Report: Weldon-Breit-Wigner formula I

M. Shokri
(Dated: March 16, 2016)

I. INTRODUCTION

In ordinary particle physics at zero temperature resonances in observables appear due

to a blow up in propagators. Consider some process whose amplitude is proportional to
1
,
s − m2
√
which m is the mass of some intermediate particle and s is the CM energy of incoming
particles. As is evident from the propagators denominator a peak in amplitude and thus
√
the observable (e.g. cross section) appears at s = m. Such peaks are studied with Breit-
Wigner formula. For a process such as ab → R∗ → cd the cross section near resonance is
proportional to
Γ∗ (R → ab)Γ(R → cd)
.
(s − m2R )2 + (mR Γtotal
R )
2

The above relation has evident physical significance. The peak happens at the mass of the
resonance. Peak’s height grows as decay widths decreases. The nominator expresses the
idea that the intermediate particle is almost on-shell. In thermal environment instead of
individual scattering events one looks through the collective production rates. Here the
focus will be on the production of dilepton pairs from a hot and dense pion gas which are
dominantly created by π + π − → e+ e− . From kinetic theory
d3 k d3 k 0
Z
+ − + − dN 0 + − + −
R(π π → e e ) ≡ 4 = 3 3 f (k)f (k )σ(π π → e e )vrel , (I.1)
d x (2π) (2π)
which in k and k 0 are momenta of incoming particles and
 k k0 
 
vrel =  − 0  .
ω ω
In the following sections I demonstrate how a generalization of Breit-Wigner formula is
found for 4dN 4 for the special case of sigma-meson resonance1 . This report will be in 3
d xd p
parts, in the first part I report the derivation of Weldon-Breit-Wigner formula for the sigma
resonance; in second part I discuss the leptonic decay of sigma-meson as well as hypothetical
relation of sigma-meson resonance to the critical temperature; the third part will be devoted
to the general Weldon-Breit-Wigner formula.

1 0
p=k+k
 II. HADRONIC DECAY

In this section I calculate decay rate of Sigma-meson to pions in presence of finite tem-
perature and density. Since I want to consider density effects I assume the pions to be
charged.

A. In vacuum

The differential decay rate in vacuum reads

λ2
dΓ(σ → π + π − ) = dΠ, (II.2)
vac 2p0
with p being four-momentum of σ in some frame of reference and λ the σππ coupling
constant. The phase space reads

4 4 0 d3 k d3 k 0
dΠ = (2π) δ (k + k − p) ,
(2π)3 2k0 (2π)3 2k00

which in k and k 0 are four-momenta of pions. A well-known trick in integration over the
phase space is to rewrite

d3 k 0 d4 k 0
Z Z
02
= 3 δ(k − m2π )θ(k00 ),
(2π) 3
2k00 (2π)

Therefore

λ2 d3 k d3 k 0
Z
+ −
Γvac (σ → π π ) = (2π)4 δ 4 (k + k 0 − p)
2p0 (2π)3 2k0 (2π)3 2k00
λ2 d3 k d4 k 0
Z
2
= (2π)4 3 3δ
4
(k + k 0 − p)δ(k 0 − m2π )θ(k00 )
2p0 (2π) 2k0 (2π)
2 Z 3
λ d k
= 2 δ(m2σ − 2p.k)θ(p0 − k0 )
8π p0 k0

For an on-shell sigma meson it is possible to boost to the CM frame which in p = (mσ , 0)
and write

λ2
Z 2
k dk dΩ 1 m
Γvac (σ → π + π − ) = 2 δ(k0 − σ )θ(mσ − k0 )
8π mσ k0 2mσ 2
2 Z
λ kk0 dk0 1 m
= δ(k0 − σ )θ(mσ − k0 )
8πmσ k0 2mσ 2
2 Z
λ m
q
= 2 dk0 k02 − m2π δ(k0 − σ )θ(mσ − k0 ),
8πmσ 2

2
which finally results to

λ2 λ2 a
q
+ − 2 2
Γvac (σ → π π ) = mσ − 4mπ = , (II.3)
16πm2σ 16πmσ
r
2
4mπ
which in a ≡ 1 − 2 .
mσ

B. In medium

In medium the phase space of bosonic decay products is enhanced as

[1 + n+ (k)][1 + n− (k 0 )] dΠ

The computations should be done in the medium’s local rest frame (LRF). I denote energies
of sigma meson and pions in LRF respectively as Ē, ω and ω 0 . Using the same trick of the
vacuum case we got

λ2 d3 k
Z Z
2
Γ(σ → π + π − ) = 2 d4 k 0 [1 + n+ (k)][1 + n− (k 0 )]δ 4 (k + k 0 − p)δ(k 0 − m2π )θ(k00 )
8π Ē 2ω
λ2
Z 2
k dk d cos θ dφ
= 2 [1 + n+ (ω)][1 + n− (Ē − ω)]δ(m2σ − 2Ēω + 2p̄k cos θ)θ(E − ω)
8π Ē 2ω
!
λ2 2Ēω − m2σ
Z
kω dω d cos θ 1
= [1 + n+ (ω)][1 + n− (Ē − ω)] δ cos θ − θ(E − ω)
8π Ē ω 2p̄k 2p̄k
!
λ2 2Ēω − m2σ
Z
= dω d cos θ [1 + n+ (ω)][1 + n− (Ē − ω)]δ cos θ − θ(E − ω)
16π Ē p̄ 2p̄k
q
which in p̄ = Ē 2 − m2σ . The delta function’s argument gives a constraint on ω as

 2Ēω − m2 2
 
σ
 ≤ 1.

2p̄k

 

3
Therefore we must have

4Ē 2 ω 2 + m4σ − 4Ēωm2σ − 4p̄2 k 2 ≤ 0

4Ē 2 ω 2 + m4σ − 4Ēωm2σ − 4p̄2 ω 2 + 4p̄2 m2π ≤ 0
4m2σ ω 2 + m2σ (Ē 2 − p̄2 ) − 4Ēωm2σ + 4p̄2 m2π ≤ 0
!
2 Ē 2 p̄2 4m2π
ω − Ēω + − 1− 2 ≤0
4 4 mσ
Ē 2 − a2 p̄2
ω 2 − Ēω + ≤0
  4
Ē + ap̄ Ē − ap̄ Ē + ap̄ Ē − ap̄
ω2 − + ω+ × ≤0
2 2 2 2
  
Ē + ap̄ Ē − ap̄
ω− ω− ≤ 0,
2 2

hence
Ē − ap̄ Ē + ap̄
ω− ≡ ≤ω≤ ≡ ω+ .
2 2
Some evident but handy relations between two limits are

ω+ + ω− = Ē
ω+ − ω− = ap̄

Now we must compute

λ2 ω+
Z
+ −
Γ(σ → π π ) = dω [1 + n+ (ω)][1 + n− (Ē − ω)],
16π Ē p̄ ω−

4
the E > ω condition is satisfied in the integration interval. Using relations proved in
appendix A

λ2 ω+
Z
+ −
Γ(σ → π π ) = dω n− (−ω)n+ (ω − E)
16π Ē p̄ ω−

λ2 eβ Ē ω+
Z
= dω [n+ (ω − E) − n+ (ω)]
16π Ē p̄ 1 − eβ Ē ω−

λ2 eβ Ē h 
−β(ω−Ē−µ)
 
−β(ω−µ)
iω+
= T ln e − 1 − ln e − 1
16π Ē p̄ 1 − eβ Ē ω−
" !#ω+
λ2 T eβ Ē eβ Ē − eβ(ω−µ)
= ln
16π Ē p̄ 1 − eβ Ē 1 − eβ(ω−µ) ω−
" ! !#
λ2 T eβ Ē eβ Ē − eβ(ω+ −µ) 1 − eβ(ω− −µ)
= ln
16π Ē p̄ 1 − eβ Ē 1 − eβ(ω+ −µ) eβ Ē − eβ(ω− −µ)
!
λ2 T eβ Ē eβ Ē − eβ(ω+ −µ) − eβ(Ē+ω− −µ) + eβ(Ē−2µ)
= ln
16π Ē p̄ 1 − eβ Ē eβ Ē − eβ(ω− −µ) − eβ(Ē+ω+ −µ) + eβ(Ē−2µ)
!
λ2 T eβ Ē eβ Ē − eβ(ω+ −µ) − eβ(Ē+ω− −µ) + eβ(Ē−2µ)
= ln
16π Ē p̄ 1 − eβ Ē eβ Ē − eβ(ω− −µ) − eβ(Ē+ω+ −µ) + eβ(Ē−2µ)
!
λ2 T eβ Ē 1 − eβ(ω+ −Ē−µ) − eβ(ω− −µ) + e−2βµ
= ln
16π Ē p̄ 1 − eβ Ē 1 − eβ(ω− −Ē−µ) − eβ(ω+ −µ) + e−2βµ
!
λ2 T eβ Ē 1 − e−β(ω− +µ) − eβ(ω− −µ) + e−2βµ
= ln
16π Ē p̄ 1 − eβ Ē 1 − e−β(ω+ +µ) − eβ(ω+ −µ) + e−2βµ
!
λ2 T eβ Ē eβµ + e−βµ − e−βω− − eβω−
= ln
16π Ē p̄ 1 − eβ Ē eβµ + e−βµ − e−βω+ − eβω+
λ2 T
 
−1 cosh βµ − cosh βω−
= ln
16π Ē p̄ 1 − e−β Ē cosh βµ − cosh βω+
2  
λ T 1 cosh βµ − cosh βω+
= ln
16π Ē p̄ 1 − e−β Ē cosh βµ − cosh βω−

Defining
 
Ē+ap̄ µ

cosh 2T − cosh T
Y ≡   ,
Ē−ap̄ µ

cosh 2T − cosh T

the final result reads

λ2 T ln Y
Γ(σ → π + π − ) = (II.4)
16π Ē p̄ 1 − e−β Ē

5
 C. Limiting behavior

The factor a has a significant physical meaning. The available phase space is proportional
to a. In vacuum this factor has a constant value but it receives temperature and density
dependency in medium. Very important region of interest is a  1 i.e. when mσ & 2mπ .
By a Taylor series expansion around a = 0 one finds

λ2 a E
sinh 2T
Γ(σ → π + π − ) =  E
 + O(a2 ), (II.5)
−
16π Ē 1 − e T cosh E − cosh µ

2T T

which reveals phase-space dependency of decay width.

III. DILEPTON PRODUCTION RATE

Starting from (I.1) I am to prove the Weldon-Breit-Wigner formula for π + π − → σ ∗ →

e+ e− . Working in π + π − CM frame one finds
q s
     4 ω 2 − m2
0
4m2π

 k k   2k   4k  π
vrel =  − 0  =   =   =
      =2 1− ,
ω ω ω M M M2
which resembles a from hadronic decay. In above formula ω is pion’s energy in CM frame
and M is the invariant mass of pions as well as di-electrons defined as
q
M ≡ (k + k 0 )2 = 2ω.

Let denote sigma meson’s momentum with p, then trivially

Z
1 = d4 p δ 4 (k + k 0 − p).

Putting found expression for vrel and multiplying RHS of (I.1) by ”‘unity”’ results
s
3 3 0
4m2π
Z Z
d k d k
R(π + π − → e+ e− ) = 2 d4 p δ 4 (k+k 0 −p) 0 + − + −
3 n+ (k)n− (k )σ(π π → e e ) 1− .
(2π)3 (2π) M2

On the other hand

Z
+ − + − dR
R(π π → e e ) = d4 p ,
d4 p

6
therefore
s
3 3 0
4m2π
Z
dR d k d k 4 0 0 + − + −
= 2 δ (k + k − p)n + (k)n − (k )σ(π π → e e ) 1 −
d4 p (2π)3 (2π)3 M2
s
4m2π d3 k d3 k 0 4
Z
+ − + −
= 2σ(π π → e e ) 1 − δ (k + k 0 − p)n+ (k)n− (k 0 )
M2 (2π)3 (2π)3
s
4m2π 0 d3 k d3 k 0
Z
+ − + − 0 4 0 0
= 2σ(π π → e e ) 1 − 2 2ω 3 0 3 4ωω δ (k + k − p)n+ (k)n− (k )
M 2ω(2π) 2ω (2π)
s
4m2π d3 k d4 k 0 4
Z
02
= 2M 2 σ(π + π − → e+ e− ) 1 − 2 3
0 0 2 0
3 δ (k + k − p)n+ (k)n− (k )δ(k − mπ )θ(k0 )
M 2ω(2π) (2π)
s
4m2π d3 k
Z
= 2M 2 σ(π + π − → e+ e− ) 1 − n+ (ω)n− (E − ω)δ(M 2 − 2p.k)θ(E − ω)
M2 (2π)6
s
2 + − + − 4m2π
= 2M σ(π π → e e ) 1 −
M2
Z 2 !
k dk d cos θ dφ 2Eω − M 2
× n+ (ω)n− (E − ω)δ cos θ − θ(E − ω)
2pk(2π)6 2ω 2pk
s
4m2π
= 2M 2 σ(π + π − → e+ e− ) 1 −
M2
!
2Eω − M 2
Z
kω dω d cos θ 2π
× n+ (ω)n− (E − ω)δ cos θ − θ(E − ω)
2pk(2π)6 2ω 2pk
s
4m2π
!
2 Z ω+
M
= 5 1− 2 dω n+ (ω)n− (E − ω) σ(π + π − → e+ e− )
2p(2π) M ω−
s
4m2π −βE
!
M2
Z ω+
= 5 1− 2 e dω n− (−ω)n+ (ω − E) σ(π + π − → e+ e− ),
2p(2π) M ω−

which have used an identity from appendix A to restore the integral we calculated in the
hadronic decay. The result is
s
dR M 2
4m2π T ln Y
= 1− σ(π + π − → e+ e− ) (III.6)
d4 p 2p̄(2π)5 M 2 eĒ/T − 1

+ − + −
A. Scattering cross section for π π →e e

The amplitude for π + π − → e+ e− is given by

2
+ − + − λ αe 0
M(π π → e e ) = 2 I 2 ū(q )γα v(q),
p − m2σ p

7
with p, q and q 0 being sigma meson and di-electrons momenta. I α is the loop connecting
sigma meson to photon. This can be rewritten as

+ − + − λM(σ ∗ → e+ e− )
M(π π → e e ) = ,
p2 − m2σ

with σ ∗ being the virtual sigma meson. The differential cross section is given by

1 λ2 X  ∗ + − 
2
dσ = M(σ → e e )  dΠdilepton , (III.7)
2ω2ω 0 vrel (p2 − m2σ )2 spins


with dilepton phase space defined as

d3 q d3 q 0
dΠdilepton = (2π)4 δ 4 (k − q − q 0 ).
(2π)3 2q0 (2π)3 2q00
Going to incoming particles CM frame

λ2
Z X
1 ∗
2
+ − 
σ= M(σ → e e ) dΠdilepton

(M 2 − m2σ )2
r 
2
2M 2 1− 4mπ
2
spins
M

From (II.3) one finds

16πm2σ Γvac (σ → π + π − )
λ2 = q ,
2 2
mσ − 4mπ

therefore
16πm2σ Γvac (σ → π + π − )
Z X
1 ∗
2
+ − 
σ= M(σ → e e ) dΠdilepton

r q
(M 2 − m2σ )2 m2σ − 4m2π
2
2M 2 1− 4mπ
2
spins
M

1 16πm2σ Γvac (σ → π + π − ).2M Γ(σ ∗ → e+ e− )

= r q
(M − mσ ) m2σ − 4m2π
2 2 2
2
2M 2 1− 4mπ
2
M

16πm2σ Γvac (σ → π + π − )Γ(σ ∗ → e+ e− )

= r q
(M − mσ ) m2σ − 4m2π
2 2 2
2
4mπ
M 1− 2
M

16πmσ Γvac (σ → π + π − )Γ(σ ∗ → e+ e− )

=
(M 2 − m2σ )2
r
2
4mπ
aM 1− 2
M

Since we are interested in cross-section near resonance we should follow the well-known
recipe for the propagator squared

(M 2 − m2σ )2 → (M 2 − m2σ )2 + (mσ Γtot

σ )
2

8
 mσ
Using above recipe and assuming M ∼ 1 we end up with

16π Γvac (σ → π + π − )Γ(σ → e+ e− )

σres (π + π − → e+ e− ) = (III.8)
(M 2 − m2σ )2 + (mσ Γtot 2
2
4mπ
1− 2 σ )
M

The above cross section depends on the available pionic phase space as 1/a thus it blows
up as a → 0.

B. Born approximation

It is instructive to compare (III.8) with what is found from a tree-level process. In

the following I assume a diagram with two vertex, one connecting incoming pions to the
intermediate photon and the other one connecting photo to massless di-electrons. The
amplitude of this process reads2

α e2
Mborn = (k α − k 0 ) ū(q 0 )γα v(q),
p2

2
I am neglecting i factors which are not important here.

9
hence working in pions CM frame which in p = (M, 0)
!2 Z
1 e2 X  α 0α 0
2 d3 q d3 q 0
σborn = (k − k )ū(q )γα v(q) (2π)4 δ 4 (k − q − q 0 )

4ωω 0 vrel p2 spins (2π)3 2q0 (2π)3 2q00
! Ward! Ward!
2
1 (4πα) z}|{ z}|{
= (2k − p )(2k − pβ )
α α β
M4
r
2
4mπ
M 22 1 − 2
M

d3 q d3 q 0
Z
× 4 qα qβ0 + qβ qα0 − gαβ (q.q 0 ) 4 4 0


3 30 (2π) δ (k − q − q )
(2π) 2q0 (2π) 2q0
32α2
d3 q d3 q 0 4
Z
= r α β
k k qα qβ0 + qβ qα0 − gαβ (q.q 0 ) 0
0 δ (k − q − q )
6 4mπ
2 2q0 2q0
M 1− 2
M
2
d3 q 4 0
Z
32α 2
α β
d q qα qβ0 + qβ qα0 − gαβ (q.q 0 ) δ(q 0 )θ(q00 )δ 4 (k − q − q 0 )


= r k k
6 4mπ
2 2q0
M 1− 2
M

16α2 d3 q
Z
kα kβ qα pβ + qβ pα − 2qα qβ − gαβ p.q δ(M 2 − 2M q0 )θ(M − q0 )


= r
4mπ
2 q0
M6 1− 2
M
q0 =|q|
z}|{
2
q02 dq0 dΩ
δ(q0 − M/2)
Z
16α
= r kα kβ qα pβ + qβ pα − 2qα qβ − gαβ p.q θ(M − q0 )
2
4mπ q0 2M
M6 1− 2
M
2 Z
32πα
kα kβ


= r q0 dq0 qα pβ + qβ pα − 2qα qβ − gαβ p.q δ(q0 − M/2)θ(M − q0 )
2
4mπ
M7 1− 2
M
2
32πα
≡ r k α k β Eαβ ,
2
4mπ
M7 1− 2
M

with
Z


Eαβ = q0 dq0 qα pβ + qβ pα − 2qα qβ − gαβ p.q δ(q0 − M/2)θ(M − q0 ).

An appropriate ansatz is

Eαβ = Epα pβ + M 2 E 0 gαβ

10
Using Ward identity pα Eαβ = 0 one finds E 0 = −E. Then contracting Eαβ with g αβ gives

1 αβ
E =− 2gEαβ
3M Z
1 αβ

=− 2g q0 dq0 qα pβ + qβ pα − 2qα qβ − gαβ p.q δ(q0 − M/2)θ(M − q0 )
3M Z
2
= q0 dq0 p.q δ(q0 − M/2)θ(M − q0 )
3M 2 Z
2
= q0 dq0 M q0 δ(q0 − M/2)θ(M − q0 )
3M 2
M
= ,
6
hence
M
Eαβ = (p p − M 2 gαβ )
6 α β
Putting this result in σborn formula gives

32πα2 M
σborn = kα kβ (p p − M 2 gαβ )
6 α β
r
2
4mπ
M7 1− 2
M
2
16πα  
= r (p.k)2 − M 2 m2π
2
4mπ
3M 6 1− 2
M
2
16πα  
= r (p0 k0 )2 − M 2 m2π
2
4mπ
3M 6 1− 2
M

16πα2
 
M 2 2 2
= r (M ) − M mπ
4mπ
2 2
3M 6 1− 2
M

4m2π
!
16πα2 M4
= 1− ,
M2
r
6 4mπ
2 4
3M 1− 2
M

and finally
s
4πα 2
4m2π
σborn = 1− (III.9)
3M 2 M2

11
 C. Weldon-Breit-Wigner formula

With (III.6) and (III.8) in hands I am in a position to prove Weldon-Breit-Wigner formula

for sigma-meson resonance
s
dR M 2
4m2π T ln Y
|res = 1− σres
d4 p 2p̄(2π)5 M 2 eĒ/T − 1
s  
2 2 + − + −
M 4m π T ln Y  16π  Γvac (σ → π π )Γ(σ → e e )
= 1 −
2p̄(2π)5 M 2 eĒ/T − 1 1 − 4m2π (M 2 − m2σ )2 + (mσ Γtot 2
2
σ )
M
2 −1/2
!
M 2
4mπ T ln Y Γvac (σ → π + π − )Γ(σ → e+ e− )
= 1 −
4p̄π 4 M2 2 2 2
eĒ/T − 1 (M − mσ ) + (mσ Γσ )
tot 2

Before going further let’s pay our attention to the factor

!−1/2
4m2π
1− ,
M2

just around resonance i.e M = mσ (1 + ∆) with ∆  1:

!−1/2 !−1/2
4m2π 4m2π
1− = 1−
M2 m2σ (1 + ∆)2
!−1/2
1 − a2
= 1−
(1 + ∆)2
 −1/2
= a2 + 2∆ + ∆2 (1 + ∆)
!!
1 ∆2
= 1−O
a a2

Now using (II.3), (II.4) and using results of [2]

Γ(ππ → σ) = e−Ē/T Γ(σ → ππ)

λ2 T ln Y
= Ē/T
16πM p̄ e −1
T ln Y 16πmσ Γvac (σ → ππ)
= Ē/T =⇒
e −1 16πM p̄a
T ln Y
Ē/T
Γvac (σ → ππ) = p̄a(1 + ∆)Γ(ππ → σ),
e −1

12
therefore
2
  
dR m2σ (1 + ∆) 3
1−O ∆
2 Γ(π + π − → σ)Γ(σ → e+ e− )
a
|res =
d4 p 4π 4 (M 2 − m2σ )2 + (mσ Γtot
σ )
2

m2 Γ(π + π − → σ)Γ(σ → e+ e− )  3

= σ4 1 + O(∆ )
4π (M 2 − m2σ )2 + (mσ Γtot
σ )
2

And finally we arrive at the desired result

m2σ Γ(π + π − → σ)Γ(σ → e+ e− )

 
dR
= (III.10)
d4 p res 4π 4 (M 2 − m2σ )2 + (mσ Γtot
σ )
2

D. Exactly on resonance

The total decay widths is given by

Γtot = Γ(σ → π + π − ) − Γ(π + π − → σ) + Γ(σ → e+ e− )

≈ Γ(σ → π + π − ) − Γ(π + π − → σ)
= Γ(σ → π + π − )(1 − e−Ē/T )

Therefore
m2σ e−Ē/T Γ(σ → π + π − )Γ(σ → e+ e− )
 
dR
= 4
d4 p 4π m2 Γ(σ → π + π − )2 1 − e−Ē/T 2
 
M =mσ σ

1 e−Ē/T Γ(σ → e+ e− )
= 4
4π Γ(σ → π + π − ) 1 − e−Ē/T 2
 

1 Γ(σ → e+ e− ) 1
= 4 + −
  
4π Γ(σ → π π ) eĒ/T 1 − e−Ē/T 1 − e−Ē/T
1 Γ(σ → e+ e− ) 1
= 4 + −
  
4π Γ(σ → π π ) eĒ/T − 1 1 − e−Ē/T

1 Γ(σ → e+ e− ) 1
= 4 + −
   ,
4π Γ(σ → π π ) e−Ē/(2T ) eĒ/T − 1 eĒ/(2T ) 1 − e−Ē/T

and finally

Γ(σ → e+ e− )
 
dR 1
= (III.11)
d4 p Γ(σ → π + π − )
 
M =mσ 16π 4 sinh2 Ē
2T

In next part I will calculate the leptonic decay in order to get results from above formula
in limit of a  1.

13
 APPENDIX A: IDENTITIES AND RELATIONS FOR DISTRIBUTION
FUNCTIONS

1. Useful identities

Identity 1.

n− (−ω) = −[1 + n+ (ω)] (A.1)

Proof.

eβ(ω−µ)
 
1
−[1 + n+ (ω)] = − 1 + =−
eβ(ω−µ) − 1 eβ(ω−µ) − 1
eβ(ω−µ) e−β(ω−µ)
=
−e−β(ω−µ) eβ(ω−µ) + e−β(ω−µ)
1 1
= −β(ω−µ)
= β(−ω+µ)
−1 + e e −1
= n− (−ω)

Identity 2.

n+ (ω − E) = −[1 + n+ (E − ω)] (A.2)

Identity 3.

[1 + n+ (ω)][1 + n− (E − ω)] = n− (−ω)n+ (ω − E) (A.3)

Identity 4.

eβE
n− (−ω)n+ (ω − E) = [n+ (ω − E) − n+ (ω)] (A.4)
1 − eβE
Proof. First observer that
1
n− (−ω) = −β(ω−µ)
=⇒
e −1
n− (−ω) 1
= −β(ω−µ) =⇒
n− (−ω) + 1 e
β(ω−µ) n− (−ω)
e = .
n− (−ω) + 1
It is also straightforward to show
 
β(ω−µ) βE 1
e =e 1+
n+ (ω − E)

14
Therefore
 
n− (−ω) βE 1
=e 1+
n− (−ω) + 1 n+ (ω − E)
n− (−ω)n+ (ω − E) = eβE [1 + n+ (ω − E)] [n− (−ω) + 1]
n− (−ω)n+ (ω − E) = eβE [1 + n+ (ω − E) + n− (−ω) + n+ (ω − E)n− (−ω)]
 
[n− (−ω)n+ (ω − E)] 1 − eβE = eβE [1 + n+ (ω − E) + n− (−ω)]

eβE
n− (−ω)n+ (ω − E) = [n+ (ω − E) − n+ (ω)]
1 − eβE

Identity 5.
1
n+ (ω)n− (E − ω) = [n+ (ω − E) − n+ (ω)] = e−βE n− (−ω)n+ (ω − E) (A.5)
1 − eβE

2. Integration formulae

Integration formula 1.
Z h i
dω n+ (ω) = T ln e−β(ω−µ) − 1 = −T ln (n− (−ω)). (A.6)

[1] H. A. Weldon, “Dilepton enhancement at 2m(pi) and chiral symmetry restoration,” Phys. Lett.
B 274, 133 (1992). doi:10.1016/0370-2693(92)90511-2
[2] H. A. Weldon, “Simple Rules for Discontinuities in Finite Temperature Field Theory,” Phys.
Rev. D 28, 2007 (1983). doi:10.1103/PhysRevD.28.2007

15