Sei sulla pagina 1di 8

Part III Quantum Field Theory

September 20, 2019

Contents
1 Symmetries and Noether’s theorem 1

2 Classical field theory and an introduction 2


2.1 Promoting discrete operators to a field . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.2 Deriving an expression for the Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . 3
2.3 Normalisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.4 Causality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4.1 Feynman propagation and time ordering . . . . . . . . . . . . . . . . . . . . . 5

3 Interacting fields 6
3.1 The S-matrix and in-out states. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

4 The Dirac Equation 7

5 Example sheet 1 7
5.1 Question 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
5.2 Question 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
5.3 Question 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

1 Symmetries and Noether’s theorem

Our concept of consered energy and momentum arise from translational symmetry in our La-
grangian. Consider a translation transformation in our spacetime coordinates

xν → xν + ν

We expcet that our symmetry here will give rise to a higher rank tensor because we have 4 linearly
independent directions arising from ν . Taylor expanding out, we have

φa (x) → φa (x) − ν ∂ν φa (x)

And this induces a similar transformation on our Lagrangian

L → L − ν ∂ν L

1
This is a symmetry because our Lagrangian changes by a total derivative.

δL = ∂µ (µ L) =⇒ F µ = µ L

By Noether’s theorem, our preserved current is therefore


∂L
j µ = ν ∂ν φa − µ L
∂µ φa

Since we have a degree of freedom here from , we can factor this out to write our conserved current
as
j µ = ν (∂ν φ)
In essence we’ve given rise to an entirely new object
∂L
(j µ )n u = ∂ν φa − δ µν
∂m uφa
we call this the energy-momentum tensor.

2 Classical field theory and an introduction

2.1 Promoting discrete operators to a field

We know from standard quantum mechanics how to find the spectrum of the quantum harmonic
oscillator H for unit mass, given by
1 1
H = p2 + mω 2 q 2 .
2 2
p, q represent our momentum and position operators respectively. We have our canonical commu-
tation relations (which we can derive by examining infinitesimal transformations)

[pa , pb ] = [qa , qb ] = 0, [qa , pb ] = iδa b .

We can intuitively promote our scalar and momentum fields in the same way,

[φa (x), π b (y)] = iδ 3 (x − y)δa b


[φa (x), φb (y)] = [πa (x), π b (y)] = 0.

We have our Lagrangian density L = 21 (∂u)2 − 12 m2 φ2 . Recall that classically we can derive our
Hamiltonian from the Lagrangian with
X
H= pi q̇i − L
i

where pi , qi represent momentum and position operators respectively and L is our full Lagrangian.
Recalling that in QFT, we promote pi to Π(x), and qi to φ(x). We can make the sum continuous
so that Z   Z  
3 3 1 2 1 2 2
H= d x Πφ̇ − L = d x Πφ̇ − (∂φ) + m φ . .
2 2

2
We can develop this even further. Our momentum Π is indeed the conjugate momentum which is
derived from Hamilton’s equations, so that
∂L
Π= =⇒ Π = φ̇
∂ φ̇
so in this case we can write H as
Z  
1 2 1 1
H= d3 x φ̇ − (∇φ)2 + m2 .
2 2 2

Recall that in quantum mechanics, we have that position and momentum obey the canonical com-
mutation relations
[qi , pj ] = iδi,j .
Similarly, we have that

[Π(y), φ(x)] = δ 3 (x − y), 0 = [φ(x), φ(y)] = [Π(y), Π(x)].

Recall that we can reduce our description of the discrete Hamiltonian via creation and annihilation
operators: r ! r !
r r
ω 1 † ω 1
a= x+i p , a = x−i p .
2 2ω 2 2ω
We can write our Hamiltonian as  
† 1
H =ω a a+ .
2
Rearranging, our expressions for the operators x, p are
r 
1  †
 ω 
x= √ a + a , p = −i a − a† .
2ω 2

We can promote these definitions to our scalar and conjugate momentum fields:

d3 p
Z
1  ix·p † −ix·p

φ(x) = a p e + ap e
(2π)3 2wp
p

d3 p
Z r
wp  
π(x) = ap eix·p − a†p e−ix·p
(2π)3 2
This gives rise to commutation relations, which we will show below.
We can also compute the analog expansions for the raising and lowering operators.
Z r r
ωx 1
ap = d x 3
φ(x)e ip·x
+i Π(x)e−ip·x
2 2ωx

With this formulation, it’s easy to see that a−p = a†p .

3
2.2 Deriving an expression for the Hamiltonian

Our Hamiltonian is given by


Z  
1 2 1 1
H= 3
d x Π + (∇φ)2 + m2 φ2
2 2 2

We go term by term. We have


Z Z 3 3 3
3 2 d xd pd q 1 
ip·x † −ix·p

iq·x † −iq·x

d x(∇φ) = − √ p · q a p e − ap e aq e − a q e
(2π)3 2 ωp ωq

which is
Z Z 3 3 3
3 2 d xd pd q 1 
i(q+q)·x † † −i(p+q)·x † i(q−p)·x † i(p−q)·x

d x(∇φ) = − √ p·q ap aq e + ap a q e − ap a q e − q a p e
(2π)3 2 ωp ωq

But recall that we can integrate over x first, and make use of the identity
Z
d3 xeia·x = (2π)3 δ(a)

This means that the above expression is equal to

d3 p p·p
Z 
† † † †
= (−1) a a
p −p − a a
p p − a a
p p + a a
p −p
(2π)3 2ωp
d3 p 2  †
Z 

= p a a
p p + a a
p p
(2π)3

It’s easy to show that Z


d3 Π2 = 0

Hence our final expression for our full Hamiltonian is


Z  
3 ωp † 1 3
H= d p ap ap + (2π) δ(0)
(2π)3 2

2.3 Normalisation

We have that our states for particles in Fock space are given by

|pi = Ca†p |0i

If we set C = 1, we find that indeed


hp|qi
is not actually Lorentz invariant. However, we have that

|pi = 2Ep a†p |0i


p

actually gives rise to a Lorentz invariant measure. why is this a Lorentz invariant measure?

4
To begin, we derive an important propertty of delta functions. Can we write the expression

δ(f (x) − f (y))

in terms of δ(x − y)? To solve this, we integrate this expression and usesubstitution in the integral
to wrangle it into a form above.
Z Z
dx
dxδ(f (x) − f (y)) = delta(g − f (y)) dg
dg
where we’ve used the substitution g = f (x). Now, the expression in the delta function is zero
anywhere where f (x) = f (y). So, assuming we have a finite amount of zeros we can split this
integral up in to the separate pieces where this holds to find that
X δ(x∗ )
i
δ(f (x) − f (y)) = 0 (x∗ )

f i
xi

where x∗i satisfies f (x∗i ) = f (y). Now in the case where f is bijective, then this implies that x = y
is the only zero and thus we have that
1
δ(f (x) − f (y)) = δ(x − y)
f 0 (y)

With this in mind, we can try to find a ’Lorentz invariant’ dot product. We simplify t his task by
working in 2 spacetime dimensions. Our Lorentz boost is given by
 0   
E 1 −β E
0 =γ
p3 −β 1 p3
1
where γ = √
1−β 2

We can see this for two reasons. Well, the first way ould be to make use of the identity

δ 3 (f (x) − f (x0 )) =

2.4 Causality

2.4.1 Feynman propagation and time ordering

We define an important object in quantum field theory, the Feynman propagator.


Definition. The Feynman propagator Df (x − y) is defined as

Df (x − y) = h0| T φ(x)φ(y) |0i ,

where we’ve defined T , the time ordered product, as


(
φ(x)φ(y) x0 > y 0
T φ(x)φ(y) =
φ(y)φ(x) y 0 > x0

We can write this expression in terms of a contour integral.

5
3 Interacting fields

3.1 The S-matrix and in-out states.

6
4 The Dirac Equation

5 Example sheet 1

Solved exercises and notes from example sheet 1.

5.1 Question 1

Show that if φ(x) is a solution to the Klein-Gordon equation, then so is φ(Λ−1 x).
We have that φ(x) satisfies
∂µ ∂ µ φ − m2 φ = 0.
Since φ(x) → φ(Λ−1 x), we have to show that

∂µ ∂ µ φ(y) − m2 φ(y), y = Λ−1 x

is equal to zero, which is equivalent to showing that the differential term remains unchanged under
the Lorentz transformation. Due to the chain rule, we have

∂µ ∂ µ φ(Λ−1 x) = ∂µ (Λ−1 )µν ∂ ν φ(Λ−1 )




= (Λ−1 )ρµ (Λ−1 )µν ∂ρ ∂ ν φ(Λ−1 x)


= (Λ−1 )ρτ gτ µ (Λ−1 )µν ∂ρ ∂ ν φ(Λ−1 x).

We make use of the identity


Λµν gνρ Λρσ = g µρ .
This identity holds equally for Λ−1 . Thus the above is

g ρν ∂ρ ∂ ν φ(Λ−1 x) = ∂µ ∂ µ φ(Λ−1 x).

The second term in the Klein-Gordon equation stays the same.

5.2 Question 2

We’re given the Lagrangian


λ ∗ 2
L = ∂µ ψ ∗ ∂ µ ψ − m2 ψ ∗ ψ − (ψ ψ) .
2
We want to show that it’s invariant under the infinitesimal transformations

δψ = iαψ, δψ ∗ = −iαψ.

We immediately notice that this is the infinitesimal representation of the transformation ψ → eiα ψ,
and that it seems to me to be enough grounds to show that this thing is invariant (since the phases
cancel out). However, if we substitute the transformation naively we have
λ
L = (1 + α2 )∂µ ψ ∗ ∂ µ ψ − m2 (1 + α2 )ψ ∗ ψ − (1 + α2 )2 (ψ ∗ ψ)2 .
2

7
The Euler-Lagrange equations for the variable ψ are
 
∂L ∂L
= ∂µ
∂ψ ∂(∂µ ψ)

and similarly for ψ ∗ .


For this Lagrangian, varying with respect to ψ yields

∂µ ∂ µ ψ ∗ = −m2 ψ ∗ − λψ ∗2 ψ.

The equation that we have when varying with respect to ψ ∗ is the same as above, just with the
variables switched around:
∂µ ∂ µ ψ = −m2 ψ − λψ ∗ ψ 2 .
We calculate the Noether current and verify it explicitly. A Noether current is a conserved quantity
under differentiation. Assuming that
L = ∂µ F µ .
(where in this case we can without loss of generality set F µ = 0), we can write down our current
as
∂L
j µ = F µ − δψ .
∂ (∂µ ψ)
However, since our Lagrangian is a function of both ψ and ψ∗, we treat these as separate variables
and modify the variation as
∂L ∂L
j µ = F µ − δψ − δψ ∗ .
∂ (∂µ ψ) ∂ (∂µ ψ ∗ )

Thus our Noether current is


j µ = iψ∂ µ ψ ∗ − iψ ∗ ∂ µ ψ.

ηρα = ηρα + α δ να ω νρ ηνµ + δ µρ ω να ηµν




= ηρα + α ηµα ω µρ + ηρν ω µα




= ηρα + α (ωρα + ωαρ )

Thus the tensor ω is antisymmetric here.

5.3 Question 3

Potrebbero piacerti anche