a

r
X
i
v
:
1
3
1
2
.
2
0
3
2
v
1


[
h
e
p
-
t
h
]


6

D
e
c

2
0
1
3
USTC-ICTS-13-17
Non-Abelian Berry phase in a semi-classical description of massive Dirac fermions
Jiunn-Wei Chen,
1
Jin-yi Pang,
2
Shi Pu,
1, 2
and Qun Wang
2
1
Department of Physics, National Center for Theoretical Sciences,
and Leung Center for Cosmology and Particle Astrophysics,
National Taiwan University, Taipei 10617, Taiwan
2
Interdisciplinary Center for Theoretical Study and Department of Modern Physics,
University of Science and Technology of China, Hefei 230026, China
We derive a semi-classical eective action for massive Dirac fermions in electromagnetic elds.
We present two approaches to the eective action the path integral approach and the Dirac
Lagrangian approach. The non-Abelian Berry phase structure emerges from two helicity states of
massive fermions with positive energy. The pseudo-spin also emerges as a vector in SU(2) helicity
space. The continuity equations for the fermion number and the helicity current are derived. The
fermion number is conserved while the helicity charge is not conserved by anomaly. Previous results
about the coecients of the chiral magnetic eect for the fermion and axial currents in the chiral
limit can be reproduced after including the anti-fermion contributions.
I. INTRODUCTION
Berry phase is the phase factor acquired by the energy eigenstate of a quantum system when the parameters of the
Hamiltonian undergo a cyclic change [1]. The Berry potential can be regarded as the induced gauge eld in parameter
space. If the energy eigenstates are degenerate, there is an internal symmetry in the Hilbert space spanned by the
degenerating states. This will lead to the non-Abelian Berry phase or non-Abelian gauge eld [24]. Ever since its
discovery, the Berry phase structure has permeated through all branches of physics especially in condensed matter
physics, see, e.g., [5] for a review.
Kinetic theory is an important tool to study the phase space dynamics of non-equilibrium systems. In such a
classical description, it is dicult to accommodate quantum eects such as chiral anomaly which has no counterpart
in classical dynamics. Recently it was proposed that the Abelian Berry potential can be introduced into the action of
chiral fermions in electromagnetic elds to accommodate the axial anomaly in semi-classical dynamics [6, 7]. We have
shown that the Covariant Chiral Kinetic Equation (CCKE) can be derived in quantum kinetic theory from the Wigner
function in electromagnetic elds [8]. The CCKE provides a semi-classical description of the phase space dynamics
of chiral fermions with the 4-dimensional Abelian Berry monopole and is connected to the axial anomaly. The chiral
kinetic equation in 3-dimension [6, 7, 9] can be obtained from the CCKE by integrating out the zero component of the
momentum and turning o the vortical terms. The Berry phase is closely related to the Chiral Magnetic and Vortical
Eect (CME and CVE) [1012]. The CME and related topics have been extensively studied in several approaches,
including AdS/CFT correspondence [1329], relativistic hydrodynamics [3033], kinetic theories [68, 3436], lattice
simulations [3741], and quantum eld theory or other eective theories [11, 4256]. Recently there have been some
developments related to the CME in Weyl semi-metals [5759]. For a recent review of the CME/CVE and related
topics, see e.g. [60].
The path integral formulation of the eective action for chiral (massless) fermions with Berry phase was given in
Ref. [7]. The canonical approach from the Hamiltonian can also be used [9]. The eective action with the Berry
phase can also be written as the Wess-Zumino-Witten form [56, 61]. In these approaches, the o-diagonal elements
of the Berry potentials have been neglected under the adiabatic expansion such that the Berry potential is Abelian.
In this paper we derive the eective action for massive Dirac fermions in background electromagnetic elds in a
rigorous way. We use two approaches, the path integral approach and the approach of integration slow modes in the
Dirac Lagrangian. The resulting eective action is then a c-number and the Berry potentials induced are non-Abelian.
As the consequence of the non-Abelian feature, a pseudo-spin becomes a degree of freedom in the phase space, and
it behaves like a vector in the SU(2) helicity space [6265]. The equation of motions for all phase space variables
including pseudo-spins can be derived from the eective action. The continuity equations for the fermion number
and the pseudo-spin can be obtained by integrating over momentum and the pseudo-spin from the kinetic equation.
The anomalous source term in the fermion number continuity equations is vanishing for massive fermions. In the
massless limit, the Berry phase becomes Abelian, and the fermion number is conserved if the anti-fermion contribution
is included. There are always anomalous source terms in the pseudo-spin continuity equation for helicity currents
for massive and massless fermions. In the massless limit, when including anti-fermions, the pseudo-spin continuity
equation becomes that of axial charge with anomaly.
2
The paper is organized as follows. In Sec. II, we introduce the derivation of the eective action for chiral (massless)
fermions in electromagnetic elds via conventional path integral approach in some approximations. In Sec. III, we
construct an improved and more rigorous path integral approach to the eective action. The non-Abelian Berry
potentials and the pseudo-spin emerge in the formalism. In Sec. IV, we give a simpler method to obtain the eective
action from the Dirac Lagrangian by separating fast and slow modes. The kinetic equations and continuity equations
will be derived in Sec. V. Finally we conclude in Sec. VI.
II. CHARGED CHIRAL FERMION IN CONVENTIONAL PATH INTEGRAL QUANTIZATION
We consider a particle moving in a background electromagnetic eld. We denote the canonical momentum as p
c
and the mechanical momentum as p, they are related by p
c
= p+eA(x). The Hamiltonian can be expressed in terms
of the canonical momentum p
c
,
H = (p
c
eA) + e(x), (1)
where (p) is the particle energy. The equation of motion can be derived by the Hamilton equation, x = H/p
c
and p
c
= H/x, as
x =
p

p
v,
p = eE+ ev B, (2)
where E = and B = A.
The path integral quantization is based on the Hamiltonian (1). We can write the transition matrix element in
path integral,
K

= x
f
[e
iH(t
f
ti)
[x
i

[Tx(t)][Tp
c
(t)]T exp
_
i

t
f
ti
dt(p
c
x H)
_
. (3)
Note that the starting and end points of the path are xed at x(t
i
) = x
i
and x(t
f
) = x
f
. Then we have
x[p
c
= exp(ip
c
x). (4)
After completing the path integral, we resume the use of p (writing p
c
in terms of p) and obtain the Euler-Lagrange
formulation of quantum mechanics,
x
f
[e
iHt
[x
i
=

[Tx(t)][Tp(t)] exp(iS), (5)

where the action is
S =

dt [eA(x) x e(x) + p x (p)] . (6)

From the Euler-Lagrange equation we can also obtain Eq. (2).
We now consider a massless and charged chiral fermion moving in the background electromagnetic eld. The
Hamiltonian can be written as
H = [p
c
eA(x)] + e(x), (7)
where are Pauli matrices = (
1
,
2
,
3
) and we emphasize that the quantum Hamiltonian should be constructed
for the canonical momentum p
c
instead of the mechanical momentum p = p
c
eA. The transition amplitude in path
integral in is given by Eq. (3) with the Hamiltonian (7).
Since the Hamiltonian (7) is a matrix, it is necessary to diagonalize it at each point of the path. To this end we
use a unitary matrix U
(pceA)
for diagonalization
U

(pceA)
HU
(pceA)
=
_
[p
c
eA[ + e(x) 0
0 [p
c
eA[ + e(x)
_
=
3
(p
c
eA) + e(x), (8)
3
where (p) [p[. One can easily verify that
U
p
= (
+
,

) =
_
e
i
cos

2
e
i
sin

2
sin

2
cos

2
_
,
U

p
=
_

_
, (9)
where and are polar angles of p as p = [p[(sin cos , sin sin , cos ), and

are eigenstates of p satisfying

H

. The transition amplitude is

K

= lim
N

_
_
N

j=1
dx
j
dp
c
j
_
_
dx
0
x
f
[x
N

_
_
N

j=1
x
j
[e
3iHt
[p
c
j
p
c
j
[e
3iHt
[x
j1

_
_
x
0
[x
i
. (10)
So we can nally obtain the action with Berry connection from the path integral quantization for massless and charged
chiral fermions in electromagnetic eld (see Appendix A),
S =

dt [p x + eA(x) x
3
(p) e(x) A(p) p] , (11)
In deriving the above action, we have used the denition A(p) = iU

p
U
p
with U
p
given by Eq. (9), whose explicit
form is
A(p) =
1
2[p[
_
e

cot

2
e

ie

+ ie

tan

2
_
. (12)
Note that the action (11) is in a matrix form. We can expand A(p) in terms of Pauli matrices (including the
unit matrix) and obtain each component as A
0
= 1/(2[p[ sin)e

, A
1
= 1/(2[p[)e

, A
2
= 1/(2[p[)e

and
A
1
= cot /(2[p[)e

.
The action (11) is based on the condition (A4). Note that due to o-diagonal elements of A(p) in Eq. (12), the
commutator of
3
and A is non-vanishing,
[
3
, A(p)] =
1
[p[
_
0 e

ie

ie

0
_
. (13)
If we neglect the o-diagonal elements of A(p), the condition (A4) is satised automatically and leads to the action
(11). This can be made diagonal into the positive and negative helicity components [6, 7],
S

dt [p x + eA(x) x (p) e(x) a

(p) p] , (14)
where we used a

(p) = A
11/22
(p). Those helicity changing process are neglected in an adiabatic expansion treatment.
III. ACTION FOR MASSIVE FERMIONS: IMPROVED PATH INTEGRAL APPROACH
In this section, we will formulate the action for massive fermions with non-Abelian Berry phase structure. To
deal with the action of the matrix form in the path integral in the previous section, we expand the state space by
introducing the pseudo-spin (sometimes we call it spin for short throughout the paper) degree of freedom in phase
space which is a vector in SU(2) space.
We now consider a transition from an initial state [x
i
, s
i
to a nal state [x
i
, s
f
, where s
i
and s
f
denote the initial
and nal pseudo-spin states, respectively. We treat the pseudo-spin space to be an internal symmetry space, which is
independent of coordinate and momentum states, [x, s = [x[s and [p, s = [p[s. A spin state in the Dirac space
is dened as
[s

= g [ , (15)
4
where g is an element of the SU(2)SU(2) representation in the Dirac space from doubling the fundamental represen-
tation in dimension 2 (i.e. it is a 44 matrix). The reference spin states [ along an arbitrary direction with positive
( = +) and negative ( = ) polarizations are 4-dimensional vectors and satisfy the orthogonal and completeness
relations: [

and

=
[ [ = 1. One can check that the same relations also hold for [s

due to gg

= 1.
For example, the form of g can be chosen as follows,
g() = exp(i
3

3
) exp(i
2

2
) exp(i
1

3
), (16)
where = diag(, ) with = (
1
,
3
,
3
) being Pauli matrices and = (
1
,
2
,
3
) are three Euler angles. But in
later discussions, we do not adopt any concrete form of g(). Eq. (15) shows that the spin state [s

can be labeled
by . From the completeness relation for the spin states [s

, we can use the following shorthand notation for the

integral over the phase space ,

d [s s[

[s

[
=

d = const.. (17)
So in the path integral we can insert

d [s s[ at dierent space-time points along the path.

The transition amplitude from an initial state [x
i
, s
i
to a nal state [x
i
, s
f
is given by
K

= x
f
, s
f
[e
iHt
[x
i
, s
i
, (18)
where H is the Hamiltonian for Dirac fermions with mass m given by
H = p + m, (19)
with
=
_
0
0
_
, =
_
1 0
0 1
_
. (20)
In the Hamiltonian we did not include the electromagnetic eld just for simplicity, we will consider it later. The
Hamiltonian H can be diagonalized by
E
p
= U
p
HU

p
= U
p
( p + m)U

p
, (21)
where E
p
=
_
[p[
2
+ m
2
and U
p
and U

p
are unitary 4 4 matrices with U
p
U

p
= 1, U

p
is given by,
U

p
= (u
+
, u

, v
+
, v

)
= N
r
_

+

a
p

+
a
p

a
p

+
a
p

_
, (22)
with N
r
=
_
(E
p
+ m)/(2E
p
) and a
p
= [p[/(E
p
+ m). Here u
e
(e = ) are positive energy eigenstates of the Dirac
equation, while v
e
are negative energy eigenstates. The helicity states are denoted by
e
which satisfy p
e
= e
e
.
For the path integral, we can insert complete sets of coordinate and spin states at N (N will be taken in the
end) time points along the space-time path, then the transition amplitude (18) becomes
K

= lim
N

_
_
N

j=1
[dx
j
][d
j
]
_
_
x
f
, s
f
[x
N
, s
N

_
_
N1

j=1
x
j+1
, s
j+1
[e
iHt
[x
j
, s
j

_
_
x
1
, s
1
[x
i
, s
i
. (23)
Each of the amplitudes between two states can be evaluated as (see Appendix B for the details of the derivation)
I
j+1,j
= x
j+1
, s
j+1
[e
iHt
[x
j
, s
j

[dp

1
][dx

2
][dp

3
] exp [ip

3
(x
j+1
x

2
)] exp[ip

1
(x

2
x
j
)]
e
iE
p

3
t
Tr[
j+1,j
(g
j+1
)
1
U

3
U
p

1
g
j
]. (24)
5
The trace in the last line can be re-written as
Trace 1 Tr[
j+1,j
(g
j+1
)
1
(g
j+1
g
j
)]
Tr[
j+1,j
(g
j+1
)
1
(p

3
p

1
) U

3
U
p

3
g
j+1
]
exp
_
tTr[
j+1,j
(g
j+1
)
1
g
j
]
itTr[
j+1,j
(g
j+1
)
1
p A(p)[
p=p

3
g
j
]
_
(25)
where we have dened A(p) = iU

p
U
p
. Combining Eqs. (23,24,25) we derive the nal form of the amplitude
K

[Tx(t)][Tp(t)][T] exp(iS), (26)

with the action
S =

t
0
dt
_
p x E
p
+ iTr[g
1
(d/dt p A)g]
_
, (27)
and boundary conditions x(0) = x
i
, s(0) = s
i
, x(t) = x
f
, and s(t) = s
f
. Note that we have neglected all irrelevant
constants in Eq. (26) and suppress the subscripts of in the action (27).
In the adiabatic expansion, we neglect negative energy eigenstates or anti-fermions. In this case, g is 2 2 matrices
and given by Eq. (16) with replacement , and one keeps the upper-left 2 2 block of the matrix A(p). As a
2 2 matrix, we can expand A(p) as A = A
a

a
/2 with a = 0, 1, 2, 3 and
0
= 1. The action (27) becomes
S =

dt
_
p x E
p
+ iTr(g
1
g) s
a
A
a
(p) p

, (28)
where s
a
is the average of
a
over a spin state,
s
a
=
1
2
s[
a
[s =
1
2
Tr(g
1

a
g). (29)
We see that s
0
= 1/2 and s
a
(a = 1, 2, 3) are functions of . Note that the Lagrangian is given by the content of
the square bracket in the action (28) and is a functional of (x, p, , x, p,

). Using Eq. (22) and the denition for the
Berry connection or potential A(p) = iU

p
U
p
, we obtain
A
0
=
1
[p[
1
sin
e

, A
1
=
m
E
p
[p[
e

,
A
2
=
m
E
p
[p[
e

, A
3
=
1
[p[
cot e

, (30)
where and are spherical angles of p = p/[p[, e

and e

are associated univectors, we have e

= p. If we set
m = 0, i.e. the massless or chiral fermion case, we recover A
0
and A
3
(up to a factor 2 from the denition of s
a
in
Eq. (29)) for the chiral fermion in Eq. (12). The dierence between Eq. (12) and (30) is: A
1,2
,= 0 from Eq. (30)
but A
1,2
= 0 from Eq. (12). In the case of Eq. (30), we really have an Abelian Berry potential. This dierence is
rooted from dierent bases of spinors used in Eq. (12) and (30). The better way is to use the bases in Eq. (22) for
the positive energy which leads to Eq. (30).
In the presence of background electromagnetic elds, we use the canonical momentum p
c
= p + eA to label a
momentum state [p
c
instead of [p. Then all p in the above should be replaced by p
c
eA, and the conjugate
relation becomes p
c
[x = e
ipcx
. Following the same procedure as in Sect. II and recovering back to p in the end,
we nally obtain the action for massive fermions in electromagnetic elds
S =

dt
_
iTr(g
1
g) +p x + eA x e E
p
s
a
A
a
p

. (31)
The rst term of the Lagrangian can also be written in such a form
iTr(g
1
g) = iTr(g
1
g

a
)

a
= G
1
ba
s
b

a
, (32)
where G
1
ba
is dened in Eq. (C4).
6
IV. ACTION FOR MASSIVE FERMIONS: SEPARATION OF FAST AND SLOW MODES
In this section, we try to derive the action (31) directly from the Lagrangian for Dirac fermions by separating the
fast modes from the slow ones. We can rewrite the Lagrangian for massive fermions in the electromagnetic eld as
L =

[i

+ ieA

) m]
=

(i
t
H), (33)
where the Hamiltonian is given by
H = [ieA] + m
0
+ e. (34)
The wave function for the positive energy can be written in the form
=

e=1
C
e
(t)e
iEt+ipx
u
e
(p), (35)
where u
e
are the positive energy solutions given in Eq. (22). Here the phase factor and u
e
(p) correspond to fast
modes, while C
e
(t) describe slow modes. We assume that C
e
satisfy the normalization condition, [C
+
[
2
+[C

[
2
= 1.
We assume that p depends on t. Substituting the above into the Lagrangian (33), we obtain
L =

d,f=1
C

d
(t)u

d
(p)(i
t
p x + e Ae)C
f
(t)u
f
(p)
=

d,f=1
_
C

d
(t)u

d
(p)u
f
(p)(i
t
p x)C
f
(t)
+C

d
(t)C
f
(t)u

d
(p)(i
t
)u
f
(p) + C

d
(t)C
f
(t)u

d
(p)(e Ae)u
f
(p)
_
= C

(t)[i
t
p x + ev
p
Ae p A(p)]C(t) (36)
where we have used u

d
(p)u
f
(p) =
df
, u

d
(p)u
f
(p) = v
p

df
, ( p + m
0
)u
d
(p) = Eu
d
(p) and p x = i
t
(e
ipx
),
we have also used the notation C(t) (C
+
(t), C

(t))
T
. We can further rewrite Eq. (36) as
L = Tr
_
C(t)C

(t)[i
t
p x + ev
p
Ae p A(p)]
_
= Tr
_
C(t)C

1
C
1
C

(t)[i
t
p x + ev
p
Ae p A(p)]
_
Tr
_
g
1
[i
t
p A(p)]g
_
+ p x + ev
p
Ae, (37)
where we have dropped in the last line the complete time derivative term d(x p)/dt. We have inserted a constant
C

1
C
1
= 1 between C(t) and C

(t), where C
1
(t) is an arbitrary normalized column vector with C

1
(t)C
1
(t) = 1 and
is an arbitrary matrix with trace 1. We have assumed g = C(t)C

1
and g

= C
1
C

(t), one can check that g is unitary,

i.e. gg

= 1. The Lagrangian (37) gives the action (31).

V. KINETIC EQUATION WITH NON-ABELIAN BERRY PHASE
The equations of motion from the above action read (see Appendix C for detailed derivation)
x = v
p
+ p s
a

a
p = eE+ e x B,
s
a
=
abc
( p A
b
)s
c
, (38)
where we have implied a = 0, 1, 2, 3 and a, b, c = 1, 2, 3 in the rst and last line respectively, we have dened
v
p

p
E
p
= p/E
p
,
abc
is the anti-symmetric tensor with
123
= 1. We observe s
a
s
a
= d(s
2
a
)/dt = 0, so
s
2
= s
2
1
+ s
2
2
+ s
2
3
is a constant. The Berry curvature in Eq. (38) are given by

0

p
A
0
,

a

p
A
a

1
2

abc
A
b
A
c
. (39)
7
From the Berry connection (30), we obtain

0
= 0,
1
=
m
E
3
p
e

2
=
m
E
3
p
e

,
3
=
1
E
2
p
p. (40)
We see that the a = 0 component does not appear in the rst line of equations of motion (38) due to the vanishing
of
0
. We dene
a
which we will use in the continuity equations later,

a
= (
ab

p
+
abc
A
c
)
b
. (41)
Substituting Eq. (39) into the above we obtain
a
= 0 (a = 1, 2, 3) for p ,= 0, where we can check
p

a
=

abc
A
c

b
=
abc

ijk
(
i
/
b
j
)/
c
k
. On the other hand, one can use explicit expressions in Eqs. (30,40) to obtain

1
=
2
= 0. For
3
, we get
p

3
= 2m
2
/([p[E
4
p
) and
3bc

b
A
c
= 2m
2
/([p[E
4
p
) for p ,= 0. One can also verify

d
3
p
3
= 0. Therefore we nally obtain
3
= 0. If we consider massless fermions, the only non-vanishing components
are A
3
,
3
,= 0 and the Berry phase is Abelian, one can check
1
=
2
= 0 and
3
=
p

3
= 4
(3)
(p). The
appearance of the delta-function is because there is a singularity at zero momentum p = 0 in the Berry curvature for
massless fermions.
The rst two equations of (38) can be simplied as

x = v
p
+ eE+ eB(v
p
),

p = eE+v
p
eB+ e
2
(E B), (42)
where we have dened s
a

a
(a = 1, 2, 3) and

= (1 +e B) as the phase space measure. Equation (42) has
a dual symmetry under the interchange of
x p, v
p
eE, eB. (43)
We have to extend the the phase space by including the spin vector s. The phase space distribution is denoted by
f(t, x, p, s) and we assume it satises the collisionless Boltzmann equation,
df
dt
=
f
t
+ x
i
f
x
i
+ p
i
f
p
i
+ s
a
f
s
a
= 0. (44)
The invariant phase space volume element is then
d =

1
(2)
3
S
0
d
3
xd
3
pd
2
s. (45)
Since s
2
= s
2
1
+ s
2
2
+ s
2
3
= 1/4, there are only two independent variables, so we add a delta-function (s
2
s
a
s
a
) in
the phase space integral and denote d
2
s d
3
s(s
2
s
a
s
a
). Note that S
0
= is a normalization constant from the
condition S
1
0

d
2
s = 1. We dene n(t, x) and J(t, x) as the fermion number density and current respectively,
n(t, x) =

d
3
pd
2
s
(2)
3
S
0

f(t, x, p, s),
J(t, x) =

d
3
pd
2
s
(2)
3
S
0

xf(t, x, p, s). (46)

Then the continuity equation for the fermion number is (see Appendix D for the derivation)
n
t
+
x
J = e
2
(E B)

d
3
pd
2
s
(2)
3
S
0

a
s
a
f. (47)
We see that the source term in the continuity equation (47) is proportional to the anomaly quantity E B and involves

a
. For massive fermions, the source term is vanishing due to
a
= 0 (a = 1, 2, 3), so the fermion number is conserved.
The physical reason for the vanishing
a
is that the Berry phase is non-Abelian and Berry curvature is analytic at
the zero momentum due to non-zero fermion mass. This reects the fact that the pseudo-spin or the helicity for
a massive fermion varies in time, so a fermions positive helicity can be changed to negative helicity and the total
fermion number with all helicity orientations is conserved.
8
However, for massless fermions the only non-vanishing component of the Berry potentials or curvatures is along
the third direction, so the Berry phase is Abelian. The only spin component s
3
is a constant in time from Eq.
(38): s
3
= 1/2 which correspond to the positive/negative helicity. Hence the spins are not continuous variables of
phase space anymore. We denote the distribution functions for positive/negative helicity fermions as f

(t, x, p)
f(t, x, p, s
3
= 1/2). We can replace the integral over s with a sum over s
3
, i.e.S
1
0

d
2
s

s3=1/2
. The only
non-vanishing component of
a
is
3
=
p

3
= 4
(3)
(p), which is singular and behaves like a monopole at the
zero momentum. Both the Berry phase and anomaly take their roles in the non-vanishing source of the continuity
equation as follows
n
t
+
x
J =
e
2
4
2
(E B)[f
+
(t, x, p = 0) f

(t, x, p = 0)]. (48)

For equilibrium Fermi-Dirac distribution at zero temperature and nite chemical potential, we have f

(t, x, p = 0) =
1, then fermion number is conserved. If there are only positive or negative helicity fermions in the system, the above
continuity equation becomes
n

t
+
x
J

=
e
2
4
2
(E B)f

(t, x, p = 0). (49)

which is identical to Eq. (23) of Ref. [8]. Here n

and J

are the fermion number densities and currents for

positive/negative helicity fermions respectively, which are obtained by integration over momenta for

f

(t, x, p)
and

xf

(t, x, p) respectively with

= 1 (e/2)
3
B. Note that the fermions we are considering in this paper
have positive energies. If we include the contribution from anti-particles and assume equilibrium distributions with
dierent chemical potentials for positive and negative helicities, the continuity equation (48) becomes the conservation
equation for fermion number in the chiral (massless) limit or Eq. (24) of Ref. [66],
n
t
+
x
J = 0, (50)
where n and J are the net fermion number density and current (fermion minus anti-fermions). Note that the source
term is vanishing because we have made the following replacement
f
+
f
+
+

f

= f
R
+

f
R
1 (at p = 0),
f

+

f

= f
L
+

f
L
1 (at p = 0), (51)
where

f

denote the the distributions for anti-fermions with positive/negative (right-handed/left-handed) helicity,
and f
R/L
and

f
R/L
denote those for fermions and anti-fermions with right-handed/left-handed (positive/negative)
chirality respectively. They are related in the massless limit by

f

=

f
L/R
. We can also reproduce the CME current
from Eq. (46), namely, J =
B
B, where
B
is the CME coecient in Eq. (22) of Ref. [66]. A systematic way of
including fermions and anti-fermions is to work in the full Dirac space with 4-dimensional Dirac spinors.
Furthermore, we can dene the helicity density and current,
n
a
(t, x) =

d
3
pd
2
s
(2)
3
S
0

s
a
f(t, x, p, s),
J
a
(t, x) =

d
3
pd
2
s
(2)
3
S
0

s
a
xf(t, x, p, s). (52)
We then derive the continuity equation for the helicity,
n
a
t
+
x
J
a
= e
2
(E B)

d
3
pd
2
s
(2)
3
S
0

b
s
b
s
a
f +

d
3
pd
2
s
(2)
3
S
0

s
a
f, (53)
see Appendix D for the details of the derivation. We see that there are two source terms in the continuity equation
(53). The rst term is vanishing for massive fermions. The second term is from the time derivative of the spin, which
can be simplied by using Eqs. (38,42),
S
1
0

d
2
s

s
a
=
abc
S
1
0

d
2
s(

p A
b
)s
c
=
1
12
e
2
(E B)
c
A
b

abc
S
1
0

d
2
s. (54)
9
If we focus on the a = 3 component, the continuity equation (53) becomes
n
3
t
+
x
J
3
=
1
6
e
2
(E B)m
2

d
3
pd
2
s
(2)
3
S
0
1
E
4
p
[p[
f(t, x, p, s), (55)
Although the source term is proportional to m
2
supercially, the above is non-vanishing in the massless limit because
the integral is singular and behaves as 1/m
2
.
Now let us look at the alternative way of taking the massless limit, i.e. we take the limit in Eq. (53), so only s
3
is
non-vanishing as a constant of time. Then the second source term of the continuity equation (53) is vanishing. The
rst source term leads to
n
3
t
+
x
J
3
=
e
2
8
2
(E B)[f
+
(t, x, p = 0) + f

(t, x, p = 0)]. (56)

From Eq. (49) we have n
3
= (n
+
n

)/2 and J
3
= (J
+
J

)/2. Including the anti-fermion distributions and using

the replacement (51), we obtain in the massless limit,
n
3
t
+
x
J
3
=
e
2
4
2
(E B), (57)
which is actually the continuity equation for the axial current with anomaly, and n
3
and J
3
are just the chiral density
and current respectively in this limit. This can be see from the fact that the axial current can be derived from
Eq. (52), namely, J
3
= J
5
/2 =
5B
B/2, where
5B
is the CME coecient for the axial current in Eq. (23) of Ref.
[66]. We note that a systematic way of including fermions and anti-fermions is to work in the full Dirac space with
4-dimensional Dirac spinors.
VI. CONCLUSIONS
We have formulated a semi-classical kinetic description of Dirac fermions in background electromagnetic elds. We
have shown that the non-Abelian Berry phase structure and the pseudo-spin emerge in such a kinetic description. We
provide two approaches to derive the eective action for Dirac fermions moving in electromagnetic elds. In the path
integral approach, we start from the Hamiltonian for the Dirac fermions in electromagnetic elds and calculate the
transition amplitude between the initial and nal states of the pseudo-spin and coordinate. The positive energy states
are chosen to build up the degenerating space. The helicity states enter the formalism and nally make the dynamical
variables for the pseudo-spin in the action. The phase space has to be enlarged by joining of the pseudo-spin. The
non-Abelian Berry potentials in momentum space appear in the action from the diagonalization of the Hamiltonian.
We also provides an alternative and much simpler approach to the eective action from the Dirac Lagrangian. We
separate the fast and slow modes of the positive energy fermionic eld and then integrate out the fast modes. The
emerging non-Abelian Berry potentials in the eective Lagrangian are given by the fast mode spinor wave functions,
while the emerging pseudo-spins are determined by the slow mode wave functions.
The equation of motions for Dirac fermions can be obtained from the eective action which involve electromagnetic
elds and non-Abelian Berry potentials and curvatures. Besides the equation of motions for x and p, there is also the
equation of motion for the pseudo-spin for helicity states whose time variation is controlled by the Berry potentials
and is like rotation or precession. We have observed a dual symmetry in the equation of motions for x and p by
interchanges x p, v
p
eE and eB. The continuity equations for the fermion number and helicity can be
derived from the equations of motions and the kinetic equation for distribution functions. Anomalous source terms
proportional to E B appear in continuity equations and involve integrals of Berry magnetic charges and helicities.
The anomalous source term in the continuity equation for the fermion number is vanishing for massive fermions,
while it is present in the continuity equation for the helicity. For massless fermions, the Berry phase becomes Abelian
and the helicity is not a continuous phase space variable. In this case, the fermion number is conserved when taking
anti-fermions into account, while the axial charge is not conserved by the presence of anomaly. The CME coecients
for the fermion and axial currents can be obtained after including the anti-fermion contributions, same as derived
before.
This work is supported by the NSFC under grant No. 11125524 and 1221504. JWC and SP are supported in part
by the NSC, NTU-CTS, and the NTU-CASTS of R.O.C.
10
Appendix A: Conventional path integral for chiral fermions
In this appendix, we will present the conventional path integral quantization of charged massless fermions in
electromagnetic. We will derive the action (11) with the Berry phase. The transition amplitude is given by (3). We
can evaluate the amplitudes inside the parenthesis by inserting complete set of states. For the rst amplitude in the
parenthesis of Eq. (10) we evaluate as
x
j
[e
3iHt
[p
c
j
=

_
4

i=1
dx
j(i)
dp
c
j(i)
_
x
j
[U
pceA
[p
c
j4
p
c
j4
[e
i(3+e)t
[x
j4
x
j4
[U

pceA
[p
c
j3

p
c
j3
[U
pceA
[x
j3
x
j3
[e
i(3+e)t
[p
c
j2
p
c
j2
[U

pceA
[x
j2

x
j2
[U
pceA
[p
c
j1
p
c
j1
[e
i(3+e)t
[x
j1
x
j1
[U

pceA
[p
c
j

_
4

i=1
dx
j(i)
dp
c
j(i)
_
exp
_
it
_
4

i=1
p
c
j(i)

x
j(i+1)
x
j(i)
t
__
U(x
j
, p
c
j4
) exp
_
it(
3
+ e)(x
j4
, p
c
j4
)

(x
j4
, p
c
j3
)U(x
j3
, p
c
j3
)
exp
_
it(
3
+ e)(x
j3
, p
c
j2
)

(x
j2
, p
c
j2
)U(x
j2
, p
c
j1
)
exp
_
it(
3
+ e)(x
j1
, p
c
j1
)

(x
j1
, p
c
j
) exp(ip
c
j
x
j1
), (A1)
where we have used the notation x
j5
x
j
. We can use the following formula to simplify the above equation,
U

(x
j2
, p
c
j2
)U(x
j2
, p
c
j1
) exp
_
iA(x
j2
, p
c
j1
) (p
c
j2
p
c
j1
)

,
U

(x
j4
, p
c
j3
)U(x
j3
, p
c
j3
) exp
_
iA(x
j3
, p
c
j3
) (eA(x
j4
) eA(x
j3
))

, (A2)
where A(p) iU

p
U
p
are called Berry connection. So the amplitude in Eq. (A1) becomes
x
j
[e
3iHt
[p
c
j
=

_
4

i=1
dx
j(i)
dp
c
j(i)
_
U(x
j
, p
c
j4
) exp
_
it
_
4

i=1
p
c
j(i)

x
j(i+1)
x
j(i)
t
__
exp
_
it[(
3
+ e)(x
j4
, p
c
j4
) A(x
j3
, p
c
j3
) (e

A(x
j4
) e

A(x
j3
))]
_
exp
_
it[(
3
+ e)(x
j3
, p
c
j2
) +A(x
j2
, p
c
j1
) ( p
c
j2
p
c
j1
)]
_
exp
_
it(
3
+ e)(x
j1
, p
c
j1
)

(x
j1
, p
c
j
) exp(ip
c
j
x
j1
). (A3)
Here we have used
exp(itC
3
) exp(itC

A) = exp
_
it(C
3
+ C

A)
1
2
(t)
2
CC

[
3
, A]
_
exp[it(C
3
+ C

A)] (A4)
where C and C

are constants. We will see that [

3
, A] ,= 0, so we have assumed the (t)
2
term is much smaller than
the t terms in Eq. (A4).
11
For the second amplitude in the parenthesis of Eq. (10), we evaluate as
p
c
j
[e
3iHt
[x
j1
=

_
4

i=1
dx

j(i)
dp
c
j(i)
_
p
c
j
[U
pceA
[x

j4
x

j4
[e
i(3+e)t
[p
c
j4
p
c
j4
[U

pceA
[x

j3

j3
[U
pceA
[p
c
j3
p
c
j3
[e
i(3+e)t
[x

j2
x

j2
[U

pceA
[p
c
j2

p
c
j2
[U
pceA
[x

j1
x

j1
[e
i(3+e)t
[p
c
j1
p
c
j1
[U

pceA
[x
j1

_
4

i=1
dx

j(i)
dp
c
j(i)
_
exp(ip
c
j
x

j4
)U(x

j4
, p
c
j
)
exp
_
it
_
4

i=1
p
c
j(i)

j(i)
x

j(i1)
t
__
exp
_
it[(
3
+ e)(x

j4
, p
c
j4
) +A(x

j3
, p
c
j4
) (p
c
j4
p
c
j3
)/t]
_
exp
_
it[(
3
+ e)(x

j2
, p
c
j3
) A(x

j2
, p
c
j2
) (eA(x

j2
) eA(x

j1
))/t]
_
exp
_
it(
3
+ e)(x

j1
, p
c
j1
)

(x
j1
, p
c
j1
) (A5)
where we have denoted x

j(0)
= x
j1
. We have also used (A4).
We observe that U

(x
j1
, p
c
j
) exp(ip
c
j
x
j1
) of Eq. (A3) and exp(ip
c
j
x

j4
)U(x

j4
, p
c
j
) of Eq. (A5) can be combined
as,
x
j
[e
3iHt
[p
c
j
p
c
j
[e
3iHt
[x
j1

exp
_
it(
3
+ e)(x
j1
, p
c
j1
)

(x
j1
, p
c
j
) exp(ip
c
j
x
j1
) exp(ip
c
j
x

j4
)U(x

j4
, p
c
j
)
= exp
_
itp
c
j

x
j1
x

j4
t
_
exp
_
it[(
3
+ e)(x
j1
, p
c
j1
)
A(x
j1
, p
c
j
) (eA(x
j1
) eA(x

j4
))/t]
_
(A6)
Finally taking the limit N , we can write the amplitude (10) into a compact form,
K

TxTp
c
U(x
f
, p
c
f
)T exp
_
i

t
f
ti
dt [p
c
x
3
(p
c
eA) e(x)
A(p
c
eA) ( p
c
e

A)
__
U

(x
i
, p
c
i
)
=

TxTpU(x
f
, p
c
f
)T exp
_
i

t
f
ti
dt [p x + eA(x) x

3
(p) e(x) A(p) p] U

(x
i
, p
c
i
). (A7)
We can read out the action (11) from above amplitude.
Appendix B: Transition amplitude in path integral for Dirac fermions
In this appendix, we give the derivation of Eq. (24). We can insert the complete set of states as follows
I
j+1,j
= x
j+1
, s
j+1
[e
iHt
[x
j
, s
j

= x
j+1
, s
j+1
[U

p
e
iEpt
U
p
[x
j
, s
j

[dp

1
][d

1
][dx

2
][d

2
][dp

3
][d

3
]
x
j+1
, s
j+1
[U

p
[p

3
, s

3
, s

3
[e
iEpt
[x

2
, s

2
x

2
, s

2
[U
p
[p

1
, s

1
p

1
, s

1
[ x
j
, s
j
. (B1)
12
Note that coordinate and momentum states are decoupled from the pseudo-spin states, i.e. [x, s = [x[s and
[p, s = [p[s. Then we can combine the spin states and obtain
I
j+1,j
=

[dp

1
][d

1
][dx

2
][d

2
][dp

3
][d

3
]
x
j+1
[ p

3
s
j+1
[U

2
[s

3
p

3
[ x

2
e
iE
p

3
t
s

3
[ s

2
[ p

1
s

2
[U
p

1
[s

1
s

1
[ s
j
p

1
[ x
j

[dp

1
][d

1
][dx

2
][d

2
][dp

3
][d

3
]
exp[ip

3
(x
j+1
x

2
)] exp [ip

1
(x

2
x
j
)] e
iE
p

3
t
s
j+1
[U

3
[s

3
s

3
[ s

2
s

2
[U
p

1
[s

1
s

1
[ s
j
. (B2)
Here we have used the fact that is commutable with g ( because[, ] = 0) in evaluating the amplitude of e
iEpt
,
so we have
p

3
, s

3
[e
iEpt
[x

2
, s

2
= p

3
[ x

2
e
iE
p

3
t
s

3
[ s

2
.
Then we can carry out integration over [d

1
][d

2
][d

3
] to remove intermediate spin states by using Eq. (17). Here we
neglect constants from the integral

d. Then we have
I
j+1,j

[dp

1
][dx

2
][dp

3
] exp [ip

3
(x
j+1
x

2
)] exp[ip

1
(x

2
x
j
)]
e
iE
p

3
t
s
j+1
[U

3
U
p

1
[s
j

[dp

1
][dx

2
][dp

3
] exp [ip

3
(x
j+1
x

2
)] exp[ip

1
(x

2
x
j
)]
e
iE
p

3
t

j+1
[(g
j+1
)
1
U

3
U
p

1
g
j
[
j

[dp

1
][dx

2
][dp

3
] exp [ip

3
(x
j+1
x

2
)] exp[ip

1
(x

2
x
j
)]
e
iE
p

3
t
Tr[
j+1,j
(g
j+1
)
1
U

3
U
p

1
g
j
] (B3)
where we have used Eq. (15) in the second equality and
j+1
[C[
j
= Tr(
j+1,j
C) in the last one.
Appendix C: Derivation of equations of motion
In this appendix, we will derive the equations of motion (38) from the action (31). We treat the Lagrangian as the
function of (x, p, , x, p,

). We will use the notation, for example, x
i
for the i-th component of the vector x. The
Euler-Lagrange equation for x is derived as
d
dt
L
x
i
= p
i
+ e
A
i
t
+ e
A
i
x
j
x
j
,
L
x
i
= e
A
j
x
i
x
j
e
A
0
x
i
,

p
i
= e
A
i
t
e
A
0
x
i
+ e
_
A
j
x
i

A
i
x
j
_
x
j
,
= eE
i
+ e
ijk
x
j
B
k
, (C1)
13
which is the second line of Eq. (38). For the Euler-Lagrange equation for p, we obtain
d
dt
L
p
i
=
ds
a
dt
/
a
i
s
a
/
a
i
p
j
p
j
L
p
i
= x
i

E
p
p
i
s
a
/
a
j
p
i
p
j

x
i
=
p
i
E
p
+ s
a
p
j
_
/
a
j
p
i

/
a
i
p
j
_

ds
a
dt
/
a
i
. (C2)
We will evaluate ds
a
/dt using the equation of motion for . Since s
0
= 1/2, there is no ds
0
/dt = 0 term in the last
line of Eq. (C2). In order to derive the equation for , we need to dene g()/
a
properly. We dene () as
exp
_
i
1
2

a
_
g() = g[()], (0) = , (C3)
with a = 1, 2, 3. Taking derivative on
a
and setting
a
= 0 we obtain
i

a
2
g() =
g()

b
N
ba
,
G
ba


b
()

=0
. (C4)
One can prove det(G) ,= 0 so the matrix N is invertible. The the Euler-Lagrange equation for is derived as
d
dt
L

a
= i
d
dt
Tr
_
g
1
g

a
_
= G
1
ca
d
dt
Tr
_
g
1

c
2
g
_
= G
1
ca
ds
c
dt
,
L

a
= iTr
_

g
1

a
g

b
_
+ iTr
_
g
1

a
g

b
_

a
Tr
_
g
1

a
2
g
_
A
a
p
= iG
1
ca
G
1
db
Tr
_
g
1
_

c
2
,

d
2
_
g
_

b
+iG
1
ca
Tr
_
g
1
_

c
2
,

a
2
_
g
_
A
a
p
= G
1
ca

cad
(A
a
p)s
d
, (C5)
which leads to
ds
a
dt
=
abc
(A
b
p)s
c
, (C6)
which is just the third line of Eq. (38). Note that s
0
does not appear in Eq. (C6), so we have implied a, b, c = 1, 2, 3.
Substituting Eq. (C6) back into Eq. (C2), we obtain the last line of Eq. (38),
x
i
=
p
i
E
p
+ s
c
p
j
_
/
c
j
p
i

/
c
i
p
j

abc
/
a
i
/
b
j
_
+s
0
p
j
_
/
0
j
p
i

/
0
i
p
j
_
=
p
i
E
p
+
ijk
p
j

c
k
s
c
, (C7)
where
c
with c = 0, 1, 2, 3 are given by Eq. (39).
14
Appendix D: Derivation of continuity equations
In this appendix, we will derive continuity equations (47,53) for the fermion number and the pseudo-spin. We use
the notation, for example, x
i
for the i-th component of the vector x. To derive the continuity equation (47) for the
fermion number, we start from taking the time derivative of n(t, x) in Eq. (46) and using Eq. (44),
n(t, x)
t
=

d
3
pd
2
s
(2)
3

t
f +

d
3
pd
2
s
(2)
3

f
t
=

d
3
pd
2
s
(2)
3

t
f

d
3
pd
2
s
(2)
3

_
x
i
f
x
i
+ p
i
f
p
i
+ s
a
f
s
a
_
. (D1)
Using the partition formula for integrals, we can rewrite the second term into the following form
I
2
=

d
3
pd
2
s
(2)
3

_
x
i
f
x
i
+ p
i
f
p
i
+ s
a
f
s
a
_
=

x
i

d
3
pd
2
s
(2)
3

x
i
f +

d
3
pd
2
s
(2)
3
(

x
i
)
x
i
f

d
3
pd
2
s
(2)
3
(

p
i
f)
p
i
+

d
3
pd
2
s
(2)
3
(

p
i
)
p
i
f

d
3
pd
3
s
(2)
3
[

(s
2
s
b
s
b
) s
a
f]
s
a
+

d
3
pd
3
s
(2)
3
(s
2
s
b
s
b
)
(

s
a
)
s
a
f
=
J
i
x
i
+

d
3
pd
2
s
(2)
3
(

x
i
)
x
i
f +

d
3
pd
2
s
(2)
3
(

p
i
)
p
i
f
+

d
3
pd
2
s
(2)
3
(

s
a
)
s
a
f, (D2)
where we have used d
2
s d
3
s(s
2
s
a
s
a
). We have dropped the complete derivatives for the momentum and the
spin whose integrals are vanishing. Note that in the fourth line, we have recovered (s
2
s
a
s
a
) because these terms
are related to /s
a
and should be handled with care, and one should pay special attention to the the second term:
we have pulled the delta function out of /s
a
since the partial derivative of the delta function gives a s
a
which
will combine with s
a
and vanishes. Applying the equation of motion (42), we obtain the following formula to further
evaluate I
2
,

t
+
x
i

x
i
= e

B + e(
x
E) = 0,
(

p
i
)
p
i
= e
2
(E B)
p
,
(

s
a
)
s
a
=
(

p
i
)
s
a

abc
/
b
i
s
c
= e
2
(E B)
abc
(
a
A
b
)s
c
, (D3)
where we have used the Maxwell equations
x
B = 0 and
x
E +

B = 0. Finally we arrive at the continuity
equation (47) for the fermion number from Eq. (D1).
Now we give the derivation of the continuity equation (53) for the pseudo-spin. To this end we follow the same
procedure by taking the time derivative of the spin density,
n
a
(t, x)
t
=

d
3
pd
2
s
(2)
3

t
s
a
f +

d
3
pd
2
s
(2)
3

s
a
f
t
=

d
3
pd
2
s
(2)
3

t
s
a
f

d
3
pd
2
s
(2)
3

s
a
_
x
i
f
x
i
+ p
i
f
p
i
+ s
b
f
s
b
_
. (D4)
15
The second term is evaluated as
I
2
=

x
i

d
3
pd
2
s
(2)
3

s
a
x
i
f +

d
3
pd
2
s
(2)
3
(

s
a
x
i
)
x
i
f
+

d
3
pd
2
s
(2)
3
(

s
a
p
i
)
p
i
f +

d
3
pd
2
s
(2)
3
(

s
b
)
s
b
s
a
f
+

d
3
pd
2
s
(2)
3

s
a
f (D5)
Using Eq. (D3) to further simplify I
2
, we obtain the continuity equation (53) for the pseudo-spin.
[1] Michael V. Berry. Quantal phase factors accompanying adiabatic changes. Proc.Roy.Soc.Lond., A392:4557, 1984.
[2] Frank Wilczek and A. Zee. Appearance of Gauge Structure in Simple Dynamical Systems. Phys.Rev.Lett., 52:2111, 1984.
doi:10.1103/PhysRevLett.52.2111.
[3] John Moody, Alfred D. Shapere, and Frank Wilczek. Simple Realizations of Magnetic Monopole Gauge Fields: Diatoms
and Spin Precession. Phys.Rev.Lett., 56:893, 1986. doi:10.1103/PhysRevLett.56.893.
[4] H.K. Lee, Maciej A. Nowak, Mannque Rho, and I. Zahed. NonAbelian Berry phases in baryons. Annals Phys., 227:175205,
1993. arXiv:hep-ph/9301242, doi:10.1006/aphy.1993.1079.
[5] Di Xiao, Ming-Che Chang, and Qian Niu. Berry Phase Eects on Electronic Properties. Rev.Mod.Phys., 82:19592007,
2010. arXiv:0907.2021, doi:10.1103/RevModPhys.82.1959.
[6] Dam Thanh Son and Naoki Yamamoto. Berry Curvature, Triangle Anomalies, and the Chiral Magnetic Eect in Fermi
Liquids. Phys.Rev.Lett., 109:181602, 2012. arXiv:1203.2697, doi:10.1103/PhysRevLett.109.181602.
[7] M.A. Stephanov and Y. Yin. Chiral Kinetic Theory. Phys.Rev.Lett., 109:162001, 2012. arXiv:1207.0747,
doi:10.1103/PhysRevLett.109.162001.
[8] Jiunn-Wei Chen, Shi Pu, Qun Wang, and Xin-Nian Wang. Berry curvature and 4-dimensional monopole in relativistic
chiral kinetic equation. Phys.Rev.Lett., 110:262301, 2013. arXiv:1210.8312, doi:10.1103/PhysRevLett.110.262301.
[9] Dam Thanh Son and Naoki Yamamoto. Kinetic theory with Berry curvature from quantum eld theories. Phys.Rev.,
D87:085016, 2013. arXiv:1210.8158, doi:10.1103/PhysRevD.87.085016.
[10] Dmitri E. Kharzeev, Larry D. McLerran, and Harmen J. Warringa. The Eects of topological charge change in
heavy ion collisions: Event by event P and CP violation. Nucl.Phys., A803:227253, 2008. arXiv:0711.0950,
doi:10.1016/j.nuclphysa.2008.02.298.
[11] Kenji Fukushima, Dmitri E. Kharzeev, and Harmen J. Warringa. The Chiral Magnetic Eect. Phys. Rev., D78:074033,
2008. arXiv:0808.3382, doi:10.1103/PhysRevD.78.074033.
[12] Dmitri E. Kharzeev and Dam T. Son. Testing the chiral magnetic and chiral vortical eects in heavy ion collisions.
Phys.Rev.Lett., 106:062301, 2011. arXiv:1010.0038, doi:10.1103/PhysRevLett.106.062301.
[13] Johanna Erdmenger, Michael Haack, Matthias Kaminski, and Amos Yarom. Fluid dynamics of R-charged black holes.
JHEP, 01:055, 2009. arXiv:0809.2488, doi:10.1088/1126-6708/2009/01/055.
[14] Nabamita Banerjee, Jyotirmoy Bhattacharya, Sayantani Bhattacharyya, Suvankar Dutta, R. Loganayagam, et al. Hydro-
dynamics from charged black branes. JHEP, 1101:094, 2011. arXiv:0809.2596, doi:10.1007/JHEP01(2011)094.
[15] Mahdi Torabian and Ho-Ung Yee. Holographic nonlinear hydrodynamics from AdS/CFT with multiple/non-Abelian
symmetries. JHEP, 08:020, 2009. arXiv:0903.4894, doi:10.1088/1126-6708/2009/08/020.
[16] Anton Rebhan, Andreas Schmitt, and Stefan A. Stricker. Anomalies and the chiral magnetic eect in the Sakai-Sugimoto
model. JHEP, 1001:026, 2010. arXiv:0909.4782, doi:10.1007/JHEP01(2010)026.
[17] Tigran Kalaydzhyan and Ingo Kirsch. Fluid/gravity model for the chiral magnetic eect. Phys. Rev. Lett., 106:211601,
2011. arXiv:1102.4334.
[18] Carlos Hoyos, Tatsuma Nishioka, and Andy OBannon. A Chiral Magnetic Eect from AdS/CFT with Flavor. JHEP,
1110:084, 2011. arXiv:1106.4030, doi:10.1007/JHEP10(2011)084.
[19] Ilmar Gahramanov, Tigran Kalaydzhyan, and Ingo Kirsch. Anisotropic hydrodynamics, holography and the chiral magnetic
eect. Phys.Rev., D85:126013, 2012. arXiv:1203.4259, doi:10.1103/PhysRevD.85.126013.
[20] Alfonso Ballon-Bayona, Kasper Peeters, and Marija Zamaklar. A chiral magnetic spiral in the holographic Sakai-Sugimoto
model. JHEP, 1211:164, 2012. arXiv:1209.1953, doi:10.1007/JHEP11(2012)164.
[21] Dmitri E. Kharzeev and Ho-Ung Yee. Chiral helix in AdS/CFT with avor. Phys.Rev., D84:125011, 2011. arXiv:1109.0533,
doi:10.1103/PhysRevD.84.125011.
[22] Antti Gynther, Karl Landsteiner, Francisco Pena-Benitez, and Anton Rebhan. Holographic Anomalous Conductivities and
the Chiral Magnetic Eect. JHEP, 1102:110, 2011. arXiv:1005.2587, doi:10.1007/JHEP02(2011)110.
[23] Anton Rebhan, Andreas Schmitt, and Stefan Stricker. Holographic chiral currents in a magnetic eld.
Prog.Theor.Phys.Suppl., 186:463470, 2010. arXiv:1007.2494, doi:10.1143/PTPS.186.463.
[24] Ho-Ung Yee. Holographic Chiral Magnetic Conductivity. JHEP, 0911:085, 2009. arXiv:0908.4189,
doi:10.1088/1126-6708/2009/11/085.
16
[25] Bindusar Sahoo and Ho-Ung Yee. Holographic chiral shear waves from anomaly. Phys.Lett., B689:206212, 2010.
arXiv:0910.5915, doi:10.1016/j.physletb.2010.04.076.
[26] A. Gorsky, P.N. Kopnin, and A.V. Zayakin. On the Chiral Magnetic Eect in Soft-Wall AdS/QCD. Phys.Rev., D83:014023,
2011. arXiv:1003.2293, doi:10.1103/PhysRevD.83.014023.
[27] Karl Landsteiner, Eugenio Megias, Luis Melgar, and Francisco Pena-Benitez. Holographic Gravitational Anomaly and
Chiral Vortical Eect. JHEP, 1109:121, 2011. arXiv:1107.0368, doi:10.1007/JHEP09(2011)121.
[28] Karl Landsteiner and Luis Melgar. Holographic Flow of Anomalous Transport Coecients. JHEP, 1210:131, 2012.
arXiv:1206.4440, doi:10.1007/JHEP10(2012)131.
[29] Shu Lin and Ho-Ung Yee. Out-of-Equilibrium Chiral Magnetic Eect at Strong Coupling. Phys.Rev., D88:025030, 2013.
arXiv:1305.3949, doi:10.1103/PhysRevD.88.025030.
[30] Dam T. Son and Piotr Surowka. Hydrodynamics with Triangle Anomalies. Phys. Rev. Lett., 103:191601, 2009.
arXiv:0906.5044, doi:10.1103/PhysRevLett.103.191601.
[31] Shi Pu, Jian-hua Gao, and Qun Wang. A consistent description of kinetic equation with triangle anomaly. Phys. Rev.,
D83:094017, 2011. arXiv:1008.2418, doi:10.1103/PhysRevD.83.094017.
[32] A. V. Sadofyev and M. V. Isachenkov. The chiral magnetic eect in hydrodynamical approach. Phys. Lett., B697:404406,
2011. arXiv:1010.1550, doi:10.1016/j.physletb.2011.02.041.
[33] Dmitri E. Kharzeev and Ho-Ung Yee. Anomalies and time reversal invariance in relativistic hydrodynam-
ics: the second order and higher dimensional formulations. Phys.Rev., D84:045025, 2011. arXiv:1105.6360,
doi:10.1103/PhysRevD.84.045025.
[34] Dam Thanh Son and Naoki Yamamoto. Kinetic theory with Berry curvature from quantum eld theories. Phys.Rev.,
D87:085016, 2013. arXiv:1210.8158, doi:10.1103/PhysRevD.87.085016.
[35] Jiunn-Wei Chen, Jian-Hua Gao, Juan Liu, Shi Pu, and Qun Wang. Parity violating quantum kinetic theory in (2+1)-
dimensions. Phys.Rev., D88:074003, 2013. arXiv:1305.1835, doi:10.1103/PhysRevD.88.074003.
[36] Cristina Manuel and Juan M. Torres-Rincon. Kinetic theory of chiral relativistic plasmas and energy density of their gauge
collective excitations. 2013. arXiv:1312.1158.
[37] M. Abramczyk, T. Blum, G. Petropoulos, and R. Zhou. Chiral magnetic eect in 2+1 avor QCD+QED. PoS,
LAT2009:181, 2009. arXiv:0911.1348.
[38] P.V. Buividovich, M.N. Chernodub, E.V. Luschevskaya, and M.I. Polikarpov. Numerical evidence of chiral magnetic eect
in lattice gauge theory. Phys.Rev., D80:054503, 2009. arXiv:0907.0494, doi:10.1103/PhysRevD.80.054503.
[39] P.V. Buividovich, E.V. Luschevskaya, M.I. Polikarpov, and M.N. Chernodub. Chiral magnetic eect in SU(2) lattice
gluodynamics at zero temperature. JETP Lett., 90:412416, 2009. doi:10.1134/S0021364009180027.
[40] P.V. Buividovich, M.N. Chernodub, D.E. Kharzeev, T. Kalaydzhyan, E.V. Luschevskaya, et al. Magnetic-Field-Induced
insulator-conductor transition in SU(2) quenched lattice gauge theory. Phys.Rev.Lett., 105:132001, 2010. arXiv:1003.2180,
doi:10.1103/PhysRevLett.105.132001.
[41] Arata Yamamoto. Chiral magnetic eect in lattice QCD with a chiral chemical potential. Phys.Rev.Lett., 107:031601,
2011. arXiv:1105.0385, doi:10.1103/PhysRevLett.107.031601.
[42] Max A. Metlitski and Ariel R. Zhitnitsky. Anomalous axion interactions and topological currents in dense matter.
Phys.Rev., D72:045011, 2005. arXiv:hep-ph/0505072, doi:10.1103/PhysRevD.72.045011.
[43] G. M. Newman and D. T. Son. Response of strongly-interacting matter to magnetic eld: Some exact results. Phys. Rev.,
D73:045006, 2006. arXiv:hep-ph/0510049, doi:10.1103/PhysRevD.73.045006.
[44] James Charbonneau and Ariel Zhitnitsky. Topological Currents in Neutron Stars: Kicks, Precession, Toroidal Fields, and
Magnetic Helicity. JCAP, 1008:010, 2010. arXiv:0903.4450, doi:10.1088/1475-7516/2010/08/010.
[45] Michael Lublinsky and Ismail Zahed. Anomalous Chiral Superuidity. Phys. Lett., B684:119122, 2010. arXiv:0910.1373,
doi:10.1016/j.physletb.2010.01.015.
[46] Masayuki Asakawa, Abhijit Majumder, and Berndt Muller. Electric Charge Separation in Strong Transient Magnetic
Fields. Phys.Rev., C81:064912, 2010. arXiv:1003.2436, doi:10.1103/PhysRevC.81.064912.
[47] Karl Landsteiner, Eugenio Megias, and Francisco Pena-Benitez. Gravitational Anomaly and Transport. Phys. Rev. Lett.,
107:021601, 2011. arXiv:1103.5006, doi:10.1103/PhysRevLett.107.021601.
[48] Defu Hou, Hui Liu, and Hai-cang Ren. Some Field Theoretic Issues Regarding the Chiral Magnetic Eect. JHEP, 1105:046,
2011. arXiv:1103.2035, doi:10.1007/JHEP05(2011)046.
[49] Siavash Golkar and Dam T. Son. Non-Renormalization of the Chiral Vortical Eect Coecient. 2012. arXiv:1207.5806.
[50] Kristan Jensen. Triangle Anomalies, Thermodynamics, and Hydrodynamics. Phys.Rev., D85:125017, 2012.
arXiv:1203.3599, doi:10.1103/PhysRevD.85.125017.
[51] Kristan Jensen, Matthias Kaminski, Pavel Kovtun, Rene Meyer, Adam Ritz, et al. Towards hydrodynamics without an
entropy current. Phys.Rev.Lett., 109:101601, 2012. arXiv:1203.3556, doi:10.1103/PhysRevLett.109.101601.
[52] E.V. Gorbar, V.A. Miransky, I.A. Shovkovy, and Xinyang Wang. Radiative corrections to chiral separation eect in QED.
Phys.Rev., D88:025025, 2013. arXiv:1304.4606, doi:10.1103/PhysRevD.88.025025.
[53] Xu-Guang Huang and Jinfeng Liao. Axial Current Generation from Electric Field: Chiral Electric Separation Eect.
Phys.Rev.Lett., 110:232302, 2013. arXiv:1303.7192, doi:10.1103/PhysRevLett.110.232302.
[54] Kristan Jensen, R. Loganayagam, and Amos Yarom. Thermodynamics, gravitational anomalies and cones. JHEP, 1302:088,
2013. arXiv:1207.5824, doi:10.1007/JHEP02(2013)088.
[55] Kristan Jensen, Pavel Kovtun, and Adam Ritz. Chiral conductivities and eective eld theory. 2013. arXiv:1307.3234.
[56] Gokce Basar, Dmitri E. Kharzeev, and Ismail Zahed. Chiral and Gravitational Anomalies on Fermi Surfaces. Phys.Rev.Lett.,
111:161601, 2013. arXiv:1307.2234, doi:10.1103/PhysRevLett.111.161601.
17
[57] Dmitri E. Kharzeev and Ho-Ung Yee. Chiral Electronics. Physical Review B88, 115119, 2013. arXiv:1207.0477,
doi:10.1103/PhysRevB.88.115119.
[58] G.E. Volovik. Spin-motive force and orbital-motive force: from magnon BEC to chiral Weyl superuids. Pisma ZhETF
98,, 539-542, 2013. arXiv:1308.6700.
[59] Gokce Basar, Dmitri E. Kharzeev, and Ho-Ung Yee. Triangle anomaly in Weyl semi-metals. 2013. arXiv:1305.6338.
[60] Dmitri E. Kharzeev, Karl Landsteiner, Andreas Schmitt, and Ho-Ung Yee. Strongly interacting matter in magnetic elds:
an overview. Lect.Notes Phys., 871:111, 2013. arXiv:1211.6245, doi:10.1007/978-3-642-37305-3_1.
[61] Ismail Zahed. Anomalous Chiral Fermi Surface. Phys.Rev.Lett., 109:091603, 2012. arXiv:1204.1955,
doi:10.1103/PhysRevLett.109.091603.
[62] A.P. Balachandran, Per Salomonson, Bo-Sture Skagerstam, and Jan-Olof Winnberg. Classical Description of Particle
Interacting with Nonabelian Gauge Field. Phys.Rev., D15:2308, 1977. doi:10.1103/PhysRevD.15.2308.
[63] A.P. Balachandran, S. Borchardt, and A. Stern. Lagrangian and Hamiltonian Descriptions of Yang-Mills Particles.
Phys.Rev., D17:3247, 1978. doi:10.1103/PhysRevD.17.3247.
[64] Michael Stone and Vatsal Dwivedi. A Classical Version of the Non-Abelian Gauge Anomaly. Physical Review D88,
045012:8pp, 2013. arXiv:1305.1955.
[65] Vatsal Dwivedi and Michael Stone. Classical chiral kinetic theory and anomalies in even space-time dimensions. 2013.
arXiv:1308.4576.
[66] Jian-Hua Gao, Zuo-Tang Liang, Shi Pu, Qun Wang, and Xin-Nian Wang. Chiral Anomaly and Local
Polarization Eect from Quantum Kinetic Approach. Phys.Rev.Lett., 109:232301, 2012. arXiv:1203.0725,
doi:10.1103/PhysRevLett.109.232301.