An Introduction to Quantum Field Theory (Peskin and Schroeder)

Solutions

Andrzej Pokraka

December 12, 2017

Contents
1 The Dirac Equation 1
1.1 Lorentz group ! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Gordon Identity ! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3 Spinor products ! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Majorana fermions ! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5 Supersymmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.6 Fierz transformations ! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.7 Discrete symmetries of the Dirac field ! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
1.8 Bound states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

1 The Dirac Equation

1.1 Lorentz group !

The Lorentz commutation relations are

[J µν , J ρσ ] = i (g νρ J µσ − g µρ J νσ − g νσ J µρ + g µσ J νρ ) .

(a) Define the generators of rotations and boots as

1 ijk jk
Li = ϵ J and K i = J 0i ,
2
where ijk is a permutation of (123). An infinitesimal Lorentz transformation can then be written

Φ → (1 − iθ · L − iβ · K) Φ.

Write the commutation relations of these vector operators explicitly. Show that the combinations
1 1
J+ = (L + iK) and J− = (L − iK)
2 2

1
commute with one another and separately satisfy the combination relation of angular momentum.
Proof: The L operator commutation relations are
! i j"
L ,L = Li Lj − Lj Li
1 ! ikl kl jmn mn "
= ϵ J ,ϵ J
4
1 ikl jmn ! kl mn "
= ϵ ϵ J ,J
4
i ikl jmn ! lm kn
g J − g km J ln − g ln J km + g kn J lm
"
= ϵ ϵ
4
i ! ikl jmn lm kn
g J − ϵikl ϵjmn g km J ln − ϵikl ϵjmn g ln J km + ϵikl ϵjmn g kn J lm
"
= ϵ ϵ
4
for the second and last term k ↔ l
i ! ikl jmn lm kn
g J − ϵilk ϵjmn g lm J kn − ϵikl ϵjmn g ln J km + ϵilk ϵjmn g ln J km
"
= ϵ ϵ
4
i ! ikl jmn lm kn
g J + ϵikl ϵjmn g lm J kn − ϵikl ϵjmn g ln J km − ϵikl ϵjmn g ln J km
"
= ϵ ϵ
4
for the third and last term m ↔ n
i ! ikl jmn lm kn
g J + ϵikl ϵjmn g lm J kn − ϵikl ϵjnm g lm J kn − ϵikl ϵjnm g lm J kn
"
= ϵ ϵ
4
i ! ikl jmn lm kn
g J + ϵikl ϵjmn g lm J kn + ϵikl ϵjmn g lm J kn + ϵikl ϵjmn g lm J kn
"
= ϵ ϵ
4
= iϵikl ϵjmn g lm J kn
= iϵikl ϵjmn −δ lm J kn
# $

= −iϵikl ϵjln J kn
= iϵikl ϵjnl J kn
i δ ij δ kn − δ in δ kj J kn
# $
=
i δ ij J kk − J ji
# $
=
= iJ ij
= iϵijk Lk

where
kk
= i δαk δβk − δβk δαk
# $
Jαβ
= 0

and
1 ijk klm lm
ϵijk Lk = ϵ ϵ J
2
1 kij klm lm
= ϵ ϵ J
2
1 # il jm
δ δ − δ im δ jl J lm
$
=
2
1 # ij
J − J ji
$
=
2
= J ij .

2
The K commutation relations are

K i, K j = K iK j − K j K i
! "

= J 0i J 0j − J 0j J 0i
= i g i0 J 0j − g 00 J ij − g ij J 00 + g 0j J i0 .
# $

This is simplified using properties of the metric g i0 = 0, g 00 = −1, g ij = −1 and the generators J 00 = 0
! i j"
K ,K = −iJ ij
= −iϵijk Lk .

Next, we need
! i j" 1 ikl ! kl 0j "
L ,K = ϵ J ,J
2
i ikl # lj k0
−g J + g kj J l0
$
= ϵ
2
i ikl # lj k0
δ J − δ kj J l0
$
= ϵ
2
i # ikj k0
ϵ J − ϵijl J l0
$
=
2
i # ijk 0k
ϵ J + ϵijk J 0k
$
=
2
= iϵijl K l .

Lastly we compute the angular momentum commutators

% &
1 1
[J+ , J− ] = (L + iK) , (L − iK)
2 2
1
= {[L, L] − i [L, K] + i [K, L] + [K, K]}
4
i
= [K, L]
2
i ! i j"
= K , L êi · êj
2
i ! i i"
= K ,L
2
= 0
% &
'
i j
( 1# i i
$ 1# j j
$
J± , J± = L ± iK , L ± iK
2 2
1 )! i j "
L , L ± i Li , K j ± i K i , Lj − K i , K j
! " ! " ! "*
=
4
1 ) ijk k
iϵ L ± i iϵijl K l ∓ i iϵjil K l + iϵijk Lk
# $ # $ *
=
4
) ijk k
iϵ L ∓ ϵijl K l ± ϵijl K l + iϵijk Lk
*
=
1 ) ijk k
iϵ L ∓ ϵijl K l
*
=
2
i ijk ) k
L ± iK k
*
= ϵ
2
i ijk k
= ϵ J±
2

3
(b) The finite-dimensional representations of the rotation group correspond precisely the to the allowed values for angular
momentum: integers or half-integers. The result of part (a) implies that all finite-dimensional representations of the
Lorentz group correspond to pairs of integers or half integers, (j+ , j− ), corresponding to pairs of representations of the
rotation group. Using the fact that J = σ/2 in the spin-1/2 representation of angular $ momentum, write explicitly the
transformation laws of the 2-component objects transforming according to the 12 , 0 and 0, 12 representations of the
# # $

Lorentz group. Show that these correspond precisely to the transformations of ψL and ψR giving in (3.37).
Proof: The representations of the Lorentz group are denoted by (m, n) ≡ πm,n where m, n are either half-integers or
integers. The irreducible representations are given by
(n) (m)
πm,n Li =
# $
I(2m+1) ⊗ Ji + Ji ⊗ I(2n+1)
+ ,
(n) (m)
πm,n K i =
# $
i I(2m+1) ⊗ Ji − Ji ⊗ I(2n+1) .
#1 $
With J = σ/2 the 2,0 representation is found to be

σi σi σi σi
- .
πm,n Li =
# $
I2 ⊗ I1 + ⊗ I1 = I2 + ⊗ I1 = ⊗ I1 =
2 2 2 2
σi σi σi σi
- . - .
πm,n K i =
# $
i I2 ⊗ I1 − ⊗ I1 = i I2 − ⊗ I1 = −i ⊗ I1 = −i .
2 2 2 2

Thus, with Li = σ i /2 and K i = −iσ i /2 the transformation law becomes

i µν
Φ( 1 ,0) → e− 2 (ωµν J ) Φ( 1 ,0)
2 2
- .
i µν
= 1 − ωµν J Φ( 1 ,0)
2 2
- .
i i
= 1 − ω0ν J 0ν − ωiν J iν Φ( 1 ,0)
2 2 2
- .
i i i i
= 1 − ω00 J 00 − ω0i J 0i − ωi0 J i0 − ωij J ij Φ( 1 ,0)
2 2 2 2 2
- .
i i i
= 1 − ω0i J 0i − ω0i J 0i − ωij J ij Φ( 1 ,0)
2 2 2 2
- .
i
= 1 − iω0i K i − ωij ϵijk Lk Φ( 1 ,0)
2 2

= (1 − iβ · K − iθ · L) Φ( 1 ,0)
2
- .
1 i
= 1 − β · σ − θ · σ Φ( 1 ,0)
2 2 2

where we have defined β i = ω0i = −ωi0 and θk = ωij ϵijk .

Now for the 0, 12 representation we have
# $

σi
- i
σi
.
σ
πm,n Li =
# $
I1 ⊗ + I1 ⊗ I2 = I1 ⊗ + I2 =
2 2 2
i
- i
σi
- . .
σ σ
πm,n K i =
# $
i I1 ⊗ − I1 ⊗ I2 = i − I2 ⊗ I1 = i ⊗ I1 .
2 2 2

4
Thus, with Li = σ i /2 and K i = iσ i /2 the transformation law becomes
i µν
Φ(0, 1 ) → e− 2 (ωµν J )
Φ(0, 1 )
2 2

= (1 − iβ · K − iθ · L) Φ(0, 1 )
2
- .
1 i
= 1 + β · σ − θ · σ Φ(0, 1 ) .
2 2 2

Upon comparison with equation (3.37) we identify

Φ( 1 ,0) = ψL and Φ(0, 1 ) = ψR .

2 2

Thus, the left- and right-handed spinor transform according to separate representations of the Lorentz group.
"
(c) The identity σ T = −σ 2 σσ 2 allows us to rewrite the ψL transformation in the unitarily equivalent form

ψ ′ → ψ ′ (1 + iθ · σ/2 + β · σ/2) ,

where ψ ′ = ψL σ . Using this law, we can represent the object that transforms as 12 , 12 as a 2 × 2 matrix that has the ψR
T 2
# $

transformation law on the left and simultaneously, the transposed ψL transforms on the right. Parametrize this matrix as
- 0
V + V 3 V 1 − iV 2
.
.
V 1 + iV 2 V 0 − V 3

Show that the object V µ transforms as a 4-vector.

Proof: Left-handed spinors transform according to
- .
1 i
ψL → 1 − β · σ − θ · σ ψL .
2 2

With ψ ′ = ψL
T 2
σ we verify the transformation law
-- . .T
1 i
ψ′ → 1 − β · σ − θ · σ ψL σ2
2 2
- .T
T 2 2 1 i
= ψL σ σ 1 − β · σ − θ · σ σ2
2 2
- .
′ 2 1 T i T
= ψ σ 1− β·σ − θ·σ σ2
2 2
- .
′ 1 i
= ψ 1+ β·σ+ θ·σ .
2 2

Now we are interested in the transformation properties of a 12 , 12 object. We parameterize the 21 , 12 object as the matrix
# $ # $

- 0
V + V 3 V 1 − iV 2
.
Φ( 1 , 1 ) = = V µ σµ
2 2 V 1 + iV 2 V 0 − V 3

5
where it will be shown that V µ is a 4-vector. Applying the transformation to Φ( 1 , 1 ) we get
2 2

- . - .
1 i 1 i
Φ( 1 , 1 ) → 1 + β · σ − θ · σ Φ( 1 , 1 ) 1 + β · σ + θ · σ
2 2 2 2 2 2 2 2
- . - .
1 1
= 1 + (β − iθ) · σ V µ σµ 1 + (β · σ + iθ) · σ
2 2
1 1
= V µ σµ + (β − iθ) · σV µ σµ + V µ σµ (β + iθ) · σ + O θ2 , β 2
# $
2 2
µ Vµ Vµ
= V σµ + β · (σσµ + σµ σ) − iθ · (σσµ − σµ σ)
2 2
µ µ
V V
= V µ σµ + β i · {σi , σµ } − iθi · [σi , σµ ]
2 2
V0 i Vj i V0 i Vj i
= V µ σµ + β · {σi , σ0 } + β · {σi , σj } − iθ · [σi , σ0 ] − iθ · [σi , σj ]
2 2 2 2
V0 i Vj i V0 i Vj i
= V µ σµ + β · {σi , I} + β {σi , σj } − iθ [σi , I] − iθ [σi , σj ]
2 2 2 2
V0 i Vj i Vj i#
= V µ σµ + iθ 2iϵijk σ k
$
β (2σi ) + β (2δij ) −
2 2 2
= V µ σµ + V 0 β i σi − V i βi + V j θi ϵijk σ k

where {σi , I} = 2σ i , [σi , I] = 0, {σi , σj } = 2δij , [σi , I] = 2iϵijk σ k . Also note that β i V j δij = −β i Vi because gij = −δij .
Recall that we have defined the anti-symmetric tensor ω0i = βi and ωij = ϵijk θk . Inserting these expressions into the
above, we have

V µ σµ + V 0 β i σi − V i βi + V j θi ϵijk σ k = V µ σµ + V 0 ω0i σ i − V i ω0i + V j ωjk σ k

= V µ σµ + V 0 ω0i σ i + V i ωi0 σ 0 + ωij V i σ j
= V µ σµ + ωµν σ ν V µ
# ν
δµ + ω νµ Vν σ µ .
$
=

We would like to show that this is identical to equation (3.19) in P&S. P&S assert that a 4-vector V µ transforms as follows
- .
i µν α
α α
V → δβ − ωµν (J )β V β
2
α
for (J µν )αβ = δαµ δβν − δαν δβµ . With this definition of (J µν )β the transformation condition becomes

i i
δβα − ωµν (J µν )α
β = δβα − ωµν (J µν )γβ g γα
2 2
i + ,
= δβα − ωµν δγµ δβν − δγν δβµ g γα
2
i
α
= δβ − (ωγβ − ωβγ ) g γα
2
= δβα − iωγβ g γα
= δβα − iωβα .

This is the identical result obtained from transforming V µ σµ . Thus, we see that V µ is indeed a 4-vector.
"

6
1.2 Gordon Identity !
Derive the Gordon identity,
p′µ + pµ iσ µν qν
% &
ū (p′ ) γ µ u (p) = ū (p′ ) + u (p) ,
2m 2m
where q = (p′ − p).
Proof:
The computation is straightforward:
i µ ν ′
σ µν qν = [γ , γ ] (pν − pν )
2
i µ ′ ′ µ i µ µ
= (γ p✁ − p✁ γ ) − 2 (γ p ✁ − pγ
✁ )
2
i i µ
= (2g µν p′ν − p ′ µ
✁γ −p
′ µ
✁ γ ) − 2 (γ p
µν µ
✁ − 2g pν + γ p)
✁
2
′µ ′ µ µ µ
= i (p − p ✁ γ ) − i (γ p ✁−p )
p′µ + pµ iσ µν qν p′µ + pµ − (p′µ − p ′ µ µ µ
% & % &
ū (p′ ) + u (p) = ū (p′ ) ✁ γ ) + (γ p✁ − p ) u (p)
2m 2m 2m
% ′ µ µ
&
p γ + γ ✁p
= ū (p′ ) ✁ u (p)
2m
mγ µ + γ µ m
% &
= ū (p′ ) u (p)
2m
= ū (p′ ) γ µ u (p)
where we have used the Fourier transformed Dirac equation and it’s adjoint equation

✁ − m) u (p) = 0 and ū (p) (p

(p ✁ − m) = 0.
"

1.3 Spinor products !

Together is Problems 5.3 and 5.6 we develop an efficient computational method for processes involving massless particles.
Let k0µ , k1µ be fixed 4-vectors satisfying k02 = 0, k12 = −1, k0 · k1 = 0. Define basic spinors in the following way: Let uL0
be
# 2the left-handed
$ spinor for for a fermion with momentum k0 . Let uR0 = k✁1 uL0 . Then, for any p such that p is lightlike
p = 0 define
1 √ 1
uL (p) = √2p·k ✁ R0 and uR (p) = 2p·k pu
pu
0 ✁ L0 . 0

This set of conventions defines the phases of spinors unambiguously (except when p is parallel to k0 ).
(a) Show that k✁0 uR0 = 0. Show that, for any lightlike p, pu
✁ R (p) = 0.
Proof:

k✁0 uR0 = k✁0 k✁1 uL0

= k0µ k1ν γµ γν uL0
= k0µ k1ν (2gµν − γν γµ ) uL0
= 2k0 · k1 uL0 − k✁1 k✁0 uL0
= 0

7
where we have used the Dirac equation for a massless particle k✁0 uL0 = 0 and the dot product k0 · k1 . Now for any lightlike
4-momentum p the definition of the spinors above satisfy the massless Dirac equation
1
pu
✁ R (p) = √ ppuL0
2p · k0 ✁✁
1
= √ p2 uL0
2p · k0
= 0.

"
(b) For the choices k0 = (E, 0, 0, −E), k1 = (0, 1, 0, 0), construct uL0 , uR0 , uL (p), and uR (p) explicitly.
Proof: For this problem we will need the operator k✁ which in the Chiral or Weyl basis is given by

k0 − k · σ
- .
0
k✁ =
k0 + k · σ 0
0 3
⎛ # 1 2
$ ⎞
0 0 #k 1− k 2 $ − k0 − ik
⎜ 0 0 − k + ik k + k3 ⎟
= ⎜ ⎝ k 0 + k 3 k 1 − ik 2
⎟.
0 0 ⎠
1 2 0 3
k + ik k −k 0 0

From equation (3.50) in P&S the Dirac spinors for are given by
- √ .
u (k0 ) = √k0 · σξ
k0 · σ̄ξ
- 5 .
0 3
= 5E (σ + σ )ξ
E (σ 0 − σ 3 )ξ
⎛ 6- . ⎞
2 0
ξ ⎟
√ ⎜ 0 0
⎜
⎟
= E ⎜ 6-
⎜ . ⎟
0 0
⎟
ξ
⎝ ⎠
0 2
⎛ - . ⎞
1 0
√ ⎜ ξ ⎟
= 2E ⎜ - 0 0 . ⎟
⎝ 0 0 ⎠
ξ
0 1
7# $T # $T 8 $T
and σ̄ = σ 0 − σ . Taking ξ = 1 0
# $ #
where ξ ∈ 1 0 , 0 1 the left-handed spinor is
⎛ - .- . ⎞ ⎛ ⎞
1 0 1 1
√ ⎜ √ ⎜
uL0 = 2E ⎜ - 0 0 .- 0 . ⎟
⎟ = 2E ⎜ 0 ⎟
⎟
⎝ 0 0 1 ⎠ ⎝ 0 ⎠
0 1 0 0
# $T
while the right-handed spinor is given by taking ξ = 0 1
⎛ - .- . ⎞ ⎛ ⎞
1 0 0 0
√ ⎜ √ ⎜
uR0 = 2E ⎜ - 0 0 .- 1 . ⎟
⎟ = 2E ⎜ 0 ⎟
⎟.
⎝ 0 0 0 ⎠ ⎝ 0 ⎠
0 1 1 1

8
The chiral spinors for any momentum p are then given by
1
uL (p) = √ puR0
2p · k0 ✁
0 3
⎛ # 1 2
$ ⎞⎛ ⎞
√ 0 0 #p 1− p 2 $ − p0 − ip 0
2E ⎜ 0 0 − p + ip p + p3 ⎟⎜ 0 ⎟
= ⎜ 0 3
⎟⎜ ⎟
2E (p0 + p3 ) ⎝ p + p p1 − ip2 0 0 ⎠⎝ 0 ⎠
5
p + ip2
1
p0 − p3 0 0 1
−p1 + ip2
⎛ ⎞
9
1 ⎜ p0 + p3 ⎟
= ⎜ ⎟
p0 + p3 ⎝ 0 ⎠
0

1
uR (p) = √ puL0
2p · k0 ✁
0 3
⎛ # 1 2
$ ⎞⎛ ⎞
√ 0 0 #p 1− p 2 $ − p0 − ip 1
2E ⎜ 0 0 3
⎜ 0 − p + ip p + p3 ⎟⎜ 0 ⎟
= 5 ⎟⎜ ⎟
2E (p + p ) ⎝ p + p
0 3 p1 − ip2 0 0 ⎠⎝ 0 ⎠
p1 + ip2 p0 − p3 0 0 0
⎛ ⎞
9 0
1 ⎜ 0 0 3 ⎟
⎜ ⎟
= 0
p +p 3 ⎝ p +p ⎠
p1 + ip2

"
(c) Define the spinor products s (p1 , p2 ) and t (p1 p2 ), for p1 , p2 lightlike, by

s (p1 , p2 ) = ūR (p1 ) uL (p2 ) and t (p1 , p2 ) = ūL (p1 ) uR (p2 ) .

Using the explicit forms for the uλ given in part (b), compute the spinor products explicitly and show that t (p1 , p2 ) =
∗
s (p1 , p2 ) and s (p1 , p2 ) = −s (p2 , p1 ). In addition, show that

|s (p1 , p2 )|2 = 2p1 · p2 .

Thus the spinor products are the square roots of 4-vector dot products.
Proof:

9
 s (p1 , p2 ) = ūR (p1 ) uL (p2 )
−p12 + ip22
⎛ ⎞⎛ ⎞
6 0 0 1 0
1 0 3
⎜ 0 0 0 1 ⎟
⎟ ⎜ p2 + p2
⎜ ⎟
0 p01 + p31 p11 − ip21 ⎜
# $
= 0 ⎟
(p01 + p31 ) (p02 + p32 ) ⎝ 1 0 0 0 ⎠⎝ 0 ⎠
0 1 0 0 0
⎛ ⎞
6 0
1 ⎜ 0 ⎟
0 p01 + p31 p11 − ip21 ⎜
# $
= 0 3 0 3
0 ⎝ −p12 + ip22 ⎠
⎟
(p1 + p1 ) (p2 + p2 )
p02 + p32
# 0
p1 + p31 −p12 + ip22 + p11 − ip21 p02 + p32
$ # $ # $ # $
= 5
(p01 + p31 ) (p02 + p32 )
−p01 p12 − p31 p12 + ip01 p22 + ip31 p22 + p11 p02 − ip21 p02 + p11 p32 − ip21 p32
= 5
(p01 + p31 ) (p02 + p32 )
p11 p02 + p32 − p01 − p31 p12 + i p01 + p31 p22 − p21 p02 + p32
# $ # $ ## $ # $$
= 5
(p01 + p31 ) (p02 + p32 )
= −s (p2 , p1 )

: p p + p3 − p0 − p3 p1 + i p0 + p3 p2 − p2 p0 + p3 : 2
: 1# 0 $ # $ ## $ # $$:
2 1 2 2 1 1 2 1 1 2 1 2 2
|s (p1 , p2 )| =
(p01 + p31 ) (p02 + p32 )
: p p + p3 − p0 − p3 p1 + i p0 + p3 p2 − p2 p0 + p3 : 2
: 1# 0 $ # $ ## $ # $$:
1 2 2 1 1 2 1 1 2 1 2 2
=
(p01 + p31 ) (p02 + p32 )
# 1# 0 $ $2 ## $$2
p1 p2 + p32 − p01 − p31 p12 + p01 + p31 p22 − p21 p02 + p32
$ # $ #
=
(p01 + p31 ) (p02 + p32 )
# 1 $2 # 0 2 $ # $2
p2 + p32 − 2p11 p02 + p32 p01 − p31 p12 − + p01 − p31 p12
$ # $# $ #
p1
=
(p01 + p31 ) (p02 + p32 )
# 0 2 2 $ # $2 # 0 $2
p1 + p31 p22 − 2 p01 + p31 p22 p21 p02 + p32 + p21 p2 + p32
$ # $ # $ #
×
(p01 + p31 ) (p02 + p32 )
# 0
p2 + p32 +# 1 $2 # 2 $2 ,
# 0
p1 + p31 +# 1 $2 # 2 $2 ,
$ $
− 2 p11 p12 + p22 p21
# $
= 0 3 p1 + p1 + 0 3 p2 + p2
(p + p1 ) (p + p2 )
# 01 3 +# $
# 20
p1 + p31 +# 1 $2 # 2 $2 ,
$ $
p2 + p2 1 2
# 2 $2 ,
− 2 p11 p12 + p22 p21
# $
= p1 + p1 + 0 p2 + p2
(p01 + p31 ) (p2 + p32 )
Since p1 and p2 are lightlike
# 1 $2 # 2 $2 # 0 $2 # 3 $2 # 0
pi + pi = pi − pi = pi + p3i p0i − p3i
$# $

we have
2 # 0
p2 + p32 p01 − p31 + p01 + p31 p02 − p32 − 2 p11 p12 + p22 p21
$# $ # $# $ # $
|s (p1 , p2 )| =
p01 p02 + p01 p32 − p31 p02 − p31 p32 + p01 p02 + p31 p02 − p01 p32 − p31 p32 − 2 p11 p12 + p22 p21
# $
=
2p01 p02 − 2 p11 p12 + p22 p21 + p31 p32
# $
=
= 2p1 · p2 .

"

10
1.4 Majorana fermions !

Recall form Eq. (3.40) that one can write a relativistic equation for massless 2-component fermion field that transforms
as the upper two components of a Dirac spinor (ψL ). Call such a 2-component field χa (x), a = 1, 2.
(a) Show that it is possible to write an equation for χ (x) as a massive field in the following way:

iσ̄ · ∂χ − imσ 2 χ∗ = 0.

That
# 2 is,2 $show, that this equation is relativistically invariant and, second, that it implies the Klein-Gordon equation,
∂ + m χ = 0. This form of the fermion mass is called Majorana mass term.
Proof:
The unitary matrix, Λ 12 , which Lorentz transforms a fermion field (spinor) is given by equation (3.30)
- . ; Λ 0
<
i µν ( 21 ,0)
Λ 12 (θ, β) = exp − ωµν S =
2 0 Λ(0, 1 )
2

where
i µ ν
S µν = [γ , γ ]
4
is the Lorentz transformation generator for spinor, β i = ω0i = −ωi0 and θk = ωij ϵijk . This transformation matrix is
block diagonal. This is seen from the block diagonal form of the generators of boost (3.26) and (3.27)
- i .
0i i σ 0
S = −
2 0 −σ i
1 ijk k
S ij = ϵ Σ
2
where the spin operator of Dirac theory is
σi
- .
k 0
Σ = .
0 σi
The blocks Λ( 1 ,0) and Λ(0, 1 ) are the left and right handed representations of the Lorentz group. A spinor which
2 2
transforms according to Λ 21 is called a Dirac spinor.
Because of the block diagonal form left and right spinors transform in different representations of the Lorentz group
; <- . ; <
Λ( 1 ,0) 0 ψL Λ( 1 ,0) ψL
Λ 21 ψ = 2
= 2
.
0 Λ(0, 1 ) ψR Λ(0, 1 ) ψR
2 2

Under infinitesimal rotations by θ and boosts β the transformation laws for the left and right handed spinors are
- .
1 i
Λ( 1 ,0) = 1− β·σ− θ·σ
2 2 2
- .
1 i
Λ(0, 1 ) = 1− β·σ+ θ·σ .
2 2 2

Now equation (3.40) tells us that we can write the relativistic field equation for a massless fermion which transforms under
Λ( 1 ,0) as iσ̄ · ∂χ = 0. Furthermore, we know how γ µ transforms. This is given by equation (3.29)
2

Λ−1 µ µ ν
1 γ Λ1 = Λ νγ .
2
2

On the left the Lorentz transformations are acting on the spinor indices of γ µ while on the right the Lorentz transformation

11
acts on the space time index µ. We can work out the transformation for σ̄ because σ̄ µ is just the off-diagonal blocks of γ µ

Λµν γ ν = Λ−1 µ
1 γ Λ1
2
⎛2 ⎞
-
0 Λµν σ ν
. 0 Λ−1 σ µ
Λ (0, 2 ) ⎠
1
( 2 ,0)
1
= ⎝ .
Λµν σ̄ ν 0 Λ−1 σ̄ µ Λ( 1 ,0) 0
(0, 12 ) 2

$ν
From the above we conclude that σ̄ µ Λ−1 µ = Λ(0, 1 ) σ̄ ν Λ−1
#
. We will also need the fact that if χtransforms according to
2 ( 21 ,0)
Λ( 1 ,0) then σ 2 χ∗ transforms under Λ(0, 1 ) (i.e., χ is left-handed and σ 2 χ∗ is right-handed). We prove this here. Applying
2 2
Λ( 1 ,0) to σ 2 χ∗ yields
2

-- . .∗
2 ∗ 2 1 i
σ χ → σ 1− β·σ− θ·σ χ
2 2
- .
1 i
2
= σ 1 − β · σ + θ · σ χ∗
∗ ∗
2 2
- .
1 i
= 1 + β · σ − θ · σ χ∗
2 2
∗
= Λ(0, 1 ) χ .
2

Lorentz transforms; review of a scalar field χ (no spatial orientation).

1. χ: Under Lorentz transformations the scalar χ transforms as χ (x) → χ′ (x′ ) = χ Λ−1 x′ where x → x′ = Λx.
# $

2. ∂µ χ: Under Lorentz transformations ∂µ χ transforms as

∂xν ∂ +#
−1 β
,
∂µ χ (x) → ∂µ′ χ′ (x′ ) = ′α
$
χ Λ x
∂x#′µ ∂x$ν α
−1 ν ′ρ
∂ Λ ρ
x ∂ +# $β ,
= ′µ ν
χ Λ−1 α x′α
∂x ∂x
# −1 $ν ρ ∂ +# $β ,
= Λ g
ρ µ ∂xν
χ Λ−1 α x′α
# −1 $ν
(∂ν χ) Λ−1 x′
# $
= Λ µ

where x → x′ = Λx. Notice that this is opposite to how the vector xµ transforms xµ → x′µ = Λµν xν . This shows
that while xµ is a contravariant vector, ∂µ is a covariant vector.
Applying the Lorentz transformation to the field equation we have
$ν
iσ̄ · ∂χ (x) − imσ 2 χ∗ (x) = 0 → iσ̄ µ Λ−1 µ ∂ν Λ( 1 ,0) χ Λ−1 x − imΛ(0, 1 ) σ 2 χ∗ Λ−1 x = 0
# # $ # $
2 2

$ν
where we know how ∂µ χ transforms. Using σ̄ µ Λ−1 µ = Λ(0, 1 ) σ̄ ν Λ−1
#
we conclude
2 ( 12 ,0)

iσ̄ · ∂χ (x) − imσ 2 χ∗ (x) = 0 → iΛ(0, 1 ) σ̄ ν Λ−1 ∂ν Λ( 1 ,0) χ Λ−1 x − imΛ(0, 1 ) σ 2 χ∗ Λ−1 x = 0
# $ # $
2 (2 )
1
,0 2 2

= Λ(0, 1 ) iσ̄ ν ∂ν χ Λ−1 x − imσ 2 χ∗ Λ−1 x = 0.

) # $ # $*
2

Thus, we see that the field equation iσ̄ · ∂χ (x) − imσ 2 χ∗ (x) = 0 is relativistically invariant.
Next, we must show that iσ̄ · ∂χ (x) − imσ 2 χ∗ (x) = 0 implies the Klein-Gordon equation. The equation −iσ̄ ∗ · ∂χ∗ (x) +
1 2
imσ 2∗ χ (x) = 0 implies χ∗ (x) = m σ σ̄ · ∂χ (x). Substitution into the complex conjugate of iσ̄ · ∂χ (x) − imσ 2 χ∗ (x) = 0

12
we have
iσ̄ µ∗ ∂µ χ∗ (x) − imσ 2∗ χ (x) = 0
1
iσ̄ µ∗ ∂µ σ 2 σ̄ ν ∂ν χ (x) − imσ 2∗ χ (x) = 0
m
σ 2 σ 2 σ̄ µ∗ σ 2 σ̄ ν ∂µ ∂ν χ (x) + m2 σ 2∗ χ (x) = 0
σ 2 σ µ σ̄ ν ∂µ ∂ν χ (x) + m2 σ 2∗ χ (x) = 0
σ µ σ̄ ν ∂µ ∂ν χ (x) − m2 χ (x) = 0
−g µν ∂µ ∂ν χ (x) − m2 χ (x) = 0
∂ 2 χ (x) + m2 χ (x) = 0
as required.
"
(b) Does the Majorana equation follow from a Lagrangian? The mass term would seem to be the variation of σ 2 ab χ∗a χ∗b ;
# $

however, since σ 2 is antisymmetric, this expression would vanish if χ (x) were an ordinary c-number field. When we go
to quantum field theory, we know that χ (x) will become an anti-commuting quantum field. Therefore, it makes sense to
develop its classical theory by considering χ (x) as a classical anti-commuting field, that is, as a field that takes as values
Grassmann numbers which satisfy
αβ = −βα for any α, β.
Note that this relation implies that α2 = 0. A Grassmann field ξ (x) can be expanded in a basis of functions as
=
ξ (x) = αn φn (x) ,
n

where the φn (x) are orthogonal c-number functions and the αn are a set of independent Grassmann numbers. Define the
complex conjugate of a product of Grassmann numbers to reverse the order:
(αβ)∗ ≡ β ∗ α∗ = −α∗ β ∗ .
This rule imitates the Hermitian conjugation of quantum fields. Show that the classical action,
% &
im # T 2
ˆ
S = d4 x χ† iσ̄ · ∂χ + χ σ χ − χ† σ 2 χ∗ ,
$
2
T
(where χ† = (χ∗ ) ) is real (S ∗ = S), and that varying this S with respect to χ and χ∗ yields the Majorana equation.
Proof:
The trick to this problem is realizing that S ∈ C and so S ∗ = S † . The complex conjugate of the action is therefore
> - .† ?
im
ˆ
†
S∗ d4 x χ† iσ̄ µ ∂µ χ + χT σ 2 χ − χ† σ 2 χ∗
# $ # $
=
2
% &
im # † 2† T †
ˆ
†
d4 x − (∂µ χ) iσ̄ µ† χ†† − χ σ χ − χ∗† σ 2† χ††
$
=
2
% &
im # † 2 ∗
ˆ
†
d4 x − (∂µ χ) iσ̄ µ χ − χ σ χ − χT σ 2 χ
$
=
2
% &
im # T 2
ˆ ' ( ˆ
†
d4 x − (∂µ χ) iσ̄ µ χ + d4 x χ σ χ − χ† σ 2 χ∗
$
=
2
% &
im # T 2
˛ ˆ ˆ
dSµ −iχ† σ̄ µ χ − d4 x − χ† iσ̄ µ ∂µ χ + d4 x χ σ χ − χ† σ 2 χ∗
! " ! # $" $
=IBP
2
% &
im
ˆ ˆ
d4 x χ† iσ̄ µ ∂µ χ + d4 x χT σ 2 χ − χ† σ 2 χ∗
!# $" # $
=
2
= S.

13
Notice that we have assumed that the field χ vanishes at the boundary of the integration region.
Recall that the action is the integral of the Lagrange density L. In our case we have
im # T 2
L = χ† iσ̄ · ∂χ + χ σ χ − χ† σ 2 χ∗
$
2
∗ µ im # 2 2 ∗ ∗
$
= χa iσ̄ab ∂µ χb + σab χa χb − σab χa χb
2
. The Euler-Lagrange equations are derived from variation of the action, requiring that to first order δχ that δS = 0.
However, we cannot just use the Euler-Lagrange equations because χ is a Grassmann field. Varying the action with respect
to χ∗ yields

δS ≡ S [χ∗ + δχ∗ ] − S [χ∗ ]

ˆ ˆ
= d4 xL (χ∗ + δχ∗ , ∂µ χ∗ + ∂µ δχ∗ ) − d4 xL (χ∗ , ∂µ χ∗ )
% &
im 2
ˆ
4 ∗ ∗ µ ∗ ∗ ∗ ∗
= d x (χa + δχa ) iσ̄ab ∂µ χb + σ (χa χb − (χa + δχa ) (χb + δχb ))
2 ab
% &
im 2
ˆ
4 ∗ µ ∗ ∗
− d x χa iσ̄ab ∂µ χb + σ (χa χb − χa χb )
2 ab
% &
im 2
ˆ
4 ∗ µ ∗ ∗ ∗ ∗
= d x δχa iσ̄ab ∂µ χb + σ (−χa δχb − δχa χb )
2 ab
% &
im # 2
ˆ
4 ∗ µ ∗ ∗ 2 ∗ ∗
$
= d x δχa iσ̄ab ∂µ χb + −σab δχa χb − σab δχa χb
2
ˆ
! µ
d4 x δχ∗a iσ̄ab 2 ∗
"
= ∂µ χb − imσab χb

= 0

Since this holds for all integration volumes

iσ̄ µ ∂µ χ − imσ 2 χ∗ = 0
which is the Majorana equation. Similarly, with the action
% &
im # † 2 ∗
ˆ
† 4 † µ T 2
$
S = S = d x − (∂µ χ) iσ̄ χ − χ σ χ −χ σ χ
2

and varying with respect to χ yields the complex conjugate equation

iσ̄ µ ∂µ χ∗ − imσ 2 χ = 0.

"
(c) Let us write a 4-component Dirac field as - .
ψL
ψ (x) = ,
ψR
and recall that the lower components of ψ transform in a way equivalent by a unitary transformation to the complex
conjugate of the representation ψL . In this way we can rewrite the 4-component Dirac field in terms of two 2-component
spinors:
ψL (x) = χ1 (x) and ψR (x) = iσ 2 χ∗2 (x) .
Rewrite the Dirac Lagrangian in terms of χ1 and χ2 and note the form of the mass term.

14
 $T $T
Proof: The Dirac Lagrange density is LD = ψ̄ (iγ µ ∂µ − m) ψ. The Dirac spinor is ψ = iσ 2 χ∗2
# #
ψL ψR = χ1 .

LD = ψ̄ (iγ µ ∂µ − m) ψ
−m iσ µ ∂µ
- .- .- .
# † $ 0 1 χ1
= χ1 −iχT2 σ 2
1 0 iσ̄ µ ∂µ −m iσ 2 χ∗2
$ −mχ1 + iσ µ ∂µ iσ 2 χ∗2
- .
−iχT2 σ 2 χ†1
#
=
iσ̄ µ ∂µ χ1 − miσ 2 χ∗2
= −iχT2 σ 2 −mχ1 + iσ µ ∂µ iσ 2 χ∗2 + χ†1 iσ̄ µ ∂µ χ1 − miσ 2 χ∗2
# $ # $

= iχT2 σ 2 σ · ∂ σ 2 χ∗2 + iχ†1 σ̄ · ∂χ1 + imχT2 σ 2 χ1 − miχ†1 σ 2 χ∗2

# $
+ ,
= iχ†1 σ̄ · ∂χ1 + iχT2 σ 2 σσ 2 · ∂χ∗2 + im χT2 σ 2 χ1 − χ†1 σ 2 χ∗2
+ ,
= iχ†1 σ̄ · ∂χ1 + iχT2 σ̄ ∗ · ∂χ∗2 + im χT2 σ 2 χ1 − χ†1 σ 2 χ∗2
+ ,
= iχ†1 σ̄ · ∂χ1 + iχT2 σ̄ ∗ · ∂χ∗2 + im χT2 σ 2 χ1 − χ†1 σ 2 χ∗2

where we have used σ 2 σ µ σ 2 = σ̄ µ∗ .

T
Since the second term, iχT2 σ̄ ∗ · ∂χ∗2 , is just a number we can simply it further using the fact that (σ̄ µ ) = −σ µ
$T
iχT2 σ̄ µ∗ ∂µ χ∗2
# T µ∗
= iχ2 σ̄ ∂µ χ∗2
= i∂µ χ†2 σ̄ µ† χ2
= i∂µ χ†2 σ̄ µ χ2 .

Thus, the Lagrange density is

+ ,
LD = iχ†1 σ̄ · ∂χ1 + iχ†2 σ̄ · ∂χ2 + im χT2 σ 2 χ1 − χ†1 σ 2 χ∗2 .

Notice that if χ1 = χ2 we have twice the Majorana Lagrange density from part (b):
+ ,
LM = iχ†1 σ̄ · ∂χ1 + iχ†1 σ̄ · ∂χ1 + im χT1 σ 2 χ1 − χ†1 σ 2 χ∗1
@ ,A
im + T 2
= 2 iχ†1 σ̄ · ∂χ1 + χ1 σ χ1 − χ†1 σ 2 χ∗1 .
2

"
(d) Show that the action of part (c) has a global symmetry. Compute the divergences of the currents

J µ = χ† σ̄ µ χ and J µ = χ†1 σ̄ µ χ1 − χ†2 σ̄ µ χ2 ,

for the theories of parts (b) and (c), respectively, and relate your results to the symmetries of these theories. Construct
a theory of N free massive 2-component fermion fields with O (N ) symmetry (that is, the symmetry of rotations in an
N -dimensional space).
Proof:
The action ˆ ' + ,(
S= d4 x iχ†1 σ̄ · ∂χ1 + iχ†2 σ̄ · ∂χ2 + im χT2 σ 2 χ1 − χ†1 σ 2 χ∗2

is invariant under the U (1) symmetry ψ → eiθ ψ.

15
To verify this we write

ψ → eiθ ψ
eiθ χ1 $
- . - .
χ1
→ ∗ .
iσ 2 χ∗2 iσ 2 e−iθ χ2
#

Thus, in terms of the χ′ s the transformation ψ → eiθ ψ is χ1 → eiθ χ1 and χ2 → e−iθ χ2 . Because in each term of the
action χ1 is partnered with either its complex conjugate or χ2 and vice versa for χ2 the action is invariant under the U (1)
symmetry χ1 → eiθ χ1 and χ2 → e−iθ χ2 .
We now compute the divergences of the currents,

Jµ = χ† σ̄ µ χ
Jµ = χ†1 σ̄ µ χ1 − χ†2 σ̄ µ χ2 ,

for the theories of part (b) and (c). To derive these currents we apply equation (2.12).
Part (c): Under the U (1) transformation the Lagrangian of part (c) is invariant. To compute the currents we need to
know the equations of motion. The equation of motion is obtained from the Euler-Lagrange equations
- .
∂L ∂L
∂µ − = 0.
∂ (∂µ χi ) ∂χi

With χi = χ1 the EL equation yields

+ ,
∂µ iχ†1 σ̄ µ − imχT2 σ 2 = 0
+ ,
∂µ χ†1 σ̄ µ − mχT2 σ 2 = 0
+ ,†
σ̄ µ† ∂µ χ†1 − mσ 2† χT2 † = 0
σ̄ µ ∂µ χ1 − mσ 2 χ∗2 = 0.

Similarly, for χi = χ2
+ , ∂ # T 2 $
∂µ iχ†2 σ̄ µ + im χ σ χ1 = 0
∂χ2 2
+ , ∂ # T 2 $
∂µ χ†2 σ̄ µ − m χ1 σ χ2 = 0
∂χ
+ , 2
∂µ χ†2 σ̄ µ − mχT1 σ 2 = 0
σ̄ µ (∂µ χ2 ) − mσ 2 χ∗1 = 0.

Now the divergence of the current can be computed

+ ,
∂µ J µ = ∂µ χ†1 σ̄ µ χ1 − χ†2 σ̄ µ χ2

= ∂µ χ†1 σ̄ µ χ1 − ∂µ χ†2 σ̄ µ χ2 + χ†1 σ̄ µ ∂µ χ1 − χ†2 σ̄ µ ∂µ χ2

† †
= (σ̄ µ ∂µ χ1 ) χ1 − (σ̄ µ ∂µ χ2 ) χ2 + χ†1 (σ̄ µ ∂µ χ1 ) − χ†2 (σ̄ µ ∂µ χ2 )
# 2 ∗ $† $†
mσ χ2 χ1 − mσ 2 χ∗1 χ2 + χ†1 mσ 2 χ∗2 − χ†2 mσ 2 χ∗1
# # $ # $
=
= mχT2 σ 2 χ1 − mχT1 σ 2 χ2 + mχ†1 σ 2 χ∗2 − mχ†2 σ 2 χ∗1
= 0.

16
Since the divergence of the current vanishes, the current is conserved under U (1) transformations.
Part (c): Under the U (1) transformation the Lagrangian of part (b) is NOT invariant
im # T 2 $† $ im +# iφ $T 2 iφ # iφ $† 2 # iφ $∗ ,
χ† iσ̄ · ∂χ + χ σ χ − χ† σ 2 χ∗ → χeiφ iσ̄ · ∂ χeiφ +
$ # #
χe σ χe − χe σ χe
2 2
im # T 2 2iφ
= χ† iσ̄ · ∂χ + χ σ χe − χ† σ 2 χ∗ e−2iφ
$
2
̸= L.
However, if m = 0 then the Lagrangian is symmetric under U (1) phase rotations.
The equations of motion are
σ̄ µ ∂µ χ − mσ 2 χ∗ = 0
∂µ χ† σ̄ µ − mχT σ 2
# $
= 0.

Thus the divergence of the current is

∂µ χ† σ̄ µ χ = ∂µ χ† σ̄ µ χ + χ† (σ̄ µ ∂µ χ)
# $ # $

mχT σ 2 χ + χ† mσ 2 χ∗
# $ # $
=
= mχT σ 2 χ + mχ† σ 2 χ∗ .
Like the Lagrangian the current is only conserved if m = 0.
For the last part of this question we construct a theory of N free massive 2-component fermion fields with O(N ) symmetry
(the symmetry of rotations in an N -dimensional space). Each free massive particle is described by the Lagrangian of part
(b)
im # T 2
La = χ†a iσ̄ · ∂χa + χa σ χa − χ†a σ 2 χ∗a
$
2
where a ∈ {1, 2, ..., N }. The total Lagrangian is the sum of the individual Lagrangians
= im # T 2
χ†a iσ̄ · ∂χa + χa σ χa − χ†a σ 2 χ∗a .
$
L=
a
2
Each fermion field satisfies the equation of motion
σ̄ µ ∂µ χa − mσ 2 χ∗a = 0
∂µ χ†a σ̄ µ − mχTa σ 2
# $
= 0.

To have O(N ) symmetry the Lagrangian must be invariant under rotations in N -dimensional space. Let the rotation
operator be denoted by Rab where a, b are the components of the operator. Applied to the bth fermion field the rotation
operator takes the bth fermion field to the ath fermion field
Rab χb = χa .
Since R is a rotation, it is an orthogonal matrix (i.e., R−1 = RT and Rab ∈ R). Application of R to the Lagrangian yields
im + ,
(Rab χb )† iσ̄ · ∂ (Rab χb ) + (Rab χb )T σ 2 (Rab χb ) − (Rab χb )† σ 2 (Rab χb )∗
=
L →
2
ab
= † im + ,
= Rab Rab χ†b iσ̄ · ∂χb + T
Rab †
Rab χTb σ 2 χb − Rab ∗ † 2 ∗
Rab χb σ χb
2
ab
im + ,
Rba Rab χ†b iσ̄ · ∂χb + Rba Rab χTb σ 2 χb − Rba Rab χ†b σ 2 χ∗b
=
=
2
ab
= † im + T 2 ,
= χb iσ̄ · ∂χb + χb σ χb − χ†b σ 2 χ∗b
2
b
= L.

17
Thus, we see that the Lagrangian for massive two-component fermions is invariant under O(N ) transformations.
"
(e) Quantize the Majorana theory of parts (a) and (b). That is, promote χ (x) to a quantum fields satisfying the canonical
anti-commutation relation 7 8
χa (x) , χ†b (y) = δab δ (3) (x − y) ,
construct a Hermitian Hamiltonian, and find a representation of the canonical commutation relations that diagonalizes
the Hamiltonian in terms of a set of creation and annihilation operators. (Hint: Compare χ (x) to the top two components
of the quantized Dirac field.)
Proof: The book suggests comparing χ to the upper two component of the quantized Dirac field
d3 p 1 =# s s
ˆ
ap u (p)e−ip·x + bs† s ip·x
$
ψ = 3
5 p v (p)e
(2π) 2Ep s
- - √ - √
d3 p p · σξ s p · ση s
. . .
1 = s
ˆ
√ −ip·x s† √ ip·x
= a e + b e .
(2π)3 2Ep p p · σ̄ξ s p − p · σ̄η s
5
s

Setting ψR = iσ 2 ψL in the Dirac Lagrangian yields the Majorana Lagrangian. Thus, we expect the Dirac field to be a
solution to the Majorana equation under certain restrictions (i.e., η s = −iσ 2 ξ). Setting χ equal to ψL , the quantized
Majorana field becomes
d3 p p · σ = # s s −ip·x
ˆ 9
+ bs† s ip·x
$
χ= 3
ap ξ e p η e .
(2π) 2Ep s

The condition of charge conjugation

∗ ∗
us (p) = −iγ 2 (v s (p)) and v s (p) = −iγ 2 (us (p))

places a restriction on the spinors, namely η s = −iσ 2 ξ. Thus, the Majorana field becomes
d3 p p · σ = # s 2 s∗ −ip·x
ˆ 9
+ bs† s ip·x
$
χ= 3
ap iσ η e p η e .
(2π) 2Ep s

We will now test our solution in the Majorana equation of motion iσ̄ µ ∂µ χ = imσ 2 χ∗ . The RHS becomes

d3 p p · σ = # s 2 s∗ −ip·x
ˆ 9
µ µ
+ bs† s ip·x
$
iσ̄ ∂µ χ = iσ̄ ∂µ 3
ap iσ η e p η e
(2π) 2Ep s
3
d p p · σ = # s 2 s∗
ˆ 9
= iσ̄ µ ap iσ η (−ipµ ) e−ip·x + (ipµ ) bs† s ip·x
$
p η e
(2π)3 2Ep s
6
d3 p √ (σ̄ · p) (p · σ) = # s 2 s∗ −ip·x
ˆ
+ ibs† s ip·x
$
= i 3
σ̄ · p ap σ η e p η e
(2π) 2Ep s
6
3
d p σ̄ · p = # s 2 s∗ −ip·x
ˆ
+ ibs† s ip·x
$
= im 3
ap σ η e p η e
(2π) 2Ep s

where we have used the fact that (σ̄ · p) (σ · p) = m2 . The LHS is

;ˆ <∗
d3 p p · σ = # s 2 s∗ −ip·x
9
2 ∗ 2 s† s ip·x
$
imσ χ = imσ ap iσ η e + bp η e
(2π)3 2Ep s
6
d3 p p · σ∗ = #
ˆ
2
−ias† 2∗ s ip·x
+ bsp η s∗ e−ip·x .
$
= imσ 3 p σ η e
(2π) 2Ep s

18
To deal with the square root of σ ∗ we note that the square root is defined as a taylor series. Inserting unity in the form
σ 2 σ 2 between each consecutive term of σ ∗ , σ ∗n = σ 2 σ̄σ 2 and the taylor series becomes one in σ̄ instead of σ ∗ . With this
in mind the LHS becomes
6
d3 p 2 p · σ̄ 2 = #
ˆ
2 ∗ 2
−ias† 2∗ s ip·x
+ bsp η s∗ e−ip·x .
$
imσ χ = imσ 3
σ σ p σ η e
(2π) 2Ep s
6
3
d p p · σ̄ = # s† s ip·x
ˆ
+ bsp σ 2 η s∗ e−ip·x .
$
= im 3
ap iη e
(2π) 2Ep s

Comparing the LHS and RHS we see that they are equal only if

bsp = asp .

Thus, the quantized Majorana field is

d3 p p · σ = # s 2 s∗ −ip·x
ˆ 9
+ as† s ip·x
$
χ= ap iσ η e p η e .
(2π)3 2Ep s

From this mode expansion of the Majorana field we can see that the Majorana particle is its own anti-particle.
To determine the commutation relations of the particle operator asp we simplify the commutator
7 8
χa (x) , χ†b (y) = δab δ (3) (x − y)

in terms of asp commutators. Simplifying, we have

7 8
δad δ (3) (x − y) = χa (x) , χ†d (y)
&1/2 % ′ &1/2
d3 p d3 p′ = p · σ
%
p ·σ
ˆ ˆ
=
(2π)3 (2π)3 sr 2Ep ab 2Ep′ ed
7# $ + ,8
, −iar† 2 r ip′ ·y ′
iasp σbc
2 s∗ −ip·x
ηc e + as† s ip·x
p ηb e p′ σf e ηf e + arp′ ηer∗ e−ip ·y
&1/2 % &1/2
d3 p d3 p′ = p · σ
%
p·σ
ˆ ˆ
=
(2π)3 (2π)3 sr 2Ep ab 2Ep ed
' 7 8
ηc σf e ηf asp , ar† −ip·x+ip′ ·y ′
2 s∗ 2 r 2 s∗ r∗
ηc ηe asp , arp′ e−ip·x−ip ·y
) *
σbc p′ e + iσbc
7 8 ) s† r * ip·x−ip′ ·y (
r† ip·x+ip′ ·y
−iηbs σf2 e ηfr as†
p , a p ′ e + η s r∗
η
b e ap , ap ′ e .

This is very complicated and it is not obvious what commutation relations for asp yields the RHS. However, since we are
using a modified Dirac field we would expect that asp satisfies the usual relations
7 8
r†
asp , arp′ as†
) *
=0= p , ap ′
7 8
3
asp , ar†
p′ = (2π) δrs δ(p − p′ ).

19
Lets test if these commutation relations are indeed the ones we require
&1/2 % &1/2
d3 p d3 p′ = p · σ
%
p·σ
7 8 ˆ ˆ
χa (x) , χ†d (y) =
(2π)3 (2π)3 sr 2Ep ab 2Ep ed
' 7 8 * ip·x−ip′ ·y (
ηc σf e ηf asp , ar†
2 s∗ 2 r −ip·x+ip′ ·y
+ ηbs ηer∗ as† r
)
× σbc p′ e p , ap ′ e
ˆ
d3 p
ˆ = % p · σ &1/2 % p · σ &1/2
3 ′
= d p
(2π)3 sr
2Ep ab 2Ep ed
' (
2 s∗ 2 r −ip·x+ip′ ·y ′
×δrs δ(p − p′ ) σbc ηc σf e ηf e + ηbs ηer∗ eip·x−ip ·y
&1/2 % &1/2
d3 p = p · σ
%
p·σ
ˆ
=
(2π)3 s 2Ep ab 2Ep ed
' (
2 s∗ 2 s −i(Ep x0 −p·x)+i(Ep y 0 −p·y)
+ ηbs ηes∗ ei(Ep x −p·x)−i(Ep y −p·y)
0 0
× σbc ηc σf e ηf e

When taken at equal times the commutator becomes

&1/2 % &1/2
d3 p = p · σ
%
p·σ
7 8 ˆ
χa (x) , χ†d (y)
! 2 s∗ 2 r ip·x−ip·y
+ ηbs ηes∗ e−ip·x+ip·y
"
= σbc ηc σf e ηf e
(2π)3 s 2Ep ab 2Ep ed
>% &1/2 &1/2 &1/2 &1/2 ?
d3 p = p · σ
% % %
p·σ p·σ s s∗ p · σ
ˆ
2 s∗ s 2 ip·(x−y) −ip·(x−y)
= σ η η σ e + η η e
(2π)3 s 2Ep ab bc c f f e 2Ep ed 2Ep ab b e 2Ep ed

To simplify this further we sum over sand use the identity,

=
ηas∗ ηbs = δab ,
s

to get
>% &1/2 &1/2 &1/2 &1/2 ?
d3 p
% % %
p·σ p·σ p·σ p·σ
7 8 ˆ
χa (x) , χ†d (y) = 2
σbc δcf σf2 e e ip·(x−y)
+ δbe e −ip·(x−y)
(2π)3 2Ep ab 2Ep ed 2Ep ab 2Ep ed
d3 p
%% & % & &
p·σ p·σ
ˆ
ip·(x−y) −ip·(x−y)
= e + e
(2π)3 2Ep ad 2Ep ad
d3 p
%% & % & &
p·σ p·σ
ˆ
= eip·(x−y) + e−ip·(x−y) .
(2π)3 2Ep ad 2Ep ad

To simplify further we need to know (p · σ)ad . Here, we work out the matrix elements of p · σ,

(p · σ)ad = (Ep − p · σ)ad

= Ep δad − pi · σad
i
.

Substitution into the commutation relation yields

d3 p Ep δad − pi · σad
i
Ep δad − pi · σad
i
7 8 ˆ % &
χa (x) , χ†d (y) = e ip·(x−y)
+ e −ip·(x−y)
(2π)3 2Ep 2Ep
i ✭✭ i ✭✭
✭✭✭
d3 p δad ip·(x−y) δad −ip·(x−y) d3 p pi · σad i
✭ ✭
% & ˆ % &
p · σ
ˆ
✭ ✭ ✭
ip·(x−y) ad −ip·(x−y)
=
(2π)3 2
e +
2
e −
(2π)✭✭✭
3 2E ✭p✭e✭ +
2Ep
e
✭ ✭ ✭
= δad δ(x − y)

20
where the second integral is odd in p.
We now turn to calculating the Hamiltonian for Majorana fields. The Hamiltonian is the Legendre transform of the
Lagrangian,
ˆ
H = d3 x (π χ̇ − L)
- .
im # T 2
ˆ
d3 x iχ† χ̇ − iχ† σ̄ · ∂χ + χ σ χ − χ† σ 2 χ∗
$
=
2
where
im # T 2
= iχ† σ̄ · ∂χ + χ σ χ − χ† σ 2 χ∗ ,
$
L
2
∂L
π = = iχ† .
∂ (χ̇)
The simplification is straight forward and messy.
We just note the result here
d3 p
ˆ =
H= 3 Ep as† s
p ap .
(2π) s

Which is exactly 1/2 the Dirac Hamiltonian,

d3 p
ˆ =#
as† s s† s
$
HD = 3 Ep p ap + b p b p ,
(2π) s

with the restriction bsp = asp .

We simplify the Hamiltonian term by term. The first term is
6
d3 p d3 p′ = # $ p·σ p′ · σ + r 2 r∗ −ip′ ·x
ˆ ˆ 9 ,
† s sT 2 ip·x s† s† −ip·x r† r ip′ ·x
iχ χ̇ = i −ia p η σ e + a p η e ∂0 a p ′ iσ η e + a p ′ η e
(2π)3 (2π)3 rs 2Ep 2Ep′
6
d3 p d3 p′ $ p · σ p′ · σ + r 2 r∗ −ip′ ·x
ˆ ˆ 9 ,
r† r ip′ ·x
=#
s sT 2 ip·x s† s† −ip·x
= i Ep ′ −ia p η σ e + a p η e a p ′σ η e + ia p ′η e
(2π)3 (2π)3 rs
2Ep 2Ep′

"

1.5 Supersymmetry
It is possible to write field theories with continuous symmetries linking fermions and bosons; such transformations are
called super-symmetries.
(a) The simplest example of a supersymmetric field theory is the theory of a free complex boson and a free Weyl fermion,
written in the form
L = ∂µ φ∗ ∂ µ φ + χ† iσ̄ · ∂χ + F ∗ F.
Here F is an auxiliary complex scalar field whose field equation is F = 0. Show that this Lagrangian is invariant (up to a
total divergence) under the infinitesimal transformation
δφ = −iϵT σ 2 χ,
δχ = ϵF − σ · ∂φσ 2 ϵ∗ ,
δF = −iϵ† σ̄ · ∂χ,

21
where the parameter ϵa is a 2-component spinor of Grassmann numbers.
(b) Show that the term % &
1 T 2
∆L = mφF + imχ σ χ + (complex conjugate)
2
is also left invariant by the transformation given in part (a). Eliminate F from the complete Lagrangian L+ ∆L by solving
its field equation, and show that the fermion and boson fields φ and χare given the same mass.
(c) It is possible to write supersymmetric nonlinear field equations by adding cubic and higher-order terms to the La-
grangian. Show that the following rather general field theory, containing the field (φi , χi ), i = 1, ..., n, is supersymmetric:

L = ∂µ φ∗i ∂ µ φi + χ†i iσ̄ · ∂χi + Fi∗ Fi

i ∂ 2 W [φ] T 2
- .
∂W [φ]
+ Fi + χi σ χj + c.c ,
∂φi 2 ∂φi ∂φj
where W [φ] is an arbitrary function of the φi , called the super-potential. For the simple case n = 1 and W = gφ3 /3,
write out the field equations for φ and χ (after elimination of F).

1.6 Fierz transformations !

Let ui , i = 1, ...4, be four 4-component Dirac spinors. In the text, we proved the Fierz rearrangement formulae (3.78) and
(3.79). The first of these formulae can be written in 4-component notation as
1 + γ5 1 + γ5 1 + γ5 1 + γ5
- . - . - . - .
µ µ
ū1 γ u2 ū3 γµ u4 = −ū1 γ u4 ū3 γµ u2 .
2 2 2 2

The above identity plays an important part in the weak interaction. Thus, we derive this equation from (3.78) and (3.79)

(ū1R σ µ u2R ) (ū3R σµ u4R ) = − (ū1R σ µ u4R ) (ū3R σµ u2R )

(ū1L σ̄ µ u2L ) (ū3L σ̄µ u4L ) = − (ū3L σ̄ µ u4L ) (ū1L σ̄µ u2L ) .

1+γ 5 1−γ 5
Denoting the handedness operators by PR = 2 and PL = 2 we have the following:

(ū1 γ µ PR u2 ) (ū3 γµ PR u4 ) = (ū1R γ µ u2R ) (ū3R γµ u4R )

= (ū1R σ µ u2R ) (ū3R σµ u4R )
= − (ū1R σ µ u4R ) (ū3R σµ u2R )
= − (ū1 γµ PR u4 ) (ū3 γµ PR u2 )

(ū1 γ µ PL u2 ) (ū3 γµ PL u4 ) = − (ū3L γµ PL u4L ) (ū1L γµ PL u2L ) .

In fact, there are similar rearrangement formulae for any product
ū1 ΓA u2 ū3 ΓB u4 ,
# $# $

where ΓA , ΓB are any of the 16 combinations of Dirac matrices listed in Section 3.4.
(a) To begin, normalize the 16 matrices ΓA to the convention
Tr ΓA ΓB = 4δ AB .
! "

This gives ΓA = 1, γ 0 , iγ j , ... ; write all 16 elements of this set.

) *

Proof: We already know that the 16 Dirac matrices are orthogonal subject to the inner product Tr [AB]. Thus, we only
have to find the normalization constant.

22
1. Scalars (ΓA = 1): Tr [1] = 4 thus, the identity is properly normalized.
2. Vectors (ΓA = γ µ ): '# $2 (
Tr γ0 = Tr [1] = 4

>; # $2 <?
− σi 0
'# $2 (
Tr γ i = Tr # $2 = −Tr [1] − 4
0 − σi
Thus, the normalized vector matrices are: γ 0 and iγ i .
3. Tensor (ΓA = σ µν ):
- .2 '
'
2
( i 2
(
Tr (σ µν ) = Tr ([γ µ , γ ν ])
2
1
= − Tr [γ µ γ ν γ µ γ ν − γ ν γ µ γ µ γ ν − γ ν γ µ γ µ γ ν + γ ν γ µ γ ν γ µ ]
4
1
= − {2Tr [γ µ γ ν γ µ γ ν ] − 2Tr [γ ν γ µ γ µ γ ν ]}
4
1
= − {2Tr [γ µ γ ν γ µ γ ν ] + 2Tr [γ µ γ ν γ µ γ ν ] − 4g µν Tr [γ µ γ ν ]}
4
1
= − {4Tr [γ µ γ ν γ µ γ ν ] − 16g µν g µν }
4
1
= − {16 [g µν g µν − g µµ g νν + g µν g µν ] − 16g µν g µν }
4
1
= − 16 [g µν g µν − g µµ g νν ]
4
= −4 {g µν g µν − g µµ g νν }
B
0 µ=ν
=
4g µµ g νν µ ̸= ν
⎧
⎨0
⎪ µ=ν
= 4 µ ̸= ν and both µ, ν are spacelike or timelike
⎪
−4 µ ̸= ν and µ is spacelike (timelike) and ν is timelike (spacelike)
⎩

Thus, the normalized tensor matrices are: −σ 0i , −σ i0 , σ 00 , σ ii .

4. Pseudo-vector (ΓA = γ µ γ 5 ):

γ µγ 5 Tr γ µ γ 5 γ µ γ 5
!# $" ! "
Tr =
Tr γ µ γ 5 γ 5 γ µ
! "
=
= Tr [γ µ γ µ ]
= 4g µµ
B
−4 µ = 0
=
4 µ=i

Thus, the normalized pseudo-vector matrices are: −γ 0 γ 5 , γ i γ 5 .

5. Pseudo-scalar (ΓA = γ 5 ):
Tr γ 5 γ 5 = 1
! "

Thus, γ 5 is already properly normalized.

23
"
(b) Write the general Fierz identity as an equation
$ = AB #
ū1 ΓA u2 ū3 ΓB u4 = C CD ū1 ΓC u4 ū3 ΓD u2
# $# $# $

C,D

with unknown coefficients C AB A

CD . Using the completeness of the 16 Γ matrices, show that
1
C AB Tr ΓC ΓA ΓD ΓB .
! "
CD =
16
Proof:
Using the completeness of the Gamma matrices we write
= = =
ΓA B
ab Γcd = C AB C
C Γad C AB D
D Γbc = C AB C D
CD Γad Γbc
C D CD

from which it follows $ = AB #

ū1 ΓA u2 ū3 ΓB u4 = C CD ū1 ΓC u4 ū3 ΓD u2 .
# $# $# $
C,D
′
D′
We now calculate C AB
C ′ D′ . Multiplying by ΓC
da Γbc we have
′ ′ =
ΓC D A B
da Γbc Γab Γcd = C AB C D C D
CD Γda Γbc Γad Γbc
C,D

which implies
' ′ ′
( = ' ′ ( ' ′ (
Tr ΓC ΓA ΓD ΓB = C AB
CD Tr ΓC ΓC Tr ΓD ΓD
C,D
= ′ ′
= 16 C AB
CD δ
CC DD
δ
C,D

= 16C AB
C ′ D′ .
1
Thus, we see that C AB ΓC ΓA ΓD ΓB .
! "
CD = 16 Tr
"
(c) Work out explicitly the Fierz transformation laws for the products (ū1 u2 ) (ū3 u4 ) and (ū1 γ µ u2 ) (ū3 γµ u4 ).
Proof:
First, the product (ū1 u2 ) (ū3 u4 ) = ū1 ΓA u2 ū3 ΓB u4 where ΓA,B = 1. Thus, we need all nonzero traces of the form
# $# $

1
C 11CD Tr ΓC 1ΓD 1
! "
=
16
1
Tr ΓC ΓD
! "
=
16
1
= δCD .
4
Thus, the product rule for ΓA,B = 1 becomes
=
C AB C
ū3 ΓD u2
# $# $
(ū1 u2 ) (ū3 u4 ) = CD ū1 Γ u4
C,D
=1
δCD ū1 ΓC u4 ū3 ΓD u2
# $# $
=
4
C,D
1 =#
ū1 ΓC u4 ū3 ΓC u2
$# $
=
4
C

24
where the index C cycles through all 16 ΓC .
Next, we examine the product (ū1 γ µ u2 ) (ū3 γµ u4 ). The coefficients C AB
CD are determined as usual

1
C µµ
Tr ΓC γ µ ΓD γµ .
! "
CD =
16
Using the matrices from our normalized basis of part (a) we define Γµ = γ 0 , iγ . In terms of Γµ we have
# $

3
B G
1
C µµ
=
C 0 D 0
Tr ΓC −iΓi ΓD −iΓi
! " ! # $ # $"
CD = Tr Γ Γ Γ Γ −
16 i=1
3
B G
1 ! C 0 D 0
" = ! C i D i"
= Tr Γ Γ Γ Γ + Tr Γ Γ Γ Γ .
16 i=1

Suppose that ΓC = γ 0 then

3
B G
1
C µµ
=
0 0 D 0
Tr Γ0 Γi ΓD Γi
! " ! "
0D = Tr Γ Γ Γ Γ +
16 i=1
3
B G
1 =
Tr ΓD Γ 0
Tr Γ0 Γ D
! " ! "
= −
16 i=1
= −δ D0 .

Similarly, if ΓC = γ j then
3
B G
1
C µµ
! j 0 D 0" = ! j i D i"
jD = Tr Γ Γ Γ Γ + Tr Γ Γ Γ Γ
16 i=1
3
B G
1 j D
=
Tr −2g ij − Γi Γj ΓD Γi
! " !# $ "
= −Tr Γ Γ +
16 i=1
3 3
B G
1 = =
−4δ Dj − 2 g ij Tr ΓD Γi − Tr Γi Γj ΓD Γi
! " ! "
=
16 i=1 i=1
3
B G
1 Dj D j
=
Tr Γj ΓD
! " ! "
= −4δ − 2Tr Γ Γ −
16 i=1
1 )
−4δ Dj − 8δ Dj − 12δ Dj
*
=
16
24
= − δ Dj
16
3
= − δ Dj
2
Thus, we see that for ΓC = γ ν we must have ΓD = γ ν and vice versa.
For ΓC,D = 1 we have C µµ
CD = 0.

The remaining matrices are the anti-symmetric matrices form the set Γ = {σ µν , γ µ γ 5 , γ 5 }we have

25
for: ΓC = γ 5
3
B G
1
C µµ
=
5 0 D 0
Tr Γ5 Γi ΓD Γi
! " ! "
5D = Tr Γ Γ Γ Γ +
16 i=1
3
B G
1 =
−Tr Γ5 ΓD −
! " ! 5 D"
= Tr Γ Γ
16 i=1
D5
= −δ

for: ΓC = γ 0 γ 5
3
B G
1
C µµ
" =
Tr Γ05 Γ0 ΓD Γ0 + Tr Γ05 Γi ΓD Γi
! ! "
ν5,D =
16 i=1
3
B G
1 ! 0 5 0 D 0" = !# 0 5 $ # i $ D i "
= Tr −γ γ γ Γ γ + Tr −γ γ iγ Γ Γ
16 i=1
3
B G
1 5 0 D
=
Tr iγ i −γ 0 γ 5 ΓD Γi
! " !# $ # $ "
= −Tr γ γ Γ +
16 i=1
3
B G
1 =
Tr Γ05 ΓD + 05 D
! " ! "
= Tr Γ Γ
16 i=1
1
= 16δ 05,D
16
= δ 05,D

for: ΓC = γ i γ 5
3
B G
1
C µµ
=
j5 0 D 0 j5 i D i
! " ! "
j5,D = Tr Γ Γ Γ Γ + Tr Γ Γ Γ Γ
16 i=1
3
B G
1 0 j5 D 0
=
Tr iγ j γ 5 iγ i ΓD Γi
! " !# $# $ "
= Tr Γ Γ Γ Γ +
16 i=1
3
B G
1 =
Tr Γj5 Γ D
iγ j iγ i γ 5 ΓD Γi
! " !# # $ $ "
= − Tr
16 i=1
3
B G
1 j5,D
=
Tr i2 2g ij − iγ i iγ j γ 5 ΓD Γi
!# # $ $ "
= 4δ −
16 i=1
3 3
B G
1 j5,D
=
ij 5 D i
=
Tr iγ i iγ j γ 5 ΓD Γi
! " !# $ "
= 4δ +2 Tr g γ Γ Γ +
16 i=1 i=1
3
B G
1 =
4δ j5,D + 2Tr γ 5 ΓD Γ j
Tr Γj5 Γ D
! " ! "
= +
16 i=1
3
B G
1 j5,D
! D j5
" = ! j5 D
"
= 4δ + 2Tr Γ Γ + Tr Γ Γ
16 i=1
24 j5,D
= δ
16
3 j5,D
= δ
2

26
for: ΓC = Γµν = −σ 0i , −σ i0 , σ 00 , σ ii
i µ ν
{σ µν , γ α } = {γ γ − γ ν γ µ , γ α }
2
i µ ν α i
= {γ γ , γ } − {γ ν γ µ , γ α }
2 2
i µ α ν i i i
= {γ , γ } γ + γ µ [γ ν , γ α ] − {γ ν , γ α } γ µ − γ ν [γ µ , γ α ]
2 2 2 2
= ig µα γ ν + iγ µ (γ ν γ α − g να ) − ig να γ µ − iγ ν (γ µ γ α − g µα )
= ig µα γ ν + iγ µ γ ν γ α − iγ µ g να − ig να γ µ − iγ ν γ µ γ α + iγ ν g µα
= 2ig µα γ ν − 2ig να γ µ + iγ µ γ ν γ α − iγ ν γ µ γ α
= 2ig µα γ ν − 2ig να γ µ + iγ µ γ ν γ α − iγ ν γ µ γ α

1.7 Discrete symmetries of the Dirac field !

This problem concerns the discrete symmetries P, C, and T .

(a) Compute the transformation properties under P, C, and T of the antisymmetric tensor fermion bilinears, ψ̄σ µν ψ, with
σ µν = 2i [γ µ , γ ν ]. This completes the table of the transformation properties of bilinears at the end of this chapter.
Proof: Starting with the parity transformation we have
i µ ν
P ψ̄σ µν ψP = P ψ̄P P [γ , γ ] P P ψP
2
i
= ηa∗ ηa ψ̄ (t, −x) γ 0 [γ µ , γ ν ] γ 0 ψ (t, −x)
2
i
= ψ̄ (t, −x) γ 0 (γ µ γ ν − γ ν γ µ ) γ 0 ψ (t, −x) .
2
Notice that µ ̸= ν otherwise the commutator vanishes. There are two cases:

1. µ ̸= ν where µ, ν = 1, 2, 3.

2 i# i j
P ψ̄σ ij ψP γ γ − γ j γ i γ 0 γ 0 ψ (t, −x)
$
= (−1) ψ̄ (t, −x)
2
= ψ̄ (t, −x) σ ij ψ (t, −x) .

2. µ ̸= ν and either µ = 0 or ν = 0. For ν = 0 the parity transformation is

i# i 0
P ψ̄σ i0 ψP ψ̄ (t, −x) γ 0 γ γ − γ 0 γ i γ 0 ψ (t, −x)
$
=
2
i# i 0
γ γ − γ 0 γ i γ 0 γ 0 ψ (t, −x)
$
= (−1) ψ̄ (t, −x)
2
= −ψ̄ (t, −x) σ i0 ψ (t, −x) .

Similarly, for µ = 0 the parity transformation is

P ψ̄σ 0j ψP = −ψ̄ (t, −x) σ 0j ψ (t, −x) .

27
The parity transformation of the bilinear tensor can be summarized by the following formula
µ ν
P ψ̄σ µν ψP = (−1) (−1) ψ̄ (t, −x) σ µν ψ (t, −x)

where B
µ 1 µ=0
(−1) =
−1 µ = 1, 2, 3.

Next we look at the time reversal transformation

i µ ν
T ψ̄σ µν ψT = T ψ̄T T [γ , γ ] T T ψT
2
# 1 3$ i µ ν ∗ # 1 3$
- .
= ψ̄ (−t, x) −γ γ [γ , γ ] γ γ ψ (−t, x)
2
$ i µ∗ ν∗
= −ψ̄ (−t, x) −γ 1 γ 3 (γ γ − γ ν∗ γ µ∗ ) γ 1 γ 3 ψ (−t, x)
# # $
2
Since γ 2∗ = −γ 2 , B#
γ 1γ 3 γ µ
$
µ∗
# 1 3$ µ=0
γ γ γ =
− γ1γ3 γµ
# $
µ = 1, 2, 3
and we can simplify the transformation law. Like the parity transformation there are two cases:

1. µ ̸= ν and µ, ν = 1, 2, 3.
$i# i j
T ψ̄σ ij ψT = − (−1)2 ψ̄ (−t, x) −γ 1 γ 3 γ 1 γ 3 γ γ − γ j γ i ψ (−t, x)
# $# $
2
= −ψ̄ (−t, x) σ ij ψ (−t, x) .

2. µ ̸= ν and either µ = 0 or ν = 0. For ν = 0 the time reversal transformation is

$ i # i∗ 0∗
T ψ̄σ i0 ψT −ψ̄ (−t, x) −γ 1 γ 3 γ γ − γ 0∗ γ i∗ γ 1 γ 3 ψ (−t, x)
# $# $
=
2
$i# i 0
− (−1) ψ̄ (−t, x) −γ 1 γ 3 γ 1 γ 3 γ γ − γ 0 γ i ψ (−t, x)
# $# $
=
2
= ψ̄ (−t, x) σ µ0 ψ (−t, x) .

Similarly, for µ = 0 the parity transformation is

T ψ̄σ 0j ψT = ψ̄ (t, −x) σ 0j ψ (t, −x) .

The time reversal transformation of the bilinear tensor can be summarized by the following formula
µ ν
T ψ̄σ µν ψT = − (−1) (−1) ψ̄ (t, −x) σ µν ψ (t, −x) .
Lastly, we work out how the bilinear tensor transforms under charge conjugation

C ψ̄σ µν ψC = C ψ̄Cσ µν CψC

$T $T
−iγ 0 γ 2 ψ σ µν −iψ̄γ 0 γ 2
# #
=
' (T
= − ψ̄γ 0 γ 2 (σ µν )T γ 0 γ 2 ψ
T
= −ψ̄γ 0 γ 2 (σ µν ) γ 0 γ 2 ψ

28
 $T $T
where in the last line we have used the fact that −iγ 0 γ 2 ψ σ µν −iψ̄γ 0 γ 2 is just a number so the transpose does not
# #

change anything. The transpose of the gamma matrices are

.T ; T
<
0 σµ (σ̄ µ )
-
µ T 0
(γ ) = = .
σ̄ ν 0 (σ ν )T 0

Since,
# 0 $T
σ = σ0
# 1 $T
σ = σ1
# 2 $T
σ = −σ 2
# 3 $T
σ = σ3

we see that B
µ T γµ µ = 0, 2
(γ ) =
−γ µ µ = 1, 3.
Therefore, # 0 2$ µ T
γ γ (γ ) = −γ µ γ 0 γ 2
# $

T
and we can show that γ 0 γ 2 (σ µν ) = −σ µν γ 0 γ 2 . Thus, the bilinear tensor under charge conjugation becomes

C ψ̄σ µν ψC = ψ̄σ µν γ 0 γ 2 γ 0 γ 2 ψ
= ψ̄σ µν ψ.

The charge conjugation transformation of the bilinear tensor can be summarized by the following formula

C ψ̄σ µν ψC = −ψ̄ (t, −x) σ µν ψ (t, −x) .

"
(b) Let φ (x) be a complex-valued Klein-Gordon field, such as we considered in Problem 2.2. Find unitary operators
P, C and an anti-unitary operator T (all defined in terms of their action on the annihilation operators ap and bp for the
Klein-Gordon particles and antiparticles) that give the following transformations of the Klein-Gordon fields:

P φ (t, x) P = φ (t, −x) ;

T φ (t, x) T = φ (−t, x) ;
Cφ (t, x) C = φ∗ (t, x) .

Find the transformation properties of the components of the current

J µ = i (φ∗ ∂ µ φ − ∂ µ φ∗ φ)

under P, C and T .
Proof: The quantized Klein-Gordon field is

d3 p1 #
ˆ
ap e−ip·x + b†p eip·x
$
φ (t, x) = 53
(2π) 2Ep
3
d p 1 # † ip·x
ˆ
φ∗ (t, x) = + bp e−ip·x
$
3
5 ap e
(2π) 2Ep

29
The natural definition for the parity operator is
P ap P = a−p
P bp P = b−p .
Let us test this definition on the KG field. Application of P yields
d3 p 1 +
ˆ ,
P φP = 3
5 P ap P e−ip·x + (P bp P )† eip·x
(2π) 2Ep
3
d p 1 +
ˆ ,
−iEp t+ip·x † iEp t−ip·x
= a −p e + b −p e .
(2π)3 2Ep
5

Defining p̃ = (Ep̃ , p̃) ≡ (Ep , −p), the above equation may be written as
ˆ 3
d (−p) 1 + ,
P φP =p↔−p 3
5 a−p e−iE−p t−ip·x + b†−p eiE−p t+ip·x
(2π) 2E−p
3
d p̃ 1 +
ˆ ,
−ip̃·(t,−x) † ip̃·(t,−x)
= 3
5 a p̃ e + b p̃ e
(2π) 2Ep̃
= φ(t, −x).
Thus, P takes φ(t, x) to φ(t, −x) as required. We note here that
P φ∗ P = φ∗ (t, −x).
This will be used later to find the transformation properties of the current.
Define the time reversal operator to be anti-linear so that when commuted past a c-number it conjugates the c-number.
The time reversal operator acting on the creation operators is defined to be:
T ap T = a−p
T bp T = b−p .
Let us test this definition on the KG field. Application of T yields
d3 p 1 +
ˆ ,
ip·x † −ip·x
T φT = 3
5 T a p T e + (T b p T ) e
(2π) 2Ep
3
d p 1 +
ˆ ,
= 3
5 a−p eip·x + b†−p e−ip·x
(2π) 2Ep
3
d p̃ 1 +
ˆ ,
ip̃·(t,−x) † −ip̃·(t,−x)
= 3
5 a p̃ e + b p̃ e
(2π) 2Ep̃
d3 p̃ 1 +
ˆ ,
= 3
5 ap̃ e−ip̃·(−t,x) + b†p̃ e−ip̃·(−t,x)
(2π) 2Ep̃
= φ(−t, x)
as required. We note here that
T φ∗ T = φ∗ (−t, x).
This will be used later to find the transformation properties of the current.
Define the time reversal operator to be anti-linear so that when commuted past a c-number it conjugates the c-number.
The time reversal operator acting on the creation operators is defined to be:
Cap C = bp
Cbp C = ap .

30
Let us test this definition on the KG field. Application of T yields

d3 p 1 +
ˆ ,
−ip·x † ip·x
CφC = 3
5 Ca p Ce + (Cb p C) e
(2π) 2Ep
d3 p 1 # −ip·x
ˆ
+ a†p eip·x
$
= 3
5 bp e
(2π) 2Ep
3
d p 1 # −ip·x
ˆ
+ a†p eip·x
$
= 3
5 bp e
(2π) 2Ep
= φ∗ (t, x)

as required. We note here that

Cφ∗ C = φ(t, x).
This will be used later to find the transformation properties of the current.
Now, we can find the transformation properties of the current. In addition to knowing how φ and φ∗ transform under P ,
T , and C, we need the transformation rules for the derivative ∂ µ . Parity takes x → −x and t → t and therefore, ∇ → −∇
and ∂0 → ∂0 . Time reversal takes x → x and t → −t and therefore, ∇ → ∇ and ∂0 → −∂0 . Charge conjugation does not
effect the derivative. With g µµ defined such that there is no sum we can write the transformations of the derivative in a
compact way,

P ∂ µP = g µµ ∂ µ
µ
T∂ T = −g µµ ∂ µ
C∂ µ C = ∂ µ.

Thus, under P , T , and C, the current becomes

P J µ (t, x)P = i (P φ∗ (t, x)P ∂ µ P φ(t, x)P − ∂ µ P φ∗ (t, x)P P φ(t, x)P )
= i (φ∗ (t, −x)g µµ ∂ µ φ(t, −x) − (g µµ ∂ µ φ∗ (t, −x)) φ(t, −x))
= (−1)µ J µ (t, −x)

T J µ (t, x)T = i (T φ∗ (t, x)T ∂ µ T φ(t, x)T − T ∂ µ T φ∗ (t, x)T T φ(t, x)T )
= i (φ∗ (−t, x) (−g µµ ) ∂ µ φ(−t, x) − ((−g µµ ) ∂ µ φ∗ (−t, x)) φ(−t, x))
µ
= − (−1) J µ (−t, x)

CJ µ (t, x)C = i (Cφ∗ (t, x)C∂ µ Cφ(t, x)C − ∂ µ Cφ∗ (t, x)CCφ(t, x)C)
= i (φ(t, x)∂ µ φ∗ (t, x) − (∂ µ φ(t, x)) φ∗ (t, x))
= i (φ(t, x)∂ µ φ∗ (t, x) − (∂ µ φ(t, x)) φ∗ (t, x))
= −J µ (t, x).

"
(c) Show that any Hermitian Lorentz-scalar local operator built from ψ (x), φ (x), and their conjugates has CP T = +1.
Proof: A Hermitian, Lorentz invariant operator must have φ paired with φ∗ and ψ paired with ψ̄. Additionally, the
Lorentz indices must be contracted to form a Lorentz scalar. Looking at table in P&S we can see that any combination
of these bilinears which yields a Hermitian, Lorentz invariant will have CP T = 1.
"

31
1.8 Bound states
Two spin-1/2 particles can combine to a state of total spin either 0 or 1. The wavefunctions for these states are odd and
even, respectively, under the interchange of the two spins.
(a) Use this information to compute the quantum numbers under P and C of all electron-positron bound states with S, P,
or D wavefunctions.
The Ps Schrodinger equation is the Schrodinger equation for a particle of electric charge e and reduced mass µ = me /2
in the Coulomb potential V (|x|) = −α/|x| where e (< 0) is the electron charge, me is the electron mass, α ≈ 1/137 is
the fine structure constant and x is the separation distance between the electron and positron in Ps. Instead of directly
solving the Ps Schrodinger equation, the position space wave functions of Ps can be obtained by taking advantage of the
similarity between the Schrodinger equation for Ps and that for hydrogen.
Specifically, the position space wave functions can be obtained by replacing the hydrogen Bohr radius with the Ps Bohr
radius in the hydrogen position space wave functions. These wave functions are characterized by the principal quantum
number (or energy quantum number), n, the orbital angular momentum quantum number, l, and the orbital angular
momentum projection quantum number, ml . The position space wave functions are
6- .
3 - .l - .
2 (n − l − 1)! −|x|/na 2|x| 2l+1 2|x|
ψnlml (x) = 3e Ln−l−1 Ylm (θ, φ), (1.1)
na 2n [(n + 1)!] na na
ml
where a = 2/me α is the Bohr radius of Ps, Lm
n−m are associated Laguerre polynomials and Yl are spherical harmonics.
Now that we have the position space wave functions of Ps, we can construct the Ps bound state. In the centre of mass
frame, the Ps bound state can be expressed by
√ d3 k 1 1
ˆ
|Ψnlml ;s,ms ⟩ = 2mPs ψnlml (k) √ √ |k; −k; s, ms ⟩, (1.2)
(2π)3 2me 2me
where s is the total spin of the Ps bound state and ms is the spin projection along the z-axis. The momentum space
wavefunction, ˆ
ψnlml (k) ≡ d3 xeik·r ψnlml (x), (1.3)

gives the amplitude for finding Ps in a given configuration where the electron has momentum k. The free state in equation
(1.2) is + 1
† −1† − 1 † 21 †
, √
|k; −k; 0, 0⟩ = ak2 b−k2 − ak 2 b−k |0⟩/ 2,
for p-Ps (s = ms = 0) and
⎧ 1† 1†
⎪ak2 b−k
2
|0⟩ for ms = 1,
√
⎪
⎨+ 1
−1† − 1 † 12 †
,
2†
|k; −k; 1, ms ⟩ = ak b−k2 + ak 2 b−k |0⟩/ 2 for ms = 0,
⎪
⎩ − 12 † − 12 †
⎪
ak b−k |0⟩ for ms = −1,

for o-Ps (s = 1) where |0⟩ is the vacuum state.

Applying the parity operator, we obtain
1 d3 k
ˆ
P |Ψnlml ;00 ⟩ = √ ψnlml (k)P |k; −k; 00⟩
me (2π)3
1 d3 k
ˆ + 1 1
− 12 † 21 †
,
2† −2†
=√ ψnlm (k)P a b
k −k − a k b −k P |0⟩
2me (2π)3 l

1 d3 k
ˆ + 1 1
− 12 † 21 †
,
2† −2†
= −|ηa |2 √ ψnlm (−k)P a b
k −k − a k b −k P |0⟩
2me (2π)3 l

32
where we have assumed that the vacuum state is even under parity. To simplify the above further, note from equation
(1.3), that the parity of ψnlml (k) is the same as ψlnml (x). Therefore, we can invert the spatial coordinate in (1.1) to
obtain the parity of the momentum space wavefunction. The only part of (1.1) that is sensitive to such an inversion is
the spherical harmonic, Ylm (θ, φ) → (−1)l Ylm (θ, φ). Thus, p-Ps transforms as

P |Ψnlml ;00 ⟩ = (−1)l+1 |Ψnlml ;00 ⟩

under parity. Similarity, o-Ps transforms as

P |Ψnlml ;1ms ⟩ = (−1)l+1 |Ψnlml ;1ms ⟩

under parity and has an P eigenvalue of (−1)l+1 . Notice that the relative phase difference between the fermion and
anti-fermion inversion phases, ηb∗ = −ηa , contributes a factor of −1 to the parity of Ps (p-Ps and o-Ps) in addition to the
parity of the wave function; this extra factor is called the intrinsic parity of a fermion--anti-fermion system.
The remaining C and T eigenvalues of Ps are obtained in a similar manner and are listed below in Table 1.

Discrete Transform Ps Eigenvalue

l+1
P (−1)
l+s
C (−1)
s+1
T (−1)

Table 1: P , C and T eigenvalues of the Ps state |Ψnlml ;sms ⟩. Here, l is the orbital angular momentum quantum number
and s is the spin angular momentum quantum number.

(b) Since the electron-photon coupling is given by the Hamiltonian

ˆ
∆H = d3 x eAµ j µ ,

where j µ is the electric current, electrodynamics is invariant to p and C if the components of the vector potential have
the same P and C parity as the corresponding components of j µ . Show that this implies the following surprising fact:
The spin-0 ground state of positronium can decay to 2 photon, but the spin-1 state must decay to 3 photons. Find the
selection rules for the annihilation of higher positronium states, and for 1-photon transitions between positronium levels.

33