Quantum Field Theory I

Problem Sets

ETH Zurich, 2019 HS

Prof. N. Beisert

c 2012–2019 Niklas Beisert, ETH Zurich
This document as well as its parts is protected by copyright.
Reproduction of any part of this work in any form without
prior written consent of ETH Zurich is permissible only for
private, scientific and non-commercial use.
The current version of this work can be found at:
http://people.phys.ethz.ch/~nbeisert/lectures/
Contents
Problem Set 1 1.1
1.1. Classical particle in an electromagnetic field . . . . . . . . . . . . . . . . . 1.1
1.2. Relativistic point particle . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1
1.3. Coherent quantum oscillator . . . . . . . . . . . . . . . . . . . . . . . . . 1.2
1.4. Coupled pendula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2

Problem Set 2 2.1

2.1. Integral definition of the step function . . . . . . . . . . . . . . . . . . . . 2.1
2.2. Discrete and continuous treatment of a 1D spring lattice . . . . . . . . . . 2.2
2.3. Classical field momentum . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2

Problem Set 3 3.1

3.1. Scalar field correlator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1
3.2. Commutator and causality . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2
3.3. Complex scalar field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2

Problem Set 4 4.1

4.1. Retarded propagator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1
4.2. Conservation of charge with complex scalar fields . . . . . . . . . . . . . . 4.1
4.3. Symmetry of the stress-energy tensor . . . . . . . . . . . . . . . . . . . . 4.2

Problem Set 5 5.1

5.1. Representations of the Lorentz algebra . . . . . . . . . . . . . . . . . . . . 5.1
5.2. Properties of gamma-matrices . . . . . . . . . . . . . . . . . . . . . . . . 5.2
5.3. Dirac and Weyl representations of the gamma-matrices . . . . . . . . . . 5.2

Problem Set 6 6.1

6.1. Spinor rotations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1
6.2. Completeness for gamma-matrices . . . . . . . . . . . . . . . . . . . . . . 6.1
6.3. Fierz identity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2
6.4. Gordon identity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2

Problem Set 7 7.1

7.1. Helicity and chirality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1
7.2. Electrodynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2
7.3. Polarisation vectors of a massless vector field . . . . . . . . . . . . . . . . 7.2

Problem Set 8 8.1

8.1. The massive vector field . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1
8.2. Hamiltonian formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1
8.3. Commutators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1
8.4. Photon propagator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2
8.5. Interacting scalar field theory and scale invariance . . . . . . . . . . . . . 8.2

Problem Set 9 9.1

9.1. Feynman propagators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1
9.2. Wick’s theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1
9.3. Interaction picture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2

3
Problem Set 10 10.1
10.1. Four-point interaction in scalar QED . . . . . . . . . . . . . . . . . . . . 10.1

Problem Set 11 11.1

11.1. Møller scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1
11.2. Compton scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1
11.3. Kinematics in 2 → 2 scattering . . . . . . . . . . . . . . . . . . . . . . . 11.2

Problem Set 12 12.1

12.1. Optical theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.1
12.2. Muon pair production . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.2
12.3. Polarised muon-electron scattering . . . . . . . . . . . . . . . . . . . . . 12.2

Problem Set 13 13.1

13.1. Feynman and Schwinger parameters . . . . . . . . . . . . . . . . . . . . 13.1
13.2. Volume of spheres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.1
13.3. Basic loop integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.2
13.4. Loop integral numerators . . . . . . . . . . . . . . . . . . . . . . . . . . 13.2
13.5. Dimensional regularisation . . . . . . . . . . . . . . . . . . . . . . . . . . 13.2

Problem Set 14 14.1

14.1. Electron self energy structure . . . . . . . . . . . . . . . . . . . . . . . . 14.1
14.2. A one-loop correction to scattering in QED . . . . . . . . . . . . . . . . 14.1

4
Quantum Field Theory I Problem Set 1
ETH Zurich, 2019 HS Prof. N. Beisert

1.1. Classical particle in an electromagnetic field

Consider the classical Lagrangian density of a particle of mass m and charge q, moving
in an electromagnetic field, specified by the electric potential φ(~x, t) and the magnetic
~ x, t):
vector potential A(~
L = 21 m~x˙ 2 + q A·
~ ~x˙ − qφ. (1.1)
Determine the following quantities, and compare the results to those for a free particle:

a) the canonical momentum pi conjugate to the coordinate xi ;

b) the equations of motion corresponding to the Lagrangian density;

c) the Hamiltonian density of the system.

1.2. Relativistic point particle

The action of a relativistic point particle is given by
Z
S = −α dτ (1.2)
P

with the proper time (relativistic time-like line element)

dτ 2 = −ηµν dxµ dxν = dt2 − dx2 − dy 2 − dz 2 (1.3)

and α a (yet to be determined) constant.

The path P between two points xµ1 and xµ2 can be parametrised by a parameter λ. With
that, the integral of the line element dτ becomes an integral over the parameter
Z τ2 r
∂xµ ∂xν
S = −α dλ −ηµν . (1.4)
τ1 ∂λ ∂λ

a) Parametrise the path by the time coordinate λ = t = x0 and take the non-relativistic
limit k~x˙ k  1 to determine the value of the constant α.

b) Derive the equations of motion by varying the action. Hint: You may want to deter-
mine the canonically conjugate momentum first.

−→

1.1
1.3. Coherent quantum oscillator
Consider the Hamiltonian of a quantum harmonic oscillator:

p2 mω 2 x2
H= + . (1.5)
2m 2

a) Introduce ladder operators to diagonalise the Hamiltonian.

b) Calculate the expectation values of the number operator N ∼ a† a as well as of the x

and p operator in a general number state |ni.

c) Calculate the variances ∆x, ∆p and ∆N in the same state |ni and use them to
determine the Heisenberg uncertainty of |ni.

d) Show that the coherent state

|αi = eαp |0i (1.6)

is an eigenstate of the annihilation operator you defined in part a).

e) Calculate the time-dependent expectation values of x, p and N ,

hα|x(t)|αi, hα|p(t)|αi, hα|N (t)|αi, (1.7)

as well as the corresponding variances to determine the uncertainty of the state |αi.
Compare your result with the result obtained in part c).

1.4. Coupled pendula

We consider a system of three identical pendula of length ` and mass m in a homogeneous
gravitational field with acceleration g. The pendula are moving in the same plane and
we denote the (small) deflection angles by θj , j = 1, 2, 3. Moreover, the pendula are
connected by massless springs whose length equals the distance of the pendula in their
gravitational ground state θj = 0. At first, consider two springs of equal spring constant
k connecting the pairs of pendula 1, 2 and 2, 3.

a) Find the equations of motion for θk and the normal modes of the system.

Now add a third spring of spring constant k 0 connecting pendula 1, 3.

b) Find the equations of motion for θk and the normal modes of the system.

c) optional: For some value of k 0 the system behaves in a special way. Can you deter-
mine k 0 ? Can you tell in what sense the system becomes special?

1.2
Quantum Field Theory I Problem Set 2
ETH Zurich, 2019 HS Prof. N. Beisert

2.1. Integral definition of the step function

In this exercise we will demonstrate that:
d
θ(x) = δ(x), (2.1)
dx
where: ( (
f (0) if 0 ∈ A,
Z
1 if x > 0,
θ(x) = dx δ(x) f (x) = (2.2)
0 if x < 0, A 0 otherwise.

To that end, consider the following function (where x, z ∈ R,  > 0):

Z +∞
1 eı̊xz
F (x, ) = dz . (2.3)
2πı̊ −∞ z −ı̊

a) Consider a semi-circular path γ± (R) of radius R in the upper/lower half of the complex
plane and ending on the real axis.

Im z
x>0
γ+
ı̊ Re z (2.4)
−R +R
γ−
x<0

Argue that:

eı̊xz
Z
lim dz = 0 if x > 0,
R→∞ γ+ (R) z −ı̊
eı̊xz
Z
lim dz = 0 if x < 0. (2.5)
R→∞ γ− (R) z −ı̊

Hint: Use integration by parts to improve convergence of the integral.

b) Consider the closed path Γ = [−R, +R] ∪ γ± (R) and make use of the Cauchy integral
formula: I
1
dz f (z) = resΓ f, (2.6)
2πı̊ Γ
where resΓ f is the sum of the residues of the poles of f surrounded by the contour
Γ , to show that:
lim+ F (x, ) = θ(x). (2.7)
→0

c) Finally, using (2.3), show relation (2.1). You will have to perform some mathemati-
cally questionable steps. Which are they precisely? Can they be justified? How?

−→

2.1
2.2. Discrete and continuous treatment of a 1D spring lattice
Consider a one-dimensional array of N particles at positions qi (t), i = 1, . . . , N connected
by elastic springs with spring force constant κ. Assume that all of the particles have mass
m, and at rest their relative distance is a.

a) Derive the Lagrangian L(qi (t), q̇i (t)) of this system and compute the Euler–Lagrange
equations.

b) Determine the continuum form of these equations by taking the limit a → 0 and
N → ∞, where the mass density µ := lima→0 (m/a) and the elastic modulus Y :=
lima→0 (κa) are kept fixed, and: qk (t) → φ(x0 + ka, t) where k ∈ Z, x0 + ka is the
position on the spring lattice, and φ is a smooth field.

c) Directly take the continuum limit of L(qi (t), q̇i (t)) and show that the Euler–Lagrange
equations of the Lagrangian density L(φ, φ0 , φ̇) are the same as those obtained in
part b).

2.3. Classical field momentum

Consider the Lagrangian of a real scalar field φ = φ(x):

L = − 21 ∂µ φ ∂ µ φ − 12 m2 φ2 . (2.8)

a) Show that the Noether charge corresponding to spatial translations is given by:
Z
P = − d~x 3 π ∂ i φ.
i
(2.9)

b) Show that the momentum reduces to the following form using Fourier modes:

d~p 3
Z
P~ = p~ a∗ (~p)a(~p). (2.10)
(2π)3 2e(~p)

c) Calculate the Poisson bracket {P i , φ(~x)} and interpret the result.

2.2
Quantum Field Theory I Problem Set 3
ETH Zurich, 2019 HS Prof. N. Beisert

3.1. Scalar field correlator

In this problem we shall consider the amplitude for a particle to be created at point y
and annihilated at point x
∆+ (y, x) :=ı̊h0|φ(y)φ(x)|0i. (3.1)

a) Use the Fourier expansionpof φ(x) to show the following integral expression for
∆+ (y, x) with p0 = e(~p) := ~p 2 + m2
d~p 3
Z
∆+ (y, x) =ı̊ eı̊p·(y−x) . (3.2)
(2π)3 2e(~p)

b) Observe that the amplitude satisfies

∆+ (y, x) = ∆+ (y − x, 0) =: ∆+ (y − x). (3.3)
What are the properties of ∆+ under translations and Lorentz transformations?
c) Use Cauchy’s residue theorem to show that ∆+ (x) can be also written as
dp4 eı̊p·x
Z
∆+ (x) = − 4 2 2
, (3.4)
C+ (2π) p + m

where the integration over the contour C+ given in the left figure of (3.7) corresponds
to the (complex) variable p0 .
d) Show that ∆+ (x) satisfies the Klein–Gordon equation, i.e.
(−∂ 2 + m2 )∆+ (x) = 0. (3.5)

e) Express ∆+ (x) for a time-like x as a single integral over the energy, and the one for
space-like x as a single integral over p = k~pk.
Hint: Use a Lorentz transformation to reduce to the cases ~x = 0 and x0 = 0 respec-
tively.
f ) For space-like x with r := k~xk, show the asymptotic behaviour of the correlator
∆+ (x) ∼ e−mr for r → ∞. (3.6)
Hint: Wrap the contour of integration around the upper branch cut of the integrand
as depicted in the right figure of (3.7). Use the fact that there is a phase shift of
eı̊π = −1 between the two sides of the branch cut.

Im p0 Im p
C+
+ı̊m
−e(~
p) Re p0 Re p
(3.7)
+e(~
p)
−ı̊m

−→

3.1
3.2. Commutator and causality
In order to know whether a measurement of the field at x can affect another measurement
at y, one may compute the commutator

∆(y − x) :=ı̊h0|[φ(y), φ(x)]|0i = ∆+ (y − x) − ∆+ (x − y). (3.8)

Show that such a commutator vanishes for a space-like separation of x and y, which proves
that causality is obeyed.

3.3. Complex scalar field

We want to investigate the theory of a complex scalar field φ = φ(x). The theory is
described by the Lagrangian density:

L = −∂µ φ† ∂ µ φ − m2 φ† φ. (3.9)

As a complex scalar field has two degrees of freedom, we can treat φ and φ† as independent
fields with one degree of freedom each.

a) Find the conjugate momenta π(~x) and π † (~x) to φ(~x) and φ† (~x) and the canonical
commutation relations. Note: we choose π = ∂L/∂ φ̇† rather than π = ∂L/∂ φ̇.

b) Find the Hamiltonian of the theory.

c) Introduce creation and annihilation operators to diagonalise the Hamiltonian.

d) Show that the theory contains two sets of particles of mass m.

e) Consider the conserved charge

Z
Q =ı̊ d~x 3 (π † φ − φ† π). (3.10)

Rewrite it in terms of ladder operators and determine the charges of the two particle
species.

3.2
Quantum Field Theory I Problem Set 4
ETH Zurich, 2019 HS Prof. N. Beisert

4.1. Retarded propagator

Consider the commutator ∆(y − x) =ı̊[φ(y), φ(x)] and define
GR (y − x) := θ(y 0 − x0 ) ∆(y − x), (4.1)
which clearly vanishes for any y 0 < x0 .
a) Show that

Im p0

dp4 eı̊p·x CR
Z
GR (x) = , Re p0 (4.2)
CR (2π)4 p2 + m2 −e(~
p) +e(~
p)

with the contour CR given in the figure.

b) Check that GR (x) is a Green function for the Klein–Gordon equation,
(−∂ 2 + m2 )GR (x) = δ 4 (x). (4.3)

4.2. Conservation of charge with complex scalar fields

Consider a free complex scalar field described by
L = −(∂µ φ∗ )(∂ µ φ) − m2 φ∗ φ. (4.4)

a) Show that the transformation

φ(x) → φ0 (x) = eı̊α φ(x) (4.5)
leaves the Lagrangian density invariant.
b) Find the conserved current associated with this symmetry.
If we now consider two complex scalar fields, the Lagrangian density is given by
L = −(∂µ φ∗a )(∂ µ φa ) − m2 φ∗a φa , a = 1, 2. (4.6)

c) Show that
φa (x) → φ0 a (x) = U a b φb (x) (4.7)
with U ∈ U(2) = {A ∈ C2×2 ; A−1 = A† = (A∗ )T } is a symmetry transformation.
d) Show that now there are four conserved charges: one given by the generalisation of
part b), and the other three given by
Z
ı̊
d~x 3 φ∗a (σ i )a b π b − πa∗ (σ i )a b φb ,


Qi = − (4.8)
2
where σ i are the Pauli matrices.
−→

4.1
4.3. Symmetry of the stress-energy tensor
Consider a relativistic scalar field theory specified by some Lagrangian L(φ, ∂φ).

a) Compute the variation of L(φ(x), ∂φ(x)) under infinitesimal Lorentz transformations

(note: ω µν = −ω νµ )
xµ → xµ − ω µ ν xν + . . . . (4.9)

b) Assuming that L(x) transforms as a scalar field, i.e. just like φ(x), derive another
expression for its variation under Lorentz transformations.

c) Compare the two expressions to show that the two indices of the stress-energy tensor
are symmetric
δL
T µν = − ∂ ν φ + η µν L = T νµ . (4.10)
δ(∂µ φ)

4.2
Quantum Field Theory I Problem Set 5
ETH Zurich, 2019 HS Prof. N. Beisert

5.1. Representations of the Lorentz algebra

The Lie algebra so(d, 1) of the Lorentz group SO(d, 1) for (d + 1)-dimensional spacetime
is given in terms of its generators M µν (the relativistic angular momentum tensor),

[M µν , M λκ ] =ı̊ η µκ M νλ + η νλ M µκ − η νκ M µλ − η µλ M νκ .


(5.1)

Any representation of the Lorentz algebra must satisfy the above commutation relation.

a) Show explicitly that the following generators J µν of the vector representation satisfy
the Lie algebra
(J µν )ρ σ :=ı̊(η µρ δσν − η νρ δσµ ). (5.2)

b) Show explicitly that the following differential operators Lµν satisfy the Lie algebra

Lµν :=ı̊(xµ ∂ ν − xν ∂ µ ). (5.3)

−→

5.1
5.2. Properties of gamma-matrices
The gamma-matrices in D dimensional spacetime satisfy the Clifford algebra

{γ µ , γ ν } = 2η µν id . (5.4)

Derive the following identities using this algebraic relation (rather than an explicit matrix
representation).
a) Prove the following contraction identities

γ µ γµ = D id ,
γ µ γ ν γµ = −(D − 2)γ ν ,
γ µ γ ν γ ρ γµ = (D − 4)γ ν γ ρ + 4η νρ id ,
γ µ γ ν γ ρ γ σ γµ = −(D − 4)γ ν γ ρ γ σ − 2γ σ γ ρ γ ν . (5.5)

b) Show that a trace of an odd number n of gamma-matrices is zero for an even number
of spacetime dimensions D

tr(γ µ1 · · · γ µn ) = 0. (5.6)

Hint: Eliminate double indices, insert id = (γ ρ )−1 γ ρ for some index value ρ (no
summation convention implied), and use cyclicity of the trace.
c) Show the following trace identities

tr(γ µ γ ν ) = tr(id) η µν ,
tr(γ µ γ ν γ ρ γ σ ) = tr(id) (η µν η ρσ − η µρ η νσ + η µσ η νρ ). (5.7)

5.3. Dirac and Weyl representations of the gamma-matrices

Using the Pauli matrices σ i , i = 1, 2, 3, together with the 2 × 2 identity matrix σ 0 ,
       
0 1 0 1 0 1 2 0 −ı̊ 3 1 0
σ := , σ := , σ := , σ := , (5.8)
0 1 1 0 ı̊ 0 0 −1
we can realise the Dirac representation of the gamma-matrices,

γD0 :=ı̊σ 3 ⊗ σ 0 , γDj := σ 1 ⊗ σ j (j = 1, 2, 3), (5.9)

where the tensor product can be written as a 4 × 4 matrix in 2 × 2 block form as follows
 
A11 B A12 B
A⊗B = . (5.10)
A21 B A22 B

We denote the Pauli matrices collectively by σ µ = (σ 0 , σ i ) and define σ̄ µ = (−σ 0 , σ i ). We

can then define the gamma-matrices in the Weyl representation
 
µ 0 σµ
γW := . (5.11)
σ̄ µ 0

Show that both representations satisfy the Clifford algebra {γ µ , γ ν } = 2η µν id.

µ
Can you show their equivalence, i.e. γW = T γDµ T −1 for some matrix T ?
µ
Hint: It may help to express the γW as tensor products of Pauli matrices.

5.2
Quantum Field Theory I Problem Set 6
ETH Zurich, 2019 HS Prof. N. Beisert
6.1. Spinor rotations
The Dirac equation is invariant under Lorentz transformations Ψ 0 (x0 ) = SΨ (x) if the
spinor transformation matrix S satisfies
Λµ ν S −1 γ ν S = γ µ . (6.1)
For an infinitesimal Lorentz transformation Λµν = ηµν + δωµν + . . . this is fulfilled if
S = 1 − 41 δωµν γ µ γ ν + . . . = 1 − 81 δωµν [γ µ , γ ν ] + . . . . (6.2)
a) Find the infinitesimal spinor transformation δS for a rotation around the z-axis, i.e.
the only non-zero components of δωµν are δω12 = −δω21 6= 0.
b) Finite transformations are obtained by exponentiation,
S = exp − 14 ωµν γ µ γ ν = exp − 81 ωµν [γ µ , γ ν ] .



(6.3)
Compute the finite rotation with angle ω12 around the same axis as before. Also
compute the finite transformation Λ = exp(ω) for vectors.
c) What happens to the individual components of a spinor under this transformation?
What is the period of the transformation in the angle ω12 ? Compare it to the finite
rotation for vectors.

6.2. Completeness for gamma-matrices

An arbitrary product of gamma-matrices for D = 4 is proportional to one of the following
16 linearly independent matrices Γ a (here a is a multi-index which specifies the type of
matrix, S, P, V, A, T, along with the corresponding spacetime indices if any)
Γ S := 1, [1]
P 5
Γ := γ , [1]
Γ V,µ := γ µ , [4]
Γ A,µ :=ı̊γ 5 γ µ , [4]
T,µν T,νµ ı̊ µ
Γ = −Γ := 2
[γ , γ ν ]. [6] (6.4)
a) Argue that any product of gamma-matrices can indeed be written as a linear combi-
nation of the above.
b) Show that tr(Γ a ) = 4δ aS .
c) Show that the trace of any product of Γ ’s is given by tr(Γ a Γ b ) = 4η ab . Here, η ab = 0
if a and b are of different type, otherwise η SS = η PP := 1, η (V,µ)(V,ν) = η (A,µ)(A,ν) := η µν
and η (T,µν)(T,ρσ) := η µρ η νσ − η µσ η νρ .
d) Argue that for any pair of multi-indices a, b there is a multi-index c and a factor α ∈ C
such that Γ a Γ b = αΓ c . Make a table of which type of Γ c you expect for the different
types a, b = S, P, V, A, T. For which products will the result be proportional to Γ S ?
e) Show that the 16 matrices are linearly independent and therefore
P forma a complete
basis of 4 × 4 spinor matrices. Hint: To do this consider a sum a βa Γ = 0. What
can be said about the coefficients?
−→

6.1
6.3. Fierz identity
a) Use the linear independence of the matrices Γ a to show that
X
δγα δδβ = 1
η (Γ a )α δ (Γ b )β γ ,
4 ab
(6.5)
a,b

where ηab is the inverse of η ab defined in problem 6.2c).

Hint: Decompose an arbitrary matrix M = a ma Γ a , find the coefficients ma , sub-
P
stitute the solution back into the original relation, and expand for the elements M α β .
Alternatively, you can contract the relationship with (Γ c )γ β .

b) Show that the above relationship can be expressed in components as

δγα δδβ = 14 δδα δγβ + 41 (γ 5 )α δ (γ 5 )β γ + 41 (γ µ )α δ (γµ )β γ

− 41 (γ 5 γ µ )α δ (γ 5 γµ )β γ − 1
32
[γ µ , γ ν ]α δ [γµ , γν ]β γ . (6.6)

Hint: It makes sense to contract

P an anti-symmetric pair of indices with a conventional
1
symmetry factor of /2 , e.g. a ma Γ = . . . + 2 mT,µν Γ T,µν , in order to account for
1 a

the fact that every summand appears twice in the contraction.

c) Use the result from part a) to show the Fierz identity:

X
X αβ Y γ δ = 1
η η tr(XΓ d Y Γ c )(Γ e )α δ (Γ f )γ β .
16 ce df
(6.7)
c,d,e,f

d) Find the Fierz transformation for the spinor products

(ū1 u2 )(ū3 u4 ) and (ū1 γ µ u2 )(ū3 γµ u4 ), (6.8)

that is, express them as Cab (ū1 Γ a u4 )(ū3 Γ b u2 ) and determine the coefficients Cab .

6.4. Gordon identity

Prove the Gordon identity,
ı̊
ūβ (~q)γ µ uα (~p) = ūβ (~q) (q + p)µ − 12 [γ µ , γ ν ](q − p)ν uα (~p).
 
(6.9)
2m
Hint: You can do this using just {γ µ , γ ν } = 2η µν and Dirac’s equation.

6.2
Quantum Field Theory I Problem Set 7
ETH Zurich, 2019 HS Prof. N. Beisert

7.1. Helicity and chirality

In four dimensions we define the chirality operator as

γ 5 :=ı̊γ 0 γ 1 γ 2 γ 3 . (7.1)

a) Show that γ 5 satisfies

2
{γ 5 , γ µ } = 0, γ5 = 1. (7.2)

b) Show that the operators

PR,L = 21 (1 ± γ 5 ), (7.3)
are two orthogonal projectors to the chiral subspaces and that they satisfy the com-
pleteness relation
PL + PR = 1. (7.4)

c) Show that the Dirac Lagrangian L = ψ̄(γ µ ∂µ − m)ψ is invariant under a chiral trans-
formation U = exp(−ı̊αγ 5 ) of the fields for m = 0, and derive the associated conserved
current. Show that a non-zero mass breaks the symmetry.

Helicity is defined to be the projection of spin along the direction of motion,

~ p
Σ·~
h(~p) = . (7.5)
k~pk

~ is the spin operator which is given in the Weyl representation by

Here, Σ
 
~ = 1 ~σ 0
Σ . (7.6)
2 0 ~σ

d) Show that chirality is not conserved for a massive fermion by computing the equations
of motions for the chiral fermions ψL and ψR , with

ψL,R = PL,R ψ. (7.7)

e) Show that helicity and chirality are equivalent for a massless spinor uα (~p).

f ) Show that the Dirac equation respects helicity.

g) Argue that helicity is not Lorentz invariant for m 6= 0.

−→

7.1
7.2. Electrodynamics
Consider the Lagrange density for electrodynamics with an external source field J µ
L(Aµ ) = − 41 Fµν F µν − J µ Aµ , where Fµν = ∂µ Aν − ∂ν Aµ . (7.8)

a) Show that the Euler–Lagrange equations are the inhomogeneous Maxwell equations.
The electric and magnetic fields are defined by Ei = F0i and εijk Bk = −Fij . What
about the homogeneous Maxwell equations?
b) for fun: Show that all Maxwell equations can summarised in the spinorial equation
γ ν γ ρ γ σ ∂ν Fρσ = 2γ ν Jν . (7.9)

For the remainder of this problem, we drop the sources, J µ = 0.

c) Construct the stress-energy tensor for this theory assuming that an infinitesimal trans-
lation δaµ transforms the gauge field according to δAρ = −δaν ∂ν Aρ . Show that the
resulting stress-energy tensor T0µν is neither symmetric nor gauge invariant.
d) Now supplement the translation by a gauge transformation with gauge transformation
parameter δα = δaν Aν and compute the resulting stress-energy tensor T µν . Convince
yourself that the latter is symmetric and gauge invariant.
e) Show that the invariant stress-energy tensor T µν leads to the standard formulae for
the electromagnetic energy and momentum densities
~2 + B
E = 21 (E ~ 2 ), S~ = E×
~ B.
~ (7.10)

7.3. Polarisation vectors of a massless vector field

Each Fourier mode in the plane wave expansion of a massless vector field has the form
A(λ)
µ (~p; x) = N (~p)(λ)
µ (~p) eı̊p·x . (7.11)
(λ)
Without loss of generality the polarisation vectors µ (~p) can be chosen to form a four-
dimensional orthonormal system satisfying
(λ) (~p)·(κ) (~p) = η λκ . (7.12)

a) Show that the following choice of polarisation vectors is orthonormal

(0)
µ (~p) = nµ , (1)
µ (~p) = (0,~ (1) (~p)),
pµ + nµ (p·n)
(3)
µ (~p) = , (2)
µ (~p) = (0,~ (2) (~p)), (7.13)
|p·n|
where nµ = (1, ~0) and p~·~ (k) (~p) = 0 as well as ~ (k) (~p)·~ (l) (~p) = δ kl with k, l = 1, 2.
b) Use the polarisation vectors to verify the completeness relation
3
X
ηλλ (λ)
µ (~p) (λ)
ν (~p) = ηµν . (7.14)
λ=0

c) Show for the physical modes of the photon that

2
X pµ nν + pν nµ pµ p ν
(λ)
µ (~p) (λ)
ν (~p) = ηµν − − . (7.15)
λ=1
p·n (p·n)2

7.2
Quantum Field Theory I Problem Set 8
ETH Zurich, 2019 HS Prof. N. Beisert

8.1. The massive vector field

Consider the Lagrangian for the free massive vector field Vµ :
L = − 12 ∂ µ V ν ∂µ Vν + 12 ∂ µ V ν ∂ν Vµ − 12 m2 V µ Vµ . (8.1)

a) Derive the Euler–Lagrange equations of motion for Vµ .

b) By taking a derivative of the equation, show that Vµ is a conserved current.
c) Show that Vµ satisfies the Klein–Gordon equation.

8.2. Hamiltonian formulation

The Hamiltonian formulation of the massive vector is somewhat tedious due to the pres-
ence of constraints. Let us consider the phase space.
a) Derive the momenta Πµ conjugate to the fields Vµ . Considering the space and time
components separately, what do you notice?
Your observation is related to constraints: the time component V0 of the vector field is
completely determined by Vk and Πk (without making reference to time derivatives).
b) Use the equations derived in problem 8.1 to show that
V0 = −m−2 ∂k Πk , V̇0 = ∂k Vk . (8.2)

c) Substitute this solution for V0 and V̇0 into the Lagrangian and perform a Legendre
transformation to obtain the Hamiltonian. Show that
Z
H = d~x 3 21 Πk Πk + 12 m−2 ∂k Πk ∂l Πl + 21 ∂k Vl ∂k Vl − 12 ∂l Vk ∂k Vl + 12 m2 Vk Vk .


(8.3)

d) Derive the Hamiltonian equations of motion for Vk and Πk , and compare them to the
results of problem 8.1.

8.3. Commutators
The unequal-time commutators ∆Vµν (x − y) = ı̊[Vµ (x), Vν (y)] for the massive vector field
read
−2
∆V


µν (x) = ηµν − m ∂ µ ∂ ν ∆(x), (8.4)
where ∆(x) is the corresponding function for the scalar field.
a) Show that these obey the equations derived in problem 8.1.
b) Show explicitly that they respect the constraint equations in (8.2), i.e.
 2   
m V0 (x) + ∂k Πk (x), Vν (y) = V̇0 (x) − ∂k Vk (x), Vν (y) = 0. (8.5)

c) Confirm that the equal-time commutators take the canonical form

Vk (~x), Πl (~y ) =ı̊δkl δ 3 (~x − ~y ).
     
Vk (~x), Vl (~y ) = Πk (~x), Πl (~y ) = 0, (8.6)

−→

8.1
8.4. Photon propagator
The Lagrangian density for the electromagnetic potential with gauge-fixing term reads

2
L = − 41 Fµν (x)F µν (x) − 21 ξ −1 ∂µ Aµ (x) . (8.7)

Show that the photon propagator (Green function) with arbitrary gauge parameter ξ is
given by
dp4 eı̊p(x−y)
Z  
V pµ pν
Gµν (x − y) = ηµν − (1 − ξ) 2 . (8.8)
(2π)4 p2 p
Note: we will not care about on-shell contributions to the propagator.

8.5. Interacting scalar field theory and scale invariance

Consider the real scalar field with interactions

L = − 21 ∂ µ φ ∂µ φ − 21 m2 φ2 − 61 µφ3 − 1
24
λφ4 . (8.9)
R
a) The action S = dx4 L is a dimensionless quantity. In natural units time and length
have mass dimension dx ∼ m−1 . What are the mass dimensions of the field φ and
the interaction couplings µ and λ?

b) The coordinate scaling transformation x0 = Λx with some Λ ∈ R+ can be extended

to the scalar field by φ0 (x0 ) = Λ−∆ φ(x) for some ∆ ∈ R. For which values of the
parameters {m, µ, λ, ∆} is the action scale invariant?

From now on, we consider the scale invariant theory.

c) Let φ(x) be a solution to the equation of motion of this interacting theory, show that
Λ∆ φ(Λx) is also a solution.

d) Derive the scale (or dilatation) current using Noether’s procedure for invariance under
the scaling transformation defined in part b). Show explicitly that it is conserved.

The remainder of the problem deals with a suitable form for the stress-energy tensor in a
scale invariant (or conformal) model.

e) Compute the stress-energy tensor and show that it has a non-vanishing trace. Use
the equations of motion to write the trace in the form Tµµ = ∂ 2 K.

f ) Show that the scale current can be expressed as S µ = xν T µν − ∂ µ K. Show that it is

conserved using the properties of T µν .

g) The stress-energy tensor can be improved by adding an extra term

T̃ µν = T µν + c ∂ µ ∂ ν K − η µν ∂ 2 K .


(8.10)

For which value of c is this tensor symmetric, conserved and traceless?

h) Show that the modified scale current S̃ µ = xν T̃ µν is conserved and compare it to the
original current S µ .

8.2
Quantum Field Theory I Problem Set 9
ETH Zurich, 2019 HS Prof. N. Beisert

9.1. Feynman propagators

The Feynman propagator for the real scalar field is defined as
(

ı̊h0|φ(x)φ(y)|0i for x0 > y 0 ,
GF (y − x) =ı̊h0| T φ(y)φ(x) |0i = (9.1)
ı̊h0|φ(y)φ(x)|0i for y 0 > x0 ,

where the time-ordering symbol T orders the fields within the product with decreasing
times from left to right.

a) Show that the propagator satisfies the defining relation

(−∂ 2 + m2 )GF (x) = δ 4 (x). (9.2)

Hint: write the propagator in terms of correlators and commutators.

The Feynman propagator for a Dirac field is defined by

(
ı̊h0|ψ a (x)ψ̄b (y)|0i for x0 > y 0 ,
GD a a


F b (y − x) =ı̊h0| T ψ (x)ψ̄b (y) |0i = (9.3)
−ı̊h0|ψ̄b (y)ψ a (x)|0i for y 0 > x0 .

b) Write this propagator in terms of the Feynman propagator GF for the real scalar field.

c) Show that the Feynman propagator for the Dirac field is a Green function of the Dirac
equation:
(−γ·∂ + m)a b GD b a 4
F c (x) = δc δ (x). (9.4)

d) Show that the Feynman propagator GF (x) for scalars can be expressed as the following
integral
dp4 eı̊p·x
Z
GF (x) = 4 2 2
. (9.5)
R4 (2π) p + m −ı̊

Derive a similar expression for the Feynman propagator for Dirac fields.

9.2. Wick’s theorem

Wick’s theorem relates the time-ordered product of fields φ0 (x) to the normal-ordered
product plus all possible contractions



T φ0 (x1 ) . . . φ0 (xm ) = N φ0 (x1 ) . . . φ0 (xm ) + all contractions . (9.6)

Prove this theorem by induction. What changes in the case of fermionic operators?

−→

9.1
9.3. Interaction picture
The field operator φ0 (x) in the interaction picture is related to the field operator φ(x) in
the Heisenberg picture by

φ(t, ~x) = U (t)−1 φ0 (t, ~x) U (t). (9.7)

We assume the fields to coincide at the reference time t0 , φ(t0 ) = φ0 (t0 ). The transfor-
mation operator is then given by



U (t) = exp ı̊(t − t0 )H0 exp −ı̊(t − t0 )H . (9.8)

a) Show that U (t) satisfies the differential equation

∂
ı̊ U (t) = Hint (t) U (t), (9.9)
∂t
with the initial condition U (t0 ) = 1. Determine the interaction Hamiltonian Hint (t).

b) Show that the unique solution to this equation with the same initial condition and
t > t0 can be written as
 Z t 
0 0
U (t) = T exp −ı̊ dt Hint (t ) . (9.10)
t0

c) Show that the operator is unitary U (t)† = U (t)−1 .

d) We define the time evolution operator U (t2 , t1 ) for t2 ≥ t1 as the time-ordered expo-
nential  Z t2 
U (t2 , t1 ) := T exp −ı̊ dt Hint (t) , (9.11)
t1

while for t1 ≥ t2 it is defined by U (t2 , t1 ) := U (t1 , t2 )−1 .

Show that it satisfies the composition rule (for all permutations of the times tk )

U (t3 , t2 ) U (t2 , t1 ) = U (t3 , t1 ). (9.12)

Deduce that it is related to U (t) by U (t2 , t1 ) = U (t2 )U (t1 )−1 .

9.2
Quantum Field Theory I Problem Set 10
ETH Zurich, 2019 HS Prof. N. Beisert

10.1. Four-point interaction in scalar QED

Consider a model of electrodynamics with a complex scalar field φ of charge q and mass
m which couples to the vector potential Aµ . This model in the Rξ -gauge is described by
the Lagrangian density
LSQED = −(Dµ φ)∗ Dµ φ − m2 φ∗ φ − 14 F µν Fµν − 12 ξ −1 (∂ µ Aµ )2 , (10.1)
where we set ξ = 1 for convenience. The covariant derivative Dµ and the electromagnetic
field strength tensor Fµν are given by
Dµ = ∂µ −ı̊qAµ , Fµν = ∂µ Aν − ∂ν Aµ . (10.2)
In this exercise we want to compute perturbatively for small q the first non-trivial contri-
bution to the time-ordered 4-point correlation function
h0| T φ(x1 )φ(x2 )φ† (x3 )φ† (x4 ) |0iint
 
h  R i
T
h0| T φ(x1 )φ(x2 )φ† (x3 )φ† (x4 ) exp ı̊ −T dy 4 Lint (y) |0i
= lim h  R
T
i . (10.3)
T →∞(1−ı̊)
h0| T exp ı̊ −T dy 4 Lint (y) |0i

a) Split the Lagrangian L of the model into a free part L0 and an interaction interaction
Lagrangian Lint .
b) Give a short explanation as to why we need to expand the 4-point correlation function
(10.3) to second order in q to obtain the leading non-trivial contribution.
c) Expand the denominator of (10.3) to order q 2 , then use Wick’s theorem to decompose
the time-ordered product into a sum of complete contractions between pairs of fields.
Find a pictorial representation for the different contributions.
d) Expand the numerator of the time-ordered 4-point correlation function (10.3) in the
same way. It may be useful to draw diagrams to simplify calculations.
e) Now combine the leading non-trivial contributions to (10.3) and group them according
to the graph topology. How can you interpret the various contributions?
f ) Consider the connected contributions to the 4-point correlation function. Insert the
Feynman propagators of the scalar and photon fields
dp4 eı̊p·(x−y)
Z
GF (x − y) = ,
(2π)4 p2 + m2 −ı̊
dp4 η µν eı̊p·(x−y)
Z
µν
GF (x − y) = , (10.4)
(2π)4 p2 −ı̊
and perform all elementary integrations. How do you interpret the individual terms
in the result?
g) How do you interpret the limit T → ∞ in (10.3)?
h) How will your result in part f) change if you use a different gauge? E.g. use the
Rξ -gauge with ξ 6= 1.
Hint: The gauge only affects the photon propagator Gµν
F .

10.1
Quantum Field Theory I Problem Set 11
ETH Zurich, 2019 HS Prof. N. Beisert

11.1. Møller scattering

a) Calculate the O(q 2 ) leading connected contribution to the scattering matrix element
M for Møller scattering,

e− (p1 , α) + e− (p2 , β) −→ e− (q1 , γ) + e− (q2 , δ), (11.1)

through direct evaluation in position space. In this process the spins of the electrons
are denoted by α, . . . , δ; pi and qi denote the momenta of the particles.

b) Repeat the calculation in part a) using the Feynman rules for QED in momentum
space.

11.2. Compton scattering

a) Calculate the O(q 2 ) leading connected contribution to the scattering matrix element
M for Compton scattering,

e− (p1 , α) + γ(p2 , σ) → e− (q1 , β) + γ(q2 , ρ), (11.2)

through direct evaluation in position space. In this process the spins of the electrons
are denoted by α and β, while ρ and σ denote the polarisations of the photons; pi
and qi denote the momenta of the particles. What changes if we replace the electrons
by positrons?

b) Repeat the calculation in part a) using the Feynman rules for QED in momentum
space.

−→

11.1
11.3. Kinematics in 2 → 2 scattering
Consider a 2 → 2 particle scattering process with the kinematics p1 + p2 → q1 + q2 and
masses m1 , . . . , m4 .

a) Show that in the centre-of-mass frame, the energies e(~pi ), e(~qi ) and the norms of
momenta k~p i k, k~q i k of the incoming
√ and the outgoing particles are entirely fixed by
the total centre-of-mass energy s and the particle masses mi .

b) Show that the scattering angle θ between p~1 and ~q1 is given by
!
s(t − u) + (m21 − m22 )(m23 − m24 )
θ = arccos p p , (11.3)
λ(s, m21 , m22 ) λ(s, m23 , m24 )

with the Mandelstam variables given by

s = −(p1 + p2 )2 , t = −(p1 − q1 )2 , u = −(p1 − q2 )2 , (11.4)

and the Källén function defined as

λ(x, y, z) = x2 + y 2 + z 2 − 2xy − 2xz − 2yz. (11.5)

c) Show that s + t + u = m21 + m22 + m23 + m24 .

d) Determine tmin and tmax from the condition |cos θ| ≤ 1, and study the behaviour of
tmin and tmax in the limit s  m2i .

e) Show that the general expression for the differential scattering cross section for 2 → 2
particle scattering

(2π)4 δ 4 (p1 + p2 − q1 − q2 ) d~q 31 d~q 32

d6 σ =
3 2e (~ 3 2e (~
|M |2 (11.6)
4 e1 (~p1 )~p2 − e2 (~p2 )~p1
(2π) q
3 1 ) (2π) q
4 2 )

reduces in the centre-of-mass frame to the following expression upon elimination of

the momenta fixed by momentum conservation

d2 σ k~q1 k |M |2
= . (11.7)
d2 Ω k~p1 k 64π 2 s

Furthermore, show that for equal masses mi = m, it reduces further to

d2 σ |M |2
= . (11.8)
d2 Ω 64π 2 s

11.2
Quantum Field Theory I Problem Set 12
ETH Zurich, 2019 HS Prof. N. Beisert

12.1. Optical theorem

For the non-trivial part of the scattering matrix we have that

hf |T |ii = (2π)4 δ 4 (Pi − Pf )Mf i , (12.1)

where S = 1 +ı̊T . Let us derive the optical theorem.

a) Use unitarity of the S matrix to show that T † T = −ı̊(T − T † ).

b) From this derive the optical theorem

XZ
∗ ∗
Mf i − Mif =ı̊ dΠX (2π)4 δ 4 (Pf − PX )MXi MXf . (12.2)
X
P R
Hint: Use the completeness relation 1 = X dΠX |XihX| with

Y d~kj3
dΠX = . (12.3)
j∈X (2π)3 2e(~kj )

c) Specialise to the case |ii = |f i = |Ai where A is a two-particle state.

P Use the optical
theorem to derive an expression for the total cross section σtot = X σXA in terms
of the imaginary part of the forward scattering amplitude Im MAA .

−→

12.1
12.2. Muon pair production
Follow the steps below to calculate the total cross section for muon pair production
e− (p1 )e+ (p2 ) → µ− (q1 )µ+ (q2 ). This process is described by the QED Lagrangian with
two fermion flavours f = e, µ with different masses but the same charge. The Lagrangian
therefore reads X
L= ψ̄f (γ µ Dµ − mf )ψf − 14 Fµν F µν (12.4)
f =e,µ

with Dµ = ∂µ −ı̊qAµ .

a) Draw all the connected diagrams that contribute to this process at the lowest non-
trivial order, and use the Feynman rules for QED in momentum space to obtain the
scattering amplitude M .

b) Compute |M |2 . Assuming that the particle spins are not measured, sum over the
spins of the outgoing particle, and average over those of the incoming ones. This
should help you bring your expression for |M |2 into a much simpler form.
Hint: You might find the completeness relations for spinors useful.

c) The differential cross section in the centre-of-mass frame is given by

|M |2 d~q1 3 d~q2 3
d6 σ = √ (2π)4 δ 4 (p1 + p2 − q1 − q2 ). (12.5)
4k~p1 k s (2π)3 2eµ (~q1 ) (2π)3 2eµ (~q2 )

Use the result for |M |2 obtained

R 6 in part b), and integrate over ~q1 and ~q2 to obtain the
total cross section σ = d σ.

12.3. Polarised muon-electron scattering

Let us study the process e− (p1 )µ− (p2 ) → e− (q1 )µ− (q2 ), i.e. the scattering of negative
muons on electrons. To simplify the calculation we assume both the electron and the
muon to be massless. This is a good approximation for sufficiently high beam energies.

a) As a first step, calculate |M |2 for unpolarised scattering (averaging over incoming

particle spins). We assume that the spins of the outgoing particles are not measured.
How is this result related to the one for muon pair production of problem 12.2?

b) We now assume the incoming muon beam to be fully polarised along its direction of
motion. How does the result change compared to the completely unpolarised case?
Can you predict or explain the outcome using parity invariance of QED?
Hint: Recall that the helicity and chiral eigenstates coincide in the massless case. You
can therefore make use of the chiral projectors PR,L = 21 (1 ± γ 5 ) for hµ = ± 1/2 .

c) Consider the cases of electrons and muons having equal and opposite polarisations
and calculate the additional contributions. How do the results for |M |2 depend on
the angle θ in the centre-of-mass frame? Can you identify special angles?

12.2
Quantum Field Theory I Problem Set 13
ETH Zurich, 2019 HS Prof. N. Beisert

13.1. Feynman and Schwinger parameters

In order to evaluate loop diagrams in momentum space, it is convenient to combine the
propagator factors in the denominator. This can be achieved by introducing auxiliary
integrals over so-called Schwinger parameters and Feynman parameters.

a) Prove the Schwinger parametrisation:

Z ∞
1 1
= dα αν−1 e−αA . (13.1)
Aν Γ(ν) 0

Argue how it can be used to combine denominators of Feynman diagrams such that
the momentum integrals can be performed efficiently.

b) The basic version of the Feynman parameter integral is

Z 1
1 dx
= 2
. (13.2)
AB 0 [xA + (1 − x)B]

Prove the generalisation to n propagator factors Ai raised to an arbitrary power νi

by recursion:
Pn Z 1 Yn
! n
! Qn νi −1
1 Γ( i=1 νi ) X x
Qn νi = Qn dxi δ 1 − xi Pn i=1 i Pn ν . (13.3)
i=1 Ai i=1 Γ(νi ) 0 i=1 i=1 [ xi Ai ] i=1 i
i=1
Pn−1
Hint: Consider the replacements xi = yi /(1 + yn ) and yn = z i=1 yi Ai /An .

13.2. Volume of spheres

The integrands of D-dimensional euclidean integrals for Feynman diagrams can be brought
to a spherically symmetric form F (`E ) = F (k`E k) using the results of problem 13.1. The
angular part of the integral yields the volume of the (D − 1)-dimensional sphere S D−1 .
Use the Gaußian integral Z ∞
√
dx exp(−x2 ) = π (13.4)
−∞

to derive the volume of the (D − 1)-sphere:

2π D/2
Vol(S D−1 ) = . (13.5)
Γ(D/2)

−→

13.1
13.3. Basic loop integral
When computing one-loop Feynman diagrams, one frequently encounters the following
basic loop integral in D-dimensional Minkowski space:
d`D
Z
2 1
In (µ ) := . (13.6)
(2π) (` + µ2 )n
D 2

Here, n is a positive integer and µ = µ(pi ) is an effective mass that is a scalar function of
the external momenta pi but independent of the loop momentum `.
a) Perform a Wick rotation to euclidean space and transform the integral to spherical
coordinates.
b) Find a criterion for divergence of the above integral in the UV region (k`E k → ∞).
For which values of n is the integral divergent in D = 4 dimensions?
c) Perform the radial integral to show that:

2 ı̊µD−2n Γ(n − D/2)

In (µ ) = . (13.7)
(4π)D/2 Γ(n)
d) Demonstrate that the derivative (d/dµ2 )I2 (µ2 ) in D = 4 dimensions is finite. Use
this feature to show that the subtracted loop integral I2 can be written as
ı̊ µ2
I2 (µ2 ) − I2 (µ20 ) = − log . (13.8)
16π 2 µ20

13.4. Loop integral numerators

More elaborate loop integrals can carry some loop momenta `µ in the numerator, e.g.:
d`D `µ d`D `µ `ν
Z Z
µ 2 µν 2
In (µ ) := , In (µ ) := . (13.9)
(2π)D (`2 + µ2 )n (2π)D (`2 + µ2 )n
Relate these integrals to the scalar integral of problem 13.3 to show that:
η µν ı̊µD−2n+2 Γ(n − 1 − D/2)
Inµ (µ2 ) = 0, Inµν (µ2 ) = . (13.10)
2 (4π)D/2 Γ(n)

13.5. Dimensional regularisation

The UV behaviour of a loop integral depends on its dimensionality D.
a) For which values of n is the integral In in problem 13.3 UV divergent for D = 3
dimensions? In how many dimensions is the integral I1 UV finite?
In the dimensional regularisation scheme, the dimension of the momentum loop integra-
tion is shifted from D = 4 to a suitable number to make the integral finite. Of course, the
physical result eventually requires the correct dimensionality for the loop integrals. The
crucial idea is to generalise D to be a complex number rather than an integer, and assume
the results of problems 13.2, 13.3 and 13.4 to remain valid as functions of D. We can
then analytically continue the integrals back to their physical dimension in the complex
plane and thus effectively use the dimensionality D of spacetime as a regulator.
b) Argue why we can use D as a regulator for the class of loop integrals in problems 13.3
and 13.4. Can you say how divergences manifest in this scheme?

13.2
Quantum Field Theory I Problem Set 14
ETH Zurich, 2019 HS Prof. N. Beisert

14.1. Electron self energy structure

In QED, the electron two-point function hψ(p)ψ̄(q)i =: −ı̊(2π)D δ D (p + q)K(p) receives
contributions from self energy diagrams.

a) Draw the Feynman diagrams corresponding to the one- and two-loop contributions.
Which of these diagrams are one-particle irreducible?

b) Write down the one-loop contribution K (2) (p) to the electron two-point function using
the QED Feynman rules in momentum space and argue why this self energy loop
integral is divergent.

c) Using the Feynman parametrisation, show that the self energy loop integral in the
first order contribution K (2) (p) = −K (0) (p)Σ(p)K (0) (p) to the two-point function is
given by Z 1
2
dx −(D − 2)xp·γ + Dı̊m I2 µ(x)2 ,



Σ(p) = −q (14.1)
0

where µ(x) := x(1 − x)p + (1 − x)m2 and I2 is the scalar loop integral defined in
2 2

problem 13.3.

d) Explain why one can make the general ansatz

Σ(p) = p·γΣV + mΣS , (14.2)

where ΣV,S are scalar functions. Write down integral expressions for them.

14.2. A one-loop correction to scattering in QED

The aim of this exercise is to gain an insight into the calculation of loop corrections to
scattering amplitudes. To this end consider the one-loop corrections to e− e− → e− e−
scattering in QED.

a) Draw all amputated and connected graphs that would contribute to this process. You
should find ten different diagrams.
(2)
b) How does the field strength renormalisation factor for the spinors, Zψ = 1 + Zψ + . . .,
contribute at this perturbative order? How does the field strength renormalisation of
the photon ZA contribute to the process? Can you sketch suitable Feynman graphs?

Now focus on the following diagram:

p0
p0 − k q0

k (14.3)
p−k q
p

−→

14.1
c) Write the scattering matrix element corresponding to the amputated Feynman graph,
and bring it to the following form
dk D Aµ
Z
0 1
4
ı̊M = −ı̊q v̄(~q)γµ v(~q ) 0 2 D
v̄(~
p) v(~p 0 ). (14.4)
(p − p ) −ı̊ (2π) B

d) Use a suitable Feynman parametrisation to rewrite the denominator B as

Z 1 Z 1 Z 1
1 δ(1 − x − y − z)
=2 dx dy dz , (14.5)
B 0 0 0 C3
where
C = k 2 − 2k·(xp + yp0 ) −ı̊. (14.6)
Complete the square and show that C can be written as

C = k 02 + (1 − z)2 m2 + xy(p − p0 )2 −ı̊, k 0 = k − xp − yp0 . (14.7)

e) Show that the numerator Aµ can be brought to the form

Aµ = −ı̊ k 02 + 2(1 − 4z + z 2 )m2 − 2(z + xy)(p − p0 )2 γ µ

 

+ z(1 − z)m[γ µ , γ ν ](p − p0 )ν . (14.8)

To do so, use:
• the anti-commutation relations for gamma-matrices

(p·γ)γ µ = 2pµ − γ µ (γ·p), (14.9)

• the Dirac equation,

v̄(~p)(γ·p) =ı̊mv̄(~p), (p0 ·γ)v(~p 0 ) =ı̊mv(~p 0 ), (14.10)

• the symmetry of the integration over k 0 , which allows the following tensorial re-
placements in the numerator
1 µν 02
k 0µ → 0, k 0µ k 0ν → η k , (14.11)
D
• the symmetry of the integral under the interchange x ↔ y,
• the Gordon identity
ı̊
v̄(~p)γ µ v(~p 0 ) = v̄(~p) −(p + p0 )µ + 21 [γ µ , γ ν ](p − p0 )ν v(~p 0 ).
 
(14.12)
2m
For the remainder of this problem, you may assume that the virtuality of the photon is
small, |(p − p0 )2 |  m2 .
f ) Using the results obtained in part d) and e), integrate over the loop momentum k 0 .
Note: Split off a divergent contribution, and cut off the integral as discussed in the
lecture. Can you interpret the residual dependence on the cutoff?
g) Integrate over the Feynman parameters x, y and z.
Note: Cut off the integral if needed. Can you interpret the residual dependence on
the cutoff?

14.2