Professional Documents
Culture Documents
q
__
b
W
(q, p);
(we use the natural units:
h = c = 1). We have also that the reverse is true: each function, a
W
(q, p), in phase space, ,
is associated with an operator, A, defined in the Hilbert space, H. This is a result of the WeylWigner formalism [2]. The
result is a non-commutative structure in , that has been explored in different ways [219]. In particular, from
W
, one
can introduce the operator
A = a
W
acting on functions b
W
, such that
A(b
W
) = a
W
b
W
; and so one can study unitary
representations of Lie groups in phase space using the Moyal product as defined by the operators
Corresponding author. Tel.: +55 (61) 3307 2900; fax: +55 61 3307 2363.
E-mail addresses: ronniamorim@yahoo.com.br (R.G.G. Amorim), khanna@phys.ualberta.ca (F.C. Khanna), asantana@fis.unb.br (A.E. Santana),
david@ufba.br (J.D.M. Vianna).
0378-4371/$ see front matter 2009 Published by Elsevier B.V.
doi:10.1016/j.physa.2009.06.013
3772 R.G.G. Amorim et al. / Physica A 388 (2009) 37713778
symplectic fields, going beyond some previous results in the literature [18,21]: we use functional methods for the case of
bosons and fermions and introduce the diagrammatic Feynman rules. As an example, we calculate at the one-loop level the
Wigner function for a quartic boson-interaction model.
The paper is organized in the following way. In Section 2, we present a brief outline of the symplectic representation
for the Poincar algebra. In Section 3 we develop functional methods for (free) boson and fermion fields. A self-interacting
boson field in phase space is studied in Section 4; and in Section 5, our final concluding remarks are presented.
2. Symplectic Lie-group representation
Using the -operators,
A = a
W
, we define the 4-momentum and 4-position operators, respectively, by
= p
= p
exp
_
i
2
_
__
= p
i
2
= q
= q
exp
_
i
2
_
__
= q
+
i
2
.
Then we introduce
M
=
Q
. These operators
and
M
] = i(g
),
[
] = i(g
+ g
).
The Casimir invariants are calculated by using the PauliLubanski matrices,
W
=
1
2
, where
is the
LeviCivita symbol. The invariants are
P
2
=
and
W
2
=
W
,
where
P
2
stands for the mass shell condition and
W
2
for the spin. The scalar representation [22] gives rise to the equation
_
p
ip
1
4
_
(q
, p
) = m
2
(q
, p
),
which is the KleinGordon equation in phase space. This equation is derived from the Lagrangian
=
1
4
(q
, p
)
q
(q
, p
)
q
+
1
2
ip
(q
, p
)
(q
, p
)
q
(q
, p
(q
, p
)
q
_
(p
m
2
)(q
, p
(q
, p
).
The association of this representation with the Wigner function formalism is given by
f
W
(q
, p
) = (q
, p
(q
, p
),
where f
W
(q
, p
_
p
i
2
_
(q
, p
) = m(q
, p
),
which is the Dirac equation in phase space, where the
) = 2g
.
This Dirac equation is derived from the Lagrangian
=
i
4
_
(q
, p
)
q
(q
, p
) (q
, p
(q
, p
)
q
_
(q
, p
)(m
)(q
, p
).
where (q
, p
) =
(q
, p
)
0
, with
(q
, p
, p
, p
) = (q
, p
) (q
, p
).
In the following we use functional methods to proceed with the quantization of these two fields, preparing an approach to
study interactions.
R.G.G. Amorim et al. / Physica A 388 (2009) 37713778 3773
3. Path integral in phase space: Free fields
For the complex scalar field in phase space in the presence of sources J (q
, p
) and J
(q
, p
), we define
1
= Z
0
_
J (q
, p
), J
(q
, p
)
_
,
so that we have
1
=
_
D(q
, p
)
_
D
(q
, p
) exp i
_
d
4
q
d
4
p
[
0
((q
, p
),
(q
, p
))
J (q
, p
(q
, p
) J
(q
, p
)(q
, p
)],
where
_
D(q
, p
) and
_
D
(q
, p
, p
) and
(q
, p
), while
_
d
4
q
d
4
p
is the
integration in the full relativistic phase space. Although we are in phase space, the usual machinery for treating functional
integrals and derivatives is still valid. Using
_
D(x)D
(x) exp
_
(x)A(x)dx
_
= (det A)
1
,
we obtain
1
=
_
det
_
1
4
q
+ ip
(p
m
2
+ i)
__
1
exp
_
i
_
J
(q
, p
)
_
1
4
q
+ ip
(p
m
2
+ i)
_
1
J (q
, p
)dq
dp
dq
dp
_
,
where
q
=
2
q
.
In the absence of sources, we have
Z
0
[0] =
_
det
_
1
4
q
+ ip
(p
2
m
2
+ i)
__
1
.
The Green function satisfies the equation
L
1
F
(q
, p
) = (q
)(p
),
where
L
1
=
_
1
4
q
+ ip
(p
2
m
2
+ i)
_
.
In this sense, we have
F
(q
, p
) =
_
1
4
q
+ ip
(p
2
m
2
)
_
1
.
Therefore, the generating functional for the free scalar field is
1
=
_
det
_
1
4
q
+ ip
(p
2
m
2
)
__
1
exp
_
i
_
J
(q
, p
)
F
(q
, p
).
J (q
, p
)dq
dp
dq
dp
_
.
Using this generating functional, the two-point function is given by
G
0
(q
, p
) =
2
Z
0
[J (q
, p
), J
(q
, p
)]
J
(q
, p
)J (q
, p
J =J
=0
= 0|T[(q
, p
(q
, p
)]|0,
where T is the time-ordering operator. The relation between
F
(q
, p
) and G
0
(q
, p
) is [22]
G
0
(q
, p
) = i
F
(q
, p
).
We obtain a physical interpretation for this formalism by showing that G
0
(q
, p
, p
) = lim
q
,p
,p
exp
_
i
q
_
G
0
(q
, p
).
3774 R.G.G. Amorim et al. / Physica A 388 (2009) 37713778
With this result, we prove that the average of an observable A(q
, p
A =
_
A(q
, p
)
_
lim
q
,p
,p
G
0
(q
, p
)
_
dq
dp
.
The functional derivative of Z
0
[J (q
, p
), J
(q
, p
)] with respect to J
(q
, p
) is
1
i
1
J
(q
, p
)
=
_
dq
dp
F
(q
, p
)J (q
, p
) exp
_
i
_
J
(q
, p
)
F
(q
, p
)
J (q
, p
)dq
dp
dq
dp
_
.
Using
_
1
4
q
+ ip
(p
m
2
)
_
, we find the differential equation for Z
0
[J (q
, p
), J
(q
, p
)], i.e.
L
1
1
= J (q
, p
)
1
.
Let us now analyze the quantization of the free Dirac field. The functional methods in phase space for fermions are
introduced by defining the Grassmann algebra in phase space, which is similar to the usual case. Then we write the
generating functional for the free Dirac field as
2
= Z
0
[(q
, p
), (q
, p
)],
such that
2
=
1
N
_
D(q
, p
)D(q
, p
) exp i
_
__
i
4
_
(q
, p
)
q
(q
, p
) (q
, p
(q
, p
)
q
_
(q
, p
)(m
)(q
, p
) + (q
, p
)(q
, p
) + (q
, p
)(q
, p
)
_
dq
dp
_
.
Defining
_
q
, we have
2
=
1
N
_
D(q
, p
)D(q
, p
) exp
_
i
_
_
(q
, p
)
_
i
4
(m
)
_
(q
, p
)
+(q
, p
)(q
, p
) + (q
, p
)(q
, p
)
_
dq
dp
_
,
where (q
, p
) and (q
, p
, p
) and (q
, p
(m
2
),
the generating functional becomes
2
=
1
N
_
D(q
, p
)D(q
, p
) exp
_
i
_
_
(q
, p
)S
1
(q
, p
) + (q
, p
)(q
, p
)
+(q
, p
)(q
, p
)
_
dq
dp
_
.
This expression leads to,
2
=
1
N
exp
_
i
_
(q
, p
)S(q
, p
)(q
, p
)dq
dp
dq
dp
_
det(iS
1
).
Since N = det(iS
1
), we have,
Z
0
[(q
, p
), (q
, p
)] = exp
_
i
_
(q
, p
)S(q
, p
)dq
dp
dq
dp
_
.
The two-point function is
S(q
, p
) =
2
Z
0
[(q
, p
), (q
, p
)]
(q
, p
)(q
, p
==0
.
The relation of this function with the Wigner function is given by
f
w
(q
, p
) = lim
q
,p
,p
exp
_
i
q
_
S(q
, p
).
This provides the physical meaning for the representation [21].
R.G.G. Amorim et al. / Physica A 388 (2009) 37713778 3775
4. Path integral in phase space: Interacting fields
We consider nowthe generating functional of the interacting complex KleinGordon field in phase space. The Lagrangian
is =
0
+
int
(,
), where
0
is the free field Lagrangian and
int
is the interaction Lagrangian. The normalized generating
functional is defined by
3
= Z
_
J (q
, p
), J
(q
, p
)
_
,
such that
3
=
1
Z
0
_
D(q
, p
)D
(q
, p
) exp
_
iS i
_
d
4
q
d
4
p
(J (q
, p
(q
, p
) + J
(q
, p
)(q
, p
))
_
,
where S =
_
d
4
q
d
4
p
. Then we get
1
i
3
J
(q
, p
)
=
1
Z
0
_
D(q
, p
)D
(q
, p
)
exp
_
iS i
_
dqdp(J (q
, p
(q
, p
) + J
(q
, p
)(q
, p
))
_
. (1)
Introducing the functional,
Z[(q
, p
),
(q
, p
)] =
exp iS
Z
0
,
it follows that
3
=
_
D(q
, p
)D
(q
, p
Z[(q
, p
),
(q
, p
)]
exp
_
i
_
d
4
q
d
4
p
(J (q
, p
(q
, p
) + J
(q
, p
)(q
, p
))
_
;
that is, a functional Fourier transform.
Differentiating
Z[(q
, p
),
(q
, p
)] with relation to
(q
, p
), we get
i
Z[(q
, p
),
(q
, p
)]
(q
, p
)
= L
1
(q
, p
) +
int
(q
, p
Z[(q
, p
),
(q
, p
)]. (2)
Multiplying this equation by
exp
_
i
_
dq
dp
(J (q
, p
(q
, p
) + J
(q
, p
)(q
, p
))
_
and performing an integral over (q
, p
) and
(q
, p
), we find
_
1
4
q
+ ip
(p
m
2
+ i)
_
1
i
3
J
(q
, p
int
(q
, p
)
_
1
i
(q
, p
)
,
1
i
J (q
, p
)
_
3
= I[ J ],
where
I[J ] = i
_
D(q
, p
)D
(q
, p
Z[(q
, p
),
(q
, p
)]
(q
, p
)
exp
_
i
_
d
4
q
d
4
p
(J (q
, p
(q
, p
) + J
(q
, p
)(q
, p
))
_
.
Integrating by parts, we obtain the differential equation,
J (q
, p
)
3
= L
1
1
i
3
J
(q
, p
int
(q
, p
)
_
1
i
(q
, p
)
,
1
i
J (q
, p
)
_
3
.
We show that the solution of this equation is
3
= N exp
_
i
_
d
4
q
d
4
p
int
_
1
i
,
1
i
J
__
1
, (3)
where N is a normalization factor. In order to demonstrate this result, we use the BakerHausdorff formula,
exp(A)B exp(A) = B + [A, B] +
1
2
[A, [A, B]] + ,
3776 R.G.G. Amorim et al. / Physica A 388 (2009) 37713778
with the properties
_
J (q
, p
),
1
i
J (q
, p
)
_
= i(q
)(p
),
and
_
J
(q
, p
),
1
i
(q
, p
)
_
= i(q
)(p
).
Thus it follows that
FJ (q
, p
)F
= J (q
, p
) +
int
(q
, p
)
_
1
i
J (q
, p
)
,
1
i
(q
, p
)
_
,
where F = exp
_
i
_
d
4
q
d
4
p
int
_
and F
= exp
_
i
_
d
4
q
d
4
p
int
_
. Then, we have
J (q
, p
)
3
= NJ (q
, p
) exp
_
i
_
d
4
q
d
4
p
int
_
1
i
(q
, p
)
,
1
i
J (q
, p
)
__
1
.
Using Eq. (2) this expression becomes
L[J ] = N exp
_
i
_
d
4
q
d
4
p
int
_
1
i
(q
, p
)
,
1
i
J (q
, p
)
__ _
J (q
, p
)
+
(q
, p
int
_
1
i
(q
, p
)
,
1
i
J (q
, p
)
__
Z
0
[J (q
, p
), J
(q
, p
)], (4)
where,
L[J ] = J (q
, p
)Z[J (q
, p
), J
(q
, p
)].
Withthe differential equationfor Z
0
[J (q
, p
), J
(q
, p
)] andchanging
int
andexp i
_
int
, Eq. (4) reduces tothe differential
equation for
3
= Z[J (q
, p
), J
(q
, p
)].
As an application, consider the complex scalar field in phase space, , with self-interaction
4
(q
, p
). This describes
relativistic bosons with contact interaction. Or in another words, we have a kinetic equation with a local Boltzmann-like
collision term. Our aim is to calculate the Wigner function describing this process at the one-loop level. With this example,
we also show that the Feynman rules for this symplectic field theory is similar to the usual field theory in the configuration
space. The first step is to calculate the two-point function. Consider the Lagrangian density
=
1
4
(q
, p
)
q
(q
, p
)
q
+
1
2
ip
(q
, p
)
(q
, p
)
q
(q
, p
(q
, p
)
q
_
(p
m
2
)(q
, p
(q
, p
) +
4!
((q
, p
(q
, p
))
2
. (5)
Expanding the interacting term in the generating functional as a power series in , we obtain up to second order
3
= N
_
1 + i
4
_
d
4
q
d
4
p
_
1
i
(q
, p
)
_
2
_
1
i
J (q
, p
)
_
2
2
16
_
d
4
q
d
4
p
_
1
i
(q
, p
)
_
2
_
1
i
J (q
, p
)
_
2
_
1
i
(q
, p
)
_
2
d
4
q
d
4
p
_
1
i
J (q
, p
)
_
2
+ O(
3
)
_
Z
0
[J (q
, p
), J
(q
, p
)].
It is important to observe that to order
0
, we have Z
0
[J (q
, p
), J
(q
, p
0
=
_
1
i
(q
, p
)
_
2
_
1
i
J (q
, p
)
_
2
Z
0
[J (q
, p
), J
(q
, p
)].
The first derivative is
_
1
i
1
J (q
, p
)
_
=
_
J
(q
, p
)
F
(q
q, p
p)d
4
q
d
4
p
exp
_
i
_
J
(q
, p
)
F
(q
, p
) J (q
, p
)d
4
q
d
4
p
d
4
q
d
4
p
_
.
R.G.G. Amorim et al. / Physica A 388 (2009) 37713778 3777
The second derivative is given by
_
1
i
J (q
, p
)
_
2
1
=
__
J
(q
, p
)
F
(q
, p
)d
4
q
d
4
p
_
2
exp
_
i
_
J
(q
, p
)
F
(q
, p
) J (q
, p
)d
4
q
d
4
p
d
4
q
d
4
p
_
.
Writing
0
[J
(q
, p
), J (q
, p
)] =
_
1
i
(q
, p
)
__
1
i
J (q
, p
)
_
2
1
,
we have
0
=
F
(q
, p
) 2
__
J
(q
, p
)
F
(q
, p
)
_
exp
_
i
_
J
(q
, p
)
F
(q
, p
)J (q
, p
)
_
+
__
J
(q
, p
)
F
(q
q, p
)d
4
q
d
4
p
_
2
__
F
(q
, p
)J (q
, p
)d
4
q
d
4
p
_
exp
_
i
_
J
(q
, p
)
F
(q
, p
)J (q
, p
)
_
.
Writing
0
[J
(q
, p
), J (q
, p
)] =
_
1
i
(q
, p
)
_
2
_
1
i
J (q
, p
)
_
2
1
,
we obtain
0
=
_
2
2
F
(q
, p
) + 4
F
(q
, p
___
J
(q
, p
)
F
(q
, p
)d
4
q
d
4
p
_ __
F
(q
, p
)J (q
, p
)d
4
q
d
4
p
__
+
__
J
(q
, p
)
F
(q
, p
)d
4
q
d
4
p
_
2
__
d
4
q
d
4
p
F
(q
, p
)J (q
, p
)
_
2
_
1
.
The two-point function is
G(q
1
, q
2
; p
1
, p
2
) =
2
Z[J
(q
, p
), J (q
, p
)]
J (q
1
, p
1
)J
(q
2
, p
2
)
J =J
=0
.
Up to order , we have
G(q
, q
; p
, p
) = G
0
(q
, q
; p
, p
)
i
2
G
0
(0)
_
dq
dp
G
0
(q
, p
)
G
0
(q
, p
) + O(
2
),
where we have changed q
1
and p
1
by q
and p
and q
2
and p
2
by q
and p
, p
) = lim
q
,p
,p
exp
_
i
q
_
G(q
, q
; p
, p
).
Applying the limit and expanding the star product up to second order (which is equivalent to going up to the second order
in the Planck constant), we obtain,
f
w
(q
, p
) = f
0
w
(q
, p
)
2
G
0
(0)
_
d
4
q
d
4
p
G
0
(q
, p
)G
0
(q
, p
)
+
1
2
lim
(q
,p
,p
)
_
2
G
0
(q
, p
)
q
2
G
0
(q
, p
)
qp
_
+
4
G
0
(0) lim
(q
,p
,p
)
(
q
p
q
p
)
_
d
4
q
d
4
p
G
0
(q
, p
)G
0
(q
, p
) + ,
3778 R.G.G. Amorim et al. / Physica A 388 (2009) 37713778
where
f
0
w
(q
, p
) = lim
q
,p
,p
exp
_
i
q
_
G
0
(q
, p
)
corresponds to the zero-order term in and describes free particles moving in phase space. Up to the first order in ,
however, f
w
(q, p) describes bosons in a contact interaction at a point (q, p).
5. Concluding remarks
In short, we have set forth a perturbative method for symplectic fields, using functional methods in phase space, . One
basic result is the proof that the two-point function in is related to the Wigner function, for both cases of bosons and
fermions. This shows the association of the present approach with others in the literature [6,7]. As an application we have
analyzed the phase space complex scalar field with quartic interaction, corresponding to a kinetic equation with a local
Boltzmann-like collision term. In this case the Feynman rules are similar to those of the usual
4
-theory. The Wigner
function is calculated at the one-loop level, by expanding the -operator up to second order. This method provides a
perturbative approach to treat the relativistic Wigner function, proper for the relativistic kinetic theory. At present we are
analyzing other aspects of this theory as the construction of the effective potential and processes of symmetry breaking in
phase space. These results will be presented in another work.
Acknowledgements
This work was partially supported by CNPq of Brazil and NSERC of Canada.
References
[1] E. Wigner, Phys. Rev. 40 (1932) 749.
[2] M. Hillery, R.F. O Connell, M.O. Scully, E.P. Wigner, Phys. Rep. 106 (1984) 121.
[3] Y.S. Kim, M.E. Noz, Phase Space Picture and Quantum MechanicsGroup Theoretical Approach, W. Scientific, London, 2001.
[4] T. Curtright, D. Fairlie, C. Zachos, Phys. Rev. D 58 (1998) 25002.
[5] H. Weyl, Z. Phys 46 (1927) 1;
J.E. Moyal, Proc. Camb. Phil. Soc. 45 (1949) 99.
[6] S.A. Smolyansky, A.V. Prozorkevich, G. Maino, S.G. Mashnic, Ann. Phys. (NY) 277 (1999) 193.
[7] T. Curtright, C. Zachos, J. Phys. A 32 (1999) 771.
[8] D. Galetti, A.F.R. de Toledo Piza, Physica A 214 (1995) 207.
[9] L.P. Horwitz, S. Shashoua, W.C. Schive, Physica A 161 (1989) 300.
[10] P.R. Holland, Found. Phys. 16 (1986) 701.
[11] M.C.B. Fernandes, J.D.M. Vianna, Braz. J. Phys. 28 (1999) 2.
[12] M.C.B. Fernandes, A.E. Santana, J.D.M. Vianna, J. Phys. A: Math. Gen. 36 (2003) 3841.
[13] A.E. Santana, A. Matos Neto, J.D.M. Vianna, F.C. Khanna, Physica A 280 (2000) 405.
[14] D. Bohm, B.J. Hiley, Found. Phys. 11 (1981) 179.
[15] M.C.B. Andrade, A.E. Santana, J.D.M. Vianna, J. Phys. A: Math. Gen. 33 (2000) 4015.
[16] C.K. Zachos, Int. J. Mod. Phys. A 17 (2002) 297.
[17] I. Galaviz, H. Garca-Compen, M. Przanowski, F.J. Turrubiates, WeylWignerMoyal for Fermi Classical Systems, arXiv: hep-th/0612245v1.
[18] J. Dito, J. Math. Phys. 33 (1992) 791.
[19] M.A. de Gosson, J. Phys. A: Math. Gen. 38 (2000) 1.
[20] M.D. Oliveira, M.C.B. Fernandes, F.C. Khanna, A.E. Santana, J.D.M. Vianna, Ann. Phys. (NY) 312 (2004) 492.
[21] R.G.G. Amorim, M.C.B. Fernandes, F.C. Khanna, A.E. Santana, D.M Vianna, Phys. Lett. A 361 (2007) 464.
[22] C. Itzykson, J.B. Zuber, Quantum Field Theory, McGraw Hill, NY, 1980.