Solution to Problems in

Quantum Field Theory

by Franz Mandl & Graham Shaw

Sanha Cheong October 4, 2017

sanha@stanford.edu
Stanford University

1 Photons and the Electromagnetic Field

1. The free-radiation field inside a cubic enclosure is given by the state:
 ∞
1 2 X cn

|ci = exp − |c| √ |ni
2 n=0 n!

where c = |c|eiδ is any complex number and |ni is the state with n photons with
some specific wavevector k and transverse polarization r, omitted in the expression.
(Such a state |ci is called a coherent state and is extremely useful in quantum
optics and other bosonic quantum field theory.)
(i) |ci is normalized: hc|ci = 1.
∞ ∞
! !
∗m n

2
 X c X c
hc|ci = exp −|c| hm| √ √ |ni
m=0 m! n=0 n!
∞ X ∞ 2 i(n−m)δ
2
−|c|
X |c| e
=e √ √ hm|ni
m=0 n=0 m! n!

Page 1 / 34
 At this point, we invoke the orthonormality hm|ni = δmn to obtain:
∞ X ∞
2
−|c|
X |c|2 ei(n−m)δ
=e √ √ δmn
m=0 n=0 m! n!
∞ 2
2 X |c|
= e−|c|
n=0
n!
2 2
−|c| |c|
=e e
=1
(ii) |ci is an eigenstate of ar (k) with the complex eigenvalue c: ar (k) |ci = c |ci.
 ∞
1 2 X cn

ar (k) |ci = exp − |c| √ ar (k) |ni
2 n=0 n!
∞
" #
n √
 
1 X c
= exp − |c|2 0+ √ n |n − 1i
2 n=1 n!

If we redefine the index by m = n − 1, we get:

 ∞
1 2 X cm+1

= exp − |c| √ |mi
2 m=0 m!
= c |ci
(iii) The expected number of photons in the enclosure N ≡ hc|N |ci is equal to |c|2 .

N ≡ hc|N |ci
= hc|a† a|ci
 2
= a |ci
 

Since a |ci = c |ci as shown in (ii), we get:

 2
= c |ci
 

= |c|2

Page 2 / 34
(iv) The RMS fluctuation of the photon number is given by:

(∆N )2 ≡ hc|N 2 |ci − N 2 = |c|2

Recall that ar (k), a†s (k0 ) = δrsδkk0 . Hence, in our case with some specific
 

wavevector k and polarization r, a, a† = 1. Then,

N 2 = a† a a† a

= a† (aa† )a
= a† (N + 1)a
= a† N a + N

Then, the RMS fluctuation becomes:

(∆N )2 ≡ hc|N 2 |ci − N 2

= hc|a† N a + N |ci − N 2
= hc|a† N a|ci + N − N 2
= |c|2 hc|N |ci + N − N 2
= |c|4 + |c|2 − |c|4
= |c|2
(v) The expected value of the electric field E is given by:
r
~ωk
E ≡ hc|E|ci = −εr (k)2 |c| sin(k • x − ωk t + δ)
2V
where V is the volume of the cubic enclosure.
Since we are only dealing with free radiation (i.e. no charge distribution,
φ = 0), we can let the longitudinal field EL be zero in the Coulomb gauge.
Then, the remaining electric field is purely transverse, which is given by:
1 ∂A (x, t)
E (x, t) = −
c r ∂t
~ωk h i(k x−ωk t)
• † −i(k x−ωk t)
•
i
= iεr (k) ar (k)e − ar (k)e
2V
At this point, we define the phase of the electric field: θ ≡ k • x − ωk t.

Page 3 / 34
 The mean electric field E is then:

E ≡ hc|E|ci
r
~ωk 
hc|ar (k)|ci eiθ − hc|a†r (k)|ci e−iθ

= iεr (k)
2V

From (ii), ar (k) |ci = c |ci and, similarly, hc| a†r (k) = hc| c∗ . Also, recall that
c = |c|eiδ . Hence, we get:
r
~ωk  iθ
ce − c∗ e−iθ

= iεr (k)
r 2V
~ωk 
2i Im ceiθ
 
= iεr (k)
2V
r
~ωk
= −εr (k)2 |c| sin(k • x − ωk t + δ)
2V
(vi) The RMS fluctuation ∆E in the electric field is given by:
~ωk
(∆E)2 ≡ hc|E 2 |ci − hc|E|ci2 =
2V
Note that the polarization vector εr has been omitted and the electric field here
is treated as a scalar quantity, since we are dealing with a definite pure polar-
ization.
First, we compute the operator E 2 :

E 2 ≡ EE
"r #2
~ωk
aeiθ − a† e −iθ


= i
2V
~ωk  2 i2θ
a e + a†2 e−i2θ − aa† − a† a

=−
2V
~ωk  2 i2θ
a e + a†2 e−i2θ − 2a† a − 1

=−
2V

Page 4 / 34
 Then, the RMS fluctuation ∆E is:

(∆E)2 ≡ hc|E 2 |ci − hc|E|ci2

~ωk  i2θ
e hc|a2 |ci + e−i2θ hc|a†2 |ci − 2 hc|a† a|ci − hc|ci

=−
2V
~ωk 2 2
−4 |c| sin (θ + δ)
2V
~ωk h 2 i2θ ∗2 −i2θ 2 2 2
i
=− c e +c e − 2|c| − 1 + 4|c| sin (θ + δ)
2V
~ωk h 2 i2θ 2 i2θ ∗

2 2 2
i
= 1−c e − c e + 2|c| − 4|c| sin (θ + δ)
2V
~ωk h 2
 n
i2(θ+δ)
o
2
i
= 1 − |c| 2 Re e − 2 + 4 sin (θ + δ)
2V
~ωk h 2 2 2 2

i
= 1 − |c| 2 cos (θ + δ) − 2 sin (θ + δ) − 2 + 4 sin (θ + δ)
2V
~ωk
=
2V
From the above calculations, we observe that, in the coherent state |ci which is
a superposition of states with all possible number of photons (we sum from n = 0 to
n = ∞), the relative fluctuations in both the mean number of photons and the mean
electric field strength vanish:
∆N
= |c|−1 = N −1/2
N
∆E
∝ |c|−1 = N −1/2
E
as N → ∞. Hence, in this limit, the state |ci behaves like a classical electromagnetic
field. In other words, classical beam of electromagnetic radiation is retained from
quantum electrodynamics by considering a coherent state in the limit of infinite mean
number of photons.

2. The Lagrangian of a point particle of mass m and electric charge q moving in an

electromagnetic potential (φ, A) is given by:
1 q
L (x, ẋ) = mẋ2 + A • ẋ − qφ
2 c

Page 5 / 34
(i) The momentum p conjugate to x is:
∂L q
p≡ = mẋ + A
∂ ẋ c
The Euler-Lagrange equations become:
 
d ∂L ∂L d  q  q ∂(Aj ẋj ) ∂φ
− = mẋi + Ai − +q
dt ∂ ẋ ∂x dt c c ∂xi ∂xi
 
d q ∂Ai ∂Ai q ∂Aj ∂φ
= m ẋi + + ẋj − x˙j + q
dt c ∂t ∂xj c ∂xi ∂xi
   
d 1 ∂Ai ∂φ q ∂Ai ∂Aj
= m ẋi + q − − + ẋj −
dt c ∂t ∂xi c ∂xj ∂xi
 
d 1 ∂A ∂φ q
= m ẋ + q − − + ẋ • (∇ × A)
dt c ∂t ∂x c
where we have used the index notation; repeated indices imply summation.
The electromagnetic fields E, B are described in terms of the potentials
as:
∂φ 1 ∂A
E=− − , B=∇×A
∂x c ∂t
and therefore the Euler-Lagrange equations give:
 
d 1
m ẋ = q E + ẋ • B
dt c
(ii) Given a Lagrangian L (x, ẋ), the Hamiltonian H(x, p) is:

H(x, p) = ẋ • p − L

Hence, in our case, the Hamiltonian becomes:

 
 q  1 q
H(x, p) = ẋ • mẋ + A − mẋ2 + A • ẋ − qφ
c 2 c
1
= mẋ2 + qφ
2
and since mẋ = p − qc A, we get the desired Hamiltonian:

1  q 2
= p − A + qφ
2m c

Page 6 / 34
 With this Hamiltonian, the Hamiltonian equations of motion give:
∂H
ẋ =
∂p
1  q 
= p− A
m c
from which we retain:
q
p = mẋ + A
c
and also
 
d q ∂Ai ∂Ai
ṗ = m ẋi + + x˙j
dt c ∂t ∂xj
∂H
=−
∂x
1  q  ∂  q  ∂φ
=− pj − A j − Aj − q
m c ∂xi c ∂xi
q ∂Aj ∂φ
= + ẋj −q
c ∂xi ∂xi
from which we retain:
 
d 1
m ẋ = q E + ẋ • B
dt c

3. Consider a Thomson Scattering process such that an unpolarized photon with a wavevec-
tor k collides with an electron, and a new photon with a wavevector k0 linearly po-
larized in a specific direction is emitted at an angle θ relative to the incoming photon
(k̂ • k̂0 = cos θ).
The differential cross-section for a Thomson scattering of photons with definite
initial and final polarizations α, β is given by:
2
σα→β dΩ = r02 εα • ε0β dΩ


2
where εα , ε0β are the initial and final polarization vectors respectively, and r0 ≡ 4πmc
e
2

is the classical electron radius.

To apply this result to our case, we average over the initial polarization α and
fix a definite final polarization, say β = 1. That is,
1 X
σunp.→1 dΩ = σα→1 dΩ
2 α=1,2

Page 7 / 34
 Since ε1 , ε2 , and k̂ form an orthonormal coordinate system,
X
2  2
0 0
εα • εβ = 1 − k̂ • εβ
α=1,2

Similarly, ε01 , ε02 , and k̂0 also form an orthonormal coordinate system. In this coordi-
nate system, we can write:
k̂ = (sin θ cos φ, sin θ sin φ, cos θ)
Therefore, we obtain:
1 X 2 2
σunp.→1 dΩ = r0 [εα • ε01 ] dΩ
2 α=1,2
1
= r02 1 − sin2 θ cos2 φ dΩ


2
If we consider both of the two perpendicular final polarizations, we obtain the
unpolarized differential cross-section:
σunp. dΩ = (σunp.→1 + σunp.→2 ) dΩ
1
= r02 2 − sin2 θ cos2 φ − sin2 θ sin2 φ dΩ


2
1 2
= r0 1 + cos2 θ dΩ


2
which is identical Eq. (1.69a).
Now, consider the case when the scattering angle θ = 90◦ . Since the choice of
the final state polarization vectors ε01 , ε02 are arbitrary within a plane perpendicular
0
to k̂ , we choose them such that:
k̂ × k̂0
ε01 =  
0
k̂ × k̂ 


and naturally ε02 = k̂0 ×ε01 for an orthonormal basis. By construction, k̂ = (cos φ, sin φ, 0)
is perpendicular to ε01 and therefore φ = π/2. Then,
σα→2 = 0
which implies that the final state polarization must be 100% in the ε01 direction, which
is the direction normal to the plane of scattering.

Page 8 / 34
 2 Lagrangian Field Theory
1. Show the the transformation
L 0 (φr , φr,α ) = L (φr , φr,α ) + ∂α Λα (x)
where Λα (x), α = 0, ... , 3 are arbitrary functions of the fields φr (x) does not alter
the equations of motions.
First, since Λα are functions of φr (x) only, we have:
∂Λα ∂φr ∂Λα ∂Λα
= = φ r,α
∂xα ∂xα ∂φr ∂φr
Then, the Euler-Lagrange equations with the new Lagrangian density L 0 be-
come:
∂L 0 ∂L 0 ∂Λα
    
∂ ∂ ∂L ∂
− = α + φs,α
∂xα ∂φr,α ∂φr ∂x ∂φr,α ∂φr,α ∂φs
∂Λα
 
∂
− L + φs,α
∂φ ∂φs
 r
∂Λα ∂ 2 Λα

∂ ∂L ∂L
= α + δrs − − φs,α
∂x ∂φr,α ∂φs ∂φr ∂φr ∂φs
 
∂ ∂L ∂L
= α −
∂x ∂φr,α ∂φ
 α r
∂ ∂Λ ∂ 2 Λα
+ α − φs,α
∂x ∂φr ∂φr ∂φs
 
∂ ∂L ∂L
= α −
∂x ∂φr,α ∂φr
2 α
∂φs ∂ Λ ∂ 2 Λα
+ α − φs,α
∂x ∂φs ∂φr ∂φr ∂φs
 
∂ ∂L ∂L
= α −
∂x ∂φr,α ∂φr
=0

2. The real Klein-Gordon field is described by the Hamiltonian density:

1 2 2 2

H (x) = c π (x) + (∇φ) + µ φ 2 2
2
Page 9 / 34
and the Hamiltonian of the field is:
Z
H≡ d3 x H (x)

The commutation relations of the fields are:

[φ(x, t), π(x0 , t)] = i~δ (x − x0 ) ,

[φ(x, t), φ(x0 , t)] = [π(x, t), π(x0 , t)] = 0

Before computing the commutation relations involving the Hamiltonian H, let

us first compute some simpler commutators.
 2 0
π (x , t), φ(x, t) = π(x0 , t)π(x0 , t)φ(x, t) − φ(x, t)π(x0 , t)π(x0 , t)


= π(x0 , t)π(x0 , t)φ(x, t)

 
− π(x , t)φ(x, t) + [φ(x, t), π(x , t)] π(x0 , t)
0 0

= π(x0 , t) [π(x0 , t), φ(x, t)] − [φ(x, t), π(x0 , t)] π(x0 , t)
= −2i~δ(x − x0 )π(x0 , t)

and
 2 0
φ (x , t), π(x, t) = φ(x0 , t)φ(x0 , t)π(x, t) − π(x, t)φ(x0 , t)φ(x0 , t)


= φ(x0 , t)φ(x0 , t)π(x, t)

 
− φ(x , t)π(x, t) + [π(x, t), φ(x , t)] φ(x0 , t)
0 0

= φ(x0 , t) [φ(x0 , t), π(x, t)] − [π(x, t), φ(x0 , t)] φ(x0 , t)
= 2i~δ(x0 − x)φ(x0 , t)
 
0 ∂ ∂ ∂
Also, define ∇ ≡ ∂x0 , ∂y0 , ∂z 0 . (This is to emphasize that the partial deriva-
tives are taken with respect to x0 , and ∇0 therefore does not act on functions of x,
e.g. ∇0 φ(x, t) = 0.) Then, we have:
h i
0 0 2
(∇ φ(x , t)) , φ(x, t) = 0

Page 10 / 34
and
h i  
0 0 2 0 0 0 0
(∇ φ(x , t)) , π(x, t) = ∇ φ(x , t) ∇ φ(x , t) π(x, t)
•

 
0 0 0 0
− π(x, t) ∇ φ(x , t) • ∇ φ(x , t)
 
0 0 0 0
= ∇ φ(x , t) ∇ φ(x , t)π(x, t)
•

 
− ∇ π(x, t)φ(x , t) • ∇0 φ(x0 , t)
0 0
 
0 0 0 0 0 0
= ∇ φ(x , t) π(x, t)∇ φ(x , t) + ∇ [φ(x , t), π(x, t)]
•

 
− ∇ φ(x , t)π(x, t) − ∇ [φ(x , t), π(x, t)] • ∇0 φ(x0 , t)
0 0 0 0

= ∇0 φ(x0 , t) • i~∇0 δ(x0 − x) + i~∇δ(x − x0 ) • ∇0 φ(x0 , t)

= 2i~∇0 δ(x − x0 ) • ∇0 φ(x0 , t)

Now, we simply combine the above results to compute the commutator relations
with the Hamiltonian:
Z  Z 
[H, φ(x)] = d3 x0 H (x0 , t) φ(x, t) − φ(x, t) d3 x0 H (x0 , t)
Z
3 0 1
 
2 2 0 0 0 2 2 2 0
= dx c π (x , t) + (∇ φ(x , t)) + µ φ (x , t) φ(x, t)
Z 2
3 0 1 2 2 0 0 0 2 2 2 0

− d x φ(x, t) c π (x , t) + (∇ φ(x , t)) + µ φ (x , t)
Z 2
1 
= d3 x0 c2 π 2 (x0 , t), φ(x, t)

Z 2
1
= d3 x0 c2 (−2i~δ(x − x0 )π(x0 , t))
2
2
= −i~c π(x)

and
Z
3 0 1 h 0 0 2
i
2
 2 0 
[H, π(x)] = d x (∇ φ(x , t)) , π(x, t) + µ φ (x , t), π(x, t)
Z 2
3 0 1
 
0 0 • 0 0 2 0 0
= dx 2i~∇ δ(x − x ) ∇ φ(x , t) + 2i~µ δ(x − x)φ(x , t)
2
= i~(µ2 − ∇2 )φ(x)

Page 11 / 34
3. Consider the Lagrangian density given by:
1  α β  1   µ2
L = − ∂α φβ (x) ∂ φ (x) + ∂α φ (x) ∂β φ (x) + φα (x)φα (x)
α β

2 2 2
where φα (x) is a real vector field.
The corresponding Euler-Lagrange equations are:
!
∂ ∂L ∂L ∂  
β α − α = β − φα + δα ∂γ φ − µ2 φα
,β β γ
∂x ∂φ ,β ∂φ ∂x
= −∂β ∂ β φα + ∂α ∂γ φγ − µ2 φα
= −  φα + ∂α ∂β φβ − µ2 φα
=0

Rewriting φα = gαβ φβ and rearranging, we get:

gαβ  +µ2 − ∂α ∂β φβ (x) = 0




Now, if we take apply an additional derivative ∂ α to the field equations, we get:

∂ α gαβ  +µ2 − ∂α ∂β φβ (x) = ∂β  φβ + µ2 ∂β φβ −  ∂β φβ




= µ 2 ∂β φβ
=0

For non-zero µ (i.e. the fields are massive), this implies the Lorenz Gauge condition:

∂α φα (x) = 0

4. The 3-momentum operator of the fields is defined as:

Z
∂φr (x)
P ≡ d3 x πr (x)
j
∂xj
This operator satisfies the following commutation relations:
 j
P , φr (x) ≡ P j φr (x) − φr (x)P j


∂φr (x0 , t) ∂φr (x0 , t)

Z
3 0 0 0
= d x πr (x , t) φr (x, t) − φr (x, t)πr (x , t)
∂x0j ∂x0j

Page 12 / 34
because the derivatives are with respect to x0 and does not affect φr (x, t),

∂φr (x0 , t)
Z
= d3 x0 [πr (x0 , t), φr (x, t)]
∂x0j
Z 0
3 0 0 ∂φr (x , t)
= − d x i~δ (x − x )
∂x0j
∂φr (x)
= −i~
∂xj
and, similarly,
∂φr (x0 , t) ∂φr (x0 , t)
Z
3 0 0 0
 j 
P , πr (x) = d x πr (x , t) πr (x, t) − πr (x, t)πr (x , t)
∂x0j ∂x0j
" #
Z 0
∂φr (x , t)
= d3 x0 πr (x0 , t) , πr (x, t)
∂x0j
Z
∂
= d3 x0 πr (x0 , t) 0 [φr (x0 , t), πr (x, t)]
∂xj
Z
∂
= i~ d3 x0 πr (x0 , t) 0 δ (x0 − x)
∂xj
∂πr (x)
= −i~
∂xj
Consider an arbitrary operator F (x) = F (φr (x), πr (x)) that can be written as
a power series of φr (x) and πr (x):
∞
X
F (x) = an φnr (x) + bn πrn (x)
n=0

where an ’s and bn ’s are arbitrary complex numbers.

Using the commutator identity

[A, B n ] = nB n−1 [A, B]

Page 13 / 34
 we se that F (x) satisfies:
∞
X
 j
am P j , φm
    j n 
P , F (x) = r (x) + bn P , πr (x)
m,n=0
X∞
am mφm−1 (x) P j , φr (x) + bn nπrn−1 (x) P j , πr (x)
   
= r
m,n=0
∞
X ∂φr (x) ∂πr (x)
= −i~ am mφm−1
r (x) + bn nπrn−1 (x)
m,n=0
∂xj ∂xj
∂F (x)
= −i~
∂xj
Define the 4-momentum of the fields as: P α = (H/c, P j ), α = 0, ... , 3 where
H is the Hamiltonian of the fields. Then, combining the above result with the Heisen-
berg equation of motion:
∂F (x)
[H, F (x)] = −i~c
∂x0
gives the covariant equations of motion:
∂F (x)
[P α , F (x)] = −i~
∂xα

5. Consider a scalar field φr (xα ) invariant under translations xα −→ x0α = xα + δα

where δα is a constant 4-vector. i.e.,
φ0 (x0α ) = φ(xα )

or, equivalently,

φ0 (xα ) = φ(xα − δα )
In the limit of an infinitesimal transformation (δα  1), we can write the trans-
formation as:
φ0 (xα ) = φ(xα − δα )
∂φ
+ O δ2


= φ(xα ) − δα
∂xα
1
≈ φ(xα ) + δα [P α , φ(x)]
i~

Page 14 / 34
 where we have used the result from the previous problem: [P α , F (x)] = −i~ ∂F (x)
∂xα .
Let us write the corresponding unitary transformation U as:
α
U = eiδα T
where T α is a Hermitian operator (T α† = T α ). In terms of U , the same translation
transformation can also be written as:
φ0 (x) = U φ(x)U †
≈ (1 + iδα T α )φ(x)(1 − iδα T α )
= φ(x) + iδα [T α , φ(x)] + O(δ 2 )
Comparing the two results, we see that T α = −P α /~ and therefore:
α
U = e−iδα P /~

3 The Klein-Gordon Field

1. A real Klein-Gordon field can be written as:
φ(x) = φ+ (x) + φ− (x)
s
X ~c2 
a(k)e−ikx + a† (k)eikx

=
2V ωk
k

where k is the wavenumber of the field and k 0 = ωk /c.

Then, we have:
∂φ(x)
φ̇(x) ≡
∂ts
X ~c2 h ωk  −ikx
 ω 
k † ikx
i
= −i c a(k)e + i c a (k)e
2V ωk c c
k
s
X ~c2
ωk a(k)e−ikx − a† (k)e−ikx
 
= −i
2V ωk
k
and therefore:
s
X ~c2
iφ̇(x) + ωk φ(x) = ωk 2a(k)e−ikx
2V ωk
k

Page 15 / 34
 To re-write the expression in the k-space, use Fourier transform, i.e. apply
R 0
V d3 x eik x on both sides. Then, we get:
r
Z
0
  Z X 2~c2 ωk 0
d3 x eik x iφ̇(x) + ωk φ(x) = d3 x a(k)e−ikx eik x
V V V
k
r
Z X 2~c2 ωk
= d3 x a(k)δkk0
V V
k
p
= 2~c2 V ωk0 a(k0 )
Rewriting this, we get:
Z  
2

−1/2 3 ikx
a(k) = 2~c V ωk d xe iφ̇(x) + ωk φ(x)
V

We define x = (x, t) and x0 = (x , t) to denote two different points in space.

0

Then, we have the following commutator relationship:

Z Z
† 0 1 0 0
d x d3 x0 ei(kx−k x )
3
 
a(k), a (k ) = 2
√
2~c V ωk ωk0 V
h V  
0 0
× iφ̇(x) + ωk φ(x) −iφ̇(x ) + ωk φ(x )0

  i
0 0
− −iφ̇(x ) + ωk0 φ(x ) iφ̇(x) + ωk φ(x)

since φ(x) is a real field and thus a Hermitian operator after quantization.
Given the commutator relationship [φ(x, t), φ̇(x0 , t)], this then becomes:
Z Z
1 0 0
a(k, a† (k0 )) = d3 x d3 x0 ei(kx−k x )
 
2
√
2~c V ωk ωk0 V V
h i
0 0
× iωk0 [φ̇(x), φ(x )] − iωk [φ(x), φ̇(x )]
Z Z
1 0 0
= 2
√ d3 x d3 x0 ei(kx−k x )
2~c V ωk ωk0 V V
h i
2 0 2 0
× ωk0 ~c δ (x − x) + ωk ~c δ (x − x )
~c2 (ωk + ωk0 )
Z
3 i(k−k 0 )x
= √ d x e
2~c2 V ωk ωk0 V
ωk + ωk 0
= √ V δkk0 ei(ωk −ωk0 )t
2V ωk ωk0
= δkk0

Page 16 / 34
 Also, we see that:
Z Z
0 1 0 0
[a(k), a(k )] = 2
√ d x d3 x0 ei(kx+k x )
3
2~c V ωk ωk0 V
h V  
0 0
× iφ̇(x) + ωk φ(x) iφ̇(x ) + ωk φ(x ) 0

  i
0 0
− iφ̇(x ) + ωk0 φ(x ) iφ̇(x) + ωk φ(x)
Z Z
1 3 3 0 i(kx+k 0 x0 )
= √ d x d xe
2~c2 V ωk ωk0 V V
h i
0 0
× iωk0 [φ̇(x), φ(x )] + iωk [φ(x), φ̇(x )]
Z Z
1 3 3 0 i(kx+k 0 x0 )
= √ d x d xe
2~c2 V ωk ωk0 V V
h i
2 0 2 0
× + ωk0 ~c δ (x − x) − ωk ~c δ(x − x )
Z
1 3 i(k+k 0 )x
h i
= √ d xe ωk 0 − ω k
2V ωk ωk0 V
1
= √ δk,−k0 (ωk0 − ωk )
2 ωk ωk 0
However, ωk = ω−k . Hence, when the delta function is non-zero (i.e. k = −k0 ),
ωk − ωk0 = 0. Therefore,

[a(k), a(k0 )] = 0

and, similarly,
 †
a (k), a† (k0 ) = 0


2. A pair of complex Klein-Gordon fields φ(x) and φ† (x) can be described by two
independent real fields φr (x), r = 1, 2:
1 1
φ = √ (φ1 + iφ2 ) , φ† = √ (φ1 − iφ2 )
2 2
and each of the real fields can be expanded as:
s
X ~c2 
ar (k)e−ikx + a†r (k)eikx

φr (x) =
2V ωk
k

Page 17 / 34
 Now, suppose we wish to expand the complex field φ(x) as:
s
X ~c2 
a(k)e−ikx + b† (k)eikx

φ(x) =
2V ωk
k

This expansion, in terms of the real fields φr (x), requires:

s
X ~c2 1 h −ikx † ikx −ikx † ikx
i
= √ a1 (k)e + a1 (k)e + ia2 (k)e + ia2 (k)e
2V ωk 2
k
s  
X ~c2 1
−ikx 1 † †
ikx
= √ a1 (k) + ia2 (k) e + √ a1 (k) + ia2 (k) e
2V ωk 2 2
k

Therefore, this expansion of the complex field φ(x) requires:

1
a(k) = √ (a1 (k) + ia2 (k))
2
and
1
b(k) = √ (a1 (k) − ia2 (k))
2
From here on, let x = (x, t) and x0 = (x0 , t). In other words, we will not
consider commutators of operators at different times in this problem.
Now, from the commutator relationships of the real field operators, we can com-
pute those of the complex field operators.
h i  1 1  
φ(x), φ̇† (x0 ) = √ (φ1 (x) + iφ2 (x)) , √ φ̇1 (x0 ) − iφ̇2 (x0 )
2 2
1 h
0
i h
0
i
= φ1 (x), φ̇1 (x ) + φ2 (x), φ̇2 (x )
2
= i~c2 δ (x − x0 )

since [φr (x), φr (x0 )] = i~c2 δ (x − x0 ) for any real field φr .

As one would expect based on the commutators of real fields, we see that:

φ(x), φ(x0 ) = φ̇(x), φ̇(x0 )

   

= φ† (x), φ† (x0 ) = φ̇† (x), φ̇† (x0 ) = 0

   

Page 18 / 34
Also, because φ1 , φ2 commute (at a given time) and φ, φ† are merely the linear com-
binations, we also see that:
 † 0
 h † 0
i
φ(x), φ (x ) = φ̇(x), φ̇ (x ) = 0

(In this sense, φ and φ† are independent fields.)

One less intuitive (and different from the properties of real fields) result is that:
h i  1 1  
φ(x), φ̇(x0 ) = √ (φ1 (x) + iφ2 (x)) , √ φ̇1 (x0 ) + iφ̇2 (x0 )
2 2
1 h
0
i h
0
i
= φ1 (x), φ̇1 (x ) − φ2 (x), φ̇2 (x )
2
=0
h i
† † 0
and, similarly, φ (x), φ̇ (x ) = 0.
Furthermore, we see that:
 
 † 0
 1
1 † 0 † 0


a(k), a (k ) = √ a1 (k) + ia2 (k) , √ a1 (k ) − ia2 (k )
2 2
1 
† 0
  † 0

= a1 (k), a1 (k ) + a2 (k), a2 (k )
2
= δkk0

and, similarly, b(k), b† (k0 ) = δk,k0 .

 

It is easy to see that

a(k), a(k0 ) = b(k), b(k0 ) = a(k), b(k0 ) = 0

     

since these will involve annihilation operators only, and the two real fields are inde-
pendent. The same holds for their Hermitian conjugates (creation operators only).
Lastly, we also can compute:
 
1 1
a(k), b† (k0 ) = √ a1 (k) + ia2 (k) , √ a†1 (k0 ) + ia†2 (k0 )
 


2 2
1 

† 0
  † 0

= a1 (k), a1 (k ) − a2 (k), a2 (k )
2
=0

Page 19 / 34
3. The Feynman ∆-function can be written as:
e−ikx
Z
1 4
∆F (x) = dk 2
(2π)4 k − µ2 + i
where we let  tend to zero after the
integration.
Applying the operator  +µ2 to this function, we get:

e−ikx
Z
2

2

1 4
 +µ ∆F (x) =  +µ dk 2
(2π)4 k − µ2 + i
Z
1 4 1 2

−ikx
= d k  +µ e
(2π)4 k 2 − µ2 + i
Z
1 4 1 2 2

−ikx
= d k −k + µ e
(2π)4 k 2 − µ2 + i
−ikx
4 −e
Z
1
= dk
(2π)4 1 + iκ
= −δ (4) (x)

where we have defined κ = /(k 2 − µ2 ) and let it tend to zero.

Hence, the Feynman ∆-function satisfies the inhomogeneous Klein-Gordon
equation.

4. Let φ be a complex Klein-Gordon field, representing a charged meson. It can be

written as:
s
−
X ~c2 
+
a(k)e−ikx + b† (k)eikx

φ(x) = φ (x) + φ (x) =
2V ωk
k

where the operators a and b follow the commutator relationships derived earlier.
Recall from Problem 2 that:

a(k), a(k0 ) = b† (k), b† (k0 ) = a(k), b† (k0 ) = 0

     

and therefore, unlike the commutator of a real field, we now have:

[φ(x), φ(x0 )] = 0

for any two different points x, x0 in spacetime.

Page 20 / 34
 Instead, for complex fields, we see that:

φ(x), φ† (x0 ) = φ+ (x), φ†− (x0 ) + φ− (x), φ†+ (x0 )

     

is non-zero.
The first commutator is:
 ~c2 X 1  0
φ (x), φ†− (x0 ) = a(k), a† (k0 ) e−i(kx−k x)
 + 
√
2V 0
ωk ωk 0
k,k
2
~c X 1 0
= √ δkk0 e−i(kx−k x)
2V 0
ωk ωk 0
k,k
0
~c2 X e−ik(x−x )
=
2V ωk
k

which, upon taking the limit V → ∞, becomes:

−ik(x−x0 )
~c2
Z
3 e
= dk
2(2π)3 ωk
+ 0
≡ i~c∆ (x − x )

where Z 3
+ 0 −ic d k −ik(x−x0 )
∆ (x − x ) ≡ e
2(2π)3 ωk
is defined the same way as it was defined for real field Klein-Gordon fields (neutral
mesons). Similarly, the second commutator is:
 −
φ (x), φ†+ (x0 ) = i~c∆− (x − x0 )


Therefore, we have:

φ(x), φ† (x0 ) = i~c ∆+ (x − x0 ) + ∆− (x − x0 ) ≡ i~c∆(x − x0 )

 

where ∆(x) ≡ ∆+ (x) + ∆− (x).

Now, define the Feynman ∆-function of a complex Klein-Gordon field φ as:

i~c∆F (x − x0 ) ≡ h0|T φ(x)φ† (x0 ) |0i



where T { } is the time-ordered product.

Page 21 / 34
 Then, this Feynman ∆-function, or the Feynman propagator, turns out to be of
the same form as that of a real Klein-Gordon field:
∆F (x) = θ(t)∆+ (x) − θ(−t)∆− (x)
e−ikx
Z
1 4
= d k
(2π)4 k 2 − µ2 + i
where θ(t) is the Heaviside
 step† function.
If t < t, then T φ(x)φ (x0 ) = φ(x)φ† (x0 ), and the Feynman propagator
0

becomes:
i~c∆F (x − x0 ) = h0|φ(x)φ† (x0 )|0i
= h0|φ+ (x)φ†− (x0 )|0i
which is interpreted as: a particle is created by a† (k) at x0 , then propagates to x which
is a later point in time, and is absorbed (annihilated) by a(k) at x.
When t0 > t, then T φ(x)φ† (x0 ) = φ† (x0 )φ(x), and therefore:
i~c∆F (x − x0 ) = h0|φ† (x0 )φ(x)|0i
= h0|φ†+ (x0 )φ− (x)|0i
which is interpreted as: an anti-particle is created by b† (k) at x, then propagates to
x0 , and is absorbed (annihilated) by b(k) at x.

5. The charge conjugation operator C is defined by:

φ(x) −→ C φ(x)C −1 = ηc φ† (x)
where C is a unitary operator which leaves the vacuum invariant (i.e., C |0i = |0i),
and ηc is a phase factor.
First, note that:
φ† (x) → C φ† (x)C −1

and, since C is unitary, we can rewrite this as:

= (C −1 )† φ† (x)C †

†
= C φ(x)C −1

†
= ηc φ† (x)
= ηc∗ φ(x)

Page 22 / 34
 Then, under
 the charge conjugation, the complex Klein-Gordon Lagrangian
density L = N (∂α φ† )(∂ α φ) − µ2 φ† φ becomes:
 
L = N (ηc ∂α φ)(ηc ∂ φ ) − µ (ηc φ)(ηc φ )
0 ∗ α † 2 ∗ †
 
∗ α † 2 †
= (ηc ηc ) N (∂α φ)(∂ φ ) − µ φφ

Since all commutators inside the normal product vanish, this simplifies to:
 
† α 2 †
= N (∂α φ )(∂ φ) − µ φ φ
=L
Thus, the Lagrangian density is invariant under the charge conjugation.
On the other hand, the charge-current density
α q  α † † α

s = −i N φ∂ φ − φ ∂ φ
~
becomes:
0α q  † α ∗ ∗ α †

s = −i N (ηc φ )∂ (ηc φ) − (ηc φ)∂ (ηc φ )
~
q ∗ 
† α α †

= −i (ηc ηc )N φ ∂ φ − φ∂ φ
~ 
q α † † α

= +i N φ∂ φ − φ ∂ φ
~
= −sα
Thus, the charge-current density changes sign under the charge conjugation.
Writing φ(x), φ† (x) out in their expansion forms in terms of the absorption and
creation operators, we see that:
s
X ~c2 
C φ(x)C −1 = C a(k)C −1 e−ikx + C b† (k)C −1 eikx

2V ωk
k
= ηc φ† (x)
s
X ~c2  †
a (k)eikx + b(k)e−ikx

= ηc
2V ωk
k

which implies that:

C a(k)C −1 = ηc b(k) , C b(k)C −1 = ηc∗ a(k)

Page 23 / 34
 Then, we also see that, under the charge conjugation, single-particle states trans-
form as:
C |a, ki = C a† (k) |0i
= C a† (k)C −1 |0i

†
= C a(k)C −1 |0i
= ηc∗ b† (k) |0i
= ηc∗ |b, ki

and, similarly,

C |b, ki = ηc |a, ki
where we have used the fact that the vacuum state |0i is unchanged by C or C −1 .
Thus, the charge conjugation C interchanges the particles (a-particles) and the
anti-particles (b-particles).

6. The parity transformation (i.e. spatial inversion) of the Hermitian (real) Klein-Gordon
field φ(x) is defined by:
φ(x, t) −→ Pφ(x, t)P −1 = ηp φ(−x, t)
where the parity operator P is unitary which leaves the vacuum invariant (i.e.,
P |0i = |0i), and ηp = ±1 is called the intrinsic parity of the field. Note that
P only influences the spatial argument of the field, but not the time t.
Under the given parity transformation,
the Lagrangian density of the real Klein-
Gordon field L = 21 (∂α φ)(∂ α φ) − µ2 φ2 transform as:
   
1 ∂φ(−x, t) ∂φ(−x, t)
L 0 (x, t) = ηp α
ηp − µ2 ηp2 φ2 (−x, t)
2 ∂x ∂xα
2 
ηp 1 ∂ 2 φ(−x, t)

= 2 2
− (∇φ(−x, t))2 − µ2 φ2 (−x, t)
2 c ∂t
2
 
1 1 ∂ φ(x, t)
= 2 2
− (−1)2 (∇φ(x, t))2 − µ2 φ2 (−x, t)
2 c ∂t
1
(∂α φ(x, t))(∂ α φ(x,t)) − µ2 φ2 (−x, t)
 
=
2
where we have used the chain rule for ∇ and the fact that ηp2 = +1.

Page 24 / 34
 Also, φ2 is necessarily an even function of each component of x. i.e.,
φ2 (x, t) = φ2 (−x, t)
Thus, the Lagrangian density L is invariant under the parity transformation:
1
L 0 (x, t) = (∂α φ(x, t))(∂ α φ(x, t)) − µ2 φ2 (x, t) = L (x, t)

2
Next, using the field operator φ(x) in the expansion form in terms of the ab-
sorption and creation operators, we can write:
s
X ~c2  0 −i(ωk0 t+k0 x) 0
+ a† (k0 )ei(ωk0 t+k x)
• •

φ(−x, t) = a(k )e
0
2V ωk0
k

Define k = −k0 . Note that ωk0 = ω−k = ωk . Then, we can write:

s
X ~c2 
a(−k)e−i(ωk t−k x) + a† (−k)ei(ωk t−k x)
• •

=
2V ωk
k

Then, the parity transformation of φ(x) can be written as:

s
X ~c2 
Pφ(x, t)P −1 = Pa(k)P −1 e−i(ωk t−k x) + Pa† (k)P −1 ei(ωk t−k x)
•
 •

2V ωk
k
s
X ~c2 
a(−k)e−i(ωk t−k x) + a† (−k)ei(ωk t−k x)
•

•
= ηp
2V ωk
k

Therefore, we derive the transformation laws for the absorption and creation
operators under P:
Pa(k)P −1 = ηp a(−k) , Pa† (k)P −1 = ηp a† (−k)
Then, we can derive how an arbitrary n-particle state transforms:
P |k1 , ... , kn i = Pa† (k1 ) ... a† (kn ) |0i
= Pa† (k1 ) ... a† (kn )P −1 |0i
= Pa† (k1 )P −1 ... Pa† (kn )P −1 |0i

= ηpn a† (−k1 ) ... a† (−kn ) |0i

= ηpn |−k1 , ... , −kn i

Page 25 / 34
 Now, let A and B arbitrary operators, and define:
B0 = B, Bn = [A, Bn−1 ] , for n = 1, 2, ...
Then, the following holds identically:
∞
iαA −iαA
X (iα)n
e Be = Bn
n=0
n!

We choose B = B0 = a(k).
(i) First, we define:
P1 ≡ eiα1 A1

where
π X
α1 = − , A1 = a† (k)a(k)
2
k

Then, we can calculate the commutators [A1 , Bn−1 ]:

" #
X
[A1 , B0 ] = a† (k0 )a(k0 ), a(k)
0
Xk 
a† (k0 ), a(k) a(k0 )

=
k0
X
=− δk0 k a(k0 )
k0
= −a(k)
and, by induction,
[A1 , Bn−1 ] = (−1)n a(k)
Therefore, we get:
P1 a(k)P1−1 = eiα1 A1 a(k)e−iα1 A1
∞
X 1  π n
= +i a(k)
n=0
n! 2
= eiπ/2 a(k)
= ia(k)

Page 26 / 34
(ii) Next, we define:

P2 ≡ eiα2 A2

where
π X
α2 = η p , A2 = a† (k)a(−k)
2
k

Then, the commutators [A2 , Bn−1 ] are:

[A2 , Bn1 ] = (−1)n a(−k)

Therefore, we get:

P2 a(k)P2−1 = eiα2 A2 a(k)e−iα2 A2

∞
X 1  π n
= −i ηp a(−k)
n=0
n! 2
= e−iηp π/2 a(−k)
= −iηp a(−k)

Since αr ’s and Ar ’s are real for both r = 1, 2, Pr = eiαr Ar are unitary. There-
fore, the product P1 P2 is unitary as well. Note that the product operator satisfies
the following identity:

(P1 P2 ) a(k) (P1 P2 )−1 = P1 P2 a(k)P2−1 P1−1

= P1 (−iηp a(−k)) P1−1
= +ηp a(−k)

which was derived earlier in this problem as the parity transformation law for the
absorption operator a(k).
Thus, P1 P2 gives an explicit form for the parity operator P.

Page 27 / 34
 4 The Dirac Field
1. For any given Dirac field ψ, we have the anti-commutator relationship:
  
ψ(x), ψ(y) ≡ ψ(x), ψ(y) + = iS(x − y)
where
 
∂ mc
S(x) = S + (x) + S − (x) = iγ µ µ + ∆(x)
∂x ~
Z 3
+ − −c dk
∆(x) = ∆ (x) + ∆ (x) = 3
sin(kx)
(2π) ωk
First, note that we have:
−c d3 k ∂
Z
∂
∆(x) = sin(kν xν )
∂xµ (2π)3 ωk ∂x µ

−c d3 k
Z
= kµ cos (ωk t − k • x)
(2π)3 ωk
Now, consider two points x, y at a same given time. That is,
x = (ct, x) and y = (ct, y)
Then, the anti-commutator above becomes:
 
ψ(x), ψ(y)  0 0 = iS(0, x − y)
x =y
Z 3 h
−ic dk µ i
= iγ kµ cos (−k • (x − y)) + sin (−k • (x − y))
(2π)3 ωk
Note that, if x − y 6= 0, the integrand is an odd function of k. Therefore, the
anti-commutator vanishes if x0 = y 0 and xi 6= y i . However, if x = y, the above
expression becomes:
Z 3 h Z 3 h
−ic dk µ i −ic d k 0 ωk i
i
iγ kµ cos 0 + sin 0 = i γ − γ ki
(2π)3 ωk (2π)3 ωk c
Z
0 1
= +γ 3
d3 k
(2π)
where the odd part of the integrand γ i ki vanishes again.
Therefore, the equal-time commutator for the Dirac field ψ is:
ψ(x), ψ(y) ≡ ψ(x), ψ(y) + = γ 0 δ(x − y)
  

Page 28 / 34
2.
    
∂ mc ∂ mc ∂ mc
iγ µ µ − S(x) = iγ µ µ − iγ ν ν + ∆(x)
∂x ~ ∂x ~ ∂x ~
  mc 2 
µ ν
= −γ γ ∂µ ∂ν + ∆(x)
~
Recall that γ µ γ ν = g µν . Defining µ ≡ mc/~, we get:

= −  +µ2 ∆(x)

=0
because ∆(x) satisfies the Klein-Gordon equation, which was mentioned in Chap-
ter 3.
Similarly, we also have:
    
∂ mc ∂ mc ∂ mc
iγ µ µ − SF (x) = iγ µ µ − iγ ν ν + ∆F (x)
∂x ~ ∂x ~ ∂x ~
= −  +µ2 ∆F (x)

= +δ (4) (x)
where the last step is the result of Chapter 3 Problem 3.
Hence, the functions S(x) and SF (x) satisfy the homogeneous and inhomoge-
neous Dirac equations respectively (because ∆(x) and ∆F (x) satisfy the homoge-
neous and inhomogeneous Klein-Gordon equations).

3. The charge-current density operator of the Dirac field ψ(x) is given by:
sµ (x) = −ecψ(x)γ µ ψ(x)
Let x, y be two different points at equal times. In other words, we have x0 =
y 0 = ct, but x 6= y. Then, thanks to the results from Problem 1, we have:

ψ(x), ψ(y) = 0
Also, we have the general anti-commutator relationships:
 
ψ(x), ψ(y) = ψ(x), ψ(y) = 0

SKIPPING FOR NOW. PHYSICALLY INTUITIVE, COMPUTA-

TIONALLY ANNOYING
Page 29 / 34
4. Recall the expansion of the real Klein-Gordon field φ(x):
s
−
X ~c2 
+
a(k)e−ikx + a† (k)eikx

φ(x) = φ (x) + φ (x) =
2V ωk
k

Now, supposes that we impose the anti-commutator relationships, instead of the

normal commutator relationships, on the absorption and creation operators:
a(k), a† (k0 ) = δkk0


{a(k), a(k0 )} = a† (k), a† (k0 ) = 0



This is equivalent to imposing Fermi-Dirac statistics, instead of Bose-Einstein statis-

tics.
The following identity will be useful in this problem.
  
A, B = AB − BA = 2AB − A, B}
Now, consider two points x, y in spacetime with a space-like separation (x − y).
First, we have:
 ~c2 X 1 h  † 0
−i(kx−k0 y)  † 0
i(kx−k0 y) i
φ(x), φ(y) = √ − a(k), a (k ) e − a (k), a(k ) e
2V 0
ωk ωk 0
k,k
~c2 X 1 
−i(kx−k 0 y) i(kx−k 0 y)

=− √ δkk0 e +e
2V 0
ω k ω k 0
k,k
~c2 X 1
=− cos (k(x − y))
V ωk
k

Since this quantity is Lorentz-invariant, it suffices to show its behavior for any
x, y with a space-like separation. Hence, we choose x, y such that they are at equal
time, but with non-zeros spatial separation x − y. Then, each cosine term is a non-
zero even function of k. Hence, the sum is non-zero.
Similarly, we can see that:
  
φ(x), φ(y) = 2φ(x)φ(y) − φ(x), φ(y)
~c2 X 1 
−i(kx−k 0 y) i(kx−k 0 y)

= 2φ(x)φ(y) + √ δkk0 e +e
2V 0
ω k ω k 0
k,k
2
~c X 1
= 2φ(x)φ(y) + cos (k(x − y))
V ωk
k

Page 30 / 34
 is also non-zero.
Therefore, having imposed the anti-commutator relationships on a(k), a† (k),
we see that neither the commutator nor the anti-commutator of the Klein-Gordon
field φ can be zero for x, y with a space-like separation.
This will prevent the commutator of any bilinear physical observable of φ at
x, y to be zero, which is a violation of microcausality. Thus, we observe that the
absorption and creation operators of the Klein-Gordon field must be quantized using
the commutator relationships (i.e. Bose-Einstein statistics).

5. For a Dirac field ψ, we define the chiral phase transformations as:

0
ψ(x) → ψ 0 (x) = exp(iαγ5 )ψ(x) , ψ † (x) → ψ † (x) = ψ † (x) exp(−iαγ5 )
where γ5 ≡ iγ0 γ1 γ2 γ3 and α is an arbitrary real parameter.
Note that γ5 anti-commutes with all γµ ’s (µ = 0, 1, 2, 3). That is, each time the
order of product between γ5 and a γµ is exchanged, the term gains a negative sign.
Under this transformation, the Lagrangian density
 
∂
L = cψ(x) i~γ µ µ − mc ψ(x)
∂x
transforms to:
 
∂
L 0 = cψ̄ 0 (x)γ 0 i~γ µ µ − mc ψ 0 (x)
∂x
 
† 0 µ ∂
= cψ (x) exp(−iαγ5 )γ i~γ − mc exp(iαγ5 )ψ(x)
∂xµ
The first term has two powers of the Dirac γ-matrix: the γ 0 originating from the
adjoint spinor ψ 0 and the γ µ in front of the derivative. Hence, the first term can be re-
ordered such that the two exponential factors due to the chiral phase transformation
cancel each other. However, the second term (i.e. the mass term) has only one γ 0 ,
and thus gains a negative sign after such re-ordering. Hence,
 
µ ∂
L = cψ(x) i~γ
0
+ mc ψ(x)
∂xµ
Therefore, the Lagrangian density is invariant only if the mass is zero.
The infinitesimal chiral phase transformation is obtained in the limit α  1:
ψ 0 (x) = ψ(x) + δψ(x) = ψ(x) + iαγ5 ψ(x)
0
ψ † (x) = ψ † (x) + δψ † (x) = ψ † (x) − iαψ † (x)γ5

Page 31 / 34
Because the massless Lagrangian density is invariant under this transformation, we
have the corresponding conserved current:
 
µ ∂L
f (x) = δψ(x)
∂ψ,µ
= cψ(x)i~γ µ iαγ5 ψ(x)

= −α~cψ(x)γ µ γ5 ψ(x)

Re-writing without the constants, we see that the axial vector current

jAµ (x) = ψ(x)γ µ γ5 ψ(x)

is conserved under the chiral phase transformation of a massless Dirac field.

Now, for a given massive Dirac field ψ(x), consider two new fields:
1 1
ψL (x) ≡ (1 − γ5 ) ψ(x) , ψR (x) ≡ (1 + γ5 ) ψ(x)
2 2
The subscripts L,R stand for “left-handed” and “right-handed” components of the
Dirac field ψ. They are also called “negative/positvie helicity” components, respec-
tively.
Note that:

ψL + ψR = ψ , ψR − ψL = γ5 ψ


Applying the operator i~∂/ − mc to the left of ψL,R gives their equations of
motion:
 

µ ∂ 1
i~∂/ − mc ψL (x) = i~γ − mc (1 − γ5 ) ψ(x)
∂xµ 2
    
1 ∂ ∂
= i~γ µ µ − mc ψ(x) − i~γ µ µ − mc γ5 ψ(x)
2 ∂x ∂x
   
1 ∂
= 0 + γ5 i~γ µ µ + mc ψ(x)
2 ∂x
   
1 ∂
= γ5 i~γ µ µ − mc ψ(x) + 2mcγ5 ψ(x)
2 ∂x
= mcγ5 ψ(x)
= mc (ψR (x) − ψL (x))


where we have used that ψ(x) satisfies the Dirac equation: i~∂/ − mc ψ(x) = 0.

Page 32 / 34
 Similarly, we also have:

1

i~∂/ − mc ψR (x) = i~∂/ − mc (1 + γ5 )ψ(x)
2
1

= γ5 −i~∂/ − mc ψ(x)
2
= −mcγ5 ψ(x)
= −mc (ψR (x) − ψL (x))
If the mass m of the Dirac field ψ(x) was zero, then the equations of motion for
ψL , ψR would be decoupled:
∂
i~γ µ ψL,R = 0
∂xµ
Now, consider the Lagrangian density:
LL = i~cψ L (x)γ µ ∂µ ψL (x)
where
1 − γ5
ψL = ψ
2
The corresponding Euler-Lagrange equations are:
 
∂ ∂LL ∂LL ∂  µ

− = µ i~cψ L γ
∂xµ ∂ψL,µ ∂ψL ∂x
 
∂ 1 − γ5
= i~c µ ψ † γ 0 γµ
∂x 2
 
∂ † 1 + γ5 0
= i~c µ ψ γ γµ
∂x 2
∂
= i~c µ ψR γ µ
∂x
=0
and
!
∂ ∂LL ∂LL ∂
− = −i~cγ µ ψL
∂xµ ∂ψ L,µ ∂ψ L ∂xµ
=0
Therefore, the given Lagrangian LL describes the right-handed massless anti-
fermions and the left-handed massless fermions only.

Page 33 / 34
5 Photons: Covariant Theory

Page 34 / 34