Electric Conductivity From The Solution of The Relativistic Boltzmann Equation
Electric Conductivity From The Solution of The Relativistic Boltzmann Equation
Electric Conductivity From The Solution of The Relativistic Boltzmann Equation
Department of Physics and Astronomy, University of Catania, Via S. Sofia 64, I-95125 Catania, Italy and
Laboratorio Nazionale del Sud, INFN-LNS, Via S. Sofia 63, I-95125 Catania, Italy
(Dated: September 9, 2014)
We present numerical results of electric conductivity el of a fluid obtained solving the Relativistic
Transport Boltzmann equation in a box with periodic boundary conditions. We compute el using
two methods: the definition itself, i.e. applying an external electric field, and the evaluation of the
Green-Kubo relation based on the time evolution of the current-current correlator. We find a very
good agreement between the two methods.
We also compare numerical results with analytic formulas in Relaxation Time Approximation
(RTA) where the relaxation time for el is determined by the transport cross section tr , i.e. the
differential cross section weighted with the collisional momentum transfer. We investigate the electric conductivity dependence on the microscopic details of the 2-body scatterings: isotropic and
anisotropic cross-section, and massless and massive particles. We find that the RTA underestimates
considerably el ; for example at screening masses mD T such underestimation can be as large
as a factor of 2. Furthermore, we study a more realistic case for a quark-gluon system (QGP)
considering both a quasi-particle model, tuned to lQCD thermodynamics, as well as the case of a
pQCD gas with running coupling. Also for these cases more directly related to the description of
the QGP system, we find that RTA significantly underestimate the el by about a 60 80%.
I.
INTRODUCTION
Relativistic Heavy Ions Collision experiments performed by Relativistic Heavy Ion Collider (RHIC) at
BNL and Large Hadron Collider (LHC) at CERN have
produced a Little Bang, meaning that the same conditions of temperature and energy density of early universe
have been generated. A system of strongly interacting
particles above the critical temperature Tc 160 M eV
[1, 2] is expected to undergo to a phase transition from
hadron matter to Quark-Gluon Plasma (QGP) [3, 4].
The collective behaviour observed in experiments [5] and
theoretical and phenomenological studies of viscous hydrodynamics [69] and parton transport [1016] have
confirmed that QGP behaves like a fluid with a very
small shear viscosity to entropy density ratio /s close to
the lower bound 1/4 predicted by AdS/CFT [17]. This
suggests that QGP could be a nearly perfect fluid with
the smallest viscous dynamics ever observed, even less
dissipative than the ultra cold matter created by magnetic traps [18, 19]. Being the QGP created at HICs
a system far from equilibrium, the study of its transport coefficients is attracting a great interest. The has
been studied extensively [2029]. Only very recently the
electric conductivity, that represents the response of a
system to the applied electric field, has captured a significative importance in the field of strongly interacting
matter for many motivations. Electric conductivity el
can be computed on the Lattice from correlation function. On the experimental side, Heavy Ion Collisions are
able to produce very strong electric and magnetic fields
puglisia@lns.infn.it
salvatore.plumari@ct.infn.it
greco@lns.infn.it
2
that RTA is a quite good approximation for the simplest
case of massless particles interacting via isotropic crosssection even if it overestimates numerical results of about
30% for temperature T > 0.4 GeV . For anisotropic crosssection and massive particles RTA underestimates more
considerably electric conductivity than shear viscosity.
Being interested in a more realistic case and in a comparison with Lattice QCD results, we employed a quasiparticle model to take into account the thermodynamics
of strongly interacting matter [42]. We consider a system of massive quasi-quarks and quasi-gluons interacting with a pQCD-like cross-section temperature and energy dependent. We also calculate electric conductivity
in the limit of massless particles with a running coupling
constant from pQCD calculation to have an asymptotic
value of el for T Tc and to study the dependence on
a different coupling.
The paper is organized as follows. In Sec. II we briefly
recall Transport Theory and Relaxation Time Approximation to obtain the analytical formula of electric conductivity el and also Green-Kubo relation. In Sec. III
we present the setup of our numerical simulations solving
numerically the Relativistic Transport Boltzmann equation and how to compute el in this framework. In Sec.
IV we show numerical results of el for the most simple
system: massless particles interacting with isotropic and
constant cross-section. In Sec. V we consider a more
general case of massive particles and anisotropic scatterings for both el and . In Sec. VI we discuss el results
for a more realistic system of a quasi-particle model, that
describes the thermodynamics of Lattice QCD. Furthermore we show the pQCD results.
II.
ELECTRIC CONDUCTIVITY
p u
(f feq )
(4)
(5)
1
~ p~
q E
T
p0
(7)
(8)
Using of the previous equation, considering the definition of electric conductivity and generalizing to a system
of different charged particles one obtains [45, 46]:
Z
e2 X 2
d3 p p~2
e2 p~2
el =
q
q q
j feq =
3T j=q,q j
(2)3 E 2
3T E 2
(9)
where qj is the quarks charge (1/3, 2/3), j is the re~
~j = el E.
(1)
laxation
time for quarks, q is the quark density, e2 =
P
2
2
2
e
j=
q ,q qj = 4e /3. We notice that in the classical
The starting point of our calculation is the Relativistic
limit Eq. (9) simplifies to the well known Drude forBoltzmann Transport (RBT) equation that in the presmula e2 /m while in the ultrarelativistic limit becomes
ence of an external field can be written as [43, 44]:
e2 /3T .
Transport coefficients can be computed in a more gen
eral way in the framework of Linear Response The(2)
p f (x, p) + qF p f (x, p) = C[f ]
p
ory where each coefficients can be related, according to
Green-Kubo relations, to correlation functions of the corwhere f (x, p) is the distribution function, F is the elecresponding flux or tensor in thermal equilibrium. Greentromagnetic field strength tensor, C[f ] is the collision inKubo formula for electric conductivity has the following
tegral which, considering only 2 2 scatterings, can be
form:
written as follows
Z
V
V
Z Z Z
el =
dt hjz (t)jz (0)i = hjz (0)jz (0)iel (10)
4
T
T
C(x, p) =
(f1 f2 f1 f2 ) |M1 2 12 | (p1 +p2 p1 p2 )
0
2
(3)
where M is the transition matrix for the elastic process
linked to the differential cross-section |M|2 = 16s(s
3
V factor, the right hand side is obtained using the fact
that the time-correlation function for a system in thermal equilibrium is a decreasing exponential exp(t/ )
[47]. Such a behaviour has been carefully checked to be
satisfied by our numerical solution of RBT, similarly to
Ref. [28] for the shear viscosity. We notice that the
numerical evaluation of Green-Kubo correlator is a powerful method for computing, within the validity of Linear
Response Theory, transport coefficients without any kind
of approximation being the only requirement the thermal
equilibrium. In Ref. [28] we have showed the numerical
convergency of Green-Kubo method in the framework of
transport code.
III.
BOX SETUP
tot t
Ntest Vcel
(11)
(14)
(12)
results are independent on volume size. Particles are distributed uniformly in space and according to Boltzmann
distribution function, f (p) = eE/T , in momentum.
In Fig. 1 we show an example of x, y, z components
of the electric current j(t) as a function of time. In
this simulation we consider a system of massive quarks,
antiquarks and gluons (m = 0.4 GeV ) at thermal and
chemical equilibrium interacting with isotropic crosssection tot = 10 mb and we have applied an electric field
eE = 0.05 GeV /f m in the z direction. As we can see the
z component (black solid line) reaches a saturation value
while x and y components fluctuate around the equilibrium value zero.
In Fig. 2 is shown the ratio el /T as a function of the
applied electric field eE for two different temperatures
T = 0.2 GeV dark circles and T = 0.4 GeV green squares
for a system of massive particles with m = 0.4 GeV interacting via isotropic cross-section with tot = 10 mb:
el /T is independent on the applied electric field which
confirms the validity of its definition. Dashed lines are
RTA predictions of Eq. (9). As we can see from Fig.
2 thermal fluctuations affect the uncertainties on electric conductivity because a greater temperature produces
greater fluctuations in the saturation value of electric current. The increasing of the electric field has the effect of
developing a more stable electric current that is easily
noticeable in the decreasing of error bars. However, the
electric field cannot be increased arbitrarily because one
has to guarantee the linear response of the system: with
a very high value of eE the system could not reach any
equilibrium value of electric current and the definition
itself of el becomes meaningless. We have checked the
correct behaviour of electric current for several value of
electric field, temperature and cross-section presented in
this work.
The other method we used to compute el is the evaluation of Green-Kubo formula Eq (10). The correlation
function hjz (t)j(0)i can be written as follows:
*
+
Z Tmax
1
hjz (t)j(0)i =
lim
dt jz (t + t )j(t ) =
Tmax Tmax 0
(13)
1
NTmax
NTmax
X
j=1
(15)
where Tmax is the maximum time chosen in our simulations, NTmax = Tmax /t represents the maximum number of time-steps and it = t, while h i denotes the
average over events generated numerically. In Fig. 3
we plot an example of Green-Kubo correlation functions
normalized to the initial value hjz (0)jz (0)i for a system
of massless quarks and gluons interacting via isotropic
cross-section (tot = 3 mb) for several temperatures in
the range T = 0.1 0.6 GeV . Correlation functions
hjz (t)jz (0)i behave like decreasing exponential exp(t/ )
as it should be for a system in thermal equilibrium. As
<jz(t)jz(0)> / <jz(0)jz(0)>
0,02
-3
j(t) [e fm ]
0,015
0,01
jz
jx
jy
0,005
0
0
20
40
60
t [fm/c]
80
-1
/T
-2
10
-3
-4
10 0
0,02
0,1
10
-1
2
t [fm/c]
T = 0.2 GeV
T = 0.4 GeV
10
T = 0.1 GeV
T = 0.2 GeV
T = 0.3 GeV
T = 0.4 GeV
T = 0.5 GeV
T = 0.6 GeV
IV.
10
100
10
0,12
(17)
5
-1
10
10
-2
10
-3
-3
10
Green-Kubo
RTA
E field method
-2
el/T
el/T
Gree-Kubo
RTA
E field
0,2
0,4
0,6
10 0,1
0,8
1
mD [GeV]
T [GeV]
Green-Kubo
RTA
-1
10
[GeV ]
FIG. 4. Electric conductivity el /T as a function of Temperature T for a system of massless particles interacting via
isotropic scatterings (tot = 3 mb). Circles are Green-Kubo
results while blue squares represent the E-field method results. Red dashed line is Relaxation Time Approximation
Eq. (17).
10
-2
10 0,1
1
mD [GeV]
10
where h(a) = 4a(1 + a) (2a + 1)ln(1 + 1/a) 2 and
a = m2D /s. For mD the function h(a) 2/3 and
tr = (2/3)tot we recover the isotropic limit, see the
previous Sec., while for finite value of mD the function
h(a) < 2/3.
The transport relaxation time in this case can be written as:
X
ij
1
i = tot hvrel h(a)i (q + q + g )
hj ijtr vrel
tr,i
=
j
(20)
K2 m
,
beT
We performed
m 2
T
4
/RTA
(el/T)/RTA
0
0,1
1
mD [GeV]
10
simulations of a system of quarks, antiquarks and gluons with m = 0.4 GeV , that is the same value obtained in quasi-particle model in the range of temperature T = 0.2 0.4 GeV . However at this point it represents only a systematic study: in the next section we will
present quasi-particle model results with temperature dependent masses for quarks and gluons. In Fig. 5 we show
el /T as a function of mD for a system of quarks and gluons interacting via the same cross-section described Eq.
(18) with tot = 10 mb and T = 0.4 GeV : open circles
represent Green-Kubo results while blue squares are obtained using the E-field method (eE = 0.05 GeV /f m);
red dashed line is RTA using Eq. (20) for the relaxation time; dotted line represent the isotropic limit, i.e.
tr = 32 tot . We can see that RTA underestimates numerical results of el by a factor of about 40% already at
mD = 1 GeV while in the isotropic limit mD > 5 GeV it
slightly overestimates numerical estimations as we know
from the simple case of massless particles of the previous
section.
The behaviour of el /T as a function of mD is similar to the one already observed for shear viscosity in
the case of massless gluons in [28]. We show for the
corresponding plot in Fig. 6 where are shown results
for the same system studied for electric conductivity.
We compute shear viscosity using Green-Kubo relation = V /T h xy (0)2 i , as already done for single component system [25, 27, 28]. In Fig. 6 open circles are
Green-Kubo results, dashed line is RTA and dotted line
represents the isotropic limit. As found in [28] RTA is
a good approximation for only for isotropic scatterings
while for mD < 1 GeV it underestimates Green-Kubo
results by a factor of 25%.
If we look in details RTA estimations for both trans-
2s
s
m2D s + m2D
(21)
48 2
h
(11Nc 2Nf ) ln TTc
Ts
Tc
i2
(22)
10
QP Green-Kubo
pQCD E-field method
QP E-field method
2,5
-1
el/T
(el/T) / RTA
10
QP Green-Kubo
pQCD E-field method
-2
10
QP E-field method
pQCD RTA
QP RTA
2
1,5
1
0,5
-3
10
T/Tc
FIG. 8. Electric conductivity el /T as a function of T /Tc .
Dark open circles are Green-Kubo results for QP model, red
open squares are QP model results computed with the E-field
method, violet open diamonds are pQCD results calculated
with the E-field method; red line and violet line are RTA
predictions respectively for the QP model and pQCD case.
Symbols are Lattice data: grey squares [59], violet triangles
[60], green circle [61], yellow diamonds [62], red diamonds[63]
and orange square[64].
0
1
1,5
2,5
T/Tc
3,5
VII.
CONCLUSION
8
formula of el is a quite good approximation for massless
and massive particles interacting with isotropic cross section. We observe that it overestimates numerical results
by about 30% for temperature T > 0.5 GeV . For massive
quarks, antiquarks and gluons interacting via anisotropic
cross-section we have calculated el and shear viscosity
showing the differences between the two transport coefficients: RTA underestimates both el and . However
we show that RTA prediction is worse for el than for
meaning that relaxation time el is not exactly equal to
.
We find two main results. The first is that GreenKubo and the E-field method to estimate the electric
conductivity agree quite well in all the range of temperature explored that are those relevant for the hot QGP
physics. The second result is that the RTA prediction of
el based on relaxation time given by the transport crosssection always underestimates the conductivity. Such
a discrepancy increases for more forward peaked crosssection (small mD ). We find for mD T that such a
discrepancy can be about a factor of 2. Instead for more
realistic description with running strong coupling we find
a discrepancy of about 60% for both the pQCD and QP
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
description.
Our results is of wide impact and shows that when a
microscopic description is employed to estimate the el
of the QGP, or of hadronic matter, if the tr is used then
one may expect to significantly underestimate el . Our
results is quite general but it has been found considering
only a system that scatters with 2 2 collisions. It is
very interesting to have similar information also when
2 3 collisions are included as in BAMPS [27] and also
to compute bulk viscosity [65] and heat conductivity [66]
that are the other two transport coefficients of interest
in an expanding quark-gluon plasma.
VIII.
ACKNOWLEDGMENTS
9
[36] S. Turbide, R. Rapp and C. Gale, Phys. Rev. C 69 (2004)
014903.
[37] O. Linnyk, W. Cassing and E. Bratkovskaya, Phys. Rev.
C 89 (2014) 034908.
[38] M. Ruggieri, F. Scardina, S. Plumari, and V. Greco,
Phys. Rev. C 89, 054914 (2014).
[39] F. Scardina, M. Colonna, S. Plumari, V. Greco,
Phys.Lett. B 724 296-300 (2013).
[40] S. Plumari, V. Baran, M. Di Toro, G. Ferini and V.
Greco, Phys. Lett. B689 (2010) 18.
[41] V. Greco, M. Colonna, M. Di Toro, and G. Ferini, Prog.
Part. Nucl. Phys. 65, 562 (2009).
[42] S. Plumari, W. M. Alberico, V. Greco and C. Ratti, Phys.
Rev. D84 (2011) 094004.
[43] K. Yagi, T. Hatsuda, Y. Miake, Quark-Gluon Plasma:
from big bang to little bang, Cambridge University Press,
2005.
[44] C. Crecignani and G. M. Kremer, The Relativistic Boltzmann Equation: Theory and Applications, Boston; Basel;
Berlin: Birkhiiuser, 2002.
[45] D. Fernandez-Fraile and A. Gomez Nicola, Phys. Rev. D
73 (2006) 045025.
[46] A. Puglisi, S. Plumari and V. Greco, arXiv:1407.2559.
[47] L. D. Landau, E. M. Lifshits and L. P. Pitaevskij, Fisica
Statistica - Parte Prima, Editori Riuniti (1986).
[48] W. Cassing, O. Linnyk, and T. Steinert, Phys. Rev. Lett.
110, 182301 (2013)
[49] B. Zhang, M. Gyulassy, C.M. Ko, Phys. Lett. B455 45
(1999).
[50] Z. -W. Lin, C. M. Ko, B. -A. Li, B. Zhang and S. Pal,
Phys. Rev. C 72 (2005) 064901
[51] B. Zhang, M. Gyulassy and C. M. Ko, Phys. Lett. B 455,
45 (1999).