Quantum Field Theory 1

Contents

1 Lecture 1 2

1
1 Lecture 1

Quantum field theory

– a technique for dealing with a quantum system of many particles

– necessary when the number of particles is not conserved, e.g. in

• in a system of photons

• in high energy collision of two electrons which can produce electron positron
pairs besides two electrons

– also a convenient tool when the number of particles is large but conserved

We begin with a familiar system where the number of particles is conserved.

First consider the simplest case: A single particle in 3 dimensions moving under
the influence of a potential V (~r).

Schrodinger equation:
∂ψ b
i~ = hψ
∂t
~2 ~ 2
hψ = −
b ∇ ψ + U (~r) ψ
2m
h: Hamiltonian
b

One way to find the general solution is to first find the eigenstates of b
h.

h un (~r) = en un (~r)
b

en : eigenvalues, un : eigenfunctions

2
{un }: complete basis of states, normalized as
Z
d3 r u∗n (~r) um (~r) = δmn

General solution:
X
ψ(t) = an (t)un (~r), an (t) = an (0) e−ien t/~
n

Now consider a system of N particles of same mass, each moving under the
same potential, and with no mutual interaction

~ri : position of the i’th particle

~ i : the gradient operators with respect to the i’th particle coordinate

∇

The wave function ψ(~r1 , · · · , ~rN ) satisfies Schrodinger equation:

N
∂ψ X
i~ =H
b N ψ, H
bN = hi
b
∂t i=1

Choose basis of states:

Wn1 ,···,nN (~r1 , · · · , ~rN ) = un1 (~r1 ) · · · unN (~rN )

General solution:
X
ψ(t; ~r1 , · · · , ~rN ) = an1 ,···,nN (t) Wn1 ,···,nN (~r1 , · · · , ~rN )
n1 ,···,nN

an1 ,···,nN (t) = e−i(en1 +···+enN ) t/~ an1 ,···,nN (0)

Note: If each un (~r) is normalized to 1, then so is Wn1 ,···,nN

3
Now consider the case when the particles are identical and bosonic.

ψ must be symmetric under ~ri ↔ ~rj for every pair i, j

We can still expand ψ in the basis {Wn1 ,···,nN } but it is redundant since only
symmetric functions need to be expanded

For this reason it will be natural to choose the basis states to be also symmetric
under the exchange of any pair of particles.

Define new basis:

1 X
un1 ,···,nN (~r1 , · · · , ~rN ) ≡ √ un1 (~r1 ) · · · unN (~rN )
N!
Permutations of ~r1 , · · · ~rN
Note: This basis is also automatically symmetric under the exchange of the
ni ’s

e.g. u1,2 and u2,1 should not be counted as separate basis states

For this reason we can label the subscripts in a fixed order e.g. in the order of
increasing energy and/or other quantum numbers

e.g. u1,2,4 , u1,1,2,4 , u1,2,2 etc.

4
 1 X
un1 ,···,nN (~r1 , · · · , ~rN ) ≡ √ un1 (~r1 ) · · · unN (~rN )
N!
Permutations of ~r1 , · · · ~rN
Now let us check the normalization of the basis states.
Z
d3 r1 · · · d3 rN un1 ,···,nN (~r1 , · · · , ~rN )∗ ul1 ,···,lN (~r1 , · · · , ~rN )

•: Unless li = ni for every i, the result vanishes

•: If li = ni for every i and all the ni ’s are different, then the result is 1

– the 1/N ! in the overall normalization cancels the N ! contributions, each giving
1

– all cross terms vanish, e.g.

Z
d3 r1 d3 r2 u∗n1 (~r1 )u∗n2 (~r2 )un2 (~r1 )un1 (~r2 ) = 0

if n1 6= n2 .

The situation is more complicated if li = ni for every i, and some of the ni ’s

are equal, e.g. u1,1,2

In this case some of the terms in the expression for un1 ,···,nN are equal since
permuting identical indices will not change the term.

To understand what happens in the case, it is useful to use a slightly different

way of labelling the basis states.

Instead of saying what states are occupied using the labels n1 , · · · , nN , possibly
with some labels repeated several times, we specify how many times the label
1 appears, how many times the label 2 appears etc.

5
e.g. u1,1,2 will be described as (2, 1, 0, 0, · · ·) – occupancy number

Occupancy number mi tells us how many particles occupy the i-th state ui

Occupancy number (m1 , m2 , · · ·) state mean that the n’th state appears mn
times
P∞
For an N particle state n=1 mn =N

Now let us consider the state

1 X
un1 ,···,nN (~r1 , · · · , ~rN ) ≡ √ un1 (~r1 ) · · · unN (~rN )
N!
Permutations of ~r1 , · · · ~rN
with occupancy numbers (m1 , m2 , · · ·)

Inside the square bracket not all N ! terms are different

– they are grouped into N !/(m1 !m2 ! · · ·) groups, each group containing m1 !m2 ! · · ·
identical terms:
1 N!
un1 ,···,nN (~r1 , · · · , ~rN ) = √ m1 !m2 ! · · · (t1 + t2 + · · · tp ), p=
N! m1 !m2 ! · · ·
Different terms ti are orthonormal since at least for one of the ~ri ’s the label
differs

The total norm is:

1
× (m1 !m2 ! · · ·)2 × p = m1 !m2 ! · · ·
N!
Therefore the state un1 ,···,nN (~r1 , · · · , ~rN ) is not normalized to 1.

We shall keep it this way.

6
Example:
1
u1,1,2 = √ [u1 (~r1 )u1 (~r2 )u2 (~r3 ) + permutations of ~r1 , ~r2 , ~r3 ]
3!
1
= √ [2 u1 (~r1 )u1 (~r2 )u2 (~r3 ) + 2 u1 (~r1 )u1 (~r3 )u2 (~r2 ) + 2 u1 (~r2 )u1 (~r3 )u2 (~r1 )]
6
Z
1
d3 r1 d3 r2 d3 r3 u∗1,1,2 u1,1,2 = × 22 × 3 = 2
6
– agrees with m1 !m2 ! · · · = 2! = 2

Next we shall analyze the action of N particle Hamiltonian on these basis

states.
N
X
H
bN = hi
b
i=1

hi un1 (~r1 ) · · · unN (~rN ) = eni un1 (~r1 ) · · · unN (~rN )

b
N
!
X
b N un (~r1 ) · · · un (~rN ) =
H eni un1 (~r1 ) · · · unN (~rN )
1 N
i=1

Note: The eigenvalue of H b N does not change under permutation of the ~ri ’s on
both sides.
N
!
X
⇒Hb N un ,···,n (~r1 , · · · , ~rN ) =
1 N
en un ,···,n (~r1 , · · · , ~rN )
i 1 N
i=1

If the state has occupancy numbers (m1 , m2 , · · ·) then the energy eigenvalue is
∞
X
mn en
n=1

7
We shall now forget about this system and consider another quantum system

– a collection of infinite number of one dimensional harmonic oscillators with

angular frequencies ω1 , ω2 , · · ·

Choose ωn ’s such that:

ωn = en /~
⇒ for each single particle energy eigenstate of the original theory, we have a
harmonic oscilator.

We can introduce creation and annihilation operators an , a†n for each harmonic
oscillator:
[an , ap ] = 0, [a†n , a†p ] = 0, [an , a†p ] = δnp
Hamiltonian of this system:
∞   ∞
X 1 X
H
b= ~ ωn a†n an + = ~ ωn a†n an + C
n=1
2 n=1

C: An overall constant which will have no physical significance

(We shall only be interested in differences in energy eigenvalues of this system)

Ground state |0i is defined as

an |0i = 0 for every n

Other states are created by applying arbitrary combinations of a†n on the vac-
uum state.

Goal: We shall show that this quantum theory is related to the system consid-
ered earlier under certain identificaions.

8
Identification of states:

un1 ,n2 ,···,nN ↔ a†n1 · · · a†nN |0i

Note that like un1 ,n2 ,···,nN , the right hand side is automatically symmetric under
ni ↔ nj

Occupation number representation:

(a†1 )m1 (a†2 )m2 · · · |0i

Next compute inner product of the states (a†1 )m1 (a†2 )m2 · · · |0i and (a†1 )m1 (a†2 )m2 · · · |0i
0 0

0 0
h0|(a1 )m1 (a2 )m2 · · · (a†1 )m1 (a†2 )m2 · · · |0i

We can analyze this by taking ai ’s to the right using the commutator between
ai and a†i .

• Unless m1 = m01 , m2 = m02 etc. the result vanishes

Any excess an or a†n will annihilate either |0i or h0|.

Ex. Check that:

h0|(a1 )m1 (a2 )m2 · · · (a†1 )m1 (a†2 )m2 · · · |0i = m1 !m2 ! · · ·

Example:

h0|(a1 )2 (a†1 )2 |0i = h0|a1 a1 a†1 a†1 |0i = h0|a1 (a†1 a1 + 1)a†1 |0i

Now replace every a1 a†1 in this expression by a†1 a1 + 1 and use

h0|a†1 = 0, a1 |0i = 0, h0|0i = 1

Result: 2 = 2!

Conclusion: a†n1 · · · a†nN |0i has the same norm as un1 ,···,nN in the first theory.

9
This gives a map between the Hilbert spaces of the two theories.

Let us compare the energy eigenvalues in the two quantum systems.

∞
!
X
b †n · · · a†n |0i =
Ha ~ωn a†n an a†n · · · a†n |0i
1 N 1 N
n=1
Use
∞
" #
X
~ωn a†n an , a†p = ~ ωp a†p
n=1

b †n · · · a†n |0i = ~(ωn + · · · + ωn ) a†n · · · a†n |0i

Ha 1 N 1 N 1 N

= (en1 + · · · + enN ) a†n1 · · · a†nN |0i

Note: we have dropped the additive constant C

This agrees with the energy eigenvalue of un1 ,n2 ,···,nN

Eigenstates of H
b N get mapped to eigenstates of H
b with the same eigenvalue.

Later we shall see that every operator in the first theory can be mapped to an
operator in the second theory.

However the converse is not true.

Specifying the first system requires fixing a value of N – the total number of
particles

This is not the case in the second system, e.g.

a†n |0i is a single particle state

a†n1 a†n2 |0i is a two particle state etc.

10
In the second system, the information on N is encoded in the eigenvalue of a
certain operator, called the number operator:
∞
X h i
N
b= a†n an , b , a†p = a†p
N
n=1

b a†n · · · a†n |0i = N a†n · · · a†n |0i

N 1 N 1 N

Therefore in the second system we can have operators that can change the
number of particles in the system.

Example: a†n1 a†n2 an3 increases the number of particles by 1

– removes a particle in state un3 and creates a pair of particles in state un1 and
un2 .

Such operators do not exist in the first description since we work with a fixed
number of particles.

Due to this, the second description has certain advantages.

If experiments show that particle number is not conserved, we can try to explain
this by adding to the Hamiltonian operators that do not conserve particle
number.

Later we shall see that there is a third description that is equivalent to the
second description

– a quantum theory of fields

11