg , tokyo , japan , july 2007 .
a. dmitrienko , a .- r .
sadeghi , s. tamrakar , and c. wachsmann . .
in s. katzenbeisser , e. weippl , l. camp , m. volkamer , m. reiter , and x. zhang , editors , _ trust and trustworthy computing _ , volume 7344 of _ lncs _ , pages 219238 .
springer berlin heidelberg , vienna , austria , 2012 .
y. dodis . .
in y. desmedt , editor , _ public key cryptography - pkc 2003 _ , volume 2567 of _ lncs _ , pages 117 .
springer berlin heidelberg , miami , fl , usa , 2003 .
y. dodis and a. yampolskiy . .
in s. vaudenay , editor , _ public key cryptography - pkc 2005 _ , volume 3386 of _ lncs _ , pages 416431 .
springer berlin heidelberg , diablerets , switzerland , 2005 .
ekberg and s. tamrakar . .
in l. chen , m. yung , and l. zhu , editors , _ third international conference on trusted systems _ , volume 7222 of _ lncs _ , pages 4865 , beijing , china , 2012 .
springer berlin heidelberg .
t. el gamal . .
in g. r. blakley and d. chaum , editors , _ advances in cryptology - crypto 84 _ , volume 196 of _ lncs _ , pages 1018 , santa barbara , california , usa , 1985 .
springer berlin heidelberg .
m. eznack , j .-
warry , c. loiseaux , g. dufay , r. atoui , n. herbreteau , j. pieniazek , and f. thabaret . .
http://www.ssi.gouv.fr/img/certificat/anssi-cc-cible_pp-2010-04en.pdf , june 2010 .
a. fiat and a. shamir . .
in a. m. odlyzko , editor , _ advances in cryptology - crypto 86 _ , volume 263 of _ lncs _ , pages 186194 , santa barbara , california , usa , 1987 .
springer berlin heidelberg .
p. fouque and j. stern .
fully distributed threshold rsa under standard assumptions . in c.
boyd , editor , _ advances in cryptology - asiacrypt 2001 _ , volume 2248 of _ lncs _ , pages 310330 , gold coast , australia , 2001 .
springer berlin heidelberg . .
http://www.frost.com/sublib/display-report.do?id=9843-00-13-00-00 , january 2011 .
g. fuchsbauer , d. pointcheval , and d. vergnaud . .
in j. garay , a. miyaji , and a. otsuka , editors , _ cryptology and network security _ , volume 5888 of _ lncs _ , pages 226247 .
springer berlin heidelberg , kanazawa , japan , 2009 .
s. d. galbraith , k. g. paterson , and n. p. smart . .
, 156(16):31133121 , 2008 .
doi : http://dx.doi.org/10.1016/j.dam.2007.12.010[10.1016/j.dam.2007.12.010 ] .
r. gennaro , s. jarecki , h. krawczyk , and t. rabin . .
in m. joye , editor , _ topics in cryptology - ct - rsa 2003 _ , volume 2612 of _ lncs _ , pages 373390 .
springer berlin heidelberg , san francisco , ca , usa , 2003 .
c. p. l. gouva , l. b. oliveira , and j. lpez . .
, 2(1):1929 , 2012 .
doi : http://dx.doi.org/10.1007/s13389-012-0029-z[10.1007/s13389-012-0029-z ] . . .
september 2012 .
t. s. heydt - benjamin , h .- j .
chae , b. defend , and k. fu . .
in g. danezis and p. golle , editors , _
6th international conference on privacy enhancing technologies - pet06 _ , volume 4258 of _ lncs _ , pages 119 , cambridge , uk , 2006 .
springer berlin heidelberg .
e. hufschmitt and j. traor . .
in t. takagi , t. okamoto , e. okamoto , and t. okamoto , editors , _ pairing - based cryptography - pairing 2007 _ , volume 4575 of _ lncs _ , pages 268292 , tokyo , japan , july 2007 .
isern - deya , a. vives - guasch , m. mut - puigserver , m. payeras - capella , and j. castella - roca . .
, 56(10):11981215 , apr .
doi : http://dx.doi.org/10.1093/comjnl/bxs033[10.1093/comjnl/bxs033 ] . . .
a. menezes , s. vanstone , and t. okamoto . .
23rd annual acm symposium on theory of computing - stoc 91 _ , pages 8089 , new orleans , louisiana , usa , 1991 .
moscow . http://moscow.ru
/ fr / guide / trip_planning / inner_transport / transport / metro/. . .
http://nfc-forum.org/wp-content/uploads/2013/12/nfc-in-public-transport.pdf , 2011 .
p. paillier . .
in _ second international workshop on practice and theory in public key cryptography , pkc 99 _ , volume 1560 of _ lncs _ , pages 223234 , kamakura , japan , mar .
springer berlin heidelberg .
t. pedersen . . in j. feigenbaum , editor , _ advances in cryptology - crypto 91
_ , volume 576 of _ lncs _ , pages 129140 .
springer berlin heidelberg , 1992 .
d. pointcheval and j. stern . .
in u. maurer , editor , _ advances in cryptology - eurocrypt 96 _ , volume 1070 of _ lncs _ , pages 387398 .
springer berlin heidelberg , saragossa , spain , 1996 .
d. pointcheval and j. stern .
security arguments for digital signatures and blind signatures .
, 13(3):361396 , 2000 .
doi : http://dx.doi.org/10.1007/s001450010003[10.1007/s001450010003 ] .
a. rupp , g. hinterwlder , f. baldimtsi , and c. paar . .
sadeghi , editor , _ financial cryptography and data security _ ,
volume 7859 of _ lncs _ , pages 205212 .
springer berlin heidelberg , okinawa , japan , 2013 .
a. sadeghi , i. visconti , and c. wachsmann . .
in c. bettini , s. jajodia , p. samarati , and x. s. wang , editors , _ international workshop on privacy in location - based applications - pilba 2008 _ , volume 397 , malaga , spain , oct . 2008 . ceur . c. schnorr . . , 4(3):161174 , 1991
doi : http://dx.doi.org/10.1007/bf00196725[10.1007/bf00196725 ] . .
proximity mobile payments : leveraging nfc and the contactless financial payments infrastructure .
http://www.smartcardalliance.org/resources/lib/proximity_mobile_payments_200709.pdf , 2007 .
szczechowiak , l. oliveira , m. scott , m. collier , and r. dahab . .
in r. verdone , editor , _ wireless sensor networks _ , volume 4913 of _ lncs _ , pages 305320 .
springer berlin heidelberg , bologna , italy , 2008 .
s. tamrakar and j .- e .
6th international conference on trust & trustworthy computing _ ,
volume 7904 of _ lncs _ , pages 115132 , london , united kingdom , 2013 .
springer berlin heidelberg . .
public transport in paris .
http://en.parisinfo.com/practical-paris/how-to-get-to-and-around-paris/public-transport-paris .
g. arfaoui , j .- f .
lalande , j. traor , n. desmoulins , p. berthom , and s. gharout . .
d. boneh , b. lynn , and h. shacham .
short signatures from the weil pairing . , 17(4):297319 , 2004 .
doi : http://dx.doi.org/10.1007/s00145-004-0314-9[10.1007/s00145-004-0314-9 ] .
s. brands . .
, 2000 .
fouque and d. pointcheval . .
in c. boyd , editor , _ advances in cryptology - asiacrypt 2001 _ , volume 2248 of _ lncs _ , pages 351368 .
springer berlin heidelberg , gold coast , australia , 2001 .
m. geuss . .
http://ars.to/13t01bc , july 2013 .
v. shoup . . , 2004 , 2004
in the following , we give further details about the payment protocol for public transport proposed by rupp et al . at fc 2013 @xcite .
then , we present an attack that enables a malicious transport authority to break users privacy . at first , the user buys a trip authorization token ( @xmath236 ) which consists of an ( extended ) coin in brands scheme @xcite .
this token enables the user to perform one trip .
then , the user receives a fresh refund token ( @xmath237 such that @xmath238 ) , sets a variable @xmath48 to 1 and @xmath239 to 0 .
@xmath48 will be used to aggregate blinding factors and @xmath239 is the refund amount . at the entry turnstile , the user shows a @xmath236 and proves its ownership .
right after this , the user receives a refund calculation token ( @xmath240 ) which consists of a mac on @xmath236 , a timestamp and the reader i d . at the turnstile of the exit ,
the user shows his @xmath241 and proves the ownership of the @xmath236 contained within it .
in addition , he collects refund on his @xmath242 , as described in figure [ tab : refundcollect ] .
a refund value @xmath243 is represented as a @xmath243-times bls signature @xcite @xmath244 such that @xmath245 is the secret bls signature key of the transport authority .
later , the user redeems his @xmath242 .
this consists on sending a blinded version of his @xmath242 ( @xmath246 ) , @xmath247 , @xmath48 and the collected amount @xmath239 to the transport authority .
the latter checks the validity of both @xmath247 and the signature : @xmath248 in @xcite , rupp et al .
assume that , for efficiency reasons , users do nt verify their rts .
indeed , verifications of refund tokens ( step 1 in figure [ tab : refundcollect ] ) is too costly and can not be handled by constrained devices such as sim cards .
this lack of verification could lead to serious privacy weaknesses : a malicious transport authority could link users trips .
> p0.3- > p0.2- > p0.45- + * user*&&*exit turnstile * + & & determine the refund @xmath243 based on the received information in @xmath241 + & & + & & @xmath249 * ( 1 ) * + & & + & & + + * user*&&*transport authority * + & + & & check validity of @xmath247 , + & & @xmath250 , + & + more precisely , for a given station , the transport authority is able to identify all the users who left at this station . to this end , in step ( 1 ) in figure [ tab : refundcollect ] , instead of computing @xmath249 , the transport authority chooses a variable @xmath251 and computes @xmath252=@xmath253 . the user will not detect that @xmath252 is improperly calculated because there is no verification on the user s side .
the refund token @xmath242 will then carry on the exponent @xmath176 till the redeeming refund step , i.e. , @xmath254 where @xmath255 corresponds to the number of times the user left the targeted station , @xmath48 is the product of all the variables @xmath256 and @xmath239 is the sum of all the refunds @xmath243 . at the redeeming refund step , upon receiving the serial number @xmath247 , the blinded version of the refund token @xmath257 , the refund amount @xmath239 , and the aggregate blinding factor @xmath48 , the transport authority will as usual check the validity of @xmath247 and whether the amount is within the allowed range @xmath258 $ ] .
now , in order to verify the signature and distinguish users who exited at the relevant station , the transport authority starts by checking the relation 1[eq1 ] ) .
if equation ( [ eq1 ] ) holds , this implies that the user did not left the targeted station . otherwise , the transport authority checks : @xmath259 to do so , the transport authority will repeat the computations with different values of @xmath255 until the relation ( [ eq2 ] ) holds .
when @xmath260 equation ( [ eq2 ] ) holds , this implies that the user left the targeted station @xmath255 times . in order to monitor several stations at the same time
, the transport authority will have a list of variables @xmath176 , e.g. , @xmath261 for station a , @xmath262 for station b , etc . at the redeeming refund step ,
the transport authority will test different values of @xmath176 until it finds the ones that satisfy the relation ( [ eq3 ] ) : @xmath263 where @xmath264 characterizes a station @xmath123 and @xmath265 represents the number of exits at the station @xmath123 .
for instance , if the user left the station a four times , the station b twice and the station m once , relation ( [ eq3 ] ) will be as follows : @xmath266 . in this way
, the transport authority will be able to identify the users that left a pool of targeted stations but also the number of times they exited at these stations : this clearly implies that users trips are _
linkable_. however , solving relation ( [ eq3 ] ) and finding the right tuples @xmath267 , becomes quickly a complex and unmanageable task when the number of targeted stations increases ( more than three for example ) . to mitigate this ,
a malicious transport authority could only monitor two stations at the same time and periodically change these monitored stations . in this way
, the transport authority will be able to monitor a large number of stations and will have a global overview of the users journeys which breaks the privacy property of rupp et al . system .
we would like to emphasize that privacy holds in p4r ( rupp et al .
system @xcite ) if we assume that the transport authority is `` honest but curious '' , i.e. , it will not deviate from the refund protocol , but such an assumption is too strong in the context of transport systems @xcite . in order to mitigate this issue
, the user should check the received refund ( step ( 1 ) in figure [ tab : refundcollect ] ) .
this verification implies pairing computations .
however , in constraint environments such as the sim card , this is not feasible yet .
a straightforward solution would be to delegate these computations to the smartphone .
such a solution has two main drawbacks .
on the one hand , delegating computations to an untrusted environment , i.e. , the smartphone which could be compromised by spywares , will affect the user s privacy . on the other hand , verifying the received refund token would lead to inefficient transactions at exit turnstiles .
* sketch of proof . *
the * completeness * of the protocol follows by inspection .
the * soundness * follows from the extraction property of the underlying proof of knowledge and the unforgeability of the underlying signature scheme .
in particular , the extraction property implies that for any prover @xmath268 that convinces @xmath269 with probability @xmath270 , there exists an extractor which interacts with @xmath268 and outputs a witness @xmath271 with probability poly(@xmath270 ) .
moreover , if we assume that the extractor input consists of two transcripts , i.e. , \{@xmath272 @xmath273 @xmath274}. the witness can be obtained by computing ( all the computations are done mod p ) : @xmath275 and @xmath276 ; @xmath277 ; the extractor succeeds when @xmath278 is invertible in @xmath279 .
if the following check @xmath149 holds , this implies that : @xmath280 .
therefore , @xmath281 . let us denote by @xmath282 .
note that we necessarily have @xmath283 , otherwise this would imply that the prover knows the secret value @xmath124 ( which would be equal to @xmath284 ) .
we therefore have @xmath285 and consequently that @xmath188 .
so the prover knows a valid bb - signature on the value @xmath64 .
this implies that @xmath286 .
also note that if @xmath149 then this implies that the prover @xmath287 only knows one representation of @xmath148 in the base @xmath288 and @xmath97 .
otherwise this would imply that @xmath287 knows the private key @xmath124 . indeed , suppose that @xmath287 knows two representations of @xmath148 in the base @xmath288 and @xmath97 .
let us denote by @xmath289 and @xmath290 these two representations .
since @xmath85 is a prime order group and @xmath291 and @xmath292 , this implies that @xmath293 and @xmath294 . since @xmath295
, this implies that @xmath296 and that @xmath297 .
let us denote by @xmath298 . then @xmath299 .
since @xmath85 is a prime order group this implies that @xmath300 . if @xmath301 , then @xmath268 can be directly used to mount a weak - chosen attack against the bb - signature scheme with probability poly(@xmath270 ) of succeeding
. thus @xmath270 must be negligible .
finally , to prove * ( honest - verifier ) zero - knowledge * , we construct a simulator _ sim _ that will simulate all interactions with any ( honest verifier ) @xmath302 . 1 .
_ sim _ retrieves @xmath140 and @xmath143 from @xmath302 .
sim _ randomly chooses @xmath286 and @xmath193 and computes @xmath303 , @xmath304 and @xmath305 .
3 . _ sim _ randomly chooses @xmath306 and computes @xmath307 and @xmath308 .
sim _ outputs @xmath309 @xmath310 .
since @xmath85 is a prime order group , then the blinding is perfect in the first two steps and @xmath311 and @xmath302 s view of the protocol are statistically indistinguishable .
* sketch of proof . *
the * completeness * follows by inspection . *
soundness : * roughly speaking , the whole proof @xmath202 can be divided in three sub - proofs @xmath312 , @xmath313 and @xmath314 , where : @xmath315 @xmath316 @xmath317 if the verifier accepts the proof @xmath202 this means that : @xmath198 ( 1 ) and @xmath318 ( 2 ) . note that the proofs @xmath312 , @xmath313 and @xmath314 are classical variants of schnorr s proof of knowledge . using the `` extractors '' of these proofs of knowledge , we can retrieve @xmath319 . from ( 1 )
we have : @xmath320 .
this implies that @xmath281 .
let us denote by @xmath282 .
( note that @xmath321 , otherwise this would imply that the prover knows the secret value @xmath124 ) .
we therefore have @xmath285 and therefore that @xmath188 .
so the prover knows a valid bb - signature on the value @xmath64 .
this implies that @xmath322 $ ] . from ( 2 ) we have that @xmath323 .
this implies that @xmath324 .
let us denote by @xmath325 , this implies that @xmath326 ( note that @xmath327 , otherwise this would imply that the prover knows the secret value @xmath233 ) .
so @xmath328 .
the prover therefore knows a valid permission token ( @xmath329 . in conclusion
, the prover knows a permission token @xmath330 and a value @xmath64 such that @xmath331 , @xmath332 $ ] and @xmath333 is an el gamal encryption of @xmath334 . *
( honest - verifier ) zero - knowledge : * since @xmath312 , @xmath313 and @xmath314 are classical zkpk , we can easily construct simulators @xmath335 , @xmath336 and @xmath337 ( respectively ) for such proofs .
from these simulators , it is straightforward to construct a simulator @xmath338 that will simulate all interactions with any ( honest ) verifier @xmath302 . since @xmath85 is a prime order group , then the blinding would be perfect and the output of @xmath338 and @xmath302 s view of the protocol would be statistically indistinguishable .
we used javacard 2.2 for developing all the cryptographic computations into the sim card . in particular , the sim card has access to the functions handling elliptic curves operations .
the sim card handles requests from the validator using the ncf contactless interface .
the validator has been developed in java and scala .
the java software uses a native library for ec scalar multiplications and pairings .
this library depends on libgmp for big integers computations and benefits from its assembly optimizations .
furthermore , computations are distributed between threads ( at jvm level ) to benefit from the multi - core architecture of the pc .
the pc is an intel(r ) xeon(r ) with a e5 - 1620 cpu with 4 cores running at 3.70ghz under a 64-bit linux os .
the nfc reader of the validator is an omnikey 5321 dual interfaces .
the curve and pairing parameters that we used in the implementation of our m - ticketing system are : q = 8243401665430090752057404098378368203946728292799613002 4655912292889294264593 p = 8243401665430090752057404098378368203918016968090658713 6896645255465309139857 b = 5
* discrete logarithm ( dl ) . * for any multiplicative group @xmath91 of prime order @xmath88 , the dl assumption states that given a random generator @xmath92 and a candidate @xmath339 , it is hard to find an integer @xmath340 such that @xmath341 .
* one - more discrete logarithm ( omdl ) @xcite . * for any multiplicative group @xmath91 of prime order @xmath88 , the omdl assumption states that given a random generator @xmath92 , a challenge oracle that produces a random group element @xmath342 when queried and a discrete logarithm oracle ( with respect to @xmath97 ) , it is hard to find , after @xmath176 queries to the challenge oracle ( where @xmath176 is chosen by the adversary ) and at most @xmath343 queries to the discrete logarithm oracle , the discrete logarithms of all @xmath176 elements @xmath342 . *
q - strong diffie - hellman i ( q - sdh - i ) . * for any multiplicative group @xmath85 of prime order @xmath88 , the q - sdh - i assumption states that given three random generators ( @xmath344 , @xmath125 , @xmath345)@xmath346 , a value @xmath347 , an oracle @xmath348 that on input a value @xmath178 outputs a pair @xmath349 with @xmath350 , it is hard to output a new triplet @xmath351 , with @xmath352 such that @xmath353 has never been queried to @xmath348 .
* remark : * the triplet @xmath354 corresponds to a permission token of our m - ticketing protocol . in the sequel , we will call the oracle @xmath348 , a bb - signature oracle and the value @xmath353 the ( permission ) token secret . * forking lemma .
* we use the forking lemma @xcite in our proofs , to prove that an adversary @xmath52 is not able to produce a new valid m - ticket @xmath35 ( respectively a schnorr s signature @xmath355 , see figure 6 ) unless he knows all the underlying secrets @xmath356 ( respectively the secret @xmath357 ) .
@xmath359=(@xmath118 , @xmath360 , @xmath361 , @xmath362 , @xmath363 , @xmath364 , @xmath191 , @xmath200 , @xmath148 , @xmath365 , @xmath366 , @xmath367 , @xmath368 , @xmath369 , @xmath370 , @xmath371 , @xmath372 , @xmath373 , @xmath374 , @xmath375 , @xmath225 ) , @xmath345=@xmath376 , @xmath360 , @xmath361 , @xmath362 , @xmath363 , @xmath364 , @xmath191 , @xmath200 , @xmath148 , @xmath365 , @xmath366 , @xmath367 , @xmath368 , @xmath369 , @xmath370 , @xmath371 , @xmath372 , @xmath373 , @xmath374 , @xmath375 , @xmath225 ) and * sketch of proof .
* we assume that the challenger @xmath387 in the experiment @xmath388 @xmath389 receives a random instance ( @xmath390 , @xmath391 , @xmath383 , @xmath392 ) of the one - more dl problem where @xmath97 is a random generator in @xmath85 .
@xmath387 will run the adversary @xmath52 as a subroutine and act as @xmath52 s challenger in the non - frameability experiment . because @xmath387 is against the one - more dl assumption , he has access to a dl oracle .
@xmath387 picks three random values @xmath393 ,
@xmath394 and @xmath395 in @xmath279 and computes @xmath396 , @xmath397 and @xmath398 .
the other generators of the m - ticketing system are randomly chosen .
@xmath52 chooses the keys @xmath78 for the transport authority and @xmath44 for the revocation authority and @xmath387 chooses the key @xmath399 for the paillier encryption scheme .
@xmath387 is therefore able to construct the public parameters @xmath12 for @xmath52 and can answer to its requests in the following way : * @xmath55 requests : @xmath387 answers using his input of the one - more dl problem .
@xmath387 puts @xmath400= @xmath401= @xmath402 * @xmath58 requests : @xmath387 does not need to do anything . * @xmath59 requests : @xmath387 uses is dl oracle to give the corresponding @xmath403 to the adversary . * @xmath61 requests : @xmath387 first uses his input of the one - more dl problem to compute the commitment @xmath191 : @xmath387 puts @xmath191= @xmath404= @xmath405 .
if the protocol does not abort then we put @xmath217=@xmath406+@xmath407 and @xmath408 , where @xmath406 is unknown to @xmath387 and @xmath379 is provided by @xmath52 . in the random oracle model , @xmath387 is able to perfectly simulate the proof of knowledge @xmath312 as well as the schnorr s signature @xmath355 .
* @xmath63 requests : all the values involved in the computation of an m - ticket @xmath35 can be easily simulated by @xmath387 except @xmath363 and @xmath36 . to compute @xmath409 where @xmath410 is a random value chosen by @xmath387 , it proceeds as follows : @xmath363 = @xmath411 = @xmath412 .
as @xmath387 does not know the value @xmath217 it can not compute or simulate the value @xmath331 .
it will therefore choose a random value @xmath48 and define @xmath36 as @xmath48 . in the random oracle model , @xmath387 can simulate the proof of knowledge @xmath202 using standard techniques .
the simulation is not perfect since we replace the value @xmath36 by a random value @xmath48 .
however @xmath52 will not detect this change unless he can solve the q - ddhi problem . * @xmath68 requests : @xmath387 proceeds as in an @xmath65 request for each unused m - ticket .
now , we assume that the adversary @xmath52 manages to produce a valid m - ticket @xmath35 such that it breaks the non - frameability of our m - ticketing protocol and it mades @xmath176 requests to the @xmath59 oracle .
this means that this m - ticket has not been obtained on a @xmath65 query and the @xmath75 algorithm on @xmath35 outputs an identifier @xmath21 which is in @xmath53 along with a schnorr s signature @xmath355 that proves that this m - ticket comes from a permission token obtained by this user can be publicly verifiable , a wrong @xmath75 procedure is statistically negligible ( even for a powerful adversary ) . ] .
@xmath359=(@xmath118 , @xmath360 , @xmath361 , @xmath362 , @xmath363 , @xmath364 , @xmath191 , @xmath200 , @xmath148 , @xmath365 , @xmath366 , @xmath367 , @xmath368 , @xmath369 , @xmath370 , @xmath371 , @xmath372 , @xmath373 , @xmath374 , @xmath375 , @xmath225 ) , @xmath345=@xmath376 , @xmath360 , @xmath361 , @xmath362 , @xmath363 , @xmath364 , @xmath191 , @xmath200 , @xmath148 , @xmath365 , @xmath366 , @xmath367 , @xmath368 , @xmath369 , @xmath370 , @xmath371 , @xmath372 , @xmath373 , @xmath374 , @xmath375 , @xmath225 ) and @xmath377=(@xmath378 , @xmath379 , @xmath380 , @xmath381 , @xmath382 , @xmath383 , @xmath384 ) then it can produce two valid m - tickets @xmath358 and @xmath385 such that @xmath386 . using the technique of replay and the soundness property of the proof @xmath202 ,
one is able to extract all the secret values ( @xmath162 , @xmath217 , @xmath256 , @xmath139 , @xmath64 ) . *
first case : * the value @xmath217 corresponds to a permission token obtained during an @xmath61 on @xmath413 ( i.e. , @xmath334 is equal to the value @xmath414 produced during such a request ) .
@xmath387 outputs the @xmath176 values @xmath403 that comes from the requests to the dl oracle plus the value @xmath415 and so breaks the one - more dl assumption . *
second case : * the value @xmath217 does not correspond to a permission token obtained during an @xmath61 on @xmath413 ( meaning that all the values @xmath414 generated during such requests are different from @xmath334 ) . we know by the definition of the experiment that no @xmath59 oracle query ( and consequently no dl oracle query ) has been made on this identity
. therefore the public key @xmath400 corresponding to @xmath413 is in the one - more dl problem input .
it follows from the forking lemma that if @xmath52 is sufficiently efficient to produce such a signature @xmath355 , then there exists an algorithm a which can produce two schnorr s signatures with non - negligible probability . using the techniques of replay and soundness , @xmath387 is able to extract the private key @xmath357 used to generate the signature @xmath355 .
@xmath387 outputs the t values @xmath403 , coming from the requests to the dl oracle , plus
the value @xmath416 and so breaks the one - more dl assumption .
we prove the non - frameability under the q - ddhi and one - more discrete logarithm assumptions @xcite .
we use in fact the omdl assumption to get a better reduction , but the proof can also be done under the discrete logarithm assumption . as the q - ddhi assumption implies the dl one , we can conclude that our m - ticketing protocol is non - frameable , in the random oracle model , under the q - ddhi assumption .
* sketch of proof .
* let @xmath52 be an adversary who breaks the unforgeability requirement of our m - ticketing protocol with non - negligible probability .
we will construct an algorithm @xmath417 , using @xmath52 as an oracle , which breaks the q - sdh - i assumption .
@xmath417 receives on input from its challenger ( @xmath85 , @xmath344 , @xmath125 , @xmath418 @xmath345 , @xmath347 ) the public parameters of the q - sdh - i challenge and has access to a bb - signature oracle .
the other generators of the m - ticketing system are randomly chosen .
@xmath417 also chooses the keys for the paillier encryption scheme .
it sends @xmath419 and the public key of the paillier scheme to @xmath52 .
the private and public keys of the el gamal cryptosystem can be chosen either by @xmath52 or @xmath417 .
@xmath417 is therefore able to construct the public parameters @xmath12 for @xmath52 and can answer to its requests as follow : * @xmath55 requests : @xmath417 randomly chooses @xmath420 and computes @xmath421 . * @xmath58 requests : @xmath417 does not need to do anything .
* @xmath59 requests : @xmath417 gives @xmath357 to @xmath52 .
* @xmath62 requests : @xmath417 plays as follows : @xmath417 first receives a paillier encryption @xmath422 .
it decrypts it ( recall that @xmath417 chose the private key for the paillier encryption scheme ) and retrieve the corresponding plaintext @xmath378 .
it then queries the bb - signature oracle on @xmath217 where @xmath423 and @xmath379 is randomly chosen by @xmath417 .
the bb - signature oracle sends back a pair @xmath424 with @xmath350 , and @xmath417 transmits it to @xmath52 along with the value @xmath379 .
so @xmath417 perfectly simulates the @xmath62 for @xmath52 .
* @xmath63 requests : since @xmath417 knows the values of all the tokens @xmath330 that have been issued during @xmath62 requests , it can easily answer to @xmath65 queries .
the simulation of this oracle is perfect . *
@xmath68 requests : @xmath417 proceeds as in an @xmath65 request for each unused m - ticket .
* * type-1 forger * : an adversary that outputs a valid m - ticket @xmath35 which can not be linked to a registered user ( corrupted or not ) .
* * type-2 forger * : an adversary that outputs more valid m - tickets than he obtained .
we show that both forgers can be used to solve the q - sdh - i problem .
however , the reduction works differently for each forger type .
therefore , initially @xmath417 chooses a random bit @xmath425 that indicates its guess for the type of forgery that @xmath417 will output . * if @xmath426 = 1 * : eventually , @xmath417 outputs , with non - negligible probability , a valid m - ticket @xmath427 such that the algorithm @xmath75 , on the input @xmath428 , returns an unknown identifier @xmath429 ( i.e. , does not appear in @xmath26 ) . *
first case * : @xmath430 ( the plaintext encrypted in @xmath431 ) has been queried during an @xmath62 request but no corresponding signature @xmath355 has been produced .
this means that the adversary did not receive the value @xmath379 ( where @xmath432 ) from the @xmath62 oracle .
we know from the forking lemma and the proofs @xmath202 and @xmath312 that @xmath52 necessarily knows @xmath353 and @xmath379 .
since only @xmath433 has been revealed by this oracle during the @xmath62 , @xmath52 could be used to extract discrete logarithms . therefore under the dl assumption
, this first case could only occur with negligible probability . * second case * : @xmath431 is an encryption of a commitment @xmath430 of a value @xmath353 that has not been queried during a @xmath62 request
. therefore @xmath353 has not been queried to the bb - signature oracle either . using the forking lemma and the soundness property of the proof @xmath312 , @xmath417 is able to extract with non - negligible probability the secrets @xmath434 underlying the m - ticket @xmath428 .
@xmath417 outputs the triplet @xmath354 to its challenger of the q - sdh - i assumption and therefore breaks it . * if @xmath426 = 2 * : eventually , @xmath52 outputs , with non - negligible probability @xmath435 , @xmath436 valid m - tickets @xmath437 , where @xmath73 is the number of calls to the @xmath62 oracle , @xmath10 is the number of authorized m - tickets by token and @xmath438 . let us denote by ( @xmath378 , @xmath379 , @xmath383 , @xmath439 ) the @xmath73 token secrets submitted to the bb - signature oracle .
w.l.o.g , we may assume that all these values are different ( recall that a token secret @xmath217 is jointly computed by @xmath52 and @xmath417 ) .
we therefore have @xmath440 distinct token pairs @xmath441 with @xmath442 and @xmath443 .
let @xmath444 be the set containing all these token pairs and @xmath445 the m - ticket corresponding to the token pair @xmath446 . among the @xmath447 m - tickets output by @xmath52 ,
there is at least one m - ticket @xmath448 for which the corresponding token pair * ( @xmath449 , @xmath450 ) does not belong to @xmath444*. otherwise this would mean that two m - tickets among these @xmath447 m - tickets , e.g. , @xmath451 and @xmath452 , have the same token pair ( since @xmath453 ) .
let us denote by ( @xmath454 , @xmath455 ) ( respectively ( @xmath456 , @xmath457 ) ) the token pair corresponding to @xmath451 ( respectively @xmath452 ) .
therefore the serial number @xmath458 of @xmath451 would be equal to @xmath459 the one of @xmath452 : @xmath458= @xmath460= @xmath461 = @xmath459 .
this case can not occur since all duplicates ( i.e. m - tickets which have the same serial numbers ) are discarded in the experiment @xmath76 .
suppose now that * @xmath462*. since @xmath463 this implies that @xmath464 .
such a case will happen with negligible probability under the q - sdh assumption ( see theorem 2 ) .
therefore @xmath465 and consequently has not been queried to the bb - signature oracle ( @xmath52 is in fact a type-1 forger ) .
@xmath417 then picks a random m - ticket @xmath428 among the @xmath447 m - tickets output by @xmath52 . with probability @xmath466
, it has chosen @xmath448 . using the forking lemma and the soundness property of the proof @xmath202 , @xmath417 is able to extract with non - negligible probability the secrets @xmath434 underlying the m - ticket @xmath428 .
@xmath417 outputs the triplet @xmath354 and therefore breaks the q - sdh - i assumption with non - negligible probability @xmath467 . to complete the proof , we need to clarify why we discard duplicates in @xmath76 .
we consider that @xmath35 = @xmath468 and @xmath428 = @xmath469 are duplicates if their serial numbers are equal .
let us denote by @xmath470 ( respectively @xmath471 ) the token pair corresponding to the ticket @xmath35 ( respectively @xmath428 ) .
since @xmath36 = @xmath472 , we have @xmath473 = @xmath474 . _ case 2.2 _ : @xmath217 and @xmath353 @xmath100 @xmath477 .
this implies that @xmath64 and @xmath478 @xmath100 @xmath479 .
otherwise @xmath52 could be used to break the q - sdh assumption ( see the proof of theorem 2 ) . since
@xmath217 and @xmath353 have been randomly chosen ( they are jointly computed by @xmath52 and @xmath417 ) , the probability that @xmath473 = @xmath474 is negligible .
therefore case 2.2 . will occur with negligible probability either .
* sketch of proof .
* we prove the * unlinkability * under a slightly weaker model than the one presented in section [ unlink ] .
indeed , we consider a slightly different experiment in which the adversary can not query the revocation oracle @xmath480 . we rename this new requirement
* unlinkability**. we would like however to emphasize that the access to revocation functionalities will likely be carefully controlled in a real deployment of an m - ticketing system .
therefore , unlinkability * is a reasonable model to consider .
anyway , in order to satisfy the original unlinkability requirement , we just need to replace the elgamal encryption scheme , which only satisfies ind - cpa security , by an ind - cca2 encryption scheme .
it is well - known that by double encrypting the same message under an ind - cpa scheme and using a _ simulation sound proof of equality _ of plaintexts , we obtain an ind - cca2 scheme .
therefore , by using the double elgamal encryption scheme , which was proved ind - cca2 in @xcite , our m - ticketing protocol would satisfy the original unlinkability requirement .
let @xmath52 be an adversary who breaks the unlinkability requirement of our m - ticketing protocol .
we will consider that a query to the @xmath481 oracle on ( @xmath21 , _ view _ ) for the report of @xmath255 unused m - tickets is equivalent to @xmath255 queries to the @xmath65 on ( @xmath21 , _ view _ ) .
let @xmath130 = @xmath10 .
we will say that an adversary @xmath52 against our unlinkability experiment @xmath482 is a type-(@xmath483 , @xmath255 ) distinguisher , with @xmath484 and @xmath485 , if @xmath52 after having received its challenge ( from its @xmath482-challenger ) makes at most @xmath483 queries to the @xmath65 oracle on ( @xmath486 , @xmath487 ) and @xmath255 queries to the @xmath65 oracle on ( @xmath80 , @xmath488 ) .
we can remark that a type-(@xmath483 , @xmath255 ) distinguisher , with @xmath484 and @xmath485 , is also a type-(@xmath489 , @xmath489 ) distinguisher .
we may therefore , without loss of generality , restrict our attention to type-(@xmath489 , @xmath489 ) distinguishers . in the sequel , we will thus assume that @xmath52 is such an adversary .
more precisely , we will suppose that after receiving @xmath490 and @xmath491 , @xmath52 arbitrarily queries the @xmath55 , @xmath58 , @xmath59 , @xmath492 and @xmath61 oracles and then queries the @xmath481 oracle on ( @xmath486 , @xmath487 ) and ( @xmath80 , @xmath488 ) . we give a security proof as a sequence of games , using shoup s methodology @xcite
. we will only give a rather high - level description of the initial game ( game 0 ) and brief descriptions of the modifications between successive games .
the challenger @xmath387 randomly chooses @xmath97 , @xmath344 , @xmath125 , @xmath177 , @xmath155 , @xmath418 , @xmath345 , @xmath91 , @xmath190 nine generators of @xmath85 and @xmath126 , @xmath184 two generators of @xmath86 .
it also chooses the keys for the paillier encryption scheme as well as the ones for the transport authority and revocation authorities .
@xmath387 sends @xmath17 and @xmath78 to @xmath52 .
@xmath387 answers to @xmath52 s requests as follows : * @xmath55 requests : @xmath387 randomly chooses @xmath420 and computes @xmath493 = @xmath494 . * @xmath58 requests : @xmath387 does not need to do anything . * @xmath59 requests : @xmath387 gives @xmath357 to @xmath52 .
* @xmath61 requests : @xmath387 chooses a random value @xmath378 to obtain a valid permission token @xmath330 and uses @xmath357 to generate the signature @xmath355 . * @xmath63 requests : @xmath387 uses its permission token @xmath330 and a fresh index @xmath64 that has not been used in a previous query of @xmath65 on @xmath21 and _ view _ and computes a ticket @xmath35 = @xmath468 . * @xmath495 requests : @xmath387 computes the m - tickets @xmath35 based on the secret @xmath217 corresponding to the user identifier @xmath21 and gives them to @xmath52 . * @xmath68 requests : @xmath387 proceeds as in a @xmath65 request for each unused m - ticket .
the adversary chooses two honest users @xmath79 and @xmath80 and two indexes @xmath496 and @xmath455 .
@xmath387 runs the protocol @xmath497 with @xmath79 and @xmath80 and outputs the corresponding views , @xmath487 and @xmath488 , to @xmath52 .
it then generates two valid m - tickets : @xmath490 = @xmath498 for @xmath79 and @xmath491 = @xmath499 for @xmath80 and send them to @xmath52 .
the goal of @xmath52 is to guess the value of @xmath163 .
after having received its challenge , @xmath52 can queries the @xmath55 , @xmath58 , @xmath59 , @xmath61 , @xmath65 , @xmath492 and @xmath481 oracles but with the following restrictions : it can not query the @xmath59 oracle on @xmath79 or @xmath80 or the @xmath492 oracle on ( @xmath486 , @xmath487 ) or ( @xmath80 , @xmath488 ) and can not query the @xmath500 @xmath501 oracle on ( @xmath486 , @xmath487 ) or ( @xmath80 , @xmath488 ) if both users did not validate the same number of m - tickets ( otherwise it can easily win this game ) . at the end of the game , the adversary outputs a bit @xmath502 , its guess .
let @xmath503 be the event that @xmath504 in this game and @xmath505 the event that @xmath504 in game i. we have : @xmath506 - 1/2|$ ] = @xmath507 = @xmath508- 1/2|$ ] let @xmath509 denote the @xmath489 unused ( reported ) m - tickets of @xmath79 and @xmath510 = ( @xmath511 , @xmath512 , @xmath513 the ones of @xmath80 , where the @xmath514 s and the @xmath515 s belongs to @xmath516 .
* * game(0,b ) * : this is the same game as * game 0 * except that we replace the el gamal ciphertext @xmath517 by an encryption of a random value and then simulate the proof @xmath518 .
such a simulated proof can easily be done in the random oracle model using standard techniques . under the ddh assumption , @xmath52 can not detect this change . indeed , we can easily construct a ddh distinguisher @xmath209 with ddh - advantage satisfying : + @xmath506- pr[s_{(0,b)}]|$ ] @xmath519 @xmath520 ( 1 ) + where @xmath520 is the ddh - advantage of @xmath209 in solving the ddh problem . in the sequel
, we will use the simplified notation @xmath521 ( respectively @xmath522 ) to denote the ddh - advantage ( respectively the q - ddhi advantage ) of some efficient algorithm in solving the ddh ( respectively q - ddhi ) problem . * * game(0,1-b ) * : this is the same game as * game(0,b ) * except that we replace the el gamal ciphertext @xmath523 by an encryption of a random value and then simulate the proof @xmath524 . such a simulated proof can easily be done in the random oracle model using standard techniques . under the ddh assumption , @xmath52 can not detect this change . indeed , we can easily construct a ddh distinguisher d with ddh - advantage satisfying : + @xmath525- pr[s_{(0,1-b)}]|$ ] @xmath519 @xmath521 ( 2 ) * * game(0,0,b ) * : this is the same game as * game(0,1-b ) * except that we replace the serial number @xmath526 by a random value and then simulate the proof @xmath518 .
such a simulated proof can easily be done in the random oracle model using standard techniques . under the q - ddhi assumption , @xmath52 can not detect this change . indeed , we can easily construct a q - ddhi distinguisher @xmath209 with q ddhi - advantage satisfying : + @xmath527- pr[s_{(0,0,b)}]|$ ] @xmath519 @xmath522 ( 3 ) * * game(0,0,1-b ) * : this is the same game as * game(0,0,b ) * except that we replace the serial number @xmath528 by a random value and then simulate the proof @xmath524 .
such a simulated proof can easily be done in the random oracle model using standard techniques . under the q - ddhi assumption , @xmath52 can not detect this change . indeed , we can easily construct a q - ddhi distinguisher @xmath209 with q ddhi - advantage satisfying : + @xmath529- pr[s_{(0,0,1-b)}]|$ ] @xmath519 @xmath522 ( 4 ) combining ( 1 ) , ( 2 ) , ( 3 ) and ( 4 ) , we obtain : + @xmath506- pr[s_1]|$ ] @xmath519 @xmath530 + @xmath531 + we then proceed similarly for each unused m - ticket of @xmath79 and @xmath80 : @xmath509 and @xmath510 = ( @xmath511 , @xmath512 , @xmath513 . for @xmath532 to @xmath489
, we define the following game . * * game(i , b ) * : this is the same game as * game i * = * game(i-1,0,1-b ) * except that we replace the el gamal ciphertext @xmath533 by an encryption of a random value and then simulate the proof @xmath534 . *
* game(i,1-b ) * : this is the same game as * game(i , b ) * except that we replace the el gamal ciphertext @xmath535 by an encryption of a random value and then simulate the proof @xmath536 . *
* game(i,0,b ) * : this is the same game as * game(i,1-b ) * except that we replace the serial number @xmath537 by a random value and then simulate the proof @xmath538 .
* * game(i,0,1-b ) * : this is the same game as * game i,0,b ) * except that we replace the serial number @xmath539 by a random value and then simulate the proof @xmath536 . using our notation ,
* game m * is the same game as * game(m-1,0,1-b)*. in the latter game , the unlinkability - challenger gives no information ( in a strong information theoretic sense ) to @xmath52 about the bit @xmath163 ( since all the el gamal ciphertexts have been replaced by encryptions of random values and all serial numbers have been replaced by random values ) .
therefore we have : @xmath541 = 1/2 $ ] .
we can now give an upper bound for @xmath507 : + @xmath507 = @xmath508- 1/2|$ ] = @xmath506- pr[s_m]|$ ] + we have : + @xmath506- pr[s_m]|$ ] @xmath519 @xmath542- pr[s_{j+1}]|$ ] @xmath519 @xmath543 + @xmath544 + therefore under the ddh and q - ddhi assumptions , the advantage @xmath507 of a type-(@xmath489 , @xmath489 ) distinguisher is negligible ( and consequently the one of a type-(@xmath483 , @xmath255 ) distinguisher , with @xmath484 and @xmath485 , is also negligible ) . | to ensure the privacy of users in transport systems , researchers are working on new protocols providing the best security guarantees while respecting functional requirements of transport operators . in this paper[multiblock footnote
omitted ] , we design a secure nfc m - ticketing protocol for public transport that preserves users anonymity and prevents transport operators from tracing their customers trips . to this end , we introduce a new practical set - membership proof that does not require provers nor verifiers ( but in a specific scenario for verifiers ) to perform pairing computations .
it is therefore particularly suitable for our ( ticketing ) setting where provers hold sim / uicc cards that do not support such costly computations .
we also propose several optimizations of boneh - boyen type signature schemes , which are of independent interest , increasing their performance and efficiency during nfc transactions .
our m - ticketing protocol offers greater flexibility compared to previous solutions as it enables the post - payment and the off - line validation of m - tickets . by implementing a prototype using a standard nfc sim card ,
we show that it fulfils the stringent functional requirement imposed by transport operators whilst using strong security parameters . in particular , a validation can be completed in @xmath0 when the mobile is switched on , and in @xmath1 when the mobile is switched off or its battery is flat . |
dust grains absorb stellar ultraviolet ( uv)optical light and reprocess it into far - infrared ( fir ) , thereby affecting the energetics of interstellar medium ( ism ) ( e.g. * ? ? ?
the fir luminosity is known to be a good indicator of star formation rate ( sfr ) in galaxies @xcite .
this can be explained if a large part of dust grains are heated by uv light from massive stars .
the fir spectral energy distribution ( sed ) of dust grains reflects various information on the grains themselves and on the sources of grain heating ( e.g. * ? ? ?
* ; * ? ? ?
dust temperature , which can be estimated from wavelength at the peak of fir sed , is determined by the intensity of the interstellar radiation field ( isrf ) and the optical properties of dust ( i.e. how it absorbs and emits light ) .
since dust grains absorb uv light efficiently , the fir luminosity and the dust temperature mostly reflect the uv radiation field @xcite . the fir sed of the galaxy ( milky way ) the `` nearest '' galaxy has been investigated by various authors .
for example , @xcite and @xcite explain the galactic fir sed as well as the extinction curve .
@xcite provide a unified explanation for the galactic polarization and extinction curve by adopting grains composed of silicate cores and organic refractory mantles , small carbonaceous particles , and polycyclic aromatic hydrocarbons ( pahs ) .
although there are some variations for assumed species , three grain components are usually assumed : silicate , graphite and pahs . besides the grain composition and the isrf , the size distribution also affects the fir sed .
the temperature of silicate and graphite grains stays at an equilibrium temperature determined by the radiative equilibrium if the size is roughly larger than 0.01 @xmath6 m @xcite .
such large grains are called large grains ( lgs ) .
since the equilibrium temperature is usually @xmath730 k in galactic environments , the emission peak lies around wavelengths of @xmath8@xmath9 m . if grains are smaller , they are transiently ( stochastically ) heated to a large temperature and emit photons with wavelengths shorter than @xmath10 m @xcite .
those grains are called very small grains ( vsgs ) . the other component ,
pah , only contributes to wavelengths shorter than @xmath11 m .
since we are interested in wavelengths longer than @xmath12 @xmath6 m , we do not treat pahs in this paper .
so far , a major part of works have focused on fir emission from high galactic - latitude cirrus or from some limited regions in the galactic plane ( e.g. * ? ? ?
* ; * ? ? ?
because of the complication in the galactic plane , which has large inhomogeneity with a variety of isrf intensity and with a large range of gas ( dust ) density , etc . , one may expect that a unified understanding of fir sed in the galactic plane is impossible . however , ( * ? ? ?
* hereafter h06 ) show that the fir colour - colour relation ( 60100 @xmath6 m vs. 140100 @xmath6 m ) defined by data points with galactic latitudes of @xmath13 has a clear sequence .
a large fraction ( @xmath1495% ) of the data points are on a colour - colour relation which they call `` main correlation '' . a minor part ( @xmath1510% ) of points
belong to another sequence called `` sub - correlation '' ( see also * ? ? ?
* ; * ? ? ?
the sub - correlation tends to be found in regions with large radio intensity , which indicates a strong isrf .
the `` main correlation '' of the galactic plane is also consistent with the fir colours of a sample of nearby galaxies compiled by @xcite ( see fig . 8 of h06 ) .
in particular , the fir colour - colour relation ( 60100 @xmath6 m vs. 140100 @xmath6 m ) of the magellanic clouds is located at the extension of the main correlation of the galactic plane . as stated by h06
, this implies that the optical properties of grains in fir are common among the galaxy and the magellanic clouds .
h06 also show that the fir colours predicted by @xcite based on `` standard '' grain optical properties @xcite can not reproduce the main correlation . indeed
, some observational and experimental results show that the fir grain properties may be different from the `` standard '' ones . based on the data of the far - infrared absolute spectrophotometer ( firas ) aboard the _ cosmic background explorer _
( _ cobe _ ) , @xcite suggest that the galactic dust has more enhanced emissivity at @xmath16 m than expected before .
some isophot observations have also reached a similar conclusion @xcite , although those observation target dense regions , where the dust properties may be modified by grain coagulation to form fluffy aggregates and/or by coating of ice mantles .
recently , @xcite have shown that the extinction at @xmath17 band derived from the fir dust emission is higher than that derived from the optical digitized sky survey data , suggesting a previous underestimation of dust emissivity at fir wavelengths .
an enhanced fir emissivity is also derived for extragalactic objects @xcite .
@xcite find experimentally that some kinds of amorphous silicate show emissivities larger than those found in @xcite .
moreover , the wavelength dependence suggested by those authors is different from that presented by @xcite . therefore , it is worth reexamining the fir emissivity of dust by using the new observational results provided by h06 .
the aim of this paper is to further investigate the origin of the main correlation both observationally and theoretically .
since the same main correlation is defined for the magellanic clouds , we also treat the fir emission in these galaxies .
furthermore , we extend the analysis of h06 to high - galactic - latitude regions to examine if the universality of the main correlation found in the galactic plane holds over the entire galaxy . since the isrf in high galactic latitudes
is considered to be more uniform than in the galactic plane , the dependence on the galactic latitude provides us with a key to understand the effect of inhomogeneous isrf . a theoretical framework for the fir
sed developed by ( * ? ? ?
* hereafter dl01 ) is adopted for theoretical analysis of the main correlation .
the sub - correlation is also investigated .
this paper is organized as follows .
first , in section [ sec : model ] , we describe the models adopted to calculate the fir sed .
then , in section [ sec : data ] , we briefly review the observational results by h06 , presenting also some additional analyses . in section [ sec : results ] , the fir colour - colour relation predicted by our models is compared with the observational main correlation . here , the sub - correlation is also discussed .
section [ sec : sum ] is devoted to the conclusion .
we adopt the theoretical framework of dl01 to calculate the sed of dust emission .
since we are interested in a wavelength range of @xmath18140 @xmath6 m , we neglect pahs , which contribute to the emission at @xmath19 m ( @xcite ; dl01 ) .
our assumptions on the relevant quantities are described below .
we assume a grain to be spherical with a radius of @xmath34 .
the absorption cross section of the grain is expressed as @xmath35 , where @xmath36 is called absorption efficiency .
the optical constants of silicate and graphite grains are taken from @xcite . for the uv regime , we adopt `` smoothed astronomical silicate '' from @xcite for silicate grains . for graphite grains ,
two - thirds of the particles have an dielectric function @xmath37 and one - third @xmath38 ( see * ? ? ?
the number density of grains with sizes between @xmath34 and @xmath39 is denoted as @xmath40 , where the subscript @xmath41 denotes a grain species ( @xmath42 stands for silicate and graphite , respectively ) .
we assume a power - law form for @xmath43 : @xmath44 where @xmath45 and @xmath46 are the upper and lower cutoffs of grain size , respectively , and @xmath47 is the normalizing constant .
@xcite find that the extinction curve observed in the solar neighborhood is reproduced by a composite model of graphite and silicate grains with @xmath48 .
we assume @xmath49 and @xmath50 m for both graphites and silicates @xcite , although @xcite adopt more elaborate functional form .
we have confirmed that the difference does not affect the following discussion on the fir colours .
the mixing ratio between silicate and graphite is uncertain and different even among models for the galactic dust @xcite .
according to @xcite , the ratio between those two species is different among the galaxy , the large magellanic cloud ( lmc ) and the small magellanic cloud ( smc ) . in the smc ,
@xcite suggest a silicate - dominated dust composition , while @xcite argue against such a composition from a study of the depletion pattern ( but see * ? ? ?
considering such large uncertainty in the grain composition , we present the fir sed of each grain species separately to avoid the uncertainty in the mixing ratio between the species .
the effects of mixing two species are briefly mentioned in section [ subsec : mixing ] . as long as we treat the fir colours of a single species ( section [ subsec : colour ] ) , the constant @xmath47 cancels out .
thus , it is not necessary to determine the absolute abundance of graphite and silicate . the relative ratio of @xmath47 between species is important only when we consider a mixture of various species ( section [ subsec : mixing ] ) . in this case , @xmath47 is determined by @xmath51 where @xmath52 is the mass density of species @xmath41 in the ism , and @xmath53 is the material density of species @xmath41 .
we assume @xmath54 g @xmath55 and @xmath56 g @xmath55 @xcite .
in fact , we require only the ratio @xmath57 ( or @xmath58 ) to calculate the fir colors in a mixture of the two species .
we adopt multi - dimensional debye models for the heat capacity of grains .
the heat capacity of a graphite grain per unit volume is ( dl01 ; @xcite ) @xmath59\ , , \end{aligned}\ ] ] @xmath60 where @xmath61 @xmath55 @xcite is the number of carbon atoms contained in the unit volume of the grain , @xmath62 is the boltzmann constant , and @xmath63 is the grain temperature .
the heat capacity of a silicate grain per unit volume is ( dl01 ) @xmath64\ , , \end{aligned}\ ] ] where @xmath65 @xmath55 is the number of atoms contained in the unit volume of the grain @xcite .
the total fir intensity ( per frequency per hydrogen nucleus per solid angle ) of dust emission at a wavelength @xmath21 can be estimated as @xmath66 where @xmath41 specifies the grain species and @xmath67 is the temperature distribution function of the grains .
we assume that the fir radiation is optically thin .
thus , the observed colour directly reflects the ratio of the above intensity . here the colour at wavelengths of @xmath68 and @xmath69 is defined as @xmath70 , which is called @xmath71 colour .
the distribution function @xmath67 is calculated by the formalism of dl01 ( see their section 4 ) based on the quantities in sections [ subsec : isrf][subsec : cap ] .
the enthalpy of a grain is divided into discrete bins , and the transition probability between the bins is treated by taking into account the heating by interstellar radiation and the cooling by emission ( see also * ? ? ?
the data analyzed by h06 are adopted in this paper .
h06 used the zodi - subtracted mission average ( zsma ) taken by the diffuse infrared background experiment ( dirbe ) of the _ cobe_. they analyzed the data at three fir wavelengths : @xmath72 m , 100 @xmath6 m , and 140 @xmath6 m .
then , they examined the relation between two colours , @xmath73 and @xmath74 .
following h06 , we adopt the intensity after colour correction for the 100-@xmath6 m and 140-@xmath6 m bands . for the galactic plane , they treated regions with galactic latitudes of @xmath75 .
the fir data of the magellanic clouds are also analyzed . in order to avoid the uncertainty in the subtraction of the zodiacal component
, they only select the region with @xmath76 mjy sr@xmath23 .
the error of @xmath77 is less than 10% . in figure
[ fig : hibi_fig2]a , we show the distribution of the galactic - plane data in the colour - colour diagram .
h06 found strong correlations between the two colours .
more than 90% of the data lie on a strong correlation called main correlation , which can be fitted as ( h06 ) @xmath78 the main correlation also explains the fir colours of the lmc and the smc ( h06b ) .
there is another correlation sequence , called sub - correlation in h06 ( see also * ? ? ?
* ) : @xmath79 h06 also examined the dependence on the galactic longitude @xmath80 , showing that the fir colours tend to shift downwards along the main correlation as @xmath80 increases from 0@xmath81 to 180@xmath81 . since the isrf in the region toward the galactic centre tends to be higher than that in the opposite direction @xcite , the colour shift along the main correlation can be caused by the difference in isrf ; that is , the fir colours of dust grains irradiated by a high isrf tends to be situated in an upper part of the main correlation on the colour - colour diagram .
we extend our analysis to high galactic latitudes ( @xmath82 ) .
the procedure of the data analysis and the selection criterion for the intensity ( i.e.@xmath76 mjy sr@xmath23 ) are the same as h06 .
most of the data are taken from @xmath83 , since the criterion for the intensity is hard to be satisfied in higher galactic latitudes . in figure
[ fig : hibi_fig2]b , we show the results .
although the data distribution is shifted slightly downward compared with the galactic plane data , we clearly see the main correlation .
moreover , the sub - correlation is not clear in the high galactic latitudes .
this supports the idea of h06 that the sub - correlation is produced by the contamination of high - isrf regions , which tend to reside in the galactic plane .
the data presented in figure [ fig : hibi_fig2 ] indicate that the main correlation is robust against the change of galactic latitudes ( @xmath84 ) .
h06 also show that the fir colours shifts along the main correlation by the variation of galactic longitudes ( @xmath80 ) .
these facts mean that the main correlation is independent of the complexity of the galactic structure .
since the inhomogeneity of isrf should be smaller in high galactic latitudes than in the galactic plane , the robustness of the main correlation over a variety of @xmath84 suggests that a detailed modeling of galactic structures and of inhomogeneous isrf are not necessary to understand the main correlation . as mentioned in the previous subsection
, h06 argued that the main correlation is produced by a sequence of varying isrf : the fir colours tend to be located in an upper part of the main correlation for a high isrf .
however , this was only suggested from the longitude dependence but was not theoretically demonstrated .
thus , in section [ sec : results ] , we investigate how the fir colour - colour relation changes for various isrf intensities . by examining
if the observed colour - colour relation is explained or not , we will finally be able to test what kind of grain species is consistent with the observational fir colour - colour relation ( section [ subsec : fir_opt ] ) .
it is worth noting that the main correlation can also explain the fir colours in the magellanic clouds .
since the lmc has a face - on geometry , the inhomogeneity in the isrf in a line of sight may be small .
the fact that the fir colour - colour relation of the lmc shifts upward relative to that of the galaxy suggests that the location in the main correlation reflects the strength of isrf , since the lmc is generally believed to have a higher isrf than the galaxy ( e.g. * ? ? ?
* ; * ? ? ?
it is surprising that the fir colours of the lmc and the smc are located on the main correlation defined by the galactic dust grains , and it is interesting to give a unified understanding of the fir colours for those three galaxies .
we also tackle this problem in sections [ sec : results ] and [ sec : discussion ] .
h06 show that the sub - correlation can be reproduced by summing two colours in the main correlation .
thus , they argue that the main correlation is fundamental .
we also follow their argument and concentrate on explaining the main correlation .
the sub - correlation is investigated in section [ subsec : mixture ] .
from various observational evidence such as extinction curves and interstellar depletion patterns , dust is believed to be mainly composed of two species : silicate and carbonaceous dust @xcite .
for those two species , @xcite proposed astronomical silicate and graphite , which reproduce the observational extinction curve and fir emission spectrum ( see also * ? ? ?
based on the models described in section [ sec : model ] , we examine how the fir colours change by the variation of isrf ( @xmath32 ) . following h06 , we focus on the relation between the 60 @xmath6m100 @xmath6 m colour [ @xmath73 ] and the 140 @xmath6m100 @xmath6 m colour [ @xmath74 ] . in figures [ fig : clr_std ] , we plot the colour - colour relations of graphite and astronomical silicate for @xmath87 , 1 , 3 , 10 , and 30 .
the discrepancy between the observational main correlation and the theoretical prediction is significant .
some mechanism that shifts the colour - colour relation upward and/or rightward should be included to explain the observational colour correlations .
vs.@xmath88 ) relation for silicate and graphite grains ( large crosses and large squares , respectively ) .
the different points correspond to @xmath87 , 1 , 3 , 10 , and 30 from low to high @xmath88 colours .
the same contour as shown in figure [ fig : hibi_fig2]a as well as the main and sub - correlations ( dashed and dotted lines , respectively ) are also presented .
[ fig : clr_std],width=302 ] it should be noticed that both silicate and graphite grains trace similar tracks on the fir colour - colour relation although the uv absorption coefficients and heat capacities are different between those two species .
this strongly suggests that the track on the fir colour - colour diagram is determined mainly by the fir absorption coefficients of grains .
indeed , both have dependence roughly described as @xmath89 .
thus , we suspect that the fir colour - colour relation is not sensitive to various factors other than the wavelength dependence of the fir absorption coefficient . the colour - colour relation shifts upwards if the contribution from vsgs relative to that of lgs increases , since the emission from stochastically heated vsgs contributes significant to the emission at @xmath72 m .
indeed , grains with @xmath90 contribute to the 60 @xmath6 m intensity @xcite .
a change of the isrf spectrum and/or of the grain size distribution can change the relative contribution from the stochastic vsg emission .
first we vary the isrf spectrum .
a hard spectrum could relatively enhance the emission from vsgs , since the cross section of a vsg is sensitive to the change of wavelength in the uv regime ( e.g. * ? ? ?
we calculate the fir colours by assuming a `` hard '' spectrum without the optical nir bump ; i.e.we adopt @xmath91 for @xmath92 instead of the expression in equation ( [ eq : isrf ] ) .
the fir colour - colour relation calculated with this hard spectrum is shown in figure [ fig : clr_enhanced60]a .
the colour - colour relation shifts upward and approaches the main correlation compared with the relation in figure [ fig : clr_std ] .
this is due to relative enhancement of the stochastic heating of vsgs , which contributes to the 60 @xmath6 m emission .
however , the change of the fir colours for various @xmath32 does not follow the main correlation .
moreover , such a hard spectrum as assumed here is unrealistic for high galactic latitude , where the main correlation is also seen .
thus , we conclude that the change of the uv spectrum can not explain the observational fir colours . the enhancement of the stochastic heating occurs also if the relative number of vsgs increases .
thus , we expect that the colour - colour relation shifts upward if we assume a grain size distribution with an enhanced number of vsgs . the number of vsgs may be increased by shattering @xcite .
the size distribution shown by @xcite can be roughly fitted by @xmath93 ( i.e. @xmath94 in equation [ eq : size_dist ] ) .
we also examine @xmath95 as an intermediate case . in both cases ,
we adopt the same values of @xmath45 and @xmath46 as assumed in section [ subsec : size ] . in figure
[ fig : clr_enhanced60]b , we show the colour - colour relation for @xmath95 and 4.0 .
we observe that the 60 @xmath6m100 @xmath6 m colour is sensitive to the grain size distribution .
however , the overall trend of the main correlation is not reproduced neither by @xmath95 nor by @xmath94 .
the 60 @xmath6m100 @xmath6 m colour is clearly out of the observed range for @xmath94 .
although @xmath95 explains the most concentrated part of the data with @xmath961 , the trend along the main correlation is not reproduced . the upper and lower cutoffs of grain size ( @xmath46 and @xmath45 ) can also be changed . however , the 140 @xmath6m100 @xmath6 m colour is not sensitive to the change of @xmath45 because it is determined by the equilibrium temperature of large grains .
it is not largely affected by the change of @xmath46 as long as @xmath97 m .
the 60 @xmath6m100 @xmath6 m colour monotonically increases if the fraction of vsgs increases . according to @xcite , grains with @xmath90
contribute to the 60 @xmath6 m flux , while a large fraction of the 100 @xmath6 m flux comes from grains with @xmath98 .
thus , as @xmath45 decreases and/or @xmath46 decreases , the 60 @xmath6m100 @xmath6 m colour monotonically increases . however , this colour is not sensitive to the grain size distribution at @xmath99 , and the abundance of such very small particles is constrained better in mid - infrared ( e.g. * ? ? ?
we have confirmed after calculations with our models that resulting colour - colour relations predicted by decreased @xmath45 and/or @xmath46 are similar to those in figure [ fig : clr_enhanced60 ] and that the observed main correlation can never be explained by changes of @xmath45 and @xmath46 . in summary , it is difficult to explain the overall main correlation by simply enhancing the stochastic heating component , which contributes to the 60 @xmath6 m intensity .
heat capacities and grain optical constants are also concerned with the fir sed .
however , as shown in appendix [ app : mat ] , difference in the heat capacity does not change the fir colour - colour relation .
there , we also show that different materials with different uv near - infrared absorption coefficients does not explain the observational fir colour - colour relation .
we note that all the materials examined in appendix [ app : mat ] have the fir spectral index ( which is defined as @xmath100 , i.e. @xmath101 is approximated to be proportional to @xmath102 ) nearly equal to 2 .
we are left with changing the fir optical properties of grains .
the following subsection is devoted to examining the possibility that the fir spectral index is different from 2 .
a spectral index @xmath103 is the most frequently used value , which is supported by some experiments and observations ( e.g. * ? ? ?
* ; * ? ? ?
on the other hand , an analysis of the firas data by @xcite indicates that the galactic dust has @xmath104 .
some other observational and experimental results also indicate that the fir emissivity index may be smaller than 2 @xcite . ) , but even if the colour is corrected by assuming @xmath105 , the difference in the colours is less than 10% .
thus , the prediction with @xmath106 can be compared with the results of h06 within an uncertainty of @xmath107% . ]
the range of @xmath108 is roughly from @xmath109 to 1.6 as suggested for the fir colours of the galaxy and the magellanic clouds @xcite .
some species indeed show such a low emissivity index @xcite : @xmath110 for @xmath111 ( amorphous ) .
since there is no available information on the optical constants of @xmath111 in the wavelengths shorter than fir , we can not include this species consistently into our models .
fortunately , the resulting fir colours are not sensitive to the detailed wavelength dependence of the uv absorption coefficient as shown in appendix [ app : mat ] .
thus , we adopt the emissivity of astronomical silicate or graphite at @xmath112 m and change the spectral index at @xmath113 m .
that is , we choose the following absorption efficiency for @xmath114 : @xmath115 where we adopt the value @xmath116 from @xcite for astronomical silicate and graphite . in figure
[ fig : clr_beta ] , we show the colour - colour relation for various fir spectral indexes . here
we call the species according to the adopted optical properties at @xmath117 although we modify artificially @xmath101 at @xmath114 .
we observe from figure [ fig : clr_beta ] that @xmath106 is roughly consistent with the main correlation if the isrf intensity is between @xmath118 and @xmath119 ( @xmath120 may also be permitted considering the uncertainty in the data ) .
it is worth noting that the colour - colour relation is just along the main correlation .
moreover , the region of the largest concentration of the data ( see the contours in fig .
[ fig : clr_beta ] ) is reproduced with a standard isrf @xmath121 .
thus , we conclude that a fir emissivity index of @xmath1181.5 is successful in reproducing the main correlation . for the absorption efficiency at @xmath113 m
, we also examine a hybrid functional form proposed by @xcite : @xmath122^{1/6}}\ , , \label{eq : reach}\end{aligned}\ ] ] which behaves like equation ( [ eq : beta ] ) with @xmath106 for @xmath123 and @xmath103 at @xmath124 ( see also * ? ? ?
we assume that @xmath125 m , following @xcite . for @xmath112 m
, we adopt the same absorption efficiency of astronomical silicate and graphite as used in section [ subsec : std ] .
thus , we set the value of @xmath126 so that the continuity of @xmath101 at @xmath127 m is satisfied . in figure
[ fig : clr_reach ] , we show the colour - colour relation for this absorption efficiency .
the resulting relation is similar to that in figure [ fig : clr_beta ] with @xmath106 .
thus , the 60 @xmath6m100 @xmath6 m vs.140 @xmath6m100 @xmath6 m colour - colour relation is determined by the spectral index around @xmath128m140 @xmath6 m and is not sensitive to the details of the absorption coefficient at the other wavelengths . in reality , both silicate and graphite contribute to the fir sed at the same time .
thus , it is worth examining the fir colour - colour relation predicted by mixtures of those two species .
as stated in section [ subsec : size ] , there is a large uncertainty in the mixing ratio among models .
thus , we examine a variety of mixing ratios . in figure
[ fig : clr_beta]c , we show the fir colour - colour relation calculated by a graphite - silicate mixture .
first , the mass ratio of those two species is assumed to be @xmath129 according to @xcite . with this ratio
, graphite dominates the fir sed and the fir colour - colour relation becomes almost the same as that of graphite ( see figure [ fig : clr_beta]b ) .
a similar graphite - dominated fir sed can be seen in @xcite .
as the fraction of silicate becomes larger , the fir colour - colour relation approaches that shown in figure [ fig : clr_beta]a . in order to show this
, we also examine a silicate - dominated composition as proposed for the lmc .
we adopt a mass ratio of @xmath130 .
the result is shown in figure [ fig : clr_beta]d , which becomes relatively similar to the silicate colours shown in figure [ fig : clr_beta]a .
however , it is evident that the colours of the mixed population are between those of single species .
we have found in the previous section that the fir colours are sensitive to the fir absorption coefficient of dust grains and are robust to the absorption coefficient in wavelengths other than the fir .
then we have suggested that the fir emissivity index of @xmath3 fits the main correlation .
an emissivity index of @xmath131 can also explain the submillimetre part of the galactic fir sed without introducing a very cold component @xcite.m intensity is larger than 3 mjy sr@xmath23 , the contamination of such a background is negligible . ] indeed it would be possible to explain the fir colour by mixing two components with different dust temperatures , but the robustness of the main correlation against the galactic latitudes would strongly require that the mixing ratio between those two components be uniform over the entire galaxy .
more natural interpretation is obviously that the main correlation reflects the fir absorption coefficient of dust grains itself .
thus , it is important to investigate what kind of the fir absorption coefficients reproduces the observational main correlation .
but for the fir emissivity shown by equation [ eq : reach ] .
[ fig : clr_reach],width=302 ] we note that we can not strongly constrain the spectral index at wavelength out of the range treated in this paper and that we do not reject the possibility that the spectral index varies as a function of wavelength as we examined in equation ( [ eq : reach ] ) . in most of the previous works on the galactic fir spectra , within the uncertainty ,
the fir sed at @xmath132 was fitted quite well both by @xmath103 and by @xmath106 with different dust temperatures .
however , since the analysis adopted in this paper ( h06 ) only uses the pixels with negligible uncertainty ( @xmath133 ) , we are able to constrain the fir dust emissivity index more strongly than previous works , based on the strong correlation called main correlation . also for extragalactic objects ,
the fir spectral index is a matter of debate .
submillimetre ( sub - mm ) seds of dust - rich galaxies such as spiral galaxies usually show an excess in the sub - mm regime relative to a single component fit to the fir regime .
this is usually interpreted as existence of very cold dust contributing to the sub - mm emission ( e.g. * ? ? ?
however , it is also important to stress that the discussion on the amount of very cold dust strongly depends on the assumed fir sub - mm spectral index .
the excess of the sub - mm emission could be explained by a small spectral index ( @xmath108 ) .
@xcite show that the fir sub - mm seds of some galaxies can be fitted with @xmath1341.5 , similar to our derived values . as shown by h06 ,
the main correlation is also consistent with the fir colours of nearby galaxies .
this implies that investigating the origin of the main correlation is of a general significance to the interpretation of the fir colours in extragalactic objects .
h06 propose that the sub - correlation shown by the dotted line in figure [ fig : hibi_fig2 ] can be explained by mixing two colours in the main correlation .
such a mixture of fir colours in a line of sight is called overlap effect in h06 .
h06 also show observationally that the sub - correlation tends to be associated with lines of sight with a high isrf intensity indicated by a strong radio continuum intensity .
thus , the sub - correlation may be reproduced by a contamination of a fir colour with a high radiation field intensity . in order to investigate the overlap effect in our framework
, we assume a two - component model , in which two fir colours calculated with two different radiation field intensities ( @xmath32 ) are mixed .
such a superposition of fir colours is treated by @xcite in a more sophisticated manner , but our simple approach has an advantage that the physical quantities are easy to interpret .
we denote the fir intensity predicted under an isrf intensity of @xmath32 as @xmath135 .
then , we examine a two - component model in which the fir intensity is composed of a mixture of two isrf @xmath136 and @xmath137 @xmath138 , @xmath139 : @xmath140 where @xmath141 is the contributing fraction of the component with @xmath142 in the line of sight . since the fir emission is optically thin , the contribution from each component is proportional to the dust optical depth .
thus , if the total dust optical depth at wavelength @xmath21 is @xmath143 , @xmath144 is occupied with a component with @xmath145 .
we adopt the fir absorption efficiency expressed by equation ( [ eq : beta ] ) with @xmath106 , since it explains the main correlation ( section [ subsec : emissivity ] ; figure [ fig : clr_beta]b ) .
we adopt the graphite species for the absorption efficiency at @xmath146 m . if we adopt the silicate absorption efficiency at @xmath146 m , we obtain similar results , and the following discussions are valid .
we fix @xmath147 as a general isrf and examine various @xmath148 and @xmath141 as a contaminating high radiation field . in figure
[ fig : highisrf ] , we show the results .
we find that the main correlation shifts rightward ( i.e. toward the sub - correlation ) on the colour - colour diagram if we mix two colours with different @xmath32 . from figure
[ fig : highisrf ] , if about 10% of the total column of a line of sight is filled with a high-@xmath32 region , the sub - correlation is reproduced . in the following , we discuss the filling factor assuming that the high radiation field originates from young massive stars .
but for the two - component models with @xmath147 and @xmath149 , 10 , and 30 ( lower to upper points ) .
the filling factors of the second component ( @xmath141 ) are assumed to be 0.3 , 0.1 , and 0.03 for the crosses , filled squares , and open squares , respectively .
[ fig : highisrf],width=302 ] we consider typical star - forming regions containing ob stars as regions with a high radiation field .
@xcite show that a typical galactic h ii region emit ionizing photons at a rate of @xmath150 s@xmath23 .
if we assume that the mean energy of the ionizing photons is 15 ev @xcite , the luminosity of the ionizing photons from the star - forming region becomes @xmath151 . assuming that the luminosity of nonionizing photons is 1.5 times as large as that of the ionizing uv photons @xcite
, we obtain the nonionizing uv luminosity of the star - forming region , @xmath152 , as @xmath153 .
we assume that the nonionizing photons contribute to the dust heating @xcite and treat the whole stellar cluster as a point source .
then , the uv radiation flux ( @xmath154 ) at a distance of @xmath155 from the centre of the star - forming region can be estimated as @xmath156 where @xmath157 is the optical depth of dust grains in the uv regime .
on the other hand , integrating equation ( [ eq : isrf ] ) over the uv regime ( @xmath1584000 ) , we obtain the uv flux as @xmath159 erg s@xmath23 @xmath22 ( we used @xmath160 for @xmath161 ) . equating this with equation ( [ eq : f_uv ] )
, we obtain @xmath162 the uv optical depth ( @xmath157 ) is related to the uv extinction in units of magnitude ( @xmath163 ) as @xmath164 .
if we take @xmath165 as a representative value of the uv wavelengths @xcite , we obtain @xmath166 for the galactic extinction curve @xcite , where @xmath167 is the colour excess defined between the @xmath168 and @xmath17 bands ( @xmath169 ) , and @xmath170 ( @xmath171 is the extinction at a wavelength of @xmath21 ) . the ratio between the hydrogen column density ( both in atomic and molecular form ) and @xmath167
is estimated as @xmath172 @xmath22 mag@xmath23 in the galactic environment @xcite . combining the above expressions
, we finally obtain the relation between @xmath157 and @xmath173 as @xmath174 .
@xcite assume that the number density of hydrogen nuclei in a star - forming region is typically @xmath175 @xmath55 .
adopting the same number density , we estimate that @xmath176 ( with @xmath175 @xmath55 ) around a star - forming region . in figure
[ fig : isrf_r ] , we show the isrf as a function of @xmath155 .
the sharp decline at @xmath177 pc is due to the dust extinction .
although @xcite focused on h ii regions with typical stellar ages less than @xmath178 yr , stars with lifetimes of @xmath179 yr ( b - type stars ) also emit uv radiation ( e.g. * ? ? ?
such uv emitting stars with long lifetimes , after dispersal of dense regions after @xmath178 yr , may be found in the field where the ism density is roughly @xmath118 cm@xmath23 .
thus , we also investigate the radiation field with @xmath181 @xmath55 . in this case
, the radiation field declines as @xmath182 and the dust extinction is negligible up to @xmath183 pc as shown in figure [ fig : isrf_r ] .
, as a function of the distance from the center of the star - forming region , @xmath155 .
the solid and dotted lines correspond to the results with @xmath175 @xmath55 and 1 @xmath55 , respectively.[fig : isrf_r],width=302 ] h06 have shown that the fir colour - colour relation follows the sub - correlation in @xmath1510% of the lines of sight in the galactic plane .
the probability ( denoted as @xmath184 ) that an arbitrary line of sight along the galactic plane passes through a region with a radius of @xmath155 can be estimated by @xmath185 , where @xmath186 is the path length , @xmath187 is the scale height of the galactic disk , and @xmath188 is the number of the regions in the entire galaxy .
taking @xmath189 , we obtain @xmath190 this probability is possibly dependent on the line of sight through the density structure in the galactic plane such as spiral arms , vertical stratification , etc .
however , the above probability is valid if we take a large number of lines of sight in the galactic plane as h06 have done .
the estimate of @xmath188 in terms of the galactic sfr ( or ob star luminosity ) is presented in appendix [ app : sfr ] , and the typical sfr ( or ob star luminosity ) of the galaxy gives @xmath191 .
the above probability estimated in equation ( [ eq : prob ] ) is comparable to the fraction of sightlines with the sub - correlation if @xmath192 pc .
if the sub - correlation is associated with regions near massive stars , the sub - correlation passes through a region whose typical distance from young stars is 5 pc . from figure
[ fig : isrf_r ] , the radiation field at @xmath192 pc from massive stars is typically @xmath193100 .
thus , in order to explain the sub - correlation , it would be necessary to consider a mixture of two fir colours : one with @xmath121 as a normal diffuse isrf and the other with @xmath193100 as a radiation field near massive stars .
this is roughly consistent with the mixture of the two components examined above and in figure [ fig : highisrf ] .
the observational main correlation between the 60 @xmath6m100 @xmath6 m colour and the 140 @xmath6m100@xmath6 m colour found by h06 is not explained by the previous models .
we have examined changes of quantities concerning the fir sed .
the most concentrated region in the colour - colour diagram and the trend along the main correlation can not be explained neither by the change of the isrf spectrum nor by the the variation of the grain size distribution .
the fir colour - colour is not very sensitive to the heat capacity and the detailed wavelength dependence of the uv absorption coefficient . on the other hand , we find that the fir colour - colour relation is sensitive to the fir absorption coefficient , in particular to the emissivity index @xmath108 defined in equation ( [ eq : beta ] ) .
the observational main correlation is consistent with the theoretical fir colours if we assume that @xmath1341.5 around 100200 @xmath6 m .
this is different from the spectral index often assumed for astronomical silicate and graphite species ( @xmath194 ) .
indeed some observational and experimental works suggest that the fir spectral index is significantly smaller than 2 and may be nearly 1 ( e.g. * ? ? ?
* ; * ? ? ?
the difference in the fir emissivity affects the estimate of dust mass and of dust temperature .
thus , we should further examine the fir dust properties not only of the galaxy but also of extragalactic objects by reexamining fir data .
new data taken by the _ spitzer space telescope _ @xcite and the _ akari _ @xcite , both of which have a fir band with @xmath113 m , are suitable for such a purpose .
we have also considered the origin of the sub - correlation by taking into account the overlap effect ; that is , the fir radiation comes from regions with a variety of uv radiation field intensity . for simplicity ,
we assume a mixture of two components illuminated by different uv radiation fields : one is a normal diffuse component with @xmath121 and the other is an intense field near to young massive stars ( @xmath193 ) . as a result
, we have found that the sub - correlation can be understood by the fir colour with high @xmath32 contaminated by that with normal @xmath195 .
this explains the observational fact that the sub - correlation is associated with the regions with intense radiation field ( h06 ) .
indeed , we have also shown that the observed fraction of the data points on the sub - correlation is consistent with the total luminosity ( or number ) of ob stars in the galaxy .
we are grateful to s. bianchi , r. schneider , i. sakon , y. okada , t. ootsubo , m. kawada , and the anonymous referee for their helpful comments .
hh has been supported by grants - in - aid for scientific research of the ministry of education , culture , sports , science and technology ( nos .
18026002 and 18740097 ) .
we fully utilized the nasa s astrophysics data system abstract service ( ads ) .
99 agladze , n. i. , sievers , a. j. , jones , s. a. , burlitch , j. m. , & beckwith , s. v. w. 1996 , apj , 462 , 1026 aguirre , j. e. , et al .
2003 , apj , 596 , 273 bianchi , s. , davies , j. i. , & alton , p. b. 1999 , a&a , 344 , l1 buat , v. , & xu , c. 1996 , a&a , 306 , 61 bohlin , r. c. , savage , b. d. , & drake , j. f. 1978 , apj , 224 , 132 dasyra , k. m. , xilouris , e. m. , misiriotis , a. , & kylafis , n. d. 2005 , a&a , 437 , 447 del burgo , c. , laureijs , r. j. , brahm , p. , & kiss , cs .
2003 , mnras , 346 , 403 dsert , f .- x . ,
boulanger , f. , & puget , j. l. 1990 , a&a , 237 , 215 dobashi , k. , uehara , h. , kandori , r. , sakurai , t. , kaiden , m. , umemoto , t. , & sato , f. 2005 , pasj , 57 , s1 dopita , m. a. , et al .
2005 , apj , 619 , 755 dorschner , j. , begemann , b. , henning , th . ,
jger , c. , & mutschke , h. 1995 , a&a , 300 , 503 draine , b. t. , & anderson , n. 1985 , apj , 292 , 494 draine , b. t. , & lee , h. m. 1984 , apj , 285 , 89 draine , b. t. , & li , a. 2001 , apj , 551 , 807 ( dl01 ) draine , b. t. , & li , a. 2007 , apj , 657 , 810 dwek , e. , et al .
1997 , apj , 475 , 565 edo , o. 1983 , phd thesis , university of arizona fukui , y. , mizuno , n. , yamaguchi , r. , mizuno , a. , & onishi , t. 2001 , pasj , 53 , l41 guhathakurta , p. , & draine , b. t. 1989 , apj , 345 , 230 hibi , y. 2006 , phd thesis , nagoya university hibi , y. , shibai , h. , kawada , m. , ootsubo , t. , & hirashita , h. 2006 , pasj , 58 , 509 hildebrand , r. h. 1983 , qjras , 24 , 267 hirashita , h. , buat , v. , & inoue , a. k. 2003 , a&a , 410 , 83 hirashita , h. , & ferrara , a. 2002 , mnras , 337 , 921 hirashita , h. , inoue , a. k. , kamaya , h. , & shibai , h. 2001 , a&a , 366 , 83 hirashita , h. , nozawa , t. , kozasa , t. , ishii , t. t. , & takeuchi , t. t. 2005 , mnras , 357 , 1077 huffman , d. r. , & stapp , j. l. 1973 , in greenberg j. m. , van de hulst h. c. , eds .
, iau symp .
52 , interstellar dust and related topics , reidel , dordrecht , p. 297
iglesias - pramo , j. , buat , v. , donas , j. , boselli , a. , & milliard , b. 2004 , a&a , 419 , 109 inoue , a. k. , hirashita , h. , & kamaya , h. 2000 , pasj , 52 , 539 jones , a. p. 2000 , j. geophys .
res . , 105 , 10257 jones , a. p. , tielens , a. g. g. m. , & hollenbach , d. j. 1996 , apj , 469 , 740 kennicutt , r. c. , jr . 1998 , ara&a , 36 , 189 kiuchi , g. , ohta , k. , sawicki , m. , & allen , m. 2004 , aj , 128 , 2743 lagache , g. , abergel , a. , boulanger , f. , & puget , j .- l .
1998 , a&a , 333 , 709 li , a. , & draine , b. t. 2001 , apj , 554 , 778 li , a. , & greenberg , j. m. 1997 , a&a , 323 , 566 li , a. , misselt , k. a. , & wang , y. j. 2006 , apj , 640 , l151 mathis , j. s. 1990 , ara&a , 28 , 37 mathis , j. s. , mezger , p. g. , & panagia , n. 1983 , a&a , 128 , 212 mathis , j. s. , rumpl , w. , & nordsieck , k. h. 1977 , apj , 217 , 425 matsuhara , h. , et al .
2006 , pasj , 58 , 673 murakami , h. 2004 , proc .
spie , 5487 , 330 nagata , h. , shibai , h. , takeuchi , t. t. , & onaka , t. 2002 , pasj , 54 , 695 onaka , t. , tokura , d. , sakon , i. , tajiri , y. y. , takagi , t. , & shibai , h. 2007 , apj , 654 , 844 pei , y. c 1992 , apj , 395 , 130 okada , y. , onaka , t. , nakagawa , t. , shibai , h. , tomono , d. , & yui , y. y. 2006 , apj , 640 , 383 reach , w. t. , et al .
1995 , apj , 451 , 188 sakon , i. , onaka , t. , ishihara , d. , ootsubo , t. , yamamura , i. , tanab , t. , & roellig , t. l. 2004 , apj , 609 , 203 sakon , i. , et al .
2007 , apj , 651 , 174 shibai , h. 2004 , advances in space research , 34 , 589 smith , l. f. , biermann , p. , & mezger , p. g. 1978 , a&a , 66 , 65 stevens , j. a. , amure , m. , & gear , w. k. 2005 , mnras , 357 , 361 takagi , t. , vanseviius , v. , & arimoto , n. 2003 , pasj , 55 , 385 takeuchi , t. t. , hirashita , h. , ishii , t. t. , hunt , l. k. , & ferrara , a. 2003 , mnras , 343 , 839 takeuchi , t. t. , ishii , t. t. , nozawa , t. , kozasa , t. , & hirashita , h. 2005 , mnras , 362 , 592 tumlinson , j. , et al .
2002 , apj , 566 , 857 weingartner , j. c. , & draine , b. t. 2001 , apj , 548 , 296 welty , e. w. , lauroesch , j. t. , blades , j. c. , hobbs , l. m. , & york , d. g. 2001 , apj , 554 , l75 werner , m. w. , et al .
2004 , apjs , 154 , 1 wright , e. l. , et al .
1991 , apj , 381 , 200
among silicate species , olivine and pyroxene are often considered as promising candidates of astronomical grains ( e.g. * ? ? ?
since the optical constants are different among species , we examine the robustness of the predicted fir colour - colour relation against the change of materials . as for olivine ,
the optical constants of two compositions are available : mg@xmath196fe@xmath197sio@xmath198 with @xmath199 and 0.5 for @xmath200500 @xmath6 m @xcite .
the optical constants at @xmath201 m do not affect our results because of small absorption efficiency , but those at @xmath202 m may be important , since the absorption of uv radiation determines the dust temperature . for both @xmath199 and 0.5 , we adopt the optical constants of olivine derived by @xcite at @xmath202 .
we calculate the fir colour - colour ( @xmath203 vs. @xmath204 ) relation by adopting the above optical constants .
the other quantities adopted are the same as those in the text . as a result
, we find that the fir colour - colour relation of the above materials is the same as that of astronomical silicate within the error of the colour - colour relation .
thus , olivine is not successful in reproducing the main correlation . if we adopt the optical constants of pyroxene found in @xcite ( but we use the olivine optical constants in @xcite for @xmath205 m ) , the main correlation is not reproduced .
thus , as far as we examined here , we conclude that silicate does not explain the observational fir colour - colour relation .
we have examined other materials ( feo , fe@xmath206o@xmath198 , sio@xmath207 , and mgo ) that can be included in silicates by using the optical constants adopted in @xcite .
we also investigate amorphous carbon instead of graphite .
the optical constants measured by @xcite are used ( also adopted by * ? ? ?
however , all of those optical constants do not change the resulting fir colour - colour significantly and do not explain the observational main correlation .
the wavelength dependence of the absorption efficiency @xmath101 is given in @xcite and @xcite for the above species .
vs.@xmath88 ) relation for various heat capacities .
the other quantities adopted for calculations are the same as those in figure [ fig : clr_std ] .
the filled squares represent the colour - colour relation with the heat capacity given in the text for graphite , and the crosses and open squares show the results with heat capacities multiplied by 1/2 and 2 , respectively .
the contours and the dashed and dotted lines are the same as those in figure [ fig : clr_std ] .
[ fig : varc],width=302 ] in our calculation , the optical constants from uv to fir are required .
the above materials have such data and are suitable for our work . for completeness
, we should examine all kinds of possible species , but this is obviously impossible .
however , the above robustness of the predicted fir colour - colour relation implies that the detailed wavelength dependence of the optical constants is not reflected in the fir colour - colour relation .
we also note that the wavelength dependence in the fir absorption efficiency ( emissivity ) is roughly proportional to @xmath208 ( i.e. @xmath194 ) for all the above species .
as shown in text , the fir colour - colour relation is sensitive to the fir emissivity index .
a similar wavelength dependence of fir emissivity is the reason why the above species produce similar fir colour - colour relations .
the difference in the heat capacity may affect the temperature of the vsgs .
the lg temperature is not affected by the heat capacity since it is determined by the equilibrium temperature independent of the heat capacity ( e.g. * ? ? ?
thus , we expect that the difference in heat capacity may cause a variation of the 60 @xmath6 m intensity . here
we test heat capacities whose values are 1/2 and 2 times of the value adopted in the text for graphite . the resulting colour - colour relations are shown in figure [ fig : varc ] . as expected above , the 140 @xmath6m100 @xmath6 m colour , which reflects the equilibrium temperature , is not affected by the change of heat capacity .
this also supports our assumption that the contribution from vsgs is negligible at @xmath209 m . the 60 @xmath6m100 @xmath6 m flux ratio becomes higher for smaller heat capacity , since the temperature of a vsg rises more when a photon is absorbed . at the same time , however , the grains rapidly cool if the heat capacity is small .
those two effects compensate , and the 60 @xmath6m100 @xmath6 m colour does not drastically change . moreover , since the contribution from lgs in radiative equilibrium at the 60 @xmath6 m band becomes larger for higher isrf , the difference in the 60 @xmath6m100 @xmath6 m for @xmath210 is negligible . thus , we conclude that the change of grain heat capacity does not affect our results in the text .
in section [ subsec : mixture ] , we introduce the number of regions with young massive stars , @xmath188 , to explain the fraction of sightlines associated with the sub - correlation .
the number of such regions reflects the recent star formation activity of the galaxy .
thus , we relate @xmath188 with the galactic star formation rate ( sfr ) .
we assume that the galactic disk has a radius of @xmath211 with thickness of @xmath212 ( we assume @xmath213 kpc and @xmath214 pc ) . for simplicity , we assume that the regions with massive stars distribute uniformly in the galactic plane .
then , @xmath188 can be estimated as @xmath215 .
the total uv luminosity in the entire galaxy can be estimated as @xmath216 , where @xmath152 is the uv luminosity of a region with massive stars .
this can be converted to the sfr as @xcite ) , we identified @xmath217 in inoue et al .
( 2000 ) with @xmath216 . ]
@xmath218 since @xmath219 , we obtain @xmath220 .
the total luminosity of the galactic ob stars is estimated to be @xmath221 @xcite . equating this with @xmath222
, we obtain @xmath223 ( or @xmath224 yr@xmath23 ) . if this number is used in equation ( [ eq : prob ] )
, we obtain the probability that a line of sight is associated with a high radiation field of @xmath193 . | a recent data analysis of the far - infrared ( fir ) map of the galaxy and the magellanic clouds has shown that there is a tight correlation between two fir colours : the @xmath0 and @xmath1 colours .
this fir colour relation called `` main correlation '' can be interpreted as indicative of a sequence of various interstellar radiation fields with a common fir optical property of grains . in this paper
, we constrain the fir optical properties of grains by comparing the calculated fir colours with the observational main correlation . we show that neither of the `` standard '' grain species ( i.e. astronomical silicate and graphite grains ) reproduces the main correlation .
however , if the emissivity index at @xmath2 m is changed to @xmath31.5 ( not @xmath4 as the above two species ) , the main correlation can be successfully explained .
thus , we propose that the fir emissivity index is @xmath31.5 for the dust in the galaxy and the magellanic clouds at @xmath5 .
we also consider the origin of the minor correlation called `` sub - correlation '' , which can be used to estimate the galactic star formation rate .
[ firstpage ] dust , extinction galaxies : ism galaxy : stellar content infrared : galaxies infrared : ism magellanic clouds |
post - agb stars are in a transition stage between the asymptotic giant branch ( agb ) and the planetary nebulae ( pn ) stage . during the early stage of post - agb evolution
the star is obscured by material expelled during the agb phase ( the agb ejecta ) . as this ejecta slowly moves away from the central star , the optical depth decreases and
the star can be detected in the optical .
when the star reaches high enough temperatures , the agb ejecta is ionized and is observable as a planetary nebula .
we have studied optically bright post - agb stars ( spectral type a to g supergiants ) which show circumstellar molecular line absorption ( c@xmath0 and cn , or ch@xmath1 ) or emission ( ch@xmath1 ) in their optical spectra .
the radial velocities and low excitation temperatures of the molecules ( bakker et al .
1996a , b ) identify them as circumstellar rather than photospheric or interstellar ( see also hrivnak 1995 ) .
the excitation of c@xmath0 is generally described by rather high temperatures , @xmath6 k , whereas the cn excitation is found to be much lower , @xmath7 k. the reason for this difference is discussed in section 2.2 .
the observed abundances and excitation of the molecules can lead to a determination of the mass - loss rate .
we have used the c@xmath0 ( @xmath8 ) and cn ( @xmath9 ) data of hd 56126 to study optical depth effects by means of the curve of growth , and we present the first results of a survey to observe these optical molecular bands at high spectral resolution ( @xmath10 ) . our primary goal is to resolve the line profiles and to determine the doppler parameter @xmath11 and the chemical ( e.g. abundances ) and physical ( e.g. expansion velocities and temperatures ) conditions of the circumstellar shell .
a curve of growth ( cog ) has been empirically determined for the @xmath12c@xmath12c phillips ( @xmath8 ) ( @xmath13 , @xmath14 , @xmath15 ) and the @xmath12c@xmath16n red system ( @xmath9 ) bands ( @xmath13 , @xmath14 , @xmath17 ) of hd 56126 ( figure 1 ) .
our motive for this investigation is to decide whether the assumption of optically thin lines for the weaker bands ( e.g. 3,0 ) is allowed .
the equivalent widths are taken from the work of bakker et al .
( 1996a , b ) .
the range of line oscillator strengths is @xmath18 to @xmath19 and @xmath20 to @xmath21 for c@xmath0 and cn , respectively .
because three c@xmath0 phillips bands have been observed , the cog has for each lower energy level @xmath22 up to nine transitions ( a p , q , and r branch for each band ) .
this redundancy allows the determination of the cog .
all transitions from a given @xmath22 level are fitted to the cog by changing the column density of that level @xmath23 . for the cn red system
the redundancy is higher due to spin - doublet splitting of the lower and upper electronic state and @xmath24-type doubling of the upper electronic state .
there are twelve transitions from each @xmath25 level : six main and six satellite branches .
three bands were used , which gives in total 36 allowed transitions per @xmath25 level ( @xmath26@xmath27 and @xmath28 ) . after all observed @xmath22 ( or @xmath25 ) levels
have been fitted to the cog ( fig . 1 , left panels )
an optical - depth - corrected absolute rotational diagram ( fig . 1 , middle panels ) gives the absolute population for each @xmath22 ( or @xmath25 ) level . under the assumption of a boltzmann distribution , a linear fit to the diagram gives the average rotational temperature @xmath29 .
finally the rotational temperature from two successive energy levels can be determined as a function of the lower energy level ( fig
. 1 , right panels ) .
the parameters derived from this analysis are given in table 1 .
the empirical cog ( fig .
1 ) shows clearly that the observations cover the optically - thin and saturated parts of the cog .
the theoretical cog has a doppler parameter @xmath30 km s@xmath3 and @xmath31 is reached at an equivalent width of @xmath32 m .
since the strongest lines in the ( 3,0 ) transitions of c@xmath0 and cn have equivalent width of 38 and 34 m , respectively , the optically thin approximation is only valid for the weaker lines in these bands .
we note that all the lines of the c@xmath0 ( @xmath33 ) = ( 3,0 ) band seem to be offset with respect to the other lines .
this might indicate that the band oscillator strength used is somewhat too low .
.results from the curve of growth analysis of c@xmath0 and cn in the optical spectrum of hd 56126 [ cols="<,>,>,<",options="header " , ] hd 235858 has been observed at a resolution of @xmath34 .
figure 3 shows a part of the spectrum which contains three different categories of molecular features .
the broad absorption lines are due to photospheric cn ( @xmath35 k ) , the narrow absorption lines are circumstellar cn ( @xmath36 = 20k and @xmath37 p @xmath38 ) , and the three strongest features are telluric h@xmath0o .
there is one photospheric atomic line ( n@xmath39 ) present in this spectrum .
the circumstellar cn lines are not resolved . since zas et al .
( 1995 ) did not notice photospheric cn absorption in their spectrum , it seems that this pulsating star has only photospheric molecular absorption when the stellar effective temperature is sufficient low ( when the star is largest ) .
in an earlier report on the detection of circumstellar cn in iras 08005 - 2356 we reported the presence of only one broad component ( bakker et al .
the new spectra ( fig . 3 , @xmath40 ) have resolved the broad component into two separate resolved components with a velocity separation of @xmath5 km s@xmath3 .
the two absorption components could be due to wind material moving at two different velocities ( multiple shells ) , or to two photodissociation fronts at different expansion velocities possibly one due to the stellar and one to the interstellar radiation field .
we have investigated optical depth effect for the circumstellar c@xmath0 and cn lines in the optical spectrum of hd 56126 . from a curve of growth analysis
we find a doppler parameter @xmath2 km s@xmath3 .
cn ( 2,0 ) lines of iras 080052356 and hd 235858 and ch@xmath1 of the red rectangle have been observed at a resolution @xmath41 .
ch@xmath1 has been resolved .
the cn lines of iras 080052356 are resolved into two separate resolved components , and the cn of hd 235858 has not been resolved . | recent optical spectra of post - agb stars show the presence of c@xmath0 , cn , and ch@xmath1 originating in the circumstellar shell .
we present here new , higher resolution spectra which provide constraints on the physical parameters and information on the line profiles .
an empirical curve of growth for the c@xmath0 phillips and cn red system lines in the spectrum of hd 56126 yields @xmath2 km s@xmath3 .
ch@xmath1 ( 0,0 ) emission lines in the spectrum of the red rectangle have been resolved with a fwhm @xmath4 0.8 kms@xmath3 .
the circumstellar cn lines of iras 080052356 are resolved into two separate components with a velocity separation of @xmath5 km s@xmath3 .
the line profiles of cn of hd 235858 have not been resolved . |
in the past decade and a half , over two hundred and fifty extra - solar planets have been discovered . with the data expected to be produced by satellites such as corot @xcite and kepler @xcite
, it will not only be possible to find smaller planets , but moons of those planets as well @xcite . as a result
, the detectability of extra - solar moons is starting to be explored in terms of their effect on planetary microlensing @xcite and transit lightcurves @xcite .
upper limits have already been placed on the mass and radius of putative moons of the planets hd 209458 b @xcite , ogle - tr-113 b @xcite and hd 189733 b @xcite . while the limitations of microlensing and the transit technique for detecting moons have been discussed and used in the literature , the limitations of other techniques such as the time - of - arrival ( toa ) technique have not .
this technique involves determining the variations in line - of - sight position to the host star , usually a pulsar , using the observed time of periodic events associated with that host .
the aim of this analysis is to explore exactly what the toa signal of a planet - moon pair is , and relate it to the planetary systems that can give the most precise timing information , the systems around millisecond pulsars .
the first planetary system outside the solar system was detected around the millisecond pulsar psr 1257 + 12 @xcite .
this detection was made by investigating periodic variations in the time of arrival of its radio pulses using a timing model .
an example timing model for the case in which the planet s orbit around the pulsar is circular is : @xmath13 where @xmath14 and @xmath15 are the times at which the initial and @xmath16 pulses are emitted in the pulsar s frame , @xmath17 and @xmath18 are the times the initial and @xmath16 pulses are received in the observatory s frame , and the term @xmath19 acts to change the frame of reference from the observatory on earth to the barycenter of the pulsar system ( see * ? ? ?
* for a more complete discussion of the components of @xmath20 ) .
the final term represents the effect of a planet on the motion of the pulsar , where @xmath9 is the planet - pulsar distance , @xmath10 is the angle between the normal of the planet - pulsar orbit and the line - of - sight and @xmath21 is the initial angular position of the planet measured from the @xmath22-axis , about the system barycenter . in addition , @xmath23 and @xmath6 are the mass of the pulsar and the planet , respectively .
currently , four planets around two millisecond pulsars have been discovered , three around psr 1257 + 12 @xcite and one around psr b1620 - 26 @xcite .
these four planets include one with mass @xmath24 earth masses , the lowest mass extra - solar planet known .
this low - mass detection threshold , in addition to measurements of orbital perturbations ( e.g. , the 2:3 orbital resonance between psr 1257 + 12c and psr 1257 + 12d , * ? ? ?
* ) , demonstrate the sensitivity of the toa technique .
indeed , as a result of their high rotation rate ( and thus large number of sampled pulses ) and low level of noise activity , millisecond pulsars make optimal targets for high precision toa work ( see e.g. , * ? ? ?
in order to investigate the perturbation caused by planet - moon binarity , the timing model presented in equation ( [ toa_example ] ) must be updated to include effects due to the presence of the moon . for simplicity , we consider here only systems in which both the orbit of the planet and moon around their common barycenter , and the orbit of the planet - moon barycenter around the pulsar , are both circular and lie in the same plane .
an example model taking this assumption into account is : @xmath25 we have explicitly modified @xmath26 to indicate that it depends on the combined planet - moon mass , and included another term , @xmath27 , to account for planet - moon binarity . here
@xmath7 is the mass of the moon , @xmath8 is the distance between the planet and the moon , and @xmath28 is the initial angular position of the planet measured from the planet - moon barycenter .
the quantities @xmath9 , @xmath8 , @xmath10 , @xmath29 and @xmath30 are shown in figure [ coordsystem ] .
the functional form of @xmath27 can be derived from @xmath31 , the vector between the system barycenter and the pulsar , using : @xmath32 where @xmath11 is the speed of light and @xmath33 is a unit vector pointing along the line of sight , the only direction along which quantities can be measured .
the governing equation for @xmath34 can be written as the sum of the zeroth order term , which describes @xmath26 , and the tidal terms , which describe @xmath27 : @xmath35,\label{toa - d2rs}\end{aligned}\ ] ] where the tidal terms have been collected into square brackets and noting that @xmath36 , @xmath37 and @xmath38 , and where the vectors @xmath31 , @xmath39 , @xmath40 , @xmath41 and @xmath42 are also shown in figure [ coordsystem ] . using the coordinate system in figure [ coordsystem ]
, it can be seen after some algebra that : @xmath43\mathbf{i } + \left[r\sin \phi_b - \frac{m_p}{m_p+m_m}r \sin \phi_p \right]\mathbf{j}\\ \label{toa - vec - rp } \mathbf{r } + \mathbf{r}_p & = & \left[r\cos \phi_b + \frac{m_m}{m_p+m_m}r \cos \phi_p\right]\mathbf{i } + \left[r \sin \phi_b + \frac{m_m}{m_p+m_m}r \sin \phi_p \right]\mathbf{j}\end{aligned}\ ] ] where @xmath44 and @xmath45 are defined in figure [ coordsystem ] , and @xmath44 is the direction to the line - of - sight , projected onto the plane of the orbit
. as the orbits are both circular and coplanar , we have that @xmath46 , and @xmath47 where @xmath3 and @xmath2 are the constant mean motions of the two respective orbits . substituting equations
( [ toa - vec - rm ] ) and ( [ toa - vec - rp ] ) into the coefficients of the last two terms of equation ( [ toa - d2rs ] ) , assuming @xmath48 and using the binomial expansion to order @xmath49 gives : @xmath50 \label{toa - expnd - ms}\end{aligned}\ ] ] @xmath51\label{toa - expnd - ps}.\end{aligned}\ ] ] substituting equations ( [ toa - vec - rm ] ) , ( [ toa - vec - rp ] ) , ( [ toa - expnd - ms ] ) and ( [ toa - expnd - ps ] ) into ( [ toa - d2rs ] ) gives , after simplification : @xmath52 .\label{toa - d2rs2}\end{aligned}\ ] ] from figure [ coordsystem ] it can be seen that : @xmath53 substituting equations ( [ toa - d2rs2 ] ) and ( [ toa - ndef ] ) into equation ( [ toa - toapertpmdef ] ) , gives : @xmath54 .\label{toa - pert1}\end{aligned}\ ] ] the @xmath55 term in equation ( [ toa - pert1 ] ) has the same frequency as the signal of a lone planet and it can be shown that it acts to increase the measured value of @xmath56 derived from @xmath26 by @xmath57 . consequently , this term can be neglected as it will be undetectable as a separate signal .
also , the edge of the stability region for a prograde satellite of the low - mass component of a high - mass binary can be approximated by @xmath58 for the case of circular orbits , where @xmath59^{1/3}$ ] is the secondary s hill radius @xcite .
when @xmath8 is equal to this maximum stable radius @xmath60 .
as @xmath61 is likely , we have that the denominators of the @xmath62 and @xmath63 terms will never approach zero .
this , in addition to the assumption of zero eccentricities , means that resonance effects can be neglected .
consequently , equation ( [ toa - pert1 ] ) can be simplified by neglecting @xmath3 in the denominators , giving : @xmath64 .\label{toa - pert2}\ ] ] writing @xmath2 in terms of @xmath8 , using kepler s law , gives : @xmath65 .\label{toa - pert}\ ] ] from equation ( [ toa - pert ] ) , we have that the size of the perturbation varies as @xmath66[r / r]^5 $ ] times the system crossing time , @xmath5 .
so , the best hope of a detectable signal occurs when the planet - moon pair are close to the parent pulsar , widely separated from each other , both quite massive , and very accurate timing data is available .
our result is consistent with a similar study done by @xcite , who calculated the radial velocity perturbation on one component of a binary star system for the case in which the other component was an unresolved pair . converting their radial velocity perturbation to a timing perturbation , setting the mass of the planet and moon equal to each other , as in the case investigated by @xcite , and noting that @xmath8 in this work is equivalent to their @xmath67 , our formula and that of @xcite agree .
to investigate whether or not it is possible to detect moons of pulsar planets , we simplify equation ( [ toa - pert ] ) by summing the amplitudes of the sinusoids , giving the maximum possible amplitude : @xmath68 thus , for a given @xmath69 , the maximum amplitude increases linearly with @xmath9 .
so , for planet - moon pairs that are far from each other and their parent pulsar , detection may be possible .
for example , a 0.1au jupiter - jupiter binary located 5.2au from a host pulsar would produce a @xmath27 of amplitude 960ns , which compares well with the 130ns residuals obtained from one of the most stable millisecond pulsars , psr j0437 - 4715 @xcite . to demonstrate this method ,
the expected maximum signals from a moon orbiting each of the four known pulsar planets was explored .
it was found that in the case of psr b1620 - 26 b , signals that are in principle detectable could confirm or rule out certain configurations of moon mass and orbital parameters ( see figure [ pulsarplanetmoon ] ) . in the particular case of psr b1620 - 26 b
, the perturbation signal will not exactly match the signal shown in equation ( [ toa - pert ] ) due to the effect of its white dwarf companion .
the effect of this companion was investigated as a side project and it was found that its effect was to introduce additional perturbations on top of the @xmath26 and @xmath27 calculated . consequently , the detection threshold represents an upper limit to the minimum detectable signal and thus the analysis is still valid . .
the mass of the host was set at 1.7 solar masses ( the sum of the mass of the pulsar and its white dwarf companion ) .
the @xmath70 detection threshold was calculated assuming the @xmath71 timing residuals given in @xcite are uncorrelated and that similar accuracy toa measurements of psr b1620 - 26 continue to the present day .
the stability region was estimated as 0.36 times the hill sphere of the planet .
, width=368,height=288 ] unfortunately , there are practical limits to the applicability of this method .
they include discounting other systems that could produce similar signals , sensitivity limits due to intrinsic pulsar timing noise , and limits imposed by moon formation and stability . first , other systems that could produce similar signals need to be investigated .
possible processes include pulsar precession ( see e.g. , * ? ? ?
* ) , periodic variation in the ism @xcite , gravitational waves @xcite , unmodelled interactions between planets @xcite and other small planets . to help investigate the last two options
, we plan on completing a more in - depth analysis of the perturbation signal of an extra - solar moon , including the effects of orbital inclination and orbital eccentricity .
second , the noise floor of the system needs to be examined .
the suitability of pulsars for signal detection is limited by two main noise sources , phase jitter and red timing noise ( for example , * ? ? ?
phase jitter is error due to pulse - to - pulse variations and leads to statistically independent errors for each toa measurement .
phase jitter decreases with increasing rotation rate ( decreasing @xmath72 ) due to the resultant increase in the number of pulses sampled each integration .
red timing noise refers to noise for which neigbouring toa residuals are correlated .
red timing noise has been historically modeled as a random walk in phase , frequency or frequency derivative ( for example , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
red noise is strongly dependent on @xmath73 .
this relationship can be understood from the theoretical standpoint that red noise is due to non - homogeneous angular momentum transport either between components within the pulsar ( e.g. * ? ? ? * ) or between it and its environment ( e.g. * ? ? ?
* ) . to illustrate the effect of these two noise sources on toa accuracy ,
an estimate of their combined residuals as a function of @xmath72 and @xmath73 is shown in figure [ pulsarplot ] .
for comparison , the values of @xmath72 and @xmath73 of every pulsar as of publication are also included .
the atnf pulsar catalogue @xcite was used to provide the pulsar data for this plot .
third , whether or not moons will be discovered depends on whether or not they _ exist _ in certain configurations , which depends on their formation history and orbital stability .
recent research suggests that there are physical mass limits for satellites of both gas giants @xcite and terrestrial planets @xcite .
also , tidal and three - body effects can strongly affect the longevity of moons @xcite .
finally , while this method was investigated for the specific case of a pulsar host , this technique could also be applied to planets orbiting other clock - like hosts such as pulsating giant stars @xcite and white dwarfs @xcite .
akgn t. , link b. , wasserman i. , 2006 , , 365 , 653 arzoumanian z. , nice d. j. , taylor j. h. & thorsett s. e. , 1994 , , 422 , 671 atobe k. , & ida s. , 2007 , , 188 , 1 auvergne m. , et al . , 2003 , , 4854 , 170 backer , d. c. , 1993 , , 36 , 11 backer d. c. , foster r. s. , & sallmen s. , 1993 , , 365 , 817 barnes j. w. & obrien d. p. , 2002
, , 575 , 1087 basri g. , borucki w. j. , & kock d. , 2005 , , 49 , 478 boynton , p. e. , groth , e. j. , hutchinson , d. p. , nanos , g. p. , jr . , partridge , r. b. , & wilkinson , d. t. , 1972 , , 175 , 217 brown t. m. , charbonneau d. , gilliland r. l. , robert l. , noyes r. w. & burrows a. , 2001 , , 552 , 699 canup r. m. & ward w. r. , 2006 , , 441 , 834 cheng k. s. , 1987 , , 321 , 805 cordes j. m. , 1980 , , 237 , 216 cordes j. m. , 1993 , , 36 , 43 detweiler , s. , 1979 , , 234 , 1100 domingos r. g. , winter o. c. & yokohama t. , 2006 , , 373 , 1227 gillon m. , pont f. , moutou c. , bouchy f. , courbin f. , sohy s. & magain p. , 2006
, , 459 , 249 groth , e. j. , 1975 , , 29 , 443 han c. , & han w. , 2002 , , 580 , 490 holman m. j. & wiegert p. a. , 1999 , , 117 , 621 jones p. b. , 1990 , , 246 , 364 kopeikin s. m. , 1997 , , 288 , 129 konacki m. , & wolszczan a. , 2003 , , 591 , l147 laughlin g. , & chambers j. e. , 2001 , , 551 , l109 manchester r. n. , hobbs g. b. , teoh a. , & hobbs m. , 2005 , , 129 , 1993 mullally f. , winget d. e. , & kepler s. o. , 2006 , , 352 , 265 pont f. et al . , 2007 , , 476 , 1347 sartoretti p. , & schneider j. , 1999 , , 134 , 553 scherer k. , fichtner h. , anderson j. d. , & lau e. l. , 1997 , science , 278 , 1919 schneider j. & cabrera j. , 2006 , , 445 , 1159 sigurdsson s. , richer h. , hansen b. , stairs i. & thorsett s. , 2003 , science , 301 , 193 silvotti et al . , 2007 , , 449 , 189 szab gy .
m. , sztmary k. , divki zs .
, & simon a. , 2006 , , 450 , 395 thorsett s. e. , arzoumanian z. , camilo f. , lyne a. g. , 1999 , , 523 , 763 van straten w. , bailes m. , britton m. , kulkarni s. r. , anderson s. b. , manchester r. n. , & sarkissian j. , 2001 , , 412 , 158 wada k. , & kokubo e. , 2006 , , 638 , 1180 wolszczan a. , 1994 , science , 264 , 538 wolszczan a. , & frail d. a. , 1992 , , 355 , 145 | the perturbation caused by planet - moon binarity on the time - of - arrival signal of a pulsar with an orbiting planet is derived for the case in which the orbits of the moon and the planet - moon barycenter are both circular and coplanar .
the signal consists of two sinusoids with frequency @xmath0 and @xmath1 , where @xmath2 and @xmath3 are the mean motions of the planet and moon around their barycenter , and the planet - moon system around the host , respectively .
the amplitude of the signal is equal to the fraction @xmath4 [ r / r]^5 $ ] of the system crossing time @xmath5 , where @xmath6 and @xmath7 are the the masses of the planet and moon , @xmath8 is their orbital separation , @xmath9 is the distance between the host pulsar and planet - moon barycenter , @xmath10 is the inclination of the orbital plane of the planet , and @xmath11 is the speed of light .
the analysis is applied to the case of psr b1620 - 26 b , a pulsar planet , to constrain the orbital separation and mass of any possible moons .
we find that a stable moon orbiting this pulsar planet could be detected , if the moon had a separation of about one fiftieth of that of the orbit of the planet around the pulsar , and a mass ratio to the planet of @xmath12 5% or larger . |
the active manipulation of spin - dependent electron transport is the principal task in spintronics @xcite .
one of the effective way to achieve spin polarization of electrons injected into semiconductors is spin - dependent tunnelling through a barrier in ferromagnetic metal / insulator / semiconductor heterostructures @xcite .
the maximum of the spin injection efficiency reaches 52% at 100 k and 32% at 290 k for a mgo barrier on gaas @xcite .
high electrical injection of spin - polarized electrons from a fe film through an al@xmath0o@xmath1 tunnel barrier into si has been demonstrated in @xcite . however , in si the electron spin polarization was observed at low temperatures
30% at 5 k , with polarization extending to at least 125 k. although important results in the spin injection have been obtained , high values of the spin polarization of injected electrons at room temperature has not been achieved .
therefore , it is crucial to find a new method of the spin polarization that allows us to achieve high spin - injection efficiency . in this paper
, we present new method of the electron spin polarization in ferromagnetic metal / insulator / semiconductor heterostructure , which based on spin - dependent backward scattering on exchange - splitted levels in a @xmath2 quantum well ( qw ) and on the electron capture by this qw .
the qw is formed in the semiconductor near the insulator / semiconductor interface .
the insulator layer is thin in order to provide electron tunneling from the ferromagnetic metal and to split levels in the qw by the exchange interaction . for the case ,
when one of exchange - splitted levels lies in the top region of the qw , the backscattering process of injected electrons becomes dependent on their spins .
the capture of backscattered spin - polarized electrons by the qw leads to an additional coulomb repulsion for electrons tunneling from the ferromagnetic metal and to a considerable spin - polarized decrease of the current flowing in the heterostructure . in this way ,
high values of the spin polarization and manipulation of injected electron spins by the charge of the qw can be achieved . by analogy with the spin polarization in ferromagnetic metal / insulator / semiconductor heterostructures , spin polarization of electron current
can be observed in heterostructures consisted of semiconductor substrates and granular films with ferromagnetic metal nanoparticles in an insulator matrix .
ones of these heterostructures are sio@xmath0(co)/gaas heterostructures , where the sio@xmath0(co ) is the granular sio@xmath0 film with co nanoparticles @xcite .
the @xmath2 qw ( accumulation electron layer ) with exchange - splitted levels is formed at the interface in the gaas @xcite . in sio@xmath0(co)/gaas heterostructures
extremely large magnetoresistance and the current reduction on the temperature dependence are observed at room temperature .
the paper is organized as follows . in the next section
, we study the electron backscattering process on exchange - splitted levels in @xmath2 qw . in sec .
iii we consider the capture of backscattered electrons by the qw and the current reduction caused by the qw charging .
the spin polarization of electron current caused by the backscattering process and , consequently , by the current reduction is described in sec .
finally , we summarize the results in sec . v.
let us consider the backward scattering of injected electrons on exchange - splitted levels of the qw in a ferromagnetic metal / insulator / semiconductor heterostructure ( fig .
[ bar2 ] ) .
exchange interaction between electrons in the ferromagnetic metal and electrons in the qw through the thin insulator layer splits electron levels in the qw .
electrons on exchange - splitted sublevels ( sublevels @xmath3 and @xmath4 in fig .
[ bar2 ] ) have opposite spin orientations .
difference of their energies is equal to the exchange energy @xmath5 . for clarification of the main features of scattering dependencies
we restrict our consideration on the backscattering process on one of exchange - splitted levels the top sublevel @xmath3 with certain spin orientation and neglect insignificant parts of the electron wavefunction on the sublevel @xmath3 outside the qw . .
@xmath6 is the ferromagnetic metal , @xmath7 is the insulator layer , @xmath8 is the applied electric field . ]
the electron wavefunction on the sublevel @xmath3 is the product of the spatial function @xmath9 and the spin function @xmath10 @xmath11 where @xmath12 is the electron spin . in the wkb
( wentzel - kramers - brillouin ) approximation @xcite the spatial function in the qw can be written as @xmath13 where @xmath14 is the wavevector of the electron on the sublevel @xmath3 in the zero approximation with respect to @xmath15 , @xmath16 is the electron mass , @xmath17 is the energy counted from the qw bottom , @xmath18 is the width of the qw , @xmath19 is the normalization coefficient , @xmath20 is the number of the sublevel @xmath3 .
the wavefunction of injected electron flying over the qw has the form of the product of the spatial function @xmath21 and the spin function @xmath22 @xmath23 where @xmath24 @xmath25 , @xmath26 is the energy counted from the qw bottom , @xmath27 is the normalization coefficient . for the interaction @xmath28 between the injected electron and the electron localized on the sublevel @xmath3 , in the first approximation with respect to @xmath28 the probability of the backscattering per unit time is @xcite @xmath29 where @xmath30 is the density of final states at the energy @xmath31 , @xmath32 is the final wavefunction and @xmath33 is the initial wavefunction combined of injected and localized electrons . if electrons form the singlet spin configuration ( @xmath34 , @xmath35 or @xmath36 , @xmath37 ) , then spatial parts of wavefunctions have the symmetric combination @xmath38 @xmath39 where @xmath40 is the wavefunction of the backscattered electron described by eq . ( [ 2bar ] ) with the substitution @xmath41 . for the singlet spin state
the backscattering probability ( [ 3bar ] ) is equal to @xmath42 where @xmath43 @xmath44 if electrons form the triplet spin configuration ( @xmath34 , @xmath37 or @xmath36 , @xmath35 ) , then spatial parts of wavefunctions are antisymmetric @xmath45 @xmath46 for the triplet state the probability ( [ 3bar ] ) can be written as @xmath47 magnitudes @xmath48 and @xmath49 in relations ( [ 4bar ] ) and ( [ 5bar ] ) are functions of wavevectors @xmath50 and @xmath51 . besides , the wavevector @xmath51 depends on the number @xmath52 of localized level : @xmath53 .
taking into account wavefunction forms ( [ 1bar ] ) and ( [ 2bar ] ) , for the uniform interaction @xmath54 we obtain @xmath55\ ] ] @xmath56 @xmath57 ^ 2.\ ] ] probabilities @xmath58 ( [ 4bar ] ) and @xmath59 ( [ 5bar ] ) strongly depend on the difference of wavevectors @xmath60 and , consequently , on the difference @xmath61 between the energy of injected electron @xmath62 and the energy of localized electron @xmath17 in the qw . for @xmath63 the energy difference is @xmath64 singlet and triplet backscattering probabilities versus the normalized wavevector difference @xmath65 for @xmath66 are shown in fig .
probabilities are normalized by the magnitude of the singlet probability @xmath67 with @xmath68 and @xmath69 . in accordance with relation ( [ x1 ] ) , the singlet and triplet probabilities 1s and 1 t ( @xmath68 ) are shown as functions of the variable @xmath61 ( upper axis ) . in this case the qw contains only one exchange - splitted level .
calculations have been done for @xmath70 nm and @xmath71 mev .
it is necessary to notice that the probability of the singlet backscattering @xmath58 ( curves 1s , 2s ) is higher than the triplet backscattering probability @xmath59 ( curves 1 t , 2 t ) the backward scattering becomes dependent on spins of injected electrons . for scattering of injected electrons on the level with @xmath68 and with the wavevector @xmath72 , the ratio of singlet and triplet probabilities leads to the relation @xmath73 on the qw versus the normalized wavevector difference @xmath74 .
@xmath75 is the wavevector of localized electron , @xmath52 is the number of level .
probabilities are normalized by the magnitude of the singlet probability @xmath67 with @xmath68 and @xmath69 .
1s , 1 t are singlet and triplet backscattering on the first level ( @xmath68 ) ; 2s , 2 t are singlet and triplet backscattering on the second level ( @xmath76 ) , respectively . for
the qw with the width @xmath70 nm and the energy depth @xmath71 mev the singlet and triplet probabilities 1s and 1 t are shown as functions of the difference @xmath61 between the energy of injected and localized electrons ( upper axis ) . ] the backscattering probability strongly reduces with growth of @xmath77 and @xmath61 .
the greatest magnitude of backscattering is achieved for the level with @xmath68 .
thus , the backscattering process becomes important , if ( 1 ) the qw contains only exchange - splitted level with @xmath68 , ( 2 ) one sublevel of the exchange - splitted level with certain spin orientation lies at the top of the qw and ( 3 ) the energy of injected electrons is closed to the energy of localized electron on this sublevel .
the capture of backscattered electrons by the qw leads to the considerable current reduction dependent on spin orientation of injected electrons .
if the fermi level lies below localized electron levels in the qw , then at a finite temperature these levels are partially filled by electrons .
backscattered electrons are captured by the qw and , in accordance with their spin orientation , they occupy different localized levels .
we suppose that the spin relaxation time is greater than the storage time of additional electrons in the qw .
then , for the singlet scattering process backscattered electrons occupy the sublevel @xmath4 with spin orientation opposite to spin orientation of the sublevel @xmath3 ( fig .
[ bar2 ] ) .
the sublevel @xmath4 lies below the sublevel @xmath3 . on the contrary , for the triplet case backscattered electrons fall on the sublevel @xmath3 .
the storage time @xmath78 of the presence of additional electrons in the qw depends on the electron - hole recombination , on temperature activation processes , and on the electron tunneling into the conduction band . for underlying levels
the storage time @xmath78 is greater than the storage time of electrons on overlying ones .
the additional charge in the qw leads to the electrostatic blockade of injected electrons and to the current reduction . in this way
, the current flowing in ferromagnetic metal / semiconductor heterostructure with qw , which contains exchange - splitted levels , is unstable .
this current instability is accompanied by the charge accumulation in the qw and by the current reduction depended on spin orientations of injected electrons .
let us calculate the reduction of the current . for clarity , we consider the current reduction caused by the singlet backscattering . in the triplet case ,
the consideration is analogous .
the current density flowing over the qw is equal to @xmath79 where @xmath80 is the electron charge , @xmath81 is the electron mobility , @xmath8 is the electric field , @xmath82 is the electron concentration over the qw , @xmath83 is the electron concentration without an electric field , @xmath84 is the potential of the field of additional localized electrons in the qw , @xmath85 is the boltzmann constant , and @xmath86 is the temperature . in the singlet backscattering case ,
the additional charge accumulates on the sublevel @xmath4 ( fig .
[ bar2 ] ) .
the potential @xmath84 of the field caused by this additional charge is determined by the equation @xcite @xmath87 where @xmath88 is the dielectric permittivity of the semiconductor in the qw region ; @xmath89 and @xmath90 are electron concentrations on the sublevel @xmath4 in the electric field and without a field , respectively . if the additional concentration of the charge @xmath91 is uniformly distributed over the qw width , then the solution of eq .
( [ 7bar ] ) is given by @xmath92 injected electrons must surmount the additional barrier with the energy height @xmath93 taking into account relations ( [ 6bar ] ) and ( [ 8bar ] ) , we obtain the current density of electrons incoming on the sublevel @xmath4 @xmath94.\ ] ] release of additional electrons from the sublevel @xmath4 is determined by the time @xmath95 and the current density of outgoing electrons can be written as @xmath96 for the equilibrium process @xmath97 and @xmath98=\frac{(n_b - n_b^{(0)})d}{\tau_b}. \label{9bar}\ ] ] relation ( [ 9bar ] ) is the equation in the unknown additional electron concentration @xmath91 .
taking into account relation ( [ 6bar ] ) , we find the current reduction caused by the singlet electron backscattering on the qw @xmath99.\label{10bar}\ ] ] current reductions @xmath100 versus the applied electric field @xmath8 for different times @xmath95 are shown in fig .
calculations are performed for @xmath101 , width of the qw @xmath102 2 nm , permittivity @xmath103 1 , @xmath104 300 k , @xmath105 @xmath106/v@xmath107s , and @xmath108 @xmath109 . from the presented dependencies
we can see that backscattering of injected electrons on exchange - splitted levels and accumulation of electrons in the qw leads to considerable reduction of the current depended on its spin orientation .
caused by the singlet electron backscattering in ferromagnetic metal / insulator / semiconductor heterostructure with quantum well ( qw ) contained exchange - splitted levels versus the applied electric field @xmath8 for different storage time @xmath95 of additional electrons in the qw .
the backscattering probability @xmath101 , width of the qw @xmath102 2 nm , temperature @xmath104 300 k , and the electron concentration over the qw @xmath108 @xmath109 . ]
the current reduction depends on the electron concentration @xmath83 in the semiconductor . for small values of @xmath83 the additional concentration @xmath91 in eq .
( [ 9bar ] ) leads to zero and the reduction is small , @xmath110 . for great values of the concentration @xmath83 ( for example , close to metal concentrations ) the qw contains filled levels and
the additional charge in the qw is impossible . as a result of this
, there is no any reduction of the current . for the triplet backscattering case ,
backscattered electrons accumulate on the level @xmath3 .
the current reduction @xmath111 is determined by relation ( [ 10bar ] ) , in which we must perform the substitution @xmath112 .
the additional electron concentration @xmath113 is the solution of eq .
( [ 9bar ] ) with substitutions @xmath114 and @xmath115 . in comparison with the singlet case , for @xmath116 and @xmath117 the current reduction @xmath111 caused by the triplet backscattering and by the accumulation of electrons in the qw is insignificant .
the qw with exchange - splitted levels can be regarded as spin filter for injected electrons .
let us consider the spin current flowing over the qw with square modulation of the spin projection @xmath118 ( fig .
[ bar5]a ) @xmath119 and @xmath120 are currents with spin polarization @xmath121 and @xmath122 , respectively , after electron backward scattering on the qw .
( b ) spin polarization @xmath123 of the electron current caused by the electron backscattering versus the electric field @xmath8 for different values of the storage time @xmath95 . ] where @xmath52 is the electron concentration , @xmath81 is the mobility , @xmath8 is the electric field , @xmath124 .
we suppose that the modulation period @xmath125 is much greater than the storage times @xmath126 and @xmath95 : @xmath127 . without spin - dependent
backscattering and charge accumulation in the qw the spin current @xmath128 is not modified . in the presence of singlet and triplet backscattering and accumulation of backscattered electrons in the qw ,
the spin current @xmath128 decreases to @xmath129 and the reduction becomes dependent on its spin projection @xmath118 . the magnitude of the reduction of the current @xmath130 caused by the singlet backscattering on the sublevel @xmath3 ( fig .
[ bar2 ] ) is higher than the magnitude of the reduction of the current @xmath131 caused by the triplet scattering process .
taking into account relations ( [ 6bar ] ) , ( [ 10bar ] ) and ( [ 11bar ] ) , for time regions far from pulse edges we can write the spin polarization as @xmath132 the spin polarization @xmath123 versus the electric field @xmath8 has been calculated for different values of the storage time @xmath95 for backscattering probability @xmath133 , width of the qw @xmath102 2 nm , permittivity @xmath103 1 , temperature @xmath104 300 k , electron mobility @xmath134 @xmath106/v@xmath107s , concentration @xmath108 @xmath109 , and time @xmath135 10 ns ( fig . [ bar5]b ) .
one can notice that the spin polarization @xmath123 increases with growth of the electric field @xmath8 and the storage time @xmath95 .
the backward scattering of injected electrons on exchange - splitted levels of quantum wells in ferromagnetic metal / insulator / semiconductor heterostructures can be used as the effective way of the spin polarization of the current . the necessary condition to obtain high values of the spin polarization is : one of the exchange - splitted levels must be in the top region of the qw . if the energy of injected electrons is close to the energy of localized electrons , the backward scattering becomes dependent on spins of injected electrons on singlet or triplet spin configurations .
it is found that the probability of the singlet backscattering @xmath58 is higher than the triplet backscattering probability @xmath59 .
the capture of backscattered electrons by the qw leads to an additional coulomb repulsion for electrons and to the considerable spin - dependent reduction of the current flowing in the heterostructure .
the spin polarization @xmath123 of the current increases with growth of the applied electric field and the storage time of electrons in the qw and its high values can be achieved at room temperature . in this way
, the qw with exchange - splitted levels in ferromagnetic metal / insulator / semiconductor heterostructures can be regarded as spin filter .
this work was supported by the government of russia ( project no .
14.z50.31.0021 , leading scientist m. bayer ) . 777 s.a .
wolf , d.d .
awschalom , r.a .
buhrman , j.m .
daughton , s. von molnar , m.l .
roukes , a.y .
chtchelkanova and d.m .
treger , science * 294 * , 1488 ( 2001 ) .
g. schmidt , j. phys .
phys . * 38 * , r107 ( 2005 ) .
i. utic , j. fabian , and s. das sarma , rev .
phys . * 76 * , 323 ( 2004 ) .
hammar and m. johnson , appl .
letters * 79 * , 2591 ( 2001 ) .
x. jiang , r. wang , r.m .
shelby , r.m .
macfarlane , s.r .
bank , j.s .
harris , and s.s.p .
parkin , phys .
letters * 94 * , 056601 ( 2005 ) .
li , g. kioseoglou , o.m.j .
van t erve , m.e .
ware , d. gammon , r.m .
stroud , and b.t .
jonker , appl .
letters * 86 * , 132503 ( 2005 ) .
x.y . dong , c. adelmann , j.q .
xie , c.j .
palmstrm , x. lou , j. strand , p.a .
crowell , j .-
barnes , and a.k .
petford - long , appl .
letters * 86 * , 102107 ( 2005 ) .
jonker , g. kioseoglou , a.t .
hanbicki , c.h .
li , and p.e .
thompson , nature physics * 3 * , 542 ( 2007 ) .
li , g. kioseoglou , o.m.j .
van t erve , p.e .
thompson , and b.t .
jonker , appl .
letters * 95 * , 172102 ( 2009 ) .
g. kioseoglou , a.t .
hanbicki , r. goswami , o.m.j .
van t erve , c.h .
li , g. spanos , p.e .
thompson , and b.t .
jonker , appl .
letters * 94 * , 122106 ( 2009 ) .
h. saito , j.c .
le breton , v. zayets , y. mineno , s. yuasa , and k. ando , appl .
letters * 96 * , 012501 ( 2010 ) .
lutsev , a.i .
stognij , and n.n .
novitskii , jetp letters * 81 * , 514 ( 2005 ) .
lutsev , a.i .
stognij , n.n .
novitskii , and a.a .
stashkevich , j. magn .
magn . mater . *
300 * , e12 ( 2006 ) .
lutsev , a.i .
stognij , and n.n .
novitskii , phys .
b * 80 * , 184423 ( 2009 ) .
l. lutsev , _ giant injection magnetoresistance .
experimental and theoretical study , perspectives of applications_. ( lambert academic publishing , saarbrucken , germany , 2013 ) , isbn 978 - 3 - 659 - 43964 - 3 [ in russian ] .
lutsev , j. phys . : condens .
matter * 18 * , 5881 ( 2006 ) .
davydov , _ quantum mechanics _ ( pergamon press , oxford , 1976 ) . l.d .
landau and e.m .
lifshitz , _ the classical theory of fields_. vol .
2 ( 4th ed . ) ( butterworth - heinemann , 1975 ) . | the backscattering process of injected electrons on exchange - splitted levels of quantum well ( qw ) in ferromagnetic metal / insulator / semiconductor heterostructure is studied .
it is found that , if one of the exchange - splitted levels lies in the top region of the qw and the energy of injected electrons is close to the energy of localized electron on this level , the backward scattering becomes dependent on spins of injected electrons .
accumulation of backscattered electrons in the qw leads to considerable reduction of the current depended on its spin orientation .
the spin polarization increases with growth of the applied electric field and the storage time of electrons in the qw .
high values of the spin polarization can be achieved at room temperature . in this way , the qw with exchange - splitted levels in ferromagnetic metal / insulator / semiconductor heterostructure can be used as effective spin filter . |
to characterize the metallicity and age of an old stellar system , iron abundance [ fe / h ] ( the logarithm of the iron - to - hydrogen abundance ratio with respect to its solar value ) and the stellar effective temperature must be determined together for its main - sequence turnoff ( msto ) stars .
these are dwarfs and subgiants of types f and early g with @xmath0 .
their spectra are dominated by fei lines , whose strength increases as declines . for both metallicity and age , then , a reliable derivation of is necessary .
this can be problematical even for an individual star .
the photometric colors often used to derive are sensitive to reddening by the interstellar medium and to the modeling of convection ( castelli , gratton , & kurucz 1997 ) , as is the more recently developed infrared flux method @xcite .
methods using balmer - line profiles are reddening - independent but still very model - dependent @xcite .
methods relying on the excitation equilibrium of fei are reddening - independent and model - independent if lines are weak , but are susceptible to systematic errors in laboratory measurements of gf - values @xcite and to possible departures from local thermodynamic equilibrium ( lte ; * ? ? ? * ) .
currently , values derived for individual msto stars show uncomfortably large spreads : values of @xcite and @xcite average 100
200 lower than those found by @xcite and @xcite , with those of @xcite and @xcite in between . not only is overall metallicity affected , but also the ratios of the abundance with respect to iron of certain elements . among these
are the and ratios so critical to uncovering the nucleosynthesis history of the early universe @xcite and deducing the age of globular clusters from the msto luminosity @xcite .
when alone is raised by 100 , modeled values of change by + 0.09dex @xcite ; changes by @xmath1dex , + 0.03dex , or + 0.16dex depending on whether the high - excitation oi 7775 triplet , the [ oi ] 6300 ground - state line , or oh molecular lines near 3100 are used @xcite . not surprisingly , the oxygen abundance in extremely metal - poor stars remains uncertain , as summarized during a workshop devoted to the problem @xcite .
@xcite and @xcite , among others , have suggested that a higher scale for metal - poor turnoff stars might eliminate the systematic differences in seen from its different diagnostics . for all these reasons , it seems important to reduce uncertainties in the determination of in msto stars .
fitting the slope of their uv flux distribution might help , since the fluxes of f
g stars peak in the optical so that uv fluxes vary rapidly with .
although fitting a uv slope is also reddening - dependent , reddening should be very low for the brightest metal - poor msto stars , which are typically within 50pc ( e.g. , * ? ? ?
reliable calculations of msto mid - uv spectra would be especially valuable in constraining the age and metallicity of a spatially unresolved old stellar system such as an elliptical galaxy or extragalactic globular cluster .
for the mid - uv flux of an old stellar population is dominated by turnoff stars @xcite , but the light redward of 4500 is dominated by cool giants in a 17 population of solar metallicity @xcite . to date , however , mid - uv studies of galaxies and globulars ( e.g. , * ? ? ?
* ; * ? ? ?
* ) have largely been empirical , comparing extragalactic spectra against observed spectra of individual f and g stars @xcite .
thus the metallicity of the extragalactic systems has generally been assumed to be solar , although age is very sensitive to this assumption @xcite .
furthermore , nonsolar abundance ratios have not been included , despite the fact that abundance enhancements of light elements such as magnesium are generally seen in metal - rich extragalactic systems @xcite .
modeling the mid - uv region is difficult , though , because of the dramatic increase in line absorption at short wavelengths .
line blending in the solar spectrum suppresses the optical continuum blueward of @xmath2 4500 @xcite , and becomes even more severe in the near- and mid - uv .
especially in the uv , many lines are `` missing '' although they appear as absorption features in spectra of solar - type stars , they lack laboratory identification , and are thus not matched by spectral synthesis calculations @xcite . as a result ,
the normalization of the solar uv continuum is controversial @xcite . with few exceptions ( e.g. , * ?
* ) , calculations of the mid - uv region are based on kurucz model flux distributions .
@xcite compared these to international ultraviolet explorer ( iue ) mid - uv spectra of nearby metal - poor halo stars , illustrating that is generally recovered when parallaxes are known and reddening negligible .
@xcite have used kurucz fluxes as generated by @xcite to examine the effect of abundance on the mid - uv light of old stellar populations , but found that the uv fluxes of cool stars were poorly reproduced .
this is not surprising given the known limitations of the opacity distribution functions on which the kurucz flux distributions are based .
they statistically treat lines in a list that includes `` predicted '' lines whose wavelengths are uncertain by @xmath210 ( because one or both energy levels have not been measured in the laboratory ) , limiting the resolution of the flux distributions to 10 .
moreover , they assume a solar iron abundance @xmath20.15 dex higher than the value currently accepted @xcite , and a solar ratio of light - element abundances = 0 .
in contrast , metal - poor stars of the galactic halo typically show = 0.3 @xcite , a factor - of - two enhancement in mg , si , ca , and ti .
this affects the fit in the mid - uv , since it includes the very strong resonance doublet of mgii at 2800 and a bound - free opacity edge of mgi at 2512 , plus many
weaker mg , si , and ti lines .
in this work we calculate the mid - uv spectrum line - by - line at high resolution , independent of the kurucz flux distributions .
our line list does not include predicted lines , but rather is based on an updated laboratory line list augmented by manual addition of missing lines where necessary . to ensure the reliability of these calculations , we match as best we can the individual line blends observed in the mid - uv echelle spectra of real near - turnoff stars , spanning a wide metallicity range and including both solar and enhanced light - element ratios .
our purpose is to reproduce mid - uv spectra of turnoff stars accurately enough to determine and to predict the composite mid - uv spectra of old stellar systems over their entire range of overall metallicity and light - element abundance ratio . in this first step ,
we describe this calibration procedure . because the calibrations depend significantly on the stellar temperatures adopted , we have carried out new determinations in seven metal - poor turnoff stars and one mildly metal - poor hotter star .
they are based on simultaneously matching both mid - uv spectra ( 2280 3120 ) and the profile and the strengths of high - excitation atomic lines in optical spectra . we thus do not rely on any photometry or parallax information in the determination .
iron abundances and ratios were adopted as derived from both weak and strong lines in the optical spectra . to assist in filling in the mid - uv line list ,
we have included mid - uv analyses of two stars with mid - uv spectra of very high resolution , the sun and 128620 (= ) , a well - studied nearby southern star with a greater - than - solar metallicity @xcite , for which was taken from the literature . in fitting each mid - uv spectrum , we adopted a single continuum normalization constant , the scale factor by which the observed mid - uv spectral flux was multiplied to match the flux predicted by theoretical calculation .
this constant proved to agree to an average of 6% @xmath3 5% with that expected from the ratio of stellar distance ( found from the parallax ) to stellar radius ( found from the model assuming a reasonable stellar mass ) .
model , , and colors all agreed extremely well with those observed , confirming errors of @xmath450 .
our mid - uv and optical echelle spectra are described in 3 . as discussed in 4 ,
was initially established from the optical spectra , and compared to the mid - uv fit .
for the most metal - poor stars , this fit did not require the addition of missing lines , but did require the use of kurucz models in which convective overshoot was turned off , in contrast to the kurucz models in general use . to reproduce mid - uv line strengths and to match the pseudocontinuum of the mid - uv spectra of more metal - rich stars , mid - uv atomic line parameters were changed and missing lines were added .
transitions especially sensitive to the surface temperature were still poorly reproduced ; this mismatch was reduced by adopting a mild enhancement of temperatures in shallow photospheric layers of the models of one - tenth solar metallicity and higher . the degree
to which the resulting calculations match both mid - uv and optical spectra of turnoff stars is shown and discussed in 5 .
the match is very good for 2900 3120 at all metallicities , and generally good throughout the mid - uv for metallicities at and below one - third solar
. it could be improved at solar metallicities by obtaining additional high - resolution mid - uv spectra of a few well - chosen standards .
our results are presented in 6 , where comparisons are made of model fluxes and colors with observations .
we conclude in 7 with a summary and a discussion of the prospects for constraining age and metallicity of quiescent stellar systems by comparing such calculations to spectra of their integrated light .
we begin by listing the properties of the stars considered in table 1 .
the first five columns following the hd number give the observed @xmath5 magnitude , colors , and the parallax from simbad .
the next five columns , discussed in 6 , show comparisons of observed properties of each star with those calculated from the models .
the parameters of the model used to calculate both mid - uv and optical spectra of each star are given in the next four columns .
in all cases but 128620 = and the sun , all the model parameters were redetermined here as described below .
the final columns show values determined previously in several works .
of the eight stars whose temperatures have been redetermined here from optical and uv spectra , seven are metal - poor main - sequence turnoff ( msto ) stars .
the eighth , 128167 , is a mildly - metal poor star lying above the halo main - sequence turnoff , either younger or a blue straggler .
it was included to give temperature leverage to the mid - uv comparisons , as discussed below .
likewise the sun and were included to represent the high - metallicity end .
the solar parameters were considered so well determined that no re - evaluation was necessary .
the star was not reanalyzed because no optical spectrum was available : it is inaccessible from the northern hemisphere .
its and abundance have been determined to moderate accuracy from extensive abundance analyses , and its gravity as = 4.33 @xmath3 0.02 as found from its visual and radial - velocity motion about its companion @xcite .
the iron abundance listed in table 1 was adopted to best match its mid - uv spectrum .
optical spectra were obtained for all stars in table 1 but the sun and .
the lick observatory shane 3 m echelle spectrograph @xcite was used with a tek @xmath6 ccd , at a fwhm resolution of 38,000 .
the spectra of 84937 , 94028 , and 106516 were kindly provided by @xcite .
exposure times on the 3 m were typically 10min for all stars except 128167 ( 40s ) .
complete spectral coverage over 3900
7800 and s / n @xmath7 100/pixel were obtained in all cases but one , although fringing compromises the accuracy of the data at the long - wavelength end of the range .
for the one exception , 114762 , the blue spectrum was obtained in two one - hour exposures by tony misch at lick with the coude auxiliary telescope , using the same echelle spectrograph configuration .
for the sun , the flux spectrum of @xcite was downloaded from the kurucz web site at http://cfaku5.harvard.edu ; the full - resolution version was matched at 150,000 resolution .
the mid - uv spectral observations of the sun modeled here are the rocket spectra of @xcite , recorded on film .
both passes of the center of the solar disk were used .
gaussian broadenings of 1.5 for macroturbulence and 80,000 fwhm for other broadening sources provided a good match .
a constant normalization scale factor of 1.10 was adopted , in keeping with the stated 10% uncertainty of its flux calibration .
the stellar uv spectra were obtained with the space telescope imaging spectrograph ( stis ) of hubble space telescope ( hst ) from mid-1998 to mid-1999 .
salient features are noted here ; full details may be found from the hst archive at http://archive.stsci.edu . to obtain the mid - uv spectra of most stars ,
the near - uv mama detector was used with the stis e230 m grating over its default range of 2280
3120 at a fwhm resolution @xmath225,000 .
106516 (= hr4657 ) was observed by @xcite in program # 7433 with the @xmath8 aperture .
mid - uv spectra for 19445 , 84937 , 94028 , 114762 , 184499 , and 201891 were acquired in snapshot program # 7402 by @xcite , with two exposures obtained for 19445 , 84937 , and 114762 .
the @xmath9 aperture was used and exposure times were 6 10min .
consequently , s / n is lower , but flux levels are less dependent on pointing .
fluxes good to 1% for these six stars are indicated by previous observations of mgii line profiles in each star , made using the hst goddard high - resolution spectrograph ( ghrs ) and the @xmath10 aperture in program # 5869 @xcite .
mid - uv spectra of the stars 128167 ( = hr5447 ) , from # 7433 , and 128620 (= hr5459 =) , from # 7263 @xcite , were recorded in multiple exposures with the e230h grating .
spectral coverage is 2380
3160 for 128167 and @xmath112280 3160 for 128620 . a fwhm resolution of @xmath260,000
was achieved with the @xmath12 aperture for 128167 , and the @xmath13nd aperture for 128620 . for neither star do fluxes always reproduce exactly in orders obtained in more than one exposure ; differences as large as 10% are seen .
similar uncertainties are anticipated for 106516 , where the @xmath12 aperture was also used .
rotation is detected in at least two stars and possibly in two more , from the extra broadening required to match line profiles . for 128167 , 106516 , 84937 , and 114762 ,
line breadths were matched in the spectral synthesis by adopting rotational velocities = 9 , 7 , 5 , and 4 respectively .
thus the resolution of the mid - uv spectrum of 128167 is effectively degraded to that of the stars observed at lower resolution , affecting the discernment of mid - uv lines .
this star nonetheless has one of the sharpest - lined spectra among those obtained in program # 7433 .
spectrum analysis was based on visually matching each observed spectrum to _ ab initio _ radiative - transfer calculations assuming static models in lte .
the program synthe @xcite , the @xcite and @xcite grids of atlas9 models , and the kurucz line lists for atomic species and molecular hydrides were downloaded from the kurucz web site , where further details on each of these can be found . as described below , all models , including that of the sun , were ultimately drawn from the @xcite grid alone .
the more metal - rich ones were modified by introducing enhanced temperatures in the shallow layers .
the synthe program accepts as input both line lists and a model atmosphere with temperature and other quantities tabulated for each depth , and calculates an array of spectra at individual emergent angles .
these may then be used individually ( as for spectra of the center of the solar disk ) or convolved with a zero or positive rotational velocity to generate an emergent flux spectrum . in either case
, the resulting spectrum can be broadened by gaussians ( or other profile shapes ) to incorporate macroturbulence and instrumental broadening .
synthe was modified to run under unix on a sun ultra-30 at the university of virginia .
mid - uv spectra from 2280 to 3160 were calculated at a resolution of 330,000 ( 500,000 for the sun and 128620 ) , and optical spectra at a resolution of 500,000 .
a gaussian macroturbulent broadening of 1.5 was assumed for all stars as well as for the sun .
spectra for some stars were rotationally broadened as listed in the last paragraph of 3 , as was the solar flux spectrum , for which a mean rotational velocity of 2 was adopted .
the spectra were then broadened by gaussians corresponding to the values of instrumental resolution listed individually in 3 . the line lists and the source code of the opacity routines available at the kurucz web site
give specifics on the wide variety of opacity sources treated by synthe .
for example , the significant absorption by the bound - free absorption edge of mgi near 2512 is treated both as line opacity , with the line list including the bound - bound transitions to @xmath14 , and as continuous opacity blueward of the bound - free edge at @xmath15 = 39759.8 @xmath16 , by the subroutine mg1op .
near @xmath15 , this routine adopts a cross section of 40 mbarn with a @xmath17 frequency dependence , based on the laboratory measurement of 46 @xmath3 12 mbarn by @xcite and the convergence of high - level lines .
where this formalism falls below the theoretical cross section , the program adopts the latter , with a cross section of 20 mbarn and a @xmath18 frequency dependence ( e.g. , * ? ? ?
* ) . for calculations redward of 5000 ,
the line list of @xcite based on the solar spectrum was adopted , with modifications due to the use here of a flux , rather than central intensity , spectrum for the sun and to the incorporation of an approximate chromosphere in the solar model .
this led to the deduction of larger gf - values for the strongest lines except where laboratory gf - values were adopted , as for low - excitation fei lines with furnace measurements @xcite .
blueward of 5000 , the kurucz 25 may 1998 atomic line lists _
gf_0300.100 , _
gf_0400.100 , and _ gf_00500.100 were used .
these differ significantly from those used in the kurucz flux distribution calculations in including only atomic lines identified in the laboratory ( with laboratory gf - values where available ) like those of @xcite but with the addition of many lines of fei , feii , and other species whose upper energy levels were newly determined in near - uv laboratory spectra ( e.g. , * ? ? ?
* ) . because the strengths or profiles of many lines were mismatched when using line parameters directly from this list , their gf - values and damping constants were changed where necessary , as described shortly .
since many lines appear in the stis mid - uv spectra that were not modeled by lines in the list , such missing lines were added blueward of 3120 .
redward of this , no missing lines were added , but gf - values were changed to match stronger lines in the sun in the regions depicted in fig .
4 below .
the , , [ fe / h ] , , and values for each reanalyzed star were initially established from and weak lines of neutral species in the optical spectra , reassessed from optical lines of once - ionized species , then confirmed or revised from the uv . for the two most metal - poor stars
, mid - uv spectra provided an immediate check on .
for the others , the following steps were iterated until all diagnostics converged .
first the profile and an assumed were used to estimate .
[ fe / h ] was adjusted until weak fei lines in the 5100
5400 region were well matched ; their strengths are largely gravity - independent . the strength of weak tii , mgi , and sii lines determined .
microturbulent velocity was set by matching weak and strong iron and titanium lines .
gravity was then checked by demanding matches for mgi , mgii , and the wings of the lines , for fei and feii , for tii and tiii , and for sii and siii .
this also constrains , for the mgii and siii doublets arise from highly excited lower levels ( @xmath78ev ) , while the optical tiii and feii lines originate from much lower levels ( @xmath113ev ) .
roughly two dozen fei lines , a dozen tii lines , and a dozen total tiii and feii lines were detected for the most metal - poor stars , with double this number typical of the more metal - rich stars .
the mgii and siii doublets at 4481 and at 6347.1 and 6371.3 were seen in all cases , but siii only marginally so in 84937 and 19445 .
once the diagnostics began to converge , iterative adjustments were made to the kurucz mid - uv atomic - line parameters .
first , gf - values were modified for atomic lines identified in the laboratory .
the relative contribution of identified lines to a blend was ascertained from the behavior of the blend in the hot star 128167 versus the cooler ones , and in the higher - gravity stars such as 19445 versus the lower - gravity ones .
damping constants were changed in some cases to match line profiles .
damping constants rather than gf - values were changed for the strong resonance lines well - observed in the interstellar medium for which gf - values are tabulated by @xcite .
next , for lines lacking identification , the laboratory line list was searched for transitions close enough in wavelength to perhaps be responsible .
the gf - value of a nearby identified line was increased by up to 4.5 in the log , and the calculation repeated .
this line was dropped if still not strong enough ; otherwise its gf - value was adjusted to match .
the high resolution of the and solar spectra showed that most such identifications were very well matched in wavelength , and thus likely to be correct , when based on increases of @xmath19 2.5 in the log .
for the two most metal - poor stars , these steps were sufficient to enable the continuum to be defined unambiguously throughout most of the mid - uv region .
its definition in more metal - rich stars was hampered by line crowding , yielding a pseudocontinuum that the modeling did not properly reproduce .
much of this pseudocontinuum mismatch appeared to be due to individual transitions missing from the line list .
this was indicated by the very common occurrence throughout the mid - uv of unmodeled absorption features found at the same wavelengths in all stars , features whose strength steadily increased in going from weak - lined to strong - lined spectra .
most of these transitions are weak , consistent with the fact that the intrinsically strongest transitions of each species are the ones most likely to have been detected in the laboratory . except just redward of the mgi edge at 2512 ,
most missing lines are probably due to iron .
its abundance is high and the spectra are rich in lines of fei and feii , which are well - populated species in solar - temperature stars . of the elements with similar first ionization potentials , only magnesium and silicon have comparable abundances ; the mid - uv spectra of their low - ionization stages
are generally much less complex , as are those of the more abundant cno elements . while iron - peak elements such as cr , mn , and ni have many transitions throughout the mid - uv , iron dominates because it is twenty or more times as abundant . in most halo stars above one - tenth solar metallicity , the relative abundances of these elements with respect to iron are within 0.2 dex of those of the sun .
the mismatch that will result from assigning missing lines of these elements to iron should be mild . in the region
immediately redward of the bound - free mgi opacity edge at 2512 , the probability is significant that missing lines may be due to magnesium rather than iron .
redward of the edge there are a very large number of transitions to known highly excited levels of the mgi atom , and presumably a very large number of similar transitions to closely - spaced highly excited levels that are as yet unidentified in the laboratory . if many weak transitions redward of 2512 are due to magnesium but have been assigned to iron instead , our calculations will underestimate the absorption in this region in metal - rich alpha - enhanced populations .
we see below that this region is poorly modeled in any case once metallicity exceeds one - tenth solar .
the modeling then proceeded iteratively by assigning tentative identifications and transition probabilities to each missing feature , then recalculating until an acceptable match was achieved or deemed unattainable with existing data ( as for the regions just redward of 2550 and 2640 in the solar spectrum ) .
features were attributed to either low - excitation fei , moderate - excitation fei , or feii , whose line strengths depend differently on stellar temperature and gravity ( pressure ) because of the boltzmann and saha equations . with a @xmath21000 higher than those of other moderately weak - lined stars , 128167 proved exceptionally useful in making these assignments .
the higher resolution of the spectra of the strong - lined stars 128620 = and the sun was essential in discerning the wavelengths of the many features emerging above one - third solar metallicity .
a brief examination was made of the gf - value changes for low - excitation fei lines , namely those identified in the laboratory with ( air ) wavelengths falling between 2600 and 3000 , and with a lower level of 6000
10000 and an upper level of 40000 50000 .
sixty - five such fei lines have gf- values from the laboratory measurements of @xcite ; 52 were not changed , while 13 were changed by from @xmath20 to @xmath21 dex .
nine lines have gf - values from the laboratory measurements of @xcite ; five were unchanged , and four changed by from @xmath22 to @xmath23dex .
thus the mid - uv gf - values found in this work generally concur with the mid - uv laboratory measurements .
in contrast , fifteen lines have theoretical gf - values from @xcite ; four were not changed , but 11 were changed by from @xmath24 to @xmath25dex , all but one decreased .
the theoretical gf - values of @xcite are seen to be subject occasionally to very large errors , usually overestimates .
this can be attributed to the finite size of the cray computer used in those calculations , which dictated that the number of levels included be truncated , and so upper levels were preferentially omitted from the model atom . because the sum of the gf - values of transitions arising from a given lower state is normalized to unity
, gf - values for the transitions included in the calculation will be overestimated because of the exclusion of others arising from the same level . thus the calculations will overestimate the gf - value by increasing amounts as the fraction decreases of the actual possible upper states that are included in the atomic model .
this fraction decreases to the blue , because higher upper levels are involved for a given lower level .
indeed , the severity of overestimates appears to increase to the blue : all the above changes made to the @xcite gf - values were @xmath26 or lower below 2650 , but redward of this only one change this negative was required .
the overestimates among the @xcite fei gf - value calculations are thus seen to be due to incompleteness and not to more fundamental difficulties .
undoubtedly many of the weaker mid - uv transitions are missing entirely from the predicted list . for the hydrides ,
separate treatment was required , as the lines are naturally associated into bands , and the kurucz hydride calculations are now several decades old . for oh , the theoretical calculations of @xcite ( communicated in august 2000 by m. bessell ) were used wherever possible . for ch , the lifbase data ( e.g. , * ? ?
* ) were used ( communicated in august 2000 by m. bessell with lines cross - identified ) . since these cover only the strongest bands
, calculations must still rely on the kurucz values for the others , notably the 4 - 3 band of oh with many weak but discernible lines in the sun .
a comparison of the kurucz oh transition probabilities against those of @xcite showed strong trends , with discrepancies increasing at high rotation number j. to place them on approximately the same scale , the kurucz transition probabilities for the oh and ch lines were decreased by an amount @xmath27 = @xmath28dex .
once missing atomic lines were largely filled in , a comparison was feasible of calculated oh and ch line strengths against those of the sun and stars .
because solar oh lines were otherwise modeled too strong , all the transition probabilities were reduced by an additional 0.15 dex uniformly . in part because weak oh lines in the sun were reasonably well modeled but strong ones were still too strong , a chromospheric mitigation of the surface temperature drop was invoked as described shortly . as illustrated below in figs .
3@xmath29 and @xmath30 , these changes suffice to reproduce both strong and weak 0 - 0 oh features in stars as well as the sun .
the calculations were first run using @xcite models , but the very metal- poor stars showed a deduced from as much as 200 higher than that from fitting the mid - uv .
that discrepancy was removed by adopting @xcite models , models in which convective overshoot has been turned off .
we did not use the alpha - enhanced @xcite models , as they are available for only a single value of the alpha - element abundance enhancement . like the @xcite models
, the @xcite models we used were calculated with = 2 and a logarithmic solar iron abundance of @xmath31 with respect to the abundance by number of hydrogen plus helium , which amounts to @xmath32 with respect to hydrogen alone .
these choices are about 1 higher in than the @xmath21 typically found for the solar flux spectrum , and about 0.15dex higher in iron abundance than indicated by the meteoric iron abundance and by several recent determinations for the sun ( e.g. , * ? ? ?
* ; * ? ? ?
thus both sets of models were calculated with higher line blanketing than expected for a solar - type star of the nominal model metallicity , since iron lines dominate the uv and optical spectrum , and the higher increases the strengths of moderately strong lines of every species .
for each star including the sun , a @xcite model was interpolated to the nearest 50 in ( 25 for the sun ) , 0.1dex in , and 0.05dex in [ fe / h ] .
the microturbulent velocity was changed at all depths to the value indicated for each star in the table and figures below . to minimize the overblanketing effect , our models were interpolated to a nominal grid abundance 0.10 0.15dex lower than that desired . to set abundances ,
the abundance of the interpolated model was then raised to the desired value and the relative abundance of iron dropped by 0.15 dex .
the specific light - element abundance elevations required by the optical observation were adopted , or assumed to be zero for the sun and , and all species equilibria were recalculated .
this procedure incorporates the specific and iron and light - element abundances set for the model in calculating both the ionization and molecular equilibria and the line and continuous opacities for its spectrum .
concerning the model structure , the procedure compensates for the effects of overblanketing due to the high choice of iron abundance , but not for the light - element ratio nor for the microturbulence .
we anticipate very mild errors in model colors and temperature structure until spectra become dominated by molecular lines or otherwise are considerably stronger - lined than any of these .
the comparison of model and observed colors ( 6 ) bears out this expectation .
for the sun , a model with = 5775 , = 4.4 , [ fe / h ] = 0 , = 0 , and = 1.0 , was interpolated in the above way . for
128620 = we adopted a similarly interpolated model with = 5800 , = 4.3 , [ fe / h ] = + 0.15dex , = 0 , = 1.0 .
while @xcite found = 5830 and [ fe / h ] = 0.25dex , most other analyses have derived somewhat lower and [ fe / h ] values . at = 5800 and = 1.0
, [ fe / h ] = 0.15dex provides a better match to the mid - uv spectrum of . even with the use of @xcite models interpolated as above , another major type of mismatch persisted in all strong - lined stars : cores of lines formed near the surface appeared too strong in the calculations , especially in extremely strong mid - uv lines such as those of mgii and mgi or feii and fei .
this is definitely not due to wayward gf - values or damping constants , because of the trend seen in the mismatch .
profiles calculated for the strongest lines in the most metal - poor spectra generally matched or were somewhat too high in the core , but were somewhat too low in the core in stars of intermediate line strength , and became too low across the the central ten ngstroms or more in the strongest - lined stars .
( see fig .
3@xmath33 below for residual trends of this nature . )
as noted above , a related mismatch was seen within the oh spectrum itself : in any given strong - lined spectrum , notably that of the sun for which the model parameters are extremely well known , weak oh lines were matched at a higher oxygen abundance than strong lines .
the possibility discussed in 5 that the solar continuum may be drawn too low would lead to a mismatch of opposite sign . to the extent that the gf - values are reliable
, then , this indicates an error in the temperature structure of the model .
the most obvious possibility is a chromosphere , which is not incorporated in flux - constant , static models such as those of synthe .
the solar chromosphere , for example , is seen in images of the limb and evident from emission reversals in the cores of strong spectral lines such as mgii .
it is generally modeled by an increase in temperature in relatively shallow layers of the atmosphere ( e.g. , * ? ? ?
* ) . any of several non - radiative heating sources extending from the photosphere through the chromosphere into the corona could be responsible ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we consequently changed the temperature of individual points in each interpolated stellar model with [ fe / h ] @xmath34 , beginning just below the point where the model becomes optically thin at 5000 ( typically at depth number 51 or 52 ) .
( we tried this for the more metal - poor models as well , but the profiles of all the strong lines became much too high near the core , so we left these models as they were . )
to best fit both optical and mid - uv spectra simultaneously , we increased the temperature at model depths 28 through 48 by 10 , 20 , 30 , 40 , 50 , 60 , 70 , 80 , 90 , 100 , 110 , 120 , 130 , 140 , 150 , 150 , 140 , 120 , 90 , 60 , and 30 respectively .
this temperature enhancement may be considered an approximate chromosphere , though it is based entirely on empirical fits of one - dimensional spectral line calculations to observed spectral line profiles .
it incorporates neither energy transport considerations nor matches to spatially resolved structures , both of which lie beyond the scope of this work .
the degree to which our mid - uv and optical spectral calculations match spectra of real stars is illustrated in the figures .
the first three show mid - uv spectra , and the last , optical . within a figure ,
panels are ordered by wavelength .
flux ( central intensity for the solar mid - uv spectrum ) is plotted versus wavelength in air . the spectrum calculated for each star ( light line )
is plotted on top of that observed ( heavy line ) . in figure 1 , the global mid - uv fit is shown for the nine stars with absolute fluxes .
all spectra have been gaussian smoothed to a fwhm resolution of 500 ( 550 for 128127 and ) .
each spectrum is displaced vertically by @xmath220% from the one below . for each star a single mid - uv continuum normalization constant
was adopted , as listed on the left . on the right
, the hd number of each star appears in bold .
below are found the model - atmosphere parameters adopted for the corresponding calculation : ( in ) , c wherever the chromosphere was modified as described at the end of 4 , log @xmath35 , [ fe / h ] , ( in ) . on the following line
appear the values in dex of the light - element abundance enhancements [ element / fe ] ( in dex ) , first that of oxygen , then those for mg , si , ca , and ti .
( the latter four are grouped together wherever they are the same . )
while the fits are generally good , serious discrepancies are noted in at the bottom of the figure near 2550 2600 and 2640 2710 , regions of relatively little line absorption in the more metal - poor spectra .
figure 2 illustrates the effect on mid - uv fits when model parameters are changed somewhat .
two of the stars of fig .
1 are depicted : the moderately weak - lined 94028 ( top two spectra ) and the rather strong - lined 184499 ( bottom three ) . star and model identifications and normalization constants are indicated as in figure 1 . the lower plot for each star represents its best fit , a replot of that star s comparison in figure 1 . comparing the bottom plot to the middle plot shows the effect for 184499 of decreasing by 0.3dex to the solar ratio .
although minimized by the adoption of a normalization constant 11% higher , the mismatch is still evident .
the strengths of the wings of the strong mgii doublet at 2800 ( and of mgi at 2852 ) are poorly reproduced , as is the overall pseudocontinuum level near and below the bound - free edge of mgi at 2512 .
similar calculations run for 94028 show a more dramatic effect in the mgii and mgi line wings , which are weak enough that little renormalization is possible , but a smaller effect near 2512 , where the h@xmath36 opacity is dominant because of the reduced magnesium abundance . comparing the plot just above the best fit for each star to the best fit
itself shows the degeneracy of the mid - uv spectra with respect to a simultaneous change in , and [ fe / h ] designed to preserve overall line strength and ionization equilibrium . for each plot just above the best fit ,
was decreased by 100 , [ fe / h ] and both concurrently decreased by 0.1dex , and the spectrum renormalized accordingly .
this change is very difficult to detect from fitting the mid - uv slope alone at these metallicities and this resolution .
the increased metallicity reduces the blue pseudocontinuum relative to the red at the same time that the increased raises it .
the change is marginally detectable in 94028 from the mismatch of low - opacity windows , but not in 184499 because the overall match is less satisfactory .
such a change does lead to a reduction in the normalization parameter , by 11% for 94028 and 17% for 184499 . in principle
this might be detected by other means , but only if is properly modeled . for as seen above , taking = 0 ( as in the kurucz flux distributions ) can largely compensate in stars such as 184499 that show a light - element enhancement .
our conclusion , then , is that if the mid - uv normalization constant is treated as a completely free parameter ( as here ) , the mid - uv flux distribution by itself can currently be used to set only in msto stars with [ fe / h ] @xmath37 .
while the mid - uv was useful in demonstrating the need for models with no convective overshoot and with an approximate chromosphere , our determinations of for the more metal - rich msto stars have relied on the optical spectra instead .
we show in fig .
4 below that the balmer line profiles and the strengths of other high - excitation lines are well matched in the optical spectra of all stars at the same as the best - fit mid - uv spectrum , and in 5 that model colors match observations as well , so that is indeed very well established here .
however , our goal of determination for turnoff stars of all metallicities from mid - uv spectra alone is not yet met .
we are not able to calculate the low - opacity windows blueward of 2900 accurately enough at near - solar metallicity , as we now demonstrate .
figure 3 plots at higher resolution and higher scale the same mid - uv spectral comparisons as are shown in figure 1 , plus that of the sun .
panel @xmath33 shows 2627.5 2630.5 , dominated by strong lines .
the other panels show regions of locally low line blanketing : @xmath38 , 2642.5 2645.5 ; @xmath39 , 2900.0 2903.0 ; @xmath29 , 3082.5 3085.5 ; @xmath30 , 3100.0 3103.0 . in all plots in fig . 3 ,
the spectrum of 184499 is offset by 30% from that of 128620 = , which in turn is offset by 50% from that of the sun , for which 1.0 marks the height of the true continuum .
the same normalization constant as in fig .
1 has been adopted for each stellar spectrum , as was a normalization for the sun 1.10 times that given in the @xcite atlas , as noted on the left .
identifications are provided at the top for the strongest lines as calculated in the solar spectrum .
manually added lines are flagged with a colon following the decimal digits of the wavelength in ngstroms .
next appears the species , the lower excitation potential in ev of atomic transitions or the band and line of molecular ones , the digits following the decimal of the residual flux in the unbroadened line of the solar spectrum calculation , and the log gf - value adopted . in fig .
3@xmath30 of the 3100.0 3103.0 region , calculated and observed spectra agree reasonably well in all stars , including and the sun .
no `` missing '' lines are strong enough to merit explicit identification at the top of the figure .
true continuum is reached in each spectrum , but fleetingly , near 3102.5 ; its placement is sensitive to the treatment of adjacent weak lines .
an unblended 0 - 0 oh line appears at 3101.230 , next to a near- continuum high point ; both are modeled well in all spectra .
the strong atomic lines are well reproduced , as are two moderately strong ch lines at 3100.120 and 3100.170 .
3@xmath29 offers a further picture of how well both continuum and oh features are modeled .
another continuum window appears near 3084.7 , but again with a weak line superimposed in the solar and spectra . the relatively unblended oh 0 - 0 band line at 3084.896 , used in several oxygen - abundance investigations , is well reproduced over the factor - of-100 metallicity range of the stars in whose spectra it is detected , including the sun itself .
nearby at 3085.2 is a blend dominated by a stronger 0 - 0 oh line ; it too is reasonably well matched in all stars .
recall that to achieve this agreement , the @xcite theoretical gf - values for oh were reduced by @xmath40dex ; otherwise these features and that of fig . 2@xmath30 , plus the overwhelming majority of other oh lines , would be too strong in the sun . that is , unless the solar continuum were raised .
the solar continuum appears to be well defined , and the oh lines of all bands matched to @xmath20.1dex , throughout the 3000 3100 region .
because there are so few continuum windows , however , we can not rule out a systematic underestimate of the continuum in and the sun of perhaps 5% or 10% . roughly speaking , raising the mid - uv continuum
would allow it to be modeled at an [ fe / h ] value more consistent with that of @xcite , while raising the solar continuum might eliminate the present need to reduce the theoretical oh gf - values by 0.15dex . as noted at the end of 6
, it would also bring the mid - uv normalization constants of the sun and closer to their expected values . a definite answer
could be obtained by deriving the stellar parameters of from an optical spectrum in the same way as for the other stars , and from modeling an improved spectrum of the solar near - uv .
random continuum variations approaching 5% occur for two , and possibly all three , of the stars that were noted in 3 as observed through a smaller aperture . for
the continuum seems appropriately placed for all stars in fig . 3@xmath30 , but in fig .
3@xmath29 the continuum of the 106516 observed spectrum appears too low , and in fig .
3@xmath39 that of 128167 may also be too low .
such variations might also be present in the 128620 = spectrum as well , but remain undetected because of the difficulty of matching its multitude of lines .
@xmath41 shows another low - opacity window at 2900 . here
the number of missing lines added remains small , and agreement is nearly as good .
oh lines are less important , being generally minor constituents of blends .
however , windows of true continuum have disappeared from the solar and spectra , and are fleeting at best in the echelle spectra of lower resolution that were acquired for mildly metal - poor stars .
continuum placement depends not so much on oh features but rather on whether all missing lines have been spotted .
clearly that is not the case at solar metallicity , where weak missing lines apparently overlap to the extent that they can not be distinguished as individuals . although the cumulative effect of their neglect on the total absorption is small , is it significant for the continuum placement , which consequently can not be determined as well as was the case above from fits such as this .
it is placed here and for the remaining panels by adopting the same normalization factors required to match figs .
3@xmath29 and @xmath30 .
by 2645 , in fig . 3@xmath38 ,
the fits have deteriorated considerably .
this is especially true at and above solar metallicity , where the numbers of missing and identified lines noted at the top are comparable .
true continuum still appears in the weakest - lined spectra near the top of the panel , but elsewhere it is gone , lost to a forest of weak lines . although the fit to the metal - poor stellar spectra is still reasonable
, the high - resolution spectra of the and the sun reveal that too few missing lines have been added , often too strong and in the wrong places . for the adjacent region from 2645 to 2700 , the sheer number of missing lines precluded their reliable identification at the low resolution of the low - metallicity stellar spectra . because many were simply omitted , the calculations overestimate the spectral flux in that region mildly in moderately metal - poor stars but seriously at and above solar metallicity , according to fig . 1 . from fig . 3@xmath33
, it is clear that these difficulties are reduced in regions dominated by strong atomic lines . there
the fit is dictated more by the ability of the calculations to match the cores and the damping wings of strong lines , and not by modeling the weak or missing features .
these are largely suppressed : although the region shown is blueward of that of panel @xmath38 , only one missing line is noted in it . the wings of the strong lines are reproduced well in the four top spectra but begin to be somewhat too broad near the core in 106516 and 201891 .
the entire @xmath31 about the core is too deep in and the sun .
this could perhaps be remedied by continuing the temperature enhancement to shallower surface layers than those indicated in 4 .
figure 4 shows fits to the optical spectra of the sun and all stars except .
identifications are given for the strongest lines in the solar spectrum .
since the optical spectra were not fluxed , the continuum was normalized to unity .
the same models were used to calculate the synthetic spectra as in figures 1 and 2 .
the wavelength regions in panels @xmath33
@xmath30 depict a variety of magnesium lines and those of other heavier elements , to illustrate the extent to which the spectral calculations match features of a wide range of excitation , ionization , and strength .
figure 4@xmath42 shows , to illustrate the satisfactory choice of .
the normalization of the profile was aided by its position near the center of an echelle order , and was done by interpolating the normalization functions of the two adjacent orders , centered @xmath375 away .
this was straightforward , for the line is much narrower than this and the continuum is very well defined by a spline fit of low order ; moreover , the normalization functions for the adjacent orders were very similar .
note that most absorption lines in this panel are telluric , not stellar , as evidenced by their shift in apparent wavelength from one star to the next .
included in fig .
4@xmath42 above the best fit for 94028 is the spectrum for the cooler 94028 model whose mid - uv spectrum is shown in fig .
this illustrates how the fit to deteriorates when models of 100 lower are adopted .
the profile is relatively insensitive to gravity in this case : calculations adopting a model that differed from the best - fit model only in having 0.2dex larger produced a profile that deviated only near 6560 and 6565 , by about the thickness of the heavy line . similar calculations showed no sensitivity of the profile to the chromospheric elevation of temperature , nor to changes of 0.3dex in iron abundance or light - element ratio .
however , @xcite models required a 200 hotter , thanks to their inclusion of convective overshoot .
panels @xmath33 @xmath30 of fig . 4 show that all the optical magnesium features are well matched using the same magnesium abundance and the same model as for the best fit to the mid - uv and to .
panel @xmath33 plots the region near the very high - excitation mgii doublet at 4481.2 ; @xmath38 , the ground - state mgi line at 4571.1 ; @xmath39 and @xmath29 , the lines at 5167.3 , 5172.7 , and 5183.6 ; and @xmath30 , the high - excitation mgi line at 5711.1 .
these panels also demonstrate that the calculations match strengths and profiles of other atomic lines regardless of species , excitation , or strength . among the lines depicted
are tii at 5173.7 , tiii at 5185.9 , ground - state fei at 5166.3 , feii at 4576.3 , and sii at 5708.4 .
these panels also contain missing features : not modeled at all in the calculations are the absorption lines seen near 4480.8 , 5170.7 , and 5181.3 .
they are not telluric , since they appear at the same position in the stellar spectra , and are especially visible in the two stronger - lined stars .
our results for the parameters of the eight metal - poor turnoff stars analyzed here are listed in columns 12 15 of table 1 .
the values found for these stars by works cited in 1 are included for reference .
our values are seen to fall in the middle of the range of previous results , preferring neither the hotter nor the cooler values .
unfortunately , the number of stars is too small and the scatter too large for more definitive conclusions .
the reliability of our values is supported by a comparison of model colors to those observed . for each star including the sun
, we interpolated the @xmath43 and colors and @xmath5 magnitudes calculated by @xcite to the , , and [ fe / h ] values of each model we used .
we have made no reddening corrections , nor any corrections for the temperature enhancement invoked in the more metal - rich models as described at the end of 4 .
the ninth , tenth , and eleventh columns of table 1 show the difference in , , and between the observed and calculated colors for each star .
the agreement is extremely good in all colors over the entire metallicity range .
the mean difference is @xmath440.010 , @xmath440.018 , and 0.000mag in , , and respectively .
the standard deviation of an individual difference is 0.010 , 0.026 , and 0.015mag , marginally larger than typical observational errors alone .
the color differences are small even for , for which both metallicity and are influential ( we interpolated to a model [ fe / h ] = + 0.05 ) , and for which we have no optical spectrum and so did not derive parameters . for the other stars , and
are sensitive primarily to @xmath45 ; their deviations correspond to @xmath345 with no allowance for observational error .
their observed colors are slightly bluer than those of models , but the size of the differences points to a scale error in of less than 50 .
we thus estimate an uncertainty of @xmath350 in each determination , from this and the poorer quality evident in figs . 2 and 4@xmath42 of the fits to mid - uv fluxes and profiles as is increased by 100 , plus the deterioration of the fit shown in fig .
4@xmath33 to mgii . from plots like those of fig . 4 at other wavelengths
, we estimate the uncertainty in [ fe / h ] to be @xmath30.05 dex at fixed , increasing to @xmath30.1 dex when the uncertainty is included .
the mid - uv slopes and profiles are only weakly sensitive to , so our results for are susceptible ( as are those of other groups relying on ionization equilibria ) to errors in gf - value scales and to possible non - lte effects .
we have not yet taken full advantage of the technique of matching damping wings of strong lines , incorporating improved treatments of hydrogen broadening such as those of @xcite and @xcite .
it would be valuable to do so given the constraints this would offer on alternative measurements of the fundamental properties of these standard stars , for example parallaxes from hipparcos and other space missions . however , the following comparisons of observed and model fluxes indicate that the values are good to @xmath30.1dex as they stand .
we first compare the uv normalization constant with that expected . according to @xcite ,
the ratio of observed to emitted stellar flux is @xmath46 , where the angular diameter @xmath47 = 2@xmath48/@xmath49 , @xmath48 being the stellar radius and @xmath49 its distance . for each star
we calculate @xmath50 in solar units as @xmath51 , assuming masses @xmath52 = 0.8 for for the metal - poor stars based on stellar isochrones @xcite and 1.15 for 128167 , and adopting 1.16 for as derived from its orbit by @xcite .
the distance in solar units is found from the parallax , and the theoretical flux normalization constant follows from @xmath53 times 4.2545 @xmath54 to convert to square steradians .
the ratio of this theoretical constant to that actually used to scale the observations in figs . 1 and 3 is listed in column 7 of table 1 .
agreement is good for all stars but one . for
, the ratio is 0.46 , suggesting an observed flux that is too low .
indeed , @xcite saw this in their 1995 mid - uv hst observations with ghrs : relative to earlier ghrs and iue observations of , `` our fluxes are lower ( by factors of 5.6 and 4.8 for the mgii and feii spectra , respectively ) , presumably because of the star s being misplaced at the edge of the small aperture . '' from a visual comparison with fig . 1 of @xcite , the fluxes we are using for the @xcite stis spectrum should be raised by a factor of 2.1 .
this brings the ratio to 0.96 , and results in an average of the nine stellar ratios of 1.06 @xmath3 0.05 , with an individual standard deviation of 0.15 . for the sun ,
the ratio is 0.91 given the scale factor of 1.10 noted above , which is within the stated normalization uncertainty but still low .
these and solar ratios would be closer to unity were the continuum levels for their spectra drawn higher by 5% 10% , a possibility raised in 5 .
as matters stand , errors in continuous opacity are limited to @xmath1110% nonetheless . the apparent @xmath5 magnitude
may also be compared to that expected from the model . from the @xmath5 magnitudes tabulated with the model colors , we calculated the difference in the stellar model @xmath5 with respect to the solar model , added this and @xmath55 to the observed solar @xmath5 = @xmath56 , then subtracted this from the observed stellar @xmath5 magnitude .
again , no reddening or other corrections were applied .
the resulting differences are shown in column 8 of table 1 .
the agreement is reasonable : the theoretical magnitudes average 0.09 @xmath3 0.06mag brighter than those observed , with a @xmath3 0.18mag individual standard deviation .
in brief , we have calculated high - resolution mid - uv spectra from first principles , after modifying line parameters and assigning line identifications for missing lines based on changes seen among spectra of stars spanning a range of well - determined values of @xmath45 , log @xmath35 , and [ fe / h ] .
the resulting calculations provide a good fit for turnoff stars of all metallicities redward of 2900 .
previous discrepancies in fitting the solar pseudocontinuum ( noted in 1 ) are attributed not to missing continuous opacity , which is good to @xmath1110% near 3100 , but rather to missing or miscalculated line opacity . to improve the accuracy of the mid - uv gf - values , damping constants , and line identifications derived here , we have determined temperatures , gravities , microturbulence , and abundances for eight stars by calculating fits to optical as well as uv spectra . for stars more metal - poor than one - thirtieth solar ,
the fit to the mid - uv spectrum was satisfactory using laboratory lines alone , and can by itself determine . for the more metal - rich stars , lines missing from the laboratory list compromise the goodness of fit of the mid - uv spectrum . a 100 increase in is seen to be difficult to detect from mid - uv spectra alone , as long as [ fe / h ] and are increased by 0.1dex to compensate .
this -[fe / h ] degeneracy parallels the age - metallicity degeneracy in elliptical galaxies @xcite . as with galaxies , it is reduced here by considering the balmer lines , with the advantage that stellar models need not include balmer - line emission nor composite populations . here was set for stars with [ fe / h ] @xmath57 by matching the line profile and the strengths of other high - excitation features . for all the metal - poor stars ,
the excellent match between calculations and observations of both optical and mid - uv spectra , and between model and observed colors , indicates an uncertainty of @xmath350 in an individual determination , and a similar uncertainty in the scale .
moreover , the stellar @xmath5 magnitudes generally agree to 0.1mag , and the mid - uv normalization constants to 10% , with those expected from the stellar angular diameter found from the observed parallax and the model , assuming reasonable masses . in this process
, it was found that these optical diagnostics and the slope of the mid - uv continuum could only be fit simultaneously when using models of @xcite in which convective overshoot has been turned off .
our work thus agrees with that of @xcite , who found that their models better reproduce the balmer profiles in the sun than do those of @xcite . downloading the @xcite models from the kurucz web site at http://cfaku5.harvard.edu
is recommended for all work where the results depend upon the temperature structure in deep continuum - forming regions .
residual discrepancies in fitting the profiles of very strong atomic lines , and the strengths of strong versus weak solar oh lines in the 3000 3100 region , are reduced here by mimicking a chromosphere towards the surface of every model with [ fe / h ] @xmath58 @xmath59 .
as described at the end of 4 , the model temperature was raised in the region just above the depths where optical continuum and weak lines are formed . while this procedure was necessary to match strong - line cores and oh line strengths over a very broad range of metallicity with a single set of gf - values , it is definitely an oversimplification .
energy and stability considerations remain to be examined .
fits might be improved with a different choice of temperature enhancement .
indeed , a two - stream model might be found necessary . in the meantime
, line cores also will be subject to uncertainties due to the presence of chromospheric emission .
@xcite demonstrate that mgii emission is always present in solar - temperature stars , with a strength that increases with increasing metallicity .
emission also increases with increasing stellar activity @xcite . since activity in turn
is associated with rapid rotation , stronger mgii emission appears commonly among young stars and close binaries .
consequently , no single quiescent chromosphere appears capable of modeling the very core of the profile of a strong line in all turnoff stars .
the reliability of our values and the good match achieved for oh lines enable a second look at the oxygen abundances of metal - poor stars , which are still uncertain as noted in 1 .
our values for are intermediate , suggesting that is not the sole cause of the oxygen discrepancies .
we defer a more detailed discussion to a future paper by @xcite , who compare oxygen abundances derived from the oh lines with those from the permitted and forbidden lines of atomic oi in these stars .
the manual addition of lines to the model line list , to match otherwise missing features seen in the observed spectrum , is clearly necessary for even an approximate representation of the mid - uv spectrum of solar - temperature stars of one - tenth solar metallicity and higher .
nonetheless the procedure has its important limitations .
one is that as overall line strength increases , the proportion of missing lines goes up dramatically ; disentangling them requires a broader range of spectra of high quality , as discussed below .
another is that the excitation or ionization of the transition could be misjudged ; the ramifications are discussed in the next paragraph .
a third is that many missing lines may not be due to iron as assumed , but to other elements .
as discussed in 4 , this is potentially serious redward of the bound - free absorption edge of mgi near 2512 . elsewhere
the effect should be minimal , because although other iron - peak elements contribute significant line absorption throughout the mid - uv , the relative abundances of all other iron - peak elements with respect to iron change little from one late f or early g main - sequence star to another in both halo and disk populations in our galaxy . to be sure ,
the list generated by this procedure can not be expected to reproduce the spectrum of a peculiar a or f star , for example a cr - enhanced star , in which the iron - peak ratio has been dramatically altered @xcite . more generally , the list is not suitable for modeling stars whose parameters lie well outside the range of those analyzed here . in such cases ,
any difference becomes important between the assigned and the actual excitation ( and ionization ) potential of the transition responsible for a line , leading to a significant misrepresentation of its strength .
consequently , the present list can not be expected to reproduce the spectra of stars hotter than early f nor cooler than early g , nor giants , as well as stars of metallicity substantially higher than solar for which lines are still missing .
the list could be made suitable for those stellar types by extending the present analysis to include such stars , provided high - resolution , high s / n mid - uv spectra are available in which mismatches induced by errors in the list can be visually identified and corrected .
we discuss this further below .
conversely , the list as it stands should do as well as indicated by fig . 1 and fig .
3 in reproducing the mid - uv spectra of any other late f or early g star of solar metallicity or less , provided a suitable model atmosphere is chosen .
this holds because once the gf - value and damping constant of a line transition are set , elemental abundance and the boltzmann and saha equilibria at each point in the atmosphere fully determine line strength in quiescent stars such as these .
dynamical effects can be ignored , and non - equilibrium effects ( if any ) are slowly - varying functions of spectral type , hence absorbed into the gf - value determinations .
the only exception is the chromospheric emission noted above .
were the line list improved to the point that a reliable match is achieved to the mid - uv spectra of f and g stars of solar metallicity and above , we could expect to use the mid - uv slope to at least constrain age and metallicity , regardless of its overall level or the appropriate light - element abundance ratio , as these could be incorporated in a grid of calculations covering the entire relevant parameter space .
spectral indexes offer an alternative means by which mid - uv spectral calculations alone might constrain metallicity and of a star or age of a stellar system .
the principle is to construct indexes based on spectral regions dominated by features whose behavior breaks the age / metallicity degeneracy and elucidates the alpha enhancement , examples of which are the balmer lines and ratios of fei / feii and mgi / mgii line strengths .
spectral indexes have indeed been used as diagnostics of age and metallicity in ellipticals by several of the groups mentioned in 1 .
for example , @xcite have compared mid - uv and optical line indexes they calculated from kurucz flux distributions against mean values observed for the stars considered by @xcite .
unfortunately , poor agreement was seen for all but two mid - uv indexes , the slope of the pseudocontinuum between 2600 and 3100 , and blended absorption near 2538 .
we should be able to improve upon some of those index calculations immediately .
as figs .
1 and 3 illustrate , the current spectral calculations should be reasonably reliable at all metallicities in spectral regions redward of 2900 , and for those dominated by strong lines blueward of that .
this includes several indexes already in use , notably the mgii and mgi lines at 2800 and 2852 and line blends in the 3000 + region .
we plan to begin such index calculations shortly , and make them generally available .
blueward of 2900 , however , our calculations are not yet reliable enough .
1 shows that our calculated spectra of solar - metallicity stars substantially overestimate the true fluxes in those wavelength regions not dominated by strong lines .
somewhat ironically , our calculations thus suggest that neither of well - behaved @xcite indexes is currently reliable at near - solar metallicity .
the problem undoubtedly is lines still missing from the line list . as seen in fig . 3
, missing lines often overlap in the solar spectrum to such an extent that they could not be individually discerned , and so their species and wavelength assignments became arbitrary .
this was true especially in the 2550 2600 and the 2640 2710 regions , where we did not complete the process .
this is the reason for the excess flux seen in the calculations at solar metallicity in these regions .
this could be remedied were spectra obtained of the same quality as that of for stars of the line strength of 184499 that span a range of properties , notably @xmath60 .
for example , the spectra of 184499 and a hyades f star with @xmath26500 should have about the same overall line strength : the hyades star s higher abundance , [ fe / h ] @xmath2 + 0.15 , is offset by its higher @xmath45 at the rate of 0.1dex per 100 , as fig . 2 illustrates .
because line crowding is much lower than in , comparing high - resolution mid - uv spectra of 184499 and a sharp - lined hyades f star could show where the strongest lines are that are missing at solar metallicity and suggest their identities from the changes observed with @xmath61 .
the and solar spectra could then be used to fill in the rest . analyzing an additional star closer in to 184499 but
with = 0 would show explicitly which transitions are due to magnesium in the 2550 region , and whether oh has been properly characterized throughout the 2900 3100 domain . at that point
, the full arsenal of indexes representing the strengths of mid - uv features could be calculated from the spectra , and their behavior with , , [ fe / h ] , and assessed explicitly .
this is a major advantage of calculations from first principles over empirical libraries , which are confined to the spectra of stars available nearby .
it is a major goal of this project .
indexes should result with significant diagnostic power for the integrated - light spectrum of an old stellar system .
we are indebted to s. allen of u. c. santa cruz for writing the initial scripts to convert the vax version of kurucz codes into unix versions .
we are grateful to m. bessell of mount stromlo and siding spring observatories for communicating the @xcite and lifbase gf - value calculations and programs for their comparison against the kurucz values , to j. valenti of the space telescope science institute for providing the 128620 spectral reductions , to t. misch of lick observatory for obtaining the blue spectrum of 114762 , and to j. fulbright of the dominion astrophysical observatory for contributing his optical echelle spectra .
we thank fulbright and r. p. kraft of lick observatory , r. l. kurucz of the harvard - smithsonian center for astrophysics , t. m. lanz and s. r. heap of the nasa / goddard space flight center , d. c. morton of the herzberg institute for astrophysics , and r. w. oconnell of the university of virginia for useful discussions . this work has made use of the simbad database , operated at cds , strasbourg , france .
support for this work was provided by nasa astrophysics data program contract s-92512-z , stsci grants go-07395 , go-07402 , and ar-8371 , and nsf grants ast 99 - 00582 and ast-0098725 , to astrophysical advances , and by nasa adp grant nag 5 - 7104 , nasa ltsa grant nag 5 - 6403 , and stsci grant go-06607 to the university of virginia .
linsky , j. l. , pagano , i. , valenti , j. , & gagne , m. 2000 , in _ cool stars , stellar systems and the sun : eleventh cambridge workshop _
garca lpez , r. j. , rebolo , r. , & zapatero osorio , m. , a. s. p. conf .
ser . , 223 , in press peterson , r. c. , & schrijver , c. j. 2000 , _ cool stars , stellar systems and the sun : eleventh cambridge workshop _ , ed
. garca lpez , r. j. , rebolo , r. , & zapatero osorio , m. , a. s. p. conf .
223 , in press plots are shown comparing observed ( heavy line ) and calculated ( light line ) spectra for the nine stars with fluxed mid - uv spectra included in this study . to the right ,
the name of each star in bold is accompanied by the model parameters used for the corresponding calculation . at the left
is the mid - uv normalization constant for each comparison .
the same as fig . 1 , but on a larger scale in 3 spectral regions .
at the bottom is the sun : two scans of the center of the solar disk are superimposed . at the very top , the strongest lines in the solar spectrum calculation are identified as described in the text .
the same mid - uv normalization constants as in fig .
1 are adopted for the top nine stars .
@xmath33 ) 2627.5 2630.5 .
@xmath38 ) 2642.5 2645.5 .
@xmath39 ) 2900 - - 2903 .
@xmath29 ) 3082.5 3085.5 .
@xmath30 ) 3100 3103 .
comparisons as in fig . 1 are shown for optical spectral regions containing magnesium or balmer lines in eight of the stars plus the sun . each observed spectrum ( heavy line )
has been normalized to unity as described in the text .
line identifications as described for fig .
3 are shown for the strongest lines in the solar calculation .
@xmath33 ) mgii .
@xmath38 ) low - excitation mgi .
@xmath39 ) , @xmath29 ) the lines .
@xmath30 ) high - excitation mgi .
@xmath42 ) . in this panel
only , the third plot from the top represents a non - optimal choice of model parameters .
it illustrates the degradation of the fit to when a is chosen 100 cooler than that of the plot immediately below .
rcccccccccccccccccccc 19445 & + 8.06 & 0.45 & @xmath62 & 0.349 & 25.85 & 1.37 & @xmath63 & @xmath64 & @xmath64 & @xmath65 & 6050 & 4.5 & @xmath66 & 1.0 & 6007 & 5842 & 6050 & 6066 & & 5825 + 84937 & + 8.33 & 0.39 & @xmath67 & 0.303 & 12.44 & 1.21 & + 0.09 & @xmath64 & @xmath64 & @xmath68 & 6300 & 4.0 & @xmath69 & 1.5 & 6314 & 6220 & & 6350 & 6202 & 6375 + 94028 & + 8.22 & 0.47 & @xmath70 & 0.343 & 19.23 & 1.07 & + 0.25 & + 0.00 & + 0.00 & @xmath65 & 6050 & 4.2 & @xmath71 & 1.3 & 6047 & 5898 & & 6060 & 5998 & 5900 + 106516 & + 6.11 & 0.45 & @xmath72 & 0.317 & 44.34 & 0.99 & @xmath1 & @xmath64 & @xmath73 & @xmath74 & 6250 & 4.3 & @xmath75 & 1.0 & 6221 & 6118 & 6208 & 6267 & 6233 & 6200 + 114762 & + 7.31 & 0.53 & @xmath76 & 0.365 & 24.65 & 0.96 & + 0.30 & @xmath77 & + 0.01 & @xmath78 & 5850 & 4.0 & @xmath79 & 1.0 & & 5790 & & 5941 & 5904 & 5800 + 128167 & + 4.46 & 0.36 & @xmath80 & 0.254 & 64.66 & 0.92 & @xmath81 & + 0.01 & @xmath64 & @xmath82 & 6850 & 4.3 & @xmath83 & 2.0 & & & 6707 & 6734 & 6737 &
+ 128620 & + 0.00 & 0.67 & + 0.23 & 0.438 & 742.2 & 0.46 & @xmath76 & + 0.00 & + 0.01 & @xmath84 & 5800 & 4.3 & + 0.15 & 1.0 & & & & & & + 184499 & + 6.62 & 0.58 & + 0.00 & 0.390 & 31.29 & 1.11 & + 0.02 & @xmath77 & @xmath77 & @xmath85 & 5750 & 4.0 & @xmath86 & 1.2 & 5734 & 5726 & 5750 & 5750 & 5773 & 5700 + 201891 & + 7.37 & 0.51 & @xmath87 & 0.358 & 28.26 & 0.92 & + 0.40 & @xmath77 & @xmath88 & @xmath78 & 5900 & 4.1 & @xmath89 & 1.0 & 6014 & 5817 & & 5974 & 5918 & 5825 + sun & @xmath56 & 0.65 & + 0.13 & & & 0.91 & & @xmath64 & @xmath63 & & 5775 & 4.44 & + 0.00 & 1.0 & & & & & & + | determining the properties of remote globular clusters and elliptical galaxies using evolutionary population synthesis requires a library of reliable model stellar fluxes .
empirical libraries are limited to spectra of stars in the solar neighborhood , with nearly solar abundances and abundance ratios .
we report here a first step towards providing a flux library that includes nonsolar abundances , based on calculations from first principles that are calibrated empirically .
because the mid - ultraviolet spectrum of an old stellar system is dominated by the contribution from its main - sequence turnoff stars , we have started by modeling these .
we have calculated mid - ultraviolet spectra for the sun and nine nearby , near - main - sequence stars spanning metallicities from less than 1/100 solar to greater than solar , encompassing a range of light - element abundance enhancements .
we first determined temperatures of eight of the stars by analyzing optical echelle spectra together with the mid - ultraviolet .
both could be matched at the same time only when models with no convective overshoot were adopted , and only when an approximate chromosphere was incorporated near the surface of relatively metal - rich models .
extensive modifications to mid - uv line parameters were also required , notably the manual assignment of approximate identifications for mid - uv lines missing from laboratory linelists . without recourse to additional missing opacity , these measures suffice to reproduce in detail almost the entire mid - uv spectrum of solar - temperature stars up to one - tenth solar metallicity , and the region from 2900 to 3100 throughout the entire metallicity range .
ramifications for abundance determinations in individual metal - poor stars and for age - metallicity determinations of old stellar systems are briefly discussed , with emphasis on the predictive power of the calculations . |
so far , @xmath0 carinae is the only luminous blue variable ( lbv ) known to be surrounded by dense molecular gas in its own ejecta nebula ( smith 2006 , 2002a ; smith & davidson 2001 ) .
nearly all other lbv nebulae show bright infrared [ fe ii ] emission from dense partially - ionized gas ( smith 2002b ; smith & hartigan 2006 ) and many have detectable dust shells ( e.g. , clark et al .
this owes to the extreme youth , high density , and extremely high mass ( more than 10 m@xmath4 ) of the bipolar homunculus nebula ( smith et al .
thus , it is a unique laboratory in which to study the behavior of molecules bathed in fuv radiation ( e.g. , ferland et al .
2005 ) .
while red supergiant ( rsg ) winds are familiar sites for molecule detection , molecular gas in the ejecta of luminous hot stars is rare .
some wolf - rayet ( wr ) stars are surrounded by molecular gas , but its origin is not always clear . the molecular gas traced by h@xmath2 and co in ngc 2359 , ngc 6888 , and the nebula around wr 134 is thought to be swept - up from the surrounding interstellar medium ( st - louis et al.1998 ; cappa et al .
2001 ; rizzo et al . 2003 ) .
several polyatomic molecules have been detected in this swept - up material around ngc2359 , including ammonia ( rizzo et al .
2001 ) . the molecular gas surrounding wr 16 , wr 18 ( ngc 3199 ) and
wr 75 ( rcw 104 ) , on the other hand , is thought to have formed from stellar ejecta in a previous rsg evolutionary phase ( marston et al .
1999 ; marston 2001 ; welzmiller et al .
in particular , hcn , hco@xmath3 , cn , and hnc have been detected in the ngc 3199 ring nebula around wr 18 ( marston 2001 ) .
the molecular gas in the homunculus , in stark contrast , is known to have formed out of ejecta from the 19th century outburst .
if a wr evolutionary phase will follow the current lbv phase of @xmath0 car , then its molecular gas may be relevant to the molecular shells around wr stars
i.e. , it hints that they may have formed in giant lbv eruptions like @xmath0 car s 19th century outburst , not just in rsg winds .
several m@xmath4 of material is present in these molecular wr nebulae as well as the homunculus ( marston et al .
1999 ; welzmiller et al . 1998 ; smith et al .
2003 ) . surrounding the homunculus are the shock - heated `` outer ejecta '' , which are ionized and well - suited for nebular abundance studies , revealing nitrogen enriched and c+o depleted material with n(n)/n(o ) of roughly 20 ( davidson et al . 1982 ; dufour et al . 1997 ;
smith & morse 2004 ) .
the chemical composition of the humongulus itself , however , is unknown because it is mostly neutral or molecular .
no co has been detected ( e.g. , cox & bronfman 1995 ) , but its near - ir emission lines of h@xmath2 are bright ( smith 2002a , 2004 , 2006 ) .
therefore , if it is composed of nitrogen - rich cno ash like the other ejecta around @xmath0 car , one might expect to find emission from ammonia .
we searched for emission from the nh@xmath1 ( j , k)=(3,3 ) inversion transition instead of the ( 1,1 ) or ( 2,2 ) transitions for two reasons .
first , h@xmath2 emission in the homunculus while narrow at any single position is spread over velocities of @xmath5500 km s@xmath6 ( smith 2002a , 2006 ) .
this would blend the ( 1,1 ) and ( 2,2 ) transitions , as they are separated by only 0.029 ghz or 367 km s@xmath6 .
second , the energy above ground for the ( 3,3 ) transition ( roughly 130 k ) is similar to the observed dust temperature of 140 k in the outer h@xmath2 skin ( smith 2006 ; smith et al .
2003 ) .
lcccc 2004 oct 30 & 1.5c & 12 h & 4.01 & 0@xmath75@xmath80@xmath74 + 2005 mar 29 & 6a & 12 h & 2.47 & 0@xmath74@xmath80@xmath73 + 2005 jul 21 & h75 & 6 h & 7.39 & 21@xmath817 +
we obtained observations of nh@xmath1(3,3 ) emission toward @xmath0 carinae using three configurations of the australia telescope compact array ( atca ) between 2004 october and 2005 june ( table 1 ) .
a bandwidth of 128 mhz was centered on 23.870 ghz , close to the rest frequency of the ammonia nh@xmath1(3,3 ) transition ( @xmath9 = 23.870129 ghz ) , and divided into 64 channels , giving an effective 2-channel velocity resolution of @xmath1050 km
s@xmath6 and velocity bandwidth of @xmath101200 km s@xmath6 . the flux was calibrated by observing pksb1934638 and adopting a value of 0.77 jy .
we estimate the final flux calibration to be better than 20% .
the bandpass and phase calibrators were pks1921 - 293 and pks b104562 , respectively .
data were edited and calibrated using the miriad software package according to standard procedures . for each individual dataset
, continuum was subtracted in the _ uv _ plane .
the resultant line data were then combined and fourier - transformed into a data cube ( @xmath11 , @xmath12 , @xmath13 ) .
each channel image was deconvolved using the clean algorithm and restored with synthesized beams of either 3@xmath72@xmath82@xmath79 ( fig . 1 ; natural weighting ) or 1@xmath70@xmath80@xmath77 ( fig
. 2 ; uniform weighting ) .
the rms noise taken from an emission - free channel is @xmath100.4 mjybeam@xmath6 per 25 km s@xmath6 . in the case of the continuum emission ,
each of the three datasets were fourier - transformed separately and the images were deconvolved using the clean algorithm .
the restoring synthesized beams as well as the values for the measured integrated total emission ( within the 10@xmath14 contour level ) are listed in table 1 .
the fluxes vary depending on the different array configurations as well as the intrinsic millimeter variability of @xmath0 car during its 5.5 yr cycle ( e.g. , white et al .
2005 ) . from the sed of @xmath0 car
, we expect a continuum flux of 23 jy at 12 mm ( brooks et al .
2005 ) , which is roughly consistent with the observed values listed in table 1 .
we do not expect the ammonia emission or absorption to vary as much as the continuum emission during the 5.5 yr cycle , because the molecular gas is shielded from the weak changing lyman continuum , and no significant change in the near - ir h@xmath2 emission is seen ( smith 2006 ) .
figure 1 shows a position - velocity diagram and spectral tracings comparing nh@xmath1 ( 3,3 ) to h@xmath2 @xmath13=10 s(1 ) from smith ( 2004 ) .
both lines have been shifted to a heliocentric velocity frame ( @xmath15=@xmath16 + 11.6 km s@xmath6 ) .
figure 2 shows the spatial distribution of ammonia compared to optical and ir dust emission ( smith et al .
2004 ; chesneau et al .
our observations are only sensitive to the brightest ammonia emission associated with the warmest dust features near the star .
we highlight three main velocity components in nh@xmath1 ( the numbers below are the numbered features in figs . 1 and 2 ) : \1 ) the brightest ammonia emission is found at + 56 km s@xmath6 in figure 1 and centered @xmath100@xmath72 e / se of the star in figure 2 , corresponding to the point where the back wall of the approaching se polar lobe meets the equator in the tightly - pinched waist of the bipolar nebula .
the ammonia emission overlaps exactly with the brightest h@xmath2 emission in velocity , and is therefore behind the continuum source ( see smith 2006 ) .
this indicates that polyatomic molecules can survive very close to the star ( within a few hundred au ) in dusty clumps seen in high resolution images ( chesneau et al .
2005 ; smith et al .
the required shielding is evidently from a column of @xmath1710@xmath18 @xmath19 and @xmath204 mag ( smith 2006 ) .
chesneau et al
. found unusual aluminum - rich dust composition in these blobs , which may be related to their molecular chemistry and the c and o depletion .
if this gas is at 140250 k ( see below ) , then the emission optical depth of this feature is 0.045 to 0.09 .
\2 ) the brightest h@xmath2 emission is seen at 73 km s@xmath6 and 1 nw of the star in figure 1a .
it is the blushifted counterpart of component 1 .
this blueshifted feature is not seen in nh@xmath1 emission , but it is seen weakly in _ absorption _ at the same velocity .
figure 2 shows that this molecular absorption coincides with the brightest equatorial dust condensations near the star ( chesneau et al .
the ammonia absorption also overlaps spatially with bright ionized blobs called the weigelt knots ( see smith et al .
these knots are known to be in the equator on the near side of the star .
thus , the ammonia molecules probably absorb the direct 12 mm free - free emission from their own ionization fronts .
the strength of the absorption feature is about 3.5 mjy ( fig .
1b ) , which would correspond to an optical depth of about 0.002 if it is absorbing the @xmath102 jy continuum source .
however , the actual optical depth depends on how much of the extended 12 mm continuum is covered by the absorbing condensations , which is very uncertain at our spatial resolution . in order for @xmath21 in component 2 to match @xmath22 for component 1
, the covering factor would be 48% .
\3 ) we also detect a third ammonia feature not seen in h@xmath2 emission , located at 110 km s@xmath6 in figure 1 and @xmath101 nw of the star in figure 2 .
its velocity clearly does not match the wall of the nw polar lobe seen in h@xmath2 .
instead , it probably originates from the so - called `` strontium filament '' ( see fig .
2a and hartman et al .
unusual low - ionization atomic emission lines like [ sr ii ] , [ ca ii ] , [ ni ii ] , etc . , are seen to be localized at this same position and velocity ( davidson et al .
2001 ; smith 2002a ; hartman et al .
the presence of nh@xmath1 emission combined with the lack of infrared h@xmath2 emission here may give important clues to the excitation of the `` strontium filament '' .
in particular , it indicates that there is warm molecular gas in the equatorial plane that is shielded from the fuv radiation which pumps the lyman - werner bands of h@xmath2 .
the fact that the nh@xmath1 emission in component 3 is found where the visual - wavelength emission lines from the `` strontium filament '' seem to end is probably due to our line - of - sight extinction this coincides with the point where equatorial material closer to the star is blocked by dust in the edge of the se polar lobe .
figure 1 may indicate a related component at positive velocities as well , at roughly + 185 km s@xmath6 .
the nh@xmath1 data also show a possible absorption feature at 500 to 550 km s@xmath6 along the line of sight to the star ( not shown ) , but it is difficult to judge the reality of this feature since it was near the edge of our observed bandpass .
we might expect absorption at that velocity from the approaching wall of the se polar lobe of the homunculus ( gull et al .
2005 ; smith 2002a , 2006 ) .
the homunculus is the only lbv ejecta nebula known to contain molecular gas bandhead emission associated with hr car ( mcgregor et al .
1988 ) , but in both cases the authors conclude that the co emission arises in the immediate circumstellar environment of the star , and not in the detached nebulae resolved in optical emission - line images . ] , and our detection of nh@xmath1 is the first detection of emission from a polyatomic molecule in the ejecta around any lbv .
the detection of ammonia raises interesting questions about how complex molecules can form rapidly in the ejecta of a hot star .
did these molecules grow directly in the gas phase , or did they accumulate on grain surfaces first ? were they expelled from icy grain mantles by shocks ?
polyatomic molecules are harder to form than simple diatomic molecules in an extreme irradiated environment , so these molecules likely formed when the ejecta were very dense and self - shielding .
the major dust formation episode happened @xmath1015 yr after the great eruption s 1843 peak , judging by the sharp drop in the visual brightness .
if molecules formed at the same time , then scaling back from the currently - observed density in the h@xmath2 skin of the homunculus ( smith 2006 ) , the density and temperature in the ejecta at that time were 10@xmath2310@xmath24 @xmath25 and 10001500 k. while bright near - ir emission from h@xmath2 has clearly been detected in the homunculus ( smith 2002a , 2006 ) , co has never been detected , either in 2.3 @xmath26 bandhead emission ( smith 2002a ) or in @xmath27co j=21 ( cox & bronfman 1995 ) .
the abscence of co combined with a clear detection of nh@xmath1 provides a qualitative indication that the molecular gas in the homunculus is indeed nitrogen rich and has been processed through the cno cycle .
this is not necessarily surprising in light of the other cno - processed ejecta around @xmath0 car ( davidson et al .
1982 ; dufour et al . 1997 ;
smith & morse 2004 ) , but it is the first actual observation that the homunculus itself is also n - rich because atomic abundances for the neutral homunculus can not be obtained easily . calculating an ammonia abundance from our current data is frought with uncertainty , which can be reduced with future observations of other ammonia transitions . for component 1 in figure 1
, we measure a total flux integrated over the line of @xmath101.08 jy km s@xmath6 , or a beam - averaged brightness temperature of @xmath28 11.3 k ( with an adopted beam filling factor of 1 ) .
the column density of nh@xmath1 ( 3,3 ) is given by ( e.g. , rohlfs & wilson 2004 ) where @xmath30=2.56@xmath810@xmath31 s@xmath6 ( ho & townes 1983 ) , and @xmath32=1.99 mhz ( 25 km s@xmath6 ) .
the observed intensity of component 1 indicates n@xmath3310@xmath34 @xmath19 . with only one transition
, we do nt know the gas temperature .
however , at @xmath35 @xmath25 in the homunculus walls ( smith 2006 ; gull et al.2005 ) , it is likely that the gas and dust temperatures are in equilibrium , so we take the range of observed dust temperature of 140250 k ( smith et al .
( we also neglect the possibility of non - thermal amplification . )
if we adopt the expression for the partition function given in equation a15 of ungerechts et al .
( 1986 ) , we find a total ammonia column density of 10@xmath36 @xmath19 assuming a temperature of 140 k , or 10@xmath37 @xmath19 assuming 250 k. the column density of h@xmath2 as we look perpendicularly through a wall of the homunculus should be several @xmath810@xmath38 @xmath19 ( smith 2006 ; ferland et al .
2005 ) . taken together
, we therefore find a rough estimate for the ammonia abundance of n(nh@xmath1)/n(h@xmath2)=(1.62)@xmath810@xmath31 .
this is almost a factor of 10 higher than ammonia abundances derived recently in a similar way for a sample of infrared dark clouds by pillai et al .
( 2006 ) , and our low estimate is higher than the highest value in their sample . because of differences in beamsize and velocity range , the limits to co emission given by cox & bronfman ( 1995 ) are not enough to place meaningful constraints on the n(nh@xmath1)/n(co ) abundance when compared with our data ; renewed attempts to detect co in @xmath0 car would be useful . unfortunately , we can not provide a quantitative estimate of the total nitrogen abundance in the ejecta from ammonia observations alone , as it depends upon grain chemistry , whether or not ammonia has been liberated from icy grain mantles , and the ammonia survival probability at various locations in the nebula .
furthermore , most of the nitrogen in the molecular zones is expected to be locked up in unobservable n@xmath2 molecules ( ferland et al .
. it will be interesting if future observations of additional transitions with higher sensitivity can constrain the ammonia abundance as a function of position throughout the homunculus , and if future observations can detect additional molecules that we might expect in nitrogen - rich molecular ejecta .
for instance , n@xmath2h@xmath3 may be expected as part of the path to nh@xmath1 .
we thank bob sault for assistance with the october 2004 observations , and e. keto for helpful discussions .
the july 2005 data were obtained through director s discretionary time .
we also thank the atnf distinguished visitor program , providing us with an opportunity to collaborate on this project in australia .
n.s . was supported through grant hf-01166.01a from the space telescope science institute , operated by the aura , inc .
, under nasa contract nas 5 - 26555 .
rizzo , j.r .
, martin - pintado , j. , & desmurs , j.f .
2003 , in iau symp .
212 , a massive star odessey , from main sequence to supernova , ed .
van der hucht , a. herrero , & c. esteban ( san francisco : asp ) , 740 | we report the first detection of ammonia in the homunculus nebula around @xmath0 carinae , which is also the first detection of emission from a polyatomic molecule in this or any other luminous blue variable ( lbv ) nebula .
observations of the nh@xmath1 ( j , k)=(3,3 ) inversion transition made with the australia telescope compact array reveal emission at locations where infrared h@xmath2 emission had been detected previously , near the strongest dust emission in the core of the homunculus .
we also detect ammonia emission from the so - called `` strontium filament '' in the equatorial disk .
the presence of nh@xmath1 around @xmath0 car hints that molecular shells around some wolf - rayet stars could have originated in prior lbv eruptions , rather than in cool red supergiant winds or the ambient interstellar medium . combined with the lack of any co detection , nh@xmath1 seems to suggest that the homunculus is nitrogen rich like the ionized ejecta around @xmath0 car .
it also indicates that the homunculus is a unique laboratory in which to study unusual molecule and dust chemistry , as well as their rapid formation in a nitrogen - rich environment around a hot star .
we encourage future observations of other transitions like nh@xmath1 ( 1,1 ) and ( 2,2 ) , related molecules like n@xmath2h@xmath3 , and renewed attempts to detect co. |
our results have impact on three fields : model theory , universal algebra , and theoretical computer science .
we therefore have a three - fold introduction . in model theory , reducts of a relational structure @xmath0
are usually considered _ up to first - order interdefinability_. to be more precise , one considers the reducts of the expansion of @xmath0 by all first - order definable relations , and two reducts @xmath1 and @xmath2 are considered to be the same if and only if there is a first - order definition of @xmath1 in @xmath2 and vice versa .
it is well - known that there is a close connection between classifications of reducts up to first - order interdefinability and the theory of infinite ( closed ) permutation groups . in 1976
, cameron @xcite showed that the highly set - transitive permutation groups are exactly the automorphism groups of the reducts of @xmath3 .
he also showed that there are exactly five reducts of @xmath3 up to first - order interdefinability . recently ,
junker and ziegler gave a new proof of this fact , and showed that @xmath4 , the expansion of @xmath3 by a constant @xmath5 , has 116 reducts @xcite .
thomas showed that the first - order theory of the random graph also has exactly five reducts @xcite and conjectured that every structure with quantifier elimination in a finite relational signature has a finite number of reducts , up to first - order interdefinability . in this article
, we initiate the systematic study of reducts up to _ primitive positive interdefinability_. a first - order formula is _ primitive positive _ if and only if it is of the form @xmath6 for atomic formulas @xmath7 . clearly , classifications up to primitive positive interdefinability are harder to obtain , compared to first - order interdefinability , since there are far more reducts to distinguish .
the simplest structure where such a classification can be studied is the structure that has no structure at all except for the equality relation , which is considered part of first - order logic .
our result classifies the reducts of the structure @xmath8 , where @xmath9 is a countably infinite set , up to primitive positive definability . while this might look trivial at first sight , we prove that the number of such reducts equals the continuum .
it turns out that the partial order ( in fact , the lattice ) of these reducts is quite complex , but can be described reasonably well .
we remark that the classification of the reducts of @xmath10 up to primitive positive interdefinability is the same as the classification for a structure of the form @xmath11 , where @xmath12 is a set of arbitrary infinite cardinality .
to show our classification result , we use a concept that is called _
clone _ in universal algebra .
clone _ is a set of operations on some fixed domain that is closed under compositions and contains all projections .
the clones we are interested in are moreover _ locally closed _ , which means that they are , similarly to closed permutation groups used for the study of reducts up to first - order interdefinability , closed sets in the natural topology on the operations . for @xmath13-categorical structures @xmath0 , there is a one - to - one correspondence between the reducts of @xmath0 and the locally closed clones that contain the automorphisms of @xmath0 . in order to classify the reducts of @xmath10 , we thus classify those locally closed clones which contain all permutations of the domain .
we believe that several of the results and techniques presented in this article can be used to also classify reducts of other @xmath13-categorical structures up to primitive positive definability .
such future classifications are likely to combine existing results on first - order closed reducts , such as the above - mentioned results on the dense linear order and the random graph , with the results of the present paper . for universal algebraists ,
the title of our paper could be `` the locally closed clones above the permutations '' .
in fact , in this article we describe the lattice of all locally closed clones on a countably infinite set that contain all permutations of this set .
traditionally , most work on clones was done for clones on a finite domain , and only occasionally clones on infinite sets were studied .
however recently , the use of methods from mathematical logic , in particular from set theory , has allowed for the production of a considerable number of new results on clones over infinite sets .
classical results on clones , mostly on finite domains , can be found in the books @xcite , @xcite , and @xcite , while the recent survey paper @xcite contains much of what is known on clones over infinite sets .
the clones on a finite domain which contain all permutations were completely determined in @xcite ; it turns out that the number of such clones is finite .
if one moves on to infinite domains , there are two possible clone lattices which can be considered : the lattice of local clones , which is closer to the clone lattice on a finite domain in both size and methodology that can be employed for its study , and the lattice of all clones , the investigation of which often involves set - theoretical constructions ( as in @xcite , for example ) . in the full clone lattice over an infinite set , the interval of those clones which contain all permutations has been subject to investigation . for example , its dual atoms have been completely described on domains of regular cardinality ( @xcite , @xcite ) . moreover , its atoms were determined for all infinite domains in @xcite .
however , the number of clones above the permutations has been shown to be huge ( @xcite , @xcite , @xcite ) , and it seems unlikely that all such clones on an infinite set will ever be classified .
as our results will show , the situation is quite different for local clones which contain the permutations .
these clones had not yet been subject to explicit investigation and were largely unknown except for a few implicit results , such as partial results on local clones containing all unary operations in @xcite .
it turns out that although the lattice of locally closed clones which contain the permutations is uncountable , one can obtain a reasonable understanding of it .
we remark here that the lattice of all local clones has been shown to be quite complicated ; in particular , already on a countably infinite base set , the class of lattices which are completely embeddable into it properly contains the class of all algebraic lattices with a countable number of compact elements ( @xcite , @xcite ) .
the satisfiability problem for propositional logic is one of the most fundamental problems in computer science , and among the hardest problems in the complexity class np .
many restrictions of this problem stay np - hard
e.g. the problems not - all - equal-3sat and one - in-3-sat . in a groundbreaking paper in 1978 , schaefer studied a large class of natural restrictions of the satisfiability problem @xcite . for a fixed structure @xmath0 over the boolean domain @xmath14 , the computational problem is to decide whether a given sentence of the form @xmath15 is true , where @xmath16 is of the form @xmath17 for a @xmath18-ary relation @xmath19 in @xmath0 .
this problem is now known as csp@xmath20 , the _ constraint satisfaction problem _ for the _ constraint language _ @xmath0 .
( it is easy to see that 3sat , not - all - equal-3sat , and one - in-3-sat can be modelled as csp@xmath20 for appropriate boolean constraint languages @xmath0 . )
schaefer proved that for boolean constraint languages @xmath0 the problem csp@xmath20 can either be solved in polynomial time , or is np - complete , and he gave a complete description of the tractable and the hard boolean constraint languages .
schaefer s theorem inspired many other complexity classification projects for computational problems that are ` parametrized ' by a constraint language , and similar results have been obtained not only for constraint satisfaction , but also for other computational tasks such as _ quantified constraint satisfaction _ , _ learning of propositional formulas _ , _ counting solutions _ , and the _ maximum solution problem_. it turns out that for all these computational problems the computational complexity does not change if we add relations to the constraint language @xmath0 that have a primitive positive definition in @xmath0 .
this implies that the complexity of these problems only depends on the so - called _ clone _ consisting of all _ polymorphisms _ of the constraint language , i.e. , the set of those operations that preserve all relations in @xmath0 .
the lattice of clones on a two - element set was determined already quite some time before schaefer , namely by post in 1941 @xcite . and indeed , schaefer s classification can be formulated in terms of post s lattice , see e.g. @xcite .
our result can be viewed as an analog of post s lattice for the logic of equality ( rather than for propositional logic ) . in this analogy
, boolean relations correspond to relations that have a first - order definition in @xmath8 , where @xmath9 is an arbitrary infinite set .
sets of such relations have been called _ equality constraint languages _ @xcite . hence , in this paper we study the lattice of polymorphism clones of equality constraint languages .
note that equality constraint languages are exactly those constraint languages that are preserved by all permutations of the domain ; in this sense , we study those constraint languages up to primitive positive interdefinability that have the maximal possible degree of symmetry .
the constraint satisfaction problem for equality constraint languages has been investigated in @xcite .
it was shown that csp@xmath20 is np - complete if the polymorphism clone of the constraint language is the smallest element in the lattice ; otherwise , it is polynomial - time solvable .
the _ quantified constraint satisfaction problem ( qcsp ) _ for equality constraint languages @xmath0 has been investigated in @xcite .
we make the notions of the introduction precise , and fix some definitions and notation from logic and universal algebra .
moreover , we explain how to translate our classification problem from logic to algebra and vice - versa : that is , we demonstrate in what way reducts correspond to local clones .
consider the relational structure consisting of a countably infinite domain and the relation of equality only .
for notational simplicity , we will assume that this domain is the set @xmath21 of natural numbers . by a _
reduct _ of @xmath22 we mean a structure @xmath23 , where @xmath24 is a set of finitary relations on @xmath21 all of which can be defined by a first - order formula in the language of equality .
let @xmath25 be a first - order formula over the relational signature @xmath26 .
then @xmath25 is called _ primitive positive _ or @xmath27 iff it is of the form @xmath28 , where @xmath29 are atomic @xmath26-formulas . for a set @xmath24 of relations on @xmath21 , we write @xmath30 for the set of all relations on @xmath21 which can be defined from @xmath24 by pp formulas ( over the signature appropriate for @xmath24 ) .
we say that a structure @xmath31 is _ pp - definable _ from a structure @xmath32 iff @xmath33 .
we say that @xmath34 are _ pp - interdefinable _ iff they are pp - definable from each other .
the structure @xmath35 is _ pp - closed _
iff @xmath36 .
the notion of pp - definability imposes a quasiorder on the reducts of @xmath37 , and the relation of pp - interdefinability is an equivalence relation on this order .
we are interested in the partial order which is obtained by factoring the quasiorder of reducts by this equivalence relation .
that is , we consider two reducts the same iff they are pp - interdefinable , or equivalently , we consider the partial order of pp - closed reducts .
in fact , the pp - closed reducts form a complete lattice , where the meet of a family of structures @xmath38 is the structure @xmath39 , and the join is @xmath40 .
fix a countably infinite base set ; for convenience , we take the set @xmath41 of natural numbers . for
all @xmath42 , denote by @xmath43 the set @xmath44 of @xmath45-ary operations on @xmath41 .
then @xmath46 is the set of all finitary operations on @xmath41 .
a _ clone _
@xmath47 is a subset of @xmath48 satisfying the following two properties : * @xmath47 contains all projections , i.e. , for all @xmath49 the operation @xmath50 defined by @xmath51 , and * @xmath47 is closed under composition , i.e. , whenever @xmath52 is @xmath45-ary and @xmath53 are @xmath54-ary , then the operation @xmath55 defined by @xmath56 also is an element of @xmath47 .
since arbitrary intersections of clones are again clones , the set of all clones on @xmath41 , equipped with the order of inclusion , forms a complete lattice @xmath57 . in this paper , we are interested in certain clones of @xmath57 which satisfy an additional topological closure property : equip @xmath41 with the discrete topology , and @xmath58 with the corresponding product topology ( tychonoff topology ) , for every @xmath42 .
a clone @xmath47 is called _ locally closed _ or just _ local _ iff each of its @xmath45-ary fragments @xmath59 is a closed subset of @xmath43 .
equivalently , a clone @xmath47 is local iff it satisfies the following interpolation property : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for all @xmath42 and all @xmath60 , if for all finite @xmath61 there exists an @xmath45-ary @xmath52 which agrees with @xmath62 on @xmath63 , then @xmath64 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ again , taking the set of all local clones on @xmath41 , and ordering them according to set - theoretical inclusion , one obtains a complete lattice , which we denote by @xmath65 : this is because arbitrary intersections of clones are clones , and because arbitrary intersections of closed sets are closed .
the join of a family @xmath66 is calculated as follows : first , consider the set of all operations on @xmath41 which can be obtained by composing operations from @xmath67 ; this set is a clone , but might not be locally closed .
for this reason , the topological closure of this set has to be formed in addition in order to arrive at the join in @xmath65 . for a set of operations @xmath68 , we write @xmath69 for the smallest local clone containing
@xmath70 ; if @xmath47 is a local clone and @xmath71 , then we say that @xmath47 is _ generated _ by @xmath70 .
observe that @xmath69 is just the intersection of all local clones containing @xmath70 , or equivalently the topological closure of the set of term operations that can be built from @xmath70 .
using this notation , the join of the family @xmath66 is simply @xmath72 .
this paper investigates , and in some sense classifies , those local clones on @xmath41 which contain all permutations of @xmath41 .
it turns out that the number of such clones is uncountable .
let @xmath73 and let @xmath74 be a relation .
we say that @xmath75 _ preserves _
@xmath76 iff @xmath77 whenever @xmath78 , where @xmath79 is calculated componentwise .
this notion of preservation links finitary relations on @xmath41 to finitary operations and is the prime tool for studying the reducts of @xmath37 , and indeed of all other @xmath13-categorical structures , up to primitive positive interdefinability . for a set of relations @xmath24
, we write @xmath80 for the set of those operations in @xmath48 which preserve all @xmath81 .
the operations in @xmath80 are called _ polymorphisms _ of @xmath24 .
the following is folklore in universal algebra , see e.g.@xcite .
[ prop : polrislocalclone ] @xmath80 is a local clone for all sets of relations @xmath24 .
moreover , every local clone is of this form .
we have seen how to assign sets of operations to sets of relations ; likewise , we can go the other way . for an operation @xmath73 and a relation @xmath74
, we say that @xmath76 is _ invariant _ under @xmath75 iff @xmath75 preserves @xmath76 .
given a set of operations @xmath68 , we write @xmath82 for the set of all relations which are invariant under all @xmath83 .
the following is well - known .
[ prop : invfisppclosed ] @xmath82 is a pp - closed set of relations for every @xmath68 . using the galois connection defined by the operators @xmath84 and @xmath85 which link operations and relations , we obtain the following well - known alternative for describing the hull operator which assigns to a set of operations the local clone which this set generates ( confer @xcite ) . [ prop : localclosure ] let @xmath68
. then @xmath86 .
the reader will now expect that our notion of closure for structures @xmath0 , namely closure under primitive positive definitions , will coincide with the closure operator @xmath87 of our galois connection .
although this is not true for infinite structures in general , the following theorem from @xcite states that it holds if @xmath0 is not too far from the finite .
( in the following , we often identify structures @xmath0 with their sets of relations , e.g. , when writing @xmath88 . )
[ thm : ppdefinability ] let @xmath0 be @xmath13-categorical
. then @xmath89 . ordering locally closed permutation groups ( just like clones , a permutation group is called locally closed iff it is a closed subset of @xmath48 ) by set - theoretical inclusion , one obtains a complete lattice since arbitrary intersections of local permutation groups are local permutation groups again .
similarly , the first - order closed structures on @xmath41 form a complete lattice . using our notation
, we can state the well - known connection ( see e.g. @xcite ) between locally closed permutation groups and first - order closed reducts of @xmath13-categorical structures as follows .
[ thm : autinv ] let @xmath0 be an @xmath13-categorical structure .
the mapping @xmath90 sending every locally closed permutation group @xmath91 containing the automorphism group @xmath92 of @xmath0 to its set of invariant relations is an antiisomorphism between the lattice of such groups and the lattice of those reducts of @xmath0 which are closed under first - order definitions .
utilizing local clones , we obtain an analogous statement to theorem [ thm : autinv ] for reducts up to primitive positive interdefinability .
[ thm : galoisconnectionppreductslocalclones ] let @xmath0 be an @xmath13-categorical structure .
then we have : 1 .
the operator @xmath84 maps every reduct of the first - order expansion of @xmath0 to a local clone above @xmath92 .
two reducts @xmath93 are mapped to the same local clone if and only if they are equivalent with respect to primitive positive interdefinability .
every local clone above @xmath92 is the polymorphism clone of a reduct of @xmath0 .
4 . the mappings @xmath84 and @xmath85 are antiisomorphisms beween the lattice of local clones above @xmath92 and the lattice of those reducts of the first - order expansion of @xmath0 which are closed under primitive positive definitions .
( 1 ) : let @xmath1 be the expansion of @xmath0 by all first - order definable relations and let @xmath94 be a reduct of @xmath1 .
its is easy to see and well - known that @xmath95 .
thus @xmath96 .
by proposition [ prop : polrislocalclone ] , @xmath88 is a local clone . + ( 2 ) :
if @xmath97 , then @xmath98 by theorem [ thm : ppdefinability ] . on the other hand , if @xmath99 , then @xmath100
thus , @xmath101 by proposition [ prop : localclosure ] , and hence @xmath102 .
theorem [ thm : ppdefinability ] then implies @xmath103 .
+ ( 3 ) : given a local clone @xmath104 , consider @xmath105 . by proposition
[ prop : localclosure ] we have @xmath106 .
it remains to show that @xmath94 is a reduct of @xmath0 , i.e. , @xmath107 , where @xmath1 is the first - order expansion of @xmath0 .
but this is obvious , since @xmath104 implies @xmath108 .
+ ( 4 ) : by ( 1 ) , ( 2 ) , and ( 3 ) , @xmath84 is a bijective mapping from the pp - closed reducts of @xmath0 onto the local clones containing @xmath92 .
it is obvious from its definition that this mapping is antitone . by theorem [ thm : ppdefinability ]
, the restriction of @xmath85 to the local clones containing @xmath92 is the inverse of @xmath84 .
the above theorem tells us that classifying reducts of @xmath13-categorical structures up to primitive positive interdefinability really amounts to understanding the lattice of local clones containing its automorphisms ; in our case , since clearly all permutations are automorphisms of @xmath37 , we have to investigate local clones containing all permutations of @xmath41 .
via the galois correspondence inv pol , it is possible to describe the reducts of @xmath37 either on the relational or on the operational side
. we will now give the classification on the clone side in as much detail as is gainful at this point of the paper ; since we still want to spare the reader the technical details , we will have to be informal at times .
an operation is called _ essentially unary _ iff it depends on at most one of its variables .
unary clone _ is a clone all of whose operations are essentially unary .
clearly , unary clones are just disguised monoids of transformations on @xmath41 .
the first part of our result describes the locally closed unary clones containing the set @xmath109 of all permutations of @xmath41 .
[ thm : unaryclones ] let @xmath110 be the lattice of locally closed unary clones containing @xmath109 .
then : 1 .
@xmath110 is countably infinite .
@xmath110 is isomorphic to the lattice of order ideals of a certain partial order @xmath111 on the set of finite increasing sequences of positive natural numbers .
3 . in particular , both @xmath110 and its inverse order are algebraic ( that is , isomorphic to the subalgebra lattice of an algebra ) , and @xmath110 is a distributive lattice .
4 . @xmath110 is a well - quasi - order , that is , there exist no infinite descending chains and no infinite antichains in @xmath110 .
( in fact , @xmath110 is a better - quasi - order . )
all elements of @xmath110 are finitely generated over @xmath112 , i.e. , for every @xmath113 there exists a finite set @xmath68 such that @xmath114 . theorem [ thm : unaryclones ] will be proven in section [ sect : monoids ] ; it is also there that the partial order @xmath111 is defined .
we remark that @xmath110 is an exceptionally well - behaved part of @xmath65 , since @xmath65 is far from satisfying any statement of the theorem : @xmath115 , @xmath65 is not algebraic ( not even upper continuous ) , and @xmath65 does not satisfy any non - trivial lattice identities @xcite . having described the unary clones , we proceed as follows : consider any unary clone .
then this clone has only essentially unary operations , and therefore differs only formally from a monoid @xmath116 of transformations .
now it turns out that the set of all local clones @xmath47 which have @xmath116 as their unary fragment , i.e. , which satisfy @xmath117 , forms an interval @xmath118 of the lattice @xmath65 ; intervals of this form are called _ monoidal_. the smallest element of @xmath118 is the unary clone we started this argument with , namely the clone of those essentially unary operations which are `` elements '' of @xmath116 ; this clone is just @xmath119 .
the largest element of @xmath118 is called @xmath120 and contains all @xmath121 satisfying @xmath122 whenever @xmath123 .
( this notation is consistent with our previous use of @xmath84 , if one thinks of the elements of @xmath116 as @xmath41-ary relations . )
clearly , the monoidal intervals constitute a natural partition of @xmath65 .
our strategy for describing the local clones containing @xmath109 is to determine the monoidal interval @xmath118 for each locally closed monoid @xmath116 containing @xmath109 ; confer figure [ figure : monoidalintervals ] . ]
the following theorem describes @xmath118 for all monoids @xmath116 which contain an operation which is neither injective nor constant .
we refer the reader to section 5 , which contains the proof of the theorem , for the definition of quasilinearity .
[ thm : monoidalintervalsoflargemonoids ] let @xmath116 be a locally closed monoid containing @xmath109 as well as a non - constant and non - injective operation .
then : 1 . if @xmath124 , then @xmath118 is a chain of order type @xmath125 with largest element @xmath48 . + its smallest element is the clone @xmath126 of all essentially unary operations . + its second smallest element is burle s clone @xmath127 of all operations which are either essentially unary or quasilinear . + for @xmath128 , its @xmath45-th smallest element is the clone @xmath129 of all operations which are either essentially unary or whose range contains less than @xmath45 elements . 2 . if @xmath130 , then there exists a maximal natural number @xmath131 such that @xmath116 contains all unary operations which take at most @xmath18 values . + if @xmath132 , then @xmath118 has only one element @xmath119 .
+ otherwise , @xmath118 is a finite chain of length @xmath133 , and : + its smallest element is @xmath119 . + its second smallest element @xmath134 consists of @xmath119 plus all quasilinear operations . + for @xmath135 , its @xmath45-th smallest element @xmath136 consists of @xmath119 plus all operations whose range is smaller than @xmath45 .
we now turn to the monoid @xmath137 locally generated by @xmath109 .
this , as a quick check shows , consists of all injections in @xmath138 .
its monoidal interval is the hardest to understand , and we need a few definitions before stating the theorem describing it .
[ def : hornclone ] let @xmath139 be the set of operations which are , up to fictitious variables , injective .
[ defn : barclone ] let @xmath140 and let @xmath141 .
if there exists an injection @xmath142 such that @xmath143 for all @xmath144 and @xmath145 for all @xmath146 , then we call @xmath75 a _ @xmath18-bar function_. let @xmath147 be the clone generated ( in the sense of [ sect : notation ] ) by any bar function , i.e. , the smallest local clone containing that bar function and all permutations of @xmath41 ( we will see in section [ subsect : bar ] that this definition makes sense ) .
[ def : injectiveinonedirectionandrichard ] let @xmath148 .
we call an operation @xmath73 _ injective in the @xmath149-th direction _ iff @xmath150 whenever @xmath151 and @xmath152 .
we say that @xmath73 is _ injective in one direction _
iff there exists @xmath148 such that @xmath75 is injective in the @xmath149-th direction .
let @xmath24 be the set of all operations which are injective in one direction .
[ def : oddclone ] let @xmath153 any operation satisfying the following : * @xmath154 , @xmath155 , @xmath156 , and * for all other arguments , the function arbitrarily takes a value that is distinct from all other function values .
we set @xmath157 to be the clone generated by @xmath158 , i.e. , the smallest local clone containing @xmath158 and @xmath109 .
the following theorem summarizes the highlights of the monoidal interval corresponding to the monoid @xmath137 generated by @xmath109 ; confer figure [ figure : monoidalintervali ] .
more detailed descriptions of the clones of the theorem as well as other clones in that interval can be found in section [ sect : infiniterangeops ] .
[ figure : monoidalintervali ] ] [ thm : monoidalintervalofs ] the monoid locally generated by @xmath109 is the monoid @xmath137 of injections , and : 1 .
the largest element of @xmath159 , @xmath160 , equals @xmath161 , where @xmath162 is the ( binary ) inequality relation .
@xmath139 is the unique cover of @xmath163 in @xmath159 , and all elements of @xmath159 except @xmath163 contain @xmath139 .
moreover , @xmath139 is generated by any binary injection , and @xmath164 consists of all relations definable by a horn formula .
@xmath147 is the unique cover of @xmath139 in @xmath159 , and all elements of @xmath159 except @xmath163 and @xmath139 contain @xmath147 .
@xmath165 are incomparable clones in @xmath159 and every clone in @xmath159 is either contained in @xmath157 or contains @xmath24 .
the number of elements of @xmath159 containing @xmath24 equals the continuum : in fact , the power set of @xmath13 , ordered by reverse inclusion , has an order embedding into the interval @xmath166 $ ] .
in particular , the same holds for the interval @xmath159 , as well as for the set of local clones above @xmath109 .
the last statement of theorem [ thm : monoidalintervalofs ] is among the hardest to prove in this paper , and has strong consequences for pp classification projects , so that it deserves an own corollary .
let @xmath0 be any relational structure .
then the number of its pp - closed reducts is uncountable .
in fact , there exists an order embedding of the power set of @xmath13 into the lattice of pp - closed reducts of @xmath0 .
it is for this reason that we can not expect to completely characterize the pp - closed reducts of any relational structure . in our case
, we obtain a complete characterization of the closed monoid lattice and of all monoidal intervals except for those corresponding to the monoids @xmath137 and @xmath167 ( see below ) , where we must content ourselves with some insights on the structure of those intervals .
it remains to describe the monoidal intervals of those monoids which contain all injections , and some constant operations .
clearly , there is only one such monoid , namely the monoid @xmath167 consisting of all constants and all injections . in general , for any set of operations @xmath68 , write @xmath168 for @xmath70 plus all constant operations , and @xmath169 for @xmath70 without all constant operations .
it turns out that @xmath170 is a complete sublattice of @xmath171 , as described in the following theorem : [ thm : monoidalintervalofsplusconst ] let @xmath167 be the monoid of all injective and of all constant operations .
then : 1 .
if @xmath172 , then @xmath173 is a local clone in @xmath159 .
2 . @xmath174 and @xmath175 .
3 . the mapping from @xmath170 into the subinterval @xmath176 $ ] of @xmath159 which sends every clone @xmath47 to @xmath173 is a complete lattice embedding which preserves the smallest and the largest element .
if @xmath177 , then @xmath178 iff @xmath179 .
5 . for all @xmath180 which do not contain @xmath24 , @xmath181 is a local clone in @xmath170 .
all clones @xmath47 in @xmath170 are of this form , as @xmath182 .
@xmath184 is the unique cover of @xmath185 in @xmath170 .
@xmath186 is the unique cover of @xmath184 in @xmath170 .
we remark that the mapping that sends every @xmath172 to @xmath173 is not surjective onto the interval @xmath176 $ ] ; see the remark after proposition [ prop:+notinjective ] .
the following sections contain the proof of our result , and of course more detailed definitions of the structures involved .
each theorem corresponds to a section : theorems @xmath187 , @xmath188 , @xmath189 and @xmath190 are proven in sections @xmath191 , @xmath192 , @xmath193 , and @xmath187 , respectively .
in addition to the notation introduced so far , we establish the following conventions .
if @xmath121 , since we are only interested in local clones containing @xmath109 , we abuse the notation @xmath194 and write @xmath195 for the local clone generated by @xmath75 together with @xmath109 .
for @xmath196 , we say that @xmath75 _ generates _
@xmath62 iff @xmath197 .
similarly for @xmath198 and @xmath68 , we say that @xmath70 _ generates _
@xmath62 iff @xmath199 .
if @xmath121 , we write @xmath200 for its range . for a relation @xmath19 , we will usually write @xmath201 instead of @xmath202 for the set of all operations which preserve @xmath19 . if @xmath5 is an @xmath45-tuple
, then we refer by @xmath203 to the @xmath149-th component of @xmath5 , for all @xmath148 .
we prove theorem [ thm : unaryclones ] describing all unary clones containing @xmath109 .
recall that a unary clone consists only of operations depending on at most one variable , and is therefore a disguised monoid of transformations . for convenience
, we therefore only deal with unary operations and monoids in this section . in particular , we adjust the meaning of certain notations for this section : for example , @xmath69 refers to the local _ monoid _ ( rather than the local clone ) generated by a set @xmath204 together with @xmath112 .
the various statements of theorem [ thm : unaryclones ] are obtained in propositions [ prop : unary : isotoideallattice ] , [ prop : unary : ideallatticewqo ] , [ prop : unary : finitelgenerated ] and corollaries [ cor : unary : latticedistributive ] and [ cor : unary : cloneswqo ] . in a first lemma
, we officially state what we already observed in the last section , namely that @xmath112 locally generates all unary injections .
let @xmath205 be a locally closed monoid .
then @xmath116 contains the monoid @xmath137 of all unary injective operations .
clearly , on any finite set every injection can be locally interpolated by a suitable permutation .
the following lemma implies that except for the full transformation monoid @xmath138 , all closed monoids above @xmath112 consist of the injections plus some finite range operations .
[ lem : unary : noninjectiveinfrangegeneratesall ] let @xmath206 have infinite image , and assume it is not injective
. then @xmath75 generates all unary operations .
we skip the fairly easy proof , and refer the reader to the very similar ( first part of the ) proof of lemma [ lem : unary : kerneltuplesorderisgeneratingorder ] .
we thus wish to know , given a finite range operation , which other finite range operations it generates . for that , we need the following concept
. let @xmath206 have finite range , and write @xmath207 .
enumerate the kernel classes of @xmath75 by @xmath208 in such a way that their sizes are increasing .
the _ kernel tuple _ @xmath209 of @xmath75 is the @xmath45-tuple @xmath210 .
note that the last entry of a kernel tuple @xmath211 always equals @xmath13 since @xmath75 must have at least one infinite kernel class .
having assigned a finite sequence with positive values in @xmath125 to every finite range operation , we are ready to order such sequences and give the definition of @xmath212 .
* let @xmath213 .
for @xmath214 and @xmath215 we write @xmath216 iff @xmath217 and the following holds : there exists a partition @xmath218 of @xmath219 into @xmath18 classes such that @xmath220 for all @xmath221 .
* we write @xmath222 for the partial order of finite increasing sequences of non - zero elements of @xmath125 ordered by @xmath223 .
* we write @xmath212 for the partial order of the finite increasing sequences of positive natural numbers ( not of values in @xmath125 ! ) ordered by @xmath223 .
observe that for finite increasing sequences @xmath224 of the same length we have @xmath216 iff @xmath225 for all @xmath148 .
the following lemma justifies our definition of @xmath223 .
[ lem : unary : kerneltuplesorderisgeneratingorder ] let @xmath226 have finite range
. then @xmath75 generates @xmath62 iff @xmath227 .
assume first that @xmath227 .
let @xmath211 have length @xmath45 and @xmath228 have length @xmath217 .
if @xmath229 , then @xmath230 for all @xmath148 . it is then not hard to see that for any finite set @xmath231 , there exist permutations @xmath232 such that @xmath233 agrees with @xmath62 on @xmath63 . if @xmath234 , then let @xmath235 be the partition provided by the definition of @xmath223 .
enumerate the kernel classes of @xmath75 by @xmath208 and in such a way that @xmath236 contains @xmath237 elements , for all @xmath238 .
now take any @xmath239 which are distinct but equivalent with respect to the partition @xmath235 . by composing @xmath75 with a permutation
, we may assume that @xmath75 maps the classes @xmath240 and @xmath236 into the class @xmath241 , and all other classes into classes @xmath242 in such a way that no two classes are mapped into the same class .
then @xmath243 is a function with @xmath244 values in its range , and @xmath245 .
proceeding like this , we arrive after @xmath246 steps at an operation @xmath247 which takes @xmath18 values and which satisfies @xmath248 for all @xmath221 .
thus we are back in the case @xmath229 , and the proof of this direction of the lemma is finished .
for the other direction , assume that @xmath75 generates @xmath62 .
let @xmath249 be as before .
since the local clone generated by @xmath75 is the topological closure of the set of term operations generated by @xmath75 , we have that for every @xmath250 , there exists a term @xmath251 consisting of permutations and @xmath75 which agrees with @xmath62 on the finite set @xmath252 .
we can write each @xmath251 as @xmath253 , where @xmath254 consists of permutations and @xmath75 , and @xmath255 is a permutation .
thus in every term @xmath251 , certain classes of @xmath75 are joined by the application of @xmath254 ( and shifted by @xmath255 , which we do not care about for the moment ) ; since there are only finitely many possibilities of joining classes of @xmath75 , there is one constellation which appears for infinitely many @xmath256 .
since @xmath257 implies that @xmath258 agrees with @xmath62 on @xmath252 , by replacing terms we may assume that the same classes are joined for all @xmath250 . naturally , this partition of classes induces a partition @xmath259 on @xmath219 via @xmath211 .
if @xmath260 are equivalent modulo the kernel of @xmath62 , then the kernel classes of @xmath75 containing @xmath261 and @xmath262 , respectively , are equivalent with respect to the partition @xmath259 . on the other hand , if @xmath260 are not in the same class of @xmath62 , then @xmath261 and @xmath262 will not lie in equivalent @xmath75-classes .
thus , taking @xmath256 large enough so that @xmath252 meets all kernel classes of @xmath62 , we can assign to every @xmath62-class ( with , say , index @xmath263 ) an equivalence class @xmath264 in an injective way ; in particular , @xmath265 . for infinitely many @xmath256 , this assignment is the same ; again , by replacing terms where necessary , we may assume it is always the same .
since for large enough @xmath256 arbitrarily large parts of the kernel classes of @xmath62 are hit , we must have @xmath266 for all @xmath263 , where @xmath267 is the class assigned to @xmath268 . by joining some classes @xmath269
, we can obtain @xmath270 without changing the latter fact .
[ lem : unary : unboundedfiniterange ] let @xmath204 .
if there is no finite bound to the sizes of the ranges of the finite range operations in @xmath70 , then @xmath70 generates @xmath138 .
let @xmath75 be any finite range operation , and let @xmath231 be finite .
then there exists a finite range function @xmath62 which agrees with @xmath75 on @xmath63 and whose kernel sequence @xmath228 has only one entry equal to @xmath13 .
now there exists a finite range operation @xmath271 in @xmath70 with @xmath272 , so @xmath62 is generated by @xmath70 .
this proves that @xmath75 is generated by @xmath70 , and hence @xmath70 generates all finite range operations .
clearly , any operation in @xmath138 can be interpolated on any finite set by a finite range operation , which implies our assertion .
the following is a consequence of lemma [ lem : unary : kerneltuplesorderisgeneratingorder ] ; it says that if finitely many finite range operations join forces , the joint generating power is not more than the sum of the generating powers of the single operations .
[ prop : unary : closure ] let @xmath204 be finite . then @xmath273 .
the non - trivial direction is to show @xmath274 . if @xmath275 is any term made of operations in @xmath70 , then it is of the form @xmath276 , where @xmath277 is a term and @xmath83 .
clearly , @xmath278 , implying @xmath279 .
thus , @xmath280 contains all terms that can be built from @xmath70 .
this implies that the union is a monoid .
being a finite union of ( topologically ) closed sets @xmath280 is itself closed , and hence contains even @xmath69 .
we now assign ideals of @xmath212 to local monoids containing @xmath112 . for a local monoid @xmath205 , we set @xmath281 for a sequence @xmath282 , write @xmath283 for the sequence obtained by gluing @xmath13 to the end of @xmath284 .
now set @xmath285 by lemma [ lem : unary : kerneltuplesorderisgeneratingorder ] , @xmath286 is always an ideal of @xmath212 .
conversely , we show in the following how to get closed monoids from ideals of @xmath212 .
[ defn : unary : limit ] let @xmath141 and let @xmath287 be an ascending sequence of @xmath18-tuples in @xmath222 .
we write @xmath288 for the smallest ( according to @xmath223 ) @xmath18-tuple @xmath284 in @xmath222 satisfying @xmath289 for all @xmath290 .
+ for an ideal @xmath291 , we let @xmath292 contain all operations in @xmath138 which are either injective , or which have finite range and whose kernel sequence is a limit of an ascending sequence of @xmath18-tuples of the form @xmath293 , where @xmath294 .
[ lem : unary : moniismonoid ] let @xmath291 be an ideal . then @xmath292 is a local monoid containing @xmath112 .
observe that if @xmath295 is a finite range function , and if @xmath296 is a finite range function such that @xmath227 , then @xmath292 contains also @xmath62 .
the range of @xmath75 have @xmath45 elements , let the range of @xmath62 have @xmath217 elements , and let @xmath235 be the partition of @xmath219 provided by the definition of @xmath223 .
a quick check shows that we may assume @xmath297 .
let @xmath298 be the sequence of @xmath299-tuples in @xmath300 such that @xmath301 .
set @xmath302 , for all @xmath303 and all @xmath304 . since @xmath305 , all @xmath306 are elements of @xmath300 , provided they are actually increasing tuples
. fixing @xmath149 , define inductively @xmath307 , for all @xmath304 .
it is not hard to see that the @xmath308 are in @xmath300 , too , and that @xmath228 is the limit of the increasing sequence @xmath309 , so @xmath310 . using lemma [ lem : unary : kerneltuplesorderisgeneratingorder ] and proposition [ prop : unary : closure ] , one now readily derives from this that @xmath292 is indeed a monoid .
it remains to show that @xmath292 is local .
let @xmath206 , and assume it can be interpolated on all finite sets by operations from @xmath292 ; assume also that it is not injective .
if @xmath75 has infinite range , then @xmath292 must contain non - injections of arbitrarily large finite range , which in turn implies that @xmath300 contains tuples of arbitrary length .
the definition of @xmath212 then shows that @xmath311 , a contradiction .
thus , all non - injections in the local closure of @xmath292 have finite range .
assume again that @xmath75 is such a non - injection , and assume @xmath211 has length @xmath18 .
fix for every set @xmath312 an operation @xmath313 which agrees with @xmath75 on this set . from some @xmath45 on
, the @xmath314 will have to take at least @xmath18 values , so we take the liberty of assuming that all @xmath314 have this property .
an easy manipulation of the @xmath314 using the fact that @xmath310 whenever @xmath315 allows us to assume that every @xmath314 takes exactly @xmath18 values and that @xmath316 for all @xmath290 . by thinning out the sequence
, we may also assume that the kernel sequences of the @xmath314 are increasing with respect to @xmath223 .
we then have @xmath317 .
now replace each @xmath18-tuple @xmath318 by a @xmath319-tuple @xmath320 with @xmath321 in such a way that @xmath322 and that the sequence @xmath287 is still ascending .
clearly , @xmath323 for all @xmath290 , and so @xmath324 implies @xmath295 .
hence , @xmath292 is locally closed .
[ prop : unary : isotoideallattice ] the mapping @xmath325 is an isomorphism from the lattice @xmath110 of locally closed monoids that contain @xmath112 onto the lattice of ideals of @xmath212 . that @xmath286 is an ideal of @xmath212 for all monoids @xmath116 follows directly from lemma [ lem : unary : kerneltuplesorderisgeneratingorder ] . from lemma [ lem :
unary : unboundedfiniterange ] we know that @xmath326 iff @xmath124 , and obviously @xmath327 iff @xmath116 contains only injections .
we have seen in lemma [ lem : unary : moniismonoid ] that for any proper ideal @xmath300 of @xmath212 , @xmath292 is a local monoid , and a straightforward verification shows @xmath328 , thus @xmath329 is onto .
also , an easy check using lemma [ lem : unary : kerneltuplesorderisgeneratingorder ] shows that @xmath330 for every local monoid which contains @xmath109 and all of whose non - injections have finite range , so @xmath329 is injective .
it is obvious that both @xmath329 and @xmath331 are order - preserving .
recall that a lattice is _ algebraic _ iff it is isomorphic to the subalgebra lattice of an algebra .
[ cor : unary : latticedistributive ] the lattice @xmath110 of local monoids above @xmath112 is distributive .
moreover , @xmath110 and its dual order are algebraic . by the preceding proposition , @xmath110 is the lattice of ideals of a partial order .
the assertions then follow from @xcite .
a partial order is called a _
well - quasi - order _ iff there are no infinite descending chains and no infinite antichains in it .
+ we call a sequence @xmath332 in a partial order with order relation @xmath333 _ bad _ iff for no @xmath334 we have @xmath335 . a standard application of the infinite ramsey s theorem shows that a partial order is well - quasi - ordered iff it contains no bad sequence ( confer e.g. @xcite ) .
[ lem : kerneltupleswellfoundednoinfantichain ] the set of finite sequences with values in @xmath125 ordered by @xmath223 is a well - quasi - order . in particular , its suborders @xmath222 and @xmath212 are well - quasi - orders .
assume that @xmath332 were a bad sequence of such finite sequences .
for every @xmath290 , let @xmath336 be the number of occurrences of @xmath13 in the tuple @xmath337 .
if the sequence @xmath338 were unbounded , then we could find @xmath290 such that @xmath336 is greater than the length of @xmath339 , implying @xmath340 , a contradiction .
thus , we can thin out the sequence in such a way that all @xmath336 are equal .
let @xmath341 be the tuple obtained from @xmath337 by leaving away the @xmath336 components equal to @xmath13 , for all @xmath290 . clearly , the @xmath341 form a bad sequence of tuples with values in @xmath13 . now
if the sequence of lengths of the @xmath341 were unbounded , then we could find some @xmath303 such that the length of @xmath342 is greater than the sum of all components of @xmath343 , hence @xmath344 , a contradiction .
thin out the sequence @xmath345 is such a way that all tuples have the same length @xmath18 .
now for all @xmath346 , we thin out our sequence so that the sequence consisting of the @xmath347-th component of the @xmath341 is increasing ; we can do this since @xmath13 is well - ordered .
the remaining sequence of @xmath341 is ascending , a contradiction . in general , the ideal lattice of a given well - quasi - order need not be a well - quasi - order .
certain well - quasi - orders which satisfy a certain strong combinatorial property and which are called _ better - quasi - orders _ , however , do have the property that their ideal lattice is well - quasi - ordered .
although giving the definition of a better - quasi - order ( see e.g. @xcite ) would be out of scope of the present paper , we remark that it follows from the basic theory of better - quasi - orders that our well - quasi - order @xmath212 is in fact a better - quasi - order ( as kruskal states in @xcite : `` all naturally known well - quasi - ordered sets which are known are better - quasi - ordered . '' ) .
therefore , the lattice of ideals of @xmath212 is a well - quasi - order ( and , in fact , even a better - quasi - order as well ) . in order to spare the reader the pain of reading the definition of a better - quasi - order
, we prove the following [ prop : unary : ideallatticewqo ] the lattice of ideals of @xmath212 is a well - quasi - order .
suppose that @xmath348 were a bad sequence of ideals . by taking away the first element of the sequence
, we may assume that @xmath349 for all @xmath290 .
consider for all @xmath290 the set @xmath350 consisting of @xmath351 plus all limits of ascending chains of tuples of the same length in @xmath351 ( as in definition [ defn : unary : limit ] ) : so @xmath352 for all @xmath290 . in @xmath350 ,
every chain is bounded from above : if @xmath353 is a chain in @xmath350 , then there is some @xmath354 such that all tuples in @xmath353 have length at most @xmath18 ; for otherwise , @xmath351 contains sequences of arbitrary length , implying @xmath355 contrary to our assumption .
but if the elements of @xmath353 all have length at most @xmath18 , then @xmath353 is bounded by construction of @xmath350 ( i.e. , adding the limits of ascending chains ) . applying zorn s lemma
, we get that every element of @xmath351 is below some maximal tuple of @xmath350 . by construction of @xmath350 , every tuple in @xmath212 which is below a maximal tuple of @xmath350
also is an element of @xmath351 .
the maximal elements of @xmath350 form an antichain with respect to @xmath223 . by lemma [ lem : kerneltupleswellfoundednoinfantichain ] ,
the set of sequences in @xmath125 , equipped with the order @xmath223 , is a well - quasi - order . in particular
, the antichain of maximal elements of @xmath350 is finite . for every @xmath356
, there exists a tuple in @xmath357 .
thus , there is a maximal tuple of @xmath358 which is not in @xmath350 .
we can thin out our sequence of @xmath351 so that this witnessing maximal tuple is the same for all @xmath356 ; denote it by @xmath359 .
now we do the same for @xmath360 and all @xmath361 , obtaining a maximal tuple @xmath362 which is not in any @xmath350 with @xmath361 .
we continue inductively in this fashion , obtaining a sequence @xmath287 . by construction , this sequence is a bad sequence in the order of finite @xmath363-valued sequences with @xmath223 , a contradiction .
[ cor : unary : cloneswqo ] the lattice of local monoids above @xmath112 is well - quasi - ordered .
[ prop : unary : finitelgenerated ] every local monoid @xmath116 above @xmath112 is finitely generated over @xmath112 , i.e. , there exists a finite @xmath364 such that @xmath365 .
moreover , the number of such monoids is countable .
if @xmath124 , then it is generated by any non - injective operation with infinite range , by lemma [ lem : unary : noninjectiveinfrangegeneratesall ] .
assume henceforth that @xmath116 contains only injections and finite range operations .
set @xmath366 to consist of all kernel sequences of operations in @xmath116 . by local closure and
what is by now a standard argument , one sees that every chain in @xmath366 has an upper bound in @xmath366 ( confer e.g. the proof of lemma [ lem : unary : moniismonoid ] ) .
hence , zorn s lemma implies that every element of @xmath366 is below a maximal element of @xmath366 .
the maximal elements of @xmath366 form an antichain with respect to @xmath223 , and therefore are finite in number by lemma [ lem : kerneltupleswellfoundednoinfantichain ] .
pick for each maximal tuple one corresponding operation in @xmath116 .
the set @xmath70 of operations thus chosen is as desired , by lemma [ lem : unary : kerneltuplesorderisgeneratingorder ] .
the above argument shows that every local monoid containing @xmath112 is determined by a finite set of finite sequences with values in @xmath125 .
there are only countably many possibilities for such finite sets , so the number of such monoids is countable .
having understood the structure of the lattice of local monoids containing @xmath112 , we move on to describe the monoidal interval corresponding to each such monoid . in this section
, we will prove theorem [ thm : monoidalintervalsoflargemonoids ] , which deals with those monoids which contain a non - constant and non - injective operation .
an operation @xmath75 on a set @xmath12 is called _ essential _ iff it is not essentially unary , i.e. , it depends on at least two of its variables .
we will see that except for @xmath48 , all clones whose unary fragment contains a non - constant and non - injective operation contain only essential operations with finite range , and none with infinite range ; hence the title of this section .
it turns out that the monoidal intervals of the monoids under consideration here are all chains which can be described nicely .
this is essentially a consequence of a theorem from @xcite for clones on finite sets , and the power of local closure . in order to state that theorem , we need the following definition .
let @xmath12 be any set .
we call an @xmath45-ary operation @xmath75 on @xmath12 _ quasilinear _ iff there exist functions @xmath367 and @xmath368 such that @xmath369 , where @xmath370 denotes the sum modulo @xmath371 . [ thm : s5:haddad - rosenberg ] let @xmath12 be a finite set of at least three elements and let @xmath75 be an essential operation on @xmath12 . set @xmath372 . then
, writing @xmath373 for the set of all permutations on @xmath12 , we have : * if @xmath374 , then @xmath75 together with @xmath373 generate all operations on @xmath12 which take at most @xmath18 values . *
if @xmath375 and @xmath75 is not quasilinear , then @xmath75 together with @xmath373 generate all operations on @xmath12 which take at most two values . * if @xmath375 , @xmath75 is quasilinear , and @xmath376 is odd , then @xmath75 together with @xmath373 generate all quasilinear operations on @xmath12 .
observe that there is no such thing as local interpolation on finite @xmath12 , so `` generates '' in the theorem refers to the term closure .
the following lemma is the infinite local version of this theorem .
[ lem : s5:haddadrosenberg : forinfinite ] let @xmath73 be essential , and assume that @xmath377 , where @xmath378 . 1 .
if @xmath374 , then @xmath75 generates all operations which take at most @xmath18 values .
if @xmath375 and @xmath75 is not quasilinear , then @xmath75 generates all functions which take at most @xmath18 values .
3 . if @xmath375 and @xmath75 is quasilinear , then @xmath75 generates all quasilinear operations .
( 1 ) : let any operation @xmath62 on @xmath379 be given .
it suffices to show that for every @xmath290 , @xmath75 generates an operation @xmath271 which agrees with @xmath62 on @xmath312 .
choosing @xmath45 large enough , we may assume that the ranges of both @xmath75 and @xmath62 are contained in @xmath312 . also , again by making @xmath45 larger , we may assume that the restriction @xmath380 of @xmath75 to @xmath312 is essential and takes @xmath18 values
. then @xmath380 is an essential operation on @xmath312 which takes @xmath18 values , and hence generates all such functions by theorem [ thm : s5:haddad - rosenberg ] .
in particular , @xmath380 generates a function @xmath381 which agrees with @xmath62 on @xmath312 .
the permutations which appear in the term which represents @xmath381 can be extended to @xmath9 by the identity ; occurrences of @xmath380 can be replaced by @xmath75 .
the resulting term @xmath271 is a function on @xmath9 which still agrees with @xmath62 on @xmath312 .
+ ( 2 ) : again , let @xmath62 and @xmath45 be given .
enlarge @xmath45 as before , if necessary , so that the restriction of @xmath75 to @xmath312 is not quasilinear .
now we argue as in the preceding proof .
+ ( 3 ) : the proof works as before ; @xmath45 only has to be chosen odd in order to allow application of theorem [ thm : s5:haddad - rosenberg ] . with the preceding lemma
, we see that it is quite easy to understand what happens when we add an essential finite range operation to a monoid .
we will now show that to the monoids relevant for this section , we in fact can not add an essential infinite range operation without generating @xmath48 . to establish this , we distinguish between those operations which preserve the binary inequality relation @xmath162 , and those which do not .
the latter case can be eliminated right away : [ prop : s5:essentialwithinfiniteimagepreservesneq ] let @xmath75 be an essential operation with infinite image .
then @xmath75 preserves @xmath162 , or it generates all operations
. the proposition will follow from the following lemma .
[ lem : s5:constantpaste ] let @xmath73 have infinite image , and assume it does not preserve @xmath162
. then @xmath75 generates a unary non - injective function that has infinite range . before proving the lemma
we show how the proposition follows from it : by lemma [ lem : s5:constantpaste ] , @xmath75 generates a unary non - injective operation with infinite range ; lemma [ lem : unary : noninjectiveinfrangegeneratesall ] then implies that @xmath75 generates all unary operations .
let @xmath382 be arbitrary .
we can find a unary finite range operation @xmath383 such that @xmath384 is essential and takes at least @xmath18 values .
by lemma [ lem : s5:haddadrosenberg : forinfinite ] , this implies that @xmath75 generates all operations which take not more than @xmath18 values . since @xmath18 was arbitrary and by local interpolation , this implies that @xmath75 generates @xmath48 .
we only have to prove something if @xmath75 is essential , for @xmath75 is itself a unary non - injective operation with infinite range otherwise .
we also assume @xmath75 to depend on all of its variables .
+ _ case 1 : _ there exist injective functions @xmath385 such that @xmath386 is injective .
choose @xmath387 infinite such that @xmath388 and @xmath389\cup\ldots \cup g_n[s])$ ] are infinite .
since @xmath75 does not preserve @xmath162 , there exist @xmath390 such that @xmath391 for all @xmath148 and such that @xmath392 .
we may assume that those values are not in @xmath393\cup\ldots\cup g_n[s]$ ] .
for all @xmath148 , set @xmath394 on @xmath395 and @xmath396 and @xmath397 . write @xmath398 .
on @xmath399 , @xmath400 is a function with infinite range which is not injective .
furthermore , the @xmath401 can be extended to permutations on @xmath9 since they are injective and have co - infinite domains and ranges .
this completes the first case .
+ _ case 2 : _
@xmath386 is not injective for all injective @xmath402 . since @xmath75 is not constant , there
exist injections @xmath403 such that @xmath404 is not constant . by assumption for this case
we have that @xmath404 is not injective .
thus , @xmath404 generates a function @xmath405 which is constant except for one argument , at which it takes another value . to see this ,
just observe that for such an @xmath277 , the kernel sequence @xmath406 satisfies @xmath407 for any non - constant @xmath408 , and apply lemma [ lem : unary : kerneltuplesorderisgeneratingorder ] .
consider arbitrary @xmath385 such that @xmath386 is injective .
such functions exist since @xmath75 has infinite range .
pick an infinite @xmath387 on which each @xmath409 is either injective or constant .
observe that it is impossible that all those restrictions of @xmath409 are constant .
say the restriction of @xmath409 to @xmath395 is constant with value @xmath410 for @xmath221 , where @xmath411 .
+ since @xmath75 depends on its first variable , there exist @xmath412 and @xmath413 such that @xmath414 . for otherwise the value of @xmath75
is determined by the values of the arguments @xmath415 , contradicting that @xmath75 depends on @xmath416 .
choose @xmath417 such that @xmath418 is distinct from both @xmath419 and @xmath420 ; this is possible as @xmath421 takes infinitely many values . set @xmath422 and @xmath423 if @xmath424 , for @xmath221 .
let moreover @xmath425 , and @xmath426 if @xmath427 , for all @xmath428 .
all @xmath429 and @xmath430 are generated by the two - valued function @xmath277 and hence by @xmath75 .
set @xmath431 .
we have that @xmath432 and @xmath433 and @xmath434 are pairwise distinct ; also , @xmath435 has finite range .
hence , @xmath75 generates a unary function @xmath435 that takes exactly three values .
+ pick @xmath151 such that @xmath436 for all @xmath148 and such that @xmath437 . for
all @xmath221 , let @xmath438 satisfy @xmath439 , @xmath440 , and @xmath441 for all @xmath442 .
clearly , @xmath443 is generated by the unary function @xmath435 with three values , and hence also by @xmath75 .
for all @xmath428 , let @xmath444 be any permutation that maps @xmath445 to @xmath203 and @xmath446 to @xmath447 .
now we set @xmath448 and have that @xmath449 and @xmath450 takes infinitely many values .
we are thus left with essential infinite range operations which do preserve the inequality relation .
the crucial theorem here has been shown in @xcite .
[ thm : s5:bininj ] every essential operation @xmath451 generates a binary injective operation .
able to produce a binary injection , we use this operation in order to enlarge ranges of unary operations : [ lem : s5:binaryinjectionplusfiniterange ] let @xmath140 be injective , and let @xmath452 be a non - constant function with finite range . then @xmath75 and @xmath62 together generate @xmath48 . write @xmath453 .
either @xmath62 is itself ( essentially ) unary , or it generates a non - constant unary operation which takes @xmath18 values by lemma [ lem : s5:haddadrosenberg : forinfinite ] . pick such a unary operation @xmath454 , and set @xmath455 .
the operation @xmath456 takes @xmath457 values .
hence , @xmath75 and @xmath62 together also generate a unary operation @xmath458 which takes @xmath459 values .
continuing in this fashion , and since @xmath460 , we obtain unary operations of all finite ranges , and hence also @xmath138 by lemma [ lem : unary : kerneltuplesorderisgeneratingorder ] . but clearly , for every operation @xmath461 , there exists a unary @xmath462 such that @xmath463 .
thus , since @xmath271 was arbitrary , @xmath75 and @xmath62 generate @xmath464 , and in turn also @xmath48 , as it is well - known and easy to see that @xmath465 .
we are thus ready to prove theorem [ thm : monoidalintervalsoflargemonoids ] .
we are given a monoid @xmath116 which contains an operation which is neither constant nor injective .
it follows from lemma [ lem : s5:binaryinjectionplusfiniterange ] that no clone properly contained in @xmath48 and containing @xmath116 can have a binary injection .
therefore , by theorem [ thm : s5:bininj ] , it can not have any essential operation which preserves the inequality relation . nor
can it contain an essential operation with infinite range which does not preserve the inequality , by proposition [ prop : s5:essentialwithinfiniteimagepreservesneq ] .
thus , it can not contain any essential operation with infinite range .
we now distinguish the two cases of the theorem : + ( 1 ) : if @xmath124 , then the only clones above @xmath138 can be the ones mentioned in the theorem , by lemma [ lem : s5:haddadrosenberg : forinfinite ] . that these sets of operations are actually clones is a straightforward verification and left to the reader .
+ ( 2 ) : if @xmath466 , then there is a largest natural number @xmath18 such that @xmath116 contains all unary operations which take at most @xmath18 values ; this follows from lemma [ lem : unary : unboundedfiniterange ] .
if @xmath132 , then no clone having @xmath116 as its unary fragment can have an essential operation , since this operation would generate all quasilinear operations by lemma [ lem : s5:haddadrosenberg : forinfinite ] , and hence all unary operations with at most two values .
consider thus the case @xmath467 .
again by lemma [ lem : s5:haddadrosenberg : forinfinite ] , the only clones in @xmath118 can be @xmath119 , @xmath134 and the @xmath468 , where @xmath469 .
it is easy to verify that these are indeed clones .
this section deals with the monoidal interval of @xmath159 , the monoid of all injections , and contains the proof of theorem [ thm : monoidalintervalofs ] .
before we start with an outline of this section , let us observe that as a straightforward verification shows , the largest element of this interval , @xmath160 , equals @xmath470 ( this is statement ( 1 ) of theorem [ thm : monoidalintervalofs ] ) . in particular , all operations of clones in this interval have infinite range ; hence the title of this section .
in subsection [ subsect : horn ] we prove that @xmath139 is the unique cover above @xmath471 in @xmath159 , and that @xmath164 is the set of all relations definable by a horn formula ( theorem [ thm : monoidalintervalofs ] , part ( 2 ) ) .
in subsection [ subsect : bar ] we prove that @xmath147 is the unique cover of @xmath139 in @xmath159 ( theorem [ thm : monoidalintervalofs ] , part ( 3 ) ) . in subsection [ subsect :
bar - odd ] we present an infinite strictly decreasing chain of clones that contain @xmath147 , and are contained in @xmath157 . to this end
, we also give relational descriptions of @xmath147 and of @xmath157 . in subsection
[ subsect : richard ] we show that @xmath24 and @xmath157 are incomparable , and that every clone in @xmath159 is either contained in @xmath157 or contains @xmath24 ( theorem [ thm : monoidalintervalofs ] , part ( 4 ) ) . in subsection [ subsect :
richardsmanyfriends ] we show that the power set of @xmath13 embeds into the intervals of those clones in @xmath159 that contain @xmath24 ; in particular , the size of this interval equals the continuum ( theorem [ thm : monoidalintervalofs ] , part ( 5 ) ) . in this subsection
we show that @xmath163 has the cover @xmath139 in the monoidal interval @xmath159 .
recall from definition [ def : hornclone ] that @xmath139 is the clone consisting of all operations @xmath75 that are essentially injective , i.e. , all operations @xmath75 that are the composition @xmath472 of an injective operation @xmath149 with projections @xmath473 ( it is straightforward to verify that this set of operations indeed forms a locally closed clone ) .
the clone @xmath139 is also called the _ horn clone _ ; the reason for this name will be given in proposition [ prop : hornchar ] .
let @xmath474 be a quantifier - free first - order formula where all atomic subformulas are of the form @xmath475 .
then @xmath474 is called _ horn _ if @xmath25 is in conjunctive normal form ( henceforth abbreviated cnf ) and each clause in @xmath474 contains at most one positive literal .
let @xmath476 be a formula in cnf .
we call @xmath25 _ reduced _ iff it is not logically equivalent to any of its subformulas , i.e. , there is no formula @xmath477 obtained from @xmath25 by deleting literals or clauses such that @xmath478 iff @xmath479 for all @xmath480 .
clearly , for every formula @xmath25 in cnf there exists a logically equivalent reduced formula . in the following proposition , the equvalence of ( 1 ) and ( 7 ) proves item ( 2 ) of theorem [ thm : monoidalintervalofs ] , stating that the horn clone is the unique cover of @xmath163 in @xmath159 .
moreover , items ( 5 ) and ( 6 ) provide finite relational generating systems of @xmath164 , and ( 2 ) provides a finite operational generating system ( in fact : a continuum of such systems ) of @xmath139 .
recall that a formula is _ horn _ iff it is in conjunctive normal form and each of its clauses contains at most one positive literal .
item ( 4 ) gives us a syntactical description of the formulas defining relations in @xmath164 .
[ prop : hornchar ] for all relations @xmath19 with a first - order definition in @xmath481 the following are equivalent . 1 .
@xmath19 is preserved by an essential operation that preserves @xmath162 .
@xmath19 is preserved by a binary injective operation .
3 . every reduced definition of @xmath19 is horn .
4 . @xmath19 has a horn definition .
@xmath19 has a pp definition in @xmath482 where @xmath483 6 .
@xmath19 has a pp - definition in @xmath484 where @xmath485 7 .
@xmath19 is preserved by @xmath139 . the implication from ( 1 ) to ( 2 )
is theorem [ thm : s5:bininj ]
. for the implication from ( 2 ) to ( 3 ) , suppose that @xmath474 is a reduced formula that defines @xmath19 but is not horn .
then there exists a clause @xmath477 of @xmath474 which contains two equalities @xmath486 and @xmath487 .
construct @xmath488 from @xmath477 by removing the equation @xmath486 , and @xmath489 by removing @xmath487 .
since @xmath474 is reduced , there exist @xmath490 such that @xmath478 but not @xmath491 , and @xmath492 but not @xmath493 .
clearly , @xmath494 , @xmath495 , @xmath496 , and @xmath497 .
set @xmath498 , where @xmath75 is the binary injection preserving @xmath19
. then @xmath499 , @xmath500 , and in fact @xmath501 does not hold . hence @xmath19 is not preserved by @xmath75 , a contradiction .
the implication from ( 3 ) to ( 4 ) is trivial . for the implication from ( 4 ) to ( 5 ) ,
let @xmath474 be a horn formula .
it suffices to show that all clauses @xmath477 of @xmath474 have a pp definition in @xmath502 .
if @xmath477 is of the form @xmath503 , consider the following pp formula .
@xmath504 assume that @xmath505 for all @xmath506 . in this case
, the pp formula implies that @xmath507 , and hence also implies that @xmath508 .
now , if @xmath509 for some @xmath506 , then for all choices of values for the other free variables the formula can be satisfied by setting @xmath510 to values that are distinct from all other values .
hence , the formula is a pp definition of @xmath477 in @xmath511 .
if @xmath477 does not contain a positive literal , consider the formula @xmath512 , which is equivalent to @xmath477 ( we assume that @xmath268 and @xmath513 are fresh variables ) .
we have seen above that the term in brackets is equivalent to a pp formula .
it is then straightforward to rewrite the whole expression as a pp formula . for the implication from ( 5 ) to ( 6 )
, it suffices to show that @xmath300 and @xmath162 have pp definitions in @xmath484 . for @xmath162 ,
this is obvious . to express @xmath514 ,
consider the pp formula @xmath515 if @xmath516 , the variables @xmath517 must denote the same value , and hence the formula implies @xmath518 . if @xmath519 , then for all choices of values for @xmath520 and @xmath521 it is possible to select values for @xmath517 that satisfy the formula .
hence , the above formula is a pp definition of @xmath300 in @xmath484 .
it is straightforward to verify the implication from ( 6 ) to ( 7 ) , because @xmath522 is preserved by projections and by injective operations .
the implication from ( 7 ) to ( 1 ) is immediate , since every at least binary injective operation is essential and has infinite image .
we show that the bar clone @xmath147 ( confer definition [ defn : barclone ] ) is the smallest clone below @xmath470 that strictly contains @xmath139 , thus proving item ( 3 ) of theorem [ thm : monoidalintervalofs ] .
a smallest non - empty @xmath18-ary relation that is preserved by all permutations of @xmath523 is called an _ orbit of @xmath18-tuples_. it is clear that every relation that is preserved by all permutations is the union of a finite number of orbits of @xmath18-tuples .
the following lemma will be useful here and in the following subsections , and a proof can be found in @xcite .
[ lem : op - arity ] let @xmath19 be a @xmath18-ary relation that consists of @xmath54 orbits of @xmath18-tuples . then every operation @xmath75 that violates @xmath19 generates an @xmath54-ary operation that violates @xmath19 .
the relation @xmath522 has been introduced in proposition [ prop : hornchar ] ; we recall it for the convenience of the reader : @xmath522 is the four - ary relation @xmath524 .
[ lem : binabovehorn ] let @xmath525 .
then @xmath75 generates a _
binary _ operation that violates @xmath522 .
the operation @xmath75 must violate the relation @xmath522 .
indeed , otherwise we would have @xmath526 , since all relations in @xmath164 have a primitive positive definition in @xmath484 by proposition [ prop : hornchar ] a contradiction .
since @xmath522 consists of two orbits of tuples , lemma [ lem : op - arity ] implies that @xmath75 generates a binary operation that violates @xmath522 .
clearly , the binary operation from lemma [ lem : binabovehorn ] is essential and not injective .
recall from definition [ defn : barclone ] that a binary function @xmath75 is called a @xmath18-bar function iff there exists an injection @xmath142 such that @xmath143 for all @xmath144 and @xmath145 for all @xmath146 ; recall also that @xmath147 is the clone generated by any bar function .
we show next that @xmath147 is well - defined .
clearly , for fixed @xmath141 , the @xmath18-bar functions generate each other .
it is also easy to see that if @xmath527 , then any @xmath18-bar function generates all @xmath45-bar functions .
[ lem : barfunctionsequvalent ] let @xmath141 , and let @xmath140 be a @xmath18-bar function and @xmath461 be a @xmath528-bar function
. then @xmath75 generates @xmath271 .
let @xmath75 be a @xmath18-bar function , and @xmath62 be a variant of the @xmath446-bar which satisfies @xmath529 for all @xmath530 , and which is injective otherwise .
clearly , @xmath62 is generated by @xmath75 .
set @xmath531 .
let @xmath532 .
then @xmath533 for all @xmath530 since @xmath534 for all @xmath535 .
also , @xmath536 for all @xmath537 . moreover , if @xmath538 and @xmath539 and @xmath540 , then it is easy to see that @xmath541 .
hence , @xmath271 is essentially a @xmath528-bar function , the only difference to a @xmath528-bar function being the value of @xmath542 , which can be undone by composing any permutation which swaps @xmath133 and @xmath543 with @xmath271 .
the next lemma characterizes the binary operations in @xmath147 .
recall the definition of injective in one direction " from definition [ def : injectiveinonedirectionandrichard ] .
[ lem : binarybarfunctionsdescription ] let @xmath140 .
then @xmath544 iff @xmath75 is injective in one direction and for all @xmath545 , each of the unary operations @xmath546 and @xmath547 is either constant or injective .
assume that @xmath75 is of that form , and let @xmath548 be given .
we can say without loss of generality that @xmath75 is injective in the first direction . write @xmath312 as a disjoint union @xmath549 in such a way that @xmath550 is injective for all @xmath551 , and constant for all @xmath552 . if @xmath553 , then there exist injections @xmath554 and a @xmath347-bar function @xmath62 such that @xmath555 agrees with @xmath75 on @xmath312 .
hence , @xmath544 .
+ we use induction over terms to show the other direction .
the statement is true for all bar - functions , including the projections .
assume it holds for @xmath556 , and consider @xmath557 .
then @xmath275 is injective in one direction ( this is easy to see , and also follows from lemma [ lem : risaclone ] in subsection [ subsect : richard ] ) .
let @xmath558 be arbitrary and consider @xmath559 .
by induction hypothesis , we have that @xmath560 and @xmath561 are constant or injective .
if both @xmath562 and @xmath563 are injective , then so is @xmath399 , since @xmath75 is injective in one direction .
if on the other hand both @xmath562 and @xmath563 are constant , then @xmath399 is constant as well .
assume therefore without loss of generality that @xmath562 is injective and @xmath563 is constant with value @xmath521 .
now if @xmath564 is constant , then so is @xmath399 , and if @xmath435 is injective , then the same holds for @xmath399 again , and we are done .
[ lem : producinginjectivefields ] let @xmath565 be essential and non - injective , and let @xmath566 . then @xmath75 generates a binary function @xmath62 which is not injective but injective on @xmath567 . without loss of generality
we assume @xmath568 for some @xmath569 and distinct @xmath570 .
pick any binary injection @xmath271 . by lemma
[ prop : hornchar ] , @xmath271 is generated by @xmath75 .
. then @xmath380 satisfies @xmath572 , and is injective in the @xmath371-nd direction .
therefore , replacing @xmath75 by @xmath380 , we may assume that @xmath75 is injective in the @xmath371-nd direction .
we use induction over @xmath18 to prove the assertion of the lemma .
+ for the induction beginning , let @xmath375 .
we may assume that there exist @xmath573 and distinct @xmath574 with @xmath575 .
indeed , otherwise @xmath75 is injective on the infinite square of pairs from @xmath576 , in which case we get the assertion by applying permutations .
since @xmath75 is essential , we may assume @xmath577 .
we have that @xmath578 and @xmath579 as @xmath75 is injective in the @xmath371-nd direction .
therefore , the only thing in our way to an injection on @xmath14 is that possibly @xmath580 . in that case
, we let a permutation @xmath581 map @xmath582 to @xmath445 , @xmath583 to @xmath284 , and set @xmath584 .
it is easy to check that @xmath380 is injective on @xmath371 and satisfies @xmath575 .
+ for the induction step , let @xmath75 be injective on @xmath18 , where @xmath566 .
we generate a function @xmath62 that is injective on @xmath133 and satisfies @xmath585 for some @xmath586 with @xmath587 .
again , we pick a binary injection @xmath271 and set @xmath571 .
clearly , @xmath380 is still injective on @xmath18 .
moreover , @xmath380 satisfies @xmath572 , and is injective in the @xmath371-nd direction .
therefore , replacing @xmath75 by @xmath380 , we may henceforth assume that @xmath75 is injective in the @xmath371-nd direction .
+ as in the induction beginning , we may assume that there exist @xmath573 and distinct @xmath574 with @xmath575 . since @xmath75 is injective in the @xmath371-nd direction we have @xmath588 $ ] .
+ if @xmath589 for some @xmath590 and @xmath591 , then we do the following : choose a permutation @xmath581 that exchanges @xmath18 with some @xmath592 for which @xmath593 , and which is the identity otherwise .
let @xmath594 be a permutation that maps @xmath595 into @xmath18 , and @xmath583 to @xmath284 .
now set @xmath596 .
we have : * all points @xmath597 that had different values under @xmath75 still have different values under @xmath380 ; in particular , @xmath380 is injective in the @xmath371-nd direction , and injective on @xmath18 .
* @xmath598 .
* @xmath599 .
if we repeat this procedure for all @xmath591 for which there exists @xmath590 with @xmath589 , then in the end @xmath600 for all @xmath601 and all @xmath591 .
now the only non - injectivity that could be left on @xmath133 could be that @xmath602 for some distinct @xmath603 .
one gets rid of this setting a last time @xmath604 , where @xmath594 is a permutation that maps @xmath605 into @xmath18 , and @xmath583 to @xmath284 .
[ lem : everybinarygeneratestheonebar ] let @xmath565 be essential and non - injective . then it generates a @xmath446-bar function .
fix a @xmath446-bar function @xmath606 .
we show that for arbitrary @xmath566 , @xmath75 generates a function which agrees with @xmath606 on @xmath18 .
define a finite sequence of natural numbers @xmath607 by setting @xmath608 , and @xmath609 . by lemma [ lem : producinginjectivefields ] and
the use of permutations , we can produce a function @xmath62 which is injective on @xmath610 and satisfies @xmath611 for some distinct @xmath612 .
moreover , we may assume that @xmath613\subseteq m_k$ ] .
we first want to produce a function @xmath614 which is injective on @xmath615 and satisfies @xmath616
. there is nothing to show if @xmath617 .
otherwise , we may assume @xmath618 , and set @xmath619 ; it is easy to check that @xmath614 has the desired properties .
replace @xmath62 by @xmath614 .
by repeating this procedure we can produce a function @xmath614 which is injective on @xmath620 and satisfies @xmath621 .
doing the same @xmath18 times , we get @xmath614 which is injective on @xmath622 and satisfies @xmath623 .
clearly , there exists a permutation @xmath581 such that @xmath624 agrees with @xmath606 on @xmath625 .
the following summarizes the results about @xmath147 obtained so far .
relational descriptions of @xmath147 will be given in proposition [ prop : relbarchar ] .
observe that the equivalence between ( 1 ) and ( 5 ) is exactly statement ( 3 ) of theorem [ thm : monoidalintervalofs ] .
[ prop : barchar ] let @xmath19 be a relation with a first - order definition in @xmath481
. then the following are equivalent . 1 .
@xmath19 is preserved by an operation from @xmath626 , i.e. , @xmath627 .
@xmath19 is preserved by a non - injective operation that depends on all its arguments and preserves @xmath162 .
@xmath19 is preserved by an operation from @xmath628 that violates @xmath522 .
@xmath19 is preserved by a @xmath446-bar function .
@xmath19 is preserved by @xmath147 , i.e. , @xmath629 .
given an operation from @xmath626 , we obtain an operation as described in ( 2 ) by leaving away all fictitious arguments , so ( 1 ) implies ( 2 ) .
an operation that is non - injective and depends on all its arguments is not generated by an injective operation , and therefore lemma [ lem : binabovehorn ] shows the implication from @xmath630 to @xmath631 . any binary operation that violates @xmath522 is necessarily essential and non - injective , and hence the conditions of lemma [ lem : everybinarygeneratestheonebar ] are satisfied , which shows the implication from @xmath631 to @xmath632 .
the equivalence of @xmath632 and @xmath633 is lemma [ lem : barfunctionsequvalent ] , and the implication from @xmath632 to @xmath630 is immediate .
the equivalence of @xmath633 and @xmath634 is by definition . in this subsection
, we explore those clones in the monoidal interval @xmath159 which contain the bar clone @xmath147 , but which are not contained in @xmath24 ( the clone of operations which are injective in one direction , cf . definition [ def : injectiveinonedirectionandrichard ] ) .
it will be necessary to first give a relational description of @xmath147 .
in the following it will be convenient to work with first - order formulas whose atomic formulas are of the form @xmath475 , _ true _ , or _
false_. the _ graph _ of a quantifier - free formula @xmath474 with such atomic formulas is the graph where the vertices are the variables of @xmath474 , and where two vertices @xmath635 are adjacent iff @xmath474 contains the sub - formula @xmath475 or the sub - formula @xmath636 .
we recall standard terminology .
if a formula @xmath474 is in conjunctive normal form , the conjuncts in @xmath474 are called _ clauses _ , and the disjuncts in a clause are called _
literals_. hence , literals are formulas that are either atomic , in which case they are also called _ positive _ , or the negation of an atomic formula , in which case they are also called _
negative_. let @xmath474 be a quantifier - free first - order formula where all atomic formulas are of the form @xmath475 or _
false_. then @xmath474 is called * _ horn _ iff @xmath474 is in conjunctive normal form and each clause in @xmath474 contains at most one positive literal ; * _ connected horn _
iff @xmath474 is horn and the graph of a clause @xmath477 in @xmath474 has at most two connected components when @xmath477 has no positive literals , and is connected when @xmath477 has positive literals . *
_ extended horn _
iff @xmath474 is a conjunction of formulas of the form @xmath637 for @xmath638 where @xmath639 are atomic formulas ( and hence of the form @xmath475 or _ false _ ) . * _ connected extended horn _
iff @xmath474 is extended horn , and if a conjunct @xmath477 of @xmath474 is connected whenever the right - hand side of the implication in @xmath477 contains a literal of the form @xmath475 and no literal _
false_. clearly , every connected horn formula can be written as a connected extended horn formula .
but there are connected extended horn formulas that are not equivalent to any connected horn formula ; we will even see that there are connected extended horn formulas that do not have a pp definition by connected horn formulas ( lemmas [ lem : gkpresrk-1 ] and [ lem : gkviolatesrk ] ) .
the conjuncts in an extended horn formula are also called _ extended horn clauses _ , and the literals on the left - hand side ( right - hand side ) of the implication of an extended horn clause are called the _ negative literals _ ( _ positive literals _ ,
respectively ) of the extended horn clause .
note that extended horn formulas can always be translated into ( standard ) horn formulas : if @xmath640 is an extended horn clause , then we can replace this clause by the conjunction of horn clauses @xmath641
. it will be convenient to say that a formula is _ preserved _ by an operation ( or a set of operations ) iff the relation that is defined by the formula is preserved by the operation ( or by the set of operations , respectively ) .
we want to give a syntactic description of the formulas that are preserved by @xmath147 .
we show that every such formula is equivalent to a connected extended horn formula .
in fact , we show the stronger result that every _ expanded horn formula _
preserved by @xmath147 is itself a connected extended horn formula .
[ def : expand ] a extended horn formula @xmath474 is called _ expanded horn _
iff it contains all connected extended horn clauses on the same set of variables as @xmath474 that are implied by @xmath474 , and for every disconnected clause @xmath642 in @xmath474 where @xmath643 does not contain _ false _ and for all variables @xmath635 in @xmath474 * adding an atomic formula of the form @xmath475 to @xmath644 , or * adding an atomic formula of the form @xmath475 to @xmath643 , or * setting @xmath643 to _ false _ results in a formula that is not equivalent to @xmath474 .
[ lem : expand ] every extended horn formula is equivalent to an expanded horn formula .
let @xmath474 be any given extended horn formula .
we construct the expanded horn formula that is equivalent to @xmath474 , as follows .
we first add to @xmath474 all connected extended horn clauses that are implied by @xmath474 .
then , if @xmath474 contains an extended horn clause @xmath645 that is not connected and where @xmath643 does not contain _ false _ , and @xmath474 is equivalent to a formula @xmath646 obtained from @xmath474 by adding an atomic formula of the form @xmath475 to @xmath644 or to @xmath643 for variables @xmath635 in @xmath474 , or by setting @xmath643 to _ false _ , then we replace @xmath474 by @xmath646 . if @xmath474 does not contain such a clause ( which will always happen after a finite number of steps , because there is only a finite number of formulas of the form @xmath475 for variables @xmath635 in @xmath474 ) , then @xmath474 is expanded horn , and clearly equivalent to the formula we started with .
[ lem : connexthorn ] if @xmath474 is an expanded horn formula , and @xmath474 defines a relation @xmath19 that is preserved by @xmath147 , then @xmath474 is connected extended horn .
suppose for contradiction that @xmath647 contains an extended horn clause @xmath477 of the form @xmath642 whose graph @xmath450 contains at least two connected components , and where @xmath643 does not contain _ false _ and contains at least one positive literal @xmath486 .
let @xmath648 be a variable from a connected component of @xmath450 that does not contain @xmath649 and @xmath650 .
let @xmath651 be the components of @xmath652 , where @xmath653 is the graph of @xmath643 .
observe that @xmath649 and @xmath650 are in distinct components : otherwise , if we replace the clause @xmath477 by @xmath654 , we obtain an equivalent formula . because @xmath477 is disconnected , this is in contradiction to the assumption that @xmath474 is expanded .
assume for the sake of notation that @xmath655 is the component of @xmath649 , and @xmath656 the one of @xmath650 .
we claim that there is a tuple @xmath284 for which @xmath657 is true , for which @xmath658 , and for which @xmath659 whenever @xmath660 are both in @xmath655 or are both in @xmath656 .
otherwise , @xmath474 implies @xmath661 , which is a connected expanded horn formula and therefore contained in @xmath474
. then @xmath474 must be equivalent to the formula @xmath646 obtained from @xmath474 by replacing the disconnected clause @xmath477 by @xmath662 , again in contradiction to the assumption that @xmath474 is expanded . to show the equivalence
, it clearly suffices to prove that @xmath646 implies @xmath474 .
so suppose that @xmath284 satisfies @xmath646 and @xmath644 .
the clause @xmath661 , which is contained in @xmath474 and in @xmath646 , shows that @xmath663 .
but then the premise of the new extended horn clause of @xmath646 is satisfied as well , and therefore @xmath284 satisfies @xmath643 , which is what we had to show .
next , we claim that there is a tuple @xmath275 with @xmath664 satisfying @xmath474 and where @xmath665 if @xmath660 are from the same component @xmath666 for some @xmath667 .
otherwise , @xmath474 implies @xmath668 .
but then the formula @xmath646 obtained from @xmath474 by replacing @xmath477 by @xmath669 is clearly equivalent to @xmath474 .
this again contradicts the assumption that @xmath474 is expanded .
let @xmath75 be a binary operation such that @xmath670 is constant for all entries @xmath513 of @xmath275 except for @xmath671 , and which is injective otherwise .
clearly , modulo permutations acting on its arguments , @xmath75 is an @xmath45-bar operation , and hence @xmath672 .
now , consider the tuple @xmath673 .
note that @xmath674 , because @xmath275 satisfies @xmath474 and @xmath275 satisfies the premise of the clause @xmath477 that is contained in @xmath474 ; since the conclusion of @xmath477 contains @xmath486 we have that @xmath674 .
it is straightforward to verify that @xmath277 does not satisfy @xmath474 , because @xmath277 satisfies @xmath644 , but @xmath675 , and therefore @xmath477 is violated .
the following relations play an important role in the relational description of @xmath147 and the clones above @xmath147 . for @xmath676 ,
let @xmath677 , @xmath678 , and @xmath679 be the relations defined by @xmath680 note that all three relations can be defined by connected extended horn formulas .
observe also that the expressive power of the relations in each of these sequences increases with increasing @xmath45 : for example , @xmath681 is a pp definition of @xmath677 from @xmath682 . [ prop : relbarchar ] let @xmath19 be a relation with a first - order definition in @xmath481 .
then the following are equivalent . 1 .
@xmath19 is preserved by @xmath147 , i.e. , @xmath629 .
every expanded horn formula that defines @xmath19 is connected extended horn .
@xmath19 can be defined by a connected extended horn formula .
4 . there exists an @xmath45 such that @xmath19 has a pp - definition in @xmath683 . the implication from @xmath630 to @xmath631
is shown in lemma [ lem : connexthorn ] . for the implication from @xmath631 to @xmath632 ,
recall that @xmath147 contains @xmath139 and hence the relation @xmath19 has a horn definition . by lemma [ lem :
expand ] , every horn formula has an equivalent expanded horn formula , and by assumption this formula is a connected extended horn definition of @xmath19 . to show that @xmath632 implies @xmath633 ,
let @xmath19 be a relation with a connected extended horn definition @xmath474 .
we first show that every extended horn clause @xmath642 from @xmath474 can be pp - defined in @xmath684 , for a sufficiently large @xmath45 , if @xmath643 is non - empty and does not contain _
let @xmath450 be the graph of @xmath644 where we add isolated vertices for each variable that appears in @xmath643 but not in @xmath644 ; in other words , let @xmath450 be the graph of @xmath642 where we remove an edge between @xmath685 and @xmath686 if @xmath475 is a literal of @xmath643 and not of @xmath644 .
let @xmath687 be the components of the graph @xmath450 .
we claim that @xmath642 can be pp - defined by @xmath678 where @xmath688 .
a pp definition is obtained from the formula @xmath689 by existentially quantifying @xmath690 , where @xmath691 . in this pp definition , the variables @xmath692 are the variables in the component @xmath693 from @xmath642 . to see that this is correct , consider a satisfying assignment to the variables of @xmath642 .
set for all @xmath149 the variables @xmath694 to the same value as @xmath695 .
suppose first that there exist two variables @xmath696 from the same component @xmath240 which are mapped to different values
. then @xmath487 can not be equal to both @xmath697 and @xmath698 , and hence either @xmath699 or @xmath700 .
hence , the conjunct @xmath701 is satisfied , and the assignment satisfies the given pp formula .
suppose now that otherwise all variables which have the same component are mapped to the same value .
since @xmath642 is connected horn , this implies by transitivity of equality that all variables that appear in @xmath642 must have the same value . but
this assignment clearly satisfies the given pp formula as well , so we are done with one direction of the correctness proof . for the other direction ,
consider a satisfying assignment for the pp formula .
if all variables are set to the same value , then @xmath642 will be satisfied by this assignment as well , because @xmath643 does not contain _
false_. otherwise , @xmath699 for some @xmath702 because of the conjunct @xmath701 .
let @xmath693 be the component of @xmath697 in @xmath450
. then @xmath703 .
but since @xmath693 is a connected component , this implies that one of the equalities in @xmath644 must be falsified , showing that @xmath642 is satisfied by the assignment . now , suppose that @xmath474 contains a clause @xmath642 where @xmath643 contains _ false _ ; i.e. , the clause is logically equivalent to a disjunction of literals of the form @xmath636 .
we proceed just as in the case above , but use the relation @xmath704 instead of @xmath705 with two new existentially quantified variables @xmath706 and @xmath707 at the two additional arguments , and we also add the conjunct @xmath708 to the formula .
we leave the verification that the resulting formula is equivalent to @xmath642 to the reader .
as we have seen , any clause from a connected horn formula can be pp - defined in @xmath709 , and thus @xmath632 implies @xmath633 .
finally , @xmath633 implies @xmath630 : it is easy to see that for all @xmath676 the relations @xmath710 and the relation @xmath162 are preserved by @xmath446-bar operations . in this section
we show that the interval between @xmath147 and @xmath470 contains an infinite strictly decreasing sequence of clones whose intersection is @xmath147 .
[ def : fk ] for all @xmath566 , fix an operation @xmath711 satisfying the following : @xmath712 moreover , for all other arguments , the function arbitrarily takes a value that is distinct from all other function values .
observe that @xmath158 appeared already in definition [ def : oddclone ] .
we use the @xmath713 to define another sequence of operations .
[ def : gk ] for all @xmath374 , fix an operation @xmath714 which satisfies @xmath715 and which takes distinct values for all other arguments .
observe that each of the operations @xmath383 ( and similarly @xmath713 ) generates @xmath147 : obviously , @xmath383 depends on all arguments , is non - injective , and preserves @xmath162 . by proposition [ prop : barchar ] , all relations that are preserved by @xmath383 are also preserved by @xmath147 , and hence @xmath383 generates @xmath147 . we now show that the operations @xmath383 , for increasing @xmath18 , generate smaller and smaller clones .
[ lem : gkpresrk-1 ] for @xmath716 , the operations @xmath383 and @xmath713 preserve @xmath717 .
we show the lemma for @xmath383 ; the proof for @xmath713 is similar and a bit easier , and left to the reader .
let @xmath718 be @xmath719-tuples that all satisfy @xmath720 .
let @xmath721 , and suppose that @xmath722 . by the definition of @xmath383
, this implies that @xmath723 , and that for all @xmath724 the tuples @xmath725 and @xmath726 have equal entries except for at most one position , call it @xmath727 .
since the function @xmath728 takes at most @xmath729 values , there must be a @xmath730 which is not in the range of @xmath728 .
for this @xmath347 we thus have @xmath731 . because @xmath732 , this implies that @xmath733 .
hence , the tuple @xmath734 is constant as well ; therefore , @xmath383 preserves @xmath720 .
[ lem : gkviolatesrk ] for @xmath716 , the operation @xmath713 preserves neither @xmath735 nor @xmath736 .
the operation @xmath383 does not preserve @xmath735 . if we apply @xmath713 to the @xmath737-tuples @xmath738 that all satisfy @xmath739 ( and all satisfy @xmath736 ) , then we obtain @xmath740 , which does not satisfy @xmath741 ( and does not satisfy @xmath736 ) . for @xmath383 , we define @xmath742 , and apply @xmath383 to @xmath743 .
we have thus seen that @xmath744 is a strictly decreasing chain of clones ( it is decreasing by the remark after the definition of these relations , and strictly so by the preceding lemma ) . by proposition [ prop : relbarchar ]
its intersection equals @xmath147 .
a clone of special interest ( see theorems [ thm : monoidalintervalofs ] and [ thm : monoidalintervalofsplusconst ] ) is the clone @xmath157 generated by @xmath158 , which will be called the _ odd clone _ ( cf . definition [ def : oddclone ] ) .
we will now describe the relations invariant under @xmath157 by providing a ( finite ) generating system of these relations , as well as a syntactic characterization of the formulas defining such relations .
let @xmath745 be the ternary relation @xmath746 [ prop : schar ] let @xmath19 be a relation with a first - order definition in @xmath481
. then the following are equivalent . 1 .
@xmath19 can be defined by a connected horn formula .
2 . @xmath19 has a pp definition by @xmath747 and @xmath162 .
3 . @xmath19 has a pp definition by @xmath745 and @xmath162 .
4 . @xmath748 , i.e. , @xmath19 is preserved by @xmath158 .
@xmath749 , but there is no definition of @xmath750 in @xmath751 . before proving the proposition , we need two proof - theoretical lemmas observing restrictions on what can be derived from connected extended horn formulas . [
lem : restrsem ] suppose @xmath474 is a connected extended horn formula that does not imply @xmath516 , but @xmath752 implies @xmath516 .
then @xmath752 also implies @xmath753 .
we rewrite @xmath474 as a conjunction @xmath754 of ( standard ) horn clauses as described above .
when for an extended clause @xmath755 for some @xmath756 we have @xmath757 , then we first remove all positive literals except @xmath16 from the extended clause , which leads to an equivalent formula .
it is well - known that the restricted form of resolution for predicate logic that requires that at least one of the parent clauses in a resolution step has to be positive , also known as _ p - resolution _ , is complete ( see e.g. @xcite , p. 102 ) .
when we apply resolution on a formula , we assume that there are always clauses @xmath758 , @xmath759 , and @xmath760 for universally quantified variables @xmath517 ( that axiomatize that the symbol @xmath761 for equality is an equivalence relation ) , and we also assume that all other variables in the formula are in fact constant symbols ( i.e. , they are existential variables before skolemization ) . since @xmath762-resolution is complete , we have in particular that @xmath763 implies @xmath516 if and only if @xmath516 can be derived by p - resolution from @xmath763 . we show that @xmath763 implies @xmath753 for all formulas @xmath516 that are not implied by @xmath754 but implied by @xmath763 .
suppose for contradiction that there is a literal @xmath516 that can be derived by p - resolution from @xmath763 but can not be derived from @xmath754 , and where @xmath763 does not imply @xmath753 .
select a literal @xmath516 with these properties whose @xmath762-derivation from @xmath763 is shortest possible .
the literal @xmath516 has either been derived from a clause @xmath764 in @xmath754 by resolving with the literals @xmath765 , or by resolving with of the three clauses having universally quantified variables ( after instantiating the universal variables ) . in the second case ,
suppose that the universal clause was @xmath766 , and @xmath516 has been derived by resolving with @xmath767 and @xmath768 , for some @xmath520 ( for the other two universal clauses , the argument is analogous ) .
then at least one of the literals @xmath767 and @xmath768 can not be derived from @xmath754 alone , because otherwise @xmath516 can be derived from @xmath754 alone , in contradiction to our assumptions .
suppose the literal that can not be derived from @xmath754 alone is @xmath767 ( the other case is analogous ) . by the choice of @xmath516 , the formula @xmath763 implies @xmath769 . at the same time
, we have that @xmath763 implies @xmath516 , and hence @xmath754 implies @xmath753 , a contradiction .
in the first case , at least one of the literals in the clause @xmath764 , say @xmath770 , can not be derived from @xmath754 alone , because otherwise @xmath516 can be derived from @xmath754 alone , in contradiction to our assumptions . by the choice of @xmath516 , @xmath763 implies @xmath771 .
let @xmath25 be the extended horn clause in @xmath474 that corresponds to the horn clause @xmath477 in @xmath754 ( i.e. , @xmath477 is among the clauses that have replaced the clause @xmath25 when creating @xmath754 from @xmath474 ) .
since @xmath25 is a connected extended horn clause , there is a sequence of variables @xmath772 of @xmath25 with @xmath773 and @xmath774 such that @xmath775 for @xmath776 is either a sub - formula of the right - hand side or of the left - hand side of the implication @xmath25 . by construction of @xmath754 , for each literal @xmath775 on the right - hand side there is the clause @xmath777 in @xmath754 .
but then it is clear that @xmath778 implies that @xmath779 , and hence also that @xmath780 .
therefore , @xmath763 implies @xmath753 , a contradiction . the following is a straightforward consequence .
[ lem : restrsem2 ] let @xmath474 be a connected extended horn formula and let @xmath477 be a conjunction of equalities @xmath781 .
if @xmath782 implies @xmath783 , but @xmath784 does not imply @xmath783 for any subconjunction @xmath488 of @xmath477 , then @xmath782 also implies @xmath785 .
we apply lemma [ lem : restrsem ] to @xmath786 , @xmath787 , and @xmath788 to derive that @xmath789 implies @xmath790 .
we can argue analogously for @xmath791 instead of @xmath787 for all @xmath792 , and the statement follows . to show that ( 1 ) implies ( 2 ) , suppose that @xmath474 is a connected horn definition of @xmath19 .
we show that each clause @xmath477 of @xmath474 has a pp definition in @xmath793 .
first , consider the case that @xmath477 contains a positive literal @xmath783 .
the graph of @xmath477 without the edge between @xmath794 and @xmath795 has at most two connected components . in the case
that the graph is still connected even after removing the edge between @xmath794 and @xmath795 , the clause @xmath477 is a tautology and can be ignored .
otherwise , let @xmath796 be the variables of the connected component of @xmath794 , and let @xmath797 be the variables of the connected component of @xmath686 .
consider the formula @xmath488 defined by @xmath798 we show that a tuple satisfies @xmath477 if and only if it satisfies @xmath488 .
if the tuple satisfies no negative literal of @xmath477 , then the only situation where @xmath488 is true is @xmath799 ; hence in this case , @xmath488 holds if and only if the tuple satisfies @xmath783 , which in turn is the case if and only if @xmath477 holds .
if on the other hand the tuple satisfies some negative literal of @xmath477 , then either there is @xmath800 such that @xmath801 or @xmath802 such that @xmath803 ; assume wlog the first is the case .
then we set @xmath804 to @xmath805 and @xmath806 to @xmath807 , @xmath808 to @xmath809 , and set @xmath810 to arbitrary pairwise disjoint values that are also distinct from all the values taken by any other variable of the clause .
it is easy to check that this assignment of variables is a satisfying assignment for the existentially quantified variables in the formula @xmath488 .
hence , @xmath488 is a pp definition of @xmath477 . if @xmath477 does not contain a positive literal , then the graph of @xmath477 consists of at most two connected components , because @xmath25 is connected horn .
assume first it does have two components .
let @xmath796 be the variables in the first and @xmath797 be the variables in the second connected component .
as above , it is easy to verify that the following formula is equivalent to @xmath477 .
@xmath811 if the graph of @xmath477 has only one component , then the following formula is equivalent to @xmath477 : @xmath812 for the implication from @xmath631 to @xmath632 , note that @xmath813 can be pp defined by @xmath814 we check that this formula is indeed equivalent to @xmath815 : let @xmath275 be a @xmath191-tupel that satisfies @xmath813 ; we assign values to the @xmath816 such that the formula above is satisfied . if @xmath817 then set @xmath818 to @xmath819 . if @xmath820 then set @xmath821 to @xmath822 .
note that if both @xmath817 and @xmath820 then our choice for @xmath823 is well - defined , since @xmath813 then implies that @xmath824 . if @xmath825 then set @xmath826 to distinct values that are also distinct from all entries of @xmath275 .
similarly , if @xmath827 , then set @xmath828 to distinct values that are also distinct from all entries of @xmath275 . if both @xmath825 and @xmath827 , then set @xmath823 to any value that is distinct from the values for @xmath829 .
it is straightforward to verify that these assignments satisfy all conjunct in the given formula .
conversely , if a @xmath191-tuple @xmath275 satisfies the formula , and @xmath817 and @xmath820 , then @xmath824 , and hence @xmath275 satisfies @xmath813 . for the implication from @xmath632 to @xmath633 it suffices to verify that @xmath158 preserves @xmath745 .
every 3-tuple @xmath275 from @xmath745 either satisfies @xmath830 or @xmath831 .
to violate @xmath745 , we would have to find three such tuples @xmath832 such that @xmath833 has two equal entries that are distinct from a third entry , say wlog . @xmath834 .
the only two ways to have @xmath835 are when @xmath835 is from @xmath836 , or when this is not the case and the tuples @xmath837 and @xmath838 are equal .
consider the first case , and assume wlog .
. then we know that @xmath840 and @xmath841 .
but then , since @xmath606 and @xmath520 are in @xmath745 , we have @xmath842 as well , and hence @xmath843 , a contradiction . the second case is similar and left to the reader .
hence , @xmath158 preserves @xmath745 . for
the implication from @xmath633 to @xmath634 , containment in @xmath844 follows from the fact that @xmath158 generates all operations in @xmath147 ( see also the remark after definition [ def : gk ] ) . by lemma [ lem : gkpresrk-1 ] , @xmath158 preserves @xmath747 and
@xmath158 also clearly preserves @xmath162 , so it suffices to show that @xmath158 violates @xmath845 . if we apply @xmath158 to the 6-tuples @xmath846 , @xmath847 , @xmath848 , which all satisfy @xmath845 , we obtain @xmath849 , which does not satisfy @xmath845 .
finally , we show the contraposition of the implication from @xmath634 to @xmath630 : if @xmath19 can not be defined by a connected horn formula , but is preserved by @xmath147 , then there is a pp - definition of @xmath850 in @xmath751 for some @xmath851 . because @xmath845 clearly has a pp - definition by @xmath850 for all
@xmath851 this will show the statement .
let @xmath852 be an expanded horn formula that defines @xmath19 . by proposition [ prop : relbarchar ] , we know that @xmath474 is a connected extended horn formula . we first claim that if for all extended horn clauses @xmath642 in @xmath474 the graph of @xmath644 has at most two connected components , then @xmath474 has a connected horn definition , contradicting our assumption on @xmath474 . to see this ,
we show that every such clause @xmath853 has a connected horn definition . if @xmath643 contains _ false _ or is empty then there is nothing to show ; assume thus that this is not the case .
if the graph of @xmath644 has exactly two components @xmath655 and @xmath656 , we know that @xmath643 contains a conjunct @xmath475 where @xmath854 and @xmath855 , because @xmath474 is connected extended horn . then @xmath642 is equivalent to the conjunction of the connected horn clauses @xmath856 and @xmath857 for every other conjunct @xmath858 of @xmath643 .
the case that the graph of @xmath644 is connected is easier and left to the reader .
therefore , @xmath474 must contain a clause @xmath642 such that @xmath644 consists of @xmath851 connected components @xmath859 .
we show that @xmath860 has the following pp - definition by @xmath19 , @xmath747 and @xmath162 .
@xmath861 to prove the equivalence , assume that @xmath862 .
we want to show that @xmath275 does not satisfy @xmath863 . if @xmath275 violates the second line of the formula @xmath863 , we are done .
otherwise , it must hold that @xmath864 .
suppose for contradiction that there are elements @xmath865 as specified in @xmath863 .
then for all @xmath792 the sub - formula @xmath866 implies that all variables @xmath809 from @xmath240 must have the same value as @xmath791 .
then the clause @xmath642 in @xmath474 implies that @xmath867 ( because @xmath868 is connected ) , which contradicts the second line of formula @xmath863 .
now assume that @xmath869 .
we want to show that @xmath275 satisfies @xmath863 . if @xmath275 satisfies @xmath870 for all @xmath871 , then any tuple from @xmath19 gives a assignment to the existentially quantified variables in @xmath863 that shows that @xmath275 satisfies @xmath863 . for
the sake of notation , let us thus assume that there is a @xmath18 with @xmath872 such that @xmath275 satisfies @xmath791 for @xmath873 , and @xmath870 for all @xmath874 .
we claim that there exists a tuple @xmath875 such that for all @xmath876 and all @xmath877 , @xmath878 it holds that @xmath879 if and only if @xmath880 . otherwise , the formula @xmath881 is unsatisfiable . since p - resolution
( also see the proof of lemma [ lem : restrsem ] ) is complete , there is a p - resolution proof of unsatisfiability .
this proof uses at most once a clause of the form @xmath882 from this formula , since it resolves only with positive literals and since therefore every resolution step using @xmath882 must be the last one , deriving _
false_. therefore , @xmath883 implies @xmath884 .
let @xmath25 be the smallest sub - formula of @xmath885 such that @xmath886 implies @xmath884 .
if @xmath25 is empty , then the clause @xmath884 is contained in @xmath474 because @xmath474 is expanded ; but then , adding @xmath887 to @xmath477 in the clause @xmath642 results in an equivalent formula , a contradiction to @xmath474 being expanded .
if @xmath25 is not empty , we can apply lemma [ lem : restrsem2 ] and obtain that @xmath886 implies that all variables @xmath888 that appear in @xmath25 have the same value as the value of @xmath884 . because @xmath474 is expanded , it has to contain the connected extended horn clause @xmath889 ; again , this shows that adding @xmath887 to @xmath644 in the clause @xmath642 results in an equivalent formula , a contradiction .
hence , we have shown the existence of a tuple @xmath875 with the properties as claimed above .
the tuple @xmath277 gives witnesses for the existentially quantified variables in @xmath863 that show that @xmath275 satisfies @xmath863 . in this section
we show that @xmath24 and @xmath157 are incomparable clones , and that every clone in @xmath159 is either contained in @xmath157 or contains @xmath24 .
we start with the fundamental observation that the set of operations @xmath24 ( see definition [ def : injectiveinonedirectionandrichard ] ) is indeed a clone .
[ lem : risaclone ] @xmath24 is a clone .
clearly , @xmath24 contains the projections .
let @xmath890 and @xmath891 .
there exist @xmath148 and @xmath892 such that @xmath75 is injective in the @xmath149-th direction and @xmath409 is injective in the @xmath347-th direction .
it is readily verified that @xmath386 is injective in the @xmath347-th direction .
let @xmath25 be a quantifier - free first - order formula where all atomic subformulas are of the form @xmath475 .
then @xmath25 is called _ negative _ if @xmath25 is in conjunctive normal form and each clause in @xmath25 either consists of a single positive literal , or does not contain positive literals .
the following theorem has been stated in equivalent form in ( * ? ? ?
* lemma 8.6 ) . for the convenience of the reader , and for completeness
, we also present a simplified proof here .
[ thm : negative ] let @xmath19 be a relation having a reduced horn definition that is not negative
. then @xmath745 has a pp definition from @xmath19 and @xmath162 .
let @xmath647 be a reduced horn definition of @xmath19 which is not negative ; assume moreover that among such definitions of @xmath19 , @xmath474 is one with the minimal possible number of occurrences of variables . since @xmath19 is not negative , there must be a clause @xmath477 in @xmath474 of the form @xmath893 for @xmath894 .
since @xmath474 is reduced , it contains a tuple @xmath895 with @xmath896 for all @xmath897 : otherwise , @xmath474 would imply @xmath898 , and we would obtain an equivalent formula by removing the literal @xmath899 from @xmath477 .
moreover , @xmath19 contains a tuple @xmath900 with @xmath901 , @xmath902 , and @xmath903 for all @xmath904 : otherwise , we could remove the literal @xmath905 from @xmath477 .
now consider the formula @xmath906 note that because @xmath19 is reduced , we have @xmath907 .
we first look at the case where @xmath908 . by appropriately renaming the variables of @xmath474
, we may assume that @xmath909 , and that @xmath910 and @xmath911 .
let @xmath912 be the ternary relation defined by existentially quantifying all variables in ( * ) except for @xmath649,@xmath913,@xmath914 . because @xmath915 satisfies @xmath916 for all @xmath917 , it satisfies ( * ) .
since @xmath918 , the relation @xmath919 contains a triple with three equal entries .
moreover , because @xmath920 satisfies ( * ) too , the relation @xmath919 contains a tuple @xmath921 with @xmath922 and @xmath923 .
assume first that @xmath919 contains a triple with three distinct entries .
we know that @xmath919 does not contain any tuple of the form @xmath924 for @xmath636 , because every @xmath45-tuple @xmath5 with @xmath925 and @xmath926 violates @xmath927 . hence , the relation defined by @xmath928 only contains the tuples with three distinct entries and the tuples with three equal entries , and we have indeed found a pp - definition of the relation @xmath745 in @xmath929 . striving for a contradiction , we suppose now that @xmath930 does not contain a triple with three distinct entries . then it is impossible that there is a tuple @xmath931 where @xmath932 : this is because having a horn definition , @xmath19 is preserved by a binary injection , by proposition [ prop : hornchar ] , and so is @xmath919 .
but the tuple obtained by applying a binary injective operation to @xmath277 and @xmath921 has three distinct entries , which contradicts our assumption for this case .
thus the clause @xmath933 is entailed by @xmath474 .
hence , we could replace the clause @xmath477 in @xmath474 by this shorter clause and obtain a formula that is equivalent to @xmath474 , because conversely , @xmath934 implies @xmath935 ( recall that @xmath936 ) .
this contradicts the choice of @xmath474 .
now take the case where @xmath937 ; say wlog @xmath938 .
let @xmath939 be the relation defined by @xmath863 by existentially quantifying all variables except for @xmath940 . as before we observe that @xmath941 and @xmath942 .
if @xmath943 , then the tuple obtained by applying a binary injective operation to @xmath944 and @xmath945 has four distinct entries ; this tuple is an element of @xmath19 , by proposition [ prop : hornchar ] .
it is easy to verify that then the formula @xmath946 is a pp definition of the relation @xmath947 given by @xmath948 the relation @xmath949 has the following pp - definition .
@xmath950 now suppose that @xmath951 and hence @xmath944 is the tuple with four equal entries . assume for a moment that there exist @xmath952 and @xmath953 such that @xmath954 implies @xmath955 .
then the relation obtained from @xmath954 by existentially quantifying @xmath956 is such that we can proceed as in the case where @xmath957 .
thus , we may henceforth assume that @xmath955 is not entailed for any @xmath958 and @xmath953 .
for every such pair @xmath959 , fix a tuple in @xmath919 witnessing that @xmath960 can be satisfied . now applying a @xmath192-ary injection to these four witnessing tuples and @xmath945
, we obtain a tuple all of whose entries are distinct .
this tuple is in @xmath919 as @xmath919 is preserved by @xmath139 , by proposition [ prop : hornchar ] .
consider the formula @xmath961 defined by the following formula .
@xmath962 where @xmath963 denotes the symmetric group on @xmath964 .
we claim that @xmath965 defines @xmath745 .
note that tuples of the form @xmath966 and @xmath967 , for distinct @xmath968 , are not contained in @xmath919 since every @xmath45-tuple @xmath521 with @xmath969 and @xmath970 violates @xmath927 .
therefore , the relation @xmath965 does not contain tuples with two values where one value appears three times , or tuples with three values where one value appears twice .
but certainly @xmath965 does contain all tuples with four equal values , and all tuples with four pairwise distinct values .
if some tuple where two values occur twice is excluded from @xmath919 , then @xmath965 does not contain any tuple with two values .
hence in this case , @xmath971 is a pp definition of @xmath949 , and we are done .
otherwise , @xmath965 contains all tuples where two values occur twice .
similarly as before it is easy to verify that then the expression @xmath972 above where @xmath919 has been replaced by @xmath965 is a pp definition of @xmath947 .
we have already seen that @xmath745 has a pp definition from @xmath399 , so it has a pp definition in @xmath973 as well .
the following proposition describes the clone @xmath24 : item ( 1 ) shows that @xmath24 is finitely generated ( namely , by any binary operation which violates @xmath745 but is injective in one direction ) ; and ( 2 ) and ( 3 ) provide a syntactical description of the formulas defining relations in @xmath974 . [
prop : disjofineq ] let @xmath19 be a relation with a first - order definition in @xmath481
. then the following are equivalent . 1 .
@xmath19 is preserved by a binary operation from @xmath470 that violates @xmath745 .
every reduced definition of @xmath19 is negative .
3 . @xmath19 has a negative definition .
@xmath19 is preserved by @xmath24 . to see that @xmath630 implies @xmath631 , fix any reduced definition of @xmath19 . since @xmath19 is preserved by a binary operation from @xmath470 , proposition [ prop : hornchar ]
implies that this definition is horn .
as there is a binary operation preserving @xmath19 and @xmath162 which violates @xmath745 , there is no pp definition of @xmath745 from @xmath19 and @xmath162 .
hence , our reduced definition must be negative , by theorem [ thm : negative ] .
the implication from @xmath631 to @xmath632 is trivial . to show that @xmath632 implies @xmath633 , fix a negative definition @xmath474 of @xmath19 .
let @xmath890 and @xmath975 be arbitrary . assume without loss of generality that @xmath75 is injective in the @xmath446-st direction .
then @xmath976 satisfies all inequalities that were satisfied by @xmath977 , which is enough to satisfy @xmath474 as @xmath474 is negative .
hence , @xmath75 preserves @xmath19 .
the implication from @xmath633 to @xmath630 is witnessed by any binary operation which is not in @xmath147 but injective in one direction , as is easily verified .
the following proposition establishes statement ( 4 ) of theorem [ thm : monoidalintervalofs ] .
[ prop : rsantichain ] the clones @xmath24 and @xmath157 form a maximal antichain in the interval @xmath978 $ ] .
in fact , every clone in this interval is either contained in @xmath157 or contains @xmath24 . in particular
, @xmath24 is a cover of @xmath979 .
first , we have to show that @xmath24 and @xmath157 are indeed incomparable . by proposition [ prop : disjofineq ]
it is clear that @xmath24 is not a subset of @xmath157 , because it contains an operation that violates @xmath745 . to show that @xmath157 is not a subset of @xmath24
, it suffices to show that the operation @xmath158 from definition [ def : oddclone ] is not injective in any direction .
indeed , this holds as @xmath980 and @xmath981 and @xmath982 for all @xmath983 .
therefore , @xmath984 and hence @xmath24 and @xmath157 form an antichain .
now , let @xmath47 be any clone from the interval @xmath978 $ ] which is not contained in @xmath157 .
then it contains an operation in @xmath470 that does not preserve @xmath745 .
since @xmath745 has two orbits , @xmath47 contains a binary operation with these properties , by lemma [ lem : op - arity ] .
thus , proposition [ prop : disjofineq ] implies that @xmath47 is above @xmath24 .
hence , the antichain that consists of @xmath24 and @xmath157 is maximal , and every clone of the interval not contained in @xmath157 does contain @xmath24 .
we now prove that there exists an antitone and injective mapping from the power set of @xmath13 into the interval @xmath985 $ ] ; in particular , this interval has cardinality @xmath986 .
this proves the last statement ( 5 ) of theorem [ thm : monoidalintervalofs ] . for all @xmath128 , write @xmath987 for all @xmath988 with @xmath989 , writing @xmath990 with @xmath991 ,
we set @xmath992 set @xmath993 observe that the @xmath994 are all negative formulas ; hence , by proposition [ prop : disjofineq ] , polymorphism clones of sets of relations defined by such @xmath994 will always contain @xmath24 . in the rest of this section , we will prove that no fixed @xmath994 can be defined from the others by primitive positive definitions .
therefore , distinct sets of such formulas define distinct clones above @xmath24 , which is what we want .
we now enumerate tuples @xmath995 as @xmath996 .
if @xmath997 is a formula , then we say that @xmath520 satisfies @xmath25 iff @xmath998 holds ; this is to say that we insert @xmath999 for @xmath1000 , where @xmath1001 and @xmath148 .
moreover , if @xmath1002 , we denote the `` @xmath1003-th column tuple '' @xmath1004 by @xmath1005 . for all @xmath128 ,
we set @xmath241 to be the @xmath1006-ary relation defined by @xmath994
. the operations of the following definition will separate the polymorphism clones of the @xmath241 . for all @xmath128 , the _ hubie - violator _
@xmath1007 is an operation defined as follows : enumerate @xmath1008 , that is , those elements of @xmath241 which have only entries in @xmath1009 , by @xmath1010 .
note that @xmath1011 since @xmath241 always contains the tuple where @xmath1012 and @xmath1013 , for all @xmath1014 .
now we define @xmath1015 by @xmath1016 , for all @xmath238 ; in other words , we set @xmath1017 . for every other input tuple in @xmath1018 , @xmath1007 takes a unique value distinct from all other values in its range .
+ let @xmath128 .
then the hubie - violator @xmath1007 violates @xmath241 . by definition , @xmath1017 , a tuple which does not satisfy @xmath1019 and which is therefore not an element of @xmath241 .
+ let @xmath1020 .
then the hubie - violator @xmath1007 preserves @xmath1021 .
write @xmath54 for the arity of @xmath1007 .
let @xmath1022 and set @xmath1023 .
we must show @xmath1024 .
we first check that @xmath284 satisfies the clauses @xmath1025 .
suppose it does not , and pick any @xmath1026 with @xmath1027 witnessing this .
we may assume wlog that @xmath1028 , where @xmath1029 , and write @xmath1030 .
set @xmath1031 , and @xmath1032 , and @xmath1033 .
if @xmath1034 were empty , then since for every value in @xmath1035 there is only a unique tuple that @xmath1007 sends to that value , we have that the column @xmath1036 equals the column @xmath1037 , for all @xmath1038 ( where we set @xmath1039 ) .
thus in that case , no row @xmath1040 would satisfy @xmath1025 , a contradiction . assuming that @xmath1035 were empty , we now show that there exists @xmath892 such that the row @xmath1040 does not satisfy @xmath1041 .
since @xmath1042 , we have @xmath1043 , for some @xmath1001 .
similarly , @xmath1044 , @xmath1045 , and @xmath1046 , where @xmath1047 .
consider first the case where @xmath1048 and @xmath1049 ( @xmath1050 may well equal @xmath1051 ) . pick @xmath1052 such that @xmath1053 , @xmath1054 , and @xmath1055 .
this is without doubt possible by defining all other entries to be very unequal . because in the definition of the hubie - violator all tuples of @xmath1008 appear , there is @xmath892 with @xmath1056 .
we have : @xmath1057 , @xmath1058 , @xmath1059 , and @xmath1060 .
hence , @xmath1061 does not satisfy @xmath1041 , a contradiction .
the case where @xmath1062 or @xmath1063 is even easier : say wlog @xmath1062 , and pick @xmath1064 such that @xmath1065 , and @xmath1066 .
we have : @xmath1067 , @xmath1058 , @xmath1059 , and @xmath1068 .
hence , in this case @xmath1069 does not satisfy @xmath1041 either , a contradiction . finally , consider the case where neither @xmath1034 nor @xmath1035 are empty .
assume wlog that @xmath1070 and @xmath1071 .
we have @xmath1072 , for some @xmath1073 .
let @xmath1074 be minimal with the property that @xmath1075 ; write @xmath1076 .
then @xmath1077 for some @xmath1001 .
pick @xmath1064 such that @xmath1078 .
we then have that @xmath1040 does not satisfy @xmath1079 : indeed @xmath1080 , and @xmath1081 for all @xmath1082 as columns since @xmath1083 . a contradiction .
it remains to show that @xmath284 satisfies @xmath1084 .
suppose it does not ; we claim that under this assumption , there is @xmath892 such that @xmath1040 does not satisfy @xmath1084 either . set @xmath1085 to consist of all values that appear in @xmath284 , @xmath1032 , and @xmath1086 .
write @xmath728 for the cardinality of @xmath1034 ; wlog we may assume @xmath1087 . observe that since @xmath284 has length @xmath737 and since it does not satisfy @xmath1084 , we have @xmath1088 .
consider the @xmath1006-tuple @xmath1089 .
using @xmath1090 , one readily checks that @xmath1091 , hence @xmath1092 for some @xmath892 .
thus , since @xmath1093 appears in the @xmath347-th row of the definition of the hubie - violator @xmath1007 , we have that for all @xmath1094 and all @xmath1038 , if @xmath1095 , then @xmath1096 .
in particular , this holds for the vectors @xmath1097 and @xmath1098 for which @xmath1099 , and we have @xmath1100 for those columns . but for the other columns @xmath1101 whose ( equal ) image under @xmath1007 is in @xmath1035 , we have @xmath1100 anyway since values outside @xmath219 are taken by at most one tuple . summarizing these two observations , we have that @xmath1102 for all @xmath1103 , so @xmath1040 does not satisfy @xmath1084 . a contradiction .
we thus know that for all @xmath128 , @xmath241 has no pp - definition from @xmath1104 . for
otherwise , the polymorphisms of @xmath1105 would be contained in those of @xmath241 , which is not the case by the preceding two lemmas .
it remains to show that @xmath241 can not be defined from the @xmath1021 with @xmath1106 either .
let @xmath1020 .
then the hubie - violator @xmath1107 preserves @xmath241 .
denote the arity of @xmath1107 by @xmath54 .
as before , let @xmath1108 and set @xmath1109 . to see that @xmath284 satisfies the clauses @xmath1025 , one proceeds as in the preceding proof , with @xmath18 and @xmath45 exchanged .
what is more difficult here is to check that @xmath284 satisfies @xmath1019 .
suppose it does not , and define @xmath1110 , as well as @xmath1111 , as before ( only with @xmath18 and @xmath45 switched ) .
now we distinguish two cases , @xmath1112 and @xmath1113 . in the first case ,
we argue like at the end of the proof of the preceding lemma : consider the tuple @xmath1114 .
one readily checks that @xmath1115 , so @xmath1092 for some @xmath892 .
thus , since @xmath1093 appears in the @xmath347-th line of the definition of @xmath1107 , we have that for all @xmath1094 and all @xmath1038 , if @xmath1116 , then @xmath1096 .
in particular , this holds for those vectors @xmath1097 and @xmath1098 which are sent by @xmath1107 to a value in @xmath1034 . since @xmath1117 for the other columns ( i.e. , those with value in @xmath1035 )
, this implies that @xmath1102 for all @xmath1118 , so @xmath1040 does not satisfy @xmath1019 , a contradiction .
to finish the proof , consider the case where @xmath1113 .
observe first that for every @xmath221 , there exists exactly one @xmath1118 such that @xmath1119 .
there is at least one such @xmath277 as @xmath1113 ; suppose there were distinct @xmath1120 with this property
. then @xmath1121 are all elements of @xmath1122 .
pick any tuple @xmath1123 with @xmath1124 .
we have that @xmath1125 are all equal , hence @xmath1040 does not satisfy @xmath1126 , a contradiction .
we may therefore wlog assume that @xmath284 looks like this : @xmath1127 , where all entries starting from the question mark are elements of @xmath1035 .
if we had @xmath1128 for some @xmath1129 , then we could derive a contradiction just like in the case @xmath1112 , by taking @xmath149 out of @xmath1034 .
so assume this is not the case .
then the set of vectors @xmath1130 equals @xmath1131 .
we prove that there exists @xmath892 such that @xmath1040 does not satisfy @xmath1132 .
let @xmath329 be the permutation on the indices of the tuple @xmath1133 which sends this tuple to @xmath1134 .
observe that @xmath329 only switches @xmath1135 and @xmath1136 , if necessary . now consider the tuple @xmath1137 , and apply @xmath1138 to the indices of this tuple , obtaining a new tuple @xmath1139 .
one readily checks @xmath1140 , so @xmath1141 for some @xmath1142 .
thus the tuple @xmath1143 , which is obtained by applying @xmath329 to @xmath1144 , equals @xmath606 . but @xmath606 does not satisfy @xmath1132 , so @xmath1145 , a contradiction .
we end this section by the following proposition , which completes the proof of theorem [ thm : monoidalintervalofs ] and of the corollary after .
the mapping @xmath329 from the power set of @xmath13 into @xmath1146 which sends every @xmath1147 to the local clone @xmath1148 is injective and antitone ( with respect to inclusion , on both sides ) . in particular
, the number of local clones containing @xmath112 equals the continuum .
since @xmath84 is antitone , it follows that the same holds for @xmath329 .
now let @xmath1149 be unequal , say wlog that there is @xmath1150 .
then the hubie - violator @xmath1151 violates @xmath1152 but preserves all @xmath1153 , where @xmath1154 .
thus @xmath1155 , and @xmath329 is indeed injective .
it remains to investigate one last monoidal interval , namely the one corresponding to the monoid @xmath167 consisting of all unary operations which are either injective or constant .
we will thus be concerned with the proof of theorem [ thm : monoidalintervalofsplusconst ] in this section .
the crucial ingredients to the proof will be proposition [ prop : rsantichain ] and the following lemma ; the definition of @xmath157 was given in definition [ def : oddclone ] .
[ lem : s+locallyclosed ] @xmath1156 .
we have to show that @xmath1157 is a local clone .
let @xmath1158 be an arbitrary non - constant operation .
we claim that @xmath1159 . to see this ,
let @xmath63 be any finite subset of @xmath1160 , where @xmath45 is the arity of @xmath75 .
extend @xmath63 to a finite set @xmath1161 such that @xmath75 is non - constant on @xmath1162 . since @xmath1158 , and since @xmath157 is by definition generated by @xmath158 , there exists a term @xmath275 using @xmath158 and the constants which agrees with @xmath75 on @xmath1162 .
since @xmath275 is non - constant , it can as well be written without constants , using @xmath158 only : this is because inserting a constant as an argument of @xmath158 gives us an injection or essentially a bar operation , both of which are generated by @xmath158 without using constants .
thus , @xmath75 can be interpolated on @xmath1162 , and hence also on @xmath63 , by a term in @xmath157 .
this proves @xmath1159 .
recall that @xmath24 consists of all operations which are injective in one direction ( definition [ def : injectiveinonedirectionandrichard ] ) . whereas @xmath1157 is a local clone
, we can not add constants to @xmath24 without generating all operations : @xmath183 .
let @xmath1163 be arbitrary , and let @xmath61 be finite .
define @xmath1164 as follows : if @xmath1165 and @xmath1166 , then @xmath62 returns @xmath1167 . for every other input tuple
, @xmath62 returns a unique value distinct from all other values in its range .
clearly , @xmath62 is injective in the @xmath1168-st direction , so @xmath1169 .
moreover , @xmath1170 agrees with @xmath75 on @xmath63 , implying @xmath1171 as @xmath63 was arbitrary .
note that the preceding proposition proves item ( 7 ) of theorem [ thm : monoidalintervalofsplusconst ] .
we have found the largest element of our monoidal interval @xmath170 : [ lem : s+isthelargest ] @xmath1172 .
clearly , the unary fragment of @xmath1157 equals @xmath167 , hence @xmath1157 is an element of the corresponding monoidal interval @xmath170 .
suppose there was a local clone @xmath47 in this interval which properly contains @xmath1157 .
take any @xmath1173 . clearly , @xmath75 must be essential .
since the only non - injections in the unary fragment of @xmath47 are constant , lemma [ lem : s5:haddadrosenberg : forinfinite ] implies that @xmath75 must have infinite range . by proposition [ prop : s5:essentialwithinfiniteimagepreservesneq ]
, @xmath75 preserves @xmath162 .
thus , @xmath1174 , so @xmath75 generates @xmath24 by proposition [ prop : rsantichain ] .
hence @xmath1175 by the preceding proposition .
the following proposition proves statement ( 1 ) of theorem [ thm : monoidalintervalofsplusconst ] .
. then @xmath1176 . in particular
, @xmath173 is a local clone in @xmath159 . by lemma [ lem : s+isthelargest ]
, @xmath47 is a subclone of @xmath1157 ; hence intersecting @xmath47 with @xmath157 just means taking away the constants , which proves our assertion . since the intersection of local clones is a local clone ,
so is @xmath173 .
clearly , the unary fragment of @xmath173 equals @xmath137 , hence @xmath173 is an element of @xmath159 .
observe that the second statement of the theorem is trivial .
we turn to the proof of item ( 3 ) .
the mapping from @xmath170 into the subinterval @xmath176 $ ] of @xmath159 which sends every clone @xmath47 to @xmath173 is a complete lattice embedding which preserves the smallest and the largest element .
it is clear that this mapping preserves the order and by the second statement of theorem [ thm : monoidalintervalofsplusconst ] , it sends the smallest ( largest ) element of @xmath170 to the smallest ( largest ) element of @xmath176 $ ] .
it is also obvious that the mapping is injective .
let @xmath177 .
clearly , @xmath1177 , hence the mapping preserves finite meets ; larger meets work the same way .
since @xmath1178 contains both @xmath173 and @xmath1179 , it contains also their join , so @xmath1180 . for the other inclusion ,
let @xmath75 be any non - constant operation in the ( not necessarily local ) clone generated by @xmath1181 .
then @xmath75 can be written as a term over @xmath1181 .
we can assume that in this term , thinking of it as a tree , the constants appear only as leaves .
we may also assume that except for leaves , no subtree of this tree is constant .
that just means that the nodes above the leaves are operation symbols in @xmath1182 , some of whose variables are set to constant values .
but these operations in @xmath1182 with some variables set to constants are again elements of @xmath1182 , as they are non - constant by assumption and since @xmath47 and @xmath1183 are clones containing all constants .
summarizing , we can write @xmath75 as a term over @xmath1182 without the use of constants . hence , @xmath1184 .
this proves @xmath1185 , and hence our mapping preserves finite joins .
infinite joins work the same way .
observe that item ( 4 ) of theorem [ thm : monoidalintervalofsplusconst ] is clear from the definitions .
we prove ( 5 ) . for all @xmath180 which do not contain @xmath24
, @xmath181 is a local clone in @xmath170 . by proposition [ prop : rsantichain ] ,
@xmath47 is contained in @xmath157 .
hence , @xmath181 is contained in @xmath1157 by lemma [ lem : s+locallyclosed ] , which equals @xmath1186 by lemma [ lem : s+isthelargest ] .
therefore @xmath181 is indeed a local clone in @xmath170
. item ( 6 ) is trivial from the definitions .
we finish the proof of the theorem by the following proposition , which restates items ( 8) and ( 9 ) : @xmath184 is the unique cover of @xmath185 and @xmath186 is the unique cover of @xmath184 in @xmath170 .
let @xmath47 be in @xmath170 , and assume it has an essential operation @xmath75
. then @xmath1159 by lemma [ lem : s+isthelargest ] , thus @xmath75 generates @xmath139 by proposition [ prop : hornchar ] .
it is easy to see that @xmath184 is a local clone , which proves the first assertion .
similarly , if @xmath47 is in @xmath170 and properly contains @xmath184 , then any operation witnessing this generates @xmath147 , by proposition [ prop : barchar ] .
again , it is straightforward to check that @xmath186 is a local clone .
we want to close this section with the remark that the mapping which sends every clone @xmath47 in the interval @xmath176 $ ] of @xmath159 to @xmath181 is not injective . in particular , it collapses all clones between @xmath1187 and @xmath157 .
[ prop:+notinjective ] @xmath1188 the operation @xmath1189 ( definition [ def : gk ] ) is injective in the first argument , and hence contained in @xmath24 .
by lemma 59 , @xmath1189 preserves @xmath1190 , and , being an element of @xmath24 , it also preserves @xmath162 .
proposition [ prop : schar ] then implies that @xmath1189 is contained in @xmath157 : otherwise , it would violate a relation preserved by @xmath157 ; that relation would have a pp definition from @xmath1190 and @xmath162 , a contradiction . therefore , the operation @xmath1189 is in @xmath1187 .
the operation given by @xmath1191 equals by definition the operation @xmath158 ( definition [ def : oddclone ] or [ def : fk ] ) , which by definition generates @xmath157 .
hence , @xmath1192 contains @xmath1157 .
the converse containment is trivial .
observe that this proposition implies that the complete lattice embedding that sends every @xmath172 to @xmath173 is not surjective onto the interval @xmath176 $ ] : if there existed a clone @xmath172 such that @xmath1193 , then @xmath1194 since @xmath182 .
thus @xmath1195 would be a clone and hence equal to @xmath1157 , implying @xmath1196 , a contradiction .
in fact the same argument shows that a clone @xmath1197 $ ] is in the range of this embedding iff @xmath1198 is a clone .
we have to leave the complete description of the monoidal intervals corresponding to @xmath137 and @xmath167 open ; we do not know if a reasonable description of these intervals is possible at all .
two particular problems seem most important to us at this point : * can one effectively decide whether for a given sequence @xmath1199 of relations that are first - order definable in @xmath1200 there is a pp - definition of @xmath1201 in the structure @xmath1202 ? * what is the cardinality of the monoidal interval corresponding to @xmath167 ?
m. bodirsky and j. kra .
the complexity of equality constraint languages .
, 3(2):136158 , 2008 . a conference version of the paper appeared in the proceedings of the international computer science symposium in russia ( csr06 ) . | we initiate the study of reducts of relational structures up to primitive positive interdefinability : after providing the tools for such a study , we apply these tools in order to obtain a classification of the reducts of the logic of equality .
it turns out that there exists a continuum of such reducts .
equivalently , expressed in the language of universal algebra , we classify those locally closed clones over a countable domain which contain all permutations of the domain . |
kontsevich solution for deformation quantization of poisson manifolds @xcite is based on a specific qft @xcite , yet universal : use one quantization to find them all .
the coefficients of the star - product , expressed as integrals on some configuration spaces , are the corresponding amplitudes .
in @xcite it was shown that the coefficients represent a cocycle of the cobar construction of the dg - coalgebra of kontsevich graphs , and it was conjectured that there should exist a direct combinatorial approach , as an alternative to the above analytical approach . in this article the deformation problem for associative algebras is pulled back at the level of graphs . as a generalization , generic associative deformations of associative algebras in a generic direction , not necessarily poisson , are also considered .
the corresponding cocycles are `` unitary characters '' , i.e. satisfying @xmath0 ( inverse relative to convolution ) .
the `` initial value ( deformation ) problem '' at tree - level ( no loops ) has a unique solution . in the case of linear coefficients , e.g. for @xmath1 , with @xmath2 a lie algebra , the solution is expected to correspond to the hausdorff series , which is the `` forefather '' of all quantizations .
this is possible since the kontsevich rule ( a kind of a dual feynman rule ) , is a cohomological pairing , allowing to pullback gerstenhaber composition from the hochschild complex .
besides its value as a self contained approach ( at the mathematical level ) to the quantization problem , the present approach is relevant to the understanding of renormalization techniques .
not surprisingly , since the coefficients may be interpreted as feynman integrals , the `` correction analysis '' of the star - product terms in order to yield associativity is similar to bphz renormalization via `` counter - terms method '' @xcite .
connections between deformation theory and renormalization are expected to emerge . to make the article accessible to a wider audience ,
the next section is a brief recall on deformation quantization and the associated combinatorial problem ( students could perhaps start with sections [ s : dq ] , [ s : ca ] and [ s : scr ] ) .
section [ s : gc ] continues with a short recall on graph cohomology from @xcite , the reduction from formality to star - products leading to section [ s : sp ] where the relevant graph operations are discussed .
examples are computed in section [ s : coa ] ( see also appendix ) , which rephrases the `` correction analysis '' of @xcite in the context of the hopf algebra of graphs . the reader in deformation theory and renormalization is invited to skim through section [ s : sp ] before considering where the correspondence between associativity and the unitarity condition for the coefficients is explained , before considering the main point stated in [ s : dgla ] . the relation with renormalization is briefly considered as part of the concluding section .
comments are most welcome !
in classical physics observables like position @xmath3 and momentum @xmath4 commute . in the quantum world
they do nt .
they are usually modeled by linear operators on hilbert spaces , satisfying the celebrated heisenberg uncertainty relation : @xmath5 .
the `` quantization problem '' consists in finding a correspondence between classical and quantum variables , in order to pass ( and benefit ) from a classical description to the quantum description .
a broad and `` simple '' approach , the so called deformation quantization approach , is based on the idea of keeping the observables `` as is '' while deforming the commutative product into a non - commutative one : a `` star - product '' satisfying : @xmath5 .
recall that the state space of a classical system is modeled by a poisson manifold @xmath6 , where @xmath7 denotes a two tensor : the poisson structure .
assume for simplicity that @xmath8 .
in the case of a constant coefficients poisson bracket , there is a nice exponential ( moyal s ) formula for such a star - product : @xmath9 where @xmath10 denote functions on @xmath11 . in 1997 , in a ground breaking paper @xcite , kontsevich devised a formula for the star - product in the general case ( variable coefficients ) : @xmath12 the above star - product formula is to be thought of as a taylor series , and where the terms , due to the non - commutativity of the bidifferential operators @xmath13 involved , correspond to a certain class of graphs @xmath14 ( not just uni - valent trees ) .
the correspondence between graphs and bidifferential operators can be thought as an instance of a `` feynman rule '' from quantum field theory ( qft ) .
it allows the use of such a graphical representation , as a tool to tame the combinatorics lurking in the background .
the difficult part in the kontsevich formula for the star - product is to come up with a formula for the coefficients @xmath15 , so that the product would be associative .
the analytical solution of kontsevich for the coefficients @xmath15 is based on integrals over some configuration spaces @xmath16 .
their interpretation as feynman integrals was given in @xcite . in this sense kontsevich solution
uses _ one _ quantization ( with underlying feynman path integral method ) to provide the adequate weights for the quantization of _ all _ poisson manifolds . what about a `` pure '' mathematical solution ?
a tentative combinatorial approach is the primary concern of the present author s work .
the condition corresponding to associativity can be rewritten as an integral over the boundary of the compactification of the above mentioned spaces .
it was noticed in @xcite that the various portions of the boundary in this multi - facet boundary correspond to a certain `` forest formula '' used in renormalization , in order to tame the infinities appearing when computing the physicist s version of the feynman integrals . with the same purpose ,
kreimer @xcite introduced on the vector space of graphs the structure of a coproduct ( see also @xcite ) . if a product `` merges '' the factors to get a result , the coproduct : @xmath17 encodes the internal structure of the graph @xmath18 , representing a list of pairs consisting of a subgraph together with the result of collapsing it in the original bigger graph .
it may be thought of as a list of ways to build the bigger graph from two pieces .
this simple technique corresponding to an almost obvious idea ( well known in combinatorics ) , clarified the renormalization problem ( roughly speaking ) .
the present author used this technique in connection with the formality morphism @xcite , and obtained that the coefficients are ( co)cycles determined by a closely related coproduct ( see also @xcite ) .
this opened a new research avenue of finding a combinatorial formula for these ( or another family of ) coefficients , satisfying the cocycle condition : @xmath19 .
it would not only simplify the computations of the deformation quantization formula , but would have applications in the theory of lie algebras .
the campbell - baker - hausdorff formula , which may be thought of as the forefather of all quantizations , is in need of a `` reasonable explicit expression '' for its coefficients ( @xcite , p.7 ) .
its terms represent various ways of bracketing lie algebra elements , corresponding to @xmath20 , where @xmath18 is a binary tree with @xmath21 vertices , and @xmath7 is the lie bracket .
consider the coaugmented counital dg - coalgebra of kontsevich graphs @xmath22 , with @xmath23 the k - vector space generated by orientation classes of labeled directed graphs @xmath24\in{{\calg}}$ ] . for additional details ( not essential for what follows ) , we direct the reader to @xcite .
the subset @xmath25 of @xmath26 will consist in graphs with @xmath21 internal vertices , @xmath27 boundary vertices and @xmath28 edges , with no loops ( including no circuits ) .
@xmath29 recall that the cocycles condition refers to the dual @xmath30 of the cobar total differential : @xmath31 where @xmath32 and @xmath33 are coderivations on the reduced tensor coalgebra @xmath34 , extending the graph differential : @xmath35)=\sum_{e\in \gamma^{(1)}_{int } } [ \gamma / e , l / e ] , \quad |\gamma^{(1)}_{int}|>0,\ ] ] and the reduced coproduct @xmath36 : @xmath37=\sum_{\gamma\to \gamma\to \gamma ' , \
\gamma\cap\partial\gamma\ne \emptyset } [ \gamma , l|\gamma]\otimes[\gamma',l/\gamma].\ ] ] the above sum is over all _ normal subgraphs _ of @xmath18 meeting the boundary @xmath38 , i.e. such that collapsing @xmath39 to a ( boundary ) vertex yields a graph from the given class @xmath26 .
we begin by introducing a special subclass of graphs .
let @xmath40 denote the orientation class of the labeled _ left bernoulli graphs _
( see also @xcite ) : @xmath41 correspondingly , the _
right bernoulli graph _
@xmath42 is built out of wedges with all right legs grounded to the boundary point @xmath43 .
it may be obtained from the binary tree corresponding to the right parenthesized product of @xmath21 letters , by collapsing all left descending terminals to the boundary point @xmath44 .
the unique terminal which is a right descendent will be denoted by @xmath43 .
a few properties of bernoulli graphs are listed next .
let @xmath45 denote a left ( right ) bernoulli graph .
then ( i ) @xmath45 are primitive elements of the hopf algebra of graphs .
( ii ) @xmath46 .
( iii ) the graph differential of a bernoulli graph is : @xmath47 the graphs @xmath45 are `` simple '' , i.e. their only normal subgraphs are the trivial ones . as a consequence @xmath48 , and ( i ) is established . rephrasing ( i ) gives ( ii ) . in order to prove ( iii ) , note that @xmath49 equals the number of internal edges , and that by collapsing an internal edge @xmath50 , essentially the `` previous '' bernoulli graphs is obtained : @xmath51 .
the claim follows after taking into account the orientation which determines the alternating sign ( by definition @xmath52 ) .
graph cocycles @xmath15 of the cobar construction applied to the above dg - coalgebra determine l - infinity morphisms @xmath53 @xcite .
if @xmath53 is such a morphism and @xmath7 is a poisson structure , thus associative and satisfying maurer - cartan equation @xmath54=0 $ ] , with @xmath55 , then : @xmath56 satisfies the corresponding maurer - cartan equation ( see @xcite , p.12 ) .
the corresponding restriction to a product on @xmath57 ( poisson algebra of functions ) : @xmath58 is associative , and thus yields a star - product .
the above `` philosophy '' can be applied to the campbell - baker - hausdorff formula , viewed as a deformation quantization of the abelian structure underlying the vector space structure , in the direction of the lie bracket ( so called cbh - quantization @xcite ) : @xmath59+\epsilon^2 \frac1{12}([x,[x , y]]+[[x , y],y])+ .... \ ] ] the graphs involved are the _ sym - admissible graphs _
@xcite , for which the corresponding prime graphs are the _ lie admissible graphs _ @xmath60 .
it turns out that since @xmath61 is a lie algebra morphism , then the star - product is determined by the restriction to @xmath62 ( see @xcite ) , for which the corresponding graphs are the bernoulli graphs . in
what follows we will study the associativity condition from the point of view of the above cocycle condition .
given @xmath63 and a system of weights @xmath15 , define the following product : @xmath64 where @xmath57 is the associative algebra @xmath65 for simplicity . from this point on ,
we will consider directed graphs with a fixed order on the boundary vertices , and without additional labels , since the same polyvector field @xmath7 will be attached to all its internal vertices , and the labeling of the edges , used to define the `` state - sum '' , is here irrelevant .
the missing symmetry factors represent the price paid for this simplification .
the graphs with @xmath66 internal vertices from @xmath67 are ( `` prime '' ) bernoulli graphs @xmath68 , or `` products '' of bernoulli graphs @xmath69 ( see [ d : prod ] ) .
the product alluded to above , introduced by @xcite , p.22 , corresponds to the disjoint union of graphs with the same boundary / external structure .
[ d : prod ] if @xmath70 define @xmath71 as the graph obtained from the disjoint union of the two graphs by identification of the boundary vertices .
extend the product to all pairs of graphs , so that all other products are zero .
for example @xmath72 , therefore justifying the notation used .
note that graphs have a unique factorization relative to the above product .
the graph is called _ prime _ if it is not a non - trivial product of graphs .
a few graphs from @xmath73 will be used in examples .
the graphs with @xmath74 internal vertices are denoted by @xmath75 : the following graphs for @xmath76 will be denoted by @xmath77 , @xmath78 ( prime graphs ) and @xmath79 .
the formula [ e : star - product ] produces a star - product , i.e. it is associative , iff for any @xmath80 and functions @xmath81 : @xmath82=0.\ ] ] using gerstenhaber composition @xcite , the above equations can be more compactly written as : @xmath83.\ ] ] the above composition of operators can be replaced with the operator associated to a certain product of graphs .
[ d : star1 ] the _ composition of graphs _ is defined as follows : @xmath84 where @xmath85 is the insertion of @xmath86 at the @xmath87 boundary vertex of @xmath88 using `` leibniz rule '' i.e. summing over all possible graphs where the `` @xmath87 legs '' of @xmath88 lend on vertices of @xmath86 .
if @xmath89 $ ] denotes the set of incoming edges at the @xmath87 boundary point of @xmath88 and @xmath90 $ ] denotes the set of internal vertices of @xmath86 , then the `` leibniz rule '' at the @xmath87 vertex yields : @xmath91\to [ n']}\gamma_1\circ_i^f\gamma_2,\ ] ] where @xmath92 denotes the operation of replacing the vertex @xmath93 by the set @xmath90 $ ] and the edges @xmath94 $ ] by @xmath95 .
the insertion operation is compatible with the product of graphs .
the linear extension of the insertion operation @xmath85 is an algebra morphism : @xmath96 the lemma amounts to the fact that gluing data @xmath97\cup[i^2_i]\to [ n']$ ] is determined by @xmath98\to [ n ' ] , j=1,2 $ ] . for the case
@xmath99 ( see @xcite , p.16 ) : @xmath100 where the indices @xmath101 and @xmath102 where replaced by the more suggestive ones @xmath44 and @xmath43 , representing the left and right boundary points respectively .
note that in this case the product of two graphs from @xmath103 and @xmath104 is a graph of @xmath105 .
a few examples are in order : @xmath106=0,\quad b_1\circ b_1=t_2^r - t_2^l+c_2^l - c_2^r.\ ] ] the last computation of @xmath107 is represented below : this product is essentially the pullback of gerstenhaber composition via the `` feynman rule '' @xmath63 .
it is similar to the pre - lie operation on graphs defined in @xcite , p.27 .
, except insertions are allowed only at boundary vertices .
the _ associator _ of graph composition is : @xmath108 [ d : assoc ] the associated graded ( lie ) bracket is defined by : @xmath109=\gamma_1\circ\gamma_2-(-1)^{(m-1)(m'-1)}\gamma_2\circ\gamma_1 , \quad \gamma_1\in g_{n , m } , \
\gamma_2\in g_{n',m'}.\ ] ] the associator satisfies the following equation : @xmath110 therefore the composition of graphs is a pre - lie composition with associated lie bracket @xmath111 $ ] . it is well know that the above antisymmetry of the associator is equivalent with the associated bracket satisfying the jacobi identity @xcite , p.5 .
a proof similar to the one for the gerstenhaber composition of hochschild cochains @xmath112 establishes the above equation .
the only difference is that here @xmath113 , which is the analog of @xmath112 , is a sum of graphs .
otherwise , the signs and symmetry properties are identical .
[ l : loops ] the composition of two graphs without circuits is a sum of graphs without circuits . since all the arrows connecting the first graph enter the second graph , @xmath114 has no circuits .
consider the basis in @xmath115 dual to @xmath116 , consisting in functionals @xmath117 .
the composition of graphs induces a `` convolution '' on functionals : @xmath118 it will be shown that @xmath119 it is the convolution corresponding to the coproduct @xmath120 introduced initially in @xcite , which is therefore dual to @xmath119 . [
d : coprod ] the _ reduced graph coproduct _ is : @xmath121 where the sum corresponds to connected `` normal '' non - trivial subgraphs with a `` connected '' intersection with the boundary of @xmath18 ( @xmath122 consists of consecutive points of @xmath38 ) .
the sign is determined by @xmath123 , which denotes the number of boundary points to the left of @xmath124 .
the corresponding _ coproduct _ will be denoted by : @xmath125 the `` normal '' subgraphs are those for which the quotient graphs belongs to the same class of graphs . as noted before , one may choose to further restrict the class @xmath116 , by discarding the graphs in the kernel of @xmath63 or with zero contribution in the definition of the star - product .
if @xmath126 , with @xmath127 ( our case of interest ) , then the terms of @xmath128 belong to @xmath129 , where @xmath130 .
the negative sign corresponds to collapsing the two boundary points from the right @xmath131 , while the positive terms correspond to subgraphs @xmath86 containing the boundary points @xmath132 .
again a few examples of reduced coproducts are in order , dual to the above graph compositions : @xmath133 @xmath134 the relation between the constructive definition [ d : star1 ] and the concept of `` graph extension '' and the above coproduct , is the content of the following lemma .
\(i ) for any graphs @xmath135 : @xmath136 where the sum is over the graphs @xmath137 such that @xmath86 is a connected subgraph whose intersection with the boundary of @xmath18 is not empty and connected .
\(ii ) the operation @xmath119 [ e : convol ] , dual to the pre - lie operation @xmath138 , is the convolution corresponding to the coproduct [ e : coprod ] : @xmath139 the coefficient of @xmath126 in @xmath114 is the algebraic sum of _ signed extensions _
@xmath140 of @xmath88 by @xmath86 , such that @xmath86 is a subgraph of @xmath18 satisfying the restrictions imposed on the coproduct [ e : coprod ] and determining the sign @xmath141 , while @xmath88 is the corresponding quotient .
this establishes the first claim ( i ) .
equivalently , ( i ) can be interpreted as stated in ( ii ) by viewing the inner product as a duality and graphs as functionals , concluding the proof .
the requirement that any component of the subgraph should intersect the boundary as a connected subset is necessary to ensure that the collapsing operation is well - defined , and the order of the points on the boundary is preserved .
[ d : bred ] the _ @xmath87 boundary reduction _ of @xmath142 is : @xmath143 where @xmath144 is the pair of boundary points @xmath145 . for example ( with @xmath146 notation for @xmath147 ) : @xmath148 represent the `` top left degree '' terms of the coproduct , and it is related with the `` merger operation '' from @xcite , p.17 ( see definition [ d : merger ] ) .
the restriction of the reduced coproduct to prime graphs plays a special role in defining the logarithm of the star - product ( sum over prime graphs ) .
the _ lie algebra of graphs _
@xmath60 is the pre - lie subalgebra of prime graphs ( e.g. @xmath149 , but not @xmath69 ) .
cobracket _ of @xmath60 is : @xmath150 where the sum is over prime graphs @xmath151 .
[ ck ] note that a normal subgraph of a prime graph is prime and the corresponding quotient is still a prime graph . as
a consequence @xmath152 is the restriction of @xmath120 and is therefore the dual of the pre - lie operation @xmath138 on prime graphs @xmath60 .
this case is similar to case of the connes - kreimer hopf algebra @xcite , p.35 - 37 ) where the dual of the pre - lie operation is the part of the coproduct corresponding to `` prime elements '' ( simple cuts and trees ) , so that the hopf algebra is the uea of the la of `` prime elements '' . as supporting evidence to the above relation we will reformulate the coproduct of @xmath153 in terms of subgraphs rather than using cuts .
the coproduct of @xmath153 has the following structure : @xmath154 where the non - trivial subgraphs @xmath39 are normal subgraphs whose components meet the boundary of the tree @xmath155 ( its terminal points ) .
the terms of the coproduct correspond to cuts which define subgraphs ( forests ) . excepting the empty set and @xmath155 itself , the proper subgraphs are _ normal _ if the quotient is still a tree .
the terms of the cobracket representing the reduction of the coproduct to the lie algebra of prime graphs correspond to simple cuts .
returning to the study of the pre - lie operation @xmath138 itself , the crucial property of the `` dual feynman rule '' @xmath63 is the compatibility between the composition of ( labeled ) graphs and associated operators ( gerstenhaber composition ) .
[ l : pre - lie ] @xmath63 is compatible with the pre - lie compositions : @xmath156 it is enough to note the corresponding relation at the level of individual insertions of labeled graphs : @xmath157 where @xmath158 and @xmath159 ( see @xcite for details ) .
we will need here ( unlabeled graphs case ) the above result for the special case when the same fixed tensor is associated to all the internal vertices .
[ d : kr ] a _ kontsevich rule _ with _ propagator _
@xmath7 ( a fixed tensor of rank @xmath21 ) , associates to an n - valent graph @xmath160 the hochschild cochain @xmath161 , also denoted @xmath162 . with this restriction and notation ,
the above lemma specializes to the following .
[ l : kr ] a kontsevich rule is a morphism of pre - lie algebras : @xmath163 from the above lemma , the pullback of the associativity condition for the star - product follows .
[ p : cc1 ] the star - product [ e : star - product ] is associative @xmath164 iff : @xmath165 since @xmath63 is linear , equation [ e : assoc1 ] reduces conform lemma [ l : pre - lie ] to a constraint equation at the level of graphs : @xmath166 with coefficients determined by @xmath15 via the coproduct : @xmath167 where in the last equality the multiplicative extension of @xmath15 to the tensor algebra of @xmath168 is still denoted @xmath15 . before investigating this `` cocycle condition '' [ e : assoc2 ] in the section [ s : ca ]
( see also @xcite ) , we will comment on some consequence relevant to the deformation theory of associative algebras .
the pre - lie algebra of graphs @xmath169 has a distinguished element @xmath170 which is `` associative '' : @xmath171 .
it therefore defines a differential @xmath172 $ ] , where @xmath111 $ ] is the corresponding graded lie bracket .
the situation is completely analogous with the hochschild dgla which controls the deformation theory of associative algebras .
this is due to the fact that graph composition is the pullback of gerstenhaber composition via the `` kontsevich rule '' @xmath63 ( a kind of a dual feynman rule ) .
this approach follows the current trend of interpreting the linear structures as representations of some discrete geometric objects / categories , e.g. algebras as being representations of operads ( bialgebras and props , cobordism categories and tqfts etc . )
the essential point is that the kontsevich rule is a cohomological pairing , associating to graphs and states on graphs polydifferential operators ( similar to feynman rules and feynman integrals ) : @xmath173 the _ graph pre - lie algebra _ is @xmath174 .
the associated _ graph deformation dgla _ is @xmath175,d)$ ] , with differential @xmath176 $ ] determined by the _ unit _
@xmath170 of @xmath60
. the graph deformation dgla @xmath177,\delta)$ ] controls the deformation of associative algebras .
the _ kontsevich representations _ of @xmath168 ( kontsevich rules for a fixed 2-tensor @xmath7 , not necessarily poisson : definition [ d : kr ] ) , are the hochschild dglas .
the graph deformation dgla is the dual picture to the hopf algebra of graphs ( @xmath138 dual to @xmath152 ) , which is essentially the uea of @xmath60 ( see @xcite ) . in view of the deformation theory
, a star - product corresponds to differentials @xmath178 $ ] : @xmath179 which are perturbations of @xmath180 .
equivalently , if @xmath181 , then @xmath182 is a solution of the maurer - cartan ( deformation ) equation in @xmath168 : @xmath183=0.\ ] ] [ t : deform ] associative deformations are kontsevich representations of deformations of the unit @xmath170 in the direction of the universal cocycle @xmath184 ( equation [ e : comp1 ] ) : @xmath185=0 , \quad \delta(b_1)=[b_0,b_1]=0,\ ] ] satisfying the maurer - cartan equation [ e : mc ] . in view of the above cohomological pairing
, a star - product may be represented as : @xmath186 of course @xmath187\ne 0 $ ] ( @xmath184 is not a solution of maurer - cartan equation ) , but it will be shown that there is no `` primary obstruction '' for the deformation equation : @xmath188 and that in the linear case , with no loops ( at tree - level ) , there is a canonical `` semi - classical '' solution , we call the _ moyal element_. the coefficients of its logarithm @xmath189 ( the _ hausdorff element _ - expansion over prime graphs @xmath190 ) are expected to be the universal coefficients of the hausdorff series of a lie algebra .
properties of the bernoulli graphs , graph composition and its dual coproduct will be investigated next .
linearity of @xmath63 allows to pullback the associativity condition of @xmath191 at the level of graphs : @xmath192 where @xmath193 is the dual of the character @xmath15 : @xmath194 then @xmath195 translates as @xmath196 , which in terms of the coproduct is the equation from proposition [ p : cc1 ] , or in terms of the associated convolution : @xmath197
. we will start with an example specializing the above statements to the case of the star - product modulo terms quadratic in @xmath198 , and rephrase the correction analysis from @xcite , 3.2 , using our tools introduced above . in improvement also
worth to be noted consists in using the gerstenhaber bracket @xmath199 $ ] , rather then the individual terms , e.g. @xmath200 .
it will be also useful to associate a planar forest to graphs without loops , by `` cutting '' the boundary points @xmath44 and @xmath43 . the corresponding operation , as well as its linear extension to linear combinations of graphs , will be denoted as @xmath201 . the ( no loops ) graph is prime iff the associated forest is a tree . at this point
we will introduce the _ global super symmetry _ on graphs : @xmath202 where _ transpose _ graph @xmath203 is obtained by flipping the graph in the plane to revers the order of the boundary points . in our special case @xmath99 , this amounts to the exchange of the `` left '' and `` right '' boundary points .
for example : @xmath204 the global symmetry is an involution @xmath205 compatible with the coproduct and associated convolution @xmath119 ( composition @xmath138 ) .
\(i ) @xmath206 , ( ii ) @xmath207 , ( iii ) @xmath208 .
( direct check ) .
for example ( equations [ e : co1 ] , [ e : co2 ] ) : @xmath209 to exemplify the operations introduced so far we rephrase the correction analysis from @xcite .
consider the linear part of the star - product , i.e. : @xmath210 then modulo quadratic terms ( with notation @xmath211 etc . ) : @xmath212\epsilon= w_0 ^ 2(b_0\circ b_0)+w_0 w_1(b_0\circ b_1+b_1\circ b_0)\epsilon=0\ ] ] where we used the previous computations ( equations [ e : comp1 ] ) .
therefore irrespective of the first two coefficients , @xmath213 starts with quadratic terms .
it is natural to normalize the series in order to correspond to a deformation of multiplication in the direction of the poisson bracket : @xmath214 . as a simpler case to investigate ,
consider first a constant poisson structure ( see @xcite , p.16 )
. then @xmath63 vanishes when evaluated on @xmath77 , and ( equation [ e : comp1 ] ) : @xmath215 to `` compensate '' this term , one has to add the graph @xmath69 , obtained `` merging '' two of the three boundary vertices of @xmath79 .
this prompts for an operation `` inverse to cutting '' the boundary vertices , reducing their number by one .
[ d : merger ] the _ merger operation _ on graphs , @xmath216 , is defined by : @xmath217 where @xmath218 ( @xmath219 ) consists from the two left ( right ) boundary points , while @xmath220 is the unit for @xmath138 .
note that the merger operation is related with the boundary reduction operation ( definition [ d : bred ] ) .
[ r : rec1 ] this operation will be relevant to the recursive formula for the coefficients of @xmath15 , since it corresponds to the `` leading term '' in the coproduct : @xmath221 where the degrees of @xmath39 and @xmath222 are smaller then the degree of @xmath18 .
the main point to keep in mind ( to be established later ) is that one of the last two terms belongs to the kernel of @xmath63 , allowing to solve for the other corresponding coefficient .
[ r : ck ] our graphs without circuits correspond to binary forests under the operation of gluing all left and right terminals respectively : @xmath223 the relation with the connes - kreimer hopf algebra of rooted trees ( with the corresponding coefficients ) will be investigated elsewhere . the correction term to be added to @xmath224
must be a multiple of @xmath225 , and corresponds to the coproduct : @xmath226 let @xmath227 .
then the new quadratic terms produced are : @xmath228=-(c_2^l - c_2^r)\ ] ] and modulo cubic order terms : @xmath229)\epsilon^2= \epsilon^2 ( w_1 ^ 2-w_0\ w_2)(c_2^l - c_2^r).\ ] ] with @xmath230 , @xmath231 gives an associative product up to cubic terms .
note that the multiplicativity of @xmath15 requires @xmath232 , and the equation @xmath233 is already satisfied since @xmath234 : @xmath235 the independent equations @xmath236 seems to correspond to @xmath126 a non - symmetric prime graph ( see lemma [ l : symm ] ) , but at this point it is not clear if there are any `` obstructions '' . as mentioned at the beginning , a symmetry factor of 2
was lost in the above computation . with a labeled @xmath69 ,
@xmath237 $ ] would be twice @xmath238 ( see section [ s : symm ] ) . in the constant coefficient case , at the @xmath239 level stage
the needed correction term is @xmath240 ( @xmath241 for labeled graphs ) , yielding the exponential function in the moyal product @xcite , p.18 ( see also section [ s : cps ] ) .
[ r : kernel ] note that for constant or linear poisson coefficients , one should work at the level of graphs modulo the kernel of @xmath63 .
also note that the correction terms appearing via merger of boundary points form a subclass of graphs ( posets ) , corresponding to the class of binary trees .
more precisely , they belong the class of graphs @xmath242 of graphs generated by forests of binary trees under the merger and composition operations : @xmath243 ( see remark [ r : ck ] ) . in the general case ,
the coefficient of the quadratic ( first non - zero ) term of @xmath244 is : @xmath245 = w_1 ^ 2\ b_1^{\circ 2}+ w_0w_2[b_0,b_1 ^ 2 ] + w_0w_2^l[b_0,b_2^l]+w_0w_2^r[b_0,b_2^r],\ ] ] where : @xmath246 a direct computation gives : @xmath247=t_2^l+c_2 , \quad [ b_0,b_2^r]=-(t_2^r+c_2).\ ] ] to cancel the common symmetric term @xmath78 , one must choose symmetric coefficients : @xmath248 then equation [ e : quadr ] becomes : @xmath249 assuming @xmath15 multiplicative and normalized , and therefore @xmath250 and @xmath251 , the equation [ e : quadr ] simplifies to : @xmath252 it vanishes if @xmath253 , and therefore @xmath193 yields a star - product modulo cubic terms .
note that the above correction did not use the jacobi identity ( see @xcite , p.20 ) : @xmath254 which at the level of graphs is : @xmath255 as a direct computation shows .
taking into account [ e : comp1 ] it is equivalent to : @xmath256 note that the failure of @xmath184 to satisfy maurer - cartan equation : @xmath257=b_1\circ b_1,\ ] ] is mapped by @xmath63 to the associator of @xmath7 . the jacobi identity in turn can be written using the `` graph associator '' under gerstenhaber composition without incorporating the leibniz rule : @xmath258 as noted in remark [ r : kernel ] , one may look for coefficients @xmath259 such that @xmath260 modulo the kernel of @xmath63 , which contains the above relation . in what follows
we will look for solutions without the above relation , i.e. look for generic associative deformations of associative operations in an arbitrary direction . to restore the symmetry factors , consider the map @xmath261 forgetting the labels of graphs of the hopf algebra of labeled graphs . in order to match graph composition @xmath262 of @xmath263 with graph composition @xmath138 of @xmath168
, the following scaling must be applied .
if @xmath264 then @xmath265 is a morphism of hopf algebras .
for example : @xmath266 where @xmath267 denotes a labeled version of the graph @xmath18
. then the composition of labeled graphs exhibits some symmetry factors .
for instance : @xmath268= \tilde{b}_0\circ_l \tilde{b}_1 ^ 2+\tilde{b}_1 ^ 2\circ_l b_0= 2(\tilde{c}_2^r-\tilde{c}_2^l).\ ] ] the correct coefficients for a star - product may be obtained as follows : @xmath269 where @xmath15 is a solution of @xmath270 in @xmath115 .
for example , modulo cubic terms , and omitting @xmath271 : @xmath272 @xmath273=t_2^r - t_2^l,\ ] ] @xmath247+[b_0,b_2^r]=t_2^l - t_2^r,\ ] ] @xmath274+[b_0,b_2^l]+[b_0,b_2^r])\epsilon^2=0.\ ] ] therefore the above @xmath193 defines an associative star - product up to cubic terms .
the above correction analysis has similarities with the bphz renormalization procedure , both relying at some point on the hopf algebra of trees as a device encoding the combinatorics of graph composition .
the first few coefficients of @xmath15 may be found solving the cocycle condition @xmath236 . since the bernoulli graphs @xmath45 are primitive elements , @xmath275 , and there are no constraints for @xmath276 .
we need @xmath277 and the normalization @xmath278 will be assumed for simplicity .
writing explicitly the reduced coproduct for prime graphs with @xmath76 yields the previous values determined using `` correction analysis '' : @xmath279 @xmath280 to understand the structure of the coproduct , introduce the following .
for any @xmath281 , let @xmath282 ( @xmath283 ) denote the number of paths in @xmath18 ending at the left ( right ) boundary point `` l '' ( respectively `` r '' ) of @xmath18 .
the _ hight of @xmath18 _ is @xmath284 .
for example @xmath285 , so @xmath286 .
in general @xmath287 . for all @xmath126 : @xmath288 where @xmath289 and @xmath290 are the subgraphs consisting of the two left / right boundary points of @xmath18 . moreover
, if both terms are present , then : @xmath291 note that there are only two subgraph @xmath292 without interior points : @xmath290 . and any other
subgraph will have a positive degree , decreasing accordingly the degree of the quotient .
the last part follows from the definition of @xmath293 , since if the terms are not equal , there will be at least one path in @xmath18 ending at the middle point of its boundary .
note that the quotients @xmath294 are just the graphs obtained by `` merging '' boundary points in the correction analysis section .
[ l : symm ] let @xmath126 .
\(i ) if @xmath18 is symmetrical ( @xmath295 ) , then @xmath128 is anti symmetrical . consequently , if @xmath15 is equivariant , then : @xmath296 \(ii ) if @xmath18 has no `` middle legs '' ( arrows lending on the second boundary point ) , then @xmath297 .
otherwise [ e : h ] holds and [ e : w ] may be solved for the coefficient @xmath15 of @xmath298 .
\(i ) the subgraphs @xmath292 containing boundary points can not be symmetrical , as have to contain either `` l '' or `` r '' , but not both .
therefore the terms of the coproduct come in mirror pairs with opposite signs . when applying a symmetrical @xmath15 ,
the result is zero .
\(ii ) follows by inspection ( see [ e : sigma ] ) . for example @xmath299 , and @xmath300 will not impose a constraint on the coefficients @xmath15 .
[ p : bern ] the coefficients of a star - product at tree - level are determined by the coefficients of the bernoulli graphs . by induction on the number of internal vertices and `` hight function '' @xmath293 the coefficients @xmath15 are determined by the values on the graphs which minimize the height function : the left bernoulli graphs . to investigate existence and uniqueness of a `` semi - classical '' solution ( graphs without loops / `` tree level '' ) , we will consider the relation with the antipode , aiming for an explicit recursion relation .
the associativity condition may be interpreted in terms of the antipode , establishing a relation ( to be explored ) with the renormalization procedure using the `` counter terms '' method ( bph and zimmermann s formulas , as well as dyson - salam procedure @xcite ) .
let @xmath301 be the group of characters of @xmath168 , equivariant with respect to the global symmetry @xmath302 .
the inverse of a character @xmath259 is @xmath303 , where @xmath304 is the antipode of @xmath168 .
a few values of the antipode can be computed readily using the _ recursive formula _ : @xmath305 where the terms in the above sum corresponds to the reduced coproduct . for the primitive elements
@xmath306 , @xmath307 . moreover , for prime elements with @xmath76 internal vertices we have : @xmath308 the right handed versions of the above formulas may be obtained using the global symmetry : @xmath309 there is an _ explicit formula _ for the antipode , which is obtained by expressing the antipode as a geometric series @xcite , p.3 : @xmath310 where @xmath311 is the unit of @xmath168 and @xmath312 the counit .
the unit in @xmath313 is @xmath314 and @xmath315 .
if @xmath15 is extended to a character @xmath316 on @xmath168 ( @xmath312 counit ) , then the associativity condition for the star - product [ e : assoc2 ] corresponds to : @xmath317 for instance by using the recursive relation for @xmath304 , where @xmath318 is the `` conjugate character '' .
the constant and linear poisson structures will be revisited . in the constant poisson structure case
the `` constraint equation '' is identically satisfied and the well - known moyal formula is recovered .
recall that all internal vertices of graphs in @xmath319 have only two descendents ( @xmath7 is a degree two tensor ) , but may have several `` parents '' ( @xmath320s acting on the corresponding coefficient of @xmath7 ) . for a constant poisson structure
@xmath321 the non - trivial graphs @xmath322 ( in @xmath67 , but not in the kernel of @xmath63 ) have bi - valent internal vertices with only outgoing edges , therefore ending on boundary vertices : @xmath323 they are generated by the only non - trivial bernoulli graph @xmath184 .
the prime graphs of @xmath324 are @xmath75 , with coproducts : @xmath325 then equation @xmath236 is identically satisfied on , and @xmath326 is determined by our initial conditions @xmath214 . with explicit symmetry factors
the moyal element is : @xmath327 the corresponding hausdorff element ( its logarithm ) is @xmath328 .
representing @xmath193 yields moyal s formula : @xmath329 .
this is the commutative case corresponding to the abelian lie algebra with one generator @xmath184 .
the `` general case '' from the point of view of lie bialgebras @xcite and corresponding to linear poisson structures , will be studied next . in the linear poisson structure case ( e.g. dual of a lie algebra @xcite , p.5 )
the equation can be solved recursively , yielding a canonical semi - classical ( tree - level ) star - product .
the coefficients are essentially those of the hausdorff series ( cbh - quantization @xcite ) . for a linear poisson structure : @xmath330 the non - trivial graphs have only tri - valent internal vertices , corresponding to trees .
the number of edges of @xmath331 is @xmath332 . if @xmath281 is such that @xmath333 , then each internal vertex has only one parent .
equivalently , if @xmath18 has at least one n - valent internal vertex with @xmath334 , then @xmath335 .
internal vertices with more then one parent will contribute a zero factor to the corresponding product in any term of the state - sum .
the correspondence between trees and the relevant class of graphs @xmath26 not belonging to the kernel of @xmath63 : @xmath336 is defined by collecting together the left terminal points to the boundary point `` l '' and all right terminal points to the boundary point `` r '' . it extends to a linear map between the corresponding hopf algebras @xmath153 and @xmath337 .
it has an inverse @xmath338 ( spanning tree / forest ) , obtained by `` cutting '' the boundary points `` l '' and `` r '' ( letting lose the terminal edges ) .
now the ( left ) bernoulli graphs , corresponding to left parenthesized expressions / trees , minimize the height function .
moreover the star - product is determined by the value of @xmath15 on bernoulli graphs .
these in turn are determined recursively by noting that for each @xmath45 there is a graph @xmath339 such that @xmath340 for example @xmath341 is the following graph : @xmath342 as a consequence the values of @xmath15 are determined inductively .
since only tree - like graphs have been considered , the corresponding star - product may be thought off as a semi - classical solution , i.e. not including quantum corrections due to loops / circuits .
this star - product corresponds to the hausdorff series , except our coefficients neglect the symmetry factors , yielding @xmath343 ( ! ? ) .
its logarithm ( a sum over prime graphs ) , is expected to correspond to the hausdorff series .
a linear poisson structure has a unique semi - classical solution of the initial value problem for the star - product at tree - level @xmath344 the coefficients of the associated moyal element are determined inductively from the equation [ e : assoc2 ] .
it is expected that the antipode may be used to derive further formulas for the coefficients .
deformation quantization based on star - products require finding coefficients which are cocycles in some sense . to quantize any poisson manifold ,
kontsevich used a simple string theory ( non - linear sigma model on the poincare disk ) to write explicit formulas for the ( finite ) feynman integrals / amplitudes .
then the resulting cocycle was used in the general case . in the present article a combinatorial approach
is sketched , leading to a self contained mathematical solution .
it was shown that the cocycles reside in the deformation dgla of graphs controlling the deformation theory of associative algebras .
star - products appear as `` kontsevich representations '' elements @xmath193 such that @xmath196 . when restricted to graphs without circuits a canonical solution of the deformation equation exists (
@xcite ) : the `` moyal element '' . in the special case of poisson structures with constant coefficients ,
its representation yields moyal s formula . for linear coefficients
, it is expected to correspond to the hausdorff series .
the hopf algebra techniques have proven as a powerful conceptual interface to the combinatorial practices of renormalization @xcite .
a direct connection with renormalization is likely to be established via the hopf algebra of graphs ( connes - kreimer , loday - ronco etc . ) . for a linear poisson structure ,
the class of non - trivial ( binary ) prime graphs @xmath26 ( not in the kernel of the `` feynman rule '' @xmath63 ) is in 1:1 correspondence with binary trees ( see equation [ e : corresp ] ) . note that in this hierarchic case ( no multiple parents ) , there is no need for considering labeled trees , as in the renormalization of feynman graphs .
a correspondence between the corresponding hopf algebras is expected to originate from a correspondence at the level of pre - lie algebras .
further considerentions regarding the relations with renormalization will be addressed elsewhere .
the first few compositions and coproducts are given below .
the right - handed version may be obtained by symmetry , applying the hopf involution @xmath345 .
a. connes and d. kreimer , renormalization in quantum field theory and the riemann - hilbert problem i : the hopf algebra structure of graphs and the main theorem .
* 210 * ( 2000 ) , no .
1 , 249273 . hep - th/9912092 . | graph cocycles for star - products are investigated from the combinatorial point of view , using connes - kreimer renormalization techniques .
the hochschild complex , controlling the deformation theory of associative algebras , is the `` kontsevich representation '' of a dgla of graphs coming from a pre - lie algebra structure defined by graph insertions ( gerstenhaber composition with leibniz rule ) .
properties of the dual of its uea ( an odd parity analog of connes - kreimer hopf algebra ) , are investigated in order to find solutions of the deformation equation .
the solution of the initial value deformation problem , at tree - level , is unique . for linear coefficients
the resulting formulas are relevant to the hausdorff series .
xypic epsf # 1by # 1 # 1by # 1 0.5 in 6.5 in 8.5 in [ section ] [ section ] [ section ] [ section ] [ section ] [ section ] [ section ] [ section ] |
in the last years a great deal of interest arose for understanding the formation of the first structures in the universe at the end of the so called dark age epoch .
this interest is motivated not only by the possibilities of direct measurements of the physical conditions prevailing in these ( proto)structures and the constraining of cosmological parameters at very high redshifts , but also by the measurement of the primordial abundances of key elements ( e.g. , d and li ) in pregalactic epochs as a direct signature of the big bang nucleosynthesis ( bbn ) . from the observational side , the re emission or absorption of cosmic microwave background radiation ( cmbr ) photons in resonant lines from h@xmath3 , lih , and hd , among other primordial molecules ,
have been suggested as a viable way to detect cosmic protostructures ( dubrovich 1977,1983 ; maoli , melchiorri & tosti 1994 ; maoli et al.1996 ) .
resonant absorption by neutral h at its 21 cm transition has also been proposed to trace protostructures at early cosmic times ( hogan & rees 1979 ; for more recent studies see barkana & loeb 2005 and more references therein ) .
the first attempt to detect lih emission and doppler induced anisotropies in the cmbr from the protostructures at redshift @xmath4 was made with the 30 meter iram radio telescope ( de bernardis et al .
1993 ) . only upper limits for the abundance of lih were obtained .
it was believed that the relative abundance of lih , [ lih / h ] , at @xmath5 could be rather adequate ( up to @xmath6 ) to produce detectable spectral features due to the optical depth in the lines of the lih rotational structure .
however , more detailed calculations of the primordial lih relative abundance have shown that , at the discussed epochs ( @xmath5 ) , it should have been actually very small , [ lih / h ] @xmath7 ( stancil , lepp & dalgarno 1996 ; bougleux & galli 1997 ) .
therefore the optical depth in the rotational lines of lih in the primordial gas is too small to produce observational features ( galli & palla 1998 ; puy & signore 2001 ; maoli et al .
2004 ) .
the other primordial molecule , potentially interesting to detect features of precollapse structures from the dark epoch , is deuterated hydrogen , hd . the dipole moment of the hd molecule is actually much smaller than the one of lih : @xmath8 debyes ( abgrall et al . 1982 ) and @xmath9 debyes .
however , for [ lih / h]@xmath10 and [ hd / h]=@xmath11 , one obtains in a first approximation @xmath12 [ hd]d@xmath13 [ lih ] d@xmath14 , concluding that hd is a species more suitable for detecting the primordial structure signatures than lih .
besides , the hd molecule is also able to amplify its emission due to ro
vibrational luminescence effects ( see e.g. , dubrovich 1997 ) . at the moment , the most remote of the detected absorbers in rotational lines of hd molecule is at redshift @xmath15 in the spectrum of the quasar @xmath16 ( varshalovich et al .
nevertheless , the emission of this molecule should be observable even from the pregalactic epoch . in the intermediate epochs at @xmath17 , where nonlinear collapse takes place for the mass scales of interest
, the hd molecule can reemit the kinetic energy of gas out of equilibrium . the emission from primordial molecular clouds in the pure rotational structure of hd molecule owing to cooling processes
has been calculated first time by shchekinov ( 1986 ) and more accurate by kamaya & silk ( 2003 ) .
recently , lipovka , nez lpez & avila reese ( 2005 ) have revisited the cooling function of this molecule by taking into account its ro vibrational structure . at temperatures above @xmath18k and at high densities ,
the cooling function is 12 orders of magnitude higher than reported before e.g. , ( flower et al . 2000 ) .
if photo ionization or shocks heat the dense collapsing clouds to temperatures above @xmath19k , these new calculations show that the gas will be able to cool efficiently again with further hd resonant line emission , and the intensity of the collapse energy reemission by the hd molecule calculated by kamaya & silk ( 2003 ) will increase probably due to the ro - vibrational structure of this molecule .
recently , it was also suggested that the h@xmath20 and hd fractions can be substantially large in relic hii regions ( johnson & bromm 2006 ; yoshida 2006 ) , as well as in large mass haloes in which the gas is collisionally ionized ( oh & haiman 2003 ) . at the linear stage of evolution of the first structures of interest ( @xmath21 ) , the elastic resonant scattering of the cmbr by primordial molecules in the evolving protostructures can damp cmbr primary anisotropies or , if the protostructures have peculiar velocities , secondary anisotropies in emission ( spectral spatial fluctuations ) can be produced ( dubrovich 1977,1993,1997 ) . in these cases ,
internal energy sources are not required to produce the observational spectral and spatial features
. therefore , _ primordial resonant lines are ideal to study the evolution of cosmic structures during their linear phase or the first phases of their collapse , before star formation triggers_. the main goal of the present work is to study the possible observational signatures from these phases previous to the collapse of luminous objects in the universe .
the hd molecule have at least two advantages for this task : 1 ) it is formed rather fast in the homogeneous primordial gas at @xmath22 , through channels similar to those of molecular hydrogen , and it may reach relatively high abundances at those epochs . indeed , at @xmath4 , the relative abundance of hd is already @xmath23 of the presend
2 ) the redshifted wavelength range of the rotational structure transitions of the hd molecule is very comfortable for detection .
the first , second , and third rotational transitions at the wavelength of 112.1 @xmath24 , 56.2 @xmath24 , and 37.7@xmath24 can be detected at 0.83.5 mm if the lines are produced at redshifts around 2030 , typical of the epochs when the first stars probably started to form according to numerical and semi analytical results ( for a review see bromm & larson 2004 ) .
several facilities planned or under construction will cover the submillimetric and millimetric wavelength ranges with high sensitivity , for example the large millimetric telescope ( lmt , or gtm for its acronym in spanish ) in mexico , whose radiometers cover the range from 0.8 to 3 mm .
galli & palla ( 2002 ) presented results from detailed calculations of chemical kinetics in thermally evolving primordial clouds , though they did not follow the hydrodynamical evolution .
the relative abundance of the hd molecule that these authors found is very high in the cooled gas clouds and , as it will be shown later here , must be observable , so , more accurate calculations of the protocloud and cloud formation process and emission by the hd molecule are required .
the aim of this paper is twofold : to recall the importance of the hd molecule for investigating the very high redshift universe , and to calculate the parameters of hd lines of interest .
the latter is a crucial step for planning future observational strategies .
we will estimate the optical depth for protoclouds in the rotational lines of the primordial hd molecule based on the @xmath1 cold dark matter ( @xmath1cdm ) scenario of cosmic structure formation . for alternative scenarios ( e.g. , khlopov & rubin 2004 )
, the predictions will probably be different . with the aim to obtain useful estimates , here
we calculate detailed chemical kinematics evolution for hd molecule in the expanding homogeneous medium ( 2 ) .
then we calculate the properties of high
density @xmath1cdm mass perturbations ( protohaloes ) at their maximum expansion ( 3.1 ) , and estimate the optical depths in the three lowest rotational line transitions of the hd molecule fraction in these protohaloes ( 3.2 ) .
we bear in mind that these opacities will slightly decrease in the first phases of collapse when the line widths increase and become potentially observable as secondary spatial frequency anisotropies in the cmbr .
we estimate the amplitudes and angular sizes of these anisotropies assuming the linear model for the peculiar velocities of the mass perturbations , and explore the possibility to detect them by the lmt / gtm submillimeter telescope under construction in mexico ( 3.3 )
. finally , we present a discussion of the implications of our results ( 4 ) .
following , we calculate the evolution of the hd abundance in the homogeneous pregalactic medium . this is a good approximation to the hd abundance in mass perturbations at their linear evolution phases as yet .
we use the concordance cosmological model ( @xmath1cdm ) with @xmath25 , @xmath26 , @xmath27 and @xmath28 kms@xmath29mpc@xmath29(e.g . , spergel et al . 2003 ) .
standard bbn yields for the light elements are assumed . in the early epochs , before the first cosmic objects are formed , the thermodynamics of the primordial gas is out of equilibrium .
therefore , in order to describe the molecular dynamics , we need to solve the kinetic equation system , which can be written as follows : @xmath30where @xmath31 is the cosmic time , @xmath32 is the density of the species @xmath33 , @xmath34 are the rates of the collisional processes @xmath35 , as functions of the kinetic temperature @xmath36 , @xmath37 are the rates of the radiative processes ( formation and destruction of the molecule by the cmbr photons characterized by the radiative temperature @xmath38 ) . in eq .
( [ kinetic ] ) the positive and negative terms correspond to the formation and destruction processes of the species @xmath32 , respectively .
to solve the equations system ( [ kinetic ] ) , it is convenient to introduce new variables : @xmath39 , where @xmath40 and @xmath41 is the present - day baryon number density , @xmath42 .
it is also convenient to substitute the cosmic time variable with the redshift , by using the appropriate relation for the flat cosmology with cosmological constant adopted here : @xmath43with these new variables , eqs .
( [ kinetic ] ) take the form : @xmath44 .
\label{17}\end{aligned}\ ] ] as was mentioned above , the reaction rates @xmath45 and @xmath46 are functions of the radiative temperature @xmath38 and the kinetic temperature @xmath36 , which depend on many parameters . the radiative temperature can be taken as in the case of the adiabatic expansion , @xmath47 , but in order to find the kinetic temperature we have to solve the differential equation for @xmath36 : @xmath48 where the first term in the right hand part describes the change on the temperature due to the universe expansion , the second one corresponds to the thermodynamics of the primordial molecules , and the third one is thomson scattering .
the second term can be written as follows : @xmath49where @xmath50 and @xmath51 are the heating and cooling functions , @xmath52 is the energy gain ( loss ) due to the chemical reactions , characterized by their enthalpy @xmath53 , @xmath54the summation is carried out over all chemical reactions .
the factor @xmath55 in the last term of eq .
( secondterm ) corresponds to the change in the primordial gas density due to the chemical reactions , which do not conserve the number of initial species ( for example @xmath56 , or @xmath57 ) .
the calculations of the molecular abundances were carried out for a wide range of redshifts for the following species involved in hd molecule formation : h , h@xmath58 , h@xmath59 , h@xmath3 , h@xmath60 , e@xmath60 , h@xmath61 , he , he@xmath58 , heh@xmath58 , d , hd , h@xmath3d@xmath58 , d@xmath58 , d@xmath60 , and hd@xmath62 .
other species are negligible in the formation processes of the hd molecule due to their small abundances .
the following reactions are of particular importance in the formation / destruction processes of the hd molecule in the primordial gas : @xmath63 more detailed consideration can be found in the many papers ( see for example galli & palla 1998,2002 ) .
the results of our molecular abundance calculations are shown in fig . 1 .
one can see that the relative abundance ratio [ hd]/[h@xmath3 ] at low redshifts is slightly larger than @xmath65 , whereas at high redshifts this value tends to the assumed primordial ratio [ d]/[h]=3 10@xmath66 .
this fact is explained as the deuteration process , d@xmath67h@xmath68 h@xmath67hd and d @xmath69 h@xmath70 hd @xmath69 h , that leads to relatively high abundance of the molecule , up to [ hd]/[h@xmath3]@xmath71 at @xmath72 in collapsed clouds ( see e.g. , galli & palla 2002 ) .
we are now interested in obtaining estimates related to the observability of the line features of the primordial hd molecule in protostructures in their linear or quasi linear evolution regime , i.e. before gravitational collapse , and first star formation .
as mentioned in the introduction , previous works have shown that the largest resonant line opacities in homogeneous mass perturbations are produced when they are at their maximum expansion or turn around epoch .
based on these conclusions , here we will calculate the hd resonant line opacities at the turnaround epoch of cdm high peak mass perturbations . after the first bursts of star formation ( population iii stars ) in the high peak ( @xmath73 ) collapsed perturbations , re ionization ,
feedback and metal enrichment processes change drastically the properties of the medium in the lower density mass perturbations .
this is why we will focus on the protostructures emerging from high 36@xmath74 peaks .
the interaction of cmbr photons with the primordial molecules inside these protostructures will produce signatures related to the pristine conditions of gas before the first star formation in the universe .
@lrlllrlllrrl@ @xmath76 & @xmath77 & @xmath78 & @xmath79 & @xmath80 & @xmath81 & @xmath82 & @xmath83 & @xmath84 & @xmath85 & @xmath86 & @xmath87 + @xmath88 & & @xmath89 & @xmath90 & @xmath91 & & @xmath89 & @xmath90 & @xmath91 & @xmath92 & & @xmath93 + + @xmath94 & 58.52 & 3.02 @xmath95 & 2.67 @xmath96 & 2.50 @xmath97 & 36.49 & 2.40 @xmath98 & 2.12 @xmath99 & 1.26 @xmath97 & 13.58 & 32.65 & 2.12 @xmath100 + @xmath101 & 46.26 & 1.51 @xmath95 & 1.34 @xmath96 & 1.46 @xmath102 & 28.77 & 1.20 @xmath98 & 1.06 @xmath99 & 7.34 @xmath97 & 19.20 & 25.72 & 1.06 @xmath100 + @xmath103 & 34.55 & 6.44 @xmath104 & 5.70 @xmath105 & 9.03 @xmath102 & 21.40 & 5.11 @xmath95 & 4.52 @xmath96 & 4.53 @xmath102 & 29.43 & 19.10 & 4.52 @xmath94 + @xmath106 & 24.17 & 2.28 @xmath104 & 2.02 @xmath105 & 5.92 @xmath107 & 14.85 & 1.81 @xmath95 & 1.60 @xmath96 & 2.97 @xmath107 & 49.40 & 13.23 & 1.60 @xmath94 + @xmath108 & 15.18 & 6.07 @xmath109 & 5.37 @xmath65 & 4.28 @xmath110 & 9.19 & 4.82 @xmath104 & 4.27 @xmath105 & 2.14 @xmath110 & 95.78 & 8.15 & 4.27 @xmath111 + @xmath112 & 7.92 & 1.02 @xmath109 & 8.99 @xmath113 & 3.60 @xmath114 & 4.61 & 8.10 @xmath109 & 7.16 @xmath65 & 1.80 @xmath114 & 233.69 & 4.03 & 7.16 @xmath115 + + @xmath94 & 118.04 & 2.42 @xmath98 & 2.14 @xmath99 & 1.25 @xmath97 & 73.99 & 1.92 @xmath116 & 1.70 @xmath117 & 6.28 @xmath118 & 4.80 & 66.30 & 1.70 @xmath101 + @xmath101 & 93.53 & 1.21 @xmath98 & 1.07 @xmath99 & 7.32 @xmath97 & 58.55 & 9.60 @xmath98 & 8.49 @xmath99 & 3.67 @xmath97 & 6.79 & 52.45 & 8.49 @xmath100 + @xmath103 & 70.10 & 5.15 @xmath95 & 4.56 @xmath96 & 4.52 @xmath102 & 43.79 & 4.08 @xmath98 & 3.61 @xmath99 & 2.26 @xmath102 & 10.40 & 39.20 & 3.61 @xmath100 + @xmath106 & 49.34 & 1.83 @xmath95 & 1.62 @xmath96 & 2.96 @xmath107 & 30.71 & 1.45 @xmath98 & 1.28 @xmath99 & 1.48 @xmath107 & 17.47 & 27.46 & 1.28 @xmath100 + @xmath108 & 31.37 & 4.86 @xmath104 & 4.30 @xmath105 & 2.14 @xmath110 & 19.39 & 3.86 @xmath95 & 3.41 @xmath96 & 1.07 @xmath110 & 33.86 & 17.30 & 3.41 @xmath94 + @xmath112 & 16.86 & 8.16 @xmath109 & 7.22 @xmath65 & 1.80 @xmath114 & 10.25 & 6.48 @xmath104 & 5.73 @xmath105 & 9.02 @xmath110 & 82.62 & 9.10 & 5.73 @xmath111 + to calculate the optical depth of primordial protoclouds by scattering of cmbr photons on hd molecules we need a model for these protoclouds .
we use the spherical top hat approach to calculate the epoch , density , and size of cdm mass perturbations at their maximum expansion for the @xmath1cdm cosmology adopted here .
this approach is justified for the kind of estimates in which we are interested ( e.g. , tegmark et al .
1997 ) , and for the quasi lineal phases of gravitational evolution .
it is assumed that gas is equally distributed as cdm in the mass perturbation but with a density @xmath119 times lower . according to the spherical top
hat approach , the overdensity of a given mass perturbation , @xmath120 , grows with @xmath121 proportional to the so called growing factor , @xmath122 , until it reaches a ( linearly extrapolated ) critical value , @xmath123 , after which the perturbation is supposed to collapse and virialize at redshift @xmath124 ( for example see padmanabhan 1993 ) : @xmath125the convention is to fix all the quantities to their linearly extrapolated values at the present epoch ( indicated by the subscript `` 0 '' ) in such a way that @xmath126 .
the redshift of turn around , @xmath127 , is calcualted by using the top hat sphere result of @xmath128 . following the evolution equation for the top hat sphere , one finds for our flat @xmath1cdm cosmology that the average sphere density at @xmath127 is given by @xmath129 , i.e. similar to the einstein de sitter case ( wang & steinhardt 1998 ; horellou & berge 2005 ) . the background density at @xmath127 is @xmath130 g @xmath131 .
the corresponding numerical gas density is @xmath132 , where @xmath133 is the h atom mass .
finally , the radius of the sphere of mass @xmath76 at the turn around or maximum expansion is given by @xmath134 ^{\frac{1}{3}}. \label{radmax}\ ] ] we need now to connect the top
hat sphere results to a characteristic mass of the perturbation , @xmath76 , produced within the @xmath1cdm structure formation scenario .
the primordial fluctuation field , fixed at the present epoch , is characterized by the mass variance , @xmath135 , which is the rms mass perturbation smoothed on a given scale @xmath136 corresponding to the mass @xmath76 . for our calculations ,
we use the power spectrum shape given by bardeen et al .
( 1986 ) normalized to @xmath137 .
thus , the mass perturbation overdensity linearly extrapolated to @xmath138 can be defined as : @xmath139 where @xmath76 and @xmath140 are the perturbation mass and peak height , respectively . for average perturbations , @xmath141 , while for rare , high density perturbations , from which emerged the first structures , @xmath142 . by introducing eq .
( [ d0cdm ] ) into eq .
( [ delta ] ) one may infer @xmath143 , and therefore @xmath127 , @xmath144 , @xmath145 , and @xmath146 for @xmath147 mass perturbations from the @xmath1cdm primordial fluctuation field . in table 1
we present these quantities for @xmath148 perturbations of different masses . as mentioned above
, we are interested in hd line signatures from the protoclouds before first luminous objects in the universe formed and started to re ionize it , and therefore only high peak low mass perturbations will be considered .
cosmological n
body + hydrodynamics simulations showed that the first stars may have formed indeed at redshifts as high as @xmath149 in ( @xmath150 ) cdm mini
haloes of m@xmath151 ( abell et al . 1998 ; 2002 ; fuller & couchman 2000 ; bromm et al .
2002 ; yoshida et al .
related to this is the fact that haloes of any given mass that collapse first ( high peaks ) do not populate `` typical '' regions at all , but rather `` protocluster '' regions ( white & springel 2000 ; barkana & loeb 2002 ) . in a recent paper , gao et al .
( 2005 ) used a novel technique to follow with unprecedentedly high resolution the growth of the most massive progenitor of a supercluster region from @xmath152 to @xmath138 . by using this simulation , reed et al .
( 2005 ) have found that , when the mass of the progenitor halo is @xmath153h@xmath154 at @xmath155 , it undergoes baryonic collapse via h@xmath20 cooling , triggering star formation at this early epoch . of course
, this halo should emerge from a very rare fluctuation peak , @xmath156 ( reed et al .
notice that rare peaks are strongly clustered , so that the probability to find a cluster of these ( proto)halos is high . in the second part of table 1 we present the same as in the first part , but for @xmath75 haloes .
the optical depth for a protocloud of size @xmath157 is given by : @xmath158where @xmath159 is the absorption coefficient , and integration is carried out over the line of sight . in the case of a mass perturbation at its turn around epoch , when the physical parameters of the gas are approximately equal for all parts of the spherical region , the integration can be reduced to a more simple expression : @xmath160 .
so , the optical depth in this case can be written as : @xmath161where @xmath162 is the wavelength , @xmath163 is the rotational quantum number , @xmath164 is the relative abundance of the hd molecule , @xmath165 is the einstein coefficient , and @xmath166 and @xmath167 are the baryon and @xmath163th rotational level population numerical densities at epoch @xmath121 , respectively . as to @xmath168 , it is well known that at the low density limit ( @xmath169@xmath170 ) , the population differs from the boltzmann one .
so , in the general case , when the gas is out of thermodynamical equilibrium , the correct values of the ro vibrational levels population should be calculated with the detailed balance equation : @xmath171where @xmath172 is the population of the ro
vibrational level @xmath173 , @xmath174 and @xmath175 are the probabilities of the radiative and collisional transitions , respectively . by using the expressions ( [ tau2 ] ) and ( [ pop ] ) , we calculate the hd line optical depths corresponding to protohaloes of several masses at their turnaround redshifts ( table 1 ) . the results corresponding to three ground rotational line transitions ( 10 ) , ( 21 ) and ( 32 ) for the @xmath148 and @xmath75 protohaloes are plotted in figs . 2 and 3 , respectively .
as one can see , the values of the optical depth for the hd molecule lines are rather large as compared with those reported for the lih molecule ( bougleux & galli 1997 ) .
even the transition @xmath176 reaches values as high as @xmath177 for the redshifts @xmath178 . the estimates presented above correspond to the linear evolution phase of fluctuations .
the calculation of chemical kinematics evolution of hd molecule during the gas cooling and collapse inside the virializing dark haloes is very difficult since it is linked to the thermal and hydrodynamical evolution . under several assumptions and simplifications , galli & palla ( 2002 ) presented results of hd abundances after the gas cooling inside virialized cdm haloes constructed according to tegmark et al .
the result of galli & palla is that the hd relative fraction , @xmath179 , increases dramatically , approximately by a factor of 200 , with respect to the initial one at the collapse epoch of the halo . in the following
, we will use this result to get a rough estimate of the hd relative fraction in collapsed @xmath1cdm haloes and the corresponding @xmath180 .
we have calculated above the evolution of the hd relative fraction , @xmath179 , in the homogeneous pregalactic medium . for a given mass fluctuation @xmath181 , we have obtained @xmath164 at @xmath127 .
let us assume that after the halo collapsed , at @xmath143 , @xmath182 increased only proportionally to the halo density increasing .
therefore , @xmath183 .
then , we assume that the gas cools rapidly and falls to the halo center in a dynamical time , @xmath184where @xmath185 is the halo radius .
now , according to the galli & palla result , we assume that at the epoch when the gas cooled and collapsed , @xmath186 , the hd relative fraction increased by a factor of @xmath187 , i.e. @xmath188 . the collapsing factor of the gas is assumed to be 10 , so that the number density increases by a factor of 1000 .
we further assume that the kinetic temperature of the cooled gas is equal to the cmbr radiation ( see fig .
10 in galli & palla ) .
the populations corresponding to the main ground line transition ( @xmath189)=(10 ) actually depend weakly on @xmath190 . in fig .
4 we show the hd ( @xmath189)=(10 ) line transition optical depths corresponding to 3 and @xmath75 haloes of several masses at their `` cooling '' redshifts . the radii used to calculate
@xmath191 were 1/10th of the corresponding halo radii @xmath192 .
for the redshifts of interest , i.e. before massive star formation ( @xmath193 ) , the optical depths of the lines are significantly high , with values from several times @xmath105 to several times @xmath194 , up to @xmath195 .
as shown by sunyaev & zeldovich ( 1970 ) ( see also dubrovich 1977 ; zeldovich 1978 ; maoli et al .
1994 ) the amplitude of secondary anisotropies in the cmbr temperature due to doppler effect of resonance scattered photons in protoclouds with peculiar motions is : @xmath196for @xmath197 , where @xmath198 is the peculiar velocity of the protocloud with respect to the cmbr at epoch @xmath121 , @xmath199 is the light speed , and @xmath200 is the optical depth at frequency @xmath140 through the protocloud at epoch @xmath121 ( to be the protocloud @xmath127 in our analysis ) .
the linear theory of gravitational instability shows that the peculiar velocity of every mass element grows with the expansion factor as @xmath201 .
an accurate approximation to this expression for a flat universe with cosmological constant is ( lahav et al .
1991 ; carroll et al .
1992 ) : @xmath202where @xmath203 , and the functions @xmath204 , @xmath205 and @xmath206 were defined in 3.1 . for the einstein
de sitter universe , eq . ( [ vpec ] ) reduces to the known expression @xmath207 .
the peculiar velocity field at @xmath138 , traced by galaxies and galaxy clusters , has been measured in a large range of scales ( for a review see e.g. , zaroubi 2002 ) . for our problem ,
the velocity field traced by galaxy clusters is probably the interesting one because , as mentioned above , the first luminous objects in the universe formed in the densest , most rare regions , i.e. those that today are clusters of galaxies . various data sets lead to different rms peculiar velocities for scales @xmath208h@xmath29mpc . here
we will adopt a value of @xmath209km / s , in agreement with results from lauer & potsman ( 1994 ) , hudson et al .
( 1999,2004 ) , and willick ( 1999 ) .
this value is apparently a factor of @xmath210 larger than the one corresponding to @xmath1cdm numerical simulations for the scales of interest . by using eq .
( [ vpec ] ) normalized to @xmath211km / s and the hd optical depths calculated in 3.2 , we may estimate with eq .
( [ deltat ] ) the corresponding temperature fluctuation of the secondary anisotropies . in the range of @xmath212 , for the case of the @xmath213 protohaloes
, we obtain that @xmath214 in the first rotational transition , while for the @xmath215 protohaloes , @xmath216 @xmath217 . in the same redshift range , for the collapsed baryonic clouds ,
whose hd abundances were estimated using results from galli & palla ( 2002 ) , we calculate @xmath218 and @xmath219 for the 3 and @xmath75 cases , respectively .
now let us estimate the integration time required for the detection of the secondary anisotropies with millimeter and submillimeter facilities under construction as for example : the 50 m lmt / gtm telescope in sierra negra , mexico ; the `` combined array for research in millimeter wave astronomy '' ( carma ) in usa ; the `` atacama large millimeter array '' ( alma ) in chile .
the observational ( integration ) time @xmath220 can be estimated from the equation @xmath221where @xmath222 is the bandwidth , @xmath223 is the detector noise temperature and @xmath224 is the amplitude of the temperature fluctuation calculated above .
for the three facilities mentioned , @xmath225 k , and we may assume @xmath226 . in the case of gtm / lmt telescope , its angular resolution limits the observability of several cases ( see below ) . by using eq .
( [ dt ] ) , for the @xmath213 and @xmath215 protohaloes , we find that the integration times required to get an observable signal of the first rotational transition of hd molecules are too large to be observed . in the case of collapsed gas clouds ( previous to star formation triggering ) inside virialized @xmath75 haloes , we estimate @xmath227 hours for @xmath228 and @xmath229 k. this integration time can be attained at different observational sessions of several hours each .
let us discuss another important parameter of the potentially observed protoclouds in molecular resonant line emission : their apparent angular size or diameter .
apparent angular size for an object of the linear size @xmath157 is given by the hoyle s formula : @xmath230where @xmath231 is the angular - diameter distance .
the angular
diameter distance for a flat universe with cosmological constant is given by : @xmath232where the hubble parameter in this case is : @xmath233and the radiation density term was neglected . now eq .
( [ hoyle ] ) can be written as follows : @xmath234 we use eq .
( [ angsize ] ) to calculate the angular size @xmath235 of @xmath236 and @xmath215 protoclouds reaching their maximum expansion in the redshift range of @xmath237 ( corresponding roughly to mass ranges of @xmath238 and @xmath239 for the @xmath213 and @xmath75 cases , respectively ) .
we calculate also the corresponding redshifted frequencies , @xmath240 , for the same three lowest rotational transitions of hd molecule considered in figs . 2 and 3 .
our results are shown in fig . 5 , along with the expected coverage in this @xmath241 plane for the lmt / gtm and carma facilities ( boxes ) .
the alma facility covers the whole plotted plane .
we also include in this plot the curves corresponding to the cold clouds within collapsed @xmath236 and @xmath75 haloes for only the first rotational transition line ( thin lines with square dots ) .
these clouds could be resolved by alma and only partially by carma .
it should be noted that high peak halos are strongly clustered ( gao et al .
2005 ; reed et al .
. it may thus be more plausible to find a cluster of such objects in a patch of the sky even if individual objects are very small .
moreover , the clustering increases the amplitude of the spectra spatial fluctuations in the cmbr temperature .
figure 5 is rather general and can be applied to any other resonant lines of similar frequencies in which the gas in protohaloes or haloes could be detected .
the hope to detect signals from the end of the so called dark age has prompted a flurry of theoretical and instrumental activity .
the observation of high redshift protostructures before they collapse and star formation triggers will certainly help us to constrain structure formation scenarios .
one possibility of such observations is based on the spectral spatial fluctuations in the cmbr temperature produced by elastic resonant scattering of cmbr photons on molecules located in protostructures moving with peculiar velocity .
this kind of secondary anisotropies search for lih molecule has been fruitless ( de bernardis et al . 1993 ) . as bougleux & galli ( 1997 ) showed , the abundances of lih molecule and their optical depths in the rotational lines are too small to produce detectable cmbr temperature fluctuations . in this paper
we have investigated the spectral spatial fluctuations in the the cmbr temperature due to elastic resonant scattering by hd molecules . for this molecule ,
the main contribution to the optical depth comes from the ground rotational transition ( @xmath189)=(10 ) , because of small populations with high rotational levels of the hd molecule at early epochs . in the case of the lih molecule ,
most of the contribution to the optical depth comes from rotational transitions between high @xmath163 levels .
the wavelenght at rest of the hd ( @xmath189)=(10 ) transition is 112.1 @xmath24 .
therefore , for @xmath212 it should be observed at the wavelengths @xmath242 mm . for higher rotational transitions ,
the wavelengths are smaller .
we have carried out calculations in order to estimate the hd optical depths and the corresponding amplitudes of the spectral spatial fluctuations in the cmbr temperature produced by moving protostructures .
first we calculated detailed chemical kinematic evolution for hd molecule in the expanding and adiabatically cooling homogeneous medium , assuming standard bbn yields .
we further calculated the optical depths in hd pure rotational lines for mass perturbations ( protostructures ) at their maximum expansion under the assumption that the hd relative fractions , @xmath179 , within these protostructures are similar to the fractions calculated for the homogeneous medium at a given @xmath121 .
the spherical top hat approach and the concordance @xmath1cdm cosmology were used to calculate the corresponding turn around redshifts , sizes , and peculiar velocities for different masses and peaks .
once the optical depths for the first three ground rotational transitions were calculated , the corresponding fluctuations @xmath243 were estimated .
we also estimated @xmath244 and @xmath243 for cold gas clouds inside collapsed cdm haloes , but under several assumptions and using results on the hd relative fraction from a crude calculation by galli & palla ( 2002 ) .
the main results from our study are the following : @xmath245 the relative fraction of hd molecule , @xmath246 , increases in the homogeneous expanding medium by a factor of @xmath247 since @xmath248 to low redshifts ( fig .
1 ) . at redshifts
@xmath249 , @xmath250 .
these fractions are expected to strongly increase during the cooling and collapse of the gas within cdm haloes .
@xmath245 the optical depth of the hd first pure rotational line in high peak protostructures at their maximum expansion increases with time . for the second and third lines ,
the corresponding @xmath180 s attain a maximum at high redshifts and then strongly decrease for lower redshifts .
the ranges of @xmath180 values of the ( 10 ) transition lines in 3 and @xmath75 protostructures at their maximum expansion in the @xmath237 redshift interval are approximately @xmath251 and @xmath252 , respectively ( figs . 2 and 3 ) . for the crude model of cold gas clouds inside collapsed 3 and @xmath75 cdm haloes , the ranges of @xmath180 in the same redshift interval are @xmath253 and @xmath254 , respectively ( fig .
the optical depths of hd rotational lines in protostructures are much higher than those of the lih molecule .
@xmath245 the ranges of redshifted frequencies and typical angular sizes in the sky of spectral spatial fluctuations due to hd molecule rotational lines in high peak protostructures and gas clouds in collapsed haloes fall partially within the observational windows of submillimeter telescope facilities under construction , as gtm / lmt , carma and alma ( fig . 5 ) .
the range of redshifts under consideration was @xmath237 ; after this epoch the first baryonic structures are probably already in site with strong emitting stellar sources .
the observational search of the spectral spatial fluctuations studied here will be important for testing models of cosmic structure formation as well as the primordial abundance of deuterium predicted in standard and non - standard bbn theories .
@xmath245 for the @xmath1cdm scenario , the amplitudes of the spectral spatial fluctuations produced by hd molecule at @xmath237 in protostructures , even as rare as @xmath255 density peaks , are too faint for reaching the flux sensitivity of submillimeter / millimeter facilities under construction . for the case of cold gas clouds within collapsed @xmath1cdm haloes emerging from @xmath256 density peaks , the estimated amplitudes , in particular for the lowest rotational transition line , are much larger and could be detected , but their estimated angular sizes in the sky are too small for telescopes such as gtm / lmt , but possibly clusters of such objects could be resolved ( high peak halos are strongly clustered ) .
the spectral spatial fluctuations predicted for very rare high redshift cold clouds in the @xmath1cdm scenario will fall within the observational capabilities of other facilities as carma and alma . for non - standard bbn theories , the [ d / h ]
abundance can be much larger than the used here .
for example , in inhomegeneous bbn models , [ d / h ] is estimated to be ten times larger than in the standard bbn ( e.g. , lara 2005 ) .
therefore the hd line optical depths presented here could be a factor of ten higher , making already detectable the corresponding spectral
spatial fluctuations produced at the maximum expansion of protohaloes . a more detailed study of spectral spatial fluctuations due to hd molecules in collapsing gas clouds as well as the inclusion of alternative bbn models is necessary in order to get more precise results and predictions than the ones presented here . in this paper
we have carried out preliminary calculations that show the viability of using hd molecule lines to search for signatures of early structure formation in the universe and/or to constrain cosmological theories .
we acknowledge the referee , dr .
n. yoshida , for his accurate report and useful comments and suggestions .
we are grateful to j. benda for grammar corrections to the manuscript .
this work was supported by promep grant 12507 0703027 to a.l . and dgapa - unam in107706 - 3 grant to v.a .
a conacyt phd fellowship is acknowledged by r.n - l .
99 abel t. , anninos p. , norman m. , zhang y. , 1998 , apj , 508 , 518 abgrall h. , roueff e. , viala y. , 1982 , a&as , 50 , 505 bardeen j.m . , bond j.r . , kaiser n. , szalay a.s . , 1986 , apj , 304 , 15 barkana r. , loeb a. , 2002 , apj , 578 , 1 barkana r. , loeb a. , 2005 , mnras , 363 , l36 bougleux e. , galli d. , 1997 , mnras , 288 , 638 bromm v. , copi p.s . , larson r.b . , 2002 , apj , 564 , 23 bromm v. , larson r.b . , 2004 , araa , 42 , 79 carroll s.m . , press w.h .
, turner e.l . , 1992 ,
araa , 30 , 499 de bernardis p. et al . , 1993 , a&a , 269 , 1 dubrovich v.k . , 1977 ,
, 128 dubrovich v.k . , 1983 ,
bull . spec .
north caucasus , 13 , 31 dubrovich v.k .
, 1993 , astronomy letters , 19 , 53 dubrovich v.k . , 1997 , a&a , 324 , 27 flower d.r . , le bourlot j. , pineau des forets g. and roueff e. , 2000 , mnras , 314 , 753 fuller t.m . ,
couchman h.m.p .
, 2000 , apj , 544 , 6 galli d. , palla f. , 1998 , a&a , 335 , 403 galli d. , palla f. , 2002 planetary and space science , 50 , issue 12 - 13 , 1197 gao l. , white s.d.m .
, jenkins a. , frenk c.s . , springel , v. , 2005 , mnras , 363 , 379 hogan c.j .
, rees m.j . , 1979 , mnras , 188 , 791 horellou c. , berge j. , 2005 , mnras , 360 , 1393 hudson m.j . , smith r.j , lucey j.r . , schlegel d.j .
, davis r.l . , 1999 ,
apj , 512 , l79 hudson m.j . ,
smith r.j . ,
lucey j.r . ,
branchini e. , 2004 , mnras , 352 , 61 johnson j.l . ,
bromm v. , 2006 , mnras , 366 , 247 kamaya h. , silk j. , 2003 , mnras , 339 , 1256 khlopov m. yu . , rubin s.g . , 2004 , `` cosmological pattern of microphysics in the inflationary universe '' , kluwer academic publishers , dordrecht lara j. f. , 2005 , phys.rev .
d72 , 023509 lahav o. , lilje , p.b . ,
primack j.r .
, rees m.j . , 1991 , mnras , 251 , 128 lauer t.r . , potsman m. , 1994 , apj , 425 , 418 lipovka a. , nez
lpez r. , avila reese v. , 2005 , mnras , 361 , 854 maoli r. , melchiorri f. , tosti d. , 1994 , apj , 425 , 372 maoli r. , ferrucci v. , melchiorri f. , signore m. , tosti d. , 1996 , apj 457 , 1 maoli r. et al , 2004 , esa sp-577 , 293 , preprint ( astro - ph/0411641 ) oh s.p .
, haiman z. , 2003 , mnras , 346 , 456 padmanabhan t. , 1993 , structure formation in the universe . cambridge university press .
puy d. , signore m. , 2002 , aip conf.pro .
, 616 , 346 , astro - ph/0108512 reed d.s . , bower r. , frenk c.s . , gao l. , jenkins a. , theuns t. , white s.d.m . , 2005 , mnras , 363 , 393 shchekinov yu . 1986 , sval , 12 , 211 spergel d.n .
et al . , 2003 , apjss , 148 , 175 stancil p.c . , lepp s. , dalgarno a. , 1996 , apj , 458 , 401 sunyaev r.a . , zeldovich ya.b . , 1970 ,
ap&ss , 7 , 3 tegmark m. , silk j. , rees m.j . , blanchard a. , abel t. , palla f. , 1997 , apj , 474 , 1 varshalovich d.a . ,
ivanchik a.v . , petitjean p. , srianand r. , ledoux c. , 2001 , astronomy letters , 27 , 683 wang l. , steinhardt p.j .
, 1998 , apj , 508 , 483 white s.d.m .
, springel v. , 2000 , in `` the first stars '' , fsproc.mpa/eso , p. 327
( astro - ph/9911378 ) willick j.a . , 1999 ,
apj , 522 , 647 yoshida n. , abel t. , hernquist l. , sugiyama n. , 2003 , apj , 592 , 645 yoshida n. , 2006 , new astronomy reviews , 50 , 19 zaroubi s. , 2002 , preprint ( astro - ph/0206052 ) zeldovich ya.b . , 1978 , sva lett . , 4 , 88 | possible detection of signatures of structure formation at the end of the dark age epoch ( @xmath0 ) is examined .
we discuss the spectral spatial fluctuations in the cmbr temperature produced by elastic resonant scattering of cmbr photons on hd molecules located in protostructures moving with peculiar velocity .
detailed chemical kinematic evolution of hd molecules in the expanding homogeneous medium is calculated .
then , the hd abundances are linked to protostructures at their maximum expansion , whose properties are estimated by using the top hat spherical approach and the @xmath1cdm cosmology .
we find that the optical depths in the hd three lowest pure rotational lines for high peak protohaloes at their maximum expansion are much higher than those in lih molecule .
the corresponding spectral spatial fluctuation amplitudes however are probably too weak as to be detected by current and forthcoming millimeter
telescope facilities .
we extend our estimates of spectral spatial fluctuations to gas clouds inside collapsed cdm haloes by using results from a crude model of hd production in these clouds .
the fluctuations for the highest
peak cdm haloes at redshifts @xmath2 could be detected in the future .
observations will be important to test model predictions of early structure formation in the universe .
cosmology : first stars galaxies : formation molecular processes cosmology : theory dark matter |
determination of dynamical masses of spiral galaxies is a decisive step in establishing empirically the clustering scale of dark matter . the cold dark matter model predicts a hierarchy of scales , from a dwarf sub - halo scale of @xmath5 of solar masses up to a galaxy cluster size .
there is a strong evidence from the @xmath6-ray observations of galaxy clusters , and from detecting dark matter with gravitational lensing @xcite for a @xmath7 scale , which is the galaxy cluster scale . at the galactic size scale ( of the order of @xmath8 ) , there are contradictory results .
dark matter was not observed in some elliptical galaxies @xcite . as for the spiral galaxies ,
it was pointed out that the amount of dark matter in the disk of the milky way galaxy near the sun is insignificant @xcite
. moreover , mass estimates of spiral galaxies are model - dependent @xcite . in the following
we propose an iterative spectral method to reconstruct mass distribution in the spiral galaxies .
the method is based on the thin disk model @xcite which offers the possibility to obtain mass density distribution from the observed rotation curve @xcite . with the iterative spectral method
this reconstruction is unique , provided there is observational data about the hydrogen density in the regions outside of rotation velocity measurements . without such data one can also extrapolate hydrogen measurements by assuming its falloff profile that would be physically acceptable .
one should also take into account the abundance of helium which is proportional to that of hydrogen . in this way
we obtain a complete set of data in the complementary regions of a given galaxy .
complete data are required to eliminate cutoff errors that normally result from cutting - off integration in an integral that relates rotational velocity to disk surface mass density at the radius where measurements of a rotation curve end .
we shall call the radius the cutoff radius and denote it by @xmath9 . in fact , in early days of galactic research
the thin disk model was applied to infer the galactic masses from the measured rotation curves . however , large errors of measurements of rotational velocities practically made the surface mass density reconstruction impossible @xcite . to avoid uncertainties in the thin disk mass reconstruction it became customary to a priori divide spiral galaxies into three major components : bulge , disk and spherical halo .
then a several parameter fit was made determining the properties of those components .
although this approach is physically appealing , the amount of arbitrariness in such a division is unacceptable . in particular , the most massive component of this construction - the spherical halo - stands on the weakest empirical footing
in this paper we address the problem of determination of masses of spiral galaxies with rotation curves violating the sphericity condition at large radii .
one expects that the disk model should approximate mass distribution in such galaxies better than models assuming a massive spherical halo .
the reconstructed surface mass density found for galaxy ngc 4736 approaches the hydrogen density at the outskirts of the galaxy , the rotation curve is perfectly matched in the measured range , and its continuation is predicted further out into the keplerian regime . thus the disk model performs much better in this case than models including the spherical dark halo . to see why the assumption of the existence of spherical halo seems to be unjustified for a class of spiral galaxies , we consider as an example the galaxy ngc 4736 .
it was found ( @xcite ) that most of the mass ( @xmath10 ) in this galaxy is concentrated in a spherical dark halo , mainly at larger radii . at these distances surfaces of constant gravitational potential
are almost spherical .
therefore one would expect in this region the rotation curve predicted by the kent s analysis to satisfy the sphericity condition . for a spherically symmetric mass distribution the keplerian mass function , @xmath11 ,
corresponding to rotational velocity @xmath12 , should be a nondecreasing function of the radial distance from the center @xmath13 , that is , @xmath14 the experimental rotation curve of ngc 4736 violates this condition for @xmath15 , see figure [ fig : f1 ] , whereas the best - fitting solution for the rotation curve found by kent satisfies this condition . at large radii
the best - fitting rotational velocity is also greater from the observed by a factor of @xmath16
. the total mass should be therefore @xmath17 rather than @xmath18 as found in @xcite ( at the currently accepted distance @xmath19 to ngc 4736 the mass is even less , @xmath20 ) .
negative sphericity test for spiral galaxy ngc 4736 .
the figure shows experimental rotation curve ( solid line ) and the corresponding keplerian mass function defined by @xmath21 ( dashed line ) .
the keplerian mass function is not increasing at all distances , therefore a massive spherical halo in the approximation of circular orbits is excluded and should be replaced with an extended flattened component . ] this shows that the mass distribution in the galaxy can not be dominated by a spherical component and that it would be more appropriate to assume more extended disk components in similar galaxies in place of massive spherical halos .
the global disk model has been used recently @xcite , however the procedure of cutting - off integrals applied therein lead to overestimation of surface density at larger radii for galaxies with declining rotation curves .
for example , surface density in the galaxy ngc 4736 could be reliably determined only within a radius of less than @xmath22 .
but the experimental data available for this galaxy enable one to find a much better determination of mass distribution as far as @xmath23 from the center .
an improved approach was presented in @xcite , where rotation curves were assumed keplerian in the unobserved regions .
this procedure is still unsatisfactory , as velocities of test particles on circular orbits encircling a disk - like object need not be keplerian . besides no comparison with the observed mass density has been done .
let @xmath24 denote the last measured value of rotational velocity at cutoff radius @xmath9 .
we define also a dimensionless function @xmath25 for describing rotation curve , where @xmath26 measures distances in units of the cutoff radius ( we work in cylindrical coordinates @xmath27 ) .
below we shall derive equations that are fundamental for the iteration method we present later on .
let the @xmath28 plane represent the galactic disc .
it is assumed that most of galactic mass is concentrated in its vicinity .
a @xmath29-symmetric and cylindrically symmetric vacuum gravitational potential is given by @xmath30 a discontinuity of @xmath31 in the direction normal to the @xmath28 plane is interpreted as a surface mass density @xmath32 of the galactic disk @xmath33 thus @xmath34 is the fourier - bessel transform of @xmath32 @xmath35 modelling of galactic rotation curves usually assumes that orbits of stars are circular and concentric , then @xmath36 , therefore , @xmath37 by substituting ( [ eq : sigmaintegral ] ) to ( [ eq : u2 ] ) one can show that and @xmath38}\xi^{-1 } , & 0<\xi < x,\\ { \left[k{\left(x/\xi\right)}-e{\left(x/\xi\right)}\right]}x^{-1 } , & 0<x<\xi , \end{array}\right.\end{aligned}\ ] ] where the complete elliptic functions are defined by @xmath39 ] @xmath40}{\mathrm{d}\chi}\right ) , \label{eq : vfromsig}\end{aligned}\ ] ] where @xmath41 stands for the `` principal value integral '' .
the simple pole @xmath42 dominates the logarithmic divergence in @xmath43 , thus the principal value integral is easily tractable numerically and can be calculated with high accuracy .
the inverse of the above integral reads ) is deliberately presented in the form containing only squares of rotational velocity .
much more familiar is its equivalent form given in the @xcite handbook , and therefore commonly used , however it has the deficiency of containing derivatives of @xmath44 which are determined experimentally with very low accuracy . ]
@xmath45}{\mathrm{d}\chi}\right),\end{aligned}\ ] ] unfortunately , this nice formula is not very useful for determining @xmath32 directly from @xmath12 , as the observed rotation curves have naturally a cut - off , and therefore the integration can not be carried out .
for example , the closer @xmath13 is to cutoff radius @xmath9 , the more @xmath32 is underestimated by cutting off the integration at @xmath9 . in order to overcome this unpleasant feature of disk geometry one may use another , complementary set of the experimental data concerning mass distribution at @xmath46 .
a global mass density and the corresponding global rotation curve can be found iteratively that would both satisfy the relations ( [ eq : vfromsig ] ) and ( [ eq : sigmamoja ] ) and be consistent with the observations .
below we shall find a first approximation to @xmath32 directly from the observed part of rotation curve . a function @xmath47 defined on the unit interval @xmath48 ,
can be expressed in terms of a series of independent functions .
in particular , if @xmath49 is integrable over the interval , a complete set of orthogonal functions on @xmath48 can be constructed from the bessel function @xmath50 @xcite . then @xmath51 the summation
is taken over all consecutive zeros @xmath52 of the bessel function @xmath53 . by multiplying both sides by @xmath54 and integrating over @xmath48
, we find that @xmath55 the amount of information encoded in the enumerable sequence @xmath56 is insufficient for determining the transform @xmath34 which is a continuum of numbers .
therefore , @xmath32 can not be determined accurately from the observed rotation curve alone , even in the region @xmath57 .
however , the elements of the sequence can be still linearly combined to approximate @xmath34 .
note , that every such approximation is equivalent to choosing a particular extrapolation of a rotation curve beyond its cutoff radius .
there are some mathematical subtleties regarding this extrapolation .
let us assume first that we put @xmath58 for @xmath46 and @xmath59 at a single point @xmath60 .
then we can use the representation ( [ eq : vseries ] ) for @xmath61 , and for @xmath62 we set @xmath63 . from the inverse
transform of ( [ eq : u2 ] ) , which reads @xmath64we then obtain a continuous spectrum when @xmath65 , the components under the summation sign are continuous functions of @xmath66 @xmath67 however , for handling the observational data more tractable are discrete spectra .
for instance , by extending formally ( [ eq : vseries ] ) to the region @xmath62 , we get @xmath68 now , the corresponding @xmath69 , calculated from equation ( [ eq : u2 ] ) , can behave quite odd for @xmath62 and become even negative !
nevertheless , both ( [ eq : contin ] ) and ( [ eq : discretespect ] ) , when substituted to ( [ eq : u2 ] ) , lead to the same @xmath69 for any @xmath48 . although the corresponding surface mass densities differ from each other also for @xmath57 , the difference is , in practice , negligible close to the centre . we may therefore accept ( [ eq : discretespect ] ) as a first approximation of @xmath34 . by substituting ( [ eq : discretespect ] ) to ( [ eq : sig_from_hat_sig ] ) we obtain @xmath70 despite that close to @xmath9 equation ( [ eq : sigmaseries ] ) is not a trustworthy approximation ( the sum may become negative ) , it almost perfectly agrees ( usually for @xmath71 ) with the surface densities corresponding to several model rotation curves that were cutting - off at different @xmath9 . from the above ,
it is evident that the total mass of the galactic disk within the cutoff radius can not be determined from its rotation law only .
however , if additional bounds for the surface mass density are known outside the cutoff radius , one may find a reliable global mass distribution by iterations . as before we denote by @xmath9 the cutoff radius , to which a rotation curve has been measured .
we denote this part of rotation curve by @xmath72 . the distribution of matter , which extends far beyond the cutoff radius and which is mostly composed of neutral and molecular hydrogen , we denote by @xmath73 .
if such data are not available , one may try to devise a physically plausible profile of baryonic matter outside @xmath9 . at the very
beginning we define @xmath74 .
we calculate from ( [ eq : sigmaseries ] ) an initial surface density @xmath75 where the corresponding spectral coefficients @xmath76 are calculated from ( [ eq : expanscoeff ] ) ( we put @xmath75 and @xmath76 in place of @xmath77 and @xmath78 , respectively ) .
let @xmath79 be the first value of @xmath13 where @xmath80 .
the first approximation @xmath81 to the sought global surface density for @xmath82 is defined by setting @xmath83 , and for @xmath84 we set @xmath85 . with this @xmath81
we begin the first iteration step ( @xmath86 ) which is described below for arbitrary @xmath87 .
the @xmath88 iteration step : : by setting @xmath89 in equation ( [ eq : vfromsig ] ) we calculate the corresponding approximation @xmath90 to a global rotational velocity .
afterwards we find the resulting discrepancy function @xmath91 defined for @xmath57 .
the function measures an approximation error in the @xmath88 iteration step and is used also to determine a correction @xmath92 to the actual global surface mass density @xmath93 . by substituting @xmath94 in place of @xmath69 in equation ( [ eq : expanscoeff ] ) we find spectral coefficients @xmath95 where now @xmath78 stands for @xmath96 .
next we calculate for @xmath57 the correction @xmath92 from equation @xmath97 which is analogous to equation ( [ eq : sigmaseries ] ) .
let @xmath98 be the first value of @xmath13 where @xmath99 .
the corrected global surface mass density @xmath100 is found by setting @xmath101 for @xmath102 and @xmath103 for @xmath104 .
now the @xmath105 iteration step follows .
we repeat this procedure until @xmath106 in the @xmath107 iteration becomes acceptably small .
as an example we choose the spiral galaxy ngc 4736 .
its rotation curve excludes the massive spherical halo , therefore the global disk model should better approximate the gravitational field in this galaxy than models assuming such a halo .
there is also the additional information about distribution of hydrogen observed in the remote region of this galaxy , where the rotation curve has not been determined .
this provides the lower bound for the surface density .
no other luminous matter has been observed there .
the amount of hydrogen in this region is comparably small .
its presence allows us to reconstruct the global mass density and predict the rotation curve beyond the presently measured range .
a high resolution rotation curve of the galaxy was adopted from @xcite .
it was measured out to @xmath108 which corresponds to cutoff radius @xmath109 ( the distance of the galaxy is @xmath110 and inclination angle @xmath111 as in @xcite ) .
the rotational velocity has been measured with the maximum error @xmath112 ( attained at @xmath113 ) and the average velocity is @xmath114 .
thus the total dynamical mass should be determinable with the accuracy better than @xmath115 .
we use also the @xmath116 resolution neutral hydrogen surface density published in @xcite ( inclination was corrected from @xmath117 to @xmath118 ) , which is consistent with @xcite . a @xmath119-band surface brightness @xmath120 , where @xmath121 , @xmath122 or @xmath123 is calculated from @xmath124}= \frac{\cos{\left(i\right)}}{\tan^2{\left(\mathrm{arcsec}\right)}}10^{0.4{\left ( m_{\beta,\sun}-\mu_{\beta}-5\right)}}.\ ] ] it is done using the respective luminosity profiles , corrected for galactic extinctions @xmath125}$ ] , taken from @xcite .
the reference absolute magnitudes for the sun are : @xmath126 @xcite , @xmath127 @xcite and @xmath128 @xcite . below we shall apply the iteration scheme described in section [ sec : it ] .
we use the same notational conventions as introduced there .
a third order interpolating function @xmath129 ( @xmath130 @xmath131 ) is used to find the spectral coefficients @xmath76 .
the @xmath132 coefficients in ( [ eq : expanscoeff ] ) fall off quickly to zero , eg .
@xmath133 , and the functional sequence of partial sums in ( [ eq : sigmaseries ] ) converges quickly .
the summation can thus be terminated at some finite @xmath134 ( we put @xmath135 ) .
this truncated sum is taken as the surface mass density @xmath136 .
next , we set @xmath137 for @xmath82 , @xmath138 for @xmath139 and @xmath140 elsewhere , where @xmath141 is the azimuthally averaged neutral hydrogen surface density , @xmath79 was defined in sec .
[ sec : it ] and @xmath142 , cf .
line @xmath81 in figure [ fig : f2 ] . a fragment of surface mass density of galaxy ngc 4736 obtained in consecutive iterations ( solid lines @xmath81 , @xmath143 , @xmath144 , @xmath145 ) and the observed column mass density of h ( @xmath73 solid circles ) . ]
the integrated mass of @xmath146 is @xmath147 and should not exceed the total mass @xmath148 more than @xmath149 , where @xmath150 .
the corresponding first approximation @xmath151 to the global rotation curve agrees with @xmath152 very well for the small radii , where @xmath146 is reliable .
however , there is still a discrepancy at greater radii , cf .
line @xmath153 in figure [ fig : f3 ] . in order to reduce it
, we find a correction @xmath154 to @xmath146 in the region @xmath155 .
the needed spectral coefficients @xmath156 were determined by finding a least square fit to a first order interpolation of @xmath157 , with @xmath158 used as basis functions . )
, the values of @xmath78 determined by the least square method would be equal to those calculated from equation ( [ eq : expanscoeff ] ) ] having found @xmath154 we calculate the corrected surface density in the second approximation @xmath159 , with the reservation we set @xmath160 for @xmath161 and @xmath162 for @xmath163 ( the line @xmath143 in figure [ fig : f2 ] ) .
now , the corresponding global rotation curve @xmath164 agrees with @xmath152 almost perfectly ( the line @xmath165 in figure [ fig : f3 ] ) , @xmath166 are rotational velocities calculated from equation ( [ eq : vfromsig ] ) for global surface densities obtained in the four consecutive steps of iteration .
the thin line @xmath167 was shown also outside the observational range ( experimental points for ngc 4736 are indicated by circles ) . ] but it still needs to be corrected a little in the vicinity of , @xmath9 , since @xmath168 .
the corresponding mass is @xmath169 and is underestimated , but @xmath170 , where @xmath171 .
similarly , we carry out the third step , to find @xmath172 and @xmath173 .
the integrated mass in this step is @xmath174 and @xmath175 .
the error has not changed substantially compared with @xmath176 , therefore we terminate iterations after the fourth step . in this way
we obtain the global surface density @xmath177 shown in figure [ fig : f4 ] . ) .
this mass distribution leads to rotation velocity that agrees perfectly with the observed rotation curve . for comparison , the surface mass density found in @xcite
is also shown ( the thin line @xmath178 ) .
the observed density profile of hydrogen h is marked with solid circles .
luminosity profiles @xmath179 ( at inclination angle @xmath118 ) in the @xmath123-band ( dotted triangles ) , the @xmath122-band ( dotted squares ) and the @xmath0-band ( dotted circles ) were corrected for galactic extinction @xcite .
the @xmath0-band luminosity was corrected also for the internal extinction ( marked with stars).[fig : f4 ] ] it gives the total mass @xmath180 ( it includes the observed hydrogen h out to @xmath181 ) , and the corresponding rotation curve @xmath182 its predicted continuation beyond the cutoff radius can be tested with future observations ( see figure [ fig : f5 ] ) .
the observed rotation curve of galaxy ngc 4736 ( circles ) and the rotation curve ( solid thick line @xmath167 ) corresponding to surface mass density @xmath145 found in the fourth iteration and calculated from equation ( [ eq : vfromsig ] ) .
the solid thin line @xmath183 is the corresponding keplerian asymptote @xmath184 where @xmath185 is the total galaxy mass .
the solid thin line @xmath186 is the rotational velocity which one would obtain from a spherically symmetric mass distribution with the same mass function as that for ngc 4736 . ] in the previous section we have obtained the surface density in galaxy ngc 4736 that is consistent with distribution of h beyond the cutoff radius @xmath9 , and perfectly agrees with the observed rotation curve . since h mass outside @xmath9 is only @xmath187 , the results ( e.g density profile ) would not be affected significantly if there was an amount of baryonic matter outside @xmath9 comparable to that of h. the global surface mass density of galaxy ngc 4736 found in the fourth iteration is shown in figure [ fig : f4 ] ( the @xmath145 line ) .
the corresponding global rotation curve calculated from equation ( [ eq : vfromsig ] ) agrees perfectly with the experimental rotation curve @xmath72 , see figure [ fig : f5 ] ) .
the @xmath122 and @xmath123-band luminosities are smaller than the @xmath0 band luminosity @xmath188 , cf .
table [ tab:1 ] ( @xmath189 and @xmath190 ) , .[tab:1 ] physical parameters obtained with the help of the iterative spectral method for spiral galaxy ngc 4736 .
( at the distance @xmath191 $ ] and inclination angle @xmath118 ) . [ cols="^,^ " , ] which may indicate the presence of dust .
the mass - to - light ratio profile ( surface density to surface luminosity ratio ) is shown in figure [ fig : f6 ] .
the predicted ratios of mass luminosity surface densities for galaxy ngc 4736 with luminosities corrected for galactic extinction in @xmath123-band ( triangles ) , @xmath122-band ( squares ) , @xmath0-band ( circles ) and corrected additionally for internal extinction @xmath0-band ( marked with stars ) . ] from the three bands , the @xmath0-band should be regarded as the most reliable tracer of the luminous matter , as it most accurately maps the amount of the radiated energy .
luminosity in @xmath122- and @xmath123-bands is reduced by the presence of dust , it also does not include the radiation from low mass stars . for galaxy ngc 4736
the @xmath0-band mass - to - light ratio is small : @xmath192 .
also its distribution , i.e. density to luminosity ratio which is of the order of @xmath193 @xmath194 ( cf .
figure [ fig : f6 ] ) is typical to stars of slightly sub - solar mass .
the total galaxy mass is surprisingly small .
it is comparable to the mass predicted in the framework of the mond model , which gives @xmath195 , or in the framework of the metric skew tensor gravity , which gives @xmath196 @xcite .
we obtain @xmath180 for the same rotation curve in the framework of newtonian physics . in figure [
fig : f7 ] is shown the mass function @xmath197 it is shown together with the keplerian mass function @xmath198 , corresponding to the observed rotation curve @xmath152 .
mass function for ngc 4736 normalized with respect to total mass @xmath180 ( thick solid line ) compared with keplerian mass function @xmath199 ( @xmath200 the thin solid line ) defined as @xmath201 , where @xmath152 is the observed rotation curve .
] the odd behaviour of the latter function was our argument against applying dark halo dominated models to galaxy ngc 4736 .
such a model used in @xcite predicted @xmath3 for the total galaxy mass at a distance @xmath19 .
the best fitting curve in that model gives rotational velocity at the cutoff radius @xmath202 times greater than the observed value .
as this leads to overestimation of the total mass by a factor of some @xmath203 , one should expect rather @xmath204 for the total mass , which is very close to our result obtained in the global disk approximation .
we note also a related and trivial observation : keplerian rotation law @xmath205 for a spherically symmetric matter distribution with the mass function numerically equal to @xmath206 , gives smaller rotation velocities at larger radii than the disk model .
also more matter would be required by a spherical model to account for observations , see figure [ fig : f5 ] .
the comparison of mass functions and rotation laws at the end of the previous section , illustrates the fact that the models with flattened mass distributions are more efficient than the commonly used models assuming spherical halo .
the former are better in accounting both for high rotational velocities as well as for low scale structure of rotation curves and with noticeably less amount of matter than the latter ( the relation between rotation and mass distribution in the disk model is very sensitive for gradients of a rotation curve ) .
the use of the disk model is justified for galaxies with rotation curves violating the sphericity condition .
this is necessary ( although not sufficient ) condition for a spherical mass distribution .
rotates with velocity @xmath207 .
this rotation low satisfies the sphericity condition : @xmath208 . ]
rotation of the spiral galaxy ngc 4736 can be fully understood in the framework of newtonian physics .
we have found a mass distribution in the galaxy that agrees perfectly with its high - resolution rotation curve , agrees with the @xmath0-band luminosity distribution giving low mass - to - light ratio of @xmath202 in this band at total mass of @xmath2 , and is consistent with the amount of h observed in the remote parts of the galaxy , leaving not much room ( if any ) for dark matter .
remarkably , we have achieved this consistency without invoking the hypothesis of a massive dark halo nor using modified gravities .
there exist a class of spiral galaxies , similar to ngc 4736 , that are not dominated by spherical mass distribution at larger radii .
most importantly , in this region rotation curves should be reconstructed accurately in order not to overestimate the mass distribution . for a given rotation curve it can be easily determined whether or not a spherical halo may be allowed at large radii by examining the keplerian mass function corresponding to the rotation curve ( the so called sphericity test ) . by using complementary information of mass distribution , independent of rotation curve
, we overcame the cutoff problem for the disk model , that for a given rotation curve , a mass distribution could not be found uniquely as it was dependent on the arbitrary extrapolation of the rotation curve .
55 binney , j. , and tremaine , s. 1987 , galactic dynamics ( princeton univ . press ) dutton , a. a. , van den bosch , f. c. , dekel , a. , and courteau , s. 2006 , preprint ( astro - ph/0604553v1 ) bosma , a. , van der hulst , j. m. , and sullivan iii , w. t. 1977 , , 57 , 373 brownstein , j. r. , and moffat , j. w. 2006 , , 636 , 721 carignan , c. 1985 , , 58 , 107 kent , s. m. 1987 , , 93 , 816 kostov , v. 2006 , preprint ( astro - ph/0604395v1 ) kuijken , k. , and gilmore , g. , 1989 mnras , 239 , 605 , 651 lebiediev n. n. , 1953 , special functions and their applications ( moscow ) massey , r. 2007 , nature , 445 , 286 mulder , p. s. , and van driel , w. , 1993 a&a , 272 , 63 munoz - tunon , c. , prieto , m. , beckman , j. , and cepa , j. , 1989 , 156 , 3001 nordsieck , k. h. 1973 , , 184 , 735 romanowsky , a. j. 2003 , science , 301 , 1696 ryzhik , i. m. , and gradshtein , i. s. 1951 , tables of integrals , sums , series and products ( moscow , leningrad ) sofue , y. , and rubin , v. 2001 39 , 137 sofue , y. , tutui , y. , honma , m. , tomita , a. , takamiya , t. , koda , j. , and takeda , y. 1999 , , 523 , 136 .
takamiya , t. , and sofue , y. 2000 , , 534 , 670 toomre , a. 1963 , , 138 , 385 | an iterative method for reconstructing mass distribution in spiral galaxies using a thin disk approximation is developed .
as an example , the method is applied to galaxy ngc 4736 ; its rotation curve does not allow one to employ a model with a massive spherical halo .
we find a global mass distribution in this galaxy ( without non - baryonic dark matter ) that agrees perfectly with the high resolution rotation curve of the galaxy .
this mass distribution is consistent with the @xmath0-band luminosity profile with the mean mass - to - light ratio @xmath1 , and also agrees with the amount of hydrogen observed in the outermost regions of the galaxy .
we predict the total mass of the galaxy to be only @xmath2 .
it is very close to the value predicted by the modified gravity models and much less than the currently accepted value of @xmath3 ( with @xmath4 of the mass in a dark matter halo ) . |
liquid polyamorphism , i.e. , the existence of distinct liquid phases of a pure substance , was proposed over 40 years ago @xcite to explain the occurrence of unusual melting properties of certain substances , such as melting - curve maxima ( as in p , cs , ba , etc ) and negative - sloped melting lines ( as in water , si , ge , ga , etc ) .
@xcite this so - called two - state model predicts the existence of a first - order phase transition between different liquid forms of the same substance , commonly referred to as the high - density liquid ( hdl ) and low - density liquid ( ldl ) , with a coexistence line that ends at a second critical point . although it is still an intensely debated issue,@xcite there are indications for the existence of such llpts in a number of substances . aside from experimental observations of llpts in elemental phosphorus @xcite , @xmath0 , @xcite the molecular liquids triphenyl phosphite @xcite and n - butane , @xcite and the compound ass .
@xcite other experimental data and computer simulations suggest the occurrence of llpts in other substances such as water , @xcite silicon , @xcite gallium @xcite and nitrogen .
@xcite in addition to studies considering the structural characteristics associated with the transition between the hdl and ldl forms , there has been an increasing interest in the liquid dynamics across llpts .
@xcite in the case of silicon , sastry and angell @xcite found that , across an interval of 15 k through the llpt , the intermediate scattering function develops a pronounced plateau characteristic of the so - called cage effect and the self - diffusion coefficient decreases by two orders of magnitude .
a similar result was found in the case of gallium .
@xcite for silicon , the development of the pronounced caging plateau for ldl is accompanied by a pronounced dip in the intermediate scattering function just before the plateau .
@xcite this dip has been linked to the appearance of an excess of low - frequency vibrational modes compared to the typical debye model vibrational density of states ( vdos ) of a crystal .
this excess has been associated with a boson peak ( bp ) , @xcite although this is still rather controversial .
@xcite this observation , in turn , has led to the suggestion of a connection between the structural llpt in si and an accompanying transition in the dynamics between a fragile and strong liquid .
strong liquids are those whose transport properties , e.g. , the shear viscosity , display an arrhenius behavior as a function of temperature , whereas fragile liquids are characterized by a manifestly non - arrhenius behavior .
empirically , the presence of a pronounced bp has been viewed as a signature of a liquid s strong character , @xcite such as in the case of @xmath1 .
@xcite in line with this perspective , the sudden appearance of a dip in the intermediate scattering function across the hdl - to - ldl transition in silicon , both for the stillinger - weber ( sw ) model @xcite as well as in _ ab initio _ calculations , has been interpreted in terms of a fragile - to - strong transition ( ftst ) .
@xcite this interpretation remains controversial , however , @xcite given that there is still no physical basis directly correlating the presence of a possible bp to liquid fragility .
indeed , as the fragility of a liquid is defined in terms of the temperature - dependence of transport properties , the verdict as to whether the appearance of an excess in the vdos across the llpt is accompanied by a ftst should be based on the analysis of transport properties such as the shear viscosity , which often appears in the so - called angell plot @xcite distinguishing between strong and fragile behavior . here
we conduct a further investigation into the possible correlation between a change in dynamics and the occurrence of a llpt . for this purpose
we analyze the dynamics of liquid silicon and gallium as described by the sw potential and modified embedded - atom model ( meam ) , @xcite respectively , both of which exhibit well - characterized llpts . for both systems
we compute the intermediate scattering functions , the vibrational densities of states as well as the temperature - dependence of the shear viscosities .
the remainder of the paper is organized as follows .
[ sec1 ] provides the technical aspects of the computer simulations used in this study .
the results are presented and discussed in sec .
[ sec2 ] . finally , in sec . [ sec3 ]
we summarize the main results of our work .
the computer simulations performed in this work were carried out using the molecular dynamics method ( md ) as implemented in the lammps package @xcite .
isobaric - isothermal ensemble ( npt ) simulations were employed to obtain the hdl and ldl phases , whereas canonical ensemble ( nvt ) simulations were used to compute correlation functions .
the integration of the nos - hoover equations of motion was carried out using a time step of 1 fs , and temperature and pressure damping parameters of 70 fs and 1 ps , respectively , were employed .
the interatomic interactions in si were modeled using the stillinger - weber ( sw ) @xcite potential , while ga was described using the modified embedded atom method ( meam ) proposed by baskes , chen , and cherne .
@xcite the sw potential is known for providing an accurate description of liquid si and its melting point temperature and it has been extensively used in the investigation of the llpt in si .
@xcite although the meam potential for ga overestimates the quantitative value of the melting point temperature at ambient pressure , it correctly reproduces essential qualitative features of the phase diagram such as the negative slope of the melting line of the @xmath2-ga phase , the positive slope of melting line of the gaii phase and the correct nine - fold coordination of the stable liquid . moreover , the meam for ga predicted a previously unknown crystalline metastable phase of ga , which was confirmed by @xmath3 calculations .
@xcite all calculations were based on simulation cells subject to periodic boundary conditions . in the cases of gallium and silicon ,
supercells containing 1152 and 1000 atoms were employed , respectively .
larger cells were used to evaluate finite - size effects on the dip in the intermediate scattering function in si .
starting from the well - equilibrated liquid above the melting point temperature , the hdl and the ldl phases were obtained according to the following protocols . in the case of si
, the liquid was quenched starting at 1700 k at a cooling rate of 50 k / ns .
for ga , the quench started at 450 k at a cooling rate of 20 k / ns . in both cases
the quenches were carried out at zero pressure , which is the same as that used in previous md studies of these systems .
@xcite the llpt takes place around 1060 k in the case of si , @xcite while for ga it occurs at a temperature of 356 k. @xcite the canonical - ensemble simulations of the hdl and ldl forms of si were carried out at 1070 k and 1050 k , respectively . for the case of ga the hdl and ldl forms were considered at 362 k and 350 k , respectively .
as mentioned earlier , particular features in the intermediate scattering function , which describes the decay of density fluctuations on a given length scale,@xcite have been interpreted as evidence for the occurrence of ftst in the case of the llpt in si . here
we compute the self- ( or incoherent ) intermediate scattering function , @xmath4 , of si and ga . for a system containing @xmath5 particles ,
this function is defined as @xmath6}\rangle , \\ & = & \frac{1}{n } \sum_{i=1}^n f_{s , i}(\boldsymbol{k},t),\end{aligned}\ ] ] where @xmath7 stands for the position of particle @xmath8 at time @xmath9 and @xmath10 is the wave vector associated with the length scale of interest .
the magnitude of @xmath10 is chosen to be @xmath11 , where @xmath12 corresponds to the position of the first peak in the radial distribution function . in the calculations for si , we used the following values of @xmath12 : 2.39 , 2.43 , 2.46 , and 2.52 , for the temperatures , 1050 k ( ldl ) , 1070 k ( hdl ) , 1200 k , and 1700 k , respectively . for ga , the distances , 2.68 ,
2.66 , 2.68 , and 2.68 , corresponding to the temperatures 350 k ( ldl ) , 362 k ( hdl ) , 400 k , and 450 k , respectively , were employed .
all results reported below correspond to averages @xmath13 of @xmath4 over three mutually perpendicular @xmath10-directions for a given magnitude @xmath14 . fig .
[ fig1 ] presents the results for @xmath13 of si and ga for different temperatures above and below the llpt .
[ fig1 ] displays the characteristic dip just at beginning of the plateau of @xmath13 for ldl - si , in agreement with previous studies .
@xcite yet , as pointed out by horbach _ et al . _
@xcite , it is possible that the dip in the intermediate scattering function is an artifact of finite size effects . to verify this possibility
we compute @xmath13 for ldl - si as a function of increasing cell sizes .
the results are reported in fig .
[ fig2 ] and clearly display convergence for cell sizes above 2700 atoms , indicating that the appearance of the dip is indeed a robust effect . returning to fig .
[ fig1 ] , in contrast to the case of si , the llpt in ga is not accompanied by the appearance of a dip in @xmath13 . according to the discussion in ref .
this difference between the relaxation behaviors of ldl - si and ldl - ga , would indicate the presence of an excess in the vibrational density of states at lower frequencies with respect to the debye model in the former and its absence in the latter .
this would suggest significant differences between the llpts in si and ga from a dynamical point of view . to further investigate this issue
, we compute the vibrational density of states ( vdos ) of both ldl liquids , as determined by the fourier transform of the normalized velocity autocorrelation function @xmath15 where @xmath16 is the velocity of the j - th particle at time @xmath9 .
given that the vdos described by the debye model increases @xmath17 , the reduced vdos , defined as , @xmath18 ( rvdos ) decays monotonically with increasing angular frequency .
therefore , an excess of vibrational modes at lower frequencies should appear in the form of peaks in the rvdos . fig .
[ fig3 ] depicts the results of the rvdos and vdos for liquid si at the same four temperatures shown in fig.[fig1 ] .
it is evident that for 1700 k , 1200 k , and 1070 k ( hdl ) , the rvdos decreases monotonically with increasing @xmath19 . for the ldl liquid at 1050 k ,
on the other hand , it displays two distinct peaks at 7.5 thz and 11.0 thz , and a diffuse peak around 17.5 thz .
this indicates that the llpt in the case of si , gives rise to an excess of low frequency modes , which could be associated with the bp .
these results are in good agreement with those obtained by jakse and pasturel using _ ab initio _ simulations . @xcite fig .
[ fig4 ] shows the results for the rvdos and vdos for liquid ga at the same four temperatures considered in fig[fig1 ] .
similar to the case of si , for the temperatures 450 k , 400 k , and 362 k ( hdl ) , the rvdos decreases monotonically as @xmath19 increases . in addition , as for the ldl form of si , the rvdos of ldl - ga at 350 k also exhibits two peaks between 2.5 thz and 4.0 thz , and a broad peak around 7.0 thz .
this result is unexpected , however , if one assumes that an excess of vibrational modes at lower frequencies manifests itself as a dip at the beginning of the plateau in the intermediate scattering function @xmath13 , as has been suggested by horbach and co - workers .
@xcite in this view , the absence of a dip in @xmath13 for ldl - ga in fig .
[ fig1 ] should indicate the absence of an excess of vibrational modes at lower frequencies .
[ fig4 ] shows that this is not the case for ga , inconsistent with the proposed correlation between a dip in @xmath13 and an excess of vibrational modes at low frequencies .
indeed , the identification of peaks in the rvdos as evidence of the bp in the liquid is somewhat controversial . in the case of si , for instance
, there is experimental evidence that the bp is absent in amorphous si , @xcite which appears at odds with the notion of the existence of such a peak in the precursor supercooled liquid phase . finally , we examine the suggested relationship between the appearance of dip in @xmath13 traversing through a llpt and a fragile - to - strong transition . for this purpose
we compute the shear viscosity @xmath20 for both liquids as a function of temperature .
the considered temperature ranges include conditions in which the liquids are thermodynamically stable , supercooled and below the llpt temperature .
the shear viscosity can be determined using a green - kubo - type relation involving the autocorrelation function of the off - diagonal component of the stress tensor , @xcite @xmath21 where @xmath22 is the volume of the system , @xmath23 is the temperature , @xmath24 is the boltzmann constant , and @xmath25 is an off - diagonal component of the stress tensor . aside from the off - diagonal components
@xmath26 , @xmath27 , and @xmath28 , there are , due to rotational invariance , two other independent components @xmath29 and @xmath30 .
@xcite in the calculations , @xmath20 is estimated using an average over these five independent components . figs . [ fig5 ] and [ fig6 ] show results for the average autocorrelation functions @xmath31 and the corresponding cumulative integrals of eq .
( [ eta ] ) for the hdl and ldl forms in si and ga , respectively .
both graphs display the convergence of the computed shear viscosities to plateau values and clearly show the dramatic scale changes occuring across the llpts .
the temperature - dependent results for liquid si are depicted in fig.[fig7 ] , showing the logarithm of the shear viscosity as a function of the inverse temperature .
starting at 3000 k , the obtained shear viscosity values , which are in agreement with experimental data,@xcite increase steadily with decreasing temperature . the increase is also manifestly nonlinear , with @xmath20 increasing more rapidly as the llpt approaches .
this point is further highlighted by the two curves obtained by fitting the temperature - dependence of the shear viscosity below the llpt to the mode - coupling - theory ( mct ) expression @xcite @xmath32 where @xmath33 is the critical temperature at which the viscosity becomes singular , and the phenomenological vogel - fulcher - tammann ( vft ) @xcite model @xmath34 where the vogel temperature @xmath35 describes the temperature at which the viscosity diverges . in the case of si
the least - squares regressions give @xmath36 k and @xmath37 k , respectively , which affirm the rather strong non - arrhenius behavior .
this is consistent with the picture that si is a fragile liquid for temperatures above llpt . upon passing through the first - order llpt
, @xmath20 undergoes a discontinuous increase , leading to a shear viscosity of that is two orders of magnitude larger than that of hdl - si .
this result is consistent with the calculations of sastry and angell @xcite that show an increase of two orders of magnitude of the diffusion coefficient upon transforming from the hdl to the ldl liquid forms .
[ fig8 ] displays the same results for the case of ga , including the curves obtained by fitting the data to the mct and vft expressions .
the obtained values are consistent with available experimental data .
@xcite once more , ga displays behavior that is similar to the case of si . above the llpt
the shear viscosity is manifestly non - arrhenius , with critical temperature values of @xmath38 k and @xmath39 k , respectively .
this implies that , as in the case of si , ga may be considered to be a fragile liquid for temperatures above the llpt .
furthermore , the occurrence of the llpt is also accompanied by an abrupt change of the viscosity , increasing about 2 orders of magnitude upon transforming from the hdl to the ldl . to verify whether the abrupt increase of the shear viscosities of both ldl forms is indeed accompanied by a character change from fragile to strong
, one should analyze the temperature dependences of both ldl shear viscosities .
unfortunately , such an explicit assessment is prohibitively costly in this case due to the extreme sluggishness of the dynamics of both ldl liquids on the time scale of the md simulations
. therefore we adopt an indirect approach , comparing our ldl viscosity data for a single temperature ( i.e. , 1050 k and 350 k for si and ga , respectively ) to those of typical fragile and strong liquids by plotting them in angell s plot .
@xcite this plot depicts the logarithm of the viscosity as function of the inverse of temperature scaled by the material s glass - transition temperature @xmath40 and shows two distinct branches for strong and fragile liquids , respectively , in an essentially universal manner .
@xcite this is shown in fig .
[ fig9 ] , which displays a number of experimental viscosity data sets extracted from ref . for various typical strong and fragile liquids , plotted as a function of @xmath41 where @xmath42 is the experimental glass - transition temperature @xmath43 . in principle , to plot the simulation results for si and ga on the same graph we need to determine the glass transition temperatures for the hdl and ldl phases of both substances . unfortunately , however , an explicit determination of these @xmath43-values is unfeasible . for the hdl form it is obscured by the occurrence of the llpt , whereas for the ldl it is inaccessible due to the extreme sluggishness of the dynamics .
for this reason we need to infer the @xmath43 s by some indirect route . for the hdl phase we can establish a lower
bound in terms of the kauzmann temperature @xmath44 at which the excess entropy of the liquid with respect to the crystal vanishes .
@xcite below @xmath45 , the entropy of the liquid would be lower than that of the crystal , leading to the so - called entropy crisis , which is then avoided by the occurrence of the glass transition .
in this view , @xmath45 can be regarded as a lower bound for @xmath40 .
in addition , it is well known that , for many substances , @xcite the singularity temperature @xmath35 in the vft equation is very close to @xmath44 .
we therefore use the divergence temperature @xmath46 of the vft model determined from the hdl data in figs .
[ fig9 ] and [ fig8 ] as estimates for @xmath43 of the hdl phases in si and ga .
[ fig9 ] shows the corresponding results for the shear viscosities of hdl si and ldl ga as a function of @xmath47 .
the results clearly show that both hdl phases of si and ga appear on the fragile branch of angell s plot together with other fragile liquids such as toluene and o - terphenyl .
indeed , it is quite striking that the results for hdl si and ga , despite their very different structural and cohesive properties , almost superpose on the plot .
an accurate placement of the ldl data on the plot is more challenging . due to the extreme sluggishness of the dynamics below the llpt it has proved prohibitive to obtain temperature - dependent data for the shear viscosity . for this reason we plot the ldl data using the divergence temperature @xmath46 obtained for the corresponding hdl phases to scale the inverse temperature . with this choice , despite the 2-order of magnitude increase of the shear viscosities across the llpts , the data points clearly remain in the fragile branch of the angell plot .
this suggests that , for both si and ga , the llpt does not constitute a ftst .
evidently , this conclusion is based on the particular placement of the ldl results using the @xmath46 values obtained for the hdl phases as an indirect estimate for @xmath43 of the ldl forms
. however , it seems plausible to assume that the hdl vft divergence temperatures not only provide lower bounds for @xmath43 of the hdls but also for the glass transition temperatures of the ldl phases .
this is the case because the ldls are much more viscous than their hdl counterparts , even for temperatures that are essentially the same .
thus it seems reasonable expect that the ldls vitrify at higher temperatures than their hdl counterparts . in this line of reasoning
, the representation of the ldl results in terms of the actual glass transition temperatures @xmath43 instead of @xmath46 would lead to a shift to the right in fig .
accordingly , the availability of the correct value of @xmath43 would still lead to the same conclusion reached based on the use of the vft divergence temperatures @xmath46 : despite the 2-order of magnitude increase of the shear viscosities across the llpts , the ldl phases still appear to belong to the fragile branch of angell s plot .
these findings are inconsistent with the occurrence of a ftst associated with the llpts in both si and ga .
in fact , it appears that the llpts are accompanied by a transition from a fragile liquid to a less fragile liquid , regardless of the appearance or not of a dip in the intermediate scattering function .
this seems at odds with the suggestion that the appearance of a dip in the intermediate scattering function through the llpt signals a change of character from fragile to strong .
we have performed a series of md simulations to study the dynamic properties of supercooled liquid si and ga with the goal of investigating correlations between liquid dynamics and the presence of a structural llpt .
the results indicate that , consistent with previous calculations , the intermediate scattering function of the ldl phase of si at 1050 k exhibits a dip at the beginning of the plateau characteristic of the @xmath48-relaxation process which is absent for the hdl phase at 1070 k. for the case of the llpt in ga , on the other hand , the transition between the hdl to ldl forms is not accompanied by the development of a dip in the intermediate scattering function .
previous work @xcite has suggested that the appearance of a dip in @xmath13 across the llpt is a manifestation of the development of an excess of vibrational states at low frequencies . in the context of our results for @xmath13
this would imply a distinct difference between the dynamic behaviors of ldl si and ga , with the former characterized by an excess of low - frequency modes that should be missing for the latter .
our rvdos results are at odds with this picture , however . despite the absence of a dip in @xmath13 for ldl - ga , its rvdos
does display peaks characteristic of an excess of low - frequency modes .
in addition , we also consider the purported connection between the development of the dip in @xmath13 and a transition in the character of liquid dynamics from fragile to strong . within this view
, the results obtained for @xmath13 should point at a ftst accompanying the hdl - ldl structural transition for the case of si and the absence of such a changeover in the case of ga . although the @xmath43 values for the ldl forms were estimated in an indirect manner , our results for the temperature dependence of the shear viscosities for both substances do not corroborate this prediction , indicating a view in which the hdl as well as ldl phases of both substances are fragile in character .
in fact , it appears that the structural llpt is accompanied by a transition from a fragile liquid to a less fragile liquid , regardless of the appearance or not of a dip in the intermediate scattering function .
we gratefully acknowledge support from the brazilian agencies cnpq , fapesp and capes .
part of the calculations were performed at ccjdr - ifgw - unicamp and cenapad - sp .
62ifxundefined [ 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 10.1063/1.1841150 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/19/41/415101 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.103.225702 [ * * , ( ) ] http://www.sciencedirect.com/science/article/pii/s0022309310005272 [ * * , ( ) ] http://dx.doi.org/10.1039/c3cp51209f [ * * , ( ) ] link:\doibase 10.1063/1.3643333 [ * * , ( ) ] link:\doibase 10.1063/1.4769126 [ * * , ( ) ] http://dx.doi.org/10.1038/35003143 [ * * , ( ) ] http://dx.doi.org/10.1038/369633a0 [ * * , ( ) ] link:\doibase 10.1126/science.1160766 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.92.025701 [ * * , ( ) ] link:\doibase 10.1126/science.1103073 [ * * , ( ) ] http://stacks.iop.org/0953-8984/17/i=27/a=l01 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.100.145701 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.102.115901 [ * * , ( ) ] http://dx.doi.org/10.1038/360324a0 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physreve.55.727 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.95.117802 [ * * , ( ) ] link:\doibase 10.1140/epjst / e2008 - 00747 - 2 [ * * , ( ) ] link:\doibase 10.1073/pnas.1100238108 [ * * , ( ) ] link:\doibase 10.1038/nmat994 [ * * , ( ) ] link:\doibase 10.1063/1.1755653 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.205702 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.075701 [ * * , ( ) ] http://dx.doi.org/10.1063/1.3663387 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevb.74.024116 [ * * , ( ) ] link:\doibase 10.1063/1.3154424 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.102.015701 [ * * , ( ) ] link:\doibase 10.1021/jp204889 m [ * * , ( ) ] http://dx.doi.org/10.1063/1.3684550 [ * * , ( ) ] http://dx.doi.org/10.1063/1.3692610 [ * * , ( ) ] http://stacks.iop.org/0295-5075/11/i=1/a=008 [ * * , ( ) ] http://stacks.iop.org/0034-4885/65/i=8/a=203 [ * * , ( ) ] http://books.google.com.br/books?id=leowizefsp4c[__ ] ( , ) http://dx.doi.org/10.1063/1.3154368 [ * * , ( ) ] http://dx.doi.org/10.1063/1.3151970 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.71.2062 [ * * , ( ) ] link:\doibase 10.1016/0009 - 2541(95)00169 - 7 [ * * , ( ) ] , http://link.aps.org/doi/10.1103/physrevlett.78.2405 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physreve.71.061501 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physreve.54.r5897 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevb.31.5262 [ * * , ( ) ] http://dx.doi.org/10.1063/1.2970084 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevb.62.3728 [ * * , ( ) ] link:\doibase 10.1016/0022 - 3093(88)90133 - 0 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevb.66.104107 [ * * , ( ) ] link:\doibase 10.1006/jcph.1995.1039 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1063/1.1631388 [ * * , ( ) ] http://dx.doi.org/10.1038/nphys1993 [ * * , ( ) ]
http://link.aps.org/doi/10.1103/physrevb.80.045209 [ * * , ( ) ] http://books.google.com.br/books?id=uhm87wzbnxec[__ ]
( , ) link:\doibase 10.1021/jp205090s [ * * , ( ) ] http://books.google.com.br/books?id=o32vxb9e5p4c[__ ] , oxford science publications ( , ) http://link.aps.org/doi/10.1103/physrevlett.81.5161 [ * * , ( ) ] link:\doibase 10.1016/s0022 - 0248(99)00437 - 6 [ * * , ( ) ] http://books.google.com.br/books?id=5acvaqaaiaaj[__ ] , edited by , , and , ( , ) p. @noop * * , ( ) link:\doibase 10.1111/j.1151 - 2916.1925.tb16731.x [ * * , ( ) ] @noop * * ( ) http://stacks.iop.org/0959-5309/48/i=2/a=308 [ * * , ( ) ] link:\doibase 10.6028/jres.102.013 [ * * , ( ) ] | using molecular dynamics simulations we analyze the dynamics of two atomic liquids that display a liquid - liquid phase transition ( llpt ) : si described by the stillinger - weber potential and ga as modeled by the modified embedded - atom model ( meam ) .
in particular , our objective is to investigate the extent to which the presence of a dip in the self - intermediate scattering function is a manifestation of an excess of vibrational states at low frequencies and may be associated with a fragile - to - strong transition ( ftst ) across the llpt , as suggested recently .
our results do not lend support to these suggestions . specifically , in the case of ga we observe the appearance of an excess of vibrational states at low frequencies , even in the absence of the appearance of a dip in the self - intermediate scattering function across the llpt .
furthermore , studying the behavior of the shear viscosities traversing the llpts we find that , despite the development of a dip in the self - intermediate scattering function for the case of si and its absence in ga , both substances are fragile in character above and below their respective llpt temperatures . |
in this paper we consider the heating of the solar corona in regions where the large - scale magnetic field deviates significantly from a ( current - free ) potential field .
evidence for such nonpotential structures comes from a variety of sources .
measurements of photospheric vector fields in active regions often show deviations from the potential field .
x - ray observations show _ sigmoids _ , which are bright s - shaped or inverse s - shaped structures inside or between active regions .
these sigmoids contain coronal loops that are highly sheared with respect to the photospheric polarity inversion line ( pil ) , and are are often associated with h@xmath0 filaments .
@xcite proposed a differential magnetic shear model in which the degree of shear decreases with height in the observed active region .
@xcite studied the nonpotentiality of active regions using extreme - ultraviolet ( euv ) images from the _ transition region and coronal explorer _ ( trace ) .
they found that significant nonpotentiality occurs when new magnetic flux has recently emerged into the corona , or when rapidly evolving , opposite polarity flux concentrations are in close contact .
@xcite tested various models of coronal heating using potential , linear force - free , and magnetostatic models .
of course , nonpotential magnetic fields also play an important role in solar flares and coronal mass ejections .
the nonpotential structures in the corona have been modeled as _ flux ropes _ in which the field lines twist about an axial field line ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the presence of a coronal flux rope implies the existence of electric currents that pass through the photosphere and into the corona .
the corona of an active region is a magnetically dominated plasma .
therefore , except during a flare or eruption , the corona is nearly force free , @xmath1 , where @xmath2 is the magnetic field and @xmath3 is the so - called torsion parameter . a magnetic field with this property is called a nonlinear force - free field ( nlfff ) .
many authors have developed nlfff models of active regions , either by extrapolating photospheric vector fields into the corona , or using a flux rope insertion method .
active regions exhibit transient brightenings at euv and soft x - ray wavelengths on time scales of a few minutes .
these single- or multi - loop brightenings are thought to be due to the release of magnetic energy by small reconnection events in the complex coronal field of an active region .
this may be a two - step process : magnetic energy may first be converted into kinetic energy of reconnection outflows , waves or energetic particles , and then be converted into thermal energy of the plasma @xcite .
it seems likely that x - ray sigmoids are also heated by such reconnection events , with many events occuring simultaneously at different sites within the sigmoid .
the events may occur so frequently that the coronal plasma does not have time to cool significantly between heating events , keeping the plasma at temperatures of 3 - 5 mk ( e.g. , * ? ? ?
* ; * ? ? ?
this would explain why sigmoids generally do not show up in euv spectral lines formed at 1 - 1.5 mk .
what is the origin of the energy released in small reconnection events ?
one possibility is that the energy is injected into the corona by small - scale random footpoint motions , which lead to twisting and braiding of coronal field lines .
however , in the highly sheared magnetic field of a sigmoid there is another possibility , namely , that the energy for coronal heating is derived from the energy of the flux rope itself . in this case
the energy is already stored in the corona and does not need to be injected at the photospheric footpoints .
generally , a large amount of free energy is available in the coronal flux rope ( @xmath4 erg ) , sufficient to explain both the heating of the coronal plasma and the occurrence of flares and coronal mass ejections .
the second scenario leads us to propose a new theory for the heating of coronal flux ropes .
the basic ideas are illustrated in figure [ fig1 ] .
the lower part of the figure shows a flux rope in the low corona . inside the flux rope
is a magnetic field that runs more or less parallel to the flux rope axis .
the flux rope is held down by an overlying coronal arcade ( not shown ) , which has a magnetic field that runs perpendicular to the flux rope axis .
hence , going from the inside to the outside of the flux rope the direction of the mean magnetic field rotates through an angle of about 90 degrees .
the upper part of the figure shows the fine - scale structure of the magnetic field at the boundary between the flux rope and its surroundings .
the magnetic field has both mean and fluctuating components , as indicated by the wiggles in the field lines .
the total magnetic field @xmath2 is assumed to be stochastic in nature @xcite .
the magnetic fluctuations are assumed to be produced by some unspecified mhd process that ultimately derives its energy from the free magnetic energy of the coronal flux rope .
@xcite proposed that turbulent fluctuations could be produced by tearing modes in coronal current sheets , but the present paper does not depend on that particular mechanism .
we only assume that reconnection occurs intermittently at many sites in the stochastic field .
the reconnection events produce outflows and energetic particles that ultimately result in heating of the coronal plasma .
the above process is an example of a `` dynamo '' process because it involves turbulent diffusion of the mean magnetic field of the coronal flux rope .
dynamo activity , defined as the amplification and/or maintenance of large - scale magnetic fields , occurs in many astrophysical and laboratory plasmas , and magnetic helicity evolution plays an important role in such activity ( e.g. , * ? ? ?
* ; * ? ? ?
@xcite and @xcite considered turbulence in low-@xmath5 plasmas , and proposed that the evolution of the mean magnetic field can be described in term of _ hyperdiffusion _ , a type of magnetic diffusion in which the magnetic helicity of the mean field is conserved .
@xcite proposed that tearing modes can produce turbulence in three - dimensional sheared magnetic fields , and showed that such turbulence causes hyperdiffusion , which may explain both fast reconnection in solar flares and the heating of the solar corona . @xcite
considered the constraints on the dynamo mechanism from helicity conservation and the back - reaction of the magnetic field on plasma dynamics .
they showed that the results of kinematic dynamo theory are strongly modified by the production of hyperdiffusion in the mean field equation .
hyperdiffusion has also been included in models for decaying active regions . in the present paper
we consider the plasma heating associated with hyperdiffusion . at present , it is unclear to what extent stochastic effects dominate the solar magnetic field .
our purpose here is not to develop a complete theory of such fields , but rather to determine some of the observable consequences of their existence .
therefore , more detailed questions , such as how the stochastic field is produced and the nature of the mhd instabilities , will remain unanswered . the main goal is to demonstrate that the existence of stochastic fields is consistent with existing observations of coronal heating in active regions .
the theory of hyperdiffusion was first proposed by @xcite and @xcite , and further developed by @xcite , @xcite , @xcite and others . in this section
we summarize the results of these analyses as they apply to the solar corona , i.e. , when the plasma is magnetically dominated .
the total magnetic field @xmath6 is a sum of mean and fluctuating components , and is described in terms of the vector potential @xmath7 , where @xmath8 .
the evolution equation for the vector potential is @xmath9 where @xmath10 is the electric field , @xmath11 is the scalar potential , and @xmath12 is the speed of light . inserting ohm s law and ampere s law , we obtain @xmath13 where @xmath14 is the plasma velocity , and @xmath15 is the ohmic diffusivity .
we assume that @xmath16 , where @xmath17 is the system size and @xmath18 is the time scale for the evolution of the mean magnetic field .
then reconnection occurs only in highly localized regions ( e.g. , current sheets ) that make up only a small fraction of the volume .
following @xcite , we use the coulomb gauge , @xmath19 , so that the scalar potential is given by @xmath20 the magnetic helicity equation follows from equation ( [ eq : dadt ] ) : @xmath21 where @xmath22 is the helicity density , and @xmath23 is the helicity flux : @xmath24 @xcite has shown that in the limit @xmath16 the right - hand side of equation ( [ eq : heli ] ) can be neglected , so this equation describes the conservation of magnetic helicity . the relative magnetic helicity is defined by @xmath25 ~ dv$ ] , where subscript @xmath26 refers to the potential field , and the integration includes the coronal and subsurface regions .
therefore , in the limit @xmath16 , if no helicity is injected into the corona at the photosphere or ejected into the heliosphere , @xmath27 is constant in time .
let @xmath28 be the mean magnetic field smoothed over some intermediate spatial scale @xmath29 , and let @xmath30 be the fluctuating field on scales less than @xmath29 .
similar decompositions are made for other physical quantities . applying the smoothing operator to equations ( [ eq : e ] ) and ( [ eq : dadt ] ) ,
we obtain : @xmath31 where @xmath32 , @xmath33 , @xmath34 and @xmath35 are the mean vector potential , scalar potential , electric field and plasma velocity , and @xmath36 is the `` mean electromotive force '' describing the effect of the fluctuations on the mean magnetic field @xcite .
here we neglected the ohmic diffusion of the mean field .
the helicity equation for the mean magnetic field can be derived from equation ( [ eq : dadt0 ] ) : @xmath37 = 2 { \cal e}_0 \cdot { \bf b}_0 .
\label{eq : heli1}\ ] ] on the other hand , applying the smoothing operator directly to equation ( [ eq : heli ] ) with @xmath38 , we obtain : @xmath39 = 0 , \label{eq : heli2}\ ] ] where @xmath40 is the contribution of the fluctuations to the helicity flux : @xmath41 we assume that the large - scale structures dominate the helicity budget , so that the helicity of the total field can be approximated by the helicity of the mean field , @xmath42 @xcite .
subtracting equation ( [ eq : heli1 ] ) from equation ( [ eq : heli2 ] ) yields the following relationship between @xmath43 and @xmath40 : @xmath44 the electromotive force can be written as a sum of components parallel and perpendicular to the mean magnetic field , @xmath45 @xmath46 .
the parallel component is given by equation ( [ eq:2eb ] ) , and the perpendicular component can be written in terms of a transport velocity @xmath47 , which is defined by @xmath48 . then the electromotive force is given by @xmath49 we assume that the magnetic field evolves slowly in response to hyperdiffusion and changes in the boundary conditions . furthermore , we assume that the magnetic pressure is much larger than the plasma pressure , so that the mean magnetic field evolves through a series of force - free equilibrium states , @xmath50 .
the force - free condition can also be written as @xmath51 where @xmath52 is the torsion parameter of the mean field . for a linear force free field ( @xmath52 = constant )
the helicity flux @xmath40 should vanish because there is no magnetic free energy available to drive the magnetic fluctuations .
therefore , following @xcite and @xcite we assume that @xmath40 is proportional to the gradient of @xmath52 : @xmath53 where @xmath54 is the so - called hyperdiffusivity . inserting expression ( [ eq : hyper ] ) into equations ( [ eq : e0 ] ) and ( [ eq : dadt0 ] ) yields the following diffusion equation for the mean magnetic field : @xmath55 where @xmath56 is the total transport velocity .
the second term on the right - hand side of equation ( [ eq : dadt1 ] ) describes the effect of the magnetic fluctuations in diffusing the mean magnetic field .
the effect is such that the magnetic helicity of the system is conserved .
hyperdiffusion describes the tendency of the magnetic field to relax to a state of constant @xmath52 @xcite .
the heating of the coronal plasma resulting from taylor relaxation was first modeled by @xcite .
the energy equation for the mean magnetic field can be derived from equations ( [ eq : dadt0 ] ) , ( [ eq : e0 ] ) and ( [ eq : fff0 ] ) : @xmath57 and inserting equation ( [ eq : hyper ] ) we can write @xmath58 where @xmath59 is the total transport velocity .
the quantity on the rhs of equation ( [ eq : ener2 ] ) is negative and can be interpreted as the rate at which mean magnetic energy is converted into heat @xcite .
assuming there is no inflow of energy at the boundaries , the total magnetic energy decreases monotonically with time : @xmath60
the above theory predicts a particular form of the mean - field diffusion equation , but does not explain why the helicity flux @xmath40 is proportional to @xmath61 , nor does it predict the value of the hyperdiffusivity @xmath62 contained in equation ( [ eq : hyper ] ) . to estimate this quantity , a detailed understanding of the processes that drive the magnetic fluctuations would be required .
@xcite assumed the fluctuations are driven by tearing mode instability , but in the present paper we make no assumption about the nature of the instability .
we only assume that ( 1 ) the driving process results in stochastic magnetic field lines @xcite , as illustrated in fig .
[ fig1 ] ; ( 2 ) reconnection occurs intermittently at many sites within this stochastic field ; ( 3 ) the reconnection events result in alfvnic perturbations that travel along the stochastic field lines until they are damped as a result of turbulent cascade .
therefore , the magnetic field is not completely force free . in a non - force - free field
, the magnetic torsion parameter @xmath0 can be defined as @xmath63 , and we assume that this quantity varies in a time - dependent way along the field lines .
the topology of the field is continually changing as a result of the reconnection processes , with new magnetic connections being formed and other connections being severed .
consider a small volume near the edge of the coronal flux rope where the mean torsion @xmath64 can be approximated as a linear function of position .
let @xmath65 be the coordinate perpendicular to the surfaces of constant @xmath52 ( we assume @xmath66 ) .
we assume that at time @xmath67 magnetic reconnection occurs at some localized site , creating a new magnetic connection between two neigboring surfaces @xmath65 = constant .
let @xmath68 and @xmath69 be two points on the newly formed field line , such that @xmath68 is located on @xmath70 and @xmath69 is located on @xmath71 .
the displacement @xmath72 between these two points as measured along the field line is assumed to be much larger than the distance @xmath73 between the two surfaces .
the displacements are given by @xmath74 where @xmath75 and @xmath76 are the parallel- and perpendicular length scales of the turbulent motions ( relative to the mean magnetic field ) .
the torsion parameters @xmath77 and @xmath78 at the two points differ by an amount @xmath79 we assume that this difference in torsion results in the launch of alfvn waves that travel in both directions along the newly formed field line . in appendix
a we show that these waves carry a certain amount of magnetic helicity
. the net helicity flux due to the waves can be estimated by inserting expression ( [ eq : dalpha ] ) into equation ( [ eq : flux ] ) : @xmath80 where @xmath81 is the alfvn speed , and we assume that the radius @xmath82 of the newly formed flux tube is determined by the outer scale of the turbulence , @xmath83 .
the quantity @xmath84 is a dimensionless constant , and we estimate it to be rather small , @xmath85 ( see appendix ) .
the @xmath65-component of the helicity flux is @xmath86 where we take into account the angle of the field lines with respect to the @xmath65-direction , and we use equations ( [ eq : fluxp ] ) and ( [ eq : disp ] ) .
this shows that the helicity flux @xmath40 due to the fluctuations is proportional to the local gradient of @xmath52 , as assumed in equation ( [ eq : hyper ] ) .
furthermore , we obtain the following estimate for the hyperdiffusivity : @xmath87 note that @xmath54 depends strongly on the perpendicular length scale of the turbulent motions associated with the reconnection in the stochastic magnetic field .
the coronal magnetic field is anchored in the photosphere where the field is highly fragmented @xcite .
the photospheric field consists of vertically oriented , kilo - gauss _ flux tubes _ separated by nearly field - free plasma .
these flux elements are continually buffeted by convective flows on the scale of the solar granulation , which results in random twisting of the flux tubes and a random walk of the flux elements over the solar surface ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . both types of motions
cause disturbances that propagate upward along the field lines in the form of alfvn waves ( e.g. , * ? ? ?
* ) . in coronal loops ,
such disturbances produce small - scale twisting and braiding of the coronal field lines and associated field - aligned electric currents .
@xcite argues that the observed widths of coronal loops are determined by these footpoint motions .
the random walk of the footpoints is predicted to drive a cascade of magnetic energy from large to small spatial scales @xcite .
the existence of such a cascade is confirmed by results from 3d numerical simulations of the parker problem with incompressible boundary flows .
however , the twisting motions of the flux tubes may be more important than the lateral motions , as the twist velocity is amplified by the expansion of the flux tubes @xcite . at the smallest scales in the corona , impulsive reconnection events ( `` nanoflares '' )
play an important role in the release of the magnetic energy @xcite .
the actual dissipation may be a multi - step process involving reconnection outflows , waves , and energetic particles .
@xcite have modeled the coronal heating that ultimately results from granule - scale footpoint motions . in this paper
we assume that the coronal heating rate @xmath88 has contributions from two different sources : ( 1 ) disspation of energy that was directly injected into the corona by granule - scale footpoint motions , as described above ; ( 2 ) dissipation of energy from a large - scale coronal flux rope via the process of hyperdiffusion .
the former uses energy propagated upward along the field lines , while the latter draws its energy from large - scale magnetic shear already present in the corona .
the direct contribution to the heating is approximately given by @xmath89 where @xmath90 is the position along the coronal loop , @xmath91 is the heating rate per unit volume , @xmath92 is the magnetic field strength , @xmath93 is the velocity of the perpendicular motions , @xmath94 is the correlation time of the photospheric motions , @xmath17 is the loop length , and @xmath95 .
this expression is a generalization of equation ( 61 ) in @xcite , using a magnetic reynolds number @xmath96
. equation ( [ eq : eps_direct1 ] ) assumes that the alfvn - travel time along the coronal loop is short compared to the correlation time , @xmath97 .
the perpendicular velocity is given by @xmath98 where @xmath99 is the root - mean - square velocity in the photosphere , @xmath100 is the photospheric field strength , and we assume that the magnetic flux associated with the perturbations is constant along the field lines from the base of the photosphere into the corona . combining the above equations , the direct heating rate is @xmath101 similar to equation ( 67 ) of @xcite . in the present paper
we use @xmath102 kg , @xmath103 @xmath104 , and @xmath105 s , which yields a photospheric diffusion constant @xmath106 @xmath107 [ see equation ( 62 ) of @xcite ] , consistent with the observed dispersal coefficient that characterizes the granular random walk up to several hours @xcite
. the heating rate due to hyperdiffusion follows from equations ( [ eq : ener2 ] ) and ( [ eq : eta4 ] ) : @xmath108 where we assume that the parallel length scale of the fluctuations is determined by the loop length , @xmath109 .
the perpendicular length scale @xmath110 of the turbulence is assumed to vary along the field lines in a manner consistent with magnetic flux conservation : @xmath111 where @xmath112 is the corresponding length scale in the photosphere .
however , the turbulence is not driven by the footpoint motions , therefore @xmath112 does not necessarily reflect any property of the photospheric velocity field .
we can not provide a more detailed expression for @xmath112 because the nature of the physical processes that produce the stochastic field are still unclear ; instead , we treat @xmath112 as a free parameter of the model . combining the above expressions , we find @xmath113 the total heating rate is assumed to be the sum of expressions ( [ eq : eps_direct ] ) and ( [ eq : eps_hyper ] ) : @xmath114 in the next section we use these expressions to compute the coronal temperature and density , which provides a link to solar observations .
@xcite developed empirical models for nlfffs in two active regions observed with trace .
our initial goal was to compute temperatures and densities for these models .
however , we found that small deviations from the force - free condition can cause significant errors in the distribution of @xmath61 along the field lines , so that the heating rate @xmath115 can not yet be accurately computed from such empirical models .
therefore , in the present paper we apply the above expressions for the heating rate to a semi - analytic model of an active region .
the model describes the _ mean _ magnetic field @xmath116 in the corona ; fluctuating fields are not modeled , nor do we consider the evolution of the mean field .
the nlfff model is constructed as follows .
we use a cartesian coordinate system @xmath117 with @xmath65 being the height above the coronal base .
following @xcite , we assume that the magnetic field has cylindrical symmetry with respect to an axis that lies horizontally below the coronal base at @xmath118 , @xmath119 .
let @xmath120 be a cylindrical coordinate system with @xmath121 the distance from the cylinder axis and @xmath122 the distance along the axis .
we consider a finite domain @xmath123 $ ] and @xmath124 $ ] , where the inner radius of the domain is given by @xmath82 , and @xmath125 .
we initially use dimensionless units such that @xmath126 .
the force free field is written as : @xmath127 where @xmath128 is the flux function , and @xmath129 describes the azimuthal component of the field .
the flux function satisfies the following non - linear equation : @xmath130 where @xmath131 . in this paper
we use boundary conditions such that @xmath132 @xmath133 at the inner boundary , and @xmath134 at the outer boundaries .
the azimuthal field is given by @xmath135 .\ ] ] where @xmath5 is a parameter describing the magnitude of the azimuthal field .
equation ( [ eq : cfff ] ) is solved using a relaxation technique . in essense
, we solve the following time - dependent equation : @xmath136 , \label{eq : cfff2}\ ] ] until a stationary solution @xmath128 is obtained ( here @xmath137 is a dimensionless diffusion constant ) . in each time step , @xmath5 is adjusted such that the total azimuthal current has a fixed value , @xmath138 . once a stationary solution is obtained , the spatial scale is changed to @xmath139 mm , and the magnetic field is rescaled by a factor 100 , producing a maximum value of @xmath140 of 55 g. the quantities @xmath141 , @xmath142 and @xmath143 are computed on a grid of @xmath144 points in the @xmath145 plane .
the magnetic vector @xmath146 in the cartesian frame can then be found by evaluating these quantities at @xmath147 .
figure [ fig2 ] shows the magnetic flux distribution at the coronal base ( @xmath148 ) and selected field lines in the flux rope .
the stability of the @xcite model has been considered by @xcite .
they find that the @xmath149 kink mode is stable provided the average twist in the flux rope is less than @xmath150 .
this condition is satisfied in the present case , so we assume that the field shown in fig .
[ fig2 ] is stable .
the nlfff model provides the spatial distribution of @xmath151 , which is needed to compute the heating rate @xmath88 .
we trace a large number of field lines that cross the vertical plane @xmath152 , and solve the energy balance equation for each field line .
we use a modified version of the iterative method described by @xcite .
the code assumes steady - state heating , and includes thermal conduction and optically thin radiative losses .
the contribution to the heating from hyperdiffusion is determined from equation ( [ eq : eps_hyper ] ) with the alfvn speed @xmath153 taken from the previous iteration .
we assumed that the perpendicular length scale of the turbulence is characterized by @xmath154 km .
the chromosphere - corona transition region ( tr ) is assumed to be located at @xmath155 , where the temperature @xmath156 k. the solution of the energy balance equation yields the temperature @xmath157 and electron density @xmath158 as functions of position @xmath90 along a field line .
figure [ fig3 ] shows modeling results for one field line at the edge of the coronal flux rope .
starting at @xmath159 and @xmath160 mm , the field line is traced forward and backward through the 3d magnetic model until the level @xmath148 is reached .
the temperature @xmath157 and density @xmath158 along this field line are shown in fig .
[ fig3]a , and the heating rate @xmath161 is shown in fig .
[ fig3]b ( _ full _ curve ) .
the dashed curve in fig .
[ fig3]b shows the `` direct '' contribution to the heating from photospheric footpoint motions . note that this contribution is relatively small ,
so for this particular field line the heating is dominated by hyperdiffusion . however ,
for other field lines in the nlfff model ( e.g. , along the axis of the flux rope ) the hyperdiffusive heating rate is smaller than the direct heating rate .
we computed temperature and density profiles for a large number of field lines that cross the vertical plane @xmath152 .
figure [ fig4 ] shows the temperature @xmath162 and density @xmath163 in the plane @xmath152 for two different heating models .
the top panels show the results when both direct and hyperdiffusive heating are applied together ( @xmath154 km ) .
note that the temperature and density are enhanced in a thick shell where the gradient of @xmath52 is largest .
the density distribution shows a `` cavity '' that is round at the top and v - shaped at the bottom .
the temperature is also enhanced at the bottom of the v - shaped cavity because the field lines in this region are relatively long , so the conductive cooling is reduced .
the bottom panels of figure [ fig4 ] show the temperature and density distributions in the @xmath118 plane when only the direct heating is applied ( i.e. , no hyperdiffusion ) .
the highest temperatures now occur within the flux rope where the field lines are longest and conductive losses are weakest .
the lowest densities occur in the v - shaped lower part of the flux rope , but the density in the upper part is nearly the same as in the surroundings , and there is no dome - shaped cavity . on the sun , cavities are frequently observed in the k - corona emission , and these cavities have been linked to the presence of coronal flux ropes .
we suggest that the observed cavities are evidence for a spatial distribution of coronal heating that is consistent with hyperdiffusion .
we propose a new mechanism for coronal heating relevant for magnetic structures in which the large - scale magnetic field is highly sheared and/or twisted .
we assume that the magnetic field in such coronal flux ropes is stochastic in nature @xcite , and contains small - scale perturbations superposed on a mean magnetic field .
the magnetic free energy associated with the flux rope can be converted into heat via small - scale reconnection processes .
we describe this process in terms of _ hyperdiffusion _ , a type of magnetic diffusion in which the magnetic helicity is conserved .
the effect of hyperdiffusion on the mean field depends on the gradient of the mean torsion parameter @xmath164 .
therefore , hyperdiffusion describes the tendency of a magnetic system to relax to a taylor state @xcite .
in section 3 we consider helicity transport in a stochastic magnetic field , and we show that small - scale reconnection events produce a net helicity flux @xmath40 consistent with hyperdiffusion .
we derive an expression for the hyperdiffusivity @xmath54 in terms of the parallel and perpendicular length scales of the magnetic perturbations .
the calculation is based on a simple model for the alfvn waves produced by reconnection events ( appendix a ) .
we find that the hyperdiffusivity depends strongly on the perpendicular length scale of the turbulence , @xmath165 .
the spatially averaged heating rate @xmath166 due to hyperdiffusion is proportional to @xmath167 .
this heating is distinct from the `` direct '' heating due to small - scale random footpoint motions . in section 4
we derive an expression for the total heating rate @xmath88 , which is a sum of direct and hyperdiffusive contributions .
our analysis of the dissipation process is very approximate and makes many simplifying assumptions .
for example , we can not identify a specific mechanism by which the stochastic magnetic field is produced and maintained .
we simply treat the perpendicular length scale @xmath112 of the fluctuating field as a free parameter to be determined from observations or more complete theories of turbulent reconnection . also , it is unclear exactly what drives the small - scale reconnection events necessary for hyperdiffusion .
the photospheric magnetic field is highly fragmented , and there are separatrix surfaces in the corona corresponding to the interfaces between photospheric flux tubes .
we speculate that the reconnection events may occur at these separatrix surfaces . in section 5 we develop a 3d model of an active region containing a magnetic flux rope , and we compute the temperature and density distributions with and without hyperdiffusion .
the results show that hyperdiffusion can have a significant effect on the temperature structure , provided the perpendicular length scale of the magnetic fluctuations @xmath168 km at the coronal base .
if @xmath112 were of order 100 km or less , the hyperdiffusive heating would not have a significant impact on the total heating rate for any field line .
for the nlfff model considered here , the heating is concentrated in a thick shell surrounding the flux rope .
the heating at the axis of the flux rope is weaker , producing a `` cavity '' similar to the coronal cavities observed on the sun . at present
there is no direct observational evidence to confirm or refute the existence of stochastic magnetic fields with perpendicular scales @xmath169 km .
therefore , it is unclear exactly how coronal flux ropes are heated .
coronal images obtained with the _ solar dynamics observatory _ ( sdo ) may provide new observational constraints on stochastic fields and the physical processes by which coronal flux ropes are heated .
we thank amitava bhattacharjee for discussions about hyperdiffusion .
this work was inspired by observations from hinode / xrt .
the work was supported by nasa ltsa grant nng04ge77 g , and by nasa lws grant nnx06ag95 g .
the transport of magnetic helicity plays an important role in the theory of hyperdiffusion . here
we consider a simple model for helicity transport due to reconnection events within the stochastic field .
the average time @xmath170 between reconnection events on a given flux tube is assumed to be a few times the alfvn travel time along that tube .
let us assume that at time @xmath67 a new section of magnetic field has just been created by reconnection .
the new section connects two points @xmath68 and @xmath69 that may be located at opposite ends of a coronal loop .
we neglect the loop curvature and approximate the new section as a cylindrical flux tube with radius @xmath82 .
we assume that in the initial state the plasma in the tube is stationary , but the torsion parameter @xmath0 varies along the tube , so the field is not force free .
the variable torsion causes the launch of counter - propagating alfvn waves that travel in opposite directions along the tube .
we assume that the wave length @xmath75 of the waves is large compared to the tube radius .
the medium surrounding the flux tube is assumed to be unaffected by the waves .
we use a cylindrical coordinate system @xmath171 where @xmath90 is the distance along the tube ( @xmath172 ) .
the plasma is assumed to be fixed at the end points @xmath68 ( @xmath173 ) and @xmath69 ( @xmath174 ) , so that the waves reflect at these points .
the magnetic field @xmath175 inside the tube is written as @xmath176 { \mbox{\boldmath $ \hat{\phi}$}},\ ] ] where @xmath177 is the axial field ( constant in space and time ) , @xmath178 is the unperturbed azimuthal field at the cylinder wall ( @xmath179 ) , and @xmath180 is the perturbation of the azimuthal field due to the waves .
the torsion parameter @xmath0 is defined by @xmath63 , and the perturbation of this quantity is @xmath181 the vector potential @xmath182 is obtained by solving the equations @xmath183 and @xmath184 .
continuity with the surrounding medium requires that @xmath185 is independent of @xmath90 and @xmath186 . in the limit @xmath187 ,
the solution is given approximately by @xmath188 \hat{\bf s } , \ ] ] where @xmath189 and we neglect a small correction @xmath190 needed to make @xmath191 exactly zero . the velocity of the waves is assumed to be purely azimuthal , @xmath192 , where @xmath193 is the velocity at the cylinder wall .
inserting this expression into equation ( [ eq : phi2 ] ) , we obtain the electric potential : @xmath194 where we assumed @xmath195 and required @xmath196 to ensure continuity with the surrounding medium .
the axial component of the helicity flux follows from equation ( [ eq : h ] ) : @xmath197 \cdot \hat{\bf s } = - \frac{r^2 b_0 ^ 2 } { r } \delta v_\phi ( s , t ) .\ ] ] the average of @xmath198 over the cross - section of the tube is @xmath199 note that the helicity flux is independent of the unperturbed azimuthal field @xmath178 .
we now compute the helicity flux for damped alfvn waves . in reality
the damping would likely involve a turbulent cascade of wave energy to smaller spatial scales , but here we use a simple model based on an effective `` frictional '' damping .
the wave equations are @xmath200 where @xmath201 is a constant density and @xmath18 is the frictional damping time .
the boundary conditions are @xmath202 , which implies @xmath203 at the boundaries .
we assume the plasma is initially stationary , @xmath204 .
the initial perturbation of the torsion parameter , @xmath205 , is assumed to be a nearly linear function of position : @xmath206 \approx \delta \alpha \left ( \frac{s}{\delta s } - \frac{1}{2 } \right ) , \label{eq : init}\ ] ] where @xmath207 is the perturbation amplitude ( we use @xmath208 ) .
the solution of the above equations is @xmath209 , \label{eq : wb}\end{aligned}\ ] ] where @xmath210 ; @xmath211 is the amplitude of the magnetic perturbation ; @xmath212 is the effective damping rate ; @xmath213 is the mode frequency ; and @xmath214 is the frequency for an undamped mode .
the above solution is valid for @xmath215 .
the helicity flux at the mid - point of the tube ( @xmath216 ) follows from equations ( [ eq : hpara ] ) and ( [ eq : wv ] ) : @xmath217 where @xmath218 in the absence of damping ( @xmath219 ) , the helicity would oscillate back and forth between the two halves of the tube , and the time average of @xmath198 would vanish . with damping there is a net transport of helicity from the region with high @xmath0 to the region with low @xmath0 : @xmath220 where @xmath84 is the time average of @xmath221 .
we assume that the next reconnection event occurs at a time @xmath170 which is only a few times the alfvn travel time , say , @xmath222 .
then @xmath84 is given by a time average over the period @xmath223 .
for a damping time @xmath18 such that @xmath224 , the time average of @xmath221 is given by @xmath85 . in section 3 we use the above expressions to estimate the hyperdiffusivity in a stochastic field . | a theory for the heating of coronal magnetic flux ropes is developed .
the dissipated magnetic energy has two distinct contributions : ( 1 ) energy injected into the corona as a result of granule - scale , random footpoint motions , and ( 2 ) energy from the large - scale , nonpotential magnetic field of the flux rope .
the second type of dissipation can be described in term of hyperdiffusion , a type of magnetic diffusion in which the helicity of the mean magnetic field is conserved .
the associated heating rate depends on the gradient of the torsion parameter of the mean magnetic field .
a simple model of an active region containing a coronal flux rope is constructed .
we find that the temperature and density on the axis of the flux rope are lower than in the local surroundings , consistent with observations of coronal cavities .
the model requires that the magnetic field in the flux rope is stochastic in nature , with a perpendicular length scale of the magnetic fluctuations of order 1000 km . |
within the context of the heterotic superstring and heterotic m - theory @xcite , there have been a number of vacuum states whose four - dimensional low energy effective field theory @xcite has the exact spectrum of the mssm with or without right - handed neutrino chiral multiplets and , to prohibit rapid proton decay , contains r - parity @xcite either as a discrete symmetry or as a subgroup of an anomaly free @xmath4 extension of the standard model gauge group @xcite .
one such vacuum was presented in @xcite and will be referred to as the @xmath0 mssm .
be that as it may , this is only the first step in finding a realistic heterotic string vacuum .
any such theory must also also be compatible with all presently observed low - energy phenomenology ; that is , it must spontaneously break electroweak ( ew ) symmetry at the observed scale , must be compatible with the newly discovered higgs particle with mass @xmath5 gev @xcite , have all sparticle masses above the present observational lower bounds and assuming r - parity is contained in an additional @xmath4 symmetry
spontaneously break that abelian group with an associated gauge boson mass in excess of the present experimental lower bound . in a series of papers @xcite ,
the @xmath0 mssm was examined in detail using a random statistical sampling of the initial set of soft supersymmetry ( susy ) breaking parameters and the results confronted with these phenomenological requirements .
it was shown , _ within a restricted region of the compactification moduli space , _ that the @xmath0 mssm easily passed each of these requirements for a large and basically uncorrelated set of initial conditions . furthermore , this analysis led to a series of low energy predictions for example , directly relating the lightest stop decay channels and branching ratios to the neutrino mass hierarchy and mixing angles thus linking lhc experimental results to neutrino measurements .
that is , the @xmath0 mssm is a possible candidate for a phenomenologically acceptable theory of the real world a statement that will be directly testable as its structure and predictions are confronted with upcoming data from the lhc , neutrino experiments and cosmological observations . for these reasons , in this paper we extend the results of @xcite to a _
more general , and more natural , region _ of the @xmath0 mssm moduli space . to quantify this
, we should briefly discuss the structure of the @xmath0 mssm .
it arises in the observable sector of heterotic m - theory compactified to four - dimensions on a shoen calabi - yau ( cy ) threefold @xcite with first homotopy group @xmath6 .
this manifold admits a specific slope - stable holomorphic vector bundle @xcite with structure group @xmath7 , as well as two wilson lines each wrapped over a two - cycle associated with a different @xmath8 homotopy factor
this theory has three , in principle distinct , mass scales@xmath9 at which the gauge bundle spontaneously breaks @xmath10 to @xmath11 and two wilson line mass scales , which we denote by @xmath12 and @xmath13 , associated with the inverse radii of their respective two - cycles . in our previous analysis , we worked _ in a restricted region of cy moduli space _ where the radius of one two - cycle is distinctly smaller than that of the other ; that is , we chose @xmath14 . by taking the scale of separation of the two wilson lines to be approximately an order of magnitude , one can exactly unify all gauge coupling parameters thus specifying boundary conditions in the renormalization group equations ( rges ) .
however , this specificity comes at the cost of introducing an _
additional scaling regime_. between @xmath13 and @xmath12 the effective theory is that of the `` left - right '' model @xcite with gauge group @xmath15 and a specific particle spectrum that can be computed from string theory .
it is only for energy - momentum below the lightest wilson line mass @xmath12 that one obtains the spectrum and @xmath16 gauge symmetry of the @xmath0 mssm . in this paper , we generalize and simplify the phenomenological analysis of the @xmath0 mssm by working in a _
generic _ region of cy moduli space where the radii of the two wilson lines and the average radius of the cy manifold are all approximately equal : that is , with @xmath17 .
this generalization is significant in that 1 ) _ the region of moduli space is much larger and more `` natural '' _ than that used previously and 2 ) _ the `` left - right '' scaling region is eliminated _ , with the @xmath0 mssm emerging immediately below the compactification scale thus simplifying the scaling regimes .
of course , the four @xmath0 mssm gauge couplings will no longer unify near the scale of the cy radius .
this does somewhat complicate the rg analysis .
however , it opens the door for a discussion of unification of all gauge couplings with the gravitational coupling at the `` string scale''as has been discussed by many authors in @xcite .
more specifically , such unification should take place at tree level .
however , at the one - loop ( and higher ) level one expects such unification to be split by `` threshold '' corrections .
these are due to several effects , such as the inclusion of field theory thresholds at each of the susy , @xmath0 and `` unification '' scales , and genus - one string theory corrections .
since in this analysis the latter is expected to be the largest , we will focus exclusively on them . by running the four @xmath0 mssm gauge parameters up to the string scale
, we will 3 ) _ statistically compute the heavy string threshold corrections for each gauge coupling_. furthermore , we will statistically compute the hypercharge gauge threshold and , by subtracting various thresholds , analyze the moduli dependent sub - component of each .
finally , there is yet another important benefit of analyzing the @xmath0 mssm at the generic region of its cy moduli space
although we will not pursue this in the present paper . in our previous work and the present analysis ,
the scale of the soft susy breaking parameters is chosen to be in the tev region .
this is done in order for our low energy phenomenological predictions to be lhc accessible . however , unlike in the case of unified gauge couplings enforced by splitting the wilson line masses discussed in @xcite which , for reasons elucidated in the conclusion , is essentially restricted to the tev region for the simultaneous wilson line masses discussed in this paper , the susy breaking mass scale can be taken to be arbitrarily large .
this has a number of important applications , both in particle phenomenology and in early universe cosmology @xcite .
we will pursue this in future work @xcite .
the present paper is structured as follows . in section
ii we review the salient parts of the @xmath0 mssm theory , presenting the spectrum , the supersymmetric and the soft susy breaking lagrangians , discussing the generic structure of spontaneous @xmath0 and ew symmetry breaking and setting our notation .
section iii is devoted to defining the exact meaning of the `` simultaneous '' wilson line analysis presented in this paper as opposed to the `` split '' wilson line approach in our previous work .
it then discusses , in detail , the four relevant mass scales from `` unification '' to electroweak symmetry breaking .
the statistical definitions of the `` unification '' mass and gauge coupling are given in our present context . in section iv ,
the three scaling regimes for both the `` right - side - up '' and `` upside - down '' scenarios along with the associated gauge coupling beta function parameters are presented .
the rg running of the yukawa couplings , including their transition at the susy scale , is discussed .
section v gives a brief review of the `` statistical '' approach to setting the initial soft supersymmetry breaking parameters at the `` unification '' scale presented in detail in our previous work . in section vi , the experimental constraints on the sparticle masses , the heavy vector boson mass and the lightest neutral higgs mass are presented .
we then solve the rges for randomly chosen initial soft susy breaking parameters sequentially from the `` unification '' scale down through the ew breaking scale subject to these constraints .
plots and the exact number of the initial points that sequentially satisfy the experimental constraints are given ; ending with the robust number of phenomenologically `` valid '' black points that satisfy all experimental constraints .
a brief analysis of both the lsp and non - lsp spectra is then given . in section vii ,
we briefly discuss fine - tuning in the @xmath0 mssm . in section viii ,
we statistically calculate the heavy string threshold corrections for each of the four @xmath0 mssm gauge couplings , as well as the hypercharge gauge threshold , and analyze the moduli dependent differences of these quantities .
finally , in section ix , we present our conclusions .
in this section , we briefly review the minimal anomaly free extension of the mssm with gauge group @xmath18 whose structure was motivated by heterotic string theory in @xcite and by phenomenological considerations in @xcite .
although this model has been discussed in our previous papers @xcite , we outline its main features in this section for specificity and to set our notation .
the abelian gauge factors @xmath19 can be rotated into physically equivalent charge bases , such as @xmath20 . however , as shown in @xcite , this comes at the cost of introducing kinetic mixing between the gauge fields .
we therefore prefer to work in the basis @xmath19 .
the gauge covariant derivative is d =- ig_3ri_3rw_3r - ig_blb^ , where @xmath21 , @xmath22 and @xmath23 , @xmath24 are the generators and couplings for the @xmath25 and @xmath26 groups respectively .
the gauge boson associated with @xmath26 is denoted @xmath27 to distinguish it from the gauge boson associated with @xmath28 , which is normally denoted @xmath29 .
the factor of @xmath30 in the last term is introduced by redefining the gauge coupling @xmath24 , thus simplifying many equations .
a radiatively induced vacuum expectation value ( vev ) of a right - handed sneutrino will break the abelian factors @xmath31 to @xmath28 , in analogy with the way the mssm higgs fields break @xmath32 to @xmath33 .
this process is referred to as `` @xmath0 '' symmetry breaking , although technically it breaks a specific combination of the groups generated from @xmath21 and @xmath22 , leaving invariant the usual hypercharge group generated by y = i_3r + .
the particle content of the model is simply that of the mssm plus three right - handed neutrino chiral multiplets .
this amounts to three generations of matter superfields @xmath34 along with the usual two higgs supermultiplets @xmath35 where we have displayed their @xmath36 quantum numbers .
the superpotential of the @xmath0 mssm is given by @xmath37 where both generational and gauge indices have been suppressed . in principle , the yukawa couplings are three - by - three complex matrices .
however , the observed smallness of the ckm mixing angles and cp - violating phase imply that the quark yukawa matrices can be approximated as diagonal and real for the purposes of rg evolution in this paper .
the charged lepton yukawa coupling can be made diagonal and real by moving the pmns angles and phases into the neutrino yukawa couplings .
the small size of neutrino masses implies that the neutrino yukawa couplings can be neglected for the purposes of rg evolution in this paper .
the smallness of first- and second - generation fermion masses implies that first and second - generation yukawa quark and charged lepton yukawa couplings can also be neglected .
the @xmath38-parameter can be chosen to be real without loss of generality .
the soft supersymmetry breaking lagrangian is @xmath39 where generation and gauge indices have been suppressed .
the @xmath40-parameters and sfermion soft - mass terms can , in principle , be hermitian matrices in family space .
however , this tends to lead to unobserved cp violation .
therefore , we proceed assuming that they are diagonal and real . furthermore , as discussed in section v , we assume that the a - parameters are proportional to the yukawa couplings .
this implies that all the @xmath40-parameters can be neglected except for the ( 3,3 ) component of the quark and charged lepton @xmath40-parameters .
the @xmath41-parameter can be chosen to be both real and positive without loss of generality .
although the gaugino soft masses can be complex in principle , this tends to lead to unobserved flavor and cp violation .
therefore , we proceed assuming that they are real . the @xmath0 symmetry is spontaneously broken by the vev in a right - handed sneutrino , which carries the appropriate @xmath21 and @xmath42 charges to break those symmetries while preserving hypercharge symmetry .
this vev is brought about by a sneutrino soft - mass term becoming tachyonic at the tev scale due to the rge evolution .
as discussed in @xcite , this vev will be purely in one of the three right - handed sneutrino generations not in a linear combination of them .
furthermore , the three generations of right - handed sneutrinoes can be relabeled without loss of generality .
therefore , we henceforth assume that it is the third - generation right - handed sneutrino that acquires a vev .
electroweak symmetry is broken by vevs in the neutral components of the up and down higgs multiplets .
the electroweak breaking vevs and the @xmath0 breaking vev together lead to small vevs in all three generations of left - handed sneutrinos .
the above vevs will be denoted by v_r , _ i , v_u , v_d , where @xmath43 is the generation index .
the neutral gauge boson that becomes massive due to @xmath0 symmetry breaking is referred to as @xmath44 .
defining @xmath45 , and assuming that @xmath46 , @xmath47 acquires to leading order a mass of m_z_r^2 = ( g_3r^2+g_bl^2 ) v_r^2 .
[ eq:237 ] the hypercharge gauge coupling is given by [ eq : y.3r.bl ] g_y = g_3r _ r = g_bl_r , where _ r = . the smallness of the neutrino masses implies , first , that the neutrino yukawa couplings are small and , second , that the left - handed sneutrino vevs are much smaller than the electroweak scale . in this limit , the minimization conditions of the potential simplify to @xmath48 noting from above that @xmath49 , equations and can be combined to give [ eq : mzr.mnuc ] m_z_r^2 = -2 m_^c_3 ^ 2 .
the vev in the third - generation right - handed sneutrino induces spontaneous bilinear r - parity violation through the operators @xmath50 where @xmath51 bilinear r - parity violation has been discussed extensively , including its relevance to neutrino masses .
see , for example , some early works @xcite .
the lagrangian of this model contains additional bilinear terms due to the sneutrino vevs : @xmath52 \\ & -\frac{1}{2 } v_r \left[-g_r \nu_3^c \tilde w_r + g_{bl } \nu_3^c \tilde b ' \right]+ \text{h.c . }
\end{split}\end{aligned}\ ] ] the r - parity violating terms in this model have a variety of interesting consequences that have been studied in a number of different contexts .
these include lhc studies @xcite , predictions for neutrinos @xcite , and connections between the two @xcite .
it has been shown that the r - parity violation can give rise to majorana neutrino masses , with the lightest left - handed neutrino being massless .
there is also a pair of sterile right - handed neutrinos that can have cosmological implications @xcite . the minimal supersymmetric @xmath0 model , reviewed in this section , will be referred to simply as the @xmath0 mssm throughout the rest of this paper .
we now turn to connecting the phenomenology of the @xmath0 mssm to its high - scale origins .
specifically , we are considering the possibility that the @xmath0 mssm is the observable sector of the low - energy effective theory of an @xmath53 heterotic string theory . in this context , the @xmath0 mssm gauge group unifies into an @xmath11 gauge group , which is itself the commutant of the @xmath54 structure group of the observable sector @xmath10 vector bundle on the cy threefold .
we have previously studied the @xmath0 mssm in this context @xcite . in this paper , however , we will study the effects of string threshold corrections on gauge unification .
this requires a discussion of gauge unification to which we now turn .
this section outlines the scales and scaling regimes associated with the evolution of the @xmath0 mssm from `` unification '' to the electroweak scale .
compactification to four dimensions yields a unified gauge group , @xmath11 , at mass scale @xmath9 .
this unified gauge group is broken by two abelian wilson lines , denoted by @xmath55 and @xmath56 .
the mass scales associated with these wilson lines , @xmath12 and @xmath13 respectively , depend on the inverse radii of the 2-cycles over which they are wrapped .
these , in turn , depend on the chosen point in the cy moduli space .
generically , one expects that the two wilson line masses are approximately the same and close to the @xmath11 unification scale .
that is , one `` naturally '' expects @xmath57 over a wide region of the cy moduli space .
however , as one moves away from these generic points the wilson line mass scales need not remain the same .
this leads to an intermediate regime between the two scales associated with the wilson lines .
the particle content and gauge group in this intermediate regime depends on which wilson line has a higher associated mass .
if @xmath14 , the particle content and gauge group of the intermediate regime is that of a `` left - right '' model . if @xmath58 , the particle content and gauge group of the intermediate regime is similar to that of a``pati - salam '' model
in each case , the lower - mass wilson line breaks the model in the intermediate regime to the @xmath0 mssm .
in fact , it was shown in @xcite that exact gauge coupling unification at one - loop _ requires _ that these scales be different .
for specificity of the rge calculation , it was convenient to _ impose _ precise gauge coupling unification .
hence , in @xcite we studied the two cases with separated wilson line masses even though this can occur only in special regions of moduli space . under the assumption that the soft susy breaking masses be of tev order to assure that sparticle masses potentially be lhc accessible we found that gauge coupling unification dictates that the wilson line scales must be separated by less than / approximately an order of magnitude in either case .
additionally , we found that both cases lead to similar low energy phenomenology .
hence , for specificity , we carried out our analysis using the first of these symmetry breaking patterns ; that is , the intermediate regime containing the `` left - right '' model .
we refer the reader to @xcite for that analysis . here , for concreteness , we simply show in figure [ fig : mu.mi.msusy ] the relationship of the @xmath59 unification scale to that of the mass @xmath12 of the the second wilson line in the `` left - right '' model case .
this is plotted as a function of @xmath60 defined below in eq . .
unification mass and @xmath12 as functions of the susy scale in the `` left - right '' scenario . ] in this paper , we turn to the analysis of the _ generic _ region of moduli space where equation , that is , @xmath17 , is satisfied , thereby _ giving up exact gauge unification_. be that as it may , to enable direct comparison of our new simultaneous wilson line results with those from the split wilson lines analyzed in @xcite , we continue to use the same notation for all quantities .
in particular , it is important to use identical notation for the @xmath0 gauge coupling .
thus far in this paper , we have discussed the gauge parameter @xmath24 , which couples to the @xmath61 generator . however , as was discussed in @xcite , this gauge coupling has to be properly normalized so as to unify with the other gauge parameters in the split wilson line scenarios .
the appropriate coupling was denoted @xmath62 and defined by g_bl = g_bl .
even though the four gauge couplings , including @xmath63 , will not unify in the simultaneous wilson line scenario in this paper , we will continue to use this parameter when appropriate .
note that @xmath62 couples to the @xmath64 generator and will appear in the rges . for quantities of physical interest , such as physical masses , @xmath24 will be used . to fully understand the evolution of this model from `` unification '' to the electroweak scale ,
it should be noted that there are four relevant mass scales of interest .
all four are described in the following : + * @xmath65 : the unification and the first and second wilson lines mass scale .
* since , as discussed above , exact unification of the four gauge couplings no longer occurs for simultaneous wilson lines , it is essential to give a justification and an explicit definition of what we mean by the `` unification mass '' in the present context . in @xcite ,
every phenomenologically valid point in the space of randomly chosen initial soft supersymmetry breaking parameters corresponds to an explicit unification mass @xmath9 and unified coupling @xmath66 .
both the unification scale and unification parameter vary for different valid points .
the associated statistical histograms for these quantities are shown in figures [ fig : a ] and [ fig : b ] respectively , along with their average values .
these are found to be gev . ] . ]
@xmath67 in this paper , we will refer to the average values @xmath68 and @xmath69 as the `` unification '' mass and `` unified '' gauge coupling and rg scale the gauge parameters between this scale and the electroweak scale .
the values of the four diverse couplings @xmath70 and @xmath71 at @xmath68 will be determined for each statistical choice of soft supersymmetry breaking parameters .
henceforth , for specificity , we will always take this unification scale and both wilson line masses to be strictly identical ; that is @xmath72 + * @xmath73 : the mass at which the right - handed sneutrino vev triggers @xmath74 symmetry breaking .
* physically , this corresponds to the mass of the neutral gauge boson @xmath44 of the broken symmetry and , therefore , the scale of @xmath44 decoupling . specifically m_z_r
= m_b - l .
[ eq:249 ] note that @xmath75 itself depends on parameters evaluated at @xmath73 .
this results in a transcendental equation that can be solved using for @xmath73 using numerical methods .
the boundary condition relating the hypercharge coupling to the gauge couplings of @xmath25 and @xmath26 at this scale is nontrivial .
it is given by @xmath76 where _
[ home1 ] as with the @xmath0 gauge coupling , the hypercharge coupling has been rescaling to allow for unification in the split wilson line scenarios . the rescaled hypercharge gauge coupling , @xmath77 , is defined by g_1 = g_y . [ home2 ] + * @xmath78 : the soft susy breaking scale .
* this is the scale at which all sparticles are integrated out , with the exception of the right - handed sneutrinos , which are associated with @xmath0 breaking and , therefore , are integrated out at the @xmath0 scale @xcite .
while the sparticles do not all have the same mass , we use the scale of stop decoupling as a representative scale associated with all sparticles .
that is , m _ = . [ eq:358 ] the scale of stop decoupling is the best choice for the susy scale since the stops give the dominant radiative corrections to phenomenologically important quantities such as the electroweak scale and the higgs mass .
see , for example , @xcite for more details .
note that the physical stop masses depend on quantities evaluated at @xmath79 .
therefore , this equation must be solved using iterative numerical methods for the correct value of @xmath79 . +
* @xmath80 : the electroweak scale . *
this is the well - known scale associated with the @xmath81 and @xmath82 gauge bosons of the standard model ( sm ) .
we identify this scale with the mass of @xmath81 boson , as is conventional .
that is , m _ = m_z .
having defined the relevant mass scales , we turn to a brief discussion of rg evolution that occurs between them .
the gauge coupling rges are _ a^-1 = - , where @xmath40 indexes the associated gauge groups .
the slope factors @xmath83 are different in the different scaling regimes .
* @xmath84 : we refer to this regime as the `` @xmath0 mssm regime '' because the particle content and gauge group are the @xmath0 mssm .
the @xmath83 factors are @xmath85 note that the hierarchy between the susy and @xmath0 scales depends on the point chosen in the initial parameter space .
the remaining two regimes depend on which of the following two cases occurs : @xmath86the `` right - side - up '' hierarchy and @xmath87the `` upside - down '' hierarchy .
+ : * @xmath88 : in this regime , the gauge group and particle content is that of the mssm plus two right - handed neutrino supermultiplets .
the gauge couplings in this regime evolve with the slope factors @xmath89 we refer to this regime as the `` mssm '' regime .
* @xmath90 : in this regime , the sparticles are integrated out , leaving the sm with an additional two sterile neutrinos .
it has the well - known slope factors @xmath91 we refer to this regime as the `` sm '' regime .
: * @xmath92 : in this regime , sparticles , with the exception of the third - generation right - handed sneutrino , are integrated out . but @xmath0 is still a good symmetry .
this yields a non - susy @xmath93 model , which also includes three generations of right - handed sneutrinos
the third of which acts as the @xmath0 higgs .
the slope factors are @xmath94 * @xmath95 : this regime is identical to the sm regime with slope factors given in eq .
( [ eq:644 ] ) .
the boundary conditions imposed on the gauge couplings are that the three @xmath96 coefficients of the sm take their experimental values at @xmath97 @xcite : @xmath98 these experimental values will then be scaled up through the various regimes : @xmath99 , @xmath100 ( for the right - side - up hierarchy ) or @xmath101 , @xmath102 ( for the upside - down hierarchy ) , followed by scaling through the @xmath0 mssm regime to @xmath2 using the beta functions listed above . the `` splitting '' of @xmath103 to @xmath104 and @xmath105 at @xmath73 is achieved using the boundary conditions , . in previous work @xcite , exact unification conveniently specified @xmath106 .
however , in the present scenario we are not requiring exact unification . hence , this specificity is lost and @xmath107 is a free parameter .
we proceed by simply setting @xmath108 in order to make the results of this paper more directly comparable to those of @xcite .
an example of the running of the gauge couplings from the electroweak scale to @xmath2 , as well as the values of the couplings @xmath109 and @xmath110 , is presented in figure [ fig : c ] using a phenomenologically acceptable point in the space of initial soft susy breaking parameters .
gev , @xmath111 gev and @xmath112 . in this example , @xmath113 , @xmath114 , @xmath115 , and @xmath116 . ] similarly to the gauge parameters , the yukawa couplings run differently under the rg through each of the above scaling regimes . before discussing them
, we must first decide which yukawa couplings are relevant to our analysis . as discussed in ,
we begin by inputting the experimentally determined yukawa couplings derived from the fermion masses at the electroweak scale . for the purposes of this paper , the sm yukawa couplings , which are three - by - three matrices in flavor space , can all be approximated to be zero except for the three - three elements which give mass to the third - generation sm fermions .
the experimentally determined initial conditions are y_t=0.955,y_b=0.0174,y_=0.0102 .
[ home3 ] for details on relating fermion masses to yukawa couplings , see @xcite .
we use lower case @xmath117 to denote yukawa couplings in the non - susy regime .
the one - loop rges for these yukawa couplings were presented in appendix a of , to which we refer the reader .
the yukawa couplings in eqn . can be evolved to the @xmath2 scale as follows . for the right - side - up scenario ,
the rges from @xmath99 are given in eqs .
( a5 ) - ( a7 ) .
there are non - trivial boundary conditions at the susy scale given by @xmath118 from @xmath100 , these parameters evolve as in ( a12 ) - ( a13 ) .
the boundary condition at the @xmath0 scale is trivial .
finally , from @xmath119 the @xmath0 mssm rges are given in eqs .
( a14 ) - ( a16 ) . for the upside - down case ,
one evolves from @xmath101 using ( a5 ) - ( a7 ) , as previously .
however , between @xmath102 the rges are given in eqs .
( a8 ) - ( a10 ) .
finally , above the susy scale one uses the @xmath0 mssm equations given in ( a14)-(a16 ) .
the remaining parameters of the @xmath0 mssm are the massive coefficients appearing in eqn . which are responsible for softly breaking supersymmetry .
their rges in each physical regime were presented in detail in and wo nt be discussed in this paper . here , we simply note that flavor and cp - violation experimental results place well - known limits on these quantities .
generically , the implication of these constraints are , approximately , as follows : * soft sfermion mass matrices are diagonal . *
the first two generations of squarks are degenerate in mass . *
the trilinear @xmath40-terms are diagonal . *
the gaugino masses and trilinear @xmath40-terms are real .
it is typically assumed that the soft trilinear @xmath40-terms are proportional to the yukawa couplings .
that is , @xmath120 for each fermions species .
each @xmath121 is real and associated with the susy scale .
each @xmath122 factor is a dimensionless matrix in family space .
this condition effectively makes all non - third - generation trilinear terms insignificant .
the above constraints are summarized as @xmath123 these constraints can be implemented at the scale @xmath2 , since rg evolution to the susy scale will not spoil these relations .
note that we do not assume that the first and second generation slepton masses are degenerate , unlike the squark masses , since this is not required by experiments .
the degeneracy or non - degeneracy of the first and second generation sleptons will not , however , greatly effect the results of this paper .
we now turn to the input values for the susy breaking parameters . unlike the cases of the gauge and yukawa couplings , these soft susy breaking parameters are not experimentally determined . in
, we introduced a novel way to analyze the initial parameter space of a susy model .
we will follow the same approach in the present analysis of simultaneous wilson lines . specifically
, we run a statistical scan of input parameters at the scale @xmath124 .
the randomly generated input parameters are then rg evolved to the susy scale .
we conduct an analysis of which of these high - scale initial conditions lead to realistic physics .
although the soft susy breaking lagrangian contains over 100 dimensionful parameters , the phenomenologically motivated assumptions discussed briefly above only allow significant values for 24 of them .
these , along with @xmath125 and the sign of certain parameters , are presented in the first column of table [ tbl : scan ] .
the high - scale initial values of the 24 relevant dimensionful susy breaking parameters are determined as follows .
we make the assumption that there is only one overall scale associated with susy breaking , requiring that these parameters be separated by no more than an order of magnitude , or so , from each other . to quantify this
, we demand that any dimension one soft susy breaking parameter be chosen at random within the range ( , mf ) , where @xmath126 is the overall scale of susy breaking and @xmath127 is a dimensionless number satisfying @xmath128 .
we will further insist that any such parameter be evenly scattered around @xmath126 ; that is , that @xmath126 be the average of the randomly generated values . in
, we found that in the case of split wilson line masses , the maximal number of phenomenologically acceptable `` valid '' initial points were obtained by statistically scattering within the interval defined by @xmath129 to allow direct comparison of the results of this paper to those of , we will continue to use these values in the present context .
this is shown in the second column of table [ tbl : scan ] , along with the scattering interval associated with tan@xmath130 and the allowed signs of various parameters .
.the parameters and their ranges scanned in this study .
the ranges for the soft susy breaking parameters are taken to be those of . [
cols="^,^,^",options="header " , ] the most common lsp in our main scan is the lightest neutralino , @xmath131 .
however , not all @xmath131 states are created equal .
lhc production modes for the lightest neutralino depend significantly on the composition of the neutralino a bino lsp can not be directly produced at the lhc , but the other neutralino lsp s can .
this is the basis we use for the division of these states .
the state @xmath132 designates a mostly rino or mostly blino neutralino , @xmath133 a mostly wino neutralino and @xmath134 a mostly higgsino neutralino . here
, the subscript mostly indicates the greatest contribution to that state . as an unrealistic example ,
if @xmath131 is 34% wino , 33% bino and 33% higgsino , it is still labeled @xmath133 .
the chargino lsp s are similarly separated into wino - like and higgsino - like charginos , and the stops and sbottom divisions are as in our earlier papers , references @xcite .
note that this notation for the stops , @xmath135 and @xmath136 , are only used to describe stop lsp s .
for non - lsp stops , we use the conventional notation @xmath137 and @xmath138 . to make fig .
[ fig:1039 ] more readable , we have made an effort to combine bins that have similar characteristics .
the first and second generation left - handed sneutrinos are combined into one bin , where about 50% of the lsp s are first generation sneutrinos .
the same holds true for the first and second generation right - handed sleptons , while the first generation right - handed sneutrino is always chosen to be lighter than the second generation right - handed sneutrino .
this similarity between the first and second generation sleptons is expected , since their corresponding yukawa couplings are not large enough to distinguish them through the rg evolution . for
both sleptons and squarks , more lsp s exist for the third - generation as expected from the effects of the third - generation yukawa couplings , which tend to decrease sfermion masses in the rge evolution .
the myriad of possible lsp s leads to a rich collider phenomenology .
this phenomenology is not the main focus of this paper , but it is worthwhile to briefly review it here . in models where r - parity violation is parameterized by bilinear
r - parity breaking , such as the @xmath0 mssm , susy particles are still pair produced and cascade decay to the lsp . at this point ,
the bilinear r - parity violating terms allow the lsp to decay .
while only a few studies have been done on the phenomenology of the minimal @xmath0 mssm @xcite , there have been several works on the phenomenology of explicit bilinear r - parity violation , which has some similarities to this model .
see @xcite for general discussions .
table [ tbl : lsp ] provides some basic information on the most probable decay modes of each of the possible lsp s .
note that @xmath139 signifies a charged lepton of any generation and @xmath140 a jet implying a light quark .
some interesting aspects of table [ tbl : lsp ] were discussed in . + * the non - lsp spectrum : * to get a sense of the non - lsp spectrum , we produce histograms of the masses of the sparticles associated with the valid points in the main scan . in the following histograms
, there will be quite a few pairs of fields that will be highly degenerate ; these will be represented by only one curve .
this includes @xmath141 sfermion partners , which are only split by small electroweak terms .
first generation squarks are also degenerate with second generation squarks with the same isospin , due to phenomenological constraints .
a consequence of this is that all first and second generation left - handed squarks are highly degenerate .
doublets , and the first- and second - family squarks must be degenerate , all four of these squarks have nearly identical mass and the histograms coincide .
the first- and second - family right - handed squarks are shown in the top - right panel .
the right - handed down squarks are generally lighter than their up counterparts because of the effect of the @xmath25 charge in the rges .
the third family squarks are shown in the bottom panel .
, title="fig : " ] doublets , and the first- and second - family squarks must be degenerate , all four of these squarks have nearly identical mass and the histograms coincide .
the first- and second - family right - handed squarks are shown in the top - right panel .
the right - handed down squarks are generally lighter than their up counterparts because of the effect of the @xmath25 charge in the rges .
the third family squarks are shown in the bottom panel .
, title="fig : " ] doublets , and the first- and second - family squarks must be degenerate , all four of these squarks have nearly identical mass and the histograms coincide .
the first- and second - family right - handed squarks are shown in the top - right panel .
the right - handed down squarks are generally lighter than their up counterparts because of the effect of the @xmath25 charge in the rges .
the third family squarks are shown in the bottom panel .
, title="fig : " ] is always lighter than @xmath142.,title="fig : " ] is always lighter than @xmath142.,title="fig : " ] is always lighter than @xmath142.,title="fig : " ] .
the @xmath143 and @xmath143 are typically higgsinos.,title="fig : " ] .
the @xmath143 and @xmath143 are typically higgsinos.,title="fig : " ] .
the @xmath143 and @xmath143 are typically higgsinos.,title="fig : " ] .
the @xmath143 and @xmath143 are typically higgsinos.,title="fig : " ] figure [ fig:1052 ] shows histograms of the squark masses . because they come in @xmath144 doublets and the first- and second - family squarks must be degenerate , all four of the first- and second - family left - handed squarks have nearly identical mass and the histograms coincide .
the degeneracy of first- and second - family squarks is also evident in the right - handed squark masses .
the first and second family right - handed down squarks are generally lighter than their up counterparts because of the effect of the @xmath25 charge in the rges .
figure [ fig : hist.sleptons ] shows histograms of the masses of the sneutrinos and sleptons .
the third - family sleptons and left - handed sneutrinos tend to be the lighter because of the influence of the @xmath145 yukawa coupling .
the right - handed sneutrinos are labeled such that @xmath146 is always lighter than @xmath147 .
figure [ fig:1054 ] presents histograms of the cp - even component of the third - generation right - handed sneutrino , the heavy higgses , the neutralinos , the charginos , and the gluino .
the cp - even component of the third - generation right - handed sneutrino is degenerate with @xmath44 .
it is always heavier than 2.5 tev because we have imposed the collider bound on @xmath44 .
the neutralinos and charginos are labeled from lightest to heaviest as is canonical in susy models .
the @xmath143 and @xmath148 are typically higgsinos . we emphasize that all of the above histograms are calculated using our main scan ; that is , for the choice of @xmath149 gev and @xmath150 .
we remind the reader that these values were chosen in so as to maximize the number of valid points and repeated in this paper so as to enable simple comparison with the split wilson mass results .
however , the mass scale of these histograms is heavily dependent on the choice of @xmath126 .
smaller ( larger ) values for @xmath126 will move the above distributions distinctly toward lighter ( heavier ) sparticle masses .
plots of the physical particle spectra for two valid points are presented in fig .
[ fig:156 ] .
these two points are selected from the pool of valid points from the main scan based on the simple criteria that they are the valid points with the largest right - side - up and upside - down hierarchy respectively ; that is , the largest splittings between the @xmath0 and susy scales in the two possible hierarchies
. scale is represented by a black dot - dash - dot line .
the susy scale is represented by a black dashed line .
the electroweak scale is represented by a solid black line .
the label @xmath151 is actually labeling the nearly degenerate @xmath151 and @xmath152 masses .
the labels @xmath153 , @xmath154 and @xmath155 are similarly labeling the nearly degenerate first- and second- family masses.,title="fig : " ] scale is represented by a black dot - dash - dot line .
the susy scale is represented by a black dashed line .
the electroweak scale is represented by a solid black line .
the label @xmath151 is actually labeling the nearly degenerate @xmath151 and @xmath152 masses .
the labels @xmath153 , @xmath154 and @xmath155 are similarly labeling the nearly degenerate first- and second- family masses.,title="fig : " ] plots of the high - scale boundary values for two sample valid points from our main scan are presented in fig .
[ fig:160 ] . while these look like figs .
[ fig:156 ] , they do not correspond to physical masses but , rather , mass parameters at @xmath2 .
these two valid points are selected from the pool of valid points from the main scan based on a simple criterion .
the two plots show the valid points with the largest and smallest amount of splitting in the initial values of the scalar soft mass parameters .
the amount of splitting is defined as the standard deviation of the initial values of the 20 scalar soft mass parameters .
for the two valid points with the largest and smallest amount of splitting .
the label @xmath156 is actually labeling the nearly degenerate @xmath156 and @xmath157 soft masses .
the labels @xmath158 and @xmath159 are similarly labeling the nearly degenerate first and second family masses.,title="fig : " ] for the two valid points with the largest and smallest amount of splitting .
the label @xmath156 is actually labeling the nearly degenerate @xmath156 and @xmath157 soft masses .
the labels @xmath158 and @xmath159 are similarly labeling the nearly degenerate first and second family masses.,title="fig : " ]
mssm . similarly , the green line specifies the amount of fine - tuning necessary for the valid points of the r - parity conserving mssm computed using the same statistical procedure as for the @xmath0 mssm with @xmath149 gev and @xmath150 .
the @xmath0 mssm shows slightly less fine - tuning , on average , than the mssm . ] a detailed discussion of the the little hierarchy problem , fine - tuning and the barbieri - giudice ( bg ) method of quantifying the degree of fine - tuning was presented in @xcite . here
, we simply give the results in the simultaneous wilson line scenario discussed in this paper . unlike the quantitites presented above , which can differ substantially from the split wilson line results in @xcite , the bg fine - tuning histogram for simultaneous wilson lines in the @xmath0 mssm is very similar to that of the split wilson line scenario .
be that as it may , for completeness , we present it here along with the fine - tuning histogram for the r - parity conserving mssm in fig .
[ fig : finetuning ] .
note that the highest percentage of valid points require fine - tuning of the order of 1/4,000 1/5,000 .
however , there remain a small number of points with fine - tuning less than 1/1,000 . as in the split wilson line scenario ,
the simultaneous wilson line @xmath0 mssm manifests somewhat less fine - tuning than the r - parity conserving mssm .
as discussed in the introduction and section iii , and graphically illustrated for a valid initial point in figure [ fig : c ] , the four gauge couplings of the @xmath0 mssm do not unify at @xmath2 for simultaneous wilson line masses ; that is , when @xmath160 however , as described in the introduction , the @xmath0 mssm arises on the observable orbifold plane of heterotic @xmath126-theory compactified on a schoen calabi - yau threefold with @xmath6 and a holomorphic vector bundle with @xmath161 structure group .
that is , the @xmath0 mssm is a low energy effective theory of heterotic string theory .
hence , as discussed in numerous papers @xcite , it is expected that at _ string tree level _ all four gauge couplings , along with the dimensionless gravitational parameter @xmath162 where @xmath163 is newton s constant and @xmath164 is the string regge slope , unify to a single parameter @xmath165 at a `` string unification '' scale @xmath166 the string coupling parameter @xmath165 is set by the value of the dilaton , and is typically of @xmath167 .
a common value in the literature , see for example @xcite , is @xmath168 which , for specificity , we will use henceforth .
therefore , we take @xmath169 and the string unification scale to be @xmath170 respectively .
note that @xmath171 is approximately an order of magnitude larger than @xmath2 .
below @xmath172 however , the couplings begin to evolve according to the rges of effective field theory .
this adds another
fourth scaling regime to the three discussed at the beginning of section iv .
this new regime is * @xmath172 @xmath2 : the effective field theory in this regime remains that of the @xmath0 mssm with the couplings @xmath173 , @xmath174 and the slope factors @xmath175 as in eqn . .
however , the rges are now altered to become on the right - hand side , where @xmath176 , @xmath177 are the associated string affine levels . however , these are all unity for the scaled gauge couplings of the @xmath0 mssm . ]
@xmath178 note that the one - loop running couplings no longer unify exactly at @xmath172 .
rather , they are `` split '' by dimensionless threshold effects . these arise predominantly from massive genus - one string modes that contribute to the correlation function @xmath179 and , hence , to the @xmath173 gauge couplings .
this is depicted graphically in figure [ fig : rehan ] . on
the genus - one string worldsheet is a typical example of heavy string threshold correction terms that need to be calculated . ]
recall that in this paper we have found @xmath180 valid initial points in the space of soft supersymmetry breaking dimensionful couplings each of which satisfies all low energy phenomenological criteria . for _ each _ of these points
, we can calculate by scaling from the electroweak scale to @xmath2 the four gauge couplings @xmath109 and @xmath110 .
note that in this analysis , we have defined an `` average '' susy scale @xmath181 , the @xmath0 breaking scale @xmath73 , as well as an `` average '' unification scale @xmath2 , in ( 28 ) , ( 24 ) and ( 22 ) respectively .
the rges have been scaled through the requisite intermediate regimes with the appropriate beta - function coefficients .
that is , we have already taken into account the predominant threshold effects associated with each of these scales .
for statistically `` average '' valid initial points the vast majority of the phenomenologically acceptable initial soft susy breaking parameters
possible additional threshold effects arising from the `` splitting '' of particle masses around these scales are expected to be relatively small and will be systematically ignored relative to the heavy string thresholds .
that is , the @xmath182 , @xmath183 parameters in ( 50 ) will closely approximate the four heavy string gauge thresholds . with this input , using eqn . , @xmath184gev and @xmath169 , @xmath172 given in , one can calculate the associated heavy string thresholds from ; that is , @xmath185 for each @xmath174 .
of course , these thresholds are expected to differ for each different valid initial point .
it follows that one should analyze the thresholds statistically graphing the dispersion of each as one runs over the @xmath180 valid initial points .
the histograms associated with each of these four thresholds are presented in figure [ fig : thresh1 ] . , @xmath186 arising from the @xmath180 phenomenologically valid points of our statistical survey . each threshold value is plotted against the percentage of valid points giving rise to it .
the bin width is 0.1.,title="fig : " ] , @xmath186 arising from the @xmath180 phenomenologically valid points of our statistical survey .
each threshold value is plotted against the percentage of valid points giving rise to it .
the bin width is 0.1.,title="fig : " ] , @xmath186 arising from the @xmath180 phenomenologically valid points of our statistical survey .
each threshold value is plotted against the percentage of valid points giving rise to it .
the bin width is 0.1.,title="fig : " ] , @xmath186 arising from the @xmath180 phenomenologically valid points of our statistical survey .
each threshold value is plotted against the percentage of valid points giving rise to it .
the bin width is 0.1.,title="fig : " ] to better understand the relationship of these different thresholds , we find it useful to plot all four of them in a single histogram .
this is presented in figure [ fig : thresh2 ] .
combined into a single graph to elucidate their relative occurrence and values . ]
it is also useful to calculate the string threshold associated with the abelian hypercharge coupling @xmath103 defined , using and , by mssm gauge couplings , this scaled hypercharge coupling has string affine level @xmath187 . ]
@xmath188 the associated statistical histogram is given in figure [ fig : thresh3 ] .
arising from the @xmath180 phenomenologically valid points of our statistical survey .
each threshold value is plotted against the percentage of valid points giving rise to it .
the bin width is 0.1 . ]
it is well - known @xcite that each string threshold breaks into two parts , @xmath189 where @xmath190 is a `` universal '' piece independent of the gauge group and @xmath191 records the contributions of all massive string states as they propagate around the genus - one string worldsheet torus diagram shown in figure [ fig : rehan ] .
note again that all string affine levels are unity in our normalization . as discussed in @xcite , an explicit calculation of the universal piece @xmath192
is difficult due to the presence of infrared divergences .
however , the @xmath191 threshold terms , although moduli dependent , can be directly calculated from string theory using a formulation given by v. kaplunovsky in @xcite and by kaplunovsky and louis in @xcite .
such calculations are heavily model dependent @xcite and , to date , have not been carried out in the @xmath0 mssm context .
be that as it may , it is useful to present our experimental predictions for @xmath193 , @xmath194 , and @xmath195 from which information about @xmath196 , @xmath197 and @xmath198 can be inferred .
the statistical results for these three quantities are presented in figure [ fig : thresh4 ] . , @xmath194 , and @xmath195 .
the third of these plots looks different because the quantity @xmath199 falls in a very narrow range .
the bin width in all three plots is 0.1.,title="fig : " ] , @xmath194 , and @xmath195 .
the third of these plots looks different because the quantity @xmath199 falls in a very narrow range .
the bin width in all three plots is 0.1.,title="fig : " ] , @xmath194 , and @xmath195 .
the third of these plots looks different because the quantity @xmath199 falls in a very narrow range .
the bin width in all three plots is 0.1.,title="fig : " ] it would be very interesting to compare these results to direct calculations using @xcite in the @xmath0 mssm context .
we will not attempt that in the present paper .
our previous work on the @xmath0 mssm used the constraint of exact gauge unification .
that is , we chose @xmath200 and , by appropriately separating the wilson line masses , forced the gauge couplings to unify exactly at the scale @xmath65 .
this hypothesis enhanced the specificity of the calculation .
for example , it set all gauge couplings to a unified value @xmath66 at @xmath9 and determined @xmath107 at the susy transition mass .
however , the splitting of the wilson lines limits the analysis to a restricted region of cy moduli space where the associated two - cycles have considerably different radii .
furthermore , it introduces a new scaling regime between @xmath201 and @xmath202in our example a `` left - right '' model with a specific spectrum . for soft susy breaking masses in the tev range , this constraint was reasonable since the separation between the wilson line masses and , hence , the difference in the two - cycle radii was less than an order of magnitude .
however , if one tries to take larger values for the soft susy breaking masses , the difference in the wilson line masses grows rapidly .
for example ,
for @xmath203 tev soft masses , the wilson lines must be separated by a factor of @xmath204reducing the calculation to an extremely unnatural region of cy moduli space . for even larger values of the soft masses ,
the calculation breaks down completely .
it follows that if one wishes to discuss the @xmath0 mssm for large soft susy breaking masses see our discussion below then it becomes necessary to analyze the theory for the more natural case of simultaneous , or nearly simultaneous , wilson lines .
hence , in this paper , we have carried out the analysis of the @xmath0 mssm under the more natural hypothesis of equal wilson line scales ; that is , we assumed that @xmath205 .
although this hypothesis does not allow exact gauge unification , it is well - known see the previous section that string threshold corrections can be responsible for such non - unification , providing a consistent theoretical framework for this analysis .
our results indicate that a substantial region of the initial soft susy breaking parameter space is consistent with all low - energy experimental data see figs .
[ fig:1204]-[fig:1148 ] .
we also presented our results for important physical quantities ; histograms of the lsp species see fig .
[ fig:1039]histograms of the sparticle masses figs .
[ fig:1052]-[fig:1054]plots of sample mass spectra at the susy scale see fig .
[ fig:156]and plots of sample initial mass data at the average unification scale fig .
[ fig:160 ] .
although different in some details from the previous exact gauge unification analysis , the results presented in this paper share many common features . in addition , they show that all low - energy phenomenological constraints can be satisfied for a wide range of initial soft susy breaking parameters over a more natural and generic region of cy moduli space .
the fundamental difference between the calculations in this paper from those of our previous work is that the gauge couplings of the @xmath0 mssm can no longer unify at an average mass associated with the inverse cy radius .
rather , the gauge couplings remain split at that scale a splitting that allows one to compute the heavy string threshold corrections due to heavy states on the genus - one string worldsheet .
these string threshold corrections and their differences are analyzed statistically , and predict what one should obtain using a a more direct worldsheet formalism .
it would be interesting to see the results of such formal string calculations done in the heterotic @xmath0 mssm context . finally , having achieved these more generic results , we are now in a position to arbitrarily raise the mass scale of the soft susy breaking parameters without encountering the above mentioned difficulties . for example
, it is now possible to raise the soft susy breaking scale up to @xmath206 or @xmath207 gev values consistent with the mass scale in `` split supersymmetry '' theories @xcite .
as we will show in a subsequent publication , this freedom will allows us to discuss a new theory of inflation within the context of the @xmath0 mssm .
this work will appear elsewhere @xcite .
we would like to thank sogee spinner for helpful conversations .
r. deen and b.a .
ovrut are supported in part by the doe under contract no .
de - sc0007901 .
a. purves was supported by doe contract no .
de - sc0007901 when much of this work was carried out .
a. lukas , b. a. ovrut and d. waldram , `` on the four - dimensional effective action of strongly coupled heterotic string theory , '' nucl .
b * 532 * , 43 ( 1998 ) doi:10.1016/s0550 - 3213(98)00463 - 5 [ hep - th/9710208 ] .
a. font , l. e. ibanez and f. quevedo , `` does proton stability imply the existence of an extra z0 ?
, '' phys .
b * 228 * , 79 ( 1989 ) .
s. p. martin , `` a supersymmetry primer , '' adv .
direct . high energy phys . * 21 * , 1 ( 2010 ) [ hep - ph/9709356 ] .
s. p. martin , `` some simple criteria for gauged r - parity , '' phys . rev .
d * 46 * , 2769 ( 1992 ) [ hep - ph/9207218 ] .
l. m. krauss and f. wilczek , `` discrete gauge symmetry in continuum theories , '' phys .
lett . * 62 * , 1221 ( 1989 ) . c. s. aulakh , a. melfo , a. rasin and g. senjanovic , `` seesaw and supersymmetry or exact r - parity , '' phys .
b * 459 * ( 1999 ) 557 [ hep - ph/9902409 ] . c. s. aulakh , b. bajc , a. melfo , a. rasin and g. senjanovic , `` so(10 ) theory of r - parity and neutrino mass , '' nucl .
b * 597 * ( 2001 ) 89 [ hep - ph/0004031 ] .
k. s. babu and r. n. mohapatra , `` minimal supersymmetric left - right model , '' phys .
b * 668 * , 404 ( 2008 ) [ arxiv:0807.0481 [ hep - ph ] ] .
d. feldman , p. fileviez perez and p. nath , `` r - parity conservation via the stueckelberg mechanism : lhc and dark matter signals , '' arxiv:1109.2901 [ hep - ph ] .
p. fileviez perez and m. b. wise ,
`` low energy supersymmetry with baryon and lepton number gauged , '' phys . rev .
d * 84 * , 055015 ( 2011 ) [ arxiv:1105.3190 [ hep - ph ] ] .
c. s. aulakh and r. n. mohapatra , `` neutrino as the supersymmetric partner of the majoron , '' phys .
b * 119 * , 136 ( 1982 ) .
m. j. hayashi and a. murayama , `` radiative breaking of @xmath208 gauge symmetry induced by broken @xmath209 supergravity in a left - right symmetric model , '' phys .
b * 153 * , 251 ( 1985 ) .
r. n. mohapatra , `` mechanism for understanding small neutrino mass in superstring theories , '' phys .
* 56 * ( 1986 ) 561 .
a. masiero and j. w. f. valle , `` a model for spontaneous r parity breaking , '' phys .
b * 251 * ( 1990 ) 273 .
v. braun , y. h. he , b. a. ovrut and t. pantev , `` vector bundle extensions , sheaf cohomology , and the heterotic standard model , '' adv .
* 10 * , no .
4 , 525 ( 2006 ) doi:10.4310/atmp.2006.v10.n4.a3 [ hep - th/0505041 ] .
m. ambroso and b. a. ovrut , `` the mass spectra , hierarchy and cosmology of b - l mssm heterotic compactifications , '' int .
j. mod .
a * 26 * , 1569 ( 2011 ) doi:10.1142/s0217751x11052943 [ arxiv:1005.5392 [ hep - th ] ] .
g. aad _ et al . _ [ atlas collaboration ] , `` measurement of the higgs boson mass from the @xmath210 and @xmath211 channels with the atlas detector using 25 fb@xmath212 of @xmath213 collision data , '' phys .
d * 90 * , 052004 ( 2014 ) [ arxiv:1406.3827 [ hep - ex ] ] .
s. chatrchyan
_ et al . _
[ cms collaboration ] , `` observation of a new boson with mass near 125 gev in pp collisions at @xmath214 = 7 and 8 tev , '' jhep * 1306 * , 081 ( 2013 ) [ arxiv:1303.4571 [ hep - ex ] ] .
b. a. ovrut , a. purves and s. spinner , `` wilson lines and a canonical basis of su(4 ) heterotic standard models , '' jhep * 1211 * , 026 ( 2012 ) doi:10.1007/jhep11(2012)026 [ arxiv:1203.1325 [ hep - th ] ] .
z. marshall , b. a. ovrut , a. purves and s. spinner , `` spontaneous r - parity breaking , stop lsp decays and the neutrino mass hierarchy , '' phys .
b * 732 * , 325 ( 2014 ) doi:10.1016/j.physletb.2014.03.052 [ arxiv:1401.7989 [ hep - ph ] ] .
z. marshall , b. a. ovrut , a. purves and s. spinner , `` lsp squark decays at the lhc and the neutrino mass hierarchy , '' phys .
d * 90 * , no . 1
, 015034 ( 2014 ) doi:10.1103/physrevd.90.015034 [ arxiv:1402.5434 [ hep - ph ] ] .
b. a. ovrut , a. purves and s. spinner , `` a statistical analysis of the minimal susy bl theory , '' mod .
a * 30 * , no .
18 , 1550085 ( 2015 ) doi:10.1142/s0217732315500856 [ arxiv:1412.6103 [ hep - ph ] ] .
k. r. dienes , a. e. faraggi and j. march - russell , `` string unification , higher level gauge symmetries , and exotic hypercharge normalizations , '' nucl .
b * 467 * , 44 ( 1996 ) doi:10.1016/0550 - 3213(96)00085 - 5 [ hep - th/9510223 ] .
e. kiritsis , c. kounnas , p. m. petropoulos and j. rizos , `` universality properties of n=2 and n=1 heterotic threshold corrections , '' nucl .
b * 483 * , 141 ( 1997 ) doi:10.1016/s0550 - 3213(96)00550 - 0 [ hep - th/9608034 ] .
h. p. nilles and s. stieberger , `` string unification , universal one loop corrections and strongly coupled heterotic string theory , '' nucl .
b * 499 * , 3 ( 1997 ) doi:10.1016/s0550 - 3213(97)00315 - 5 [ hep - th/9702110 ] .
v. barger , p. fileviez perez and s. spinner , `` minimal gauged
u(1)(b - l ) model with spontaneous r - parity violation , '' phys .
* 102 * , 181802 ( 2009 ) doi:10.1103/physrevlett.102.181802 [ arxiv:0812.3661 [ hep - ph ] ] .
l. l. everett , p. fileviez perez and s. spinner , `` the right side of tev scale spontaneous r - parity violation , '' phys .
d * 80 * , 055007 ( 2009 ) doi:10.1103/physrevd.80.055007 [ arxiv:0906.4095 [ hep - ph ] ] .
d. k. ghosh , g. senjanovic and y. zhang , `` naturally light sterile neutrinos from theory of r - parity , '' phys .
b * 698 * ( 2011 ) 420 [ arxiv:1010.3968 [ hep - ph ] ] .
v. barger , p. fileviez perez and s. spinner , `` three layers of neutrinos , '' phys .
b * 696 * ( 2011 ) 509 [ arxiv:1010.4023 [ hep - ph ] ] .
b. mukhopadhyaya , s. roy and f. vissani , `` correlation between neutrino oscillations and collider signals of supersymmetry in an r - parity violating model , '' phys
. lett .
b * 443 * , 191 ( 1998 ) [ hep - ph/9808265 ] .
e. j. chun and j. s. lee , `` implication of super - kamiokande data on r - parity violation , '' phys .
d * 60 * , 075006 ( 1999 ) [ hep - ph/9811201 ] .
e. j. chun and s. k. kang , `` one loop corrected neutrino masses and mixing in supersymmetric standard model without r - parity , '' phys .
d * 61 * , 075012 ( 2000 ) [ hep - ph/9909429 ] .
m. hirsch , m. a. diaz , w. porod , j. c. romao and j. w. f. valle , `` neutrino masses and mixings from supersymmetry with bilinear r parity violation : a theory for solar and atmospheric neutrino oscillations , '' phys . rev .
d * 62 * , 113008 ( 2000 ) [ erratum - ibid .
d * 65 * , 119901 ( 2002 ) ] [ hep - ph/0004115 ] .
p. fileviez perez and s. spinner , `` the minimal theory for r - parity violation at the lhc , '' arxiv:1201.5923 [ hep - ph ] .
p. fileviez perez and s. spinner , `` supersymmetry at the lhc and the theory of r - parity , '' phys .
b * 728 * , 489 ( 2014 ) [ arxiv:1308.0524 [ hep - ph ] ] .
g. gamberini , g. ridolfi and f. zwirner , `` on radiative gauge symmetry breaking in the minimal supersymmetric model , '' nucl .
b * 331 * , 331 ( 1990 ) .
olive et al .
( particle data group ) , chin .
c , 38 , 090001 ( 2014 ) . a. djouadi , `` the anatomy of electro - weak symmetry breaking .
i : the higgs boson in the standard model , '' phys .
* 457 * , 1 ( 2008 ) [ hep - ph/0503172 ] .
w. porod , m. hirsch , j. romao and j. w. f. valle , `` testing neutrino mixing at future collider experiments , '' phys .
d * 63 * , 115004 ( 2001 ) [ hep - ph/0011248 ] .
m. hirsch and w. porod , `` neutrino properties and the decay of the lightest supersymmetric particle , '' phys .
d * 68 * , 115007 ( 2003 ) [ hep - ph/0307364 ] .
p. w. graham , d. e. kaplan , s. rajendran and p. saraswat , `` displaced supersymmetry , '' jhep * 1207 * , 149 ( 2012 ) [ arxiv:1204.6038 [ hep - ph ] ] .
n. arkani - hamed and s. dimopoulos , `` supersymmetric unification without low energy supersymmetry and signatures for fine - tuning at the lhc , '' jhep * 0506 * , 073 ( 2005 ) doi:10.1088/1126 - 6708/2005/06/073 [ hep - th/0405159 ] .
g. f. giudice and a. romanino , `` split supersymmetry , '' nucl .
b * 699 * , 65 ( 2004 ) erratum : [ nucl .
b * 706 * , 487 ( 2005 ) ] doi:10.1016/j.nuclphysb.2004.11.048 , 10.1016/j.nuclphysb.2004.08.001 [ hep - ph/0406088 ] . | + [ 10 mm ] * abstract : * in previous work , we presented a statistical scan over the soft supersymmetry breaking parameters of the minimal susy @xmath0 model .
for specificity of calculation , unification of the gauge parameters was enforced by allowing the two @xmath1 wilson lines to have mass scales separated by approximately an order of magnitude .
this introduced an additional `` left - right '' sector below the unification scale . in this paper , for three important reasons , we modify our previous analysis by demanding that the mass scales of the two wilson lines be simultaneous and equal to an `` average unification '' mass @xmath2 .
the present analysis is 1 ) more `` natural '' than the previous calculations , which were only valid in a very specific region of the calabi - yau moduli space , 2 ) the theory is conceptually simpler in that the left - right sector has been removed and 3 ) in the present analysis the lack of gauge unification is due to threshold effects particularly heavy string thresholds , which we calculate statistically in detail . as in our previous work ,
the theory is renormalization group evolved from @xmath2 to the electroweak scale being subjected , sequentially , to the requirement of radiative @xmath0 and electroweak symmetry breaking , the present experimental lower bounds on the @xmath0 vector boson and sparticle masses , as well as the lightest neutral higgs mass of @xmath3125 gev .
the subspace of soft supersymmetry breaking masses that satisfies all such constraints is presented and shown to be substantial . |
quantum entanglement is a key ingredient in quantum information processing@xcite . in practice
, entanglement can be easily destroyed by environment noise@xcite .
quantum illumination(qi ) can benefit from entanglement in target detection even it is under these entanglement - destroying noise@xcite .
the aim of quantum illumination is to detect low - reflective target which is embedded in a bright background thermal bath .
half of a pair of entangled optical beams is sent out to interrogate the target region . then the returned and the retained signal beam are used to decide the presence or absence of the object . even though the fragile entanglement is easily destroyed by the bright thermal noise
, quantum - illumination protocol still has remarkable advantage over classical probe protocol@xcite .
several experiments have realized the practical protocol in quantum sensing@xcite .
recently , weedbrook et al.@xcite show that quantum discord exhibits the role of preserving the benefits of quantum illumination while entanglement is broken .
it is advantageous to operate the frequency of the signal which interrogates the object in microwave region@xcite .
but so far there is no efficient way to detect single photon in microwave region while in the visible light frequency region , ultrasensitive detection of single - photon have been achieved .
such that the detection of microwave signals via the detection of their entangled optical signals is a more efficient way@xcite .
optomechanical systems could be a good candidate to realize the entanglement of microwave and optical fields by using mechanical motion@xcite .
however , the existing proposal is supposed to work with rotating wave approximation(rwa ) which requests weak many - photon couplings of optomechanical systems@xcite .
in addition , it is extremely difficult to achieve strong coupling in optical regime for a system to generate entanglement between optical and microwave modes .
moreover , plenty of photons in cavities can induce bistability which may give rise to experimental difficulties@xcite .
therefore , it is meaningful to study the generation of entanglement in weak interaction regime . also , the existing study is limited to the situation that the output signals are monochromatic .
a study for the more practical case of finite bandwidth is needed . in this paper
, we propose a novel realization of quantum illumination in weak coupling regime based on multimode optomechanical systems .
the transmitter is an optomechanical system consisting of a two - mode microwave cavity coupling with a two - mode optical cavity via a mechanical resonator . in this system ,
the output microwave signal and the optical signal of the two cavities are entangled .
the receiver is a similar optomechanical system which converts the reflected microwave signal into optical signal .
the retained optical mode of the transmitter and the optical mode of the receiver are then sent to the photon - detectors to make a joint measurement .
we use a two - mode and off - resonant(with frequency detuning @xmath0 ) process to minimize the mechanical thermal noise .
consequently , our method can achieve a significant reduction of error probability than classical system of the same transmitted energy .
our method works in the weak many - photon coupling regime where rotating - wave approximation works well .
moreover , we optimize the delay time of the microwave signal s filter function .
this makes it possible for our method to work with a finite bandwidth of signals . in this case
, we can still achieve an improvement of signal - to - noise ratio by @xmath1 than that of classical detection given that signals - bandwidth @xmath2 which equals cavities - bandwidth @xmath3 .
the paper is organized as follows . in sec .
ii we describe the system and derive the quantum langevin equations for the hamiltonian of our proposed system . in sec .
iii we calculate the error probability of the proposed qi system , give an analytical expression for the delay time and analyze the limited bandwidth situation .
finally , sec .
iv summarizes the main conclusions of this work .
+ as illustrated in fig .
[ fig1 ] ( a ) , we consider an optomechanical system including a single - mode mechanical resonator coupled to both a two - mode microwave cavity and a two - mode optical cavity .
the two microwave modes are at frequencies @xmath4 , @xmath5 and two optical modes are at frequencies @xmath6 , @xmath7 respectively .
the mechanical mode is at frequency @xmath8 .
cavity modes @xmath9 and @xmath10 are resonantly driven by lasers .
the hamiltonian of the system is@xcite @xmath11 where @xmath12 , @xmath13 and @xmath14 are the intensities and phases of the driving lasers respectively . given the frequency mismatch fulfilling @xmath15 , under the resolved - sideband approximation and rotated - wave approximation , the hamiltonian can be linearized as@xcite @xmath16 where @xmath17 , @xmath18 , @xmath19 and @xmath3 is the cavity decay rate .
coupling terms @xmath20 and @xmath21 can be chosen to be real and positive by adjusting the laser phases @xmath22 and @xmath23 respectively .
the hamiltonian shows that the emitted photons from modes @xmath24 and @xmath25 could be entangled via intermediate phonon mode @xmath26 . taking cavity dissipation @xmath3 and mechanical dissipation @xmath27 into consideration , the langevin equations of eq .
( [ ham1 ] ) are@xcite @xmath28 the equations can be solved after fourier transformations where @xmath29=\int_{-\infty}^{\infty}\hat{o}(t)e^{i\omega t}dt$](with @xmath30 ) . applying the input - output relations @xmath31(with @xmath32)@xcite
, we obtain @xmath33=&a_{+}\hat{a}_{+,in}[\omega]-b\hat{a}_{-,in}^\dagger[-\omega]+c_{+}\hat{b}_{in}[\omega],\\ \hat{a}_{-}[-\omega]=&b^{*}\hat{a}_{+,in}^\dagger[\omega]+a_{-}\hat{a}_{-,in}[-\omega]+c_{-}\hat{b}_{in}^\dagger[\omega ] .
\end{aligned}\end{aligned}\ ] ] the correlation functions of input mode satisfy @xmath34 where @xmath35 , @xmath36 and @xmath37 are thermal photon numbers of the two cavity modes and phonon thermal number of mechanical mode respectively . we obtain @xmath38&=\langle\hat{a}_{+,out}^\dagger[\omega]\hat{a}_{+,out}[\omega^\prime]\rangle = 2\pi[4n_{b}^{t}\frac{g_1 ^ 2\gamma\kappa}{|g_1 ^ 2-g_2 ^ 2+[\gamma + 2 i ( \delta -\omega ) ] ( \kappa -2 i \omega ) |^2}\nonumber\\&+4\frac{(g_1g_2\kappa)^2 } { |[g_1 ^ 2-g_2 ^ 2+(\gamma + 2 i ( \delta -\omega ) ) ( \kappa -2 i \omega ) ] ( \kappa -2 i \omega ) |^2}]\delta(\omega-\omega^\prime),\\ n_{-}[-\omega,-\omega^\prime]&=\langle\hat{a}_{-,out}^\dagger[-\omega]\hat{a}_{-,out}[-\omega^\prime]\rangle = 2\pi[4(1+n_{b}^{t})\frac{g_2 ^ 2\gamma\kappa}{|g_1 ^ 2-g_2 ^ 2+[\gamma + 2 i ( \delta -\omega ) ] ( \kappa -2 i \omega ) |^2}\nonumber\\&+4\frac{(g_1g_2\kappa)^2 } { |[g_1 ^ 2-g_2 ^ 2+(\gamma + 2 i ( \delta -\omega ) ) ( \kappa -2 i \omega ) ] ( \kappa -2 i \omega ) |^2}]\delta(\omega-\omega^\prime),\\ x[\omega,-\omega^\prime]&=\langle\hat{a}_{+,out}[\omega]\hat{a}_{-,out}[-\omega^\prime]\rangle=2\pi\delta(\omega-\omega^\prime ) [ -2g_1g_2\kappa\nonumber\\ & \frac{g_1 ^ 2(\kappa-2i\omega ) + ( \kappa + 2i\omega)(g_2 ^ 2-(\gamma + 2i(\delta -\omega ) ) ( \kappa -2i\omega))+2 ( 1+n_{b}^{t})\gamma(\kappa ^2 + 4\omega ^2)}{(g_1 ^ 2-g_2 ^ 2+(\gamma + 2 i
( \delta -\omega ) ) ( \kappa -2 i \omega ) ) ( g_1 ^ 2-g_2 ^ 2+(\gamma -2 i(\delta -\omega ) ) ( \kappa + 2 i \omega ) ) ( \kappa ^2 + 4 \omega ^2 ) } ] . \ ] ] here we assume that @xmath39 . in practice , in order to minimize the error probability of output signals , an increase of repeat rate is needed .
therefore , we need to consider output signals within certain frequency bandwidth . after projecting output signals onto wave - packet modes , @xmath40 . here
@xmath41 are filter functions which satisfy @xmath42 .
we use logarithmic negativity to measure the entanglement of two cavity modes .
the logarithmic negativity is given by@xcite @xmath43 where @xmath44 is the smallest partial transposed symplectic eigenvalue of matrix @xmath45 .
the components of the covariance matrix have the form @xmath46 where @xmath47^{t}$ ] with @xmath48 and @xmath49 .
thus we have @xmath50 where @xmath51|^2n_{+}[\omega]d\omega,\nonumber\\ v_{33}&=&\langle\hat{a}_{-,f}^\dagger\hat{a}_{-,f}\rangle=\int_{-\infty}^{\infty}|f_{-}[\omega]|^2n_{-}[\omega]d\omega,\nonumber\\ v_{13}&=&\mathrm{re}(\langle\hat{a}_{+,f}\hat{a}_{-,f}\rangle)=\mathrm{re}(\int_{-\infty}^{\infty}f_{+}^{*}[\omega]f_{-}^{*}[-\omega]x[\omega]d\omega),\nonumber\\ v_{14}&=&\mathrm{im}(\langle\hat{a}_{+,f}\hat{a}_{-,f}\rangle)=\mathrm{im}(\int_{-\infty}^{\infty}f_{+}^{*}[\omega]f_{-}^{*}[-\omega]x[\omega]d\omega),\nonumber\end{aligned}\ ] ] with @xmath52=\frac{1}{2\pi}\int_{-\infty}^{\infty}n_{+}[\omega,\omega^\prime]d\omega^\prime,\nonumber\\ n_{-}[\omega]=\frac{1}{2\pi}\int_{-\infty}^{\infty}n_{-}[\omega,\omega^\prime]d\omega^\prime,\nonumber\\ x[\omega]=\frac{1}{2\pi}\int_{-\infty}^{\infty}x[\omega,-\omega^\prime]d\omega^\prime.\nonumber\end{aligned}\ ] ] $ ] around @xmath53 and around @xmath54 where @xmath55 , @xmath56 , @xmath57 , @xmath58(@xmath59 , @xmath60 ) , @xmath61 , @xmath62 .
( b ) the spectrum of logarithmic - negativity @xmath63 .
( c ) the spectrum of logarithmic - negativity over output photon number @xmath64.,scaledwidth=45.0% ] we draw the spectrum of @xmath65 $ ] , @xmath66 $ ] and @xmath66/n_{+}[\omega]$ ] in fig . [ en ] . as shown in fig .
[ en](a ) , output photon number @xmath65 $ ] has two peaks .
one is at @xmath67 with width @xmath68 and the other is at @xmath69 with width @xmath70 .
when @xmath0 approaches @xmath71 , the two peaks emerge together . in this case ,
@xmath72\gg1 $ ] and the maximum entanglement appears at @xmath53 , and the ratio(@xmath73 ) of the entanglement to the photon number goes to the minimum .
this is the reason why the optimum ratio @xmath73 in the previous literature ref .
@xcite appears at @xmath74 but not at the optimum entanglement point where parameter @xmath75 . in our case where @xmath76 , even though the ratio @xmath73 is the second minimum valley at around @xmath53 , it still have a considerable value .
the input - output relation for the receiver is @xmath77=b^{*}\hat{a}_{+r , in}^{\dagger}[\omega]+a_{-}\hat{a}_{-,in}^\prime[-\omega]+c_{-}\hat{b}_{in}^{\prime\dagger}[\omega],\end{aligned}\ ] ] where @xmath78 is the signal reflected from interrogated target region .
we have @xmath79 when there is no low reflective object in the target region(hypothesis @xmath80 ) and @xmath81 when there contains low reflective object in the target region(hypothesis @xmath82)@xcite . here
@xmath83 is the background thermal mode in the interrogated region .
the mean photon number is @xmath84 or @xmath85 providing the absence or presence of low reflective object .
we measure the photon numbers of mixed modes @xmath86 .
the object s absence or presence is determined by the photon number difference which is @xmath87 where @xmath88 and @xmath89 is the independent , identically distributed ( iid ) signal - idler - mode pairs .
thus the error probability of receiver is given by@xcite @xmath90 where the signal - to - noise ratio satisfies @xmath91 ^ 2 } { ( \sqrt{(\langle\delta\hat{n}_{c,+}-\delta\hat{n}_{c,-}\rangle_{h_0})^2}+\sqrt{(\langle\delta\hat{n}_{c,+ } -\delta\hat{n}_{c,-}\rangle_{h_1})^2})^2},\label{snr}\end{aligned}\ ] ] where @xmath92,\label{12b}\\ ( \langle\delta\hat{n}_{c,+}-\delta\hat{n}_{c,-}\rangle_{h_0})^2&=\langle\hat{n}_{c,+}\rangle_{h_0}(\langle\hat{n}_{c,+}\rangle_{h_0 } + 1)+\langle\hat{n}_{c,-}\rangle_{h_0}(\langle\hat{n}_{c,-}\rangle_{h_0 } + 1 ) - \frac{(\langle\hat{a}_{-\eta , f}^\dagger\hat{a}_{-\eta , f}\rangle_{h_0 } -\langle\hat{a}_{-,f}^\dagger\hat{a}_{-,f}\rangle)^2}{2},\\ ( \langle\delta\hat{n}_{c,+}-\delta\hat{n}_{c,-}\rangle_{h_1})^2&=\langle\hat{n}_{c,+}\rangle_{h_1}(\langle\hat{n}_{c,+}\rangle_{h_1 } + 1)+\langle\hat{n}_{c,-}\rangle_{h_1}(\langle\hat{n}_{c,-}\rangle_{h_1 } + 1 ) - \frac{(\langle\hat{a}_{-\eta , f}^\dagger\hat{a}_{-\eta , f}\rangle_{h_1 } -\langle\hat{a}_{-,f}^\dagger\hat{a}_{-,f}\rangle)^2}{2}\nonumber\\ & -2(\sqrt{\eta}\mathrm{im}[\langle b\hat{a}_{+,f}\hat{a}_{-,f}\rangle])^2.\end{aligned}\ ] ] in fig .
[ p ] ( a ) , we plot @xmath93 with cooperative parameters @xmath94 and @xmath95 in resonant frequency @xmath53 . here @xmath96 is the signal - to - noise ratio of classical microwave transmitter which is equivalent to @xmath97@xcite . from the figure
, we can see that the maximum @xmath98 appears at @xmath99 and a relative large @xmath98 could be obtained with weak many - photon optomechanical coupling where @xmath100 .
photon number @xmath101 $ ] is shown in fig .
[ p](b ) with @xmath102 and @xmath103 . versus @xmath102 and @xmath103 .
parameter settings are @xmath56 , @xmath57 , @xmath58 , @xmath61 , @xmath62 , @xmath104 and @xmath105(room temperature @xmath106 ) .
( b ) photon number @xmath72 $ ] versus cooperative parameters @xmath102 and @xmath103 . for @xmath107 and @xmath108 the stability condition could be @xmath109@xcite.,title="fig:",scaledwidth=50.0% ] + let us now discuss the situation where the output signals have a bandwidth @xmath2 , i.e. , filter functions @xmath110 $ ] and @xmath111 $ ] have a bandwidth @xmath2 .
the filter functions have the form @xmath110=e^{i\omega t_d}/\sqrt{\sigma}$ ] , @xmath111=1/\sqrt{\sigma}$ ] for @xmath112 $ ] and @xmath113=0 $ ] for @xmath114 or @xmath115 . here
$ ] has a time - delay term @xmath116 as discussed in ref .
@xcite . to find optimal @xmath117 at which the signal - to - noise ratio decreases slowly with the bandwidth
, we revisit eq .
( [ 12b ] ) .
@xmath118=\frac{1}{\sigma}\int_{-\frac{\sigma}{2}}^{\frac{\sigma}{2}}\nonumber\\ & & [ e^{-i\omega t_d}|x'[\omega]|e^{i(\varphi(0)+\varphi^{\prime}(0)\omega+ ... )}],\end{aligned}\ ] ] where @xmath119=[a_{+}|b|^2+c_{+}c_{-}b(n_b^t+1)]$ ] and @xmath120 is the phase of @xmath119 $ ] .
the elimination of integrand s phase dependence on @xmath121 would improve the value of signal - to - noise ratio .
if bandwidth @xmath2 is small , then @xmath122 could be the optimal delay time which reads @xmath123,\end{aligned}\ ] ] where we assume that @xmath124 and @xmath125 . in fig .
[ width ] ( a ) , we show @xmath126 with bandwidth @xmath2 for delay time @xmath127 and @xmath128 . it can be easily seen that the signal - to - noise ratio indeed decreases slowly with @xmath2 increases for @xmath128 compared to that for @xmath127 .
figure of merit @xmath98 with @xmath2 is shown in fig .
[ width ] ( b ) .
the advantage of qi over classical system decreases rapidly for @xmath127 while qi always has advantage over classical ones even for relative large @xmath2 for @xmath128 .
the performance of qi is about @xmath1 better than classical system at @xmath129 . versus bandwidth @xmath130 .
parameter settings are @xmath131 , @xmath132 , @xmath133 , @xmath134 , @xmath135 , @xmath136 , @xmath137 , @xmath138 , @xmath104 .
( b ) @xmath98 versus bandwidth @xmath130.,title="fig:",scaledwidth=35.0% ] + fig .
[ errorp ] and @xmath139 versus @xmath140 for @xmath141 .
[ width ] for the other parameter values .
, title="fig:",scaledwidth=35.0% ] + plots error probability @xmath142 and @xmath139 versus @xmath140 for @xmath129 .
microwave(optical ) photon number @xmath143(@xmath144 ) .
qi has orders of magnitude better performance in error probability than classical system .
in summary , we have proposed a scheme to achieve optomechanical qi in weak coupling regime .
the frequency mismatch @xmath0 minimizes the mechanical thermal noise and increases the ratio @xmath73 . as a result
, we achieve a substantial reduction of error probability than classical system at @xmath145 compared to that of ref.@xcite at @xmath146 .
consequently , our quantum illumination proposal works in weak coupling regime .
furthermore , we optimize the delay time of microwave signal s wave packets . finally , we find a @xmath1 improvement of signal - to - noise ratio for @xmath129 .
we acknowledge the financial support in part by the @xmath147-plan of shandong province ( taishan scholars ) , nsfc grant no .
@xmath148 and @xmath149 , open research fund program of the state key laboratory of low - dimensional quantum physics grant no .
@xmath150 ; and the key @xmath151 plan project of shandong province , grant no .
@xmath152 .
99 m. a. nielsen and i. l. chuang , quantum computation and quantum information ( cambridge university press , cambridge , england , 2000 ) .
c. weedbrook et al .
mod . phys . * 84 * , 621 ( 2012 ) .
m. m. wilde , quantum information theory , ( cambridge university press , cambridge , england , 2013 ) .
m. f. sacchi , phys .
a * 72 * , 014305 ( 2005 ) .
m. piani and j. watrous , phys .
lett . * 102 * , 250501 ( 2009 ) .
s. lloyd , science * 321 * , 1463 ( 2008 ) .
tan , et al .
* 101 * , 253601 ( 2008 ) .
j. h. shapiro and s. lloyd , new .
* 11 * , 063045 ( 2009 ) .
s. guha and b. i. erkmen , phys .
a * 80 * , 052310 ( 2009 ) .
s. barzanjeh , et al .
* 114 * , 080503 ( 2015 ) .
c. weedbrook , s. pirandola , j. thompson , v. vedral , and m. gu , new . j. phys . * 18 * , 043027(2016 ) .
e. d. lopaeva , et al .
lett . * 110 * , 153603 ( 2013 ) .
z. zhang , m. tengner , t. zhong , f. n. c. wong , and j. h. shapiro , phys .
lett . * 111 * , 010501 ( 2013 ) .
z. zhang , s. mouradian , f. n. c. wong , and j. h. shapiro , phys .
lett . * 114 * , 110506 ( 2015 ) .
g. romero , j. j. garca - ripoll , and e. solano , phys .
lett . * 102 * , 173602 ( 2009 ) . y .-
f . chen , et al .
lett . * 107 * , 217401 ( 2011 ) .
l. c. g. govia , et al .
a * 90 * , 062307 ( 2014 ) .
k. zhang , f. bariani , y. dong , w. zhang , and p. meystre , phys .
. lett . * 114 * , 113601 ( 2015 ) .
f. marquardt , s. m. girvin , physics , * 2 * , 40 ( 2009 ) ; p. meystre , ann .
( berlin ) * 525 * , 215 ( 2013 ) ; m. aspelmeyer , t. kippenberg , and f. marquardt , rev . mod . phys . * 86 * , 1391 ( 2014 )
. m. paternostro , et al .
lett . * 99 * , 250401 ( 2007 ) . c. wipf , t. corbitt , y. chen , and n. mavalvala , new j. phys .
* 10 * , 095017 ( 2008 ) .
barzanjeh , d. vitali , p. tombesi , and g. j. milburn , phys .
a * 84 * , 042342 ( 2011 ) .
s. g. hofer , w. wieczorek , m. aspelmeyer , and k. hammerer , phys .
a * 84 * , 052327 ( 2011 ) .
barzanjeh , m. abdi , g. j. milburn , p. tombesi , and d. vitali , phys .
109 * , 130503 ( 2012 ) .
l. tian , phys .
lett . * 110 * , 233602 ( 2013 ) . y .-
wang and a. a. clerk , phys .
lett . * 110 * , 253601 ( 2013 ) .
m. c. kuzyk , s. j. van enk , and h. wang , phys .
a * 88 * , 062341 ( 2013 ) .
wang , s. chesi , and a. a. clerk , phys .
a * 91 * , 013807 ( 2015 ) .
a. mari and j. eisert , phys .
lett . * 108 * , 120602 ( 2012 ) .
z. j. deng , x .- b .
yan , y .- d .
wang , and c .- w .
wu , phys .
a * 93 * , 033842 ( 2016 ) .
z. j. deng , s. j. habraken , f. marquardt , new .
* 18 * , 063022(2016 ) . c. gardiner and p. zoller , quantum noise , 3rd ed .
( springer , new york , 2004 ) ; d. f. walls and g. j. milburn , quantum optics ( springer , berlin , 1994 ) . g. vidal and r. f. werner , phys . rev . a * 65 * , 032314 ( 2002 ) . | we propose to realize microwave quantum illumination in weak coupling regime based on multimode optomechanical systems . in our proposal
the multimode together with a frequency - mismatch process could reduce mechanical thermal noise .
therefore , we achieve a significant reduction of error probability than conventional detector in weak coupling regime . moreover , we optimize the signal - to - noise ratio for limited bandwidth by tuning the delay time of entangled wave - packets .
= 0 pt |
in the last years the enumerative geometry of straight lines on quartic surfaces in @xmath1 has been studied by many authors . unlike smooth cubic surfaces , which always contain 27 lines
, one can prove by a standard dimension count that a general quartic surface does not contain any line at all .
one is therefore led to study the maximal number of lines that a quartic surface can contain .
historically , the main focus has been on smooth complex quartic surfaces , which are well - known examples of algebraic k3 surfaces . in 1882
f. schur @xcite discovered a surface with 64 lines that now carries his name , given by the following equation @xmath2 the fact that 64 is indeed the highest number of lines that a smooth complex quartic surface can contain was proven by b. segre in 1943 @xcite . around 70 years later , though , s. rams and m. schtt @xcite discovered a flaw in segre s argument and fixed his proof , extending it to smooth quartic surfaces defined over a field of characteristic different from 2 and 3 . at the same time , a. degtyarev , i. itenberg and a. s. sertz @xcite spurred by a remark of w. barth @xcite were tackling the same problem using the theory of k3 surfaces and nikulin s theory of discriminant forms .
their work resulted in a complete classification of large configurations of lines on smooth complex quartic surfaces up to projective equivalence .
they also proved that schur s quartic is the only one containing 64 lines ( this holds not only over @xmath3 , but also over any field of characteristic different from 2 and 3 , see @xcite ) .
segre rams
schtt s theorem was generalized by the author @xcite to k3 quartic surfaces
i.e. , quartic surfaces containing at worst rational double points as singularities defined over a field of characteristic different from 2 and 3 .
the bound of 64 lines is still not known to be sharp for k3 quartic surfaces .
degtyarev ( private communication ) has proven the existence of a complex k3 quartic surface with 52 lines and 2 singular points of type @xmath4 , whose equation is not known .
non - k3 complex quartic surfaces have been studied by gonzlez alonso and rams @xcite .
if they are not ruled by lines , they can contain at most 48 lines .
they also conjecture that the actual bound is 31 . given a line @xmath5 on a k3 quartic surface @xmath0 , the pencil of planes in @xmath1 containing @xmath5
induces a genus 1 fibration on the minimal desingularization of @xmath0 , i.e. an elliptic or a quasi - elliptic fibration : the latter case can only appear in characteristic 2 and 3 , by a theorem of j. tate @xcite .
heuristically , quasi - elliptic fibrations can carry a higher number of fiber components than elliptic fibrations , thus allowing for a higher number of lines on @xmath0 .
in characteristic 2 , @xmath0 can contain at most 68 lines @xcite .
moreover , if @xmath0 contains 68 lines , then @xmath0 is projectively equivalent to a member of a certain 1-dimensional family discovered by rams and schtt .
on the other hand , the genus 1 fibration induced by a line @xmath5 can be quasi - elliptic only if @xmath5 contains a singular point .
thus , if @xmath0 is smooth , then the fibration is always elliptic .
this fact was proven by rams and schtt in @xcite , where they also exhibited a smooth surface with 60 lines .
degtyarev @xcite subsequently proved that 60 is indeed the maximum possible .
in this paper we are concerned with k3 quartic surfaces defined over a field of characteristic 3 .
the fermat quartic surface defined by @xmath6 contains 112 lines ( see , for instance , @xcite ) .
our main result is the following theorem .
[ thm : char3 ] if @xmath0 is a k3 quartic surface defined over a field of characteristic @xmath7 , then @xmath0 contains @xmath8 lines or at most @xmath9 .
moreover , if @xmath0 contains @xmath8 lines , then @xmath0 is projectively equivalent to the fermat quartic surface .
this theorem extends a similar result by rams and schtt @xcite on smooth quartic surfaces .
the bound of 67 lines is not known to be sharp .
degtyarev @xcite has proven a bound of 60 lines for smooth quartic surfaces not projectively equivalent to the fermat quartic surface , which is not known to be sharp either .
the number 60 , though , could only be attained by a non - supersingular surface .
he also showed that smooth supersingular quartic surfaces have 112 lines or at most 58 , and that there are three different possible configurations of 58 lines . in examples [ ex:58 - 1st - conf ] , [ ex:58 - 2nd - conf ] and [ ex:58 - 3rd - conf ]
we provide explicit equations of three 1-dimensional families of smooth quartic surfaces containing these configurations .
the first and second families had already been found also by degtyarev , while the third one is new , to our knowledge .
the three families were found while examining the proof of proposition [ prop : star ] ( for the third one , also taking advantage of degtyarev s explicit configuration of lines , which was communicated to us by email ) , which states that if @xmath0 contains a star , i.e. , four ( coplanar ) lines intersecting at a smooth point , then @xmath0 contains at most 58 lines .
this provides evidence for the following conjecture . a k3 quartic surfaces defined over a field of characteristic @xmath7 contains @xmath8 lines or at most @xmath10 .
the bound of 67 lines could possibly be improved by distinguishing between supersingular and non - supersingular k3 quartic surfaces . for the former
, a deep lattice - theoretical analysis seems necessary . for the latter ,
our geometrical approach could probably lead to a bound of 64 lines .
the paper is structured as follows .
[ sec : k3-quartic ] : : we set the notation and present some general facts on k3 quartic surfaces defined over a field of any characteristic .
[ sec : char3-elliptic ] : : we study elliptic lines .
not all arguments in characteristic 0 carry over to characteristic 3 , mainly because they were concerned with @xmath7-torsion sections , which are now not so well behaved .
[ sec : char3-qe ] : : we examine a new phenomenon that does not appear in characteristic 0 , namely quasi - elliptic lines : these are very important because they exhibit particularly high valencies .
[ sec : char3-proof ] : : we carry out the proof of theorem [ thm : char3 ] .
we prove a better estimate 58 lines under the hypothesis that @xmath0 contains a star and is not projectively equivalent to the fermat quartic surface , see proposition [ prop : star ] .
[ sec : char3-examples ] : : we discuss some examples of k3 quartic surfaces with many lines . in particular
, we present three 1-dimensional families of smooth surfaces with 58 lines and a surface with 8 singular points and 48 lines .
i wish to thank my supervisor matthias schtt , alex degtyarev and vctor gonzlez alonso for many helpful comments and discussions .
given a field @xmath11 of characteristic @xmath12 , a _
k3 quartic surface _ is a surface in @xmath1 of degree @xmath13 with at most rational double points as singularities . in this section
we fix our notation while recalling some general facts about k3 quartic surfaces . for more details the reader
is referred to @xcite .
henceforth , @xmath0 will denote a fixed k3 quartic surface , @xmath14 the set of ( isolated ) singular points on @xmath0 and @xmath15 the minimal desingularization of @xmath0 . from now on
, @xmath5 will denote a line lying on a k3 quartic surface @xmath0 with minimal desingularization @xmath16 .
any divisor in the complete linear system defined by @xmath17 will be called a _ hyperplane divisor _ ( and often denoted by @xmath18 , too ) .
the strict transform of @xmath5 will be denoted by @xmath19 .
the pencil of planes @xmath20 containing the line @xmath5 induces a genus 1 fibration @xmath21 . a line @xmath5 is said to be _ elliptic _
( respectively _ quasi - elliptic _ ) if it induces an elliptic ( respectively _ quasi - elliptic _ ) fibration .
a fiber @xmath22 of @xmath23 ( @xmath24 ) is the pullback through @xmath25 of the residual cubic @xmath26 contained in @xmath27 .
we denote the restriction of @xmath23 to @xmath19 again by @xmath23 .
if the morphism @xmath28 is constant , we say that @xmath19 has degree @xmath29 ; otherwise , the _ degree _ of @xmath5 is the degree of the morphism @xmath28 . a line @xmath5 is said to be _ separable _
( respectively _ inseparable _ ) if the induced morphism @xmath30 is separable ( respectively inseparable ) .
the _ singularity _ of a line @xmath5 is the number of singular points of @xmath0 lying on @xmath5 .
given a separable line @xmath5 , we will say that a point @xmath31 on @xmath5 is a point of ramification @xmath32 if the corresponding point on @xmath19 has ramification index @xmath33 and @xmath34 .
we recall that if @xmath35 does not divide @xmath33 , then @xmath36 and can be omitted , whereas if @xmath35 divides @xmath33 , then @xmath37 . for the proof of the following lemma , see ( * ? ? ?
* lemma 2.7 ) .
[ lemma : linesthroughsingularpoint ] if @xmath31 is a singular point on a k3 quartic surface @xmath0 , then there are at most @xmath38 lines lying on @xmath0 and passing through @xmath31 . given a k3 quartic surface @xmath0 , we will denote by @xmath39 the number of lines lying on @xmath0 .
building on an idea of segre @xcite , we will usually be interested in finding a _ completely reducible plane _
, i.e. , a plane @xmath40 such that the intersection @xmath41 splits into the highest possible number of irreducible components , namely four lines @xmath42 ( not necessarily distinct ) .
if a line @xmath43 not lying on @xmath40 meets two or more distinct lines @xmath44 , then their point of intersection must be a singular point of the surface .
it follows that @xmath39 is bounded by @xmath45 it will then be a matter of finding a bound for the second and third contribution .
the former will be usually dealt with using lemma [ lemma : linesthroughsingularpoint ] . as for the latter ,
it is natural to introduce the following definition .
the _ valency _ of @xmath5 , denoted by @xmath46 , is the number of lines on @xmath0 distinct from @xmath5 which intersect @xmath5 in smooth points .
most of the time we will express the latter contribution in terms of @xmath47 , and much of the work will be dedicated to finding a bound for these quantities .
of course , not all k3 quartic surfaces admit a completely reducible plane , in which case we will turn to other techniques , such as the ones presented in @xcite .
a _ 3-fiber _ is a fiber whose residual cubic splits into three lines , whereas a _ 1-fiber _ is a fiber whose residual cubic splits into a line and an irreducible conic . a line @xmath5 is said to be _ of type @xmath48 _ , @xmath49 , if in its fibration there are @xmath50 fibers of the former kind and @xmath51 fibers of the latter kind .
the ( local ) valency of a fiber @xmath52 , denoted by @xmath53 , is the number of lines distinct from @xmath5 contained in the plane corresponding to @xmath52 that meet @xmath5 in a smooth point . when it is clear from the context , we will simply write @xmath54 . obviously , @xmath55 and the sum is actually a finite sum .
it follows from this formula that if @xmath5 has type @xmath48 , then @xmath56 let @xmath57 be the coordinates of @xmath1 . up to projective equivalence , we can suppose that the line @xmath5 is given by the vanishing of @xmath58 and @xmath59 , so that the quartic @xmath0 is defined by @xmath60 where @xmath61 are non - negative integers .
[ rmk : res - cub ] we will usually parametrize the planes containing @xmath5 by @xmath62 , @xmath63 , where of course @xmath64 denotes the plane @xmath65 . two equations which define the _ residual cubic _ @xmath26 contained in @xmath27 are the equation of @xmath27 itself and the equation @xmath66[x_1,x_2,x_3]_{(3)}$ ] obtained by substituting @xmath58 with @xmath67 in and factoring out @xmath59 .
an explicit computation shows that the intersection of @xmath5 with @xmath26 is given by the points @xmath68 $ ] satisfying @xmath69 given a line @xmath5 of positive degree , a crucial technique to find bounds for @xmath46 is to count the points of intersection of the residual cubics @xmath26 and @xmath5 which are inflection points for @xmath26 . by _
inflection point _ we mean here a point which is also a zero of the hessian of the cubic .
in fact , if a residual cubic @xmath26 contains a line as a component , all the points of the line will be inflection points of @xmath26 .
writing out the equation of a cubic in @xmath70 explicitly and computing the determinant of its hessian matrix , one can also check the following lemma .
[ lemma : line&tangentconic ] let @xmath71 be a reducible cubic in @xmath70 that is the union of an irreducible conic and a line @xmath43 . then
, the locus of inflection points of @xmath71 is exactly @xmath43 . supposing that the surface @xmath0 is defined as in equation , the hessian of the equation @xmath72 defining the residual cubic @xmath26 ( see remark [ rmk : res - cub ] ) restricted on the line @xmath5 is given by @xmath73[x_2,x_3]_{(3)},\ ] ] which is a polynomial of degree @xmath74 in @xmath75 , with forms of degree @xmath7 in @xmath76 as coefficients ( if @xmath77 , this definition has to be slightly modified , see @xcite ) .
we want now to find the number of lines intersecting @xmath5 by studying the common solutions of and .
it is convenient to extend segre s nomenclature @xcite .
the resultant @xmath78 with respect to the variable @xmath75 of the polynomials and is called the _ resultant _ of the line @xmath5 .
we say that a line @xmath5 of positive degree is a line of the _ second kind _ if its resultant is identically equal to zero .
otherwise , we say that @xmath5 is a line of the _
first kind_. a root @xmath79 $ ] of @xmath78 corresponds to a point @xmath80 $ ] on @xmath5 ; if @xmath31 is a smooth surface point , then it is an inflection point for the residual cubic passing through it .
[ prop:1stkind ] if @xmath5 is a line of the first kind , then @xmath81 .
see ( * ? ? ?
* proposition 2.18 ) .
we follow here the nomenclature of @xcite .
in particular , the _ line graph _ of a k3 quartic surface @xmath0 is the dual graph of the strict transforms of its lines on @xmath82 .
the line graph @xmath83 of a k3 quartic surface @xmath0 is a graph without loops or multiple edges . by definition ,
the number of its vertices is equal to @xmath39 . a dynkin diagram ( resp .
extended dynkin diagram ) is also called an _ elliptic _
graph ( resp .
_ parabolic _ graph ) . a k3 quartic surface @xmath0 is called _ triangle free _ if its line graph contains no triangles , i.e. , cycles of length @xmath7 .
in other words , a k3 quartic surface @xmath0 is triangle free if there are no triples of lines on @xmath0 intersecting pairwise in smooth points .
[ lem : elliptic - triangle - free ] if @xmath5 is an elliptic line on a triangle free k3 quartic surface , then @xmath84 .
[ lem : triangle - conf ] if three lines on @xmath0 form a triangle , then they are contained in plane @xmath40 such that the intersection of @xmath40 and @xmath0 has one of the configurations pictured in figure [ fig : triangle - conf ] . for a proof , see ( * ? ? ? * lemmas 2.22 and 2.23 ) .
the nomenclature for completely reducible planes introduced in figure [ fig : triangle - conf ] will be used also in the next sections .
( -2.5,-0.87) ( 0.5,4.33 ) ; ( -2.13,2.23) ( 2.87,-0.5 ) ; ( -3,0) ( 3,0 ) ; ( 0,4.46) ( 0,-1 ) ; ( 2.5,3.5 ) node @xmath85 ; ( -2.5,-0.87) ( 0.5,4.33 ) ; ( -2.13,2.23) ( 2.87,-0.5 ) ; ( -3,0) ( 3,0 ) ; ( 0,4.46) ( 0,-1 ) ; ( -2,0 ) circle ( 3pt ) ; ( 2.5,3.5 ) node @xmath86 ; ( -2.5,-0.87) ( 0.5,4.33 ) ; ( -2.13,2.23) ( 2.87,-0.5 ) ; ( -3,0) ( 3,0 ) ; ( 0,4.46) ( 0,-1 ) ; ( -2,0 ) circle ( 3pt ) ; ( 0,0 ) circle ( 3pt ) ; ( 2.5,3.5 ) node @xmath87 ; ( -2.5,-0.87) ( 0.5,4.33 ) ; ( -2.13,2.23) ( 2.87,-0.5 ) ; ( -3,0) ( 3,0 ) ; ( 0,4.46) ( 0,-1 ) ; ( -2,0 ) circle ( 3pt ) ; ( 0,0 ) circle ( 3pt ) ; ( 2,0 ) circle ( 3pt ) ; ( 2.5,3.5 ) node @xmath88 ; ( -2.5,-0.87) ( 0.5,4.33 ) ; ( 2.5,-0.87) ( -0.5,4.33 ) ; ( -3,0) ( 3,0 ) ; ( 0,4.46) ( 0,-1 ) ; ( 2.5,3.5 ) node @xmath89 ; ( -2.5,-0.87) ( 0.5,4.33 ) ; ( 2.5,-0.87) ( -0.5,4.33 ) ; ( -3,0) ( 3,0 ) ; ( 0,4.46) ( 0,-1 ) ; ( -2,0 ) circle ( 3pt ) ; ( 2.5,3.5 ) node @xmath90 ; ( -2.5,-0.87) ( 0.5,4.33 ) ; ( 2.5,-0.87) ( -0.5,4.33 ) ; ( -3,0) ( 3,0 ) ; ( 0,4.46) ( 0,-1 ) ; ( -2,0 ) circle ( 3pt ) ; ( 0,0 ) circle ( 3pt ) ; ( 2.5,3.5 ) node @xmath91 ; ( -2.5,-0.87) ( 0.5,4.33 ) ; ( 2.5,-0.87) ( -0.5,4.33 ) ; ( -3,0) ( 3,0 ) ; ( 0,4.46) ( 0,-1 ) ; ( -2,0 ) circle ( 3pt ) ; ( 0,0 ) circle ( 3pt ) ; ( 2,0 ) circle ( 3pt ) ; ( 2.5,3.5 ) node @xmath92 ; ( -1.5 , -0.6) ( 1.5 , 4.6 ) ; ( -1.5,4.6) ( 1.5,-0.6 ) ; ( -3,2) ( 3,2 ) ; ( 0,5) ( 0,-1 ) ; ( 2.5,3.5 ) node @xmath93 ; ( -3,-1 ) rectangle ( 3,5 ) ; ( -2.5,-0.87) ( 0.5,4.33 ) ; ( 2.5,-0.87) ( -0.5,4.33 ) ; ( -3,-0.1) ( 3,-0.1 ) ; ( -3,0.1)(3,0.1 ) ; ( 2.5,3.5 ) node @xmath94 ; ( -1,0 ) circle ( 3pt ) ; ( 0,0 ) circle ( 3pt ) ; ( 1,0 ) circle ( 3pt ) ; ( -3,-1 ) rectangle ( 3,5 ) ; ( -2.5 , -0.6) ( 0.5 , 4.6 ) ; ( -2.5,4.6) ( 0.5,-0.6 ) ; ( -3,1.9) ( 3,1.9 ) ; ( -3,2.1) ( 3,2.1 ) ; ( 2.5,3.5 ) node @xmath95 ; ( 0,2 ) circle ( 3pt ) ; ( 1,2 ) circle ( 3pt ) ; ( 2,2 ) circle ( 3pt ) ;
from now on we assume that the ground field @xmath11 is algebraically closed and has characteristic 3 .
in this section we study elliptic lines , especially separable elliptic lines .
inseparable lines ( both elliptic and quasi - elliptic ) will be analyzed in
[ sec : char3-qe ] .
the results of this section are summarized in table [ tab : char3-elliptic ] .
[ tab : char3-elliptic ] .known bounds for the valency of a separable elliptic line according to its kind , degree and singularity .
sharp bounds are marked with an asterisk . [ cols="^,^,^,<",options="header " , ] for each case , we consider the lattice generated by @xmath19 , a general fiber @xmath52 , the fiber components of the degenerate fibers and the cuspidal curve @xmath96 ( which must be different from @xmath19 , since @xmath5 is separable ) .
all intersection numbers are univocally determined ( @xmath97 because @xmath5 has degree @xmath7 ) , except for @xmath98 but @xmath99 can only take up the values @xmath100 on account of lemma [ lem : char3-d=3-degk ] .
we check that this lattice has rank bigger than @xmath101 in all cases , except for case 6 with @xmath102 ( i.e. , @xmath103 ) . on the other hand
, this case does not exist .
in fact , suppose that @xmath5 is as in case 6 with @xmath104 ; in particular , @xmath5 has no ramified fibers with multiple components and , since @xmath105 , @xmath5 is of the second kind .
it follows that * if @xmath5 has a point of ramification @xmath106 , then by lemma [ lemma:2ndkind - ram2 ] , the ramified fiber must be a cusp , i.e. , @xmath96 intersects @xmath19 so @xmath107 ; * if @xmath5 has ramification @xmath108 , then by lemma [ lemma:2ndkind - ram3_4 ] the ramified fiber must be either a cusp or the union of three distinct lines ; in both cases , @xmath107 .
the following surface contains a separable quasi - elliptic line @xmath109 of degree @xmath7 with valency 21 , thus attaining the bound of proposition [ prop : qe - deg3 ] : @xmath110 it contains only one singular point @xmath111 $ ] of type @xmath112 .
[ prop : qe - deg2 ] if @xmath5 is a quasi - elliptic line of degree 2 , then @xmath113 . by proposition [ prop:1stkind ]
, we can assume that @xmath5 is of the second kind .
let @xmath31 be the singular point on @xmath5 and let @xmath52 be a fiber of @xmath5 with @xmath114 and @xmath115 its corresponding residual cubic .
the cubic @xmath115 is reducible , because it contains at least a line .
suppose that @xmath115 splits into a line @xmath116 and an irreducible conic ( which must be tangent to each other because fibers of type @xmath117 are not admitted in a quasi - elliptic fibration ) .
since @xmath114 , the line @xmath116 meets @xmath5 in a smooth point ; hence , the following three configurations may arise : ( 0,2.5 ) arc ( -270:60:2 and 1.5 ) ; ( -3,0) ( 3,0 ) ; at ( 3,0 ) @xmath5 ; ( 2,-1) ( 2,3.5 ) ; ( -1.5,0 ) circle ( 3pt ) ; at ( -1.5,0 ) @xmath31 ; at ( 2,3 ) @xmath116 ; ( 0,3 ) arc ( -270:60:2 and 1.5 ) ; ( -3,0) ( 3,0 ) ; at ( 3,0 ) @xmath5 ; ( 2,-1) ( 2,3.5 ) ; ( 0,0 ) circle ( 3pt ) ; at ( 0,0 ) @xmath31 ; at ( 2,3 ) @xmath116 ; ( 0,2.5 ) arc ( -270:60:2 and 1.5 ) ; ( -3,1) ( 3,1 ) ; at ( 3,1 ) @xmath5 ; ( 2,-1) ( 2,3.5 ) ; ( -2,1 ) circle ( 3pt ) ; at ( -2,1 ) @xmath31 ; at ( 2,3 ) @xmath116 ; all three configurations are impossible for the following reasons : * in the first configuration , the conic meets @xmath5 in a non - inflection point ( the smooth surface point ) , by lemma [ lemma : line&tangentconic ] ; * the second configuration can be ruled out by an explicit parametrization ( in a line of the second kind , either the point @xmath31 is a ramification point , or the cubic passing twice through @xmath31 is singular at @xmath31 ) ; * the third configuration gives rise to a fiber of type @xmath118 , which is not a quasi - elliptic fiber .
thus , @xmath115 must split into three ( not necessarily distinct ) lines and at least one of them should pass through @xmath31 .
since there can be at most 8 lines through a singular point ( lemma [ lemma : linesthroughsingularpoint ] ) , there can be at most 7 such reducible fibers , each of them contributing at most 2 to the valency of @xmath5 , whence @xmath113 .
the bound given by proposition [ prop : qe - deg2 ] is sharp .
in fact , the following quartic surface contains a quasi - elliptic line @xmath109 of degree @xmath106 and valency @xmath119 : @xmath120 the quartic contains two singular points , @xmath121 $ ] of type @xmath4 and @xmath122 $ ] of type @xmath112 .
the line @xmath5 has 7 fibers of type @xmath123 and one fiber of type @xmath124 corresponding to the plane containing @xmath125 . if @xmath5 is a quasi - elliptic line of degree @xmath126 , then @xmath127 .
the fibration induced by the line @xmath5 has at most 10 reducible fibers , each of which contributes at most 1 to its valency .
in this section we employ the notation and the ideas of [ subsec : triangle - free ] and @xcite .
[ prop : char3-alex ] let @xmath128 be the line graph of a triangle free k3 quartic surface . if @xmath128 contains a parabolic subgraph @xmath129 , then @xmath130 the subgraph @xmath128 induces a genus 1 fibration , which can be elliptic or quasi - elliptic ( * ? ? ?
* , theorem 1 ) .
the vertices in @xmath131 are fiber components of this fibration .
if the fibration is elliptic , there can not be more than 24 components , on account of the euler
poincar characteristic .
if the fibration is quasi - elliptic , we obtain from formula that @xmath132 a fiber of type @xmath123 can contain at most 2 lines , since there are no triangles .
hence , from and we deduce @xmath133 [ lem : char3-triangle - free ] if @xmath134 is a line on a triangle free k3 quartic surface , then @xmath84 .
thanks to lemma [ lem : elliptic - triangle - free ] , we can suppose that @xmath5 is quasi - elliptic .
we can prove that @xmath135 for any degenerate fiber @xmath52 , which together with the euler poincar characteristic formula and equation yields @xmath127 . to see this , note that a reducible fiber has @xmath136 , so is obvious for a 1-fiber . on the other hand , a 3-fiber can not be of type @xmath123 , in virtue of the triangle free hypothesis ; hence , it must induce a fiber of type @xmath124 or @xmath137 , for which @xmath138 and holds . [
prop : char3-triangle - free ] a triangle free k3 quartic surface can contain at most 64 lines .
one can adapt the proof in ( * ? ? ?
* proposition 5.9 ) using proposition [ prop : char3-alex ] and lemma [ lem : char3-triangle - free ] .
we will suppose from now on that @xmath0 has a triangle formed by the lines @xmath139 , which are necessarily coplanar .
the plane on which they lie must contain a fourth line @xmath140 ( which might coincide with one of the former ) .
we will start our analysis with the following special configuration .
a _ star _ on a quartic surface @xmath0 is the union of four distinct lines meeting in a smooth point .
( -1.5 , -0.6) ( 1.5 , 4.6 ) ; ( -1.5,4.6) ( 1.5,-0.6 ) ; ( -3,2) ( 3,2 ) ; ( 0,5) ( 0,-1 ) ; since the four lines in a star are necessarily coplanar , a star is the same as a configuration @xmath93 ( figure [ fig : triangle - conf ] ) .
the lines have necessarily degree 3 because there are no singular points on the plane containing them
. we will be able to prove in proposition [ prop : star ] that if @xmath0 contains a star and is not projectively equivalent to the fermat surface , then @xmath141 , which is a sharp bound .
we will first need a series of lemmas . in all of them
, we will parametrize the surface @xmath0 as in in such a way that the star is contained in the plane @xmath142 and the lines meet at @xmath143 $ ] , i.e. , setting the following coefficients equal to @xmath29 : @xmath144 if necessary , we will parametrize a second line in the star @xmath43 as @xmath145 , by further assuming @xmath146 .
[ lemma : star-1st - kind ] if @xmath5 is a line of the first kind in a star , then @xmath147 .
it can be checked by an explicit computation that the resultant of @xmath5 has a root of order @xmath148 at the center of the star ; this implies that there are at most @xmath149 lines meeting @xmath5 not contained in the star .
[ lemma : star-(3(3),2 ) ] if @xmath5 is a separable line of ramification @xmath150 contained in a star , then it is of the first kind . by a change of coordinates , we can assume that the point of ramification index 2 is @xmath151 $ ] , and that ramification occurs at @xmath65 .
imposing that @xmath5 is of the second kind leads to a contradiction ( @xmath5 can not be separable ) .
[ lemma : star-3_4 ] if three lines in a star are separable and at least two of them have ramification @xmath108 , then the third one also has ramification @xmath108 . beside @xmath152 and @xmath153
, we can suppose without loss of generality that a third line is given by @xmath154 , setting @xmath155 .
the conditions for @xmath5 , @xmath43 or @xmath156 to be of ramification @xmath108 are @xmath157 , @xmath158 and @xmath159 , respectively .
clearly , two of them imply the third one .
[ lemma : star - no-3x3_4 ] if three lines in a star are separable , then at most two of them can be of the second kind .
we parametrize @xmath5 , @xmath43 and @xmath156 as in the previous lemma .
imposing that all three of them are of the second kind leads to a contradiction ( at least one of them must be inseparable ) .
[ lemma : star-1st - kind+3_4 ] let @xmath5 and @xmath43 be two lines in a star ; if @xmath5 is a separable line of the second kind , and @xmath43 is a line of the first kind of ramification @xmath108 , then @xmath160 .
this can be checked again by an explicit computation of the resultant of @xmath43 , which has now a root of order @xmath161 at the center of the star .
[ lemma : star - not - cuspidal+cuspidal ] let @xmath5 and @xmath43 be two lines in a star ; if @xmath5 is a cuspidal line , and @xmath43 is not cuspidal , then @xmath160 .
we parametrize @xmath109 as in corollary [ cor : family - c ] . by virtue of lemma [ lemma : inseparable=>cuspidal ]
, we can suppose that @xmath153 is separable .
an explicit computation shows that @xmath43 can not be of the second kind , and that its resultant has a root of order @xmath161 in @xmath162 .
[ lemma : star - not - cuspidal+2xcuspidal ] let @xmath5 , @xmath43 and @xmath156 be three lines in a star ; if @xmath5 and @xmath43 are cuspidal , and @xmath156 is not cuspidal , then @xmath163 .
we parametrize @xmath152 and @xmath153 as in corollary [ cor : family - c ] , i.e. , we suppose that @xmath0 is given by the family @xmath164 where the following coefficients are set to zero : @xmath165 by a further rescaling we put @xmath166 and we consider @xmath167 .
the line @xmath156 is inseparable and we can compute its polynomial @xmath168 as in formula in the proof of lemma [ lemma : inseparable=>cuspidal ] ( by parametrizing the pencil with @xmath169 ) , which turns out to be @xmath170 this means that @xmath156 has only one singular fiber in @xmath171 ( namely a fiber of type @xmath123 with the maximum possible index of wild ramification ) , unless @xmath172 and @xmath173 , in which case @xmath156 is cuspidal . [ lemma : cuspidal=>two - stars ] if @xmath5 is a cuspidal line which is not contained in at least two stars , then @xmath174 . on account of lemma [ cor : cuspidal - degenerate - fiber ] , the number of stars in which @xmath5 is contained is exactly equal to the number of fibers of type @xmath123 in its fibration ; moreover , @xmath175 if @xmath52 is of type @xmath124 or @xmath137 , yielding @xmath176 recalling formula , we deduce that if @xmath177 then @xmath174 .
[ prop : star ] if @xmath0 contains a star and is not projectively equivalent to the fermat surface , then @xmath0 contains at most @xmath10 lines .
let @xmath178 be the lines contained in the star .
we will always use the bound , which takes the form @xmath179 1 .
suppose first that all lines @xmath44 are not cuspidal .
* if @xmath180 for @xmath181 , then @xmath182 * if @xmath183 , then by lemmas [ lemma : star-1st - kind ] and [ lemma : inseparable=>cuspidal ] , @xmath184 must be separable of the second kind ; hence @xmath185 ; if @xmath186 for @xmath187 , then @xmath188 * if @xmath183 and @xmath189 , then by the same token both @xmath184 and @xmath190 are separable lines of the second kind .
on account of lemma [ lemma : star-(3(3),2 ) ] , they both have ramification @xmath108 .
we claim that both @xmath191 and @xmath192 are not greater than @xmath193 .
in fact , if @xmath194 is separable , then by lemmas [ lemma : star-3_4 ] and [ lemma : star - no-3x3_4 ] it must be of the first kind and have ramification @xmath108 , which in turn implies that @xmath195 , because of lemma [ lemma : star-1st - kind+3_4 ] ; if @xmath194 is inseparable , then @xmath84 by lemma [ lemma : inseparable=>cuspidal ] .
the same applies to @xmath140 .
hence , we conclude that @xmath196 2 .
assume now that exactly one of the lines , say @xmath184 , is cuspidal , so that @xmath197 . on account of lemma [ lemma : star - not - cuspidal+cuspidal ]
we have @xmath198 3 .
suppose then that both @xmath184 and @xmath190 are cuspidal .
if @xmath194 and @xmath140 are not cuspidal , then by lemma [ lemma : star - not - cuspidal+2xcuspidal ] @xmath199 4 .
finally , suppose that @xmath184 , @xmath190 and @xmath194 are cuspidal .
* by a local computation it can be seen that @xmath140 is also necessarily cuspidal . *
thanks to the bound of lemma [ lemma : cuspidal=>two - stars ] , we can suppose that at least two lines , say @xmath184 and @xmath190 , are part of another star .
* pick two lines @xmath200 and @xmath201 , each of them in another star containing @xmath184 respectively @xmath190 , which intersect each other ( necessarily in a smooth point ) . *
perform a change of coordinates so that @xmath184 , @xmath190 , @xmath200 and @xmath201 are given respectively by @xmath202 , @xmath145 , @xmath203 and @xmath204 .
* impose that @xmath184 , @xmath190 and @xmath194 are cuspidal lines : the resulting surface is projectively equivalent to fermat surface . in this section
we study the case in which @xmath0 admits a triangle .
the three lines forming the triangle need to be coplanar , and we will denote by @xmath40 the plane on which they lie . obviously , the plane @xmath40 intersects @xmath0 also in a fourth line , which might coincide with one of the first three .
we first consider this degenerate case .
[ prop : triangle - mult ] if @xmath0 admits a completely reducible plane @xmath40 with a triangle and a multiple component , then @xmath0 contains at most @xmath205 lines . by lemma [ lem : triangle - conf ] ,
@xmath0 admits configuration @xmath94 or @xmath95 . in order to bound @xmath39 ,
we use as usual formula .
let @xmath206 be the double line in the plane @xmath40 containing one of the two configurations , and let @xmath184 and @xmath190 be the two simple lines .
lines meeting @xmath206 different from @xmath184 and @xmath190 must pass through the singular points ; hence , by lemma [ lemma : linesthroughsingularpoint ] there can be at most @xmath207 of them .
note that @xmath184 and @xmath190 can not be cuspidal because of corollary [ cor : cuspidal - degenerate - fiber ] . in the fibrations induced by @xmath184 and
@xmath190 the plane @xmath40 corresponds to a fiber with a multiple component , hence with euler
poincar characteristic at least 6 ; therefore , if @xmath184 and @xmath190 are both elliptic , there can be at most @xmath208 more lines meeting them , yielding @xmath209 suppose that @xmath184 is quasi - elliptic .
the plane @xmath40 corresponds to a fiber of type @xmath124 or @xmath137 ; hence , there can be only one singular point on @xmath206 : in fact , by inspection of the dynkin diagrams , a component of multiplicity @xmath106 in these fiber types meets at most @xmath106 other components ( and one of them is the strict transform of @xmath190 ) .
the lines @xmath184 and @xmath190 not being cuspidal , we know that they have valency at most @xmath210 .
it follows that @xmath211 [ lemma : a0a1 ] let @xmath5 and @xmath43 be two lines of degree @xmath7 in configuration @xmath85 or @xmath86 . if @xmath212 , then @xmath213 .
let @xmath40 be the plane containing @xmath5 and @xmath43 .
both lines are separable , since otherwise the respective residual cubics would intersect them in one point .
we suppose that also @xmath214 and look for a contradiction
. since both lines have valency greater than @xmath208 , they must be lines of the second kind with ramification @xmath215 or @xmath108 .
in particular , they must have a point of ramification @xmath7 ( let us call it @xmath216 and @xmath217 ) , which does not lie on @xmath40 . up to change of coordinates , we can assume the following : * @xmath40 is the plane @xmath142 ; * @xmath5 and @xmath43 are given respectively by @xmath202 and @xmath145 ; * @xmath31 is given by @xmath151 $ ] and @xmath218 by @xmath219 $ ] ; * ramification in @xmath31 ( resp .
@xmath218 ) occurs in @xmath65 ( resp .
@xmath162 ) .
this amounts to setting the following coefficients equal to @xmath29 : @xmath220 furthermore , @xmath221 , since the two residual lines in @xmath40 do not contain @xmath143 $ ] , the intersection point of @xmath5 and @xmath43 ; we set @xmath222 after rescaling one variable .
two necessary condition for @xmath5 and @xmath43 to be lines of the second kind are @xmath223
this means that the residual conic in @xmath224 is given explicitly by @xmath225 this conic splits into two lines by hypothesis ; hence , it has a singular point .
computing the derivatives , one finds that the following condition must be satisfied : @xmath226 substituting into , one finds that the conic degenerates to a double line : @xmath227 thus , we have neither configuration @xmath85 nor @xmath86 . [
prop : triangle - not - star ] if @xmath0 admits a triangle but not a star , then @xmath0 contains at most @xmath9 lines .
the proof is a case - by - case analysis on the configurations that are given by lemma [ lem : triangle - conf ] , except configurations @xmath164 ( a star , treated in proposition [ prop : star ] ) , @xmath94 and @xmath95 ( treated in proposition [ prop : triangle - mult ] ) . beside the fact that there are at most @xmath38 lines through a singular point and the bounds on the valency of [ sec : char3-elliptic ] and [ sec : char3-qe ]
, one should observe that in configurations of type @xmath228 , the three lines meeting at the same ( smooth ) point must be of the first kind by lemma [ lemma:2ndkind - ram2 ] .
for configurations @xmath85 and @xmath86 , one uses lemma [ lemma : a0a1 ] .
for instance , let us prove the proposition for configuration @xmath86 .
let @xmath184 and @xmath190 be the lines through the singular point , and @xmath194 and @xmath140 the other two lines .
we know that @xmath229 , @xmath230 , whereas lemma [ lemma : a0a1 ] applies to @xmath194 and @xmath140 , yielding @xmath231 it follows from that @xmath232 we leave the remaining cases to the reader .
we have treated the case of triangle free surfaces in proposition [ prop : char3-triangle - free ] , the star case in proposition [ prop : star ] and the star free triangle case in proposition [ prop : triangle - not - star ] , so the proof is now complete .
in this last section , we present examples of k3 quartic surfaces with many lines .
in particular , we provide explicit equations for three 1-dimensional families of surfaces with 58 lines .
most of the examples including the first two families ( examples [ ex:58 - 1st - conf ] and [ ex:58 - 2nd - conf ] ) were found during the proof of the theorem , especially of proposition [ prop : star ] .
note that the first two families had already been discovered independently by a. degtyarev , who was also aware of the existence of a third configuration with 58 lines .
we found the third family ( example [ ex:58 - 3rd - conf ] ) after we were informed of his work ; as far as we know , this family is new .
if @xmath236 , the surface contains a star ( in @xmath142 ) formed by two cuspidal lines ( of valency 30 ) and two elliptic lines with no other singular fibers than the star itself ( hence , of valency 3 ) .
the remaining @xmath237 lines are of type @xmath238 .
the surface contains exactly 19 stars .
more precisely , as long as @xmath241 , the surface contains one cuspidal line ( given by @xmath202 ) which intersects 12 lines of type @xmath242 , and 18 lines of type @xmath243 ; the remaining 27 lines are of type @xmath244 ( for instance , @xmath204 ) .
the surface contains exactly 10 stars . for @xmath234
we find again a model of the fermat surface , whereas for @xmath245 the surface contains 20 lines and a triple point .
for @xmath239 we obtain the union of two planes and a quadric surface . more precisely , if @xmath248 and @xmath249 , then the surface contains exactly one star in the plane @xmath250 the star is formed by two lines of type @xmath251 and two lines of type @xmath243 , whose equations can be explicitly written after a change of parameter @xmath252 .
each line of type @xmath251 meets 18 lines of type @xmath253 , and each line of type @xmath243 meets 9 lines of type @xmath244 .
all lines are elliptic . if @xmath256 , then the surface contains 9 points of type @xmath4 and 40 lines . the star in the plane is formed by two elliptic lines of type @xmath242 and two quasi - elliptic lines of type @xmath243 .
each line of type @xmath242 intersects 9 lines of singularity 2 and valency 6 , while each line of type @xmath243 intersects 9 lines of singularity 1 and valency 9 .
[ ex : shimada - shioda ] the reduction modulo 3 of shimada
shioda s surface @xmath260 ( see @xcite ) can be written @xmath261 where @xmath262 it contains 8 singular points of type @xmath4 and 48 lines .
so far , this is the example known to us with highest number of lines and at least one singular point . | we prove that a k3 quartic surface @xmath0 defined over a field of characteristic 3 can contain at most 112 lines .
if @xmath0 contains 112 lines , then it is projectively equivalent to the fermat quartic surface ; otherwise , @xmath0 contains at most 67 lines .
if @xmath0 contains a star , then it has 112 lines or at most 58 .
we provide explicit equations of three 1-dimensional families of smooth quartic surfaces with 58 lines . |
since the first discovery of an x - ray pulsar in the cen x@xmath13 system , it has been generally accepted that accretion disk plays a key role in the evolution of accretion - powered x - ray binary pulsars ( axbps ) .
the instantaneous luminosity and spin rates are mainly determined by accretion rates from the surrounding disk through the release of gravitational potential energy and the transfer of angular momentum . if the precessing accretion disk is tilted with respect to the orbital plane of the binary and periodically lies in a plane containing the line - of - sight to the pulsar , the long - term super - orbital motions detected from a few axbps can also be explained . in spite of its critical importance and extensive previous studies ,
however , the detailed properties of the accretion disk around axbp are still poorly known . one useful phenomenon for studying the accretion disk around axbp is quasi - periodic oscillations ( qpos ) which have been considered to be related to the motion of the accretion disk , especially the innermost part .
about 10 axbps have been detected with x - ray qpos with frequency of 5220 mhz to date ( boroson et al . 2000 , and references therein ) .
according to the beat - frequency model ( bfm ) , which has been the canonical model for qpos in x - ray binaries , the observed qpo frequency ( @xmath4 ) is a beat frequency between the coherent spin frequency of the pulsar ( @xmath5 ) and the keplerian frequency ( @xmath6 ) of the innermost disk , @xmath4 = @xmath6 @xmath1 @xmath5 , at the magnetosphere boundary of the pulsar @xcite .
another model often used in explaining the x - ray binary qpos is the keplerian - frequency model ( kfm ) , which interprets the qpos as a result of absorption of x - rays by inhomogeneities in the inner edge of the accretion disk , and , subsequently , the observed qpo frequency as the frequency of the keplerian motion , @xmath4 = @xmath6 @xcite . since the kfm requires @xmath5
@xmath7 @xmath4 = @xmath6 ( otherwise the accretion could be prohibited by the so - called propeller effect " [ illarinov & sunyaev 1975 ] ) , which is not always true for the case of axbps , a modification is needed to explain the observed frequencies of qpos from axbps ( e.g. , kommers , chakrabarty , & lewin 1998 , hereafter kcl98 ) . hence
, while the bfm explains qpos as luminosity oscillations , the kfm explains qpo as beaming oscillations .
lmc x4 is a persistent , disk - fed , high - mass axbp in the large magellanic cloud with a pulsational period of @xmath013.5 s. the average x - ray ( 2@xmath125 kev ) luminosity is @xmath02 @xmath2 10@xmath8 ergs s@xmath9 and the optical companion is a 14th magnitude o - type star of @xmath015 @xmath10 .
the @xmath01.4 day orbital period was found by several different studies including observations of periodic x - ray eclipses and optical brightness variations .
the long - term intensity of lmc x4 varies with a period of @xmath030.3 days , indicating the possible existence of a precessing accretion disk tilted relative to the orbital plane .
heindl et al .
( 1999 ) pointed out that , if this long - term variation is due to reprocessing of the x - ray beam by a disk , thomson scattering by a hot ionized disk must be the dominant mechanism of reprocessing .
heemskerk & van paradijs ( 1989 ) estimated the average x - ray albedo of the accretion disk to be @xmath00.93 from analysis of the optical light curve .
lmc x4 has exhibited large x - ray flares which uniquely distinguish it from other axbps . according to the results of levine , rappaport , & zojcheski ( 2000 , hereafter lrz00 ) , the x - ray ( 2@xmath125 kev ) luminosity increases up to @xmath010@xmath11 ergs s@xmath9 during the flares and the modulation factor of the pulsations increases substantially .
individual pulsations are clearly visible from the raw data of the flares .
two relatively long - term temporal structures of the flares were identified with characteristic time scales of @xmath0150 and @xmath12 400 s , respectively .
the period of the flaring episodes is not well estimated , but is roughly equal to the orbital period .
although it is very likely that the flaring episodes are related to increased rates of accretion , no proposed mechanism successfully explains the origin of the flares ( e.g. , lrz00 ) .
rxte detected several large flares from lmc x4 on 1996 august 19 during @xmath036-hour continuous observations of the source .
we transformed the photon arrival times from the good xenon data of the proportional counter array to the solar system barycenter using the jpl de400 ephemeris .
four data segments , each with @xmath04.5-ks length , containing the flaring episode were obtained .
light curves of the first and second data segments , obtained with the all five proportional counter units in the 2@xmath15 kev range , are shown in fig .
1 with a time resolution of 4 s. several large flares are clearly identified , and we shall call the flares of the first data segment flare 1 " and the second data segment flare 2 . "
the phase of the @xmath035-d super - orbital motion , based on the results of ilovaisky et al .
( 1984 ) , for the flares is 0.66 @xmath13 0.23 , while the @xmath01.4-d binary orbital phases , based on the ephemeris of lrz00 , for flare 1 and flare 2 are @xmath00.26 and @xmath00.30 , respectively .
the difference between the start time of data segment 1 and data segment 2 is @xmath06000 s. three different temporal structures during the flares are evident in fig . 1 : ( 1 ) strong burst - like structures with @xmath07001500 s period ( we shall call these big qpos " [ bqpos ] ) ( 2 ) relatively weak oscillations with @xmath050500 s period ( we shall call these medium qpos " [ mqpos ] ) , and ( 3 ) @xmath013.5-s coherent pulsations in the small window of figure 1a . fig .
2 shows the leahy - normalized power density spectra ( pds ) of flare 1 and flare 2 obtained with a time resolution of 2@xmath14 s. the main windows show the pds around the coherent pulse frequency , @xmath00.074 hz . the small windows on the left sides show the low - frequency ( mhz ) pds in the logarithmic frequency scale , while the small windows on the right sides show the detailed structures of the sidebands around the coherent frequency .
at least three characteristics are evident in the pds of fig . 2 : ( 1 ) a strong peak at the coherent pulse frequency , ( 2 ) well - developed significant sidebands around the coherent frequency , and ( 3 ) several strong components at mhz frequencies .
the low - frequency powers due to the bqpos and mqpos are identified in in the left panels .
the frequency of the bqpos decreases from @xmath01.35 to @xmath00.65 mhz from flare 1 to flare 2 , which is evident in fig .
1 . the mqpos in fig . 1 are the several strong excess power features at @xmath02@xmath120 mhz .
the sidebands of flare 1 are very close to a symmetric distribution with respect to the coherent frequency , while the sidebands of flare 2 show significant asymmetry between the intensities of the lower - frequency ( lf ) and the higher - frequency ( hf ) sidebands , with the lf @xmath02 times stronger than the hf sidebands .
table [ tbl-1 ] lists fractional root - mean - square ( frms ) amplitudes , obtained by the methods described by van der klis ( 1989 ) , of the qpos and coherent pulsations in figure 2 .
3 shows the same pds as fig . 2 , but with the energy range of 513 kev .
we see almost identical structures to the 25 kev range . as in fig .
2 , the structure of the sidebands around the coherent pulse frequency closely resembles the structure of the low - frequency powers in the mhz frequency range . in order to effectively compare the sideband structures with the power distribution of the mhz qpos ,
the powers of the mhz qpos were overlaid onto those of the sidebands : first the pds of the mhz qpos were shifted to the frequency of the hf sidebands , and then a mirror image , with respect to the coherent frequency , of the shifted pds onto the frequencies of the lf sidebands was made .
finally , the intensities of the shifted pds of the mhz qpos were scaled to match those of the sidebands
4 compares the pds of the sidebands ( solid histogram ) with the shifted , scaled pds ( crosses ) of the mhz qpos in flare 1 ( fig .
4a ) and flare 2 ( fig .
4b ) , respectively .
they show almost identical distributions ( except for the excess emission of the lf sidebands of flare 2 ) , indicating that the amplitudes of the coherent pulsations is modulated by those of the mhz qpos .
qpo - related sidebands around a coherent pulse frequency were previously detected from the low - mass x - ray binary pulsar 4u 1626@xmath167 ( kcl98 ) , where @xmath048 mhz qpos modulate @xmath0130 mhz coherent pulsations . in the low - energy pds of this source ,
the lf sidebands are significantly stronger than their hf complements .
kcl98 suggested that a structure ( or blob " ) orbiting at the keplerian radius of the qpo frequency produces ( 1 ) symmetric sidebands by crossing the line - of - sight once per orbit and thus attenuating the pulsar beam and ( 2 ) excess emission in the lf sidebands of the low - energy pds by reprocessing of the pulsar beam at the beat frequencies between the pulsational frequency and the qpo when the structure is not in the line - of - sight .
the model of kcl98 explains the observed pds of 4u 1626@xmath167 , although the nature of the orbiting structure and the reprocessing mechanism are not clear . since the direct comparison between the pds of the qpos and those of the sidebands
are not given for this source , it is not clear whether the coherent pulsations are modulated exactly by the amplitude of the qpos .
lmc x4 shows at least two apparently different features from 4u 1626@xmath167 .
first , the mhz qpos only appear during the large x - ray flares , while the @xmath048 mhz qpos of 4u 1626@xmath167 seem to have existed for more than a decade .
next , only the mqpos of flare 2 show the significant excess emission of the lf sidebands , and do so in both the low- and high - energy bands .
we discuss the origin of the mhz qpos from lmc x4 via a combination of the bfm and a modified version of the kfm .
first , the bqpos can be explained by the bfm as follows : suppose that the observed x - ray flares are induced by big clumps accreting onto the surface of the neutron star at the inner edge of the accretion disk .
the beating between the coherent pulsar frequency and the keplerian frequency of the clumps can result in beat frequencies at the observed qpo frequencies , @xmath4 = @xmath6 @xmath5 , implying that the keplerian radius corresponding to the bqpos lies very close to , but inside of , the corotation radius for accretion to occur ( bildsten et al . 1997 ) .
this is consistent with the results of previous observations @xcite which predict that the magnetospheric radius of lmc x4 is comparable to the corotation radius , i.e. , lmc x4 is in spin equilibrium . the expected corotation radius for an 1.4 @xmath10 neutron star is @xmath0 9.7 @xmath2 10@xmath15 cm .
next , the mqpos can be explained by a modified version of the kfm as follows : suppose that the observed frequencies represent the keplerian orbital frequencies of clumps ( or blobs " ) , implying keplerian orbital radii of @xmath02@xmath110 @xmath2 10@xmath16 cm .
the observed sidebands of the mqpos , then , can be explained similarly to the model of kcl98 , where the excess emission of the lf sidebands is generated by reprocessing on the clump when the clump is not in the observer s line - of - sight to the pulsar .
the model of kcl98 , however , explains the excess emission of the lf sidebands only in the soft - energy band , which is not the case for lmc x4 .
this difference can be explained by the difference in reprocessing mechanim of accretion disks around the two sources .
if the reprocessing mechanims on the accretion disk around lmc x4 is less energy sensitive , such as reflection and scattering , the excess emission in the lf sidebands of lmc x4 in both soft- and hard - energy bands can be accounted for .
this argument is indeed consistent with the results of heemskerk & van paradijs ( 1989 ) , where a very high x - ray albedo ( 0.93 ) of the accretion disk around lmc x4 was obtained .
heindl et al .
( 1999 ) also suggested the possibility that scattering is the dominant reprocessing mechanism of the accretion disk around lmc x4 . in conclusion , it seems to be likely that the bqpos originate near the corotation radius , while the mqpos originate from the orbital motions of clumps in the accretion disk outside the corotation radius . in fig . 2 and fig
3 , the excess emission of the lf sidebands appears only in flare 2 .
this may indicate the existence of the excess free electrons produced by the intense radiation from flare 1 .
these excess free electrons then may effectively enhance the scattering efficiency for the radiation from flare 2 . on the other hand , the apparent highest frequency of the mqpos increases from @xmath010 to @xmath020 mhz in fig 2 . and figure 3 .
since higher frequency implies a smaller keplerian radius , this increase may indicate a radial infalling motion in the accretion disk during the large x - ray flares . if this is true , the inferred average velocity is @xmath02 km s@xmath9 .
if the bqpos are due to the large accreting clumps as explained above , it is evident that they are unique feature caused by the large x - ray flares . during the normal state ( i.e.
, non - flaring state ) , the power at the frequencies of the mqpos are significantly less than the power expected from the frms amplitude of the large x - ray flares , suggesting the possibility that the mqpos are also unique features of the flaring episodes .
although the combination of the bfm and a modified version of the kfm seems to account for the observed temporal properties of the large x - ray flares , this does not rule out , at this point , the possibility of different mechanisms for the origin of the mhz qpos .
for example , titarchuk & osherovich ( 2000 ) proposed global oscillations in the accretion disk for the origin of qpos .
more observations , especially long - term monitoring of the flare activities , are needed to investigate these scenarios more thoroughly .
we summarize our results as follows : * we have discovered mhz qpos and qpo - amplitude - modulated coherent pulsations in large x - ray flares from lmc x4 . *
the aperiodic variabilities of the flares can be classified as bqpos ( @xmath00.651.35 mhz ) and mqpos ( @xmath0220 mhz ) . *
the bqpos can be explained by the beating between the coherent pulse frequency of the neutron star and the keplerian frequency of large accreting clumps near the corotation radius * the mqpos can be explained by attenuation and reprocessing from the orbital motions of clumps in the accretion disk outside the corotation radius . *
it is very likely that reflection and scattering are the dominant reprocessing mechanisms on the accretion disk around lmc x4 , and the increase of the highest qpo frequencies may indicate an infalling motion of the accretion disk inner edge as the flares continue . * the results of this paper
present some of best and most extreme examples of coupling between periodic and aperiodic variabilities of axbps .
we would like to thank the referee , bram boroson , for comments and suggestions .
dsm acknowledges wynn ho for his comments and careful reading of manuscript .
this research has made use of data obtained from the _ high energy astrophysics science archive research center _ ( heasarc ) , provided by nasa s goddard space flight center .
dsm is supported by nsf grant ast-9986898 .
se is supported in part by an nsf faculty early careeer development ( career ) award ( nsf-9983830 ) .
alpar , a. , & shaham , j. 1985 , nature , 316 , 239 bildsten , l. , et al .
1997 , apjs , 113 , 367 boroson , b. , obrien , k. , horne , k , kallman , t. , still , m. , boyd , p. t. , quaintrell , h. , & vrtilek , s. d. 2000 , apj , 545 , 399 heemskerk , m. h. m. , & van paradijs , j. 1989 , a&a , 223 , 154 heindl , w. a. , gruber , d. e. , rothschild , r. e. , crannel , c. j. , lang , f. l. , & kaplan , l. 1999 , baas , 31 , 715 illarionov , a. f. , & sunyaev , r. a. 1975 , a&a , 39 , 185 ilovaisky , s. a. , chevalier , c. , motch , c. , pakull , m. , van paradijs , j. , & lub , j. 1984 , a&a , 140 , 251 kommers , j. m. , chakarbarty , d. , & lewin , w. h. g. 1998 , apj , 497 , l33 ( kcl98 ) lamb , f. k. , shibazaki , n. , alpar , a. , & shaham , j. 1985 , nature , 317 , 681 la barbera , a. , burderi , l. , di salvo , t. , iaria , r. , & robba , n. 2000 , baas , 32 , 715 levine , a. m. , rappaport , s. a. , & zojcheski , g. 2000 , apj , 541 , 194 ( lrz00 ) naranan , s. , et al .
1985 , apj , 290 , 487 titarchuk , l. , & osherovich , v. 2000 , apj , 542 , l111 van der klis , m. 1989 , in timing neutron stars , ed . h. @xmath17gelman and e. p. j. van den heuvel ( kluwer : dordrecht ) , 27 van der klis , m. , stella , l. , white , n. , jansen , f. , & parmar , a. n. 1987 , apj , 316 , 411 woo , j. w. , clark , g. w. , levine , a. m. , corbet , r. h. d. , & nagase , f. 1996 , apj , 467 , 811 | we report the discovery of milli - hertz ( mhz ) quasi - periodic oscillations ( qpos ) and qpo - modulated pulsations during large x - ray flares from the high - mass x - ray binary pulsar lmc x4 using data from the _ rossi x - ray timing explorer _ ( rxte ) .
the lightcurves of flares show that , in addition to @xmath074 mhz coherent pulsations , there exist two more time - varying temporal structures at frequencies of @xmath00.651.35 and @xmath0220 mhz .
these relatively long - term structures appear in the power density spectra as mhz qpos and as well - developed sidebands around the coherent pulse frequency as well , indicating that the amplitudes of the coherent pulsation is modulated by those of the mhz qpos .
one interesting feature is that , while the first flare shows symmetric sidebands around the coherent pulse frequency , the second flare shows significant excess emission in the lower - frequency sidebands due to the @xmath0220 mhz qpos .
we discuss the origin of the qpos using a combination of the beat - frequency model and a modified version of the keplerian - frequency model . according to our discussion
, it seems to be possible to attribute the origin of the @xmath00.651.35 and @xmath0220 mhz qpos to the beating between the rotational frequency of the neutron star and the keplerian frequency of large accreting clumps near the corotation radius and to the orbital motion of clumps at keplerian radii of 2@xmath110 @xmath2 10@xmath3 cm , respectively . |
considerable efforts have been made in the understanding of localization properties of electronic states of disordered systems since the work by anderson.@xcite being the simplest one to investigate the above phenomena , the anderson model is formulated through a tight binding hamiltonian with diagonal disorder and hopping @xmath0 to nearest neighbors .
the site energies are randomly distributed within some range @xmath1 , and @xmath2 is the relevant parameter to distinguish between the strong and weak localization regimes .
anderson showed that disorder , in three dimensional systems , drives a metal - insulator transition , with a mobility edge separating localized from extended electronic states .
the one parameter scaling theory , introduced by abrahams _
_ , @xcite predicts that all states are localized for dimensions smaller than or equal to two .
dimension two then appears as a marginal case .
the one - parameter scaling hypothesis assumes that the conductance of a finite sample scales in a universal way , and has been supported by several experiments done in the early 1980s,@xcite and by a large number of numerical works.@xcite however , it has been suggested that the theory breaks down when electron - electron interactions are included , or when an external magnetic field is applied.@xcite remarkable new experiments on 2d electron systems,@xcite in samples with much higher electron mobilities than the previous ones , do show a clear evidence for a metal - insulator transition at low electron densities , in the absence of a magnetic field .
the effect has been ascribed to the dominant role of coulomb interactions , and a new scaling theory has been proposed for interacting electrons,@xcite leaving open the possibility of a 2d metal . within the one - electron scheme , it was well known that in 2d , the effect of back scattering from impurities is enhanced , leading to an insulating behavior at zero temperature .
electron trajectories interfere constructively in the backward direction , and one gets localization for any degree of disorder .
transport properties with an external magnetic field are useful probes to test the above effects .
quantum interference of electron trajectories , modified by the applied field , may change the localization properties of the electron wave function .
one evidence of this fact is the observation of magnetoresistance oscillations with an applied field in thin disordered metallic cylinders.@xcite with this geometry , and orienting the field along the cylinder axis , one can select the phase of the electron orbits in order to observe the phenomenon macroscopically . for two - dimensional systems , like the ones realized in semiconductor heterostructures , it has been observed a metal - insulator transition induced by the magnetic field.@xcite samples that , at sufficiently low temperatures and zero field , exhibits a localized electronic behavior , present evidence of extended states at strong magnetic fields .
other evidences for a transition to a metallic regime have been observed in experimental studies of the magnetoresistance and magnetoconductance.@xcite as a general rule , quantum interference effects are invoked to explain the delocalization of the electronic wave function with the field .
@xcite the reentrant behavior of the magnetoresistance found by some authors at very high magnetic fields,@xcite is associated with the shrinkage of the electronic wave function.@xcite one effect marginally taken into account , and that may have important consequences , even within the one - electron picture , is the inclusion of short - range order ( sro ) correlations in the electronic properties of disordered systems.@xcite real disordered systems are not necessarily random , but may present different degrees of short - range order . in a binary alloy , for given concentrations of the atomic species ,
one may get configurations ranging from segregation to complete order as a function of the sro properties . in this work
we investigate the role of the above effects and its interplay with the action of an applied magnetic field , in the quantum site percolation limit .
the latter regime corresponds to the case where the two site energies , for a binary alloy , are very far apart in the energy scale , and _ percolation _ of the wave function is only possible through atoms of the same species .
thus , this limit favors localization when short range correlates atoms of different kind as nearest neighbors .
we remark that this case is difficult to calculate within other approximations of nonlocal character.@xcite we calculate the average electronic density of states ( dos ) using a previously developed method reported in ref .. our approach relies on one dimensional numerical techniques , and is efficiently applied to long stripes and bars of finite cross section with the idea that two and three dimensional lattices may be obtained by joining infinite linear chains . from grounds based on statistical mechanics results,@xcite this geometry , generically called as _ cylindric
_ , is expected to converge more rapidly to the bulk limit than the simpler block geometry .
indeed , for the disordered case , the bulk regime can be extrapolated with a relatively small number of chains .
sro effects are included following the kikuchi method,@xcite where correlations are systematically handled .
the last step in the calculation involves a configurational average , where the probability of a given configuration is obtained in terms of the short and long range order parameters .
the electronic dos calculated through the above procedure , in spite of being an average , is rich in structures , and as we will show below , is highly susceptible to changes in local properties .
magnetic field effects are considered with the introduction of a phase in the electronic hopping term@xcite _ via _ the so called peierls substitution.@xcite this phase is related to the magnetic flux across an elementary cell , as given by the well known aharonov - bohm effect.@xcite the lattice model approach adopted in this paper is known to accurately describe the lowest landau level in the situation where the disorder broadening is such that no overlap of magnetic sub - bands occurs .
we are also neglecting electron - electron interactions and electron spin ( spin - split landau levels ) , and all the effects found in the present contribution are exclusively due to disorder in a one - particle lattice model .
we found that magnetic field effects are stronger near the segregation limit , where one expects weak localization properties and a higher degree of coherence in the electronic wave function .
an oscillatory behavior of the electronic dos is observed as a function of the field , with surprising effects on the structure of the dos .
our paper is organized as follows : in the next section we present the method to calculate the dos . as illustrative examples ,
we show results for one - component non - disordered samples in two and three dimensions . the two - dimensional disordered case including sro effects is also shown for completeness . in section
iii we extend our calculation to three - dimensional disordered systems in the presence of an applied magnetic field . finally , in section iv we present and
discuss numerical results for several examples of the latter .
the computational scheme that will be employed here , was developed previously by the authors to study the electronic dos in two - dimensional disordered binary alloys.@xcite for completeness , we discuss in this section some details of the method and present the extension to three dimensions in the next section .
a tight - binding hamiltonian for a single @xmath3-atomic - like orbital , in a hypercubic lattice structure in @xmath4 dimensions , yields the following dispersion relation in @xmath5-space : @xmath6 where @xmath7 is the atomic energy , @xmath8 is the hopping ( kinetic ) term for nearest neighbors in a pure system , and @xmath9 is the lattice parameter . the dos for @xmath4-dimensional systems can be obtained recursively from the dos in @xmath10 dimensions , in the form of a convolution with the dos of a linear chain @xmath11 this result is shown analytically in the appendix for two dimensions . for a finite system with the geometry of a _ hyper - cylinder
_ , like a strip with finite width , the dos is obtained as a discrete convolution of the dos of the linear chain with the distribution of eigenvalues for the finite transverse section of the cylinder @xmath12 in finite systems , strong oscillations are present in the dos , specially near the points where van hove singularities are developed in the infinite limit ( in analogy with the so called gibbs phenomenon for fourier series ) .
the above oscillations are typical size effects , with the amplitude being damped when the size of the finite section is increased.@xcite in fig.[fig1 ] we display the dos for the square lattice for wide but finite width stripes , along with the infinite two dimensional case ( uppermost figure ) .
the corresponding three dimensional case ( cubic lattice ) is depicted in fig.[fig2 ] .
we compare finite bar geometries with the infinite three dimensional limit .
the energy scale is given in units of @xmath8 , the hopping integral .
the large amplitude of oscillations are due to coherence effects in pure samples , and as we will show later , size effects decay more rapidly in disordered systems . in addition , as remarked in the introduction , the _ cylindric _ geometry , with an infinite dimension along the axis , is expected to converge considerably faster to the bulk limit than the block geometry .
this latter fact allows the use of relatively small samples to extrapolate the dos of disordered alloys .
one comment is in order here . as shown in expression ( [ findos ] ) , the dos for a finite system is obtained as a weighted superposition of one - dimensional densities .
the interchain coupling displaces the center of gravity of the one - dimensional dos to the _ renormalized _ value @xmath13 , which takes into account the true coordination number and the degrees of freedom of the transverse section of the system .
the weight in ( [ findos ] ) for each chain is exactly the discrete density of the eigenvalues @xmath14 .
the above exact analytic results are extended heuristically to treat disordered systems . in this section
we show the procedure for two dimensions , @xcite where the transverse part of the system is a finite chain .
local order effects are handled with a single variable , the so called cowley s
parameter @xmath15 @xcite : @xmath16 with @xmath17 being the probability of having an @xmath18 pair , and @xmath19 and @xmath20 the concentrations for species @xmath21 and @xmath22 respectively . with the above definition , @xmath15 ranges from @xmath23 ( limit of complete local order ) to @xmath24 ( limit of segregation ) , passing through intermediate cases with variable local order that include the random limit for @xmath25 . for markovian chains ,
sro effects were embodied in a numerically exact algorithm based on the so called negative factor counting ( nfc ) method@xcite to calculate the dos for such systems .
our scheme thus relies on exact calculations for @xmath26 , the dos of disordered infinite chains with local order described by @xmath15 . to get the analogous of expression ( [ findos ] ) for a disordered two - dimensional stripe of width @xmath27
, we generate all the configurations of a finite chain of @xmath27 sites , and get the corresponding eigenvalues and their multiplicities . for a given configuration , the dos is obtained as @xmath28 where the @xmath29 is an infinite chain dos with the center of gravity displaced to @xmath13 , the eigenvalues of the finite transverse chain for the given configuration .
each term of the sum ( [ strip ] ) is weighted by @xmath30 , the discrete distribution of the eigenvalues @xmath14 .
the final dos is an average over all the possible configurations : @xmath31 with @xmath32 equal to the probability of a given configuration @xmath33 .
the latter is obtained , for markov chains,@xcite as a weighted product of pair probabilities and depends only on @xmath15 and the concentrations . in fig.[fig3 ] we show the lowest sub - band for 50 - 50% alloying with different degrees of sro in two dimensional strips of width @xmath34 . in spite of the finite size , the three cases already display two dimensional features characteristic of the infinite limit .
the dos structure changes dramatically as a function of the local order , with highly localized states near the random and ordered limits .
the upper figure is close to the segregation limit and resembles the pure case shown in fig.[fig1 ] . for this regime
, the electronic wave function may percolate through atoms of the same type in very large islands .
in contrast with one dimension , the above properties are typical of the two dimensional topology , with many different percolation paths for extended electron trajectories . for three dimensions ,
the finite section is of two dimensional character and sro properties are handled with more parameters that include the possibility of closed path correlations .
this is done using the kikuchi method@xcite and is described in the next section .
following ideas developed in finite - size scaling methods , @xcite our system is taken as a bar in three dimensions , with finite dimensions in the @xmath35 and @xmath36 directions , and infinite extension along the @xmath37 direction .
periodic boundary conditions are imposed for the finite section of the system .
the magnetic field is applied parallel to the @xmath37 axis .
we model our system using an alloy tight binding ( atb ) hamiltonian : @xmath38 where off - diagonal elements are taken different from zero only for nearest neighbors , and @xmath39 is a phase that represents the effect of the magnetic field in an elementary cell with the lattice parameter as a side ( peierls substitution@xcite ) . for simplicity , we choose the hopping amplitude as real along the @xmath35 axis ( @xmath40 ) , and complex along the @xmath36 axis .
the phase difference between two successive bonds is related to the magnetic flux enclosed in an elementary cell , @xmath41 where @xmath42 is the associated electromagnetic vector potential .
periodic boundary conditions for the finite section of the system induce a quantization of the magnetic flux that depends on the size of the sample .
one can get rid of the above constraint by choosing free ends , but in this case finite size effects decay more slowly .
as mentioned in the previous section , we studied the limit corresponding to the quantum site percolation case , where @xmath43 .
sro is treated using the kikuchi method with closed square diagrams as highest order correlations.@xcite this approximation is known to correctly describe the two dimensional topology . for each square
we have 16 different configurations .
these configurations are displayed in fig.[fig4 ] , with the respective probabilities ( @xmath44 ) and degeneracies ( @xmath45 ) .
normalization conditions require : @xmath46 for concentrations , @xmath47 for pairs , and @xmath48 for squares . in all , we get 5 independent parameters , which are chosen to be @xmath17 , @xmath49 , @xmath50 , @xmath51 and @xmath52 , with the last two defined as : @xmath53 while the first three parameters ( @xmath54 ) are related to correlations of short range , the last two ( [ lro ] ) handle long range order .
dependent variables are written as : @xmath55 within the approximation adopted , a given configuration of the finite section is obtained by joining squares . the corresponding probability is obtained as a product of conditional square probabilities , where the constraint is that two squares are linked by a common bond .
this approach has been devised in analogy to markovian chains,@xcite but with the square as the basic correlation . as it will be seen in the next section , our method allows to include local properties under the coupled effects of variable disorder and application of a magnetic field .
in the examples that follow , we consider 50 - 50% alloying and several degrees of local order .
the site energies are taken to be @xmath56 and @xmath57 , in units of the hopping term @xmath8 .
in all the figures , we only display the lowest sub - band , the one centered at @xmath58 .
fig.[fig5 ] shows examples calculated in the absence of a magnetic field : ( a ) near the ordered limit ; ( b ) random disorder ; and ( c ) near the limit of segregation .
the behavior is similar to the one shown previously for two dimensions . near the segregated limit
the dos for each sub - band resembles the one of the pure case , suggesting the presence of extended electronic states . on the other hand ,
when the system is near the ordered limit , the dos shows the typical structure of localized states , as it should be expected for two kinds of atoms with very different site energies . in this instance , the probability for one state to propagate is nearly zero , due to almost infinite barriers .
the authors are currently calculating other cases where the centroids approach to each other , and the bands overlap .
in this situation , quantum diffusion is possible by means of tunneling through high but finite barriers , and a delocalization onset is expected as a function of the band separation.@xcite to illustrate the role of local order , in fig.[fig6 ] we display the dos values for three nearby energy levels as a function of the correlation @xmath59 , in two dimensions .
one of the levels corresponds to a sharp peak in the middle of the band in the lowest part of fig.[fig3 ] , for @xmath60 .
size effects are pictured in fig.[fig7 ] for the random alloy .
it is worth noting that the dos of finite systems already present the whole structure of the infinite case , even for small transverse sections .
this seems to be characteristic of the bar geometry , where one of the dimensions is infinite and the bulk limit is extrapolated with not too big cross sections . due to numerical limitations , the next examples in the presence of a magnetic field ,
are calculated in finite geometries only ( sections of 12 atoms ) .
we now analyze the role of the magnetic field . in fig.[fig8 ] we show the dos for a case near the segregation limit .
an oscillatory behavior of the dos with the field is observed , and the figure depicts cases up to half the period in the flux .
the values are adjusted to periodic boundary conditions .
the effect is paramount at the center of the sub - band , with the appearance of a gap for @xmath61 .
the phenomenon is due to destructive interference of the electronic paths in the presence of the field .
the band width is also affected , with a striking reduction at half a period . on the other hand , tiny effects
are observed near the random and ordered limits , with no sensitive changes in the structure of the dos as a function of the field .
other cases with different concentrations were also calculated . as a general rule ,
the dos structure is more affected by the field near the so called _ weak localization _ regime , when the concentration of one of the species is small .
localization and delocalization of electronic states may be interpreted as pure quantum effects due to destructive or constructive interference of the different electronic paths .
this interference is shifted by the application of a field . in fig.[fig9 ] , we show the effect of the magnetic field on a particular level as a function of short - range correlations .
the figure displays a trajectory in our parameter space , ranging from the local ordered limit to the segregated case .
this trajectory , for short , is labeled with the symbol @xmath59 .
the figure pictures in a different way our comments above : the dos oscillates as a function of the field , but the effect predominates in the vicinity of the segregated limit ( for @xmath62 in the figure ) . in summary
, it has been shown that the method proposed is numerically efficient , reproduces the dos in important limits , and predicts new behaviors as function of the sro and magnetic field . to theoretically investigate localization properties
, one needs to calculate the electron wave function , or get indirect information _ via _ the so called localization length . in two and three dimensions , most of the works in the literature concentrate around simulations or other numerical techniques . as a common characteristic
, they are based on finite size scaling methods , but most of the cases investigated correspond to random disordered systems , represented generally by the anderson model or the quantum percolation model@xcite our results , in spite that we calculate the average _ dos _ , suggest that _ sro effects may play an important role in the localization of electronic states_. a phase diagram in which the degree of localization is displayed , may depend not only on the concentrations , but also on the sro parameters . in order to illustrate this issue
, we suggest numerical computations of the localization length in systems like the ones considered here , with equal concentrations of the species , very different site energies , and variable local order properties .
changes in the localization properties will then be entirely ascribed to short - range correlations .
the applied magnetic field strongly changes the electronic properties in the _ weak localization _ limit . on the other hand , in the opposite regime , the effects of the field on the average dos are negligible .
huge magnetic fields are involved in the above phenomena due to the small size of the elementary cell .
however , the effects predicted here may be brought to a range susceptible of observation in the case of superlattice structures , or in geometries where the electronic paths enclose macroscopic areas that contribute to the flux .
we note that the present state of the art in nanostructure technology allows the fabrication of modulated 2d systems with lattice parameters of the order of 100 nm or more ( see for instance ref . ) .
the authors are grateful to professors d. gottlieb , b. laks , n. majlis , and p. schulz for helpful discussions .
this work was partially financed by a grant from _
( brazil , project no .
b-11487/2b9221 ) .
one of the authors ( g.g.c . ) also acknowledges support from _ conselho de desenvolvimento cientfico e tecnolgico _
( cnpq , brazil , project no.301221/77 - 4 ) .
the dos for the pure cubic lattice in @xmath4 dimensions was given in section ii by eq.([dos ] ) .
we show explicitly this result for the two - dimensional case .
if we consider that states @xmath63 represent the basis in the site representation , where the indexes are integers along the @xmath35 and @xmath36 axes respectively , the hamiltonian reads : @xmath64 taking the fourier transform in relation to the @xmath35 direction : @xmath65 we get : @xmath66 the green s function in the basis @xmath67 is given by : @xcite @xmath68 where @xmath69 are the eigenvalues : @xmath70 with : @xmath71 and @xmath72 corresponding to the eigenvalue of an _ effective _ chain with site energy @xmath73 and hopping @xmath74 between first neighbors along the @xmath36 direction .
the dos and the green s function satisfy the relation@xcite : @xmath77 with : @xmath78 and subsequently the dos is given by : @xmath79 where @xmath80 is the dos of an infinite linear chain centered at @xmath72 , and @xmath81 is the degeneracy of the eigenvalue . | electronic properties of disordered binary alloys are studied _ via _ the calculation of the average density of states ( dos ) in two and three dimensions .
we propose a new approximate scheme that allows for the inclusion of local order effects in finite geometries and extrapolates the behavior of infinite systems following _ finite - size scaling _ ideas .
we particularly investigate the limit of the quantum site percolation regime described by a tight - binding hamiltonian .
this limit was chosen to probe the role of short range order ( sro ) properties under extreme conditions .
the method is numerically highly efficient and asymptotically exact in important limits , predicting the correct dos structure as a function of the sro parameters .
magnetic field effects can also be included in our model to study the interplay of local order and the shifted quantum interference driven by the field .
the average dos is highly sensitive to changes in the sro properties , and striking effects are observed when a magnetic field is applied near the _ segregated _ regime .
the new effects observed are twofold : there is a reduction of the band width and the formation of a gap in the middle of the band , both as a consequence of destructive interference of electronic paths and the loss of coherence for particular values of the magnetic field .
the above phenomena are periodic in the magnetic flux . for other limits that imply strong localization ,
the magnetic field produces minor changes in the structure of the average dos . |
the main object of this paper is a stochastic differential equation @xmath0 , \end{gathered}\ ] ] where @xmath1 is a standard wiener process ; @xmath2 is a fractional brownian motion ( fbm ) with the hurst parameter @xmath3 ; @xmath4 is a poisson measure with finite intensity measure .
such equation can be used to model processes with jumps on financial markets , where two principal random noises influence the prices .
one noise is coming from economical background and has a long - range dependence , which is modeled by the fbm .
another noise is intrinsic to the stock exchange , where millions of agents act independently and behave irrationally sometimes ; this is a white noise and it is modeled by a wiener process . although equation were not studied before , many authors considered some particular examples . for a pure fractional stochastic equation without wiener component and jumps @xmath5 existence and uniqueness of a solution to such equation
was proved first in @xcite . in @xcite
this result was proved under weaker assumptions on coefficients and integrability of the solution was established for @xmath6 . for a homogeneous drift - less equation ( i.e. @xmath7 , @xmath8 ) ,
the integrability was shown in @xcite for all @xmath9 .
mixed stochastic differential equation without jumps @xmath10 was first considered in @xcite , where unique solvability was proved for time - independent coefficients and zero drift , i.e. @xmath11 .
later , in @xcite , existence of solution to was proved under less restrictive assumptions , but only locally , i.e. up to a random time . in @xcite global existence and uniqueness of solution to was established under the assumption that @xmath1 and @xmath2 are independent .
the latter result was obtained in @xcite and @xcite without the independence assumptions , and it was also shown in @xcite that all moments of the solution are finite for @xmath6 .
it is also worth mentioning that article @xcite contains related results , which imply , in particular , that has unique solution for @xmath12 .
recently , equation without brownian component ( i.e. @xmath13 ) was considered in @xcite , where the existence of solution is proved under weaker conditions on regularity of @xmath14 ( only hlder continuity in the second variable ) and without the assumptions @xmath15 , @xmath9 . on the other hand , in that paper @xmath16 and @xmath17 is independent of @xmath18 , which are much stronger assumptions than those of the present paper . in this paper
we show that has a unique solution .
the main result is existence of moments of the solution , which is an important property for applications .
the existence of moments is proved for all @xmath9 in inhomogeneous case , which is a novelty in comparison to the results of @xcite .
the paper is organized as follows .
section [ sec : prelim ] gives basic definitions . in section [ sec : exuniq ] we prove existence and uniqueness of the solution to .
section [ sec : exmom ] contains results on integrability of solutions to .
let @xmath19 be a complete filtered probability space satisfying the usual assumptions .
let also @xmath20 be an @xmath21-wiener process and @xmath22 be an @xmath23-adapted fractional brownian motion ( fbm ) , i.e. a centered gaussian process with the covariance @xmath24}=\frac{1}{2}(s^{2h}+t^{2h}-|t - s|^{2h})$ ] .
let also @xmath4 be an @xmath23-adapted poisson measure with intensity measure @xmath25 , i.e. @xmath26}=\pi(dy)dt.\ ] ] we will assume that the intensity measure @xmath25 is finite : @xmath27 it is well known that @xmath2 has a modification with almost surely continuous paths ( even hlder continuous of any order up to @xmath28 ) , and further we will assume that it is continuous itself .
now we define how we understand the integrals in . the integral with respect to the wiener process
@xmath1 is the standard it integral , and it is well defined as long as @xmath29 almost surely . the integral with respect to @xmath4 is defined as usual . since @xmath15 , the process
@xmath30 is well defined and it is a pure jump process , which almost surely has finite number of jumps ; we can assume that it is cadlag .
so the integral with respect to @xmath4 is just a finite sum @xmath31 } q(s , x_{s-},\delta l_s),\ ] ] where @xmath32 .
as for the integral with respect to the fbm @xmath2 , we use the generalized lebesgue stieltjes integral ( see @xcite ) .
its construction uses the fractional derivatives @xmath33 assuming that @xmath34 , \ d_{b-}^{1-\alpha}g_{b-}\in l_\infty[a , b]$ ] , where @xmath35 , the generalized ( fractional ) lebesgue - stieltjes integral @xmath36 is defined as @xmath37 it follows from hlder continuity of @xmath2 that @xmath38 $ ] a.s .
then for a function @xmath39 with @xmath40 $ ] we can define integral with respect to @xmath2 through : @xmath41 note that in the case where @xmath39 is piecewise hlder continuous with exponent @xmath42 , this integral is just a limit of forward integral sums @xmath43 .
( this fact is proved in @xcite for hlder continuous functions @xmath39 , but it is easily checked that the proof works for piecewise hlder continuous functions as well . ) hence , for such functions , all usual properties of integral hold : linearity , additivity etc . throughout the paper ,
the symbol @xmath44 will denote a generic constant , whose value is not significant and can change from one line to another . to emphasize its dependence on some parameters ,
we will put them into subscripts .
in this and the following sections we impose the following assumptions on the coefficients of : 1 . the function @xmath45 is differentiable in @xmath18 and for all @xmath46 , @xmath47 $ ] latexmath:[\[\begin{gathered } 2 . the functions @xmath14 , @xmath49 and @xmath50 are lipschitz continuous in @xmath18 : @xmath51 for all @xmath52 , @xmath47 $ ] .
the functions @xmath14 , @xmath49 and @xmath50 are hlder continuous in @xmath53 : for some @xmath54 and for all @xmath55 $ ] , @xmath56 @xmath57 we do not impose any assumptions on the function @xmath58 except joint measurability in all arguments . we will say that a process @xmath59 is a solution to if it is cadlag and has the following properties : * for some @xmath60 @xmath61 for all @xmath62 such that @xmath63,{{\mathbb r } } ) = 0 $ ] ( i.e. @xmath59 is hlder continuous between jumps of the process @xmath64 defined in ) ; * equation holds almost surely for all @xmath47 $ ] . from
it follows that , almost surely , @xmath64 has finitely many jumps on @xmath65 $ ] , so a solution @xmath59 is piecewise hlder continuous of order @xmath60 , consequently , the integral @xmath66 is well defined for all @xmath47 $ ] .
it also follows that @xmath67 a.s .
, thus the integrals @xmath68 and @xmath69 are well defined too . [ thm_ex_sol ] equation has a unique solution .
for @xmath70 , consider the following equation : @xmath71 where @xmath72 is a wiener process , @xmath73 is a process with almost surely hlder continuous paths of order @xmath74 .
it was proved in @xcite that such equation has a unique solution in the class of hlder continuous processes of order @xmath75 .
we denote this solution by @xmath76 .
let @xmath77 be the moment of the @xmath78th jump of process @xmath64 .
define a sequence of processes @xmath79 recursively as follows .
let the process @xmath80 be the solution to .
if for @xmath81 the process @xmath82 is constructed , set @xmath83 , @xmath84 , @xmath85 , @xmath86 . on the stochastic basis @xmath87 with @xmath88 , @xmath89\}$ ] , the process @xmath90 is a wiener process , and @xmath91 is hlder continuous of any order @xmath92 , hence we can define @xmath93 , @xmath94 .
now we put @xmath95 and show that this process solves . indeed , for @xmath96 , @xmath97 , we have @xmath98 thus , for any @xmath99 we can write @xmath100 i.e. @xmath59 solves .
uniqueness follows from a similar reasoning : from uniqueness of solution for and the strong markov property of @xmath1 we get , that for @xmath96 , @xmath97 , the solution of satisfies @xmath101 .
on the other hand , @xmath102 , hence any solution of coincides with the one constructed above .
we start by making pathwise estimates of the solution to equation without jumps .
we fix some @xmath103 and introduce the following notation : @xmath104 } = \sup_{s\le u < v < t } \left(\frac{{\left\vertf(v)-f(u)\right\vert}}{(v - u)^{1-\alpha } } + \int_u^v \frac{{\left\vertf(u)-f(z)\right\vert}}{(z - u)^{2-\alpha}}dz\right).\end{gathered}\ ] ] observe that it follows from that @xmath105}\\\times \int_a^b \left({\left\vertf(s)\right\vert}(s - a)^{-\alpha } + \int_a^s { \left\vertf(s)-f(u)\right\vert}(s - u)^{-1-\alpha}du \right)ds . \end{gathered}\ ] ] [ pathwiseestimatemixed ] for the solution @xmath59 of , the following estimate holds : @xmath106}^{1/(1-\alpha)}\right\}}\left(1+{\left\lverti_b\right\rvert}_{\infty;t}\right),\ ] ] where @xmath107 , @xmath47 $ ] . a similar estimate ( naturally , without @xmath108 )
was obtained for the pure fractional equation in @xcite , but with exponent @xmath109 instead of @xmath110 here . in our proof
we will use methods similar to those of @xcite , but we modify them as follows . while in @xcite , the sum @xmath111 is estimated and a version of the gronwall lemma is used , here we will estimate these terms separately and then use a kind of two - dimensional gronwall lemma .
we start by estimating @xmath115 : @xmath116 where @xmath117 is as above , @xmath118 , @xmath119 .
estimate @xmath120 , @xmath121 by , @xmath122 summing up , we have @xmath123 whence @xmath124 where we have used the estimate @xmath125 further , we estimate @xmath126 : @xmath127 where @xmath128 here to estimate @xmath129 we used the following computation : @xmath130 combining all estimates , we get @xmath131 whence @xmath132 thus , we get the following system of inequalities : @xmath133 with some constant @xmath134 ( which can be assumed to be greater than 1 without loss of generality ) . putting @xmath135
, we get from the first inequality that @xmath136 we remark that @xmath137 and @xmath138 are almost surely finite by the results of @xcite . plugging this to the second inequality and making simple transformations
, we arrive at @xmath139 with some constant , which is no longer of interest . substituting this to
, we get @xmath140 finally , @xmath141}^{1/(1-\alpha)}\right)\right\}}(1+j_b)\\\le c \exp{\left\{c{\left\lvertb^h\right\rvert}_{0;[0,t]}^{1/(1-\alpha)}\right\}}(1+j_b),\end{gathered}\ ] ] as required .
thanks to lemma [ pathwiseestimatemixed ] , it is enough to prove that all moments of @xmath146}\right\}}$ ] and of @xmath108 are finite .
the first follows from the fact that @xmath147}$ ] is an almost surely finite supremum of a gaussian family , and @xmath148 , since @xmath149.the first is proved as in ( * ? ? ?
* lemma 2.3 ) , but the proof is short and for completeness we repeat it here .
denote @xmath150 and write @xmath151}\le c_p(i_b ' + i_b''),\ ] ] where @xmath152}{\left\vert\int_0^{t } b_s dw_s\right\vert}^p \right ] } \le c_p{\mathsf{e}\left[\left(\int_0^t { \left\vertb_s\right\vert}^2ds\right)^{p/2}\right ] } < \infty,\\ i_b '' = { \mathsf{e}\left[\sup_{t\in[0,t]}\left(\int_0^t { \left\vert\int_s^t b_z dw_z\right\vert } ( t - s)^{-1-\alpha}ds\right)^p \right]}.\end{gathered}\ ] ] by the garsia
rodemich rumsey inequality ( * ? ? ? * theorem 1.4 ) , for arbitrary @xmath153 , @xmath154 $ ] @xmath155 where _
( t ) = ( _ 0^t _ 0^t dxdy)^/2 .
@xmath156 for @xmath157 by the hlder inequality @xmath158}\le c_{p,\eta}\int_0^t \int_0^t \frac{{\mathsf{e}\left[{\left\vert\int_x^y b_v dw_v\right\vert}^{p}\right ] } } { { \left\vertx - y\right\vert}^{p/2}}dx\,dy\\ \le c_{p,\eta}\int_0^t \frac{{\mathsf{e}\left[\left(\int_x^y ( b_v)^2 dv\right)^{p/2}\right ] } } { { \left\vertx - y\right\vert}^{p/2 } } dx\,dy<\infty,\end{gathered}\ ] ] whence @xmath159}\sup_{t\in[0,t]}\left(\int_0^t ( t - s)^{-1/2-\eta-\alpha}ds\right)^p < \infty,\ ] ] and the statement follows .
now turn to equation . to prove existence and uniqueness of its solution
, we did not make any assumptions about @xmath4 and @xmath58 , except that @xmath4 has finite activity : @xmath15 .
to prove existence of moments , we will make further assumptions on the measure @xmath4 and the coefficient @xmath58 in addition to h1h4 .
let @xmath77 be the moment of the @xmath78th jump of process @xmath64 .
as in the proof of [ thm_ex_sol ] , consider for @xmath70 the equation @xmath71 where @xmath72 is a wiener process , @xmath73 is an adapted process with almost surely @xmath164-hlder continuous paths for some @xmath74 , and denote the unique solution to this equation by @xmath76 .
a reasoning similar to that used in the proof of lemma [ pathwiseestimatemixed ] gives @xmath165}\le k{\left\verty\right\vert } \exp{\left\{k{\left\lvertz\right\rvert}_{0;[0,t]}^{1/(1-\alpha)}\right\}}(1+{\left\lverti_{b , y}\right\rvert}_{\infty;t}),\ ] ] where @xmath166 , the constant @xmath134 depends only on constants in assumptions h1h4 , without loss of generality we assume @xmath167 .
hence we get for @xmath168 @xmath169}^{1/(1-\alpha)}\right\}}\left(1 + { \left\lverti_{b}\right\rvert}_{\infty;\tau_1}\right),\end{gathered}\ ] ] where @xmath170 . for convenience denote @xmath171 and assume without loss of generality that @xmath172 , @xmath56 . then @xmath173}^{1/(1-\alpha)}\right\ } } ( 1+j_{b}),\ ] ] where @xmath174 further , @xmath175}^{1/(1-\alpha)}\right\ } } ( 1+j_{b}).\end{gathered}\ ] ] using inequality consequently on @xmath176 , @xmath177 , @xmath178 and estimating jumps of @xmath59 as in , we arrive at @xmath179}{\left\vertx_{t}\right\vert}\le \sup_{t\in[0,t]}{\left\vertx_{t}\right\vert}_1\le ( 3k)^{n(t)}(1+j_{b})^{n(t)+1}{\left\vertx_0\right\vert}_1\exp{\left\{k{\left\lvertb^h\right\rvert}_{0;[0,\tau_1\wedge
t]}^{1/(1-\alpha)}\right\}}\\ \times \prod_{n:\tau_n\le t}g(\delta l_{\tau_n})\exp{\left\{k{\left\lvertb^h\right\rvert}_{0;[\tau_n,\tau_{n+1}\wedge t]}^{1/(1-\alpha)}\right\}}\\ \le c(6k)^{n(t)}\left(1+j_{b}^{n(t)+1}\right)\exp{\left\{k{\left\lvertb^h\right\rvert}_{0;[0,\tau_1\wedge t]}^{1/(1-\alpha)}\right\}}\\ \times \prod_{n:\tau_n\le t}g(\delta l_{\tau_n})\exp{\left\{k{\left\lvertb^h\right\rvert}_{0;[\tau_n,\tau_{n+1}\wedge t]}^{1/(1-\alpha)}\right\}}\end{gathered}\ ] ] where @xmath180 denotes the number of jumps of the process @xmath64 on @xmath65 $ ] . hence , by the hlder inequality , @xmath181}{\left\vertx_{t}\right\vert}^{p}\right]}\le c_p\left(\vphantom{{\mathsf{e}\left[\prod_{n=1}^{n(t ) } g(\delta l_{\tau_n})^{4p}\right]}}{\mathsf{e}\left[(6k)^{4pn(t)}\right ] } \left(1+{\mathsf{e}\left[j_{b}^{4p(n(t)+1)}\right]}\right)\right.\\ \times\left.{\mathsf{e}\left[\prod_{n=1}^{n(t ) } g(\delta l_{\tau_n})^{4p}\right]}{\mathsf{e}\left[\prod_{n=0}^{n(t)}\exp{\left\{4kp{\left\lvertb^h\right\rvert}_{0;[\tau_n,\tau_{n+1}\wedge t]}^{1/(1-\alpha)}\right\}}\right]}\right)^{1/4 } , \end{gathered}\ ] ] with @xmath182 .
since the jumps of @xmath64 are jointly independent and do not depend on @xmath180 , which has a poisson distribution , we have @xmath183}<\infty$ ] and @xmath184 } = \exp{\left\{({\mathsf{e}\left[g(\delta l_{\tau_1})^{4p}\right]}-1)\pi({{\mathbb r}})t\right\}}\\ = \exp{\left\{\left(\int_{{{\mathbb r}}}g(y)^{4p}\pi(dy)-1\right)\pi({{\mathbb r}})t\right\}}<\infty.\end{gathered}\ ] ] further , write @xmath185 } = { \mathsf{e}\left[{\mathsf{e}\left[j_{b}^{4p(n(t)+1)}\,\middle\vert\ , l\right]}\right]}\end{gathered}\ ] ] from the formula for the solution to , obtained in the proof of theorem [ thm_ex_sol ] , we get that @xmath186 , where @xmath187 is certain non - random measurable function .
thus , we can write @xmath188 , where @xmath189 is a non - random bounded function . therefore , since @xmath1 and @xmath2 do not depend on @xmath64 , we obtain @xmath190 } = { \mathsf{e}\left[\sup_{0\le s\le t\le t } { \left\lvert\int_s^\cdot g(u , w , b^h , l ) dw_u\right\rvert}_{\alpha;[s , t]}^{4pk}\right ] } \bigg|_{l = l , k= n(t)+1}.\end{gathered}\ ] ] abbreviate @xmath191 and estimate @xmath192}\\\le { \left\vert\int_0^t g_u(l ) dw_u\right\vert } + { \left\vert\int_0^s g_u(l ) dw_u\right\vert } + \int_s^t { \left\vert\int_u^t g_z(l ) dw_z\right\vert}(t - u)^{-1-\alpha } du.\end{gathered}\ ] ] then @xmath193}^{4pk}\right]}\le 2 ( i_1 + i_2),\end{gathered}\ ] ] where @xmath194 } { \left\vert\int_0^t g_u(l ) dw_u\right\vert}^{4pk}\right ] } \le { \mathsf{e}\left[\left(\int_0^tg_u(l)^2 du\right)^{2pk}\right]}\le k_1^k,\\ i_2 = { \mathsf{e}\left[\sup_{0\le s\le t\le t } \left(\int_s^t { \left\vert\int_u^t g_z(l ) dw_z\right\vert}(t - u)^{-1-\alpha } du\right)^{4pk}\right]},\end{gathered}\ ] ] with some constant @xmath195 independent of @xmath196 .
the term @xmath197 is estimated as in the proof of theorem [ moments - mixed ] : from the garsia rodemich rumsey inequality we get for any @xmath153 , @xmath154 $ ] @xmath198 with @xmath199 now for @xmath200 @xmath201}\le c_{\eta}^{4pk}t^{2 - 4pk\eta } \int_0^t \int_0^t \frac{{\mathsf{e}\left[{\left\vert\int_x^y g_v(l ) dw_v\right\vert}^{4pk}\right ] } } { { \left\vertx - y\right\vert}^{2pk}}dx\,dy\\ \le c_{\eta}^{4pk}t^{2 - 4pk\eta } \int_0^t \frac{{\mathsf{e}\left[\left(\int_x^y g_v(l)^2 dv\right)^{2pk}\right ] } } { { \left\vertx - y\right\vert}^{2pk } } dx\,dy \le k_2^{k},\end{gathered}\ ] ] where the constant @xmath202 is independent of @xmath196 .
consequently , @xmath203 collecting all estimates , we get @xmath193}^{4pk}\right]}\le k_3^{k}\end{gathered}\ ] ] with @xmath204 independent of @xmath196 .
therefore , @xmath185 } \le { \mathsf{e}\left[k_3^{n(t)+1}\right]}<\infty.\end{gathered}\ ] ] consider the last multiple in .
denote @xmath205}^{1/(1-\alpha)}$ ] . by the hlder inequality , @xmath206}\le \prod_{n=0}^{n(t ) } { \mathsf{e}\left[e^{q_n z_n}\right]}^{1/q_n}= \prod_{n=0}^{n(t)}{\mathsf{e}\left[{\mathsf{e}\left[e^{q_n z_n}|l\right]}\right]}^{1/q_n},\end{gathered}\ ] ] where @xmath207 using the independence of @xmath2 and @xmath64 , write @xmath208 } = { \mathsf{e}\left[\exp{\left\{4kps ( b - a)^{-\kappa } { \left\lvertb^h\right\rvert}_{0;[a , b]}^{1/(1-\alpha)}\right\}}\right]}\big\vert_{a= \tau_n , b=\tau_{n+1}\wedge t , s=\sum_{n=0}^{n(t ) } \delta_n^{\kappa}}.\ ] ] in view of the self - similarity property of @xmath2 , it is easy to check that @xmath209}^{1/(1-\alpha)}\overset{d}{=}{\left\lvertb^h\right\rvert}_{0;[0,1]}^{1/(1-\alpha)}.\ ] ] consequently , @xmath206}\le { \mathsf{e}\left[{\mathsf{e}\left[\exp{\left\{4kp s { \left\lvertb^h\right\rvert}_{0;[0,1]}^{1/(1-\alpha)}\right\}}\right ] } \big\vert_{s=\sum_{n=0}^{n(t ) } \delta_n^{\kappa}}\right]}\end{gathered}\ ] ] it is easy to check that @xmath210
. therefore , @xmath211 this implies @xmath206}\le { \mathsf{e}\left[\exp{\left\{4kp t^{\kappa } { \left\lvertb^h\right\rvert}_{0;[0,1]}^{1/(1-\alpha)}\right\}}\right]},\end{gathered}\ ] ] which is finite as @xmath212}$ ] is a finite supremum of a gaussian family and @xmath148 .
the proof is now complete .
bai , l. and ma , j. _ stochastic differential equations driven by fractional brownian motion and poisson point process _ , preprint ( 2012 ) , arxiv : math.pr/1206.2710 , available at http://arxiv.org/pdf/1206.2710v1.pdf .
y. hu and d. nualart . _ differential equations driven by hlder continuous functions of order greater than 1/2 _ , in _ stochastic analysis and applications
_ , _ abel symp .
_ , vol . 2 .
springer , berlin , 2007 , pp .
399413 .
m. l. kleptsyna , p. e. kloeden , and v. v. anh . _ existence and uniqueness theorems for stochastic differential equations with fractal brownian motion _ , problemy peredachi informatsii 34 ( 1998 ) , no . 4 , pp .
5161 .
y. s. mishura and g. m. shevchenko .
_ stochastic differential equation involving wiener process and fractional brownian motion with hurst index @xmath214 _ , comm .
theory methods 40 ( 2011 ) , no .
1920 , pp . 34923508 .
y. s. mishura and g. m. shevchenko .
_ mixed stochastic differential equations with long - range dependence : existence , uniqueness and convergence of solutions_. comput .
64 ( 2012 ) , no . 10 , pp . 31173127 | in this paper , we consider a stochastic differential equation driven by a fractional brownian motion ( fbm ) and a wiener process and having jumps . we prove that this equation has a unique solution and
show that all moments of the solution are finite .
fractional brownian motion ; wiener process ; poisson measure ; stochastic differential equation ; moments primary 60g15 ; secondary 60g22 , 60h10 , 60j65 . |
the behavior of qcd at finite temperature is of great theoretical and experimental interest .
traditionally , finite temperature calculations of a field theory are performed in euclidean space with periodic boundary conditions imposed in the time direction .
however , the same information can be extracted from a minkowski space calculation by examining the density of states in the spectrum as a function of energy . in particular , an exponentially increasing density of states
implies a finite critical temperature @xmath1 @xcite .
there is some experimental evidence of this exponential increase in the hadron spectrum @xcite .
recently , there has been renewed interest in the lightcone formulation of thermal field theory in minkowski space @xcite . in this letter
we investigate the aspects of a lightcone formulation of qcd relevant for finite temperature and , in particular , its implications for the string picture of qcd . to leading order of the @xmath2 expansion of qcd @xcite , we compute the density of glueball states per unit mass via the transverse lattice approach to non - abelian gauge theory @xcite .
this is a hamiltonian method that combines light - front quantization in the two `` longitudinal '' spacetime directions , @xmath3 , with a lattice in the remaining `` transverse '' directions , @xmath4 .
it has been successfully used to compute the properties of the lightest glueballs and mesons ( for a review , see ref .
@xcite ) . in these calculations ,
effective hamiltonians on coarse lattices were tuned to optimize lorentz covariance of the lightest boundstates .
the method also yields the spectrum of heavier boundstates .
a rapid exponential rise in the density of states suggests that there may be enough states at relatively low masses , where the coarse lattice calculation is still accurate for aggregate quantities , to observe this behaviour unambiguously .
using the detailed methods described in refs .
@xcite , we demonstrate that in an intermediate mass range , @xmath5 times the lightest glueball mass , the density of states does indeed show a degree of universality with respect to cutoffs .
we find an exponentially increasing density of states , implying a finite - temperature hagedorn transition .
although the method is not well - suited to producing an accurate prediction of the critical temperature , we make a numerical estimate and produce analytic bounds on it .
one application of interest is to the string picture of large-@xmath6 gauge theory .
results from previous euclidean lattice computations @xcite indicate a @xmath1 lower than that of the nambu - goto string model @xcite , suggesting that a qcd string in the large-@xmath6 limit has more worldsheet degrees of freedom than this free bosonic string .
our calculations suggest that this is due at least in part to longitudinal oscillations .
this is also consistent with the nambu - goto string description appearing in a distinct , unstable phase of transverse lattice gauge theory discovered by klebanov and susskind @xcite .
the transverse lattice is formulated in the following manner .
two coordinates @xmath7 , @xmath8 , are continuous while two directions @xmath9 are discretized as a lattice , spacing @xmath10 .
the longitudinal continuum gauge potentials @xmath11 lie at sites @xmath12 .
the transverse flux link fields @xmath13 , @xmath14 , lie on the link between @xmath12 and @xmath15 , representing the gauge fields polarized in the @xmath12 directions . in general
there are also fermi fields @xmath16 , but we will not need these for the glueball spectrum at large-@xmath6 . for a coarse transverse lattice ,
the strategy is to perform a color - dieletric expansion @xcite of the most general lightcone gauge hamiltonian , renormalisable with respect to the continuum coordinates @xmath7 , in powers of @xmath17 . provided these link fields are chosen sufficiently heavy , one can truncate this expansion to study the low lying hadron boundstates dominated by just a few particles of these fields .
the remaining couplings in the effective hamiltonian can then be accurately constrained by optimizing symmetries broken by the cutoffs in low energy observables .
typically this method is viable in a window between small lattice spacings , where the fields become too light to justify the color - dieletric expansion , and large lattice spacings where lorentz covariance breaks down uncontrollably . for completeness
, we give the lagrangian , used in ref .
@xcite , that we shall employ here .
it contains all allowed terms up to order @xmath18 for the large-@xmath6 limit : @xmath19[\mbox { h. c. } ] \right\ } \nonumber \\ & & + { \beta \over n_c a^{2 } } { { \rm \ , tr \!}}\left\ { m_{1 } m_{2 } m_{1}^{{\dagger } } m_{2}^{{\dagger } } \right\ } + \mbox{h .
c. } \nonumber \\ & & - \frac{1}{a^{2 } } \sum_r\left [ \frac{\lambda_1}{n_c } { { \rm \ , tr \!}}\left\ { m_r m_r^{{\dagger } } m_r m_r^{{\dagger } } \right\ } \right . \nonumber \\ & & \left
. + \frac{\lambda_2}{n_c } { { \rm \ , tr \!}}\left\ { m_r m_r m_r^{{\dagger } } m_r^{{\dagger } } \right\ } + \frac{\lambda_3 } { n_{c}^2 } \left ( { { \rm \ , tr \!}}\left\ { m_r m_r^{{\dagger } } \right\ } \right)^2\right ] \nonumber\\ & & - { \lambda_4 \over a^{2 } n_c } \sum_{\sigma=\pm 2 , \sigma^\prime = \pm 1 } { { \rm \ , tr \!}}\left\ { m_\sigma^{{\dagger } } m_\sigma m_{\sigma^\prime}^{{\dagger } } m_{\sigma^\prime } \right\ } \nonumber\\ & & - { 4 \lambda_5 \over a^{2 } n_{c}^2 } { { \rm \ , tr \!}}\left\ { m_1 m_1^{{\dagger } } \right\}{{\rm \ , tr \!}}\left\ { m_2 m_2^{{\dagger } } \right\ } \ ; . \label{pot1 } \label{ferlag}\end{aligned}\ ] ] where @xmath20 is the continuum field strength in the @xmath21 planes at each site @xmath12 .
we have defined @xmath22 and hold @xmath23 finite as @xmath24 . in the @xmath24 limit ,
glueball states consists of connected loops of transverse flux links and are absolutely stable to decay .
the reader is referred to ref .
@xcite for details of the construction of the lightcone hamiltonian and fock space , the renormalisation , and the determination of various couplings appearing in eqn .
( [ ferlag ] ) .
these couplings are accurately constrained by optimizing covariance of low - lying glueball eigenfunctions and rotational invariance of the heavy - source potential .
the relatively large number of couplings means the hamiltonian is highly ` improved ' and can give cut - off independent results on quite coarse lattices . taking the fundamental scale to be the string tension @xmath25 , refs .
@xcite investigated individual glueball masses up to about 1.5 times the lightest glueball mass and the method was shown to produce accurate results , even for lattice spacings of order @xmath26 .
we investigate here whether any part of the higher excited spectrum , probed at finite temperature , is similarly under control .
@xmath27 + @xmath28 @xmath27 + @xmath28
we employ a dlcq cutoff @xmath29 in the @xmath30 direction @xcite to obtain a finite fock space basis at finite transverse lattice spacing @xmath10 .
therefore , we must study the stability of results under variation of both @xmath29 and @xmath10 . figs .
[ fig1 ] and [ fig2 ] display the distribution of glueball masses @xmath31 , in units of the string tension @xmath32 , for the @xmath33 state in the spectrum .
the results in fig.[fig1 ] are shown for fixed @xmath34 , the lattice spacing at which the best covariance of the low - lying spectrum was obtained for lagrangian ( [ ferlag ] ) , and increasing values of the dlcq cutoff @xmath35 , where @xmath36 is the limit of infinite dimensional fock space .
the largest basis at @xmath37 contains 654,948 states in total , though only the lowest 20,000 are plotted for file storage reasons . at the largest masses displayed , the dlcq cutoff is introducing non - universal artifacts because it cuts off the maximum number of @xmath38-particles in a state . at lower masses
however , there is universal behavior of the slope of the distribution .
we repeated the calculation for @xmath37 at four different lattice spacings , shown in fig.[fig2 ] , the couplings of ( [ ferlag ] ) re - optimized for covariance of the low - lying spectrum at each . again , for the range of masses plotted , the slope is relatively universal .
the approximately universal linear growth observed in this region implies that the density of states per unit mass , @xmath39 , grows exponentially with mass @xmath38 @xmath40 this exponential behaviour establishes itself already in the first 100 or so states , occuring at less than twice the lightest glueball mass , which is within the expected range of validity of the method @xcite .
it is perhaps surprising from the figures that the universal growth appears to extend to much higher masses .
one would expect the individual highly excited glueball masses to shift around relative to their true values due to cutoff dependence . however , an aggregate quantity like the density of states appears to be less sensitive to cutoff dependence .
it is well known @xcite that exponential growth of the density of states leads to a diverging canonical partition function for temperature @xmath41 . to determine an accurate value for @xmath1
, one needs to know the density for asymptotically high masses @xmath38 , when power corrections to the expression ( [ exp ] ) at lower masses can be neglected . in principle ,
knowledge of the functional form of power corrections could enable one to extract @xmath1 accurately from lower masses , but in practice the form is unknown for qcd .
therefore , although exponential growth is clearly occuring , we can only extract a rough estimate for @xmath1 from the numerical data , based on linear fits to figs.[fig1 ] and [ fig2 ] .
there are three systematic sources of error in our estimate .
the first two come from the variation of the data in figs .
[ fig1 ] and [ fig2 ] , which we take as indicative of the error due to residual dependence on the cutoffs @xmath29 and @xmath10 .
the third comes from the linear fits themselves , since power corrections to eqn .
( [ exp ] ) in the actual data will make the fit result for @xmath1 depend upon the range of masses chosen for the fit .
those intervals are naturally limited above and below : for all the data , smooth linear behaviour does not start until the block of points beginning at @xmath42 ; dlcq artifacts limit the maximum useful value of @xmath43 , roughly @xmath44 for the @xmath35 data respectively . within these limits
, we performed linear fits to the data sets in the figures for a variety of different mass intervals .
the maximum variation of @xmath1 with interval chosen was of order @xmath45 , which we take as indicative of the error made by using simplified fitting function ( [ exp ] ) .
as a result , we estimate @xmath46 , where the first error is due to @xmath29 , second due to @xmath10 , and third from the fits
. we can also deduce analytic bounds on @xmath1 . in ref .
@xcite , we derived in transverse lattice gauge theory an approximate analytic formula for the number @xmath47 of highly excited glueball states of mass squared @xmath48 , @xmath49 integer .
the derivation was based on the assumption that the glueball masses were governed only by longitudinal dynamics of the transverse link fields and a fixed number of such links in a state . assuming a transverse degeneracy factor @xmath50 for the number of distinct states with @xmath51 links , the formula of ref.@xcite trivially generalises to @xmath52}\ } \ , \label{analytic}\ ] ] where @xmath53 is the dilogarithm function .
this leads to a hagedorn temperature @xmath54 } \ , \label{tc}\ ] ] the maximum degeneracy one could expect is from a long free chain of @xmath51 links , which on a square transverse lattice gives @xmath55 .
this provides an upper bound on the density of states and hence a lower bound @xmath56 .
conversely , setting @xmath57 corresponds to neglecting transverse structure altogether .
this is a lower bound on the density of states , implying @xmath58 is an upper bound .
the numerical solution investigated above is somewhere between these extremes ; in reality there are non - trivial longitudinal _ and _ transverse dynamics . recent large-@xmath6 extrapolations of euclidean lattice computations @xcite are more suited to providing accurate determinations of the critical temperature , obtaining @xmath59 .
this is less than the result , @xmath60 , for the non - interacting nambu - goto string @xcite .
this would suggest that the large-@xmath6 gauge theory flux string has more worldsheet degrees of freedom than a simple bosonic one , which has only transverse oscillations . in making this comparison based on @xmath1 with a free string , it is important to use results from the large-@xmath6 limit , since only in this limit do lightcone flux strings not split and join . at finite @xmath6 , by examining only @xmath1 one can not separate physically distinct world - sheet effects from string interaction effects proportional to @xmath61 .
also , the hagedorn determination of @xmath1 directly from the density of states is strictly correct only for a gas of free hadrons / strings .
although euclidean lattice results provide an accurate value for @xmath1 , they do not explain why it differs from the bosonic string value .
transverse lattice gauge theory can provide an explanation with the help of an older analysis by klebanov and susskind @xcite .
they noted that choosing a sufficiently tachyonic link field mass @xmath62 ( see eqn.[ferlag ] ) would drive the theory into a new phase , where glueball states were unstable to becoming infinitely long chains of transverse links , each link with frozen longitudinal dynamics . in this phase
, there are only transverse dynamics ( this is the complete opposite of the scenario used above to derive the bounds ) .
the spectrum of this phase is exactly that of the nambu - goto string @xcite .
thus , at least some of the extra degrees of freedom present in the qcd phase of transverse lattice gauge theory , which can lower the critical temperature , appear to be longitudinal oscillations of the flux string .
et al . _ ,
b * 521 * , 33 ( 2001 ) ; + s. mattiello _
_ , few body syst .
* 31 * , 159 ( 2002 ) ; + v. s. alves , a. das and s. perez , phys .
d * 66 * , 125008 ( 2002 ) ; + h. a. weldon , phys .
d * 67 * , 085027 ( 2003 ) ; _ ibid .
_ 128701 ( 2003 ) ; + a. das and x. zhou , phys .
d * 68 * , 065017 ( 2003 ) ; + a. n. kvinikhidze and b. blankleider , phys .
d * 69 * , 125005 ( 2004 ) ; + j. raufeisen and s. j. brodsky , hep - th/0408108 y. nambu , proc .
conf . on symm . and quark model , wayne state univ .
1969 , gordon and breach , london 1970 ; phys .
d*10 * , 4262 ( 1974 ) ; + l. susskind , phys .
d*1 * , 1182 ( 1970 ) ; + t. goto , prog .
* 46 * , 1560 ( 1971 ) . | numerical computations are performed and analytic bounds are obtained on the excited spectrum of glueballs in @xmath0 gauge theory , by transverse lattice hamiltonian methods .
we find an exponential growth of the density of states , implying a finite critical ( hagedorn ) temperature .
it is argued that the nambu - goto string model lies in a different universality class . |
the investigation of two - dimensional ( 2d ) qcd in the context of the schwinger - dyson ( sd ) dynamical equations of motion has been initiated by the pioneering paper of t hooft @xcite .
he used the free gluon propagator in the light - cone gauge , which is free from ghost complications .
he used also the large @xmath3 ( the number of colors ) limit technique in order to make the perturbation ( pt ) expansion with respect to @xmath4 reasonable . in this case the planar diagrams are reduced to quark self - energy and ladder diagrams , which can be summed .
the bound - state problem within the bethe - salpeter ( bs ) formalism was finally obtained free from the infrared ( ir ) singularities .
the existence of a discrete spectrum only ( no continuum in the spectrum ) was demonstrated . since this pioneering paper 2d qcd continues to attract attention ( see , for example , review @xcite and recent papers @xcite and references therein ) . despite its simplistic vacuum structure
it remains a rather good laboratory for the modern theory of strong interactions , which is four - dimensional ( 4d ) qcd @xcite . in our previous publications
@xcite we have investigated 2d qcd in the arbitrary covariant gauge for the first time . in these works a new , nonperturbative ( np ) solution ( using neither large @xmath3 limit technique explicitly nor a weak coupling regime , i.e. , ladder approximation ) to 2d qcd in the covariant gauge
is obtained in the context of the above - mentioned sd equations , complemented by the corresponding slavnov - taylor ( st ) identities .
it is well known , however , that covariant gauges , in general , are complicated by the ghost contributions .
nevertheless , we have shown that ghost degrees of freedom can be considerable within our approach @xcite .
the ghost - quark sector contains a very important piece of information on quark degrees of freedom themselves through the corresponding quark st identity .
this is just the information which should be self - consistently taken into account . in this way
a close set of equations has been derived for the quark propagator @xcite .
the main purpose of this letter is to exactly solve the obtained system of equations in the chiral limit and to develop analytical methods of its solution in the general case , i.e. , for the nonzero current quark masses .
let us emphasize in advance that we have found a general solution for the heavy quark propagator , which by no means could be reduced to the free one .
all of this will provide the necessary basis for future numerical calculations as well .
the final system of equations , obtained in ref .
@xcite for the quantities in the quark sector , are presented by the quark sd equation and the quark st identity as follows ( euclidean signature ) :
@xmath5 for simplicity , here we remove an overbar from the definitions of the renormalized green s functions , retaining it only for the coupling constant @xmath6 ( which has the dimensions of mass ) in order to distinguish it from initial ( `` bare '' ) coupling constant .
it contains all known finite numerical factors .
@xmath7 is obviously the proper quark - gluon vertex at zero momentum transfer .
the euclidean version of our parametrization of the quark propagator is as follows : @xmath8 .
it is convenient to introduce the dimensionless variables and functions as @xmath9 .
performing further some tedious algebra of the @xmath10 matrices in 2d euclidean space , the system ( 2.1 ) can be explicitly reduced to a system of a coupled , nonlinear ordinary differential equations of the first order for the @xmath11 and @xmath12 quark propagator form factors , namely @xmath13 where @xmath14 , @xmath15 , and the prime denotes the derivative with respect to the euclidean dimensionless momentum variable @xmath16 . for the dimensionless current quark mass ,
we introduce the notation @xmath17 .
the formal exact solution of the system ( 2.2 ) for the dynamically generated quark mass function is @xmath18 and @xmath19 is the constant of integration .
not losing generality , it can be fixed as @xmath20 , where @xmath21 is some constant momentum squared , and @xmath22 with @xmath23 then the equation determining the @xmath11 function becomes @xmath24 as was emphasized in refs.@xcite , the important observation is that the formal exact solution ( 2.3 ) exhibits the algebraic branch point at @xmath25 , which completely @xmath26 at any finite point on the real axis in the @xmath16-complex plane whatever the solution for the @xmath11 function might be .
thus the solution for the quark propagator can not be presented as the expression having finally a pole - type singularity at any finite point @xmath27 ( euclidean signature ) , i.e. , @xmath28 certainly satisfies thereby the first necessary condition of quark confinement , formulated at the fundamental quark level as the absence of a pole - type singularity in the quark propagator @xcite .
it is well known that such kind of unphysical singularity ( algebraic branch point at @xmath25 ) is due to the inevitable ghost contributions in the covariant gauge qcd . however , as was explained in refs .
@xcite , it will not cause any problems within our approach in order to calculate truly np quantities , such as quark condensate .
the absence of a pole - type singularities in the quark propagator as a criterion of confinement at the microscopic level is only first necessary condition .
the second sufficient condition of this criterion , formulated at the macroscopic ( hadron ) level , is the existence of a discrete spectrum only ( no continuum in the spectrum ) in the bound - state problem within the corresponding bs formalism @xcite .
its discussion is obviously beyond the scope of the present letter .
from a coupled system of the differential equations ( 2.2 ) it is easy to see that this system @xmath29 , @xmath30 and @xmath31 , @xmath32 thus any nontrivial solution automatically breaks the @xmath33 invariance of the quark propagator , and therefore @xmath34 leads to the spontaneous chiral symmetry breakdown at the fundamental quark level ( @xmath35 , dynamical quark mass generation ) . in all previous investigations
a chiral symmetry preserving solution always exists . for simplicity , we do not distinguish between @xmath12 and @xmath36 , calling both dynamically generated quark mass functions .
a few remarks are in order .
a nonzero , dynamically generated quark mass function defined by condition ( 2.8 ) is the order parameter of dbcs at the fundamental quark level . at the phenomenological level
the order parameter of dbcs is the nonzero chiral quark condensate determined as @xmath37 within our approach ( see ref .
@xcite and @xmath38 is explicitly given below in eq .
( 2.11 ) ) . in general , it can be formally zero , even when the mass function is definitely nonzero .
thus the nonzero , dynamically generated quark mass is a much more appropriate condition of dbcs than the quark condensate .
one may say that this is the first necessary condition of dbcs , while the nonzero chiral quark condensate is the second sufficient one
. in the chiral limit ( @xmath39 ) the system ( 2.2 ) can be solved exactly .
the solution for the @xmath11 function is @xmath40 it has thus the correct asymptotic properties ( see fig .
it is regular at small @xmath16 and asymptotically approaches the free propagator at infinity ( @xmath41 ) , which can be formally achieved by the two ways : @xmath42 at fixed @xmath43 and/or by @xmath44 as well .
let us note that the last limit is known as the pt one .
for the dynamically generated quark mass function @xmath12 the exact solution is @xmath45 where @xmath46 is an arbitrary constant of integration and @xmath47 is some constant momentum squared for the chiral limit case .
it is regular at zero .
in addition , it also has algebraic branch points at @xmath48 and at infinity ( at fixed @xmath49 , i.e. , when @xmath43 is fixed ) . as in the general ( nonchiral ) case ,
these unphysical singularities are caused by the inevitable ghost contributions in the covariant gauges ( for general behavior of this solution see fig .
as was mentioned above , @xmath50 automatically has a correct behavior at infinity ( it does not depend on the constant of integration , since it was specified in order to get regular at zero solution ) . in the pt limit ( @xmath44 )
, the constant of the integration @xmath49 and the variable @xmath16 go to infinity uniformly ( @xmath51 ) , so the dynamically generated quark mass function ( 2.11 ) identically vanishes in this limit , in accordance with the vanishing current light quark mass in the chiral limit .
obviously , we have to keep the constant of integration @xmath49 in eq .
( 2.11 ) arbitrary but finite in order to obtain a regular at zero point solution .
the problem is that if @xmath52 , then the solution ( 2.11 ) does not exist at all at any finite @xmath16 , in particular at @xmath53 .
this is valid , of course , for the general solution ( 2.3 ) as well .
let us formulate and develop now the calculation scheme , which gives the solution of the system ( 2.2 ) step by step in powers of the light nonzero current quark masses , as well as in the inverse powers of the heavy quark masses .
for this purpose , it is much more convenient to start from the ground system itself , eqs .
( 2.2 ) , rather than to investigate the general solution ( 2.3 ) . for this purpose ,
let us rewrite the ground system ( 2.2 ) as follows : @xmath54 as was mentioned above , we are interested in the solutions which are @xmath55 and asymptotically approach the free quark case at infinity . because of our parametrization of the quark propagator , its asymptotics have to be determined as follows ( euclidean signature ) : @xmath56 , and neglecting @xmath57 in the denominators for light quarks .
the ground system ( 3.1 ) is very suitable for numerical calculations .
let us now develop the above - mentioned analytical formalism , which makes it possible to find solution of the ground system ( 3.1 ) step by step in powers of the light ( @xmath58 ) nonzero current quark masses , the so - called chiral perturbation theory at the fundamental quark level .
for this purpose it is convenient to present the quark propagator form factors @xmath59 and @xmath60 as follows : @xmath61 where it is formally assumed that @xmath62 . substituting these expansions into the ground system ( 3.1 ) and omitting some tedious algebra , one obtains @xmath63 and for @xmath64 , one obtains @xmath65 where @xmath66 is is obvious that the system ( 3.3 ) describes the ground system ( 3.1 ) in the chiral limit ( @xmath67 ) .
as we already know , it can be solved exactly ( see below as well ) . the first nontrivial correction in powers of small @xmath68
is determined by the following system , which follows from eqs .
( 3.4 ) and ( 3.5 ) , and it is @xmath69 where we omit the dependence on the argument @xmath16 , for simplicity . in the similar way can be found the system of equations to determine terms of order @xmath70 in the solution for the quark propagator and so on .
let us present a general solution to the first of eqs .
( 3.4 ) , which is @xmath71 it is always regular at zero , since all @xmath72 are regular as well .
the advantage of the developed chiral perturbation theory at the fundamental quark level is that each correction in the powers of small current quark masses is determined by the corresponding system of equations which can be formally solved exactly .
let us write down the system of solutions approximating the light quark propagator up to first correction , i.e. , @xmath73 this system is @xmath74 @xmath75 and @xmath76 @xmath77.\ ] ] in physical applications we also need @xmath78 , so we have @xmath79 + ... , \end{aligned}\ ] ] and the relation between constants of integration @xmath49 and @xmath80 remains , in general , arbitrary .
however , there exists a general restriction , namely @xmath81 and it should be real , which may lead to some bounds for the constants of integration , while @xmath82 always remains valid . for heavy quarks ( @xmath83 )
it makes sense to replace @xmath84 . in this case
it is convenient to find solution for heavy quark form factors @xmath59 and @xmath60 as follows : @xmath85 and for heavy quark masses it is formally assumed that @xmath86 , i.e. , the inverse powers are small . substituting these expansions into the first equation of the ground system ( 3.1 ) and omitting some tedious algebra , one finally obtains @xmath87 and @xmath88 in the same way , by equating terms at equal powers in the inverse of heavy quark masses , from second of the equations of the ground system ( 3.1 ) , one finally obtains @xmath89 and @xmath90 where @xmath91 are again given by eqs .
( 3.5 ) . solving eqs .
( 3.17 ) and taking into account eq . ( 3.15 ) , one obtains @xmath92 so the final system to be solved further becomes @xmath93 it is possible to show that all odd terms are simply zero , i.e. , @xmath94 . the explicit solutions for a few first nonzero terms are @xmath95 @xmath96 @xmath97 thus our solutions for the heavy quark form factors look like @xmath98 where @xmath99 and @xmath100 where @xmath101 summing up , one obtains @xmath102 in terms of the euclidean dimensionless variables , the quark propagator is @xmath103 using our solutions , obtained above , it can be written down as follows : @xmath104 where @xmath105 is nothing else but the free quark propagator with the substitution @xmath106 , i.e. , @xmath107 since @xmath108 is of order @xmath109 , then eq . ( 3.30 ) , becomes @xmath110 i.e. , it is reduced to the free quark propagator up to terms of order @xmath111 .
it is instructive to show explicitly that our solution ( 3.32 ) possesses the heavy quark flavor symmetry @xcite .
we will show that the quark propagator to leading order ( up to terms of order @xmath2 ) in the inverse powers of the heavy quark mass will not depend on it , i.e. , it is a manifestly flavor independent to the leading order of this expansion . for this purpose , we must take into account that the argument @xmath16 , which is the dimensionless momentum of the heavy quark , contains itself the heavy quark mass @xmath1 .
in other words , a standard heavy quark momentum decomposition should be used , namely @xmath112 as well as @xmath113 where @xmath114 is the four - velocity with @xmath115 ( euclidean signature ) .
it should be identified with the four - velocity of the hadron .
the `` residual '' momentum @xmath116 is of dynamical origin . in these terms
the euclidean dimensionless dynamical momentum variable @xmath117 then becomes @xmath118 where we denote @xmath119 with @xmath120 and @xmath121 .
substituting these expressions into the eq .
( 3.31 ) and taking into account only the leading order term in the inverse powers of @xmath1 , one finally obtains @xmath122 where @xmath123 which is exactly the heavy quark propagator @xcite .
thus our propagator does not depend on @xmath1 to leading order in the heavy quark mass limit , @xmath124 , i.e. , in this limit it possesses the heavy quark flavor symmetry , indeed .
it is easy to understand that the chiral perturbation theory at the fundamental quark level developed for light quarks in subsection a completely coincides with the general solution ( 2.3 ) , complemented by eqs . (
2.4 ) , ( 2.5 ) and ( 2.6 ) .
we use these equations in the chiral limit as input information in the expansions ( 3.2 ) .
however , things are not so straightforward in the case of heavy quarks . developing the chiral perturbation theory in the inverse powers of the heavy quark masses in subsection b
, we do not use the general solution ( 2.3 ) , only the system ( 2.2 ) itself . in this subsection
we will show that the general solution ( 2.3 ) , complemented by eqs .
( 2.4 ) , ( 2.5 ) and ( 2.6 ) , for the heavy quark mass function possesses much more information than the direct solution of the system ( 2.2 ) on account of the expansions ( 3.14 ) provides at all . starting from the expansion ( 3.14 ) for the @xmath11 function , which contributes into the quark wave function renormalization only , and using exact eqs .
( 2.5 ) and ( 2.6 ) , it is possible to show explicitly that it is determined by the solution ( 3.28 ) , i.e. , it is @xmath125 where the @xmath126 function is given in eq . ( 3.25 ) .
thus in the case of the @xmath11 function the straightforward solution of the initial system ( 2.2 ) completely coincides with exact solution , indeed .
unfortunately , things are not so simple for the heavy quark mass function @xmath12 , which should be determined with the help of the exact solution ( 2.3 ) , on account of the solution ( 4.1 ) . substituting it first into the relation ( 2.5 ) and then using the relation ( 2.4 ) and doing some tedious algebra , one finally obtains @xmath127 where @xmath128 so the general solution ( 2.3 ) becomes @xmath129 where @xmath130 the first two integrals can be explicitly expressed in terms of the corresponding integral exponential function @xmath131 , so one has @xmath132 \bigr ] \nonumber\\ & + & 2 \bar m_q^2 e^{-2x } \bigl [ -{e^{2c } \over 2(c+\bar m_q^2)^2 } - { e^{2c } \over ( c+\bar m_q^2 ) } + 2e^{-2\bar m_q^2 } ei(2(c+\bar m_q^2 ) ) \nonumber\\ & + & { e^{2x } \over 2(x+\bar m_q^2)^2 } + { e^{2x } \over ( x+\bar m_q^2 ) } - 2e^{-2\bar m_q^2 } ei(2(x+\bar m_q^2 ) ) \bigr ] + \bar d_b(c , \bar m_q;x).\end{aligned}\ ] ] it is convenient to separate the dependence on the constant of integration @xmath19 , so after some algebra one obtains @xmath133 where @xmath134e^{2(c - x)}.\end{aligned}\ ] ] evidently , by no means the exact solution ( 4.7 ) can be reduced to the free quark propagator solution ( 3.28 ) for the @xmath12 function , except , maybe , of the asymptotical regime ( se below ) .
it is regular at zero .
it identically vanishes in the pt limit @xmath44 , as it should be , since in this case @xmath135 uniformly ( see general solution ( 2.3 ) ) . in the heavy quark mass infinite limit ( @xmath124 and @xmath43 fixed )
, the quark momentum also goes to infinity ( i.e. , @xmath136 see , for example eq . ( 3.33 ) ) . in this case
the constant of integration @xmath19 remains finite , and therefore the composition @xmath137 also vanishes .
using further the asymptotics of the integral exponential function @xmath138 as follows : @xmath139 , \quad z \rightarrow \infty,\ ] ] from eq .
( 4.7 ) one obtains @xmath140 choosing negative sign in the square root , one finally obtains @xmath141^{1 \over 2},\ ] ] where @xmath142 does not depend on @xmath19 and its explicit expression is not important here .
thus , only in the uniform limit @xmath143 the heavy quark propagator may become the free one up to the composition @xmath144 , similar to eqs .
( 3.28 ) .
it is instructive to present explicitly a few first terms of the expansion ( 3.14 ) for the @xmath12 function by substituting it directly into the general solution ( 2.3 ) , on account of the known already expansion for the @xmath11 function . omitting all the tedious algebra one obtains @xmath145 b_0^{-1}(x),\ ] ] and all odd terms are zero . in order to reproduce the free quark propagator case
, one has to go again to the limit @xmath136 and fixed @xmath19 . neglecting then the exponentially suppressed terms and choosing the negative sign in front of the square root , i.e. , @xmath146 , one obtains @xmath147 and so on .
summing up , one gets the free quark propagator solution ( 3.28 ) , indeed .
however , even in this case there is a solution with opposite sign , corresponding apparently to the free heavy anti - quark propagator .
concluding , let us remind that it is a general feature of nonlinear systems , like the initial system ( 2.2 ) , that the number of independent solutions is not fixed @xmath148 .
we have shown that the quark propagator in 2d covariant gauge qcd reveals several desirable and promising features , so our conclusions are : 4 ) .
we develop an analytical formalism , the so - called chiral perturbation theory at the fundamental quark level , which allows one to find solution for the quark propagator in powers of the light quark masses .
each correction satisfies the differential equation , which can be formally solved exactly .
we develop also an analytical formalism , which allows one to find solution for the quark propagator in the inverse powers of the heavy quark masses .
it coincides with the free quark propagator up to terms of order@xmath149 .
so this solution automatically possesses the heavy quark flavor symmetry up to terms of order @xmath150 .
we have proved that the exact solution for the @xmath11 function coincides with the direct solution of the initial system ( 2.2 ) , obtained by using the above - mentioned expansions in the inverse powers of the heavy quark masses .
7 ) . at the same time
, the exact solution ( 4.7 ) for the heavy quark mass function @xmath12 by no means can be reduced to the free heavy quark propagator .
so it is not coincided with the solution obtained by the expansions in the inverse powers of the heavy quark masses .
there is no doubt left that by using the expansions in the inverse powers of the heavy quark masses from the very beginning , we are loosing some piece of the important information on the structure of the nonlinear system ( 2.2 ) itself .
so its direct solution , obtained by using the straightforward expansions in the inverse powers of the heavy quark masses , has a particular character .
evidently , such straightforward solution does not take into account the response of the vacuum , which determines the modification of the quark propagator , while the exact solution ( 4.7 ) does take this response into the consideration .
the general solution ( 2.3 ) does not demonstrate the principal difference in the analytical structure for light and heavy quarks propagators .
also at the fundamental quark level the heavy quark mass limit is not lorentz covariant .
that is why in the case of heavy quarks we will use the general solution ( 4.7 ) rather than solutions ( 3.30 ) and ( 3.36 ) . to take into account
the vacuum s response is important even for heavy quarks .
our approach makes it possible to calculate physical observables from first principles .
all results will depend only on the renormalized coupling constant ( which has the dimensions of mass ) and the corresponding constant of integration .
a physically well - motivated scale - setting scheme is only needed to fix them .
our general conclusion is that 2d covariant gauge qcd implies quark confinement and dynamical breakdown of chiral symmetry without explicitly involving some extra degrees of freedom .
the only dynamical mechanism responsible for them , which can be thought of in 2d covariant gauge qcd , is the direct interaction of massless gluons @xcite .
this interaction is a main dynamical effect not only in 2d qcd but in 4d qcd as well , i.e. , in qcd itself .
however , to directly generalize the quark confinement mechanism of 2d covariant gauge qcd to 4d qcd is impossible .
the problem is that in former theory the coupling constant , having the dimensions of mass , plays the role of a mass gap , which was introduced and discussed by jaffe and witten ( jw ) in ref .
@xcite . in latter theory the coupling constant is dimensionless , and there is none of the characteristic scales in qcd lagrangian .
if qcd confines then such a characteristic scale is very likely to exist , and it is not @xmath151 , of course , which can be considered as responsible for the nontrivial pt dynamics in qcd ( scale violation , asymptotic freedom ) . a possible way how to introduce a mass gap responsible for the nontrivial np dynamics in qcd has been described in ref .
its possible relation to the above - mentiond jw mass gap is also discussed there .
g. t hooft , nucl .
b 75 ( 1974 ) 461 .
e. abdalla , m. c. b. abdalla , phys .
265 ( 1996 ) 253 .
a. armoni , y. frishman , j. sonnenschein , nucl .
b 596 ( 2001 ) 459 .
u. trittman , nucl .
b 587 ( 2000 ) 311 .
k. aoki , k. ito , phys .
d 60 ( 1999 ) 096004 .
w. marciano , h. pagels , phys . rep .
c 36 ( 1978 ) 137 .
v. gogohia , phys .
b 531 ( 2002 ) 321 .
v. gogokhia , gy .
kluge , phys .
d 66 ( 2002 ) 056013 .
g. preparata , phys .
d 7 ( 1973 ) 2973 .
h. georgi , in : perspectives in the standard model , proc .
tasi-91 , ( boulder , colorado , 1991 ) , eds .
ellis et.al .
, ( ws , singapore , 1992 ) . m. neubert , phys .
245 ( 1994 ) 260 .
a. jaffe , e. witten , quantum yang - mills theory , | we have investigated a closed set of equations for the quark propagator , which has been obtained earlier within a new , nonperturbative approach to two - dimensional covariant gauge qcd .
it is shown that this theory implies quark confinement ( the quark propagator has no poles , indeed ) , as well as dynamical breakdown of chiral symmetry ( a chiral symmetry preserving solution is forbidden ) .
the above - mentioned set of equations can be exactly solved in the chiral limit .
we develop an analytical formalism , the so - called chiral perturbation theory at the fundamental quark level , which allows one to find solution for the quark propagator in powers of the light quark masses .
each correction satisfies the differential equation , which can be formally solved .
we develop also an analytical formalism which allows one to find solution for the quark propagator in the inverse powers of the heavy quark masses .
it coincides with the free heavy quark propagator up to terms of order @xmath0 , where @xmath1 is the heavy quark mass .
so this solution automatically possesses the heavy quark flavor symmetry up to terms of order @xmath2 . at the same time
, we have found a general solution for the heavy quark propagator , which by no means can be reduced to the free one . |
in this article , the term cdw is used to indicate a periodic modulation of the charge density @xmath2 , irrespective of the process behind its generation .
this modulation is described , in the unperturbed system and for the simplest form of cdw , as @xmath3 , where @xmath4 is the intensity and @xmath5 the characteristic wavelength .
perturbations to cdws have been studied extensively @xcite , but most studies are concerned either with uniform perturbations ( e.g. an external electric field ) or point - like perturbations ( e.g. pinning by defects ) , and often consider one - dimensional models , appropriate for quasi - one - dimensional materials , where the coherence length in the perpendicular directions is smaller than the atomic distance .
what is considered here , instead , is the effect of a localized but extended perturbation of typical length scale similar to the cdw wavelength , acting on a material where the coherence length is macroscopic in more than one dimension . starting from the standard fukuyama - lee - rice model @xcite for cdw
, the charge modulation is described through a ginzburg - landau theory as a classical elastic medium .
the complex order parameter @xmath6 is considered , taking into account the amplitude degree of freedom @xmath7 , as well as the phase @xmath8 . the charge density can be expressed as @xmath9 , so for the unperturbed system @xmath10 and @xmath11 are constant ( at least locally ) , and the free energy reads : @xmath12=\int\left[-2f_0{\left|\psi({\mathbf{r}})\right|}^2+f_0{\left|\psi({\mathbf{r}})\right|}^4+\kappa{\left|\nabla\psi({\mathbf{r}})\right|}^2\right]\mathrm{d}{\mathbf{r}},\ ] ] where @xmath13 sets the energy scale and @xmath14 represents the elastic energy contribution . considering the external perturbation by the afm tip , generally described as a potential @xmath15 coupling to the charge density , adds an interaction term to the free energy : @xmath16=\int\left[v({\mathbf{r}})\mathrm{re}(\psi({\mathbf{r}})e^{i{\mathbf{q}}\cdot{\mathbf{r}}})\right]\mathrm{d}{\mathbf{r}}.\ ] ] a well studied case is the impurity one @xcite , where @xmath17 . in that case
phase oscillations are often enough to describe the ground state of the system , which comes from the balance of elastic and potential energy .
these point - like perturbations , however , only impose a likewise point - like constraint on the order parameter , and therefore can not lead to a phase slip in the absence of an external driver @xcite .
i will therefore consider the case where @xmath15 has a finite width of the order of the wavelength @xmath5 , and minimize the total free energy @xmath18 given a specific shape of @xmath15 .
following in the steps of the impurity model and considering only phase perturbations , the functional would be @xmath19=\int\left[\kappa{\left|\nabla\phi({\mathbf{r}})\right|}^2+v({\mathbf{r}})\rho_0\cos({\mathbf{q}}\cdot{\mathbf{r}}+\phi({\mathbf{r}}))\right]\mathrm{d}{\mathbf{r}}.\ ] ] this model , however , is inadequate to describe this specific effect .
in fact , considering a purely one - dimensional model would lead to a linear behavior of @xmath8 where the potential is zero . since we expect a decay far from the perturbation ,
this is a clearly unphysical result .
moreover , due to the nature of the phase , which is defined modulo @xmath20 , given some boundary conditions the solution is not univocally defined unless the total variation of @xmath8 is also specified . assuming the phase to have the unperturbed value @xmath21 far from the perturbation
, we can define the integer _ winding number _
@xmath22 of a solution as the integral @xmath23 taken along the cdw direction @xmath24 ( with @xmath25 typically representing the unperturbed case ) .
since any change in the winding number along the @xmath24 direction would extend to the whole sample and unnaturally raise the energy of such a solution , to recover a physical result one needs to take into account the amplitude degree of freedom , which will allow for the presence of dislocations and local changes in the winding number .
for these reasons , the minimization will be performed in two dimensions , with the complete complex order parameter and in a subspace with a defined winding number .
the final result is expected to be similar to what previously considered in the wider context of phase slip @xcite and more specifically in the case of localized phase slip centers @xcite .
namely , the local strain induced by the perturbation on the phase will reduce the order parameter amplitude , to the point where a local phase slip event becomes possible . in more than one dimension
, the boundary between areas with different winding number will be marked by structures such as _
vortices_. from this preliminary analysis ,
the mechanism responsible for the dissipation peaks can be understood : as the tip approaches the surface , it encounters points where the energies of solutions with different winding number undergo a crossover . at these points
the transition between manifolds is not straightforward , due to the mechanism required to create the vortices , therefore the oscillations lead to jumps between different manifolds , resulting in hysteresis for the tip and ultimately dissipation .
to asses the validity of the proposed mechanism , numerical simulations of the tip - surface interaction were performed .
a two - dimensional model is considered , since it takes into account the relevant effects while keeping the simulation simpler ; moreover , the experimental substrate has a quasi - two - dimensional structure , so that volume effects can be expected to be negligible . differently from the experimental system
@xcite , the cdw is characterized by a single wave vector @xmath24 , leading to a simpler order parameter and a clearer effect . to represent the effect of the tip , the shape of a van der waals potential
@xmath26 is integrated over a conical tip at distance @xmath27 from the surface .
the result can be reasonably approximated in the main area under the tip by a lorentzian curve @xmath28 where @xmath29 is the distance in the plane from the point right below the tip and the parameters are found to scale like @xmath30 and @xmath31 .
free energy @xmath32 as a function of tip distance @xmath27 for subspaces with different winding number @xmath22 .
results from simulations on a 201@xmath33201 grid with parameters ( see text ) @xmath34 ev / nm , @xmath35 ev , @xmath36 nm@xmath37 , @xmath38 ev@xmath39 nm , @xmath40 nm@xmath37 and boundary conditions @xmath41 . insets : charge density @xmath2 ( full lines ) and phase @xmath8 ( dashed lines ) along the line passing right below the tip ( indicated by the vertical dashed line ) for different winding number @xmath22 , at the positions indicated by the dots on the energy curves . ] knowing the shape of the perturbation , the total free energy @xmath18 is minimized numerically on a square grid of points with spacing much smaller than the characteristic wavelength of the cdw , imposing a constant boundary condition @xmath42 on the sides perpendicular to @xmath24 , while setting periodic boundary conditions in the other direction to allow for possible phase jumps .
the minimization is carried out through a standard conjugated gradients algorithm @xcite .
the parameters employed are order of magnitude estimates of the real parameters , which reproduce the relevant experimental effects in a qualitative fashion .
the charge density @xmath2 and phase @xmath8 profile along the line passing right below the tip is shown in the insets of fig .
[ fig : encurves ] for solutions with different winding number @xmath22 .
the position of the attractive tip is indicated by the vertical dashed line , and the phase slip occurred under it , leading to and increase in charge density , is clearly visible .
parallel lines far from the tip always revert to the @xmath25 manifold , through the creation of vortices .
the main curves in fig . [ fig : encurves ] represent the minimum energy at defined @xmath22 as a function of the distance @xmath27 .
since the solution with a given winding number lies in a local minimum , it is possible to use the minimization algorithm to find solutions in a certain subspace , even when this is not the global minimum for a specific distance , by starting from a reasonable configuration ( namely , the minimum at close distance ) .
two different crossing points can be seen , which would give rise to two peaks in the experimental dissipation trace .
of course a more complex cdw configuration or different parameters would give rise to more peaks .
force as a function of distance for evolutions with @xmath43 nm , @xmath44 nm and different values of @xmath45 with @xmath46 ev@xmath39s .
inset : total work @xmath47 as a function of oscillation frequency @xmath45 . ] to completely justify the validity of the dissipation mechanism , we need to look into the dynamics of the cdw , to guarantee that the evolution through a crossing point does not lead to immediate relaxation between different manifolds . to do this ,
the time evolution of the system was simulated , following the time - dependent ginzburg - landau equation @xcite @xmath48 this equation can be interpreted as an overdamped relaxation , with a coefficient @xmath49 , of the order parameter towards the equilibrium position . integrating this equation ( through a standard runge - kutta algorithm @xcite ) , the force as a function of the distance can be computed for a tip performing a full oscillation perpendicular to the surface according to the law @xmath50 .
[ fig : fcurves ] shows the force evolution during such oscillations at different frequencies . as we can see
the tip suffers a hysteresis even at low frequencies , since the decay from one manifold to the other happens far from the crossing point .
the area of the loops represents directly the dissipated energy per cycle @xmath47 , as reported in the inset .
based on these simulations , the theory hereby presented is shown to be compatible with the experimental findings : it accounts for the existence of multiple peaks , their appearance far away from the surface and their nature being related to the cdw structure of the sample .
it is interesting to notice that the vortices appearing in our simulations have been described before in the context of cdw conduction noise @xcite , where their creation and movement justifies the phase slip near the cdw boundaries . in this sense
, this theory lies in between these macroscopic effect and the simple one - dimensional model of defect pinning and phase slip @xcite , as is appropriate for a localized but extended perturbation .
to conclude , i have presented a mechanism to explain peaks in the dissipation of a tip oscillating at specific distances above a cdw surface : these occur around instability points corresponding to the crossing of energy levels characterized by different winding numbers .
numerical simulations in a system of reduced complexity support the validity of this mechanism
. it would be interesting to investigate the same effect in other systems displaying cdw or even spin density waves , as well as systems where the origin of charge modulation is related to the fermi surface and not to other effects .
_ acknowledgements _ the author thanks his collaborators g.e . santoro and e. tosatti .
he acknowledges research support by snsf , through sinergia project crsii2 136287/1 , by erc advanced research grant n. 320796 modphysfrict , and by miur , through prin-2010llkjbx_001 vanossi a. , manini n. , urbakh m. , zapperi s. , and tosatti e. , rev .
phys . * 85 * 529 ( 2013 ) .
langer m. , kisiel m. , pawlak r. , pellegrini f. , santoro g.e .
, buzio r. , gerbi a. , balakrishnan g. , baratoff a. , tosatti e. , and meyer e. , nature mater .
* 13 * 173177 ( 2014 ) .
grner g. , rev .
phys . * 60 * 1129 ( 1988 ) .
fukuyama h. , and lee p.a .
b * 17 * 535 ( 1978 ) .
lee p.a . , and rice t.m . ,
b * 19 * 3970 ( 1979 ) .
tucker j.r .
b * 40 * 5447 ( 1989 ) .
ttt i. , and zawadowski a. , phys .
b * 32 * 2449 ( 1985 )
. maki k. , phys .
a * 202 * 313 ( 1995 ) .
gorkov l.p .
, zh . eksp .
86 * 1818 ( 1984 ) .
mcmillan w.l .
b * 12 * 1187 ( 1975 ) .
press w.h . ,
teukolsky s.a .
, vetterling w.t . , and flannery b.p .
_ numerical recipes : the art of scientific computing ( 3rd ed . )
_ , cambridge university press ( 2007 ) .
, verma g. , and maki k. , phys .
lett . * 52 * 663 ( 1984 ) . | a mechanism is proposed to describe the occurrence of distance - dependent dissipation peaks in the dynamics of an atomic force microscope tip oscillating over a surface characterized by a charge density wave state .
the dissipation has its origin in the hysteretic behavior of the tip oscillations occurring at positions compatible with a localized phase slip of the charge density wave .
this model is supported through static and dynamic numerical simulations of the tip surface interaction and is in good qualitative agreement with recently performed experiments on a nbse@xmath0 sample .
the study of the microscopic mechanisms leading to energy dissipation and friction has very important theoretical and practical implications . in recent years
, experiments have started to single out the effects of microscopic probes in contact or near contact with different surfaces , and much theoretical effort has been devoted to the full understanding of such experiments @xcite . in particular ,
the minimally invasive non - contact experiments offer a chance to investigate delicate surface properties and promise to bring new insight on localized effects and their interaction with the bulk .
recently , a non - contact atomic force microscopy ( afm ) experiment @xcite on a nbse@xmath0 sample has shown dissipation peaks appearing at specific heights from the surface and extending up to @xmath1 nm far from it .
these peaks were obtained with tips oscillating both parallel and perpendicular to the surface , and in a range of temperatures compatible with the surface charge density wave ( cdw ) phase of the sample . in this paper ,
a model is proposed explaining in detail the mechanism responsible for these peaks : the tip oscillations induce a charge perturbation in the surface right under the tip , but , due to the nature of the cdw order parameter , multiple stable charge configurations exist characterized by different `` topological '' properties .
when the tip oscillates at distances corresponding to the crossover of this different manifolds , the system is not allowed to follow the energy minimum configuration , even at the low experimental frequencies of oscillation , and this gives rise to a hysteresis loop for the tip , leading to an increase in the dissipation .
while the idea behind this dissipation mechanism has been proposed by the author and collaborators in the original paper @xcite , this article expands on the technical aspects of the model , highlighting details to appear in a future publication . |
there has been considerable amount of work on learning when labeling is expensive , such as techniques on transductive inference and active learning . with the emergence of crowdsourcing services , like amazon mechanical turk , labeling costs in many applications
have dropped dramatically .
large amounts of labeled data can now be gathered at low price . due to a lack of domain expertise and misaligned incentives ,
however , labels provided by crowdsourcing workers are often noisy . to overcome the quality issue ,
each item is usually simultaneously labeled by several workers , and then we aggregate the multiple labels with some manner , for instance , majority voting .
an advanced approach for label aggregation is suggested by dawid and skene[1 ] .
they assume that each worker has a latent confusion matrix for labeling .
the off - diagonal elements represent the probabilities that a worker mislabels an arbitrary item from one class to another while the diagonal elements correspond to her accuracy in each class .
worker confusion matrices and true labels are jointly estimated by maximizing the likelihood of observed labels .
one may further assume a prior distribution over worker confusion matrices and perform bayesian inference [ 2][3][4 ] .
the method of dawid - skene ( 1979 ) implicitly assumes that a worker performs equally well across all items in a common class . in practice , however , it is often the case that one item is more difficult to label than another . to address this heterogeneous issue ,
zhou et al.(2012)[5 ] propose a minimax entropy principle for crowdsourcing .
it results in that each item is associated with a latent confusion vector besides a latent confusion matrix for each worker .
observed labels are determined jointly by worker confusion matrices and item confusion vectors through an exponential family model .
moreover , it turns out that the probabilistic labeling model can be equivalently derived from a natural assumption of objective measurements of worker ability and item difficulty
. such kinds of objectivity arguments have been widely discussed in the literature of mental test theory [ 6][7 ] .
all the above approaches are for aggregating multiclass labels and in many scenarios , the labels are ordinal .
( 2014)[8 ] proposed a work to aggregate votes using minimax conditional entropy for ordinal labels .
most of the methods use statistical methods to aggregate the observed labels by transforming them to some probability or entropy measures .
but , there has been no work that operates directly on the observed labels .
we present a method to learn the aggregates of the votes using clustering .
we first show the formulation that allows us to use clustering as an approximation of the vote aggregation stratagem .
we first draw a parallel between the restricted boltzmann machine ( rbm ) learning and the expectation maximization ( em ) algorithm of the david - skene algorithm and then show that gaussian - softmax rbms[9 ] can be approximated by a gaussian mixture model ( gmm ) , whose specific conditions lead to a direct mapping to the traditional k - means algorithm[10][11 ] .
to then elucidate the clustering paradigm , we perform clustering using the @xmath0-rbm model as proposed in [ 14 ] .
the restricted boltzmann machine is a bipartite , undirected graphical model with visible ( observed ) units and hidden ( latent ) units .
the rbm can be understood as an mrf with latent factors that explains the input visible data using binary latent variables .
the rbm consists of visible data @xmath1 of dimension @xmath2 that can take real values or binary values , and stochastic binary variables @xmath3 of dimension @xmath0 .
the parameters of the model are the weight matrix @xmath4 that defines a potential between visible input variables and stochastic binary variables , the biases @xmath5 for visible units , and the biases @xmath6 for hidden units . when the visible units are real - valued , the model is called the gaussian rbm , and its joint probability distribution can be defined as follows : @xmath7 @xmath8 where z is a normalization constant . the conditional distribution of this model can be written as : @xmath9 @xmath10 where @xmath11 is the sigmoid function , and @xmath12 is a gaussian distribution . here , the variables in a layer ( given the other layers ) are conditionally independent , and thus we can perform block gibbs sampling in parallel .
the rbm can be trained using sampling - based approximate maximum - likelihood , e.g. , contrastive divergence approximation [ 12 ] .
after training the rbm , the posterior ( equation 2 ) of the hidden units ( given input data ) can be used as feature representations for classification tasks .
we define the gaussian - softmax rbm as the gaussian rbm with a constraint that at most one hidden unit can be activated at a time given the input , i.e. , @xmath13 .
the energy function of the gaussian - softmax rbm can be written in a vectorized form as follows : @xmath8 subject to @xmath13 .
for this model , the conditional probabilities of visible or hidden units given the other layer can be computed as : @xmath14 @xmath15 where @xmath16 is the @xmath17-th column of the @xmath18 matrix , often denoted as a `` basis '' vector for the @xmath17-th hidden unit . in this model , there are @xmath19 possible configurations ( i.e. , all hidden units are 0 , or only one hidden unit @xmath20 is 1 for some @xmath17 ) .
the k - means clustering is an unsupervised algorithm that assigns clusters to data points .
the algorithm can be written as * randomly choose k cluster centers , @xmath21 .
* assign an incoming data point @xmath22 to the nearest cluster center @xmath23 * @xmath24 becomes centroid of the cluster .
@xmath25 the procedure is repeated till convergence , that is all points have been assigned the best cluster centers and over many trials @xmath26 , we take the best possible solution as the cluster assignment .
the gaussian mixture model is a directed graphical model where the likelihood of visible units is expressed as a convex combination of gaussians .
the likelihood of a gmm with @xmath27 gaussians can be written as follows : @xmath28 for the rest of the paper , we denote the gmm with shared spherical covariance as gmm(@xmath29 ) , when @xmath30 for all @xmath31 . for the gmm with arbitrary positive definite covariance matrices
, we will use the shorthand notation gmm(@xmath32 ) .
k - means can be understood as a special case of gmm with spherical covariance by letting @xmath33 [ 13 ] .
compared to gmm , the training of k - means is highly efficient ; therefore it is plausible to train k - means to provide an initialization of a gmm . then the gmm is trained with em algorithm .
the em algorithm learns the parameters to maximize the likelihood of the data as described by equation [ gmm ] .
the following section outlines the proof for vote aggregation as a special case of clustering problems , when trying to model the problem using rbms .
rbms are learned in a manner so as to minimize the negative log - likelihood of the data . in a vote aggregation setup ,
the data is the observed labels .
thus , we can see that learning from rbms are similar to aggregating votes from the dawid - skene algorithm which also minimizes the negative log - likelihood of the observed labels .
but in the training of rbms , we often encounter the normalization constant @xmath34 , which is intractable and this makes it difficult to train an rbm , and we need to approximate @xmath34 to learn the ideal parameters for the same . hence , it becomes difficult to directly apply rbms to aggregate votes . in this section , we show that a gaussian rbm with softmax hidden units can be converted into a gaussian mixture model , and vice versa .
this connection between mixture models and rbms with a softmax constraint completes the chain of links between k - means , gmms and gaussian - softmax rbms and helps us to visualize vote aggregation as a clustering problem .
as equation [ gsrbm ] shows , the conditional probability of visible units given the hidden unit activations for gaussian - softmax rbm follows a gaussian distribution . from this perspective
, the gaussian - softmax rbm can be viewed as a mixture of gaussians whose mean components correspond to possible hidden unit configurations . in this section ,
we show an explicit equivalence between these two models by formulating the conversion equations between gmm(@xmath35 ) with @xmath27 gaussian components and the gaussian - softmax rbm with @xmath0 hidden units .
* proposition * the mixture of @xmath36 gaussians with shared spherical covariance of @xmath37 is equivalent to the gaussian - softmax rbm with k hidden units .
we prove the following conversions by showing the following conversions .
* 1 . from gaussian - softmax rbm to gmm(@xmath29 ) : * we begin by decomposing the chain rule : @xmath38 where @xmath39 since there are only a finite number of hidden unit configurations , we can explicitly enumerate the prior probabilities : @xmath40 if we define @xmath41 , then we have @xmath42 .
in fact , @xmath43 can be calculated analytically , @xmath44 using this definition , we can show the equivalence as , @xmath45 * 2 . from gmm(@xmath35 ) to gaussian - softmax rbm : * we will also show this by construction .
suppose we have the following gaussian mixture with @xmath27 components and the shared spherical covariance @xmath37 : @xmath46 this gmm can be converted to a gaussian - softmax rbm with the following transformations : @xmath47 @xmath48 @xmath49 it is easy to see that the conditional distribution @xmath50 can be formulated as a gaussian distribution with mean @xmath51 , which is identical to that of the gaussian - softmax rbm .
further , we can recover the posterior probabilities of the hidden units given visible units as the follows : @xmath52 therefore , a gaussian mixture can be converted to an equivalent gaussian rbm with a softmax constraint .
gmms learn a density function over the data , while trying to maximize its likelihood . from maximum likelihood estimation , the equation
a gmm tries to learn is @xmath53 .
but since we do not know @xmath54 s , we maximize the marginal likelihood , which is given by @xmath55 , where @xmath56 is the number of clusters . from the gaussian bayes classifier , @xmath57 , that is , @xmath58
p(y = i)\ ] ] when @xmath59 is spherical , with same @xmath60 for all classes , @xmath61 $ ] .
if each @xmath22 belongs to one class @xmath62 , marginal likelihood is given by : @xmath63\ ] ] for estimating the parameters , we maximize the log - likelihood with respect to all clusters and this gives , @xmath64\ ] ] equivalently , minimizing the negative log - likelihood gives , @xmath65\ ] ] which is the same as the k - means function . we thus show that the vote aggregation methodolody when applied from an rbm model perspective can be viewed as a clustering problem , one of k - means specifically .
thus , we can consider vote aggregation learned by maximizing the likelihood of observed labels to be a clustering problem .
our framework uses @xmath0 component rbms .
each component rbm learns one non - linear subspace .
the visible units @xmath66 , @xmath67 correspond to an @xmath68 dimensional visible ( input ) space and the hidden units @xmath69 correspond to a learned non - linear @xmath70-dimensional subspace .
the @xmath0-rbm model has @xmath0 component rbms .
each of these maps a set of sample points @xmath71 to a projection in @xmath72 .
each component rbm has a set of symmetric weights ( and asymmetric biases ) @xmath73 that learns a non - linear subspace .
note that these weights include the forward and backward bias terms .
the error of reconstruction for a sample @xmath74 given by the @xmath56th rbm is simply the squared euclidean distance between the data point @xmath74 and its reconstruction by the @xmath56th rbm , computed using .
we denote this error by @xmath75 . the total reconstruction error @xmath76 in any iteration @xmath77 is given by @xmath78 .
the @xmath0 rbms are trained simultaneously . during the rbm training ,
we associate data points with rbms based on how well each component rbm is able to reconstruct the data points .
a component rbm is trained only on the training data points associated with it .
the component rbms are given random initial weights @xmath79 .
as in traditional k - means clustering , the algorithm alternates between two steps : ( 1 ) computing association of a data point with a cluster and ( 2 ) updating the cluster parameters . in @xmath0-rbms , the @xmath80th data point
is associated with the @xmath56th rbm ( cluster ) if its reconstruction error from that rbm is lowest compared to other rbms , i.e. if @xmath81 .
once all the points are associated with one of the rbms the weights of the rbms are learnt in a batch update . in hard clustering
the data points are partitioned into the clusters exhaustively ( i.e. each data point must be associated with some cluster ) and disjointly ( i.e. each data point is associated with only one cluster ) .
in contrast with k - means where the update of the cluster center is a closed form solution given the data association with clusters , in @xmath0-rbms the weights are learned iteratively .
in this section , we report empirical results of our method on real crowdsourced data .
we consider the @xmath82 error metric .
let us denote by @xmath83 the true rating and @xmath84 the estimate .
the error metrics is defined as : ( 1 ) @xmath82 = @xmath85 .
all the research ( code and datasets ) is reproducible and is available at : https://github.com/gupta-abhay/deep-voteaggregate . *
web search relevance rating * the web search relevance rating dataset contains 2665 query - url pairs and 177 workers[5 ] .
give a query - url pair , a worker is required to provide a rating to measure how the url is relevant to the query .
the rating scale is 5-level : perfect , excellent , good , fair , or bad . on average ,
each pair was labeled by 6 different workers , and each worker labeled 90 pairs .
more than 10 workers labeled only one pair .
* dog image labeling * we chose the images of 4 breeds of dogs from the stanford dogs dataset [ 8 ] : norfolk terrier ( 172 ) , norwich terrier ( 185 ) , irish wolfhound ( 218 ) , and scottish deerhound ( 232 ) .
the numbers of the images for each breed are in the parentheses .
there are 807 images in total .
a worker labeled an image at most once , and each image was labeled 10 times .
there are four architectures considered for both the datasets .
we consider two rbms , binary - binary rbms and gaussian - binary rbms .
the architectures are the following : 1 .
binary - binary rbm with 30 visible units and 5 hidden units . 2 .
binary - binary rbm with 18 visible units and 3 hidden units . 3 .
gaussian - binary rbm with 6 visible units and 5 hidden units .
gaussian - binary rbm with 6 visible units and 3 hidden units . 1 .
binary - binary rbm with 40 visible units and 4 hidden units . 2 .
binary - binary rbm with 20 visible units and 2 hidden units . 3 .
gaussian - binary rbm with 10 visible units and 4 hidden units .
gaussian - binary rbm with 10 visible units and 2 hidden units .
we report the results , both @xmath82 and @xmath86 errors of the architectures considered in tables [ web - label ] and [ dog - label ] .
the @xmath82 error of the dawid - skene model on the web search data is 0.17 and the error on the dog data is 0.21 .
.error metrics for web - search data [ cols="^,^,^,^,^",options="header " , ] all the results are done over and average of 20 runs .
we see from the results that the results of one - hot encodings are the best among all the proposed architectures , for both the web and dog datasets
. this can be because rbms capture binary data and thus it is able to capture the one - hot encodings in a good manner .
also , we see that in the web data , when we use gaussian binary rbms , we get 100% error .
this may be because gaussian sampling of the data is not ideal for this dataset .
on trying cd - k above @xmath87 , we get huge reconstruction errors for every data point . however , between cd-1 and cd-2 , cd-2 outperforms cd-1 .
also , pcd gives huge reconstruction errors for the web dataset , but gave results comparable to cd-1 for the dog dataset .
we give a plot for the average reconstruction error per sample as the rbm proceeds for the web dataset in figure [ web - sample ] .
[ 1 ] dawid , a. p. and skene , a. m. maximum likeihood estimation of observer error - rates using the em algorithm .
journal of the royal statistical society , 28(1):2028 , 1979 .
[ 6 ] rasch , g. on general laws and the meaning of measurement in psychology . in proceedings of the 4th berkeley symposium on mathematical statistics and probability , volume 4 , pp .
321 - 333 , berkeley , ca , 1961 .
[ 8 ] zhou , d. , liu , q. , platt , j. and meek , c. aggregating ordinal labels from crowds by minimax conditional entropy . in proceedings of the 31st international conference on machine learning pp .
262 - 270 .
[ 10 ] sohn , k. , jung , d.y . ,
lee , h. and hero iii , a.o . , 2011 ,
efficient learning of sparse , distributed , convolutional feature representations for object recognition . in computer vision ( iccv ) , 2011 ieee international conference on ( pp .
2643 - 2650 ) . ieee .
[ 14 ] chandra , s. , kumar , s. and jawahar , c.v . , 2013 .
learning multiple non - linear sub - spaces using k - rbms . in proceedings of the ieee conference on computer vision and pattern recognition ( pp . 2778 - 2785 ) . | an important way to make large training sets is to gather noisy labels from crowds of non experts .
we propose a method to aggregate noisy labels collected from a crowd of workers or annotators .
eliciting labels is important in tasks such as judging web search quality and rating products .
our method assumes that labels are generated by a probability distribution over items and labels .
we formulate the method by drawing parallels between gaussian mixture models ( gmms ) and restricted boltzmann machines ( rbms ) and show that the problem of vote aggregation can be viewed as one of clustering .
we use @xmath0-rbms to perform clustering .
we finally show some empirical evaluations over real datasets . |
magnetic fields are ubiquitous .
all astrophysical objects are known to have magnetic fields of different magnitudes , e.g . , 1 gauss at the stellar scale to @xmath2 gauss at the galactic scale @xcite .
the origin of such fields ( _ primordial field _ ) is not very clear - there are several competing theories which attempt to describe this @xcite .
however , a finite magnetic field in any physical system undergoes a temporal decay due to the finite conductivity of the medium .
so , for steady magnetic fields to occur in astrophysical bodies , there has to be a mechanism of regeneration of the magnetic fields , which takes place due to the dynamo process @xcite .
most astrophysical bodies are thought to have _ fast _ dynamo operating within themselves ( there are exception to this , e.g. , the moon , venus and mars in our solar system ) resulting into exponential growth of the magnetic fields .
this mechanism requires a turbulent velocity background @xcite [ though non - turbulent velocity fields too can make a seed ( initial ) magnetic field to grow ( for details see @xcite ) , we will not consider such cases here ] . since the dynamo equation , in the linear approximation ( see below ) gives unbounded exponentially growing solutions for the long wavelength ( large scale ) part of the magnetic fields , it is linearly unstable in the low wavenumber limit .
however , one does not see a perpetual growth of magnetic fields in the core of the earth or in the sun .
for example , geomagnetic fields ( @xmath3 1 gauss ) are known to be stable for about @xmath4 years @xcite .
thus , the physically realisable solutions of the dynamo equations can not be unstable in the long time limit .
it is now believed that the non - linear feedback due the lorentz force term in the navier - stokes equation is responsible for the saturation of the magnetic field growth ( see , e.g. , @xcite ) . the study of this problem has already been the subject of previous work by many groups .
for example pouquet , frisch and lorat @xcite studied the connections between the dynamo process and the inverse cascade of magnetic and kinetic energies within a eddy damped quasi - normal markovian approximation .
moffatt @xcite has examined the back reactions due to the lorentz force for magnetic prandtl number @xmath5 by linearising the equations of motion of three - dimensional ( @xmath6 ) magnetohydrodynamics ( mhd ) .
vainshtein and cattaneo @xcite discussed several nonlinear restrictions on the generations of magnetic fields .
et al _ @xcite discussed nonlinear @xmath0-effects within a two - scale approach .
rogachevskii and kleeorin @xcite studied the effects of an anisotropic background turbulence on the dynamo process .
brandenburg examined non - linear @xmath0-effects in numerical simulation of helical mhd turbulence @xcite .
in particular , he examined the dependences of dynamo growth and the saturation field on the magnetic prandtl number @xmath7 ( the ratio of the magnetic- to the kinetic- viscosities ) .
bhattacharjee and yuan @xcite studied the problem in a two - scale approach by linearising the equations of motion .
dynamo mechanism has two competing processes at work : amplification of the magnetic field by the dynamo process and ohmic dissipation due to finite resistivity of the medium concerned .
which one among these two effects will dominate depends on the case in study . in some specific models ,
however , one can analyze this completely .
a good example of such models is the kraichnan - kazantzev dynamo @xcite where the velicity field is assumed to be gaussian - distributed , delta - correlated in time and the magnetic field is governed by the induction equation @xcite . in this model
the statistics of the velocity field is taken to be parity invariant so that the @xmath0-effect is ruled out .
the main results from this model include i)the existence of dynamo in the infinite magnetic reynolds number limit for a particular choice of the variance of the velocity distribution @xcite and ii)the existence of a critical magnetic reynolds number only above which dynamo growth is possible @xcite .
however , not much is known about this when invariance due to parity is broken and when the velocity field is not temporally delta - correlated . in a recent simulations @xcite the authors found , in a model simulation for the solar convection zone , a monotonic increase of the horizontal @xmath0-effect with rotation .
et al _ showed , in numerical simulations , that unless magnetic hyperviscosity is less than a critical value , magnetic fields did not grow @xcite , confirming the existence of a critical magnetic reynolds number ( @xmath8 ) .
our studies generalize the existing results . in this paper
we use a minimal model of @xmath0-effect ( see below ) to study dynamo with @xmath0-effect to calculate the @xmath0 coefficient for arbitrary correlations and viscosities , and ask the following questions : 1 .
do the turbulent dynamo growth and the saturation processes require any turbulent background ? or do they function with arbitrary parity - breaking and fluctuating velocity and initial magnetic field correlations ? for the kinetic and magnetic energies and cascades of appropriate quantities ; if there is no mean magnetic field then the energy spectra is expected to be k41-type - see ref.@xcite . ] .
2 . what is the _ hydrodynamic limit _ ( long wavelength limit ) of the dynamo problem ? by this
we ask how the magnetic field correlations scale in the infra red limit during the initial - growth regime .
3 . can arbitrarily large magnetic viscosity prevent dynamo growth ?
in other words , is there a critical magnetic reynolds number @xmath8 above which the dynamo growth sets in ? to study the above mentioned questions we employ a diagrammatic perturbation theory , which has been highly successful in the contexts of critical dynamics @xcite , driven systems @xcite , etc .
this can be easily extended to higher orders in perturbation expansion and is very suitable for handling continuous kinetic and magnetic spectra , unlike the two - scale approximation .
this was first used to study stationary , homogeneous and isotropic mhd in ref.@xcite .
we use this method to study non - stationary statistical states ( dynamo growth ) which facilitates studies on the hydrodynamic limit of the dynamo problem in a renormalisation group framework .
we use diagrammatic perturbation theory to calculate expressions for the @xmath0 coefficiants for arbitrary background velocity and initial magnetic field correlations and magnetic prandtl number @xmath7 for both the early growth and the late time saturation . with our expressions for @xmath0
we examine the three issues mentioned above .
we investigate these for arbitrary correlations and magnetic prandtl number @xmath7 with no approximations other than the existence a perturbation theory .
our principal results are : * we calculate the @xmath0-coefficients for arbitrary correlations and viscosities . *
we examine the hydrodynamic limit in the kinematic regime and predict the existence of a critical @xmath8 or rotation above which dynamo growth will occur for certain correlations with infra red singularity . in our all our studies , we do not assume any variance for the velocity field .
instead , we use the navier - stokes equation to describe the dynamics of the velocity field .
this allows us to use a renormalisation group framework to study the hydrodynamic limit . the first question that we investigate is phenomenologically very important because different systems may have different velocity and initial magnetic field spectra .
therefore , it is important to understand the dependence of the dynamo on these spectra .
we explicitly demonstrate that the nonlinear feedback of the magnetic fields on the velocity fields in the form of the lorentz force stabilises the growth for arbitrary velocity and initial magnetic field correlations .
this demonstrates that the basic features of the dynamo mechanism are qualitatively independent of the velocity and magnetic field spectra and , essentially , are properties of the @xmath6mhd equations .
details ( e.g. , the values of the @xmath0-coefficients ) of course , depend upon the actual forms of the spectra .
our renormalization group analysis indicates that dynamo growth takes place only if the ekman number @xmath9 ( for a given @xmath8 ) when the velocity and the initial magnetic field spectra are sufficiently singular in the long wavelength limit .
the structure of this paper is as follows : in sec.[dyn ] we discuss the general dynamo mechanism within the standard linear approximation for arbitrary velocity and initial magnetic field correlations and viscosities . in sec.[halt ] we show that beyond the linear approximation non - linear effects lead to the eventual saturation of magnetic field growth for arbitrary background kinetic energy and initial magnetic energy spectra , and viscosities .
we elucidate how different background kinetic energy and initial magnetic energy spectra affect the values of the @xmath0-coefficients . in sec.[renor ] we analyze the initial dynamo growth in a renormalization group framework . we show that for sufficiently singular velocity and magnetic field spectra the ekman number must be @xmath10 1/2 for the magnetic fields to grow . for velocity and magnetic field spectra which go to zero in the long wavelength limit
there are no such restrictions . in sec.[summ ] we present our conclusions .
in the kinematic approximation @xcite , i.e. , in the early - time regime , when the magnetic energy is much smaller than the kinetic energy ( @xmath11 , where @xmath12 and @xmath13 are the velocity and magnetic fields respectively ) the lorentz force term of the navier stokes equation is neglected . in that weak magnetic field limit , which is reasonable at an early time , the time evolution problem for the magnetic fields is a linear problem as the induction equation @xcite is linear in magnetic fields @xmath14 : @xmath15 where @xmath16 is the magnetic viscosity .
the velocity field is governed by the navier - stokes equation @xcite ( in the absence of the lorentz force ) @xmath17 here @xmath18 is the fluid viscosity , @xmath19 an external forcing function , @xmath20 the pressure and @xmath21 the density of the fluid .
we take @xmath19 to be a zero mean , gaussian stochastic force with a specified variance ( see below ) . in a two - scale @xcite
approach one can then write an _ effective _ equation for @xmath22 , the long - wavelength part of the magnetic fields
@xcite : @xmath23 where the _ electromotive force _ @xmath24 .
@xmath25 is the large scale component of the velocity field @xmath26 .
an _ operator product expansion _ ( ope ) is shown to hold @xcite which provides a gradient expansion in terms of @xmath27 for the product @xmath24 @xcite @xmath28 for homogenous and isotropic flows ( @xmath29 ) eq.([ope ] ) gives , @xmath30 which is the standard turbulent dynamo equation . here
@xmath16 now is the _ effective _ magnetic viscosity which includes both the microscopic magnetic viscosity and the turbulent diffusion , represented by @xmath31 in eq.([ope ] ) .
@xmath0 depends upon the statistics of the velocity field ( or , equivalently , the correlations of @xmath19 ) . retaining only the @xmath0 -term and dropping all others from the rhs of eq.([dyna ] ) , the equations for the cartesian components of @xmath27 become ( we neglect the dissipative terms proportional to @xmath32 as we are interested only in the long wavelength properties ) + @xmath33 the eigenvalues of the matrix is @xmath34 . thus depending on the sign of the product @xmath35 , one mode grows and the other decays .
the third mode is unphysical , because the corresponding eigenfunction is proportional to @xmath36 and hence in conflict with @xmath37 .
since growth rate is proportional to @xmath38 and dissipation is proportional to @xmath32 , large scale fields continue to grow leading to long wavelength instability .
thus in the long time limit effectively only the growing mode remains .
growth rate @xmath0 is a pseudo - scalar quantity , i.e. , under parity transformation @xmath39 ,
@xmath40 @xcite .
since @xmath0 depends upon the statistical properties of the velocity field , its statistics should not be parity invariant .
this can happen in a rotating frame , where the angular velocity explicitly breaks reflection invariance .
the navier - stokes ( ns ) ( including the lorentz force ) and the induction equation in an inertial frame in @xmath41 space take the form @xmath42 and @xmath43 here , @xmath44 and @xmath45 are the fourier transforms of @xmath46 and @xmath47 respectively ,
@xmath48 is the projection operator , which appears due to the divergence - free conditions on the velocity and magnetic fields ( we consider incompressible fluid for simplicity ) . equations ( [ nsk ] ) and ( [ indk ] ) have to be supplemented by appropriate correlations of @xmath49 and initial conditions on @xmath50 .
we choose @xmath51 and @xmath52 to have zero mean and to be gaussian distributed with the following variances : @xmath53 @xmath54 where @xmath55 and @xmath56 are some functions of @xmath57 ( to be specified later ) . in a rotating frame with a rotation velocity @xmath58 the eq.([nsk ] ) takes the form @xmath59 whereas eq.([indk ] ) has the same form in the rotating frame .
@xmath60 is the coriolis force .
the centrifugal force @xmath61 is a part of the _ effective pressure_=@xmath62 which does not contribute to the dynamics of incompressible flows .
the bare propagator @xmath63 ( obtained from the linearized version of eq.([nskr ] ) ) of @xmath64 + @xmath65 + such that @xmath66 where @xmath26 is the column vector @xmath67 + one can verify that with the form of the bare propagator given above , an odd - parity part in the velocity auto - correlator @xmath68 appears which is proprotional to the rotation @xmath69 .
notice that @xmath70 is different from @xmath71 - this is just the consequence of the fact that @xmath69 distinguishes the @xmath72-direction as a preferred direction in space , making the system anisotropic .
however for frequencies @xmath73 or length scales @xmath74 ( here @xmath72 is the dynamical exponent ) isotropy is restored . in that regime , to @xmath75 the role of the global rotation is to introduce a non - zero odd - parity part in @xmath76 proportional to @xmath69 .
this can be also seen by noting that in the inertial frame the correlation @xmath77 is of the form @xmath78 [ cf .
eq.([varf ] ) ] where @xmath79 is a scalar function of @xmath57 and hence in the rotating frame the correlator is proportional to @xmath80 where @xmath81 s are appropriate rotation matrices ( we have suppressed the indices ) .
similarly , initial magnetic field correlations , given by eq.([varb ] ) transforms accordingly in the rotating frame . since rotation matrices act on @xmath82 and eq.([varb ] ) in the same way , magnetic field auto - correlator @xmath83 has an odd parity part in the rotating frame with the same sign as the odd parity part in the velocity correlator .
thus the effects of rotation can be _ modeled _ ( to the lowest order ) by introducing parity breaking parts in eqs.([varf ] ) and ( [ varb])@xcite @xmath84 in conjunction with the eqs.([nsk ] ) and ( [ indk ] ) , where @xmath85 is the totally antisymmetric tensor in @xmath6 .
this way of modeling rotation effects is , of course , only approximate , but suffices for our purposes as this explicitly incorporates parity breaking .
one can , however , construct experimental set ups @xcite which are described correctly by eqs.([varbi ] ) .
the parity breaking parts in the noise correlations or initial conditions ensure that the velocity and the initial magnetic field correlators have non - zero odd parity parts , as would happen in a rotating frame .
an important dimensionless number is the _ ekman number _ @xmath86 which can be related to @xmath87 by equating the parity braking parts of the velocity correlator calculated from ( linearized ) eq.([nskr ] ) and eq.([varf ] ) with that from eqs .
( [ nsk ] ) and ( [ varbi ] ) .
this gives @xmath88 .
now , one may ask what is the relative sign between @xmath87 and @xmath89 ?
since the parity breaking parts of the correlators of the velocity and the magnetic fields have same sign and are proportional to @xmath90 respectively , @xmath87 and @xmath89 must have same sign .
as already noted , introduction of parity breaking terms in the force / initial correlations is well - known in the literature , we , nevertheless , give the analysis in details in order to emphasise on the fact that fluid and magnetic helicities must have the same sign .
furthermore , for a complete description of the effects of rotation , in addition to the coriolis force , a forcing with a preferred direction is also required .
we , however , do not include all these details as introduction of parity - breaking correlations is sufficient for our purposes . in this sense , this can be thought of as a _ reduced _ or a _ minimal model _ for dynamo .
one may note that a nonzero kinetic helicity is required for the @xmath0-effect as the @xmath0-coefficient is proportional to the kinetic helocity . even though a global rotation explicity breaks the parity invariance of the system under space reversal ,
rotation alone is not enough to yield a non - zero helicity .
this is because the helicity is pseudo - scalar and , therefore , can be constructed only out of an axial vector ( here , rotation @xmath91 ) and a polar vector . in typical astrophysical settings ,
the latter one could be provided by , say , a density inhomogeneity . even though this is not contained in eq .
( [ nsk ] ) , our minimal model , nevertheless , produces a finite helicity due to the helical nature of the forcing function .
thus , our minimal model is able to capture both the breakdown of parity due to the rotation and the generation of helicity due to the rotation and any other preferred direction . in the kinematic approximation , which neglects the lorentz force term of the navier - stokes equation
, the time evolution of the magnetic fields follows from the linear induction equation ( [ indeq ] ) .
we assume , for the convenience of calculations , that the velocity field ( @xmath26 ) has reached a statistical steady state .
this is acceptable as long as the loss due to the transfer of kinetic energy to the magnetic modes by the dynamo process is compensated by the external drive . in the kinematic ( i.e. , linear ) approximation , we work with the eqs.([nsk ] ) ( without the lorentz force ) and ( [ indk ] ) .
we choose @xmath92 to be a zero - mean , gaussian random field with correlations @xmath93 our initial conditions for the magnetic fields are @xmath94 since we are interested to investigate the dynamo process with arbitrary statistics for the velocity and magnetic fields we work with arbitrary @xmath95 and @xmath96 .
for k41-type spectra , we require @xcite @xmath97 and @xmath98 .
these choices ensure that under spatial rescaling @xmath99 , @xmath100 which is the kolmogorov scaling @xcite . note that both the force correlations in the eq.([nsk ] ) and the initial conditions on eq.([indk ] ) have parts that are parity breaking , in conformity with our previous discussions
we now calculate the @xmath0-term .
we use an iterative perturbative method which is very similar to and discussed in details in ref.@xcite . in this method ,
terms in each order of the perturbation series can be represented by appropriate feynman diagrams @xcite .
even though , for simplicity , we confine ourselves to the lowest order in the perturbation theory ( represented by the _ tree level diagrams _ ) , which is sufficient for our purposes , higher order calculations represented by higher order digrams can be done in a straight forward manner .
below we give the expression for @xmath0 in the kinematic approximation ( which we call the ` direct ' term - responsible for growth ) in the lowest order of the perturbation theory ( see fig.[diag]a ) : @xmath101 from which one can read the @xmath0-term : @xmath102\ ] ] giving @xmath103 for large @xmath104 .
the suffix @xmath105 refers to _ growth _ or the _ direct _ term , as opposed to _ feedback _ which we discuss in the next sec.[halt ] .
the growth term is proportional to @xmath38 and diffusive decay proportional to @xmath32 .
the angular brackets represent averaging over the noise and initial - condition ensembles .
when the magnetic fields become strong , it is no longer justified to neglect the feedback of the magnetic fields in the form of the lorentz force .
so we need to work with the _ full _ eqs.([nsk ] ) and ( [ indk ] ) . the ideas of ope as elucidated in sec.[dyn ] are still valid for the full non - linear problem .
but the value of @xmath0 is expected to change from its value in the linear problem . in presence of the lorentz force
there is an additional contribution to @xmath0 ( fig.1b ) . to evaluate that
, we follow a diagrammatic perturbation approach similar to that described in the previous section . here
also we restrict ourselves to the lowest order only ( i.e. , the tree level diagrams ) though extension to higher orders is straight forward .
we obtain @xmath106 which gives ( @xmath107 refers to feedback ) @xmath108 which , after some simplifications , yields , @xmath109 where @xmath110\tilde{d}_2(q)$ ] is a growing function of time for small wavenumbers . as before ,
angular brackets refer to averaging over noise and initial - condition ensembles .
thus @xmath111 grows in time . since , at any finite time @xmath104 , when the non - linear feedback on the velocity field due to the lorentz force is nolonger negligible , both @xmath112 and @xmath111 are non - zero and we get @xmath113}},\nonumber \\
\alpha_f&=&{2s_3\over 3}{4\over 15}\int { d^3q\over ( 2\pi)^3}{\tilde{d}_2(q , t)q^2 \over { |(\alpha_d+\alpha_f)q|-2\mu q^2 } } , \label{divexp}\end{aligned}\ ] ] with @xmath114\tilde { d}_2(q ) . \label{selfd2}\ ] ] equations ( [ divexp ] ) and ( [ selfd2 ] ) are to be solved self - consistently @xcite .
thus the net growth rate is proportional to @xmath115 for the mode @xmath116 .
the expressions ( [ divexp ] ) have apparent divergences at finite @xmath117 ; so in perturbative calculations one should treat the @xmath0-terms as perturbations which remove these divergences .
this problem is akin to that in kuramoto - shivashinsky equation for flame front propagation @xcite .
so the expressions for @xmath112 and @xmath111 are @xmath118 which do not have any finite wavevector singularity .
expressions ( [ finexp ] ) are obtained , as mentioned before , by truncating the perturbation series at the tree level .
extensions to higher orders are straight forward .
illustrative examples of higher order diagrams have been shown in fig.[oneloop ] .
let us now consider various @xmath57 dependences of @xmath119 and @xmath96 .
when the background velocity field is driven by the navier - stokes equation with a conserved noise ( thermal noise ) one requires that @xmath120 , giving @xmath121 .
if we assume similar @xmath57-dependences for @xmath122 then we require @xmath123 constant and @xmath124 .
these choices yield @xmath125\over 2\mu |q| } , \label{alphathermal}\end{aligned}\ ] ] which remain finite even if the system size diverges .
a fully developed turbulent state , characterised by k41 energy spectra , is generated by @xmath126 and @xmath127 .
in addition if we assume that the initial magnetic fields correlation also have k41 scaling then @xmath128 and @xmath129 .
if one starts with a k41-type initial correlations for the magnetic fields , then at a later time the scale dependence for the magnetic field correlations are likely to remain same ; only the amplitudes grow .
notice that the spectra diverge as wavevector @xmath130 , i.e. , as the system size diverges .
this is a typical characteristic of fully developed turbulence .
for such a system we find @xmath131 the notable difference between the expressions eqs.([alphathermal ] ) and ( [ alphak41 ] ) for the @xmath0 coefficients is that the @xmath0 coefficients diverge with the system size if the energy spectra are singular in the infra red limit ( as in for fully developed turbulence ) .
these divergences are reminiscent of the divergences that appear in critical dynamics @xcite which are handled by renormalisation group methods .
in general , at early times ( small @xmath111 ) , @xmath111 increases exponentially in time .
the growth rate of @xmath111 decreases with time .
since @xmath112 and @xmath111 have different signs , @xmath132 ) and eq.([alphak41 ] ) suggest that the early - time growth and late time saturation of magnetic fields take place for different types of background velocity correlations and initial magnetic field correlations .
therefore dynamo instability and its saturation are rather intrinsic properties of the @xmath6mhd equations with broken reflection invariance .
one may also note that for @xmath133-type of correlations ( singular in the infrared limit ) one has forward cascade of kinetic energy @xcite : this is because energy is fed into the system mostly in the large scale ( i.e. , for small @xmath57 ) whereas , dissipation acts primarily in the small scales ( large @xmath57 ) , resulting into a cascade of energy from the large- to small- scales . on the other hand , for correlations smooth in the infra red limit , there is no such cascade .
these results indicate that the existence of the dynamo mechanism does not require any special background velocity field spectrum , though the value of the @xmath0-coefficient depends upon it .
our results also suggest that these processes may take place for varying magnetic prandtl number @xmath134 .
the above analysis crucially depends on the fact that @xmath111 and @xmath112 have opposite signs , which , in turn , imply that @xmath87 and @xmath89 have same signs .
we have already seen that in a physically realisable situation where parity is broken entirely due to the global rotation , @xmath87 and @xmath89 indeed have the same sign .
in the _ first order smoothing approximation _ @xcite in the kinematic limit , to calculate @xmath135 one considers only the induction equation as @xmath26 is supposed to be given .
however when one goes beyond the kinematic approximation , one has to consider the navier - stokes equation as well .
thus in the first - order smoothing approximation one writes the equations for the fluctuations @xmath26 and @xmath14 as ( to the first order ) @xmath136 and @xmath137 where the ellipsis refer to all other terms in the navier - stokes equation and @xmath138 and @xmath139 are the large scale ( _ mean field _ ) part of the velocity and magnetic fields @xcite . with these we can write @xmath140 here
the ellipsis refer to non-@xmath0 terms in the expansion of @xmath141 ( see eq.(4 ) ) .
thus for isotropic situations @xmath142 $ ] where @xmath143 is a correlation time .
thus @xmath0 is proportional to the difference in the fluid and magnetic _ torsalities_@xcite , ( fluid helicity being the same as fluid torsality and magnetic helicity being proportional to magnetic torsality ) a result obtained in @xcite using other methods and approximations .
note that eqs .
( 21 ) and ( [ finexp ] ) are very similar to but not exactly the one that were obtained in @xcite ( in our notations @xmath87 is proportional to fluid torsality ( or fluid helicity ) and @xmath89 is proportional to magnetic torsality ) .
we ascribe this difference to the essential difference between a two - scale approach and our dirgrammatic perturbation theory which , we believe is more suitable for handling continuous kinetic and magnetic spectra .
we have seen that in eqs.([alphak41 ] ) the @xmath0-coefficients diverge in the hydrodynamic ( @xmath130 ) limit which calls for a renormalisation group ( rg ) analysis as a natural extension of our diagrammatic perturbative calculations . in fully developed @xmath6mhd , in the steady state ,
correlation and response functions exhibit dynamical scaling with the _ dynamic exponent _
@xmath144 @xcite ( for a different approach see @xcite ) , which means renormalised viscosities ( kinetic as well as magnetic ) diverge @xmath145 for a wavenumber @xmath57 belonging to the inertial range . even for decaying mhd with initial k41-type correlations
this turns out to be true @xcite where equal time correlations exhibit dynamical scaling with @xmath144 .
the question is , what it is in the initial transient of dynamo growth ( @xmath146 saturation time ) .
we examine this in a renormalization group framework .
since we are interested in the early growth , we neglect the lorentz force and work with eq.([indk ] ) inconjunction with the initial magnetic field correlations and noise correlation given by eq.([varbi ] ) . as before
, we assume a statistical steady state for the velocity field .
it is well - known that correlations @xmath147 exhibit scaling form @xmath148 where @xmath149 and @xmath72 are the spatial scaling and dynamical exponents respectively @xcite where @xmath150 is a scaling function .
the galilean invariance of the mhd equations constraints these exponents to obey the relation @xmath151 @xcite .
in addition to that , for fully developed turbulence due to non - renormalization of the noise - correlators [ cf .
eq.([varf ] ) ] the exponents are fully determined : @xmath152 , which means the renormalised fluid viscosity diverges as @xmath153 in the limit wavevector @xmath130 . during early growth , equal - time magnetic field correlations @xmath154 are expected to exhibit a scaling form @xmath155 ( @xmath146 saturation time ) where @xmath156 and @xmath157 are the magnetic spatial scaling and dynamical exponents respectively , and @xmath158 is a scaling function .
similar conditions arising from the galilean invariance and non - renormalization of the initial k41-like magnetic field spectrum determines @xmath159 and @xmath160 .
we perform a renormalization group analysis following @xcite .
as mentioned earlier , the @xmath0-term is treated as a perturbation . in a renormalisation - group transformation ,
one integrates out a shell of modes @xmath161 , and and simultaneously rescales length scales , time intervals and fields through @xmath162 .
this has the effect that the nonlinearities are affected only by naive rescaling ( this , a consequence of the galilean invariance of the @xmath6mhd equations , essentially implies that the diagrammatic corrections to the nonlinearties vanish in the long wavelength limit ) .
the variances eq.([varf ] ) , which diverge at low wavenumbers remain unrenormalised and thus affected only by rescaling .
there are however fluctuations corrections to @xmath16 and @xmath112 which we evaluate at the lowest order .
the resulting rg flow equations for @xmath16 and @xmath112 , obtained in a one - loop calculation are @xmath163,\\ \label{flow1 } { d\alpha_d\over dl}&=&\alpha_d[z_b-1+a_2{\tilde{d}_1\over \alpha_d \nu ( \nu+\mu)\lambda^{3 } } ] , \label{flow2}\end{aligned}\ ] ] where @xmath164 are numerical constants
. equations ( [ flow1 ] ) and ( [ flow2 ] ) are similar to those presented in ref.@xcite [ eqs .
( 10.13 ) and ( 10.14 ) ] but not exactly same .
the differences arise mainly ( apart from some detail technical differences in the perturbation theories involed ) from the fact that in ref.@xcite the expressions for the @xmath0-coefficients were derived for a given variance of the velocity field . in contrast , we use the navier - stokes equation , driven by a stochastic force of given variance , in place of a given velocity variance . by substituting the value of the exponents in eqs.([flow1 ] ) and
( [ flow2 ] ) we find _ renormalized _
( i.e. , wavevector dependent ) @xmath165 in the hydrodynamic ( @xmath130 ) limit .
thus in that limit , the effective dynamo equation takes the form @xmath166 where the ellipsis refer to non - linear terms and @xmath167 refers to the growing mode .
thus , in the hydrodynamic limit , there is growth of the magnetic fields only if @xmath168 .
this can happen only if the renormalised magnetic viscosity is less than a critical value , set by @xmath112 , i.e. , the kinetic helicity . in terms of the ekman number
@xmath169 this condition means @xmath1 for anti dynamo , i.e. , no growth , equivalently @xmath9 for growth of the magnetic fields .
this can be achieved in two ways , namely by increasing rotation , keeping the magnetic viscosity ( or the magnetic reynolds number ) constant , or decreasing the magnetic viscosity ( i.e. , increasing the magentic reynolds number ) for a constant rotation .
this conclusions are in good agreement with the numerical results of ref.@xcite .
since renormalised magnetic viscosity increases with its bare ( microscopic ) value , it suggests that bare magnetic viscosity must be less than a critical value for growth to be possible .
thus our rg results qualitatively explain the numerical results of kida _ et al _ @xcite who they found that unless magnetic hyperviscosity was less than a critical value there was no growth ( it can be easily argued that a hypermagnetic viscosity gives rise to a magnetic viscosity in the longer scale and hence their result in effect imposes a critical value of the magnetic viscosity ) . in our model @xmath0-effect
is proportional to @xmath170 which in turn is proportional to the global rotation frequency .
hence our results suggest that @xmath0-effect is likely to grow with increasing rotational speed which is in agreement with the results of ref.@xcite . on the other hand ,
if the background velocity and the initial magnetic field correlators do not have an infra red singularity ( i.e. , when the correlators @xmath171 ) there is no fluctuation correction to the magnetic viscosity and to the @xmath0-coefficient resulting in the fact that the growth term ( @xmath172 ) dominates over the dissipation ( @xmath173 ) for sufficiently small wavenumber @xmath57 , leading to growth even for arbitrarily large magnetic viscosity .
therefore , there is no critical @xmath8 .
thus the effects of the infrared divergences that appear in the expressions for the @xmath0-coefficients [ eq .
( [ alphak41 ] ) ] are quite significant : they indicate , as for the driven diffusive nonequilibrium systems with diverging kinetic coefficients in the hydrodynamic limit @xcite , divergence of time - scales in the hydrodynamic limit . since , the @xmath0-term in eq .
( [ indk ] ) is proportional to wavenumber @xmath57 , the time - scale of growth of the mode with wavenumber @xmath57 is @xmath174 .
this remains true , even in the hydrodynamic limit , for the case when there is no divergence in the @xmath0-coefficients .
in contrast , when the @xmath0-coefficient diverge in the infra red limit , the growth rate changes _ qualitatively _ from its linear dependence on wavenumber @xmath57 in the hydrodynamic limit .
for example , with the the background velocity correlations and the initial magnetic field correlations given by eq .
( [ varf ] ) , the @xmath0 coefficients diverge as @xmath175 in the long wavelength limit .
hence , the effective growth rate is changed to @xmath176 .
a full self - consistent calculation ( when feedback due to the lorentz force can not be neglected ) for the @xmath0-coefficients require simultaneous solutions of the self - consistent expressions for magnetic prandtl number , magnetic- to kinetic- energy ratio and the @xmath0-coefficients which can be handled in our scheme of calculations .
the self - consistent solutions are expected to be influenced by the degree of crosscorrelations between the velocity and magnetic fields @xcite .
so far , we have assumed that both @xmath177 and @xmath119 have the same infra red singularity ( @xmath178 and @xmath179 ) .
this need not be the case always .
however , if @xmath119 is non - singular then @xmath112 does not diverge . as a result ,
the growth rate is just @xmath180 even in the hydrodynamic ( long wavelength ) limit .
effective dissipation , however , will still be @xmath181 and thus it will dominate over @xmath182 growth . therefore , there will be no growth in the hydrodynamic limit .
thus , our analyses suggest that in any fully developed turbulent system with @xmath0-effect , helicity spectrum ( given by @xmath119 ) should be as singular as the kinetic energy spectrum ( given by @xmath177 ) .
in conclusions , we have calculated expressions for the @xmath0-coefficients in a diagrammatic perturbation theory on a minimal model for arbitrary background velocity and initial magnetic field correlations , and fluid and magnetic viscosities .
we show that the parity breaking parts of the velocity and magnetic field variances must have the same sign , which is the case in any physical system .
we explicitly show that the processes of early growth and late - time saturations may take place independent of any special velocity and initial magnetic field correlations .
even though our explicit calculations were done by using simple initial conditions for the calculational convenience , the results that we obtain are general enough and it is apparent that the feedback mechanism is qualitatively independent of the details of the initial conditions and force correlations .
one may note that for one of the force / initial correlations there is no kinetic energy cascade in the conventional sense but we still find dynamo action .
it is quite reasonable to expect that our results should be valid for more realistic initial conditions also . in effect
we have explicitly demonstrated the robustness and generality of the dynamo mechanism and that the dynamo mechanism is an intrinsic property of the @xmath6mhd equations .
we have also shown , within our rg analysis , that the magnetic viscosity should be less than a critical value for growth of magnetic fields a result which was previouly observed in numerical simulations .
we conclude the existence of a critical ekman number for k41-type correlations : we find growth only when @xmath9 , confirming recent numerical results .
this is easily understood in our framework .
the issue of divergent effective viscosities in the inertial range assumes importance as it may help to overcome some of the non - linear restrictions as discussed by vainshtein and cattaneo @xcite . a system of magnetohydrodynamic turbulence in a rotating frame , after the saturation time ( i.e. , after which there is no net growth of the magnetic fields ) belongs to the universality class of usual three - dimensional magnetohydrodynamic turbulence in a laboratory .
this can be seen easily in both the lab and the rotating frames ; the critical exponents characterising the correlation functions can be calculated exactly by using the galilean invariance and noise - nonrenormalisation conditions @xcite . an important question , which remains open for further investigations , is the multiscaling properties of the velocity and the magnetic field structure functions at various stages of the growth of the magnetic fields . in what concerns an experimental observation of our results
, one should add that even though it is not easy to verify our results in an experimental set up , numerical simulations of eqs.([nsk ] ) and ( [ indk ] ) with the variances ( [ varbi ] ) with different @xmath57-dependences can be performed to check these results .
the author wishes to thank j. k. bhattacharjee for drawing his attention to this problem , and r. pandit , j. santos and the anonymous referee for many fruitful comments and suggestions .
the author thanks the alexander von humboldt foundation , germany for financial support .
h.k moffatt , _ magnetic field generation in elctrically conducting fluids _ , cambridge university press , cambridge ( 1978 ) .
p. p. kronberg , _ rep .
phys . _ * 57 * , 325(1994 ) ; r. beck , a. brandenburg , d. moss , a. shokorov , and d. sokoloff , _ annual review of astronomy and astrophysics _ , * 34 * , 155 ( 1996 ) ; k. enqvist , _ int . j. mod
. phys . _ * d7 * , 331 ( 1998 ) .
a. basu , _ phys rev .
e _ * 56 * , 2869 ( 1997 ) .
a. pouquet , u. frisch and j. lorat , _ j. fluid mech .
_ * 77 * , 321 ( 1976 ) .
h. k. moffatt , _ j. fluid mech . _ * 53 * , 385 ( 1972 ) .
s. l. vainshtein and f. cattaneo , _ astrophysical jl . _ * 393 * , 165 ( 1992 ) . g. b. field , e. g. blackman and h. chou , _ apj _ , * 513 * , 638 ( 1999 ) .
i. rogachevskii and n. kleeorin , _ phys .
, * 64 * , 056307 ( 2001 )
. a. brandenburg , _ apj _ , * 550 * , 824 ( 2001 ) .
a. bhattacharjee and y. yuan , _ apj _ , * 449 * , 739 ( 1995 ) .
kraichnan , _ phys . fluids _ , * 11 * , 945 ( 1968 ) .
a. p. kazantsev , _ sov .
* 26 * , 1031 ( 1968 ) .
m. vargassola , _ phys .
e _ , * 53 * , r3021 ( 1996 ) .
d. vincenzi , _ jl .
_ , * 106 * , 1073 ( 2002 ) .
r. arlt and a. brandenburg , _ astron . &
_ , * 376 * , 713 ( 2001 ) .
a. basu , a. sain , s. k. dhar and r. pandit , _ phys .
_ , * 81 * , 2687 ( 1998 ) .
see also : e. g. blackman and a. brandenburg , _ apj _ , * 579 * , 359 ( 2002 ) .
p. c. hohenberg and b. i. halperin , _ rev .
_ , * 49 * , 435 ( 1977 )
. d. forster , d. r. nelson , and m. j. stephen , _ phys .
rev a _ , * 16 * , 732 ( 1977 ) .
l. w. lee , _ ann .
_ , * 32 * , 292 ( 1965 ) .
a. basu and j. k. bhattacharjee , _ europhys .
lett . _ * 46 * , 183 ( 1997 ) .
j d jackson , _ classical electrodynamics _ ,
2nd edn . ,
wiley eastern , new delhi , ( 1975 ) .
u. frisch , _ turbulence : the legacy of a.n .
( cambridge university press , cambridge , 1995 ) .
v yakhot and s a orzag , _
_ , * 1 * , 1 ( 1986 ) . v. s. lvov and i. procaccia , _ phys
_ , * 69 * , 3543 ( 1992 ) .
a. raichudhuri , _ the physics of fluids and plasmas _ , cambridge university press , cambridge ( 1998 ) .
j - d fournier , p - l sulem and a. pouquet , _ j. phys
* 15 * , 1393 ( 1982 ) ; a. pouquet , j - d fournier and p - l sulem , _ j. phys .
( paris ) _ * 39 * , l199 ( 1978 ) ; l.l .
kichatinov , _ magnetohydrodynamics _ * 21 * , 105 ( 1985 ) ; r. g. kleva , _ phys . fluid _ * 29 * , 2882 ( 1986 ) ; y. zhou and g. vahala , _ j. plasma phys .
_ * 39 * , 511 ( 1988 ) ; d. w. longcope and r. n. sudan , _ phys . fluid b _ * 3 * , 1945 ( 1991 ) .
a. basu , j. k. bhattacharjee and s. ramaswamy , _ eur .
j b _ , * 9 * , 425 ( 1999 ) .
m. k. verma , _ phys . plasma _ * 6 * , 1455 ( 1999 ) .
a. basu , _ phys .
e _ , * 61 * , 1407 ( 2000 ) .
s. kida , s. yanase and j mizushima , _ phys .
fluids a _ , * 3 * , 457 ( 1991 ) .
h. k. moffatt , _ rep .
_ , * 46 * , 621 ( 1983 ) .
a. basu , unpublished . | we analyze the effects of the background velocity and the initial magnetic field correlations , and viscosities on the turbulent dynamo and the @xmath0-effect . we calculate the @xmath0-coefficients for arbitrary magnetic and fluid viscosities , background velocity and the initial magnetic field correlations .
we explicitly demonstrate that the general features of the initial growth and late - time saturation of the magnetic fields due to the non - linear feedback are qualitatively independent of these correlations .
we also examine the hydrodynamic limit of the magnetic field growth in a renormalization group framework and discuss the possibilities of suppression of the dynamo growth below a critical rotation .
we demonstrate that for kolmogorov- ( k41 ) type of spectra the _ ekman number _
@xmath1 for dynamo growth to occur . |
the field of numerical relativity has matured to a stage where it is possible to simulate realistic systems of astrophysical interest . in this paper , we examine the global effects of viscosity on differentially rotating , relativistic stars . viscosity can have significant effects on the stability of neutron stars .
for example , it can drive a secular bar instability in rapidly rotating neutron stars , as shown in newtonian gravitation @xcite and in general relativity @xcite .
viscosity can suppress the @xmath1-modes @xcite and other gravitational - radiation driven instabilities , including the secular bar modes @xcite .
viscosity also destroys differential rotation , and this can cause significant changes in the structure and evolution of differentially rotating massive neutron stars . differentially rotating neutron stars can support significantly more rest mass than their nonrotating or uniformly rotating counterparts , making `` hypermassive '' neutron stars possible @xcite .
such hypermassive neutron stars can form from the coalescence of neutron star binaries @xcite or from rotating core collapse .
the stabilization arising from differential rotation , although expected to last for many dynamical timescales , will ultimately be destroyed by magnetic braking and/or viscosity @xcite .
these processes drive the star to uniform rotation , which can not support the full mass of the hypermassive remnant .
this process can lead to `` delayed '' catastrophic collapse to a black hole , possibly accompanied by some mass loss .
such a delayed collapse might emit a delayed gravitational wave signal detectable by laser interferometers .
moreover , the collapse , together with any residual gas in an ambient accretion disk , could be the origin of a gamma - ray burst ( grb ) .
both magnetic fields and viscosity can destroy differential rotation in a rapidly rotating star @xcite .
simple estimates show that the magnetic braking ( alfvn ) timescale for a laminar field is much shorter than the timescale of molecular ( neutron ) viscosity in a typical massive neutron star .
hence magnetic fields are expected to be the principal mechanism driving neutron stars toward rigid rotation .
phase mixing arising from magnetic braking @xcite , or other possible magnetohydrodynamic instabilities @xcite might stir up turbulence .
turbulent shear viscosity could then dominate the subsequent evolution . in this paper , we are primarily interested in identifying the global evolutionary consequences of shear viscosity in a relativistic star , independent of the detailed nature or origin of the viscosity .
to explore the consequences of the loss of differential rotation in equilibrium stars , we study the secular evolution of differentially rotating relativistic stars in the presence of a shear viscosity . viscosity and magnetic fields
have two things in common : ( 1 ) they both change the angular velocity profiles of a differentially rotating star , and ( 2 ) they both act on _ secular _ timescales , which can be many rotation periods .
the latter inequality poses a severe challenge to numerical simulations using a hydrodynamic code .
it is too taxing for a _ hydrodynamic _ code using an explicit differencing scheme to evolve a star for physical realistic secular timescales . to solve this problem
, we artificially amplify the strength of viscosity so that the viscous timescale is short enough for numerical treatment .
however , we keep the viscous timescale substantially longer than the dynamical timescale of the stars , so that the evolution of the star remains quasi - stationary .
we then check the validity of our results by reducing the viscosity on successive runs and testing that the viscosity - induced physical behavior is unchanged ; rather , only the timescale changes and does so inversely with the strength of viscosity .
a more detailed discussion of the expected scaling is presented in section [ justification ] @xcite . to study viscous evolution
, we need to perform long simulations in full general relativity . typically , we evolve the stars in axisymmetry .
this allows us to follow the secular evolution of the stars with high resolution in a reasonable amount of time .
viscosity can , however , drive nonaxisymmetric instabilities when a star is rapidly rotating . to test for such instabilities
, we also perform lower - resolution , three - dimensional ( 3d ) simulations on the most rapidly rotating stars we consider . for non - hypermassive neutron stars that are slowly and differentially rotating
, we find that viscosity simply drives the whole star to rigid rotation .
if the non - hypermassive neutron star is rapidly and differentially rotating , however , viscosity drives the inner core to rigid rotation and , at the same time , expels the material in the outer layers .
the final system in this case consists of a rigidly - rotating core surrounded by a low - density , ambient disk in quasi - stationary equilibrium .
our most interesting results concern the fate of hypermassive neutron stars .
we numerically evolve four models with different masses and angular momenta .
we find that in all cases , viscosity drives the cores to rigid rotation and transports angular momentum outwards into the envelope . as a result
, the core contracts in a quasi - stationary manner , and the outer layers expand to form a differentially rotating torus .
of the four models we have studied , the star with the highest mass collapses to a black hole , with about 20% of the rest mass leftover to form a massive accretion disk . on the contrary , the other three stars do not collapse to black holes , but form star + disk systems , similar to the final state of the rapidly rotating non - hypermassive neutron stars described above . as will be discussed in section [ rad_cooling ]
, viscosity generates heat so that the stars do not evolve adiabatically in general .
the extra thermal pressure due to viscous heating helps to support the stars .
we also consider the limit of rapid cooling , whereby the heat generated by viscosity is immediately removed from the stars .
of the three stars which do not collapse to black holes in the no - cooling limit , we found that the one with the lowest angular momentum undergoes catastrophic collapse in the rapid - cooling limit .
about 10% of the rest mass is leftover to form an accretion disk in this case . to test the validity of the axisymmetric results , we perform 3d simulations to check for any nonaxisymmetric instabilities .
we do not find any unstable nonaxisymmetric modes and the 3d results agree with the axisymmetric results .
our results suggest that viscous braking of differential rotation in a hypermassive neutron star can , but does not always , lead to catastrophic collapse .
when catastrophic collapse does occur , the remnant is a black hole surrounded by a massive accretion disk .
this outcome is very different from that of the collapse of an unstable , rigidly - rotating `` supramassive '' neutron star , in which the whole star collapses to a black hole , leaving only a tiny amount of material to form a disk @xcite .
many models for grbs require a massive disk around a rotating black hole to supply energy by neutrino processes @xcite .
our results suggest that viscous forces in a hypermassive star could lead to the formation of a massive disk around such a black hole .
the structure of this paper is as follows . in section [ setup ] , we derive the relativistic navier - stokes equations containing shear viscosity in a 3 + 1 form suitable for numerical integration , and describe how we evolve them in both axisymmetry and full 3 + 1 dimensions .
we then describe in section iii several tests that we perform to check our code .
we present the results of our simulations on five selected stars in section iv . finally , we briefly summarize and discuss our conclusions in section v.
throughout this paper , latin indices denote spatial components ( 1 - 3 ) and greek indices denote spacetime components ( 0 - 3 ) .
we adopt geometrized units , so that @xmath2 .
we evolve the 3-metric @xmath3 and the extrinsic curvature @xmath4 using the bssn formulation @xcite .
the fundamental variables for bssn evolution are @xmath5\ , \\
\tilde\gamma_{ij } & \equiv & e^{-4\phi}\gamma_{ij}\ , \\
k & \equiv & \gamma^{ij}k_{ij}\ , \\
\tilde a_{ij } & \equiv & e^{-4\phi}(k_{ij } - { 1\over 3}\gamma_{ij}k)\ , \\ \tilde\gamma^i & \equiv & -\tilde\gamma^{ij}{}_{,j}\ .\end{aligned}\ ] ] the evolution and constraint equations for these fields are summarized in @xcite ( hereafter paper i ) . in the presence of matter , these evolution equations contain the following source terms : @xmath6 where @xmath7 is the stress tensor , and @xmath8 is the future - directed unit normal to the time slice .
one must impose gauge conditions which specify the lapse @xmath9 and the shift @xmath10 .
we use a k - driver lapse and gamma - driver shift , as described in paper i. the numerical implementation of the equations is discussed in paper i , with some improvements to enhance stability described in @xcite .
the latter are particularly relevant for the post - collapse version of our code that we implement with black - hole excision .
we treat the matter in our neutron stars as an imperfect fluid with a shear viscosity , but no bulk viscosity and no heat conduction .
the stress tensor for the fluid is @xmath11 here , @xmath12 , @xmath13 , @xmath14 , and @xmath15 are the rest - mass density , specific internal energy , pressure , and fluid four - velocity , respectively .
the quantity @xmath16 is the coefficient of viscosity and is related to the kinematic viscosity @xmath17 by @xmath18 .
the shear tensor @xmath19 is defined by @xcite @xmath20 where @xmath21 is the fluid 4-acceleration .
we assume a @xmath22-law equation of state @xmath23 our fundamental fluid variables are @xmath24 where @xmath25 is the specific enthalpy . the conservation of stress - energy @xmath26 and the law of baryon number conservation @xmath27 give the relativistic continuity , energy , and navier - stokes equations @xmath28 where @xmath29 is the 3-velocity .
the quantity @xmath30 is determined by the normalization condition @xmath31 , which yields @xmath32^{-2}\ , \ ] ] where @xmath33 . the stress - tensor @xmath34 generates the following source terms in the field evolution equations : @xmath35 many of the systems we evolve possess and maintain symmetry about their rotation axis , which we set to be the @xmath36-axis
. then we can eliminate one dimension and simplify the equations .
we utilize axisymmetry and follow @xcite to evolve the field and hydrodynamic variables on the @xmath37 plane .
the data off of this plane can be obtained by rotating the data on this plane .
as we explain in section [ axisym_methods ] , we find it advantageous when performing 2 + 1 simulations to evolve the hydrodynamic equations ( [ evolve_rhostar])-([evolve_stilde ] ) in cylindrical coordinates but on a cartesian ( @xmath38 ) grid . on the @xmath37 plane ,
the cylindrical coordinates @xmath39 , @xmath36 , and @xmath40 are related to the cartesian coordinates @xmath41 , @xmath42 , and @xmath36 as follows : @xmath43 , @xmath44 , @xmath45 . using these relations , eqs .
( [ evolve_rhostar])-([evolve_stilde ] ) in cylindrical coordinates can be written @xmath46 \nonumber\\ & & \partial_t ( \tilde s_y - 2\alpha e^{6\phi}\eta\sigma^0_y ) \label{eulery}\\ & & \qquad\ \ + { 1 \over x^2}\partial_b(x^2 \tilde s_y v^b - 2\alpha e^{6\phi}\eta x^2 \sigma^a_y)=0 \nonumber \\ & & \partial_t e_{\star } + { 1 \over x}\partial_b(e_{\star } x v^b ) \label{energy } \\ & & \qquad\ \ = { 2\over\gamma } \alpha e^{6\phi}\eta ( \rho_0\epsilon)^{(1-\gamma)/\gamma}\sigma^{\alpha\beta}\sigma_{\alpha\beta } \ , \nonumber\end{aligned}\ ] ] where the indices @xmath47 and @xmath48 run over @xmath41 and @xmath36 ( c.f .
eq.(2.10)-(2.13 ) of @xcite ) .
there are two dynamical timescales for a rotating star .
gravity provides the free - fall timescale @xmath49 @xmath50 where @xmath0 is the gravitational mass of the star ( or merged binary remnant ) and @xmath51 is the radius .
if the star is rotating , its rotation period @xmath52 provides another important timescale : @xmath53 dynamical instabilities ( e.g. instability to radial collapse or to dynamical bar formation ) will act on the above timescales .
the stars we study are dynamically stable initially , so their structure is altered on secular timescales .
rotating compact stars may be secularly unstable to gravitational - radiation driven instabilities .
the strongest instabilities of this kind are the ( nonaxisymmetric ) @xmath1-modes and the bar mode .
the timescale of the @xmath54 @xmath1-mode instability is given by @xcite @xmath55 the gravitational - radiation driven ( dedekind ) bar - mode instability occurs if the star is rapidly rotating so that @xmath56 , where @xmath57 is the ratio of kinetic to gravitational potential energy .
the threshold @xmath58 for a newtonian star ( @xmath59 ) , and _
decreases _ as the compactness of the star ( i.e. @xmath60 ) increases @xcite .
the timescale of this instability is estimated to be @xcite @xmath61 viscosity alone can also drive a ( jacobi ) bar - mode instability .
the threshold is identical for a newtonian star ( @xmath58 ) but _ increases _ as the compaction increases @xcite .
the relevant timescale is ( see @xcite , p.99 ) @xmath62 this is comparable to the viscous timescale @xmath63 discussed below .
magnetic fields coupled to the matter will redistribute angular momentum .
in fact , even an initially small magnetic field frozen into the matter will be wound up and can destroy differential rotation in the star on an alfvn timescale @xcite : @xmath64 viscosity will also redistribute angular momentum on a viscous timescale @xmath65 .
one form of viscosity present in neutron stars is due to the transport of energy and momentum of neutrons .
this viscosity acts on a timescale @xcite @xmath66 where @xmath67 , and @xmath68 is the characteristic temperature .
it is widely believed that , for cold neutron stars ( @xmath69 ) , the neutron fluid in the inner crust condenses into a superfluid of @xmath70 cooper pairs @xcite , while in the interior , the neutrons could form a @xmath71 superfluid @xcite ( although this is less certain ) , and the protons a @xmath70 superfluid . in the case of neutron superfluidity ,
@xmath72 will vanish , and the dominant viscosity will be due to electron - electron scattering @xcite @xmath73 electron and proton fluids are forced to move together in the mhd limit @xcite .
differences in velocity between the neutron and proton - electron fluids are damped fairly quickly by mutual friction @xcite .
viscosity can be used as a model for turbulence in certain situations .
turbulence may occur in young neutron stars as a result of pure hydrodynamic effects or magnetic instabilities @xcite .
turbulence is often modeled by the `` @xmath9-disk '' law , in which a shear stress @xmath74 is added to the hydrodynamic equation ( see e.g. @xcite , chap .
14 ) . here
@xmath9 is a nondimensional constant ( which should not be confused with the lapse function ) with values in the range @xmath75 .
the viscosity in this model roughly corresponds to @xmath76 , where @xmath77 is the velocity of turbulent cells relative to the mean fluid motion , @xmath78 is the size of the largest turbulent cell , and @xmath79 is the sound speed .
the corresponding timescale is @xmath80 which is much shorter than all the other secular timescales .
hence turbulent viscosity , if present , is likely to dominate the secular evolution of differentially rotating stars .
thermal energy is radiated away primarily by neutrinos .
for hot neutron stars ( @xmath81 ) , the cooling is dominated by the direct urca process , and the star cools on a timescale @xmath82 s ( see @xcite and @xcite , chap .
11 ) . for cooler neutron stars ,
the cooling is dominated by the modified urca process , and the star cools on a timescale @xmath83 s @xcite .
depending on the temperature and the nature of the viscosity , the cooling timescale may be greater than or less than the viscous timescale .
if @xmath84 , then the heat generated by viscosity will build up inside the star .
otherwise , it will be radiated away as quickly as it is generated .
we study both limits in this paper .
secular effects will in general take many rotation periods to significantly affect the structure or velocity profile of a differentially rotating star .
this poses a challenge to the numerical treatment of these changes .
because of the short courant timestep required for numerical stability , it would be computationally prohibitive to evolve a star for such a long time using an explicit finite differencing scheme . the use of a fully implicit scheme for the finite differencing can allow stable evolutions with larger @xmath85 .
each timestep is , however , much more computationally expensive as it involves matrix inversion .
moreover , no fully implicit routine for the coupled einstein field and relativistic hydrodynamic equations exists at present .
the secular timescales are so much longer than the dynamical timescales that the star can be thought of as evolving quasi - statically .
therefore , it might be possible to treat the secular evolution in the quasistatic approximation , as in typical stellar evolution ( henyey ) codes , by constructing a sequence of equilibrium configurations up to the moment that stable equilibrium can no longer be sustained .
this approach has been used to study the viscous evolution of differentially rotating white dwarfs @xcite .
however , building the required equilibrium models in full general relativity is a nontrivial task
. it would be particularly difficult to identify the meridional currents and core - halo bifurcation that often arise in rapidly rotating configurations .
more significantly , it would not be possible to follow the evolution of the configuration with a quasistationary approach if a dynamical instability ( i.e. collapse ) is triggered during the secular evolution .
instead , we use our relativistic hydrodynamic code and artificially amplify the strength of viscosity so that the viscous timescale is short enough to make numerical treatment tractable .
however , we keep the viscous timescale sufficiently long that the hierarchy of timescales is maintained , and the secular evolution still proceeds in a quasi - stationary manner . the behavior of the real system can then be determined by rescaling the time variable to adjust the viscous timescale to its physical value .
the characteristic viscous timescale is @xmath86 where @xmath87 is an averaged value of @xmath16 across the star .
suppose we evolve the same star , once with @xmath88 and once with @xmath89 . if both @xmath90 and @xmath91 are large enough so that they do not compete with the dynamical timescale , but shorter than any other secular timescale ( see , e.g. section [ timescales ] ) , then the configuration of the star with viscosity @xmath90 at time @xmath92 will be the same as the configuration of the star with viscosity @xmath91 at time @xmath93 . by varying @xmath65 over a wide range and corroborating this scaling , we are confident that the physical behavior we observe is real
. we can then scale the result of numerical simulations to the appropriate strength of viscosity , provided that the physically relevant viscous timescale is much shorter than all the other secular timescales ( e.g. , @xmath94 , @xmath95 ) . as discussed in the previous subsection ,
turbulent viscosity is likely to dominate the secular evolution .
we adopt the stress tensor as in eq .
( [ tuv ] ) and specify the viscosity @xmath16 suitable for modeling turbulence .
we consider the turbulent viscosity described in @xcite : @xmath96 ( this viscosity law is also used in some accretion - disk models @xcite . ) typically , @xmath77 is proportional to the sound speed @xmath79 .
hence @xmath97 . for simplicity , we assume that @xmath98 is constant inside the star .
then we have @xmath99 where @xmath100 is a constant , and is related to the coefficient of kinematic viscosity @xmath17 by @xmath101 .
in our numerical simulations we specify the value of @xmath102 for each run .
we model the initial stars as rotating polytropes with polytropic index @xmath103 , so that @xmath104 .
it is convenient to rescale all quantities with respect to @xmath105 .
since @xmath106 has dimensions of length , we can define the following nondimensional variables @xcite @xmath107 where the spacetime coordinates are @xmath108 .
however , to simplify our notation , we will drop all the bars .
hereafter , all variables are understood to be in `` @xmath109 units . '' using eq .
( [ tvis_general ] ) , we can see that @xmath65 scales with @xmath51 , @xmath110 , and @xmath102 as @xmath111 which for @xmath112 becomes @xmath113 for definiteness , we take @xmath65 to be @xmath114 where @xmath115 is the equatorial radius , @xmath116 is the maximum value of @xmath12 in the star , and @xmath117 is a dimensionless constant .
we use the constant @xmath117 to approximately match @xmath65 to the rate of decay of differential rotation observed by our simulations . with the appropriate @xmath65 , the value of @xmath118 , which is proportional to the rate of energy dissipation
[ see eq .
( [ evolve_estar ] ) ] , is expected to decay like @xmath119 we measure @xmath65 by numerically evolving a given star , and observing the decay with time of @xmath120 , the average value of @xmath121 throughout the star , weighted by rest density .
we determine the value of @xmath117 by requiring that @xmath65 roughly corresponds to the e - folding time of the decay of @xmath122 . carrying out this measurement on a sampling of the stars used below , we find that @xmath123 in all cases .
thus we use @xmath124 to define @xmath65 in the sections below .
we note that the turbulent viscosity adopted above for our numerical treatment is roughly equivalent to an `` @xmath9 '' model provided we identify @xmath125 .
( [ tvis_p ] ) for @xmath65 is then equivalent to eq .
( [ t_turb ] ) .
our stars do not evolve isentropically .
viscosity heats the matter on a timescale @xmath65 , as shown by eq .
( [ evolve_estar ] ) . at the same time , neutrino radiation carries away heat , cooling the star on a timescale @xmath126 .
we will carry out simulations below in two opposite limits , which we describe in some detail in appendix [ cooling_app ] .
in the _ no - cooling _ limit , @xmath127 , so we ignore radiative cooling and simply evolve eqs .
( [ evolve_rhostar])-([evolve_stilde ] ) . in
the _ rapid - cooling _
limit , @xmath128 , so we evolve eqs .
( [ evolve_rhostar])-([evolve_stilde ] ) as before , but _ without including _ the viscous heating term in the energy equation [ eq . ( [ evolve_estar ] ) ] .
this will allow net heating by adiabatic compression but not by viscosity .
the viscous heat is assumed to be ( instantaneously ) lost by radiation in this limit , while viscous braking proceeds .
the emitted radiation will carry off some momentum as well as energy , causing a modification of eq .
( [ evolve_stilde ] ) , but this will have a much smaller effect provided the luminosity does not exceed the ( neutrino ) eddington luminosity .
baumgarte and shapiro @xcite investigated the loss of angular momentum in binary neutron star merger remnants due to radiation , and they found it to be fairly small . we therefore feel justified in ignoring radiative corrections to eq .
( [ evolve_stilde ] ) .
our hydrodynamical scheme employs van leer - type advection with artificial viscosity to handle shocks .
we also use a `` no - atmosphere '' approach , in which the density at any point on our grid can fall exactly to zero .
our hydrodynamical algorithms are described in detail in paper i. we have evolved the above equations both in two dimensions , assuming axisymmetry , and in three dimensions . using axisymmetry saves us computational time and allows us to use higher resolution . however
, 3d runs must still be carried out for rapidly rotating systems in order to check for the occurrence of nonaxisymmetric instabilities .
there are several ways to evolve in axisymmetry .
one could write the field and hydrodynamic evolution equations in cylindrical coordinates ( @xmath129,@xmath36,@xmath130 ) and evolve in this coordinate system .
this has the advantage that one can explicitly remove the dependence of the variables on @xmath130 .
unfortunately , there are singularities in the cylindrical coordinate system which can make the evolution of the field equations in these coordinates difficult . instead
, we choose to evolve the metric variables ( @xmath131 , @xmath132 , @xmath133 , @xmath134 , @xmath135 , @xmath9 , and @xmath10 ) in axisymmetry using the cartoon method @xcite . in this approach ,
variables are evolved on a cartesian grid consisting of three planes corresponding to @xmath136 , @xmath137 , and @xmath138 .
then the middle ( @xmath137 ) plane is evolved using the 3d evolution equations in cartesian coordinates .
each time the middle plane is evolved forward one timestep , the @xmath136 and @xmath138 planes are updated by applying the assumption of axisymmetry .
thus , the value of a tensor * f * at location @xmath139 on @xmath140 is equal to * f * at @xmath141 rotated by ( coordinate ) tensor transformation about the @xmath36-axis by angles @xmath142 . since an arbitrary point @xmath141
will generally not coincide with any gridpoint on the @xmath37 plane , interpolation in @xmath41 is necessary to apply this update .
we use third - order polynomial interpolation , so that we do not lose second - order accuracy . with the hydrodynamic evolution equations
, we also have the choice of either evolving in cylindrical coordinates or evolving in cartesian coordinates using the cartoon prescription . like shibata in his work on axisymmetric star collapse @xcite
, we choose to evolve the fluid variables in cylindrical coordinates , i.e. we use eqs .
( [ continuity])-([energy ] ) .
this is superior to using the cartoon method because eq .
( [ continuity ] ) can be finite differenced in such a way that the total rest mass will be exactly conserved ( except for flow beyond the outer boundaries ) . in the absence of viscosity ,
angular momentum also becomes a numerically conserved quantity .
we have found that evolving the fluid variables in cylindrical coordinates gives significantly more accurate runs than evolving via cartoon hydrodynamics .
the drawback of using 2d evolutions is the instability caused by the coordinate singularity on the @xmath143 axis .
( this instability is also present if we use the 3d cartesian navier - stokes equations together with the cartoon boundary condition . )
there are several ways of removing this instability .
shibata @xcite adds a small artificial shear viscosity @xmath144 where @xmath145 is the flat - space laplacian , and @xmath146 is the momentum variable evolved by the code of @xcite ( instead of @xmath147 ) .
we have confirmed that adding such a term to eq .
( [ evolve_stilde ] ) can stabilize our code .
however , since we will be studying the effects of real shear viscosity , we instead choose to remove the instability using a higher - order dissipation scheme .
namely , we add a small kreiss - oliger dissipation term @xcite @xmath148 we use @xmath149 for all the simulations reported in this paper . in both 2d and 3d simulations
, we assume that our system preserves equatorial symmetry across the @xmath150 plane , and we therefore only evolve the @xmath151 portion of the grid . in 3d runs , we make the added assumption of @xmath152-symmetry , which allows us to evolve only half of the remaining grid , which we choose to be the @xmath153 half . when performing 2d runs , we evolve only the region @xmath154 since the values of variables in the @xmath155 region can be deduced from the values on the @xmath154 region from the assumed axisymmetry .
we compute all spacial derivatives using standard centered differencing .
we integrate forward in time using a 3-step iterated crank - nicholson scheme .
so , for example , we evolve the equation @xmath156 from timestep @xmath157 at time @xmath92 to timestep @xmath158 at time @xmath159 by the following algorithm : + ` ( i ) predict : { } ^1f^{n+1 } \equiv f^n + \delta t \dot f(f^n ) ` + ` ( ii ) 1st correct : ~ { } ^2f^{n+1 } \equiv f^n + \delta t [ 0.4\dot f(f^n)+0.6\dot f({}^1f^{n+1 } ) ] ` + ` ( iii ) 2nd correct : ~ f^{n+1 } = f^n + \delta t [ 0.4\dot f(f^n)+0.6\dot f({}^2f^{n+1 } ) ] ` + as discussed in paper i , the coefficients 0.4 and 0.6 were chosen to improve stability . in the presence of viscosity ,
the time - differencing is not entirely straightforward , due to the presence of time derivatives of @xmath15 in @xmath19 , and of time derivatives of @xmath19 in the navier - stokes equations .
thus , @xmath160 ( which gives @xmath161 ) is an expression which itself contains @xmath161 and @xmath162 .
since the viscosity is a small perturbing force on the fluid motion , we find that it is sufficient to split off the viscous terms and integrate them separately ( operator splitting ) .
in particular , we compute @xmath161 and @xmath163 appearing in the viscous terms in a non - time centered way .
consider the @xmath164 evolution equation for the viscous piece : @xmath165 , where we have suppressed all indices . to evolve this equation from timestep @xmath157 to timestep @xmath166
, we need to know the time derivatives of @xmath167 and @xmath168 . when performing the predictor step ,
these time derivatives are approximated by subtracting values of the fields on the timestep @xmath157 from those on the previous timestep , @xmath169 .
+ ` ( i ) before predictor step , ~ compute \dot u^{n-1/2 } = [ u^n - u^{n-1}]/\delta t , ~ \sigma^n = \sigma(u^n,\dot u^{n-1/2 } ) , ~ \dot \sigma^{n-1/2 } = [ \sigma^n - \sigma^{n-1}]/\delta t ` + note that these time derivatives are centered at @xmath170 .
we then carry out the predictor step . + `
( ii ) predict : { } ^1\tilde s^{n+1 } = \tilde s^n + \delta t \dot{\tilde s}(\sigma^n,\dot\sigma^{n-1/2 } ) ` + from the predicted values of @xmath167 and @xmath168 , we construct time derivatives centered at @xmath171 and use these in the corrector step . + ` ( iii ) compute { } ^1u^{n+1 } from { } ^1\tilde s^{n+1 } ~ { } ^1\dot u^{n+1/2 } = [ { } ^1u^{n+1 } - u^{n}]/\delta t , ~ { } ^1\sigma^{n+1 } = \sigma({}^1u^{n+1},{}^1\dot u^{n+1/2 } ) , ~ { } ^1\dot \sigma^{n+1/2 } = [ { } ^1\sigma^{n+1 } - \sigma^{n}]/\delta t ` + ` ( iv ) 1st correct : ~ { } ^2\tilde s^{n+1 } = \tilde s^n + \delta t[0.4\dot{\tilde s}(\sigma^n,\dot\sigma^{n-1/2 } ) ~ + 0.6\dot{\tilde s}({}^1\sigma^{n+1},{}^1\dot\sigma^{n+1/2 } ) ] ` + ` ( iii ) compute { } ^2u^{n+1 } from { } ^2\tilde s^{n+1 } ~ { } ^2\dot u^{n+1/2 } = [ { } ^2u^{n+1 } - u^{n}]/\delta t , ~ { } ^2\sigma^{n+1 } = \sigma({}^2u^{n+1},{}^2\dot u^{n+1/2 } ) , ~ { } ^2\dot \sigma^{n+1/2 } = [ { } ^2\sigma^{n+1 } - \sigma^{n}]/\delta t ` + ` ( iv ) 2nd correct : ~ \tilde s^{n+1 } = \tilde s^n + \delta t[0.4\dot{\tilde s}(\sigma^n,\dot\sigma^{n-1/2 } ) ~ + 0.6\dot{\tilde s}({}^2\sigma^{n+1},{}^2\dot\sigma^{n+1/2 } ) ] ` by differencing the equations in this way , the dominant nonviscous terms in the evolution equations are accurate to second order ( except for small effects due to the use of the coefficients 0.4 and 0.6 in the corrector steps ) , but small viscous terms involving time derivatives are only accurate to first order in @xmath85 .
numerical convergence tests show that our code is nearly second - order convergent in space and time .
we find this to be sufficient for our purposes . when computing @xmath19 , we first use eq .
( [ sigma_def ] ) to get the spatial components @xmath172 .
the remaining components @xmath173 are then obtained from the conditions @xmath174 . as in most other eulerian hydrodynamic codes ,
high velocities can easily develop in the low - density regions near the surfaces of our stars .
the method for evolving such regions in the absence of viscosity is described in paper i. since calculation of the shear tensor involves taking derivatives of the velocity field , we are unable to calculate it accurately in the very low - density regions .
to ensure stability , we set @xmath175 in regions where @xmath176 .
since these low - density regions contain an insignificant amount of rest mass , this prescription should not affect our evolutions .
we confirm this by varying the cutoff density in several test problems and checking that the effect is negligible .
our most important diagnostics are the total mass - energy @xmath0 and angular momentum @xmath177 .
these are both defined by surface integrals at infinity @xcite : @xmath178 using gauss law , these surface integrals can be converted to volume integrals : @xmath179 in axisymmetry , the angular momentum integral simplifies to @xcite @xmath180 henceforth , we drop the subscript @xmath36 since all angular momentum is in the @xmath36-direction . the mass @xmath0 and angular momentum @xmath177 in our grid should be strictly conserved only in the absence of radiation ( although gravitational radiation carries no angular momentum in axisymmetry ) . since the energy and angular momentum emitted in gravitational waves are negligible in our runs , this means that @xmath0 and @xmath177 should be conserved in the no - cooling limit .
they thus serve as useful code checks in this limit . in the rapid - cooling limit
, the mass computed by the volume integral ( [ mfin ] ) over the numerical grid will not be conserved
thermal energy is carried off of the grid by thermal radiation .
the expected rate of mass - energy decrease due to thermal energy loss can be computed by differentiating eq .
( [ mfin ] ) with respect to time .
to lowest order , we can ignore the effects of the quasi - stationary loss of thermal energy on the spacetime , and so ignore the time derivatives of field variables .
then only the first term in eq .
( [ mfin ] ) will be important .
@xmath181 where @xmath182 is the component of the time derivative of @xmath110 caused by loss of internal energy due to cooling .
this quantity may be computed by applying the chain rule to eq .
( [ rho_source ] ) : @xmath183 changes in @xmath30 are ignored because they represent a higher - order influence on @xmath184 .
the rate of change in @xmath185 due to cooling is given by the effective balance of heating and cooling that characterizes the rapid - cooling limit ( see appendix a ) .
thus , @xmath186 is minus the value of @xmath187 due to viscous heating , i.e. @xmath188 [ see eq .
( [ evolve_estar ] ) ] . from equations ( [ chain ] ) and ( [ e_cool ] ) , it is straightforward to construct @xmath189\ .\end{aligned}\ ] ] finally , the quantity which is nearly conserved in the rapid - cooling limit ( up to losses due to gravitational radiation ) is @xmath190 in both the no - cooling and rapid - cooling runs , we can divide @xmath0 into its constituent pieces : the rest mass @xmath191 , internal energy mass @xmath192 , kinetic energy @xmath68 , and gravitational potential energy @xmath193 , defined by @xcite @xmath194 where @xmath195 is the proper 3-volume element . to study the effects of heating , it is useful to break up the internal energy @xmath13 into its `` cold '' component @xmath196 , and its `` thermal '' component @xmath197
. then we can break up @xmath192 into cold and hot components @xmath198 note that in the rapid - cooling limit , @xmath199 .
finally , we also compute the circulation along closed curves . for a closed curve @xmath200 with tangent vector @xmath201
, the circulation is defined to be @xmath202 where @xmath203 parameterizes points on @xmath200 [ i.e. @xmath204 . according to the kelvin - helmholtz theorem
, the circulation @xmath205 will be conserved in the absence of viscosity if @xmath200 moves with the fluid and if the fluid is barotropic [ @xmath206 . when viscosity is present or the equation of state is more general , as in the case of nonisentropic flow , the navier - stokes equations give @xmath207 d\zeta\ .
\label{evolve_circ}\ ] ] if @xmath175 , the second term in the integrand vanishes , and if @xmath208 , the remaining term is an exact differential .
then @xmath209 integrates to zero , in accord with the kelvin - helmoltz theorem . in axisymmetry , we choose to evaluate @xmath205 on circular rings on the equatorial @xmath150 plane , so that @xmath210 .
the ring @xmath200 intersects our 2d grid at a point on the @xmath41-axis . by our symmetries ,
the curve @xmath200 will always remain circular and always remain at @xmath150 , so the lagrangian point representing @xmath200 only moves in @xmath41 .
evolution in @xmath92 and in @xmath211 are simply related by @xmath212 .
since the system is axisymmetric , the integrand , being a scalar , is constant along @xmath200 , so eqs .
( [ circ_define ] ) and ( [ evolve_circ ] ) simplify to @xmath213 \label{circ_fin}\ .\end{aligned}\ ] ] hence the quantity @xmath214 given by @xmath215 is conserved , even in the presence of viscosity . finally , we compute the hamiltonian and momentum constraint violations [ given by equations ( 16 ) and ( 17 ) of paper i ] .
we monitor the l2 norm of the violation of each constraint .
we compute the l2 norm of a gridfunction @xmath216 by summing over every gridpoint @xmath217 : @xmath218 the constraint violations are normalized as described in paper i [ eqs .
( 59 ) and ( 60 ) ] .
0.07 in [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] @xmath219 the time at which the simulation was terminated .
+ @xmath220 this quantity corresponds to an average of @xmath221 over the final configuration of the star weighted by rest - mass density at the end of the simulation .
thermal pressure generated by viscous heating causes @xmath222 ( recall @xmath223 ) .
we find that the viscous heating is much more significant in the low - density region than in the core .
+ @xmath224 these values are obtained by solving eqs .
( [ eq : jisco])([eq : q ] ) .
+ @xmath225 the quantity @xmath57 is undefined when the star undergoes a dynamical collapse .
the number given here is an approximate value before the star becomes dynamically unstable .
[ tab : results ] four of the five stars we consider are hypermassive , and we expect viscosity to change their structure significantly . star i is the most hypermassive star ( @xmath226 , where @xmath227 is the mass limit for uniformly rotating @xmath112 polytropes , i.e. for stars at the mass - shedding limit @xcite ) .
we find that this star eventually collapses to a black hole , but a substantial amount of rest mass is leftover to form a massive accretion disk .
we consider three other hypermassive models ( stars ii , iii and iv ) to study whether or not all hypermassive neutron stars will collapse in the presence of viscosity .
stars ii , iii and iv have the same rest mass ( @xmath228 ) which is slightly smaller than that of star i but different angular momenta @xmath177 . we find that stars iii and iv never collapse , but evolve in a quasi - stationary manner to a uniformly rotating core surrounded by a low - density , disk - like envelope .
star ii eventually collapses to a black hole if we impose rapid cooling . in the no - cooling limit , however , this model forms a uniformly rotating core surrounded by a substantial disk .
star v is the only non - hypermassive model .
as expected , this star does not collapse under the action of viscosity .
however , viscosity can not drive the whole star to rigid rotation , because the angular momentum of the star exceeds the maximum angular momentum allowable for a rigidly - rotating star having the same rest mass .
instead , viscosity again leads to a uniformly rotating core and a differentially rotating disk - like envelope .
the final outcomes of the five models are summarized in table [ tab : results ] .
we also performed 3d simulations on stars i and iv to search for unstable , nonaxisymmetric secular modes .
a nonaxisymmetric bar instability usually develops when a star is rotating rapidly , i.e. has a sufficiently large @xmath57 .
of the five models we study , stars i and iv have the highest @xmath57 .
we do not find any nonaxisymmetric instabilities in these two models , and the 3d results roughly agree with the axisymmetric results .
we discuss the results of our simulations in detail in the following subsections .
0.5 cm star i is the most massive neutron star we study .
we first perform an axisymmetric calculation with no cooling . at @xmath229 ,
the star has a toroidal density profile , i.e. , the maximum density occurs off center ( see the upper left panel of fig . [
fig : conti ] ) .
as viscosity gradually brakes differential rotation , the star readjusts to a monotonic density profile .
figure [ fig : rho_alpha_i ] shows the maximum rest - mass density and the minimum value of the lapse as a function of time .
figure [ fig : conti ] shows the meridional rest - mass density contours at various times .
we see that a meridional current is built up in the process .
however , the magnitude of the meridional velocity ( @xmath230 ) is much smaller than the typical rotational velocity ( @xmath231 ) . and the rest mass of the disk @xmath232 after the appearance of the apparent horizon at @xmath233 .
note that time is plotted in units of @xmath0 ( @xmath234 ) .
black hole excision is employed to track this late evolution.,width=302 ] viscosity destroys differential rotation and transfers angular momentum to the outer layers . in the early phase of the evolution , the core contracts and the outer layers expand in a quasi - stationary manner .
as the core becomes more and more rigidly - rotating , it approaches instability because the star is hypermassive and can not support a massive rigidly - rotating core . at time @xmath235
, the star becomes dynamically unstable and collapses .
an apparent horizon appears at time @xmath236 . without black hole excision , the code crashes about @xmath237 after the horizon appears because of grid stretching .
about 30% of rest mass remains outside the apparent horizon at this point .
we then continue the evolution using the excision technique described in @xcite .
we are able to extend the evolution reliably for another @xmath238 .
the system settles down to a black hole surrounded by a massive ambient disk .
the rest mass @xmath232 and angular momentum @xmath239 of the disk can be calculated by integrating the rest - mass and angular momentum density over the volume outside the apparent horizon [ c.f .
( [ def : m_0 ] ) and ( [ jaxi ] ) ]
. the mass of the black hole can be estimated by the proper circumference of the horizon in the equatorial plane : @xmath240 ( assuming that the spacetime can be described by a kerr metric ) .
figure [ fig : excision ] shows the evolution of @xmath241 and the rest mass of the disk @xmath232 .
we estimate the final values of @xmath241 , @xmath232 , and @xmath239 by fitting these curves to analytic functions of the form @xmath242 and extrapolating these fitting functions to @xmath243 @xcite .
we estimate that the mass of the final black hole is @xmath244 .
the asymptotic rest mass and angular momentum of the ambient disk are found to be @xmath245 and @xmath246 .
we can infer from the conservation of angular momentum that the final angular momentum of the black hole is @xmath247 .
hence we find @xmath248 .
the formation of a massive disk is mainly due to the fact that viscosity transports angular momentum from the inner core to the outer layers .
the material in the outer region is unable to fall into the black hole because of the centrifugal barrier .
the final mass of the black hole and disk can also be estimated independently from the conservation of specific angular momentum using a method developed by shapiro and shibata @xcite , which we apply below . during the dynamical collapse ,
the effect of viscosity is negligible . since the spacetime is axisymmetric ,
the specific angular momentum @xmath249 of a fluid particle is conserved . for a kerr black hole of mass @xmath241 and angular momentum @xmath250 , the specific angular momentum of a particle at the innermost stable circular orbit ( isco ) @xmath251
is given by @xmath252 where @xmath253 .
the isco radius is @xmath254 \ , \ ] ] where @xmath255 \ \ ] ] and @xmath256 the rest mass and angular momentum of the escaping matter in the envelope with @xmath257 is given by @xmath258 we assume that the energy radiated by gravitational waves is negligible so that the total mass - energy of the system is approximately conserved .
hence we have m & = & m_h + m_disk , [ eq : meq ] + j & = & j_h + j_disk .
[ eq : jeq ] for a bound system , the contribution to the mass - energy of the matter in the disk , @xmath259 , is smaller than its rest mass @xmath260 .
we write @xmath261 we consider two opposite limits for @xmath262 : @xmath263 and @xmath264 .
the value @xmath265 is a good approximation in the weak gravity regime . in the limit where @xmath266
, we have @xmath267 and @xmath268 . hence in this limit , @xmath269 , which is 0.92 for star
i ( see table [ tab : models ] ) .
we expect that the correct @xmath262 lies somewhere between these two extremes , which are not very different .
the mass and angular momentum of the black hole can be estimated by solving the system of transcendental eqs .
( [ eq : jisco])([eq : q ] ) at a particular time slice during the pre - excision dynamical collapse phase , including the time slice at the onset of dynamical collapse ( where @xmath262 is close to unity ) .
we find that the values of @xmath241 and @xmath250 are insensitive to @xmath262 .
they are also insensitive to which time slice we choose to do the calculation .
we find @xmath270 and @xmath271 ( @xmath272 ) .
the rest mass in the ambient disk is found to be @xmath273 .
it should be noted that this calculation is based on the assumptions that the spacetime around the disk can be approximated by a kerr metric , and that the disk is moderately thin and lies in the equatorial plane of the hole .
this approximation is not reliable when the disk is massive ( @xmath274 ) . in our case
, we find this calculation agrees rather well with the actual asymptotic values determined by the dynamical simulation with excision .
henceforth we will use eqs .
( [ eq : jisco])([eq : q ] ) to estimate @xmath241 , @xmath250 , @xmath232 and @xmath239 whenever an apparent horizon forms .
0.5 cm since viscosity is small in the low - density region , it takes longer to remove the differential rotation in the outer layers .
figure [ fig : omegai ] shows the angular velocity profiles at various times .
we see that by the time the inner core collapses , the material in the outer layers is still differentially rotating .
after the dynamical collapse , viscosity will cause some of the remaining material to slowly accrete onto the black hole .
we monitor the conserved quantities and the constraints during the entire evolution .
since our finite - difference scheme preserves the rest mass , the variation of @xmath191 can only come from material flowing out of the grid .
we find that @xmath191 is conserved to 0.01% , and angular momentum is conserved to 0.1% .
the hamiltonian constraint is violated by less than 0.3% before the dynamical collapse occurs .
it increases to 3% by the time an apparent horizon appears .
the momentum constraints are violated by less than 1% before the dynamical collapse occurs , and increase to 6% by the time an apparent horizon appears . 0.5 cm we next perform an axisymmetric simulation of star i with rapid cooling .
the dashed lines in fig .
[ fig : rho_alpha_i ] show the time evolution of the maximum rest - mass density and the minimum value of the lapse . as in the no - cooling case ,
the inner core contracts and the outer layers expand in a quasi - stationary manner .
the core then collapses dynamically to a black hole and leaves behind a massive accretion disk .
since there is no viscous heating , the whole process occurs more quickly than in the no - cooling case .
the dynamical collapse occurs at time @xmath275 and the apparent horizon appears at @xmath276 .
figure [ fig : conti_cool ] shows the meridional rest - mass density contours at various times .
we estimate , by solving eqs .
( [ eq : jisco])([eq : q ] ) at @xmath277 with @xmath263 , that the mass and angular momentum of the final black hole are @xmath270 and @xmath278 ( @xmath279 ) .
about 20% of rest mass escapes capture to form an accretion disk .
] so that the curves again coincide , which indicates that viscosity plays an insignificant role during the dynamical collapse phase.,title="fig:",width=302 ] 0.5 cm in section [ visctest ] , we demonstrated that when the viscous timescale is significantly longer than the dynamical timescale , the secular rates of change of all physical quantities scale inversely with viscosity .
the secular evolution of star i with rapid cooling is short enough for us to perform another detailed scaling test .
figure [ fig : scaling ] demonstrates this scaling behavior by evolving star i with four different strengths of viscosity @xmath280 0.4 , 0.2 , 0.1 , and 0.05 .
( the curves in fig .
[ fig : rho_alpha_i ] correspond to @xmath281 . )
we see that the scaling holds until dynamical collapse at time @xmath282 .
when @xmath283 , the evolution of the system is no longer driven by viscosity .
we therefore expect that the collapse is independent of the strength of viscosity as long as the viscous timescale is much longer than the dynamical timescale . in the lower panel of fig .
[ fig : scaling ] , we demonstrate that it is possible to shift the time axes ( @xmath284 ) so that the four viscosity runs yield the same result when @xmath285 , which indicates that viscosity is insignificant during the dynamical collapse .
the values of @xmath286 are determined by requiring that the scaling relation @xmath287 holds , and that the four curves be aligned when plotted against the shifted time @xmath288 .
we found that @xmath289 6.1 , 12.0 , 24.08 and 47.75 respectively for @xmath280 0.4 , 0.2 , 0.1 , and 0.05 .
the fact that we are able to find @xmath290 that satisfies these requirements validates our physical interpretation of the two phases of evolution . of the particles at time
@xmath229 is chosen so that the initial fraction of rest mass @xmath291 interior to the cylinder of radius @xmath292 is , from left to right , @xmath293=0.03 , 0.1 , 0.2 , 0.3 , 0.4 , 0.5 , 0.6 , 0.7 , 0.8 and 0.9 .
the cross in the diagram denotes the location of the apparent horizon at the end of the simulation.,title="fig:",width=302 ] 0.5 cm to better visualize the effects of viscosity , we follow the motions of ten selected lagrangian fluid elements .
figure [ fig : worldlines ] shows the worldlines of these particles .
we choose the particles to be in the equatorial plane of the star .
equatorial symmetry implies that the particles will remain in the equatorial plane at all times .
the position of a fluid particle @xmath292 satisfies the equations @xmath294 we label the particles by the initial fraction of rest mass interior to the cylinder of radius @xmath292 .
we see that the particles with the initial mass fraction @xmath295 move toward the center and ultimately move inside the apparent horizon , while those with @xmath296 move away from the center and remain outside the apparent horizon .
this agrees with our estimates of the rest mass of the ambient disk . of star i in the rapid - cooling limit .
the mass is not conserved since the thermal energy generated by viscous heating is removed .
however , the sum of the remaining mass and the mass carried away by `` radiation '' , @xmath297 , is approximated conserved.,title="fig:",width=302 ] 0.5 cm since there is rapid cooling , the mass is not conserved because the thermal energy generated by viscous heating is removed , as discussed in section [ diagnostics ] . however ,
when we account for the mass - energy carried away by thermal radiation , @xmath297 , the total mass @xmath298 should be conserved approximately [ see eq .
( [ mtot ] ) ] .
figure [ fig : mi ] shows @xmath0 and @xmath299 as a function of time before an apparent horizon appears .
the total mass is well - conserved except near the end of the simulation , where the numerical error arising from the grid stretching causes a few percent drop in the mass .
we monitored the violations of the constraints during the evolution .
the violation of the hamiltonian constraint is @xmath300 before the dynamical collapse occurs , and goes up to 7% at the time when the apparent horizon appears .
the violations of the momentum constraints are also before the dynamical collapse occurs , and increase to 8% at the time when the apparent horizon appears . 1.0 cm 0.5 cm the evolution of star ii due to viscosity is different from that of star i.
although it is still hypermassive , the mass of star ii is smaller than that of star i. when evolved in the absence of cooling , the star does not collapse to a black hole , but forms a rigidly - rotating core with a low - density disk - like envelope . when evolved in the rapid - cooling limit
, however , the star collapses to a black hole .
figure [ fig : cdenii ] shows the evolution of the central density . 0.5 cm and the lower graph shows the contours at the end of the simulation ( @xmath301 ) .
the contours are labeled as in fig .
[ fig : conti].,title="fig : " ] 0.5 cm 0.5 cm 0.5 cm for star ii evolved without cooling.,title="fig:",width=302 ] 0.5 cm 0.5 cm in the no - cooling case , star ii has not collapsed to a black hole by the end of simulation ( @xmath302 ) , but is settling to a uniformly rotating core surrounded by a massive torus .
figure [ fig : contii ] shows the meridional density contours at the beginning and at the end of the simulation .
figure [ fig : omega_ii ] shows the angular velocity profiles at various times .
viscosity drives the star to a quasi - equilibrium , rigidly - rotating core surrounded by a low - density disk .
we can not exclude the possibility that some of the outer material will slowly accrete onto the uniformly rotating inner core , eventually triggering collapse to a black hole .
however the star acquires enhanced pressure support against collapse from viscous heating ( i.e. @xmath303 where @xmath304 ) .
hence it may no longer be hypermassive with respect to this new `` hot '' equation of state , as the simulation suggests . during the entire simulation
, the star loses 1.2% of its rest mass and 4.5% of its angular momentum due to material flowing out of the grid .
figure [ fig : constraints ] shows the l2 norms of the hamiltonian and momentum constraints defined by eqs .
( 59 ) and ( 60 ) of paper i. we see that the violation of all the constraints are smaller than 1% during the entire evolution of @xmath305 .
the values of the ratio of kinetic to gravitational potential energy , @xmath57 , for all of the stars we studied decrease with time .
figure [ fig : towii ] shows the evolution of @xmath57 for star ii evolved without cooling .
viscosity transforms part of the rotational kinetic energy into heat .
it also changes the equilibrium configuration of the star significantly , causing a redistribution of various energies .
figure [ fig : energies ] shows the time evolution of various energies defined in eqs .
( [ def : m0])-([mih ] ) .
the mass of the system decreases by 1.4% due to a small amount of mass flowing out of the grid ( not visible in the graph ) .
the rotational kinetic energy @xmath68 decreases slightly .
the contraction of the core raises the gravitational binding energy @xmath306 , as well as the adiabatic part of the internal energy @xmath307 .
viscous heating generates the thermal energy @xmath308 , which prevents the star from undergoing catastrophic collapse . in the rapid - cooling case
, star ii collapses dynamically at time @xmath309 .
an apparent horizon appears at @xmath310 .
the mass and angular momentum of the final black hole are estimated by solving eqs .
( [ eq : jisco])([eq : q ] ) : @xmath311 and @xmath312 ( @xmath313 ) .
about 10% of rest mass is left as an accretion disk .
the situations for stars iii and iv are similar .
the inner core contracts in a quasi - stationary manner while the outer layers expand .
each system evolves into a rigidly - rotating core surrounded by a disk - like envelope .
the stars do not collapse to black holes at the end of the simulations whether or not rapid cooling is imposed .
again , we do not rule out the possibility that they might collapse to black holes when enough material accretes onto the inner core . from table [ tab : results ] , we see that at the end of the simulation , a large amount of angular momentum is transported to a massive disk . for stars
iii and iv , the rest mass of the core @xmath314 is smaller than the rest - mass limit of a supramassive star @xmath315 . for star ii ,
@xmath314 is slightly smaller than @xmath315 in the absence of cooling , but is slightly greater than @xmath315 in the rapid - cooling limit . for star i ,
@xmath316 in both the rapid - cooling and no - cooling cases .
this suggests that the fate of a hypermassive neutron star depends on whether viscosity can create a rigidly - rotating core with @xmath316 , in which case it will collapse . both viscous heating and the star s initial angular momentum play an important role in the final outcome . a hypermassive neutron star with higher mass and lower angular momentum is prone to collapse , whereas viscous heating tends to suppress the collapse . finally , we study the effect of viscosity on star v , which is non - hypermassive .
as expected , the star does not collapse to a black hole , irrespective of cooling .
the star eventually evolves into a rigidly - rotating core surrounded by a disk .
the fact that the star does not simply become rigidly rotating without shedding mass is due to the fact that viscosity conserves @xmath191 and @xmath177 . for a given @xmath191 and equation of state
, there is a maximum value of angular momentum @xmath317 above which a star can no longer maintain rigid rotation without shedding mass at the equator . in the case of star
v , it is apparent that @xmath318 .
hence viscosity can not force the whole star into rigid rotation .
similar results were found in studies of viscous evolution of differentially rotating white dwarfs assuming newtonian gravitation @xcite .
the results of the axisymmetric runs described above will not be physically relevant if the models are secularly unstable to bar formation .
we evolved stars i and iv in 3d to check for the formation of bars .
these models were chosen because , of our five models , they have the highest values of @xmath57 . viscous heating can lead to expansion and , hence , a decrease in @xmath57 .
thus , to further increase the chance of bar formation , we performed these runs in the rapid - cooling limit .
we superimposed @xmath319 perturbations on the initial data according to eq .
( [ seedbar ] ) for both stars with magnitude @xmath320 , so that @xmath321 initially .
both runs were performed in @xmath152-symmetry with a uniform grid of size @xmath322 and outer boundaries in the @xmath41-@xmath42 equatorial plane at @xmath323 for star i and @xmath324 for star iv . to reduce computational costs
, the extent of the grid in the @xmath36-direction is only half as large as in the @xmath41- and @xmath42-directions .
this setup is feasible because these models are initially highly flattened due to rapid rotation and because their evolution results in an expansion which is largely horizontal .
for star i , we find that @xmath325 decreases in magnitude until the code terminates due to the collapse , when @xmath326 . before the collapse ,
all constraints are satisfied to within @xmath327 while @xmath0 and @xmath177 are conserved to within @xmath328 . for star iv , @xmath325 also decreases , reaching 0.0023 after @xmath329 , when the simulation is terminated . in this case , the constraints are satisfied to within @xmath330 while @xmath0 and @xmath177 are conserved to within @xmath331 .
we also find that the rest density contours remain nearly axisymmetric throughout the evolutions of both stars .
these results indicate that both stars are stable against secular bar formation on the viscous timescale .
we have simulated the evolution of rapidly rotating stars in full general relativity including , for the first time , shear viscosity .
our findings indicate that the braking of differential rotation in hypermassive stars always leads to significant structural changes , and often to delayed gravitational collapse .
the rest mass , angular momentum , and thermal energy all play a role in determining the final state .
we performed axisymmetric numerical simulations of five models to study the influence of these parameters . in the presence of shear viscosity , the most hypermassive model which we studied ( star i ) , collapses to a black hole whether we evolve by ignoring cooling , or by assuming rapid cooling of the thermal energy generated by viscosity
however , the viscous transport of angular momentum to the outer layers of the star results in mass outflow and the formation of an appreciable disk .
next , we considered three hypermassive models ( stars ii , iii , and iv ) with the same rest mass @xmath191 , but different values of the spin parameter @xmath332 .
these models have smaller @xmath191 than star i , and are therefore less prone to collapse .
star ii , which has @xmath333 , collapses when evolved in the rapid - cooling limit , leaving behind a disk .
but without cooling , this model evolves to a stable , uniformly rotating core with a differentially rotating massive disk .
the additional thermal pressure support provided by viscous heating prevents collapse in this case .
in contrast , stars iii and iv , which have @xmath334 and 1.1 , respectively , do not collapse even in the rapid - cooling limit .
this is sensible because these models have a smaller rest mass than star i , but larger angular momenta than star ii .
though the cores of stars iii and iv contract , they are prevented by centrifugal support from reaching the necessary compaction to become dynamically unstable . in both cases , we find low - density disks surrounding uniformly rotating cores
. however , our simulations do not rule out the possibility that slow accretion of the disk material could eventually drive the uniformly rotating cores to collapse . disk formation also occurs for star v , which is differentially rotating but non - hypermassive .
since there exist stable , uniformly rotating models with the same rest mass , the braking of differential rotation in this case does not result in collapse .
however , differentially rotating stars can support larger @xmath57 than uniformly rotating stars . in the case of star
v , there does not exist a uniformly rotating star with the same ( high ) angular momentum and rest mass , so that mass shedding must take place as viscosity drives the star to uniform rotation . in the final state , we find a rigidly - rotating core surrounded by a low - density , disk . since results obtained from our axisymmetric code are physically reliable only for models which are not subject to nonaxisymmetric instabilities , we evolved stars i and iv in 3d to check for such instabilities .
previous studies in newtonian gravity have found that the secular , viscosity - driven bar instability in uniformly rotating stars should set in when @xmath335 @xcite . when general relativity is taken into account
, the threshold value can be somewhat higher @xcite .
thus , of all of our models , stars i and iv have the best chances of developing bars since they have the highest @xmath57 .
we introduced an initial bar - shaped perturbation and ran these cases in the rapid - cooling limit .
we found that , in both cases , the small initial perturbation decays and no bar is formed .
this is somewhat surprising since @xmath57 is well above 0.14 in both of these cases .
we plan to address this issue in a future report . for the evolution of each of our five models , we find that a massive disk or torus forms in the final state .
the disk typically carries @xmath336 of the rest mass of the initial configuration .
viscosity transports angular momentum from the interior of the star to the more slowly rotating exterior .
the exterior regions then expand to accommodate the additional centrifugal force , forming a low - density disk .
the inner core becomes rigidly rotating and , in some cases , undergoes gravitational collapse .
the disk , however , remains differentially rotating since viscosity acts much more slowly in low - density regions . for cases in which black holes are formed ,
the mass of the disk may be estimated by integrating the rest - mass density for those fluid elements which have specific angular momentum @xmath337 greater than the value at the isco , @xmath251 [ see eq .
( [ eq : deltam0 ] ) ] .
the estimates obtained in this way agree reasonably well with the results of our numerical simulations .
particularly good agreement was found for the case of star i with no cooling , for which we were able to extend the evolution some @xmath338 beyond the first appearance of an apparent horizon . the rest mass and angular momentum of the disk surrounding the rotating black hole
could then be calculated directly and agreed well with the estimates .
we expect that excision techniques will continue to be crucial in establishing the final fate of systems involving matter surrounding black holes . in a recent paper , shibata @xcite numerically simulated collapses of marginally stable , supramassive stars .
these supramassive models were constructed using polytropic equations of state with @xmath339 and rotate at the mass - shedding limit with @xmath340 .
shibata found that the collapse of these stars results in kerr black holes and that no more than @xmath341 of the initial rest mass remains outside of the hole .
this result is quite different from our finding that disks are usually present following collapse .
however , the initial data for the two calculations are quite different , as well as our inclusion of viscosity
. the analysis of @xcite takes uniformly rotating , unstable configurations as initial data and follows their dynamical evolution .
our calculations begin with differentially rotating , stable configurations and follow both their secular ( viscous ) and dynamical evolution .
viscosity drives our configurations to uniform rotation .
we find that massive disks usually form as by - products of the formation of uniformly rotating cores .
this is due primarily to the transport of angular momentum from the inner to the outer layers .
in addition , all of our models have @xmath342 .
( large angular momentum is required to generate a hypermassive neutron star in equilibrium . )
since this range is higher than that considered in @xcite , our models more naturally produce disks @xcite . all of the phenomena observed in our simulations follow from the braking of differential rotation in strongly relativistic stars .
this may be accomplished by viscosity as shown here , but magnetic fields are likely to be more important .
the fate of the hypermassive remnants of binary neutron star mergers may crucially depend on these effects .
the loss of differential rotation support in such a remnant may lead to delayed gravitational collapse .
this collapse could in turn lead to a delayed gravitational wave burst following the quasi - periodic inspiral and merger signal @xcite .
our results indicate that if the remnant is not sufficiently hypermassive , collapse may not occur , at least not until the star cools by radiating away its thermal energy . understanding the evolution of such merger remnants could aid the interpretation of signals observed by ground based gravitational wave detectors , such as ligo , virgo , geo , and tama . in addition
, short - duration grbs are thought to result from mergers of binary neutron stars or neutron star - black hole systems @xcite . in this scenario ,
the grb may be powered by accretion from a massive torus or disk surrounding a rotating black hole .
we have demonstrated that such disks are easily produced during the evolution of hypermassive neutron stars .
the braking of differential rotation may also be important in neutron stars formed in core collapse supernovae .
nascent neutron stars are probably characterized by significant differential rotation ( see , e.g. , @xcite and references therein ) .
conservation of angular momentum during the collapse is expected to result in a large value of @xmath57 .
however , uniform rotation can only support small values of @xmath57 without shedding mass ( @xcite , chap .
thus , nascent neutron stars from supernovae probably rotate differentially .
if the induced differential rotation is strong enough , hypermassive neutron stars can form . their subsequent evolution and final fate then depends on the presence of viscosity or magnetic fields .
such considerations may be important for long - duration grbs in the collapsar model @xcite . in this model , the grb
is powered by accretion onto the central black hole formed through core collapse in a massive star .
several interesting astrophysical systems undergo secular evolution in strongly gravitating environments . in this paper
, we have shown that it is possible to study secular effects that occur over many dynamical timescales using hydrodynamic computations in full general relativity .
we consider this an important step toward future numerical explorations of secular effects in other contexts .
in particular , we plan to incorporate mhd into our evolution code , as magnetic braking probably acts more quickly than viscosity to destroy differential rotation in many systems , like neutron stars or supermassive stars @xcite .
our results have also raised the following interesting question : under what circumstances are differentially rotating , compressible neutron stars with high @xmath57 unstable to nonaxisymmetric modes ?
we plan to address this issue in a future report .
the calculations for this paper were performed at the national center for supercomputing applications at the university of illinois at urbana - champaign ( uiuc ) .
this paper was supported in part by nsf grants phy-0090310 and phy-0205155 and nasa grant nag 5 - 10781 at uiuc .
in this appendix , we describe the thermal properties of our configurations .
the dissipation of rotational energy by viscosity heats the stars , but they may be cooled by radiation ( e.g. neutrino radiation ) .
the presence of radiation contributes a term @xmath343 to the stress tensor : @xmath344 , with @xmath34 given by eq .
( [ tuv ] ) .
this modifies the equations of motion ( [ eom ] ) to @xmath345 where @xmath346 is the 4-force density due to radiation ( see @xcite ) .
the specific entropy @xmath347 of a fluid element with temperature @xmath68 and number density @xmath348 changes when there is heating and cooling according to @xmath349 where @xmath211 is the proper time along the element s worldline . here
, we have separated the contributions from viscosity and radiation to the heating rate .
the quantity @xmath350 is the viscous heating rate per unit volume , which comes from the @xmath351 term in eq .
( [ evolve_estar ] ) . in terms of the thermal energy density @xmath352 and viscous timescale @xmath65 ,
@xmath350 is roughly @xmath353 in general , a fluid can be heated and cooled by the presence of a radiation field .
the energy equation becomes @xmath354 where the last equality arises from evaluating @xmath355 in the comoving orthonormal frame of the fluid , where @xmath356 .
then @xmath357 where the integral is evaluated in the comoving frame and @xmath358 , @xmath359 , and @xmath360 are the opacity , intensity , and emissivity at frequency @xmath17 @xcite . for applications of interest here , radiation mediates _ net _ cooling of the viscous - heated fluid .
hence , we can set @xmath361 for simplicity . the first law of thermodynamics @xmath362 and eq .
( [ ideal_p ] ) give @xmath363 here we define the entropy parameter @xmath105 by @xmath364 , where , in general , @xmath365 .
the form of @xmath366 will depend on the details of the neutron star s microphysics , but it must have the property that @xmath367 when @xmath368 , where @xmath369 is the value of @xmath105 for the unheated fluid ( i.e. no emission from a zero - entropy fluid ) .
accordingly , we replace eq .
( [ g0eqn ] ) for @xmath366 by the following illustrative form @xmath370/\tau_{\rm cool}\ , \ ] ] where @xmath371 is a constant and @xmath126 is the radiation timescale . combining equations ( [ dsdtau])-([cooling_rate ] )
, we find @xmath372}{\tau_{\rm cool}}\right\ } \ . \label{evolve_kappa}\ ] ] in the limit @xmath127 , radiative cooling is unimportant and @xmath105 increases due to viscous heating .
we refer to this regime as the _ no - cooling _ limit . if @xmath128 , then the first term on the right hand side of eq . ( [ evolve_kappa ] ) may be dropped in relation to the second , giving @xmath373 thus , @xmath105 is exponentially driven to @xmath369 .
we refer to this regime as the _ rapid - cooling _
limit , whereby the thermal energy generated by viscosity is radiated immediately and does not heat the gas . in effect , @xmath374 in this limit . in practice
we implement this limit by omitting the @xmath375 in eq .
( [ evolve_estar ] ) .
though we consider only these two limits in our analysis , we expect that , in reality , heating will dominate in some regimes and cooling in others .
we treat both limiting cases in our simulations . w. h. press , and s. a. teukolsky , astrophys .
j. , * 181 * , 513 ( 1973 ) ; d. lai , f. a. rasio , and s. l. shapiro , astrophys .
j. supp .
, * 88 * , 205 ( 1993 ) ; d. skinner , and l. lindblom , astrophys . j. , * 461 * , 920 ( 1996 ) .
s. bonazzola , j. frieben , and e. gourgoulhon , astrophys .
j. , * 460 * , 379 ( 1996 ) ; s. l. shapiro , and s. zane , astrophys .
j. supp . , * 117 * , 531 ( 1998 ) ; t. d. girolamo , and m. vietri , astrophys .
j. , * 581 * , 519 ( 2002 ) ; d. gondek - rosinska , and e. gourgoulhon , phys .
d , * 66 * , 044021 ( 2002 ) .
n. stergioulas , and j. l. friedman , astrophys .
j. , * 492 * , 301 ( 1998 ) ; s. morsink , n. stergioulas , and s. blattning , astrophys .
j. , * 510 * , 854 ( 1999 ) ; s. yoshida , l. rezzolla , s. karino , and y. eriguchi , astrophys .
j. , * 568 * , l41 ( 2002 ) .
h. o. kreiss and j. oliger , _
methods for the approximate solution of time dependent problems _ , garp publication series no .
10 ( world meteorological organization , geneva , 1973 ) ; a. m. abrahams , g. b. cook , s. l. shapiro , and s. a. teukolsky , phys .
d * 49 * , 5153 ( 1994 ) ; m. w. choptuik , e. w. hirschmann , s. l. liebling , and f. pretorius , class .
quant . grav .
* 20 * , 1857 ( 2003 ) .
what appears to be a drift in @xmath241 due to numerical error becomes visible towards the end of the simulation .
this makes the extrapolated value of @xmath241 less reliable than the values of @xmath232 and @xmath239 .
g. s. bisnovatyi - kogan , ya .
b. zeldovich , and i. d. novikov , sov .
astron . * 11 * , 419 ( 1967 ) ; t. w. baumgarte and s. l. shapiro , astrophys .
j. * 526 * , 941 ( 1999 ) ; k. c. b. new and s. l. shapiro , astrophys .
j. * 548 * , 439 ( 2001 ) . | viscosity and magnetic fields drive differentially rotating stars toward uniform rotation , and this process has important consequences in many astrophysical contexts .
for example , merging binary neutron stars can form a `` hypermassive '' remnant , i.e. a differentially rotating star with a mass greater than would be possible for a uniformly rotating star .
the removal of the centrifugal support provided by differential rotation can lead to delayed collapse of the remnant to a black hole , accompanied by a delayed burst of gravitational radiation .
both magnetic fields and viscosity alter the structure of differentially rotating stars on secular timescales , and tracking this evolution presents a strenuous challenge to numerical hydrodynamic codes . here , we present the first evolutions of rapidly rotating stars with shear viscosity in full general relativity . we self - consistently include viscosity in our relativistic hydrodynamic code by solving the fully relativistic navier - stokes equations .
we perform these calculations both in axisymmetry and in full 3 + 1 dimensions . in axisymmetry ,
the resulting reduction in computational costs allows us to follow secular evolution with high resolution over dozens of rotation periods ( thousands of @xmath0 ) .
we find that viscosity operating in a hypermassive star generically leads to the formation of a compact , uniformly rotating core surrounded by a low - density disk .
these uniformly rotating cores are often unstable to gravitational collapse .
we follow the collapse in such cases and determine the mass and the spin of the final black hole and ambient disk .
however , viscous braking of differential rotation in hypermassive neutron stars does not always lead to catastrophic collapse , especially when viscous heating is substantial .
the stabilizing influences of viscous heating , which generates enhanced thermal pressure , and centrifugal support prevent collapse in some cases , at least until the star cools . in all cases studied , the rest mass of the resulting disk is found to be 10 - 20% of the original star , whether surrounding a uniformly rotating core or a rotating black hole .
this study represents an important step toward understanding secular effects in relativistic stars and foreshadows more detailed , future simulations , including those involving magnetic fields . |
the main aim of the shapley supercluster survey ( shass ) is to quantify the influence of hierarchical mass assembly on galaxy evolution and to follow such evolution from filaments to cluster cores , identifying the primary location and mechanisms for the transformation of spirals into s0s and des .
the most massive structures in the local universe are superclusters , which are still collapsing with galaxy clusters and groups frequently interacting and merging , and where a significant number of galaxies are encountering dense environments for the first time .
the shapley supercluster ( hereafter ssc ) was chosen because of i ) the peculiar cluster , galaxy and baryon overdensities @xcite ; ii ) the relative dynamical immaturity of this supercluster and the possible presence of infalling dark matter haloes as well as evidence of cluster - cluster mergers ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) ; iii ) the possibility that it is the most massive bound structure known in the universe , at least in the 10mpc central region ( see * ? ? ?
these characteristics make the ssc an ideal laboratory for studying the impact of hierarchical cluster assembly on galaxy evolution and to sample different environments ( groups , filaments , clusters ) .
furthermore , the redshift range of this structure ( [email protected] * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) makes it feasible to measure the properties of member galaxies down to the dwarf regime , providing that the observations reach the suitable depth . a detailed discussion of the scientific aspects of the survey is given in merluzzi et al .
( @xcite ) .
although , the ssc has been investigated by numerous authors since its discovery ( shapley @xcite ) both for its cosmological implications ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) and for studies of cluster - cluster interactions ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , none of them could systematically tackle the issue of galaxy evolution in the supercluster environment due to the lack of accurate and homogeneous multi - band imaging covering such an extended structure .
shass aims to fill this gap measuring the integrated [ magnitudes , colours , star formation rates ( sfrs ) ] and internal ( morphological features , internal colour gradients ) properties of the supercluster galaxies .
shass will map a region of @xmath0 260 h@xmath12 mpc@xmath1 ( at z=0.048 ) , centred on the ssc core , which is constituted by three abell clusters : a3558 ( z=0.048 , melnick & quintana @xcite , metcalfe , godwin & spenser @xcite ; abell richness @xmath13=4 , abell , corwin & olowin @xcite ) , a3562 ( z=0.049 , @xmath13=2 ) and a3556 ( z=0.0479 , @xmath13=0 ) ; and two poor clusters sc1327 - 312 and sc 1329 - 313 .
the present survey covers also six other abell clusters : a3552 , a3554 , a3559 , a3560 , as0724 , as0726 , as shown in fig .
the survey boundaries are chosen not only to cover all 11 clusters and the likely connecting filaments , but also to extend into the field and to map the structures directly connected to the ssc core . in @xcite we derived the stellar mass density distribution based on supercluster members showing that all the clusters in the shass area are embedded in a common network and identified a filament connecting the ssc core and the cluster a3559 as well as the less pronounced overdensity extending from the ssc core towards a3560 .
the data set of the survey includes optical ( @xmath2 ) and nir ( @xmath14 ) imaging acquired with ( vst ) and visible and infrared survey telescope for astronomy ( vista ) respectively , and optical spectroscopy with aaomega . at present
the @xmath10-band imaging and aaomega spectroscopic surveys are completed , while the other observations are ongoing . in addition , the recent public release of data from the wide - field infrared survey explorer ( @xmath15 , * ? ? ?
* ) provides photometry at both near - ir ( 3.4,4.6@xmath16 m ) and mid - ir ( 12,22@xmath16 m ) wavelengths , allowing independent measurements of stellar masses down to @xmath17=10@xmath18m@xmath19 at 10@xmath3 and sfr down to 0.46m@xmath20yr@xmath21 at 10@xmath3 ( 0.2m@xmath20yr@xmath21 at 5@xmath3 ) . finally , in the 23deg@xmath1 of the ssc core , panoramic imaging in the uv ( galaxy evolution explorer , galex ) , optical ( eso wide field imager , wfi ) , nir ( ukirt / wfcam ) and mid - infrared ( spitzer / mips ) are also available @xcite .
the optical survey , whose coverage is indicated by the 1deg@xmath1 boxes in fig . [ fig1 ] , will enable us to i ) derive accurate morphologies , structural parameters ( @xmath22_log_@xmath23 and @xmath24 ) as well as detect some of the observational signatures related to the different processes experienced by supercluster galaxies ( e.g. extraplanar material ) ; ii ) estimate accurate colours , photo - zs ( @xmath25 , see * ? ? ?
* ) and stellar masses ; iii ) evaluate the sfrs and resolve the star forming regions at least for the subsample of brighter galaxies .
the survey depth enables global and internal physical properties of shapley galaxies to be derived down to @xmath26 + 6 . in the first case of obtaining accurate measurements of aperture photometry and colours , we require signal - to - noise ratios ( snr ) of 20 in all four bands for ssc galaxies down to @xmath26 + 6 . secondly ,
for the morphological analysis and resolving internal properties and structures there is a more stringent requirement of snr@xmath0100 ( in a 3@xmath4 diameter aperture , see * ? ? ?
* ; * ? ? ?
* ) for the deeper @xmath6-band imaging .
for this reason we are collecting the @xmath6-band imaging under the best observing conditions , with a full width at half - maximum ( fwhm)@xmath00.8arcsec or better , corresponding to 0.75kpc at z=0.048 .
additionally , the @xmath6 imaging is fundamental to our weak lensing analysis , to ensure a sufficient density of lensed background galaxies with shape measurements . with these data it will be possible to separate the different morphological types , trace ongoing sf ,
reveal recent interaction or merging activities and thus obtain a census of galaxies whose structure appears disturbed by the environment ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) .
in order to achieve the scientific goals of the survey , accurate photometry is required .
this implies a _
clean _ source catalogue containing , together with the measured photometric properties , indicators of the reliability of these measurements .
this paper describes the methodology used to produce the photometric catalogues and the adopted procedures .
observations are overviewed in sect .
[ sec:2 ] . in sect .
[ sec:3 ] we describe the construction of the catalogues , the criteria to classify spurious objects and unreliable detections , the flags adopted in the catalogues and the procedure for star / galaxy separation . the accuracy and completeness of the derived photometry
is discussed in sect .
[ sec:4 ] .
each parameter of the released @xmath10-band catalogues is detailed in sect .
[ sec:5 ] and the summary is given in sect .
[ sec:6 ] . throughout the paper , we assume a cosmology with @xmath27= 0.3 , @xmath28= 0.7 and h@xmath29= 70 km s@xmath21 mpc@xmath30 .
according to this cosmological model 1 arcmin corresponds to 56.46 kpc at z=0.048 .
the magnitudes are given in the ab photometric system .
the optical survey ( pi : p. merluzzi ) is being carried out using the italian inaf guaranteed time of observations ( gto ) with omegacam on the 2.6 m eso vlt survey telescope located at cerro paranal ( chile ) .
the camera has a corrected field of view of 1@xmath311@xmath32 , corresponding to @xmath0 [email protected] h@xmath12 mpc@xmath34 at the supercluster redshift , sampled at 0.21arcsec per pixel with a 16k@xmath3316k detector mosaic of 32 ccds , with gaps of 25 - 85arcsec in between chips . each field
is observed in four bands : @xmath2 . to achieve the required depth , total exposure times for each pointing are 2955s in @xmath8 , 1400s in @xmath9 , 2664s in @xmath6 and 1000s in @xmath10 . to bridge the gaps a diagonal dither pattern of five exposures in @xmath8 , @xmath9 and @xmath10 , and nine exposures in @xmath6
is performed , with step size of 25@xmath4 ( 15@xmath4 for the @xmath6 band ) in x , and 85@xmath4 ( 45@xmath4 in @xmath6 ) in the y - direction .
the total area is covered by 23 contiguous vst pointings overlapping by @xmath03@xmath35 as shown in fig .
[ fig1 ] , where dots denote the spectroscopic supercluster members ( @xmath36kms@xmath21 ) available from literature at the time of the survey planning .
the x - ray centres are indicated by crosses for all the known clusters except as0726 , whose centre is derived by a dynamical analysis .
= 13500 - 16000 km s@xmath30 , taken from literature .
note that the literature redshift coverage is not uniform across the region shown .
black crosses show the cluster centres . ]
the survey started in 2012 february and will be completed in 2015 ( spanning eso periods p88-p95 ) , provided that all the foreseen observations are carried out . at present
, the survey coverage differs for each band with only the @xmath10-band observations available for the whole area .
this implies that the results concerning the quality of the photometry ( depth , completeness , accuracy ) are based on a representative subsample of the final catalogues for @xmath37 ( 48% , 43% and 61% , respectively ) bands and for the whole catalogue in @xmath10 band .
the data are collected on clear and photometric nights with good and uniform seeing conditions . in fig .
[ fig2 ] we plot the seeing values of 11 , 10 , 14 and 23 fields in @xmath2 , respectively . out of the observed fields about 80% ( @xmath38 ) and 90% ( @xmath6 ) are acquired with fwhm@xmath39 0.8arcsec , with the median seeing in @xmath6 band equal to 0.6arcsec , corresponding to 0.56 kpc h@xmath7 at z=0.048 .
the @xmath8 band is characterized by a slightly poorer seeing .
we discuss the effect of the seeing on the aperture magnitudes in sect .
[ sec:3.1.2 ] .
the data reduction is already described in merluzzi et al .
@xcite . here
we summarize the main steps for the reader s convenience .
images are reduced and combined using the vst tube imaging pipeline ( grado et al .
@xcite ) , developed _ ad hoc _ for the vst data .
the pipeline follows the standard procedures for bias subtraction and flat - field correction .
a normalized combination of the dome and twilight flats , in which the twilight flat is passed through a low - pass filter first , were used to create the master flat .
a gain harmonization procedure has been applied , finding the relative ccd gain coefficients which minimizes the background level differences in adjacent ccds .
a further correction is applied to account for the light scattered by the telescope and instrumental baffling .
this is an additional component to the background , which , if not corrected for , causes a position - dependent bias in the photometric measurements .
this component is subtracted through the determination and the application of the illumination correction ( ic ) map .
the ic map is determined by comparing the magnitudes of photometric standard fields with the corresponding sdss - dr8 ( sloan digital sky survey - data release 8) psf magnitudes . for the @xmath10 band
a correction is required because of the fringe pattern due to thin - film interference effects in the detector from sky emission lines .
the fringing pattern is estimated as the ratio between the super - flat and the twilight sky flat , where super - flat is obtained by overscan and bias correcting a sigma clipped combination of science images .
the fringe pattern is subtracted from the image , applying a scale factor which minimizes the absolute difference between the peak and valley values ( maximum and minimum in the image background ) in the fringe corrected image .
the photometric calibration onto the sdss photometric system is performed in two steps : first a relative photometric calibration among the exposures contributing to the final mosaic image is obtained through the comparison of the magnitudes of the bright unsaturated stars in the different exposures , using the software scamp ( bertin @xcite ) ; then the absolute photometric calibration is computed on the photometric nights comparing the observed magnitude of stars in photometric standard fields with sdss photometry . for those fields observed on clear nights ,
we take advantage of the sample of bright unsaturated stars in the overlapping region between clear and photometric pointings and by using scamp , each exposure of the clear fields is calibrated on to the contiguous photometrically calibrated field . the magnitude
is then calibrated by adopting the following relation : @xmath40 where @xmath41 is the magnitude of the star in the standard system , @xmath42 is the instrumental magnitude , @xmath43 is the coefficient of the colour term , @xmath44 is the colour of the star in the standard system , @xmath45 is the extinction coefficient , @xmath46 is the airmass and @xmath47 is the zero - point .
the results are reported in table [ tab1 ]
. .absolute photometric calibration coefficients .
[ cols= " < , < , < , < , < " , ]
the shass will map an area of @xmath023 deg@xmath1 ( @xmath0 260 h@xmath12 mpc@xmath1 at z=0.048 ) of the ssc with the principal aim being to quantify the influence of hierarchical mass assembly on galaxy evolution , and to follow this evolution from filaments to cluster cores .
shass will provide the first homogeneous multi - band imaging covering the central region of the supercluster , including optical ( @xmath2 ) and nir ( @xmath14 ) imaging acquired with vst and vista , which allows accurate multiband photometry to be obtained for the galaxy population down to m*+6 at the supercluster redshift .
in particular , the @xmath6-band images are collected with a median seeing equal to 0.6arcsec , corresponding to 0.56 kpc h@xmath7 at z=0.048 and thus enabling the internal properties of supercluster galaxies to be studied and distinguish the impact of the environment on their evolution . in this article
, we described the methodology for producing photometric catalogues .
the optical survey is ongoing and the analysis presented in this work is performed on 11 , 10 , 14 and vst fields out of 23 in @xmath37 bands , nevertheless in all bands the considered subsamples are representative of the final whole sample .
the @xmath10-band imaging is instead complete .
the catalogues are produced using the software sextractor @xcite in conjuction with psfex ( @xcite ) , and a careful analysis of the software outputs .
we were able to obtain a robust separation between stars and galaxies up to the completeness limit of the optical data , through a progressive approach using : i ) the _ stellarity index _ ( ) ; ii ) the half - light radius ( ) ; iii ) the new sextractor classifier ; iv ) the peak of the surface brightness above background ( ` \mathrm{\mu_{max } } ` ) ; v ) a final visual inspection for objects classified as galaxies but with @xmath480.90 .
the shass catalogues reach average 5@xmath3 limiting magnitudes inside a 3@xmath4 aperture of @xmath2=[24.4,24.6,24.1,23.3 ] and a completeness limit of @xmath2=[23.8,23.8,23.5,22.0 ] , which corresponds to @xmath0 m*@xmath5 + 8.5 at the supercluster redshift .
these values correspond to the survey expectations .
the @xmath10-band catalogue is released to the community through the use of the virtual observatory tools .
it includes 734 319 sources down to @xmath10=22.0 over the whole area .
the catalogue is 93% complete at this magnitude limit and 34% of the sources are galaxies .
the service is publicly accessible via browser at the address http://shass.na.astro.it .
the shass data base is also publicly available within the euro - vo registry framework , under the inaf - dame astronomical archive identification authority ( ivo://dame.astro.it / shass - i ) .
the authors thank the referee , r. smith , for his constructive comments and suggestions . this work is based on data collected with the eso - vlt survey telescope with omegacam ( eso programmes 088.a-4008 , 089.a-0095 , 090.a-0094 ,
091.a-0050 ) using italian inaf guaranteed time observations .
the data base has made use of svocat , a vo publishing tool developed in the framework of the spanish virtual observatory project supported by the spanish mineco through grant aya @xmath49 and the cosadie fp7 project ( call @xmath50 research infrastructures , project @xmath51 ) .
svocat is maintained by the data archive unit of the centro de astrobiologa ( csic -inta ) .
the research leading to these results has received funding from the european community s seventh framework programme ( fp7/2007 - 13 ) under grant agreement number 312430 ( opticon ; pi : p. merluzzi ) and prin - inaf 2011 : _ galaxy evolution with the vlt surveys telescope ( vst ) _
( pi a. grado ) .
cph was funded by conicyt anillo project act-1122 .
pm thanks m. petr - gotzens for her support in the vst observations .
am and mb acknowledges financial support from prin - inaf 2014 : _ glittering kaleidoscopes in the sky , the multifaceted nature and role of galaxy clusters _ ( pi m. nonino ) .
pm and gb acknowledge financial support from prin - inaf 2014 : _ galaxy evolution from cluster cores to filaments _ ( pi b.m .
poggianti ) .
bertin , e. 2011 , in evans i. n. , accomazzi a. , mink d. j. , rots a. h. , eds , asp conf .
442 , automated morphometry with sextractor and psfex , p. 435
bouy h. , bertin e. , moraux e. , cuillandre j. c. , bouvier j. , barrado d. , solano e. , bayo a. , 2013 , a&a 554 , a101 mohr , j. j. et al .
2012 , the dark energy survey data processing and calibration system , publicated in software and cyberinfrastructure for astronomy ii .
proceedings of the spie , volume 8451 , article i d . | the shapley supercluster survey is a multi - wavelength survey covering an area of @xmath023 deg@xmath1 ( @xmath0 260 mpc@xmath1 at z=0.048 ) around the supercluster core , including nine abell and two poor clusters , having redshifts in the range 0.045 - 0.050 .
the survey aims to investigate the role of the cluster - scale mass assembly on the evolution of galaxies , mapping the effects of the environment from the cores of the clusters to their outskirts and along the filaments .
the optical ( @xmath2 ) imaging acquired with omegacam on the vlt survey telescope is essential to achieve the project goals providing accurate multi - band photometry for the galaxy population down to m*+6 .
we describe the methodology adopted to construct the optical catalogues and to separate extended and point - like sources .
the catalogues reach average 5@xmath3 limiting magnitudes within a 3@xmath4 diameter aperture of @xmath2=[24.4,24.6,24.1,23.3 ] and are 93% complete down to @xmath2=[23.8,23.8,23.5,22.0]mag , corresponding to @xmath0 m*@xmath5 + 8.5 .
the data are highly uniform in terms of observing conditions and all acquired with seeing less than 1.1arcsec full width at half - maximum .
the median seeing in @xmath6-band is 0.6arcsec , corresponding to 0.56kpch@xmath7 at z=0.048 .
while the observations in the @xmath8 , @xmath9 and @xmath6 bands are still ongoing , the @xmath10-band observations have been completed , and we present the @xmath10-band catalogue over the whole survey area .
the latter is released and it will be regularly updated , through the use of the virtual observatory tools .
this includes 734,319 sources down to @xmath10=22.0mag and it is the first optical homogeneous catalogue at such a depth , covering the central region of the shapley supercluster .
[ firstpage ] methods : data analysis - methods : observational - catalogues - virtual observatory tools - galaxies : clusters : general - galaxies : clusters : individual : a3552 - galaxies : clusters : individual : a3554 - galaxies : clusters : individual : a3556 - galaxies : clusters : individual : a3558 - galaxies : clusters : individual : a3559 - galaxies : clusters : individual : a3560 - galaxies : clusters : individual : a3562 - galaxies : clusters : individual : as0724 - galaxies : clusters : individual : as0726 - galaxies : clusters : individual : sc1327 - 312 - galaxies : clusters : individual : sc1327 - 313 - galaxies : photometry . |
understanding the vacuum structure of qcd is one of the most interesting and intriguing field of research for about thirty years now .
new signals of non - trivial topological fluctuations in qcd have been proposed in recent times , like chiral magnetic effect @xcite and anomaly induced transport @xcite .
it is an ongoing topic of research in trying to understand appropriate experimental signatures of chiral magnetic effect in heavy ion collision experiments at rhic at bnl and at cern .
recent studies of hydrodynamic evolution in presence of chiral anomaly has also provided hints towards charge separation of different pion species @xcite .
however this is not a unique signature of non - trivial topological fluctuations in qcd .
effects of topological fluctuations would also show up in ratios of thermodynamic observables near the chiral phase transition at vanishingly small baryon densities if indeed the strongly interacting matter created in heavy - ion collision thermalizes . as known from the pioneering paper @xcite
that @xmath0 though an anomalous symmetry in qcd , may affect the nature of phase transition in qcd with two light quark flavors .
if the @xmath0 anomaly is not effectively restored at the chiral crossover transition then in the limit of vanishingly small light quark mass , the phase transition is second order with @xmath1 critical exponents , otherwise it could be either first order or second order with @xmath2 universality class @xcite .
remarkably , it is known from model quantum field theories in the @xmath1 universality class that the ratio of higher order baryon fluctuations like @xmath3 should have a very characteristic behavior at the chiral crossover transition @xcite .
measuring such observables in experiments would allow us to understand whether @xmath0 is still effectively broken and thus indirectly find out the topological fluctuations responsible for it . our earlier work @xcite has lead to the hints that @xmath0 may not be effectively restored at @xmath4 , @xmath5 is the chiral crossover transition temperature and weakly interacting instantons may be responsible for its breaking at such temperatures . in this talk ,
i present our latest results on our ongoing studies about understanding the mechanism for the @xmath0 breaking at temperatures between @xmath6-@xmath7 .
we investigate the nature of the topological structures that cause @xmath0 breaking at these temperatures from the eigenvalue spectrum of the qcd dirac operator and from the topological susceptibility using lattice techniques .
i conclude the discussion with the possible consequences of presence of such non - trivial topological structures at these temperatures for the experiments .
effects of anomalous @xmath0 symmetry on qcd phase transition with two light quark flavors is known from model quantum field theories with same symmetries as qcd .
however the @xmath0 breaking is implemented through an effective term in the lagrangian whose strength can only be determined from first principles lattice qcd studies .
however lattice studies have their own share of systematic uncertainties .
one needs to choose a sufficiently large lattice volume to contain enough of these topological structures .
second and the most crucial aspect is to control the lattice cut - off effects .
most lattice fermion discretizations break the continuum chiral symmetry and the flavor singlet @xmath0 explicitly .
only two known lattice fermion discretizations , the overlap and the domain wall fermions in the limit of infinite large fifth dimensional extent on which it is defined , are known to realize chiral symmetry explicitly on the lattice .
these operators also satisfy an exact index theorem , thus the @xmath0 breaking can be related to the non - trivial gauge field topology .
we use the gauge configurations generated with mbius domain wall fermion discretization and iwasaki gauge action with dislocation suppressing determinant .
the details of the gauge configurations are in @xcite .
the light and the strange quark masses are physical such that @xmath8 .
the small residual chiral symmetry breaking for these domain wall fermion configurations is @xmath9 and the index is not exactly realized .
hence we use overlap fermions as valence quarks to count the zero modes of the gauge configurations and measure the eigenvalue spectrum .
the overlap operator @xmath10 $ ] , consists of a sign function of wilson dirac operator @xmath11 , which is the lattice discretization of the continuum dirac operator with an irrelevant term and a domain wall height parameter @xmath12 .
the sign function was measured to a precision of about @xmath13 and the violation of the exact chiral symmetry given by the ginsparg - wilson relation was of the same order of magnitude for each gauge configuration . on each configuration
, @xmath14 eigenvalues of @xmath15 were measured using the kalkreuter - simma ritz algorithm @xcite .
we analyzed about @xmath16 configurations at each temperature .
our lattice volumes are sufficiently large enough , @xmath17 essential for such a study . in the left panel of fig .
[ fig : dweig ] we show our results for the eigenvalue density of the qcd dirac operator as a function of temperature .
as the temperature increases one observes a gradual appearance of a small near - zero mode peak separating from the bulk eigenvalues whose contribution diminishes at @xmath7 as shown in the right panel of fig .
[ fig : dweig ] .
one needs a careful study of the near - zero modes since its origin may be due to residual chiral symmetry breaking effects of the domain wall fermions @xcite .
one of the measures of @xmath0 breaking is the non - vanishing value of the difference between integrated pion and delta meson correlation functions , @xmath18 .
from the left panel of fig .
[ fig : chipd ] it is evident that this quantity is non - zero at even @xmath7 and a large contribution comes from the first @xmath14 eigenvalues .
it is also known that if the analytic(bulk ) part of the eigenvalue spectrum goes as @xmath19 , then @xmath0 breaking effects are invisible in upto 6-point correlation functions in the scalar and pseudo - scalar sectors @xcite .
we observe that the analytic part of the renormalized eigenvalue spectrum has a behavior quite different from @xmath20 at @xmath7 irrespective of the lattice discretization ( right panel of fig .
[ fig : chipd ] ) , by comparing our eigenspectrum with those previously measured by us using improved staggered ( hisq ) fermions @xcite .
we conclude that @xmath0 is not effectively restored at @xmath7 and the contribution of the analytic part of the eigenspectrum to its breaking is quite robust unaffected by lattice cut - off effects . for more details
see @xcite .
breaking in the scalar - pseudoscalar sector measured through the observable @xmath18 ( left panel ) shows that @xmath0 is not effectively restored at @xmath7 .
the black data points are the contribution from the lowest 50 eigenvalues to the total value denoted by the red points taken from ref .
@xcite . in the right panel ,
the renormalized eigenvalue spectrum is plotted for different fermion discretizations show quite a robust behavior for the bulk part.,title="fig : " ] breaking in the scalar - pseudoscalar sector measured through the observable @xmath18 ( left panel ) shows that @xmath0 is not effectively restored at @xmath7 .
the black data points are the contribution from the lowest 50 eigenvalues to the total value denoted by the red points taken from ref . @xcite . in the right panel ,
the renormalized eigenvalue spectrum is plotted for different fermion discretizations show quite a robust behavior for the bulk part.,title="fig : " ]
the @xmath0 breaking at @xmath21 leads to a finite @xmath22 mass and is related to the physics of instantons @xcite .
near the chiral crossover , the qcd vacuum properties are not yet well understood .
instanton liquid model @xcite provides explanation for chiral symmetry breaking and properties of hadronic correlations @xcite .
it is also known that qcd instantons at finite temperature with non - trivial holonomy has substructures called dyons @xcite which carry both electric and magnetic charges . in order to understand the existence and interactions between these different topological structures , we measure the topological susceptibility in qcd , defined as @xmath23 $ ] .
it measures the local fluctuations of the topological charge @xmath24 though @xmath25 in the thermodynamic limit .
since we use overlap fermions with an exact index theorem , we for the first time , report the results for @xmath26 in lattice qcd with exact chiral symmetry at high temperatures .
the @xmath26 shows sensitivity to the quark mass as seen in the left panel of fig .
[ fig : topsusc ] . for light quark masses about twice the physical values ,
@xmath26 is enhanced by factor two near @xmath5 .
fitting @xmath27 to a function @xmath28 , we get an exponent @xmath29 for physical quark masses and @xmath30 for @xmath31 . for a dilute gas of instantons , including first order quantum fluctuations over the classical gauge configurations
gives temperature dependence of @xmath32 .
from the temperature dependence of our data for @xmath26 , it is evident that dilute gas picture has not yet set in at @xmath7
. it also gives a hints about possible existence of dyons in the qcd medium just above @xmath5 , though the trend could be also consistent with the instanton liquid model .
a still better probe is to look into the higher moments of topological charge fluctuations . in the right panel of fig .
[ fig : topsusc ] , we show our results for @xmath33 . while near @xmath5 , the data is consistent with an interacting instanton ensemble that scales as @xmath34 , the dilute instanton gas limit do not set in even at @xmath7
. there are strong residual interactions between the finite temperature instantons or they dissociate into dyonic components .
this is consistent with our earlier observation that dilute gas limit may set in only around @xmath35 @xcite .
our study has comprehensively shown that @xmath0 for qcd with two light quark flavors is explicitly broken at @xmath7 . though the near - zero part of the qcd dirac eigenspectrum still needs to be studied in detail , the contribution of the analytic ( bulk ) part of the spectrum to @xmath0 breaking is quite robust insensitive to lattice cut - off effects .
we also looked into the topological constituents that contribute dominantly to @xmath26 just above @xmath5 .
we speculate about the presence of dyons from our data .
it has been recently reported that magnetic monopoles may explain the jet - quenching phenomena in the low @xmath36 region in the range @xmath37gev@xmath38 gev @xcite .
dyons with non - zero magnetic charges could also possibly contribute to the jet quenching phenomenology in the strongly coupled qgp .
our first results for @xmath26 in the region near @xmath5 , also opens up the possibility of accounting for the chiral imbalance in anomalous hydrodynamics through @xmath26 as an input instead of a non - zero axial chemical potential .
@xmath39 the work has been partially supported through the doe under contract no .
de - sc0012704 and the bmbf under grants 05p12pbcta and 05p15pbcaa .
the gpu codes used in our work were in part based on some publicly available quda libraries @xcite .
99 k. fukushima , d. kharzeev and h. warringa , phys . rev .
d 78 , 074033 ( 2008 ) .
k. fukushima , d. kharzeev and h. warringa , phys .
104 , 212001 ( 2010 ) .
y. hirono , t. hirano , d. e. kharzeev , [ arxiv:1412.0311 ] ; r. d. pisarski and f. wilczek , phys .
d 29 , 338 ( 1984 ) .
a. pelissetto , e. vicari , phys .
d 88 , 105018 ( 2013 ) ; y. nakayama and t. ohtsuki , phys .
d 91 , 021901 ( 2015 ) .
b. friman et .
al . , eur .
j. c 71 , 1694 ( 2011 ) . v. dick et .
al . , phys .
d 91 , 094504 ( 2015 ) .
m. i. buchoff et .
al . , phys .
d 89 , 054514 ( 2014 ) ; t. bhattacharya et .
al . , phys . rev
113 , 082001 ( 2014 ) .
t. kalkreuter and h. simma , comput .
93 , 33 ( 1996 ) .
g. cossu et .
al . , jlqcd collaboration [ arxiv:1510.07395 [ hep - lat ] ] .
s. aoki , h. fukaya , y. taniguchi , phys .
d 86 , 114512 ( 2012 ) . v.
al . , in preparation .
g. thooft , phys .
d 14 , 3432 ( 1976 ) .
e. v. shuryak , nucl .
b 203 , 93 ( 1982 ) .
e. -m .
ilgenfritz and e. v. shuryak nucl .
b 319 , 511 ( 1989 ) .
t. kraan and p. van baal , nucl .
b 533 , 627 ( 1998 ) ; k. -m .
lee and c. -h .
lu , phys .
d 58 , 025011 ( 1998 ) .
j. xu , j. liao , m. gyulassy , chin .
32 , 092501 ( 2015 ) .
m. a. clark et .
al . , comput .
181 , 1517 ( 2010 ) . | the nature of chiral phase transition for two flavor qcd is an interesting but unresolved problem .
one of the most intriguing issues is whether or not the anomalous u(1 ) symmetry in the flavor sector is effectively restored along with the chiral symmetry .
this may determine the universality class of the chiral phase transition . since the physics near the chiral phase transition is essentially non - perturbative ,
we employ first principles lattice techniques to address this issue .
we use overlap fermions , which have exact chiral symmetry on the lattice , to probe the anomalous u(1 ) symmetry violation of 2 + 1 flavor dynamical qcd configurations with domain wall fermions .
the latter also optimally preserves chiral and flavor symmetries on the lattice , since it is known that the remnant chiral symmetry of the light quarks influences the scaling of the chiral condensate in the crossover transition region .
we observe that the anomalous u(1 ) is not effectively restored in the chiral crossover region .
we perform a systematic study of the finite size and cut - off effects since the signals of u(1 ) violation are sensitive to it .
we also provide a glimpse of the microscopic topological structures of the qcd medium that are responsible for the strongly interacting nature of the quark gluon plasma phase .
we study the effect of these microscopic constituents through our first calculations for the topological susceptibility of qcd at finite temperature , which could be a crucial input for the equation of state for anomalous hydrodynamics .
chiral fermions ; axial anomaly ; gauge field topology ; topological susceptibility ; instantons |
wave interaction problems are of great interest from a theoretical point of view as well as for possible applications in applied sciences . in this framework ,
a prominent role has been played by model evolution equations whose canonical structure allows for exact solutions describing relevant wave profiles as it happens for @xmath0 hyperbolic systems involving two dependent and two independent variables .
as well known , these mathematical models can be recast into a form expressing the evolution of a privileged set of field variables ( riemann invariants ) along the related characteristic curves . moreover , in the homogeneous case , the quasilinear system can be reduced to linear form through the classical hodograph transformation which , in principle , can be solved by using the riemann method @xcite . therefore @xmath1 homogeneous models represent a prototype for determining classes of systems whose canonical structure allows for exact solutions that facilitate a full understanding of the interaction process of hyperbolic waves . furthermore , for such a kind of models , the wave dynamics is only ruled by the behaviour of the solutions along the associated families of characteristic curves and , by making use of the special class of simple wave solutions @xcite , it is possible to understand in detail the role played by different families of characteristic curves inside the interaction region .
within such a theoretical framework , the riemann method was extended to the nonhomogeneous case in @xcite and a large new classes of solutions to quasilinear systems of pdes have been obtained in @xcite . furthermore a combined use of the hodograph method and of the differential constraints technique was considered in @xcite and quite recently was used in order to study nonlinear wave interactions @xcite as well as discontinuous initial value problems @xcite for homogeneous and nonhomogeneous @xmath1 systems . in a different way from @xmath1 models of first order pdes , for strictly hyperbolic systems involving @xmath2 dependent variables and two independent variables , the riemann invariants in general
do not exist so that a detailed description of wave interactions in terms of exact and closed form solutions to initial value problems is a hard task . in such a context
the construction of solutions describing regular interactions of simple waves , the conditions for the superposition of riemann waves in terms of initial data as well as the identification of the regions of interactions have been studied by several authors @xcite .
however , within the latter context a remarkable role is played by homogeneous systems of hydrodynamic type with semi - hamiltonian structure which can be diagonalized in terms of suitable field variables which are , in fact , riemann invariants . actually , via the generalized hodograph method
a general solution of these systems can be obtained @xcite .
recently , in @xcite , the approach worked out in @xcite for @xmath1 hyperbolic systems , has been enlarged to these class of diagonalizable semi - hamiltonian homogeneous hyperbolic systems in order to perform an accurate description of the associated hyperbolic wave interaction processes .
within such a theoretical framework , here our main aim is to extend to the three dimensional case the procedure proposed in @xcite for hydrodynamic models involving two independent variables .
in particular we will consider the two commuting systems @xcite @xmath3 where @xmath4 and @xmath5 which admit
the conservation laws @xmath6 hydrodynamic type systems ( [ chrom1 ] ) and ( [ chrom2 ] ) are nothing but the chromatography system written in the lagrangian and the euler coordinates respectively .
in passing we notice that the three dimensional two component quasilinear system ( [ hydro ] ) can be written in a form of a single quasilinear equation of second order in two alternative forms @xmath7 where @xmath8 or @xmath9 where @xmath10 in this paper we describe nonlinear wave interactions for a special class of solutions of the quasilinear system ( [ hydro ] ) and , in fact , we extend the study of nonlinear hyperbolic wave interactions to three dimensional case .
furthermore we determine the control parameters for the simple wave deformation in the interaction region in terms of the initial / boundary conditions .
the explicit evaluation of such parameters has proved to be a useful tool for describing special ( soliton - like ) simple wave interactions @xcite as well as to perform quantitative measures or predictions of interest to engineering applications @xcite . in particular , we will focus our attention on the two commuting hydrodynamic type systems ( [ chrom1 ] ) and ( [ chrom2 ] ) , whose general solution is determined in @xcite by the generalized hodograph method @xmath11 which characterizes , through ( [ a - b ] ) , a special class of exact solutions to ( [ hydro ] ) useful for describing nonlinear @xmath12 wave interactions . in ( [ sol ] ) @xmath13 are arbitrary functions .
the paper is organized as follows . in section [ inivalue ] the general solution ( [ sol ] ) is rewritten in terms of the characteristic parameters associated to the characteristic curves of the systems ( [ chrom1 ] ) , ( [ chrom2 ] ) and the explicit expression of the functions @xmath14 there involved , in terms of the initial / boundary data , is obtained . in section [ nwave ] a detailed analytical description of two different @xmath12 wave
interaction problems is presented .
then , in order to validate the analytical results there obtained , in section [ numerical ] , corresponding numerical integrations of ( [ hydro ] ) are shown .
some final comments are given in section [ conclusion ] .
here our aim is to calculate the functions @xmath15 involved in the general solution ( [ sol ] ) once initial / boundary conditions are given . such a result will be useful in the next section in order to study nonlinear @xmath16 wave interactions described by ( [ chrom1 ] ) , ( [ chrom2 ] ) and in turn , through ( [ a - b ] ) , admitted by ( [ hydro ] ) .
first we consider an arbitrary initial / boundary value problem .
let @xmath17 be a smooth curve in the euclidean space of the independent variables @xmath18 @xmath19 and we assume the following boundary data for @xmath20 @xmath21 following a procedure outilined in @xcite , from ( [ sol ] ) evaluated on @xmath22 , we obtain @xmath23 therefore it is straightforward to ascertain that multiplying by @xmath24 the corresponding equation of the set ( [ ini2 ] ) , by taking the sum over @xmath25 , a further integration allows us to express @xmath26 in terms of @xmath27 as follows @xmath28 along with @xmath29 by inserting the expressions ( [ iint ] ) into relations ( [ ini2 ] ) for @xmath30 , multiplying each equation by @xmath31 , taking the sum over @xmath32 and integrating the resulting relation we are able to obtain @xmath33 .
whereupon a further iterative procedure gives rise to @xmath34 where @xmath35 are obtained from @xmath36 with @xmath37 finally , from ( [ a - h ] ) , after some algebra , the function @xmath38 are determined in terms of the boundary data ( [ inicurve ] ) , ( [ ini1 ] ) @xmath39 @xmath40 in ( [ wavelets11 ] ) , ( [ wavelets22 ] ) and in what follows the prime denotes the derivative with respect to the indicated argument .
next , we make use of the general expressions ( [ wavelets11 ] ) , ( [ wavelets22 ] ) in order to describe nonlinear wave interactions prescribed by the quasilinear system ( [ hydro ] ) .
first of all , being solution ( [ sol ] ) implicitily expressed in terms of the functions @xmath41 which remain constant along the appropriate characteristic curves @xmath42 of the system ( [ chrom1 ] ) , ( [ chrom2 ] ) , we introduce the characteristic parameters @xmath43 which are ruled by the following equations @xmath44 so that @xmath45 const on its own characteristic .
therefore , bearing in mind the properties of @xmath46 quasilinear hyperbolic systems @xcite , let us assume @xmath47 which allows to obtain classes of solutions representing superposition of simple waves @xcite .
such solutions are of particular interest from a physical point of view because they cover a wide range of wave phenomena arising in gas dynamics , electrodynamics , chemical engineering , etc . @xcite . therefore ,
owing to ( [ omegak ] ) , the general solution ( [ sol ] ) can be expressed in terms of the characteristic parameters @xmath48 as follows @xmath49 where @xmath50 and , given a function @xmath51 , we denote @xmath52 now , in view of investigating initial / boundary value problems associated to ( [ chrom1 ] ) and ( [ chrom2 ] ) ( i. e. to ( [ charact ] ) ) , although the qualitative features apply to the general case , for simplicity we normalize @xmath48 as follows @xmath53 so that , taking ( [ omegak ] ) into account , the initial / boundary data ( [ inicurve ] ) , ( [ ini1 ] ) for @xmath54 take the form @xmath55 and , from ( [ wavelets11 ] ) , ( [ wavelets22 ] ) , the arbitrary functions @xmath56 as well as @xmath57 specialize as follows @xmath58 @xmath59 therefore , once the initial / boundary data ( [ inidata ] ) are specified , insertion of ( [ wavelets1 ] ) and ( [ wavelets2 ] ) into ( [ wavelets ] ) allows us to investigate in detail the evolution of the resulting wave pulses . finally we remark that , although smooth initial data are prescribed , the solution of quasilinear hyperbolic systems are subject to nonlinear breakdown @xcite so that our analysis will be valid until the blow up of the solution along the characteristic curves @xmath60 does not occur . more precisely , if the initial / boundary data are choosen small enough on a region of the @xmath61plane , then there exists a space interval @xmath62 $ ] in which the gradient catastrophe for solutions of the hydrodynamic type system ( [ chrom1])-([chrom2 ] ) does not occur @xcite . under this assumption
it is possible to select the initial / boundary data ( [ inidata ] ) in such a way that every characteristic @xmath63 has a tangent plane with inclination ( measured with respect to the positive direction of the @xmath64axis ) smaller than any characteristic of the family @xmath65 @xcite .
in such a case , owing to the invariance of @xmath41 along the associated characteristic curve @xmath60 , it is straightforward to ascertain that if we consider initial / boundary data ( [ inidata ] ) with compact support then the @xmath66space of indipendent variables will be divided into disjoint regions of constant states , simple waves and collision regions @xcite . within this analytical framework , in the next section , two different interaction processes will be described and the resulting regions highlighted .
the aim of this section is to give an exact quantitative description of nonlinear @xmath12 wave interactions prescribed by ( [ wavelets ] ) along with ( [ wavelets1 ] ) and ( [ wavelets2 ] ) .
therefore let us consider the initial / boundary data @xmath67 where @xmath68 are real numbers , @xmath69 are smooth functions and @xmath70 @xmath71 are arbitrary constants . in passing we notice that the initial / boundary value problem ( [ rho_k ] ) describes @xmath12 waves localized at @xmath72 in the interval @xmath73 $ ] which , according to ( [ charact ] ) , propagate through regions which are adjacent to constant states so that the pulses in point after a finite space @xmath74 separate themselves and become in fact simple waves .
it results that in the @xmath75space there are @xmath16 distinct simple wave regions @xmath76 ( see figure [ dinteract1 ] and figure [ dinteract3 ] where , for the sake of simplicity , the cases @xmath77 and @xmath78 are shown , respectively , in the plane @xmath79constant and in @xmath80space ) where each characteristic parameter @xmath81 can be explicitly calculated so that , once the initial / boundary data ( [ rho_k ] ) are specified , by using ( [ wavelets ] ) the behaviour of the emerging simple waves can be fully investigated . in particular , for a fixed value @xmath32 , the simple wave travelling along the characteristic curve @xmath82 corresponds to the region @xmath83 and from ( [ wavelets ] ) , ( [ wavelets1 ] ) , ( [ wavelets2 ] ) and ( [ rho_k ] ) , after some algebra , we obtain the characteristic wave parameters in each simple wave region @xmath76 @xmath84 t \rbrace = \alpha _ { k}+\lambda _ { k}\left ( \alpha _ { k}\right ) \label{i}\ ] ] where @xmath85 are given by @xmath86 with @xmath87 the interaction terms @xmath85 represent a quantitative measure of the distortion of the simple wave travelling along @xmath88 which depends on the initial / boundary data ( [ rho_k ] ) and vanish if there is only the localized pulse travelling along @xmath89 , that is @xmath90 next we are interested in studying the interaction between @xmath91 waves and a single pulse initially localized in disjoint intervals .
therefore we consider the following initial / boundary data @xmath92 where @xmath93 denote real constants , @xmath94 are smooth functions and @xmath95 are arbitrary constants . in such a case the pulse travelling along
the characteristic curve @xmath96 traverses region @xmath97 where it is a simple wave , interacts with the pulses travelling along the @xmath89 ( @xmath98 ) characteristic curves and emerges in region @xmath99 as simple wave ( see figure [ dinteract2 ] for @xmath100 ) . in the @xmath101space
there are @xmath102 distinct simple wave regions where each characteristic parameter @xmath103 can be explicitly calculated ( see figure [ dinteract4 ] for @xmath78 ) .
in particular from ( [ wavelets ] ) , ( [ wavelets1 ] ) and ( [ wavelets2 ] ) , along with ( [ data_3 ] ) , we obtain the following expressions for the characteristic wave parameters in each simple wave region region @xmath97 @xmath104t\rbrace = \alpha_{1 } \cr \mbox{with } \quad y_{1}\leq \alpha _ { 1}\leq y_{2},\quad y_{2}\leq \alpha _ { k}\leq y_{3}\ ; \ ; \left ( k\geq 2\right ) } \label{region_i}\ ] ] region
@xmath99 @xmath105t\rbrace = \alpha_{1}+\theta_{1}\left(\alpha_{1}\right ) \cr \mbox{with } \quad y_{1}\leq \alpha _
{ 1}\leq y_{2},\quad \alpha _ { k}\geq y_{4}\quad \left ( k\geq 2\right ) } \label{region_iv}\ ] ] region @xmath106
@xmath107t\rbrace = \alpha_{k}+\theta_{k}\left(\alpha_{k}\right ) \cr \mbox{with } \quad y_{3}\leq
\alpha_{k}\leq y_{4},\quad \alpha _ { 1}\leq y_{1},\quad \alpha _ { j}\geq y_{4 } \ ; \ ; ( j > k),\quad \alpha _ { j}\leq y_{3 } \ ; \ ; ( j < k ) \label{region_ii}}\ ] ] in ( [ region_iv ] ) , ( [ region_ii ] ) the functions @xmath108 measure the @xmath109 wave parameters distortion due to the interaction with the pulses travelling along the @xmath110 characteristics and are given by @xmath111 @xmath112 with @xmath113 we notice that the pulses generated by the localized data @xmath114 , interact with the fastest simple wave travelling along the @xmath115 characteristic curve and emerge , after interaction , with altered profiles .
the parameters distortion @xmath108 are given in terms of the initial / boundary data and determine a quantitative measure of this alteration . therefore the resulting wave s distortion depend strongly by the set of initial / boundary values under interest .
@xmath116{figura1.eps } } \end{array } $ ] plane of the interaction process generated by initial / boundary data ( [ data_3 ] ) for @xmath117 .
the colored ( orange ) region represents the interaction region .
@xmath118 are the simple wave regions .
the dashed lines would correspond to the choice @xmath119 for @xmath120 and point out that the emerging pulses are distorted by the interaction process . , scaledwidth=60.0% ] @xmath121{figura3.ps } } \put(-60,90){\fontsize{12}{12}$x $ } \put(-150,2){\fontsize{12}{12 } $ y_{1}$ } \put(-200,-1){\fontsize{12}{12 } $ y_{2}$}\put(-270,140){\fontsize{12}{12}$i_{1}$ } \put(-355,115){\fontsize{12}{12}$i_{2}$ } \put(-335,-5){\fontsize{12}{12}$y$ } \put(-110,40){\fontsize{12}{12}$t$ } \put(-65,10){\fontsize{12}{12}$0 $ } & \end{array}$ ] @xmath122{figura4.ps } } \put(-365,80){\fontsize{12}{12}$x $ } \put(-340,50){\fontsize{12}{12}$t $ } \put(-120,-10){\fontsize{12}{12}$y $ } \put(-155,90){\fontsize{12}{12 } $ ii$ } \put(-300,-2){\fontsize{12}{12 } $ i$ } \put(-330,-10){\fontsize{12}{12 } $ y_{1}$ } \put(-293,-10){\fontsize{12}{12 } $ y_{2}$ } \put(-240,140 ) { \fontsize{12}{12}$i_{2}$ } \put(-240,-10){\fontsize{12}{12}$y_{3}$ } \put(-220,-10){\fontsize{12}{12}$y_{4}$ } \put(-360,-10){\fontsize{12}{12}$0 $ } & \end{array}$ ] @xmath123{figura5_a } & \includegraphics[width=0.4\textwidth]{figura5_b } \\
\includegraphics[width=0.50\textwidth]{figura5_c } & \includegraphics[width=0.4\textwidth]{figura5_d}\\ \includegraphics[width=0.50\textwidth]{figura5_e } & \includegraphics[width=0.4\textwidth]{figura5_f } \end{array } \put(-170,130){\fontsize{12}{12}$t$ } \put(-380,130){\fontsize{12}{12}$t$ } \put(-170,-30){\fontsize{12}{12}$t$ } \put(-380,-30){\fontsize{12}{12}$t$ } \put(-170,-190){\fontsize{12}{12}$t$ } \put(-380,-190){\fontsize{12}{12}$t$ } \put(-280,-60){\fontsize{12}{12}$y$ } \put(-60,-60){\fontsize{12}{12}$y$ } \put(-280,100){\fontsize{12}{12}$y$ } \put(-60,100){\fontsize{12}{12}$y$ } \put(-280,-210){\fontsize{12}{12}$y$ } \put(-60,-210){\fontsize{12}{12}$y$ } \put(-205,-91){\fontsize{12}{12}$b(60,y , t)$ } \put(-205,75){\fontsize{12}{12}$b(30,y , t)$ } \put(-205,225){\fontsize{12}{12}$b(0,y , t)$ } \put(-410,225){\fontsize{12}{12}$a(0,y , t)$ } \put(-410,75){\fontsize{12}{12}$a(30,y , t)$ } \put(-410,-91){\fontsize{12}{12}$a(60,y , t)$ } & \end{array}$ ] @xmath124{figura6_a } \ , & \includegraphics[width=0.45\textwidth]{figura6_b } \medskip\\ \includegraphics[width=0.45\textwidth]{figura6_c } \ , & \includegraphics[width=0.45\textwidth]{figura6_d}\\ \end{array } \put(-45,165){\fontsize{12}{12}$x$ } \put(-280,165){\fontsize{12}{12}$x$ } \put(-45,-30){\fontsize{12}{12}$x$ } \put(-280,-30){\fontsize{12}{12}$x$ } \put(-300,-160){\fontsize{12}{12}$y$ } \put(-90,-160){\fontsize{12}{12}$y$ } \put(-300,30){\fontsize{12}{12}$y$ } \put(-90,30){\fontsize{12}{12}$y$ } \put(-205,-15){\fontsize{12}{12}$b(x , y,5)$ } \put(-205,175){\fontsize{12}{12}$b(x , y,0)$ } \put(-420,175){\fontsize{12}{12}$a(x , y,0)$ } \put(-420,-15){\fontsize{12}{12}$a(x , y,5)$ } \end{array}$ ] @xmath116{figura7.eps } } \put(-220,-10){\fontsize{12}{12}$y $ } \put(-300,100){\fontsize{12}{12 } $ \lambda_{2}(\alpha_{2})$ } \put(-300,30){\fontsize{12}{12 } $ \lambda_{3}(\alpha_{3})$ } \put(-240,140 ) { \fontsize{12}{12}$\lambda_{1}(\alpha_{1})$ } \end{array } $ ] @xmath125{figura8_a } \ , & \includegraphics[width=0.42\textwidth]{figura8_b } \medskip\\ \includegraphics[width=0.42\textwidth]{figura8_c}\ , & \includegraphics[width=0.42\textwidth]{figura8_d}\medskip\\ \includegraphics[width=0.42\textwidth]{figura8_e}\ , & \includegraphics[width=0.42\textwidth]{figura8_f } \end{array } \put(-175,140){\fontsize{12}{12}$t$ } \put(-380,140){\fontsize{12}{12}$t$ } \put(-175,-30){\fontsize{12}{12}$t$ } \put(-380,-30){\fontsize{12}{12}$t$ } \put(-175,-200){\fontsize{12}{12}$t$ } \put(-380,-200){\fontsize{12}{12}$t$ } \put(-280,-60){\fontsize{12}{12}$y$ } \put(-60,-60){\fontsize{12}{12}$y$ } \put(-280,110){\fontsize{12}{12}$y$ } \put(-60,110){\fontsize{12}{12}$y$ } \put(-280,-235){\fontsize{12}{12}$y$ } \put(-60,-235){\fontsize{12}{12}$y$ } \put(-205,-100){\fontsize{12}{12}$b(80,y , t)$ } \put(-205,70){\fontsize{12}{12}$b(20,y , t)$ } \put(-205,245){\fontsize{12}{12}$b(0,y , t)$ } \put(-400,245){\fontsize{12}{12}$a(0,y , t)$ } \put(-400,70){\fontsize{12}{12}$a(20,y , t)$ } \put(-400,-100){\fontsize{12}{12}$a(80,y , t)$ } & \end{array}$ ] @xmath124{figura9_a } \ , & \includegraphics[width=0.45\textwidth]{figura9_b } \medskip\\ \includegraphics[width=0.45\textwidth]{figura9_c } \ , & \includegraphics[width=0.45\textwidth]{figura9_d}\\ \end{array } \put(-45,165){\fontsize{12}{12}$x$ } \put(-280,165){\fontsize{12}{12}$x$ } \put(-45,-30){\fontsize{12}{12}$x$ } \put(-270,-42){\fontsize{12}{12}$x$ } \put(-300,-175){\fontsize{12}{12}$y$ } \put(-90,-170){\fontsize{12}{12}$y$ } \put(-300,30){\fontsize{12}{12}$y$ } \put(-90,30){\fontsize{12}{12}$y$ } \put(-205,-19){\fontsize{12}{12}$b(x , y,5)$ } \put(-205,180){\fontsize{12}{12}$b(x , y,0)$ } \put(-420,175){\fontsize{12}{12}$a(x , y,0)$ } \put(-420,-25){\fontsize{12}{12}$a(x , y,5)$ } \end{array}$ ] @xmath124{figura10_a } \ , & \includegraphics[width=0.45\textwidth]{figura10_b } \\ \end{array } \put(-320,-70){\fontsize{12}{12}$y $ } \put(-100,-70){\fontsize{12}{12}$y $ } \put(-180,-10){\fontsize{12}{12 } $ \theta_{2}(\alpha_{2})$ } \put(-180,50){\fontsize{12}{12 } $ \theta_{3}(\alpha_{3})$ } \put(-390,36 ) { \fontsize{12}{12}$\theta_{1}(\alpha_{1})$ } \end{array } $ ]
in this section , in order to validate the analytical results previously obtained as well as to get a deeper insight into the interaction / separation processes described hitherto , we integrate numerically the set of equations ( [ hydro ] ) . to this aim the @xmath74 coordinate is considered as the `` evolution '' variable so that the computation is performed for @xmath127 $ ] , over the domain @xmath128 \times \left[0,t_{end}\right ] $ ] . according to ( [ a - b ] )
the `` initial '' data for @xmath129 and @xmath130 are obtained from @xmath131 where @xmath132 and @xmath133 $ ] , @xmath134 are constants .
as far as the `` boundary '' data @xmath135 and @xmath136 are concerned , since the relations ( [ sol ] ) evaluated at @xmath137 represent the general solution of the homogeneous system of hydrodynamic type ( [ chrom2 ] ) in @xmath138 variables @xcite , then in the numerical procedure at hand we assume the following `` boundary '' conditions @xmath139 where @xmath140 are given by numerical integration of ( [ chrom2 ] ) for @xmath127 $ ] over the spatial domain @xmath141 $ ] with initial conditions @xmath142 . in figures [ figura5 ] and [ figura6 ] , by considering the set of parameteres @xmath143
we show spatial and temporal evolution of an initial pulse characterized by ( [ boundarydata1 ] ) , ( [ boundarydata ] ) and ( [ bd ] ) .
it should be noticed that the behaviours of the field variables @xmath144 and @xmath145 support the qualitative analysis depicted in figures [ dinteract1 ] and [ dinteract3 ] .
in fact the choice @xmath146 simulates a pulse localized at @xmath147 in a stripe of the @xmath148plane ( as illustrated in figures [ figura5 ] and [ figura6 ] ) which after a finite distance @xmath74 ( interaction region ) separates into three simple waves travelling along different characteristic curves @xmath60 and become separate by regions of constant states .
moreover in figure [ figura7 ] we plot the control parameters @xmath149 associated to ( [ boundarydata ] ) and ( [ parameters1 ] ) for the simple waves deformation in the interaction region . in this case
the boundary data provide @xmath150 and @xmath151 therefore , according to ( [ interpro ] ) , the profiles of the characteristic curves @xmath152 and @xmath153 originating at @xmath154 and @xmath155 respectively , do not distort being @xmath156 .
moreover , as predicted by ( [ interpro ] ) , all the @xmath149 become constant for @xmath157 $ ] .
next we choose the following set of parameters @xmath158 in order to simulate a simple wave and a pulse localized at @xmath147 in different stripes of the @xmath148plane ( as illustrated in figures [ figura8 ] and [ figura9 ] ) . in this case , by integrating the system of pdes ( [ hydro ] ) , we notice that the behaviours of the field variables @xmath144 and @xmath145 confirm the qualitative analysis depicted in figures [ dinteract2 ] and [ dinteract4 ] .
in particular figures [ figura8 ] and [ figura9 ] show that the simple wave , travelling along the fastest characteristic curve @xmath152 , interacts with the pulses travelling along @xmath159 and @xmath160 , affects their profiles and emerges again as simple wave .
moreover , after a finite distance @xmath74 , the initial pulse separates into three simple waves .
finally in figure [ figura10 ] we plot the control parameters @xmath161 associated to ( [ boundarydata ] ) and ( [ parameters2 ] ) for the simple waves deformation . in this case
the boundary data provide @xmath162 , @xmath163 , @xmath164 and @xmath165 therefore , according to ( [ interpro_iii ] ) , ( [ interpro_iv ] ) , @xmath166 and @xmath167 become constant for increasing values of @xmath168 and @xmath169 whereas @xmath170 is constant if @xmath171 $ ] .
the existence of riemann invariants plays a fundamental role in studying two dimensional nonlinear wave interactions as well as for some physically interesting examples of multi component hydrodynamic systems as in chromatography or electrophoresis models which can be also written in a diagonal form .
however in multi dimensional case nonlinear wave interactions are not so deeply understood because in such a case the governing systems do not possess , in general , riemann invariants .
nevertheless , in the theory of integrable multi dimensional quasilinear systems of first order , every hydrodynamic reduction admittes riemann invariants .
since any integrable multi dimensional quasilinear system of first order possesses infinitely many multi component hydrodynamic reductions , one can select the appropriate hydrodynamic reduction according to given boundary conditions or cauchy initial problems . by this reason in this paper we considered a three dimensional two component integrable quasilinear system of first order which has infinitely many hydrodynamic reductions . for simplicity
we restricted our attention to multi - component reductions which already are known as the chromatography system , because in this particular case we already know how to construct a general solution .
for such a model we gave an exact analytical description of three dimensional nonlinear wave interactions .
such a theoretical problem was usually considered for two - dimensional hyperbolic models @xcite , while , in the case of more space variables , for special evolution processes ruled by an auxiliary @xmath172 hyperbolic subsystem @xcite .
the crucial point of the procedure at hand is the existence of @xmath12 riemann invariants for each particular hydrodynamic reduction ( see details in @xcite ) of three dimensional quasilinear system ( [ hydro ] ) .
therefore the method can be applied to the class of semi - hamiltonian homogeneous systems of hydrodynamic type which admit a diagonalized form @xcite . in this case
a general solution can be obtained by means of the generalized hodograph method @xcite .
although the study developed herein can be performed for any diagonalizable semi - hamiltonian homogeneous model which can be solved by the generalized hodograph method , the procedure was illustrated for two commuting systems ( [ chrom1 ] ) and ( [ chrom2 ] ) associated to the three dimensional system ( [ hydro ] ) .
two different situations were illustrated .
first we described the evolution of @xmath12 waves initially localized in a closed interval , next we studied the interaction between @xmath91 pulses and a further single wave . in both cases
the resulting waves behave like simple waves and after the interaction region they are distorted . the amount of the produced distortion is analytically computed through the @xmath173 and @xmath174 terms which depend on the initial data given , respectively , by ( [ rho_k ] ) and ( [ data_3 ] ) . as it was pointed out in section [ inivalue ] ,
although in section [ nwave ] we studied nonlinear wave interactions concerning the examples characterized by the initial / boundary conditions ( [ rho_k ] ) and ( [ data_3 ] ) , we remark that our procedure can be in principle applied to describe hyperbolic nonlinear wave interactions corresponding to any set of initial / boundary data . finally ,
in order to validate the analytical results herein obtained , the numerical integration of the governing system ( [ hydro ] ) was performed .
the resulting three dimensional figures confirm the behaviours which have been analytically described by means of the proposed procedure .
33 grundland a m 1974 , riemann invariants for nonhomogeneous systems of first - order partial quasilinear differential equations algebraic aspects .
examples from gasdynamics . _ arch .
. stos . ) _ * 26 * 271295 .
grundland a m and zelazny r 1983 , simple waves in quasilinear hyperbolic systems .
i. theory of simple waves and simple states .
examples of applications _ j. math .
* 24 ( 9 ) * 23052314 , doi.org/10.1063/1.525979 .
grundland a m and zelazny r 1983 , simple waves in quasilinear hyperbolic systems .
riemann invariants for the problem of simple wave interactions _ j. math .
phys . _ * 24 ( 9 ) * 23152329 , doi : 10.1063/1.525980 .
curr c , fusco d and manganaro n 2012 hodograph transformation and differential constraints for wave solutions to @xmath1 quasilinear hyperbolic nonhomogeneous systems _ j. phys . a : math .
theor . _ * 45 ( 19 ) * 195207 doi:10.1088/1751 - 8113/45/19/195207 .
curr c , fusco d and manganaro n 2011 a reduction procedure for generalized riemann problems with application to nonlinear transmission lines
_ j. phys . a : math .
theor . _ * 44 ( 33 ) * 335205 doi:10.1088/1751 - 8113/44/33/335205 .
curr c and fusco d 2013 nonlinear wave interactions for quasilinear hyperbolic @xmath1 _ aapp , atti della accademia peloritana dei pericolanti classe di scienze fisiche , matematiche e naturali _ * 91 ( 1 ) * a6 , doi : 10.1478/aapp.91s1a6 . | three dimensional nonlinear wave interactions have been analytically described .
the procedure under interest can be applied to three dimensional quasilinear systems of first order , whose hydrodynamic reductions are homogeneous semi - hamiltonian hydrodynamic type systems ( i.e. possess diagonal form and infinitely many conservation laws ) .
the interaction of n waves was studied . in particular
we prove that they behave like simple waves and they distort after the collision region .
the amount of the distortion can be analytically computed .
october 2016 _ keywords : _ nonlinear wave interactions , hydrodynamic integrable systems , generalized hodograph method |
magnetic reconnection is a fundamental process of breaking and topological rearrangement of magnetic field lines in magnetized plasmas .
reconnection converts magnetic energy into kinetic and thermal energy and is believed to be responsible for many phenomena observed in the laboratory and cosmic plasmas @xcite . because electrical resistivity is very low in hot plasmas , magnetic reconnection due to resistive dissipation of magnetic field is typically a very slow process @xcite .
reconnection can become much faster in the case when resistivity is anomalously high due to local plasma instabilities @xcite .
another possibility is fast reconnection made possible because of two - fluid plasma effects @xcite , which require a two - fluid magnetohydrodynamics ( mhd ) description of plasma . in the limit of zero electron - to - ion mass ratio ,
two - fluid mhd equations simplify and reduce to hall - mhd equations .
the later include the hall and electron pressure terms in the ohm s law , in addition to the resistivity term present in single - fluid mhd . because of its relative simplicity ,
hall - mhd description of plasma has been extensively used in numerical simulations of magnetic reconnection @xcite . however , to the best of our knowledge , a simple self - consistent analytical model of hall reconnection , similar to the classical sweet - parker model of resistive reconnection , has not been constructed . in this paper
we consider hall - mhd equations and present a theoretical model of hall reconnection .
full two - fluid calculations for magnetic reconnection will be considered elsewhere .
the analytical derivations of this paper are similar to the derivations done by @xcite for the case of resistive single - fluid mhd reconnection .
for simplicity and brevity , we use physical units in which the speed of light @xmath3 and four times @xmath4 are replaced by unity , @xmath5 and @xmath6 . to rewrite our equations in the cgs units
, one needs to make the following substitutions : magnetic field @xmath7 , electric field @xmath8 , electric current @xmath9 , electrical resistivity @xmath10 , and the proton electric charge @xmath11 .
we assume the plasma is non - relativistic , with both phase and physical velocities much smaller than the speed of light .
we neglect electron inertia for the hall - mhd description of plasma .
the generalized ohm s law is @xcite @xmath12 where @xmath13 is the ion mass , @xmath14 is the plasma density , @xmath15 is the plasma velocity ( equal to the ion velocity ) , @xmath16 is the tensor of the electron pressure .
the first two terms on the right - hand side of eq .
( [ ohms_law ] ) are the single - fluid mhd terms , the third and fourth terms are the hall and electron pressure terms .
the equation of plasma motion is @xcite @xmath17 where @xmath18 is the tensor of the total pressure ( equal to the sum of the electron and ion pressure tensors ) , and we neglect plasma viscosity .
equation ( [ motion_law ] ) appears exactly the same as in the case of single - fluid mhd .
note that @xmath19 , and , for non - relativistic plasma , @xmath20 .
we consider hall magnetic reconnection in the classical two - dimensional sweet - parker - petschek reconnection layer , shown in fig .
[ figure_layer ] .
the layer is in the x - y plane with the x- and y - axes being perpendicular to and along the layer respectively .
all @xmath21 derivatives are zero .
the thickness of the reconnection layer is @xmath22 , which is defined as the thickness of the out - of - plane current ( @xmath23 ) profile across the layer .
note that @xmath22 is approximately equal to the electron layer thickness , while the ion layer thickness @xmath24 can be much larger .
velocity @xmath25 is the plasma inflow velocity in the upstream region at point @xmath26 , outside the electron layer .
the magnetic field @xmath27 at point @xmath26 is in the y - direction .
the out - of - plane field @xmath28 is assumed to have a quadrupole structure ( see fig .
[ figure_layer ] ) , in agreement with numerical simulations and laboratory experiments of two - fluid reconnection @xcite ( a nearly uniform `` guide '' field component of @xmath28 is taken to be zero ) .
the reconnection layer is assumed to have a point symmetry with respect to its geometric center point @xmath29 in fig .
[ figure_layer ] and reflection symmetries with respect to the axes @xmath30 and @xmath31 .
thus , the x- , y- and z - components of @xmath15 , @xmath32 and @xmath33 have the following simple symmetries : @xmath34 , @xmath35 , @xmath36 , @xmath37 , @xmath38 , @xmath39 , @xmath40 , @xmath41 and @xmath42 . in the derivations presented below
we will extensively use these symmetries .
now let us list four assumptions that we make in this study .
first , we assume that the plasma flow is incompressible inside the reconnection layer , @xmath43 .
second , we assume that the electrical resistivity @xmath44 is constant and very small , so that the lundquist number is very large .
third , we assume that the reconnection process is slow and quasi - stationary , so that we can neglect all time - derivatives @xmath45 in eqs .
( [ ohms_law])([motion_law ] ) and below .
this assumption is satisfied if the reconnection rate is slow , @xmath46 , and there are no plasma instabilities in the reconnection layer .
fourth , we assume that the electron and ion pressure tensors are isotropic , so that the pressure tensors in eqs .
( [ ohms_law ] ) and ( [ motion_law ] ) are scalars . using ampere s law and neglecting the displacement current , we find the x- and y - components of the current to be @xmath47 and @xmath48 . here and
below we use convenient notations @xmath49 , @xmath50 , @xmath51 , @xmath52 , and etc .
the z - component of the current at the reconnection layer central point @xmath29 is @xmath53 where we use estimates @xmath54 and @xmath55 at point @xmath29 .
next , neglecting the time derivative in eq .
( [ motion_law ] ) , the equation for acceleration of plasma along the reconnection layer , in the y - direction , is @xmath56 .
we calculate the first order partial derivative @xmath57 of this equation at point o and obtain @xmath58 here , we used the fact that the pressure term is @xmath59 . thus , the drop of pressure @xmath18 along the layer is equal to the magnetic pressure drop of the parallel field component outside the layer .
this result follows from the force balance condition for the plasma across the reconnection layer ( in analogy with the sweet - parker derivations for slowly inflowing plasma ) , and its rigorous proof can be found in @xcite .
the last expression in eq .
( [ acceleration ] ) is obtained by defining the magnetic field external scale as @xmath60 this is the scale of magnetic field just outside the reconnection electron layer ( at point m ) and can be interpreted as the length of the layer .
the @xmath61 derivative on the left - hand - side of eq .
( [ acceleration ] ) can be estimated from plasma incompressibility condition at the o - point , @xmath62 where we use an estimate @xmath63 .
next , the faraday s law @xmath64 for the x- and y - components of a quasi - stationary magnetic field in two dimensions is @xmath65 and @xmath66 .
therefore , the electric field z - component @xmath67 is constant in space , and from generalized ohm s law ( [ ohms_law ] ) we obtain @xmath68 now , we use eq .
( [ ohms_law_z ] ) to calculate @xmath67 at points o and m ( see fig .
[ figure_layer ] ) . at point
o we have @xmath69 . at point
m we have @xmath70 , where we use an estimate @xmath71 .
equating these two expressions for @xmath67 at points o and m and neglecting the resistive term outside the reconnection layer at point m , @xmath72 , we obtain @xmath73 next , we calculate the second order partial derivative @xmath74 of eq .
( [ ohms_law_z ] ) at point
o. we have @xmath75 here , to obtain the final expression , we use the fact that the y - scale of the current @xmath23 , to a factor of order unity , is about the same as the y - scale of the outside magnetic field , @xmath76 .
this result can be understood by taking the @xmath74 partial derivative of the ampere s law equation @xmath77 , see eq .
( [ amperes_law ] ) , while keeping @xmath78 constant because the partial derivative in @xmath31 is to be taken at constant @xmath79 ( for details see @xcite ) .
next , we use the z - component of the faraday s law .
we have @xmath80 .
we calculate the @xmath81 partial derivative of this equation at point o and use ohm s law ( [ ohms_law ] ) for @xmath82 and @xmath83 . after tedious but straightforward derivations
, we obtain @xmath84 \nonumber\\ & & -\left[(\partial_y b_x)_o(\partial_{xx}^2v_z)_o+ ( \partial_x b_y)_o(\partial_{yy}^2v_z)_o\right ] \nonumber\\ & & + ( m_p/\rho e)\left[(\partial_y b_x)_o(\partial_{xx}^2j_z)_o+ ( \partial_x b_y)_o(\partial_{yy}^2j_z)_o\right ] \nonumber\\ & \approx & 2\eta(\partial_{xy}^2 b_z)_o/\delta^2 \nonumber\\ & & -2(m_p/\rho e)\left[(j_o/\delta^2)(\partial_y b_x)_o+ j_o^2/l^2\right ] .
\label{e_xy}\end{aligned}\ ] ] here , to derive the final expression , we use @xmath85 , @xmath86 , @xmath87 , @xmath88 and @xmath89 .
we also use formula @xmath90 . to prove it ,
let us consider the z - component of plasma motion equation ( [ motion_law ] ) , which is @xmath91 .
we calculate the @xmath92 and @xmath74 derivatives of this equation at point o and obtain @xmath93 and @xmath94 .
thus , we have @xmath95 , @xmath96 and @xmath90 . now , we have six equations ( [ amperes_law ] ) , ( [ acceleration ] ) , ( [ v_derivative ] ) , ( [ e_z_perpendicular ] ) , ( [ e_z_parallel ] ) and ( [ e_xy ] ) , and we have six unknowns : @xmath97 , @xmath78 , @xmath25 , @xmath61 , @xmath98 and @xmath99 .
thus , we can solve for all unknown quantities .
for convenience of presentation , we express the solution in terms of the alfven velocity @xmath100 , the ion inertial length @xmath101 and the lundquist number @xmath102 .
the solution is @xmath103 the last equation gives the reconnection rate @xmath67 . in the limit @xmath104 and @xmath105 the reconnection rate is slow , @xmath46 , and our assumption of a quasi - stationary reconnection process is self - consistent .
equation ( [ vy_y ] ) implies that the ions are accelerated up to approximately alfven velocity @xmath106 along the reconnection layer of length @xmath2 . at the same time , the rate of electron acceleration along the layer at point o is @xmath107 , where @xmath108 denotes the electron velocity .
the ion layer thickness @xmath24 can be estimated as follows . in the upstream region outside the ion layer at @xmath109 ideal single - fluid
mhd applies .
therefore , at @xmath109 and @xmath110 the resistive and hall terms in eq .
( [ ohms_law_z ] ) can be neglected and @xmath111 , where @xmath112 is the plasma inflow velocity outside the ion layer .
velocity @xmath113 is called the reconnection velocity .
it can also be estimated as @xmath114 .
thus , @xmath115 we discuss the solution ( [ j_o])-([delta_ion ] ) in the next section .
now , let us make an important remark .
our analytical derivations involve an approximate solution of the hall - mhd equations in the infinitesimal neighborhood of line om across the reconnection electron layer ( see fig .
[ figure_layer ] ) .
all physical quantities , @xmath97 , @xmath78 , @xmath25 , @xmath61 , @xmath98 , @xmath116 , @xmath27 and @xmath2 , are defined either at point o ( the layer center ) or at point m ( the upstream region ) . in other words , all our derivations involve only the upstream region and the layer center , and we do not need to consider the downstream region for estimation of the reconnection rate . this `` local '' equations approach was first developed in @xcite for single - fluid mhd reconnection with anomalous electrical resistivity , and this approach works for hall - mhd reconnection as well .
note however , that we define the field scale @xmath2 by eq .
( [ l ] ) , and its exact value , as well as the value of field @xmath27 in the upstream region , depend on the `` global '' solution of the hall - mhd equations outside the reconnection layer . both @xmath2 and @xmath27 enter our model as parameters .
determination of their values requires numerical simulations of the global field configuration and is not considered here .
when @xmath117 ( @xmath118 is the sweet - parker layer thickness ) , the solution ( [ j_o])-([delta_ion ] ) reduces to the sweet - parker solution : @xmath119 , @xmath120 , @xmath121 , @xmath122 , @xmath123 , @xmath124 and @xmath125 . in the opposite limit ,
when @xmath126 and reconnection is in a collisionless hall regime , we have @xmath127 the electron acceleration rate along the reconnection layer at o - point is @xmath128 .
fast electron outflow along the layer creates the quadrupole field @xmath28 @xcite . from eq .
( [ ez_collisionless ] ) we find that the rate of collisionless hall reconnection , @xmath129 , is independent of the electrical resistivity @xmath44 @xcite .
the reconnection velocity is @xmath130 . from eq .
( [ delta_ion_collisionless ] ) we find @xmath131 for the ion layer thickness , which is in agreement with experiment @xcite . in it
noteworthy that in the absence of collisions , when @xmath132 , it is an anisotropic electron pressure that balances @xmath67 field at point o. in this case electrons are accelerated by @xmath67 field during time @xmath133 , while they are unmagnetized and are traveling with thermal speed @xmath134 inside the electron layer of length @xmath2 @xcite . as a result ,
the effective resistivity becomes @xmath135 , where @xmath136 is the electron inertial length ( @xmath137 is the electron mass ) .
thus , pressure anisotropy and electron inertia become important ( they will be considered elsewhere ) . in the end of this paper , let us briefly discuss possible mechanisms of fast reconnection , which is independent of the macroscopic size of the reconnecting system .
first , note that the collisionless hall reconnection rate @xmath129 is high when the reconnection electron layer length @xmath2 is microscopically small and comparable to @xmath0 .
although the value of @xmath2 can not be determined in our model , numerical simulations find that @xmath2 can indeed be much smaller than the macroscopic system size @xcite .
for example , @xmath2 can be as small as @xmath138 during quasi - stationary collisionless reconnection , resulting in a very fast reconnection rate @xmath139 @xcite .
another possible mechanism of fast reconnection can result from the dependence of the reconnection rate on the density @xmath14 and temperature @xmath140 of the plasma .
we have @xmath141 @xcite , @xmath142 , @xmath143 , @xmath144 .
therefore , @xmath145 , and the collisionless hall reconnection rate is @xmath146 .
thus , even if initially @xmath147 , as reconnection proceeds , the plasma is heated up by joule heating , the plasma temperature rises , the plasma density drops , and the reconnection becomes more and more collisionless and faster and faster .
this run - away reconnection process can possibly operate in solar corona and earth s magnetosphere .
the dependence @xmath146 also supports a model of self - regulation heating in solar corona @xcite .
i am especially grateful to russell kulsrud for many very helpful discussions and for pointing out to me the importance of electron pressure anisotropy .
i would also like to thank ellen zweibel , fausto cattaneo , masaaki yamada , boon chye ( bc ) low , eugene parker , alex obabko , dmitri uzdensky , hantao ji , viacheslav titov , amitava bhattacharjee and the anonymous referee for their interest in this work , discussions and for useful comments .
this work was supported by the nsf center for magnetic self - organization ( cmso ) in laboratory and astrophysical plasmas at the university of chicago .
+ _ a note added after publication : _ after the publication of his paper , the author has became aware that simakov and chacon ( 2008 ) have independently obtained a hall reconnection rate formula , which is the same as eq .
( 17 ) up to numerical factors of order unity @xcite .
although the final result is basically the same , the two derivations are significantly different .
the author has used a rigorous local analysis , in which the hall - mhd equations are solved only in the upstream region and in the center of the reconnection layer . by contrast
, simakov and chacon ( 2008 ) have used a different approach that also considers the downstream region . in particular
, they have assumed that the thickness of the ion layer is given by @xmath148 , and obtained their reconnection rate formula as an approximate interpolation between the rates in the collisional and collisionless regimes .
there are also minor differences in the problem formulation ; that of simakov and chacon ( 2008 ) includes electron viscosity , which is neglected by the author , and neglects electron pressure , which is included by the author .
expression @xmath149 for the reconnection velocity in the collisionless regime was obtained earlier by stanley w. h. cowley ( 1985 ) @xcite . | the rate of quasi - stationary , two - dimensional magnetic reconnection is calculated in the framework of incompressible hall magnetohydrodynamics ( mhd ) , which includes the hall and electron pressure terms in the ohm s law .
the hall - mhd equations are solved in a local region across the reconnection electron layer , including only the upstream region and the layer center . in the case
when the ion inertial length @xmath0 is larger than the sweet - parker reconnection layer thickness , the dimensionless reconnection rate is found to be independent of the electrical resistivity and equal to @xmath1 , where @xmath2 is the scale length of the external magnetic field in the upstream region outside the electron layer , and the ion layer thickness is found to be @xmath0 . |
what follows is essentially a verbatim transcript of the introductory talk as given at adass , in quebec , on the morning of the first session which happened to be the day before the 2008 u.s
. presidential elections .
good morning everybody .
bonjour , tout le monde . everyone awake ?
no jet lag ?
it s very nearly tomorrow morning in australia , which is confusing for me , so if i drift off topic , just talk amongst yourselves . because that s what this is all about .
talking amongst ourselves .
i want to try to encourage some discussion about whether we can communicate better amongst ourselves .
this seems particularly worthwhile in the context of an adass meeting , because this is one of the few occasions we do all get to meet together . in many ways ,
_ you _ do nt need to hear this talk but do nt try to leave , there are webcams on all the doors and _ we _ know your e - mail addresses .
you do nt need it because you _ are _ the communicators in our field .
you ve come to adass .
you ve probably struggled with your organisation s bureaucracy and expenses software .
( have you noticed that lots of places seem to have accounting software that ca nt possibly have been used anywhere else first except possibly enron and lehman brothers ? ) you ve booked the tickets , carried on your laptops how many people wrote their talks on the way over ?
oh , all still working on them , are you
you ve explained to your colleagues that it s not just a jolly .
it s cold in quebec in winter .
you re prepared to miss what s going on back home . tomorrow
s a big day in some places .
november 4th . in australia , they re running the melbourne cup , the biggest race of the year . and
this year , i m not going to lose money on it for the first time , because i m here . and tomorrow , something important is going on south of the border , right ?
adass is just about the main communications channel in astronomical software .
the people who run it deserve some applause . but adass on its own is nt enough .
we should mention the spie meetings .
spie is covering software now , especially at what i think of as the sharp end telescopes and instrument control . and
this year , spie had a very interesting session on software sharing .
i ll come back to that later .
outside the conferences , and outside individual projects , things are much more limited . and conferences do nt cover everything they naturally concentrate on the cutting edge work .
they re expensive to get to , and what people learn at them can be quite hard to pass on to others .
they probably are nt going to get cheaper ` peak air travel ' and ` peak oil ' are probably going to happen at about the same time .
now , i m here because a few months ago i scattered around an e - mail bemoaning the state of communications in astronomical software . and
this happened because i suddenly realised how little i knew about what was going on .
i was asked to prepare an estimate for a proposed antarctic telescope .
produce a software structure design for the whole system telescope , instruments , data reduction , communications , everything .
costs accurate to ten percent would be nice , and you can do it in about two weeks , ca nt you ? suddenly , i needed to know how you would structure the software for a new telescope .
i learned something a long time ago as a physics student doing lab experiments . if you re having trouble with something , find out who did it the week before and ask them .
back then , this was called ` cheating '
. now we call it ` consulting the knowledge base ' .
( i ve always assumed this was the real driver behind projects like seti : you want to know about the rest of the universe ? find someone who lives there and ask them . `
excuse me , you know of any earth - like planets in your neighbourhood ? ' .
much easier than trying to find them yourself , and you do nt need all those expensive telescopes . )
so , i wanted to consult the knowledge base for our field .
and i was nt sure where it was .
incidentally , at the other end of the food chain , the astronomers do this quite well .
( i m not sure just which end they are , but it s the other end . )
software people do nt use libraries much , because it always seems they only have the out of date books : ` 8086 architecture ' , ` vms programming ' , ` field guide to dinosaurs ' , etc .. but astronomers are in there reading the annual review of astronomy and astrophysics , where pundits publish review papers quite a big job , but they get cited a lot . on the web , pre - prints appear in astro - ph .
conferences have review talks .
there is nt an annual review of astronomical software , so how do you find out what people are using ?
what middleware ?
acs , corba , .net , drama , taros ?
can you really run everything from labview ?
how do you string together a data reduction pipeline ?
i found myself wanting some central web site with all this arranged for me . with a search box where i could type ` observatory middleware ' .
some if this stuff is available , but not all , and google is only a rather awkward window into it .
information needs an index and a table of contents , and producing really good ones is still a job for people . in the past , there have been web sites that have tried to provide what i , at least , ca nt find now an overview of our whole field .
some of you have been directly involved in them .
asds , the astronomical software and documentation service , started with the idea of helping you find astronomical software packages , and expanded to cover telescope and instrument manuals .
but i m afraid that after years of useful service , asds is now just a set of links to a machine that no longer responds
. one of these links is in astroweb , another site that seems to be on the backburner now .
obviously , there are reasons why sites are allowed to fade away .
things have changed in our world over the years .
some changes are good for communication the way the web itself has changed is one of these .
collaborative tools have emerged : twikis , blogs , even youtube and secondlife i ll get back to these .
but there also has to be a will to use them , and things have changed there too .
a lot of us are more ` managed ' these days .
time spent has to be justified and accounted for .
and you are nt supposed to spend your time giving away what your organisation perceives as its competitive advantage . to me ` ip ' stands for ` internet protocol ' and is a good thing .
unfortunately , it also means ` intellectual property ' , and that may or may not be a good thing .
ok , i can see why this matters .
organisations have to pay the rent : ` spare a dollar ?
telescope and ageing instrument suite to support ' , ` every cent donated goes to our latest design study ' . and
security is an issue .
i actually found it hard to get all the e - mail addresses i wanted , even for people i knew , because many places do nt publicise e - mail addresses any more . because if your address gets out , you start to get such interesting e - mail .
someone in singapore regularly sends me a list of prices for second - hand road - building equipment . `
caterpillar excavator , good condition , engine recently rebuilt . '
how they found out i need one is beyond me .
none of this helps .
it is hard to find the time to maintain web pages for the public good , even if you re part of that public , and to justify it to management .
but , at the same time , the web is making things much easier .
and i suspect most of you are already using some of these new tools at least within specific projects .
you can see why public interest , general purpose , web pages fade .
if all changes have to go through one or two central people , making these changes will become dull , burdensome , unprofitable .
but this is the new millennium .
20th century fox may have decided to stay in the last century ( and a lot of marketing dollars must have gone into that decision ) but the web has moved on .
the point about all the ` web 2.0 ' technologies is that their content is no longer centrally provided .
wikipedia , youtube , flickr , are just repositories for user - created content .
lots of you use twikis for your collaborative projects .
people hold video conferences over skype .
way back when international phone conferences were a very big deal , i once sat in on a meeting in the uk from my office in australia using e - mail .
i had a contact at the meeting , and when they wanted my opinion , a couple of e - mails went back and forth .
it actually worked pretty well , given that it was late afternoon in the uk and sometime after midnight in australia .
and then , around three - thirty in the morning , i realised i d not heard anything for quite some time , and nobody was replying to me any more .
i waited some time , then gave up and went home .
later , i discovered they d finished the meeting , packed up , and gone to the pub for a drink .
not only did nt they invite me , they did nt even bother to tell me
skype is _ much _ better . and
people really do hold virtual conferences in secondlife , and presumably their avatars all go off to a virtual pub afterwards and get virtually drunk . and
it s these extra meetings , of course , where you learn a lot at conferences , but which never gets into the proceedings . in various ways ,
individual projects are , individually , making good use of the web 2 tools .
some of them . one point made more than once in the replies to my grumbling e - mail was that the vo people did a better job of communicating than do the people at my end of things the instrument and telescope control people .
i suspect , although it might be a topic for discussion later , that vo is set up as a really big collaborative project , and that in such projects internal communication is expected , budgeted for , supported , has a microsoft project task code , and so on .
so even writing a project blog is a legitimate activity , as it should be .
but outside individual projects , if you want to see the whole range of activity across our field , what is there ? not a lot .
when i sent my e - mail around , i did nt know anything about the planned software sharing initiative at spie this year .
a paper on software sharing was presented , and an open discussion followed .
some of the people involved are here alan bridger , kim gillies , steve wampler , should all be in this room somewhere .
you can get a copy of the paper ` enabling technologies and constraints for software sharing in large astronomy projects ' ( chiozzi et al .
2008 ) , if you know what to google for .
it shows up on the eso web pages , and you can get it from spie itself if you re a member , or know someone who is .
it s an interesting paper .
it identifies a lot of what makes software sharing difficult .
i was interested to see that they quote the example of the portable telescope control system , ptcs .
they comment that it has been adopted by a number of telescopes , but ` we could find no information about recent developments ' . and , as i read that , we were commissioning the replacement control system for the aat , the australian 4 meter , completely based on ptcs .
so there certainly had been recent developments .
it s an example of poor communication , i guess .
i do nt think ptcs has a public web presence .
i agree it s hard to share software although , obviously , people do .
but i think it s much easier to share experiences .
even if you could find ptcs for download on the web , what you _ really _ want first would be to see comments from a number of people who ve done telescope control systems , explaining what they did and how well they thought it worked . as a result of this software sharing initiative ,
there is , of course , now a twiki . right now , it s brand new , fairly empty . getting content wo nt happen by magic .
i suspect it will need people to solicit content i heard a suggestion for a rotating ` editor ' position , for example .
here s a idea . if someone from each institution here just listed the software they used
instrument control environments , middleware , data reduction pipelines even their project management software that would be a lot of organised information that could go in a twiki .
i d like to ride a personal hobby horse here .
documentation does nt have to be beautiful to be useful .
some things geometric algorithms , say need complex diagrams , and microsoft word is nt very good at doing them . they take ages to do on a computer . or a few minutes on a white board or a piece of paper , then a digital camera or a scanner and you have a jpeg diagram .
video yourself drawing it on the white board , and you can explain the algorithm as you go .
our one aao expert on the vlt instrumentation software left us , and during his last week i sat him down with a white board and a terminal , in front of a video camera , and got him to explain the system internals to me .
i d ask him about what i did nt understand , he d explain the short cuts that usually do nt get documented .
i learned a lot i learned you should turn your mobile phone off before doing this stuff and we now have some permanent documentation . you can put your documentation on youtube if you like .
you can certainly put videos into a twiki .
one reply to my original e - mail put it bluntly : ` you have to answer the question : what s in it for me ? '
well , apart from the fame , the glory , the adulation , the hundreds of screaming fans oh , sorry , that s a different career path .
but you know the answer . it s why you really are the right audience for this talk , and i m glad you stayed for it .
you know what you get from communication , from sharing experiences .
it s why you ve come all this way with your laptops and no more than 100 ml of liquids in your hand baggage . | meetings such as adass demonstrate that there is an enthusiasm for communication within the astronomical software community .
however ,
the amount of information and experience that can flow around in the course of one , relatively short , meeting is really quite limited .
ideally , these meetings should be just a part of a much greater , continuous exchange of knowledge . in practice , with some notable
but often short - lived exceptions , we generally fall short of that ideal . keeping track of what is being used , where , and how successfully , can be a challenge .
a variety of new technologies such as those roughly classed as web 2.0 are now available , and getting information to flow ought to be getting simpler , but somehow it seems harder to find the time to keep that information current .
this paper looks at some of the ways we communicate , used to communicate , have failed to communicate , no longer communicate , and perhaps could communicate better .
it is presented in the hope of stimulating additional discussion and possibly even a little action aimed at improving the current situation . |
at the moment , there are many models of thermonuclear explosion of a star , that lead to the event we know as a type ia supernova ( sn ia ) .
some of them were discussed in the talk by j.niemeyer @xcite .
only a few parameters , such as kinetic energy and total @xmath0ni production , can be derived directly from the explosion modelling and compared with the observational values .
the subsequent evolution of the exploded star gives us much more possibilities to compare models and to decide which one fits observations better by reproducing more details in sn ia light curves and spectra .
we will focus here on the broad - band ubvi and bolometric light curve computations for sn ia models .
there are several effects in sne physics which lead to difficulties in the light curve modelling of any type of sne .
for instance , an account should be taken correctly for deposition of gamma photons produced in decays of radioactive isotopes , mostly @xmath0ni and @xmath0co .
after being emitted , gamma photons travel through the ejecta and can finish up in either thermalization or in non - coherent scattering processes . to find this one has to solve the transfer equation for gamma photons together with hydrodynamical equations .
full system of equations should involve also radiative transfer equations in the range from soft x - rays to infrared for the expanding medium .
there are millions of spectral lines that form sn spectra , and it is not a trivial problem to find a convenient way how to treat them even in the static case . the expansion makes the problem much more difficult to solve : hundreds or even thousands of lines give their input into emission and absorption at each frequency . on the first glance , modelling of sne ia seems easier than of other types of supernovae ,
since the hydro part is very simple : coasting stage starts very early , there are no shocks , and no additional heating from them .
on the other hand , much more difficulties arise in the radiation part .
sne ia becomes almost transparent in continuum at the age of a few weeks .
this means that nlte effects are stronger than for other types of supernovae .
radiation is decoupled from matter within the entire sn ia ejecta even before maximum light ( which occurs around the 20th day after explosion ) , see e.g. @xcite or fig . 2 in @xcite .
in this case one can not ascribe to radiation the gas temperature , or any other temperature , since sn ia spectrum differs strongly from a blackbody one .
one has to solve a system of time - dependent transfer equations in many energy groups instead , with an accurate prescription for treatment of a huge number of spectral lines , which are the main source of opacity in this type of sn @xcite .
we compute broad - band ubvi and bolometric light curves of sne ia with a multi - energy radiation hydro code stella .
time - dependent equations for the angular moments of intensity in fixed frequency bins are coupled to lagrangean hydro equations and solved implicitly .
thus , we have no need to ascribe any temperature to the radiation : the photon energy distribution may be quite arbitrary . while radiation is nonequilibrium in our approximation , ionization and atomic level populations are assumed to be in lte .
the effect of line opacity is treated as an expansion opacity according to eastman & pinto 1993 @xcite and to our new recipes @xcite .
we have compared the results and found that infrared bandpass ( in which the ejecta are the most transparent ) is more sensitive to the treatment of opacity than ubv ( see fig . 4 in @xcite for the comparison of sn ia light curves calculated with these two approaches ) ,
so one should be very careful on this point . to simulate nlte effects we used the approximation of the absorptive opacity in spectral lines .
nlte results @xcite and etla approach @xcite demonstrate that fully absorptive lines gives us an acceptable approximate description of nlte effects .
we treat gamma - ray opacity as a pure absorptive one , and solve the @xmath1-ray transfer equation in a one - group approximation following @xcite .
the heating by decays @xmath0ni @xmath2 @xmath0co@xmath2 @xmath0fe is taken into account . in the calculations of sn
ia light curves we use up to 200 frequency bins and up to @xmath3 zones as a lagrangean coordinate .
in our previous work @xcite we studied the behaviour of the broad - band optical and ultraviolet light curves for classical 1d models of sne ia : chandrasekhar - mass models w7 ( deflagration ) @xcite and dd4 ( delayed detonation ) @xcite , and sub - chandrasekhar - mass models la4 ( helium detonation ; actually , an averaged 2d model ) @xcite and wd065 ( detonation with low @xmath0ni production ) @xcite . here
we concentrate on a new 3d deflagration model by m.reinecke et al .
@xcite ( mr , hereafter ) . working in more dimensions they have to involve less free parameters than was needed in 1d , so they get their model almost from the `` first principles '' .
.parameters of sn ia models [ cols="<,<,<,<",options="header " , ] the main features of the 3d model are compared with the ones of classical 1d models in the table . from the first glance it seems that the light curves for the models which differ so much could not be similar . nevertheless , fig .
[ sorokina.figw7mr0 ] demonstrates that they are similar in many details .
the possible reasons for this can be understood if one has a look at the element distribution over the ejecta .
the compositions for w7 and mr models are shown in fig .
[ sorokina.figchem ] . at the moment of our light curve computation
the full nuclide yields for mr were not yet obtained .
therefore , the model consisted of the elements , which were chosen as representative examples for the energy release calculation , namely `` fe '' for iron group elements that were divided onto 80% of @xmath0ni and 20% of @xmath0fe , `` mg '' for intermediate mass elements , and unburned c and o in equal proportion . the instabilities that have developed in the 3d model were not supposed to be so huge in approximate 1d models of explosion .
this has led to the differences in the nickel distribution over the ejecta : it is mixed to the outermost layers in the 3d model .
these layers became much more opaque than in the 1d models , and , despite having less than a half of kinetic energy , the 3d model has a photospheric velocity comparable to that in the 1d models .
it is probably still a bit too low , so the light curve is wider , since the ejecta expand a bit slower , and photons are locked inside them for a bit longer time .
the broad - band light curves for mr model fit the observations of one of the typical sn ia , sn 1994d , in @xmath4 and @xmath5 bands surprisingly well , while classical 1d models , such as w7 and dd4 , show faster decline in the optics than it is observed .
unfortunately , the bolometric light curve for mr model ( fig . [ sorokina.figbol ] ) is somewhat too slow .
the ejecta must expand with higher speed to let photons to diffuse out faster . ;
dashed for a @xmath6 cone with maximum @xmath0ni abundance ; dotted for a @xmath6 cone with minimum @xmath0ni abundance.,scaledwidth=60.0% ] since our light curve code is 1d , we can not model 3d effects directly .
we tried to estimate them by averaging the model over different solid angles .
the main question we wished to answer was : how different can the sn ia look like for an observer from different sides . to start with , we just compared the models averaged over cones with opening angles of @xmath6 with the one averaged over the whole @xmath7 @xcite .
we tried to choose two limiting cases : the angles with minimal and maximal nickel abundances , i.e. one with nickel bubble extended to the surface , and another one between such bubbles , with nickel only in the center . since total @xmath0ni mass differs in these models , the light curves are also very different ( fig . [ sorokina.fig3ni ] ) .
they just simulate different models , but not a model observed from different sides . in reality , the total energetics one observes is defined by total @xmath0ni mass .
so we have to preserve it when averaging the model over different directions . in the next experiment
we have taken the same solid angle with low @xmath0ni abundance and enhanced the mass of @xmath0ni in every mass zone by the same ratio , so that the total @xmath0ni mass became equal to the one of the original 3d model .
the resulting light curve is compared with the fully averaged case in the fig .
[ sorokina.figmr1ni ] . since @xmath0ni mass fixes the brightness of sne ia @xcite , the light curves become equally bright , but the maximum is shifted in time due to a bit different energetics and mixing .
we believe that differences between 1d and 3d light curve calculations should not be much stronger than the difference between these two light curves , when other physical assumptions are fixed .
( red solid ) ; and for a @xmath6 cone with @xmath0ni abundance rescaled to preserve the total @xmath0ni mass ( black dashed).,scaledwidth=60.0% ] there is also another reason which allows us to believe that the mr model is better than the classical 1d models .
tycho snr is believed to be the remnant of sn ia .
we calculated in details the x - ray emission of tycho snr , allowing for time - dependent ionization and recombination , and compared the computed x - ray spectra and images in narrow filter bands with xmm - newton observations of the tycho .
our preliminary results @xcite show that all chandrasekhar - mass models produce similar x - ray spectra at the age of tycho , but they differ strongly in the predictions of how the remnant should look like in the lines of different ions due to very different distribution of elements in the ejecta for 1d and 3d models .
we found that w7 and dd4 models produce rather wide ring in fe lines , while it is narrow for mr model .
the image for the latter model is very similar to what is observed .
the new 3d sn ia model mr @xcite is very appealing .
yet it is not a final one : a detailed post - processing of nucleosynthesis changes the composition .
it has been done very recently @xcite , and it is not yet checked in the light curve calculation .
our light curve computations are also preliminary , since more work is needed on the expansion opacity .
hopefully , none of the required improvements will spoil the light curve of this model and its x - ray spectra on the snr stage , since the specific qualities of the model can be primarily explained by the enrichment of the outermost layers of sn ejecta by fe and ni .
the sn light curve modelling still has a lot of physics to be added , such as a 3d time - dependent radiative transfer , including as much as possible of nlte effects , which are especially essential for sne ia @xcite .
all this will improve our understanding of thermonuclear supernovae .
the authors are grateful to wolfgang hillebrandt , to the organizers of the conference and to the mpa staff for all their help , to stan woosley for his hospitality at ucsc where a major part of the work on expansion opacity was done , and to bruno leibundgut for providing us with the data @xcite in electronic form . | we present theoretical ubvi- and bolometric light curves of sne ia for several explosion models , computed with our multi - group radiation hydro code .
we employ our new corrected treatment for line opacity in the expanding medium .
the results are compared with observed light curves .
our goal is to find the most viable thermonuclear sn model that gives good fits not only to a typical sn ia light curve , but also to x - ray observations of young snia remnants .
it appears that classical 1d snia models , such as deflagration w7 and delayed detonation dd4 , fit the light curves not so good as a new 3d deflagration model by reinecke et al ( which is averaged over angles for our lc modelling ) .
this model seems good also in reproducing x - ray observations of tycho snr .
we believe that the main feature of this model which allows us to get correct radiation during the first month , as well as after a few hundred years , when an snr forms , is strong mixing that pushes material enriched in iron and nickel to the outermost layers of sn ejecta . |
recently , superdeformed ( sd ) rotational bands have been discovered in @xmath1ar , @xmath4ca and @xmath5ti @xcite .
one of the interesting new features of them is that they are built on excited @xmath6 states and observed up to high spin , in contrast to the sd bands in heavier mass regions where low - spin portions of them are unknown in almost all cases @xcite .
these excited @xmath6 states may be associated with multiparticle - multihole excitations from the spherical closed shells , so that we can hope to learn from such data detailed relationships between spherical shell model and sd configurations . for the mass @xmath730 - 50 region , although existence of a sd band in @xmath0s with the sd magic number @xmath8 has been expected for a long time @xcite , it has not yet been observed and remains as a great challenge @xcite . in this paper , as a continuation of the systematic theoretical search @xcite for sd bands in the mass @xmath730 - 50 region by means of the cranked skyrme - hartree - fock ( shf ) method@xcite , we would like to suggest that , in addition to @xmath0s , the neutron - rich nucleus @xmath1s and the nuclei , @xmath2s and @xmath3s , which are situated close to the neutron - drip line@xcite , are also good candidates for finding sd rotational bands in sulfur isotopes .
the appearance of the sd band in @xmath1s is suggested in connection with the sd shell structure at @xmath9 characterizing the observed sd band in @xmath4ca .
the drip - line nuclei , @xmath2s and @xmath3s , are expected to constitute a new `` sd doubly closed '' region associated with the sd magic numbers , @xmath10 for protons and @xmath11 for neutrons .
an interesting theoretical subject for the sd bands in nuclei near the neutron drip line is to understand deformation properties of neutron skins .
the calculated density distributions indeed exhibit superdeformed neutron skins .
the calculation has been carried out with the use of the three - dimensional ( 3d ) , cartesian coordinate - mesh representation without imposing any symmetry restriction@xcite . in parallel , we also carry out the standard calculations @xcite imposing reflection symmetries . by comparing the symmetry restricted and unrestricted calculations
, we can examine the stability of the sd solutions of the shf equations against reflection - asymmetric deformations . in this way
, we have found several cases where the sd minima obtained in the symmetry - restricted calculations are in fact unstable with respect to the reflection - asymmetric deformations . in general , the sd states are rather soft against reflection - asymmetric deformations , so that we need careful study about the stabilities of them against various kinds of deformation breaking the reflection symmetries . after a brief account of the cranked shf calculational procedure in section 2 , we present and discuss results of the calculation in section 3 , and give conclusions in section 4 .
we shall present deformation energy curves for sulfur isotopes from @xmath0s to @xmath3s , and focus our attention on properties of rotational bands built on the sd @xmath6 states , stabilities of the sd local minima against the reflection - asymmetric deformations , and density distributions of the sd states .
a preliminary version of this work was reported in @xcite .
since the cranked shf method in the 3d coordinate - mesh representation is well known@xcite , we here give only a minimum description about the computational procedure actually adopted .
for a recent comprehensive review on selfconsistent mean - field models for nuclear structure , see ref .
@xcite . the cranked shf equation for a system uniformly rotating about the @xmath12-axis
is given by @xmath13 where @xmath14 , @xmath15 and @xmath16 mean the hamiltonian with the skyrme interaction , the rotational frequency and the @xmath12-component of angular momentum , respectively , and the bracket denotes the expectation value with respect to a slater determinantal state .
we solve the cranked shf equation by means of the imaginary - time evolution technique@xcite in the 3d cartesian - mesh representation .
the algorithm of numerical calculation is the standard one @xcite , except that we allow for both reflection- and axial - symmetry breakings . in this case
, it is important to accurately fulfill the center - of - mass and principal - axis conditions .
this is done by means of the constrained hf procedure@xcite .
we solved these equations inside the sphere with radius @xmath17=10 fm and mesh size @xmath18=1 fm , starting with various initial configurations .
the accuracy for evaluating deformation energies with this mesh size was carefully checked by tajima@xcite and was found to be quite satisfactory .
when we make a detailed analysis of density distributions , however , we use a smaller mesh size of @xmath19 fm .
in addition to the symmetry - unrestricted cranked shf calculation , we also carry out symmetry - restricted calculations imposing reflection symmetries about the @xmath20- , @xmath21- and @xmath22-planes .
below we call these symmetry - unrestricted and -restricted cranked shf versions `` unrestricted '' and `` restricted '' ones , respectively
. solutions of the cranked shf equation give local minima in the deformation energy surface . in order to explore the deformation energy surface around these minima and draw deformation energy curves as functions of deformation parameters , we carry out the constrained hf procedure with relevant constraining operators@xcite .
for the skyrme interaction , we adopt the widely used three versions ; siii @xcite , skm@xmath23@xcite and sly4 @xcite .
figure 1 shows deformation energy curves for sulfur isotopes from @xmath0s to @xmath3s obtained with the use of the siii interaction .
solid lines with and without filled circles represent the results obtained by the unrestricted and restricted versions , respectively .
the result of calculation indicates that the sd minima ( with the quadrupole deformation parameter @xmath24 ) appear in the neutron - rich nucleus @xmath1s and the drip - line nuclei , @xmath2s and @xmath3s , in addition to the well - known case of @xmath0s . as seen in figs . 2 and 3 ,
similar results are obtained for the skm@xmath23 and sly4 interactions , except that the sd states in @xmath2s is unstable against the reflection - asymmetric deformation for the sly4 interaction ( see subsection 3.3 ) . as discussed in refs.@xcite
, the sd local minimum in @xmath0s corresponds to the doubly closed shell configuration with respect to the sd magic number @xmath25 and involves two protons and two neutrons in the down - sloping single - particle levels originating from the @xmath26 shell .
the sd local minimum in @xmath1s results from the coherent combination of the sd magic number , @xmath10 , and the neutron shell effects occurring at large deformation for @xmath9 .
the latter shell effect has been confirmed recently by the discovery of the sd rotation band in @xmath4ca @xcite .
the sd shell gap at @xmath9 is associated with the 4p-4h excitation from below the @xmath9 spherical closed shell to the @xmath26 shell . focusing our attention on the occupation numbers of such high-@xmath27 shells and distinguishing protons(@xmath28 ) and neutrons(@xmath29 ) , these sd configurations in @xmath0s and @xmath1s are denoted in figs .
1 - 3 as @xmath30 and @xmath31 , respectively .
the sd local minima in the drip - line nuclei , @xmath2s and @xmath3s , result from the coherent combination of the proton sd shell effect and the neutron shell effects occurring at superdeformation for @xmath3232 - 34 .
the neutron configurations in these sd states are similar to those in the known sd bands in @xmath33zn and @xmath34zn associated with the sd magic numbers @xmath3230 - 32@xcite .
we find that the sd shell effect is strong also for @xmath35 in the sulfur isotopes under consideration , while the sd local minimum in @xmath36s with @xmath37 is unstable against the reflection - asymmetric deformation ( see subsection 3.3 ) . in the drip - line nuclei @xmath2s and @xmath3s ,
the @xmath26 shell is fully occupied even in the spherical limit and the sd configurations involve neutron excitations from the @xmath38-shell to the @xmath39 shell .
as before , focusing our attention on the occupation numbers of the high-@xmath27 shells , let us use the notation @xmath40 for a configuration in which single - particle levels originating from the @xmath26 and @xmath39 shell are occupied by @xmath41 protons and @xmath42 neutrons , respectively
. with such notations , the sd local minima in @xmath2s and @xmath3s correspond to the @xmath43 , and @xmath44 configurations , respectively .
the appearance of the sd minimum in @xmath1s suggests that we can expect a sd band associated with the same neutron configuration to appear also in the @xmath9 isotone , @xmath45ar , situated between @xmath1s and @xmath4ca .
we examined this point and the result is shown in fig .
we find that the two local minima associated with the configurations @xmath30 and @xmath31 compete in energy and their relative energy differs for different versions of the skyrme interaction : as clearly seen in the deformation - energy curves obtained by the symmetry - restricted calculations , the former with smaller @xmath46 is slightly lower for skm@xmath23 and sly4 while the latter with larger @xmath46 is slightly lower for siii .
counting both protons and neutrons , these local minima respectively correspond to the 4p-6h and 6p-8h configurations with respect to the spherical doubly closed shell of @xmath4ca .
as we discussed in the previous papers@xcite , the two configurations can mix each other in the crossing region through the reflection - symmetry breaking components in the mean field . specifically , around the crossing point between the down - sloping @xmath47 $ ] level ( coming from the @xmath26 shell ) and the up - sloping @xmath48 $ ] level ( coming from the @xmath49 shell below the @xmath9 spherical magic number ) , the @xmath50-type non - axial octupole deformation is generated , and they mix each other through this component of the mean field ( see fig . 5 ) .
note that the matrix element of the operator @xmath50 between the two levels satisfies the selection rules , @xmath51 and @xmath52 , for the asymptotic quantum numbers @xmath53 and @xmath54 . as a result of this mixing ,
the deformation - energy curve becomes rather flat in the symmetry - unrestricted calculation .
recently , the sd band corresponding to the 4p-6h configuration was found in experiment@xcite .
the data suggest significant competition between different configurations , which requires further analysis of shape fluctuation dynamics by going beyond the static mean - field approximation .
let us focus our attention on the sd local minima shown in figs . 1 - 3 , and investigate properties of the rotational bands built on them .
excitation energies of these sd rotational bands are plotted in fig . 6 as functions of angular momentum .
these rotational bands are calculated by cranking each shf solution ( the sd local minima in figs . 1 - 3 ) and following the same configuration with increasing value of @xmath15 until the point where we can not clearly identify the continuation of the same configuration any more .
thus , the highest values of angular momentum in this figure does not necessarily indicate the band - termination points but merely suggest that drastic changes in their microscopic structure take place around there .
different slopes with respect to the angular momentum between @xmath1s and @xmath3s can be easily understood in terms of the well known scaling factor @xmath55 for the rigid - body moment of inertia .
this point can be confirmed in fig . 7 which displays the angular momentum @xmath56 , the kinematical and dynamical moments of inertia , @xmath57 and @xmath58 , and the rigid - body moments of inertia @xmath59 as functions of the rotational frequency @xmath15 .
we see that the calculated moments of inertia are slightly larger than the rigid - body values at @xmath60 , and smoothly decrease as @xmath15 increases until @xmath612.5 and 1.8 mev@xmath62 for @xmath63s and @xmath3s , respectively .
the result calculated with the sly4 interaction is shown here , but we obtained similar results also with the siii and skm@xmath23 interactions .
calculated quadrupole deformation parameters @xmath64 are displayed in the upper portion of fig .
we see that the @xmath46 values slowly decrease while the axial - asymmetry parameters @xmath65 gradually increase with increasing angular momentum for all cases of @xmath0s , @xmath1s and @xmath3s .
the variations are rather mild in the range of angular momentum shown in this figure .
single - particle energy diagrams ( routhians ) for these sd bands are displayed in fig .
8 as functions of the rotational frequency @xmath15 .
this figure indicates that level crossings take place in @xmath1s and @xmath3s if we further increase the angular momentum .
let us examine stabilities of the sd local minimum against both the axially symmetric and asymmetric octupole deformation ( @xmath66 ) .
figure 9 presents deformation energy curves as functions of the octupole deformation parameters @xmath67 for fixed quadrupole deformation parameters ( the equilibrium value of @xmath46 at the sd minimum in each nucleus and @xmath68 ) .
the computation was carried out by means of the constrained hf procedure with the use of the siii , skm@xmath23 , and sly4 interactions .
the result of calculation clearly indicates that the sd states in @xmath0s , @xmath1s and @xmath3s are stable against the octupole deformations and that they are softer for @xmath69 with lower values of @xmath70 ( i.e. , for @xmath71 and @xmath72 ) , irrespective of the skyrme interactions used .
we obtaind a similar result also for @xmath2s ( but omitted in this figure ) .
although the sd minima in @xmath0s , @xmath1s and @xmath3s are stable with respect to @xmath67 , we found several cases where the sd minima obtained in the symmetry - restricted calculations become unstable when we allow for reflection - asymmetric deformations of a more general type .
as a first example , let us discuss the sd minimum in @xmath36s which appears in the restricted calculation ( see figs . 1 - 3 ) . in this case
, the coupling between the down - sloping @xmath73 $ ] level ( associated with the @xmath26 shell ) and the up - sloping @xmath74 $ ] level ( stemming from the @xmath75 shell ) takes place in the proton configuration , when we allow for the breaking of both the axial and reflection symmetries .
thus , the sd configuration @xmath43 mixes with the @xmath76 configuration ( which lacks the proton excitation to the @xmath26 shell and has a smaller equilibrium value of @xmath46 ) . as a consequence of this mixing
, the barrier between the two configurations disappears and the sd minimum becomes unstable in the unrestricted calculations ( see figs . 1 - 3 ) .
note that the difference @xmath77 in the asymptotic quantum number @xmath53 between the two single - particle levels , @xmath73 $ ] and @xmath74 $ ] , is three , so that they can not be mixed by the octupole operator @xmath78 which transfers the asymptotic quantum numbers @xmath53 and @xmath54 by @xmath79 and @xmath80 .
thus , this mixing may be associated with the reflection - asymmetric deformation of a more higher order like @xmath81 . as a second example , we take up the sd minimum in @xmath2s . in this case ,
two configurations , @xmath43 and @xmath44 compete in energy and their relative energy differs for different versions of the skyrme interaction ( see figs . 1 - 3 ) .
when we allow for the breaking of both the axial and reflection symmetries , the coupling between the down - sloping @xmath82 $ ] level ( associated with the @xmath39 shell ) and the @xmath83 $ ] level in the @xmath38 shell takes place in the neutron configuration , so that they mix each other .
note that the @xmath82 $ ] and @xmath83 $ ] levels satisfy the selection rules , @xmath51 and @xmath52 , for the matrix elements of the octupole operator @xmath50 . in the calculation with the sly4 interaction , since the former configuration with smaller @xmath46
is situated slightly lower in energy than the latter , the barrier between the two configurations disappears as a result of this mixing .
this mixing effect in conjunction with that mentioned above for the @xmath43 configuration in @xmath36s deteriorates the sd minimum for the sly4 case .
the above examples indicate detailed microscopic mechanisms within the mean - field theory how the stability of the sd local minimum is determined by relative energies between the neighboring configurations and their mixing properties .
figure 10 displays the neutron and proton density profiles for the sd states in @xmath0s , @xmath1s and @xmath3s calculated with the use of the sly4 interaction .
we obtained similar results also for siii and skm@xmath23 . in this figure , equi - density lines with 50% and 1% of the central density in the @xmath20- and @xmath21-planes are drawn for the sd bands at @xmath84 and at high spins .
we can clearly see that superdeformed neutron skin appears in @xmath3s which is situated close to the neutron drip line .
the root - mean - square values , @xmath85 and @xmath86 , of these density distributions are listed in table 1 . to indicate the deformation properties of the neutron skin in @xmath3s , calculated values for protons and neutrons
are separately listed together with their sums and differences .
we obtained density distributions similar to those for @xmath3s also for the sd state in @xmath2s .
a similar result of theoretical calculation exhibiting the superdeformed neutron skin was previously reported in ref.@xcite for the sd state in the very neutron - rich nucleus @xmath87dy@xmath88 .
on the basis of the cranked shf calculations in the 3d coordinate - mesh representation , we have suggested that , in addition to the well - known candidate @xmath0s , the neutron - rich @xmath1s and the the drip - line nuclei , @xmath2s and @xmath3s , are also good candidates for finding sd rotational bands in sulfur isotopes .
calculated density distributions for the sd states in @xmath2s and @xmath3s , which are situated close to the neutron - drip line , exhibit superdeformed neutron skins .
the numerical calculations were performed on the nec sx-5 supercomputers at rcnp , osaka university , and at yukawa institute for theoretical physics , kyoto university .
this work was supported by the grant - in - aid for scientific research ( no .
13640281 ) from the japan society for the promotion of science .
root - mean - square values , @xmath85 and @xmath86 , of the density distributions at @xmath84 ( second column ) and at @xmath89 ( third column ) of the sd band in @xmath0s , @xmath1s and @xmath3s , calculated with the use of the sly4 interaction . neutron and proton contributions
are separately listed together with their sums ( total ) and differences ( diff . ) .
+ s to @xmath3s calculated at @xmath84 as functions of the quadrupole deformation @xmath46 by means of the constrained shf procedure with the siii interaction .
the deformation parameter is defined as @xmath90 the axial - asymmetry parameter @xmath65 is constrained to be zero .
solid curves with and without filled circles represent the results obtained by the unrestricted and restricted versions , respectively .
the notation @xmath91 indicates a configuration in which single - particle levels originating from the @xmath26 shell are occupied by @xmath41 protons and @xmath42 neutrons .
likewise , @xmath40 indicates that levels from the @xmath26 shell are occupied by @xmath41 protons and those from the @xmath39 shell by @xmath42 neutrons .
, scaledwidth=100.0% ] shell ) do not mix within the mean - field approximation ( left - hand side ) .
in contrast , when such symmetry restriction is removed , smooth crossover between the two configurations is possible via mixing of the positive- and negative - parity levels ( right - hand side ) .
octupole deformation @xmath92 of the mean field rises in the crossing region . in this figure ,
the crossing between the two levels with the asymptotic quantum numbers @xmath47 $ ] and @xmath48 $ ] is illustrated as an example .
the two levels satisfy the selection rule , @xmath51 and @xmath52 , for the matrix elements of the non - axial octupole operator @xmath50 , so that the mixing between them takes place mainly through the @xmath50 component of the mean field .
, scaledwidth=100.0% ] s , @xmath1s , and @xmath3s calculated by means of the cranked shf method .
results obtained with the use of the siii , skm@xmath23 , and sly4 interactions are plotted by solid , dashed , and dotted curves , respectively .
their shape evolutions as functions of angular momentum @xmath56 in the @xmath64 plane are displayed in the upper portions .
the @xmath46 values decrease with increasing @xmath56 . , scaledwidth=100.0% ] and the moments of inertia @xmath93 are plotted as functions of rotational frequency @xmath15 for the sd rotational bands in @xmath0s , @xmath1s , and @xmath3s .
the sly4 interaction is used .
values of the kinematical and dynamical moments of inertia , @xmath57 and @xmath58 , are plotted in unit of @xmath94/mev by solid and dashed curves , respectively . for reference ,
the rigid - body moments of inertia @xmath59 evaluated with the calculated density @xmath95 are also indicated .
, scaledwidth=100.0% ] s , @xmath1s , and @xmath3s , plotted as functions of rotational frequency @xmath15 .
the left(right)-hand side displays those for protons(neutrons ) .
the levels associated with the @xmath39 and @xmath26 shells are drawn by thick - solid and thick - dashed lines , respectively .
other occupied levels associated with the @xmath96 and @xmath38 shells are drawn by thin - solid and thin - dashed lines , respectively .
unoccupied levels are drawn by thin - dotted lines .
numbers indicate the fermi surfaces and total numbers of single - particle states below them .
the result calculated with sly4 is shown here , but we obtained similar results also with siii and skm@xmath23 . , scaledwidth=100.0% ] ) as functions of the octupole deformation parameters @xmath67 , calculated for @xmath0s , @xmath1s , and @xmath3s , by means of the constrained hf procedure with the use of the siii , skm@xmath23 and sly4 interactions .
the quadrupole deformation parameters are fixed at the equilibrium value of @xmath46 in each nucleus and @xmath68 .
one of the @xmath67 is varied while the other @xmath69 s are fixed to zero .
the deformation parameters @xmath92 and @xmath69 are defined in terms of the expectation values of the octupole operators ( see ref.@xcite for their explicit expressions ) .
, scaledwidth=100.0% ] - and @xmath22- planes of the sd band at @xmath84 in @xmath0s , @xmath1s , and @xmath3s , calculated with the use of the sly4 interaction .
neutron ( proton ) equi - density lines with 50% and 1% of the central density are shown by dashed ( solid ) lines ( the inner and outer lines correspond to the 50% and 1% lines , respectively ) .
right - hand side : same as the left - hand side but for @xmath89 for @xmath0s , @xmath1s , and @xmath3s , respectively .
, scaledwidth=100.0% ] | on the basis of the cranked skyrme - hartree - fock calculations in the three - dimensional coordinate - mesh representation , we suggest that , in addition to the well - known candidate @xmath0s , the neutron - rich nucleus @xmath1s and the drip - line nuclei , @xmath2s and @xmath3s , are also good candidates for finding superdeformed rotational bands in sulfur isotopes .
calculated density distributions for the superdeformed states in @xmath2s and @xmath3s exhibit superdeformed neutron skins .
pacs : 21.60-n ; 21.60.jz ; 27.30.+t + keywords : cranked skyrme - hartree - fock method ; superdeformation ; neutron - rich nuclei ; high - spin state ; sulfur isotopes |
the spin network quantum simulator model @xcite represents a bridge between circuit schemes for standard quantum computation and approaches based on notions from topological quantum field theories ( tqft ) @xcite . the spin network computational space , naturally modelled as a graph for any fixed number of incoming spins , supports computing processes represented by families of paths and provides , on the one hand , a consistent discretized version of the topological quantum computation approach . on the other hand , such a quantum combinatorial scheme , which essentially encodes the ( quantum deformed ) @xmath0 racah
wigner tensor algebra , turns out to be particularly suitable to address problems arising in ( low dimensional ) topology and group theory .
the guiding idea of this paper is that the exponential efficiency that quantum algorithms may achieve with respect to classical ones proves to be especially relevant in problems in which the space of solutions is characterized by a structure definable in terms of the grammar and the syntax of a language , rather than algebraic or number theoretic in nature .
the spin network setting provides a ` natural encoding ' for classes of problems which basically share the combinatorial structure of the language underlying the ( re)coupling theory of @xmath0 angular momenta @xcite . on the other hand , the jones polynomial @xcite is no doubt the most famous knot invariant in topology , a knot invariant being a function on knots ( or links , namely circles embedded in @xmath6space ) which is invariant under isotopy ( smooth deformations ) of the knot . among its many connections to various mathematical and physical areas ( see _ e.g. _
@xcite for applications in statistical mechanics ) , we are mainly interested here in its relations with tqft @xcite . in the seminal paper @xcite , witten put link invariants in a field theoretical setting , showing that jones polynomials arise as vacuum expectation values of wilson loop operators in a three dimensional @xmath0 chern simons ( topological ) quantum field theory where the fundamental representation of the gauge group @xmath0 lives on each component of the link .
such an invariant was extended to arbitrary representations living on the link components and in this paper we shall deal with such generalizations , referred to as ` extended ' or ` colored ' jones polynomials @xcite . from the ( classical ) computational side ,
it was proved that the exact evaluation of the jones polynomial of a link @xmath1 , @xmath7 at @xmath8 root of unity , can be performed in polynomial time in terms of the number of crossings of the planar diagram of @xmath1 if @xmath9 is a 2nd , 3rd , 4th , 6th root of unity .
otherwise , the problem is @xmath10hard @xcite ( the computational complexity class @xmath10hard is the enumerative analog of the @xmath11 class ) .
however , kitaev , larsen , freedman and wang @xcite showed that their ` topological ' quantum computation setting , relying on the same tqft quoted before , implicitly provides an efficient quantum algorithm for the approximation of the jones polynomial at a fifth root of unity . unfortunately , this important algorithm was never explicitly formulated .
this is particularly unfortunate since it is known that the approximation problem is * bqp*hard , and a quantum algorithm for this problem is thus of particular importance .
let us point out that recently aharonov , jones and landau proposed an efficient quantum algorithm that approximates the problem of evaluating the jones polynomial based , rather than on physical results from tqft , on the path model representation of the braid group and the uniqueness of the markov trace for the temperley
lieb algebra @xcite .
the argument is that the @xmath10hardness of the problem does not rule out the possibility of good approximations , and indeed these authors provide an efficient , explicit and simple quantum algorithm to approximate the jones polynomial at all roots of unity for both the trace and the plat closures of a braid . + our strategy is quite different from theirs , since we shall basically provide a quantum ( automaton ) system whose internal evolution can be controlled in such a way that its probability amplitude gives the desired polynomial .
as mentioned , one of the features of the jones polynomial that will be used extensively is that it can also be defined via braids ( a geometric @xmath12braid is a set of @xmath12 strands with fixed endpoints in the plane ) .
a braid can be ` closed up ' to form a link by tying its ends together . in this paper we shall be interested in one of the two ways to perform such closures , namely the plat closure of the braid , and hence consider extended jones polynomials associated with such link diagrams , _ cfr . _
[ borromeanplat ] . on a broader front ,
the study of braid groups and their applications is a field which has attracted great interest from physicists , mathematicians and computer scientists alike ( _ cfr . _
@xcite for an updated review ) . besides for its value in studying the braids in a theoretical framework
, applications to knot theory have been known for years , while applicability to the field of cryptography has been realized recently @xcite .
the analysis of algorithmic problems related to braid group has thus acquired a great practical significance , in addition to its intrinsic theoretical interest .
the approach we present here exploits a @xmath4braided version of the original spin network setting @xcite to make it accept the language of the braid group and to deal with link polynomials ( see @xcite for a presentation of some preliminary results ) . as pointed out before , the ` physical ' background provided by the 3@xmath13 quantum @xmath0 chern
simons field theory plays a prominent role , because our computational scheme is actually designed as a discretized conterpart of the topological quantum computation setting proposed in @xcite .
moreover , this framework is exactly what is needed to deal with ( normalized ) @xmath0colored link polynomials expressed as vacuuum expectation values of composite wilson loop operators , on the one hand , and with unitary representations of the braid group , on the other .
these expectation values , in turn , will provide a bridge between the theory of formal languages and quantum computation , once more having as natural arena for discussion the @xmath4braided spin network environment . we are going to implement families of finite states ( and discrete time)quantum automata capable of accepting the language generated by the braid group , and whose transition amplitudes are colored jones polynomials . more precisely , our results will be interpreted in terms of ` processing of words ' written in the alphabet given by the generators of the braid group on a quantum automaton in such a way that the expectation value associated with the internal automaton ` evolution ' is exactly the extended jones polynomial . the quantum automaton in question
will in turn correspond to a path in the @xmath4braided spin
network computational graph .
the calculation of the polynomial of ( the plat closure of ) a link @xmath1 on @xmath2 strands will be shown to be bounded from above by a linear function of the number of crossings of the link , on the one hand , and polynomially bounded in terms of the braid index @xmath2 , on the other .
notice that the growth rate of the time complexity function in terms of the integer @xmath3 appearing in the root of unity @xmath4 can be easily estimated to be ( polynomially ) bounded by resorting to the tqft background , since @xmath3 is nothing but the chern
simons coupling constant .
we shall leave as open problems the analysis of the complexity of the preparation of ( initial and final ) states as well as the efficient implementation of the individual automaton transition functions , which might be addressed by means of approximating ( classical or quantum ) algorithms .
+ in conclusion , we argue that our field theoretical approach could be further generalized , by suitable modifications of the braiding prescriptions in the spin network scheme , to deal with 2variables link polynomials such as the homfly invariant @xcite , related to the partition function of potts model @xcite .
the content of the paper is , as far as possible , self contained . in section 2
we briefly recall the definitions of classical and quantum languages and finite states automata . in section 3
we give a review of the spin network computational framework modelled on the racah wigner tensor algebra of @xmath0 . in section 4
we deal with the @xmath4braided version of the spin network simulator , which relies on the tensor algebra of @xmath5 ( at @xmath14 root of unity ) .
section 5 is splitted into two parts : in 5.1 we review the ` quantum group approach ' ( and related @xmath15-matrix ) to the study of ( unitary ) braid group representations and ` quantum ' link invariants ; in 5.2 we present the field
theoretical background ( chern simons tqft , wess zumino boundary theory , composite wilson loop operators and their expectation values ) trying to resort to geometric intuition rather than to a deep knowledge of techniques in quantum field theory . in section 6
we explain in details the automaton calculation of the extended jones polynomial .
the theory of automata and formal languages addresses in a rigorous way the notions of computing machines and computational processes .
if @xmath16 is an alphabet , made of letters , digits or other symbols , and @xmath17 denotes the set of all finite sequences of words over @xmath16 , a language @xmath18 over @xmath16 is a subset of @xmath17 .
the length of the word @xmath19 is denoted by @xmath20 and @xmath21 is its @xmath22th symbol .
the empty word is @xmath23 and the concatenation of two words @xmath24 , @xmath25 is denoted simply by @xmath26 . in the sixties
noam chomsky introduced a four level hierarchy describing formal languages according to their structure ( grammar and syntax ) : regular languages , context free languages , context sensitive languages and recursively enumerable languages .
the processing of each language is inherently related to a particular computing model ( see _ e.g. _
@xcite for an account on formal languages ) .
here we are interested in finite states
automata , the machines able to accept regular languages . a deterministic finite state automaton ( dfa )
consists of a finite set of states @xmath27 , an input alphabet @xmath16 , a transition function @xmath28 , an initial state @xmath29 and a set of accepted states @xmath30 .
the automaton starts in @xmath29 and reads an input word @xmath19 from left to right . at the @xmath22th step ,
if the automaton reads the word @xmath21 , then it updates its state to @xmath31 , where @xmath32 is the state of the automaton reading @xmath21 .
one says that the word has been accepted if the final state reached after reading @xmath19 is in @xmath33 . in the case of a non deterministic finite state automaton ( nfa ) ,
the transition function is defined as a map @xmath34 , where @xmath35 is the power set of @xmath27 . after reading a particular symbol
, the transition can lead to different states , according to some assigned probability distribution . if a nfa has @xmath36 states , for each symbol @xmath37 there is an @xmath38 transition matrix @xmath39 for which @xmath40 if and only if the transition from the state @xmath22 to the state @xmath41 is allowed once the symbol @xmath42 has been read . generally speaking , quantum finite states
automata ( qfa ) are obtained from their classical probabilistic counterparts by moving from the notion of ( classical ) probability associated with transitions to quantum probability amplitudes .
computation takes place inside a suitable hilbert space through unitary matrices and a number of different models have been proposed , see _ e.g. _ @xcite , just to mention a couple of them .
following @xcite , the measure once quantum automaton is a 5-tuple @xmath43 , where @xmath44 is a finite set of states , @xmath45 is a finite input alphabet with an end marker symbol @xmath46 and @xmath47 is the transition function .
here @xmath48 is the probability amplitude for the transition from the state @xmath49 to the state @xmath50 upon reading the symbol @xmath51 .
the state @xmath52 is the initial configuration of the system , and @xmath53 is an accepted final states .
for all states and symbols the function @xmath54 must be unitary .
marker @xmath46 is the last symbol of each input and computation terminates after reading it . at the end of the computation
the automaton measures its configuration : if it is an accepted state then the input is accepted , otherwise is rejected .
the configuration of the automaton is in general a superposition of states in the hilbert space where the automaton lives .
the transition function is represented by a set of unitary matrices @xmath55 , where @xmath56 represents the unitary transition of the automaton reading the symbol @xmath51 .
the probability amplitude for the automaton of accepting the string @xmath19 is given by @xmath57 and the explicit form of @xmath58 defines the language @xmath59 accepted by that particular automaton . if @xmath60 denotes the projector over the accepted states , the probability for the automaton of accepting the string @xmath19 is given by @xmath61 where @xmath62 .
the spin network model of computation was introduced in @xcite and worked out in @xcite as a general framework for processing information in the quantum context and is essentially modelled on the combinatorics of the racah
wigner algebra of @xmath0 .
the spin network can be seen as a collection of graphs @xmath63 parametrized by an integer @xmath36 ( @xmath64 ) , where @xmath65 is the number of incoming angular momentum variables , each associated with an irreducible representation ( irrep ) of @xmath0 , @xmath66 @xmath67 in @xmath68 units ( we choose units in which @xmath69 ) . on the physical side ,
these @xmath65 basic variables enter in the construction of different sets of ( pure angular momenta ) eigenspaces selected according to the different types of quantum interactions we whish to simulate .
the fact that physical interactions in many ( conservative ) quantum systems can be well modelled on ( combinations of ) two body interactions @xcite opens the possibility of calling into play the powerful algebraic combinatorial setting underlying @xmath0 binary coupling and recoupling theory ( _ cfr .
_ @xcite and the original references therein ) . before going into some more details on this realization of the spin network graphs ,
let us point out that the combinatorial structure encoded into the racah
wigner algebra is actually shared by other discrete structures .
a first type of realization is purely graph theoretical .
the vertex set @xmath70 of the graph @xmath63 can be identified with the set of ( rooted ) binary trees with @xmath65 labelled leaves where the leaves ( terminal nodes ) and the internal nodes of the trees are labelled by integers and half
integers @xmath71 , _ cfr . _
[ trees ] . leaves .
such trees are in one to
one correspondence with the vertex set @xmath70 of the graph @xmath72,width=226 ] undirected edges between vertices are drawn whenever a pair of vertices ( labelled trees ) are connected by two kinds of topological elementary moves , namely twist and rotation , illustrated in fig .
[ twist - rot ] . where only 30 out of 60 vertices are shown ( the picture can be completed by taking the mirror image of each tree at the antipodal vertex ) . the remaining 60 vertices are arranged into an isomorphic graph obtained by swapping one pair of labels , _ e.g. _ @xmath73 .
solid edges represent rotations and dashed edges represent twists.,width=264 ] the resulting graph , known as twist rotation graph , is depicted for @xmath74 in fig .
[ trgraph ] and its combinatorial properties are analyzed in @xcite and in appendix a of @xcite .
another realization of the spin
network is in terms of words endowed with pairs of parentheses representing a non commutative and non associative binary operation . in this case
the vertices of the graph @xmath63 are associated with words @xmath19 made of letters from the alphabet @xmath75 @xmath76 @xmath77 pairs of labelled parentheses @xmath78 @xmath79 , _
e.g. _ @xmath80 where @xmath81 is the label assigned to the root .
two vertices are connected by an edge if it is possible to switch from one to the other either by swapping the elements inside a parenthesis , @xmath82 , or by changing the parenthesization structure @xmath83 @xmath84 .
coming back to the racah wigner setting , the interpretation of the spin network graph goes on as follows .
there exists a one to
one correspondence @xmath85 between the vertices of @xmath86 and the computational hilbert spaces of the simulator
. the label @xmath87 has the following meaning : for any given pair @xmath88 , all binary coupling schemes of the @xmath65 angular momenta @xmath89 , identified by the quantum numbers @xmath90 ( summing up to a total @xmath81 ) plus @xmath91 ( corresponding to the @xmath92 intermediate angular momenta @xmath93 ) and by the brackets defining the binary couplings , provide the ` alphabet ' in which quantum information is encoded ( the rules and constraints of bracketing are instead part of the ` syntax ' of the resulting coding language ) .
the hilbert spaces @xmath94 thus generated are spanned by complete orthonormal sets of states with suitable quantum number label set such as , _
e.g. _ for @xmath95 , @xmath96 , @xmath97 .
more precisely , for a given value of @xmath36 , @xmath98 is the simultaneous eigenspace of the squares of @xmath99 hermitean , mutually commuting angular momentum operators @xmath100 with fixed sum @xmath101 , of the intermediate angular momentum operators @xmath102 and of the operator @xmath103 ( the projection of the total angular momentum @xmath104 along the quantization axis ) .
the associated quantum numbers are @xmath105 @xmath106 @xmath107 @xmath108 and @xmath109 , where @xmath110 in integer steps .
+ if @xmath111 @xmath112 @xmath113 denotes the factorized hilbert space , namely the @xmath114fold tensor product of the individual eigenspaces of the @xmath115 s , the operators @xmath116 s represent intermediate angular momenta generated , through clebsch
gordan series , whenever a pair of @xmath117 s are coupled .
as an example , by coupling sequentially the @xmath118 s according to the scheme @xmath119 @xmath120 which generates @xmath121 , @xmath122 , and so on we should get a binary bracketing structure of the type @xmath123 @xmath124 @xmath125 , where for completeness we add an overall bracket labelled by the quantum number of the total angular momentum @xmath81 .
note that , as far as @xmath126 s quantum numbers are involved , any value belonging to @xmath127 is allowed , while the ranges of the @xmath128 s are suitably constrained by clebsch
gordan decompositions ( _ e.g. _ if @xmath121 @xmath129 @xmath130 @xmath131 ) .
+ we denote a binary coupled basis of @xmath114 angular momenta in the @xmath132representation and the corresponding hilbert space as @xmath133^{\mathfrak{b}}\ , ; k_1^{\mathfrak{b}\,},\,k_2^{\mathfrak{b}\ , } , \ldots , k_{n-1}^{\mathfrak{b}}\ , ; \,jm\ , \rangle,\ ; -j\leq m\leq j \}\ ] ] @xmath134 where the string inside @xmath135^{\mathfrak{b}\,}$ ] is not necessarily an ordered one , @xmath136 indicates the current binary bracketing structure and the @xmath128 s are uniquely associated with the chain of pairwise couplings selected by @xmath136 .
+ for a given value of @xmath81 each @xmath137 has dimension @xmath138 over @xmath139 , but hilbert spaces corresponding to different bracketing schemes , although isomorphic , are not identical .
they actually correspond to ( partially ) different complete sets of physical observables , namely for instance @xmath140 @xmath141 @xmath142 @xmath143 and @xmath144 @xmath145 respectively ( in particular , @xmath146 and @xmath147 can not be measured simultaneously ) . on the mathematical side
this remark reflects the fact that the tensor product @xmath148 is an associative operation only up to isomorphisms . for what concerns unitary operations acting on the computational hilbert spaces ( [ bbspace ] )
, we shall consider here unitary transformations associated with recoupling coefficients ( @xmath149 symbols ) of @xmath0 , thought of as @xmath41gates in the present quantum computing context .
as shown in @xcite , any such coefficient can be splitted into ` elementary ' @xmath41gates , namely racah and phase transforms .
a racah transform applied to a basis vector is defined formally as @xmath150 are used here to denote generic , both incoming ( @xmath151 s in the previous notation ) and intermediate ( @xmath152 s ) spin quantum numbers ( this operation corresponds to a rotation in the twist
rotation graph , _ crf .
[ twist - rot ] , bottom and fig .
[ trgraph ] ) .
its explicit expression reads @xmath153 @xmath154^{1/2 } \left\ { \begin{array}{ccc } a & b & d\\ c & f & e \end{array}\right\}\;|(\,(a\,b)_d \,c)_f \,;m\rangle,\ ] ] where there appears the @xmath155 symbol of @xmath0 and @xmath156 plays the role of the total angular momentum quantum number .
note that , according to the wigner eckart theorem , the quantum number @xmath109 ( as well as the angular part of wave functions ) is not altered by such transformations , and that the same happens with any @xmath149 symbol . on the other hand ,
the effect of a phase transform @xmath157 ( a twist operation on the twist
rotation graph , see fig .
[ twist - rot ] , top and fig .
[ trgraph ] ) amounts to introducing a suitable phase whenever two spin labels are swapped @xmath158 these unitary operations are combinatorially encoded into the edge set @xmath159 of the graph @xmath86 : @xmath160 is just the subset of the cartesian product @xmath161 selected by the action of these unitary @xmath41gates . in the framework
described above , a computation is represented in a natural way by a collection of step by step transition rules ( gates ) , namely a family of ` elementary unitary operations ' and we assume that it takes one unit of the intrinsic discrete time variable to perform anyone of them . such prescriptions amount to select ( families of ) ` directed paths ' in the spin
network computational space @xmath162 , all starting from the same input state and ending in an admissible output state .
a single path in the given family can be interpreted as a ( finite states ) quantum automaton calculation , once we select a particular encoding scheme for the problem we wish to address . by a directed path @xmath163 with fixed endpoints we mean a ( time ) ordered sequence @xmath164 for computational states ( which are vectors expressed in the bases ) and @xmath165 is the lexicographical labelling of the states along the path . finally , @xmath166 is the length of the path @xmath167 and @xmath168 is the time required to perform the process in terms of the discrete time unit @xmath169 .
a computation consists in evaluating the expectation value of the unitary operator @xmath170 associated with the path @xmath167 , namely @xmath171 by taking advantage of the possibility of decomposing @xmath170 uniquely into an ordered sequence of elementary gates , ( [ snexp1 ] ) becomes @xmath172 with @xmath173 for short . the symbol @xmath174 denotes the ordered product along the path @xmath167 and each elementary operation is rewritten as @xmath175 @xmath176 to stress its ` one step ' character .
such expectation values are particular instances of the general expression for the quantum amplitude of a finite states automaton , once a suitable language has been encoded into the computational space of the spin network simulator .
as we shall see in the following section , the basic ingredient for addressing link invariants arising in the context of chern simons field theory is the ` tensor structure ' naturally associated with the representation ring of the lie algebra of a simple compact group which plays the role of the gauge group of the theory . in the case of @xmath0
this structure is provided by ( tensor products of ) hilbert spaces supporting irreducible representations together with unitary morphisms between them : these are exactly the objects collected into the racah
wigner algebra discussed in section 3 .
however , when dealing with ( planar diagrams of ) links we shall also have to specify the eigenvalues of the braiding matrix to be associated with the crossings of the links and this extension can be achieved by ` braiding ' the racah wigner tensor category . in the present context , it is natural to take advantage of quantum group techniques in order to ` split ' any phase transform ( [ cphase ] ) by assigning different weights depending on a deformation parameter @xmath4 to be defined below to right and left handed twists . from the combinatorial viewpoint
, this generalization corresponds to replace the spin network computational space @xmath162 with its @xmath4braided counterpart ( see fig .
[ brtrgraph ] ) @xmath177 where the ( classical ) @xmath155 symbol in any racah trasform ( [ cracah2 ] ) will become @xmath4deformed .
braided twist rotation graph @xmath178 : with respect to the unbraided situation , each twist has been splitted.,width=264 ] the tensor category we are going to introduce is associated with the quantum group @xmath5 ( @xmath179 root of unity ) , namely the universal enveloping algebra of @xmath0 endowed with additional structures which make it a quasitriangular quasi hopf*algebra ( see _ e.g. _ @xcite and other references therein ) .
@xmath180 is an associative algebra generated by elements @xmath181 , @xmath182 and @xmath103 which satisfy the commutation relations
@xmath183 = \pm j _
\pm ; \left [ { j _ + , j _ - } \right ] = \left [ { 2j_z } \right]_q , \hfill \\\ ] ] where the @xmath4integer @xmath184_q$ ] is defined as @xmath184_q \equiv ( q^{\mathbf{n}/2}-q^{-\mathbf{n}/2})/ ( q^{1/2}-q^{-1/2})$ ] .
@xmath5 is a deformation of the universal enveloping algebra of the lie algebra @xmath185 since in the limit @xmath186 the above relations reduce to the commutation relations for the @xmath185 generators .
+ a hopf algebra structure can be introduced by defining the coproduct homomorphism @xmath187 acting on @xmath181 , @xmath182 and @xmath103 according to @xmath188 @xmath189 the tensor algebra associated with @xmath5 can be worked out in practice as in the case of @xmath185 , so that we have hilbert spaces supporting irreducible representations , @xmath4clebsch gordan coefficients , @xmath4racah coefficients and so on .
the crucial difference consists in the fact that the irreps label set acquires a cut off , namely each label must be chosen in the set @xmath190 @xmath191 , where the integer @xmath3 is related to the deformation parameter @xmath4 by @xmath4 @xmath192 . denoting by @xmath193 and @xmath194 the hilbert spaces supporting two irreps @xmath195
, their ( truncated ) tensor product can be decomposed according to the clebsch
gordan series @xmath196 as happens in the classical case , the two bases associated with the eigenspaces involved in the tensor product can be connected by means of clebsch gordan coefficients according to @xmath197 where @xmath198 , @xmath199 , @xmath200 .
quantum cg coefficients @xmath201 can be suitable normalized and satisfy orthogonality relations @xcite . + the quantum racah transformation comes out when we consider different binary couplings in the tensor product @xmath202 of three irreducible representations , as done in the classical case ( _ cfr . _ and ) . for instance
@xmath203 @xmath204_q [ 2j_{23 } + 1]_q \;\right)^{-1/2}\ , @xmath4dimensions of the irreps involved .
the components of @xmath205 are the racah coefficients of the algebra @xmath206 and these symbols satisfy orthogonality relations , symmetry properties and identities which look like suitable @xmath4deformations of the corresponding classical ones ( and reduce to them in the limit @xmath207 ) @xcite .
the quantum racah coefficient and the @xmath4counterpart of the wigner @xmath155 symbol differ as usual by a phase factor , namely @xmath208 finally , we introduce the ( differently normalized ) symbol @xmath209_q \left [ { 2j_{23 } + 1 } \right]_q } } } , \ ] ] which , on the one hand , enhances the matrix character of the quantum racah transform and , on the other , is particularly suitable to be generalized to deal with more than three incoming spin labels .
link invariants are functions on links ( collections of knots , namely closed circles in @xmath6space ) which depend only on the isotopy class of the link .
an ( ambient ) isotopy can be thought of as a continuous transformation performed on the link embedded in @xmath210 which deformes at will the shape of the link without cuttings .
let us point out preliminarly that the link invariants of polynomial type we are going to address here are ` universal ' in the sense that historically distinct approaches ( r matrix representations obtained with the quantum group method , monodromy representations of the braid group in @xmath211 conformal field theories , the quasi tensor category approach by drinfeld and the @xmath212 quantum chern simons theory , see _ e.g. _
@xcite for reviews ) are indeed different aspects of the same underlying algebraic structure .
we shall focus in particular on the chern simons setting @xcite since , on the one hand , it embraces the universal structure of ( unitary ) braid group representations shared by all the models quoted above and , on the other , can be naturally encoded into the ( braided ) spin network computational scheme .
the ( colored ) link polynomials arising from @xmath0 quantum cs theory can be referred to as ` extended ' jones polynomials , since the jones polynomial @xcite is recovered by selecting the fundamental ( @xmath213 ) representation of @xmath0 on each of the link components ( or on each strand of the associated braid ) . moreover ,
the topological quantum field approach is inherently related to low dimensional geometry since , for instance , suitable combinations of these invariants can be interpreted as topological invariants of hyperbolic @xmath6manifolds , obtained by surgery along framed links in the @xmath6-sphere @xcite .
+ let us point out that the definitions of link polynomials from hecke ( or temperley
lieb ) algebra realizations of the braid group exploited in @xcite in the quantum computational context can be derived quite easily in the framework we are adopting here , since it can be shown that the associated invariants do satisfy the linear skein relations which characterize such realizations @xcite . before addressing a full fledged approach to @xmath212 chern
simons theory , we pause a little bit digressing on @xmath15matrix representations of the braid group arising from quantum groups
. the associated invariants of knots and links are commonly refereed to as ` quantum ' invariants , since they are quantities depending on the deformation parameter @xmath4 of the ` quantum group ' under consideration .
let @xmath214 be a ( semi)simple lie algebra , @xmath215 its universal enveloping algebra and @xmath70 a finite dimensional ( complex ) vector space in the associated tensor algebra ( the prototype is of course the unitary tensor algebra of @xmath216 described in details in section 4 ) .
the representation theory of any such quantum group is naturally endowed with an invertible linear operator , the so called @xmath15matrix @xmath217 which satisfies the quantum yang
baxter equation @xmath218 where both sides of the above expression are to be understood as linear transformations @xmath219 .
the general procedure for constructing quantum invariants of oriented knots ( or links ) presented as closure ( or platting ) of braids can be outlined as follows .
consider an oriented knot diagram ( namely the projection of a knot with orientation onto a fixed plane ) and insert an horizontal line as depicted in fig .
[ trefoil ] . , where each factor is chosen in order to comply with the diagram orientation.,width=188 ] to each intersection point between the line and the diagram we assign either the representation space @xmath70 or its dual @xmath220 , depending on whether the portion of the knot nearby the intersection is oriented upwards or downwards .
the whole configuration of such points on the line turns out to be associated with the tensor product of the individual vector spaces ( orderered from left to right ) .
the connection with braid groups comes out when we consider two parallel horizontal lines intersecting the knot diagram .
more precisely , the portion of the knot diagram between a pair of horizontal lines represents the geometric realization of a braid @xmath221 , which in turn is an element of the artin braid group @xmath222 , for some suitable @xmath36 .
@xmath222 has @xmath36 generators , denoted by @xmath223 plus the identity @xmath224 , which satisfy the relations @xmath225 @xmath226 an element of the braid group is a word in the standard generators of @xmath222 , _ e.g. _ @xmath227 @xmath228 @xmath229 @xmath230 ; the length @xmath231 of the word @xmath221 is the number of its letters .
the group acts naturally on topological sets of @xmath36 disjoint strands ordered from left to right in the sense that each generator @xmath232 corresponds to the over crossing of the @xmath22th strand on the @xmath233th , and @xmath234 represents the inverse operation ( under crossing ) according to @xmath235 @xmath236 .
on the other hand , when we represent @xmath222 in the tensor algebra of @xmath215 , the action of a braid @xmath221 is naturally associated with a linear operator @xmath237 connecting the vector spaces introduced above , see fig .
[ braidmap ] . is represented as a map @xmath237 between the vector spaces living on the bottom and top lines.,width=226 ] since @xmath238 is a linear representation , we can simply specify its action on the standard generators @xmath239 @xmath240 to get @xmath241 @xmath242 , and extend this action to @xmath237 by linearity .
the @xmath15matrix , namely the linear operator introduced in , is to be intended as the set of ( elementary ) crossing operators in some given representation @xmath238 , constrained by the quantum yang baxter relation .
knot theory is closely related to ( representations of ) braid groups owing to alexander s theorem @xcite , which states that every knot ( or link ) @xmath1 in the @xmath6sphere @xmath243 @xmath244 can be presented ( not uniquely ) as a closed braid for some suitable @xmath36 ( to get a knot from the open braid of fig .
[ braidmap ] we have to connect with arcs the lower and upper endpoints of each strand ) .
we might also consider the plat presentation of a knot ( characterized by the fact that the braid involved must possess an even number of strands ) , which is exactly the type of presentation depicted in fig .
[ trefoil ] for the trefoil knot ( see also fig .
[ borromeanplat ] and fig .
[ trefoilex ] in the appendix ) .
anyway , we can generate invariants of knots ( links ) for both types of presentations by taking some trace of the operator @xmath237 , where @xmath221 is the braid associated with the given knot or link .
the fact that the resulting quantities must depend only on the isotopy type of the knot can be suitably translated into the braid group setting by resorting to the notion of invariance under markov moves , and thus we should actually speak of ` markov traces ' ( _ cfr . _
@xcite for reviews on knot theory and braid group ) . summing up
, the quantum group approach provides a purely algebraic construction of link invariants as ( markov ) traces of representation matrices of the braid group in the tensor algebra of @xmath215 .
such invariants are polynomials in the deformation parameter @xmath4 and its inverse @xmath245 . in the case of @xmath216 ( @xmath4 a root of unity ) , the associated @xmath4braided racah
wigner algebra ( discussed in section 4 ) is naturally endowed with hilbert spaces and unitary operators , namely the ideal arena to address ( quantum ) computational problems concerning both link polynomials and braid group .
a topological quantum field theory ( tqft ) is a particular type of gauge theory , namely a theory quantized through the ( euclidean ) path integral prescription starting from a classical yang
mills action defined on a suitable @xmath13dimensional space(time ) .
tqft are characterized by observables ( correlation functions ) which depend only on the global features of the space on which these theories live , namely they are independent of any metric which may be used to define the underlying classical theory .
the geometrical generating functionals and correlation functions of such theories are computable by standard techniques in quantum field theory and provide novel representations of certain global invariants ( for @xmath13-manifolds and/or for particular submanifols embedded in the ambient space ) which are of prime interest .
let us recall in brief the basic axioms for a unitary tqft in @xmath246 before going through the case which is of interest here , namely @xmath0 chern simons theory @xcite
. denote by @xmath247 and @xmath248 a pair of 2dimensional manifolds and by @xmath249 a @xmath6dimensional manifold with boundary @xmath250 @xmath251 ( all manifolds here are compact , smooth and oriented ) .
a unitary @xmath6dimensional quantum field theory corresponds to the assignment of + * i ) * finite dimensional hilbert spaces ( endowed with non degenerate bilinear forms ) @xmath252 and @xmath253 to @xmath247 and @xmath248 , respectively ; + * ii ) * a map ( technically , a functor ) connecting such hilbert spaces @xmath254}\ ; \mathcal{h}_{\sigma_2}\ ] ] where @xmath249 is a manifold which interpolates between @xmath247 ( incoming boundary ) and @xmath248 ( outgoing boundary ) . without entering into details concerning a few more axioms ( diffeomorphism invariance , factorization _ etc .
_ ) we just recall that unitarity implies that + * iii ) * if @xmath255 denotes the surface @xmath45 with the opposite orientation , then @xmath256 @xmath257 , where @xmath258 stands for complex conjugation ; + * iv ) * the mappings are unitary and @xmath259=$ ] @xmath260 $ ] , where @xmath261 denote the manifold with the opposite orientation with respect to @xmath249 .
the classical @xmath0 chern simons action for the sphere @xmath243 ( which is the simplest compact , oriented @xmath6manifold without boundary ) is given by @xmath262 where @xmath16 is the connection 1form with value in the lie algebra @xmath185 of the gauge group , @xmath3 is the coupling constant , @xmath263 is the exterior differential , @xmath264 is the wedge product of differential forms and the trace is taken over lie algebra indices .
the partition function of the quantum theory corresponds to the map restricted to the case of empty boundaries and is obtained as a ` path integral ' , namely by integrating the exponential of @xmath22 times the classical action over the space of gauge invariant flat @xmath0 connections ( the field variables ) according to the formal expression @xmath265\;=\ ; \int [ da]\,\exp \left\{\frac{i\ , k}{4 \pi}\ , s_{cs}\,(a)\,\right\}\ ] ] where the coupling constant @xmath3 is constrained to be a positive integer by the gauge invariant quantization procedure and is related to the deformation parameter @xmath4 ( see below ) .
the generating functional , written for a generic compact oriented @xmath6manifold @xmath249 with @xmath266 , is a global invariant , namely depends only on the topological type @xcite .
the extension of to the case of a manifold with boundaries , @xmath267 , requires modifications of the classical action by suitable wess
type terms to be associated with each boundary component @xcite .
however , we do not need here the explicit expression of such boundary action since what we are interested in are expectation values of observables in the quantized field theory which will just require the knowledge of ( vectors belonging to ) the boundary hilbert spaces , _ cfr . _
* i ) * above . in particular
, it turns out that the gauge invariant observables in the quantum cs theory are expectation values of wilson line operators associated with oriented knots ( links ) embedded in the @xmath6manifold ( commonly referred to as wilson loop operators ) .
knots and link are ` colored ' with irreps of the gauge group @xmath0 , restricted to values ranging over @xmath268 , where the integer @xmath3 is related to the deformation parameter @xmath4 by @xmath269 ( see section 4 for details on the @xmath216 representation algebra ) . the wilson loop operator associated with a knot
@xmath270 carrying a spin@xmath41 irreducible representation is defined as ( the trace of ) the holonomy of the connection 1form @xmath16 evaluated along the closed loop @xmath270 @xmath271 , namely @xmath272=tr_j\,pexp\oint_ka,\ ] ] where @xmath273 is the path ordering . for a link @xmath1 made of a collection of @xmath32 knots
@xmath274 , each labelled by an irrep , the expression of the composite wilson operator reads @xmath275\,=\,\prod_{l=1}^s\;\mathbf{w}_{j_l}\,[k_l].\ ] ] in the framework of the path integral quantization procedure , expectation values of observables are defined as functional averaging weighed with the exponential of the classical action .
in particular , the functional average of the wilson operator is @xmath276\,=\,\frac { \int [ da]\;\mathbf{w}_{j_1 \ldots j_s}\,[l]\,e^{\,\frac{ik } { 4\pi } \,s_{cs}\,(a ) } } { \int [ da]\;e^{\,\frac{ik } { 4\pi } s_{cs}\,(a)}},\ ] ] where @xmath277 is the cs action for the @xmath6sphere given in and the generating functional in the denominator will be normalized to 1 in what follows .
it can be shown that this expectation value , which essentially$ ] and the genuine colored jones polynomial is given by @xmath278 @xmath279 @xmath280 $ ] , once suitable normalizations for the unknots have been chosen . here
@xmath281 is the writhe associated with the planar diagram @xmath282 of the link @xmath1 , defined as @xmath283 .
the summation runs over the self crossing points of @xmath282 and @xmath284 according to simple combinatorial rules ( see _ e.g. _ @xcite ) .
the writhe is easily recognized from the link diagram by simple counting arguments , so that computational problems involving both link invariants belong to the same complexity class . ]
coincides with the extended ( colored ) jones polynomial @xcite , depends only on the isotopy type of the oriented link @xmath1 and on the set of irreps @xmath285 ( note also that @xmath286 = \mathcal{e } [ \bar{l}]$ ] , where @xmath287 is obtained from @xmath1 by reversing the orientation ) .
the explicit evaluation of can be carried out in several ways , by resorting to either field
theoretic methods , quantum group approaches ( outlined above ) or through combinatorial state sum functionals . for future
convenience we just sketch here the approach which relies on the extension of cs quantum theory endowed with a wess zumino conformal field theory on its boundary to the case in which the boundary components are intersected by knots or links , namely become @xmath288manifolds with punctures ( note that this setting is closely related to the topological quantum computation approach @xcite ) .
the basic geometric ingredients can be easily visualized as in fig .
[ pant ] , where a portion of a @xmath6dimensional manifold @xmath249 ( technically , a handlebody decomposition ) is shown , together with an incoming boundary @xmath247 and an outgoing boundary @xmath248 made of two disjoint components , @xmath289 and @xmath290 .
a portion of some knot ( link ) embedded in the ambient @xmath6manifold is also depicted , and its intersections with the boundaries are ` punctures ' which inherit the irreps labels from the associated ( wilson ) lines .
manifold with one incoming boundary and two outgoing boundaries .
lines belong to some knot ( or link ) embedded in the manifold and intersect the @xmath211 boundaries in some points ( punctures).,width=226 ] according to the axioms of tqft , we may associate with each boundary a ( finite dimensional ) hilbert space , that is @xmath252 for the incoming boundary and @xmath291 ( here , for simplicity , we do not explicitate the labels of puncures ) . the chern
simons unitary functional ( see axiom * ii ) * ) is a state in the tensor product of these hilbert spaces or , more precisely , @xmath292\;:\;\mathcal{h}_{\sigma_1}\;\rightarrow \mathcal{h}_{\sigma_2}\ ] ] @xmath293\;\;\in\;\mathcal{h}_{\sigma_1}\,\otimes \mathcal{h}^{*}_{\sigma_2},\ ] ] where in the last row we have used also axiom * iii ) * since the incoming and outgoing boundaries must be endowed with opposite orientations .
moreover , such type of expression is compatible with the quantum group approach outlined in section 5.1 since the the chern
simons mapping in , when restricted to punctures , induces automatically ( unitary ) representations of the braid group in the tensor algebra of @xmath5 .
+ finally , it can be shown @xcite that the conformal blocks of the @xmath294 wess zumino field theory living on the boundaries with punctures actually provide the basis vectors for the hilbert spaces introduced above ( the level @xmath295 of the wz model is related to the deformation parameter @xmath4 according to @xmath296 , and in turn @xmath295 is related to the coupling constant @xmath297 of the cs theory in the bulk by @xmath298 ) . in the following section we shall carry on the explicit construction of such bases , which will allow us to recast the expectation value of the composite wilson operator into a form suitable to be handled for computational purposes .
as anticipated at the end of the previous section , we start with the construction of the basis vectors which will enter into the explicit expression of the expectation value of the composite wilson operator emerging from quantized @xmath212 chern simons theory .
we use here the general setting given in @xcite since it can be easily adapted to the @xmath4braided spin network scheme of section 4 .
+ consider an oriented link @xmath1 embedded in the @xmath6sphere , @xmath243 @xmath299 , endowed with a plat representation , namely presented as the closure of an oriented braid with @xmath2 strands ( _ cfr .
[ borromeanplat ] , [ trefoil ] and [ trefoilex ] ) .
if we remove two open three balls from @xmath243 we get two boundaries , @xmath247 and @xmath248 , both topologically equivalent to @xmath300 , but with opposite orientations , @xmath301 @xmath302 @xmath303 ( recall from section 5.2 that an @xmath294 wess zumino conformal field theory is naturally associated with the oriented boundary surfaces ) .
we can accomodate in such an ambient manifold , @xmath2 ` unbraided ' wilson lines carrying irreps @xmath304 , starting from the incoming ( lower ) boundary and ending into the outgoing ( upper ) one ( punctures inherit the labellings @xmath305 from the strands of the braid ) .
denote this ` identity ' colored oriented braid as @xmath306 where @xmath307 @xmath308 represents the spin @xmath309 together with an orientation @xmath310 for a strand going into or away from the boundary , while stars over the symbols represent here the opposite choice of the orientation .
+ in order to generate an arbitrary ( oriented ) braid @xmath311 out of the identity braid @xmath312 we have to apply a braiding operator , denoted by the symbol @xmath313 and written in terms of generators @xmath314 to be defined below , starting from the lower boundary . with such prescription we shall get the braid @xmath315 where the labels have been ordered according to the requirement of having a plat presentation for the associated oriented link .
our goal will consist in recasting the expectation value of the composite wilson operator , written in functional terms in , into an expression which contains a quantity of the type @xmath316 where @xmath317 is the operator associated with the oriented braid .
the shorthand notations @xmath318 and @xmath319 represent correlators of @xmath2 primary fields in the @xmath294 wz theory , to be interpreted here as states belonging to the boundary hilbert spaces associated , respectively , with the incoming and outgoing hilbert spaces of the underlying cs theory ( _ cfr . _ the axioms for tqft in section 5.2 ) .
+ the basis vectors to be associated with the incoming boundary can be denoted in general as @xmath320;\;[\mathbf{k};\,\mathbf{h}];\;0,0\right\rangle,\ ] ] where the last two entries are the quantum numbers @xmath321 @xmath302 @xmath322 for a singlet state of the total angular momentum and the first string represents the incoming spin variables ( we drop the hat on oriented objects whenever not necessary ) .
the second group of entries denotes particular sets of intermediate angular momentum labels , arising from binary couplings of the @xmath2 primary fields @xmath41 s , with a bipartite structure represented by the symbols @xmath323 and @xmath324 , chosen in order to comply with the rules described below . before addressing the latter in general ,
let us have a look at the simplest non trivial case of @xmath325 incoming spin labels .
the underlying admissible combinatorial structures are depicted in fig .
[ racahontrees ] . as labelled binary trees
. they are connected by a duality matrix.,width=226 ] in this case we have just two types of basis vectors , related by the so called duality matrix of wz theory @xmath326;[k , k;-];00 \right\rangle=\sum\limits_{k ' } \left ( \begin{array}{cc } j_1&j_2\\ j_3&j_4 \end{array } \right .
\bigg\vert \left .
\begin{array}{c } k\\ k ' \end{array } \right)_q \left| [ j_1,j_2,j_3,j_4];[k';\tfrac{1}{2}];00 \right\rangle,\ ] ] where the vectors on the left hand side do not contain any @xmath327label , namely @xmath328 $ ] @xmath329 $ ] , while on the right we have the combination @xmath330 $ ] @xmath331 $ ] .
the array @xmath332 in is the ( normalized ) @xmath4racah symbol of @xmath5 introduced in of section 4 ( with respect the notation used there , from now on we drop the subscript @xmath4 on vectors ) . ,
see and .,width=226 ] in the general case of @xmath2 incoming spin labels , the two combinatorially distinct bases which have to be involved are specializations of the vectors in to the two configurations depicted in fig . [ correlators ] .
the extension of the duality transformation to the case of an arbitrary ( even ) number of incoming spins can be done by resorting to two types of more complicated arrays , which can be represented as @xmath333 where the matrix indices to be involved in summations whenever transformations which generalize are implemented are listed in the right hand side of the arrays . + as happens in the standard racah wigner setting , it can be shown that each of these arrays , which represent the @xmath4deformed counterparts of @xmath0 @xmath149 coefficients , can be decomposed in terms of @xmath4racah transformations , _ cfr .
_ sections 3,4 of @xcite and @xcite . when the outgoing hilbert space is considered ( corresponding to the boundary @xmath334 endowed with the opposite orientation with respect to @xmath335 ) we have to introduce bra type bases which are dual ( and orthonormal ) with respect to the bases in . with an obvious choice of notations we set @xmath336;\,[\mathbf{k};\mathbf{h}]\,;0,0 \right| \left .
[ j_1, ... ,j_{2n}];[\mathbf{k}';\mathbf{h}']\,;0,0 \right\rangle = \delta_{\mathbf{k},\mathbf{k } ' } \delta_{\mathbf{h},\mathbf{h}'},\ ] ] where , as before , @xmath337 represent multi indices to be associated with the admissible configurations of binary coupled spins and there appear multiple kronecker deltas .
the discussion above was aimed to recognizing the crucial fact that the basis vectors @xmath338;[k_1, ... ,k_n;h_1, ... ,h_{n-2}];0,0\right\rangle\ ] ] are eigenfunctions of the odd braiding operators @xmath339 , while the basis vectors @xmath340;[r_1, ... ,r_{n-1};s_1, ... ,s_{n-1}];0,0\right\rangle\ ] ] are eigenfunctions of the even braiding operators @xmath341 .
+ the explicit expressions of the eigenvalues , in the odd and even case respectively , read @xmath342 @xmath343 here @xmath344 , @xmath4 is the deformation parameter , @xmath345 @xmath346 , @xmath347 is the quadratic casimir for the spin@xmath348 representation and @xmath349 denote the orientation of the strands labelled by @xmath41 and @xmath350 , respectively .
thus @xmath351 is the eigenvalue of the matrix which performs a right handed half
twist in contiguous strands with the same orientation , while @xmath352 is the eigenvalue of the matrix which performs a right handed half
twist in strands with opposite orientation .
the explicit expression of the formal expectation value given in above gives , after normalization according to the standard conventions ( _ cfr .
_ @xcite ) , the extended jones polynomial of the colored link @xmath1 associated with the braid , namely @xmath353 = \,\prod_{i = 1}^n \;\,[2j_i + 1]_q \;\,\times\ ] ] @xmath354;[0;0];0,0 \mathbf{\lvert } \,b\ , \bigl ( \begin{smallmatrix } \hat{j}_1 & \hat{j}_1^ * & \dots & \hat{j}_n & \hat{j}_n^ * \\
\hat{l}_1 & \hat{l}_1^ * & \dots & \hat{l}_n & \hat{l}_n^ * \end{smallmatrix } \bigr ) \mathbf{\rvert } \ , [ \hat{j}_1,\hat{j}_1^*, ... ,\hat{j}_n,\hat{j}_n^*];[0,0];0,0\rangle , \ ] ] where @xmath355_q$ ] is the @xmath4dimension of the @xmath5 irrep @xmath309 defined in section 4 .
the operator @xmath356 is expressed in terms of ( a finite sequence of ) the elementary braiding operators @xmath357 , suitably changed into the current odd ( even ) basis by acting with a @xmath4duality matrix , whenever an even ( odd ) vector of type ( , respectively ) is encountered . according to the expressions and for the admissible operations , the running variable of the polynomial is given by @xmath358 for any integer @xmath359 .
moreover , the above expectation value is to be interpreted as a trace over free spin labels .
this feature derives of course from the geometric construction of the plat presentation of the link @xmath1 outlined at the beginning of this section , since the colored oriented braid has to be ` closed up ' to get the associated link .
more precisely , corresponding strands must not only match pairwise with the correct orientations ( as is made manifest by our notation ) , but @xmath41 and @xmath360type labels have to be appropriately identified ( traced ) in pairs , namely the @xmath360 s in are not new independent labels .
+ coming back to the @xmath6dimensional picture , such trace procedure amounts to gluing back the two opposite oriented boundary @xmath288spheres , paying attention to the coloring of the punctures , to end up with the same @xmath6sphere we started from .
as already pointed out , the resulting link polynomial , arising as vacuum expectation value of the composite wilson loop operator in the quantum @xmath0 cs theory for @xmath249 @xmath361 , is automatically an invariant of regular isotopy ( _ cfr . _ the remarks in the footnote of section 5.2 ) .
this long technical discussion about the derivation of the extended jones polynomial is nothing but the necessary premise to address the main issue of the present paper , namely the analysis of the connections among the theory of formal languages ( section 2 ) and the spin network computational scheme ( sections 3 and 4 ) , on the one hand , and braid group and links invariants on the other .
we are going to interpret the results established so far in terms of ` processing of words ' , written in the alphabet given by the generators of the braid group , on a quantum automaton in such a way that the expectation value associated with the ` evolution ' of the automaton is precisely the extended jones polynomial . the quantum automaton in question
will in turn correspond to a path in the @xmath4braided spin
network computational graph . in order to comply with the requirements for a finite states quantum automaton described in section 2
, we have to provide explicitly the 5-tuple @xmath362 , where @xmath363 represents a set of acceptable final states .
now @xmath44 is a finite set of states belonging to the hilbert spaces of the tensor algebra of @xmath5 described in section 5.1 , whose combinatorial content was depicted in fig .
[ correlators ] .
labels of @xmath12 irreps in this quantum group are associated with the strands of the plat presentation of the link @xmath1 , and @xmath364 is the braid group to be selected .
@xmath45 is the alphabet made of the @xmath365 generators of @xmath364 : each generator ( and its inverse ) represents a letter of the alphabet , and words are written as composition of these elementary braids .
the function @xmath54 denotes a set of unitary matrices defining the transition rules and there is of course one matrix for each letter of the alphabet since we are linearly representing the braid group in the tensor algebra @xmath5 . as initial state
@xmath52 we pick up one particular binary coupled state , namely @xmath366;[0;0];0,0\right\rangle , \ ] ] where , with respect to the generic expression for an odd basis vector given in , we choose @xmath367 $ ] @xmath368 $ ] , namely we select a ` multi singlet ' intermediate state
the set of final states @xmath363 is constrained by the topological properties of the plat presentation , namely final states may differ from @xmath369 by a permutation on the string @xmath370 .
thus we can actually build up a family of @xmath371 automata out of one initial state @xmath369 . by acting with the symmetric group on the binary parenthesization structure of we
may get , for instance , the singlet final state @xmath372;[0,0];0,0 \right\rangle.\ ] ] the unitary transition rules codified in the set @xmath54 are : * if the automaton is in an even ( odd ) state and it reads an even ( odd ) braid generator , then the system evolves with the r matrix associated to the proper braid generator , @xmath373 or @xmath341 , see , and ; * if the automaton is in an odd ( even ) state and it reads an even ( odd ) braid generator , then the system evolves with the proper duality transformation ( see ) to update the actual state into the configuration consistent with the parity of the given braid generator . as pointed out before , such a transformation can be splitted into a finite sequence of elementary duality ( @xmath4racah ) transformations of the type .
once a final state @xmath374 has been selected ( the right permutation can be singled out in a fast way even by a classical machine ) the evaluation of the polynomial is carried out by the automaton in a number of steps linear in the length @xmath375 of the ` word ' @xmath313 . since @xmath375 is the sum of the numbers of elementary braiding operators and @xmath4duality transformations entering the explicit expression of @xmath313 , the length of the word is bounded from above by a linear function of the number of crossings of the plat presentation of the associated link @xmath1 . on the other hand , in the worst case we have to perform one duality transformation before applying each elementary braiding operator @xmath376 .
this happens , for instance , in the evaluation of the jones polynomial of the trefoil knot illustrated in the appendix .
this latter remark lead us to conclude that the time complexity function ( more precisely , the number of computational steps in our automaton calculation ) equals the length of @xmath313 and is bounded according to @xmath377 is a positive number depending on the number of generators of the braid group @xmath364 and @xmath378 is the number of crossings in the plat presentation of the link @xmath1 . + we can estimate @xmath379 by observing that any such automaton can be uniquely associated with a particular path @xmath380 in the @xmath4braided spin
network computational space @xmath381 for @xmath382 .
in particular , the maximum number of elementary @xmath4racah transforms entering a duality matrix of type must coincide with the number of racah transforms entering into one ( classical ) @xmath149 symbol since the combinatorics of such operations is manifestly the same .
hence we may exploit results from graph theory which tell us that the rotation graph @xmath383 obtained from @xmath384 by ignoring twists ( or braidings) has a diameter of the order @xmath385 ( the diameter is defined as the maximum over the set of distances between pairs of vertices , where the distance is the minimum number of edges connecting two given vertices ) ( _ cfr .
_ @xcite and appendix a of @xcite for a complete discussion of the spin network combinatorics ) .
clearly the ` distance ' between the current basis and the eigenbasis with the right parity can not exceed the above maximum distance , and consequently the factor @xmath379 in grows as @xmath386 namely polynomially in the number of strands of the link . a deeper connection with the @xmath4braided spin
network computational scheme comes out however when we recognize that the expectation value representing the extended jones polynomial is not only the quantum transition amplitude of a finite states
automaton , as pointed out before , but complies also with the expectation value to be associated with a path @xmath380 in the @xmath4version of the spin network computational space . in this new perspective
, what we are really doing is to ` encode ' the combinatorial structure underlying quantum @xmath0 chern
simons field theory ( and the associated wz boundary theory ) at some fixed level @xmath295 into the abstract @xmath4braided @xmath0spin
network for @xmath387 .
this does not mean , of course , that we have set up a quantum algorithm for the extended jones polynomial in the strict sense , since the encoding map could not be ` efficiently ' represented ( nor efficiently approximated ) with respect to standard models of computation ( boolean circuits , turing machines ) .
we provide , however , a quantum system whose evolution can be controlled in such a way that its probability amplitudes give the desired link polynomials .
the crucial issue of constructing a _ bona fide _ quantum algorithm is under investigation .
it will require in particular : * i * ) ( efficient ) encoding schemes for binary coupled states ; * ii * ) ( efficient ) algorithms to evaluate ( or approximate ) the basic operations , namely the racah transform and its associated @xmath155symbol for arbitrary entries on the one hand , and the elementary braiding operators on the other . as a final remark we notice that the field theoretical approach gives us automatically the rate of growth of the absolute value of extended jones polynomial with respect to the chern
simons coupling constant @xmath388 .
the absolute value of the reshetikhin
turaev @xcite quantum invariants of @xmath6manifolds , @xmath389 ( which are linear combinations of colored polynomials associated with surgery framed links @xmath390 ) are estimated to grow as @xmath391 , where the exponent @xmath263 is bounded from above by some simple function ( depending on the gauge group @xmath392 ) of the heegaard genus of the manifold ( _ cfr .
_ @xcite , ch . 7 ) . 99 a. marzuoli and m. rasetti ( 2002 ) , _ spin network quantum simulator _ ,
a306 , pp .
a. marzuoli and m. rasetti ( 2005 ) , _ computing spin network _ ,
345407 .
freedman , a. kitaev and z. wang ( 2002 ) , _ simulation of topological field theories by quantum computers _ ,
587603 .
freedman , a. kitaev and z. wang ( 2002 ) , _ a modular functor which is universal for quantum computation _
227 , pp . 605622 .
freedman , a. kitaev , m. larsen and z. wang ( 2002 ) , _ topological quantum computation _
biedenharn and j.d .
louck ( 1981 ) _ the racah wigner algebra in quantum theory _
( topic 12 . coupling of n angular momenta : recoupling theory ) , encyclopedia of mathematics and its applications , vol 9 , g c .
, addison wesley publ .
co. ( reading ma ) .
jones ( 1985 ) _ a polynomial invariant for knots via von neumann algebras _ , bull .
103111 . f.y .
wu ( 1992 ) _ knot theory and statistical mechanics _
64 , n.4 , pp .
10991131 .
m. f. atiyah ( 1989 ) _ topological quantum field theories _ , publ .
ihes 68 , pp .
175186 . e. witten ( 1989 ) _ quantum field theory and the jones polynomial _
121 , pp . 351399 .
n. reshetikhin and v. g. turaev ( 1991 ) _ invariants of 3manifolds via link polynomials and quantum groups _ , invent .
547597 .
r. kirby and p. melvin ( 1991 ) _ the 3manifold invariants of witten and reshetikhin turaev for @xmath393 _ invent . math .
473545 .
f. jaeger , d.l .
vertigan and d.j.a .
welsh ( 1990 ) _ on the computational complexity of the jones and tutte polynomials _ , math .
d. aharonov , v. jones and z. landau ( 2005 ) _ a polynomial quantum algorithm for approximating the jones polynomial _ , quant - ph/0511096 .
j.s . birman and t.e .
brendle ( 2004 ) _ braids : a survey _ , math./ gt/ 0409205 .
i. anshel , m. anshel , and d. goldfeld ( 1999 ) _ an algebraic method for public key cryptography _ , math .
lett . 6 , pp .
287291 .
s. garnerone , a. marzuoli and m. rasetti ( 2005 ) _ spin networks , quantum automata and link invariants _ , gr - qc/0511161 .
p. freyd , d. yetter , j. hoste , w. lickorish , k. millett and a. ocneanu ( 1985 ) _ a new polynomial invariants of knots and links _ , bull .
183312 .
j. e. hopcroft and j. d. ullman ( 1979 ) _ introduction to automata theory , languages and computation _ , addison wesley ( reading ma ) .
a. ambainis , j. watrous ( 2002 ) _ two - way finite automata with quantum and classical states _ theoret .
299 - 311 . c. moore and j.p .
crutchfield ( 2000 ) _ quantum automata and quantum grammars _ theor .
275306 .
j. kempe , d. bacon , d.a .
lidar and k.b .
whaley ( 2001 ) _ theory of decoherence free fault
tolerant universal quantum computation _ phys .
a 63 , 0423071/29 .
v. aquilanti and c. coletti ( 2001 ) _ 3nj symbols and harmonic superposition coefficients : an icosaedral abacus _ chem .
601611 .
c.kassel ( 1994 ) _ quantum groups _ grad .
texts in math .
155 , springer
verlag ( new york
berlin ) a.n .
kirillov and n.y .
reshetikhin ( 1988 ) _ representations of the algebra @xmath394 , @xmath4orthogonal polynomials and invariants of links _ , in ` infinite dimensional lie algebras and groups ' , ed .
kac , adv .
ser . in math .
7 , world scientific ( singapore ) , pp . 285339 .
e. guadagnini ( 1993 ) _ the link invariants of the chern
simons field theory _ ,
de gruyter ( berlin ) .
kauffman ( 1991 ) _ knots and physics _ , world scientific ( singapore ) .
j. alexander ( 1923 ) _ a lemma on a system of knotted curves _ , proc .
usa , 9 , pp . 9395 .
birman ( 1974 ) _ braids , links and mapping class groups _
, princeton univ . press ( princeton nj ) .
s. carlip ( 1998 ) _ quantum gravity in 2 + 1 dimensions _ , cambridge university press ( cambridge ) .
kaul ( 1994 ) _ chern simons theory , coloured oriented braids and link invariants _ , commun .
289319 .
v. fack , s. lievens and j. van der jeugt ( 2002 ) _ on the diameter of the rotation graph of binary coupling trees _ ,
118 . t. ohtsuki ed .
( 2002 ) _ problems on invariants of knots and 3manifolds _ , geometry and topology monographs , vol .
a simple application of the procedure described in section 6 for the evaluation of quantum link invariants is the explicit computation of the jones polynomial for the plat presentation of the trefoil knot @xmath395 depicted in fig .
[ trefoilex ] .
the four strands are labelled by a same @xmath396 , together with its opposite @xmath397 , from left to right . accordingly , the initial and final states to be associated with the quantum automaton are @xmath398;[0,0;-];0,0 \right\rangle\ ] ] and @xmath399;[0,0;-];0,0 \right\rangle\ ] ] respectively , and they comply with the prescription of being odd vectors , see . from the picture we easily recognize that the operator to be employed is @xmath400 .
since we are interested in the evaluation of the jones polynomial we set from now on @xmath401 .
moreover , in order to apply the even braiding operator @xmath402 we have to perform preliminarly a duality transformation on the odd vector @xmath403;[0;-];00 \right\rangle = \sum\limits_{l=0}^1 \left ( \begin{array}{cc } \tfrac{1}{2}&\tfrac{1}{2}\\ \tfrac{1}{2}&\tfrac{1}{2 } \end{array } \right .
\bigg\vert \left .
\begin{array}{c } l\\ 0 \end{array } \right)_q \left| [ \tfrac{1}{2},\tfrac{1}{2},\tfrac{1}{2},\tfrac{1}{2}];[l;-];00 \right\rangle,\ ] ] which converts the initial state of the automaton into eigenvectors of the braid generator @xmath402 .
the application of @xmath404 gives @xmath405;[0;-];00 \right\rangle=\sum\limits_{l=0}^1 ( \lambda_l^{(+)})^3 \left ( \begin{array}{cc } \tfrac{1}{2}&\tfrac{1}{2}\\ \tfrac{1}{2}&\tfrac{1}{2}\end{array } \right .
\bigg\vert \left .
\begin{array}{c } l\\ 0 \end{array } \right)_q \left| [ \tfrac{1}{2},\tfrac{1}{2},\tfrac{1}{2},\tfrac{1}{2}];[l;-];00 \right\rangle,\ ] ] where there appears the cube of the eigenvalue @xmath406 defined in . according to the expression of the extended jones polynomial given in and taking into account
, we get @xmath407_q\;\langle\left[\tfrac{1}{2},\tfrac{1}{2},\tfrac{1}{2},\tfrac{1}{2}];[0;-];00\right|\,(b_2)^3 \left| [ \tfrac{1}{2},\tfrac{1}{2},\tfrac{1}{2},\tfrac{1}{2}];[0;-];00 \right\rangle,\ ] ] which , by using the orthogonality relations of the duality matrices ( kaul ) , amounts to @xmath408_q\sum\limits_{l=0}^1 \lambda_l^{(+)^3}=\frac{-1+q+q^3}{q^4}\ ] ] as required . | the spin
network quantum simulator model , which essentially encodes the ( quantum deformed ) @xmath0 racah wigner tensor algebra , is particularly suitable to address problems arising in low dimensional topology and group theory . in this combinatorial framework
we implement families of finite states and discrete time quantum automata capable of accepting the language generated by the braid group , and whose transition amplitudes are colored jones polynomials .
the automaton calculation of the polynomial of ( the plat closure of ) a link @xmath1 on @xmath2 strands at any fixed root of unity is shown to be bounded from above by a linear function of the number of crossings of the link , on the one hand , and polynomially bounded in terms of the braid index @xmath2 , on the other .
the growth rate of the time complexity function in terms of the integer @xmath3 appearing in the root of unity @xmath4 can be estimated to be ( polynomially ) bounded by resorting to the field theoretical background given by the chern simons theory .
* quantum automata , braid group and link polynomials * + dipartimento di fisica , politecnico di torino , + corso duca degli abruzzi 24 , 10129 torino ( italy ) + e - mail : [email protected] + + dipartimento di fisica nucleare e teorica , universit degli studi di pavia and istituto nazionale di fisica nucleare , sezione di pavia , + via a. bassi 6 , 27100 pavia ( italy ) + e - mail : [email protected] + + dipartimento di fisica , politecnico di torino , + corso duca degli abruzzi 24 , 10129 torino ( italy ) + e - mail : [email protected] + link invariants ; braid group representations ; chern simons theory ; quantum automata ; racah wigner algebra ; spin network simulator ; topological quantum computation ; @xmath5 representation theory . |
microtubules are polar linear polymers that perform major organizational tasks in living cells @xcite . through a unique feature of microtubule assembly ,
termed dynamic instability @xcite , they function as molecular machines @xcite that move cellular structures during processes such as cell reproduction @xcite .
a surprising feature of microtubules is that they remain out of equilibrium under fixed external conditions and can undergo alternating periods of rapid growth and even more rapid shrinking ( fig .
[ l - vs - t ] ) .
these sudden polymerization changes are driven by the interplay between several fundamental processes .
microtubules grow by the attachment of guanosine triphosphate tubulin complexes ( gtp ) at one end @xcite .
structural studies indicate that the end of a microtubule must consist of a `` cap '' of consecutive gtp monomers @xcite for growth to continue @xcite .
once polymerized , the gtp of this complex can irreversibly hydrolyze into guanosine diphosphate ( gdp ) .
if all the monomers in the cap convert to gdp , the microtubule is destabilized and rapid shrinkage ensues by the detachment of gdp tubulin units .
the competition between gtp attachment and hydrolysis from gtp to gdp is believed to lead to the dynamic instability in which the gtp cap hydrolyzes to gdp and then the microtubule rapidly depolymerizes .
the stochastic attachment of gtp can , however , lead to a rescue to the growing phase before the microtubule length shrinks to zero @xcite .
the origin of this dynamic instability has been actively investigated .
one avenue of theoretical work on this dynamical instability is based on models of mechanical stability @xcite .
for example , a detailed stochastic model of a microtubule that includes all the thirteen constituent protofilaments has been investigated in ref . @xcite . by using model parameters that were inferred from equilibrium statistical physics , vanburen et al .
@xcite found some characteristics of microtubule evolution that agreed with experimental data @xcite .
the disadvantage of this detailed modeling , however , is its complexity , so that it is generally not possible to develop an intuitive understanding of microtubule evolution .
another approach for modeling the dynamics of microtubules is based on effective two - state models that describe the dynamics in terms of a switching between a growing and a shrinking state @xcite .
the essence of many of these models is that a microtubule exists either in a growing phase ( where a gtp cap exists at the end of the microtubule ) or a shrinking phase ( without a gtp cap ) , and that there are stochastic transitions between these two states . by tuning parameters appropriately , it is possible to reproduce the phase changes between the growing and shrinking phases of microtubules that have been observed experimentally @xcite . while the two - state model has the advantage of having only a few parameters , a constant rate of switching between a growing and shrinking microtubule
is built into the model .
thus switching models can not account for the stochastic avalanches and catastrophes that occur in real microtubules .
on the other hand , a minimalist model of microtubule dynamics has been proposed and investigated by flyvbjerg et al .
@xcite . in their model
, they dispense with attempts to capture all of the myriad of experimental parameters within a detailed model , but instead constructed an effective continuous theory to describe microtubule dynamics . their goal was to construct an effective theory that contained as few details as possible .
as stated in ref .
@xcite , they envision that their effective theory should be derivable from a fundamental , microscopic theory and its parameters .
this minimalist modeling is the approach that we adopt in the present work .
we investigate a recently introduced @xcite kinetic model that accounts for many aspects of microtubule evolution .
our main result is that only a few essential parameters with simple physical interpretations are needed to describe the rich features of microtubule growth , catastrophes , and rescues @xcite .
we treat a microtubule as a linear polymer that consists of gtp or gdp monomers that we denote as @xmath0 and @xmath1 , respectively . to emphasize this connection between chemistry and the model , we will write the former as gtp@xmath2 and the latter as gdp@xmath3 .
the state of a microtubule evolves due to the following three processes : 1 .
attachment : a microtubule grows by attachment of a guanosine triphosphate ( gtp@xmath2 ) monomer .
@xmath4 2 .
conversion : once part of the microtubule , each gtp@xmath2 can independently convert by hydrolysis to a guanosine diphosphate ( gdp@xmath3 ) .
@xmath5 3 .
detachment : a microtubule shrinks due to detachment of a gdp@xmath3 monomer _ only from the end of the microtubule_. @xmath6 and @xmath7 denote the terminal and the active end of the microtubule .
it is worth mentioning that these steps are similar to those in a recently - introduced model of dna sequence evolution @xcite , and that some of the results about the structure of dna sequences seem to be related to our results about island size distributions in microtubules . and attachment rates : ( a ) @xmath8 where the microtubule generally remains short , ( b ) @xmath9 , where the length fluctuates strongly , and ( c ) @xmath10 , where the microtubule grows nearly steadily.,title="fig:",scaledwidth=39.0% ] and attachment rates : ( a ) @xmath8 where the microtubule generally remains short , ( b ) @xmath9 , where the length fluctuates strongly , and ( c ) @xmath10 , where the microtubule grows nearly steadily.,title="fig:",scaledwidth=39.0% ] and attachment rates : ( a ) @xmath8 where the microtubule generally remains short , ( b ) @xmath9 , where the length fluctuates strongly , and ( c ) @xmath10 , where the microtubule grows nearly steadily.,title="fig:",scaledwidth=39.0% ] generically , the @xmath11 phase space separates into a region where the microtubule grows ( on average ) with a certain rate @xmath12 , and a compact phase where the average microtubule length is finite .
these two phases are separated by a phase boundary @xmath13 along which the growth rate @xmath12 vanishes . while the behavior of a microtubule for general parameter values is of interest , we will primarily focus on extreme values of the governing parameters where we can obtain a detailed statistical characterization of the microtubule structure . for certain properties , such as the shape of the phase diagram , we will also present results of numerical simulations of the model . in sec .
[ growth ] , we study the evolution of a microtubule under unrestricted growth conditions namely no detachment and an attachment rate that does not depend on the identity of the last monomer .
our results here are relevant to understanding the distribution of cap length and the diffusion coefficient of the tip of the microtubule in the growth phase .
the predictions of the model in this limit could also be useful in understanding the binding pattern of proteins to microtubules @xcite .
since proteins are important regulatory factors in microtubule polymerization , these results could prove useful in interpreting the effects of proteins on microtubule growth . by a master equation approach
, we will determine both the number of gtp@xmath2 monomers on a microtubule , as well as the length distributions of gtp@xmath2 and gdp@xmath3 islands ( fig .
[ cartoon ] ) .
many of these analytical predictions are verified by numerical simulations . in sec . [ constrained ] , we extend our approach to the case of constrained growth , @xmath14 , in which microtubule growth depends on whether the last monomer is a gtp@xmath2 or a gdp@xmath3 . in sec .
[ instant ] , we investigate the phenomenon of `` catastrophe '' for infinite detachment rate @xmath15 , in which a microtubule shrinks to zero length when all of its constituent monomers convert to gdp@xmath3 .
we derive the asymptotic behavior of the catastrophe probability by expressing it as an infinite product and recognizing the connection of this product with modular functions .
we also determine the asymptotic behavior of the size distribution of avalanches , namely , sequences of consecutive gdp@xmath3 detachment events . finally , in sec .
[ general ] , we discuss the behavior of a microtubule for general parameter values through a combination of numerical and analytic results . here numerical simulations are useful to extract quantitative results for parameter values that are note amenable to theoretical analysis .
several calculational details are given in the appendixes .
we define unrestricted growth as the limit of detachment rate @xmath16 , so that a microtubule grows without bound . here
we consider the special case where the attachment rate does not depend on the identity of the last monomer ; that is , the limit of @xmath17 , where the attachment is unconstrained . because of the latter condition , the number @xmath18 of gtp@xmath2 monomers decouples from the number of gdp@xmath3 , a greatly simplifying feature
. are shown dark ( blue ) and regions of gdp@xmath3 are light ( yellow ) .
the gtp@xmath2 regions get shorter further from the tip that advances as @xmath19 , while the gdp@xmath3 regions get longer.,scaledwidth=40.0% ] the average number of gtp@xmath2 monomers evolves as @xmath20 the gain term accounts for the adsorption of a gtp@xmath2 at rate @xmath21 , while the loss term accounts for the conversion events gtp@xmath2 @xmath22 gdp@xmath3 , each of which occurs with rate 1 .
thus @xmath23 approaches its stationary value of @xmath21 exponentially quickly , @xmath24 more generally , consider the probability @xmath25 that there are @xmath18 gtp@xmath2 monomers at time @xmath26 .
this probability evolves according to @xmath27 the loss term @xmath28 accounts for conversion events gtp@xmath2 @xmath22 gdp@xmath3 that occur with total rate @xmath18 , and the attachment of a gtp@xmath2 at the end of the microtubule of length @xmath18 with rate @xmath21 .
the gain terms can be explained similarly . in terms of generating function @xmath29 , eq .
can be recast as the differential equation @xmath30 introducing @xmath31 and @xmath32 , we transform eq . into the wave equation @xmath33 whose solution is an arbitrary function of @xmath34 or , equivalently ,
@xmath35 . if the system initially is a microtubule of zero length , @xmath36 , the initial generating function @xmath37 , so that @xmath38 .
thus for @xmath39 , @xmath40 , from which @xmath41 expanding this expression in a power series in @xmath42 , the probability for the system to contain @xmath18 gtp@xmath2 monomers is the time - dependent poisson distribution @xmath43^n}{n!}\,e^{-\lambda(1-e^{-t})}.\ ] ] from this result , the mean number of gtp@xmath2 monomers and its variance are @xmath44 the length distribution @xmath45 of the microtubule evolves according to the master equation @xmath46\ ] ] for the initial condition @xmath47 , the solution is again the poisson distribution @xmath48 from which the average and the variance are @xmath49 thus the growth rate of the microtubule and the diffusion coefficient of the tip are @xmath50 a more comprehensive description is provided by the joint distribution @xmath51 that a microtubule has length @xmath52 and contains @xmath18 gtp@xmath2 monomers at time @xmath26 .
this distribution evolves as @xmath53 this joint distribution does _ not _ factorize , that is , @xmath54 , because @xmath55 . to demonstrate this inequality
, we compute @xmath56 by multiplying eq . by @xmath57 and summing over all @xmath58 to give @xmath59 using eqs . and for @xmath23 and @xmath60 and integrating we obtain @xmath61 using eq .
, we have @xmath62 , so that the joint distribution is factorizable asymptotically . for completeness , we give the full solution for @xmath51 in appendix [ app - p ] . because of the conversion process gtp@xmath2
@xmath22 gdp@xmath3 , the tip of the microtubule is comprised predominantly of gtp@xmath2 , while the tail exclusively consists of gdp@xmath3 .
the region from the tip until the first gdp@xmath3 monomer is known as the _ cap _ ( fig .
[ cap - fig ] ) and it plays a fundamental role in microtubule function .
we now use the master equation approach to determine the cap length distribution .
cap of length 4 , then three gtp@xmath2 islands of lengths 1 , 3 , and 2 , and three gdp@xmath3 islands of lengths 3 , 2 , and a `` tail '' of length 5 .
the rest of the microtubule consists of gdp@xmath3.,scaledwidth=40.0% ] , @xmath63 , and @xmath17 compared to the theoretical prediction of eq . .
] consider a cap of length @xmath64 .
its length increases by 1 due to the attachment of a gtp@xmath2 at rate @xmath21 .
the conversion of any gtp@xmath2 into a gdp@xmath3 at rate 1 reduces the cap length from @xmath64 to an arbitrary value @xmath65 .
these processes lead to the following master equation for the probability @xmath66 that the cap length equals @xmath64 : @xmath67 equation is also valid for @xmath68 if we set @xmath69 .
note that @xmath70 is the probability for a cap of length zero .
we now solve for the stationary distribution by summing the first @xmath71 of eqs .
with @xmath72 set to zero to obtain @xmath73 the cumulative distribution , @xmath74 , thus satisfies the recursion @xmath75
using the normalization @xmath76 and iterating , we obtain the solution in terms of the gamma function @xcite : @xmath77 hence the cap length distribution is @xmath78 and the first few terms are @xmath79 results of direct simulation of the kinetic model are compared to the predicted cap length distribution ( eq . ) in fig .
[ capsize ] .
because of the finite length of the simulated microtubule , there is a largest cap length that is accessible numerically . aside from this limitation ,
the simulations results are in agreement with theoretical predictions .
it is instructive to determine the dependence of the average cap length @xmath80 on @xmath21 .
using @xmath81 , we rearrange @xmath82 into @xmath83 using , the above sum may be written in terms of the confluent hypergeometric series @xcite : @xmath84 we now determine the asymptotic behavior of @xmath82 by using the integral representation @xmath85 to recast the average cap length as @xmath86 where @xmath87 is the ( lower ) incomplete gamma function . in the realistic limit of @xmath88
, we use the large @xmath21 asymptotics @xmath89 to give @xmath90 thus even though the number of gtp@xmath2 monomers equals @xmath21 , only @xmath91 of them comprise the microtubule cap , as qualitatively illustrated in fig .
[ cartoon ] .
note that the average cap length is proportional to the square - root of the velocity ; essentially the same result was obtained from the coarse - grained theory of flyvbjerg et al .
@xcite . at a finer level of resolution
, we determine the distribution of island sizes at the tip of a microtubule ( fig .
[ cap - fig ] ) .
a simple characteristic of this population is the average number @xmath92 of gtp@xmath2 islands .
if all gtp@xmath2 islands were approximately as long as the cap , we would have @xmath93 .
as we shall see , however , @xmath92 scales linearly with @xmath21 because most islands are short .
a similar dichotomy arises for negative islands . to write the master equation for the average number of islands , note that the conversion gtp@xmath2 @xmath22 gdp@xmath3 eliminates islands of size 1
additionally , an island of size @xmath94 splits into two daughter islands , and hence the number of islands increases by one , if conversion occurs at any one of the @xmath95 in the interior of an island as illustrated below : @xmath96 conversely , if the cap has length 0 , attachment creates a new cap of length 1 at rate @xmath21 .
the net result of these processes is encoded in the rate equation @xmath97 with @xmath98 the average number of gtp@xmath2 islands of size @xmath64 .
we now use the sum rules @xmath99 and @xmath100 to recast as @xmath101 from which the steady - state average number of islands is @xmath102 for large @xmath21 , the number of islands approaches @xmath103 , while the number of gtp@xmath2 monomers equals @xmath21 .
thus the typical island size is 2 . nevertheless , as we now show , the gtp@xmath2 and gdp@xmath3 island distributions actually have power - law tails , with different exponents for each species .
the gtp@xmath2 island size distribution evolves according to the master equation @xmath104 this equation is similar in spirit to eq . for the cap length distribution . as a useful self - consistency check , the sum of eqs .
gives , while multiplying by @xmath64 and summing over all @xmath105 gives eq . .
, @xmath106 , and @xmath17 for the size distribution of positive islands , @xmath107 .
the solid line is the theoretical prediction of eq . .
] the stationary distribution satisfies @xmath108 using @xmath109 , we transform at @xmath110 to @xmath111 similarly , using @xmath112 we transform at @xmath113 to @xmath114 thus using and we obtain @xmath115 the same procedure gives @xmath98 for larger @xmath64 .
since the @xmath98 represent the _ average _ number of islands of size @xmath64 , they become meaningful only for @xmath116 where an appreciable number of such islands exist . in this limit , we write @xmath98 more compactly by first rearranging into the equivalent form @xmath117 we then use and the asymptotic properties of the gamma function to find that the right - hand side of eq .
is @xmath118 and is therefore negligible in the large-@xmath21 limit .
thus reduces to @xmath119 , with solution @xmath120 $ ] .
we find the amplitude @xmath121 by matching with the exact result , eq . , to give @xmath122 for large @xmath21 .
the final result is @xmath123 in the large @xmath21 limit , @xmath124 , and the above result can be re - written as @xmath125 remarkably , the size distribution of the positive islands is _ identical _ to the degree distribution of a growing network with strictly linear preferential attachment @xcite .
the results for the island size distribution in the large @xmath21 limit are compared to simulation results in fig .
[ islandsize ] .
these asymptotic results are expected to apply to island sizes @xmath64 much smaller than the size of the cap which scales as @xmath126 .
the distributions obtained from the numerical simulations should then obey the theoretical form but with a finite - size cutoff .
the results in fig .
[ islandsize ] are consistent with this picture but , interestingly , the numerical distribution rises above the theoretical curve before falling sharply below it .
this anomaly occurs in many heterogeneous growing network models , and it can be fully characterized in terms of finite - size effects @xcite . when @xmath116 , both the length of the cap and the length of the region that contains gtp@xmath2 become large . in this limit , the results from the discrete master equation can be expressed much more elegantly and completely by a continuum approach .
the fundamental feature is that the conversion process gtp@xmath2
@xmath22 gdp@xmath3 occurs independently for each monomer .
since the residence time of each monomer increases linearly with distance from the tip , the probability that a gtp@xmath2 does not convert decays exponentially with distance from the tip .
this fact alone is sufficient to derive all the island distributions .
consider first the length @xmath127 of the populated region ( fig .
[ cap - fig ] ) . for a gtp@xmath2 that is a distance @xmath128 from the tip ,
its residence time is @xmath129 in the limit of large @xmath21 .
thus the probability that this gtp@xmath2 does not convert is @xmath130 .
we thus estimate @xmath127 from the extremal criterion @xcite @xmath131 that merely states that there is of the order of a single gtp@xmath2 further than a distance @xmath127 from the tip .
since @xmath132 when @xmath21 is large , the length of the active region scales as @xmath133 the probability that the cap has length @xmath64 is given by @xmath134 the product ensures that all monomers between the tip and a distance @xmath64 from the tip are gtp@xmath2 , while the prefactor gives the probability that a monomer is a distance @xmath135 from the tip is a gdp@xmath3 . expanding the prefactor for large @xmath21 and rewriting the product as the sum in the exponent , we obtain @xmath136 a result that also can be obtained by taking the large-@xmath21 limit of the exact result for @xmath66 given in eq . .
similarly , the probability to find a positive island of length @xmath64 that occupies sites @xmath137 is @xmath138 the two prefactors ensure that sites @xmath128 and @xmath139 consist of gdp@xmath3 , while the product ensures that all sites between @xmath140 and @xmath141 are gtp@xmath2 .
most islands are far from the tip and they are relatively short , @xmath142 , so that simplifies to @xmath143 the total number of islands of length @xmath64 is obtained by summing the island density over all @xmath128 .
since @xmath88 , we replace the summation by integration and obtain @xmath144 the power law tail agrees with eq . , whose derivation explicitly invoked the @xmath116 limit .
we can also obtain the density of negative islands in this continuum description , a result that seems impossible to derive by a microscopic master equation description . in parallel with ,
the density of negative islands of length @xmath142 with one end at @xmath128 is given by @xmath145 and the total number of negative islands of length @xmath64 is @xmath146 again , we find a power - law tail for the gdp@xmath3 island size distribution , but with exponent 2 .
the total number of gdp@xmath3 monomers within the populated zone is then @xmath147 .
while this sum formally diverges , we use the upper size cutoff , @xmath148 to obtain @xmath149 .
since the length of the populated zone @xmath150,this zone therefore predominantly consists of gdp@xmath3 islands . in analogy with the cap , consider now the `` tail''the last island of gdp@xmath3 within the populated zone ( see fig . [ cap - fig ] ) .
the probability @xmath151 that it has length @xmath64 is @xmath152 using we simplify the above expression to @xmath153 hence the average length of the tail is @xmath154 which is much longer ( on average ) than the cap .
when @xmath14 , the rate of attachment depends on the state of the tip of the microtubule
attachment to a gtp@xmath2 occurs with rate @xmath21 while attachment to a gdp@xmath3 occurs with rate @xmath155 .
while this state dependence makes the master equation description for the properties of the tubule more complicated , qualitative features about the structure of the populated zone are the same as those in the case @xmath17 . in this section ,
we outline some of the basic features of the populated zone when @xmath14 , but we still keep @xmath16 .
the average number of gtp@xmath2 monomers now evolves according to the rate equation @xmath156 which should be compared to the rate equation eq . for the case @xmath17 .
the loss term on the right - hand side describes the conversion gtp@xmath157gdp@xmath3 , while the remaining terms represent gain due to attachment to a gtp@xmath2 with rate @xmath21 and to a gdp@xmath3 with rate @xmath155 . here
@xmath158 is the probability for a cap of length zero , that is , the last site is a gdp@xmath3 .
the stationary solution to is @xmath159 so we need to determine @xmath158 . by extending eq . to the case @xmath160
, we then find that @xmath158 is governed by the rate equation @xmath161 thus asymptotically @xmath162 and substituting into , the average number of gtp@xmath2 monomers is @xmath163 more generally , we can determine the distribution of the number of gtp@xmath2 monomers ; the details of this calculation are presented in appendix [ app ] .
the growth rate of a microtubule equals @xmath155 when the cap length is zero and to @xmath21 otherwise .
therefore @xmath164 for the diffusion coefficient of the tip of a microtubule , we need its mean - square position . as in the case @xmath17 , it is convenient to determine the probability distribution for the tip position
. thus we introduce @xmath165 and @xmath166 , the probabilities that the microtubule length equals @xmath52 and the last monomer is a gtp@xmath2 or a gdp@xmath3 , respectively . these probabilities satisfy @xmath167 summing these equations , the length distribution @xmath168 satisfies @xmath169 the state of the last monomer does not depend on the microtubule length @xmath52 for large @xmath52 .
thus asymptotically @xmath170 substituting into we obtain a master equation for the tubule length distribution of the same form as eq . , but with prefactor @xmath171 given by instead of @xmath21 . as a result of this correspondence , we infer that the diffusion coefficient is one - half of the growth rate , @xmath172 for large @xmath21 both the growth rate of the tip of the microtubule and its diffusion coefficient approach the corresponding expressions in eq . for the case @xmath17 .
the master equations for the cap length distribution are the same as in the @xmath17 case when @xmath173 .
the master equations for @xmath68 and @xmath110 are slightly changed to account for the different rates at which attachment occurs at a gdp@xmath3 monomer : @xmath174 solving iteratively we recover @xmath175 and also obtain @xmath176 _ etc . _
the general solution for the @xmath66 is found by the same method as in sec .
[ sec : cld ] to be @xmath177 which are valid for @xmath105 . with this length distribution ,
the average cap length is then @xmath178 and the sum can again be expressed in terms of hypergeometric series as in eq . .
rather than following this path , we focus on the most interesting limit of large @xmath21
. then the cap length distribution approaches to the previous solution for the case @xmath17 and the mean length reduces to , independent of @xmath179 .
for the distribution of island sizes , the master equation remains the same as in the @xmath17 case when @xmath173 .
however , when @xmath110 , the master equation becomes @xmath180 in the stationary state . then the average number of islands and the average number of islands of size 1 are found from @xmath181 using @xmath162 and eqs .
and we obtain @xmath182 \end{aligned}\ ] ] again , in the limit of large @xmath21 , the average island size distribution reduces to our previously - quoted results in or equivalently .
the leading behavior in the @xmath116 limit is again independent of @xmath179 .
for @xmath183 , a microtubule can recede if its tip consists of gdp@xmath3 .
the competition between this recession and growth by the attachment of gtp@xmath2 leads to a rich dynamics in which the microtubule length can fluctuate wildly under steady conditions . in this section ,
we focus on the limiting case of infinite detachment rate , @xmath184 . in this limit , any gdp@xmath3 monomer(s ) at the tip of a microtubule are immediately removed .
thus the the tip is always a gtp@xmath2 ; this means that the parameter @xmath179 become immaterial .
finally , for @xmath184 , we also require the growth rate @xmath116 to have a microtubule with an appreciable length .
this is the limit considered below .
as soon as the last monomer of the tubule changes from a gtp@xmath2 to a gdp@xmath3 , a string of @xmath64 contiguous gdp@xmath3 monomers exist at the tip and they detach immediately .
we term such an event an _ avalanche of size @xmath64_. we now investigate the statistical properties of these avalanches . the switches from a growing to a shrinking state of a microtubule are called _ catastrophes _ @xcite .
if a newly - attached gtp@xmath2 at the tip converts to a gdp@xmath3 and the rest of the microtubule consists only of gdp@xmath3 at that moment , the microtubule instantaneously shrinks to zero length , a phenomenon that can be termed a _
global catastrophe_. we now determine the probability for such a catastrophe to occur . formally , the probability of a global catastrophe is @xmath185 the factor @xmath186 gives the probability that the monomer at the tip converts to a gdp@xmath3 before the next attachment event , while the product gives the probability that the rest of the microtubule consists of gdp@xmath3 . in principle , the upper limit in the product is set by the microtubule length .
however , for @xmath187 , each factor in the product is close to 1 and the error made in extending the product to infinity is small .
the expression within the product is obtained under the assumption that the tubule grows steadily between these complete catastrophes and the smaller , local catastrophes , are therefore ignored in this calculation .
the leading asymptotic behavior of the infinite product in is found by expressing it in terms of the dedekind @xmath188 function @xcite @xmath189 and using a remarkable identity satisfied by this function , @xmath190 for our purposes , we define @xmath191 to rewrite this identity as @xmath192 where @xmath193 .
specializing to the case @xmath194 yields @xmath195 since the time between catastrophes scales as the inverse of the occurrence probability , this inter - event time becomes very long for large @xmath21 . in the instantaneous detachment limit , @xmath184 , the catastrophes are avalanches whose size is determined by the number of gdp@xmath3s between the tip and the first gtp@xmath2 island .
a global catastrophe is an avalanche of size equal to the length of the tubule , whose occurrence probability was calculated in the preceding section .
similar arguments can be used to calculate the size distribution of the smaller avalanches .
since the cap is large when @xmath21 is large , an avalanche of size 1 arises only through the reaction scheme @xmath196 where the first step occurs at rate 1 and the second step is instantaneous .
since attachment proceeds with rate @xmath21 , the probability that conversion occurs before attachment is @xmath197 ; this expression gives the relative frequency of avalanches of sizes @xmath198 .
analogously , an avalanche of size two is formed by the events @xmath199 and the probability that the first two steps occur before an attachment event is @xmath200 to lowest order . at this level of approximation ,
the relative frequency of avalanches of size equal to 1 is @xmath201 .
since we are interested in the regime @xmath88 , we shall only consider the leading term in the avalanche size distribution . generally an avalanche of size @xmath64 is formed if the system starts in the configuration @xmath202 then @xmath71 contiguous gtp@xmath2 monomers next to the tip convert , and finally the gtp@xmath2 at the tip converts to gdp@xmath3 before the next attachment event .
the probability for the first @xmath71 conversion events is @xmath203 , where the factorial arises because the order of these steps is irrelevant .
the probability of the last step is @xmath204 .
thus the relative frequency of avalanches of size @xmath64 is @xmath205 the result can also be derived by the approach of sec .
[ continuum ] .
we use the fact that the configuration @xmath206 occurs with probability @xmath207 .
multiplying by the probability that the monomer at the tip converts before the next attachment event then gives the probability for an avalanche of size @xmath64 : @xmath208 using @xmath209 we recover .
if we expand the exponent to the next order , @xmath210 , eq . becomes @xmath211
the general situation where the attachment and detachment rates , @xmath21 and @xmath15 , have arbitrary values , and where the parameter @xmath14 seems analytically intractable because the master equations for basic observables are coupled to an infinite hierarchy of equations to higher - order variables .
for example , the quantity @xmath70 , the probability for a cap of length zero , satisfies the exact equation @xmath212 and the speed of the tip is @xmath213 here @xmath214 is the probability that there is a gdp@xmath3 at the front position with a gtp@xmath2 on its left .
thus to determine @xmath158 we must find @xmath215 , which then requires higher - order correlation functions , _ etc_. this hierarchical nature prevents an exact analysis and we turn to approximate approaches to map out the behavior in different regions of the parameter space .
for @xmath216 , the conversion gtp@xmath217 gdp@xmath3 at rate one greatly exceeds the rates @xmath218 of the other three basic processes that govern microtubule dynamics .
hence we can assume that conversion is instantaneous .
consequently , the end of a microtubule consists of a string of gdp@xmath3 , @xmath219 , in which the tip advances at rate @xmath155 and retreats at rate @xmath15 .
thus from the speed of the tip is @xmath220 when @xmath221 . on the other hand , for @xmath88
, @xmath70 is small and @xmath222 is exceedingly small .
hence @xmath223 . substituting this result into and solving for @xmath158 we find
@xmath224 note that indeed @xmath225 when @xmath88 . using in
we obtain the general result for the growth velocity @xmath226 a basic characteristic of microtubule dynamics is the phase boundary in the parameter space that separates the region where the microtubule grows without bound and a region where the mean microtubule length remains finite . for small @xmath21 ,
this boundary is found from setting @xmath227 in eq . to give @xmath228 for @xmath229 .
the phase boundary is a straight line in this limit , but for larger @xmath21 the boundary is a convex function of @xmath21 ( see fig . [ phase - diag ] ) .
we can compute the velocity to second order in @xmath15 and @xmath21 by assuming @xmath230 and then using and .
this leads to the phase boundary @xmath231 that is both convex and more precise . on this phase boundary
, the average tubule length grows as @xmath232 . and ( b ) @xmath233 .
the dashed line represents the prediction that is appropriate for small @xmath15 .
the extremes of the error bars are the points for which the tubule velocities are 0.005 and 0.015 , and their average defines the data point .
, title="fig:",scaledwidth=31.5% ] and ( b ) @xmath233 .
the dashed line represents the prediction that is appropriate for small @xmath15 .
the extremes of the error bars are the points for which the tubule velocities are 0.005 and 0.015 , and their average defines the data point .
, title="fig:",scaledwidth=31.5% ] when @xmath21 is large , eq
. implies that @xmath171 is positive and it reduces to @xmath234 for @xmath235 .
this simple result follows from the fact that recession of the microtubule is controlled by the unit conversion rate .
as soon as conversion occurs , detachment occurs immediately for @xmath236 and the microtubule recedes by one step .
since advancement occurs at rate @xmath21 , the speed of the tip is simply @xmath237 .
however , for extremely large @xmath15 the prediction @xmath234 breaks down and the microtubule becomes compact . to determine the phase boundary in this limit ,
consider first the case @xmath184 .
as shown in the previous section , the probability of a catastrophe roughly scales as @xmath238 so that the typical time between catastrophes is @xmath239 . since @xmath234 , the typical length of a microtubule just before a catastrophe is @xmath240 .
suppose now the detachment rate @xmath15 is very large but finite .
the microtubule is compact if the time to shrink a microtubule of length @xmath241 is smaller than the time @xmath242 required to generate a gtp@xmath2 by the attachment @xmath243 we checked the theoretical predictions and for the phase boundary numerically ( fig .
[ phase - diag ] ) . for small @xmath21 ,
the agreement between theory and the simulation is excellent . for larger @xmath21 ,
the tubule growth is more intermittent and it becomes increasingly difficult to determine the phase boundary with precision .
nevertheless , the qualitative expectations of our theory remain valid .
finally , we study the fluctuations of the tip in the small and large @xmath21 limits . in the former case but also on the growth phase @xmath221 , the tip undergoes a biased random walk with diffusion coefficient @xmath244 for large @xmath21 , the analysis is simplified by the principle that can be summarized by : `` the leading behaviors in the @xmath116 limit are universal , that is , independent of @xmath179 and @xmath15 . ''
this is not true if @xmath179 is particularly small [ like @xmath204 ] and/or if @xmath15 is particularly large [ like @xmath246 given by ] .
but when @xmath247 and @xmath248 , the above principle is true , and @xmath249 in the leading order .
the computation of sub - leading behaviors is more challenging .
we merely state here two asymptotic results .
when @xmath250 , we again have the relation @xmath251 , with @xmath252 .
if @xmath253 , we find @xmath254 the derivation of the latter uses the probabilities @xmath165 and @xmath166 and follows similar steps as in sect . [ xy - sub ] .
we investigated a simple dynamical model of a microtubule that grows by attachment of a gtp@xmath2 to its end at rate @xmath21 , irreversible conversion of any gtp@xmath2 to gdp@xmath3 at rate 1 , and detachment of a gdp@xmath3 from the end of a microtubule at rate @xmath15 . remarkably , these simple update rules for a one - dimensional system lead to steady growth , wild fluctuations , or a steady state .
our model has a minimalist formulation and therefore is not meant to account for all of the microscopic details of microtubule dynamics .
rather , our main goal has been to solve for the structural and dynamical properties of this idealized microtubule model .
some of the quantities that we determined , such as island size distributions , have not been studied previously . thus our predictions about the cap and island size distributions may help motivate experimental studies of these features of microtubules . a rich phenomenology was found as a function of the three fundamental rates in our model .
when gtp@xmath2 attachment is dominant ( @xmath88 ) and the attachment is independent of the identity of the last monomer on the free end of the microtubule ( @xmath17 ) , the gtp@xmath2 and gdp@xmath3 organize into alternating domains , with gradually longer gtp@xmath2 domains and gradually shorter gdp@xmath3 domains toward the tip of the microtubule ( fig . [ cartoon ] ) . here , the parameter @xmath21 could naturally be varied experimentally by either changing the temperature or the concentration of tubulins ( free gtp@xmath2 ) in the solution .
the basic geometrical features in this growing phase of a microtubule can be summarized as : [ cols="^,<,^",options="header " , ] we emphasize that the island size distributions of gtp@xmath2 and gdp@xmath3 are robust power laws with respective exponents of @xmath255 and @xmath256 . in the limit of @xmath257 , in which attachment
is suppressed when a gdp@xmath3 is at the free end of the microtubule , the average number of gtp@xmath2 monomers on the microtubule asymptotically is @xmath155 , while the rest of the results in the above table remain robust in the long - time limit .
conversely , when detachment of gdp@xmath3 from the end of the tubule is dominant ( detachment rate @xmath258 , a rate that also could be controlled by the temperature ) , the microtubule length remains bounded but its length can fluctuate strongly .
when the attachment rate is also large , the strong competition between attachment and detachment leads to wild fluctuations in the microtubule length even with steady external conditions .
we developed a probabilistic approach that shows that the time between catastrophes , where the microtubule shrinks to zero length , scales exponentially with the attachment rate @xmath21 .
thus a microtubule can grow essentially freely for a very long time before undergoing a catastrophe . for the more biologically relevant case of intermediate parameter values
, we extended our theoretical approaches to determine basic properties of the tubule , such as its rate of growth , fluctuations of the tip around this mean growth behavior , and the phase diagram in the ( @xmath259 ) parameter space . in this intermediate regime ,
numerical simulations provide more detailed picture of the geometrical structure and time history of a microtubule .
introducing the two - variable generating function @xmath260 we recast into a partial differential equation @xmath261 writing @xmath262}\,\,\mathcal{q}(x , y , t)\ ] ] we transform into a wave equation for an auxiliary function @xmath263 @xmath264 whose general solution is @xmath265 the initial condition @xmath266 implies @xmath267 and therefore @xmath268}\,\,\phi(x,\ln(1-y))=1\ ] ] from which @xmath269}\ ] ] combining eqs .
, we arrive at @xmath270
for @xmath14 , we consider the distributions @xmath272 and @xmath273 , defined as the probabilities to have @xmath18 gtp@xmath2 monomers with the tip being either gtp@xmath2 or gdp@xmath3 , respectively . these probabilities satisfy a closed set of coupled equations . in the stationary state
these equations become these generating functions have seemingly compact expressions but one has to keep in mind that the @xmath286 and @xmath287 probabilities are actually infinite sums .
for instance , @xmath288 .
recalling the definition of the confluent hypergeometric function we obtain @xmath289 where @xmath290 is the rising factorial .
note that @xmath291 . computing @xmath292 one can determine @xmath293 .
some of these formulas can be simplified using the kummer relation @xmath294 for instance , @xmath295 | we investigate a simple dynamical model of a microtubule that evolves by attachment of guanosine triphosphate ( gtp ) tubulin to its end , irreversible conversion of gtp to guanosine diphosphate ( gdp ) tubulin by hydrolysis , and detachment of gdp at the end of a microtubule . as a function of rates of these processes , the microtubule can grow steadily or
its length can fluctuate wildly . in the regime where detachment can be neglected ,
we find exact expressions for the tubule and gtp cap length distributions , as well as power - law length distributions of gtp and gdp islands . in the opposite limit of instantaneous detachment , we find the time between catastrophes , where the microtubule shrinks to zero length , and determine the size distribution of avalanches ( sequence of consecutive gdp detachment events ) .
we obtain the phase diagram for general rates and verify our predictions by numerical simulations . |
realizing chiral symmetry on the lattice is a problem with a long history . in 1975
wilson decided to break chiral symmetry explicitly in order to remove the unwanted doubler fermions @xcite . as a result ,
recovering chiral symmetry with wilson fermions requires fine - tuning as well as taking the continuum limit . in 1981
it was proved by nielsen and ninomiya @xcite and in a different way by friedan @xcite that removing the doubler fermions while maintaining locality is possible only if one explicitly breaks the chiral symmetry of the continuum theory . in 1992 , using a wilson term in 4 + 1 dimensions , kaplan constructed domain wall fermions in order to realize chiral symmetry on the lattice @xcite . at the end of the same year
, narayanan and neuberger introduced the related overlap formulation of chiral lattice fermions @xcite . already in 1982
ginsparg and wilson @xcite had discussed how to maintain a remnant of chiral symmetry on the lattice .
this paper was rediscovered by hasenfratz in 1997 in relation to the perfect action approach to qcd @xcite .
in particular , perfect lattice fermions obey the by now well - known ginsparg - wilson relation .
it was immediately realized by neuberger that the same is true for overlap fermions @xcite . in 1998
lscher showed that lattice fermion actions that obey the ginsparg - wilson relation realize a new version of chiral symmetry that is natural for theories on the lattice and coincides with the usual one in the continuum limit @xcite .
this development culminated in the non - perturbative construction of chiral lattice gauge theories @xcite . in this paper
we consider another aspect of lattice chiral symmetry , namely how it can be realized nonlinearly .
nonlinear realizations of symmetries are well understood in the continuum in the context of low - energy effective theories describing spontaneous symmetry breaking .
goldstone s theorem tells us that the spontaneous breakdown of a continuous global symmetry @xmath0 to an unbroken subgroup @xmath1 implies the presence of massless particles : the number of goldstone boson fields is given by the difference between the number of generators of @xmath0 and @xmath1 @xcite .
the chiral symmetry group @xmath2 of qcd with @xmath3 massless quark flavors is spontaneously broken to the vector subgroup @xmath4 .
hence there are @xmath5 massless goldstone bosons : the three pions for @xmath6 and , in addition , the four kaons , and the @xmath7-meson for @xmath8 . the theoretical framework for describing goldstone bosons with effective fields
was developed by weinberg @xcite and by coleman , wess , and zumino @xcite , partly in collaboration with callan @xcite . as a result
, goldstone bosons are described by fields in the coset space @xmath9 and the group @xmath0 is nonlinearly realized on those fields . in qcd the pions , the kaons , and the @xmath7-meson
are accordingly described by matrix valued fields in the coset space @xmath10 = su(n_f)$ ] .
notably , the pion effective theory developed by weinberg @xcite has a systematic low - energy expansion in gasser and leutwyler s framework of chiral perturbation theory @xcite .
when matter fields are included in the effective theory of goldstone bosons , they also transform nonlinearly under chiral symmetry @xcite .
in particular , the global symmetry @xmath0 is hidden in a locally realized symmetry @xmath1 with a corresponding `` gauge '' field that is constructed from the goldstone boson field .
it must be pointed out that there is no true gauge symmetry @xmath1 , just the global symmetry @xmath0 . exploiting a nonlinearly realized chiral symmetry , gasser , sainio , and svarc incorporated a single nucleon in chiral perturbation theory @xcite .
systematic power counting schemes for non - relativistic heavy baryon chiral perturbation theory were developed by jenkins and manohar @xcite and by bernard , kaiser , kambor , and meissner @xcite . a more economical scheme that maintains explicit lorentz invariance was introduced by becher and leutwyler @xcite .
one may ask why one would want to put low - energy effective theories on the lattice .
indeed , many questions that arise in the framework of low - energy lagrangians can be solved analytically using chiral perturbation theory .
usually , such calculations are performed in the continuum using dimensional regularization .
it is certainly more tedious to work with the lattice regularization in chiral perturbation theory calculations .
still , in order to address conceptual problems , for example , related to power divergences , the lattice has already been used in magnon @xcite , pion @xcite , as well as heavy baryon chiral perturbation theory @xcite . in the context of lattice qcd ,
effective field theories provide powerful tools to keep systematic errors under control @xcite .
moreover , chiral lagrangians on the lattice have been investigated in order to facilitate the lattice determination of the gasser - leutwyler coefficients @xcite .
although interesting , this is not the main motivation for the present work .
we formulate lattice theories with nonlinearly realized chiral symmetry in order to address non - perturbative questions that arise in the context of low - energy effective theories .
it is perhaps not obvious that such problems even exist .
an example of a non - perturbative problem in the context of chiral lagrangians concerns the rotor spectrum of massless qcd in a finite box .
however , this problem reduces to one in quantum mechanics and has been solved analytically by leutwyler @xcite without any need for a lattice .
few - nucleon systems have been treated with effective field theory methods in nuclear physics .
for example , the two - nucleon system was studied by kaplan , savage , and wise @xcite and the three - nucleon system was investigated by bedaque , hammer , and van kolck @xcite .
again , the solution of these problems required non - perturbative techniques but not yet a lattice . going to larger nuclei and eventually to
nuclear matter is a non - trivial task that requires to solve more complicated non - perturbative problems in the effective theory @xcite . here
the lattice provides a natural framework in which such questions can be addressed on a solid theoretical basis .
a central observation of this paper is that the nielsen - ninomiya theorem is naturally avoided when chiral symmetry is nonlinearly realized on the lattice .
in fact , constituent quark masses do not break a nonlinearly realized chiral symmetry .
similarly , the wilson term introduced in order to remove the doubler fermions does no longer break chiral symmetry explicitly .
hence , exact chiral symmetry is naturally maintained using a simple wilson action while the fermion doubling problem is still solved .
this raises the question how anomalies manifest themselves in this context .
remarkably , in effective theories with nonlinearly realized chiral symmetry , both in the continuum and on the lattice , the fermion fields do not contribute to anomalies .
instead anomalies must be added by hand in the form of wess - zumino - witten @xcite and goldstone - wilczek @xcite terms .
constructing these terms on the lattice while maintaining their topological features is a non - trivial problem which we do not address in this paper . in this respect ,
our construction of lattice theories with nonlinearly realized chiral symmetry is still incomplete .
consequently , we limit ourselves to applications where anomalies play no major role . such applications concern , for example , models of static baryons interacting with pions at non - zero chemical potential and the chiral quark model of georgi and manohar @xcite .
dynamical fermion simulations in a lattice theory with nonlinearly realized chiral symmetry may be substantially easier than with the usual linear realization . using the nonlinear realization , even in the chiral limit ,
the fermion mass is large due to spontaneous chiral symmetry breaking .
hence , the simulated dynamical fermions should behave algorithmically like heavy wilson fermions while still maintaining manifest chiral symmetry . using a linear realization of chiral symmetry on the quark fields , explicit auxiliary pion fields
have already been explored in order to simplify qcd simulations @xcite .
it has to be mentioned that , in general , lattice theories with nonlinearly realized chiral symmetry have a potential sign problem similar to the one that arises for a single flavor of wilson fermions .
fortunately , this sign problem is a lattice artifact which should disappear in the continuum limit , in particular , for sufficiently heavy fermions .
recently , without directly addressing the potential sign problem , dynamical fermion algorithms have been developed for an odd number of flavors of wilson fermions @xcite .
we expect that similar methods are applicable to lattice fermions with nonlinearly realized chiral symmetry .
still , it should be pointed out that the feasibility of dynamical fermion simulations depends crucially on the severity of the sign problem , which should be investigated in numerical simulations .
we also need to mention that additional complex action problems arise if pseudo - vector fermion couplings or wess - zumino - witten and goldstone - wilczek terms are included . in this paper
we concentrate on the theoretical formulation of lattice theories with a nonlinearly realized chiral symmetry , while results of numerical simulations will be presented elsewhere .
interestingly , the lattice construction presented here enables us to simulate massless pions interacting with static baryons even at non - zero baryon chemical potential @xmath11 and finite temperature @xmath12 , both for two and for three flavors .
in contrast to full qcd , in the static baryon model this is possible without encountering a complex action problem .
although questions involving high temperatures or high baryon densities are beyond the applicability of a systematic low - energy expansion , this model can still be used to answer questions concerning the universal features of the critical endpoint of the chiral phase transition line in the @xmath13-plane . although one should not expect to obtain precise information on non - universal features of qcd , such as the location of the critical endpoint
, one may learn at least semi - quantitatively how the location of the endpoint changes as a function of the strange quark mass .
in addition , one will gain experience with numerical methods for locating the critical endpoint in a model that has the same global symmetry properties as full qcd but is much simpler to simulate .
a nonlinear realization of chiral symmetry was also used by georgi and manohar in the construction of their chiral quark model @xcite .
this model contains massive constituent quarks interacting with pions and gluons .
while gluons are flavor singlets , both constituent quarks and pions transform non - trivially under the nonlinearly realized chiral symmetry .
if this model provides an appropriate low - energy description for qcd , one could understand why the non - relativistic quark model works .
however , the confinement of constituent quarks is a non - perturbative issue that can not be addressed quantitatively in the continuum formulation of the model . in this paper
we present a lattice construction of the georgi - manohar model with nonlinearly realized chiral symmetry . this non - perturbative framework will allow one to decide quantitatively if the georgi - manohar model can explain the success of the non - relativistic quark model .
the rest of this paper is organized as follows . in section 2 the nonlinear realization of chiral symmetry
is reviewed in the continuum , while section 3 contains the corresponding construction on the lattice . in both cases
the relevant symmetries are discussed in detail .
in particular , effective lattice theories for pions and baryons are considered and the issue of fermion doubling is discussed . in section 4
we discuss how low - energy effective theories can be treated non - perturbatively on the lattice and how continuum results can be extracted .
section 5 explains how theories with static baryons can be simulated at non - zero chemical potential without encountering a complex action problem . in section 6
we discuss the lattice version of the georgi - manohar model and propose numerical simulations in order to decide if it indeed provides a satisfactory explanation for the success of the non - relativistic quark model .
section 7 discusses the possible form of the phase diagram of the lattice chiral quark model and the relation of this model to wilson s standard lattice qcd .
finally , section 8 contains our conclusions .
in this section we review how theories with nonlinearly realized chiral symmetry are formulated in the continuum . the fields used in this construction
are introduced with emphasis on their behavior under chiral symmetry transformations , charge conjugation , and parity .
simple examples of low - energy effective theories with nonlinearly realized chiral symmetry are given and manifestations of anomalies are being discussed . consider the chiral symmetry group @xmath14 of qcd involving @xmath3 quark flavors which breaks spontaneously to the subgroup @xmath15 . following @xcite the goldstone boson field
is then given by @xmath16 \in g / h = su(n_f)$ ] .
here @xmath17 and @xmath18 is the pion decay constant . under global chiral rotations
@xmath19 , this field transforms as @xmath20 from the goldstone boson field one derives a field @xmath21 as @xmath22 in order to avoid ambiguities in taking the square - root , we proceed as follows .
first , the field @xmath23 is diagonalized by a unitary transformation @xmath24 , i.e. @xmath25 where @xmath26 is a diagonal matrix of complex phases @xmath27,\exp[i \varphi_2(x ) ] , ... , \exp[i \varphi_{n_f}(x)]\}.\ ] ] at this point , the phases @xmath28 are defined only modulo @xmath29 . in order to fix this ambiguity
, we demand that @xmath30 this uniquely fixes the phases of the eigenvalues and defines the matrix @xmath26 up to an irrelevant permutation of its diagonal elements .
we then define @xmath31,\exp[i \varphi_2(x)/2 ] , ... , \exp[i \varphi_{n_f}(x)/2]\},\ ] ] and we obtain @xmath32 when the eigenvalues of @xmath23 are non - degenerate , the matrix @xmath33 is uniquely defined in this way and is located in the middle of the shortest geodesic connecting @xmath23 with the unit - matrix @xmath34 in the group manifold of @xmath35 . on the other hand , if two eigenvalues are degenerate , the corresponding eigenvectors are not uniquely defined . if , in addition , the corresponding phases @xmath36 and @xmath37 differ by @xmath29 , this ambiguity gives rise to a coordinate singularity in the definition of @xmath33 . in the @xmath6 case , the singularities arise for @xmath38 , which corresponds to a single point in the @xmath39 group manifold . for general
@xmath3 , the submanifold of @xmath35 matrices with two degenerate eigenvalues ( and with @xmath40 ) has dimension @xmath41 .
hence , the codimension of the singular submanifold is @xmath42 for all @xmath3 .
consequently , in a 4-dimensional space - time the singularities in @xmath33 are located on 1-dimensional lines . if one chooses the vacuum as @xmath43 , the singularities represent the world - lines of skyrmion centers .
however , it should be noted that the location of the singularities varies under those chiral rotations that do not leave the vacuum invariant .
still , the effective actions that we will construct with @xmath33 are completely chirally invariant , even in the presence of skyrmions or other large - field configurations . under chiral rotations the field @xmath33 transforms as @xmath44
the matrix @xmath45 depends on @xmath46 and @xmath47 as well as on the field @xmath23 and can be written as @xmath48^{1/2 } [ u(x)^{1/2}]^\dagger = l[l^\dagger r u(x)^\dagger]^{1/2 } u(x)^{1/2}.\ ] ] for transformations in the unbroken vector subgroup @xmath49 , @xmath50 , the field @xmath45 reduces to the global flavor transformation @xmath51 .
the local transformation @xmath45 is a nonlinear representation of chiral symmetry .
for two subsequent chiral transformations with @xmath52 and @xmath53 , one has @xmath54^{1/2 } [ u(x)^{1/2}]^\dagger , \
v_2(x ) = r_2[r_2^\dagger l_2 u(x)']^{1/2 } [ { u(x)'}^{1/2}]^\dagger,\ ] ] with @xmath55 , so that indeed @xmath56^{1/2 } r_1 [ u(x)^{1/2}]^\dagger \nonumber \\ & = & r_2 r_1 [ r_1^\dagger r_2^\dagger l_2 l_1 u(x)]^{1/2 } [ u(x)^{1/2}]^\dagger \nonumber \\ & = & r[r^\dagger l u(x)]^{1/2 } [ u(x)^{1/2}]^\dagger = v(x).\end{aligned}\ ] ] now consider a dirac spinor field @xmath57 and @xmath58 that transforms under the fundamental representation of the flavor group .
this field may represent a nucleon ( for @xmath6 ) or a constituent quark ( for any value of @xmath3 ) .
global chiral rotations @xmath59 can be realized nonlinearly on this field using the transformations @xmath60 this has the form of an @xmath35 gauge transformation despite the fact that it represents just a global @xmath61 symmetry . in order to construct a chirally invariant ( i.e. @xmath35 `` gauge '' invariant ) lagrangian
, one needs an @xmath35 flavor `` gauge '' field .
indeed , the anti - hermitean composite field @xmath62\ ] ] transforms as a `` gauge '' field @xmath63 + v(x ) u(x ) r^\dagger \partial_\mu [ r u(x)^\dagger v(x)^\dagger]\ } \nonumber \\ & = & v(x ) [ v_\mu(x ) + \partial_\mu ] v(x)^\dagger.\end{aligned}\ ] ] a hermitean composite field is given by @xmath64,\ ] ] which transforms as @xmath65 - v(x ) u(x ) r^\dagger \partial_\mu [ r u(x)^\dagger v(x)^\dagger]\ } \nonumber \\ & = & v(x ) a_\mu(x ) v(x)^\dagger.\end{aligned}\ ] ] it is interesting to investigate the transformation properties of the fields @xmath57 , @xmath58 , @xmath23 , @xmath33 , @xmath66 , and @xmath67 under charge conjugation and parity . under charge conjugation , the fermion field transforms in the usual way @xmath68 with @xmath69 .
the goldstone boson field @xmath23 transforms as @xmath70 consequently , the field @xmath33 also follows the same transformation law @xmath71^{1/2 } = [ u(x)^t]^{1/2 } = [ u(x)^{1/2}]^t = u(x)^t.\ ] ] it is easy to show that the composite fields @xmath66 and @xmath67 transform as @xmath72 under parity , the fermion field behaves again in the usual way , i.e. @xmath73 where 4 denotes the euclidean time direction .
the goldstone boson field @xmath23 transforms as @xmath74 so that @xmath75^{1/2 } = [ u(- \vec x , t)^\dagger]^{1/2 } = [ u(- \vec x , t)^{1/2}]^\dagger = u(- \vec x , t)^\dagger.\ ] ] similarly , one finds @xmath76 where @xmath77 denotes a spatial direction .
hence , @xmath66 is a vector and @xmath67 is an axial vector . at this point
we are prepared to write down the leading terms in the euclidean action of a low - energy effective theory for nucleons ( with @xmath6 ) @xmath78&=&\int d^4x \
\big\{m \overline \psi \psi + \overline \psi \gamma_\mu ( \partial_\mu + v_\mu ) \psi + i g_a \overline \psi \gamma_\mu \gamma_5 a_\mu \psi \nonumber \\ & + & \frac{f_\pi^2}{4 } \mbox{tr}[\partial_\mu u^\dagger \partial_\mu u ] - \frac{\langle \overline \psi \psi \rangle}{2 } \mbox{tr}[{\cal m } u^\dagger + { \cal m}^\dagger u]\big\}.\end{aligned}\ ] ] here @xmath79 is the nucleon mass generated by spontaneous chiral symmetry breaking , @xmath80 is the coupling to the isovector axial current , and @xmath81 and @xmath82 are respectively the chiral condensate per flavor and the mass matrix of the current quarks .
when @xmath57 represents a constituent quark , @xmath79 would be the constituent quark mass . in that case , gluons must also be included in order to confine the constituent quarks .
this is what happens in the chiral quark model of georgi and manohar @xcite ( see section 6 ) .
it is remarkable that due to the nonlinear realization of chiral symmetry the fermion mass term is chirally invariant .
this makes sense because the mass @xmath79 arises from the spontaneous breakdown of chiral symmetry even in the chiral limit , i.e. when the current quarks are massless .
in fact , the term that contains the matrix @xmath83 of current quark masses is the only source of explicit chiral symmetry breaking in the leading terms of the effective theory .
it is interesting to note that the composite field @xmath66 is coupled to the fermions as a flavor `` gauge '' field that enters the covariant derivative @xmath84 . like
@xmath85 , the axial vector field @xmath86 also transforms in the adjoint representation of the flavor group .
however , it is not a `` gauge '' field and it couples to the fermions with strength @xmath80 .
it must be noted that there is , of course , no true @xmath35 gauge symmetry , just an @xmath87 global symmetry . in particular ,
the `` gauge '' field @xmath66 is not a fundamental field and there is , for instance , no @xmath35 gauss law .
until now we have considered a fermion field @xmath57 transforming in the fundamental representation of @xmath35 . in qcd with three massless flavors ( @xmath8 ) the baryons transform as a flavor octet .
the corresponding dirac spinor field , denoted by @xmath88 and @xmath89 , is traceless and transforms in the adjoint representation @xmath90 the covariant derivative is then given by @xmath91,\ ] ] which transforms as @xmath92 under the nonlinearly realized chiral symmetry . in this case
, the euclidean action of the effective theory for pions , kaons , @xmath7-mesons , and baryons takes the form @xmath93&=&\int d^4x \
\big\{\mbox{tr}[m \overline \chi \chi + \overline \chi \gamma_\mu d_\mu \chi + i d \overline \chi \gamma_\mu \gamma_5 \{a_\mu,\chi\ } + i f \overline \chi \gamma_\mu \gamma_5 [ a_\mu,\chi ] ] \nonumber \\ & + & \frac{f_\pi^2}{4 } \mbox{tr}[\partial_\mu u^\dagger \partial_\mu u ] - \frac{\langle \overline \psi \psi \rangle}{2 } \mbox{tr}[{\cal m } u^\dagger + { \cal m}^\dagger u]\big\},\end{aligned}\ ] ] where the matrix @xmath94 now also includes the strange quark mass . the low - energy parameters @xmath95 and @xmath96 are analogous to the coupling @xmath80 of the two flavor case
. it is interesting to discuss how anomalies arise in effective theories with nonlinearly realized chiral symmetry .
let us begin with a discussion of the axial @xmath97 symmetry of qcd .
when the number of colors @xmath98 is finite , this symmetry of the classical action is anomalously broken by quantum effects due to instantons @xcite . as a result , the @xmath99-meson is not a goldstone boson . only in the large @xmath98 limit
the @xmath99-meson becomes a pseudo - goldstone boson @xcite . for @xmath100 ,
the @xmath97 symmetry is so strongly broken that it can not even be considered an approximate symmetry of qcd . in particular , the @xmath99-meson is so heavy that it is not included explicitly in the low - energy effective theory .
indeed , the effective theories that we have written down do not have a @xmath97 symmetry at all .
this follows because the goldstone boson field lives in @xmath35 and not in @xmath101 .
let us now move on to those anomalies that arise when one gauges the electroweak interactions . in effective theories with
nonlinearly realized chiral symmetries anomalies manifest themselves @xcite in the form of wess - zumino - witten @xcite and goldstone - wilczek terms @xcite .
for example , in the @xmath6 case , witten s global @xmath102 anomaly @xcite manifests itself in a topological factor @xmath103^{n_c } \in \pi_4[su(2 ) ] = \{1,-1\}$ ] that must be included in the path integral @xcite .
this ensures that , for odd @xmath98 , the skyrme soliton @xcite is quantized as a fermion representing the physical nucleon in the pure pion theory .
the @xmath102 gauge invariant version of the skyrme baryon number current is the goldstone - wilczek current @xcite . in the presence of electroweak instantons , this current is no longer conserved and reproduces the t hooft anomaly in the baryon number @xcite . in the @xmath104 case
, the wess - zumino - witten term replaces the topological factor @xmath103 $ ] in the goldstone boson theory .
this is because then @xmath105 = \{0\}$ ] and instead @xmath106 = { { \sf z \!\!\ !
the wess - zumino - witten action has the form @xmath107 = \frac{1}{240 \pi^2 i } \int_{h^5 } d^5x \ \varepsilon_{\mu\nu\rho\sigma\lambda } \mbox{tr}[(u^\dagger { \partial}_\mu u ) ( u^\dagger { \partial}_\nu u)(u^\dagger { \partial}_\rho u)(u^\dagger { \partial}_\sigma u ) ( u^\dagger { \partial}_\lambda u)],\ ] ] where @xmath108 is a 5-dimensional hemi - sphere whose boundary @xmath109 is ( compactified ) space - time .
the physical goldstone boson field @xmath110 is extended to a field @xmath111 that depends on the unphysical fifth coordinate @xmath112 $ ] such that @xmath113 and @xmath114 . of course
, the 4-dimensional physics should be independent of how the field @xmath111 is deformed into the bulk of the fifth dimension , and should only depend on @xmath23 .
this is possible because the interpolation ambiguity of @xmath115 $ ] is given by @xmath29 times a winding number in @xmath106 = { { \sf z \!\!\ !
z}}$ ] . hence , when the prefactor of the wess - zumino - witten term is quantized in integer units , the 4-dimensional physics is independent of how one interpolates the goldstone boson field into the unphysical fifth dimension .
remarkably , in the low - energy theory for qcd , the quantized prefactor is the number of colors @xmath98 so that @xmath116 ) \exp(i
n_c s_{wzw}[u]).\ ] ] for @xmath6 , one can show that @xmath117 ) = \mbox{sign}[u]^{n_c}$ ] .
the wess - zumino - witten term contains the vertex that describes the anomalous decay of the neutral pion into two photons . for general @xmath98
, this process also receives contributions from a goldstone - wilczek term @xcite .
in contrast to the @xmath6 case , for @xmath104 , the wess - zumino - witten term also breaks an unwanted intrinsic parity symmetry @xmath118 that is present in the goldstone boson action @xmath119 $ ] but not in qcd @xcite .
the full parity @xmath120 acts on the pseudo - scalar goldstone bosons @xmath121 by spatial inversion accompanied by a sign - change , i.e. @xmath122 , such that @xmath123 the intrinsic parity @xmath118 , on the other hand , leaves out the spatial inversion and takes the form @xmath124 if @xmath118 were a symmetry of qcd , the number of goldstone bosons would be conserved modulo two . while this is indeed the case for @xmath6 ( where @xmath118 is nothing but @xmath0-parity @xcite ) , for @xmath104 intrinsic parity
is not a symmetry of qcd .
for example , the @xmath125-meson can decay both into two kaons or into three pions .
however , the goldstone boson action @xmath119 $ ] is indeed invariant under @xmath118 , @xmath126 = s[u^\dagger ] = s[u].\ ] ] hence it has more symmetry than the underlying qcd action .
the wess - zumino - witten action , on the other hand , is odd under @xmath118 , @xmath127 = s_{wzw}[u^\dagger ] = - s_{wzw}[u],\ ] ] and thus reduces the symmetry of the effective theory to the one of qcd .
the next step is to ask how anomalies are represented when explicit baryon fields are present in the low - energy effective theory . in the presence of electroweak @xmath128 gauge fields @xmath129 and @xmath130 one obtains @xmath131u(x ) + u(x)[{\partial}_\mu + b_\mu(x)]u(x)^\dagger\ } , \nonumber \\
a_\mu(x ) = \frac{i}{2}\{u(x)^\dagger[{\partial}_\mu + w_\mu(x)]u(x ) - u(x)[{\partial}_\mu + b_\mu(x)]u(x)^\dagger\},\end{aligned}\ ] ] which still transform as @xmath132v(x)^\dagger , \
a_\mu(x ) ' = v(x ) a_\mu(x ) v(x)^\dagger.\ ] ] hence , the dynamical gauge fields @xmath133 and @xmath134 are incorporated in the `` gauge '' field @xmath66 constructed from the pion field @xmath23 . by varying the resulting action with respect to @xmath133 and @xmath134 ,
one obtains the left- and right - handed fermion currents @xmath135 since fermions with a nonlinearly realized chiral symmetry do not contribute to anomalies @xcite , even in the presence of fermion fields , the anomalies are still contained in goldstone - wilczek and wess - zumino - witten terms . with explicit baryon fields being present in addition to the skyrme solitons in the pion field ,
one may wonder if baryons are doubly counted in this approach .
first of all , in the power counting scheme of chiral perturbation theory one can not address questions about skyrmions because all contributions from higher order terms are then equally important .
in a non - perturbative approach to the effective theory , one can ask if , for example , in the sector with one baryon , the perturbative vacuum configuration of the pion field becomes unstable against skyrmion formation .
if so , the chiral power counting scheme is not applicable and the systematic low - energy approach fails .
hence , in the low - energy effective theory , one is restricted to the trivial topological sector and the issue of double counting does not arise . since skyrmions have no place in the systematic low - energy expansion , one can not use the goldstone - wilczek current to describe baryon number violating processes in the framework of chiral perturbation theory .
this should not be too surprising .
baryon decay as rare as it may be is a violent event that releases a lot of energy and can hence not be understood within a low - energy effective theory .
let us now construct theories with a nonlinearly realized chiral symmetry on the lattice . in analogy to the previous section , we set up the lattice fields needed for this construction such that the correct symmetry properties are respected . with these fields ,
we formulate lattice actions of low - energy effective theories in which chiral symmetry is nonlinearly realized .
in particular , we discuss how fermion doublers can be removed using a wilson term without explicitly breaking chiral symmetry and how the nielsen - ninomiya theorem is avoided .
the issue of anomalies is also addressed in the lattice formulation .
the goldstone boson field @xmath136 lives on the sites @xmath137 of a 4-dimensional hyper - cubic lattice and it transforms under global chiral rotations as @xmath138 as in the continuum , the field @xmath139 is constructed as @xmath140 and it transforms as @xmath141 again , the field @xmath142 depends on the global chiral transformations @xmath46 and @xmath47 as well as on the field @xmath143 and is given by @xmath144 the continuum results for two subsequent chiral transformations given in eqs .
( [ v1 ] ) and ( [ v2 ] ) remain completely unchanged on the lattice .
also the discussion of the coordinate singularities still applies .
however , it should be noted that the singularities generically fall between lattice points .
still , using an appropriate interpolation , it is possible to detect these topological objects even on the lattice .
next we consider the fermion field @xmath145 and @xmath146 which again lives on the lattice sites @xmath137 . as in the continuum , global chiral rotations
are nonlinearly realized on this field such that @xmath147 this field transforms in the fundamental representation of @xmath35 and can hence represent a nucleon ( for @xmath6 ) or a constituent quark ( for any value of @xmath3 ) . for @xmath8 ,
the lattice baryon field @xmath148 and @xmath149 also lives on the lattice sites but transforms in the adjoint representation @xmath150 we now proceed to the construction of @xmath151 the lattice analog of the continuum flavor `` gauge '' field @xmath66 . it should be noted that @xmath152 is a flavor parallel transporter along a lattice link in the group @xmath35 , while the continuum flavor `` gauge '' field is in the algebra @xmath153 . in analogy to the continuum expression ( [ v ] ) for @xmath66 , we construct the lattice object @xmath154.\ ] ] here @xmath155 is the unit - vector in the @xmath11-direction . by construction
, @xmath156 transforms as a parallel transporter under the nonlinearly realized chiral symmetry , i.e. @xmath157 in the classical continuum limit , @xmath158 $ ] .
however , at finite lattice spacing , @xmath156 is in general not an element of @xmath35 , just a complex @xmath159 matrix in the group @xmath160 .
hence , it is natural to project a group - valued parallel transporter @xmath151 out of @xmath156 by performing a @xmath161 coset decomposition @xmath162 here @xmath163 is a positive hermitean matrix that transforms as @xmath164 and @xmath165 ( with @xmath166\!-\pi,\pi]$ ] ) is the phase of the determinant of @xmath156 .
then , by construction , @xmath167 is indeed in @xmath35 and transforms as @xmath168 we also want to construct a lattice version of the continuum field @xmath67 defined in ( [ a ] ) . for this purpose , we first consider @xmath169.\ ] ] while in the continuum @xmath170 , the lattice field @xmath171 transforms as @xmath172 also , in contrast to the continuum field @xmath67 , the lattice field @xmath171 is in general neither traceless nor hermitean . to cure these problems ,
it is natural to introduce the field @xmath173 - \frac{1}{2n_f } \mbox{tr } [ \tilde a_{x,\mu } v_{x,\mu}^\dagger + v_{x,\mu } \tilde a_{x,\mu}^\dagger ] { { \sf 1 \!\ ! 1}},\ ] ] which , by construction , is indeed traceless and hermitean and which transforms as @xmath174 with the matrix @xmath142 located at the site @xmath137 on the left end of the link @xmath175 .
similarly , we define the object @xmath176 - \frac{1}{2n_f } \mbox{tr } [ v_{x,\mu}^\dagger \tilde a_{x,\mu } + \tilde a_{x,\mu}^\dagger v_{x,\mu } ] { { \sf 1 \!\ !
1}},\ ] ] which transforms as @xmath177 with the matrix @xmath178 located at the site @xmath179 on the right end of the link @xmath175 .
it should be noted that @xmath180 and @xmath181 are not independent , but are related by parallel transport , i.e. @xmath182 let us now consider the symmetry properties of the lattice fields under charge conjugation and parity .
the site fields @xmath145 ,
@xmath146 , @xmath148 , @xmath149 , @xmath143 , and @xmath183 transform exactly as their continuum analogs and will not be discussed again . instead , we concentrate on the fields @xmath152 , @xmath181 , and @xmath180 . under charge conjugation , they transform as @xmath184 similarly , under parity transformations , one finds @xmath185 where @xmath77 denotes a spatial direction and 4 again denotes the euclidean time direction . hence , as in the continuum , @xmath152 is a vector and @xmath181 and @xmath180 define an axial vector .
we now use the fields introduced above to construct lattice actions for baryon effective theories .
we start out with the @xmath6 case . in close analogy to the continuum action ( [ action2 ] ) , we define the lattice action as @xmath186&=&m \sum_x \overline \psi_x \psi_x + \frac{1}{2 } \sum_{x,\mu } ( \overline \psi_x \gamma_\mu v_{x,\mu } \psi_{x+\hat\mu } - \overline \psi_{x+\hat\mu } \gamma_\mu v_{x,\mu}^\dagger \psi_x ) \nonumber \\ & + & \frac{i g_a}{2 } \sum_{x,\mu } ( \overline \psi_x \gamma_\mu \gamma_5 a^l_{x,\mu } \psi_{x } + \overline \psi_{x+\hat\mu } \gamma_\mu \gamma_5 a^r_{x,\mu } \psi_{x+\hat\mu } ) \nonumber \\ & + & \frac{r}{2 } \sum_{x,\mu } ( 2 \overline \psi_x \psi_x - \overline \psi_x v_{x,\mu } \psi_{x+\hat\mu } - \overline \psi_{x+\hat\mu } v_{x,\mu}^\dagger \psi_x ) \nonumber \\ & -&\frac{f_\pi^2}{4 } \sum_{x,\mu } \mbox{tr}[u_x^\dagger u_{x+\hat\mu } + u_{x+\hat\mu}^\dagger u_x ] - \frac{\langle \overline \psi \psi \rangle}{2 } \sum_x \mbox{tr}[{\cal m } u_x^\dagger + { \cal m}^\dagger u_x ] .
\nonumber \\ \\end{aligned}\ ] ] similarly , for @xmath8 , ( where the baryon field transforms in the adjoint representation ) the lattice action corresponding to ( [ nf3action ] ) is given by @xmath187&=&m \sum_x \mbox{tr}\overline \chi_x \chi_x + \frac{1}{2 } \sum_{x,\mu } \mbox{tr } ( \overline \chi_x \gamma_\mu v_{x,\mu } \chi_{x+\hat\mu } v_{x,\mu}^\dagger - \overline \chi_{x+\hat\mu } \gamma_\mu v_{x,\mu}^\dagger \chi_x v_{x,\mu } ) \nonumber \\ & + & \frac{i d}{2 } \sum_{x,\mu } \mbox{tr } ( \overline \chi_x \gamma_\mu \gamma_5 \{a^l_{x,\mu},\chi_{x}\ } + \overline \chi_{x+\hat\mu } \gamma_\mu \gamma_5 \{a^r_{x,\mu},\chi_{x+\hat\mu}\ } ) \nonumber \\ & + & \frac{i f}{2 } \sum_{x,\mu } \mbox{tr } ( \overline \chi_x \gamma_\mu \gamma_5 [ a^l_{x,\mu},\chi_{x } ] + \overline \chi_{x+\hat\mu } \gamma_\mu \gamma_5 [ a^r_{x,\mu},\chi_{x+\hat\mu } ] ) \nonumber \\ & + & \frac{r}{2 } \sum_{x,\mu } \mbox{tr } ( 2 \overline \chi_x \chi_x - \overline \chi_x v_{x,\mu } \chi_{x+\hat\mu } v_{x,\mu}^\dagger - \overline \chi_{x+\hat\mu } v_{x,\mu}^\dagger \chi_x v_{x,\mu } ) \nonumber \\ & -&\frac{f_\pi^2}{4 } \sum_{x,\mu } \mbox{tr}[u_x^\dagger u_{x+\hat\mu } + u_{x+\hat\mu}^\dagger u_x ] - \frac{\langle \overline \psi \psi \rangle}{2 } \sum_x \mbox{tr}[{\cal m } u_x^\dagger + { \cal m}^\dagger u_x].\end{aligned}\ ] ] the wilson term proportional to @xmath188 removes the doubler fermions . in standard lattice
qcd this term breaks chiral symmetry explicitly .
remarkably , when chiral symmetry is nonlinearly realized , not only the fermion mass term ( proportional to @xmath79 ) but also the wilson term is chirally invariant .
the only source of explicit chiral symmetry breaking is the current quark mass matrix @xmath83 .
it is interesting to ask how the nielsen - ninomiya theorem @xcite has been avoided . clearly , the actions ( [ latact2 ] ) and ( [ latact3 ] ) are local . for @xmath6 ,
the corresponding dirac operator is given by @xmath189&=&m \delta_{x , y } + \frac{1}{2 } ( \gamma_\mu v_{x,\mu } \delta_{x+\hat\mu , y } - \gamma_\mu v_{x-\hat\mu,\mu}^\dagger \delta_{x-\hat\mu , y } ) \nonumber \\ & + & \frac{i g_a}{2 } ( \gamma_\mu \gamma_5 a^l_{x,\mu } \delta_{x , y } + \gamma_\mu \gamma_5 a^r_{x-\hat\mu,\mu } \delta_{x , y } ) \nonumber \\ & + & \frac{r}{2 } ( 2 \delta_{x , y } - v_{x,\mu } \delta_{x+\hat\mu , y } - v_{x-\hat\mu,\mu}^\dagger \delta_{x-\hat\mu , y}).\end{aligned}\ ] ] the nielsen - ninomiya theorem assumes that the dirac operator anticommutes with @xmath190 .
this is not the case when chiral symmetry is nonlinearly realized , and hence one of the basic assumptions of the no - go theorem is not satisfied .
interestingly , ginsparg - wilson fermions evade the nielsen - ninomiya theorem by violating the same assumption .
of course , the nonlinear realization of chiral symmetry requires an explicit pion field which is not present in the fundamental qcd lagrangian .
let us again discuss the issue of anomalies .
on the lattice , anomalies may receive contributions from the doubler fermions .
hence , removing the doublers is intimately related to anomaly matching .
it is straightforward to include electroweak gauge fields in the lattice theory as well .
for example , the parallel transporters @xmath191 of the electroweak gauge field transform as @xmath192 together with the @xmath193 gauge field @xmath194 , the field @xmath195 is incorporated in the field @xmath196,\ ] ] which is again projected on a parallel transporter @xmath197 by a @xmath198 coset decomposition .
also the fields @xmath199 and @xmath200 are constructed as in ( [ aleft ] ) and ( [ aright ] ) , now using @xmath201.\ ] ] the lattice fermion action is constructed as in the ungauged case ( [ latact2 ] ) and is now @xmath128 gauge invariant .
remarkably , not only the lattice fermion action but also the lattice fermion measure is gauge invariant . at first sight , this seems to be a severe problem because gauge invariance of both the fermion action and the fermion measure implies that lattice fermions with nonlinearly realized chiral symmetry do not contribute to anomalies .
one might even suspect that the doubler fermions have not been properly removed and thus have canceled the anomalies of the physical fermions .
fortunately , this is not the case . indeed , as we have already seen in the continuum theory , fermions with a nonlinearly realized chiral symmetry do not contribute to anomalies . instead , even in the presence of fermion fields , the anomalies
are still contained in goldstone - wilczek and wess - zumino - witten terms . consequently
, these terms should also be included in the lattice action .
it is an interesting problem to construct lattice versions of goldstone - wilczek and wess - zumino - witten terms maintaining the topological properties of the continuum theory .
since these terms do not play a central role in the applications discussed in the remainder of the paper , we defer this problem to future investigations .
it should be noted that the low - energy effective theory without the addition of goldstone - wilczek and wess - zumino - witten terms , although consistent , does not correctly describe the low - energy physics of the underlying microscopic theory with the anomalies .
we also would like to point out that the topological terms give rise to complex action problems which may severely affect the efficiency of numerical simulations .
in this section we describe how non - perturbative calculations can be performed in low - energy effective theories on the lattice .
in particular , we discuss how continuum physics can be extracted .
as we have discussed in the introduction , non - perturbative problems indeed exist @xcite but have so far not required a lattice .
this is likely to change when more complicated effective theories describing , for example , nuclear matter @xcite are studied . for simplicity ,
we first describe how the pure pion low - energy effective theory can be treated on the lattice .
it should be noted that , at present , there is no need for such calculations .
the only non - perturbative problem we are aware of in this context is the @xmath35 chiral rotor which has been solved analytically @xcite .
the leading term in the chiral lagrangian for pions reads @xmath202
= \int d^4x \
\big\{\frac{f_\pi^2}{4 } \mbox{tr}[\partial_\mu u^\dagger \partial_\mu u ] - \frac{\langle \overline \psi \psi \rangle}{2 } \mbox{tr}[{\cal m } u^\dagger + { \cal m}^\dagger u]\big\}.\ ] ] it is straightforward to discretize this theory .
a simple lattice action reads @xmath203 = - \frac{f_\pi^2}{4 } \sum_{x,\mu } \mbox{tr}[u_x^\dagger u_{x+\hat\mu } + u_{x+\hat\mu}^\dagger u_x ] - \frac{\langle \overline \psi \psi \rangle}{2 } \sum_x \mbox{tr}[{\cal m } u_x^\dagger + { \cal m}^\dagger u_x].\ ] ] of course , the bare lattice parameters @xmath18 and @xmath81 get renormalized .
a basic statement of the low - energy effective theory is that , to leading order , the dynamics of pions depends only on the renormalized parameters @xmath204 and @xmath205 . in the chiral limit , @xmath206
, the above lattice action has a phase transition at a critical bare coupling @xmath207 . for @xmath6 ,
the transition is second order , while for @xmath104 , it is expected to be first order .
for @xmath208 , one is in the symmetric phase without goldstone bosons , while for @xmath209 , one is in the broken phase in which the goldstone bosons govern the low - energy physics .
a non - perturbative calculation in the low - energy effective theory can be performed anywhere in the broken phase . in particular , there is no need to approach a second order phase transition in order to extract continuum physics .
this is because goldstone bosons are naturally massless without fine - tuning .
however , the renormalized parameter @xmath204 will in general stay at the cut - off scale . still , the low - energy physics ( which takes place far below the lattice cut - off ) is universal and depends on the details of the lattice action only through @xmath204 . in practical lattice calculations , one can choose any convenient bare coupling @xmath209 in the broken phase and determine the corresponding renormalized coupling @xmath204
. then one can choose @xmath210 such that the physical pion mass @xmath211 is correctly reproduced ( in units of @xmath204 ) .
after the low - energy parameters have been fixed to experiment , any non - perturbative calculation of a low - energy phenomenon can be performed in the lattice effective theory .
remarkably , at leading order , no lattice artifacts arise although @xmath204 is at the cut - off scale .
of course , in the lattice theory lorentz invariance is reduced to discrete translations and rotations . hence , there are terms in the low - energy theory of the lattice model which are forbidden by lorentz symmetry but not by the discrete rotation subgroup .
such terms are absent at leading order and arise only in the higher orders of the low - energy expansion .
their prefactors are additional low - energy parameters ( besides the usual gasser - leutwyler coefficients ) which must be tuned such that , to the given order of the momentum expansion , the low - energy theory is free of lattice artifacts .
non - perturbative calculations in the baryon effective theory can be performed in a similar way .
the effective theory correctly describes the low - energy pion dynamics in the sector with one baryon . in this case , the bare parameters @xmath79 and @xmath80 ( for @xmath6 ) or @xmath96 and @xmath95 ( for @xmath8 ) , as well as @xmath18 and @xmath210 need to be tuned such that the corresponding renormalized parameters match the experiment . just like @xmath204 ,
now also the baryon mass @xmath212 stays at the cut - off scale .
this is no problem because at low energies the results of the effective theory are again universal and depend on the lattice details only through the values of the low - energy constants .
there is no doubling problem as long as the mass of the doubler fermions is outside the energy range of the effective theory .
in particular , there is no need to approach a second order phase transition in order to make the physical baryon much lighter than the doubler fermions .
instead , by including higher order terms , one can push the doubler fermions to higher energies and thus extend the validity range of the effective theory .
more practical aspects of simulations of lattice theories with nonlinearly realized chiral symmetries will be discussed in the following sections .
nonlinear realizations of lattice chiral symmetry may also be useful beyond the strict validity range of low - energy effective theories . in this section
we discuss baryons in the static limit , @xmath213 , where interesting simplifications arise .
in particular , one can then perform simulations at non - zero baryon chemical potential without encountering a complex action problem .
we first discuss the problem in the continuum and then formulate the static baryon model on the lattice . within this model
we propose exploratory studies of the qcd phase diagram at non - zero baryon chemical potential . in the continuum formulation ,
static baryons at an undetermined position @xmath214 ( with @xmath215 ) are described by the spatial integral of the polyakov loop @xmath216 = \int d^3x \ \mbox{tr } \ { \cal p } \exp[\int_0^\beta dt \
v_4(\vec x , t)]\ ] ] of the flavor `` gauge '' field @xmath66 which is a function of @xmath217 .
here @xmath218 denotes path ordering and @xmath219 is the inverse temperature .
the partition function of pions in the background of a single static baryon then takes the form @xmath220 \exp(- s[u ] ) \exp(- \beta m),\ ] ] where the baryon mass @xmath79 will ultimately be sent to infinity and @xmath119 $ ] is the pure pion action given in ( [ pupi_cont ] ) .
similarly , the pion partition function in the presence of a single static antibaryon at an undetermined position is given by @xmath221^ *
\exp(- s[u ] ) \exp(- \beta m).\ ] ] in the next step , we consider a system of pions in the background of @xmath222 static baryons and @xmath223 static antibaryons which is described by the partition function @xmath224^n \frac{1}{\overline n ! } ( \phi[u]^*)^{\overline n } \exp(- s[u ] ) \exp[- \beta m
( n + \overline n)].\ ] ] the permutation factors @xmath225 and @xmath226 take into account the indistinguishability of baryons and antibaryons , respectively . introducing the baryon chemical potential @xmath11 that couples to the baryon number @xmath227 , we obtain the grand canonical partition function @xmath228 \nonumber \\ &
= & \sum_{n,\overline n } \int { \cal d}u \
\frac{1}{n ! } \phi[u]^n
\frac{1}{\overline n ! } ( \phi[u]^*)^{\overline n } \exp\{- s[u ] - \beta n(m - \mu ) - \beta \overline n(m + \mu)\ } \nonumber \\ & = & \int { \cal d}u \ \exp\{- s[u ] + \exp[- \beta ( m - \mu ) ] \phi[u ] + \exp[- \beta ( m + \mu ) ] \phi[u]^*\}. \nonumber \\ \\end{aligned}\ ] ] the present calculation for baryons is consistent only for @xmath229 .
hence , in order to obtain a non - trivial result , we also send @xmath11 to infinity keeping the difference @xmath230 finite .
then the partition function reduces to @xmath231 + \exp[- \beta ( m - \mu ) ] \phi[u]\}.\ ] ] in the @xmath6 case , the nucleons transform in the fundamental representation of @xmath39 which has a real - valued trace .
consequently , the spatial integral of the corresponding polyakov loop @xmath216 = \int d^3x \
\mbox{tr}_f { \cal p } \exp[\int_0^\beta dt \ v_4(\vec x , t)],\ ] ] takes real values only . for @xmath8 ,
the trace in the fundamental representation is in general a complex number .
however , since the baryons transform in the adjoint representation , the appropriate polyakov loop averaged over spatial positions reads @xmath216 = \int d^3x \ \mbox{tr}_a { \cal p } \exp[\int_0^\beta dt \ v_4(\vec x , t ) ] = \int d^3x \ \big\{|\mbox{tr}_f { \cal p } \exp[\int_0^\beta dt \ v_4(\vec x , t)]|^2 - 1\big\},\ ] ] which is again real .
hence , in contrast to full qcd , both for @xmath6 and for @xmath8 no complex action problem arises . using similar methods , the thermodynamics of gluons in the background of static quarks
have been investigated @xcite . in that case ,
a complex action problem does arise but , at least in the potts model approximation to qcd , could be solved with a meron - cluster algorithm @xcite .
it is straightforward to formulate the static baryon model on the lattice . for @xmath6 , the spatial integral of the polyakov loop for a static nucleon
is given by @xmath216 = \sum_{\vec x } \mbox{tr}_f \prod_t v_{\vec x , t,4},\ ] ] and , for @xmath8 , the corresponding quantity for a static baryon takes the form @xmath216 = \sum_{\vec x } \mbox{tr}_a[\prod_t v_{\vec x , t,4 } ] = \sum_{\vec x } \big\{|\mbox{tr}_f \prod_t v_{\vec x , t,4}|^2 - 1\big\}.\ ] ] then , the lattice partition function corresponding to ( [ partmu ] ) reads @xmath232 + \exp[- \beta ( m - \mu ) ] \phi[u]\}\ ] ] with the pure pion lattice action @xmath119 $ ] given in ( [ pupi_latt ] ) .
it should be noted that we have ignored the pauli principle for static baryons occupying the same lattice site .
the effect of this approximation is a lattice artifact that does not affect the continuum limit .
the static baryon model can be used in exploratory studies of the qcd phase diagram at non - zero chemical potential . at @xmath233 ,
the pure pion lattice model has a finite temperature chiral phase transition , above which chiral symmetry is restored .
for two massless flavors , this transition is second order and it is in the universality class of the 3-dimensional @xmath234-model .
this transition extends into the @xmath13-plane and is expected to turn into a first order phase transition at a tricritical point .
when non - zero current quark masses are included , the second order phase transition line is washed out to a crossover and the tricritical point turns into a critical endpoint of the remaining line of first order phase transitions .
locating the critical endpoint in the @xmath13-plane is an important issue in heavy ion collision experiments @xcite which has been addressed using lattice methods @xcite .
unfortunately , in these studies one is limited to small lattices due to a severe complex action problem .
the small-@xmath11 region has also been investigated in @xcite .
one should not expect to be able to extract the precise location of the critical endpoint using the static baryon model presented here .
however , as long as the model has the correct symmetry properties , one can at least investigate universal features of this point , which is expected to be in the universality class of the 3-dimensional ising model .
moreover , one can gain experience with numerical methods for locating the critical endpoint in a model that is similar to full qcd but much simpler to simulate
. we would like to mention that the static baryon model shares all global symmetries with qcd , but is , in addition , also invariant under the intrinsic parity transformation @xmath118 . in the @xmath6 case ,
@xmath118 is just @xmath0-parity and hence a symmetry of qcd . in the @xmath8 case
, @xmath118 is not a symmetry of qcd . at the level of the effective theory , this symmetry
is explicitly broken by the wess - zumino - witten term in the goldstone boson sector and by the @xmath80 term in the baryon sector .
since we have not included the wess - zumino - witten term or the @xmath80 term in the lattice action , as it stands , the @xmath8 static baryon model has the undesirable @xmath118 symmetry . in a study of universal properties of the @xmath8 model
one may hence want to break @xmath118 explicitly by hand in order to exactly reproduce all symmetries of qcd . with light up and down quarks and a very heavy strange quark ,
the situation is essentially like in the @xmath6 case .
as one lowers the strange quark mass , the critical endpoint is expected to move toward smaller values of @xmath11 until it reaches the temperature axis . for even smaller strange quark masses ,
the chiral phase transition at @xmath233 is first order .
lattice calculations suggest ( but do not show unambiguously ) that , for nature s quark mass values , at @xmath233 there is merely a crossover . in that case
, the critical endpoint would indeed be located in the @xmath13-plane at @xmath235 .
while one can not determine the precise location of the critical endpoint in qcd from the static baryon model , it should be possible to obtain qualitative ( and perhaps even semi - quantitative ) information concerning the location of this endpoint as a function of the strange quark mass .
in this section we use a nonlinearly realized lattice chiral symmetry in order to address non - perturbative questions concerning the confinement of constituent quarks . again
, we do not rely on the validity of a systematic low - energy expansion .
the chiral quark model of georgi and manohar @xcite offers an explanation for why the non - relativistic quark model works . however , there are some potential non - perturbative problems which we suggest to investigate on the lattice .
the chiral quark model is formulated in terms of gluons , pions , and constituent quarks @xmath57 and @xmath58 with the euclidean action @xmath236&=&\int d^4x \
\big\{m \overline \psi \psi + \overline \psi \gamma_\mu ( \partial_\mu + v_\mu + g_\mu ) \psi + i g_a \overline \psi \gamma_\mu \gamma_5 a_\mu \psi \nonumber \\ & + & \frac{f_\pi^2}{4 } \mbox{tr}[\partial_\mu u^\dagger \partial_\mu u ] - \frac{\langle \overline \psi \psi \rangle}{2 } \mbox{tr}[{\cal m } u^\dagger + { \cal m}^\dagger u ] \nonumber \\ & + & \frac{1}{2 g_s^2 } \mbox{tr}[g_{\mu\nu } g_{\mu\nu}]\big\}.\end{aligned}\ ] ] here @xmath237 is the @xmath238 gluon field , and @xmath239 is the strong coupling constant between gluons and constituent quarks ( transforming in the fundamental representations of both @xmath238 color and @xmath35 flavor ) . under gauge transformations @xmath240 the gluon field transforms as @xmath241 g(x)^\dagger.\ ] ] as usual , the field strength takes the form @xmath242.\ ] ] we limit ourselves to @xmath243 .
the @xmath244 case is exceptional because @xmath39 is pseudo - real and hence the pattern of chiral symmetry breaking is different @xcite .
the chiral quark model of georgi and manohar is based on the assumption that the energy scale for chiral symmetry breaking is larger than the one for confinement . an effective description in terms of constituent quarks which receive their mass from chiral symmetry breaking before they get confined by residual low - energy gluons should then make sense .
in fact , the phenomenological success of the non - relativistic quark model may suggest that this picture is indeed correct . in their work ,
georgi and manohar provided a framework that puts the non - relativistic quark model on a solid field theoretical basis .
in particular , their chiral quark model holds the promise to share the successes of the non - relativistic quark model . for example , making reasonable assumptions about the confinement of constituent quarks , it describes baryon magnetic moments and hyperon non - leptonic decays with about 10 percent accuracy .
in addition , in contrast to the non - relativistic quark model , the georgi - manohar model is manifestly chirally invariant and is thus automatically consistent with the predictions of chiral perturbation theory .
georgi and manohar also pointed out some potential problems of their model .
for example , the constituent quarks may form a pseudo - scalar bound state in addition to the explicit pion introduced as a fundamental field .
if this bound state is light , one would double - count the pion .
georgi and manohar argued that this should not happen .
the pseudo - scalar constituent quark bound state should either be so heavy that it lies outside the range of the low - energy theory or it might represent the next physical pseudo - scalar state above the pion .
a similar issue arises for the baryons .
in addition to the baryons built from constituent quarks , one can form baryonic skyrmions from the fundamental pion field . again
, georgi and manohar argued that double - counting should not arise because the skyrmions may either be very heavy or unstable .
a potentially more severe problem is related to the confinement scale . in the chiral quark model the strong coupling constant
is fixed by adjusting the color hyperfine splitting in the baryon spectrum .
the estimated value of @xmath239 is so small that one might expect unacceptably low - lying glueball states . for the same reason
, there might be a low - temperature deconfinement phase transition in the gluon sector significantly below the finite temperature chiral phase transition .
these potential problems are impossible to address quantitatively in the continuum formulation of the chiral quark model because they involve the non - perturbative dynamics of confinement . in order to be able to address these issues ,
we formulate the constituent quark model on the lattice . due to the nonlinear realization of chiral symmetry , global axial transformations manifest themselves as pion field dependent ( and thus local ) transformations on the constituent quark fields .
in particular , the constituent quarks couple to the pions through the composite flavor `` gauge '' field . using the lattice construction of a nonlinearly realized chiral symmetry presented in section 3 , it is now straightforward to put the georgi - manohar model on the lattice .
the resulting action takes the form @xmath245&=&m
\sum_x \overline \psi_x \psi_x + \frac{1}{2 } \sum_{x,\mu } ( \overline \psi_x \gamma_\mu v_{x,\mu } g_{x,\mu } \psi_{x+\hat\mu } - \overline \psi_{x+\hat\mu } \gamma_\mu v_{x,\mu}^\dagger g_{x,\mu}^\dagger \psi_x ) \nonumber \\ & + & \frac{i g_a}{2 } \sum_{x,\mu } ( \overline \psi_x \gamma_\mu \gamma_5 a^l_{x,\mu } \psi_{x } + \overline \psi_{x+\hat\mu } \gamma_\mu \gamma_5 a^r_{x,\mu } \psi_{x+\hat\mu } ) \nonumber \\ & + & \frac{r}{2 } \sum_{x,\mu } ( 2 \overline \psi_x \psi_x - \overline \psi_x v_{x,\mu } g_{x,\mu } \psi_{x+\hat\mu } - \overline \psi_{x+\hat\mu } v_{x,\mu}^\dagger g_{x,\mu}^\dagger \psi_x ) \nonumber \\ & -&\frac{f_\pi^2}{4 } \sum_{x,\mu } \mbox{tr}[u_x^\dagger u_{x+\hat\mu } + u_{x+\hat\mu}^\dagger u_x ] - \frac{\langle \overline \psi \psi \rangle}{2 } \sum_x \mbox{tr}[{\cal m } u_x^\dagger + { \cal m}^\dagger u_x ] \nonumber \\ & -&\frac{1}{g_s^2 } \sum_{x,\mu,\nu } \mbox{tr } [ g_{x,\mu } g_{x+\hat\mu,\nu } g_{x+\hat\nu,\mu}^\dagger g_{x,\nu}^\dagger + g_{x,\nu } g_{x+\hat\nu,\mu } g_{x+\hat\mu,\nu}^\dagger g_{x,\mu}^\dagger].\end{aligned}\ ] ] here @xmath246 ( again taking @xmath243 ) denotes the standard color parallel transporters living on the lattice links , which transform under gauge transformations @xmath247 as @xmath248 in the classical continuum limit the lattice gauge field @xmath249 $ ] is related to the continuum gauge field @xmath250 that lives in the @xmath251 algebra . for @xmath215 , the dirac operator of the chiral quark model @xmath252&=&m \delta_{x , y } + \frac{1}{2 } ( \gamma_\mu v_{x,\mu } g_{x,\mu } \delta_{x+\hat\mu , y } - \gamma_\mu v_{x-\hat\mu,\mu}^\dagger g_{x-\hat\mu,\mu}^\dagger \delta_{x-\hat\mu , y } ) \nonumber
\\ & + & \frac{i g_a}{2 } ( \gamma_\mu \gamma_5 a^l_{x,\mu } \delta_{x , y } + \gamma_\mu \gamma_5 a^r_{x-\hat\mu,\mu } \delta_{x , y } ) \nonumber \\ & + & \frac{r}{2 } ( 2 \delta_{x , y } - v_{x,\mu } g_{x,\mu } \delta_{x+\hat\mu , y } - v_{x-\hat\mu,\mu}^\dagger g_{x-\hat\mu,\mu}^\dagger \delta_{x-\hat\mu , y}),\end{aligned}\ ] ] is @xmath190-hermitean , i.e. @xmath253^\dagger = \gamma_5 d[u , g ] \gamma_5 $ ] .
thus , its spectrum contains complex conjugate pairs of eigenvalues @xmath254 and @xmath255 .
only the real eigenvalues are in general not paired .
consequently , the fermion determinant can be negative only if there is an odd number of negative ( real and thus unpaired ) eigenvalues . in the presence of a sufficiently large constituent quark mass @xmath79 ,
negative eigenvalues are very much suppressed .
hence , the potential sign problem due to a negative fermion determinant is expected to be mild .
a similar situation arises when one wants to simulate an odd number of flavors of wilson fermions , for example , when one wants to include the strange quark in a qcd simulation . in this case
not directly addressing the potential sign problem algorithms have been developed to perform numerical simulations @xcite .
it is to be expected that similar methods can be applied to the action presented here .
unfortunately , for @xmath256 the dirac operator is in general no longer @xmath190-hermitean .
hence , we can no longer argue that there is only a mild sign problem . only numerical simulations can tell if a severe sign problem arises for realistic values of @xmath80 .
if this is the case , one might want to simulate at purely imaginary values of @xmath80 and attempt an analytic continuation to the real physical value .
however , in order to address the non - perturbative questions concerning the confinement of constituent quarks , it may be sufficient to work at @xmath215
. in the chiral limit ( @xmath206 ) and for @xmath257 the lattice georgi - manohar model has four bare parameters : the constituent quark mass @xmath79 , the axial coupling @xmath80 , the pion decay constant @xmath18 , and the strong gauge coupling @xmath239 .
the chiral quark model on the lattice can be treated similarly to the low - energy baryon effective theory . in particular , one need not necessarily approach a second order phase transition but can simulate anywhere in the chirally broken phase .
for example , one can tune the bare constituent quark mass @xmath79 and the bare pion decay constant @xmath18 such that both the renormalized nucleon mass @xmath258 and the renormalized pion decay constant @xmath204 stay at the cut - off with their ratio equal to the experimental value .
similarly , the bare @xmath80 can be tuned to reproduce the observed weak decay of the neutron and @xmath239 can be adjusted to the nucleon - delta mass splitting . with the bare parameters fixed in this way , one can ask if the low - energy physics in the gluon sector is consistent with experiment .
for example , it is interesting to investigate if there is a first order deconfinement phase transition at an unusually low temperature or if there are unacceptably light glueball states .
if potential problems of this kind do not arise , the chiral quark model may indeed be able to explain the phenomenological success of the non - relativistic quark model .
unlike the baryon effective theory , we do not expect the chiral quark model to represent a systematic low - energy expansion .
in particular , one should not expect that its low - energy gluon physics is universal and depends on @xmath79 , @xmath80 , @xmath18 , and @xmath239 only .
still , despite the fact that other details of the model are expected to influence the gluon sector , it is interesting to ask if the gluon dynamics can be modeled successfully in this way . since one can not expect to obtain precise results directly relevant to qcd
, one may ask if using the machinery of lattice field theory is at all justified . clearly , this is a matter of taste .
however , even if only qualitative insight into the success of the non - relativistic quark model can be gained , this may be interesting enough .
in addition , a lattice study of the georgi - manohar model would also provide valuable experience with dynamical fermion algorithms with exact chiral symmetry in a set - up that is much simpler than lattice qcd with ginsparg - wilson fermions .
it is interesting to ask if a nonlinearly realized chiral symmetry can also lead to a new approach to lattice qcd itself . using a linear realization of chiral symmetry and an explicit auxiliary pion field ,
similar questions have been addressed in @xcite . in this section
we investigate such questions by exploring the phase structure of the lattice chiral quark model as a function of its bare parameters @xmath79 , @xmath18 , and @xmath239 . for simplicity
, we put @xmath215 and @xmath257 . in the limit of infinite constituent quark mass , @xmath259 ,
the pions and gluons decouple from each other : the model splits into an @xmath238 yang - mills theory and an @xmath260 nonlinear @xmath261-model for the pions .
the pure gluon theory is expected to have no bulk phase transition in the bare gauge coupling @xmath239 and , as a consequence of asymptotic freedom , reaches its continuum limit as @xmath262 . for large values of @xmath263 , the pion field @xmath143
is ordered and chiral symmetry is spontaneously broken to @xmath49 , while for small values of @xmath263 , it is restored . at @xmath259
the critical coupling @xmath264 is independent of @xmath239 .
consequently , there is a horizontal line connecting the points @xmath265 and @xmath266 in the phase diagram of fig . 1 .
in the @xmath6 case , the phase transition is second order , while for @xmath8 , it is expected to be first order .
next we consider the limit of infinite bare coupling @xmath267 .
the pion field @xmath268 is then frozen to a space - time independent constant @xmath269 .
of course , this does not mean that @xmath183 is a constant .
this can be illustrated as follows .
if @xmath270 for all @xmath137 , all @xmath183 can be diagonalized by the same unitary transformation @xmath271 , i.e. @xmath272 . the matrix @xmath273 is a constant diagonal matrix @xmath274 up to a diagonal matrix @xmath275 obeying @xmath276 .
the hermitean matrices @xmath277 have diagonal elements @xmath278 and determinant 1 .
hence , we can now write @xmath279 such that indeed @xmath280 is independent of @xmath137 .
let us now consider the other fields in the @xmath267 limit .
the lattice field @xmath156 takes the form @xmath154 = w s_x s_{x+\hat\mu } w^\dagger = v_{x,\mu}.\ ] ] since @xmath156 is already in @xmath35 , the @xmath161 coset decomposition yields @xmath281 .
similarly , it is straightforward to show that @xmath282 .
we now perform the transformation @xmath283 on the constituent quark field @xmath284 as a consequence , @xmath285 so that ( for @xmath267 ) the action ( [ gmaction ] ) takes the form @xmath286&=&m \sum_x \overline \psi_x ' \psi_x ' + \frac{1}{2 } \sum_{x,\mu } ( \overline \psi_x ' \gamma_\mu g_{x,\mu } \psi_{x+\hat\mu } ' - \overline \psi_{x+\hat\mu } ' \gamma_\mu g_{x,\mu}^\dagger \psi_x ' ) \nonumber \\ & + & \frac{r}{2 } \sum_{x,\mu } ( 2 \overline \psi_x ' \psi_x ' - \overline \psi_x ' g_{x,\mu } \psi_{x+\hat\mu } ' - \overline \psi_{x+\hat\mu } ' g_{x,\mu}^\dagger \psi_x ' ) \nonumber \\ & -&\frac{1}{g_s^2 } \sum_{x,\mu,\nu } \mbox{tr } [ g_{x,\mu } g_{x+\hat\mu,\nu } g_{x+\hat\nu,\mu}^\dagger g_{x,\nu}^\dagger + g_{x,\nu } g_{x+\hat\nu,\mu } g_{x+\hat\mu,\nu}^\dagger g_{x,\mu}^\dagger ] .
\nonumber \\ \\end{aligned}\ ] ] remarkably , this is nothing but the standard wilson fermion action .
hence , in the limit @xmath267 , @xmath262 , and @xmath287 ( point @xmath288 in fig . 1 ) , we reach the usual continuum limit of lattice qcd .
this point is connected to the point @xmath96 at strong coupling ( @xmath289 ) by a line of second order phase transitions .
fine - tuning to this line , a massless `` pion '' emerges as a quark - antiquark bound state .
this `` pion '' is not the goldstone boson that results from chiral symmetry breaking , except in the continuum limit .
still , in wilson s approach to lattice qcd this particle represents the physical pion .
then , recovering chiral symmetry requires fine - tuning as well as taking the continuum limit .
of course , in the chiral quark model with a nonlinearly realized chiral symmetry , there is a genuine goldstone pion .
however , this particle decouples at @xmath290 .
it is tempting to try to avoid fine - tuning and to represent the physical pion by the goldstone boson of the nonlinearly realized chiral symmetry .
this would suggest to approach the qcd point @xmath288 from finite values of @xmath263 out of the chirally broken phase .
as one approaches the point @xmath288 , one should then avoid the critical surface connecting the points @xmath288 , @xmath96 , @xmath0 , and @xmath1 because otherwise one would double - count the pion .
however , it is not clear if one will obtain a new approach to qcd in this way .
for example , it is likely that the explicit goldstone pion decouples and that the qcd pion must still be assembled as a quark - antiquark bound state .
let us now consider the @xmath291 plane in the phase diagram .
then the gluon field is completely frozen to a pure gauge @xmath292 .
it should be noted that the continuum limit of lattice qcd is taken by sending @xmath262 in a controlled way , not by simply putting @xmath291 . in particular , at @xmath291 there is no gluon field and thus no confinement . in this case
, we transform the fermion field as @xmath293 , @xmath294 , so that @xmath295 hence , for @xmath291 the action ( [ gmaction ] ) reduces to @xmath296&=&m \sum_x \overline \psi_x ' \psi_x ' + \frac{1}{2 } \sum_{x,\mu } ( \overline \psi_x ' \gamma_\mu v_{x,\mu } \psi_{x+\hat\mu } ' - \overline \psi_{x+\hat\mu } ' \gamma_\mu v_{x,\mu}^\dagger \psi_x ' ) \nonumber \\ & + & \frac{i g_a}{2 } \sum_{x,\mu } ( \overline \psi_x ' \gamma_\mu \gamma_5 a^l_{x,\mu } \psi_{x } ' + \overline \psi_{x+\hat\mu } ' \gamma_\mu \gamma_5 a^r_{x,\mu } \psi_{x+\hat\mu } ' ) \nonumber \\ & + & \frac{r}{2 } \sum_{x,\mu } ( 2 \overline \psi_x ' \psi_x ' - \overline \psi_x ' v_{x,\mu } \psi_{x+\hat\mu } ' - \overline \psi_{x+\hat\mu } ' v_{x,\mu}^\dagger \psi_x ' ) \nonumber \\ & -&\frac{f_\pi^2}{4 } \sum_{x,\mu } \mbox{tr}[u_x^\dagger u_{x+\hat\mu } + u_{x+\hat\mu}^\dagger u_x ] - \frac{\langle \overline \psi \psi \rangle}{2 } \sum_x \mbox{tr}[{\cal m } u_x^\dagger + { \cal m}^\dagger u_x ] , \nonumber \\ \\end{aligned}\ ] ] which has exactly the same form as the baryon action ( [ latact2 ] ) .
of course , here the field @xmath145 describes a colored constituent quark .
since we have now switched off the strong interactions , these quarks are no longer confined . for @xmath6 , a possible form of the phase diagram in the @xmath297-plane ( at @xmath291 )
is depicted in fig .
it also includes the aoki phase not shown in fig . 1 .
in the limit of infinite constituent quark mass , @xmath259 , the model reduces to an @xmath298 nonlinear @xmath261-model for the pions . for @xmath299 , the pion field @xmath143
is ordered and hence chiral symmetry is spontaneously broken to @xmath300 , while for @xmath301 , it is restored .
it is interesting to consider negative values of @xmath263 as well ( still keeping @xmath259 ) . on a bi - partite lattice
one can map the theory back to the one with positive values of @xmath263 by changing the sign of @xmath143 on all odd lattice sites .
hence , for @xmath302 , there is an `` antiferromagnetically '' ordered phase in which chiral symmetry is again broken from @xmath234 to @xmath303 .
hence , for @xmath259 , there are two phase transitions , one at @xmath264 separating the usual `` ferromagnetic '' chirally broken from the `` paramagnetic ''
chirally symmetric phase and another one at @xmath304 separating the `` paramagnetic '' from the `` antiferromagnetic '' broken phase . at @xmath305 and @xmath267 ( point @xmath288 in figs . 1 and 2 ) there is a second order phase transition at which the constituent quark mass vanishes .
in addition , there are unphysical second order phase transitions at @xmath306 ( with @xmath307 ) at which some of the doubler fermions become massless . in the context of wilson s lattice qcd the region of @xmath308 was explored by aoki @xcite .
he identified a rich structure including a phase in which parity and flavor are spontaneously broken .
it should be noted that for @xmath215 the model ( [ lataction ] ) has a symmetry under the replacement of @xmath79 by @xmath309 .
this replacement can be compensated for by changing the sign of @xmath145 on all even sites and the sign of @xmath146 on all odd sites .
this symmetry manifests itself in the phase diagram of fig . 2 .
our sketch of the aoki phase in fig .
2 is inspired by its form in wilson s lattice qcd .
the precise structure of the phase diagram can only be inferred from numerical simulations .
the phase diagram of fig . 2 is a possible interpolation between the limits @xmath259 and @xmath267 , partly inspired by the phase diagram of the smit - swift model and other related higgs - yukawa models @xcite .
attempts to construct chiral gauge theories as the continuum limit of such lattice models have failed because it was impossible to remove the doubler fermions , both at weak and at strong yukawa couplings @xcite .
such problems do not affect our construction of lattice effective theories with nonlinearly realized chiral symmetry because , in that case , there is no need to approach a second order phase transition to extract low - energy continuum physics .
however , attempts to construct full qcd with a manifest nonlinearly realized chiral symmetry on the lattice are likely to fail for the same reasons as the smit - swift model did . still , it is interesting to ask if the lattice chiral symmetry of ginsparg - wilson fermions can be realized nonlinearly .
we leave such questions for future investigations .
in this paper we have constructed lattice theories with baryon or constituent quark fields on which chiral symmetry is nonlinearly realized . in this framework
the fermion doublers are removed by the wilson term without breaking chiral symmetry explicitly .
this is possible because fermion fields do not contribute to anomalies in theories with nonlinearly realized chiral symmetries .
instead the anomalies must be included explicitly in the form of wess - zumino - witten and goldstone - wilczek terms .
the lattice construction of such terms remains an interesting challenge for future investigations .
the use of nonlinearly realized chiral symmetry is well established in the context of chiral perturbation theory which provides a power counting scheme for a systematic low - energy expansion .
however , also non - perturbative problems may arise in the low - energy regime .
some of these problems can be dealt with analytically @xcite . in general , however , such problems require some numerical work . while one can still proceed without monte carlo methods in investigations of few - nucleon systems @xcite , this is likely to change as one moves on to larger nuclei and ultimately to nuclear matter @xcite .
our lattice formulation provides a framework in which future numerical studies of this kind can be performed on a solid theoretical basis .
even if one leaves the framework of a systematic low - energy expansion , the nonlinear realization of chiral symmetry on the lattice may provide insight into interesting physical questions .
for example , as we have seen , in a model of static baryons at non - zero chemical potential one can investigate universal aspects of the qcd phase diagram , like the nature of the critical endpoint of the chiral phase transition line . in the static baryon approximation
, one can not expect to obtain results directly relevant to qcd for non - universal quantities such as the location of the critical endpoint .
still , one may be able to collect , for example , qualitative information on how the critical endpoint moves as the strange quark mass is varied .
another interesting example is the chiral quark model of georgi and manohar .
it holds the promise to explain the success of the non - relativistic quark model and has successfully described a variety of physical properties of hadrons . however , there is a number of potential dynamical problems that can only be addressed non - perturbatively .
our lattice formulation allows us to investigate if these potential dynamical problems do arise or not .
in particular , the confinement of constituent quarks out of reach of conventional methods can be directly addressed with lattice methods .
these investigations will allow one to decide if the chiral quark model indeed provides a satisfactory explanation for the success of the non - relativistic quark model .
along the way , one will be able to sharpen the algorithmic tools of dynamical fermion simulations in a model that shares many features with qcd most important a manifest chiral symmetry but which should be much simpler to simulate than ginsparg - wilson fermions in the chiral limit . at this stage
, it is not clear if lattice theories with a nonlinearly realized chiral symmetry can also provide a formulation of full qcd beyond low - energy effective theories . in order to answer this question ,
one may attempt to relate lattice qcd with ginsparg - wilson fermions to a lattice theory with explicit pion fields and a nonlinearly realized chiral symmetry .
if such a connection can be found analytically , one would expect that the anomalies of ginsparg - wilson fermions manifest themselves as wess - zumino - witten and goldstone - wilczek terms of the lattice pion field .
this is an interesting project for future studies which holds the promise to further deepen our understanding of the non - perturbative regularization of chiral symmetry .
we would like to thank r. c. brower , g. colangelo , j. gasser , p. hasenfratz , j. jersak , h. leutwyler , m. lscher , a. manohar , and f. niedermayer for helpful discussions .
this work was supported in part by funds provided by the u.s .
department of energy ( d.o.e . ) under cooperative research agreement doe - fg02 - 96er40945 , by the schweizerischer nationalfond ( snf ) , by the european community s human potential programme under contract hprn - ct-2000 - 00145 , and by the german academic exchange service ( daad ) . | we formulate lattice theories in which chiral symmetry is realized nonlinearly on the fermion fields . in this framework
the fermion mass term does not break chiral symmetry .
this property allows us to use the wilson term to remove the doubler fermions while maintaining exact chiral symmetry on the lattice .
our lattice formulation enables us to address non - perturbative questions in effective field theories of baryons interacting with pions and in models involving constituent quarks interacting with pions and gluons .
we show that a system containing a non - zero density of static baryons interacting with pions can be studied on the lattice without encountering complex action problems . in our formulation
one can also decide non - perturbatively if the chiral quark model of georgi and manohar provides an appropriate low - energy description of qcd .
if so , one could understand why the non - relativistic quark model works .
addtoresetequationsection _ keywords _ : chiral quark model , chiral symmetry , lattice qcd , low - energy effective theories , non - zero baryon chemical potential _ pacs numbers _ : 11.30.rd , 12.38.gc , 12.38.mh , 12.39.fe |
a mono - layer of graphite sheet , called graphene , has attracted growing interests recently.@xcite graphene exhibits a dirac cone with a linear dispersion at the corner of the first brillouin zone , resulting in a variety of novel transport phenomena of electrons .
they stimulate theoretical and experimental studies taking account of analogy to physics of relativistic electron , such as klein tunneling@xcite and zitterbewegung@xcite .
moreover , semi - infinite graphene and finite stripe of graphene with zigzag edges support a novel edge state with nearly flat dispersion.@xcite on the contrary , armchair edge does not support such an edge state .
the flat dispersion implies that the density of state ( dos ) diverges at the flat band energy , in a striking contrast to the zero dos in bulk .
so far , theoretical investigation of graphene heavily relies on the tight - binding model .
it is not clear to what extent the edge states change in other models on the honeycomb lattice . here , we study a photonic analog of graphene model,@xcite namely , two - dimensional photonic crystal ( phc ) composed of the honeycomb lattice of dielectric cylinders embedded in a background substance .
the honeycomb lattice consists of two inter - penetrating triangular lattices ( called a and b sub - lattices ) with the same lattice constant . in phc
it is not rare to have the dirac cone in the dispersion diagram .
triangular and honeycomb lattices of identical circular rods support multiple dirac cones at the corner of the first brillouin zone .
it should be noted that they have the same point group of six - fold symmetry .
doubly degenerate modes at the corner of the first brillouin zone exhibit the dirac cone owing to the point group symmetry.@xcite this fact suggests that , the symmetry is crucial and the dirac cone is not limited in the tight - binding model of electrons on the honeycomb lattice .
some perturbation breaks the symmetry of the original honeycomb lattice and causes a crucial influence on the linear dispersion . in electronic systems
the energy difference between a- and b - site atomic orbitals,@xcite periodic magnetic flux of zero average,@xcite and rashba spin - orbit interaction@xcite are such examples of the symmetry breaking .
they break at least either of time - reversal symmetry ( trs ) or space - inversion symmetry ( sis ) or parities in plane .
therefore , the point group of the original honeycomb lattice is reduced into a smaller group . as a result
, the two - dimensional irreducible representations are prohibited , and the doubly - degenerate modes are lifted .
the gap between the lifted modes is correlated with the magnitude of the symmetry breaking . the effective theory around a nearly degenerate point
is described by the massive dirac hamiltonian , where the mass gap can be controlled via the degree of the symmetry breaking . in the honeycomb lattice
phcs the trs is efficiently broken by applying a magnetic field parallel to the cylindrical axis .
nonzero static magnetic field induces imaginary off - diagonal elements in the permittivity or permeability tensors , through the magneto - optical effect .
the sis is broken if the a - site rods are different from the b - site rods.@xcite therefore , we can continuously tune the degree of the symmetry breaking in the honeycomb lattice phcs .
this tunability is a great advantage of the photonic analog of graphene model and its extension.@xcite from a theoretical point of view , the tight - binding approximation , which is commonly used in modeling of graphene , is not widely applicable for photonic band calculation .
accordingly , the non - bonding orbital in the nearest neighbor tight - binding approximation , which is responsible for flatness of the dispersion curve of the zigzag edge state , is completely absent in phcs .
for example , even in the original honeycomb phc with both the trs and the sis , the dispersion of the zigzag edge states is not flat because of the absence of the non - bonding orbital .
regarding the system with boundary , photonic system is quite distinct from electronic system . in the latter system
the electrons near fermi level are prohibited to escape to the outer region via the work function , _
i.e. _ , a confining potential , and the wave functions of the electrons are evanescent in the outer region
. therefore , to sustain an edge state , formation of the band gap in bulk is the minimum requirement . on the other hand , in the former system confining potentials for photon
are absent at the boundary .
energy of photon is always positive as in free space , and no energy barrier exists between the phc and free space .
the simplest way to confine photonic edge states in the phc is to utilize the light cone .
this restriction of the confinement makes photonic systems quit nontrivial in various aspects .
in particular , the topological relation between bulk and edge@xcite is of high interest in photonic systems without trs . in quantum hall system
nontrivial topology of bulk states leads to the emergence of chiral edge states , which are robust against localization effect .
the edge states play a crucial role in this system.@xcite recently , haldane and raghu proposed one - way light waveguide realized in phcs without trs.@xcite explicit construction of such waveguides is demonstrated by several authors.@xcite this paper also shades light to this topic , by using simpler structure than those demonstrated so far .
this paper is organized as follows .
section ii is devoted to present bulk properties of the phc with and/or without trs and sis . a numerical method to deal with edge states
is given in sec .
properties of zigzag and armchair edge states are investigated in detail in secs .
iv and v , respectively .
a one - way light transport along the edge of a rectangular - shaped phc is demonstrated in sec .
finally , summary and discussions are given in sec .
let us consider two - dimensional phcs composed of the honeycomb array of circular cylinders embedded in air .
the photonic band structure of the phcs with and without trs is shown in fig .
[ band ] for the transverse magnetic ( tm ) polarization . for comparison , the photonic band structure of the transverse electric ( te ) polarization
is also shown for the phc with trs .
the sis holds in all the cases .
( color online ) the photonic band structure of the honeycomb lattice phcs of dielectric cylinders embedded in air .
solid ( dashed ) line stands for the tm band structure of the phc with ( without ) trs .
the sis holds in both the cases .
the dielectric constant and radius of the cylinders are taken to be 12 and @xmath0 , respectively , where @xmath1 is the lattice constant .
the magnetic permeability of the cylinders is taken to be 1 for the phc with trs and is given by eq .
( [ rodmu ] ) for the phc without trs . for comparison ,
the te band structure of the phc with trs is shown by doted line .
, scaledwidth=45.0% ] here , the dielectric constants @xmath2 and radius @xmath3 of the a(b)-cylinders are taken to be 12 and @xmath0 , respectively . the magnetic permeability of the cylinders is taken to be 1 for the phc with trs , and has the tensor form given by @xmath4 for the phc without trs .
the first , second , and third rows ( columns ) stand for @xmath5 and @xmath6 cartesian components , respectively .
the cylindrical axis is taken to be parallel to the @xmath6 axis .
the imaginary off - diagonal components of @xmath7 are responsible for the magneto - optical effect and break the trs .
thus , parameter @xmath8 represents the degree of the trs breaking .
as mentioned in introduction , for the phc with trs the dirac cone is found at the k point . in particular ,
the first ( lowest ) and second tm bands are in contact with each other at the k point .
they are also in contact with the k point because of the spatial symmetry .
this property is quite similar to the tight - binding electron in graphene . as for the dirac point at @xmath9 of the tm polarization ,
the fourth band is in contact with the fifth band at k ( and k ) , whereas the former and the latter are also in contact with the third and sixth bands , respectively at the @xmath10 point .
concerning the te polarization , the dirac cones are not clearly visible , but are indeed formed between the second and third and between the fourth and fifth .
on the other hand , in the phc without trs , all the degenerate modes at @xmath10 and k are lifted .
the point group of this phc becomes @xmath11 and the point group of @xmath12 at the k point is @xmath13 .
they are abelian groups , allowing solely one - dimensional representations .
therefore , the degeneracy is forbidden .
the energy gap between the lifted modes is proportional to @xmath8 if it is small enough .
the sis breaking , @xmath14 , lifts the double degeneracy at k , but not at @xmath10 when the trs is preserved .
the energy gap between the lifted modes is proportional to @xmath15.@xcite let us focus on the gap between the first and second tm bands of the phc as a function of the sis and trs breaking parameters .
the phase diagram of the phc concerning the gap is shown in fig . [ phase ] .
phase diagram of the honeycomb lattice phcs for the tm polarization .
phase space is spanned by two parameters , @xmath16 and @xmath8 , which represent the sis and trs breaking , respectively .
the average dielectric constant and the radius of the cylinders are kept fixed to @xmath17 and @xmath18 , respectively .
, scaledwidth=45.0% ] at generic values of the parameters the gap opens . however , if we change the parameters along a certain curve in the parameter space , the gap remains to close as shown in fig .
[ phase ] .
this property implies that at finite @xmath8 the gap closes only at a certain value of @xmath15 .
although the gap opens in both the regions above and below the curve , the two regions are topologically different , and are characterized by the chern numbers of the first and second photonic bands .
the chern number is a topological integer defined by @xmath19 for each non - degenerate band . here ,
bz , uc , and @xmath20 stand for brillouin zone , unit cell , and the area of unit cell , respectively .
the envelop function @xmath21 of the @xmath22-th bloch state at @xmath12 is of @xmath23 ( _ i.e. _ , the @xmath6 component of the electric field ) . in the upper region of fig .
[ phase ] , @xmath24 and @xmath25 , whereas in the lower region @xmath26 . at the gap closing point , the chern number transfers between the upper and lower band under the topological number conservation law.@xcite we will see that the phase diagram correlates with a property of edge states in corresponding phc stripes . this correlation is a guiding principle to design a one - way light transport near phc edges.@xcite figure [ phase ] shows solely the phase diagram in the first quadrant of real @xmath15 and @xmath8 .
the mirror reflection with respect to the @xmath15 axis gives the phase diagram in the fourth quadrant , where @xmath27 and @xmath28 are interchanged due to the inversion of @xmath8 .
the phase diagram in the second and third quadrants is obtained by the mirror reflection with respect to the @xmath8 axis .
the resulting phase diagram is similar to that obtained in a triangular lattice phc with anisotropic rods.@xcite
so far , we have concentrated on properties of the phcs of infinite extent in plane .
if the system has edges , there can emerge edge states which are localized near the edges and are evanescent both inside and outside the phc . in this section
we introduce a phc stripe with two parallel boundaries .
the boundaries are supposed to have infinite extent , so that the translational invariance along the boundary still holds .
the edge states are characterized by bloch wave vector parallel to the boundary .
optical properties of the phc stripe are described by the s - matrix .
it relates the incident plane wave of parallel momentum @xmath29 to the outgoing plane wave of parallel momentum @xmath30 , where @xmath31 and @xmath32 are the reciprocal lattice vectors relevant to the periodicity parallel to the stripe.@xcite both the waves can be evanescent . to be precise ,
the s - matrix is defined by @xmath33 where @xmath34 is the plane - wave - expansion components of upward ( + ) and downward ( - ) incoming ( outgoing ) waves of parallel momentum @xmath30 , respectively . in our phcs under consideration
the s - matrix can be calculated via the photonic layer - korringa - kohn - rostoker method@xcite as a function of parallel momentum @xmath35 and frequency @xmath36 .
if the s - matrix is numerically available , the dispersion relation of the edge states is obtained according to the following secular equation : @xmath37.\end{aligned}\ ] ] strictly speaking , this equation also includes solutions of bulk states below the light line .
if we search for the solutions inside pseudo gaps ( _ i.e. _ @xmath35-dependent gaps ) , solely the dispersion relations of the edge states are obtained . in actual calculation , however , the magnitude of @xmath38 $ ] becomes extremely small with increasing size of the matrix .
the matrix size is given by the number of reciprocal lattice vectors taken into account in numerical calculation . in order to obtain numerical accuracy
, we have to deal with larger matrix .
therefore , this procedure to determine the edge states is generally unstable .
instead , we employ the following scheme .
suppose that the s - matrix is divided into two parts @xmath39 and @xmath40 that correspond to the division of the phc stripe into the upper and lower parts .
this division is arbitrary , unless the upper or lower part is not empty .
the following secular equation also determines the dispersion relation of the edge states : @xmath41.\end{aligned}\ ] ] this scheme is much stable for larger matrix .
as far as true edge states are concerned , the secular equation has the zeros in real axis of frequency for a given real @xmath35 .
here we should also mention leaky edge states ( _ i.e. _ , resonances near the edges ) , which are not evanescent outside the phc but are evanescent inside the phc .
such an edge state is still meaningful , because the dos exhibits a peak there .
the peak frequency as a function of parallel momentum @xmath35 follows a certain curve that is connected to the dispersion curve of the true edge states . to evaluate the leaky edge states , the method developed by ohtaka _ et al_@xcite is employed . in this method , the dos at fixed @xmath35 and @xmath36 is calculated with the truncated s - matrix of open diffraction channels .
the unitarity of the truncated s - matrix enables us to determine the dos via eigen - phase - shifts of the s - matrix .
a peak of the dos inside the pseudo gap corresponds to a leaky edge state .
first , let us consider the zigzag edge . figure [ pband ] shows four sets of the projected band diagram of the honeycomb phc and the dispersion relation of the edge states localized near the zigzag edges . in fig .
[ pband ] the shaded regions represent bulk states , whereas the blank regions correspond to the pseudo gap . inside the pseudo gap edge states
can emerge . in the evaluation of the edge states
, we assumed the phc stripe of @xmath42 , being @xmath43 the number of the layers along the direction perpendicular to the zigzag edges .
the projected band diagrams at point @xmath44 in the phase diagram ( fig . [ phase ] ) and
the dispersion curves of the edge states . the zigzag edge is assumed .
the shaded regions represent bulk states .
the edge states are of the phc stripes with 16 layers .
thin solid line stands for the light line .
the surface brillouin zone is taken to be @xmath45 in order to see the connectivity of the edge - state dispersion curves . , scaledwidth=45.0% ] here , we close up the first and second bands .
higher bands are well separated from the lowest two bands .
each set refers to either of four points indicated in the phase diagram of fig .
[ phase ] . in accordance with the dirac cone in fig .
[ band ] , the projected band structure of point @xmath46 also exhibits a point contact at @xmath47 and @xmath48 .
the first and second bands are separated for @xmath49 and @xmath50 , but are nearly in contact at @xmath47 for @xmath51 .
this is because @xmath51 is close to the phase boundary .
except for the lower right panel , in which the trs and the sis are broken , the projected band diagrams and the edge - state dispersion curves are symmetric with respect to @xmath52 .
this symmetry is preserved if either the trs or the sis holds .
the time - reversal transformation implies @xmath53 where @xmath54 is the eigen - frequency of the @xmath22-th bloch state at given parameters of @xmath15 and @xmath8 , and @xmath55 is the momentum perpendicular to the edge .
since @xmath8 is inverted , eq .
( [ t - transform ] ) is not a symmetry of the phc , but is just a transformation law . in the case of @xmath56 , after the projection concerning @xmath55 , the symmetry with respect to @xmath57 is obtained .
this symmetry combined with the translational invariance under @xmath58 results in the symmetry with respect to @xmath52 .
similarly , the space inversion results in @xmath59 the symmetry with respect to @xmath60 and @xmath61 is obtained at @xmath62 .
when edge states are well defined in phcs with enough number of layers , their dispersion relation satisfies @xmath63 owing to the time - reversal and space - inversion transformations , respectively . here ,
@xmath64 and @xmath65 denote the dispersion relation of opposite edges of the phc stripe . at @xmath56 , both @xmath64 and @xmath65
are symmetric under the inversion of @xmath35 .
in contrast , at @xmath62 they are interchanged .
the resulting band diagram is symmetric with respect to @xmath60 and @xmath61 as in fig .
[ pband ] .
the upper left panel of fig .
[ pband ] shows two almost - degenerate curves that are lifted a bit near the dirac point .
this lifting comes from the hybridization between edge states of the opposite boundary , owing to finite width of the stripe . the lifting becomes smaller with increasing @xmath43 , and eventually two curves merge with each other .
since @xmath46 corresponds to @xmath66 , we obtain @xmath67 owing to eqs .
( [ t - transform_edge ] ) and ( [ p - transform_edge ] ) , irrespective of @xmath35 .
as is the same with in graphene , our edge states appear only in the region @xmath68
. however , the edge - state curves are not flat , in a striking contrast to the zigzag edge state in the nearest - neighbor tight - binding model of graphene . in the upper right panel two edge - state curves
are separated in frequency and each curve terminates in the same bulk band . on the contrary , in the lower two panels the dispersion curves of the two edge states intersect one another at a particular point and each curve terminates at different bulk bands .
for instance , in the lower left panel , the curve including @xmath69 terminates at the upper band near @xmath47 and at the lower band near @xmath70 . at other points in the parameter space
, we found that the two edge - state curves are separated if the system is in the phase of zero chern number .
otherwise , if the system is in the phase of non - zero chern number , the two curves intersect one another .
the wave function of the edge state at marked points @xmath69 and @xmath71 is plotted in fig .
[ edge ] .
( color online ) the electric field intensity @xmath72 of the true edge state at @xmath69 ( left panel ) and @xmath71 ( right panel ) in fig .
[ pband ] .
the intensity maxima is normalized as 1 . in the enlarged panels the poynting vector flow
is also shown .
, scaledwidth=50.0% ] we can easily see that the edge states at @xmath69 and @xmath71 are localized near different edges . this property is consistent with the fact that at @xmath62 , @xmath64 and @xmath65 are interchanged under the inversion of @xmath35 . the field configuration at @xmath69 is identical to that at @xmath71 after the space - inversion operation ( @xmath73 rotation ) .
since the sis is preserved in this case , they are the sis partners .
it is also remarkable that the electric field intensity is confined almost in the rods forming one particular sub - lattice .
this field pattern is reminiscent of the non - bonding orbital of the zigzag edge state in graphene .
the edge state at @xmath74 has the negative ( positive ) group velocity .
moreover , no other bulk and edge states exist at the frequency .
therefore , solely the propagation from left to right is allowed near the upper edge , while the propagation from left to right is allowed in the lower edge . in this way
a one - way light transport is realized near a given edge .
the one - way transport is robust against quenched disorder with long correlation length , because the edge states are out of the light line and the bulk states at the same frequency is completely absent.@xcite this is also the case in the lower right panel of fig .
[ pband ] , although the frequency range of the one - way transport is very narrow .
it should be noted that the non - correlated disorder would cause the scattering into the states above the light line , where the energy leakage takes place .
detailed investigation of disorder effects is beyond the scope of the present paper .
the results obtained in this section strongly support the bulk - edge correspondence , which was originally proven in the context of quantum hall systems@xcite and was discussed in the context of photonic systems recently.@xcite namely , the number of one - way edge states in a given two - dimensional omni - directional gap ( _ i.e. _ @xmath35-independent gap ) is equal to the sum of the chern numbers of the bulk bands below the gap . in our case
the chern number of the lower ( upper ) band is equal to -1 ( 1 ) .
a negative sign of the sum corresponds to the inverted direction of the edge propagation .
accordingly , there is only one ( one - way ) state per edge in the gap between the first and second bands .
moreover , no edge state is found between the second and third bands .
this behavior is consistent with the chern numbers of the first and second bands , according to the bulk - edge correspondence .
finally , let us briefly comment on the edge states in @xmath46 and @xmath49 . in @xmath46
the two edge states are completely degenerate at @xmath75 . for the system with narrow width , there appear the bonding and anti - bonding orbitals , each of which has an equal weight of the field intensity in both the zigzag edges . as for the edge states in @xmath49
, the upper ( lower ) edge states are localized near the upper ( lower ) zigzag edge .
next , let us consider the armchair edge . the projected band diagram and the dispersion curves of the edge states are shown in fig .
[ pband_armchair ] .
we assumed the phc stripe with @xmath76 .
the projected band diagrams at point @xmath44 in the phase diagram ( fig . [ phase ] ) and
the dispersion curves of the edge states . the armchair edge is assumed .
the shaded regions correspond to bulk states .
the edge states are of the phc stripes with 64 layers .
thin solid line stands for the light line .
, scaledwidth=45.0% ] it should be noted that they are symmetric with respect to @xmath57 regardless of sis and trs .
this property is understood by the combination of a parity transformation and eq .
( [ t - transform ] ) . under the parity transformation with respect to the mirror plane parallel to the armchair edges , @xmath77 by combining eqs .
( [ t - transform ] ) and ( [ parity ] ) , we obtain the symmetric projected band diagram with respect to @xmath57 . concerning the edge states , the parity transformation results in @xmath78 therefore , by combining eqs .
( [ t - transform_edge ] ) and ( [ parity_edge ] ) , we can derive that @xmath64 and @xmath65 are interchanged by the inversion of @xmath35 , regardless of sis and trs : @xmath79 equation ( [ interchange ] ) results in the degeneracy between @xmath64 and @xmath65 at the boundary of the surface brillouin zone .
moreover , it is obvious from eq .
( [ parity_edge ] ) that the two edge states are completely degenerate at @xmath56 in the entire surface brillouin zone . in the armchair projection
the k and k points in the first brillouin zone are mapped on the same point @xmath57 in the surface brillouin zone , being above the light line .
therefore , possible edge states relevant to the dirac cone are leaky , unless the region outside the phc is screened .
accordingly , the dos of an armchair edge state at fixed @xmath35 shows up as a lorentzian peak , in a striking contrast to that of a zigzag edge state being a delta - function peak .
the dispersion relation of the leaky edge states depends strongly on the number of layers @xmath43
. however , if @xmath43 is large enough , the @xmath43-dependence disappears .
we found that at large enough @xmath43 , the leaky edge states correlate with the chern number fairly well . in the case as @xmath50 where the chern numbers of the upper and lower bands are nonzero , we found a segment of the dispersion curve of the leaky edge state whose bottom is at the lower band edge , as shown in the lower left panel of fig .
[ pband_armchair ] .
there also appear another segment of the dispersion curve which crosses the light line . across the phase boundary , the upper band touches to and separates from the lower band .
after the separation as the case @xmath80 , the bottom of the former segment moves from the lower band edge to the upper band edge as shown in the lower right panel of fig .
[ pband_armchair ] . by increasing @xmath15 ,
this segment hides among the upper bulk band ( not shown ) .
we should note that the dispersion curve of the leaky edge states is obtained by tracing the peak frequencies of the dos as a function of @xmath35 . if a peak becomes a shoulder , we stopped tracing the curve and indicated shoulder frequencies by dotted curve .
this is the case for @xmath50 and @xmath80 . for @xmath50 ,
the dos changes its shape from peak to shoulder at @xmath81 .
this is why the segment including @xmath82 and @xmath83 seems to terminate around there .
however , we can distinguish this shoulder in the region @xmath84 and crosses the light line . in the dos spectrum of @xmath80 , we can find two shoulders just below the peaks of bulk states in the region @xmath85 .
again , they merge each other and become an asymmetric peak for @xmath86 .
such an asymmetric peak consists of two peaks with different heights and widths , which come from the lifting of the degenerate edge states in the limit of @xmath56 .
actually , for @xmath46 and @xmath49 in which the edge states are doubly - degenerate , we can see a nearly - symmetric single peak for the leaky edge states in each case . as in the case of zigzag edge ,
the leaky edge states in the two - dimensional omni - directional gap exhibit a one - way light transport if the relevant chern number is nonzero . here
, we consider the structure with two horizontal armchair edges .
the incident wave with positive @xmath35 coming from the bottom can not excite the leaky edge state just above the lower band edge , _
e.g. _ , state @xmath83 in fig .
[ pband_armchair ] .
however , the incident wave with negative @xmath35 coming from the bottom can excite the leaky edge state , _
e.g. _ , at @xmath82 . in the latter case
, the leaky edge state has negative group velocity , traveling from right to left .
this relation becomes inverted for the plane wave coming from the top .
the incident plane wave with positive ( negative ) @xmath35 can ( can not ) excite the leaky edge state localized near the upper armchair edge .
this edge state has positive group velocity , traveling from left to right . in this way
, one - way light transport is realized as in the zigzag edge case . under quenched disorder
the one - way transport is protected against the mixing with bulk states , because no bulk state exists in the omni - directional gap . however , in contrast to the zigzag edge case , even the disorder with long correlation length could enhance the energy leakage to the outer region .
figure [ gkedgeconf ] shows the electric field intensity @xmath72 induced by the incident plane wave whose @xmath36 and @xmath35 are at the marked points ( @xmath82 and @xmath83 ) in fig .
[ pband_armchair ] .
the intensity of the incident plane wave is taken to be 1 and the field configuration above @xmath87 is omitted .
( color online ) the electric field intensity @xmath72 induced by the incident plane wave having @xmath88 at @xmath82 ( left panel ) and @xmath83 ( right panel ) in fig .
[ pband_armchair ] .
the incident plane wave of unit intensity comes from the bottom of the structure . in the enlarged panels the poynting vector flow
is also shown .
, scaledwidth=50.0% ] although , the dispersion curve is symmetric with respect to @xmath57 , the field configuration is quite asymmetric . of particular importance is the near - field pattern around the lower edge . in the left panel
the strongest field intensity of order 40 is found in the boundary armchair layer , whereas in the right panel it is found outside the phc with much smaller intensity . in both the cases ,
the transmittances are the same and nearly equal to zero .
accordingly , no field enhancement is observed near the upper edge ( not shown ) .
the remarkable contrast of the field profiles indicates that the leaky edge state with horizontal energy flow is excited in the left panel , but is not in the right panel .
if the plane wave is incident from the top , the field pattern exhibits an opposite behavior .
that is , the plane wave with @xmath36 and @xmath35 at @xmath83 from the top excites the leaky edge state localized near the upper edge , but at @xmath82 it can not excite the leaky edge state .
the property of each edge state is also understood as follows . when we scan @xmath35 from negative to positive along the dispersion curve of the leaky edge state , the localized center of the edge state transfers from one edge to the other .
the critical point is at the bottom of the dispersion curve , where the edge state merges to the bulk state of the lower band .
it is extended inside the phc , making a bridge from one edge to the other .
the entire picture is consistent with the interchange of @xmath64 and @xmath89 under the inversion of @xmath35 .
finally , let us comment on the field configuration of other edge states .
for @xmath46 and @xmath49 , the edge states are degenerate between the upper and lower edges . accordingly
, the incident plane wave coming from the bottom ( top ) of the structure excites the leaky edge states localized around the bottom ( top ) edge .
it is regardless of the sign of @xmath35 . for @xmath50 and @xmath80 , the edge - state curve that crosses the light line corresponds to an asymmetric peak in the dos , which is actually the sum of two peaks .
it is difficult to separate the two peaks , because they are overlapped in frequency .
thus , the edge states can be excited by the incident wave coming from both top and bottom of the phc . concerning the quadratic edge state around @xmath57 of @xmath80 , a similar contrast in the field configuration between positive and negative @xmath35
is obtained as in @xmath82 and @xmath83 . however , under quenched disorder this edge state readily mixes with bulk states that exist at the same frequency .
the direction of the one - way transport in the zigzag edge is consistent with that in the armchair edge .
let us consider a rectangular - shaped phc whose four edges are zigzag , armchair , zigzag , and armchair in a clockwise order .
the one - way transport found in figs .
[ pband ] and [ pband_armchair ] must be clockwise in this geometry . to verify it certainly happens
, we performed a numerical simulation of the light transport in the rectangular - shaped phc .
the multiple - scattering method is employed along with a gaussian beam incidence.@xcite we assume @xmath90 for the zigzag edges and @xmath76 for the armchair edges .
the incident gaussian beam is focused at the midpoint of the front armchair edge .
the electric field intensity @xmath72 at the focused point is normalized as 1 and the beam waist is @xmath91 .
the frequency and the incident angle of the beam are taken to be @xmath92 and @xmath93 , which corresponds to the leaky edge state very close to the @xmath82 point .
the beam waist size is chosen to avoid possible diffraction at the corner of the phc and not to excite the states near the @xmath83 point at the same time .
resulting electric field intensity @xmath72 is plotted in fig .
[ onewayconf ] .
( color online ) the electric field intensity @xmath72 induced by the time - harmonic gaussian beam coming from the left of the rectangular - shaped phc .
the four edges are either zigzag ( top and bottom ) or armchair ( left and right ) .
the beam is focused at the mid point of the left armchair edge with the unit electric field intensity and beam waist of @xmath91 .
, scaledwidth=45.0% ] the incident beam is almost reflected at the left ( armchair ) edge , forming the interference pattern in the left side of the phc .
however , as in the left panel of fig .
[ gkedgeconf ] , the leaky edge state is certainly excited there .
this edge state propagates upward , and is diffracted at the upper left corner .
a certain portion of the energy turns into the zigzag edge state localized near the upper edge .
this edge state propagates from left to right .
the energy leakage at the upper edge is very small compared to that in the left and right edges .
this zigzag edge state is more or less diffracted at the upper right corner .
however , the down - going armchair edge state is certainly excited in the right edge .
obviously , the field intensity of the right edge reduces with reducing @xmath94 coordinate .
this behavior is consistent with the energy leakage of the armchair edge state .
finally , the field intensity almost vanished at the lower right corner . in this way
, the clockwise one - way light transport is realized in the rectangular - shaped phc .
we also confirmed that the incident beam with the same parameters but inverted incident angle ( @xmath95 ) does not excite the counterclockwise one - way transport along the edges .
the incident beam is just reflected without exciting the relevant leaky edge state in accordance with the right panel of fig .
[ gkedgeconf ] .
in summary , we have presented a numerical analysis on the bulk and edge states in honeycomb lattice phcs as a photonic analog of graphene model and its extension . in the tm polarization
the dirac cone emerges between the first and second bands .
the mass gap in the dirac cone is controllable by the parameters of the sis or trs breaking . on a certain curve in the parameter space ,
the band touching takes place .
this curve divides the parameter space into two topologically - distinct regions .
one is characterized by zero chern number of the upper and lower bands , and the other is characterized by chern number of @xmath96 . of particular importance is the correlation between the chern number in bulk and light transport near edge .
non - zero chern number in bulk photonic bands results in one - way light transport near the edge .
it is quite similar to the bulk - edge correspondence found in quantum hall systems . in this paper
we focus on the tm polarization in rod - in - air type phcs .
this is mainly because the band touching takes place between the lowest two bands and they are well separated from higher bands by the wide band gap , provided that the refractive index of the rods are high enough . in rod - in - air type phcs the te polarization results in the band touching between the second and third bands .
however , the dirac cone is not clearly visible , although it is certainly formed . as for hole - in - dielectric type phcs
, an opposite tendency is found .
namely , the band touching between the lowest two bands takes place only in the te polarization . in this case
the distance between the boundary column of air holes and the phc edge affect edge states .
therefore , we must take account of this parameter to determine the dispersion curves of the edge states . concerning the trs breaking
, we have introduced imaginary off - diagonal components in the permeability tensor .
this is the most efficient way to break the trs for the tm polarization .
such a permeability tensor is normally not available in visible frequency range.@xcite however , in ghz range it is possible to obtain @xmath8 of order 10 .
such a large @xmath8 is necessary to obtain a robust one - way transport against thermal fluctuations , etc . in the numerical setup
we assume an intermediate frequency range with smaller @xmath8 .
on the other hand , in the te polarization , the trs can be efficiently broken by imaginary off - diagonal components in the permittivity tensor . in this case
the phc without the trs can operate in visible frequency range .
however , strong magnetic field is necessary in order to induce large imaginary off - diagonal components of the permittivity tensor .
thus , it is strongly desired to explorer low - loss optical media with large magneto - optical effect , in order to have robust one - way transport .
recently , another photonic analog of graphene , namely , honeycomb array of metallic nano - particles , is proposed and analyzed theoretically.@xcite particle plasmon resonances in the nano - particles act as if localized orbitals in carbon atom .
the tight - binding picture is thus reasonably adapted to this system , and nearly flat bands are found in the zigzag edge .
vectorial nature of photon plays a crucial role there , giving rise to a remarkable feature in the dispersion curves of the edge states in the quasi - static approximation .
in contrast , vectorial nature of photon is minimally introduced in our model , but a full analysis including possible retardation effects and symmetry - breaking effects has been made .
effects of the te - tm mixing in off - axis propagation are an important issue in our system .
in particular , it is interesting to study to what extent the bulk - edge correspondence is modified .
we hope this paper stimulates further investigation based on the analogy between electronic and photonic systems on honeycomb lattices . | this paper presents a theoretical analysis on bulk and edge states in honeycomb lattice photonic crystals with and without time - reversal and/or space - inversion symmetries .
multiple dirac cones are found in the photonic band structure and the mass gaps are controllable via symmetry breaking .
the zigzag and armchair edges of the photonic crystals can support novel edge states that reflect the symmetries of the photonic crystals .
the dispersion relation and the field configuration of the edge states are analyzed in detail in comparison to electronic edge states .
leakage of the edge states to free space is inherent in photonic systems and is fully taken into account in the analysis .
a topological relation between bulk and edge , which is analogous to that found in quantum hall systems , is also verified . |
large amounts of interstellar gas can be removed from the main disks of spiral galaxies by four main processes : tides due to the gravitational force of a companion , ram pressure stripping during a near - head - on collision between gas - rich galaxies , ram pressure stripping by intracluster gas , and galactic winds driven by supernovae . by ejecting processed gas into intergalactic space
, these mechanisms contribute to the metal - enrichment of the intergalactic medium . in many cases , which of these four processes is active in a given galaxy system
can be determined from the optical , radio , and/or x - ray morphology .
the signature of a tidal encounter between two galaxies is the presence of long stellar and/or hi tails and bridges ( e.g. , toomre @xmath13 toomre 1972 ) , while head - on collisions between gas - rich galaxies can produce ring galaxies ( e.g. , lynds @xmath13 toomre 1976 ; theys @xmath13 spiegel 1977 ) as well as gaseous bridges between the galaxies ( struck 1997 ) .
ram pressure stripping by intracluster gas leads to an hi deficiency ( e.g. , giovanelli @xmath13 haynes 1983 ) , truncation of the outer hi disk of a galaxy ( cayette et al .
1990 ) , and in some cases , bending of the hi disk ( kenney @xmath13 koopmann 1999 ) , but not usually removal of large quantities of molecular gas ( kenney @xmath13 young 1989 ) .
galactic winds are identified by extended extra - disk ionized gas without stellar counterparts ( e.g. , rand , kulkarni , @xmath13 hester 1990 ) .
star formation sometimes occurs in gas clouds far removed from the main disks of galaxies . of the four processes that can remove gas from disks , the best - known to trigger
the formation of young stars in the stripped gas is tidal forces : numerous examples of luminous h ii regions in tidal features have been found ( e.g. , schweizer 1978 ; mirabel et al .
1991 , 1992 ) .
young stars have also been found in extra - disk gas clouds thought to have been stripped during head - on collisions ( smith et al .
1999 ) and interstellar - intracluster encounters ( xu et al . 1999 ) . in order to better understand the processes that lead to the removal of gas from galaxies and the triggering of star formation in this gas
, it is important to make a complete inventory of the gas in these features .
this means not just the atomic gas , which has been surveyed in a large number of extragalactic tails and bridges ( e.g. , haynes et al .
1984 ; smith 1991 ; hibbard @xmath13 van gorkom 1996 ) , but also the molecular gas .
molecular gas has proved elusive in classical tidal tails ; in our earlier co survey of six tidal tails ( smith @xmath13 higdon 1994 ) , no co was found to very low levels , while only very low mass concentrations of molecular gas ( @xmath1410@xmath15 m@xmath16 ) were found in the tidal features of the nearby interacting system m81/m82/ngc 3077 ( brouillet , henkel , @xmath13 baudry 1992 ; walter @xmath13 heithausen 1999 ) .
in contrast to tidal features , gas removed from galaxy disks by ram pressure during head - on collisions between two gas - rich galaxies may be richer in co. the proto - type of this class of object is the co - rich gas concentration found outside the disk of the virgo cluster galaxy ngc 4438 by combes et al .
if the standard galactic n@xmath8/i@xmath9 ratio holds in this feature , it contains m@xmath8 @xmath10 10@xmath17 m@xmath16 .
the ring - like morphology of ngc 4438 and the proximity of the companion galaxy ngc 4435 suggests that this clump was removed during a head - on collision between the two galaxies ( kenney et al .
another apparent example of gas ram pressure - stripped during a galaxy - galaxy collision is the eastern tail of the peculiar galaxy ngc 2782 .
this feature has strong co emission , corresponding to 6 @xmath1 10@xmath18 m@xmath19 , assuming the galactic n@xmath8/i@xmath9 conversion factor ( smith et al .
1999 ) . based on morphological considerations
, we surmised that this feature was created during a near head - on collision between two galaxies , rather than a grazing encounter ( smith 1994 ) .
these results suggest that features produced in head - on collisions may differ in a fundamental way from classical tidal tails and bridges formed during grazing encounters : they may be richer in co. during a head - on collision , the material pulled out into a tail or bridge may originate in the inner disk of one of the galaxies , and so may be more metal - rich than gas pulled out from the outer disk in a more distant encounter .
the metallicity of the gas may affect the n@xmath8/i@xmath9 ratio , and therefore the detectability of the co line . both theoretical ( e.g. , maloney @xmath13 black 1988 ) and observational ( e.g. , wilson 1995 ; verter @xmath13 hodge 1995 ; arimoto , sofue , @xmath13 tsujimoto 1996 ) studies show that low metallicities can lead to enhanced n@xmath8/i@xmath9 ratios compared to the galactic value . in gas with low abundances and dust content
, ultraviolet radiation penetrates more deeply into a molecular cloud , causing a larger c@xmath20 region relative to the co core , increasing the n@xmath8/i@xmath9 ratio .
the conclusion that tidal gas differs from ram pressure stripped gas is quite uncertain , in part because of the very small sample size .
we have therefore continued our co survey of extragalactic tails and bridges , selecting systems with high hi column densities , strong star formation rates , and/or ring - like morphologies . in this paper
, we present new co data for tails and bridges in nine additional interacting galaxy systems . we compare these data with the previous results discussed above , as well as the new co data discussed in smith ( 2000 ) , braine et al .
( 2000 ) , and gao et al .
( 2000 ) . throughout this paper , we assume h@xmath21 = 75 km s@xmath11 mpc@xmath11 .
the co ( 1 @xmath0 0 ) observations were made using the 3 mm sis receiver on the national radio astronomy observatory ( nrao ) 12 m telescope during several observing runs between 1996 and 2000 .
two 256@xmath12 mhz filterbanks , one for each polarization , were used for the observations , providing a total bandpass of 1300 km s@xmath11 with a spectral resolution of 5.2 km s@xmath11 .
a nutating subreflector with an azimuthal beam throw of 3@xmath22 was used , taking care to avoid chopping on other galaxies , tails , or bridges in the system .
each scan was 6 minutes long .
the beamsize fwhm is 55@xmath4 at 115 ghz .
the pointing was checked periodically with bright continuum sources and was consistent to 10@xmath4 .
the system temperatures ranged from 200 to 500 k. calibration was accomplished using an ambient chopper wheel . a total of 45 positions in nine interacting systems were observed .
these nine systems are listed in table 1 , along with a brief description of their optical morphologies . in figure 1 , we display optical images of these systems , obtained from the digitized sky survey ( dss ) .
table 2 lists the observed positions in the galaxies and tails / bridges ; these positions and the co fwhm beamwidth are marked in figure 1 .
fourteen of the observed positions were in tails or bridges ; the rest were in the main disks of the galaxies .
a total of eleven tails and bridges were observed . in three systems ( arp 144 , ngc 2814/20 , and ngc 3628 ) , we observed multiple positions in a single tail or bridge .
table 2 also lists the central velocities of the observed bandpasses .
note that i zw 192 , ngc 3561b , and position a in stephan s quintet were all observed twice , with two different central velocities , to increase the observed bandpass ; these two sets of data were combined .
the results of the co observations are given in table 3 : the line fluxes , velocities , widths , and rms noise levels .
the summed spectra for each observed position are displayed in figure 2 .
note that the center of ngc 7828 and one of the tail positions in this system were previously observed by smith @xmath13 higdon ( 1994 ) ; the new and old data have been combined .
out of our nine interacting systems , the only unambiguous detection of co outside of the main disk of a galaxy is the star formation region in stephan s quintet . in two of our systems
( ngc 3395/6 and the taffy galaxies ) , the two galaxies in the pair are separated by only 70@xmath4 @xmath0 90@xmath4 , and the star formation regions in the bridges are unresolved from the main disks with our beamsize .
thus , although the bridge positions were detected in co , this may be emission from the galaxy disk in the 12 m beam .
higher resolution follow - up observations are needed to confirm the existence of co in these regions . in table 4 , we give the molecular gas mass for the galaxy disks , derived assuming the standard galactic n@xmath8/i@xmath9 ratio ( 2.8 @xmath1 10@xmath23 @xmath3/(k km s@xmath11 ) ; bloemen et al .
1986 ) and the emission fills the beam ( @xmath24@xmath25 = 0.82 ) . for the galaxies where more than one position in the disk was detected ,
the total co flux from the galaxy was obtained assuming a gaussian distribution . for the observed extra - disk positions in these systems ,
the molecular gas mass was derived under the same assumptions , and are recorded in table 5 .
although these assumptions may not hold in all cases , they provide a basis for comparison . in table 5
, we also provide the hi column densities averaged over the co beam , along with the implied m@xmath8/m@xmath26 ratios and the beam size @xmath27 , in kpc . for comparison , in table 6 ,
we give results for thirteen other tails , bridges , and extra - disk gas clouds previously observed in co. this includes structures in ngc 2782 ( smith @xmath13 higdon 1994 ; smith et al .
1999 ) , ngc 4438 ( combes et al .
1988 ) , ngc 4676 , ngc 7252 , and arp 143 ( smith @xmath13 higdon 1994 ) , arp 245 ( braine et al . 2000 ) , ngc 4410 ( smith 2000 ) , ngc 3561 ( braine et al .
2000 ) , m81 ( brouillet et al . 1992 ) and ngc 3077 ( walter @xmath13 heithausen 1999 ) , as well as the tentative ( 4@xmath28 ) detection of co in the southern tail of ngc 4038/9 ( gao et al . 2000 ) .
note that in this paper we present co data for the northern tail in the ngc 3561 system , while data for the southern feature are available from braine et al .
( 2000 ) . in tables 5 and 6
, we also give the h@xmath12 luminosities for these sources , when available , as well as the l@xmath29/m@xmath8 ratio .
for the sources in tables 5 and 6 which have hi data available , in figure 3 we plot the hi column density in the co beam n@xmath26 against m@xmath8/m@xmath26 .
for those with h@xmath12 fluxes also available , in figure 4 we compare l@xmath29/m@xmath8 with m@xmath8/m@xmath26 .
figures 3 and 4 show a wide range in l@xmath29/m@xmath8 and m@xmath8/m@xmath26 ratios for the features in our sample .
the most extreme cases are 1 ) the ngc 7714/5 bridge , with a very high hi column density but very low co flux , 2 ) the eastern feature of ngc 2782 , with a low l@xmath29/m@xmath8 ratio , 3 ) the molecular clump near ngc 4438 , with a very high implied m@xmath8/m@xmath26 ratio and a low l@xmath29/m@xmath8 value , and 4 ) the extra - disk star formation region in stephan s quintet , with a high l@xmath29/m@xmath8 ratio .
the global ratios for spiral galaxies are typically l@xmath29/m@xmath8 @xmath10 0.0025 @xmath0 0.1 l@xmath16/m@xmath16 and m@xmath8/m@xmath26 @xmath10 0.2 @xmath0 6 ( young et al .
the star formation region in stephan s quintet , the arp 245 feature , the ngc 4038/9 paper , and the upper limits to the ngc 4676 tail and the ngc 3561 features are consistent with these global ratios .
in contrast , the features in ngc 2782 and ngc 4438 have very low l@xmath29/m@xmath8 ratios , and the ngc 7714/5 bridge has a low implied m@xmath8/m@xmath26 ratio compared to global values for spirals .
dwarf galaxies typically have higher l@xmath29/m@xmath8 ratios and lower m@xmath8/m@xmath26 ratios than normal spirals , if the standard galactic conversion factor is applied .
for example , the 14 dwarf galaxies in the sage et al .
( 1992 ) sample have l@xmath29/m@xmath8 @xmath30 0.04 l@xmath16/m@xmath16 and m@xmath8/m@xmath26 @xmath14 0.5 , after converting to the standard galactic conversion factor , while all but two of the 25 dwarf galaxies studied by israel , tacconi , @xmath13 baas ( 1995 ) have m@xmath8/m@xmath26 ratios @xmath14 0.02 with this conversion factor .
the inferred molecular gas masses for dwarf galaxies may be low because their n@xmath8/i@xmath9 ratio is enhanced due to low metallicities ( maloney @xmath13 black 1988 ) .
note , however , that specific locations within dwarf galaxies may have high implied m@xmath8/m@xmath26 ratios ; for example , in ic 10 , at positions with n@xmath31 @xmath30 2 @xmath1 10@xmath2 @xmath3 , m@xmath8/m@xmath26 @xmath10 3 ( ohta , sasaki , @xmath13 saito 1988 ) . the tails and bridges in figure 3 that are undetected in co have upper limits to their m@xmath8/m@xmath26 ratios consistent with global values for dwarfs .
the upper limits for the ngc 3561 irregular , the ngc 7714 loop , the ngc 7715 tail , and the ngc 7714/5 bridge , are lower than typical global values for spirals ; the rest have less strict upper limits .
the subset of these undetected features with h@xmath12 fluxes available ( ngc 4676 , ngc 7714/5 , and the northern tail of ngc 3561 ) have l@xmath29/m@xmath8 lower limits consistent with values for both dwarfs and spirals .
however , the arp 245 feature , the stephan s quintet source , the ngc 4038/9 tail , the eastern ngc 2782 tail , and the ngc 4438 co clump are brighter in co relative to h@xmath12 than typical dwarf galaxies , while stephan s quintet , the eastern ngc 2782 tail , and the ngc 4438 co clump have higher inferred m@xmath8/m@xmath26 ratios than dwarfs .
the southern ngc 3561 feature has an m@xmath8/m@xmath26 ratio higher than that usually found in dwarfs , but just a lower limit to l@xmath29/m@xmath8 .
the most extreme feature in our sample in terms of inferred m@xmath8/m@xmath26 ratio is the bridge of ngc 7714/5 .
this bridge , which contains luminous h ii regions ( arp 1966 ; bernlhr 1993 ; gonzlez - delgado et al .
1995 ; smith et al . 1997 ) , was not detected in co , in spite of its high column density of atomic hydrogen ( 1.6 @xmath1 10@xmath2 @xmath3 in the co beam ; from the data in smith et al .
1997 ) . at this bridge ,
our upper limit implies n@xmath8 @xmath14 5.0 @xmath1 10@xmath32 @xmath3 and m@xmath8/m@xmath26 @xmath14 0.063 , more than three times smaller than our upper limit for the system with the second lowest m@xmath8/m@xmath26 ratio in a tail or bridge , the ngc 3561 irregular .
the hi column density in this bridge is above the h@xmath7 and co self - shielding limits for solar metallicity gas ( @xmath105 @xmath1 10@xmath23 @xmath3 and 10@xmath2 @xmath3 , respectively ; federman et al .
1979 ; van dishoeck @xmath13 black 1988 ) .
however , the nucleus of ngc 7714 is known to be low metallicity ( french 1980 ; garca - vargas et al .
1997 ) , so it is likely that the gas in the bridge would also have low abundances , leading to an enhanced n@xmath8/i@xmath9 ratio .
thus there may be more molecular gas in this system than implied by the co measurements .
we note that the mid - infrared camera on the infrared space observatory ( isocam ) did not detect this bridge ( ohalloran et al .
2000 ) , but it did detect the more distant extra - disk source in stephan s quintet ( xu et al .
this suggests a smaller warm dust component in the ngc 7714/5 bridge .
the smaller galaxy ngc 7715 , the eastern tail of ngc 7715 , and the hi loop north of ngc 7714 are also weak in co , with inferred m@xmath8/m@xmath26 ratios @xmath140.16 , @xmath140.16 , and @xmath140.24 , respectively .
these may also have enhanced n@xmath8/i@xmath9 ratios .
note that ngc 7715 , its tail , and the ngc 7714 loop have not been detected in h@xmath12 ( smith et al .
also note that ngc 7714 and ngc 7715 have absolute blue magnitudes of @xmath019.8 and @xmath018.1 , respectively ( de vaucouleurs et al . 1991 ) , compared to @xmath017.7 for the large magellanic cloud ( de vaucouleurs et al .
1991 ; sandage , bell , @xmath13 tripicco 1999 ) . at absolute magnitudes fainter than about @xmath019 , both irregular and spiral galaxies typically have less than solar metallicities ( skillman et al . 1989
; vila - costas @xmath13 edmunds 1992 ; storchi - bergmann et al .
1994 ) , thus ngc 7715 may also be metal - poor . at
the other extreme from the ngc 7714/5 bridge in terms of co brightness is the eastern tail of ngc 2782 and the extra - disk gas cloud near ngc 4438 , with low l@xmath29/m@xmath8 ratios compared to the other features ( see figure 4 ) .
these features have strong co emission , but little on - going star formation ( combes et al . 1988 ; kenney et al . 1995 ; smith et al . 1999 ) .
none of the features in our new sample are as rich in co as the ngc 2782 and ngc 4438 features .
as discussed at length in smith et al .
( 1999 ) , the gas in these features may be metal - rich material removed from the interiors of their disks by near - head - on collisions , leading to high co fluxes .
star formation may be inhibited in these features because of the collision ( smith et al .
gas clouds pushed out of a galactic disk by a high velocity collision may compress and then adiabatically expand , reducing their self - gravity .
this may decrease the rate of star formation and therefore the l@xmath29/m@xmath8 ratio . in figure 3 and
tables 5 and 6 , the structure with the highest hi column density in the co beam is the northern tail of the relatively nearby galaxy arp 245 , which has n@xmath26 = 2.3 @xmath1
10@xmath2 @xmath3 averaged over the 3.5 kpc beam ( duc et al . 2000 ) .
as discussed in duc et al .
( 2000 ) , this feature is relatively gas - rich and co - rich ( m@xmath8 @xmath10 1.5 @xmath1 10@xmath18 m@xmath16 , m@xmath26 @xmath10 4.8 @xmath1 10@xmath33 m@xmath16 , and m@xmath8/m@xmath26 @xmath10 0.31 in the co beam ) and has many properties in common with spiral galaxies .
it has an l@xmath29/m@xmath8 ratio similar to spirals , and a stellar population dominated by an underlying old population ( duc et al .
furthermore , it has a relatively high extinction ( a@xmath34 = 2.6 , from the balmer decrement ) , a high blue luminosity ( m@xmath34 = @xmath017.2 and l@xmath35 = 1.2 @xmath1 10@xmath17 l@xmath16 , uncorrected for internal extinction ) , and a metallicity of 12 + log(o / h ) @xmath10 8.6 ( duc et al . 2000 ) . with a modest correction for internal extinction ,
these values are consistent with the metallicity - luminosity relationship for late - type spirals ( zaritsky , kennicutt , @xmath13 huchra 1994 ; garnett et al .
thus it is possible that this structure may not be a tidal tail , but rather a pre - existing edge - on disk galaxy that is interacting with the other two galaxies in the system .
high resolution kinematic data may be useful in testing this hypothesis . in any case , this feature appears to be richer in co relative to hi and h@xmath12 than most dwarf galaxies .
the fact that the arp 245 feature was detected in co and many of the other features were not may be due to its high hi column density and relatively high metallicity .
there is some evidence that n@xmath8/i@xmath9 is correlated with metallicity ( wilson 1995 ; verter @xmath13 hodge 1995 ; arimoto et al .
the trends implied in these papers suggest that at the metallicity of the arp 245 feature , the n@xmath8/i@xmath9 ratio is enhanced only slightly , between 1.5 @xmath0 3 times bigger than the galactic value .
thus , at the high hi column density of the arp 245 feature , the co self - shielding limit is exceeded .
we note that ic 10 , which has a lower oxygen abundance that the arp 245 feature , shows strong co at positions where n@xmath26 exceeds 10@xmath2 @xmath3 ( ohta , sasaki , @xmath13 saito 1988 ) . at present , the metal abundances of most of the features listed in tables 5 and 6 are unknown .
the ngc 3561 system ( arp 105 ; figure 1d ) contains a spiral ( ngc 3561a ) , an s0 or e galaxy ( ngc 3561b ) , and two tidal features with on - going star formation , presumably pulled out from the spiral galaxy .
the concentration of gas and star formation in the northern tail is classified as a magellanic irregular by duc @xmath13 mirabel ( 1994 )
. the southern feature , which crosses the companion ngc 3561b , is visible as an optical knot in figure 1d , and is classified as a compact dwarf by duc @xmath13 mirabel ( 1994 ) .
this southern feature was detected in co by braine et al .
( 2000 ) , yielding a high m@xmath8/m@xmath26 ratio in the co beam of @xmath100.8 .
this ratio is higher than that of the arp 245 feature , in spite of its more modest hi column density in the co beam , @xmath10 2 @xmath1 10@xmath23 @xmath3 ( duc et al . 1997 ) and its lower oxygen abundance ( 12 + log(o / h ) @xmath10 8.4 ; duc @xmath13 mirabel 1994 ) .
this feature also has a lower blue luminosity ( m@xmath35 = @xmath016.9 ; l@xmath35 = 9.0 @xmath1 10@xmath18 l@xmath16 , uncorrected for internal extinction ) and a lower extinction ( a@xmath34 = 1.0 ) ( duc et al . 1994 ) than the arp 245 feature .
the northern tail of ngc 3561 ( the magellanic irregular ) has a peak hi column density four times higher and a blue luminosity five times larger than the southern tail ( duc et al .
1997 ) , and has a higher oxygen abundance 12 + log(o / h ) @xmath10 8.6 ( duc @xmath13 mirabel 1994 ) , but was undetected in co , with a lower m@xmath8/m@xmath26 ratio of @xmath140.2 .
this difference may be due in part to the difference in beam sizes of our co observations and the higher resolution observations of braine et al .
( 2000 ) , which are less affected by beam dilution . in the interaction scenario for this system presented by duc et al .
( 1997 ) , the spiral galaxy ngc 3561a interacts with the elliptical ngc 3561b , drawing out a long tail to the north , and a shorter countertail to the south .
since both features may have originated in ngc 3561a , it is reasonable to expect they would have similar gas properties .
the southern feature may have a relatively small angular size , high column density clump in which the self - shielding limits are exceeded . if the tentative detection of co in the ngc 4038/9 tail ( gao et al .
2000 ) is confirmed by more sensitive observations , it places this tail in the same category as the ngc 3561 and arp 245 features : it is richer in co than dwarf galaxies , relative to its h@xmath12 flux .
the metallicity in this feature has been estimated to be somewhat lower than that in the arp 245 feature , 12 + log(o / h ) @xmath10 8.4 ( mirabel et al .
1992 ) .
the tidal features of m81 and ngc 3077 have also been detected in co ( brouillet et al . 1992 ; walter @xmath13 heithausen 1999 ) . like those in arp 245 and ngc 3561 , they also have high inferred m@xmath8/m@xmath26 ratios ( table 6 ) .
however , the inferred h@xmath7 masses in the co beam are small ( @xmath1010@xmath36 m@xmath16 , using the standard galactic n@xmath8/i@xmath9 ratio ; 10@xmath37 m@xmath16 , using the virial theorem ) and there is no evidence for on - going star formation in these locations ( brouillet et al .
1992 ; henkel et al .
1993 ; walter @xmath13 heithausen 1999 ) .
the m81/m82/ngc 3077 group is very nearby , so the area subtended by the co beam ( @xmath10360 pc ) is much smaller than in the other systems listed .
the other features plotted in figure 3 have physical beam diameters ranging from 1.8 kpc ( ngc 4438 ) to 43 kpc ( i zw 192 ) . as with the arp 245 and ngc 3561 features
, the co self - shielding threshold may be exceeded locally in these features .
we note that the hi column density in the co beam is quite high for the m81 and ngc 3077 positions , 10@xmath2 @xmath3 .
a recent larger area co survey of the ngc 3077 tail ( walters @xmath13 heithausen 2000 ) shows co in two clumps covering an area of about 3@xmath22 ( 2.8 kpc ) . in this region ,
the standard galactic conversion factor gives m@xmath8 @xmath10 6 @xmath1 10@xmath36 m@xmath16 and an average h@xmath7 column density of @xmath108 @xmath1 10@xmath32 @xmath3 .
as in ic 10 , co is only detected at hi column densities above @xmath10 10@xmath2 @xmath3 . in this larger area ,
the inferred m@xmath8/m@xmath26 ratio is smaller , @xmath10 0.1 , consistent with the upper limits on the undetected tails and bridges in the sample . of the nine extra - disk locations in our new sample ,
only one was detected in co : the position in stephan s quintet ( hcg 92 ; arp 319 ) .
stephan s quintet is a compact group of four galaxies with similiar velocities ( @xmath106000 km s@xmath11 ; ngc 7317 , ngc 7318a / b , and ngc 7319 ) , and a probable foreground galaxy ( ngc 7320 ) with a much lower velocity ( @xmath10800 km s@xmath11 ) .
the outlying galaxy ngc 7320c also has a velocity of @xmath106000 km s@xmath11 , and so probably also belongs to this group .
strong gravitational interactions are clearly present in this group ( figure 1 g ) . a long optical tail stretches more than 1@xmath22 ( 26 kpc ) to the southeast of ngc 7319 ; this tail is also detected in hi ( shostak et al . 1984 ; verdes - montenegro et al . 2000 ) .
two more optical tails are seen extending to the north of the close pair ngc 7318a / b ( see figure 1 g ) .
along the easternmost of these features , x - rays ( pietsch et al .
1997 ) and radio continuum emission ( van der hulst @xmath13 rots 1981 ) are visible , suggesting a shock front . at the location
where the two optical tails of ngc 7318a / b intersect , a large concentration of ionized gas has been detected ( moles , sulentic , @xmath13 mrquez 1997 ; ohyama et al .
1998 ; xu et al . 1999 ) .
this source is also present in the optical dss image ( figure 1 g ) , in the mid - infrared map of xu et al .
( 1999 ) , and in hi ( shostak , sullivan , @xmath13 allen 1984 ; verdes - montenegro et al .
optical spectroscopy shows that the ionization mechanism for this source is young stars , while spectral signatures of shocks are present to the south , along the radio continuum and x - ray ridge ( moles , mrquez , @xmath13 sulentic 1998 ) . at the location of the extra - disk star formation region in stephan s quintet ,
we have detected two co components , at 6000 km s@xmath11 and 6700 km s@xmath11 , at the 5@xmath28 and 6@xmath28 significance level , respectively .
these two components have similar co fluxes ( table 3 ) .
the large velocity separation of these two features argues against the idea that they are caused by the double - horned signature of a rotating disk , but are instead two distinct gas clouds .
these two components each contain 1.0 @xmath1 10@xmath17 m@xmath16 of molecular gas , if the standard galactic n@xmath8/i@xmath9 ratio holds .
one of these components ( 6000 km s@xmath11 ) was detected at a 4@xmath28 level in the co interferometer data of gao @xmath13 xu ( 2000 ) .
we have also detected co in ngc 7319 and ngc 7318b .
ngc 7319 had previously been seen in co ( verdes - montenegro et al .
1998 ; yun et al . 1997 ) , however , the weaker co flux from 7318b was not seen in those less sensitive studies .
the co detection in ngc 7318b is at 6000 km s@xmath11 , a somewhat higher velocity than the optical velocity of this galaxy ( 5765 km s@xmath11 ; moles et al .
1998 ) .
both of the co components in the extra - disk star formation region have hi counterparts ( shostak et al .
they also both appear to have on - going star formation .
moles et al .
( 1998 ) obtained optical spectra of three h ii regions in this vicinity ; two have velocities of @xmath106020 km s@xmath11 and one has a velocity of 6680 km s@xmath11 .
two h@xmath12 velocity components are also indicated by the narrowband imaging of xu et al .
( 1999 ) .
the h@xmath12 luminosity , molecular gas mass , and hi mass for the extra - disk source in stephan s quintet are similar to those of many normal spiral galaxies ( young et al .
1996 ) , but unlike those of typical dwarf galaxies ( sage et al .
the implied m@xmath8/m@xmath26 ratios for both velocity components in this region are very high , @xmath100.65 , higher than values typically found for dwarf galaxies ( sage et al .
the l@xmath29/m@xmath8 ratio for the stephan s quintet source , 0.016 l@xmath16/m@xmath16 , is consistent with global values for spirals ( young et al . 1996 ) , as well as the values found for the ngc 4676 tail , the arp 245 feature , and the southern tail of ngc 3561 , but is higher than the values seen for the eastern tail of ngc 2782 and the ngc 4438 clump . in smith et al .
( 1999 ) , we suggested that star formation was inhibited in the eastern tail of ngc 2782 and the ngc 4438 co clump because of strong shocks sustained during a head - on collision between two gas - rich galaxies . if the extra - disk cloud in stephan s quintet was formed in this way , star formation is clearly not inhibited .
to explain the peculiar morphology of stephan s quintet , a number of different interaction scenarios have been suggested .
some of these scenarios involve ram pressure stripping during a head - on collision between two galaxies .
for example , peterson @xmath13 shostak ( 1980 ) suggest that a head - on collision between ngc 7318b and ngc 7319 removed the gas and caused the radio continuum ridge between them .
a different scenario was suggested by shostak et al .
( 1984 ) : a prograde tidal encounter of ngc 7320c and ngc 7319 occurred , pulling gas out of these galaxies , creating the long tail south of ngc 7319 .
this was followed by a high - speed collision between ngc 7318a and ngc 7318b .
other scenarios involve a collision between an ` intruder galaxy ' ( usually ngc 7318b ) and intragalactic gas that had previously been tidally - stripped or ram pressure - stripped in an earlier encounter between two different galaxies in the group ( van der hulst @xmath13 rots 1981 ; moles et al .
moles et al .
( 1997 ) suggest that a direct collision between the outlying galaxy ngc 7320c and ngc 7319 occurred in the recent past ( 10@xmath18 years ago ) , removing a large quantity of gas from these galaxies . after this event , ngc 7318b entered the group at high velocity , colliding with this stripped gas and triggering star formation .
our new data shed some light on the question of the origin of the extra - disk gas in stephan s quintet .
our detection of two co - rich components in the extra - disk star formation region shows that the star formation in this feature was triggered by a collision involving relatively high metallicity , molecule - rich gas concentrations , rather than metal - poor intracluster medium .
this does not rule out ram pressure stripping of ngc 7318b by an intergalactic medium , but it does require that this intergalactic medium be relatively metal - rich gas : gas previously removed from a galaxy . how this gas was originally removed from the disk is unclear : as shown in arp 245 and ngc 3561 , structures apparently created by tidal forces can be relatively co - rich .
thus this intergalactic gas may have been removed either by tidal forces , or via ram pressure stripping .
the 6700 km s@xmath11 extra - disk component may have originally come from ngc 7319 or ngc 7318a , which have similar optical velocities ( 6650 km s@xmath11 and 6620 km s@xmath11 , respectively ; moles et al .
the most likely candidate of these two is the barred spiral ngc 7319 , which shows a surprising lack of h ii regions ( arp 1973 ) , a long optical tail , and a very offset co distribution ( yun et al .
ngc 7318a also has an optical tail , however , this may have been created by another interaction .
the 6000 km s@xmath11 extra - disk gas may have originated in either the outlying galaxy ngc 7320c or in ngc 7318b .
ngc 7320c has an optical radial velocity of 6000 @xmath38 150 km s@xmath11 ( lynds 1972 ) , consistent with this velocity .
previous arguments for the involvement of ngc 7320c ( moles et al .
1997 ) cited the long tail south of ngc 7319 that points towards ngc 7320c .
alternatively , this component of the extra - disk gas may have come from ngc 7318b . as noted above
, our co observations show a faint co line in ngc 7318b at this velocity , redshifted 300 km s@xmath11 from the optical velocity .
this detection corresponds to a molecular gas mass of 1.7 @xmath1 10@xmath17 m@xmath16 , assuming the standard n@xmath8/i@xmath9 conversion factor .
we therefore suggest that the extra - disk gas undergoing star formation in stephan s quintet originated in ngc 7318b and ngc 7319 . at the present time it is unclear whether a single collision between ngc 7318b and ngc 7319 occurred , ram pressure stripping the gas and creating the shock front between the two galaxies , or whether gas had been removed from ngc 7319 earlier , by an earlier encounter with another galaxy , and then this gas was impacted by ngc 7318b .
we have obtained new co ( 1 @xmath0 0 ) observations of eleven extragalactic tails and bridges , and compared these measurements with previously - published data for another thirteen features . of these 24 structures ,
four have inferred m@xmath8/m@xmath26 upper limits ( assuming the galactic n@xmath8/i@xmath9 conversion factor ) less than global values for spiral galaxies , consistent with those found for irregular galaxies .
three of these four co - poor features are in the ngc 7714/5 system ; the most extreme case is the star forming bridge , which has a very high hi column density ( 1.6 @xmath1 10@xmath2 @xmath3 in the 55@xmath4 co beam ) , yet was undetected in co , giving an inferred m@xmath8/m@xmath26 upper limit of @xmath140.063 .
these features may have enhanced n@xmath8/i@xmath9 ratios compared to the galactic value due to low metallicities . of the 24 structures in our combined sample ,
eight have been detected in co. three of these eight ( the stephan s quintet extra - disk source , the ngc 4038/9 tail , and the arp 245 tail ) have inferred efficiencies of star formation ( l@xmath29/m@xmath8 ) consistent with global values for spiral galaxies .
in contrast , the l@xmath29/m@xmath8 ratios for the eastern tail of ngc 2782 and the ngc 4438 clump are lower than typical for disk galaxies , suggesting inhibited star formation . in none of the features
were our co limits low enough to conclusively show l@xmath29/m@xmath8 ratios as high as those found in dwarf galaxies .
of the eight detected features , two ( ngc 2782 and ngc 4438 ) have optical morphologies indicating head - on collisions .
the stephan s quintet feature may also have been formed in this manner . in these systems
, we suggest that metal - rich gas with an approximately galactic n@xmath8/i@xmath9 ratio was pulled from the inner parts of the galaxies .
a few recent studies have reported the detection of co in a number of classical tidal features ( e.g. , ngc 4038/9 , arp 245 , and ngc 3561 ) . in these structures
, the hi column density may be locally high enough that the co self - shielding limit is exceeded , in spite of modestly enhanced n@xmath8/i@xmath9 ratios .
there is a slight tendency for these tidal features to have lower m@xmath8/m@xmath26 ratios and higher l@xmath29/m@xmath8 ratios than ram pressure stripped features discussed above , but there is a large amount of scatter and the sample size is small .
more observations are needed to determine whether there is truly a statistical difference in the gaseous and star formation properties of tidal and ram pressure stripped features .
we are grateful for the help of the nrao 12 m telescope operators and staff in obtaining these data .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory under contract with nasa .
we are pleased to acknowledge partial funding for this project from a nasa grant administrated by the american astronomical society and from nsf grant int-9908542 .
optical images of the nine interacting galaxy systems in our sample , with the locations and fwhm beamsizes of the co beams marked .
all images are from the digitized sky survey ( dss ) , except when noted .
a ) arp 144 ( ngc 7828/9 ) .
there is an hi tail without a known optical counterpart to the southeast of the optical pair ( higdon 1988 ) .
b ) ngc 2814/20/mk 208 .
c ) ngc 3395/6 .
d ) arp 105 ( ngc 3561 ) .
this image has been smoothed by a 3@xmath4 fwhm gaussian to enhance low surface brightness features .
e ) the brightest galaxy in the leo triplet , ngc 3628 .
a long hi tail extends to the east of ngc 3628 ( rots 1978 ; haynes , giovanelli , @xmath13 roberts 1979 ) ; this tail has a faint optical counterpart not visible in the dss image shown here ( kormendy @xmath13 bahcall 1974 ) .
the two marked locations are in this tail .
we also observed the center of ngc 3628 .
f ) i zw 192 .
g ) stephan s quintet .
h ) a narrowband red continuum image of ngc 7714/5 ( arp 284 ) , from smith et al . ( 1997 ) .
i ) ugc 12914/5 ( taffy galaxies ) .
a comparison of n@xmath26 with m@xmath8/m@xmath26 for the extra - disk positions in tables 5 and 6 with both hi and co data . for comparison purposes ,
the standard galactic n@xmath8/i@xmath9 ratio is used to calculate the molecular gas masses . as discussed in the text
, this ratio may not hold in all cases .
the unlabeled points near m@xmath8/m@xmath26 @xmath10 0.8 @xmath0 0.9 and n@xmath26 @xmath10 10@xmath23 @xmath3 are the data for the arp 143 tail , the two positions in the arp 144 tail , the ngc 2782 western tail , and the ngc 4410 tail ( see tables 5 and 6 ) .
@xmath39all values averaged over the 55@xmath4 nrao 12 m beam , except where noted .
@xmath40calculated assuming the standard galactic n@xmath8/i@xmath9 ratio ( n@xmath8/i@xmath9 = 2.8 @xmath1 10@xmath23 @xmath3/(k km s@xmath11 ) and m@xmath8 = 1.1 @xmath1 10@xmath41 d@xmath42 @xmath43s@xmath44dv , where d is the distance in mpc ; bloemen et al . 1986 ) , h@xmath21 = 75 km s@xmath11 mpc@xmath11 , and assuming 34 jy / k for the 12 m telescope and the source fills the beam ( @xmath24@xmath25 = 0.82 ) .
as noted in the text , these assumptions may not hold for all these sources ; they are used for comparison purposes here .
@xmath45not corrected for extinction .
@xmath46as tabulated in smith @xmath13 higdon 1994 . @xmath47using the data from braine
2000 and duc et al .
2000 , in a @xmath1023@xmath4 beam .
the h@xmath12 luminosity has been corrected for [ n ii ] using h@xmath12/(h@xmath12+[n ii ] ) @xmath10 0.7 , as indicated by the spectroscopy in duc et al .
( 2000 ) . @xmath48using the data from brouillet
et al . 1992 and rots @xmath13 shane 1975 , in a @xmath1023@xmath4 beam .
no evidence of star formation at this position has been observed ( henkel et al .
the assumed distance is 3.2 mpc ( sandage @xmath13 tammann 1978 ; humphreys @xmath13 aaronson 1987 ) . @xmath49using the data from walter @xmath13 heithausen 1999 , in a @xmath1023@xmath4 beam .
there is no observed star formation at this position ( walter @xmath13 heithausen 1999 ) .
the hi column density quoted here was estimated from the hi map in walter @xmath13 heithausen 2000 .
@xmath50from data in smith et al .
these are averaged over the five detected positions in the tail , covering an area of @xmath107645 arcsec@xmath51 , the equivalent of an 87@xmath4 beam . @xmath52using the data from braine
et al . 2000 ,
duc @xmath13 mirabel 1994 , and duc et al .
1997 , in a @xmath1023@xmath4 beam .
the h@xmath12 flux was derived from spectroscopy with a 1.5@xmath4 slit , and so is an upper limit .
@xmath53the co measurement is a tentative ( 4@xmath28 ) detection from gao et al .
2000 ; note that this assumes a broader line ( 175 km s@xmath11 ) than the upper limit quoted in smith @xmath13 higdon 2000 .
the hi measurement is estimated from the hibbard et al .
( 2000 ) hi map .
the h@xmath12 measurement is from mirabel et al .
@xmath54as tabulated in smith et al .
1999 , in a 23@xmath4 beam . @xmath55using the data from smith 2000 . | we present new co ( 1 @xmath0 0 ) observations of eleven extragalactic tails and bridges in nine interacting galaxy systems , almost doubling the number of such features with sensitive co measurements .
eight of these eleven features were undetected in co to very low co / hi limits , with the most extreme case being the ngc 7714/5 bridge .
this bridge contains luminous h ii regions and has a very high hi column density ( 1.6 @xmath1 10@xmath2 @xmath3 in the 55@xmath4 co beam ) , yet was undetected in co to rms t@xmath5@xmath6 = 2.4 mk .
the hi column density is higher than standard h@xmath7 and co self - shielding limits for solar - metallicity gas , suggesting that the gas in this bridge is metal - poor and has an enhanced n@xmath8/i@xmath9 ratio compared to the galactic value .
only one of the eleven features in our sample was unambiguously detected in co , a luminous hi - rich star formation region near an optical tail in the compact group stephan s quintet .
we detect co at two widely separated velocities in this feature , at @xmath106000 km s@xmath11 and @xmath106700 km s@xmath11 .
both of these components have hi and h@xmath12 counterparts .
these velocities correspond to those of galaxies in the group , suggesting that this gas is material that has been removed from two galaxies in the group .
the co / hi / h@xmath12 ratios for both components are similar to global values for spiral galaxies . |
single - wall carbon nanotubes ( swnts ) are quasi one - dimensional wires of great interest for future electronic - devices applications
. a swnt can be constructed by rolling up a graphene sheet and its geometry is univocally specified by a pair of chiral indexes @xmath2 @xcite .
the relationship between @xmath3 and @xmath4 defines three groups of swnts : armchair @xmath5 , zigzag ( @xmath6 or @xmath7 ) and chiral nanotubes .
the indexes @xmath2 also determine the swnt electronic structure .
if @xmath8 is not a multiple of 3 , swnts of diameters in the 1 - 2 nm range are semiconductors with a band gap @xmath9 larger than 0.5 ev .
only armchair swnts were believed to be truly metallic . indeed , because of the curvature , zigzag and chiral swnts with @xmath8 multiple of 3 , present a small band gap of few tens of mev @xcite , and for this reason are called quasi - metallic ( qm ) @xcite .
however , recent transport measurements on chirality - identified tubes showed that also armchair tubes are qm with @xmath10 to @xmath11 mev , likely because of electrons correlation @xcite . the description and interpretation of the transport properties of swnt devices ,
are often based on a semi - classical approach such as the boltzmann transport equation ( bte ) @xcite .
however , the bte can not describe some cases such us strong localization regimes @xcite .
another effect that can not be described by the boltzmann transport approach is the zener tunneling ( zt ) @xcite , which is the electric - field - induced tunneling of carriers from one band to another through the forbidden energy gap @xmath9 . in absence of zt ,
if the temperature is smaller than the gap ( @xmath12 ) , only one type of carrier is responsible for charge transport : when the fermi level , @xmath13 , is in the conduction ( valence ) band , the current is carried by electrons ( holes ) .
if @xmath13 is in the gap region charge - transport does not occur . on the contrary , in presence of zt ,
the current is carried simultaneously by both electrons and holes and a non - zero current exists even if @xmath13 is in the gap region .
the critical electric field @xmath14 for which the zt affects transport properties , can be estimated considering the inter - band tunneling probability across the gap of the swnt hyperbolic bands . in absence of scattering ,
the transition probability as a function of the source - drain electric field @xmath15 is given by @xmath16 @xcite where @xmath17 and @xmath18 cm.@xmath19 is the fermi velocity of graphene ( and of qm nanotubes far from the gap ) . in semiconducting tubes ( @xmath20 mev )
, @xmath21 v/@xmath22 is unrealistically large . at experimentally - accessible electric fields ,
zt is absent , and in a diffusive regime a semi - classical picture is sufficient to describe transport properties . instead , in qm tubes ( @xmath23 mev ) @xmath24
v/@xmath25 , a value that nanotubes can sustain @xcite . in this case
zt could be observed . in this work ,
we study theoretically the current voltage characteristics , of qm nanotubes both in the linear and high field regime and we analyze the interplay between zener tunneling and defect ( or impurity ) scattering .
we consider long qm tubes with an energy band gap @xmath26 @xmath27 ( similar to the values measured in experiments @xcite ) in a diffusive regime , _
i.e. _ when the tube length is much longer than the defect - induced elastic scattering lengths ( @xmath28 ) of 50 nm and 300 nm presently used .
the approach is based on the solution of the master equation for the evolution of the density matrix .
this approach correctly describes inter - band zener tunneling .
we take into account the elastic scattering of electrons with defects ( we consider scattering of electrons by short - range potential ) and the inelastic scattering with optical phonons .
we do not consider the possible occurence of weak or strong localization effects @xcite .
elastic and phonon scattering are assumed to be incoherent and , thus , the phase coherence length @xmath29 .
scattering with acoustic phonons plays a minor role since it is characterized by scattering lengths that are much larger than the @xmath28 used here ( see _ e.g. _ ref . ) .
we suppose that the tubes are supported and that , for the regime under study , the phonon population is equilibrated at the temperature of the substrate .
we assume that the doping and the potential drop is uniform along the channel , and that the electron population of the valence and conduction bands is independent on the position in the channel .
namely , we consider the bulk transport properties of an infinite swnt under a spatially homogeneous electric - field applied along the tube axis . we remind that at high bias conditions in small gap nanotubes , the scattering with optical phonons can lead to a current saturation @xcite and to nonequilibrium optical - phonon population ( hot phonons ) @xcite . in suspended metallic tubes , this can lead to the observation of negative differential resistance @xcite .
this kind of phenomena are not included in the present model and their inclusion would lead to increase of the differential resistance for high electric field ( this point will be discussed in sects .
[ sec_transport]- [ sec_current ] ) .
electronic transport measurements in qm nanotubes were reported in .
the two papers deal with transport regimes which are quite differents from the present one . because of this
, the results of can not be directly compared with the present ones . in particular , in both ,
the elastic scattering lengths are much longer than in the present work and , in both cases , they are comparable with the length of the tubes . because of this , ref
. can observe interference effects between the contacts and ref .
can interpret the results by assuming an electronic distribution non uniform along the length of the channel . on the contrary , in the present work we are in a diffusive regime : the scattering length @xmath30 , where @xmath31 is the length of the tube , and @xmath32 , where @xmath33 is the phase coherence length . in this diffusive regime
, we can assume a uniform distribution of the electrons along the tube and that interference effects between the contacts are not observable .
the graphene hamiltonian for the @xmath34-bands close to the @xmath35 point ( dirac point ) of the brillouin zone ( bz ) is @xcite : @xmath36 where @xmath37 is the fermi velocity , @xmath38 with @xmath39 and @xmath40 the pauli matrices .
@xmath41 is the momentum operator in the two - dimensional bz of graphene .
@xmath42 is a 2@xmath432 matrix which describes the @xmath34 and @xmath44 bands .
starting from eq.([graph ] ) , one can derive an effective hamiltonian @xmath45 for a swnt : @xmath46 @xmath47 is the momentum operator in the one - dimensional bz of the nanotube ( a momentum parallel to the tube axis ) and @xmath48 is a constant fixed by the energy gap @xmath49 .
the eigenstates of @xmath45 are bloch states and have the form @xmath50 here @xmath31 is the length of the system and @xmath51 , is the periodic part of the bloch state : @xmath52 where @xmath53 .
@xmath54 indicates positive energy ( conduction band ) whereas @xmath55 describes states with negative energy ( valence band ) with respect to the dirac point .
the eigenvalues of @xmath45 read : @xmath56 we now apply a spatially homogeneous and static ( source - drain ) electric field @xmath15 along the tube axis .
the hamiltonian , in the vector - potential gauge , becomes : @xmath57 \label{nanotfield}\ ] ] where @xmath58 is the vector potential component along the tube axis , and @xmath59 the electron charge .
the instantaneous eigenstates of the time dependent hamiltonian , eq.([nanotfield ] ) , are given by @xcite : @xmath60 with eigenvalues @xmath61 .
a framework in which quantum effects such as zener tunneling appear naturally is the liouville - von neumann master equation ( me ) @xcite . within this formalism
our variable is the one - electron density - matrix operator @xmath62 .
the time evolution of @xmath62 in absence of scattering processes is given by : @xmath63 .
\label{liouville}\ ] ] since we are dealing with a spatially homogeneous electric field applied to an infinite nanotube , only the elements of the density matrix diagonal with respect to @xmath64 will be relevant @xcite .
we use the instantaneous eigenstates ( eq.([nanotfb ] ) ) as basis .
then , from eq.([liouville ] ) one finds the time evolution of @xmath65 @xcite : @xmath66 where @xmath67 and @xmath68 is defined by : @xmath69 the @xmath68 term introduces the inter - band transitions .
using perturbation theory for the derivative with respect to @xmath64 we can show that : @xmath70 where @xmath71 is the velocity operator defined by : @xmath72 = v_f\sigma_x \label{velop}\ ] ] for our system @xmath68 takes the form : @xmath73 the full evolution equation is obtained by adding the collision integral @xmath74_{\alpha\beta}$ ] to the right - hand side of eq.([me1 ] ) .
the collision integral results from the sum of two independent contributions due to inelastic electron - phonon scattering and elastic electron - defect scattering .
electron - phonon scattering is treated as in ref .
@xcite and read : @xmath75^{e - ph}_{\alpha\beta}=\frac{1}{2l } \sum _ { k ' } \sum_{\eta \nu}(w^{\nu}_{k'\eta \rightarrow k\alpha } + w^{\nu}_{k'\eta\rightarrow k\beta } ) & \\
\times \rho _ { \eta \eta } ( k')(\delta _ { \alpha\beta } -\rho_{\alpha \beta } ( k ) ) & \\-\frac{1}{2l } \sum_{k'}\sum_{\eta \nu}(w^{\nu}_{k\alpha\rightarrow k'\eta } + w^{\nu}_{k\beta\rightarrow k'\eta } ) & \\
\times \rho _ { \alpha\beta } ( k)(1-\rho _ { \eta\eta}(k ' ) ) \end{split } \label{colph}\ ] ] where @xmath76 , the electron - phonon scattering probability is evaluated using fermi golden rule @xcite : @xmath77 with @xmath78 the coupling matrix elements and @xmath79 the phonon branch . @xmath80 is the phonon energy with @xmath81 and @xmath82 is the phonon population .
the upper ( lower ) sign in the delta function corresponds to phonon emission ( absorption ) .
+ the elastic scattering by defects is supposed to be incoherent . in this case
the collision integrals for the diagonal and non - diagonal elements @xmath83 of the density matrix are given by : @xmath84_{\alpha \alpha } = \frac{1}{l } \sum_{k ' } w^{el}_{k'\alpha \rightarrow k\alpha } [ \rho_{\alpha \alpha}(k ' ) - \rho_{\alpha \alpha}(k)]\ ] ] @xmath84_{\alpha\beta } = -\frac{1}{2l } \sum_{k'}(w^{el}_{k'\alpha \rightarrow k\alpha } + w^{el}_{k'\beta \rightarrow k\beta})\rho_{\alpha \beta}(k).\ ] ] the transition probability , within the fermi golden rule , is defined by : @xmath85 where @xmath86 is the density of defects ( that is the number of defects per unit length ) and @xmath87 is the scattering potential . finally , the master equation , that describes the time evolution reads : @xmath88^{el}_{\alpha \beta } + \mathrm{i}[\rho]^{e - ph}_{\alpha\beta}. \end{split } \label{me2}\ ] ] for a given value of the electric field @xmath15 , a steady - state solution of eq.([me2 ] ) is reached for @xmath89 . in this work ,
the steady - solution is obtained by numerical integration in time using a finite - difference method . the electronic population ( that is the diagonal elements @xmath90 ) determines the density of electrons @xmath91 ( number of electrons in excess per unit length of the tube ) by : @xmath92,\ ] ] where @xmath93 and @xmath94 are respectively the spin and valley degeneracy .
@xmath91 is a constant of motion of eq.([me2 ] ) .
we will show simulations done at different values of the fermi level @xmath13 , that is by fixing the initial value of @xmath91 to : @xmath95\ ] ] where @xmath96 is the fermi - dirac function , computed here for @xmath97 k. the charge current is given by : @xmath98 where @xmath99 is defined by : @xmath100 \overline { \vert k , \beta\rangle } = \langle k,\alpha \vert \frac{i}{\hbar}[\mathcal{h},x ] \vert k , \beta\rangle\ ] ] using the definition of the velocity operator given in eq.([velop ] ) , one can show that @xmath101 finally the steady current will be evaluated for the steady - state solution .
an interesting aspect of the master equation , is that we can recover the standard boltzmann transport equation ( bte ) , i.e. the semiclassical description , setting @xmath102 of eq.([me2 ] ) to zero . in this case
the diagonal elements of the density matrix evolve according to : @xmath103^{el}_{\alpha \alpha } + \mathrm{i}[\rho]^{e - ph}_{\alpha\alpha } , \end{split } \label{boltz}\ ] ] and the off - diagonal elements of the steady solution vanish . in this section
, we derive a set of simplified equations ( which can be evaluated analytically ) valid for small values of the electric field @xmath15 within the linear regime . in the linear regime
the applied field will cause a small deviation of the equilibrium distribution @xmath104 .
the steady state density matrix ( with @xmath105 ) can be written as : @xmath106 where @xmath107 is the density matrix element in absence of field @xmath104 and @xmath108 the deviation linear in @xmath15 .
@xmath107 is the equilibrium distribution which is diagonal and equal to @xmath109 . inserting eq.([eqlin1 ] ) in eq.([me2 ] ) , and keeping all the terms linear in @xmath15 , one obtains the solution for the diagonal elements of the density matrix : @xmath110 with @xmath111 , the momentum relaxation time , defined by @xcite : @xmath112 where @xmath113 for back - scattering processes and @xmath114 for forward scattering ones . indeed , after the scattering with a defect ( or a phonon ) , the electron can either maintain its propagation direction , which is forward scattering ( fs ) , or reverse it , which is back - scattering ( bs ) .
the off - diagonal elements @xmath115 are given by : @xmath116 where @xmath117 is defined by : @xmath118 where @xmath119 , @xmath120 and @xmath121 , are respectively the elastic , hole - phonon and electron - phonon scattering rate for the equilibrium distribution and defined by : @xmath122 @xmath123 @xmath124 \label{gamephe}\ ] ] the zero - field conductivity , @xmath125 , can be obtained by inserting the steady state solutions of the linear regime , eqs.([diagdens ] ) and ( [ nondiagd ] ) in the current expression , eq.([curr ] ) : @xmath126 where @xmath127 is the conductivity obtained from the diagonal elements of the density matrix .
it is given by : @xmath128 @xmath127 is the semi - classical conductivity of the boltzmann equation and at zero temperature it reduces to : @xmath129 with @xmath130 the fermi momentum and @xmath131 the one - dimensional density of states ( dos ) per unit length . for the lowest sub - band of the qm nanotube @xcite : @xmath132 where @xmath133 is the constant dos ( per unit length ) of the lowest sub - band of a metallic tube .
@xmath134 is the step function which equals 1 for @xmath135 and @xmath114 otherwise . in the case of a metallic tube
the conductivity , eq.([condb ] ) , can be written as : @xmath136 where @xmath137 , the elastic scattering length , is given by : @xmath138 the current due to zener inter - band is given by @xmath139 , that arises from the off - diagonal part of the density matrix , eq .
( [ nondiagd ] ) : @xmath140 ^ 2 + [ \hbar\gamma_{tot}(k)]^2}(1-\delta_{\alpha \beta } ) .
\label{kubo}\end{aligned}\ ] ] notice that this expression is similar to the off - diagonal kubo formula for the optical conductivity @xcite but with a finite value of the parameter @xmath117 . when @xmath141 , the zener contribution ( @xmath142 ) vanishes and the total conductivity is equal to that of the semiclassical boltzmann approach ( @xmath127 )
. thus , in the linear regime , the zener tunnelling contributes to the transport only in presence of scattering events that limit the carrier lifetime . on the contrary , at finite field
, zener tunneling can occur in the ballistic case ( see formula right above eq.([ezener ] ) ) .
we performed calculations for a metallic tube and a quasi - metallic tube ( qm ) with an energy band gap @xmath26 @xmath27 ( similar to the values measured in experiments @xcite ) .
the tubes are subject to a spatially uniform electric field and the number of carriers ( electrons or holes ) is varied by changing the fermi level @xmath13 . in the metallic tube
the dispersion relation is linear @xmath143 , with @xmath144 @xmath145 , in a qm tube the bands are hyperbolas , eq.([nanote ] ) .
the transport properties of the qm nanotube are explored by a semi - classical and a quantum - mechanical approach . in the semi - classical case ,
we do not take into account the inter - band transitions due to the electric field by setting to zero the off - diagonal elements of the density matrix ( @xmath146 with @xmath147 ) .
the evolution of the system is then governed by the bte , eq .
( [ boltz ] ) . in the quantum approach ,
we take into account the inter - band transitions due to the electric field by considering the off - diagonal elements of the density matrix .
the evolution is then governed by the me , eq .
( [ me2 ] ) . for the metallic tube , the me that governs the quantum evolution can be obtained in the limit @xmath148 . in this case , the coupling @xmath149 [ ( eq . ( [ rzen1 ] ) ] , that promotes zener transitions , diverges at @xmath150 , and the electrons that cross the @xmath150 point tunnel with 100% probability from the valence ( conduction ) to the conduction ( valence ) band . as a consequence , for metallic tubes , the quantum - mechanical approach ( me ) reduces to the semiclassical approach ( bte ) if we change the meaning of the band indexes in eq .
[ boltz ] .
in particular , while for qm tubes we distinguish between `` valence '' and `` conduction '' bands , for metallic tubes we distinguish between bands with positive or negative velocity ( @xmath151 or @xmath152 ) , see fig.[qmb ] .
we consider short - range scatterers and the electron - defect coupling @xmath153 is taken to be a constant and @xmath64-independent . under this hypothesis in a metallic tube the elastic scattering mean free path @xmath154 is @xmath64-independent .
the value of @xmath153 is determined fixing @xmath155 nm and @xmath156 nm . in the qm tubes , @xmath153 is also assumed to be a constant equal to that of the metallic tubes .
note , however , that , in the qm case , the scattering mean free path become @xmath64-dependent .
indeed , because the presence of a gap , the density of states , that determines the scattering rate , is energy dependent .
only three optical phonons can play an important role in transport phenomena at high bias : two ( longitudinal and transverse ) correspond to the graphene optical phonons at @xmath157 with symmetry e@xmath158 , and one corresponds to the graphene phonon at @xmath159 with symmetry a@xmath160 @xcite .
the energies of these phonons are respectively @xmath161 @xmath27 and @xmath162 @xmath27 . for our calculation
the electron - phonon coupling matrix elements @xmath163 are taken constant and independent of @xmath64 and @xmath164 , with @xmath165 . the label @xmath166 stands for a back - scattering process ( @xmath167 ) , and @xmath168 stands for a forward scattering process ( @xmath169 ) . in a metallic tube ,
the corresponding scattering lengths , @xmath170 , are proportional to the tube diameter @xmath171 @xcite , that we assume to be 2 nm . in this case , @xmath172 nm , @xmath173 nm and @xmath174 . for the qm tubes we use the @xmath175 obtained in the metallic case .
the phonon population @xmath176 is set to zero ( cold phonons , only phonon emission is treated ) .
we do not consider the generation of hot phonons , that occurs at high bias in nanotubes , when the elastic scattering length is much larger than the phonon one , @xmath170 @xcite . in the present case , this is justified by the strong elastic scattering used in the present model that competes with the phonon scattering , as in graphene @xcite .
we now show the results obtained for two different values of the elastic scattering length , namely @xmath177 nm and @xmath178 nm .
the goal is to understand in which conditions the zener tunneling is experimentally observable .
we will compare the results obtained with the semi - classical ( or boltzmann ) approach with those obtained with the quantum treatment ( master equation ) . in the semi - classical approach the zener tunneling is not allowed .
thus , the semi - classical and quantum results will differ when zener tunneling is relevant .
results for metallic tubes ( for which the distinction quantum versus . semi - classical does not hold ) are shown as a comparison .
the results shown for a finite value of the source - drain electric field , @xmath15 , are obtained from numerical solution of the transport equations .
the results for @xmath179 are derived analytically in sec .
[ sec_zero_field ] fig .
[ cond5 ] reports the conductivity @xmath180 of the qm tube as a function of the fermi level @xmath13 for various values of the electric field @xmath15 . in fig .
[ cond5 ] , @xmath181 is the zero - bias ( @xmath179 ) conductivity calculated in the linear regime for the temperature @xmath97 k. @xmath182 is the component of @xmath181 due to zener tunneling ( @xmath182 is a well defined quantity , see sec .
[ sec_zero_field ] ) , and @xmath183 is the zero - bias conductivity of a metallic tube . for the metallic tube , the conductivity @xmath183 has a finite value which is proportional to the elastic scattering length @xmath28 and which does not depend on @xmath13 ( this is true as soon as @xmath13 does not meet the higher subbands of the swnt , which are not considered in this study ) . for the qm tube , the conductivity @xmath181 tends to be equal to the metallic one for @xmath184 : when the doping is sufficiently high there is only one type of carriers , the details of the electronic bands near the dirac point are not relevant , the transport is properly described by the semi - classical boltzmann approach and the qm and metallic tube behavior are indistinguishable . on the other hand , for @xmath185 the presence of the electronic gap @xmath9 acts as a barrier and the qm conductivity suddenly diminishes . in particular , from fig .
[ cond5 ] , for @xmath185 , @xmath186 meaning that the qm conductivity is entirely due to zener tunneling ( inter - band ) and that the boltzmann conductivity ( intra - band ) vanishes
[ cond5 ] also reports the finite field conductivity @xmath187 for different values of the electric field .
also at finite field , the conductivity is entirely due to zener tunneling for small values of @xmath13 .
the comparison of the results obtained for different values of the elastic scattering lengths , @xmath28 , is very remarkable . indeed , for small doping ( @xmath185 ) , the zero - bias conductivity for the smallest scattering length ( @xmath188 s.m@xmath189 for @xmath190 nm in fig .
[ cond5 ] ) , is higher than the one for the longest scattering length , ( @xmath191 s.m@xmath189 for @xmath192 nm ) .
moreover , most important , the ratio between the high doping @xmath181 ( @xmath1930.2 ev ) and the small doping @xmath181 is much higher for the smallest scattering length than for the longest one .
this counterintuitive behavior can be understood by considering the electronic scattering rate ( or broadening ) @xmath194 , defined in eq .
@xmath194 represents the uncertainty with which the energy of an electronic band can be defined .
[ fig3 ] reports the electronic bands for the qm tube : the bands are represented as lines with a finite thickness ( gray area in fig .
[ fig3 ] ) equal to @xmath194 .
the increased value of @xmath194 in the vicinity of the gap is due to the elastic scattering .
the broadened electronic bands of fig .
[ fig3 ] overlap near @xmath150 for @xmath190 nm ( low quality samples ) , meaning that , because of energy - uncertainty , the electrons do not see the gap and can tunnel from a band to the other even if the applied field is small . on the contrary , for @xmath192 nm ( better quality samples ) the bands do not overlap and for small electric field the zener tunneling can not take place . in fig .
[ fig3 ] , the sudden increase of the broadening for @xmath195 @xmath189 is due to the scattering with optical phonons .
this scattering mechanism is active only when @xmath196 and @xmath197 , which is far from the gap .
zener tunneling is , thus , activated only by the elastic scattering ( with defects ) and not by the scattering with optical phonons .
finally , to see the things from a different perspective , fig .
[ cur5 ] reports the current as a function of @xmath13 for various values of the electric field @xmath15 . for large doping ( large @xmath13 ) ,
the qm tube current calculated with the semiclassical and with the quantum approaches tend to be equal . also , both currents tend to be equal to the metallic tube one .
the semi - classical approach predicts a zero current in the gap and underestimates the current in a range of @xmath13 that increases with @xmath15 . in the region of the gap ,
the current is thus entirely due to zener tunneling .
to analyze the results as a function of the electric field @xmath15 , we consider two characteristic fields : @xmath14 and @xmath198 .
@xmath14 is defined in eq .
( [ ezener ] ) and it is the value of the electric field for which , in the absence of scattering , the zener scattering is expected to be sizable .
@xmath198 is defined as @xmath199 and it is the value of the field for which optical - phonon emission become sizable compared to elastic scattering .
@xmath198 is higher for smaller values of @xmath28 . to understand this definition
, one should consider that optical phonons have a finite energy @xmath200 . in order to emit an optical phonon an electronic carrier needs to attain a minimum activation energy corresponding to @xmath200 .
this is possible if the carrier has accelerated in the electric field @xmath15 at least through a length @xmath201 @xcite .
the process is dominated by the zone boundary phonons since @xmath202 .
for small electric fields , @xmath203 , @xmath204 and the scattering with defects prevents the carriers to accelerate and , thus , to reach the activation energy . on the other hand ,
if @xmath205 , @xmath206 and the carriers can attain the activation energy before scattering with the defects .
thus , for small electric field , elastic scattering is the dominant scattering mechanism ( in this regime the current - voltage curve of metallic tubes is linear @xcite ) . for electric fields
higher that @xmath198 the scattering with optical phonons is activated .
this is associated with a decrease of the differential conductivity ( sub - linear current - voltage curve ) which can lead to a saturation of the current at high drain - source voltages in metallic tubes with long elastic scattering lengths @xcite .
the scattering with optical phonons can lead to an anomalous increase of the optical - phonon population ( hot - phonons ) @xcite which , in turn , will further diminish the high - bias conductivity .
[ mcur5 ] shows the current as a function of the electric field for different values of @xmath13 . at high doping , the results for the semi - classical ( bte ) and of the quantum ( me ) approaches are very similar and approach the metallic results .
the two approaches provide significantly different results for sufficiently high electric fields and small values of the fermi level .
in particular , for @xmath207 the current is exactly zero within the semi - classical model , while the current has a finite value , which increases with the electric field , within the quantum approach , fig .
[ mcur5 ] . for @xmath208 and @xmath209
ev , within the quantum approach the current versus .
electric field is a super - linear curve for @xmath210 .
such a behavior was observed in graphene @xcite and in large diameter nanotubes @xcite , and identified as due to zener tunneling .
let us compare the results for the different scattering lengths @xmath28 ( fig .
[ mcur5 ] ) .
zener tunneling should be more visible for @xmath211 50 nm , than for @xmath211 300 nm , for two reasons .
first , for @xmath211 50 nm , the zener current is already relevant for electric fields smaller than @xmath14 ( this is explained in the discussion related to fig .
[ fig3 ] ) .
second , for @xmath211 50 nm @xmath212 , while for @xmath211 300 nm @xmath213 .
now , the scattering with optical phonons is associated to a sub - linear current which tend to mask the presence of the super - linear zener current . for @xmath211 50 nm
, @xmath214 means that this `` masking '' does not occur in the region of interest . for @xmath211 300 nm
, @xmath213 means that this `` masking '' is already active in the region where zener is active . beside this
, one should also consider that in the present simulations we are not including the possibility of hot phonons .
the inclusion of this effect would decrease the differential conductance for @xmath215 .
this should not change the @xmath211 50 nm curves but should further mask the zener current in the @xmath211 300 nm case .
finally , in by fig.[map3 ] and [ map5 ] we present the two - dimensional maps of the current as a function of @xmath15 and @xmath13 . in one of the panels
, we also report the difference between the current obtained within the quantum approach and that of the semi - classical calculation .
this difference is higher where the zener tunneling is more important . for metallic tubes ,
the current does not depend on the fermi level because , within the linear band approximation , the electronic density of states is a constant . note that for sufficiently high @xmath13 values the semi - classical and the quantum results for the qm tubes are almost indistinguishable and approach the results for the metallic tubes .
it is instructive to compare the steady - state distribution functions of the electronic populations associated with the semi - classical or quantum approaches .
[ disthlf ] reports the distribution functions obtained for @xmath190 nm and the fermi level @xmath216 ev , that is for conditions in which the quantum effects are relevant . in qm tubes ,
when the quantum approach is used both electron and hole contribute to the current , whereas in the semiclassical approach only electrons carry a current since the valence band is filled .
interestingly , at high field the population of qm tubes tends to that of metallic tubes only when the quantum approach is used .
indeed , in the metallic tube ( which corresponds to the limit @xmath217 ) the tunneling between valence and conduction band is total .
the metallic limit can be reached only when one includes interband transitions .
[ disthlf ] also shows the imaginary part of the off - diagonal element of the density matrix @xmath218 .
this quantity is different from zero when quantum effects are relevant .
one can remark that zener tunneling turns on near @xmath219 ( with a maximum at @xmath219 ) and decays after few oscillations for @xmath220 .
the typical wave vector associated with these oscillations is @xmath221 , @xmath9 being the electronic gap .
we have presented a theoretical model based on the master equation to explore the quantum effects ( inter - band transitions induced by electric field or zener tunneling ) on the transport properties of homogeneous quasi - metallic carbon nanotubes ( qm ) .
the presence and relevance of the zener tunneling has been highlighted by comparing with a semi - classical boltzmann transport approach , in which the inter - band transitions are neglected .
we studied qm tube with an electronic gap @xmath222 mev in the presence of impurities or defects which influence the transport by providing an elastic scattering channel for the carriers .
zener tunneling is relevant for small doping , when the fermi energy lies in or close to the forbidden gap @xmath9 . in absence of elastic scattering ( in high quality samples ) , the small size of the band gap of qm tubes enables zener tunnelling for realistic values of the the electric field ( above @xmath0 1 v/@xmath1 m ) .
however , for such electric fields the scattering with the optical phonons ( whose effect is underestimated in the present approach ) tends to mask the zener current . on the other hand
, the presence of a strong elastic scattering ( in low quality samples ) further decreases the values of the field required to observe zener tunnelling .
indeed , for elastic - scattering lengths of the order of 50 nm , zener tunnelling affects the current / voltage characteristic already in the linear regime . in other words ,
zener tunneling is made visible by defects in qm tubes .
this result is similar to what has been recently shown for graphene @xcite .
6 jean - christophe charlier , xavier blase and stephan roche , rev .
phys . , volume @xmath223 , ( 2007 ) m. ouyang , j - l .
huang , c. l. cheung , and c. m. lieber , science @xmath224 , 702 ( 2001 ) p. delaney , h. j. choi , j. ihm , s. g. louie , m. l. cohen , nature @xmath225 , 466 ( 1998 ) .
v v. deshpande , b. chandra , r. caldwell , d s. novikov , j. hone and m. bockrath , science vol .
@xmath226 , 106 ( 2009 ) .
z. yao , c l. kane and c. dekker , phys .
@xmath227 , 2941 ( 2000 ) . c. zener , proc .
( london ) @xmath228 , 523 ( 1934 ) .
a. v. andreev , phys .
@xmath229 , 247204 ( 2007 ) .
d. jena , t. fang , q. zhang , and h. xing , appl . phys .
@xmath230 , 112106 ( 2008 ) .
bushmaker , v.v .
deshpande , s. hsieh , m.b .
bockrath , and s.b .
cronin , nano lett .
* 9 * , 2862 ( 2009 ) .
m. lazzeri , s. piscanec , f. mauri , a. c. ferrari , and j. robertson , phys .
@xmath231 , 236802 ( 2005 ) .
m. lazzeri , f. mauri , phys .
b. @xmath232 , 165419 ( 2006 ) .
e. pop , d. mann , j. cao , q. wang , k. goodson , and h. dai , phys .
* 95 * , 155505 ( 2005 ) .
castro neto a. h. , guinea f. , peres n. m. r. , novoselov k. s. and geim a. k. , rev .
@xmath233 , 109 ( 2009 ) j. b. krieger and g. j. iafrate , phys .
b @xmath234 , 5494 ( 1986 ) .
fausto rossi , _ theory of semiconductor quantum devices _ , springer , 2011 .
fausto rossi and tilmann kuhn , rev .
@xmath235 , ( 2002 ) . j. b. krieger and g. j. iafrate , phys .
b @xmath236 , 9644 ( 1987 ) .
r. hbner and r. graham , phys .
b @xmath237 , 4870 ( 1996 ) .
chihiro hamaguchi , _ basic semiconductor physics _ ,
springer berlin heidelberg , 2010 .
mark lundstrom and jing guo _ nanoscale transistors _
, springer - verlag , 2006 .
kentaro nomura and a. h. macdonald , phys .
lett @xmath238 076602 ( 2007 ) j - y .
park , s. rosenblatt , y. yaish , v. sazonova , h. stunel , s. braig , t. a. arias , p. w. brouwer , and p. l. mceuen , nano .
@xmath239 , 517 ( 2003 ) .
a. barreiro , m. lazzeri , j. moser , f. mauri , and a. bachtold , phys .
@xmath240 , 076601 ( 2009 ) .
n. vandecasteele , a. barreiro , m. lazzeri , a. bachtold , and f. mauri , phys .
b @xmath241 , 045416 ( 2010 ) .
a. svizhenko , m. p. anantram , and t. r. govindan , ieee trans .
@xmath239 , @xmath242 ( september 2005 ) , 557 - 562 . | we study theoretically the impact of zener tunneling on the charge - transport properties of quasi - metallic ( qm ) carbon nanotubes ( characterized by forbidden band gaps of few tens of mev ) .
we also analyze the interplay between zener tunneling and elastic scattering on defects . to this purpose
we use a model based on the master equation for the density matrix , that takes into account the inter - band zener transitions induced by the electric field ( a quantum mechanical effect ) , the electron - defect scattering and the electron - phonon scattering . in presence of zener tunnelling the qm tubes support an electrical current even when the fermi energy lies in the forbidden band gap . in absence of elastic scattering ( in high quality samples ) , the small size of the band gap of qm tubes enables zener tunnelling for realistic values of the the electric field ( above @xmath0 1 v/@xmath1 m ) .
the presence of a strong elastic scattering ( in low quality samples ) further decreases the values of the field required to observe zener tunnelling .
indeed , for elastic - scattering lengths of the order of 50 nm , zener tunnelling affects the current / voltage characteristic already in the linear regime . in other words , in quasi - metallic tubes ,
zener tunneling is made more visible by defects . |
absorption due to highly ionized oxygen and elements of similar atomic number ( e.g. o vii , viii , ne x ) is commonly found in the soft x - ray spectra of active galactic nuclei ( agns ) ( halpern 1984 ; nandra & pounds , 1994 ; reynolds , 1997 ; george @xmath2 1998 ) . high resolution x - ray spectra from _ chandra _ and _ xmm - newton _ show that these absorption lines are always blue - shifted compared with the emission lines and so must be in outflow from the central continuum source ( collinge @xmath2 2001 , kaastra @xmath2 2002 , kinkhabwala @xmath2 2002 , blustin @xmath2 2003 , krongold @xmath2 2003 , netzer @xmath2 2003 , turner @xmath2 2004 ) .
the absorption edges and lines respond to variation of the ionizing continuum emitted by the central engine . for example , krongold et al .
( 2007 ) show that for ngc4051 the ionization state of the wa responds to the continuum flux as predicted for gas that stays close to photoionization equilibrium , so that other factors , e.g. velocity changes and shocks , do not appear to be important .
it is therefore reasonable to assume that the absorbing cloud is photoionized by the continuum radiation from the active galactic nuclei ( agn ) .
this partially ionized optically thin gas along the line of sight to the center of the agn has come to be known as the warm absorber ( hereafter wa ) .
understanding the structure of the wa will lead to better understanding of the structure of agns ; for example , identifying the location from which the wind arises will impose constraints on the physics of the wind source .
some authors find that the wa is well fitted with a model with two ionization phases ( krongold @xmath2 2003 ; krongold @xmath2 2005a ; ashton @xmath2 2006 ; collinge @xmath2 2001 ) or three phases ( netzer @xmath2 2003 ; morales , fabian & reynolds , 2000 ; kaastra @xmath2 2002 ) , in which the components are in pressure equilibrium .
other authors ( ogle @xmath2 , 2004 ) and steenbrugge @xmath2 , ( 2005 ) have argued in favor of a continuous distribution of ionization state for the wa , in objects for which others fit just a few phases .
one way to investigate which solution is more likely is to investigate the physics determining the thermal structure of wa .
the effect of various factors on the wa can be studied conveniently using the thermal equilibrium curve ( hereafter referred to as the _ stability curve _ ) of temperature ( t ) against the ratio of ionization parameter ( @xmath3 , see below for definition ) to t. gas lying off the stability curve will cool or heat until reaching the curve .
if the curve has kinks that produce multiple stable values at fixed @xmath4 , then the wa can have multiple phases in mutual equilibrium .
this effect may be responsible for the multiphase nature of the interstellar medium of galaxies ( field , 1965 ) .
ionizing continuum from agn can induce such kinks ( krolik , mckee & tarter , 1981 ) , to a larger extent than the interstellar medium .
the pronounced _ kinks _ representing wa in the stability curve lie in the intermediate temperature range , @xmath5 where photoionization heating and collisional cooling of various metals are expected to determine the thermal balance of the wa .
the details of the thermal stability is strongly affected by the shape of the ionizing continuum emitted by the central source ( i.e. their spectral energy distributions or seds ) , and by the chemical composition of the absorbing gas .
our main motivation in this paper is to improve on previous work ( hess , kahn & paerels 1997 ; komossa & mathur 2001 ; krolik , mckee & tarter 1981 , krolik & ; kriss 2001 ; reynolds & fabian 1995 ; rozanska , kowalska & goncalves 2008 ) by carrying out a systematic investigation of the stability curves and finally to give a quantitative handle on the relative stability nature of the wa as a function of the various physical parameters involved . in
we discuss the physical properties of the wa . in
we give all the constraints and assumptions used in the paper .
the simulations of the stability curves are sensitively dependent on the underlying atomic physics used by the photoionization codes .
we have shown ( chakravorty @xmath2 , 2008 , hereafter paper i ) , that dielectronic recombination rate coefficients have a large effect on the stability curve .
using recent atomic physics databases that include many more of these rates , with more accurate values , is important because it significantly changes the predictions for the physical nature of the wa .
the serious impact of this effect is addressed in further details in while we discuss the stability curve and comparison with earlier works .
we choose a specific scheme to represent the ionizing continuum ( discussed in ) and vary its parameters to create a reasonable ensemble of spectral shapes as demonstrated by a wide variety of agn which harbour wa .
the effect of these different seds on wa nature is discussed in .
rozanska ( 2008 ) have demonstrated that for sufficiently soft agn sed , the thermal balance becomes dependent on the density values if it is above a certain threshold .
this can only happen if the gas is dense enough and the ionizing continuum is appropriate to induce heating processes other than photoionization to dominate the thermal balance .
since we sample a rather large variety of ionizing continua for our investigations , the present work gives us an appropriate opportunity to study the effect of density on stability curves in .
optical and ultraviolet emission and absorption line studies suggest that central regions of agn have solar or higher metallicities ( hamann & ferland , 1999 ) and the wa temperature range of the stability curve is sensitive to the atomic interactions due to various metals .
this abundance dependence provides the motivation for our study of the influence of individual elements , and groups of elements in .
the improved updated atomic database used for the simulations provides the opportunity to go beyond the qualitative discussions and , for the first time , make quantitative comparisons of the multiphase nature and the robustness of the stability regions as a function of the ionizing sed and chemical composition of the absorber in .
our results are summarised in .
studies of wa variability in response to continuum changes show that it is reasonable to assume the wa to be in ionization and thermal equilibrium ( nicastro @xmath2 , 1999 ) .
the timescales to reach equilibrium seem to be less than a day ( nicastro @xmath2 1999 ; krongold @xmath2 2003 ; krongold @xmath2 2005a ; krongold @xmath2 2005b ; krongold @xmath2 2007 ) .
equilibrium is reached when photoionization is balanced by recombination , and ionization and compton heating are balanced by recombination and compton cooling .
the compton heating and cooling become important at high temperature which is solely determined by the shape of the illuminating continuum , and typically is @xmath6 for standard seyfert 1 spectral energy distributions , which the clouds can attain only when the ionizing continuum has significant number of hard x - ray photons .
the typical column density in was , estimated mainly using absorption edges , is @xmath7 ( see the references in ) . in some cases lines
are easily detected whereas edges are not seen ( kaastra @xmath2 2002 ; rozanska @xmath2 2004 ) .
however , the @xmath8 values derived from edges alone are systematically high by large factors ( @xmath910 ) - see krongold @xmath2 2003 , kaspi @xmath2 2002 .
modeling of the absorber allows the temperature and ionization parameter to be calculated .
typically , the temperatures are of the order of @xmath10 and the ionization parameter @xmath11 ( see definition of @xmath3 below ) . however , in some objects was are found to have additional components which exhibit higher temperatures , @xmath12 ( see the references in ) .
the estimated distance of the wa from the central source covers a wide range : from the kiloparsec scale of the extended narrow emission line region ( kinkhabwala @xmath2 2002 ) , through distances of @xmath13 , at the scale of the obscuring molecular torus ( krolik & kriss , 2001 ; crenshaw @xmath2 2003 ; blustin @xmath2 2005 ) , to within the @xmath14 broad emission line region , on the scale of the accretion disk , which is a few thousand schwartzchild radii ( krongold @xmath2 , 2007 ) , where the wa would form the base of an accretion disk wind ( murray & chiang 1995 , elvis 2000 )
. the photoionization state of the wa can be parametrised by the ionization parameter which is the ratio of the ionizing photon flux to the gas density ( tarter , tucker & salpeter , 1969 ) : @xmath15 } , \labequn{ionparm}\ ] ] where @xmath16 is the luminosity between @xmath17 rydberg .
hence @xmath18 , @xmath19 being the gas pressure in the wa .
although , in this paper we have used @xmath3 for all our discussions , in the context of warm absorbers , there is another popular definition of the ionization parameter given by @xmath20 @xmath21 ( @xmath22 ) is the rate of incident photons above 1 rydberg and c is the velocity of light . for a given shape of the sed of the ionizing continuum ,
the conversion from @xmath3 to @xmath23 is unique and can be done conveniently .
although luminosity can be constrained from separate independent observations , it is difficult to remove the degeneracy between density and distance in the product @xmath24 featuring in the denominator of the ionization parameter .
thus , the estimation of distances for wa becomes challenging ; for example , netzer 2003 and krongold 2006 put upper limits differing by up to three orders of magnitude on the wa distance in the same object ngc3783 ( but using different observations ) . as the mass loss rate is proportional to the radius such large factors of uncertainty
have a large effect on agn energetics and on feedback from the agn to the surrounding media .
krolik , mckee and tarter ( 1981 ) showed that a stable state exists at @xmath25 for a photoionized cloud which undergoes collisional cooling .
there can be an additional stable state for the absorbing cloud at @xmath26 , if the incident photoionizing continuum has a hard x - ray component as an extra heating agent .
thus they suggest that presence of a hard x - ray continuum would result in clumping , with cold clouds remaining in equilibrium with a hot medium .
however , for optically thin , ionized gas , the temperature range @xmath27 is well known to be an unstable zone , bringing into question the model of wa as gas in thermal equilibrium .
radiation in spectral lines is the primary cooling mechanism in this temperature range .
the contribution from radiative recombination and two - photon continuum emission ( gehrels & williams 1993 ) form some additional small regions of stability .
we assume a simple model of a geometrically thick , optically thin , plane parallel slab of gas illuminated on one face for the absorbing medium which is located at a certain distance @xmath28 from a central continuum source of bolometric luminosity @xmath29 .
the geometrical thickness of the gas is taken to be much less than @xmath28 .
the physical condition of the absorbing medium is determined by the flux received at the illuminated face of the gas ( @xmath30 ) , the shape of the ionizing continuum ( sed ) , the number density of the gas @xmath31 , the column density of the gas @xmath8 , and the chemical composition of the absorbing gas . to obtain the equilibrium conditions for a gas with an assumed set of parameters which specify these inputs , we have used the publicly available photoionization code cloudyversion c07.02 ( hereafter c07 ) , see ferland ( 1998 ) .
cloudy calculates the ionization equilibrium conditions by solving the energy and charge conservation equations under the assumption that all the atomic processes have had time to reach a steady state .
the density @xmath32 of @xmath33 species is then given by the balance equation @xmath34 where @xmath35 is the rate of species @xmath36 going to @xmath37 , _ source _ is the rate per unit volume of appearance of new atoms in @xmath33 species and _ sink _ is the rate that they are lost ( osterbrock & ferland , 2006 ) .
thermal equilibrium conditions are calculated by evaluating the heating and cooling rates at different temperatures for the given gas and finding the temperatures where the two rates balance each other .
for all the calculations in this paper we have used cloudy in the _ constant density _ mode which is chosen for the sake of numerical simplicity , as we probe a rather large range of parameter space to determine the properties of the wa .
we would like to note here that using the titan code ( dumont , abrassart & collin , 2000 ) , rozanska @xmath2 ( 2006 ) and goncalves @xmath2 ( 2006 ) have pioneered the _ constant total pressure _ mode of calculations for photoionization equilibrium where they relax the approximation that the density is constant .
their only assumption is that the line of sight gas is in total pressure equilibrium and they introduce additional iterations between temperature and density within the code in a self consistent manner .
this results in natural stratification of the cloud with different temperatures and densities where each layer is illuminated by the radiation already absorbed by the previous layer .
these two methods agree well with each other for optically thin clouds , but show significant differences when the column density of the absorbing cloud is high ( @xmath38 ) . in our future publications where we will restrict ourselves to study equilibrium properties of wa as functions of a smaller range of physical parameters , we will use similar methods .
estimation of temperature is independent of @xmath31 , for a gas in thermal and ionization equilibrium when the dominant processes of heating and cooling are photoionization and recombination respectively .
this is because in this regime of thermal equilibrium , density enters the temperature estimation only through the ionization parameter .
hence the stability curve would be insensitive to the value of the density used .
this gives us the liberty to choose an intermediate value for wa density , @xmath39 , leading to @xmath40 from , for typical values @xmath41 corresponding to wa .
except for , where we check the robustness of this assumption , we have assumed the value @xmath42 for @xmath43 throughout the paper . the @xmath44 part of the observed spectra of agn is usually modeled as a simple power - law with exponential lower and upper cut - offs specified by @xmath45 and @xmath46 respectively , @xmath47 where @xmath48 is the flux per unit frequency interval and @xmath49 is the energy spectral index . for most quasars ,
the x - ray spectral index lies in the range @xmath50 as shown by wilkes & elvis in 1987 and also recently by grupe @xmath2 2006 and lopez @xmath2 2006 .
however some objects show extreme values like @xmath51 or @xmath52 .
there is a systematic difference between radio - quiet and radio - loud agns in that radio - quite agns have steeper x - ray spectra ( @xmath53 ) than radio - loud ones , ( @xmath54 , wilkes & elvis 1987 ; grupe @xmath2 2006 ; lopez @xmath2 2006 ) .
ionizing continua given by have been used earlier by reynolds & fabian ( 1995 , hereafter rf95 ) for doing detailed analysis of nature of wa as a function of various relevant physical parameters .
their work is further discussed in .
the dashed and dotted curve labeled as _ early set _ in the lower panel of is drawn using with @xmath55 , @xmath56 and @xmath57 . in the lower panel of ,
the _ standard set _
continuum is compared with various other seds used or discussed in this paper and we will refer to these comparisons in more details in . in this paper we have used the simple definition for the ionizing continuum , as in , only in where we have given an introductory description of the stability curve and compared results with similar works in literature ( rf95 ) . in all other sections we have considered the more realistic _ standard set _ spectra .
all through the paper we have assumed @xmath69 , and for a given value of @xmath70 , obtained the desired value of @xmath58 by adjusting @xmath63 . in , when we investigate the changes in the nature of the stability curve with variation in the shape of the continuum , we consider a wide range of slopes , @xmath71 and @xmath72 and a wide range of higher energy cutoff , @xmath73 to incorporate various types of agn .
the stability curve , may be considered to be a phase diagram in the @xmath74 plane , with each point on the curve representing an equilibrium state of the system at a temperature @xmath75 , for ionization parameter @xmath3 .
the other parameters describing the system have specific values as mentioned below .
the solid curve in is generated using version c07 of cloudy , for an absorber having gas density @xmath76 and solar metallicity , being ionized by a power - law continuum having energy spectral index @xmath77 and extending from @xmath78 ; hereafter we will refer to this as the _ early model _ and the corresponding parameter set as the _
early set_. computations were done using a grid of models with a range of values of the ionization parameter @xmath3 , and the code is constrained to perform single zone calculations to ensure that the gas is optically thin . the ionizing continuum used in the _ early model _
is shown as the dotted - and - dashed curve in the lower panel of .
we use the _ early set _ in this section to compare results with earlier works like rf95 who have done extensive study of the stability properties of the wa as a function of various shapes of the ionizing continuum .
an isobaric perturbation of a system in equilibrium is represented by a small vertical displacement from the stability curve ; such perturbations leave @xmath79 constant , which for constant @xmath43 leave the pressure unchanged .
if the system is located on a part of the curve with positive slope , then a perturbation corresponding to an increase in temperature leads to cooling , while a decrease in temperature leads to heating of the gas .
a gas in such a state is therefore thermally stable . on the other hand ,
a gas located in a region of the curve with negative slope is thermally unstable , and may achieve a multiphase equilibrium under certain physical conditions which we will discuss in later sections . before we go on to discuss our own results with the stability curve , here , we would like to mention some of the earlier works done by various authors on wa and/or stability curves .
krolik @xmath2 ( 1981 ) obtained the stability curve for agn using parameters similar to our _ early set _ except for the ionizing continuum .
the continuum used by them is plotted as the dotted curve in .
they have argued in favour of @xmath80 stable states remaining in pressure equilibrium with the @xmath81 phases . although there are some _ narrow _ stable phases at @xmath82 , they have ignored the significance of such states .
moreover their stability curves do not show any stable states at @xmath12 . in a similar study on the absorbing medium , mathew & ferland ( 1987 )
do not find the @xmath80 and @xmath81 phases in pressure equilibrium with each other .
the continuum used by mathew & ferland ( 1987 ) is widely used as a broad band continuum for agn modeling studies and has been represented with the dashed line in .
the energy spectral index of the ionizing continuum in the _ standard _ set is very similar to that of the mathew & ferland sed in x - rays ( 0.3 - 100 @xmath83 ) .
however , it is not objective to make a comparison at this stage because , although the ionizing continuum in the _ standard _ set given by represents the overall shape of the sed between 3.6 @xmath84 to 2 @xmath83 using the definition of @xmath58 , it is unable to give information on any underlying specific features in this energy range . in this paper
we will constrain ourselves to use ionizing continuum given by and , but will investigate the details of the spectral shape in the energy range ( 1 - 1000 @xmath84 ) in subsequent publications ( chakravorty @xmath2 , in preparation ) by incorporating the effect of soft excess at @xmath85 and a disk blackbody component at @xmath86 .
reynolds and fabian ( 1995 , rf95 ) have studied the warm absorber stability conditions using physical conditions as described by our _ early set _ of parameters .
they used a simple power - law for the ionizing continuum and varied the power - law index @xmath70 to check for its effects on the wa and also examined some basic effects of adding a soft excess component to the power - law ionizing continuum . using power - law continuum with @xmath55 , they find that the stability curve is independent of density variation in the range @xmath87 hess ( 1997 ) address the discrepancy that the physical properties inferred from observations of some low - mass x - ray binaries and seyfert galaxies correspond to thermally unstable regions of the stability curve predicted by photoionization codes .
they have carried out an extensive study of the stability curve as a function of various ionizing continuum and different abundance of the absorbing gas .
they have studied the cause of the instabilities in great detail and have cautioned against the outdated fits and parameters used by the simulation code and also against the accuracy of the atomic data then used for determining the thermal balance .
komossa & meerschweinchen ( 2000 ) and komossa & mathur ( 2001 ) have extensively studied the stability curve respectively as a function of ionizing continuum and chemical composition of the absorber with the aim of explaining the wa in various narrow line seyfert 1 galaxies .
they find that steeper ionizing continuum stabilizes wa and super - solar metallicity of the gas enhances the possibility of multiphase wa . by modeling the observations of ngc4051
, indications are found that the wa in this object has super - solar abundance . in the wa temperature range
@xmath88 cooling processes are dominated by collisional recombination which can be radiative or dielectronic . these processes
lead to local regions of thermal stability in this otherwise unstable temperature range as discussed by gehrels & williams ( 1993 ) , hess , kahn & paerels ( 1997 ) and also depicted by rf95 .
we can see these effects in our stability curve generated using the _ early set _ of parameter values and shown as the solid curve in .
however the atomic physics data base has been updated considerably over the last decade and our results show significant quantitative variations from earlier work which we emphasise by making explicit comparisons with rf95 .
the dotted line in is a reproduction of the stability curve ( using version c8412a of cloudy ) in figure 3 of rf95 where they had used the _ early set _ of physical parameters .
the variations in the c07 stability curve from the c84 curve lead to different physical predictions for the wa ; for example , the @xmath89 state becomes much more stable in our work resulting in more pronounced possibility of multiphase nature of the wa .
the @xmath90 region of the stability curve , relevant for wa , sensitively depends on the detailed atomic physics of the various elements which contribute to photoelectric heating and cooling due to recombination and collisionally excited lines . in paper
i we have shown that the total recombination rates ( dielectronic + radiative ) , for the significant cooling agents which bring about the maximum variation in the stability curves in , are predicted to be larger in c07 than in c84 .
c c c c c c c c c c c c c c c c c c c & & & & & & & & + [ 0cm][0cm]model & & & & & & & & & & & & [ 0cm][0cm]low energy & & [ 0cm][0cm]@xmath31 & & [ 0cm][0cm]@xmath91 & & [ 0cm][0cm]metallicity + & & definition & & @xmath70 & & @xmath58 & & @xmath92 & & @xmath93 & & [ 0cm][0cm]cut off & & & & & & + + + + _ early _ & & & & 0.8 & & - & & - & & @xmath94 & & @xmath56 & & @xmath95 & & zone 1 & & @xmath96 + & & & & & & & & & & & & ( ) & & & & ( see text ) & & + + + + _ standard _ & & & & 0.8 & & 1.2 & & 2.0 & & @xmath97 & & & & @xmath98 & & @xmath99 & & @xmath96 + & & & & & & & & & & & & & & ( see text ) & & & & + + in the wa temperature range @xmath100 , dielectronic recombination dominates over radiative recombination for many ions ( osterbrock & ferland , 2006 ) . unlike the radiative recombination rate coefficients , dielectronic recombination rate coefficients ( drrc ) have undergone significant changes over the last decade .
the importance of revisiting the values of drrc were first pointed out by savin ( 1997 ) and further emphasised by a series of papers by badnell and coworkers ( colgan , pindzola & badnell 2004 , colgan 2003 , altun 2004 , zatsarinny 2004a , mitnik & badnell 2004 , zatsarinny 2003 , gu 2003 , zatsarinny 2004b and gu 2004 ) .
the drrc data base is more extensive now and shows that the rate coefficients for different ionization states of various elements are substantially larger than the values used previously , but the database is still not complete , especially for the lower ionization states . for ions which do not have computed drrc values
, c07 uses a solution , as suggested by ali ( 1991 ) ; for any given kinetic temperature , ions that lack data are given drrc values that are the averages of all ions with the same charge .
the advantage of this method is that the assumed rates are within the range of existing published rates at the given kinetic temperature and hence can not be drastically wrong .
however , in paper i , we have checked whether the points in the c07 stability curves have computed drrc values or guessed average values , concentrating on the parts of the curve which have multiphase solutions for the warm absorber ( @xmath101 ) and are different from the c84 stability curve .
we find that all the ions which act as major cooling agents for each of these points in the stability curve have reliable computed drrc values which are not likely to change in near future .
thus , the new data base provides a more robust measurement of the various physical parameters involved in studying the thermal and ionization equilibrium of photoionised gas .
this not only gives us the motivation to improve upon previous qualitative stability curve analysis by various authors ( hess , kahn & paerels 1997 ; komossa & mathur 2001 ; krolik , mckee & tarter 1981 ; krolik & ; kriss 2001 ; reynolds & fabian 1995 ; rozanska 2008 ) but also provides the chance to extend their work with quantitative estimates on the nature of the multiphase wa in .
the ionizing continuum used in the _ early set _ has now been shown by improved spectral uv and x - ray measurements not to correspond to observed continua ( zheng @xmath2 1997 , shang @xmath2 2005 ) .
as discussed in , we will use a more realistic sed with two power - law components given by and lower energy cut - off given by .
the fiducial value for @xmath58 for seyfert1 galaxies is 1.2 ( netzer 1993 ) .
we define a _ standard set _ of parameter values to be used in the rest of the paper where the ionizing continuum shown as the solid line in is parametrized by @xmath77 , @xmath102 and @xmath63 is appropriately adjusted to attain @xmath103 .
@xmath93 for both components is @xmath104 .
the high energy cut - off is an average of the observational values reported from beppo - sax which found hard x - ray spectral cut - off to span the range 50 to 450 @xmath83 ( matt 2004 ; tueller 2008 ) .
the absorbing gas in this _ standard _ model is assumed to have solar abundance ( @xmath105 ) as given by allende prieto , lambert & asplund ( 2002 , 2001 ) , holweger ( 2001 ) and grevesse & sauval ( 1998 ) .
the density is given by @xmath106 as in the _
early set_. in addition we assume @xmath107 , for the slab of the gas in the _
standard set_. shows a comparison of the various parameters used in the _ early set _ and the _
in this section we consider the effect on the stability curve of changing the shape of the continuum in by varying @xmath70 and @xmath58 over a range of values .
all through the section we assume that the gas has chemical composition of solar abundance .
the left panels of shows the shapes of the ionizing continua and the corresponding stability curves for different values of @xmath70 , @xmath58 is held at a constant value of @xmath108 for these curves .
we have highlighted the stable @xmath109 warm absorber states .
for very flat continua , with energy index @xmath110 or less , almost no stable state exists in the temperature range of @xmath111 , since there are no regions of the curve here with positive slope except for a very narrow one at @xmath112 .
the large number of high energy photons pushes the comptonization temperature to @xmath113 . as noted by krolik @xmath2 ( 1981 )
, there is a cooler phase at @xmath114 and a hot phase which is transparent to atomic features .
warm absorbers are thus unlikely to be found in agn which have such flat x - ray spectra .
stability curves for steeper incident spectra show additional stable states in the intermediate temperature regime .
curves corresponding to @xmath115 and @xmath77 are both stable up to @xmath116 .
the _ standard _ curve shows one extensive stable state at @xmath12 . we will discuss the effect of multiple phases further in . for steeper spectra with @xmath117 ,
the absorbing gas is predicted to be stable for all @xmath3 with the compton temperature now lying at @xmath118 .
the distinct multiple phases at different temperatures , but in near pressure equilibrium are lost . for such steep continua ,
stable states exist over a continuous distribution of @xmath4 and @xmath75 in the phase space .
in summary , agn having flat x - ray spectra with @xmath119 are not likely to have was .
although a steeper continuum can lead to physical conditions that can sustain warm absorbing gas in thermal equilibrium , multiphase nature of wa can only be seen for agn spectra with intermediate slopes of @xmath1 .
it is interesting to note that multiphase was are predicted for continua having slopes in the range @xmath50 which is the observed range for the bulk of the observed agn population ( wilkes & elvis 1987 ; grupe @xmath2 2006 ; lopez @xmath2 2006 ) .
we would like to note an interesting difference between results obtained by us and that by krolik & kriss ( 2001 , hereafter kk01 ) . using a model ionizing continuum given by kaspi ( 2001 ) ,
kk01 have obtained stability curves computed by version 2.1 of xstar ( kallman 2000 ) .
the ionizing continuum by kaspi ( 2001 ) has values of @xmath120 and @xmath121 , very similar to those used in the _ standard set_. however , kk01 find that the corresponding stability curve shows a steep near vertical rise in temperature from @xmath122 to @xmath12 for @xmath123 which gives a continuous distribution of stable states over a wide range of @xmath3 , while pressure ( @xmath4 ) remains almost constant .
for the same range of @xmath3 , ( @xmath124 ) , our _ standard _ curve , shows two distinct wa phases in pressure equilibrium with each other and the change in temperature is gradual and spans a much wider range of @xmath4 values . in the right panels of we
have shown the results of varying the euv to soft x - ray slope @xmath58 ( see ) .
the x - ray slope @xmath70 is held at a constant value of @xmath125 .
observations find @xmath58 between @xmath126 and @xmath127 as reported by strateva @xmath2 2005 and steffen @xmath2 2006 .
they find that there is an anti correlation between @xmath58 and the ultraviolet luminosity @xmath128 as was first reported by tananbaum in 1979 .
we have used the representative values of @xmath129 in addition to the _ standard _ value @xmath108 .
the wa phases are highlighted as before . from we
see that the nature of stability in the @xmath130 curve is very similar to that in the @xmath67 curve .
however , beyond @xmath108 , with a sharper drop in flux from the euv to the soft x - ray region , with steeper value of @xmath58 , there is a gradual increase in the stability of the absorbing gas at @xmath131 , and a gradual shift to higher pressure ( @xmath4 ) .
the @xmath132 curve shows three distinct stable states at @xmath133 in pressure equilibrium with each other . for steeper @xmath134 , stability at @xmath89
is increased .
the @xmath135 curve shows no unstable states . for such spectra , the wa can have a continuous distribution of pressure and density
the compton temperature for the stability curves understandably gradually drops from @xmath136 for @xmath137 to @xmath138 for @xmath139 as the ratio of euv to x - ray photons increases and comptonization is progressively dominated by lower energy photons .
to summarise , the existence of the wa is independent of the euv to x - ray slope of the ionizing continuum .
however , its multiphase nature is influenced by this slope ; the steeper the fall of flux from euv to x - ray ( larger @xmath58 , the lesser is the possibility of finding a multiphase wa .
the _ standard set _ ionizing continuum is exponentially cut off at @xmath68 ( see ) .
results from the swift / bat ( burst alert telescope ) hard x - ray sky survey ( tueller et al . , 2008 ) show that , of the brightest few dozen agns , for which @xmath93 can be determined , the lowest value is @xmath94 , while some 2/3 of the objects have @xmath140 , the highest value that bat can measure . as a result an appropriate value of @xmath93 for the _ standard set _ needs to be set well above the _ early set _ value .
we adopt @xmath68 .
the high energy cut - off is important in compton processes and hence in determining the compton temperature of a stability curve , but is not likely to affect its wa region .
however , for the sake of completeness we examine the effect on stability curves of varying the value of @xmath93 .
shows the various stability curves drawn using the different values for @xmath93 .
the curves are appropriately labeled .
all other physical parameters are as in the _
standard set_. for the lowest considered value of @xmath141 , the compton processes start dominating the thermal balance at lower temperatures resulting in enhanced stability at @xmath142
. however , the influence is still not significant enough to change the overall qualitative nature of the wa or its multiphase properties .
the majority of the swift / bat agn have @xmath143 . for such values of @xmath93 ,
the wa remains effectively insensitive to the variation in the high energy cut - off ; the only influence is in increasing the compton temperature with the increase in @xmath93 .
as stated in , we have assumed that the stability curve is insensitive to the variation in the value of hydrogen density @xmath31 of the wa gas . based on this assumption
, we have taken the liberty to choose an arbitrary fiducial value of @xmath144 for @xmath145 , the typical value of ionization parameter reported for was , thus yielding @xmath146 from , which we hold constant throughout this paper except in this section .
many authors carrying out photoionization calculations make similar assumptions ( hess , kahn & paerels , 1997 ; krolik , mckee & tarter , 1981 ; krolik & kriss , 2001 , rf95 ) and rf95 have even verified this statement over a wide range of densities @xmath147 for a simple power - law ionizing continuum with @xmath55 .
however , more recently , rozanska , kowalska & goncalves ( 2008 ) showed that the stability curve shows remarkable density dependence if the ionizing continuum has a complicated shape with relatively more soft x - ray , photons although the same exercise with a simple power - law ionizing continuum with @xmath148 yields no density dependence on the part of the stability curve . in this paper
we have studied the stability curve s behaviour as a function of complicated ionizing continua with having different ratios of the soft to hard x - ray photons .
hence it is an interesting exercise to check if the stability curves , especially in the range @xmath149 , drawn using the different ionizing continua considered in this paper show any variation if the density is changed . in this section
we relax the assumption made in that @xmath146 and for each ionizing continuum considered in , draw several stability curves by varying this ratio from @xmath150 such that the density at @xmath151 varies from @xmath152 in multiplicative steps of ten from one curve to another .
the results are shown in .
the upper panels in the figure show the stability curves for continua with different values of the x - ray slope @xmath153 , whereas the bottom panels correspond to different values of the euv to x - ray slope @xmath154 .
the determination of thermal equilibrium of the wa is seen to be independent of density in the range @xmath155 , whatever be the shape of the ionizing continuum . stability curves with @xmath156 show insignificantly small differences from the rest of the curves with lower densities as we vary @xmath70 from 0.5 through 0.8 to 1.1 this indicates that the soft x - ray slope has little or no role to play in inducing density dependence to the estimations of thermal equilibrium conditions .
the most interesting result of this exercise is that as the euv to x - ray slope @xmath157 is increased from 1.2 through 1.4 and 1.6 to 1.8 , the wa thermal equilibrium properties become significantly density dependent for @xmath158 and the dependence is more pronounced with the increase in @xmath157 i.e. with the increase in the relative number of euv photons .
@xmath159 can be considered to be an upper limit on wa density and would be consistent with clouds residing within the @xmath14 broad emission line region , on the scale of the accretion disc , a few thousand schwartzchild radii ( krongold @xmath2 , 2007 ) .
thus higher values of density are unphysical for wa .
however , in each of the panels in , we have also plotted the @xmath160 curve for comparison and it shows gross difference from the stability curves with lower densities for all the continuum shapes considered . since the @xmath139 curves show the maximum sensitivity to density variation , we select them to investigate the cause of this effect .
we choose the @xmath145 model from each of the curves with densities @xmath161 and check what are the dominant heating and cooling agents in determining the thermal balance .
the fractional contribution towards the total cooling rate by the dominant cooling agents undergoes only a gradual change as we change the density though the above mentioned range .
the remarkable change is shown by the dominant heating agent . for @xmath162 ,
the heating agents are all photoionized species and free - free heating starts to dominate for @xmath163 .
the percentage of heating by free - free absorption increases from @xmath164 for @xmath163 to @xmath165 for @xmath166 .
the ionization equilibrium conditions of the absorbing gas can be uniquely determined if the shape of the ionizing continuum and the flux received by the gas is specified .
assuming that the absorber is in ionization equilibrium , cloudy evaluates the heating and cooling rates at different temperatures for the given gas and determines the thermal equilibrium solution for the temperature where these two rates balance each other . in the wa temperature regime where photoionization is the dominant heating process and thermal equilibrium
is achieved by recombination cooling , the temperature estimates become independent of density as both the heating and the cooling terms have similar dependence on density which cancels out from both sides of the thermal balance equation . at higher temperatures ,
the thermal balance is attained by compton heating and cooling , in which case also , the density cancels out . however , the interplay between various heating and cooling processes is also a function of the ionization continuum . for spectral shapes with steep @xmath157 there maybe sufficiently large number of low energy photons to make the free - free absorption by ionized hydrogen
dominate over photoionization as the heating process .
the heating rate for this interaction has a different dependence on density and will result in making the thermal equilibrium conditions sensitive to density variations for @xmath31 greater than a certain threshold value .
the interesting point to note at the end of this exercise is that although the mode of photoionization calculations used by us and rozanska ( 2008 ) are different , our results are qualitatively consistent with each other . for sufficiently soft uv dominated ionizing continuum
the thermal equilibrium calculations become density dependent if the gas is denser than a certain threshold .
since the gas is also assumed to be in ionization equilibrium , the ionization structure will also be influenced .
the effects would be much more pronounced for _ constant pressure _
mode of calculations as shown by rozanska ( 2008 ) .
the importance of this result is that for dense gas illuminated by uv dominated ionizing continua , the density of the gas can be estimated directly by modeling the observed spectra in uv .
the behavior of the models along the stability curve , from @xmath167 to @xmath168 is sensitive to the role of atomic physics in determining the thermal state of the gas .
the photoionization cross section @xmath169 from the ground level of @xmath170 ion with the threshold @xmath171 is given by the equation @xmath172 where @xmath173 is the number density of the @xmath170 ion , @xmath174 is the number of excitations per particle to the @xmath175 state and @xmath176 is the mean intensity .
thus the function @xmath177 is a measure of the interaction between the radiation field and the @xmath170 ion . in
we have shown the dependence of @xmath178 on @xmath3 and hence on the ionization state of the absorbing gas .
we have selected three different models along the _ standard _ stability curve and plotted @xmath179 as a function of energy with the values of @xmath180 and @xmath181 indicated for each curve ; @xmath182 being the hydrogen density .
the absorption edges of various important ions including cv , cvi , ovii , oviii , fexvii and nex which are the signatures of wa have been indicated with dotted lines . as the gas becomes more ionized by factors of @xmath183 its interaction with radiation decreases by orders of magnitude , especially at energies less than @xmath184 . at any given energy
the value of @xmath185 is lower for higher @xmath3 .
however , with highly ionized gas , the interaction shifts to the higher energies . for @xmath186 ,
there is no sharp increase in @xmath185 at energies corresponding to the wa edges ; these interactions grow with @xmath3 as seen in the middle and the bottom curves corresponding to @xmath187 and @xmath188 respectively .
we see above that at values of @xmath189 relevant for was , the nature of the interaction between gas and radiation is significantly influenced by the atomic physics of the heavier elements . in the following subsections
we have studied the effect of different chemical compositions of the absorber on the stability curve .
the illuminating continuum is assumed to be the same as that in the _ standard set _ values ( table 1 ) .
we have generated the stability curves for ( i ) super - solar abundances , ( ii ) sub - solar abundances and ( iii ) other abundances where specific elements or group of elements have been removed from the solar metallicity .
the left panel of shows stability curves for gas having super - solar abundances as defined in hamann & ferland ( 1993 ) .
in addition to the @xmath190 and the @xmath191 phases , all the curves also exhibit wa states .
however the temperatures of the wa phases and their distribution in the phase space change with metallicity .
the nature of stability exhibited can be separated into three different groups showing distinctly different trends .
all the curves in the group with solar to three times solar exhibit stable gas at @xmath192 , with the extent of the stable region decreasing as the metallicity is increased .
@xmath193 for the _ standard _ curve , @xmath194 being the ionization parameter corresponding to the middle of the @xmath89 wa phase .
the range of ionization parameters for the @xmath89 gas is @xmath195 for the _ standard _ curve .
the @xmath196 drawn in the left panel of indicates that the @xmath89 wa is pushed towards higher values of @xmath3 as the metallicity increases .
another stable states appears in all the curves at @xmath197 , but this time with the increase in metallicity , the extent of the stable region is increased . @xmath198 for the _ standard _ curve , where @xmath194 is defined to be the ionization parameter corresponding to the middle of the @xmath199 stable wa phase which shows the range @xmath200 for the _ standard _ curve .
the solid black line across the @xmath199 phases of the different stability curves is the @xmath201 line .
thus , we see that with the increase of metallicity , @xmath194 is decreased slightly and the @xmath199 stable phase moves to lower @xmath4 and higher @xmath75 . a very narrow ( in the range of @xmath202 ) phase
is seen to appear at @xmath203 for metallicity of 3 @xmath204 solar .
the detailed quantitative comparisons are made in .
the curves in the next group with @xmath205 , are very similar to each other , with a very narrow stable state at @xmath131 , which is not in pressure equilibrium with the @xmath206 phase and the @xmath207 state becomes more prominent .
as abundances are increased further ( @xmath208 ) the stable state at @xmath131 disappears , while the @xmath209 states remain quite wide and in pressure equilibrium with one another .
the stability curves for the super - solar abundances thus suggest that the possibility for the warm absorbing medium to exist in discrete phases in pressure equilibrium with each other increases with increasing metallicity .
we return to these issues in with quantitative details .
the stability curves for the super - solar metallicities with @xmath210 are seen to have gas at @xmath211 to be in pressure equilibrium with the wa states at higher temperatures .
such stable states at @xmath80 are also highlighted in the corresponding stability curves in the left panel of .
gas at @xmath80 might be representing the ` belr clouds ' .
hence this phenomenon might indicate the interesting possibility of some connection between the x - ray and uv absorber , as discussed by turner @xmath2 ( 1995 ) and elvis ( 2000 ) .
however , these results have to be checked out further with more realistic agn spectra which is beyond the scope of this paper and will be addressed in our next publication ( chakravorty , in preparation ) .
we have also investigated the possibility of high metallicity gas making it possible for wa to exist even when the x - ray ionizing continuum is very flat .
the stability curves are shown in the small inset window of the left panel of . even for solar abundance
, the @xmath212 curves did not show any significant @xmath89 wa and higher metallicity diminishes the chances further . at higher temperatures of @xmath12 ,
very narrow regions of stability appear for @xmath213 .
however , even in these cases there is no multiphase exhibited by the stability curves . thus even if the gas has high metallicity , was are not likely to be present in the intervening medium , if the ionizing continuum is very flat .
gas in agn has super - solar metallicity ( hamann & ferland , 1999 ) .
however , agn illuminate low density gas on large scales which has low abundances , especially at high redshift .
sub - solar chemical composition are relevant for such line of sight absorbing systems including damped lyman-@xmath70 systems in the inter galactic medium and smaller galaxies like the magellanic clouds .
the inter - cluster medium , for example , has @xmath214 abundance and in certain systems there are even stars with an abundance of @xmath215 ( pettini , 2006 and references therein ) .
it is , therefore , interesting to examine the effects of sub - solar abundance on the stability curve .
we show the low metallicity stability curves in the right panel of . as before ,
the solid black lines across the stable wa phases are the constant @xmath3 lines .
the ionization parameter @xmath216 corresponding to the middle of the @xmath89 stable phase increases from its standard value of @xmath193 with the decrease in metallicity and there is gradual enhancement of the thermal stability of gas . on the other hand , with almost constant @xmath198 , the stable phase at @xmath217 moves towards lower temperature and higher pressure ( @xmath4 ) and its extent in the @xmath218 shrinks until such a phase disappears for @xmath219 .
detailed quantitative comparisons will be presented in . absorbing gas with _
zero metals _ might be relevant for super massive black holes ( smbhs ) or their seeds at redshift @xmath220 , which are formed by direct collapse and where no star formation precedes the radiating smbh ( volonteri & rees , 2005 ; begelman , volonteri & rees , 2006 ) .
we have shown the _
zero metals _ stability curve in the right panel of .
there is a single phase wa at @xmath89 with significantly high @xmath221 as compared to the _ standard _ @xmath222 .
has shown how the interaction between gas and radiation is sensitive to the ionization potentials of various ions of elements like c , o , fe and ne indicating that the wa states are likely to be affected not only by the overall abundance of the absorber , but also by specific elements and groups of elements .
shows the effect on the stability curves if certain elements or groups of elements are not present in the absorbing gas . in both the panels of the figure
, we have also drawn the stability curve for the _ standard set _ of parameters and _ zero metals _ for comparison .
the wa consists of absorption edges and lines of different ionization stages of oxygen , carbon , neon and iron .
we define the _ x - ray group _ constituting of only these four elements .
the left frame of shows the results of simulating various stability curves corresponding to the removal of different individual elements of the group from a solar composition gas .
setting carbon abundance to zero does affect the stability curve at @xmath114 and the deviations imply that carbon acts as a cooling agent at those temperatures .
however at higher temperatures relevant for wa , carbon has no significant effect .
the removal of oxygen and neon leads to deviations in the stability curves which indicate that for @xmath223 oxygen acts as a significant cooling agent with some contributions from neon , whereas for @xmath224 both act as heating agents .
removal of iron has the maximum effect as it renders the curve relatively featureless in the range @xmath225 .
such an effect is expected because at higher energies iron is the most abundant absorbing element and plays a major role to stabilise the gas by acting as a significant heating agent .
this result has interesting cosmological consequences .
type ia supernovae created iron when the universe was about 1 gyr old ( hamann & ferland , 1999 and the references therein ) .
hence there would be an interesting transition in the nature of the wa at this age ; was devoid of iron from before this era can not show a significant @xmath199 phase .
future missions like constellation x might be able to detect such a change .
like the removal of iron , the removal of the _ x - ray group _
( o , c , ne , fe ) as a whole also _ flattens _ out the stability curve , but to a larger extent as shown in the right panel of .
if the absorbing system is devoid of the _ x - ray group _ , the stable states at @xmath89 are pushed towards higher values of @xmath3 with @xmath226 as compared to @xmath222 for the _ standard _ curve and there are no higher temperature stable states thus nullifying the possibility of multiphase wa .
in addition to the elements of the _ x - ray group _ , it is instructive to study the effect of the @xmath70-elements ( ne , mg , si , s , ar , ca , ti ) produced together with oxygen in type ii supernovae which precede the type ia supernovae where iron and associated elements are formed . thus the relative abundance of these elements gives a handle on the stellar evolution history and the age of the galaxy .
ne is the only @xmath70-element which has been studied as a part of the _ x - ray group _
( o , c , ne , fe ) as well , and its effect on the stability curve has already been discussed above . among the remaining @xmath70-elements , magnesium , sulphur , argon , calcium and titanium have negligible effects .
only silicon is a significant cooling agent at @xmath227 , so that removal of silicon shows an almost vertical jump from @xmath89 to @xmath199 in the corresponding stability curve in the left panel of . as a group
however , @xmath70-elements suggest interesting consequences . according to the evolution history of metals in the universe ,
the @xmath70-elements are formed first , followed by the other metals in subsequent stages .
hence we investigated the nature of an absorbing gas which has only the @xmath70-elements as metals and the corresponding stability curve is shown in the right panel of labeled as ` @xmath70-elements only ' .
for @xmath228 the curve is same as the `` _ x - ray group _ removed '' curve and at lower temperatures , @xmath229 , it traces the @xmath219 curve .
these results suggest that a certain amount of evolution of metals is necessary for any absorbing systems to show wa signatures and might set an upper limit to the redshifts at which warm absorbers can form . in
we have collected the curves showing the major effects due to varying the chemical composition of the absorber . as the metallicity of the absorber
is decreased from @xmath230 to @xmath96 and to @xmath231 , the stability curve flattens gradually .
the possibility of finding multiphase wa decreases with decrease in metallicity .
this result is in conjunction with those found by komossa and mathur ( 2001 ) .
it also decreases if the absorbing medium is deficient in some elements .
if a wa lacks iron , but has all the other elements with solar abundance , it then has nearly the same stability properties as an absorber with half solar abundance .
the x - ray group of elements are the most influential in making multiple phases possible . absorbing gas at early epochs of star formation when only @xmath70-elements formed
show the same thermal properties as absorbers devoid of the x - ray group of elements , thus suggesting that the properties of wa are significantly dependent on the evolution history of metals in the universe . also shows the stability curve for @xmath232
; such a primordial gas of hydrogen and helium has a single phase of absorbing gas at @xmath89 for a significantly high ionization parameter .
no stable state at @xmath199 and hence no multiphase solutions are possible if the metallicity of the absorber is @xmath233 .
metals thus seem to be necessary for absorbing gas to exist at @xmath109 .
hess ( 1997 ) have carried out an extensive study trying to account for the unstable thermal equilibrium zones in temperature ranges relevant for agn and x - ray binaries paying attention to even individual ions of various metals responsible for causing the instabilities .
they have predicted the ionization structure and thermal states of absorbers with a wide variety of chemical compositions .
the results posed in this paper especially in agree qualitatively with their results but as expected , they vary in quantitative details . for example , like them , we also find oxygen and iron to be playing crucial roles in the ionization and thermal balance equations .
however , unlike hess in this paper we find that oxygen and iron influence distinctly different temperature zones of the stability curve .
as mentioned in their paper , such minute details would crucially depend upon the accuracy of the atomic physics incorporated which is definitely addressed much better in the current photoionization codes than in 1997 .
the determination of the particular ions responsible for the heating or cooling of the gas at wa temperatures , is very sensitive to the exact shape of the ionizing continuum between 50 to 500 ev .
such investigations and quantitative comparisons are beyond the scope of this paper and will be addressed in our subsequent publications ( chakravorty , in preparation ) .
the equilibrium curves we have derived in the previous sections often contain segments that allow phases at different temperature to occur at similar pressure @xmath4 .
this has clear physical interest as distinct phases in pressure equilibrium have been derived in a number of cases of was ( krongold @xmath2 , 2003 , 2007 , netzer @xmath2 , 2003 ) , while in others it has been claimed that there is a continuous range of ionization parameter ( ogle @xmath2 , 2004 , steenbrugge @xmath2 , 2005 ) . in this section
we quantify the extent of these multiphase regions .
however , we would like to note that for carrying out such an exercise , the atomic database needs to be accurate .
as mentioned a number of times in the previous sections , care has been taken to upgrade the underlying atomic physics in cloudy with the most recent updates . in
we present detailed quantitative estimates of the wa properties for the various parameter sets used in this paper . for convenience of representation
we introduce a few definitions for use in .
the first four columns describe the parameter set used to derive the stability curve and the wa properties .
@xmath234 in columns 5 , 9 and 13 and @xmath235 in columns 6 , 10 and 14 respectively are the ionization parameter and the temperature corresponding to the middle of the @xmath236 stable wa phase , where @xmath235 is expressed in units of @xmath236 .
@xmath237 $ ] in columns 7 , 11 and 15 give the range of @xmath202 over which stable states of wa are exhibited for a temperature of @xmath236 . for wa to exist in discrete multiple phases ,
different temperature stable states require to exhibit the same values of @xmath4 . the larger the range of these common values , higher we presume to be the probability of a multiphase wa . such intersection in the range of @xmath4
is represented by @xmath238 $ ] for @xmath89 and @xmath199 phases and by @xmath239 $ ] for the states at @xmath89 and @xmath199 and their values are noted in columns 8 and 12 respectively .
@xmath240 in the second last column is the number of stable states in the stability curve in the temperature range @xmath241 and the number of states in pressure equilibrium with each other is given by @xmath242 in the last column .
we would note that the values of the various quantities recorded in are useful for understanding trends in the variation of the physical quantities as the parameter sets are changed .
however , for more robust estimations of these numbers and specific predictions of the wa properties , we require more detailed modeling of the ionizing continuum and the gas abundances . such investigations will be described in a subsequent publication ( chakravorty in preparation ) . in , we plot @xmath243 $ ] against @xmath244 $ ] for a range of continuum shapes ( left ) and abundances ( right ) , where @xmath243 $ ] and @xmath244 $ ] are the range of @xmath75 and @xmath4 for the stable states at @xmath245 .
presumably , the higher the values of these two quantities , the better are the chances of existence of the state at @xmath236 , because any perturbed gas is more likely to find a larger stable region .
the various parameter sets used to describe wa conditions in this paper have been represented by different symbols which have been explained in the bottom panel .
the gray square represents the _ standard set _ of parameters and the rest of the symbols indicate the changes from it . from the systematic analysis
we can see the following trends : 1 .
the variation of soft x - ray slope @xmath70 of the ionizing continuum shows that was exist only if @xmath0 and will have higher temperature gas if @xmath246 .
2 . increases in the soft x - ray slope @xmath70 decrease the ionization parameter of the @xmath89 wa while decreasing its temperature , but increases both @xmath3 and @xmath75 for the @xmath199 phase . both @xmath247 $ ] and @xmath248 $ ] increases with the value of @xmath70 , which is also demonstrated by the _ square _
data points in the top and the middle left panels of accompanied with increase in @xmath249 $ ] , indicating that wa phases stabilise with steeper soft x - ray continua .
the multiphase scenario , however , is not facilitated significantly with the variation in @xmath70 .
multiphase solutions at @xmath250 are seen only for the standard value of @xmath55 .
although the @xmath251 stability curve also shows 2 phases , the value of @xmath238 $ ] is very low , making the possibility of multiple phases in equilibrium seem unlikely .
we have found very interesting difference between our results and those obtained by krolik & kriss ( 2001 ) .
they have drawn stability curves for ionizing continuum having properties similar to our _ standard set _ sed having @xmath252 and @xmath55 .
the stability curves shown in kk01 have a near vertical jump in temperature from @xmath253 to @xmath12 as @xmath3 runs from @xmath254 to @xmath255 thus exhibiting a continuous distribution of stable states in @xmath256 space .
for the same set of parameters ( ie , the _ standard set _ ) we not only get two distinct wa phases in pressure equilibrium with each other , but also do not see any drastic change in temperature over a narrow range of @xmath202 , although our @xmath3 values also run from @xmath257 to @xmath258 .
5 . as the value of the uv - x - ray slope @xmath58
is increased , the ionization parameter for the @xmath89 wa increases steadily accompanied by a small decrease in temperature .
@xmath194 also increases but with a small increase in temperature . 6 .
the range @xmath247 $ ] remains almost constant for @xmath259 , but abruptly increases for the value of 1.8 .
@xmath248 $ ] grows abruptly for @xmath260 before which it increases steadily but slowly with the increase in @xmath58 .
there is no significant wa state at @xmath199 for @xmath261 , comptonization becomes the dominant heating and cooling processes at these temperatures for such an sed .
the triangles in the left panels of show these trends . 7 .
multiphase solutions at @xmath262 are seen in the range @xmath259 . for @xmath263
, there is an additional stable state at @xmath264 , but @xmath239 $ ] is very low to claim any possibility of this state being in pressure equilibrium with the lower temperature phases .
8 . through the increase of metallicity of the absorbing gas from @xmath265 to @xmath266 ,
the @xmath89 wa maintains almost constant temperature . for sub - solar metallicity , @xmath216 decreases steadily through the increase of abundance , but this trend is inverted for super - solar metallicity .
the @xmath199 wa , on the other hand , shows a rise in temperature @xmath267 throughout the metallicity variation , while @xmath194 remains constant within a factor of 1.4 .
there is an anti - correlation between @xmath247 $ ] and @xmath248 $ ] ; the @xmath89 wa shrinks in @xmath202 until it vanishes for @xmath266 , whereas the phase at @xmath199 progressively grows in @xmath202 from @xmath268 to @xmath269 .
this trend is also depicted in the top and the middle right panels of by the _
open stars _ and the _ closed circles_. a @xmath264 phase starts appearing from @xmath270 , and @xmath271 $ ] increases with abundance . 10 .
interesting multiphase properties are exhibited by wa with varying abundance .
$ ] is significant in the range @xmath272 for which @xmath262 gas are in pressure equilibrium with each other . beyond this range for @xmath213 ,
the stability curves show the interesting feature of significant multiphase solution between the @xmath273 phases ; a property depicted by only these three stability curves among all the parameter sets considered in this paper . 11 . in the case of the parameter sets having super - solar metallicity
, we find an interesting phenomenon depicted by the stability curves .
for @xmath274 , the wa stable phases ( @xmath241 ) are seen to be in pressure equilibrium with gas at @xmath80 , which might be representing the ` broad emission line region ( belr ) clouds ' .
such a feature in the stability curve might indicate the connection between the x - ray and uv absorber .
absorption edges due to various partially ionized elements are seen in the soft x - ray spectra of about half of the seyfert1 galaxies and observed quasars .
these absorption features are due to the presence of a warm ( @xmath275 ) gas along our line of sight to the agn .
we have shown that the existence and nature of this absorption component crucially depends on the shape of the ionizing continuum from the central source which illuminates the gas , and also on the chemical composition of the absorbing medium .
we have used the stability curve as a tool to study the stability of the wa as a function of its various physical properties .
we have considered an ionizing continuum with two power - law components which together reproduce the observed x - ray spectral index as well as the observed values of @xmath58 .
systematic investigation shows that if the x - ray spectral index , @xmath70 , of the ionizing continuum intervening gas of solar metallicity is flat with @xmath276 , then the agn can not have a wa . for moderately steep x - ray spectra with @xmath277 ,
it is possible to sustain a wa .
a multiphase nature of the wa is , however , seen only for @xmath1 .
it is an interesting coincidence that most of the observed quasars also have soft x - ray slopes similar to @xmath1 and in future studies we shall investigate whether such a coincidence leads to deeper physical implications . if the ionization continuum becomes even steeper then , instead of discrete phases , the absorbing gas exhibits a continuous distribution of @xmath4 and @xmath75 .
we have also verified that the nature of the wa is not affected by the higher energy cut - off of the ionizing continuum .
temperature estimates for warm absorbing gas were generally accepted to be independent of the density of the gas , because the ionization and thermal equilibrium is believed to be achieved by the balance between photoionization and recombination .
we show in this paper , in agreement with rozanska ( 2008 ) , that the stability properties of highly dense wa might show density dependence influenced by sufficiently soft uv dominated ionizing continuum .
this is a very interesting result because it implies that for such objects , we would not require independent estimates of density from observations ; modeling the uv spectra itself can provide constraints on gas density .
the chemical composition of the absorbing medium plays a critical role in determining its multiphase nature .
the higher the metallicity of the medium , the greater is the probability of a multiphase parameter space .
the role of individual elements or certain important groups of elements were also studied . amongst the elements , iron , which is formed from type ia supernovae no earlier than 1 gyr , plays the most important role as a heating agent at higher temperatures ( @xmath278 ) while oxygen is a significant cooling agent for @xmath279 .
the x - ray group ( c , ne , o , fe ) , understandably , has significant influences , because the constituent elements are the ones which have important atomic transitions in the energy range relevant for warm absorbers ( @xmath280 ) .
if the absorber is over abundant in iron , oxygen or in the x - ray group of elements then the absorbing medium is likely to be multiphase .
a very interesting result seems to suggest itself from the study of influence of @xmath70-elements ( ne , mg , si , s , ar , ca , ti ) , which are the first metals formed in the universe along with oxygen through the supernovae of type ii . the thermal properties of the warm absorber change drastically from the time when the absorbing gas is essentially constituted of only @xmath70-elements as metals to the time when the absorber is enriched enough to have the whole x - ray group included in its chemical composition .
we have also observed that zero metal abundance gas does not produce was in agn .
our analysis on the chemical composition of the absorbing gas shows that the warm absorber temperature range is very sensitive to the exact abundance of the absorber and there is an indication that the warm absorber might have super - solar metallicity . while fitting soft x - ray data and determining the physical properties of the warm absorber in individual objects care should be taken to include the effect of super - solar metallicities .
when we consider super - solar metallicity , the stability curve shows a @xmath281 stable state is in pressure equilibrium with the wa stable phases .
this might imply some connection between the was and the belr .
we plan to take up this issue in future work .
we have used an upgraded version of cloudy ( c07.02 ) where the most recent developments in atomic physics have been incorporated .
this gives us the confidence to carry out a quantitative analysis on the multiphase nature of the wa presented in rigorous details in .
we have been able to provide specific trends of the behaviour of the wa as a function of the prevalent physical conditions like the shape of the ionizing continuum and the chemical composition of the gas .
comparison between observations of wa properties and simulations can be made more concrete by providing such quantitative measures .
we will extend this exercise to greater details in our subsequent publications .
in this paper we have used simple broken power - laws to define the ionizing continuum for the absorbing gas .
however , the soft excess in ultraviolet is often modeled using blackbody components which peak at @xmath282 .
moreover the continuum spectra of agn also has the disk blackbody component at @xmath283 .
we intend to investigate the effects of such continua on the wa properties as the next step in this systematic analysis .
sc would sincerely like to thank ranjiv misra and raghunath srianand for useful discussions and for suggestions which have helped the development of the paper .
we would like to sincerely thank the anonymous referee whose comments have significantly improved the paper .
99 ali , b. ; blum , r. d. ; bumgardner , t. e. ; cranmer , s. r. ; ferland , g. j. ; haefner , r. i. & tiede , g. p. , 1991 ,
pasp , 103 , 1182 allende prieto , c. , lambert , d.l . , & asplund , m. , 2001 , apj , 556 , l63 allende prieto , c. , lambert , d.l . , & asplund , m. , 2002 , apj , 573 , l137 altun z. , yumak a. , badnell n. r. , colgan j. , pindzola m. s. , 2004 , a&a , 420 , 775 ashton , c. e. , page , m. j. , branduardi - raymont , g. , blustin , a. j. , mnras , 366 , 521 ( 2006 ) begelman , m.c . ,
volonteri , m. , rees , m.j .
2006 , mnras , 370 , 289 blustin , a. j. @xmath2 2003 , a&a , 403 , 481 blustin , a. j. ; page , m. j. ; fuerst , s. v. ; branduardi - raymont , g. & ashton , c. e. , 2005 , a&a , 431 , 111 chakravorty , s. , kembhavi , a.k . , elvis , m. , ferland , g. & badnell , n.r .
, 2008 , mnras , 384l , 24 colgan j. , pindzola m. s. , whiteford a. d. , badnell n. r. , 2003 , a&a , 412 , 597 colgan , j. , pindzola , m.s . &
badnell , n.r . 2004 , astron . astrophys . 417 , 1183 collinge , m. j. @xmath2 apj , 557:2 , ( 2001 ) crenshaw , d. m. , kraemer , s. b. , & george , i. m. 2003 , ara&a , 41 , 117 g dumont , a .- m . , abrassart , a. , & collin , s. 2000 , a&a , 357 , 823 elvis , m. , 2000 , apj , 545 , 63 ferland , g. j. ; korista , k. t. ; verner , d. a. ; ferguson , j. w. ; kingdon , j. b. & verner , e. m. , 1998 , pasp , 110 , 761 - 778 field , g.b . , 1965 ,
apj , 142 , 531 gehrels , n and williams , e.d .
, apj 418:l25 , 1993 george , i. m. ; turner , t. j. ; netzer , hagai ; nandra , k. ; mushotzky , r. f. & yaqoob , t. , 1998 , apjs , 114 , 73 goncalves , a. c. ; collin , s. ; dumont , a .- m . ; mouchet , m. ; rozanska , a. ; chevallier , l. & goosmann , r. w. , 2006 , a&a 451 , l23 grevesse , n. , & sauval , a.j . , 1998 , space science review , 85 , 161 - 174 grupe , d. ; mathur , s. ; wilkes , b. & osmer , p. , 2006 ,
apj , 131 , 55 gu m. f. , 2003 , apj , 590 , 1131 gu , m. f. 2004 , apjs , 153 , 389 halpern , j.p . , 1984 ,
apj , 281,90 hamann , f. , ferland , g. , 1993 , apj , 418 , 11 hamann , f. , ferland , g. , annu . rev . astron . astrophys .
37:487 hess , c.j . ; kahn , s.m . ; paerels , f.b.s . , 1997 ,
apj , 478 , 94 holweger , h. , joint soho / ace workshop `` solar and galactic composition '' . edited by robert f. wimmer - schweingruber .
publisher : american institute of physics conference proceedings vol .
598 location : bern , switzerland , march 6 - 9 , 2001 , p.23 kaastra , j. s. ; steenbrugge , k. c. ; raassen , a. j. j. ; van der meer , r. l. j. ; brinkman , a. c. ; liedahl , d. a. ; behar , e. & de rosa , a. , 2002 , a&a , 386 , 427 kallman , t. r. 2000 , xstar user s guide , version 2.1 ( greenbelt , md : nasa / gsfc ) kaspi , s. , 2001 , apj , 554 , 216 kaspi , s. @xmath2 2002 , apj , 574 , 643 kinkhabwala , a. @xmath2 2002 , apj , 575 , 732 komossa & mathur 2001 , a&a , 374 , 914 komossa & meerschweinchen 2000 , a&a , 354 , 411 krolik , j. , & kriss , g. a. apj 561:684 , ( 2001 ) krolik , j. h. , mckee , c. f. , & tarter , c. b. 1981 , apj , 249 , 422 krongold , y. ; nicastro , f. ; brickhouse , n. s. ; elvis , m. ; liedahl , d. a. & mathur , s. , 2003 , apj 597 , 832 krongold , y. ; nicastro , f. ; brickhouse , n. s. ; elvis , m. & mathur , s. , 2005a , apj 622 , 842 krongold , y. ; nicastro , f. ; elvis , m. ; brickhouse , n. s. ; mathur , s. & zezas , a. , 2005b , apj 620 , 165 krongold , y. ; nicastro , f. ; elvis , m. ; brickhouse , n. ; binette , l. ; mathur , s. ; jim@xmath284nez - bail@xmath285n , e. , 2007 , apj , 659 , 1022 mathews , w.g . ; ferland , g.j .
, 1987 , apj , 323 , 456 matt , g. , nuphs , 132 , 97 m , 2004 mitnik , d.m . &
badnell , n.r . 2004 , astron . astrophys .
425 , 1153 morales , r. , fabian , a. c. & reynolds , c. s. , mnras , 315 , 149 , ( 2000 ) murray , n. ; chiang , j. , 1995 , apj , 454l , 105 nandra , k. , & pounds , k. a. 1994 , mnras , 268 , 405 netzer , h. 1993 , apj , 411 , 594 netzer @xmath2 , 2003 , apj 599 , 933 nicastro , f. ; fiore , f. ; perola , g. c. & elvis , m. , 1999 , apj 512 , 184 ogle , p. m. ; mason , k. o. ; page , m. j. ; salvi , n. j. ; cordova , f. a. ; mchardy , i. m. & priedhorsky , w. c. , 2004 , apj , 606 , 151 osterbrock d. e. , ferland g. j. , 2006 , astrophysics of gaseous nebulae and active galactic nuclei , 2nd edn .
university science books , sausalito , ca pettini , m. , the fabulous destiny of galaxies : bridging past and present .
marseille international cosmology conference ( 5 : 2005 : marseille , france ) proceedings of the vth marseille international cosmology conference , june 20 - 24 , 2005 , marseille , france . edited by v. lebrun , a. mazure , s. arnouts and d. burgarella .
isbn 2914601190 .
paris : frontier group , 2006 , p.319 reynolds , c. s. 1997 , mnras , 286 , 513 reynolds , c.s . &
fabian , a.c . , mnras , 273:1167 ( 1995 ) rozanska , a. , czerny , b. , dumont , a .- m . , collin , s. , & siemiginowska , a. 2004 , apj , 600 , 96 rozanska , a. ; goosmann , r. ; dumont , a .- m . &
czerny , b. , a&a 452 , 13 ( 2006 ) rozanska , a. , kowalska , i. & goncalves , a.c .
, 2008 , a&a , 487 , 895 savin , d.w . 1997 , apjl , 489 , 115 shang , z. @xmath2 2005 , apj , 619,41 steenbrugge , k.c .
@xmath2 a&a 434 , 569 ( 2005 ) steffen , a. t. ; strateva , i. ; brandt , w. n. ; alexander , d. m. ; koekemoer , a. m. ; lehmer , b. d. ; schneider , d. p. ; vignali , c. , 2006 , aj , 131 , 2826 strateva , iskra v. ; brandt , w. n. ; schneider , donald p. ; vanden berk , daniel g. ; vignali , cristian , 2005 , aj , 130 , 387 tananbaum , h. @xmath2 , 1979 , apj , 234 , l9 tarter , c.b . ,
tucker , w. and salpeter , e.e .
, apj 156:943 ( 1969 ) tueller , j. ; mushotzky , r. f. ; barthelmy , s. ; cannizzo , j. k. ; gehrels , n. ; markwardt , c. b. ; skinner , g. k. & winter , l. m. , 2008 , apj , 681 , 113 turner , t. j. ; nandra , k. ; george , i. m. ; fabian , a. c. ; pounds , k. a. 1993 , apj , 419 , 127 turner , a. k. , fabian , a. c. , lee , j. c. , & vaughan , s. 2004 , mnras , 353 , 319 volonteri , m. , rees , m.j .
2005 , apj , 633 , 624 wilkes , b.j . & elvis , m. , 1987 , apj , 323 , 243 zatsarinny o. , gorczyca t. w. , korista k. t. , badnell n. r. , savin d. w. , 2003 , a&a , 412 , 587 zatsarinny o. , gorczyca t. w. , korista k. t. , badnell n. r. , savin d. w. , 2004a , a&a , 417 , 1173 zatsarinny o. , gorczyca t. w. , korista k. t. , badnell n. r. , savin d. w. , 2004b , a&a , 426 , 699 zheng , w. ; kriss , g.a . ; telfer , rc . ; grimes , j.p . & davidsen , a.f . , 1997 ,
apj , 475 , 469 | signatures of warm absorbers are seen in soft x - ray spectra of about half of all seyfert1 galaxies observed and in some quasars and blazars .
we use the thermal equilibrium curve to study the influence of the shape of the ionizing continuum , density and the chemical composition of the absorbing gas on the existence and nature of the warm absorbers .
we describe circumstances in which a stable warm absorber can exist as a multiphase medium or one with continuous variation in pressure .
in particular we find the following results : i ) the warm absorber exists only if the spectral index of the x - ray power - law ionizing continuum @xmath0 and has a multiphase nature if @xmath1 , which interestingly is the spectral index for most of the observed seyfert 1 galaxies ; ii ) thermal and ionization states of highly dense warm absorbers are sensitive to their density if the ionizing continuum is sufficiently soft , i.e. dominated by the ultraviolet iii ) absorbing gas with super - solar metallicity is more likely to have a multiphase nature ; iv ) the nature of the warm absorber is significantly influenced by the absence of iron and associated elements which are produced in the later stages of star formation history in supernovae of type ia .
# 1[eq:#1 ] # 1[fig:#1 ] # 1[sec:#1 ] # 1[subsecn:#1 ] # 1[tab:#1 ] # 1equation [ eq:#1 ] # 1figure [ fig:#1 ] # 1section [ sec:#1 ] # 1section [ subsecn:#1 ] # 1table [ tab:#1 ] # 1#1 # 110^#1 -1 cm quasars : absorption lines - galaxies : active - seyfert - ism : abundances x - rays : ism |
the study of submanifolds with constant mean curvature , i.e. , @xmath4 submanifolds , and , in particular , that of @xmath4 surfaces in @xmath0-dimensional spaces , represents a very active research topic in differential geometry for more than @xmath7 years .
there are several ways to generalize these submanifolds .
for example , keeping the @xmath4 hypothesis and adding other geometric hypotheses to the submanifold or , by contrast , in the particular case of hypersurfaces in space forms , studying the hypersurfaces which are `` highly non-@xmath4 '' .
the biconservative submanifolds seem to be an interesting generalization of @xmath4 submanifolds .
biconservative submanifolds in arbitrary manifolds ( and in particular , biconservative surfaces ) which are also @xmath4 have some remarkable properties ( see , for example @xcite ) . @xmath4
hypersurfaces in space forms are trivially biconservative , so more interesting is the study of biconservative hypersurfaces which are non-@xmath4 ; recent results in non-@xmath4 biconservative hypersurfaces were obtained in @xcite .
the biconservative submanifolds are closely related to the biharmonic submanifolds .
more precisely , let us consider the _ bienergy functional _
defined for all smooth maps between two riemannian manifolds @xmath8 and @xmath9 and given by @xmath10 where @xmath11 is the tension field of @xmath12 . a critical point of @xmath13 is called a _ biharmonic map _ and is characterized by the vanishing of the _ bitension field _ @xmath14 ( see @xcite ) .
a riemannian immersion @xmath15 or , simply , a submanifold @xmath16 of @xmath17 , is called _ biharmonic _ if @xmath12 is a biharmonic map .
now , if @xmath18 is a fixed map , then @xmath13 can be thought as a functional defined on the set of all riemannian metrics on @xmath16 .
this new functional s critical points are riemannian metrics determined by the vanishing of the _ stress - bienergy tensor _ @xmath19 .
this tensor field satisfies @xmath20 if @xmath21 for a submanifold @xmath16 in @xmath17 , then @xmath16 is called a biconservative submanifold and it is characterized by the fact that the tangent part of its bitension field vanishes .
thus we can expect that the class of biconservative submanifolds to be much larger that the class of biharmonic submanifolds .
the paper is organized as follows . after a section where we recall some notions and general results about biconservative submanifolds , we present in _
section @xmath0 _ the local , intrinsic characterization of biconservative surfaces .
the local , intrinsic characterization theorem provides the necessary and sufficient conditions for an abstract surface @xmath22 to admit , locally , a biconservative embedding with positive mean curvature function @xmath23 and @xmath24 at any point .
our main goal is to extend the _
classification results for biconservative surfaces in @xmath1 , with @xmath2 and @xmath3 , to _ global _ results , i.e. , we ask that biconservative surfaces to be _ complete _ , with @xmath25 everywhere and @xmath26 on an open dense subset . in _
section @xmath27 _ we consider the global problem and construct complete biconservative surfaces in @xmath5 with @xmath25 on @xmath16 and @xmath24 at any point of an open dense subset of @xmath16 .
we determine such surfaces in two ways .
one way is to use the local , extrinsic characterization of biconservative surfaces in @xmath5 and `` glue '' two pieces together in order to obtain a complete biconservative surface .
the other way is more analytic and consists in using the local , intrinsic characterization theorem in order to obtain a biconservative immersion from @xmath28 in @xmath5 with @xmath25 on @xmath29 and @xmath30 on an open dense subset of @xmath29 ( the immersion has to be unique ) ; here , @xmath31 is a positive constant and therefore we obtain a one - parameter family of solutions .
it is worth mentioning that , by a simple transformation of the metric @xmath32 , @xmath33 is ( intrinsically ) isometric to a helicoid . in the _ last section _ we consider the global problem of biconservative surfaces in @xmath6 with @xmath25 on @xmath16 and @xmath24 at any point of an open dense subset of @xmath16 .
as in the @xmath5 case , we use the local , extrinsic classification of biconservative surfaces in @xmath6 , but now the `` gluing '' process is not as clear as in @xmath5 .
further , we change the point of view and use the local , intrinsic characterization of biconservative surfaces in @xmath6 .
we determine the complete riemannian surfaces @xmath34 which admit a biconservative immersion in @xmath6 with @xmath25 everywhere and @xmath26 on an open dense subset of @xmath29 and we show that , up to isometries , there exists only a one - parameter family of such riemannian surfaces indexed by @xmath35 .
we end the paper with some figures , obtained for particular choices of the constants , which represent the non-@xmath4 complete biconservative surfaces in @xmath5 and the way how these surfaces can be obtained in @xmath6 .
throughout this work , all manifolds , metrics , maps are assumed to be smooth , i.e. in the @xmath36 category , and we will often indicate the various riemannian metrics by the same symbol @xmath37 .
all surfaces are assumed to be connected and oriented .
a _ harmonic map _ @xmath38 between two riemannian manifolds is a critical point of the _ energy functional _ @xmath39 and it is characterized by the vanishing of its _ tension field _ @xmath40 the idea of the stress - energy tensor associated to a functional comes from d. hilbert ( @xcite ) . given a functional @xmath41 , one can associate to it a symmetric 2-covariant tensor field @xmath42 such that @xmath43 at the critical points of @xmath41 .
when @xmath41 is the energy functional , p. baird and j. eells ( @xcite ) , and a. sanini ( @xcite ) , defined the tensor field @xmath44 and proved that @xmath45 thus , @xmath42 can be chosen as the stress - energy tensor of the energy functional .
it is worth mentioning that @xmath42 has a variational meaning .
indeed , we can fix a map @xmath46 and think @xmath41 as being defined on the set of all riemannian metrics on @xmath16 .
the critical points of this new functional are riemannian metrics determined by the vanishing of their stress - energy tensor @xmath42 .
more precisely , we assume that @xmath16 is compact and denote @xmath47 for a deformation @xmath48 of @xmath49 we consider @xmath50 .
we define the new functional @xmath51 and we have the following result .
let @xmath52 and assume that @xmath16 is compact .
then @xmath53 therefore @xmath49 is a critical point of @xmath54 if and only if its stress - energy tensor @xmath42 vanishes .
we mention here that , if @xmath55 is an arbitrary isometric immersion , then @xmath43 .
a natural generalization of harmonic maps is given by biharmonic maps .
a _ biharmonic map _ @xmath38 between two riemannian manifolds is a critical point of the _ bienergy functional _ @xmath56 and it is characterized by the vanishing of its _ bitension field _
@xmath57 where @xmath58 is the rough laplacian of @xmath59 and the curvature tensor field is @xmath60 } z , \quad
\forall x , y , z\in c(tm).\ ] ] we remark that the _ biharmonic equation _ @xmath61 is a fourth - order non - linear elliptic equation and that any harmonic map is biharmonic .
a non - harmonic biharmonic map is called proper biharmonic . in @xcite
, g. y. jiang defined the stress - energy tensor @xmath19 of the bienergy ( also called _ stress - bienergy tensor _ ) by @xmath62 as it satisfies @xmath20 the tensor field @xmath19 has a variational meaning , as in the harmonic case .
we fix a map @xmath46 and define a new functional @xmath63 then we have the following result .
let @xmath46 and assume that @xmath16 is compact .
then @xmath64 so @xmath49 is a critical point of @xmath65 is and only if @xmath66 .
we mention that , if @xmath55 is an isometric immersion then @xmath67 does not necessarily vanish . _
a submanifold of a given riemannian manifold @xmath9 _ is a pair @xmath68 , where @xmath69 is a manifold and @xmath70 is an immersion .
we always consider on @xmath16 the induced metric @xmath71 , thus @xmath72 is an isometric immersion ; for simplicity we will write @xmath70 without mentioning the metrics . also , we will write @xmath73 , or even @xmath16 , instead of @xmath74 . a submanifold @xmath75 is called _ biharmonic _ if the isometric immersion @xmath12 is a biharmonic map from @xmath8 to @xmath9 . even if the notion of biharmonicity may be more appropriate for maps than for submanifolds , as the domain and the codomain metrics are fixed and the variation is made only through the maps , the biharmonic submanifolds proved to be an interesting notion ( see , for example , @xcite ) . in order to fix the notations , we recall here only the fundamental equations of first order of a submanifold in an riemannian manifold .
these equations define the second fundamental form , the shape operator and the connection in the normal bundle .
let @xmath55 be an isometric immersion .
for each @xmath76 , @xmath77 splits as an orthogonal direct sum @xmath78 and @xmath79 is referred to as the normal bundle of @xmath12 , or of @xmath16 , in @xmath17 .
denote by @xmath80 and @xmath81 the levi - civita connections on @xmath16 and @xmath17 , respectively , and by @xmath82 the induced connection in the pull - back bundle @xmath83 .
taking into account the decomposition in , one has @xmath84 where @xmath85 is called the second fundamental form of @xmath16 in @xmath17 . here
@xmath86 denotes the cotangent bundle of @xmath16 .
the mean curvature vector field of @xmath16 in @xmath17 is defined by @xmath87 , where the @xmath88 is considered with respect to the metric @xmath49 .
furthermore , if @xmath89 , then @xmath90 where @xmath91 is called the shape operator of @xmath16 in @xmath17 in the direction of @xmath92 , and @xmath93 is the induced connection in the normal bundle .
moreover , @xmath94 , for all @xmath95 , @xmath89 . in the case of hypersurfaces
, we denote @xmath96 , where @xmath97 and @xmath92 is the unit normal vector field , and we have @xmath98 ; @xmath23 is the _
( @xmath99 times ) mean curvature function_. a submanifold @xmath16 of @xmath17 is called @xmath100 if @xmath101 is parallel in the normal bundle , and @xmath4 if @xmath102 is constant
. when confusion is unlikely we identify , locally , @xmath16 with its image through @xmath12 , @xmath103 with @xmath104 and @xmath105 with @xmath106 . with these identifications in mind ,
we write @xmath107 and @xmath108 if @xmath21 for a submanifold @xmath16 in @xmath17 , then @xmath16 is called _
biconservative_. thus , @xmath16 is biconservative if and only if the tangent part of its bitension field vanishes .
we have the following characterization theorem of biharmonic submanifolds , obtained by splitting the bitension field in the tangent and normal part .
[ biharmonic_submanifold ] a submanifold @xmath69 of a riemannian manifold @xmath109 is biharmonic if and only if @xmath110 and @xmath111 where @xmath112 is the laplacian in the normal bundle . various forms of the above result were obtained in @xcite . from here
we deduce some characterization formulas for the biconservativity .
let @xmath69 be a submanifold of a riemannian manifold @xmath109 .
then @xmath16 is a biconservative submanifold if and only if : 1 .
@xmath113 ; 2 .
@xmath114 ; 3 .
@xmath115 .
the following properties are immediate .
let @xmath69 be a submanifold of a riemannian manifold @xmath109 . if @xmath116 then @xmath16 is biconservative .
let @xmath69 be a submanifold of a riemannian manifold @xmath109 .
assume that @xmath17 is a space form , i.e. , it has constant sectional curvature , and @xmath16 is @xmath100 .
then @xmath16 is biconservative .
let @xmath69 be a submanifold of a riemannian manifold @xmath109 .
assume that @xmath16 is pseudo - umbilical , i.e. , @xmath117 , and @xmath118 .
then @xmath16 is @xmath4 .
if we consider the particular case of hypersurfaces , then theorem @xmath119 becomes if @xmath69 is a hypersurface in a riemannian manifold @xmath120 , then @xmath16 is biharmonic if and only if @xmath121 and @xmath122 where @xmath92 is the unit normal vector field of @xmath16 in @xmath17 .
a hypersurface @xmath69 in a space form @xmath123 is biconservative if and only if @xmath124 any @xmath4 hypersurface in @xmath123 is biconservative .
we are interested to study biconservative surfaces which are non-@xmath4 .
we will first look at them from a local , extrinsic point of view and then from a global point of view . while by `` local '' we will mean the biconservative surfaces @xmath125 with @xmath25 and @xmath24 at any point of @xmath16 , by `` global '' we will mean the _ complete _ biconservative surfaces @xmath125 with @xmath25 at any point of @xmath16 and @xmath24 at any point of an open and dense subset of @xmath16 . in this section , we consider the local problem , i.e.
, we take @xmath125 a biconservative surface and assume that @xmath25 and @xmath24 at any point of @xmath16 .
let @xmath126 and @xmath127 two vector fields such that @xmath128 is a positively oriented orthonormal basis at any point @xmath76 . in particular
, we obtain that @xmath16 is parallelizable .
if we denote by @xmath129 the eigenvalues functions of the shape operator @xmath130 , since @xmath131 and @xmath132 , we get @xmath133 and @xmath134 .
thus the matrix of @xmath130 with respect to the ( global ) orthonormal frame field @xmath135 is @xmath136 we denote by @xmath137 the gaussian curvature and , from the gauss equation , @xmath138 , we obtain @xmath139 thus @xmath140 on @xmath16 . from the definitions of @xmath141 and @xmath127 , we find that @xmath142 using the connection @xmath143-forms , the codazzi equation and then the extrinsic and intrinsic expression for the gaussian curvature , we obtain the next result which shows that the mean curvature function of a non - cmc biconservative surface must satisfy a second - order partial differential equation . more precisely , we have the following theorem .
let @xmath125 a biconservative surface with @xmath25 and @xmath24 at any point of @xmath16
. then we have @xmath144 where @xmath145 is the laplace - beltrami operator on @xmath16 .
in fact , we can see that around any point of @xmath16 there exists @xmath146 local coordinates such that @xmath147 and @xmath148 is equivalent to @xmath149 i.e. , @xmath23 must satisfy a _
second - order ordinary differential equation_. indeed , let @xmath150 be an arbitrary fixed point of @xmath16 and let @xmath151 be an integral curve of @xmath141 with @xmath152 .
let @xmath153 the flow of @xmath127 and @xmath146 local coordinates with @xmath154 such that @xmath155 we have @xmath156 and @xmath157 if we write the riemannian metric @xmath49 on @xmath16 in local coordinates as @xmath158 we get @xmath159 , and @xmath141 can be expressed with respect to @xmath160 and @xmath161 as @xmath162 where @xmath163 , @xmath164 .
let @xmath165 .
since @xmath166 , we find that @xmath167 it can be proved that @xmath168=\frac{3\left(x_1 f\right)}{4f}x_2,\ ] ] and thus @xmath169f=0 $ ] .
on the other hand we have @xmath170 we recall that @xmath171 at any point of @xmath172 , and then @xmath173 at any point of @xmath172 . therefore , from , @xmath174 , i.e. , @xmath175 . since @xmath176 , and @xmath177 ,
we have @xmath178 , i.e. , @xmath179 in @xcite it was found an equivalent expression for @xmath148 , i.e. , @xmath180 therefore , using , relation is equivalent to .
if @xmath125 is a non-@xmath4 biharmonic surface , then , there exists an open subset @xmath172 such that @xmath25 , @xmath24 at any point of @xmath172 , and @xmath23 satisfies the following system @xmath181 as we have seen , this system implies @xmath182 which , in fact , is a ode system .
we get @xmath183 as an immediate consequence we obtain @xmath184 and combining it with the first integral @xmath185 of the first equation from , where @xmath186 is a constant , we obtain @xmath187 if we denote @xmath188 , we get @xmath189 .
thus , @xmath190 satisfies a polynomial equation with constant coefficients , so @xmath190 has to be a constant and then , @xmath23 is a constant , i.e. , @xmath191 on @xmath172 ( in fact , @xmath23 has to be zero ) .
therefore , we have a contradiction ( see @xcite for @xmath2 and @xcite , for @xmath192 ) we can also note that relation @xmath148 , which is an extrinsic relation , together with , allows us to find an _ intrinsic relation _ that @xmath193 must satisfy .
more precisely , the gaussian curvature of @xmath16 has to satisfy @xmath194 and the hypotheses @xmath25 and @xmath24 , at any point , are equivalent with the intrinsic conditions @xmath140 and @xmath195 .
formula is very similar to the ricci condition .
further , we will briefly recall the ricci problem . given an abstract surface @xmath22 , we want to find the conditions that have to be satisfied by @xmath16 such that , locally , it admits a minimal embedding in @xmath1 .
it was proved ( see @xcite ) that if @xmath22 is an abstract surface such that @xmath140 at any point of @xmath16 , where @xmath196 is a constant , then , locally , it admits a minimal embedding in @xmath1 if and only if @xmath197 condition is called the _
ricci condition with respect to @xmath198 _ , or simply the _ ricci condition_. if @xmath199 holds , then , locally , @xmath16 admits a one - parameter family of minimal embeddings in @xmath1 .
we can see that relations @xmath200 and @xmath199 are very similar and , in @xcite , the authors studied the link between them .
thus , for @xmath2 , it was proved that if we consider a surface @xmath22 which satisfies @xmath201 and @xmath202 , then there exists a very simple conformal transformation of the metric @xmath49 such that @xmath203 satisfies @xmath199 .
a similar result was also proved for @xmath204 , but in this case , the conformal factor has a complicated expression ( and it is not enough to impose that @xmath22 satisfy , but we need the stronger hypothesis of it to admit a biconservative immersion in @xmath1 ) .
unfortunately , condition @xmath201 does not imply , locally , the existence of a biconservative immersion in @xmath1 , as in the minimal case .
we need a stronger condition .
it was obtained the following local , intrinsic characterization theorem .
[ thm : char ] let @xmath22 be an abstract surface and @xmath205 a constant .
then , locally , @xmath16 can be isometrically embedded in a space form @xmath1 as a biconservative surface with positive mean curvature having the gradient different from zero at any point if and only if the gaussian curvature @xmath137 satisfies @xmath206 , @xmath207 , for any point @xmath76 , and its level curves are circles in @xmath16 with curvature @xmath208 if the surface @xmath16 in theorem [ thm : char ] is simply connected , then the theorem holds globally , but , in this case , instead of a local isometric embedding we have a global isometric immersion .
we remark that unlike in the minimal immersions case , if @xmath16 satisfies the hypotheses from theorem [ thm : char ] , then there exists a _
unique _ biconservative immersion in @xmath1 ( up to an isometry of @xmath1 ) , and not a one - parameter family .
the characterization theorem can be equivalently rewritten as below .
[ thm : char - local ] let @xmath22 be an abstract surface with gaussian curvature @xmath137 satisfying @xmath206 and @xmath207 at any point @xmath76 , where @xmath196 is a constant .
let @xmath209 and @xmath210 be two vector fields on @xmath16 such that @xmath128 is a positively oriented basis at any point of @xmath76 .
then , the following conditions are equivalent : * the level curves of @xmath137 are circles in @xmath16 with constant curvature @xmath211 * @xmath212 * locally , the metric @xmath49 can be written as @xmath213 , where @xmath214 are local coordinates positively oriented , @xmath215 , and @xmath216 ; * locally , the metric @xmath49 can be written as @xmath217 , where @xmath214 are local coordinates positively oriented , and @xmath218 satisfies the equation @xmath219 and the condition @xmath220 ; moreover , the solutions of the above equation , @xmath221 , are @xmath222 where @xmath12 is in some open interval @xmath223 and @xmath224 are constants ; * locally , the metric @xmath49 can be written as @xmath217 , where @xmath214 are local coordinates positively oriented , and @xmath218 satisfies the equation @xmath225 and the conditions @xmath220 and @xmath226 ; moreover , the solutions of the above equation , @xmath221 , are @xmath227 where @xmath12 is in some open interval @xmath223 and @xmath228 are constants , @xmath229 .
the proof follows by direct computations and by using remark 4.3 in @xcite and proposition 3.4 in @xcite . from the above theorem we have the following remarks . * if condition @xmath230 is satisfied , i.e. , the integral curves of @xmath127 are circles in @xmath16 with a precise constant curvature , then the integral curves of @xmath141 are geodesics of @xmath16 . * if condition @xmath231 is satisfied , then @xmath137 has to be a solution of the equation @xmath232 * if condition @xmath231 is satisfied and @xmath233 , then @xmath234 is a flat surface and , trivially , a ricci surface with respect to @xmath198 .
* let @xmath218 be a solution of the equation .
we consider the change of coordinates @xmath235 where @xmath236 is a positive constant and @xmath237 , and define @xmath238 then @xmath239 and @xmath153 also satisfies equation .
if @xmath218 satisfies the first integral @xmath240 where @xmath229 , then , for @xmath241 , @xmath242 satisfies @xmath243 from here , as the classification is done up to isometries , we note that the parameter @xmath244 in the solution of is not essential and only the parameter @xmath245 counts .
thus we have a one - parameter family of solutions . *
if @xmath12 is a solution of , for some @xmath198 , then @xmath246 , where @xmath247 is a real constant , is a solution of for @xmath248 . *
if @xmath2 , we note that if @xmath12 is a solution of , then also @xmath249 is a solution of the same equation , i.e , the condition @xmath230 from theorem @xmath250 is invariant under the homothetic tranformations of the metric @xmath49 .
then , we see that equation is invariant under the affine change of parameter @xmath251 , where @xmath252 .
therefore , we must solve equation up to this change of parameter and an additive constant of the solution @xmath12
. the additive constant will be the parameter that counts . in the @xmath2 case ,
the solutions of equation , are explicitly determined in the next proposition .
[ prop - c=0 ] the solutions of the equation @xmath253 which satisfy the conditions @xmath220 and @xmath254 , up to affine transformations of the parameter with @xmath252 , are given by @xmath255 we note that , when @xmath2 , we have a one - parameter family of solutions of equation , i.e. , @xmath256 , @xmath31 being a positive constant .
if @xmath257 , then we can not determine explicitly @xmath218 .
another way to see that in the @xmath204 case we have only a one - parameter family of solutions of equation is to rewrite the metric @xmath49 in certain non - isothermal coordinates .
further , we will consider only the @xmath3 case and we have the next result .
let @xmath22 be an abstract surface with @xmath258 , where @xmath221 satisfies @xmath259 where @xmath12 is in some open interval @xmath223 , @xmath260 are positive constants , and @xmath261 is a constant .
then @xmath22 is isometric to @xmath262 where @xmath263 , @xmath264 is a positive constant , and @xmath265 and @xmath266 are the positive vanishing points of @xmath267 , with @xmath268 .
let us consider @xmath269 and @xmath270 the surfaces @xmath271 and @xmath272 are isometric if and only if @xmath273 and the isometry is @xmath274 .
therefore , we have a one - parameter family of surfaces .
we note that the expression of the gaussian curvature of @xmath271 does not depend on @xmath35 .
more precisely , @xmath275 but , if we change further the coordinates @xmath276 , then we `` fix '' the domain , i.e. , @xmath271 is isometric to @xmath277 and @xmath35 appears in the expression of @xmath278 .
in this section we consider the global problem and construct complete biconservative surfaces in @xmath5 with @xmath25 everywhere and @xmath24 at any point of an open dense subset . or , from intrinsic point of view , we construct a complete abstract surface @xmath22 with @xmath202 everywhere and @xmath279 at any point of an open dense subset of @xmath16 , that admits a biconservative immersion in @xmath5 , defined on the whole @xmath16 , with @xmath25 on @xmath16 and @xmath26 on the open dense subset .
first , we recall a local extrinsic result which provides a characterization of biconservative surfaces in @xmath5 .
let @xmath280 be a surface in @xmath5 with @xmath281 and @xmath282 for any @xmath76 .
then , @xmath16 is biconservative if and only if , locally , it is a surface of revolution , and the curvature @xmath283 of the profile curve @xmath175 , @xmath284 , is a positive solution of the following ode @xmath285 in @xcite there was found the local explicit parametric equation of a biconservative surface in @xmath5 .
let @xmath280 be a biconservative surface in @xmath5 with @xmath281 and @xmath282 for any @xmath76 . then
, locally , the surface can be parametrized by @xmath286 where @xmath287 with @xmath288 a positive constant and @xmath289 .
we note that any two such surfaces are not locally isometric , so we have a one - parameter family of biconservative surfaces in @xmath5 . these surfaces are not complete .
we denote by @xmath290 the image of @xmath291 .
the `` boundary '' of @xmath290 , i.e. , @xmath292 , is the circle @xmath293 , which lies in the @xmath294 plane . at a boundary point ,
the tangent plane to the closure @xmath295 of @xmath290 is parallel to @xmath296 .
moreover , along the boundary , the mean curvature function is constant @xmath297 and @xmath298 .
thus , in order to obtain a complete biconservative surface in @xmath5 , we can expect to `` glue '' along the boundary two biconservative surfaces of type @xmath290 corresponding to the same @xmath288 ( the two constants have to be the same ) and symmetric to each other , at the level of @xmath36 smoothness . in fact , it was proved that we can glue two biconservative surfaces @xmath290 and @xmath299 , at the level of @xmath36 smoothness , only along the boundary and , in this case , @xmath300 . if we consider the symmetry of @xmath301 , with respect to the @xmath302 axis , we get a smooth , complete , biconservative surface @xmath303 in @xmath5 .
moreover , its mean curvature function @xmath304 is positive and @xmath305 is different from zero at any point of an open dense subset of @xmath303 .
the profile curve @xmath306 can be repara - metrized as @xmath307 and now @xmath308 .
the homothety of @xmath5 , @xmath309 , renders @xmath310 onto @xmath311 . in @xcite ,
there were also found the complete biconservative surfaces in @xmath5 with @xmath25 at any point and @xmath24 at any point of an open dense subset , but there , the idea was to use the intrinsic characterization of the biconservative surfaces .
more precisely , we have the next global result . [ main_th1 ]
let @xmath312 be a surface , where @xmath313 is a positive constant .
then we have : * the metric on @xmath29 is complete ; * the gaussian curvature is given by @xmath314 and therefore @xmath315 at any point of @xmath316 ; * the immersion @xmath317 given by @xmath318 is biconservative in @xmath5 , where @xmath319 the first two items follow by standard arguments . for the last part
, we note that choosing @xmath320 in and using the change of coordinates @xmath321 , where @xmath322 , the metric induced by @xmath323 coincides with @xmath32 .
then , we define @xmath324 as : for @xmath322 , @xmath325 is obtained by rotating the profile curve @xmath326 and for @xmath327 , @xmath325 is obtained by rotating the profile curve @xmath328 we consider the surface @xmath28 .
then @xmath33 satisfies the ricci condition and can be minimally immersed in @xmath5 as a helicoid or a catenoid . concerning the biharmonic surfaces in @xmath5 we have the following non - existence result .
there exists no proper biharmonic surface in @xmath5 .
as in the previous section , we consider the global problem for biconservative surfaces in @xmath6 , i.e. , our aim is to construct complete biconservative surfaces in @xmath6 with @xmath25 everywhere and @xmath24 at any point of an open and dense subset .
let @xmath280 be a biconservative surface in @xmath6 with @xmath281 i @xmath329 at any point @xmath76 .
then , locally , the surface , viewed in @xmath330 , can be parametrized by @xmath331 where @xmath332 is a positive constant ; @xmath333 are two constant orthonormal vectors ; @xmath334 is a curve parametrized by arclength that satisfies @xmath335 and , as a curve in @xmath336 , its curvature @xmath283 is a positive non constant solution of the following ode @xmath337 such that @xmath338 we consider @xmath345 and @xmath346 and change the coordinates @xmath214 in @xmath347 .
then , we get @xmath348 where @xmath349 , @xmath350 and @xmath351 are positive solutions of @xmath352 and @xmath353 with @xmath354 a constant and @xmath355 .
we note that an alternative expression for @xmath356 was given in @xcite .
if we denote @xmath360 the image of @xmath356 , then we note that the boundary of @xmath360 is made up from two circles and along the boundary , the mean curvature function is constant ( two different constants ) and its gradient vanishes .
more precisely , the boundary of @xmath360 is given by the curves @xmath361 and @xmath362 these curves are circles in affine planes in @xmath330 parallel to the @xmath363 plane and their radii are @xmath364 and @xmath365 , respectively .
at a boundary point , using the coordinates @xmath366 , we get that the tangent plane to the closure of @xmath360 is spanned by a vector which is tangent to the corresponding circle and by @xmath367 where @xmath368 or @xmath369 .
thus , in order to construct a complete biconservative surface in @xmath6 , we can expect to glue along the boundary two biconservative surfaces of type @xmath360 , corresponding to the same @xmath339 .
in fact , if we want to glue two surfaces corresponding to @xmath339 and @xmath370 along the boundary , then these constants have to coincide and there is no ambiguity concerning along which circle of the boundary we should glue the two pieces . but this process is not as clear as in @xmath5 since we should repeat it infinitely many times .
the first item follows by standard arguments . for the second item
, we note that choosing @xmath380 in and using the change of coordinates @xmath381 , the metric induced by @xmath382 coincides with @xmath383 .
the immersion @xmath384 depends on the sign @xmath389 and on the constant @xmath390 in the expression of @xmath391 . as the classification is up to isometries of @xmath6 , the sign and the constant
are not important , but they will play an important role in the gluing process . the key idea in our construction of complete biconservative surface in @xmath6 is to notice that @xmath271 is , locally and intrinsically , isometric to a surface of revolution in @xmath5 . the first step is to construct a complete surface of revolution in @xmath5 which on an open dense subset is locally isometric to @xmath271 .
we start with the next result .
[ theorem3.18 ] let us consider @xmath271 as above .
then @xmath271 is the universal cover of the surface of revolution in @xmath5 given by @xmath392 where @xmath393 , @xmath394 @xmath395 is a positive constant and @xmath396 is constant .
we note that the boundary of @xmath400 is given by the curves @xmath403 and @xmath404 these curves are circles in affine planes in @xmath5 parallel to the @xmath405 plane and their radii are @xmath406 and @xmath407 , respectively . at a boundary point , using the coordinates @xmath408 , we get that the tangent plane to the closure of @xmath400 is spanned by a vector which is tangent to the corresponding circle and by the vector @xmath409 .
thus , the tangent plane is parallel to @xmath296 axis .
geometrically , we start with a piece of type @xmath400 and by symmetry to the planes where the boundary lie , we get our complete surface @xmath410 ; the process is periodic and we perform it along the whole @xmath296 axis .
analytically , we fix @xmath35 and @xmath411 , and alternating the sign and with appropriate choices of the constant @xmath398 , we can construct a complete surface of revolution @xmath410 in @xmath5 which on an open subset is locally isometric to @xmath412 . in fact , these choices of @xmath413 and @xmath414 , and of the constants @xmath398 are uniquely determined by the `` first '' choice of @xmath413 , or of @xmath414 , and of the constant @xmath398 .
we start with @xmath413 and @xmath415 .
the profile curve of @xmath400 can be seen as the graph of a function depending on @xmath399 and this allows us to obtain a function @xmath416 such that the profile curve of @xmath410 to be the graph of the function @xmath417 depending on @xmath399 and defined on the whole @xmath296 ( or @xmath418 ) .
the function @xmath419 $ ] is periodic and at least of class @xmath420 .
[ main_th3 ] the surface of revolution given by @xmath421 is complete and , on an open dense subset , it is locally isometric to @xmath371 .
the induced metric is given by @xmath422 @xmath423 .
moreover , @xmath279 at any point of that open dense subset , and @xmath424 everywhere .
the universal cover of the surface of revolution given by @xmath426 is @xmath29 endowed with the metric @xmath427 .
it is complete , @xmath424 on @xmath29 and , on an open dense subset , it is locally isometric to @xmath271 and @xmath279 at any point .
moreover any two surfaces @xmath428 and @xmath429 are isometric .
the second step is to construct effectively the biconservative immersion from @xmath428 in @xmath6 , or from @xmath430 in @xmath6 .
the geometric ideea of the construction is the following : from each piece @xmath400 of @xmath410 we `` go back '' to @xmath271 and then , using @xmath384 and a specific choice of @xmath413 or @xmath414 and of the constant @xmath390 , we get our biconservative immersion @xmath431 .
again , the choices of @xmath413 and @xmath414 , and of the constant @xmath390 are uniquely determined ( modulo @xmath432 , for @xmath390 ) by the `` first '' choice of @xmath413 , or of @xmath414 , and of the constant @xmath390 ( see @xcite for all details ) . in the case
@xmath2 , the idea was to construct , by symmetry , a complete biconservative surface in @xmath5 starting with a piece of a biconservative surface .
we illustrate this in the following figure obtained for @xmath439 . in the case
@xmath3 , the construction of a complete biconservative surface in @xmath6 can be summarized in the next diagram , obtained for @xmath440 , @xmath415 and we started with @xmath413 in the expression of @xmath399 .
( manifold ) at ( 3.6,0 ) ; ( 4.3,-2 ) node[below ] @xmath22 ; ( band ) at ( 12,0 ) ; ( 10,0 ) ( 14,0 ) ; ( 10.5,-2.5 ) ( 10.5,2.5 ) ; ( 11.2,0 ) node[below ] @xmath265 ; ( 12.9,0 ) node[below ] @xmath266 ; ( 14.1,0 ) node[below ] @xmath441 ; ( 10.5,2.6 ) node[left ] @xmath442 ; ( 12,-2.5 ) node[below ] @xmath271 ; ( manifold) node[midway , fill = none , above ] isometry(9.5,0 ) ; ( 11,-3.3)-| ( 11,-4 ) ; ( 11,-4)-| ( 13,-4 ) ; ( 13,-4 ) ( 13,-8 ) node[midway , sloped , fill = none , below ] @xmath443 node[midway , sloped , fill = none , above ] biconservative ; at ( 13,-9.3)@xmath6 ; ( 10.7,-3.3)-| ( 10.7,-4 ) ; ( 10.7,-4 ) (3.5,-4 ) node[midway , sloped , fill = none , below ] @xmath444 node[midway , sloped , fill = none , above ] isometry ; at ( 2,-4 ) ; ( 2,-5.5 ) node[below ] @xmath445 ; ( graph6 ) at ( 5.3,-9.3 ) ; ( 5.5,-10 ) node[below ] @xmath446 ; ( 2,-6.2)(5,-8 ) node[pos=0.5,sloped , fill = none , above ] playing with the node[pos=0.5,sloped , fill = none , below ] constant @xmath398 and @xmath389 ; | we survey some recent results on biconservative surfaces in @xmath0-dimensional space forms @xmath1 with a special emphasis on the @xmath2 and @xmath3 cases .
we study the local and global properties of such surfaces , from extrinsic and intrinsic point of view .
we obtain all non-@xmath4 complete biconservative surfaces in @xmath5 and @xmath6 . |
at this point we consider the union of all small fibers . for the remainder of the paper , we consider an arbitrarily chosen but fixed @xmath22 .
we call a nonempty fiber * small * if its diameter is less than @xmath21 .
let @xmath17 be a continuous map where @xmath18 .
given three points @xmath50 such that the distance between each pair is at least @xmath21 , no more than two belong to small fibers of @xmath2 .
[ thelemma ] assume that @xmath50 all belong to small fibers .
if the distance between each pair is at least @xmath21 , then @xmath25 , @xmath26 , and @xmath27 must be distinct ; without loss of generality , suppose @xmath28 .
since the complement in @xmath8 of the open ball of radius @xmath21 centered at @xmath31 is path - connected , being homeomorphic to the product @xmath51 , a curve can be drawn from @xmath33 to @xmath34 such that the distance from @xmath31 to every point along the curve is at least @xmath21 . however , by the intermediate value theorem , for some point @xmath15 on the curve , @xmath52 .
the fiber containing @xmath31 therefore has diameter at least @xmath21 , and so can not be small .
we now show that the small fiber lemma places restrictions on the union of small fibers for real - valued functions .
let @xmath0 be a continuous map where @xmath1 .
when @xmath3 , the union of small fibers is bounded ; when @xmath4 the union of small fibers can be unbounded .
[ second theorem ] we begin with the case where @xmath3 .
recall that a fiber is small if its diameter is less than @xmath21 .
if the union of all small fibers is contained in an open ball of radius @xmath21 , then of course the union of small fibers is bounded .
if the union of all small fibers is not contained in an open ball of radius @xmath21 , then there must exist points @xmath53 such that both belong to small fibers and such that @xmath54 .
it then follows from the small fiber lemma that for any point @xmath15 belonging to any small fiber , either @xmath55 or else @xmath56 .
the set of all such points is of course bounded .
when @xmath4 , a simple example shows that the union of small fibers of @xmath2 can be unbounded .
consider the continuous map @xmath0 given by : @xmath57 note that fibers of @xmath2 are ( @xmath58)-spheres with centers along the @xmath59 axis .
the union of all small fibers is the set of points whose distance to the @xmath59 axis is less than @xmath60 .
although the union of small fibers is bounded when @xmath3 , two small fibers can be located arbitrarily far apart .
consider for example the urysohn - like function @xmath17 , for @xmath18 , given by @xmath61 for distinct points @xmath33 and @xmath31 ; an example in the case @xmath62 is illustrated in figure [ threesets ] . with one exception ,
all fibers of @xmath2 are ( @xmath63)-spheres whose centers lie on the line that passes through @xmath33 and @xmath31 , but not on the segment @xmath64 .
an additional fiber is the perpendicular bisector of @xmath64 , a hyperplane that becomes a sphere when the point at infinity is adjoined .
the spherical fibers are smallest when their centers are closest to @xmath64 , and grow as their centers move away from it .
the distance between the sets of small fibers can be made arbitrarily large by moving @xmath33 and @xmath31 arbitrarily far apart . given by ( [ urysohn - equation ] ) .
all small fibers are contained in two regions centered near @xmath33 and @xmath31 ; these regions can be located arbitrarily far apart . ]
aside from illustrating that small fibers can be located arbitrarily far apart , figure [ threesets ] also highlights a general property of continuous real - valued functions with small fibers . in particular
, we can use an approach similar to that used in proving the small fiber lemma to show the following .
let @xmath17 be a continuous function where @xmath18 .
if @xmath2 has a small fiber , then @xmath2 is bounded from above or from below .
further , if @xmath65 both belong to small fibers and @xmath54 , then @xmath2 is bounded from above and below .
[ third theorem ] if @xmath2 has a small fiber @xmath66 , then this fiber is contained in some closed ball @xmath67 of radius @xmath21 ; since @xmath67 is closed , @xmath2 is bounded on it . if @xmath2 is unbounded from above and from below , then there exist points @xmath68 such that @xmath69 . as in the proof of the small fiber lemma ,
since the complement of @xmath67 in @xmath8 is path - connected , a curve can be drawn in it from @xmath33 to @xmath31 . by the intermediate value theorem
there exists a point @xmath15 on that curve such that @xmath70 , and so the fiber over @xmath16 is not small , contradicting our assumption .
the second part of the theorem can be proven in a similar manner .
if @xmath65 both belong to small fibers and @xmath54 , then @xmath25 and @xmath26 must be distinct ; without loss of generality , suppose @xmath71 .
consider the closed ball @xmath67 of radius @xmath21 centered at @xmath31 .
since @xmath67 is closed , @xmath2 is bounded on it . if @xmath2 is unbounded from above
, then there exists a point @xmath72 such that @xmath28 . as above
, a curve can be drawn from @xmath33 to @xmath34 such that the distance from @xmath31 to every point along the curve is at least @xmath21 . by the intermediate value theorem
, there exists a point @xmath15 along that curve such that @xmath73 . since @xmath74 , @xmath31 can not belong to a small fiber , contradicting our assumption and showing that @xmath2 must be bounded from above .
an analogous argument shows that @xmath2 is bounded from below .
note that discontinuous functions , even those that are bounded , integrable , and decay to zero , need not have fibers of arbitrarily large diameter .
consider , for example , the following function @xmath75 : @xmath76 here @xmath2 takes a unique rational value on each unit square in the plane , and so every nonempty fiber of @xmath2 has diameter @xmath77 .
note that @xmath2 tends to 0 as @xmath78 tends to @xmath79 .
moreover , @xmath2 is in @xmath80 and @xmath81 , and hence in @xmath82 for all @xmath83 .
this example can be generalized for arbitrary @xmath1 , with a suitable choice of prime numbers .
the analysis above provides a cautionary tale for data science analysts .
the use of continuous maps to reduce the dimension of point sets in high - dimensional euclidean spaces entails what we might call the `` curse of continuity '' there will exist points arbitrarily far apart that are identified under such maps .
not only will knowledge of @xmath84 not allow us to recover @xmath15 exactly , but we will generally be unable to determine @xmath15 to within any finite error . under suitable restriction of the domain this issue
might be avoided , but knowledge of such restrictions is not always available a priori .
in contrast , discontinuous mappings suffer no such inherent limitations .
equation ( [ lp ] ) , for example , can be scaled such that its fibers are @xmath36-dimensional cubes of edge length @xmath85 .
the diameter of each fiber is then precisely @xmath86 .
knowledge of @xmath84 then allows us to determine @xmath15 to within a maximal error @xmath86 .
an application highlighting some limitations of continuous maps in analyzing structure in large point sets can be found in @xcite . in computational materials
science research , continuous `` order parameter '' mappings are often constructed to summarize structural information near each particle in a system of particles .
this order parameter is subsequently used to identify larger - scale structural features of the system .
the continuity of the order - parameter entails that points arbitrarily far apart in a relevant configuration space will map to the same order - parameter value .
consequently , continuous order - parameters regularly fail to distinguish structurally distinct configurations of points , making automated analysis difficult or impossible . in that paper , the authors suggest a discrete order - parameter , based on voronoi cell topology , which largely avoids this degeneracy .
we note that theorem [ first theorem ] can be obtained as a consequence of corollary 0.3 in @xcite , though the proof here is simpler .
also , a similar result for proper mappings can be found as a consequence of an exercise presented at the end of section 3.3 in @xcite .
finally , theorem [ first theorem ] can also be obtained as a simple corollary of what larry guth has called the `` large fiber lemma '' ( * ? ? ?
* section 7.6 ) , ( * ? ? ?
* section 6 ) , itself a corollary of the lebesgue covering lemma which is used in topological dimension theory .
if @xmath87^n \rightarrow \mathbb{r}^m$ ] is a continuous map where @xmath1 , then one of the fibers of @xmath2 has diameter at least 1 .
[ largefiberlemma ] the technique used in the proof of theorem [ first theorem ] can also be used to strengthen this lemma in two ways .
first , we can replace the unit @xmath36-cube with the inscribed @xmath88-sphere of unit diameter .
second , the inequality of the conclusion becomes an equality .
this paper is concerned with fibers in @xmath8 equipped with the standard euclidean metric .
it is easily shown that theorem [ first theorem ] can fail when @xmath8 is equipped with other metrics . for example , we can map @xmath89 homeomorphically to the cylinder capped by a half - sphere on one end and unbounded on the other .
if the origin is mapped to the center of the spherical cap , then fibers of the distance function to the origin are circles whose diameters are all bounded by that of the cylinder .
thus , one might wonder about necessary and sufficient conditions on the metric for theorem [ first theorem ] to hold . in this paper
, we proved that the union of small fibers can have unbounded diameter when @xmath4 .
however , one might ask if there must exist a line through the origin in @xmath8 onto which the projection of this set is bounded .
if so , is there a lower bound on the number of such orthogonal lines , depending on @xmath36 and @xmath42 ?
the authors would like to thank the anonymous referee for numerous helpful remarks on an earlier draft . | we present a surprisingly short proof that for any continuous map @xmath0 , if @xmath1 , then there exists no bound on the diameter of fibers of @xmath2 .
moreover , we show that when @xmath3 , the union of small fibers of @xmath2 is bounded ; when @xmath4 , the union of small fibers need not be bounded .
applications to data analysis are considered .
high - dimensional data sets are often difficult to analyze directly and , consequently , methods of simplifying them are important to modern data - intensive sciences .
continuous mappings @xmath5 are frequently used to reduce the dimension of large data sets .
indeed , a classic result of johnson and lindenstrauss @xcite shows that for @xmath6 points in any euclidean space , there exists an injective lipschitz function which maps these points into @xmath7 with minimal distortion in pairwise distances .
however , while continuous maps enjoy many desirable properties , the following suggests that a measure of caution should be exercised before employing them for high - dimensional data analysis .
we present a simple proof that for any continuous map @xmath0 , if @xmath1 then there exists no bound on the diameter of fibers of @xmath2 .
therefore , points can be arbitrarily far apart in @xmath8 , yet map to the same point under @xmath2 .
the * fibers * of a map @xmath9 are the preimages @xmath10 of points in @xmath11 .
the * diameter * of a set @xmath12 is the supremum @xmath13 , where @xmath14 denotes the euclidean distance between @xmath15 and @xmath16 .
we begin by considering real - valued functions .
let @xmath17 be a continuous function where @xmath18 .
then for any @xmath19 , there exists @xmath20 whose fiber has diameter greater than @xmath21 .
[ m=1 ] assume that some @xmath22 bounds all fiber diameters .
consider three points @xmath23 such that the distance between any two is @xmath24 , as in figure [ drawing ] .
as @xmath21 bounds the fiber diameters , @xmath25 , @xmath26 , and @xmath27 must be distinct ; without loss of generality , suppose @xmath28 . by the intermediate value theorem ,
the line segment @xmath29 contains a point @xmath15 such that @xmath30 .
as the distance from @xmath31 to any point on @xmath29 is greater than @xmath21 , the fiber containing @xmath31 must have diameter greater than @xmath21 , contradicting our assumption that @xmath21 bounds all fiber diameters .
such that the distance between each pair is @xmath24 .
if @xmath32 , then by the intermediate value theorem there exists a point @xmath15 on the segment connecting @xmath33 to @xmath34 such that @xmath35 , and so @xmath15 and @xmath31 belong to the same fiber .
since the distance from @xmath15 to @xmath31 is greater than @xmath21 , @xmath21 can not bound all fiber diameters . ]
the intermediate value theorem plays a central role in the proof above , and will turn up again several times in what follows .
a generalization of this proposition can be established using the borsuk - ulam theorem @xcite , a result about continuous mappings from an @xmath36-sphere @xmath37 to @xmath8 .
for every continuous map @xmath38 there exist antipodal points @xmath39 such that @xmath40 .
the borsuk - ulam theorem holds for @xmath36-dimensional spheres of any radius centered at the origin , and also for spheres centered at any point in a euclidean space of dimension greater than @xmath36 . in the latter case , the antipodal map is the symmetry about the center of the sphere .
we can now show the following .
let @xmath0 be a continuous map where @xmath1 .
then for any @xmath19 , there exists @xmath41 whose fiber has diameter greater than @xmath21 . [ first theorem ]
consider an @xmath42-sphere @xmath43 with radius @xmath21 and centered at the origin . by borsuk - ulam , there are antipodal points @xmath44 such that @xmath40 .
the points @xmath15 and @xmath45 lie in the same fiber of @xmath2 and @xmath46 , so this fiber has diameter at least @xmath24
. an analogous result will hold for any domain @xmath47 in which we can isometrically embed @xmath42-spheres of arbitrarily large diameter and any co - domain @xmath48 , where @xmath49 and @xmath11 are endowed with the subspace topology . |
the significant role played by processes involving neutrinos in core - collapse supernovae ( type ii supernovae ) is well known @xcite . until the core reaches densities of @xmath2 , a substantial amount of the gravitational energy of the collapse
is radiated by neutrinos that leave the star freely . however , at higher densities neutrino interactions with matter become important on the time - scale of the collapse , leading to neutrino trapping and thermalization .
supernova core - collapse simulations require a detailed description of neutrino transport and should in principle include all potentially important neutrino reactions .
it was first pointed by haxton @xcite that the neutral - current inelastic neutrino scattering on nuclei involving the excitation of giant resonances can lead to significant neutrino cross - sections and , therefore , this process should be incorporated into core - collapse simulations . shortly thereafter ,
this was done by bruenn and haxton @xcite .
they found that the inelastic neutrino - scattering on nuclei plays the same important role as the neutrino - electron scattering in equilibrating neutrinos with matter ( see also ref .
@xcite ) .
in their study bruenn and haxton approximated the nuclear composition of the core by a single representative nucleus , @xmath0fe .
moreover , the relevant cross - sections were calculated by assuming that only allowed gamow - teller and first - forbidden upward transitions from the nuclear ground - state contribute to neutrino scattering . however , supernova matter has a temperature of an order of 1 mev or higher and the neutrinos scatter off nuclei which are in thermally populated excited states . as was first realized in ref .
@xcite , upward and downward transitions from nuclear excited states to lower - lying states completely remove the energy threshold for the inelastic neutrino - nucleus scattering in the supernova environment and contribute to a significant enhancement of the cross - section for low - energy neutrinos .
moreover , and this is more important , due to downward transitions from nuclear excited states to lower - lying states neutrinos can gain energy after interacting with the nucleus , thereby assisting in cooling the core and reducing its entropy .
this is different to inelastic scattering with electrons where due to the degeneracy of electrons neutrinos mainly loose energy .
an explicit calculation of reaction rates and cross - sections at finite temperature can be performed by summing over boltzmann - weighted , individually determined contributions from nuclear excited states . however , for @xmath3 a state - by - state evaluation includes too many states to derive the cross section for each individual state and , hence , is unfeasible . to overcome this difficulty an approximate method to treat thermal effects on the inelastic neutrino - nucleus scattering was proposed in @xcite ( see also ref .
@xcite ) within the so - called hybrid approach @xcite . in this method
the contributions of the allowed gamow - teller transitions to the neutrino - nucleus cross - section are derived from large - scale shell - model ( sm ) calculations , while the forbidden contributions are considered within the rpa . to treat thermal effects within the hybrid approach , the gamow - teller contribution to the cross - section
is split into the neutrino down - scattering ( @xmath4 ) and neutrino up - scattering ( @xmath5 ) parts , where @xmath6 denote the neutrino energies in the initial and final states , respectively . for
the down - scattering part the brink hypothesis was applied which states that gt distributions built on nuclear excited states are the same as those for the nuclear ground state but shifted by the excitation energy . under this assumption ,
the down - scattering part of the cross - section becomes temperature independent .
the temperature dependence arises from the up - scattering part which accounts for contributions of downward transitions from nuclear excited states .
these contributions are determined by inversion of the shell - model gt distributions for the low - lying states .
large - scale shell - model calculations provide a detailed strength distribution of charge - neutral gamow - teller ( gt@xmath7 ) transitions that strongly dominate the inelastic neutrino - nucleus scattering at low neutrino energies ( @xmath8 ) . however , being applied to hot nuclei , this method has its own shortcomings mainly because it partially employs the brink hypothesis when treating gt@xmath7 transitions from nuclear excited states .
as follows from experimental studies of giant dipole resonances ( gdr ) in hot nuclei , the gdr strength function exhibits a temperature dependence ( see , e.g. , the monograph @xcite and one of the latest reviews @xcite ) , i.e. , the validity of the brink hypothesis is not obvious .
moreover , theoretical calculations performed for charge - exchange gt transitions in the framework of the shell - model monte carlo ( smmc ) method demonstrate that with increasing temperature the gt centroid shifts to lower energies and the width of the distribution increases with the appearance of low - lying states @xcite .
in addition , the present computer capabilities allow application of large - scale shell - model calculations only to iron group nuclei ( @xmath9-shell , @xmath10 ) , whereas neutrino scattering on more massive and neutron - rich nuclei also may play an important role in various astrophysical scenarios .
thus , the problem of an accurate description of inelastic neutrino - nucleus scattering in the supernova environment is not solved completely yet and alternative methods are desirable . in @xcite , we have developed such an alternative approach to treat thermal effects on inelastic neutrino - nucleus scattering cross - sections .
this approach is based on the thermal quasiparticle random phase approximation ( tqrpa ) .
we apply it in the context of the thermo field dynamics ( tfd ) formalism @xcite which enables a transparent treatment of upward and downward transitions from thermally excited nuclear states and opens possibilities for systematic improvements .
this approach was also recently used in studies of the electron capture on hot nuclei under supernova conditions @xcite . in @xcite
, the thermal effects on the inelastic neutrino scattering off the hot @xmath11fe nucleus were investigated .
it was shown that the tqrpa does not support brink s hypothesis and leads to temperature dependent strength distributions for allowed and forbidden transitions . as a result , both the up- and down - scattering parts of the cross - section are temperature dependent . despite the differences between the two approaches , the tqrpa revealed the same thermal effect as was found in @xcite .
namely , a temperature increase results in considerable enhancement of the cross - section for neutrino energies lower than the energy of the gt@xmath7 resonance . in the present paper , we extend our previous study by also considering inelastic neutrino scattering off neutron - rich nuclei beyond @xmath9-shell . in our calculations , we take into account not only the first - forbidden transitions but also contributions from higher multipoles . for the selected iron isotope , @xmath12fe
, we perform a detailed comparison of the calculated tqrpa cross - sections with the hybrid approach results and discuss the reason for the observed discrepancy at low neutrino energies .
the paper is organized as follows . in sec .
[ formalism ] we present some important features of the tfd formalism and briefly outline how to treat upward and downward transitions in a hot nucleus within the tqrpa .
the details of our approach are expounded in @xcite . in sec .
[ formalism ] , we also provide the necessary formulas to calculate inelastic neutrino - nucleus cross - sections at finite temperatures .
the results of the numerical calculations are presented and discussed in sec .
[ results ] for the sample nuclei @xmath0fe and @xmath1ge .
the conclusions are summarized in sec .
[ conclusion ] .
in the stellar environment during the core - collapse phase all nuclear reactions mediated by the strong and electromagnetic interaction are in equilibrium with their inverse @xcite . neglecting weak - interaction
mediated reactions , nuclei are in thermal equilibrium with heat and particle reservoirs and , therefore , can be described as a thermal ensemble . in tfd , such an equilibrium ensemble
is represented by a temperature - dependent state termed the thermal vacuum @xmath13 .
the thermal vacuum is determined as the zero - energy eigenstate of the thermal hamiltonian , @xmath14 , and it satisfies the thermal state condition @xmath15 in the above equations @xmath16 is the original nuclear hamiltonian and @xmath17 is its tilde counterpart acting in the auxiliary hilbert space ; an operator @xmath18 acts in the physical hilbert space , @xmath19 is its tilde partner , and @xmath20 is a phase factor .
the thermal state condition guarantees that the expectation value @xmath21 is equal to the ( grand)canonical average of @xmath18 . in this sense ,
relation is equivalent to the kubo - martin - schwinger condition for an equilibrium ( grand)canonical density matrix @xcite .
weak - interaction processes , such as inelastic neutrino scattering , induce transitions from the thermal vacuum to excited states of the thermal nuclear hamiltonian . as follows from the definition of @xmath22 ,
each of its eigenstates with positive energy has a counterpart the tilde - conjugate eigenstate with negative but same absolute value of energy .
transitions from the thermal vacuum to positive energy states ( upward transitions ) correspond to excitation of the nucleus , while transitions to negative energy states ( downward transitions ) describe the decay of thermally excited states .
let us now consider a general nuclear hamiltonian consisting of mean fields for protons and neutrons , pairing interactions , and residual two - nucleon interactions : @xmath23 to fix an average number of protons and neutrons we introduce the respective chemical potentials into @xmath24 .
the residual interaction can contain both particle - hole and particle - particle terms .
we assume a spherically symmetric nucleus , although the deformation can be easily included into the theory . within the tqrpa , to find excited states of a hot nucleus , we first introduce thermal quasiparticle creation ( @xmath25 ) and annihilation ( @xmath26 ) operators which account for pairing correlations at finite temperature .
the structure of this operators is found by diagonalizing the @xmath27 part of the thermal hamiltonian and simultaneously demanding that the vacuum of thermal quasiparticles obeys the thermal state condition .
then , to account for the long - range residual interaction , we introduce thermal phonon operators @xmath28 of given total angular momentum @xmath29 whose action on the thermal vacuum @xmath13 creates thermal excited states , while the thermal vacuum itself is the vacuum for the @xmath30 operators . the structure and the energy of thermal phonons can be found by applying either the variational principle or the equation of motion method under two additional constraints : ( i ) phonon operators commute like bosonic ones ; ( ii ) the vacuum of thermal phonons obeys the thermal state condition .
the resulting phonon operators have the following form:^j_m$ ] denotes the coupling of two single - particle angular momenta @xmath31 to the total angular momentum @xmath32 .
the bar over index @xmath33 means time - inversion . ]
@xmath34^j_m + % \widetilde\psi^{j i}_{j_1j_2}[\widetilde\beta^\dag_{\overline{\jmath_1 } } \widetilde\beta^\dag_{\overline{\jmath_2}}]^j_m
\\ + % \eta^{j i}_{j_1j_2}[\beta^\dag_{j_1 } \widetilde\beta^\dag_{\overline{\jmath_2}}]^j_m % + \phi^{j i}_{j_1j_2}[\beta_{\overline{\jmath_1}}\beta_{\overline{\jmath_2}}]^j_m \\ % + \widetilde\phi^{j i}_{j_1j_2}[\widetilde\beta_{j_1 } \widetilde\beta_{j_2}]^j_m + \xi^{j i}_{j_1j_2}[\beta_{\overline{\jmath_1 } } \widetilde\beta_{j_2}]^j_m\bigr\}\end{gathered}\ ] ] and they diagonalize the thermal hamiltonian @xmath35 within the tqrpa .
the phonon amplitudes @xmath36 as well as the phonon energies @xmath37 are the solution of the tqrpa equations .
it should be emphasized that in the zero - temperature limit the tqrpa method turns into the standard qrpa . in @xcite
, we have performed a detailed analysis of finite temperature effects on the spectrum of charge - exchange thermal phonons . here
we repeat the main conclusions which remain valid for the charge - neutral excitations as well . due to the terms in involving tilde thermal quasiparticle operators ( terms like @xmath38 and @xmath39 ) the spectrum of thermal phonons
contains negative- and low - energy states which do not exist at zero temperature .
since ( see @xcite for more details ) the creation of a tilde quasiparticle is equivalent to the annihilation of a thermally excited bogoliubov quasiparticle , the excitation of the aforementioned `` new '' phonon states can be interpreted as thermally unblocked transitions from nuclear excited states .
furthermore , both the energies of thermal quasiparticles and the interaction strength between them are temperature dependent . as a result , after solving the tqrpa equations we obtain a temperature - dependent spectrum of thermal phonons .
once the structure of thermal phonons is determined , one can evaluate transition strengths ( probabilities ) from the thermal vacuum to thermal one - phonon states . for a given transition operator @xmath40 we have @xmath41 where @xmath42 and @xmath43 are the strengths of upward and downward transitions , respectively .
they are connected by the relationship @xmath44 where @xmath45 is a positive solution of the tqrpa equations .
this relation links the probabilities to transfer and gain energy @xmath46 from a hot nucleus .
it is interesting to note that the same relationship between the upward and downward transition strengths is used in @xcite when considering the thermal strength functions for emission and absorption of neutrino - antineutrino pairs by hot nuclei . in @xcite , the relation results from the principle of detailed balance . in tfd
, it arises from the thermal state condition imposed on the thermal vacuum .
we also would like to point out that in @xcite , due to the application of brink s hypothesis , the absorption ( upward ) strength is considered to be temperature independent and only the emission ( downward ) strength depends on temperature due to the boltzmann factor @xmath47 .
in contrast , within the present approach , both the upward and downward transition strengths are temperature dependent . deriving the inelastic neutrino - nucleus cross - section at finite temperature we follow the walecka - donnelly formalism @xcite which is based on the standard current - current form of the weak interaction hamiltonian .
then the temperature - dependent differential cross - section for a transition from the thermal vacuum to the final thermal one - phonon state takes the form @xmath48 here , @xmath49 is the fermi constant of the weak interaction and @xmath50 the scattering angle .
the coulomb - longitudinal , @xmath51 , and transverse , @xmath52 , terms in eq . are given by @xmath53 @xmath54 -\tan\frac{\theta}{2}\sqrt{-\frac{q^2_\mu}{2q^2 } + \tan^2\frac{\theta}{2 } } \\
\times\bigl [ 2 \mathrm{re } \langle j i\| \hat t^\mathrm{mag}_j\|0(t)\rangle\langle j i\|
\hat t^\mathrm{el}_j\|0(t)\rangle^ * \bigr],\end{gathered}\ ] ] where @xmath55 ( @xmath56 ) is the 4-momentum transfer and the notation @xmath57 is used to denote both the non - tilde and the tilde states .
the upper sign in the above equations refers to upward transitions from the thermal vacuum to non - tilde states ( @xmath58 ) , while the lower sign corresponds to downward transitions to tilde states ( @xmath59 ) .
the multipole operators @xmath60 , @xmath61 , @xmath62 , and @xmath63 denote the charge , longitudinal , and transverse electric and magnetic parts of the hadronic current , respectively , as defined in @xcite . for the vector , axial - vector , and pseudoscalar form - factors which describe the internal structure of the nucleon we use parametrization from ref . @xcite ( see also ref .
@xcite ) . from eq .
, the total cross - section , @xmath64 , as a function of temperature and incoming neutrino energy is obtained by integrating over the scattering angle and summing over all possible final thermal excited states @xmath65 here , we follow ref . @xcite and split the total cross - section into two parts : @xmath66 describes the neutrino down - scattering process and includes only transitions to non - tilde phonon states , while @xmath67 corresponds to the neutrino up - scattering associated with transitions to tilde states . for inelastic scattering of low - energy neutrinos ,
i.e. , in the long wavelength limit ( @xmath68 ) , only two multipole operators survive , @xmath69 and @xmath70 , which contribute to @xmath71 transitions .
then the integration over the scattering angle in eq . can be performed analytically and , in view of the detailed balance principle , the low - energy cross - section can be written as @xmath72 where @xmath73 is the transition strength for the gamow - teller operator ( see eq .
the sum @xmath74 in the first , down - scattering , term implies summation over @xmath71 non - tilde thermal phonon states with the energy @xmath75 .
apparently , for vanishing neutrino energies , @xmath76 , only the second , up - scattering , term persists at finite temperatures .
we also note that although the boltzmann factor suppresses the contributions of downward transitions from high - lying thermally excited states , the phase - space factor @xmath77 acts in the opposite direction and favors them .
the formalism presented above is employed to study thermal effects on the inelastic neutrino scattering off the two sample nuclei , @xmath78 and @xmath79 .
the iron isotope is among the most abundant nuclei at the early stage of the core - collapse , while the neutron - rich germanium isotope can be considered as the average nucleus at later stages @xcite .
let us now specify the nuclear hamiltonian which will be used in the present study .
like in @xcite , we apply a phenomenological hamiltonian containing separable particle - hole residual interactions with isoscalar and isovector parts .
we neglect particle - particle interactions except for the bcs pairing forces .
this hamiltonian is usually referred to as the quasiparticle - phonon model ( qpm ) @xcite . for @xmath78 and @xmath79 ,
the single - particle energies and wave - functions are derived from an appropriate woods - saxon mean - field potential @xcite .
the depth of the wood - saxon potential as well as the pairing strength parameters are fixed in the same manner as in @xcite . in the obtained single - particle level schemes
@xmath78 has two neutron holes in the @xmath80 subshell and two protons in the @xmath81 subshell , while @xmath79 has closed @xmath82 neutron and @xmath81 proton subshells .
it is notable that the sequence of single - particle levels for @xmath79 is close to that used in ref .
@xcite for the same nucleus in spite of different woods - saxon parameterizations . solving the bcs equations at zero - temperature
we get the proton and neutron pairing gaps : @xmath83 for @xmath78 and @xmath84 for @xmath79 .
thus , the critical temperature @xmath85 above which the pairing gap collapses , according to the bcs theory ( see refs .
@xcite for more details ) , is @xmath86 for @xmath0fe and @xmath87 for @xmath1ge . in the present study ,
multipoles up to @xmath88 contributing to the neutrino - scattering cross - section are included in the calculations . to generate the thermal one - phonon excited states
, we use both multipole and spin - multipole components of the residual interaction @xmath89 here @xmath90 and @xmath91 are single - particle multipole and spin - multipole operators @xcite , and changing the sign of the isotopic index @xmath92 means changing @xmath93 .
the excitations of natural parity ( @xmath94 ) are generated by the multipole and spin - multipole @xmath95 interactions , while the spin - multipole interactions with @xmath96 are responsible for the states of unnatural parity ( @xmath97 ) . to generate @xmath98 excitations , we take into account both the particle - hole residual interaction and the particle - particle interaction stemming from the pairing part of the hamiltonian . here
we would like to emphasize that the inclusion of the particle - particle residual interaction into the hamiltonian does not affect the strength distributions and the cross - sections for temperatures above the critical one .
in contrast to @xcite , in the present study the radial form - factors of multipole and spin - multipole operators in eq . have the @xmath99 form .
we found that this form of the radial form - factors gives better agreement with results of relativistic self - consistent qrpa calculations @xcite when comparing multipole composition of the cross - sections ( see the discussion below ) .
the respective isoscalar and isovector strength parameters , @xmath100 and @xmath101 , are first roughly estimated following refs .
@xcite and then partly refined on the basis of available experimental data .
for example , in @xmath0fe the isovector strength parameters @xmath102 and @xmath103 are slightly readjusted to reproduce the experimental centroid energies of the gt@xmath104 and gt@xmath105 resonances @xcite .
we find that the isovector strength parameter @xmath106 estimated according to @xcite reproduces the experimental position of the gdr centroid ( @xmath107 ) in @xmath0fe @xcite quite well .
in addition , the isoscalar strength parameters @xmath108 for @xmath0fe and @xmath1ge are fitted to exclude the spurious @xmath109 state due to the center of mass motion of the nucleus .
fe and @xmath1ge as functions of the incoming neutrino energy @xmath110 .
the total cross - sections include contributions of @xmath111 multipoles ( solid lines ) .
the dashed lines show the cross - sections calculated when the @xmath71 contributions are omitted .
the dash - dotted lines display the @xmath71 contributions to the cross - sections calculated with the full @xmath112-dependent transition operator whereas the @xmath71 contributions calculated with the gt@xmath7 operator are shown by the dash - double dotted lines . ] before proceeding to thermal effects we consider the inelastic neutrino - nucleus cross - sections at zero temperature and perform a comparison with the available results of other approaches .
we note once again that at @xmath113 our calculations are equivalent to the qrpa .
the calculated ground - state cross - sections for @xmath0fe and @xmath1ge are shown in fig .
[ crsect_t0 ] for incoming neutrino energies @xmath114 .
as one can see in the figure , for neutrinos with @xmath115 the total cross - sections are dominated by @xmath71 transitions . due to the energy gap in the @xmath71 nuclear states the cross - sections drop
rapidly to zero as the neutrino energy approaches the reaction threshold . within the present qrpa calculations ,
the lowest @xmath71 states in @xmath0fe and @xmath1ge have energies 4.06 and 2.67 mev , respectively .
note , that the experimental energy of the first @xmath71 excited state in @xmath0fe is @xmath116 mev .
for the ground - state cross - sections we also analyze the effect due to the exploitation of the full @xmath112-dependent @xmath71 transition operator instead of its long wavelength limit . in the latter case
the @xmath71 operator reduces to the gamow - teller operator @xmath117 where @xmath118 @xcite is the ratio of the axial and vector weak coupling constants and @xmath119 denotes the zero component of the isospin operator in spherical coordinates . here
we would like to remind that within the hybrid approach @xcite , the gt contribution to the cross - section is obtained by using the large - scale shell - model calculations . therefore , to make a comparison with the hybrid approach calculation more transparent , we use the same quenching factor for the axial weak coupling constant , @xmath120 , when calculating the matrix elements of the @xmath71 transition operator .
strength in @xmath0fe and @xmath1ge . ]
let us first demonstrate the calculated qrpa ( quenched ) gt@xmath7 strength distributions .
referring to fig .
[ gt0_0 ] , at zero temperature the gt@xmath7 strength is concentrated in the resonance state around @xmath121 . according to our qrpa calculations , the main contribution to the gt@xmath7 resonance in @xmath0fe comes from the proton and neutron single - particle transitions @xmath122 . in @xmath1ge ,
the neutron transition @xmath123 also contributes to the resonance .
in addition , for both nuclei our qrpa calculations predict a weak low - lying gt@xmath7 strength ( @xmath124 ) arising from the @xmath125 single - particle transitions . for @xmath0fe , the gross - structure of the gt@xmath7 strength distribution
agrees quite well with the shell - model results @xcite , meaning that the gt@xmath7 strength is concentrated in the resonance region around 10 mev with a small bump at low energy .
the same good agreement can be found if we compare the shell - model gt@xmath7 distributions for @xmath11fe ( see ref .
@xcite ) with our previously reported qrpa result @xcite . however , although our calculations reproduce the resonance positions in @xmath126fe , it is a well - known fact that the qrpa fails to recover all nuclear correlations needed to correctly describe the full resonance width and produces only a part of it , the so - called landau width .
the latter is quite small for the gt@xmath7 resonance . as a result
, the fine structure of gt distributions in the vicinity of the resonance is not reproduced in our calculations . in this respect
the shell - model calculations are clearly advantageous . using the calculated strength distributions we apply eq . and calculate the gt contribution to the ground - state cross - sections . in fig .
[ crsect_t0 ] , these contributions are shown by the dash - double dotted lines . from the figure we conclude that for neutrino energies @xmath127 mev , when @xmath71 transitions dominate the cross - section , application of the gt@xmath7 operator instead of the @xmath112-dependent @xmath71 operator is fully justified .
however , for neutrinos with the energy @xmath128 the gt@xmath7 operator overestimates the cross - sections by about 25% .
.the cross - sections ( in units of @xmath129 ) for inelastic neutrino scattering on the ground state of @xmath0fe . the present qrpa results ( second column ) are compared with those from @xcite and with the hybrid approach results @xcite . [ cols="^,^,^,^,^",options="header " , ] [ 56fe_comparison ] in table [ 56fe_comparison ] , we compare the calculated ground - state cross - sections for @xmath0fe with those obtained with the hybrid approach @xcite , the relativistic self - consistent qrpa @xcite , and the qrpa - based framework from ref .
the range of incoming neutrino energies is @xmath130 mev .
as it follows from the table , except for low neutrino energies ( @xmath131 mev ) , the cross - sections of all four models are in good qualitative agreement .
it is interesting to note that for @xmath132 mev the present results are generally closer to the hybrid approach results than the results of the other two qrpa - based methods . to explain the discrepancy between our calculations and those of the hybrid approach at low neutrino energies
, we note that at @xmath133 the calculated cross - sections are strongly sensitive to the fine details of the gt@xmath7 distribution in the resonance region .
as it was already discussed above , the large - scale shell - model calculations adequately reproduce the fragmentation of the gt@xmath7 resonance strength whereas the qrpa calculations predict its much stronger concentration near the excitation energy @xmath134 .
for this reason , our cross - section calculated for @xmath135 is considerably smaller than the hybrid approach result . in the cross - sections for the inelastic neutrino scattering on the ground - states of @xmath0fe and @xmath1ge at incoming neutrino energies @xmath136 and 60 mev . for @xmath0fe ,
the results of the present analysis ( grey bars ) are compared with the results of relativistic self - consistent qrpa calculations ( black bars ) .
the latter are obtained from fig . 3 in ref .
@xcite . ]
we also study how relative contributions of different multipoles to the total cross - sections depend on the energy of incoming neutrinos . in fig . [ mult_contributions ] , the relative contributions are shown for @xmath137 neutrinos .
even at @xmath138 a largely dominant multipole is @xmath71 , although contributions coming from the other multipoles are not negligible . for @xmath1ge
this contribution reaches about 30% of the total cross - section .
this is due to the neutron excess which makes possible spin - dipole @xmath109 and @xmath139 @xmath140 transition at relatively low neutrino energies .
the situation is quite different for @xmath141 where the multipole transitions @xmath142 contribute about equally to the cross - sections . in fig . [ mult_contributions
] , we compare the obtained multipole composition of the cross - section for @xmath0fe with that from relativistic self - consistent qrpa calculations @xcite . although our cross - sections are somewhat smaller than those in ref .
@xcite ( compare the second and the third columns of table [ 56fe_comparison ] ) , one can observe an excellent agreement between the two models based on rather different backgrounds . in accordance with ref .
@xcite , we find that @xmath98 allowed transitions only marginally contribute to the inelastic cross - section and this finding is true for finite temperatures as well .
for this reason , in the discussion below , we will always mean only the @xmath71 multipole channel when considering the allowed transitions .
the angular distributions of the scattered neutrinos are shown in fig .
[ angular ] for two incoming neutrino energies , @xmath143 . to make the presentation more transparent , we normalize the calculated differential cross sections to their value at @xmath144 . as shown in the figure , neutrinos scatter predominately in the backward direction . for @xmath145 ,
when @xmath71 transitions dominate , the momentum transfer is small and the angular dependence of the differential cross - section essentially corresponds to @xmath146 @xcite .
the small deviation for @xmath1ge is due to a non - negligible contribution from the forbidden multipoles ( see fig . [ mult_contributions ] ) . for @xmath147
neutrinos , owing to the dominant contribution of forbidden multipoles , the backward - to - forward asymmetry of the differential cross - sections becomes more pronounced .
now we turn our discussion to thermal effects on the inelastic neutrino - nucleus scattering .
we start by considering the temperature evolution of the strength distributions for gt@xmath7 transitions which dominate low - energy neutrino scattering . in fig .
[ gt0_t ] , we display on a logarithmic scale the gt@xmath7 strength distributions at three different temperatures relevant in the supernova context @xcite : @xmath148 mev corresponds to the condition in the core of a presupernova model for a @xmath149 star ; @xmath150 mev and @xmath151 mev relate approximately to neutrino trapping and neutrino thermalization stages , respectively .
the transition energy @xmath152 refers to the excitation energy of a thermal one - phonon state and is equivalent to the neutrino energy transfer . to make the thermal effects more visible
, the ground - state gt@xmath7 distributions are displayed in fig .
[ gt0_t ] as well .
strength distributions for @xmath0fe ( left panels ) and for @xmath1ge ( right panels ) vs. transition energy .
the latter is equivalent to the neutrino energy transfer .
the arrows indicate the zero - temperature threshold @xmath153 for inelastic neutrino scattering off @xmath0fe . ] since the brink hypothesis is not valid within our approach , in fig .
[ gt0_t ] we observe a redistribution of the gt@xmath7 strength for upward transitions ( @xmath154 ) .
namely , at temperatures above the critical one no extra energy has be paid to break a cooper pair .
therefore , by virtue of the vanishing of pairing correlations , both the gt@xmath7 resonance and its low - energy tail move to lower energies .
our calculations indicate that , with increasing temperature up to 1.72 mev , the resulting resonance energy shift reaches about 1.5 mev in @xmath0fe and 1.2 mev in @xmath1ge .
it is interesting to note that within the present tqrpa calculations for @xmath0fe the low - lying gt@xmath7 strength shifts below the zero - temperature threshold ( i.e. , below the experimental energy of the first @xmath71 state ) .
furthermore , the thermal smearing of the nuclear fermi surface makes low - energy particle - particle and hole - hole transitions possible which are pauli - blocked at zero temperature .
such thermally - unblocked transitions enhance the low - lying component of the gt@xmath7 distributions and make it more fragmented . since the @xmath1ge nucleus has a larger single - particle level density near the fermi surface , the temperature - induced enhancement and fragmentation of the low - lying gt@xmath7 upward strength is more significant than in @xmath0fe .
downward strength distributions obtained using ( dashed lines ) and without using ( solid lines ) the brink hypothesis .
note , that the values are scaled by a factor of @xmath155 ( @xmath156 ) and @xmath157 ( @xmath158 ) . ] here we would like to stress that the appearance of a sizeable amount of the low - lying transition strength in nuclei at @xmath159 is predicted in all theoretical studies of hot nuclei .
for example , this was found already in one of the first papers on the subject ref.@xcite , where the distributions of the electric @xmath160 and @xmath161 transitions at @xmath159 in @xmath162pb were calculated and in many subsequent studies ( see , e.g. , @xcite . the same effect is predicted for the charge - exchange allowed and first - forbidden transitions as well @xcite . for the charge - exchange gamow - teller transitions
this feature was also obtained within the shell - model monte - carlo theory @xcite .
moreover , shell - model monte - carlo calculations demonstrate that with increasing temperature the centroid of the gt@xmath105 resonance shifts to lower energies . focusing our attention on the negative energy downward transitions we observe from fig .
[ gt0_t ] that the corresponding gt@xmath7 strength increases with increasing temperature .
this is just a consequence of detailed balance : the higher the temperature , the more substantial is the population of nuclear excited states and hence , the higher is the probability to gain energy from a hot nucleus .
note that the gt@xmath7 strength around @xmath163 can be attributed to the deexcitation of the gt@xmath7 resonance .
given the importance of thermal effects on the upward strength distributions , it is worthwhile to examine how the violation of brink s hypothesis affects the downward strength .
it is obvious that the shift of the gt@xmath7 distributions to lower energies and the appearance of low - energy transitions should magnify the strength of downward transitions .
this effect is clearly demonstrated in fig . [ gt0_running ] which shows running sums for the gt@xmath7 downward strength distributions derived by using the brink hypothesis or not .
the former are obtained from the ground - state ( @xmath113 ) distributions by multiplication with the boltzmann factor @xmath47 . referring to the figure ,
the considerable increase of the overall downward strength is mainly caused by the thermal effects on the low - energy tail in the gt@xmath7 distributions .
this is most pronounced at low temperatures ( @xmath156 ) .
however , at high temperatures ( @xmath158 ) the gt@xmath7 resonance becomes thermally populated and its shift to lower energies also contributes to the downward strength increase .
fe and @xmath1ge at three different temperatures relevant for core - collapse . for comparison
the ground - state cross - sections are also shown .
middle panels : contributions of forbidden transitions to the finite - temperature cross - sections .
lower panels : temperature dependence of the fraction of down - scattered neutrinos in the thermal enhancement of the cross - section . ]
the detailed discussion above allows us to understand better the thermal effects on the inelastic neutrino - nucleus scattering . in the top panels of fig .
[ crsect_t ] , we compare the ground - state cross - sections with those calculated at the three core - collapse temperatures .
as follows from our calculations , temperature effects are unimportant for @xmath164 mev when neutrinos have sufficiently large energy to excite the gt@xmath7 resonance and collective excitations with other multipolarities .
note that a downward shift of the gt@xmath7 resonance only marginally affects the cross - sections at such high neutrino energies .
however , as one can see from the plots , the cross - sections significantly depend on temperature for low - energy neutrinos .
namely , the reaction threshold disappears and the cross - sections are enhanced by up to two orders of magnitude when the temperature rises from @xmath165 to @xmath166 .
it is significant that all these features have pointed out in @xcite as well . in fig .
[ crsect_t ] , we also demonstrate the overall contribution of the forbidden transitions @xmath167 to the cross - sections .
in contrast to hybrid approach calculations @xcite their contributions are temperature dependent . however , comparing the upper and middle panels of fig .
[ crsect_t ] , one concludes that the enhancement of the cross - sections at finite temperatures is essentially due to thermal effects on the gt@xmath7 transition strengths . at vanishing neutrino energies , @xmath76 ,
the finite temperature cross - sections are given by the second , up - scattering , term in eq . which accounts for the gt@xmath7 downward transitions from thermally excited nuclear states .
as shown in fig .
[ gt0_t ] and pointed out previously , the strength of such transitions increases with temperature thereby enhancing the cross - sections
. however , in our approach due to the violation of the brink hypothesis , the down - scattering part of the cross - section , @xmath168 , is also temperature dependent and it increases with temperature owing to the thermally unblocked low - energy gt@xmath7 transitions and the downward shift of the gt@xmath7 resonance .
this effect is clearly shown in ref .
@xcite for @xmath11fe . to analyze relative importance of the two types of neutrino scattering processes in the thermal enhancement of the cross - section
, we introduce the ratio @xmath169 @xmath170 where the difference @xmath171 represents an overall enhancement of the cross - section due to thermal effects .
note that within the hybrid approach @xmath172 , because @xmath173 in this approach .
we plot the ratio @xmath169 in the lower panels of fig .
[ crsect_t ] as a function of @xmath110 for the selected temperatures .
as expected , the ratio @xmath174 for low - energy neutrinos and then , with increase of @xmath110 , @xmath169 gradually decreases indicating a rising contribution of the up - scattering process to the cross - section thermal enhancement .
it is seen from the plots that for @xmath175 neutrinos , both from the up - scattering and down - scattering processes contribute to the noticeable enhancement of the cross - sections , although their relative importance depends on temperature : the higher the temperature the more important is the contribution of the up - scattering process .
consequently , even for @xmath133 neutrinos , when the excitation of the gt@xmath7 resonance becomes possible , the up - scattering component of the cross - section appears to be comparable with the down - scattering one for temperatures @xmath176 .
fe calculated within the present tqrpa approach and the hybrid approach ( ref .
@xcite , fig.11 ) .
the solid and dashed lines show the present results for @xmath156 and @xmath158 , respectively .
the dash - dotted and dash - double - dotted lines show results from ref .
@xcite for the same @xmath177 values . ] in fig .
[ crsect_t_comparison ] , we compare our results for @xmath0fe with those obtained within the hybrid approach @xcite .
the comparison is made for temperatures @xmath178 .
as one can see , at @xmath179 there is noticeable disagreement between the results of the two approaches : the tqrpa cross - sections are larger by a factor of 2 to 5 than the hybrid approach ones . to understand the cause of the discrepancy
, we calculate the spectrum of outgoing neutrinos scattered off @xmath0fe at the same temperatures as in fig .
[ crsect_t_comparison ] and compare the results with the hybrid approach calculations ( see fig.13 of ref .
@xcite ) . in fig .
[ spectra ] , the spectra are shown for the same initial neutrino energies as in ref .
@xcite : @xmath180 .
note that for a clearer presentation and for comparison convenience , the tqrpa spectra are normalized to unity and folded with the breit - wigner function with a width of @xmath181 .
fe at @xmath182 and @xmath166 and three initial neutrino energies : @xmath183 ( solid line ) , @xmath184 ( dashed line , all values shifted by 0.2 ) , and @xmath185 ( dash - dotted line , all values shifted by 0.4 ) .
the triangles correspond to the energy of the incoming neutrino . ] at low temperatures , the downward transitions from the thermally excited gt@xmath7 resonance are strongly suppressed by the boltzmann factor .
therefore , for @xmath156 and @xmath186 the spectrum is dominated by neutrinos up- and down - scattered due to the low - energy gt@xmath7 transitions . in fig .
[ spectra ] , such transitions correspond to the sizable peaks in the spectrum at around @xmath187 .
the dominance of low - energy up- and down - transitions in the scattering of low - energy neutrinos off @xmath0fe at @xmath156 is also supported by the hybrid approach studies ( see the upper - middle panel of fig . 13 in ref .
however , the energy and the strength of such transitions calculated with the tqrpa and the hybrid approach are different . as discussed in detail above , within the tqrpa thermal effects shift the low - lying gt strength in @xmath0fe to energies below the zero - temperature threshold ( fig . [ gt0_t ] ) and significantly increase the strength of the corresponding inverse downward transitions ( fig .
[ gt0_running ] ) , thus favoring neutrino inelastic scattering .
no such effects are expected within the hybrid approach due to the application of the brink hypothesis .
for this reason , the low - energy ( @xmath188 ) tqrpa cross - section at @xmath156 appears to be larger than the hybrid approach one . with increasing energy of incoming neutrinos ,
the excitation of the gt@xmath7 resonance comes into play as evidenced by peaks in the spectra around @xmath189 .
consequently , the cross - section becomes less sensitive to thermal effects on the low - lying gt@xmath7 strength . as a consequence for @xmath190 we observe excellent agreement between the tqrpa and the hybrid approach results .
the situation is slightly different for the higher temperature , @xmath158 .
now downward transitions from the thermally excited gt@xmath7 resonance are possible , and owing to a large phase - space factor they can contribute significantly to the up - scattering of low - energy neutrinos . in the spectra this contribution corresponds to the peak at around @xmath191 . as indicated in fig .
[ gt0_running ] , at @xmath192 thermal effects increase the strength of downward transitions from the gt@xmath7 resonance .
although this increase is not of the same magnitude as for the low - lying gt@xmath7 strength , due to a larger phase - space factor , its contribution to the cross - section enhancement is substantial .
therefore , one can conclude that at @xmath158 the joint action of thermal changes in both the low - lying gt@xmath7 strength and the gt@xmath7 resonance enhances the absolute value of the tqrpa cross - section at @xmath193 in comparison with the hybrid approach .
like for the @xmath156 case , thermal effects become less important for higher neutrino energies , and both approaches yield very similar results for the cross - sections . considering thermal effects on angular distributions of outgoing neutrinos
we find that they are rather unimportant . for up - scattered neutrinos these distributions
are only slightly more backward peaked than those for down - scattered neutrinos owing to the larger momentum transfer .
we have studied thermal effects on the inelastic neutrino - nucleus scattering in the supernova environment .
the thermal effects were treated within the thermal quasiparticle random phase approximation .
the calculations were performed for @xmath0fe and @xmath1ge .
we find that the tqrpa calculations do not support the brink hypothesis and lead to temperature - dependent strength distributions for both allowed and forbidden transitions involved in the neutrino inelastic scattering .
it is shown that thermal effects shift the gt@xmath7 centroid to lower energies and make low - energy gt@xmath7 transitions possible . as a result , in contrast to hybrid approach calculations @xcite , both the up - scattering and down - scattering components of the cross - section exhibit a noticeable temperature dependence at low - energy neutrinos .
our calculations reveal the same thermal effects as the hybrid approach based calculations .
namely , the reaction threshold for inelastic neutrino - nucleus scattering is removed at finite temperatures and the cross - section for low - energy neutrinos is significantly enhanced .
however , the calculated cross - sections for @xmath0fe at low neutrino energies are several times larger than those obtained within the hybrid approach .
we have shown that the discrepancy is due to the violation of brink s hypothesis in our approach .
this is the main result of the present study .
in addition , it has been demonstrated that the tqrpa approach can be used to study inelastic neutrino scattering off massive neutron - rich nuclei at finite temperatures .
another advantage of our approach is that it incorporates the detailed balance principle , whereas in the hybrid approach detailed balance is violated .
there are several directions to improve our approach . at present , its predictive power is limited by the phenomenological hamiltonian with schematic separable residual interactions .
it would therefore be desirable to combine our tfd - based tqrpa method with self - consistent qrpa calculations based on more realistic effective interactions . for neutrino scattering and neutrino absorption reactions at zero temperature ,
such calculations were performed recently within the relativistic nuclear density functional theory @xcite .
this improvement would also allow to take into account the effects of nuclear deformation . for supernova electron - capture rates in @xmath9-shell nuclei ,
deformation was recently included in self - consistent qrpa calculations with the skyrme interaction @xcite .
the other direction of improvement is the inclusion of correlations beyond the tqrpa by taking into account the coupling of thermal phonons with more complex ( e.g. two - phonon ) configurations . at zero temperature
this problem was considered within the qpm @xcite by exploiting separable schematic effective interactions . however , with the separable approximation for the skyrme interaction @xcite one could consider phonon coupling at finite temperatures within a self - consistent theory . the collaboration with prof . t. s. kosmas and dr .
v. tsakstara from the university of ioannina is acknowledged .
we are also greatly indebted to prof .
g. martnez - pinedo for helpful discussions and important comments on this paper .
this work was supported in part by the heisenberg - landau program , the cnrs - rfbr grant 11 - 091054 , and the deutsche forschungsgemeinschaft grant no .
sfb 634 .
47ifxundefined [ 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 \doibase http://dx.doi.org/10.1016/j.physrep.2007.02.002 [ * * , ( ) ] , link:\doibase 10.1103/physrevlett.60.1999 [ * * , ( ) ]
link:\doibase 10.1086/170316 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.011101 [ * * , ( ) ] link:\doibase 10.1086/170317 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.nuclphysa.2004.09.005 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/s0370-2693(01)01502-7 [ * * , ( ) ] link:\doibase 10.1103/physrevc.63.025802 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/s0375-9474(01)00992-7 [ * * , ( ) ] http://books.google.ru/books?id=ceabfyskw-cc[__ ] , contemporary concepts in physics ( , ) link:\doibase 10.1140/epja / i2006 - 10116 - 7 [ * * , ( ) ] link:\doibase 10.1103/physrevc.56.3079 [ * * , ( ) ] link:\doibase 10.1134/s1063778811080059 [ * * , ( ) ] link:\doibase 10.1142/s0217979296000817 [ * * , ( ) ] http://books.google.ru/books?id=ra7vaaaamaaj[__ ] ( , ) \doibase http://dx.doi.org/10.1016/0003-4916(81)90058-0 [ * * , ( ) ] link:\doibase 10.1103/physrevc.81.015804 [ * * , ( ) ] link:\doibase 10.1142/s0218301309013695 [ * * , ( ) ] link:\doibase 10.1143/jpsj.12.570 [ * * , ( ) ] link:\doibase 10.1103/physrev.115.1342 [ * * , ( ) ] @noop ( ) in http://books.google.ru/books?id=l9ic3gt578gc[__ ] , ( , ) p.
\doibase http://dx.doi.org/10.1016/0370-1573(79)90010-3 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/s0920-5632(02)01761-9 [ * * , ( ) ] link:\doibase 10.1103/physrevc.86.035804 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0375-9474(84)90673-0 [ * * , ( ) ] http://books.google.ru/books?id=ly3_c3zg3quc[__ ] ( , ) @noop * * , ( ) , \doibase http://dx.doi.org/10.1016/0375-9474(81)90557-1 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0375-9474(83)90411-6 [ * * , ( ) ] http://stacks.iop.org/1742-6596/410/i=1/a=012172 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0370-2693(76)90526-8 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0370-1573(75)90056-3 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0375-9474(83)90632-2 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0375-9474(93)90603-u [ * * , ( ) ] link:\doibase 10.1103/physrevc.24.1940 [ * * , ( ) ] in http://books.google.ru/books?id=i1bjqgaacaaj[__ ] , ( , ) p.
\doibase http://dx.doi.org/10.1016/j.ppnp.2007.01.003 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0375-9474(84)90434-2 [ * * , ( ) ] \doibase
[ * * , ( ) ] link:\doibase 10.1103/physrevc.87.025801 [ * * , ( ) ] link:\doibase 10.1103/physrevc.87.045801 [ * * , ( ) ] link:\doibase 10.1103/physrevc.57.1204 [ * * , ( ) ] link:\doibase 10.1007/bf01323673 [ * * , ( ) ] http://stacks.iop.org/0953-8984/24/i=22/a=225304 [ * * , ( ) ] | we study inelastic neutrino scattering off hot nuclei for temperatures relevant under supernova conditions .
the method we use is based on the quasiparticle random phase approximation extended to finite temperatures within the thermo field dynamics ( tqrpa ) .
the method allows a transparent treatment of upward and downward transitions in hot nuclei , avoiding the application of brink s hypothesis . for the sample nuclei @xmath0fe and @xmath1ge
we perform a detailed analysis of thermal effects on the strength distributions of allowed gamow - teller ( gt ) transitions which dominate the scattering process at low neutrino energies . for @xmath0fe and @xmath1ge the finite temperature cross - sections are calculated by taking into account the contribution of allowed and forbidden transitions .
the observed enhancement of the cross - section at low neutrino energies is explained by considering thermal effects on the gt strength .
for @xmath0fe we compare the calculated cross - sections to those obtained earlier from a hybrid approach that combines large - scale shell - model and rpa calculations . |
this conference on the disk - halo interaction in galaxy evolution has been a showcase for many exciting new results in a broad spectrum of galactic and extragalactic research . later in this summary
i will briefly consider the significance of some of these results , and how they show the way forward .
but first a few paragraphs of historical background , for the benefit of people who are new to the field .
these concepts were implicit in many of the papers at the conference , but were not often mentioned explicitly . for astrophysicists ,
the disk - halo connection is almost as old as the interstellar medium itself . by the early 1950 s it was clear that both cosmic rays and interstellar clouds must sometimes leave the disk .
a series of papers culminating in spitzer s seminal 1956 analysis ( spitzer @xcite , shklovsky @xcite , pickelner @xcite ) , based on the gravitational potential of the disk determined much earlier by oort ( @xcite ) , predicted a hot gas halo above and below the disk .
the need for cosmic ray ( cr ) leakage into this halo implied that magnetic fields must connect it with the disk . in the decade that followed , detailed astrophysical calculations of the heating and cooling rates of the neutral and ionized gas in the interstellar medium ( ism ) showed that thermal equilibrium between different phases at very different temperatures could be achieved at a range of pressures that depends on the heavy element abundance and the radiation field intensity ( field _ et al . _
@xcite , dalgarno & mccray @xcite ) .
meanwhile , by the early 1960 s the high velocity clouds seen in 21-cm line surveys were interpreted as evidence for both a population of neutral clouds in the halo , and a surrounding hot gas through which they were falling toward the disk ( oort @xcite , @xcite ) .
dramatic observational confirmation of the existence of hot ism gas in the milky way disk came in the early 1970 s from soft x - ray emission ( williamson _ et al . _
@xcite , sanders _
@xcite ) and uv absorption ( jenkins & meloy @xcite , york @xcite ) .
this hot gas was immediately identified with that predicted by spitzer for the halo , and the cycling of gas between the disk and halo was proposed soon after by shapiro & field ( @xcite , c.f .
bregman @xcite ) .
it was soon clear that the hot phase gas occupies much of the disk and connects with the halo , in part from interpretation of the anticorrelation between nearby atomic and molecular clouds and the soft x - ray emission ( mccammon _ et al . _
@xcite , mebold _ et al . _
@xcite ) .
the theoretical framework for understanding the hot phase of the ism developed in the early 1970 s as well , as the role of old supernova remnants ( snr ) in pressurizing the disk was explored quantitatively ( cox & smith @xcite , salpeter @xcite , mckee & ostriker @xcite ) . by the mid-1980
s it was clear that the mckee - ostriker paradigm was insufficient to describe the real ism . the milky way ( mw ) and
many other spiral galaxies have star formation rates high enough to drive hot gas out of the disk , and with it energy in thermal and mechanical forms ( cox & mccammon @xcite , cowie @xcite , norman & ikeuchi @xcite , heiles @xcite ) , which is not included in the mckee - ostriker model .
meanwhile the power of computational astrophysics was brought to the problem of snr evolution ( chevalier & gardner @xcite , mansfield & salpeter @xcite , struck - marcell & scalo @xcite , maclow _ et al . _
@xcite , cioffi & shull @xcite ) , culminating in two- and finally three - dimensional simulations of the global ism ( rosen & bregman @xcite , vazquez - semadeni _ et al . _
@xcite , avillez @xcite ) .
one of the threads running through all of these works , observational and theoretical , is that we can not fully understand the ism of the mw disk without understanding also that of the halo and how the two media interact and exchange matter . in the last two decades our observational data on the ism in the mw halo
has improved dramatically .
even more dramatic progress has been made in studies of the ism of other galaxies .
we can see clearly now that galactic fountains operate in some spiral galaxies , and not in others .
we see examples of disks with infall and outflow .
it is even becoming possible to trace these processes as functions of redshift .
one of the goals of this conference is to bring our detailed knowledge of the disk - halo interaction in the mw to bear on the much larger question of galaxy evolution over cosmic time .
this conference has been ambitious in its combination of topics , including observational , theoretical , and computational studies of the mw disk and halo , the disks and halos of nearby galaxies , and gas infall and outflow in galaxies at intermediate and high redshifts .
the papers reviewing these topics allow us to look back and also ahead , to see what will become possible in the next decade or so as powerful new telescopes become available . rather than try to summarize the papers individually ,
i will consider a few themes , and group the papers accordingly . of course any such grouping is arbitrary and does not do justice to the depth and complexity of the papers or their subjects , but there is not room for much more detail .
citations given as names without dates refer to papers presented at this meeting and their written versions that appear in this volume .
it has been understood for a long time that the ism is subject to a magnetic rayleigh - taylor instability wherein the light fluids constituted by the magnetic field and the relativistic cr electrons can move through the heavier fluid constituted by the thermal phases of the gas to escape the gravitational potential of the mw disk ( parker @xcite ) . but how much impact this has on the dynamics of the gas in the disk and halo is still not known .
one approach is to map the interstellar magnetic field to see where and how the field joins the disk with the halo .
it is possible in various ways to map the field in the disk , as explained in the papers by beck and valle .
searching for a vertical component of the field in the solar neighborhood is particularly interesting , as reported by mao , based on faraday rotation surveys at high latitudes .
troland s paper explains how zeeman splitting observations at 21-cm give values for @xmath0 .
this quantity is about 0.3 for the thermal pressure ( with kinetic velocities given by the temperature ) but @xmath1 when non - thermal bulk velocities are included .
this reflects the typical mach number of about 4 for turbulent velocities in cool , neutral medium ( cnm ) clouds . thus the magnetic field is dynamically important in this phase , and probably in all other phases of the ism , but it is not completely dominant .
it is significant , however , that magnetic pressure is much higher than thermal gas pressure , as this could moderate the rate of cooling .
ferrire s paper supports troland s in these conclusions , and gives an overview on a larger scale of the magnetic field structure of the disk .
the dynamical importance of the magnetic field has recently been included in simulations of the structure and dynamics of the ism , as reported at this conference by breitschwerdt , asgekar , hanasz , and woltanski . including the cr pressure , which breitschwerdt couples to the gas through a spectrum of magneto - acoustic waves , drives the gas quite effectively out of the disk , in good agreement with the analytical prediction of everett .
the extensive review by dogiel presents several fundamental equations and observations that describe cr propagation through the disk and halo .
some well established results from this are the scale height of the cr population ( 3 kpc ) , the propagation path length ( 3 mpc ) , and hence the travel time ( 10@xmath2 yr ) .
clearly there is a lot of energy available in the form of magnetic fields and cr s ; the cr acceleration processes alone have luminosity of more than 10@xmath3 erg s@xmath4 .
further , per unit energy density , the cr s are more efficient at driving mass out of the disk and into the halo than is simple gas pressure , as everett explains . on the much larger scales of clusters of galaxies , the paper by bernet reviews the evidence for magnetic fields to at least redshift z=1.3 as traced by faraday rotation .
thus we can infer that magnetic fields and cosmic rays have been escaping from galaxies , at least some galaxies in some environments , for a very long time .
the different thermal phases of the ism are typically studied with different spectral lines or other tracers , so understanding the juxtaposition and relative abundances of the different phases is like putting together separate pieces of a puzzle .
ferrire s , jenkins , and vzquez - semadeni s papers review concepts of pressure equilibrium , thermal equilibrium , transition of gas from one phase to another , and how the mixture of phases depends on height above the plane , @xmath5 .
the scale heights of the warm and cool phases are discussed further by gaensler , reynolds , and kalberla .
the ism phases in m31 are discussed in the papers of berkhuijsen , bogdan , and braun .
the phase - mixture of the mw ism is one of the central topics of the conference , and the papers in this volume represent the considered opinions of people who have been leaders of this field for decades . yet the numbers presented here , particularly the scale heights and their variation with galactic radius , differ in some cases by 30% to 50% from estimates made a few years ago . clearly
the diffuse ism of the mw needs further research ! but the papers presented at this meeting reinforce the general consensus that we do understand the ism in the mw disk fairly well , and we can confidently extend our observing techniques and analysis tools to the halo , to nearby galaxies , and on to high redshifts .
the scale heights of the phases do not completely describe the structure of the ism in the mw halo . as lockman puts it in his presentation , in the halo
we see `` hydrogen that has a story to tell about how it got there . ''
one of the most dramatic examples is smith s cloud ( smith @xcite , lockman _
@xcite ) , a high velocity cloud ( hvc ) with mass greater than 10@xmath6 m@xmath7 and total space velocity of about 300 km s@xmath4 , approaching the galactic plane at some 75 km s@xmath4 .
this is large compared to typical hvc s described in wakker s review , but by no means the largest , as complex c is estimated to have mass 4 x 10@xmath2 m@xmath7 .
typical distances above or below the plane , @xmath8 , are in the range 5 to 15 kpc for these large hvc complexes . assuming that they represent primordial material , their mass accretion rate averages to about 0.5 to 1 m@xmath7 yr@xmath4 . as putman points out in her paper , maintaining an accretion rate of 1 m@xmath7 yr@xmath4 over the last 5 to 7 gyr would require a primordial reservoir of as many as 500 dwarf galaxies with hi mass 10@xmath2 m@xmath7 , or 5000 hvc complexes of 10@xmath6 m@xmath7 .
this is a lot , there certainly do not seem to be nearly that many such systems in the local group remaining to be accreted .
so it is important to work out the stories that the hvc complexes have to tell , by detailed studies such as those described by ford , dedes , madsen , haffner , leiter , and hill .
the larger objective of such work is to determine what their precursors were .
some of today s hvc complexes may have come from dwarf galaxies that were tidally disrupted and captured by the mw , or they may have come from tidal stripping of an existing local group member , like the magellanic stream discussed in the papers by madsen , wakker , and putman , or perhaps they have somehow been part of the mw halo since the epoch of galaxy formation . there seems to be a clear distinction between the large , 21-cm selected hvc complexes and the much lower mass clouds that dominate optical and uv surveys toward distant halo stars , described by wakker , ben - bekhti , nasoudi - shoar , and fox , and earlier by smoker _
( @xcite ) , lockman & savage ( @xcite ) , and see the historical data discussed by benjamin & danly ( @xcite ) .
the clouds seen in these surveys are similar to clouds seen in extragalactic absorption studies ; at least some of them are classical fountain - return clouds .
some of the hi hvc s are probably in this category as well . for launching the fountain ,
the large galactic chimneys and supershells are leading contenders , as described by mcclure - griffiths , sato , pidopryhora , gonalves , and dawson .
these concentrate and direct the energy provided by many supernova remnants to drive hot gas out of the plane and into the lower halo .
such structures appear to be common in nearby spiral galaxies .
the structure of the m82 superwind is reminiscent of a galactic chimney driven to the extreme by the starburst nucleus , as discussed by westmoquette and trinchieri .
one of the most successful aspects of this conference has been the seamless merging of galactic and extra - galactic observations and theory .
this shows the maturity of the field , in that the same astrophysical phenomena that we study in detail in the mw we can trace more broadly in nearby spirals , and with fair sensitivity even at intermediate to high redshift .
thus many of the themes in dettmar s review of disk - halo interaction in nearby galaxies echo those of ferrire s review of the same topic in the mw , and many other reviews include both galactic and extra - galactic examples .
in particular , the hi hvc s in the local group and nearby galaxy groups reviewed by putman , oosterloo , and chynoweth , are analogous to the large hvc complexes in the mw halo .
but their morphology can be completely different . unlike the local group
, the m81-m82-ngc3077 group has a neutral intra - group medium , probably a massive tidal feature , as described by chynowyth .
the local group is more typical , apparently cases like the m82 group are rare .
but it is critically important to get more sensitive hi surveys of many more nearby groups , in order to interpret the data from absorption surveys at intermediate and high redshifts .
this will come from new arrays with enhanced surface brightness sensitivity , like the australian square kilometre array pathfinder ( askap ) .
spiral galaxies seen very nearly edge - on are excellent subjects for study of the @xmath5 distribution of the ism phases , and the structure of the gas in ther halos .
such studies are the topics of the papers by rossa , wu , rand , heald , howk , and oosterloo s poster , and they are included in the reviews by dettmar and reynolds . from these studies
we learn what a range of fountain activity there is in spiral galaxies , depending on the star formation rate ( sfr ) .
the lower halos of the edge - on spirals that have been studied in detail show how different the scale heights of the diffuse neutral and warm ionized media can be , ranging from a few hundred pc to a few kpc .
rossa also makes the point that the morphology of the diffuse ionized gas changes with sfr , with the high sfr galaxies showing a smooth ionized halo , and the intermediate ( lower ) sfr galaxies showing many wisps that probably trace large chimneys , bubbles , and supershells .
another critical question that edge - on spirals can help answer is how the rotation velocity of the gas in the halo departs from that of the disk below .
heald and benjamin discuss this based on observations , and fraternali reviews theoretical and computational studies . here again , the sfr seems to play a critical role .
values for the gradient in rotation velocity with @xmath5 vary from -8 to -30 km s@xmath4 kpc@xmath4 , with the higher values seen in the galaxies with lower sfr s .
this may imply that high fountain activity strengthens the dynamical connection between the disk and lower halo , essentially acting as a viscosity that depends on sfr .
similar numbers for the milky way are still a bit controversial ; levine _ et al . _
( @xcite ) find -22@xmath96 km s@xmath4 kpc@xmath4 , see also the review by kalberla & kerp ( @xcite ) . a series of major surveys of nearby galaxies including sings ( kennicutt _ et al . _
@xcite ) , things ( walter _ et al . _
@xcite ) , and song ( regan _ et al . _
@xcite ) are mentioned in a few papers in this conference , including those of leiter and portas .
these surveys are primarily concerned with the disks of nearby spiral galaxies .
they are rich in information on the mixture of ism phases , and the links from the ism to star formation .
they will also prove useful for understanding the disk - halo connection .
these surveys are particularly interesting for revealing the structure of the outer disks , beyond the area of active star formation .
conditions in such outer disks are in some ways intermediate between those in the inner disk and in the halo , as discussed by kalberla and dickey for the mw .
these surveys give a hint of the even more powerful studies of nearby galaxies that will be possible in the coming decade with telescopes such as alma , jwst , and the ska .
surveys of absorption lines in the spectra of qso s at high redshift are a very powerful way to trace the abundance of gas clouds intervening along the path to the continuum source .
studies of the lyman alpha forest can detect clouds with atomic hydrogen column densities as low as 10@xmath10 @xmath11 .
ranges of progressively increasing column densities are sampled by metal line systems , lyman limit systems , and damped lyman alpha systems finally reaching column densities of 10@xmath12 @xmath11 or higher .
a remarkable result from forty years of such studies is that the abundance of absorbing systems as a function of column density , @xmath13 , can be described as a power law over this entire range of column densities , with @xmath14 , as reviewed by richter .
many of these absorption lines originate in the halos or outer disks of intervening galaxies ; we infer that some of these clouds are driven by galactic fountain processes similar to those we see in the mw and nearby galaxies .
the ism in high redshift galaxies can also be studied using emission line tracers , particularly at mm and sub - mm wavelengths , as reviewed by greve .
as alma comes on - line in the next few years , this field will grow rapidly , and we may hope to detect galaxies like the mw at redshift z@xmath151 . for now , most of the galaxies studied at such great distances are at the high end of the co and far - ir luminosity functions , typically starburst and ultra - luminous ir galaxies plus some qso s . with the power of 8 m class optical telescopes ,
absorption lines can be studied not only toward bright qso s , but also using the light of normal galaxies at high redshift .
the deep survey described by koo , and the specific examples presented by rubin , show the amazing accomplishments of high z spectroscopy at the keck telescope . both outflow and infall are seen in galaxies in the deep sample , with outflow correlating with infrared luminosity , but infall not .
results of stacked spectra give very useful numbers , such as that 10% of absorption lies beyond the escape velocities of the galaxies , and the fact that almost all galaxies have at least some outflow gas . a further result
is that it is not the dwarfs but the more massive galaxies , like the mw , that dominate the total outflow that ultimately enriches the inter - galactic medium with heavy elements . to calibrate the present abundance of absorbing clouds vs. the numbers in the early universe requires low redshift surveys of the far - uv lines , seen shifted into the optical and near - ir at high redshift , using far - uv telescopes like fuse .
the papers by fox and richter give useful comparisons among different line tracers in the optical and uv .
an interesting twist is the appearance of two populations of mg ii absorbers .
the stronger class of lines is more reliably a tracer of outflows .
theories and questions involving galaxy evolution and the history of disk - halo interactions were presented by bernet , jachym , fangano , martin and bregman .
there are some very big issues involved , such as the `` missing baryon problem '' raised by bregman , and the metal enrichment of the igm discussed in detail by martin .
martin makes the point that `` what you see is not everything that is there . ''
it is very important to remember that outflows can and do happen at temperature and density combinations for which we have no sensitive tracers .
for example , at temperature @xmath16 k ( 1 kev x - rays ) we can hardly make out the winds even from nearby galaxies , let alone high redshift systems , as trinchieri explains .
telescopes that will become available over the next decade will bring improvements in resolution and sensitivity of an order of magnitude or more in many wavebands . with such improvements over current capabilities it will become possible to study galaxies like the mw at z=1 with the same level of detail that we now apply to galaxies in nearby groups .
this conference has demonstrated that * extragalactic and galactic astrophysics merge * when the observational data get that good .
the point is , the questions we are asking now about nearby galaxies are the same ones we have been asking and answering in the mw for decades .
soon we may be doing the same for galaxies at z=1 and beyond .
experts in this field , albeit self - appointed ones like many of us attending this conference , stand to gain academic ground in studies of galaxies at high redshift , as the science shifts into our purview .
we can start thinking now about what we would do if we had images and spectra of galaxies at cosmological distances as good as those we have now for , e.g. , m81 and m82 . here are some questions that we might begin to answer : * do disks assemble by satellite accretion ?
why does nt this disrupt the disk ? does it lead to sf bursts ?
if there are sf bursts , do they denude the disk ? does sf have a negative feedback `` governor '' that limits its speed in a disk where the ism can easily be lost to a wind or fountain ? * is the intra - cluster medium ( icm ) that we see in nearby rich clusters representative of the inter - galactic medium ( igm ) , in metallicity and in the ratio of total baryon mass to galaxy light ?
if so , how and when did all the baryons escape from the galaxies ?
was this happening at the same time as the accretion that built galaxies ?
* did the magnetic field build up slowly through a dynamo process in galaxy disks , or was it rapidly generated during the epoch of galaxy formation ?
how much seed field came from the big bang itself ?
is there an igm of magnetic fields and cr s similar to those seen in the icm of rich clusters ? in the next decade
we may be confronted with the data we need to answer these questions .
`` confronted '' because we may not be ready to understand it all . for want of more profound insight , the interpretation of that data will probably follow astrophysical analysis paths similar to those discussed in this volume as applied to the disk and halo of the mw and nearby galaxies .
we can look forward to extending what we have learned about our own disk - halo connection to galaxies on much larger scales .
if this brings deeper understanding of the history of the universe , then we can hope for more conferences as interesting as this one in years to come .
avillez , m.a . , 2000 ,
mnras 315 , 479 .
benjamin , r.a .
, & danly , l. , 1997 , apj 481 , 764 .
bregman , j.n .
, 1980 , apj 236 , 577 .
chavalier , r.a . & gardner , j. , 1974 , apj 192 , 457 .
cioffi , d.f . , & shull , j.m .
1991 , apj 367 , 96 .
cox , d.p . & smith , b.w . , 1974 ,
apj 189 , 105 .
cox , d.p . & mccammon , d. , 1986 , apj 304 , 657 .
cowie , l.l . , 1987 , in interstellar processes , eds .
hollenbach & h.a.thronson , ( dordrecht : reidel ) , p. 245 .
dalgarno , a. , & mccray , r.a . , 1972 ,
ara&a 10 , 375 .
+ field , g.b . ,
goldsmith , d.w . , and habing , h.j . , 1969 , apj lett .
155 , l149 .
heiles , c. , 1990 , apj 354 , 483 .
jenkins , e.b . , and meloy , d.a .
, 1974 , apj 193 , l121 .
kalberla , p.m.w . & kerp , j. , 2008 , ara&a 47 in press .
kennicutt , r.c . ,
armus , l. , bendo , g. , calzetti , d. , dale , d.a .
_ 2003 , pasp 115 , 928 .
levine , e.s . ,
heiles , c. , & blitz , l. , 2008 , apj 679 , 1288 .
lockman , f.j .
, benjamin , r.a .
, heroux , a.j .
, & langston , g.i . , 2008 , apj 679 , l21 .
lockman , f.j .
& savage , b.d .
1995 , apjs 97 , 1 .
maclow , m .-
m . , mccray , r. , & norman , m.l . , 1989 ,
apj 337 , 141 .
mansfield , v.n .
& salpeter , e.e . , 1974 ,
apj 190 , 305 .
mccammon , d. , meyer , s.s . , sanders , w.t . , and williamson , f.o .
, 1976 apj 209 , 46 .
mckee , c.f .
& ostriker , j.p . , 1977 ,
apj 218 , 148 .
mebold , u . , cernicharo , j. , velden , l. , reif , k. , crezelius , c. , & goerigk , w. , 1985 a&a 151 , 427 .
norman , c.a . &
ikeuchi , s. , 1989 , apj 345 , 372 .
oort , j.h . , 1932 ,
oort , j.h .
1966 bull .
inst . neth .
oort , j.h .
1970 a&a 7 , 381 .
parker , e.n . , 1966 ,
apj 145 , 811 .
pickelner , s.b .
1953 , c.r .
urss , 88 , 229 .
regan , m.w . ,
thornley , m.d . ,
helfer , t.t . ,
sheth , k. , wong , t. , _ et al .
_ 2001 , apj 561 , 218 .
rosen , a. & bregman , j.n .
1995 , apj 440 , 634 .
salpeter , e.e .
, 1976 , apj 206 , 673 .
sanders , w.t . , kraushaar , w.l . ,
nousek , j.a .
, & fried , p.m. , 1977 , apj 217 , 87 .
shapiro , p.r . & field , g.b .
, 1976 apj 205 , 7625 .
shklovsky , i.s . , 1952 ,
j. u.s.s.r . , 29 , 418 .
smith , g.p . , 1963 , bull .
netherlands , 17 , 203 . smoker , j.v .
, rolleston , w.r.j . ,
kay , h.r.m . ,
kilkenny , d. , arnal , m. , _ et al .
_ 2003 , mnras , 346 , 119 .
spitzer , l. , 1956 , apj 124 , 20 .
struck - marcell , c. & scalo , j.m . , 1984 ,
apj 277 , 122 .
vazquez - semadeni , e. , passot , t. , & pouquet , a. , 1995 , apj 441 , 702 .
walter , f. , brinks , e. , de blok , w.j.g .
, bigiel , f. , kennicutt , r.c . ,
thornley , m.d . , & leroy , a. , 2008 , aj 136 , 2563 .
williamson , f.o . , sanders , w.t . ,
kraushaar , w.l .
, mccammon , d. , borken , r. , & bunner , a.n .
, 1974 , apj 193 , 133 .
york , d.g .
1974 , apj 193 , 127 . | the papers in this volume represent a broad spectrum of observational , theoretical , and computational astrophysics , sharing as a unifying core the disk - halo interaction in the milky way and other spiral galaxies .
this topic covers a wide range of galactic and extra - galactic research , built on a foundation of numerous and diverse physical processes .
this summary groups the papers according to six themes , with some historical background and finally a look to the future .
the final message is that the astrophysical techniques discussed and reviewed at this conference will grow over the next decade to answer even more fundamental questions about galaxy evolution and the history of the universe . |
the currently standard hierarchical clustering paradigm for large scale structure formation in a @xmath0cdm universe has successfully confronted a broad variety of observations , ranging from the formation of galaxy clusters , to large scale velocity fields , to power spectra of the galaxy distribution and of the microwave background , and more ( see , e.g. , @xcite ) . however , the theory of formation and evolution of galaxies is not in a very satisfactory state .
serious challenges have emerged in the last years ( @xcite ) : the excess of predicted small scale structure ; the persistent inability of even the best semi - analytic models ( @xcite ) to account for the surface density of massive galaxies at substantial redshifts detected by ( sub)-mm surveys with scuba and mambo ( @xcite ) and by deep k - band surveys ( @xcite ) ; the difficulties to account for the distribution of velocity dispersions of low - ionization damped lyman-@xmath1 systems at @xmath2 ( @xcite ) ; the low predicted specific angular momentum of galactic disks ( @xcite ) ; the observational evidence contradicting the existence of the central cusp in the dark matter distribution , predicted by numerical simulations ( @xcite ) . on larger scales ,
the observed relationship between x - ray luminosity and gas temperature for groups of galaxies strongly deviates from expectations of the simplest hierarchical clustering models .
solutions of these problems may require advances in different fields that may include particle physics ( self - interacting dark matter ? ) , deviations from a power law of the power spectrum of primordial density perturbations ( as may be suggested by wmap data , @xcite ) , or the astrophysics of galaxy formation and evolution . in any case
, a better understanding of the complex physical processes governing the galaxy formation is mandatory .
we discuss here how unique information may be provided by the sunyaev - zeldovich ( sz ) effect .
in fact , the the proto - galactic gas is expected to have a large thermal energy content , leading to a detectable sz signal , both when the protogalaxy collapses with the gas shock - heated to the virial temperature ( @xcite ) , and in a later phase as the result of strong feedback from a flaring active nucleus ( @xcite ) .
let us consider a fully ionized gas with a thermal energy density @xmath3 , within the virial radius @xmath4 where @xmath5 is the hubble constant in units of @xmath6 and @xmath7 , @xmath8 is the mean density of the universe at the virialization redshift @xmath9 .
the comptonization parameter @xmath10 , characterizing the amplitude of the sunyaev - zeldovich effect due to this gas , can be estimated as ( @xcite ) : @xmath11 here @xmath12 , @xmath13 being the black - body constant and @xmath14 the temperature of the cosmic microwave background ( cmb ) .
the amount @xmath15 of gas thermal energy transferred to the cmb through thomson scattering is : @xmath16 @xmath17 being the gas cooling time by thomson scattering . a useful reference value for the thermal energy content of the gas
, @xmath18 is its binding energy ( @xmath19 , @xmath20 being the circular velocity of the galaxy at its virial radius , cf .
the amplitude of the sz dip in the rayleigh - jeans region can be written as : @xmath21 although current estimates of the ratio of the mass in stars to the total mass in massive spheroidal galaxies yield average values of @xmath22 ( @xcite ) , a value of @xmath23 , close to the cosmic ratio between baryon and total mass density , is likely to be more appropriate for proto - galaxies before a large fraction of the initial gas mass is swept off by the combined action of supernova explosions and of quasar feedback . the sz effect in eq .
( [ deltat ] ) shows up on small ( typically sub - arcmin ) angular scales .
quasar - driven blast - waves could inject into the ism an amount of energy several times higher than the gas binding energy , thus producing larger , if much rarer , sz signals .
is the absolute value of @xmath24 [ eq . ( [ ssz ] ) ] ]
impressively many lines of evidence converge in indicating that the formation of super - massive black holes powering nuclear activity is intimately linked to the formation of their host galaxies ( @xcite .
these include : * the discovery that massive dark objects ( mdos ) , with masses in the range @xmath25@xmath26 and a mass function matching that of baryons accreted onto black holes during the quasar activity ( @xcite ) , are present in essentially all local galaxies with a substantial spheroidal component ( @xcite ) ; * the tight correlation between the mdo mass ( m@xmath27 ) and the velocity dispersion ( @xmath28 ) of stars in the host galaxy ( @xcite ) , the mass of the spheroidal component ( @xcite ) , and the mass of the dark halo @xcite ; recently @xcite have found that the m@xmath29@xmath28 correlation is already present at redshift up to @xmath30 ; * the correspondence between the luminosity function of active star - forming galaxies at @xmath31 , the b - band luminosity function of quasars , and the mass function of dark halos at the same redshift ( @xcite ) ; * the similarity between the evolutionary histories of the luminosity densities of galaxies and quasars ( e.g. @xcite ) . as a consequence , the evolutionary histories of both active nuclei and spheroidal galaxies ( or galactic bulges )
can only be understood if we properly allow for their mutual feedback .
as discussed by @xcite ) , the mutual feedback between galaxies and quasars during their early evolutionary stages may indeed be the key to overcome some of the crises of the currently standard scenario for galaxy evolution .
one ingredient of the scenario proposed by @xcite are powerful quasar driven outflows , carrying a considerable fraction of the quasar bolometric luminosity , @xmath32 , increasing as @xmath33 ( for emission close to the eddington limit ) . for large enough galaxies ( @xmath34 ) , the interstellar medium ( ism ) is eventually swept out , and , correspondingly , the star - formation is switched off , on shorter timescales for more massive objects .
the outflows are expected to be highly supersonic , and therefore liable to induce strong shocks , transiently heating the interstellar gas to temperatures exceeding the virial value . a black - hole ( bh ) accreting a mass @xmath35 with a mass to radiation conversion efficiency @xmath36 releases an energy @xmath37 .
we adopt the standard value for the efficiency @xmath38 and assume that a fraction @xmath39 of the energy is fed in kinetic form and generates strong shocks turning it into heat .
actually the fraction , @xmath40 , of energy going into heating of the interstellar gas may be a function of the bh mass , or of the bolometric luminosity of the active nucleus ( agn ) .
for example , @xcite argue that , based on the model for agn - driven outflows by @xcite , for emission at the eddington limit , the fraction of bolometric luminosity released in kinetic form increases as @xmath33 .
this conclusion is , however , model - dependent and we prefer the simpler assumption of a constant value for @xmath40 , that may be viewed as an effective , luminosity weighted value . as discussed by @xcite , @xmath41 can account for the pre - heating of the intergalactic medium in groups of galaxies .
using the recent re - assessment by @xcite of the well known correlation between the bh mass and the stellar velocity dispersion @xmath42 we get @xmath43 the amplitude of the sz dip in the rayleigh - jeans region due to quasar heating of the gas is then estimated as : @xmath44 following @xcite , we adopt an isothermal density profile of the galaxy . the virial radius , encompassing a mean density of @xmath45 , is then : @xmath46 corresponding to an angular radius : @xmath47 where @xmath48 is the angular diameter distance .
the `` flux density '' , which in the rayleigh - jeans region is actually negative , associated to the sz effect is given by ( neglecting relativistic corrections ) : @xmath49 where @xmath50 and @xmath51 with @xmath52 .
the counts of quasar - driven sz signals , and the corresponding small scale fluctuations , can be estimated from the evolving b - band luminosity function of quasars , @xmath53 , relating the total amount of energy released , @xmath54 , to the b - band luminosity , @xmath55 .
if the quasar radiates at the eddington limit @xmath56 , where @xmath57 is the bolometric correction and @xmath58 is the quasar lifetime for @xmath38 .
the lifetime , @xmath59 , of the transient sz effect due to the quasar - driven blast - waves is approximately the time for the shock to reach the outer border of the host galaxy . using the expression ( @xcite ) for the evolution with time of the radius of a self - similar blast - wave expanding in a medium with an isothermal density profile , @xmath60
, we have : @xmath61 the number density of sz sources per unit interval of sz `` flux '' @xmath24 can then be estimated as @xmath62 where @xmath63 is the blue luminosity of a quasar at redshift @xmath64 causing an sz flux @xmath24 , and @xmath65 is the duration of the optically bright phase of the quasar evolution .
figure 1 shows , as an example , the predicted counts at 30 ghz , as a function of @xmath24 , obtained adopting the exponential model for the evolving luminosity function of quasars derived by @xcite for an einstein - de sitter universe , an hubble constant of @xmath66 , and an optical spectral index of quasars @xmath67 ( @xmath68 ) .
the parameters have been set at @xmath69 , @xmath39 , @xmath70 , @xmath71 .
as discussed by @xcite quasars can be used as effective signposts for massive spheroidal galaxies in their early evolutionary phases ; this allows us to bypass the uncertainties in the normalization of the primordial perturbation spectrum affecting estimates based on the various versions of the press & schechter formalism . according to @xcite ,
the mass of their dark - matter halo , @xmath72 , is related to the mass of the central black - hole by : @xmath73 under the assumption that quasars emit at the eddington limit , the number density of sources with gas at virial temperature , producing sz dips of amplitude given by eq .
( [ deltat ] ) , can be straightforwardly related to the quasar luminosity function @xmath53 by an equation analog to eq .
( [ phisz ] ) , replacing @xmath59 with the gas cooling time , @xmath74 .
the counts shown in fig . 1
refer to the same cosmological model and the same lifetime of the quasar optically bright phase as above and to @xmath75 .
as for the cooling time , which is rather uncertain , we have made the extreme assumption that @xmath76 .
thus the counts plotted have to be taken as upper limits .
the poisson fluctuations are straightforwardly computed from the counts ; since the latter are very steep , the results are insensitive to the upper flux density cutoff . to estimate fluctuations due to clustering we adopted the spatial correlation function of quasars estimated by @xcite for an einstein - de sitter universe ( @xmath77 , with @xmath78mpc , constant in comoving coordinates ) , cut - off at @xmath79 .
the power spectra of fluctuations due to both the gas in virial equilibrium and heated by the quasar feedback are compared , in fig . 2 , with the recent measurements of arcminute scale fluctuations at 30 ghz by the cbi ( @xcite ) and bima ( @xcite ) experiments .
we stress again that our estimated of signal from the gas at the virial temperature is based on an over - simplified treatment .
in fact , the thermal history of the protogalactic gas is quite complex : only a fraction of it may be heated to the virial temperature ( @xcite ) ; cooling may be relatively rapid in the densest regions ; on the other hand , significant heating may be provided by supernovae and by the central active nucleus . also , we expect that ( @xcite ) , particularly in the case of the most massive halos virializing at high @xmath64 , the quasar feedback removes the residual interstellar gas on a timescale ranging to 0.5 to a few gyrs . thus , again , the results shown in fig . 2 are ( possibly generous ) upper limits .
evidences of statistically significant detections at 30 ghz of arcminute scale fluctuations in excess of predictions for primordial anisotropies of the cosmic microwave background ( cmb ) and also in excess of the estimated contamination by extragalactic radio sources have recently been obtained by the cbi ( @xcite ) and bima ( @xcite ) experiments .
it may be noted that , although the cbi group applied a quite elaborated treatment of radio - sources , the sensitivity of their 31 ghz point source observations does not really guarantee proper allowance for this foreground .
their strategy comprised pointed 31 ghz observations with a @xmath80 detection limit of @xmath81mjy , of all nvss sources with @xmath82mjy and a direct survey reaching a flux limit of @xmath83mjy , over an area of 1.8 square degrees .
if the slope of their 31 ghz counts keeps constant down to flux densities several times lower than @xmath84 , the poisson fluctuations due to sources below the detection limit amount to @xmath85k , and may entirely account for the measured signal .
thus , their excess signal ( 15-@xmath86k in the multipole range @xmath874000 ) may be more conservatively interpreted as an upper limit .
the bima group ( @xcite ) carried out a vla survey at 4.8 ghz of their fields down to a flux density limit of @xmath88jy to identify and remove point sources . in this case , the residual contamination is indeed likely to be small , at the cost of a loss of sensitivity . if indeed the detected signal can not be attributed to extragalactic radio sources , its most likely source is the thermal sz effect ( @xcite ) .
the sz within rich clusters of galaxies has been extensively investigated ( @xcite ) .
the estimated power spectrum was found to be very sensitive to the normalization ( @xmath89 ) of the density perturbation spectrum .
a normalization @xmath90 , somewhat higher than indicated by other data sets , is apparently required to account for the cbi data .
high resolution hydrodynamical simulations of structure formation ( @xcite ) have highlighted the presence of substantial sz structure on sub - arcmin scales .
we have carried out an analytical investigation of the sz signals associated to the formation of spheroidal galaxies .
we find that they are potentially able to account for the bima results .
proto - galactic gas heated at the virial temperature and with a cooling time comparable with the expansion time may provide the dominant contribution ; in this case we expect clustering fluctuations of amplitude comparable to poisson fluctuations , although with a flatter power spectrum .
potentially detectable sz signals can also be produced by strong feedback from the central active nucleus .
it is , in fact , becoming increasingly clear that the interplay between star - formation and nuclear activity plays a key role in shaping the early evolution of both the host galaxy and the active nucleus .
the enormous amounts of energy that the active nuclei ( agns ) release to their environment , not only in radiative but also in kinetic form , can strongly influence the gas both in the host galaxy and in the surrounding intergalactic medium , with conspicuous effects on x - ray emission and on the sz signal .
a flaring agn can drive an energetic shock , causing a transient , strong rise of the average gas pressure and , consequently , of the amplitude of the sz effect .
these signals are much rarer but can individually reach substantially higher values . high frequency surveys with sub - arcmin resolution can directly detect individual sz sources .
we expect a surface density of @xmath91 at @xmath92mjy . of course , the counts at these relatively bright fluxes are likely dominated by sz effects in clusters of galaxies , which however should be easily distinguishable because of their much larger angular size and much lower redshift .
we thus conclude that the sz effect is an effective probe of the thermal state of the gas , of its evolution and , in particular , of the agn feedback .
the main parameter governing the global effect on the ambient medium of the energy released by agns is the ratio , @xmath93 , between the amount of energy deposited in the medium and the binding energy of the receiving gas ( @xcite ) .
@xmath93 can plausibly be @xmath94 for very massive galaxies and small groups , but is @xmath95 for rich clusters .
qualitatively different effects are therefore expected in the two cases .
if @xmath96 the energy injection heats the gas and partially ejects it out of the potential well .
this may lead directly to a correlation between the black - hole mass and the velocity dispersion of stars ( or the bulge mass ) with slope and normalization in nice agreement with the data .
also , the ensuing increase of the gas temperature and decrease of the x - ray luminosity ( because of the decreased gas density ) may explain the deviation of the observed @xmath97@xmath98 from predictions of the hierarchical clustering scenario , in the presence of gravitational effects only .
ongoing and forthcoming high resolution surveys in the cm / mm region may directly test this scenario .
clearly the sz effect is a much more effective probe of the thermal energy content of the plasma than the x - ray emission .
the former is in fact an almost perfect calorimeter ( @xcite ) , being directly proportional to the thermal energy . on the contrary
, x - ray emission is very sensitive to the details of the density distribution of the emitting gas .
cavaliere a. , vittorini v. , 1998 , in the young universe : galaxy formation and evolution at intermediate and high redshift , eds .
s. dodorico , a. fontana , & e. giallongo , asp conf .
146 , p. 26 | the sunyaev - zeldovich ( @xcite ; sz ) effect has proven to be an extremely powerful tool to study the physical and evolutionary properties of rich clusters of galaxies .
upcoming sz experiments , with their much improved sensitivity and angular resolution , will provide unique information also on phases of galaxy evolution characterized by the presence of large amounts of hot proto - galactic gas .
we present a preliminary analysis of the sz signals that can be expected at the collapse of the proto - galaxy when , according to the standard scenario , the gas is heated at its virial temperature , and during episodes of strong energy injections from the active nucleus .
the contributions of such signals to excess power on arc - minute scales recently found by cbi and bima experiments are briefly discussed .
address = inaf , osservatorio astronomico di padova , vicolo dellosservatorio 5 , i-35122 padova , altaddress = sissa , via beirut 4 , i-34014 trieste address = iasf - cnr , via gobetti , 101 , i-40129 bologna address = dipartimento fisica , ii universit di roma , via ricerca scientifica 1 , 00133 roma address = sissa , via beirut 4 , i-34014 trieste address = inaf , osservatorio astronomico di padova , vicolo dellosservatorio 5 , i-35122 padova , altaddress = sissa , via beirut 4 , i-34014 trieste address = dipartimento fisica , ii universit di roma , via ricerca scientifica 1 , 00133 roma address = dipartimento di fisica , universit di milano , via celoria 16 , i-20133 milano address = inaf , osservatorio astronomico di trieste , via g.b .
tiepolo 11 , i-34131 trieste |
stimulated brillouin scattering ( sbs ) is a nonlinear and self - amplifying interaction between guided optical waves and hypersound waves in waveguides @xcite .
the scattering of light from mechanical vibrations was first predicted by brillouin in 1922 @xcite ; the stimulated version of this process was first experimentally demonstrated @xcite shortly after the invention of the laser by chiao et al . and
has since been used successfully to characterize materials at hypersonic frequencies @xcite . following the general trend of miniaturization ,
sbs has been studied and applied in ever smaller structures such as nano - structured optical fibres @xcite and even waveguides integrated on a chip @xcite . in these systems
sbs ceases to be a bulk effect and surface effects strongly come into play most prominently radiation pressure appears as a second major interaction process in addition to the bulk photoelastic effect . due to this additional coupling as well as the very tight mode confinement ,
high sbs - gains can be achieved in integrated waveguides @xcite , which makes them extremely interesting for signal processing applications @xcite as well as narrow linewidth light sources @xcite .
recently the idea of coherent phonon generation in phonon - lasers has gained considerable interest @xcite .
however , all these potential applications require or at least greatly benefit from the extremely narrow linewidth of the sbs process ; line width broadening is highly detrimental in many cases @xcite .
this motivates our investigation of the inhomogeneous line broadening due to structural variations of the waveguide and potential connections to the sbs - gain itself .
the finite phonon life time ( or equivalently in backward sbs : the phonon propagation length ) defines an intrinsic line width of the sbs - resonance .
however , it has been known for some time that various imperfections of the sbs - system can distort the resonance and inhomogeneously broaden it . some of these imperfections such as insufficient acoustic guidance in combination with moderate optical mode confinement @xcite ( one possible explanation for sbs broadening in optical fibres , but somewhat debated in this context ) or simply the finite length of suspended nanowires @xcite are inherent to some waveguide designs and can not be avoided .
in addition , local deviations from the intended ideal geometry can always lead to local variations in the stokes shift . as a result ,
the total sbs - response of a non - ideal waveguide is the convolution of the intrinsic resonance and the stokes frequency distribution along the waveguide .
incidentally , the geometry - dependence of the local stokes shift has been used to measure the homogeneity of optical fibres @xcite and to increase the sbs - threshold @xcite . however , to date no connection has been made between the sbs - gain and the sensitivity of the stokes shift , although this is suggested by the close connection between sbs and opto - mechanics @xcite , in which in turn the coupling strength is known to be closely related to the deformation - sensitivity of the optical subsystem @xcite . ) and an optical probe ( superscript @xmath0 ) excite a sound wave inside a waveguide that is assumed to be smooth , but whose geometry is smoothly and slowly perturbed along the axis of propagation , the @xmath1-axis .
other cases studied in this manuscript include forward sbs and inhomogeneous strain fields ( due to temperature gradients or lattice mismatch ) instead of the sketched geometry variations . , scaledwidth=80.0% ] the aim of this paper is to understand the impact of smooth and slowly varying perturbations of the waveguide structure ( as sketched in fig . [
fig : system ] ) .
we include variations of the waveguide geometry and internal strain fields that change smoothly along the waveguide and on a length scale greater than the acoustic decay length .
this is opposed to surface roughness , which is usually a major concern in fabrication , since it causes linear optical propagation loss via rayleigh scattering in backward sbs an analogous scattering of mechanical waves also contributes to the acoustic loss .
in contrast , adiabatic structural variations do not scatter travelling waves and do not cause linear loss . however , they locally modify the dispersion relations and thereby influence the relative phase between the three waves participating in the sbs - process . here
, we show that one mechanism for this is through the optical dispersion relation and the sensitivity is closely connected to the sbs - gain , whereas the second mechanism is via the mechanical dispersion relation and completely independent from the sbs - gain .
we show that these two mechanisms influence forward and backward sbs quite differently and that increasing the sbs - gain either by increasing the acousto - optic coupling or reducing the mechanical damping leads to a regime where the line width is dominated by structural variations .
the total sbs - response is then reduced to the square root of the intrinsic sbs - gain .
we derive our results rigorously within the classical variant of a recent hamiltonian formulation of sbs @xcite . throughout this manuscript
we adhere to the notation and conventions introduced in our earlier paper on coupled - mode theory of sbs @xcite . in section [ sec : qualitative ] , we qualitatively describe the relation between sbs and the sensitivity of the dispersion relations and the resulting impact on the sbs - resonance .
although very useful for an intuitive understanding , this discussion lacks the quantitative precision of the formal derivation provided in section [ sec : theory ] .
the results of that section are then analytically studied in section [ sec : discussion ] , focusing on the two important special cases of backward sbs and intra - mode forward sbs .
next , we study the relative importance of the two broadening mechanisms in forward and backward sbs in section [ sec : numerics ] using a family of suspended silicon nanowires as an example and find that they can compensate each other in appropriately engineered waveguides .
section [ sec : summary ] summarises and concludes the paper .
the double - arrow highlights the opto - acoustic transition that fulfils the conservation of both energy and momentum .
the solid lines represent the dispersion relations of the ideal system and the dashed lines those of the perturbed system .
the light pink lines are parallels of the respective left side optical branches and indicate which acoustic wave number is phase matched .
note that the magnitude of @xmath2 is grossly exaggerated compared to @xmath3 .
, scaledwidth=50.0% ] the acoustic properties of a waveguide are modified by static structural perturbations via mechanical nonlinearities , which are usually not relevant for the dynamics of optically generated sound waves .
more precisely , there are separate contributions from internal strain fields and from variations of the cross section .
strain fields enter via the nonlinear relation between stiffness and stress : @xmath4 where @xmath5 and @xmath6 are the strain and stress tensors , @xmath7 is the conventional linear stiffness tensor and @xmath8 is a second order nonlinear stiffness tensor ( see section [ sec : theory ] ) . at the same time , internal strain fields also modify the mass density distribution , because the trace of the strain tensor corresponds to a volume dilatation .
finally , changes in the cross section geometry cause a change in the boundary conditions ( to be more specific : in the position of the boundary , not in the type of condition imposed there ) of the acoustic helmholtz equation . as an example , a slight reduction of the waveguide width might increase or decrease the acoustic frequency @xmath9 depending on whether the acoustic mode is predominantly longitudinal or transversal and on whether the boundaries are approximately free or clamped . at the same time , structural perturbations
modify the optical properties of the waveguide .
in contrast to the aforementioned acoustic frequency shift this modification is closely related to the sbs - process , which consists of two complementary nonlinear processes : the acoustic wave is excited by optical forces that are created by the beat of the optical pump and stokes waves .
the sound wave spatio - temporally varies the waveguide s optical properties and modulates the pump wave such that the stokes side band is amplified .
it is clear that the second process can only be efficient if the waveguide s dispersion relation is sensitive to small mechanical perturbations ; if it was insensitive the acoustic wave would not modulate the pump wave .
this can not be compensated by a stronger excitation of the sound wave ( neglecting irreversible forces ) , because both processes are described by the same mode overlap integral , hence are equally strong .
however , the optical dispersion relation can not distinguish between static and dynamic ( acoustic ) mechanical perturbations .
therefore , the optical properties of any waveguide that exhibits strong sbs are intrinsically sensitive to small structural perturbations introduced in the fabrication process .
more precisely , waveguides whose sbs is mainly caused by photoelasticity are predominantly sensitive to internal strain fields , caused by inhomogeneous cooling of the sample , whereas waveguides whose sbs is mainly caused by radiation pressure are predominantly sensitive to slight variations of the waveguide cross section .
the variation of the optical dispersion relation occurs as a function of @xmath1 rather than time , breaking the conservation of optical momentum .
therefore , this variation results in a change in the optical wave numbers @xmath10 and @xmath11 while leaving the optical frequencies unchanged . as a result ,
the pump frequency @xmath3 is exactly constant along the waveguide and the frequency @xmath12 of the perfectly phase - matched stokes wave varies only as a result of variations in the acoustic frequency @xmath9 .
these optical wave number variations translate to variations of the acoustic wave number through the momentum matching condition @xmath13 therefore , the optical dispersion relation sensitivity indirectly affects the acoustic frequency .
this effect has been practically applied in certain types of sbs - based fibre sensors @xcite . the total local change in the acoustic frequency
( see also fig .
[ fig : phase_matching ] ) is given by the combination of the optical wave number variations and the direct acoustic shift @xmath14 described earlier : @xmath15 it should be noted at this point that the two contributions to the total frequency change can have opposite sign and therefore cancel ( see sec . [
sec : numerics ] for an example ) . in a waveguide that is subject to inhomogeneous variations of the geometry and strain fields
the acoustic frequency variation becomes a function of the position @xmath1 within the waveguide .
it is instructive to imagine the waveguide to be composed of many short and approximately homogeneous sections ( see fig . [
fig : broadening ] ) .
each such section acts as an independent sbs - amplifier with a stokes frequency that is offset from the ideal by @xmath16 .
all other properties such as intrinsic line width and differential sbs - gain ( the sbs gain per unit length of waveguide measured in units of @xmath17 ) can be assumed to be unchanged .
the total sbs - spectrum of the complete waveguide is then simply the superposition of the individual sbs - spectra . in the limit of very short homogeneous sections , within an undepleted pump approximation and assuming a lorentzian intrinsic resonance shape , the stokes amplification on resonance is then the integral : @xmath18 ^ 2 } , \label{eqn : qualitative_superposition}\end{aligned}\ ] ] where @xmath19 represents the phonon life time and the waveguide is assumed to extend from @xmath20 to @xmath21 . within the integral
, @xmath22 is the local ( differential ) sbs power gain measured in @xmath17 ( corresponding to the quantity @xmath23 in boyd s book @xcite ) and @xmath24 is the total amplification accumulated over the total waveguide length @xmath25 ( denoted @xmath26 by boyd ) .
if the distribution @xmath16 has a width of the order @xmath27 and especially if it is lopsided , the sbs - resonance is inhomogeneously broadened into a non - lorentzian shape as indicated in fig .
[ fig : broadening ] .
so far , we have qualitatively explained the two main effects of structural variations on the sbs - resonance : perturbations to the acoustic and the optical dispersion relation and we have furthermore argued that the magnitude of the latter effect must be closely related to the sbs - gain . for practical purposes , it is probably sufficient to compute @xmath14 and @xmath28 numerically and insert them into eqs .
( [ eqn : tot_shift],[eqn : qualitative_superposition ] ) .
however , the exact relationship between optical sensitivity and sbs - gain as well as further consequences can only be elucidated by a formal derivation .
this is therefore the topic of the next section .
we assume sbs in a lossless optical waveguide that is oriented along the @xmath1-direction and provides confinement for the optical and all relevant acoustic modes , which are furthermore assumed to be only weakly damped .
mostly out of convenience we restrict ourselves to mechanically isotropic materials .
furthermore , we assume that the optical and acoustic envelopes as well as the deviations from the ideal waveguide geometry vary slowly along the @xmath1-axis compared to the acoustic decay length .
we mostly follow the notation and conventions established in our earlier paper @xcite : upper case letters refer to physical ( real - valued ) quantities , lower case letters to eigenmodes and envelope functions , and a tilde implies that the quantity is subjected to a phase transformation ( rotating frame ) .
we furthermore introduce the symbol @xmath29 for the inverse permittivity and the superscripts `` @xmath30 '' and `` @xmath31 '' to refer to the interior of the waveguide s cross section and its boundary , respectively .
we use einstein s summation convention where it helps to simplify the notation .
we formulate this problem within a classical hamiltonian framework .
to this end , we transfer a recent comprehensive quantum - mechanical description of sbs @xcite to the classical domain by replacing operators with functions and commutators with poisson brackets .
one of the main results of ref .
@xcite is the full acousto - optic hamiltonian , which reads in classical form : @xmath32 + \sum_{ij}\int { { \rm{d}}}^3 r \ \bigg [ \frac{b_i b_i}{2 \mu_0 } + \frac{d_i \eta_{ij}({{\bf{r } } } ; { { \bf{u } } } ) d_j}{2 { \varepsilon}_0 } \bigg ] \quad = \int { { \rm{d}}}z \ { { \mathcal{h}}}(z ) \ , \label{eqn : initial_hamiltonian}\end{aligned}\ ] ] where @xmath33 is the hamiltonian @xmath1-density , the hamiltonian per unit length of waveguide .
the mechanical displacement , strain and momentum density fields and the electric and magnetic induction fields are denoted with @xmath34 , @xmath35 , @xmath36 , @xmath37 and @xmath38 , respectively .
the material properties are described by the mass density @xmath39 , the stiffness tensor @xmath40 , the inverse relative permittivity tensor @xmath29 ; the electromagnetic terms furthermore involve the permittivity @xmath41 and permeability @xmath42 of vacuum .
all material constants are assumed to depend on the mechanical displacement field @xmath34 as well as its derivatives .
the first integral describes the mechanical part of the system , whereas the second integral covers both the electromagnetic part and the acousto - optic interaction term , which is implicitly contained in the dependence of the inverse permittivity @xmath43 on the mechanical deformation field and its derivatives .
the two main contributions @xcite to this dependence are @xmath44 a further term of linear order in @xmath34 the moving polarization term described in ref .
@xcite is usually too weak to be observable and is neglected here . within this framework
we describe structural perturbations as a static contribution to the total mechanical displacement field @xmath34 : @xmath45 where the superscripts `` @xmath46 '' and `` @xmath47 '' denote the static and acoustic contributions , @xmath48 is the permittivity distribution of the ideal ( perfectly fabricated , sound - free ) system and @xmath49 is the pockels tensor that describes the photoelastic effect .
( [ eqn : ao_perturb ] ) reflects the close connection between the acousto - optic interaction and the sensitivity of the optical dispersion relation regarding structural perturbations .
similarly , the mechanical stiffness and the mass density depend on the mechanical displacement field : @xmath50 here , the trace of the strain tensor expresses the volume dilatation .
the strain - dependent stiffness tensor presented in contains the leading order nonlinearity of mechanically isotropic bulk materials ; microscopically , the sixth rank tensor @xmath51 describes three - phonon scattering .
this can also be derived within the classical framework of hyperelasticity @xcite : in that approach , the stress is introduced as the strain - derivative of a free energy that is not quadratic in the strain variable . assuming a purely reversible mechanical response , this free energy is identical to the hamiltonian , whose strain - dependence can be expanded into a taylor series .
the leading ( cubic ) anharmonic strain - term is created by the sixth rank tensor @xmath51 when inserting eq .
( [ eqn : nonlinear_stiffness ] ) into eq .
( [ eqn : initial_hamiltonian ] ) . in reality
, this three - photon process might be forbidden by symmetry and a higher - order multi - phonon scattering would constitute the leader order of mechanical bulk nonlinearity .
furthermore , we have no reliable information about the order of magnitude of static strain fields across the large number of realisable integrated waveguides .
still , we have introduced this three - phonon term to illustrate how mechanical bulk nonlinearities would impact inhomogeneous broadening , although we neglect them in our analysis in section [ sec : discussion ] .
nonlinear mechanical wave processes usually can be neglected in typical sbs setups , so we only retain the static contributions to the mechanical nonlinearity : @xmath52 these terms describe the sensitivity of the acoustic dispersion relation to structural perturbations .
we now decompose the total hamiltonian @xmath1-density into the unperturbed mechanical and electromagnetic parts ( @xmath53 , @xmath54 ) , the interaction term @xmath55 and the structural perturbations @xmath56 and @xmath57 : @xmath58 where the first three terms are derived in ref .
@xcite : @xmath59 the integrals are carried out over the entire transverse plane .
the fourth term is formed in direct analogy to @xmath55 : @xmath60 and the last term is @xmath61 the difference between @xmath55 and @xmath56 is that the former results in a nonlinear coupling term of the optical and acoustic equations of motion , whereas the latter causes a linear perturbation of the wave propagation within the optical fields . the fifth term @xmath57 does not have a counterpart , because we neglected nonlinear wave effects within the propagating mechanical waves .
next , we expand the electromagnetic and mechanical fields into the eigenmode bases of the respective idealized waveguide problems @xcite @xmath62 ^ 2 \tilde{{{\bf{e}}}}^{(i ) } ; \label{eqn : ewp_opt } \\ ( \nabla_\perp + { { \rm{i}}}q \hat z)_j c_{ijkl } ( \nabla_\perp + { { \rm{i}}}q \hat z)_k \tilde
u_l = & - \rho \omega^2 \delta_{il } \tilde u_l .
\label{eqn : ewp_ac}\end{aligned}\ ] ] in both equations the wave numbers along the waveguide serve as parameters , where the optical wave number @xmath11 has to be chosen such that @xmath3 matches the frequency of the modelled pump laser and the acoustic wave number is subsequently found through eq .
( [ eqn : momentum_matching ] ) . in very extended systems such as conventional optical fibres ,
an infinite number of acoustic basis functions can contribute to the total acoustic sbs response .
in this case the treatment by beugnot et al .
@xcite is a viable alternative to our description .
however , we here restricted ourselves to nano - scale waveguides , in which the spectral separation between acoustic modes is greater than the sbs linewidth , in which only one or very few acoustic eigenmodes can be simultaneously excited by the two optical waves .
consequently , we express the optical and acoustic fields as modulations of single basis functions @xcite @xmath63 here , @xmath64 and @xmath65 are the classical counterparts of the quickly rotating envelope functions in ref .
@xcite and are related to the more common slowly varying envelope functions @xmath66 and @xmath67 via the relations @xmath68 they fulfil the poisson bracket relations @xmath69^\ast } , \ { \tilde a^{(j)}(z',t ' ) } \big\ } } = & \frac{{{\rm{i}}}\omega^{(i)}}{{{\mathcal{e}}}^{(i ) } } \delta_{ij } \delta(z - z ' ) \delta(t - t ' ) , \\ { \big\ { { \tilde b^\ast(z , t ) } , \ { \tilde b(z',t ' ) } \big\ } } = & \frac{{{\rm{i}}}\omega}{{{\mathcal{e}}}_b } \delta(z - z ' ) \delta(t - t').\end{aligned}\ ] ]
the symbols @xmath70 denote the respective modal energies per unit length of waveguide @xcite : @xmath71^\ast { \varepsilon}^{{\text{ideal}}}_{kl } \tilde e^{(i)}_l , \\ { { \mathcal{e}}}_b = & 2 \omega^2 \int { { \rm{d}}}^2 r \ \rho^{{\text{ideal}}}|\tilde{{{\bf{u}}}}|^2.\end{aligned}\ ] ] within the expansions ( [ eqn : expansion_opt],[eqn : expansion_ac ] ) the unperturbed contributions to the hamiltonian @xmath1-density simply become @xmath72 using the fact that the acoustic eigenmodes ( indexed by superscript @xmath73 ) for a fixed wave number @xmath74 form a complete function set , we can expand the static deformations : @xmath75.\end{aligned}\ ] ] in this we have introduced separate expansions for the distribution of a static strain field and the deviation of the waveguide geometry , which contains both effects of the static strain field and variations of the geometry due to fabricational ( photolithography or etching ) imperfections . in analogy to refs .
@xcite , we now introduce separate modal overlap integrals to capture the effects of photoelasticity and of boundary displacement : @xmath76^\ast e^{(j)}_l p_{klmn } \partial_m [ u^{(p)}_n]^\ast ; \label{eqn : def_q_bulk } \\
q^{{{\text{edge}}}}_{ij;p } = & \int_{{{\text{edge } } } } { { \rm{d}}}{{\bf{r } } } \ \big [ ( { \varepsilon}_a - { \varepsilon}_b ) { \varepsilon}_0 ( \hat n \times { { \bf{e}}}^{(i)})^\ast ( \hat n \times { { \bf{e}}}^{(j ) } ) - ( { \varepsilon}^{-1}_a - { \varepsilon}^{-1}_b ) { \varepsilon}^{-1}_0 ( \hat n \cdot { { \bf{d}}}^{(i)})^\ast ( \hat n \cdot { { \bf{d}}}^{(j ) } ) \big ] ( \hat n \cdot { { \bf{u}}}^{(p)})^\ast .
\label{eqn : def_q_edge}\end{aligned}\ ] ] now , by substituting the modal expansions in ( [ eqn : expansion_opt],[eqn : expansion_ac ] ) and ( [ eqn : perturb_expansion_u],[eqn : perturb_expansion_s ] ) into the interaction terms in ( [ eqn : perturb_int],[eqn : perturb_opt ] ) , and integrating over the transverse plane , we find the interactions can be reduced to the forms @xmath77^\ast b \underbrace { \big ( q^{{{\text{bulk}}}}_{12;n_b } + q^{{{\text{edge}}}}_{12;n_b } \big)}_{q_{{\text{sbs } } } } + { { \rm{c.c.}}}\label{eqn : modal_interact } \\ { { \mathcal{v}}}^{{\text{opt}}}(z ) = & \sum_n |a^{(1)}|^2 \big ( \zeta^{(n ) } q^{{{\text{bulk}}}}_{11;n } + \xi^{(n ) } q^{{{\text{edge}}}}_{11;n } \big ) + |a^{(2)}|^2 \big ( \zeta^{(n ) } q^{{{\text{bulk}}}}_{22;n } + \xi^{(n ) } q^{{{\text{edge}}}}_{22;n } \big ) , \label{eqn : modal_perturb_opt}\end{aligned}\ ] ] where @xmath78 is the index of the phase - matched acoustic mode .
the edge contribution @xmath79 to the perturbation term @xmath56 is of course identical to the expression by johnson @xcite .
finally , we analogously express the acoustic structural perturbation term @xmath80 in terms of a bulk overlap integral @xmath81^\ast [ \partial_k u^{(n_b)}_l ] f_{ijklmn } + \frac{\omega^2 \big|u_i^{(n_b)}\big|^2}{2 \rho^{{\text{ideal } } } } \bigg\ } \big(\partial_m u^{(p)}_n + \partial_n e^{(p)}_m \big),\end{aligned}\ ] ] which directly corresponds to the photoelastic perturbation overlap @xmath82 and an edge integral @xmath83 that corresponds to the electromagnetic boundary displacement overlap @xmath79 .
the explicit expression for @xmath83 is slightly involved and requires a short derivation ; both can be found in appendix [ appx : mech_edge ] .
mathematically , this term contains the edge contributions from eq .
( [ eqn : perturb_ac ] ) , the spatial derivatives of the stiffness and mass density functions .
physically , it describes the simple fact that the transverse acoustic pattern which is a standing wave in the transversal plane is detuned very much like a drumhead upon size variations of the transverse acoustic cavity . using the poisson bracket relations , expanding the @xmath1-integral of the hamiltonian density into the appropriate taylor series @xcite and truncating the expansion to first order
, we find the equations of motion : @xmath84 we now have all tools to compute the spectrum of the steady state sbs response of an imperfect waveguide using the acoustic green function @xcite . to this end , we assume a small signal approximation and a local acoustic response approximation , for which we assume that the modal envelopes @xmath85 , @xmath67 as well as the perturbation envelopes @xmath86 , @xmath87 vary slowly compared to the acoustic decay length , which is of the order @xmath88 in backward sbs and is negligible in intra - mode forward sbs . following ref .
@xcite , we introduce the inverse acoustic decay length @xmath89 as well as an additional parameter @xmath90 that describes a wave number offset between the ideally phase - matched stokes mode and the probe wave that is injected in a seeded small - signal sbs - experiment .
the poisson brackets involving the perturbation terms ( [ eqn : modal_perturb_opt],[eqn : modal_perturb_ac ] ) are : @xmath91 expressed in terms of the structural detuning parameters @xmath92 the steady state equations then follow : @xmath93^\ast a^{(2)},\end{aligned}\ ] ] with the sbs coupling parameter @xmath94 from eq .
( [ eqn : modal_interact ] ) and the powers @xmath95 and @xmath96 of the eigenmodes as introduced in ref . @xcite .
we then solve the acoustic equation using its green function : @xmath97^\ast a^{(2)}(z - z ' ) \exp\big(-[\alpha - { { \rm{i}}}\delta q^{{\text{dir}}}(z - z ' ) ] z'\big ) \big\}. \label{eqn : b_convolution}\end{aligned}\ ] ] we assumed that the optical intensities as well as the structural envelopes vary slowly on the length scale @xmath98 , so we can assume that the @xmath99 , @xmath100 , @xmath101 and the absolute values @xmath102 and @xmath103 are constant within the convolution ; @xmath90 is simply a constant .
we may therefore approximate the product of optical envelopes at the position @xmath104 as that product at position @xmath1 modulo a phase factor that expresses the relative beat @xmath105 : @xmath106^\ast a^{(2)}(z - z ' ) \approx & [ a^{(1)}(z)]^\ast a^{(2)}(z ) \exp \big ( { { \rm{i}}}[\delta \beta^{(2)}(z ) - \delta \beta^{(1)}(z ) - \kappa ] z ' \big ) , \\ \rightarrow \quad
b(z ) \approx & - \frac{{{\rm{i}}}\omega q_{{\text{sbs}}}}{{{\mathcal{p}}}_b } [ a^{(1)}(z)]^\ast a^{(2)}(z ) l(z ) ,
\intertext{\centering with } l(z ) = & \frac{1}{\alpha - { { \rm{i}}}[\delta q^{{\text{tot}}}(z ) - \kappa ] } \\
\delta q^{{\text{tot}}}= & \delta q^{{\text{dir}}}- \delta \beta^{(1 ) } + \delta \beta^{(2)}.\end{aligned}\ ] ] we can now insert this expression for the acoustic field into the optical equations . along the lines of ref .
@xcite , we then transform them into a pair of equations for the optical powers @xmath107 : @xmath108 the local sbs - gain function is given by @xmath109 ^ 2}. \label{eqn : local_gain}\end{aligned}\ ] ] within a small signal approximation , @xmath110 , this leads to a separable equation for the stokes power @xmath111 : @xmath112,\end{aligned}\ ] ] and an expression for the total stokes amplification of the waveguide of length @xmath25 : @xmath113 ^ 2 } } \label{eqn : theory_result}\ ] ] in units of db . in the simple case of @xmath1-independent aberration coefficients @xmath86 and
@xmath87 the integral in eq .
( [ eqn : theory_result ] ) becomes equal to one if the detuning @xmath90 is chosen to be @xmath114 .
this means that to an excellent approximation the acoustic frequency and thereby the sbs - resonance is simply shifted by the frequency @xmath115 in agreement with eq .
( [ eqn : tot_shift ] ) . in the case of inhomogeneous coefficients ,
the integral in eq .
( [ eqn : theory_result ] ) expresses the linear superposition of individual sbs - resonances and therefore the inhomogeneous resonance broadening in agreement with eq .
( [ eqn : qualitative_superposition ] ) .
we will now study the two practically most relevant special cases : intra - mode forward sbs and backward sbs .
qualitatively , inter - mode scattering behaves similarly to backward sbs .
first , we study the case of forward intra - mode sbs , we assume @xmath116 in this case , the sbs coupling coefficient is @xmath117 furthermore , the optical wave number perturbations are identical : @xmath118 as a result , @xmath119 , the indirect resonance broadening via the optical dispersion relation is absent in intra - mode forward sbs . strictly speaking , this result is exact only if the optical group velocity is identical at @xmath120 and @xmath3 ;
in practice this is irrelevant except for extremely dispersive optical modes , in the slow light regime near band edges .
the remaining source of structural resonance broadening is the direct mechanical contribution @xmath121
if we assume that internal strain fields are too weak to cause noticeable resonance broadening , we may assume @xmath122 that the resonance detuning can be regarded as the simple geometric effect of changing the size of a transversal acoustic cavity .
the mechanical perturbation overlaps @xmath123 and the acousto - optic overlaps @xmath124 are in principle completely independent , so there is no fundamental reason why an increase in the sbs - gain would increase the inhomogeneous broadening in the case of intra - mode forward sbs .
this is fundamentally different in the case of backward sbs and ( to a lesser extent ) in the case of general inter - mode sbs .
the optical modes that participate in conventional backward sbs are the counter - propagating partners of each other and therefore related via complex conjugation : @xmath125^\ast ; \quad { { \mathcal{p}}}^{(1 ) } = - { { \mathcal{p}}}^{(2)}. \label{eqn : modes_bsbs}\end{aligned}\ ] ] therefore , the sbs coupling coefficient and the optical wave number perturbations are : @xmath126 while the expression for the direct acoustic detuning is identical to the case of forward sbs .
the subscript indices of the individual overlap integrals @xmath82 and @xmath79 are the mode labels of the involved two optical and the acoustic mode [ see eqs .
( [ eqn : def_q_bulk][eqn : def_q_edge ] ) ] .
it is in fact the acousto - optic forward - sbs coupling of the respective acoustic and optical modes , basically the square root of the forwards sbs gain for that combination of modes .
this time , the optical wave number perturbations add up and the total stokes shift becomes : @xmath127.\end{aligned}\ ] ] the direct mechanical contribution is qualitatively identical to the previous case .
furthermore , we can assume that the predominantly longitudinal acoustic mode contributing to backward sbs is rather insensitive to slight geometry variations .
this suggests to focus on the indirect ( optical ) contribution to the broadening and ignore the direct ( acoustic ) contribution . as a result , the general result eq .
( [ eqn : theory_result ] ) then has the explicit form ( @xmath128 ) : @xmath129 the mode coefficients @xmath130 and @xmath131 in the denominator are mode overlaps that define the forward sbs gain and are identical to the coefficients @xmath132 and @xmath133 in the numerator except for a complex conjugation of one of the optical modes courtesy of eq .
( [ eqn : modes_bsbs ] ) . the general bsbs result eq .
( [ eqn : bsbs_general_result ] ) is too convoluted to interpret directly .
we therefore restrict ourselves to a very specific limiting case that nonetheless provides some insight into the general impact of geometrical variations in bsbs in nanoscale waveguides .
we assume that the acousto - optic interaction is dominated by the edge term @xmath134 , by radiation pressure .
we furthermore assume that the contributions @xmath135 of the phase matched acoustic mode to the geometry perturbations provide a good estimate for the total optical sensitivity @xmath136 .
we finally assume that the geometry variations are completely random within a certain interval . within this approximation @xmath137
is uniformly distributed in an interval @xmath138 $ ] ; other distributions of @xmath137 clearly lead to different broadened resonance shapes , but also exhibit the general trends described below .
these assumptions reduce eq .
( [ eqn : bsbs_general_result ] ) to the much simpler form @xmath139 for the case of strong broadening ( @xmath140 ) the integral approaches the limit @xmath141 and therefore the total amplification approaches the limit @xmath142 this result consists of a combination of eigenmode constants such as frequencies and mode powers , the ratio of the intrinsic forward and backward sbs gains and a term defining how the broadened sbs amplification scales with various design parameters , especially the fabrication tolerance @xmath143 .
the gain ratio is always slightly below one , as we show in appendix [ appx : gain_ratio ] , and we can simply assume it to be a constant .
the last term finally scales with the inverse of the acoustic damping , the fabrication tolerance and @xmath144 , which can be identified as being proportional to the square root of the fsbs power gain using eq .
( [ eqn : local_gain ] ) and eq .
( [ eqn : fsbs_gain ] ) .
alternatively , we can also formulate the result as scaling with the square root of the bsbs power gain by using @xmath145 both interpretations are basically equivalent , because fsbs and bsbs gain are closely related ( appendix [ appx : gain_ratio ] ) .
if we assume a waveguide with a significant sbs contribution due to radiation pressure and we assume the tolerance @xmath143 to be given , the reduction of @xmath146 either by reducing the acoustic damping parameter @xmath89 or by increasing the fsbs acousto - optic coupling @xmath144 will lead into a regime where the bsbs amplification is defined by @xmath144 .
in practice , the acoustic loss can be reduced by reducing the contact area of a nearly suspended waveguide to its substrate @xcite ; certain slot waveguide designs @xcite on the other hand allow control of the radiation pressure coupling by varying the gap size .
crystal direction .
we study the interaction between the fundamental optical mode and two symmetry - permitted acoustic modes with lowest frequency .
the acoustic mode patterns show the in - plane displacement components as arrows and the axial displacement coordinate via a color map .
the first acoustic mode ( orange frame and graph lines ) does not have any axial contributions , whereas the second acoustic mode ( blue frame and lines ) has no in - plane contributions to the displacement field .
the relative detunings are displayed in percent for a change of the waveguide width by @xmath147 . due to the extremely small acoustic wave number in fsbs only the direct acoustic perturbation process contributes to the resonance shift
, whose relative magnitude follows the relative width change exactly for mode 2 and qualitatively for mode 1 .
this demonstrates the direct acoustic response to width changes is basically given by the detuning of the transverse mechanical cavity .
the annotations in the top left panel specify the computed sbs power gains of the two acoustic modes for the extreme waveguide widths assuming a mechanical quality factor of @xmath148 . ] as the last part of our study , we performed numerical experiments on the sensitivity of the sbs frequency shift to variations of a waveguide s geometry .
we selected a family of silicon nanowires with a height of @xmath149 and a width varying between @xmath150 and @xmath151 in steps of @xmath152 .
for each such waveguide , we characterized forward and backward sbs between the fundamental optical mode and the two lowest symmetry - permitted acoustic modes @xcite using the finite - element solver comsol with an element size of @xmath153 inside the waveguide . for each step of @xmath154
, we estimated the width sensitivity by increasing the width by @xmath147 and repeating the calculation .
the resulting relative frequency shifts are plotted as solid lines in the lower left panels of fig .
[ fig : example_fsbs ] and fig .
[ fig : example_bsbs ] . finally , we added computed sbs power gains of the respective acoustic modes as annotations to the sbs frequency graphs ; they are intended to give a rough idea of the respective acousto - optic coupling . in the case of fsbs ( fig .
[ fig : example_fsbs ] ) , the difference is entirely due to the direct ( acoustic ) perturbation effect , because the acoustic wave number is basically zero .
both acoustic modes are standing waves in the acoustic resonator , formed perpendicular to the axis of light propagation , where the first mode has contributions from both in - plane components of the mechanical displacement field and the second mode only contributions from the out - of - plane component . as a result ,
the second mode is a pure standing s - wave and consequently its eigenfrequency is entirely determined by the waveguide width and the transverse speed of sound .
this leads to the exact correspondence of the variation of @xmath9 and the relative change of the waveguide width ( red crosses in fig .
[ fig : example_fsbs ] ) .
in contrast , the first mode is a hybrid of an s - wave and a p - wave , because both in - plane components contribute . as a result ,
the relationship between waveguide width and @xmath9 is not as straightforward .
this mode for a waveguide width of @xmath151 was studied experimentally in ref .
@xcite and its sensitivity has been determined to be @xmath155 at a measured brillouin frequency of @xmath156 equivalent to a sensitivity of @xmath157 in excellent agreement with fig . [
fig : example_fsbs ] . still , fig .
[ fig : example_fsbs ] demonstrates that the geometry sensitivity of fsbs is only a response to size variations of a transverse acoustic resonator .
in addition , we have also computed broadened fsbs spectra of the first mode ( computed frequency shift : @xmath158 at a width of @xmath151 ) .
[ fig : num_resonance_shapes ] shows the impact of a width variation of @xmath159 . this is a little less than two lattice constants of the silicon crystal
, it corresponds roughly to two monolayers of atoms .
[ fig : num_resonance_shapes ] also illustrates the effect of different distributions of width variations along the waveguide .
the leftmost panel shows the unperturbed spectrum assuming a mechanical quality factor of @xmath148 , which is a realistic value for certain fsbs experiments @xcite where mechanical loss is dominated by acoustic leakage .
the center panel shows the broadened spectrum assuming that the waveguide length linearly changes along the total waveguide length ( typically several millimetres to centimetres ) ; as a result the quality factor significantly drops to @xmath160 .
the rightmost panel finally shows the broadened spectrum assuming that the waveguide width fluctuates sinusoidally along the waveguide with a period that is large compared to the acoustic decay length and the optical wave length ( the exact periodicity does not matter ) . in this case , not only is the quality factor further reduced to @xmath161 , but the resonance is also deformed into a highly non - lorentzian shape with two distinct peaks .
this example shows that high effective quality factors in integrated sbs require very consistent fabrication down to the atom level . at a mechanical quality factor of @xmath148 ,
the system studied in fig .[fig : example_fsbs ] and in ref .
the three panels show the unperturbed resonance ( left panel ) , the line for the case that the width linearly varies from @xmath162 to @xmath163 over the waveguide length ( center panel ) and for the case that the waveguide width is sinusoidally undulated between the values @xmath162 and @xmath163 . in both cases
, the perturbation reduces the quality factor by about a factor of two .
, scaledwidth=70.0% ] for bsbs , the situation is slightly more complicated , because the indirect ( optical ) perturbation contributes to the total sensitivity .
therefore , we additionally determined the impact of the direct contribution by solving the acoustic problem for the perturbed geometry and the wave number derived from the unperturbed problem ( circles in fig .
[ fig : example_bsbs ] ) ; the indirect ( optical ) contribution ( squares in fig .
[ fig : example_bsbs ] ) then follows as the difference .
both studied acoustic modes are hybrid and have an axial wave length that is comparable to the waveguide s dimensions .
therefore , the acoustic frequency is roughly determined by the bulk speed of sound and the wave number .
the displacement field of the first acoustic mode has appreciable @xmath164-components , which leads to a noticeable direct ( acoustic ) sensitivity .
in contrast , the second mode is mostly insensitive to the direct perturbation effect .
both acoustic modes are very susceptible to the indirect ( optical ) perturbation effect .
for example , the second acoustic mode responds to a width variation of @xmath147 with a resonance shift of up to @xmath165 .
correspondingly , a waveguide with a width of @xmath150 would have to be manufactured to better than @xmath166 ( less than a monolayer of atoms ) along its entire length in order to maintain an overall quality factor of @xmath167 .
evidence for this problem has been observed experimentally @xcite
. one interesting detail of fig .
[ fig : example_bsbs ] is the fact that the direct and indirect contributions to the inhomogeneous broadening can be of comparable magnitude .
furthermore , they will tend to have opposite signs , because an increase in waveguide width tunes the transverse acoustic resonator to lower frequencies via the direct ( acoustic ) perturbation mechanism , while an increase in the width also leads to higher optical mode indices , corresponding to greater wave numbers and eventually higher acoustic frequencies via the indirect ( optical ) perturbation mechanism .
this suggests the very promising idea to design a waveguide structure with a deliberately _ strong _ direct sensitivity in order to cancel the inevitable indirect sensitivity . the result would be a waveguide design with greatly relaxed fabrication tolerances with respect to the waveguide width .
the general applicability of this concept to minimise the sensitivity with respect to one design parameter ( in this case the width ) is proven by the behavior of mode 1 in fig .
[ fig : example_bsbs ] .
there , the direct and indirect contributions for width variations nearly cancel for small waveguide widths . however , the presented mode is only of academic interest because of its very low sbs power gain .
finally , it should be noted that only one specific type of perturbations ( e.g. width variations , wall angle variations or height variations ) can be compensated with this technique while potentially increasing the sensitivity of the other perturbation types . crystal direction .
in contrast to fsbs ( fig .
[ fig : example_fsbs ] ) , both studied acoustic modes involve axial and in - plane components .
furthermore , bsbs features both the direct ( acoustic ) process ( with negative sign as for fsbs ) and in addition the indirect ( optical ) process with the opposite sign .
the former can be neglected for wider waveguides ; however , it is possible that both contributions cancel each other . as in fig .
[ fig : example_fsbs ] we annotated the power gain , this time assuming a quality factor of @xmath167
. the extremely low power gain of mode 1 is caused by nearly exact u cancellation of the photoelastic and moving - boundary related coupling in this structure . ]
in conclusion , we have analysed the sensitivity of the optical and acoustic dispersion relations on structural variations of a waveguide and related this to the sensitivity of the stokes shift @xmath9 .
we have shown that variations of the optical dispersion relation are a major source of inhomogeneous broadening in backward sbs and that they are intimately related to the forward sbs coupling , because slowly and smoothly varying perturbations mainly modify the phase of an optical mode , and induce negligible back - scattering .
this leads to a counter - intuitive result in sbs dominated by radiation pressure ( see section [ sec : rad_bsbs ] ) : since the inhomogeneous broadening is proportional to the forward sbs coupling ( square root of fsbs - gain ) , the maximum of a strongly broadened sbs spectrum is proportional to the ratio of the backward gain and the square root of the forward gain . as a result
the maximum of the strongly broadened response of a backward sbs system scales only with the square root of the naively expected sbs - gain .
this means that high - gain backward sbs waveguides are intrinsically sensitive to fabricational imperfections and require fabricational tolerances matched not only to the intrinsic acoustic line width but also to the opto - acoustic coupling strength .
therefore , it may be advisable in practice to sacrifice some intrinsic sbs - gain to avoid the tolerance - dominated regime , especially since this might reduce the linear or nonlinear optical loss at the same time and thereby increase the overall figure of merit .
we have furthermore shown that the indirect resonance sensitivity is absent in forward sbs and that therefore no fundamental link between the resonance sensitivity and the sbs - gain exists in this case .
nonetheless , the direct ( acoustic ) resonance sensitivity can be appreciable and has been observed in at least one nanophotonic realisation of forward sbs @xcite .
our numerical example for fsbs is in excellent agreement with this experiment .
finally , we have argued that the direct ( acoustic ) and indirect ( optical ) contributions to the resonance sensitivity tend to have opposite sign .
this allows one to design comparatively insensitive high - gain waveguides with a strong direct sensitivity carefully engineered to compensate the intrinsic indirect sensitivity .
c.w . , m.j.s . , b.j.e . and c.g.p .
acknowledge financial support from the australian research council ( arc ) via the discovery grant dp130100832 , its laureate fellowship ( b.j.e . , fl120100029 ) program , and the arc center of excellence cudos ( ce110001018 ) .
r.v.l . acknowledges the agency for innovation by science and technology in flanders ( iwt ) for a phd grant .
in this appendix , we derive the edge contribution to the mechanical perturbation term .
this is analogous to the perturbation theory developed by johnson et al . for the perturbation of the electromagnetic dispersion relation @xcite , which is based on the continuity of the transversal electric field and of the normal electric induction field across a material boundary . in the context of continuum mechanics ,
the quantities that are continuous across a material boundary are the normal projection of the stress tensor @xmath168 and the in - plane projection of the strain tensor @xmath169 where @xmath170 is the surface normal . in analogy to the electromagnetic case , the perturbation to the hamiltonian is the normal projection of the surface displacement multiplied by the mechanical energy density expressed in terms of the conserved quantities , where we restrict ourselves to isotropic materials such as glasses or homogenised polycrystalline cubic materials ( the expressions for anisotropic materials are more complex , but of the same general form ) .
we introduce an auxiliary strain quantity @xmath171_{kjlm } t^\perp_{lm } , \label{eqn : s_eff}\end{aligned}\ ] ] where @xmath172 $ ] is the mechanical compliance tensor .
this allows us to write the acoustic perturbation overlap caused by boundary displacements in a concise form : @xmath173 ( \hat n \cdot { { \bf{u}}}).\end{aligned}\ ] ] within the modal expansion employed in section [ sec : theory ] , the edge - effect of the @xmath174-th acoustic basis function of the phase matched acoustic mode @xmath78 becomes : @xmath175 ( \hat n \cdot { { \bf{u}}}^{(p)})^\ast , \intertext{with the modal ( note the lower case ) generalisation of { eq.~(\ref{eqn : s_eff } ) } } s^{{\text{eff}}}_{ij } = & \frac{1}{2 } \bigg [ \partial_i u^{(n_b)}_j + \partial_j u^{(n_b)}_i - n_i \sum_k n_k \big ( \partial_k u^{(n_b)}_j + \partial_j u^{(n_b)}_k \big ) \\
\nonumber & \quad + n_i \sum_{klpq\alpha\beta } n_k [ c^{-1}]_{kjlp } n_l n_q c_{pb\alpha\beta } \big(\partial_\alpha u^{(n_b)}_\beta + \partial_\beta u^{(n_b)}_\alpha \big ) \bigg].\end{aligned}\ ] ]
in this appendix we investigate the ratio between the acousto - optic coupling terms of forward and backward sbs .
we find that they are always of the same order and that the radiation pressure coupling is always greater for forward sbs .
inserting @xmath176^\ast = { { \bf{e}}}$ ] into the explicit definitions eqs .
( [ eqn : def_q_bulk],[eqn : def_q_edge ] ) the forward and backward coupling terms are : @xmath177^\ast \\
\nonumber & \quad + \int_{{{\text{edge } } } } { { \rm{d}}}{{\bf{r } } } \ \big [ ( { \varepsilon}_a - { \varepsilon}_b ) { \varepsilon}_0 ( \hat n \times { { \bf{e}}}^\ast ( \hat n \times { { \bf{e } } } ) - ( { \varepsilon}^{-1}_a - { \varepsilon}^{-1}_b ) { \varepsilon}^{-1}_0 ( \hat n \cdot { { \bf{d}}})^\ast ( \hat n \cdot { { \bf{d } } } ) \big ] ( \hat n \cdot { { \bf{u}}}^{(p)})^\ast , \\ q^{{\text{bsbs}}}_p = & q^{{{\text{bulk}}},{{\text{bsbs}}}}_{p } + q^{{{\text{edge}}},{{\text{bsbs}}}}_{p } = q^{{\text{bulk}}}_{12;p } + q^{{\text{edge}}}_{12;p } \\ = & { \varepsilon}_0 \int_{{\text{bulk}}}{{\rm{d}}}^2 r \ \sum_{klmn } { \varepsilon}_r^2 e_k^\ast e_l^\ast p_{klmn } \partial_m [ u^{(p)}_n]^\ast \\ \nonumber & \quad + \int_{{{\text{edge } } } } { { \rm{d}}}{{\bf{r } } } \ \big [ ( { \varepsilon}_a - { \varepsilon}_b ) { \varepsilon}_0 ( \hat n \times { { \bf{e}}}^\ast ( \hat n \times { { \bf{e}}}^\ast ) - ( { \varepsilon}^{-1}_a - { \varepsilon}^{-1}_b ) { \varepsilon}^{-1}_0 ( \hat n \cdot { { \bf{d}}})^\ast ( \hat n \cdot { { \bf{d}}}^\ast ) \big ] ( \hat n \cdot { { \bf{u}}}^{(p)})^\ast .\end{aligned}\ ] ] the only difference is the complex conjugation of the second optical mode . due to the symmetry of eqs . ( [ eqn : ewp_opt],[eqn : ewp_ac ] ) the phases of the optical as well as the acoustic modes can be adjusted such that the in - plane field components are purely real - valued and the axial components are purely imaginary .
thus , the main effect of the complex conjugation is a sign reversal in the @xmath1-component of the second optical mode . in the photoelastic coupling term @xmath82 the axial components of one mode and the transversal components of another mode
can be nontrivially combined depending on the orientation of the principal axes of @xmath49 relative to the waveguide geometry .
however , the edge term @xmath79 can be easily analysed .
first we notice that the axial component of @xmath178 is irrelevant .
next , we decompose the transversal electromagnetic fields into components normal and parallel to the waveguide surface : @xmath179 within this notation we find : @xmath180 ( \hat n \cdot { { \bf{u}}}^{(p)})^\ast , \\ q^{{{\text{edge}}},{{\text{bsbs}}}}_p = & \int_{{{\text{edge } } } } { { \rm{d}}}{{\bf{r } } } \ \big [ ( { \varepsilon}_a - { \varepsilon}_b ) { \varepsilon}_0 ( |e_\parallel|^2 - |e_z|^2 ) + ( { \varepsilon}^{-1}_b - { \varepsilon}^{-1}_a ) { \varepsilon}^{-1}_0 |d_\perp|^2 \big ] ( \hat n \cdot { { \bf{u}}}^{(p)})^\ast.\end{aligned}\ ] ] the factors @xmath181 and @xmath182 always have the same sign ( note the reversal of subscripts ) . therefore , the radiation pressure contribution to the intra - mode forward sbs coupling is always greater than the corresponding contribution to the backward sbs coupling .
this result does not apply to the photoelastic coupling .
p. dainese , p. st . j. russell , n. joly , j. c. knight , g. s. wiederhecker , h. l. fragnito , v. laude , and a. khelif , `` stimulated brillouin scattering from multi - ghz - guided acoustic phonons in nanostructured photonic crystal fibres , '' nature phys .
* 2 , * 388392 ( 2006 ) .
r. pant , c. g. poulton , d .- y .
choi , h. mcfarlane , s. hile , e. li , l. thvenaz , b. luther - davies , s. j. madden , and b. j. eggleton , `` on - chip stimulated brillouin scattering , '' opt .
express * 19 , * 82858290 ( 2011 ) .
h. shin , w. qiu , r. jarecki , j. a. cox , r. h. o. ill , a. starbuck , z. wang , and p. t. rakich , `` tailorable stimulated brillouin scattering in nanoscale silicon waveguides , '' nature commun .
* 4 , * 1944 ( 2013 ) .
s. chin , l. thvenaz , j. sancho , s. sales , j. capmany , p. berger , j. bourderionnet , and d. dolfi , `` broadband true time delay for microwave signal processing , using slow light based on stimulated brillouin scattering in optical fibers , '' opt .
express * 18 , * 2259922613 ( 2010 ) .
b. stiller , a. kudlinski , min won lee , g. bouwmans , m. delqu , j .- c .
beugnot , h. maillotte , and t. sylvestre , `` sbs mitigation in a microstructured optical fiber by periodically varying the core diameter , '' ieee photon .
* 24 , * 667669 ( 2012 ) .
c. wolff , m. j. steel , b. j. eggleton , and c. g. poulton , `` stimulated brillouin scattering in integrated photonic waveguides : forces , scattering mechanisms and coupled mode analysis , '' phys .
a * 92 * , 013836 ( 2015 ) .
s. g. johnson , m. ibanescu , m. a. skorobagotiy , o. weisberg , j. d. joannopoulos , y. fink , `` perturbation theory for maxwell s equations with shifting material boundaries , '' phys .
e * 65 , * 066611 ( 2002 ) . | we study the impact of structural variations ( that is slowly varying geometry aberrations and internal strain fields ) on the width and shape of the stimulated brillouin scattering ( sbs ) resonance in nanoscale waveguides . we find that they lead to an inhomogeneous resonance broadening through two distinct mechanisms : firstly , the acoustic frequency is directly influenced via mechanical nonlinearities ; secondly , the optical wave numbers are influenced via the opto - mechanical nonlinearity leading to an additional acoustic frequency shift via the phase - matching condition .
we find that this second mechanism is proportional to the opto - mechanical coupling and , hence , related to the sbs - gain itself .
it is absent in intra - mode forward sbs , while it plays a significant role in backward scattering .
in backward sbs increasing the opto - acoustic overlap beyond a threshold defined by the fabrication tolerances will therefore no longer yield the expected quadratic increase in overall stokes amplification .
finally , we illustrate in a numerical example that in backward sbs and inter - mode forward sbs the existence of two broadening mechanisms with opposite sign also opens the possibility to compensate the effect of geometry - induced broadening .
our results can be transferred to other micro- and nano - structured waveguide geometries such as photonic crystal fibres . |
Subsets and Splits
Data Mining Abstracts
This query searches for and retrieves 2000 documents containing the phrase 'data mining' in the abstract, providing limited but specific document samples.