This action might not be possible to undo. Are you sure you want to continue?

BooksAudiobooksComicsSheet Music### Categories

### Categories

### Categories

Editors' Picks Books

Hand-picked favorites from

our editors

our editors

Editors' Picks Audiobooks

Hand-picked favorites from

our editors

our editors

Editors' Picks Comics

Hand-picked favorites from

our editors

our editors

Editors' Picks Sheet Music

Hand-picked favorites from

our editors

our editors

Top Books

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Audiobooks

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Comics

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Sheet Music

What's trending, bestsellers,

award-winners & more

award-winners & more

Welcome to Scribd! Start your free trial and access books, documents and more.Find out more

Hartmut Wittig

Theory Group

Deutsches Elektronen-Synchrotron, DESY

Notkestrasse 85

22603 Hamburg

Germany

Lectures presented at the School for Young High Energy Physicists,

Rutherford Appleton Laboratory, September 2003

Contents

0 Prologue 3

1 Ingredients of Quantum Field Theory 3

1.1 Basic concepts of Quantum Mechanics . . . . . . . . . . . . . . . . . . . . 3

1.2 Special Relativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.3 Relativistic wave equation: The Klein-Gordon equation . . . . . . . . . . . 6

Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2 Quantisation of the free scalar ﬁeld 8

2.1 Heisenberg picture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2.2 Plane wave solutions of the Klein-Gordon equation . . . . . . . . . . . . . 9

2.3 Quantisation of the real Klein-Gordon ﬁeld . . . . . . . . . . . . . . . . . . 10

2.4 Causality and commutation relations . . . . . . . . . . . . . . . . . . . . . 12

Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

3 Lagrangian formalism 17

3.1 Classical Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3.2 Classical ﬁeld theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

3.3 Quantum ﬁeld theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

3.4 Summary: canonical quantisation for real scalar ﬁelds . . . . . . . . . . . . 24

Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

4 Interacting scalar ﬁelds 26

4.1 The S-matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

4.2 More on time evolution: Dirac picture . . . . . . . . . . . . . . . . . . . . 28

4.3 S-matrix and Green’s functions . . . . . . . . . . . . . . . . . . . . . . . . 31

Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

5 Perturbation Theory 34

5.1 Wick’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

5.2 The Feynman propagator . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

5.3 Two-particle scattering to O(λ) . . . . . . . . . . . . . . . . . . . . . . . . 38

5.4 Graphical representation of the Wick expansion: Feynman rules . . . . . . 40

5.5 Feynman rules in momentum space . . . . . . . . . . . . . . . . . . . . . . 42

5.6 S-matrix and truncated Green’s functions . . . . . . . . . . . . . . . . . . 43

Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

6 Concluding remarks 45

Acknowledgements 45

A Notation and conventions 46

1

When I became a student of Pomeranchuk

in 1950 I heard from him a kind of joke

that the Book of Physics had two volumes:

vol.1 is “Pumps and Manometers”, vol.2

is “Quantum Field Theory”

Lev Okun

0 Prologue

The development of Quantum Field Theory is surely one of the most important achieve-

ments in modern physics. Presently, all observational evidence points to the fact that

Quantum Field Theory (QFT) provides a good description of elementary particles for a

wide range of energies, leading up to the Planck mass, E

<

∼

M

Planck

· 10

19

GeV. His-

torically, Quantum Electrodynamics (QED) emerged as the prototype of modern QFT’s.

It was developed in the late 1940s and early 1950s chieﬂy by Feynman, Schwinger and

Tomonaga, and is perhaps the most successful theory in physics: the anomalous magnetic

dipole moment of the electron is predicted by QED with an accuracy of one part in 10

10

.

The scope of these lectures is to provide an introduction to the formalism of Quan-

tum Field Theory. This is best explained by restricting the discussion to the quantum

theory of scalar ﬁelds. Furthermore, I shall use the Lagrangian formalism and canonical

quantisation, thus leaving aside the quantisation approach via path integrals. Since the

main motivation for these lectures is the discussion of the underlying formalism leading

to the derivation of Feynman rules, the canonical approach is totally adequate. The

physically relevant theories of QED, QCD and the electroweak model are covered in the

lectures by Nick Evans, Michael Kr¨ amer and Adrian Signer.

The outline of these lecture notes is as follows: in section 1 we shall review the two

main ingredients of QFT, namely Quantum Mechanics and Special Relativity. Section 2

discusses the relativistic wave equation for free scalar ﬁelds, i.e. the Klein-Gordon equa-

tion. We shall derive a consistent quantum interpretation of its solutions in terms of

the Klein-Gordon ﬁeld. In section 3 the Lagrangian formalism will be discussed, which

is based on the Principle of least Action and provides a general framework to impose

the canonical quantisation rules. The more interesting case of interacting scalar ﬁelds is

presented in section 4: we shall introduce the S-matrix and examine its relation with the

Green’s functions of the theory. Finally, in section 5 the general method of perturbation

theory is presented, which serves to compute the Green functions in terms of a power

series in the coupling constant. Here, Wick’s Theorem is of central importance in order

to understand the derivation of Feynman rules.

1 Ingredients of Quantum Field Theory

1.1 Basic concepts of Quantum Mechanics

Let us recall that physical states in Quantum Mechanics are described by vectors [ψ` in

a Hilbert space H. On the other hand, physical observables are described by hermitian

3

operators acting on the state vectors [ψ`. Common examples for operators are

position operator : ˆ x

momentum operator : ˆ p = −i∇

Hamiltonian :

ˆ

H =

ˆ p

2

2m

+ V (ˆ x) = −

2

∇

2

2m

+ V (ˆ x). (1.1)

The required hermiticity of operators implies that these have real eigenvalues, so that

their expectation values can be interpreted as the result of a measurement.

The deﬁnition of the Hamiltonian follows the expression for the total energy in clas-

sical mechanics, i.e.

T + V =

p

2

2m

+ V (x). (1.2)

If the potential energy V vanishes, then the total energy E is given by

E =

p

2

2m

(1.3)

which is the classical, non-relativistic energy-momentum relation. In Quantum Mechanics,

the operators must satisfy commutation relations, for instance,

ˆ x, ˆ p

= i (1.4)

[ˆ x, ˆ x] =

ˆ p, ˆ p

= 0. (1.5)

Equation (1.4) expresses the well-known fact that precise and independent measurements

of x and p cannot be made. To summarise the above, one may say that quantisation is

achieved by replacing physical observables by the corresponding hermitian operators and

imposing suitable commutation relations.

Quantum Mechanics can be formulated in terms of diﬀerent so-called quantum pic-

tures, which treat the time dependence of state vectors and operators in diﬀerent ways.

In the Schr¨ odinger picture state vectors are time-dependent, whereas operators describ-

ing observables are time-independent. The state vector [ψ` can be represented by the

wavefunction Ψ(x, t), which satisﬁes the Schr¨ odinger equation:

i

∂

∂t

Ψ(x, t) =

ˆ

HΨ(x, t) (1.6)

The expectation value of some hermitian operator

ˆ

O at a given time t is then deﬁned as

'

ˆ

O`

t

=

d

3

xΨ

∗

(x, t)

ˆ

OΨ(x, t), (1.7)

and the normalisation of the wavefunction is given by

d

3

xΨ

∗

(x, t)Ψ(x, t) = '1`

t

. (1.8)

Since Ψ

∗

Ψ is positive, it is natural to interpret it as the probability density for ﬁnding

a particle at position x. Furthermore one can derive a conserved current j, as well as a

continuity equation by considering

Ψ

∗

(Schr.Eq.) −Ψ (Schr.Eq.)

∗

. (1.9)

4

The continuity equation reads

∂

∂t

ρ = −∇ j (1.10)

where the density ρ and the current j are given by

ρ = Ψ

∗

Ψ (positive), (1.11)

j =

2im

(Ψ

∗

∇Ψ−(∇Ψ

∗

)Ψ) (real). (1.12)

Now that we have derived the continuity equation let us discuss the probability interpre-

tation of Quantum Mechanics in more detail. Consider a ﬁnite volume V with boundary

S. The integrated continuity equation is

V

∂ρ

∂t

d

3

x = −

V

∇ j d

3

x

= −

S

j dS (1.13)

where in the last line we have used Gauss’s theorem. Using eq. (1.8) the lhs. can be

rewritten and we obtain

∂

∂t

'1`

t

= −

S

j dS = 0. (1.14)

In other words, provided that j = 0 everywhere at the boundary S, we ﬁnd that the

time derivative of '1`

t

vanishes. Since '1`

t

represents the total probability for ﬁnding

the particle anywhere inside the volume V , we conclude that this probability must be

conserved: particles cannot be created or destroyed in our theory. Non-relativistic Quan-

tum Mechanics thus provides a consistent formalism to describe a single particle. The

quantity Ψ(x, t) is interpreted as a one-particle wave function.

1.2 Special Relativity

Before we try to extend the formalism of Quantum Mechanics to relativistic systems, let

us ﬁrst recall some basic facts about special relativity. Let x

µ

= (x

0

, x) denote a 4-vector,

where x

0

= ct

x

is the time coordinate (our conventions for 4-vectors, the metric tensor,

etc. are listed in Appendix A). Suppose that someone sends out a light signal at time y

0

and position y. In the time interval (x

0

−y

0

) the light travels a distance [x −y[, so that

(x −y)

2

= (x

0

−y

0

)

2

(x

0

−y

0

)

2

−(x −y)

2

≡ (x −y)

2

= 0. (1.15)

This deﬁnes the light cone about y. In the forward light cone (FLC) the concepts of

“before” and “after” make sense. One important postulate of special relativity states

that no signal and no interaction can travel faster than the speed of light. This has

important consequences about the way in which diﬀerent events can aﬀect each other.

For instance, two events which are characterised by space-time points x

µ

and y

µ

are said

to be causal if the distance (x−y)

2

is time-like, i.e. (x−y)

2

> 0. By contrast, two events

characterised by a space-like separation, i.e. (x −y)

2

< 0, cannot aﬀect each other, since

the point x is not contained inside the light cone about y.

5

y

space

time

(x −y)

2

< 0, space-like

(x −y)

2

> 0, time-like

(x −y)

2

= 0, light-like

Figure 1: The light cone about y. Events occurring at points x and y are said to be

time-like (space-like) if x is inside (outside) the light cone about y.

In non-relativistic Quantum Mechanics the commutation relations among operators

indicate whether precise and independent measurements of the corresponding observables

can be made. If the commutator does not vanish, then a measurement of one observables

aﬀects that of the other. From the above it is then clear that the issue of causality must

be incorporated into the commutation relations of the relativistic version of our quantum

theory: whether or not independent and precise measurements of two observables can be

made depends also on the separation of the 4-vectors characterising the points at which

these measurements occur. We will return to this important issue in section 2.

1.3 Relativistic wave equation: The Klein-Gordon equation

In analogy to the non-relativistic case we shall now derive a relativistically invariant wave

equation. Our starting point is the relativistic energy-momentum relation

E

2

= p

2

+ m

2

(1.16)

where we have set c = 1. We may try to quantise the theory by replacing observables by

the corresponding hermitian operators, which gives

−

2

∂

2

∂t

2

φ(x, t) = −

2

∇

2

φ(x, t) + m

2

φ(x, t), (1.17)

where φ(x, t) is the wavefunction. This can be rewritten as

2

∂

2

∂t

2

−∇

2

φ(x, t) + m

2

φ(x, t) = 0

⇒

2

+ m

2

φ(x, t) = 0, (1.18)

which is the Klein-Gordon equation (KGE) for a free particle. As in the non-relativistic

case one can derive a continuity equation:

φ

∗

(KGE) −φ (KGE)

∗

. (1.19)

6

The density and current in this case are given by

ρ = i

φ

∗

∂

∂t

φ −(

∂

∂t

φ

∗

)φ

(real) (1.20)

j = −i (φ

∗

∇φ −(∇φ

∗

)φ) (real) (1.21)

However, if we try to take over the probabilistic interpretation encountered in the non-

relativistic case, we run into a number of problems:

1. Although

V

ρ d

3

x is again independent of time, the quantity ρ, though real, is not

positive deﬁnite, which means that ρ cannot be interpreted as a probability density;

2. One might try to interpret φ

∗

φ as a probability density instead, but then one ﬁnds

that φ

∗

φ is not conserved, since it does not satisfy a continuity equation. In other

words, it cannot be shown that

V

φ

∗

φd

3

x is independent of time;

3. There are states of arbitrarily low energies:

E = ±

p

2

+ m

2

(1.22)

so that the system has no ground state.

Therefore we will have to abandon the interpretation which was so successful in the non-

relativistic case.

So far we have assumed that the number of particles should be constant in the

relativistic theory. If we drop this requirement and introduce the concept of particle

creation or annihilation, then the interpretation of φ as a one-particle wavefunction does

not have to be upheld. As a consequence,

V

φ

∗

φd

3

x may vary over time due to the

change in the number of particles. Although this apparently overcomes the problem

that a consistent relativistic theory for one particle cannot be formulated, the problem

of negative energies remains unsolved. As we shall see, this is ultimately possible if we

re-interpret φ as a ﬁeld operator which can destroy or create particles.

Problems

1.1 Starting from the Schr¨ odinger equation for the wavefunction Ψ(x, t),

−

2

∇

2

2m

+ V (x)

Ψ(x, t) = i

∂

∂t

Ψ(x, t)

show that the probability density ρ = Ψ

∗

Ψ satisﬁes the continuity equation eq. (1.10),

with the current j given by eq. (1.12). Verify that the continuity equation can be

written in manifestly covariant form, i.e.

∂

µ

j

µ

= 0, j

µ

= (cρ, j).

1.2 Derive the corresponding continuity equation for the Klein-Gordon equation using

a similar procedure as in Problem 1.1. Set c = 1 and verify that the continuity

equation can be written as ∂

µ

j

µ

= 0 where

j

µ

= i (φ

∗

(∂

µ

φ) −(∂

µ

φ

∗

) φ) .

7

2 Quantisation of the free scalar ﬁeld

We are now going to discuss the Klein-Gordon equation for free, i.e. non-interacting

particles in more detail, ﬁnally arriving at a consistent way to quantise it. From now on

we will always set = c = 1.

2.1 Heisenberg picture

Before we are ready to discuss solutions of the Klein-Gordon equation we have to look at

the equations of motion in more detail, and derive a result which will prove useful for the

later analysis. As mentioned earlier, the physical states of a system are represented by

vectors in a Hilbert space. We have already encountered the Schr¨ odinger picture, where

the state vectors have an explicit time dependence. In order to indicate this we shall

use the notation [t; S` for a Schr¨ odinger state at time t which satisﬁes the Schr¨ odinger

equation:

i

∂

∂t

[t; S` =

ˆ

H[t; S`. (2.1)

By contrast, operators have no explicit time dependence in the Schr¨ odinger picture. The

formal solution of eq. (2.1) in a time interval from t

0

to t is given by

[t; S` = e

−i

ˆ

H(t−t

0

)

[t

0

; S`, (2.2)

which is easily veriﬁed by inserting it into the Schr¨ odinger equation.

Let us consider the expectation value of an observable A, i.e.

't; S[

ˆ

A

S

[t; S` ≡

d

3

xΨ

∗

(x, t)

ˆ

A

S

Ψ(x, t), (2.3)

where the subscript “S” on

ˆ

A

S

denotes that the corresponding operator is deﬁned in

the Schr¨ odinger picture. It is possible to trade oﬀ the time dependence between the state

vectors and the operators, since both are merely mathematical devices to describe physical

quantities. The expectation value, however, has to remain unchanged. Using the solution

eq. (2.2) we obtain

't; S[

ˆ

A

S

[t; S` = 't

0

; S[e

i

ˆ

H(t−t

0

)

ˆ

A

S

e

−i

ˆ

H(t−t

0

)

[t

0

; S`. (2.4)

We can thus deﬁne the time-dependent operator

ˆ

A

H

(t) and the time-independent state

vector [H` through

ˆ

A

H

(t) = e

i

ˆ

H(t−t

0

)

ˆ

A

S

e

−i

ˆ

H(t−t

0

)

(2.5)

[H` ≡ [t

0

; H` = [t

0

; S`. (2.6)

Then we have

't; S[

ˆ

A

S

[t; S` = 'H[

ˆ

A

H

(t)[H` (2.7)

The above relations between

ˆ

A

S

and

ˆ

A

H

(t) and the respective state vectors deﬁne the

Heisenberg picture, in which operators and states are labelled by “H”. The state vectors

[H` of the Heisenberg picture are time-independent but may carry a time label, which

8

simply means that these vectors are identiﬁed with the corresponding Schr¨ odinger states

at that particular time (see eq. (2.6)). Using eq. (2.5) it is easy to derive

∂

∂t

ˆ

A

H

(t) = i

ˆ

H,

ˆ

A

H

(t)

, (2.8)

which is the Heisenberg equation of motion.

State vectors in both pictures can depend on a spatial coordinate x. Let us consider

a Heisenberg state [x; H` and compute

ˆ p, ˆ x

[x; H`:

[ ˆ p

j

, ˆ x

k

] [x; H` = ˆ p

j

ˆ x

k

[x; H` − ˆ x

k

ˆ p

j

[x; H`

= −i

∂

∂x

j

(ˆ x

k

[x; H`) + ix

k

∂

∂x

j

[x; H`

= −i

∂x

k

∂x

j

[x; H` −ix

k

∂

∂x

j

[x; H` + ix

k

∂

∂x

j

[x; H`

= −iδ

jk

[x; H`. (2.9)

(The indices j, k refer to the components of x and p). What we have recovered is the

uncertainty relation, [ˆ p

j

, ˆ x

k

] = −iδ

jk

. This result can easily be generalised to Heisenberg

operators

ˆ

A

H

which have an explicit dependence on x:

ˆ p

j

,

ˆ

A

H

(x, t)

= −i

∂

∂x

j

ˆ

A

H

(x, t). (2.10)

To summarise, any Heisenberg operator satisﬁes the following equations of motion:

∂

∂t

ˆ

A

H

(x, t) = i

ˆ

H,

ˆ

A

H

(x, t)

(2.11)

∂

∂x

j

ˆ

A

H

(x, t) = i

ˆ p

j

,

ˆ

A

H

(x, t)

. (2.12)

If, in accordance with the conventions in Appendix A, we use the 4-vectors

∂

∂x

µ

=

∂

∂t

, −∇

⇒ ˆ p

µ

=

ˆ

H, ˆ p

(2.13)

then we can rewrite the equations of motion in covariant form

∂

∂x

µ

ˆ

A

H

(x) = i

ˆ p

µ

,

ˆ

A

H

(x)

, x ≡ x

µ

= (t, x). (2.14)

These are called the generalised Heisenberg equations of motion.

2.2 Plane wave solutions of the Klein-Gordon equation

In covariant form the Klein-Gordon equation reads

+ m

2

φ(x) = 0. (2.15)

Let us consider real solutions, characterised by φ

∗

(x) = φ(x) and interpret φ(x) as a

classical, real ﬁeld. Here, “classical” simply means “not quantised”, and by using the

9

concept of “ﬁelds” we make the analogy with electromagnetism. It is well known that

in Maxwell’s theory the scalar and vector potentials (denoted by Φ and V , respectively)

satisfy wave equations of the type

Φ(x) = 0, V

i

(x) = 0, i = 1, 2, 3, (2.16)

and the same is true for the components of the electric and magnetic ﬁelds, i.e.

E

i

(x) = 0, B

i

(x) = 0, i = 1, 2, 3. (2.17)

These wave equations are identical to the Klein-Gordon equation, except for the mass

term, which is absent in the Maxwell theory. This illustrates the close relationship of

a relativistically invariant theory like classical electromagnetism with the Klein-Gordon

equation and motivates the interpretation of φ as a ﬁeld variable.

To ﬁnd the solutions of the Klein-Gordon equation let us consider an ansatz of plane

waves

φ(x) ∝ e

i(k

0

t−k·x)

(2.18)

The Klein-Gordon equation is satisﬁed if (k

0

)

2

−k

2

= m

2

so that

k

0

= ±

k

2

+ m

2

. (2.19)

If we choose the positive branch of the square root then we can deﬁne the energy as

E(k) =

k

2

+ m

2

> 0 (2.20)

and obtain two types of solutions which read

φ

+

(x) ∝ e

i(E(k)t−k·x)

, φ

−

(x) ∝ e

−i(E(k)t−k·x)

. (2.21)

The general solution is a superposition of φ

+

and φ

−

. Using

E(k)t −k x = k

µ

k

µ

= k

µ

k

µ

= k x (2.22)

this solution reads

φ(x) =

d

3

k

(2π)

3

2E(k)

e

ik·x

α

∗

(k) + e

−ik·x

α(k)

, (2.23)

where α(k) is an arbitrary complex coeﬃcient. From the general solution one easily reads

oﬀ that φ is real, i.e. φ = φ

∗

.

2.3 Quantisation of the real Klein-Gordon ﬁeld

Starting from the classical solution to the KGE in eq. (2.23) we will now derive a quantum

interpretation of φ. Recall that quantisation is achieved by replacing observables by

the corresponding operators, as well as the imposition of suitable commutation relations

among these operators. In electromagnetism one can indeed express one component of

10

the electric ﬁeld, E

i

(x), which is a classical observable, by the expectation value of a

ﬁeld operator, i.e.

E

i

(x) = ' [

ˆ

E

i

(x)[ `. (2.24)

classical observable exp. value of ﬁeld operator (2.25)

Historically this is how a ﬁeld theory was quantised for the ﬁrst time. Here we have used

empty bra and ket symbols to indicate that the states have to be speciﬁed in more detail

– this is discussed later. If we follow the same approach for the real Klein-Gordon ﬁeld,

then we have to replace the ﬁeld φ by the expectation value of the ﬁeld operator

ˆ

φ:

φ(x) = ' [

ˆ

φ(x)[ ` (2.26)

The condition φ = φ

∗

is obtained if

ˆ

φ

†

=

ˆ

φ. In other words, φ is a real solution of the

KGE if the operator

ˆ

φ is hermitian and satisﬁes the KGE as well. Furthermore,

ˆ

φ is a

Heisenberg operator and must therefore satisfy the generalised Heisenberg equation (c.f.

eq. (2.14)):

∂

∂x

µ

ˆ

φ(x) = i

ˆ p

µ

,

ˆ

φ(x)

. (2.27)

From the classical solution φ(x) of eq. (2.23) we can deduce the following form of

ˆ

φ(x)

ˆ

φ(x) =

d

3

k

(2π)

3

2E(k)

e

ik·x

ˆ a

†

(k) + e

−ik·x

ˆ a(k)

. (2.28)

Here ˆ a

†

(k) and ˆ a(k) are operators, whose properties can be worked out by inserting the

expression for

ˆ

φ into the generalised Heisenberg equation, which gives

∂

∂x

µ

ˆ

φ(x) =

d

3

k

(2π)

3

2E(k)

¸

e

ik·x

(ik

µ

)ˆ a

†

(k) + e

−ik·x

(−ik

µ

)ˆ a(k)

¸

=

d

3

k

(2π)

3

2E(k)

¸

e

ik·x

i[ˆ p

µ

, ˆ a

†

(k)] + e

−ik·x

i[ˆ p

µ

, ˆ a(k)]

¸

. (2.29)

By comparing the integrands we ﬁnd

[ˆ p

µ

, ˆ a

†

(k)] = k

µ

ˆ a

†

(k) (2.30)

[ˆ p

µ

, ˆ a(k)] = −k

µ

ˆ a(k) (2.31)

These relation will help later us to specify the state vectors in more detail. Let us now

consider the vacuum state, which does not contain any particles. This state is denoted

by [0`, and it we normalise it to one, i.e.

'0[0` = 1. (2.32)

Furthermore, the vacuum has zero energy and momentum, which implies that the mo-

mentum operator ˆ p

µ

has a zero eigenvalue when acting on [0`:

ˆ p

µ

[0` = 0. (2.33)

11

Now we consider [ˆ p

µ

, ˆ a

†

(k)][0`. Using eq. (2.30) we ﬁnd that

[ˆ p

µ

, ˆ a

†

(k)][0` = ˆ p

µ

ˆ a

†

(k)[0` − ˆ a

†

(k)ˆ p

µ

[0`

= k

µ

ˆ a

†

(k)[0`

⇒ ˆ p

µ

ˆ a

†

(k)[0` = k

µ

ˆ a

†

(k)[0`. (2.34)

In other words, ˆ a

†

(k)[0` is an eigenstate of ˆ p

µ

with eigenvalue k

µ

= (E(k), k), where k

µ

is

the 4-momentum of a relativistic particle with mass m, and E(k) =

k

2

+ m

2

. Therefore

it makes sense to interpret ˆ a

†

(k)[0` as a one-particle state. Similarly we ﬁnd that

ˆ p

µ

ˆ a(k)[0` = −k

µ

ˆ a(k)[0`. (2.35)

However, this means that ˆ a(k)[0` is a state with negative energy, −k

0

= −E(k) < 0. Such

states are clearly unphysical, so that we deﬁne

ˆ a(k)[0` = 0. (2.36)

Furthermore, one can show by applying eqs. (2.30) and (2.31), that

ˆ p

µ

ˆ a

†

(k

2

)ˆ a

†

(k

1

)[0` = (k

µ

1

+ k

µ

2

)ˆ a

†

(k

2

)ˆ a

†

(k

1

)[0` (2.37)

ˆ p

µ

ˆ a(k

2

)ˆ a

†

(k

1

)[0` = (k

µ

1

−k

µ

2

)ˆ a(k

2

)ˆ a

†

(k

1

)[0`. (2.38)

It is thus natural to interpret ˆ a

†

(k

2

)ˆ a

†

(k

1

)[0` as a two-particle state, and, after repeating

this procedure n times one obtains

ˆ a

†

(k

n

) ˆ a

†

(k

1

)[0` n-particle state (2.39)

The operators ˆ a

†

(k) and ˆ a(k) increase, respectively decrease the number of particles in a

given state by one. We call

ˆ a

†

(k) : creation operator

ˆ a(k) : annihilation operator. (2.40)

The states which appear in our quantised version of the Klein-Gordon equation are thus

multi-particle states. We still have to look at the normalisation of these states in detail,

and also introduce the commutation rules for our ﬁeld operator

ˆ

φ.

2.4 Causality and commutation relations

In section 1 it was mentioned that commutation relations of operators in the relativistic

case must be imposed in accordance with the principle of causality, which is central in

Special Relativity. Consider two 4-vectors, x and x

, such that

(x −x

)

2

< 0 (space-like). (2.41)

In this case x is outside the light cone about x

**and vice versa. Events occurring at x
**

and x

**respectively (for instance, measurements of some observable), cannot aﬀect each
**

12

other – they are independent. It is then clear that the ﬁeld operators

ˆ

φ(x) and

ˆ

φ(x

) must

commute

ˆ

φ(x),

ˆ

φ(x

)

= 0 for (x −x

)

2

< 0. (2.42)

This condition is sometimes called micro-causality. Equation (2.42) is an important rela-

tion in order to derive the commutation rules for the creation and annihilation operators

ˆ a and ˆ a

†

. Our starting point is

ˆ

φ(x, t),

ˆ

φ(x

, t

)

= 0, [t

−t[ < [x −x

[ = 0, (2.43)

which is just eq. (2.42) with explicit space and and time coordinates. As long as [t

−t[ <

[x −x

[, the commutator vanishes in a ﬁnite interval [t

−t[. It also vanishes for t

= t,

hence

ˆ

φ(x, t),

ˆ

φ(x

, t)

= 0,

¸

ˆ

φ(x, t),

∂

∂t

ˆ

φ(x

, t)

= 0, x = x

. (2.44)

These expressions are referred to as the equal time commutators of the ﬁeld operator

ˆ

φ.

Before we can use them in order to deduce the commutators for creation and annihilation

operators, we rewrite

ˆ

φ, by letting k →−k in the term proportional to ˆ a(k) in eq. (2.28).

ˆ

φ(x, t) =

d

3

k

(2π)

3

2E(k)

e

−ik·x

¸

e

iEt

ˆ a

†

(k) + e

−iEt

ˆ a(−k)

¸

. (2.45)

Similarly one obtains

∂

0

ˆ

φ(x, t) ≡

∂

∂t

ˆ

φ(x, t) =

d

3

k

(2π)

3

i

2

e

−ik·x

¸

e

iEt

ˆ a

†

(k) −e

−iEt

ˆ a(−k)

¸

. (2.46)

The Fourier transform of

ˆ

φ(x, t) reads

d

3

xe

ip·x

ˆ

φ(x, t) =

d

3

k

(2π)

3

2E(k)

d

3

xe

i(p−k)·x

. .. .

(2π)

3

δ

3

(p−k)

¸

e

iEt

ˆ a

†

(k) + e

−iEt

ˆ a(−k)

¸

=

1

2E

¸

e

iEt

ˆ a

†

(p) + e

−iEt

ˆ a(−p)

¸

, (2.47)

where E = E(p) in the last line. Thus we ﬁnd

2E

d

3

xe

ip·x

ˆ

φ(x, t) =

¸

e

iEt

ˆ a

†

(p) + e

−iEt

ˆ a(−p)

¸

, (2.48)

and a similar procedure applied to the Fourier transform of ∂

0

ˆ

φ(x, t) yields

−2i

d

3

xe

ip·x

∂

∂t

ˆ

φ(x, t) =

¸

e

iEt

ˆ a

†

(p) −e

−iEt

ˆ a(−p)

¸

. (2.49)

If we take the result of eq. (2.48) we ﬁnd

e

iE

1

t

ˆ a

†

(p

1

) + e

−iE

1

t

ˆ a(−p

1

), e

iE

2

t

ˆ a

†

(p

2

) + e

−iE

2

t

ˆ a(−p

2

)

= 2E

1

2E

2

d

3

x

d

3

y e

ip

1

·x

e

ip

2

·y

ˆ

φ(x, t),

ˆ

φ(y, t)

. (2.50)

13

The equal time commutator tells us that the rhs. must vanish, which gives

e

i(E

1

+E

2

)t

ˆ a

†

(p

1

), ˆ a

†

(p

2

)

+ e

−i(E

1

−E

2

)t

ˆ a(−p

1

), ˆ a

†

(p

2

)

+ e

i(E

1

−E

2

)t

ˆ a

†

(p

1

), ˆ a(−p

2

)

+ e

−i(E

1

+E

2

)t

ˆ a(−p

1

), ˆ a(−p

2

)

= 0. (2.51)

This relation is valid for all times t, and we deduce that

ˆ a

†

(p

1

), ˆ a

†

(p

2

)

=

ˆ a(−p

1

), ˆ a(−p

2

)

= 0, (2.52)

which tells us the important result that both creation and annihilation operators commute

among themselves!

Let us discuss the consequences by considering two creation operators, ˆ a

†

(p

1

) and

ˆ a

†

(p

2

), acting on the vacuum [0`. The fact that creation operators commute implies that

[p

1

, p

2

` ≡ ˆ a

†

(p

1

)ˆ a

†

(p

2

)[0` = ˆ a

†

(p

2

)ˆ a

†

(p

1

)[0` ≡ [p

2

, p

1

`, (2.53)

where the positions of the 3-momenta on the far left- and far right-hand sides refer to

particle one and two, respectively. We see that, since [p

1

, p

2

` = [p

2

, p

1

`, the state [p

1

, p

2

`

is symmetric if we exchange the two particles. We have thus derived that our particles

must be bosons ! The quantised Klein-Gordon theory describes relativistic particles which

are bosons, i.e. which have integer spin. The fact that we have considered real solutions

to the KGE means that we consider particles without charge. An example of a neutral

particle with spin zero is the neutral pion, π

0

. The formalism that we have developed so

far can be used to describe free neutral pions.

We still have to derive the commutator of

ˆ a(p), ˆ a

†

(q)

. If we solve eq. (2.48) and (2.49)

for ˆ a and ˆ a

†

then we ﬁnd

ˆ a(p), ˆ a

†

(q)

= e

i(E(p)−E(q))t

d

3

x

d

3

y e

−ip·x

e

iq·y

−iE(p)

ˆ

φ(x, t), ∂

0

ˆ

φ(y, t)

+ iE(q)

∂

0

ˆ

φ(x, t),

ˆ

φ(y, t)

¸

. (2.54)

This expression can be simpliﬁed using the commutators for the ﬁeld operators. The

equal time commutator tells us that

ˆ

φ(x, t), ∂

0

ˆ

φ(y, t)

= 0 for x = y, (2.55)

as required by causality. However, so far we have not imposed any commutation rules for

the case x = y. Let us try

ˆ

φ(x, t), ∂

0

ˆ

φ(y, t)

= iδ

3

(x −y), (2.56)

in analogy with Quantum Mechanics, where [ˆ x

j

, ˆ p

k

] = iδ

jk

. We then obtain

ˆ a(p), ˆ a

†

(q)

= e

i(E(p)−E(q))t

d

3

x

d

3

y e

−ip·x

e

iq·y

¸

E(p) δ

3

(x −y) + E(q) δ

3

(y −x)

¸

= e

i(E(p)−E(q))t

d

3

xe

−i(p−q)·x

. .. .

(2π)

3

δ

3

(p−q)

(E(p) + E(q)). (2.57)

14

Our commutation rule for ˆ a and ˆ a

†

thus reads

ˆ a(p), ˆ a

†

(q)

= (2π)

3

2E(p)δ

3

(p −q). (2.58)

Armed with this result we can now look at the normalisation of our multi-boson states.

Let [k` and [k

**` be two one-meson states, i.e.
**

[k` = ˆ a

†

(k)[0`, [k

` = ˆ a

†

(k

)[0`. (2.59)

Then we have

'k

[k` = '0[ˆ a(k

)ˆ a

†

(k)[0`

= '0[

ˆ a(k

), ˆ a

†

(k)

[0` = (2π)

3

2E(k)δ

3

(k

−k) > 0. (2.60)

We see that the way we have set up the commutation rules for

ˆ

φ and ∂

0

ˆ

φ in eq. (2.56)

ensures that our multi-meson states have positive norm, i.e.

d

3

k

(2π)

3

2E(k

)

'k

[k` =

d

3

k

E(k)

E(k

)

δ

3

(k

−k) = 1, (2.61)

and thus our states form indeed a proper Hilbert space.

To conclude this chapter, let us make a brief summary of what we have actually

done. A consistent quantisation of the free Klein-Gordon ﬁeld has been achieved by going

through the following steps:

• Abandon the interpretation of φ(x, t) as a one-particle wave function and think of

it as a ﬁeld variable.

• Replace the classical ﬁeld φ by a ﬁeld operator

ˆ

φ.

• Expand

ˆ

φ in terms of creation and annihilation operators ˆ a

†

and ˆ a.

• Impose the equal time commutators

ˆ

φ(x, t), ∂

0

ˆ

φ(y, t)

= iδ

3

(x −y),

ˆ

φ(x, t),

ˆ

φ(y, t)

= 0, (2.62)

which ensures that the resulting theory obeys causality.

• This implies the following commutation rules for ˆ a

†

, ˆ a:

ˆ a

†

(k

1

), ˆ a

†

(k

2

)

= 0 (2.63)

[ˆ a(k

1

), ˆ a(k

2

)] = 0 (2.64)

ˆ a(k

1

), ˆ a

†

(k

2

)

= (2π)

3

2E(k

1

)δ

3

(k

1

−k

2

) (2.65)

Relation (2.63) implies that our particles must be bosons. The state vectors are

interpreted as multi-boson (or multi-meson) states:

[0` : vacuum, '0[0` = 1

ˆ a

†

(k)[0` = [k` : one-meson state, positive energy

ˆ a(k)[0` = 0 : exclude states with negative energy. (2.66)

15

Furthermore, relation (2.65) implies that one-meson states must have positive norm.

The generalisation to n-meson states is achieved by repeated applications of creation

operators, e.g.

ˆ a

†

(k

1

) ˆ a

†

(k

n

)[0` = [k

1

, . . . , k

n

` . (2.67)

The resulting Hilbert space of multi-particle states is called a Fock space.

Problems

2.1 Let

ˆ

A

H

(t) and

ˆ

A

S

be operators in the Heisenberg and Schr¨ odinger pictures, respec-

tively. Using the relation

ˆ

A

H

(t) = e

i

ˆ

H(t−t

0

)

ˆ

A

S

e

−i

ˆ

H(t−t

0

)

derive the Heisenberg equation of motion, eq. (2.8).

2.2 Derive the generalised uncertainty relation, eq. (2.10), for a Heisenberg operator

ˆ

A

H

(x, t). Hint: Consider

ˆ p

j

,

ˆ

A

H

(x, t)

**[x; H`, where [x; H` is an arbitrary Heisen-
**

berg state.

2.3 Given the relativistic invariance of the measure d

4

k, show that the integration mea-

sure

d

3

k

(2π)

3

2E(k)

is Lorentz-invariant, provided that E(k) =

k

2

+ m

2

. Hint: Start from the ex-

pression

d

4

k

(2π)

3

δ(k

2

−m

2

) θ(k

0

)

and use

δ(x

2

−x

2

0

) =

1

2x

(δ(x −x

0

) + δ(x + x

0

))

What is the signiﬁcance of the δ and θ functions above?

2.4 Verify that eq. (2.23) is indeed a solution of the Klein-Gordon equation.

2.5 Use the commutator relations

ˆ p

µ

, ˆ a

†

(k)

= k

µ

ˆ a

†

(k),

ˆ p

µ

, ˆ a(k)

= −k

µ

ˆ a(k)

to derive eqs. (2.37) and (2.38).

2.6 Starting from the expression for ∂

0

ˆ

φ(x, t):

∂

0

ˆ

φ(x, t) =

d

3

k

(2π)

3

i

2

e

−ik·x

¸

e

iEt

ˆ a

†

(k) −e

−iEt

ˆ a(−k)

¸

16

invert the Fourier transform to obtain

e

iEt

ˆ a

†

(k) −e

−iEt

ˆ a(−k) = −2i

d

3

xe

ik·x

∂

0

ˆ

φ(x, t).

Use this result together with eq. (2.48) in the lecture notes, i.e.

e

iEt

ˆ a

†

(k) + e

−iEt

ˆ a(−k) = 2E(k)

d

3

xe

ik·x

ˆ

φ(x, t).

to solve for ˆ a

†

(k) and ˆ a(k). Verify that the expression for the commutator [ˆ a(p), ˆ a

†

(q)]

reads

[ˆ a(p), ˆ a

†

(q)] = e

i(E(p)−E(q))t

d

3

xd

3

y e

−ip·x+iq·y

i

E(q)

∂

0

ˆ

φ(x, t),

ˆ

φ(y, t)

−E(p)

ˆ

φ(x, t), ∂

0

ˆ

φ(y, t)

¸

.

3 Lagrangian formalism

We could have derived all results in the previous section in a much more axiomatic fash-

ion, by starting from the Lagrangian of the theory. The Lagrangian formalism is used

in all modern theories of matter, most notably in the Standard Model of Elementary

Particle Physics, and its extensions. The starting point for any theoretical treatment of

fundamental interactions is the relevant Lagrangian, for instance

L

QED

, L

QCD

, L

GSW

(3.1)

for Quantum Electrodynamics, Quantum Chromodynamics and the electroweak (Glashow-

Salam-Weinberg) theory, respectively. In this section we will introduce the Lagrangian

formalism in the context of classical mechanics, before we move on to classical ﬁeld theory

and then quantise it.

3.1 Classical Mechanics

In Classical Mechanics one considers point particles of mass m at some position x. A

typical problem is then to ﬁnd the trajectory x(t) of that particle by integrating its

equation of motion, which is

m¨ x = F ≡ −

∂V

∂x

, (3.2)

where V is the potential. This well-known result is Newton’s 2nd law. The Lagrangian

formalism provides a generalisation such that the equation of motion can, in fact, be

derived from the Principle of least Action.

The Lagrangian for a single classical particle is deﬁned as

L = T −V =

1

2

m˙ x

2

−V (x)

= L(x, ˙ x). (3.3)

17

The integral

S =

t

2

t

1

L(x, ˙ x) dt (3.4)

is called the action. The Principle of least Action then states that the motion of the

particle (i.e. its trajectory) must be such that the action S is a minimum. In other words,

the requirement that S be a minimum singles out one particular trajectory x(t). Let us

minimise S in the standard fashion by considering a variation in the path x(t):

x(t) −→x

(t) = x(t) + δx(t), δx(t) <x. (3.5)

As a boundary condition we require that the points at the boundaries stay ﬁxed, i.e.

x

(t

1

) = x(t

1

)

x

(t

2

) = x(t

2

)

⇒δx(t

1

) = δx(t

2

) = 0. (3.6)

The variation of S can then be worked out in detail by considering

S + δS =

t

2

t

1

L(x + δx, ˙ x + δ ˙ x) dt, δ ˙ x =

d

dt

δx. (3.7)

The variation δS is obtained from the Taylor expansion of the integrand about (x, ˙ x)

δS =

t

2

t

1

∂L

∂x

δx +

∂L

∂ ˙ x

δ ˙ x

dt

=

∂L

∂ ˙ x

δx

t

2

t

1

+

t

2

t

1

∂L

∂x

−

d

dt

∂L

∂ ˙ x

δxdt (3.8)

The ﬁrst term in the last line vanishes, because of the boundary conditions. If S is a

minimum, then δS = 0, and since the remaining integral in eq. (3.8) must vanish for

arbitrary variations δx, this is only possible if the integrand itself vanishes. Thus we have

derived the Euler-Lagrange equation

∂L

∂x

−

d

dt

∂L

∂ ˙ x

= 0. (3.9)

If we insert eq. (3.3) into the Euler-Lagrange equation we obtain

∂L

∂x

= −

∂V (x)

∂x

= F

d

dt

∂L

∂ ˙ x

=

d

dt

m˙ x = m¨ x

⇒ m¨ x = F = −

∂V

∂x

(Newton’s law). (3.10)

So we ﬁnd that the equation of motion (here: Newton’s 2nd law) is reproduced by the

Principle of least Action.

In this example the Lagrangian formalism and Newton’s law are equivalent, and it is

a legitimate question to ask what we have gained. The key advantage of the Lagrangian

18

formalism is that it helps us to study the implications of symmetries for a given theory.

We will return to this point below when we discuss classical ﬁeld theory.

We shall now establish the relation of the Principle of least Action with the Hamil-

tonian formalism. To this end we deﬁne the conjugate momentum p by

p ≡

∂L

∂ ˙ x

= m˙ x, (3.11)

and the Hamiltonian H via

H(x, p) ≡ p ˙ x −L(x, ˙ x)

= m˙ x

2

−

1

2

m˙ x

2

+ V (x)

=

1

2

m˙ x

2

+ V (x) = T + V. (3.12)

The Hamiltonian H(x, p) is the total energy of the system; it is a function of the position

variable x and the conjugate momentum

1

p. It is now easy to derive Hamilton’s equations

(see exercise )

∂H

∂x

= −˙ p,

∂H

∂p

= ˙ x. (3.13)

The Euler-Lagrange equations and Hamilton’s equations provide an entirely equivalent

description of the system. We can now have a look at conservation laws. Assuming that

the Hamiltonian has no explicit time dependence we ﬁnd

d

dt

H(x, p) =

∂H

∂x

dx

dt

+

∂H

∂p

dp

dt

= −˙ p ˙ x + ˙ x ˙ p = 0, (3.14)

and hence we conclude that the total energy T + V is conserved.

3.2 Classical ﬁeld theory

We can extend the Lagrangian formalism for a classical point particle to ﬁeld theory. The

rˆ ole of the particle’s trajectory x(t) is now played by the classical ﬁeld φ(x, t) = φ(x).

The “dictionary” which provides the relation to ﬁeld theory reads as follows:

Classical Mechanics: Classical ﬁeld theory:

t −→ x

µ

x(t) −→ φ(x) (ﬁeld conﬁguration)

L(x, ˙ x) −→ L(φ, ∂

µ

φ) (Lagrangian density)

(3.15)

The action is deﬁned as

S =

d

4

xL(φ, ∂

µ

φ), (3.16)

where L is usually referred to as the Lagrangian density, and the Lagrangian is given by

L(φ, ∂

µ

φ) =

d

3

xL(φ, ∂

µ

φ). (3.17)

1

It should be noted that the conjugate momentum is in general not equal to m˙ x.

19

In order to derive the Euler-Lagrange equations in this case, let us consider variations of

the ﬁeld and its derivative according to

φ →φ + δφ, ∂

µ

φ →∂

µ

φ + δ∂

µ

φ, δ∂

µ

φ = ∂

µ

δφ. (3.18)

The variation of the action becomes

δS =

d

4

x

∂L

∂φ

δφ +

∂L

∂(∂

µ

φ)

δ(∂

µ

φ)

=

∂L

∂(∂

µ

φ)

δφ

. .. .

=0 at boundaries

+

d

4

x

∂L

∂φ

−∂

µ

∂L

∂(∂

µ

φ)

δφ. (3.19)

As in the previous subsection one can argue that the integrand itself must vanish if δS = 0.

This yields the Euler-Lagrange equations for a classical ﬁeld theory:

∂L

∂φ

−∂

µ

∂L

∂(∂

µ

φ)

= 0, (3.20)

where in the second term a summation over the Lorentz index µ is implied.

Let us now consider the Lagrangian

L =

1

2

∂

µ

φ∂

µ

φ −

1

2

m

2

φ

2

(3.21)

The functional derivatives yield

∂L

∂φ

= −m

2

φ,

∂L

∂(∂

µ

φ)

= ∂

µ

φ (3.22)

so that

∂

µ

∂L

∂(∂

µ

φ)

= ∂

µ

∂

µ

φ = φ. (3.23)

The Euler-Lagrange equation then implies

(+ m

2

)φ(x) = 0, (3.24)

and one recovers the Klein-Gordon equation from a Lagrangian via the Euler-Lagrange

equation. In analogy to classical mechanics we can deﬁne a conjugate momentum π

through

π(x) ≡

∂L(φ, ∂

µ

φ)

∂

˙

φ(x)

=

∂L(φ, ∂

µ

φ)

∂(∂

0

φ(x))

= ∂

0

φ(x). (3.25)

Note that the momentum variables p, p

µ

and the conjugate momentum π are not the

same. The word “momentum” is used only as a semantic analogy to classical mechanics.

In classical mechanics we have shown that the total energy is conserved. If we want

to do the same for classical ﬁeld theory, then we ﬁrst have to have a look at what is

known as Noether’s Theorem. In a nutshell, Noether’s theorem says that the invariance

of a Lagrangian under a symmetry transformation implies the existence of a conserved

20

quantity. For instance, the conservation of 3-momentum p is associated with translational

invariance of the Lagrangian, i.e. the transformation

x →x + a, a : constant 3-vector, (3.26)

while the conservation of energy comes from the invariance of the Lagrangian under time

translations

t →t + τ, τ : constant time interval. (3.27)

For our classical ﬁeld theory one can use Noether’s theorem to derive the following relation

∂

µ

∂L

∂(∂

µ

φ)

∂

ν

φ −g

µν

L

= 0, (3.28)

where the expression in curly brackets is called the energy-momentum tensor, Θ

µν

. Equa-

tion (3.28) states that the energy-momentum tensor is conserved for every component

ν:

∂

µ

Θ

µν

≡ ∂

0

Θ

0ν

−∂

j

Θ

jν

= 0. (3.29)

Let us examine Θ

00

in more detail:

Θ

00

=

∂L

∂(∂

0

φ)

∂

0

φ −g

00

L

= π(x)(∂

0

φ(x)) −L. (3.30)

The last line is reminiscent of the deﬁnition of the Hamiltonian in eq. (3.14), and so we

can deﬁne the Hamiltonian density H as

H(π, φ) ≡ Θ

00

= π(x)(∂

0

φ(x)) −L(φ, ∂

µ

φ). (3.31)

The conservation of energy can now be shown by considering

∂

∂t

V

d

3

xΘ

00

=

V

d

3

x∂

0

Θ

00

=

V

d

3

x∂

j

Θ

j0

=

S

dS Θ

0

= 0, (3.32)

where we have used eq. (3.29) in the second line. The Hamiltonian density is a conserved

quantity, provided that there is no energy ﬂow through the surface S which encloses the

volume V . In a similar manner one can show that 3-momentum p

j

, which is related to

Θ

0j

, is conserved as well.

We have thus established the Lagrange-Hamilton formalism for classical ﬁeld theory:

we derived the equation of motion (Euler-Lagrange equation) from the Lagrangian and

introduced the conjugate momentum. We then deﬁned the Hamiltonian (density) and

considered conservation laws by studying the energy-momentum tensor Θ

µν

.

21

3.3 Quantum ﬁeld theory

We shall now quantise the theory by promoting the classical ﬁelds φ, π to ﬁeld operators

ˆ

φ(x) and ˆ π(x), and by imposing the (equal time) commutators

ˆ

φ(x, t), ˆ π(y, t)

= iδ

3

(x −y), ˆ π(x) ≡ ∂

0

ˆ

φ(x) (3.33)

ˆ

φ(x, t),

ˆ

φ(y, t)

=

ˆ π(x, t), ˆ π(y, t)

= 0. (3.34)

Both operators

ˆ

φ and ˆ π can be expanded in terms of creation and annihilation operators

ˆ a

†

and ˆ a. Let’s work out the Hamiltonian, i.e. the operator of the total energy. The

starting point is the expression for the Hamiltonian density

H = ˆ π(x)(∂

0

ˆ

φ(x)) −L(

ˆ

φ, ˆ π). (3.35)

In terms of ˆ a

†

and ˆ a one obtains

ˆ

H ≡

d

3

xH =

1

4

d

3

p

(2π)

3

ˆ a

†

(p)ˆ a(p) + ˆ a(p)ˆ a

†

(p)

=

1

2

d

3

p

(2π)

3

2E(p)

E(p)

ˆ a

†

(p)ˆ a(p) + ˆ a(p)ˆ a

†

(p)

. (3.36)

Let us now work out the energies of some of our multi-particle states. Starting with the

vacuum as the ground state we get

'0[

ˆ

H[0` =

1

2

d

3

p

(2π)

3

2E(p)

E(p)

¸

'0[ˆ a

†

(p)ˆ a(p)[0` +'0[ˆ a(p)ˆ a

†

(p)[0`

¸

. (3.37)

The ﬁrst term in curly brackets vanishes, since ˆ a annihilates the vacuum. The second can

be rewritten as

ˆ a(p)ˆ a

†

(p)[0` =

¸

ˆ a(p), ˆ a

†

(p)

+ ˆ a

†

(p)ˆ a(p)

¸

[0`. (3.38)

It is now the second term which vanishes, whereas the ﬁrst can be replaced by the com-

mutator. Thus we obtain

'0[

ˆ

H[0` = δ

3

(0)

1

2

d

3

p E(p) = ∞, (3.39)

which means that the energy of the ground state is inﬁnite! This result seems rather

paradoxical, but it is quite easy to work around it. What we are actually interested in

is the energy of multi-particle states relative to the vacuum, i.e. the ground state. In

this case it does not really matter what the absolute value of the ground state energy is.

We can simply redeﬁne the vacuum energy by setting it to zero, thereby subtracting the

vacuum contribution from expectation values of the Hamiltonian

ˆ

H. Let us make this

more explicit by considering the modiﬁed Hamiltonian

ˆ

H

ˆ

H

=

1

2

d

3

p

(2π)

3

2E(p)

E(p)

¸

ˆ a

†

(p)ˆ a(p) + ˆ a(p)ˆ a

†

(p) −'0[ˆ a

†

(p)ˆ a(p) + ˆ a(p)ˆ a

†

(p)[0`

¸

=

1

2

d

3

p

(2π)

3

2E(p)

E(p)

¸

2ˆ a

†

(p)ˆ a(p) +

ˆ a(p), ˆ a

†

(p)

−'0[

ˆ a(p), ˆ a

†

(p)

[0`

¸

.(3.40)

22

Here the subtraction of the vacuum energy is shown explicitly, and we can rewrite is as

ˆ

H

=

1

2

d

3

p

(2π)

3

2E(p)

E(p)ˆ a

†

(p)ˆ a(p)

+

1

2

d

3

p

(2π)

3

2E(p)

E(p)

¸

ˆ a(p), ˆ a

†

(p)

−'0[

ˆ a(p), ˆ a

†

(p)

[0`

¸

.

=

1

2

d

3

p

(2π)

3

2E(p)

E(p) ˆ a

†

(p)ˆ a(p) +

ˆ

H

vac

(3.41)

The operator

ˆ

H

vac

ensures that the vacuum energy is properly subtracted: if [Ψ` and [Ψ

`

denote arbitrary Fock states, then it is easy to see that 'Ψ

[

ˆ

H

vac

[Ψ` = 0. In particular we

now ﬁnd that

'0[

ˆ

H

[0` = 0, (3.42)

as it should be. A simple way to remove the vacuum contribution is to introduce

normal ordering. Normal ordering means that all annihilation operators appear to the

right of any creation operator. The notation is

: ˆ aˆ a

†

: = ˆ a

†

ˆ a, (3.43)

i.e. the normal-ordered operators are enclosed within colons. For instance

:

1

2

ˆ a

†

(p)ˆ a(p) + ˆ a(p)ˆ a

†

(p)

: = ˆ a

†

(p)ˆ a(p). (3.44)

It is important to keep in mind that ˆ a and ˆ a

†

always commute inside : :. This is true

for an arbitrary string of ˆ a and ˆ a

†

. With this deﬁnition we can write the normal-ordered

Hamiltonian as

:

ˆ

H : = :

1

2

d

3

p

(2π)

3

2E(p)

E(p)

ˆ a

†

(p)ˆ a(p) + ˆ a(p)ˆ a

†

(p)

:

=

d

3

p

(2π)

3

2E(p)

E(p) ˆ a

†

(p)ˆ a(p). (3.45)

Hence, we ﬁnd that

'Ψ

[ :

ˆ

H : [Ψ` = 'Ψ

[

ˆ

H

[Ψ`, (3.46)

and, in particular, '0[ :

ˆ

H : [0` = 0. The normal ordered Hamiltonian thus produces a

sensible result for the vacuum energy.

Let us now look at the interpretation of the Hamiltonian in more detail. In particular,

we shall be interested in the combination ˆ a

†

ˆ a, which appears in its deﬁnition. Consider

a one-meson state, [k`, of a particle with mass m and momentum k. Let us see what

happens when the operator

d

3

p

(2π)

3

2E(p)

ˆ a

†

(p)ˆ a(p) acts on [k`:

d

3

p

(2π)

3

2E(p)

ˆ a

†

(p)ˆ a(p) [k` =

d

3

p

(2π)

3

2E(p)

ˆ a

†

(p)ˆ a(p) ˆ a

†

(k) [0`

=

d

3

p

(2π)

3

2E(p)

ˆ a

†

(p)

¸

ˆ a(p), ˆ a

†

(k)

+ ˆ a

†

(k)ˆ a(p)

¸

[0`

=

d

3

p ˆ a

†

(p) δ

3

(p −k)[0` = ˆ a

†

(k)[0` = [k`. (3.47)

23

We ﬁnd that [k` is an eigenstate of

d

3

p

(2π)

3

2E(p)

ˆ a

†

(p)ˆ a(p) with eigenvalue 1. Let us now

look at a two meson state [k, k`. Using similar manipulations as above we ﬁnd that

d

3

p

(2π)

3

2E(p)

ˆ a

†

(p)ˆ a(p) [k, k` = 2 [k, k`. (3.48)

At this point the pattern becomes clear: the operator

d

3

p

(2π)

3

2E(p)

ˆ a

†

(p)ˆ a(p) (3.49)

returns the number of mesons in a given Fock state, and is therefore referred to as the

number operator. This is easily generalised to n-meson states, i.e. one can derive

d

3

p

(2π)

3

2E(p)

ˆ a

†

(p)ˆ a(p) [k, . . . , k`

. .. .

n momenta

= n [k, . . . , k`

. .. .

n momenta

. (3.50)

Note that the normal-ordered Hamiltonian diﬀers from the number operator just by a

factor of E(p). If we work out the energy of an n-meson state we get

:

ˆ

H : [k, . . . , k` =

d

3

p

(2π)

3

2E(p)

E(p)ˆ a

†

(p)ˆ a(p) [k, . . . , k` = nE(k) [k, . . . , k`, (3.51)

where E(k) =

k

2

+ m

2

> 0. This implies that the energy of the n-meson state is

positive. One can easily generalise this to a system of n mesons of equal mass but unequal

momenta k

1

, . . . , k

n

:

:

ˆ

H : [k

1

, . . . , k

n

` = ¦E(k

1

) + . . . + E(k

n

)¦ [k

1

, . . . , k

n

`. (3.52)

We conclude that the Hamiltonian :

ˆ

H : returns the total energy of an n-meson state.

Since the number operator has strictly non-negative eigenvalues, and since E(k) > 0, the

energy of any Fock state cannot be negative. This is ﬁnally the solution of the negative

energy problem encountered earlier.

3.4 Summary: canonical quantisation for real scalar ﬁelds

1. Start from the classical Lagrangian density

L(φ, ∂

µ

φ) =

1

2

(∂

µ

φ∂

µ

φ) −

1

2

m

2

φ

2

, φ(x) : classical ﬁeld (3.53)

2. The Principle of least Action implies the equation of motion

∂L

∂φ

−∂

µ

∂L

∂(∂

µ

φ)

= 0, ⇒

+ m

2

φ(x) = 0. (3.54)

3. Deﬁne a conjugate momentum π(x) and the Hamiltonian density by

π(x) =

∂L

∂(∂

0

φ(x))

= ∂

0

φ(x),

H = π(x)(∂

0

φ(x)) −L(φ, ∂

µ

φ) (3.55)

24

4. Quantise the theory by regarding φ(x), π(x) as ﬁeld operators with equal time

commutators

ˆ

φ(x, t), ˆ π(y, t)

= iδ

3

(x −y)

ˆ

φ(x, t),

ˆ

φ(y, t)

=

ˆ π(x, t), ˆ π(y, t)

= 0. (3.56)

The operators

ˆ

φ and ˆ π are expanded in terms of ˆ a

†

and ˆ a, which are used to generate

a basis of Fock states.

5. The energy of a given Fock state is given by the normal-ordered Hamiltonian,

:

ˆ

H : =

1

2

d

3

p

(2π)

3

ˆ a

†

(p)ˆ a(p), (3.57)

acting on that state. Normal ordering ensures that the vacuum energy is subtracted

and that the total energy of a given Fock state is non-negative.

As a ﬁnal remark in this section, let us note that the term “canonical quantisation” refers

to the imposition of commutation relations among ﬁeld operators. An alternative method

of quantisation, which is actually more ﬂexible, is based on the path integral formulation.

This approach is widely used for the quantisation of non-Abelian gauge theories, but a

detailed discussion is beyond the scope of these lectures. More details can be found in

standard textbooks on Quantum Field Theory.

Problems

3.1 Starting from the deﬁnition of the Hamiltonian, H(x, p) ≡ p ˙ x − L(x, ˙ x), derive

Hamilton’s equations

∂H

∂x

= −˙ p,

∂H

∂p

= ˙ x

3.2 Show that the Hamiltonian density H for a free scalar ﬁeld is given by

H =

1

2

¸

(∂

0

φ)

2

+ (∇φ)

2

+ m

2

φ

2

¸

.

Use this result to express the Hamiltonian

ˆ

H =

1

2

d

3

x

∂

0

ˆ

φ)

2

+ (∇

ˆ

φ)

2

+ m

2

ˆ

φ

2

¸

of the quantised theory in terms of creation and annihilation operators and show

that it is given by

ˆ

H =

1

2

d

3

p

(2π)

3

2E(p)

E(p)

¸

ˆ a

†

(p)ˆ a(p) + ˆ a(p)ˆ a

†

(p)

¸

3.3 Verify eq. (3.50).

25

3.4 Using the expressions for

ˆ

φ and ˆ π in terms of ˆ a and ˆ a

†

, show that the unequal time

commutator

ˆ

φ(x), ˆ π(x

)

is given by

ˆ

φ(x), ˆ π(x

)

=

i

2

d

3

p

(2π)

3

e

ip·(x−x

)

+ e

−ip·(x−x

)

Show that for t = t

one recovers the equal time commutator

ˆ

φ(x, t), ˆ π(x

, t)

= iδ

3

(x −x

)

4 Interacting scalar ﬁelds

From now on we shall always discuss quantised real scalar ﬁelds. It is then convenient to

drop the “hats” on the operators that we have considered up to now

So far we have only discussed free ﬁelds without any interaction between them. As

this does not make for a very interesting theory, let us now add an interaction Lagrangian

L

int

. The full Lagrangian L is given by

L = L

0

+L

int

(4.1)

where

L

0

=

1

2

∂

µ

φ∂

µ

φ −

1

2

m

2

φ

2

(4.2)

is the free Lagrangian density discussed before. The Hamiltonian density of the interaction

is related to L

int

simply by

H

int

= −L

int

, (4.3)

which follows from its deﬁnition. We shall leave the details of L

int

unspeciﬁed for the

moment. What we will be concerned with mostly are scattering processes, in which two

initial particles with momenta p

1

and p

2

scatter, thereby producing a number of particles

in the ﬁnal state, characterised by momenta k

1

, . . . , k

n

. This is schematically shown in

Fig. 2. Our task is to ﬁnd a description of such a scattering process in terms of the

underlying quantum ﬁeld theory.

p

1

p

2

k

1

k

2

k

n

Figure 2: Scattering of two initial particles with momenta p

1

and p

2

into n particles with

momenta k

1

, . . . , k

n

in the ﬁnal state.

26

4.1 The S-matrix

The timescales over which interactions happen are extremely short. The scattering (in-

teraction) process takes place during a short interval around some particular time t with

−∞ < t < ∞. Long before t, the incoming particles evolve independently and freely.

They are described by a ﬁeld operator φ

in

deﬁned through

lim

t→−∞

φ(x) = φ

in

(x). (4.4)

Long after the collision the particles in the ﬁnal state evolve again like in the free theory,

and the corresponding operator is

lim

t→+∞

φ(x) = φ

out

(x). (4.5)

The ﬁelds φ

in

, φ

out

are the asymptotic limits of the Heisenberg operator φ. They both

satisfy the free Klein-Gordon equation, i.e.

(+ m

2

)φ

in

(x) = 0, (+ m

2

)φ

out

(x) = 0. (4.6)

Operators describing free ﬁelds can be expressed as a superposition of plane waves (see

eq. (2.28)). Thus, for φ

in

we have

φ

in

(x) =

d

3

k

(2π)

3

2E(k)

e

ik·x

ˆ a

†

in

(k) + e

−ik·x

ˆ a

in

(k)

, (4.7)

with an entirely analogous expression for φ

out

(x). Note that the operators ˆ a

†

and ˆ a also

carry subscripts “in” and “out”.

We can now use the creation operators ˆ a

†

in

and ˆ a

†

out

to build up Fock states from the

vacuum. For instance

ˆ a

†

in

(p

1

) ˆ a

†

in

(p

2

)[0` = [p

1

, p

2

; in`, (4.8)

ˆ a

†

out

(k

1

) ˆ a

†

out

(k

n

)[0` = [k

1

, . . . , k

n

; out`. (4.9)

We must now distinguish between Fock states generated by ˆ a

†

in

and ˆ a

†

out

, and therefore we

have labelled the Fock states accordingly. In eqs. (4.8) and (4.9) we have assumed that

there is a stable and unique vacuum state:

[0` = [0; in` = [0; out`. (4.10)

Mathematically speaking, the ˆ a

†

in

’s and ˆ a

†

out

’s generate two diﬀerent bases of the Fock

space. Since the physics that we want to describe must be independent of the choice of

basis, expectation values expressed in terms of “in” and “out” operators and states must

satisfy

'in[ φ

in

(x) [in` = 'out[ φ

out

(x) [out` . (4.11)

Here [in` and [out` denote generic “in” and “out” states. We can relate the two bases by

introducing a unitary operator S such that

φ

in

(x) = S φ

out

(x) S

†

(4.12)

[in` = S [out` , [out` = S

†

[in` , S

†

S = 1. (4.13)

27

S is called the S-matrix or S-operator. Note that the plane wave solutions of φ

in

and φ

out

also imply that

ˆ a

†

in

= S ˆ a

†

out

S

†

, ˆ a

in

= S ˆ a

out

S

†

. (4.14)

By comparing “in” with “out” states one can extract information about the interaction –

this is the very essence of detector experiments, where one tries to infer the nature of the

interaction by studying the products of the scattering of particles that have been collided

with known energies. As we will see below, this information is contained in the elements

of the S-matrix.

By contrast, in the absence of any interaction, i.e. for L

int

= 0 the distinction between

φ

in

and φ

out

is not necessary. They can thus be identiﬁed, and then the relation between

diﬀerent bases of the Fock space becomes trivial, S = 1, as one would expect.

What we are ultimately interested in are transition amplitudes between an initial

state i of, say, two particles of momenta p

1

, p

2

, and a ﬁnal state f, for instance n particles

of unequal momenta. Obviously the initial state is characterised by an “in” state, whereas

the ﬁnal one must be an “out” state. The transition amplitude is then given by

'f, out[ i, in` = 'f, out[ S [i, out` = 'f, in[ S [i, in` ≡ S

ﬁ

. (4.15)

The S-matrix element S

ﬁ

therefore describes the transition amplitude for the scattering

process in question. The scattering cross section, which is a measurable quantity is then

proportional to [S

ﬁ

[

2

. All information about the scattering is thus encoded in the S-

matrix, which must therefore be closely related to the interaction Hamiltonian density

H

int

. However, before we try to derive the relation between S and H

int

we have to take a

slight detour.

4.2 More on time evolution: Dirac picture

The operators φ(x, t) and π(x, t) which we have encountered are Heisenberg ﬁelds and

thus time-dependent. The state vectors are time-independent in the sense that they do

not satisfy a non-trivial equation of motion. Nevertheless, state vectors in the Heisenberg

picture can carry a time label. For instance, the “in”-states of the previous subsection are

deﬁned at t = −∞. The relation of the Heisenberg operator φ

H

(x) with its counterpart

φ

S

in the Schr¨ odinger picture is given by

φ

H

(x, t) = e

iHt

φ

S

e

−iHt

, H = H

0

+ H

int

, (4.16)

Note that this relation involves the full Hamiltonian H = H

0

+ H

int

in the interacting

theory. We have already found solutions to the Klein-Gordon equation in the free theory,

and so we know how to handle time evolution in this case. It now turns out to be useful to

introduce a new quantum picture for the interacting theory, in which the time dependence

is governed by H

0

only. This is the so-called Dirac or Interaction picture. The relation

between ﬁelds in the Interaction picture, φ

I

, and in the Schr¨ odinger picture, φ

S

, is given

by

φ

I

(x, t) = e

iH

0

t

φ

S

e

−iH

0

t

. (4.17)

At t = −∞ the interaction vanishes, i.e. H

int

= 0, and hence the ﬁelds in the Interaction

and Heisenberg pictures are identical, i.e. φ

H

(x, t) = φ

I

(x, t) for t → −∞. The relation

28

between φ

H

and φ

I

can be worked out easily:

φ

H

(x, t) = e

iHt

φ

S

e

−iHt

= e

iHt

e

−iH

0

t

e

iH

0

t

φ

S

e

−iH

0

t

. .. .

φ

I

(x,t)

e

iH

0

t

e

iH

0

t

e

−iHt

= U

−1

(t) φ

I

(x, t) U(t), (4.18)

where we have introduced the unitary operator U(t)

U(t) = e

iH

0

t

e

−iHt

, U

†

U = 1. (4.19)

The ﬁeld φ

H

(x, t) contains the information about the interaction, since it evolves over

time with the full Hamiltonian. In order to describe the “in” and “out” ﬁeld operators,

we can now make the following identiﬁcations:

t →−∞ : φ

in

(x, t) = φ

I

(x, t) ≡ φ

H

(x, t), (4.20)

t →+∞ : φ

out

(x, t) = φ

H

(x, t). (4.21)

Furthermore, since the ﬁelds φ

I

evolve over time with the free Hamiltonian H

0

, they

always act in the basis of “in” vectors, such that

φ

in

(x, t) = φ

I

(x, t), −∞< t < ∞. (4.22)

The relation between φ

I

and φ

H

at any time t is given by

φ

I

(x, t) = U(t) φ

H

(x, t) U

−1

(t). (4.23)

As t →∞ the identiﬁcations of eqs. (4.21) and (4.22) yield

φ

in

= U(∞) φ

out

U

†

(∞). (4.24)

From the deﬁnition of the S-matrix, eq. (4.12) we then read oﬀ that

lim

t→∞

U(t) = S. (4.25)

We have thus derived a formal expression for the S-matrix in terms of the operator U(t),

which tells us how operators and state vectors deviate from the free theory at time t,

measured relative to t

0

= −∞, i.e. long before the interaction process.

An important boundary condition for U(t) is

lim

t→−∞

U(t) = 1. (4.26)

What we mean here is the following: the operator U actually describes the evolution

relative to some initial time t

0

, which we will normally suppress, i.e. we write U(t)

instead of U(t, t

0

). We regard t

0

merely as a time label and ﬁx it at −∞, where the

interaction vanishes. Equation (4.26) then simply states that U becomes unity as t →t

0

,

which means that in this limit there is no distinction between Heisenberg and Dirac ﬁelds.

29

Using the deﬁnition of U(t), eq. (4.19), it is an easy exercise to derive the equation

of motion for U(t):

i

d

dt

U(t) = H

int

(t) U(t), H

int

(t) = e

iH

0

t

H

int

e

−iH

0

t

. (4.27)

The time-dependent operator H

int

(t) is deﬁned in the interaction picture, and depends

on the ﬁelds φ

in

, π

in

in the “in” basis. Let us now solve the equation of motion for U(t)

with the boundary condition lim

t→−∞

U(t) = 1. Integrating eq. (4.27) gives

t

−∞

d

dt

1

U(t

1

) dt

1

= −i

t

−∞

H

int

(t

1

) U(t

1

) dt

1

U(t) −U(−∞) = −i

t

−∞

H

int

(t

1

) U(t

1

) dt

1

⇒ U(t) = 1 −i

t

−∞

H

int

(t

1

) U(t

1

) dt

1

. (4.28)

The rhs. still depends on U, but we can substitute our new expression for U(t) into the

integrand, which gives

U(t) = 1 −i

t

−∞

H

int

(t

1

)

1 −i

t

1

−∞

H

int

(t

2

) U(t

2

) dt

2

dt

1

= 1 −i

t

−∞

H

int

(t

1

)dt

1

−

t

−∞

dt

1

H

int

(t

1

)

t

1

−∞

dt

2

H

int

(t

2

) U(t

2

), (4.29)

where t

2

< t

1

< t. This procedure can be iterated further, so that the nth term in the

sum is

(−i)

n

t

−∞

dt

1

t

1

−∞

dt

2

t

n−1

−∞

dt

n

H

int

(t

1

) H

int

(t

2

) H

int

(t

n

). (4.30)

This iterative solution could be written in much more compact form, were it not for the

fact that the upper integration bounds were all diﬀerent, and that the ordering t

n

<

t

n−1

< . . . < t

1

< t had to be obeyed. Time ordering is an important issue, since one

has to ensure that the interaction Hamiltonians act at the proper time, thereby ensuring

the causality of the theory. By introducing the time-ordered product of operators, one

can use a compact notation, such that the resulting expressions still obey causality. The

time-ordered product of two ﬁelds φ(t

1

) and φ(t

2

) is deﬁned as

T ¦φ(t

1

) φ(t

2

)¦ =

φ(t

1

)φ(t

2

) t

1

> t

2

φ(t

2

)φ(t

1

) t

1

< t

2

≡ θ(t

1

−t

2

) φ(t

1

)φ(t

2

) + θ(t

2

−t

1

) φ(t

2

)φ(t

1

), (4.31)

where θ denotes the step function. The generalisation to products of n operators is

obvious. Using time ordering for the nth term of eq. (4.30) we obtain

(−i)

n

n!

n

¸

i=1

t

−∞

dt

i

T ¦H

int

(t

1

) H

int

(t

2

) H

int

(t

n

)¦ , (4.32)

30

and since this looks like the nth term in the series expansion of an exponential, we can

ﬁnally rewrite the solution for U(t) in compact form as

U(t) = T exp

−i

t

−∞

H

int

(t

) dt

, (4.33)

where the “T” in front ensures the correct time ordering.

4.3 S-matrix and Green’s functions

The S-matrix, which relates the “in” and “out” ﬁelds before and after the scattering

process, can be written as

S = 1 + iT, (4.34)

where T is commonly called the T-matrix. The fact that S contains the unit operator

means that also the case where none of the particles scatter is encoded in S. On the other

hand, the non-trivial case is described by the T-matrix, and this is what we are interested

in. Let us consider again the scattering process depicted in Fig. 2. The S-matrix element

in this case is

S

ﬁ

=

k

1

, k

2

, . . . , k

n

; out

p

1

, p

2

; in

=

k

1

, k

2

, . . . , k

n

; out

a

†

in

(p

1

)

p

2

; in

, (4.35)

where a

†

in

is the creation operator pertaining to the “in” ﬁeld φ

in

. Our task is now to

express a

†

in

in terms of φ

in

, and repeat this procedure for all other momenta labelling

our Fock states. As we shall see, this will relate the S-matrix element S

ﬁ

to the vacuum

expectation value of ﬁelds, i.e. a Green function, which can be computed in various ways.

The following identities will prove useful

ˆ a

†

(p) = i

d

3

x

¸

∂

0

e

−iq·x

φ(x) −e

−iq·x

(∂

0

φ(x))

¸

≡ −i

d

3

xe

−iq·x

←→

∂

0

φ(x), (4.36)

ˆ a(p) = −i

d

3

x

¸

∂

0

e

iq·x

φ(x) −e

iq·x

(∂

0

φ(x))

¸

≡ i

d

3

xe

iq·x

←→

∂

0

φ(x). (4.37)

The S-matrix element can then be rewritten as

S

ﬁ

= −i

d

3

x

1

e

−ip

1

·x

1

←→

∂

0

k

1

, . . . , k

n

; out

φ

in

(x

1

)

p

2

; in

= −i lim

t

1

→−∞

d

3

x

1

e

−ip

1

·x

1

←→

∂

0

k

1

, . . . , k

n

; out

φ(x

1

)

p

2

; in

, (4.38)

where in the last line we have used eq. (4.4) to replace φ

in

by φ. We can now rewrite

lim

t

1

→−∞

using the following identity, which holds for an arbitrary, diﬀerentiable function

f(t), whose limit t →±∞ exists:

lim

t→−∞

f(t) = lim

t→+∞

f(t) −

+∞

−∞

df

dt

dt. (4.39)

31

The S-matrix element then reads

S

ﬁ

= −i lim

t

1

→+∞

d

3

x

1

e

−ip

1

·x

1

←→

∂

0

k

1

, . . . , k

n

; out

φ(x

1

)

p

2

; in

+i

+∞

−∞

dt

1

∂

∂t

1

d

3

x

1

e

−ip

1

·x

1

←→

∂

0

k

1

, . . . , k

n

; out

φ(x

1

)

p

2

; in

.(4.40)

The ﬁrst term in this expression involves lim

t

1

→+∞

φ = φ

out

, which gives rise to a contri-

bution

∝

k

1

, . . . , k

n

; out

a

†

out

(p

1

)

p

2

; in

. (4.41)

This is non-zero only if p

1

is equal to one of k

1

, . . . , k

n

. This, however, means that the

particle with momentum p

1

does not scatter, and hence the ﬁrst term does not contribute

to the T-matrix of eq. (4.34). We are then left with the following expression for S

ﬁ

:

S

ﬁ

= −i

d

4

x

1

k

1

, . . . , k

n

; out

∂

0

¸

∂

0

e

−ip

1

·x

1

φ(x

1

) −e

−ip

1

·x

1

(∂

0

φ(x

1

))

¸

p

2

; in

.

(4.42)

The time derivatives in the integrand can be worked out:

∂

0

¸

∂

0

e

−ip

1

·x

1

φ(x

1

) −e

−ip

1

·x

1

(∂

0

φ(x

1

))

¸

= −

E(p

1

)

2

e

−ip

1

·x

1

φ(x

1

) −e

−ip

1

·x

1

∂

2

0

φ(x

1

)

= −

¸

−∇

2

+ m

2

e

−ip

1

·x

1

φ(x

1

) + e

−ip

1

·x

1

∂

2

0

φ(x

1

)

¸

, (4.43)

where we have used that −∇

2

e

−ip

1

·x

1

= p

2

1

e

−ip

1

·x

1

. For the S-matrix element one obtains

S

ﬁ

= i

d

4

x

1

e

−ip

1

·x

1

k

1

, . . . , k

n

; out

∂

2

0

−∇

2

+ m

2

φ(x

1

)

p

2

; in

= i

d

4

x

1

e

−ip

1

·x

1

x

1

+ m

2

k

1

, . . . , k

n

; out

φ(x

1

)

p

2

; in

. (4.44)

What we have obtained after this rather lengthy step of algebra is an expression in which

the ﬁeld operator is sandwiched between Fock states, one of which has been reduced to a

one-particle state. We can now successively eliminate all momentum variables from the

Fock states, by repeating the procedure for the momentum p

2

, as well as the n momenta

of the “out” state. The ﬁnal expression for S

ﬁ

is

S

ﬁ

= (i)

n+2

d

4

x

1

d

4

x

2

d

4

y

1

d

4

y

n

e

(−ip

1

·x

1

−ip

2

·x

2

+ik

1

·y

1

+···+kn·yn)

x

1

+ m

2

x

2

+ m

2

y

1

+ m

2

yn

+ m

2

0; out

T¦φ(y

1

) φ(y

n

)φ(x

1

)φ(x

2

)¦

0; in

, (4.45)

where the time-ordering inside the vacuum expectation value (VEV) ensures that causality

is obeyed. The above expression is known as the Lehmann-Symanzik-Zimmermann (LSZ)

reduction formula. It relates the formal deﬁnition of the scattering amplitude to a VEV of

time-ordered ﬁelds. The VEV in the LSZ formula for the scattering of two initial particles

into n particles in the ﬁnal state is called the (n + 2)-point Green’s function:

G

n+2

(y

1

, y

2

, . . . , y

n

, x

1

, x

2

) =

0; out

T¦φ(y

1

) φ(y

n

)φ(x

1

)φ(x

2

)¦

0; in

. (4.46)

32

In the next section we shall discuss a method how to compute the Green’s function of

scalar ﬁeld theory. This will allow for an explicit calculation of G

n+2

. The LSZ formula

then provides the link to the scattering amplitude S

ﬁ

, and [S

ﬁ

[

2

ﬁnally yields the cross

section, which can be measured in an experiment.

Before we can tackle the actual computation of Green’s function, we must take a

further step. Let us consider the n-point Green’s function

G

n

(x

1

, . . . , x

n

) = '0 [T¦φ(x

1

) φ(x

n

)¦[ 0` . (4.47)

The ﬁelds φ which appear in this expression are Heisenberg ﬁelds, whose time evolution

is governed by the full Hamiltonian H

0

+ H

int

. In particular, the φ’s are not the φ

in

’s,

which we know how to handle. In order to be useful for actual computations, we ﬁrst

have to isolate the dependence of the ﬁelds on the interaction Hamiltonian. Let us recall

the relation between the Heisenberg ﬁelds φ(t) and the “in”-ﬁelds

2

φ(t) = U

−1

(t) φ

in

(t) U(t). (4.48)

We now assume that the ﬁelds are properly time-ordered, i.e. t

1

> t

2

> . . . > t

n

, so

that we can forget about writing T( ) everywhere. After inserting eq. (4.48) into the

deﬁnition of G

n

one obtains

G

n

= '0[U

−1

(t

1

)φ

in

(t

1

)U(t

1

) U

−1

(t

2

)φ

in

(t

2

)U(t

2

)

U

−1

(t

n

)φ

in

(t

n

)U(t

n

)[0`. (4.49)

Now we introduce another time label t such that t t

1

and −t < t

1

. For the n-point

function we now obtain

G

n

=

0

U

−1

(t)

U(t)U

−1

(t

1

)φ

in

(t

1

)U(t

1

) U

−1

(t

2

)φ

in

(t

2

)U(t

2

)

U

−1

(t

n

)φ

in

(t

n

)U(t

n

)U

−1

(−t)

¸

U(−t)

0

. (4.50)

The expression in curly braces is now time-ordered by construction. An important obser-

vation at this point is that it involves pairs of U and its inverse, for instance

U(t)U

−1

(t

1

) ≡ U(t, t

1

). (4.51)

One can easily convince oneself that U(t, t

1

) provides the net time evolution from t

1

to t.

We can now write G

n

as

G

n

=

0

U

−1

(t) T

φ

in

(t

1

) φ

in

(t

n

) U(t, t

1

) U(t

1

, t

2

) U(t

n

, −t)

. .. .

U(t, −t)

¸

U(−t)

0

. (4.52)

Let us now take t → ∞. The relation between U(t) and the S-matrix eq. (4.25), as well

as the boundary condition eq. (4.26) tell us that

lim

t→∞

U(−t) = 1, lim

t→∞

U(t, −t) = S, (4.53)

2

Here and in the following we suppress the spatial argument of the ﬁelds for the sake of brevity.

33

which can be inserted into the above expression. We still have to work out the meaning

of '0[U

−1

(∞) in the expression for G

n

. In a paper by Gell-Mann and Low it was argued

that the time evolution operator must leave the vacuum invariant (up to a phase), which

justiﬁes the ansatz

'0[U

−1

(∞) = K'0[, (4.54)

with K being the phase. Multiplying this relation with [0` from the right gives

'0[U

−1

(∞)[0` = K'0[0` = K. (4.55)

Furthermore, Gell-Mann and Low showed that

'0[U

−1

(∞)[0` =

1

'0[U(∞)[0`

, (4.56)

which implies

K =

1

'0[S[0`

. (4.57)

After inserting all these relations into the expression for G

n

we obtain

G

n

(x

1

, . . . , x

n

) =

'0[ T ¦φ

in

(x

1

) φ

in

(x

n

) S¦ [0`

'0[S[0`

. (4.58)

The S-matrix is given by

S = T exp

−i

+∞

−∞

H

int

(t) dt

, H

int

= H

int

(φ

in

, π

in

), (4.59)

and thus we have ﬁnally succeeded in expressing the n-point Green’s function exclusively

in terms of the “in”-ﬁelds. This completes the derivation of a relation between the general

deﬁnition of the scattering amplitude S

ﬁ

and the VEV of time-ordered “in”-ﬁelds. The

link between the scattering amplitude and the underlying ﬁeld theory is provided by the

n-point Green’s function.

Problems

4.1 Using the deﬁnition U(t) = e

iH

0

t

e

−iHt

derive the equation of motion for U(t),

eq. (4.27).

4.2 Derive eqs. (4.36) and (4.37).

5 Perturbation Theory

In this section we are going to calculate the Green’s functions of scalar quantum ﬁeld

theory explicitly. We will specify the interaction Lagrangian in detail and use an approx-

imation known as perturbation theory. At the end we will derive a set of rules, which

34

represent a systematic prescription for the calculation of Green’s functions, and can be

easily generalised to apply to other, more complicated ﬁeld theories. These are the famous

Feynman rules.

We start by making a deﬁnite choice for the interaction Lagrangian L

int

. Although

one may think of many diﬀerent expressions for L

int

, one has to obey some basic principles:

ﬁrstly, L

int

must be chosen such that the potential it generates is bounded from below

– otherwise the system has no ground state. Secondly, our interacting theory should be

renormalisable. Despite being of great importance, the second issue will not be addressed

in these lectures. The requirement of renormalisability arises because if one computes

quantities like the energy or charge of a particle, one typically obtains a divergent result

3

.

There are classes of quantum ﬁeld theories, called renormalisable, in which these diver-

gences can be removed by suitable redeﬁnitions of the ﬁelds and the parameters (masses

and coupling constants).

For our theory of a real scalar ﬁeld in four space-time dimensions, it turns out that

the only interaction term which leads to a renormalisable theory must be quartic in the

ﬁelds. Thus we choose

L

int

= −

λ

4!

φ

4

(x), (5.1)

where the coupling constant λ describes the strength of the interaction between the scalar

ﬁelds, much like, say, the electric charge describing the strength of the interaction between

photons and electrons. The full Lagrangian of the theory then reads

L = L

0

+L

int

=

1

2

(∂

µ

φ)

2

−

1

2

m

2

φ

2

−

λ

4!

φ

4

, (5.2)

and the explicit expressions for the interaction Hamiltonian and the S-matrix are

H

int

= −L

int

, H

int

=

λ

4!

d

3

xφ

4

in

(x, t)

S = T exp

−i

λ

4!

d

4

xφ

4

in

(x)

. (5.3)

The n-point Green’s function is

G

n

(x

1

, . . . , x

n

)

=

∞

¸

r=0

−

iλ

4!

r

1

r!

0

T

φ

in

(x

1

) φ

in

(x

n

)

d

4

y φ

4

in

(y)

r

0

∞

¸

r=0

−

iλ

4!

r

1

r!

0

T

d

4

y φ

4

in

(y)

r

0

. (5.4)

In order to evaluate this expression we must expand G

n

in powers of the coupling λ. This

only makes sense if λ is suﬃciently small. In other words, the interaction Lagrangian must

act as a small perturbation on the system. As a consequence, the procedure of expanding

Green’s functions in powers of the coupling is referred to as perturbation theory.

3

This is despite the subtraction of the vacuum energy discussed in section 3.

35

5.1 Wick’s Theorem

The n-point Green’s function in eq. (5.4) involves the time-ordered product over at least

n ﬁelds. There is a method to express VEV’s of n ﬁelds, i.e. '0[T ¦φ

in

(x

1

) φ

in

(x

n

)¦ [0`

in terms of VEV’s involving two ﬁelds only. This is known as Wick’s theorem.

Let us for the moment ignore the subscript “in” and return to the deﬁnition of

normal-ordered ﬁelds. The normal-ordered product : φ(x

1

)φ(x

2

) : diﬀers from φ(x

1

)φ(x

2

)

by the vacuum expectation value, i.e.

φ(x

1

)φ(x

2

) = : φ(x

1

)φ(x

2

) : +'0[φ(x

1

)φ(x

2

)[0`. (5.5)

We are now going to combine normal-ordered products with time ordering. The time-

ordered product T¦φ(x

1

)φ(x

2

)¦ is given by

T¦φ(x

1

)φ(x

2

)¦ = φ(x

1

)φ(x

2

)θ(t

1

−t

2

) + φ(x

2

)φ(x

1

)θ(t

2

−t

1

)

= : φ(x

1

)φ(x

2

) :

θ(t

1

−t

2

) + θ(t

2

−t

1

)

+'0[φ(x

1

)φ(x

2

)θ(t

1

−t

2

) + φ(x

2

)φ(x

1

)θ(t

2

−t

1

)[0`. (5.6)

Here we have used the important observation that

: φ(x

1

)φ(x

2

) : = : φ(x

2

)φ(x

1

) :, (5.7)

which means that normal-ordered products of ﬁelds are automatically time-ordered.

4

Equation (5.6) is Wick’s theorem for the case of two ﬁelds:

T¦φ(x

1

)φ(x

2

)¦ = : φ(x

1

)φ(x

2

) : +'0[T ¦φ(x

1

)φ(x

2

)¦ [0`. (5.8)

For the case of three ﬁelds, Wick’s theorem yields

T¦φ(x

1

)φ(x

2

)φ(x

3

)¦ = : φ(x

1

)φ(x

2

)φ(x

3

) : + : φ(x

1

) : '0[T¦φ(x

2

)φ(x

3

)¦[0`

+ : φ(x

2

) : '0[T¦φ(x

1

)φ(x

3

)¦[0`+ : φ(x

3

) : '0[T¦φ(x

1

)φ(x

2

)¦[0` (5.9)

At this point the general pattern becomes clear: any time-ordered product of ﬁelds is

equal to its normal-ordered version plus terms in which pairs of ﬁelds are removed from the

normal-ordered product and sandwiched between the vacuum to form 2-point functions.

Then one sums over all permutations. Without proof we give the expression for the

general case of n ﬁelds (n even):

T¦φ(x

1

) φ(x

n

)¦ =

: φ(x

1

) φ(x

n

) :

+ : φ(x

1

)

φ(x

i

)

φ(x

j

) φ(x

n

) : '0[T¦φ(x

i

)φ(x

j

)¦[0` + perms.

+ : φ(x

1

)

φ(x

i

)

φ(x

j

)

φ(x

k

)

φ(x

l

) φ(x

n

) :

'0[T¦φ(x

i

)φ(x

j

)¦[0`'0[T¦φ(x

k

)φ(x

l

)¦[0` + perms.

+. . . +

+'0[T¦φ(x

1

)φ(x

2

)¦[0`'0[T¦φ(x

3

)φ(x

4

)¦[0` '0[T¦φ(x

n−1

)φ(x

n

)¦[0`

+ perms.. (5.10)

4

The reverse is, however, not true!

36

The symbol

φ(x

i

) indicates that φ(x

i

) has been removed from the normal-ordered product.

Let us now go back to '0[T¦φ(x

1

) φ(x

n

)¦[0`. If we insert Wick’s theorem, then we

ﬁnd that only the contribution in the last line of eq. (5.10) survives: by deﬁnition the VEV

of a normal-ordered product of ﬁelds vanishes, and it is precisely the last line of Wick’s

theorem in which no normal-ordered products are left. The only surviving contribution

is that in which all ﬁelds have been paired or “contracted”. Sometimes a contraction is

represented by the notation:

φ(x

i

)φ

. .. .

(x

j

) ≡ '0[T¦φ(x

i

)φ(x

j

)¦[0`, (5.11)

i.e. the pair of ﬁelds which is contracted is joined by the braces. Wick’s theorem can now

be rephrased as

'0[T¦φ(x

1

) φ(x

n

)¦[0` = sum of all possible contractions of n ﬁelds. (5.12)

Let us look at a few examples. The ﬁrst is the 4-point function

'0[T¦φ(x

1

)φ(x

2

)φ(x

3

)φ(x

4

)¦[0` = φ(x

1

)φ

. .. .

(x

2

)φ(x

3

)φ

. .. .

(x

4

)

+φ(x

1

)φ

. .. .

(x

2

)φ(x

3

)φ

. .. .

(x

4

) + φ(x

1

)φ

. .. .

(x

2

)φ(x

3

)φ(x

4

)

. .. .

. (5.13)

The second example is again a 4-point function, where two of the ﬁelds are also normal-

ordered:

'0[T¦φ(x

1

)φ(x

2

) : φ(x

3

)φ(x

4

) :¦[0` = φ(x

1

)φ

. .. .

(x

2

) : φ(x

3

)φ

. .. .

(x

4

) :

+φ(x

1

)φ

. .. .

(x

2

) : φ(x

3

)φ

. .. .

(x

4

) : +φ(x

1

)φ

. .. .

(x

2

) : φ(x

3

)φ(x

4

) :

. .. .

. (5.14)

In this example, though, the contraction of : φ(x

3

)φ(x

4

) : vanishes by construction, so only

the last two terms survive! As a general rule, contractions which only involve ﬁelds inside a

normal-ordered product vanish. Such contractions contribute only to the vacuum. Normal

ordering can therefore simplify the calculation of Green’s functions quite considerably, as

we shall see explicitly below.

5.2 The Feynman propagator

Using Wick’s Theorem one can relate any n-point Green’s functions to an expression

involving only 2-point functions. Let us have a closer look at

G

2

(x, y) = '0[T¦φ

in

(x)φ

in

(y)¦[0`. (5.15)

We can now insert the solution for φ in terms of ˆ a and ˆ a

†

. If we assume t

x

> t

y

then

G

2

(x, y) can be written as

G

2

(x, y) =

d

3

p d

3

q

(2π)

6

4E(p)E(q)

0

ˆ a

†

(p) e

ip·x

+ ˆ a(p) e

−ip·x

ˆ a

†

(q) e

iq·y

+ ˆ a(q) e

−iq·y

0

=

d

3

p d

3

q

(2π)

6

4E(p)E(q)

e

−ip·x+iq·y

0

ˆ a(p)ˆ a

†

(q)

0

. (5.16)

37

p

1

p

2

k

1

k

2

Figure 3: Scattering of two initial particles with momenta p

1

and p

2

into 2 particles with

momenta k

1

and k

2

.

This shows that G

2

can be interpreted as the amplitude for a meson which is created at

y and destroyed again at point x. We can now replace ˆ a(p)ˆ a

†

(q) by its commutator:

G

2

(x, y) =

d

3

p d

3

q

(2π)

6

4E(p)E(q)

e

−ip·x+iq·y

0

ˆ a(p), ˆ a

†

(q)

0

=

d

3

p

(2π)

3

2E(p)

e

−ip·(x−y)

, (5.17)

and the general result, after restoring time-ordering, reads

G

2

(x, y) =

d

3

p

(2π)

3

2E(p)

e

−ip·(x−y)

θ(t

x

−t

y

) + e

ip·(x−y)

θ(t

y

−t

x

)

. (5.18)

Furthermore, using contour integration one can show that this expression can be rewritten

as a 4-dimensional integral

G

2

(x, y) = i

d

4

p

(2π)

4

e

ip·(x−y)

p

2

−m

2

+ i

, (5.19)

where is a small parameter which ensures that G

2

does not develop a pole. This cal-

culation has established that G

2

(x, y) actually depends only on the diﬀerence (x − y).

Equation (5.19) is called the Feynman propagator G

F

(x −y):

G

F

(x −y) ≡ '0[T¦φ(x)φ(y)¦[0` = i

d

4

p

(2π)

4

e

ip·(x−y)

p

2

−m

2

+ i

. (5.20)

The Feynman propagator is a Green’s function of the Klein-Gordon equation, i.e. it

satisﬁes

x

+ m

2

G

F

(x −y) = −iδ

4

(x −y), (5.21)

and describes the propagation of a meson between the space-time points x and y.

5.3 Two-particle scattering to O(λ)

Let us now consider a scattering process in which two incoming particles with momenta

p

1

and p

2

scatter into two outgoing ones with momenta k

1

and k

2

, as shown in Fig. 3.

The S-matrix element in this case is

S

ﬁ

= 'k

1

, k

2

; out[p

1

, p

2

; in`

= 'k

1

, k

2

; in[S[p

1

, p

2

; in`, (5.22)

38

and S = 1 + iT. The LSZ formula eq. (4.45) tells us that we must compute G

4

in order

to obtain S

ﬁ

. Let us work out G

4

in powers of λ using Wick’s theorem. To make life

simpler, we shall introduce normal ordering into the deﬁnition of S, i.e.

S = T exp

−i

λ

4!

d

4

x : φ

4

in

(x) :

(5.23)

Suppressing the subscripts “in” from now on, the expression we have to evaluate order by

order in λ is

G

n

(x

1

, . . . , x

n

) (5.24)

=

∞

¸

r=0

−

iλ

4!

r

1

r!

0

T

φ(x

1

)φ(x

2

)φ(x

3

)φ(x

4

)

d

4

y : φ

4

(y) :

r

0

∞

¸

r=0

−

iλ

4!

r

1

r!

0

T

d

4

y : φ

4

(y) :

r

0

.

Starting with the denominator, we note that for r = 0 one ﬁnds

r = 0 : denominator = 1. (5.25)

If r = 1, then the expression in the denominator only involves ﬁelds which are normal-

ordered. Following the discussion at the end of section 5.1 we conclude that these contri-

butions must vanish, hence

r = 1 : denominator = 0. (5.26)

The contribution for r = 2, however, is non-zero. But then the case of r = 2 corresponds

already to O(λ

2

), which is higher than the order which we are working to. Therefore

denominator = 1 to order λ. (5.27)

Turning now to the numerator, we start with r = 0 and apply Wick’s theorem, which

gives

r = 0 : '0[T¦φ(x

1

)φ(x

2

)φ(x

3

)φ(x

4

)¦[0`

= G

F

(x

1

−x

2

) G

F

(x

3

−x

4

) + G

F

(x

1

−x

3

) G

F

(x

2

−x

4

)

+ G

F

(x

1

−x

4

) G

F

(x

2

−x

3

), (5.28)

which can be graphically represented as

x

3

x

4

x

1

x

2

+

x

3

x

4

x

1

x

2

+

x

3

x

4

x

1

x

2

But this is the same answer as if we had set λ = 0, so r = 0 in the numerator does not

describe scattering and is hence not a contribution to the T-matrix.

39

For r = 1 in the numerator we have to evaluate

r = 1 : −

iλ

4!

0

T

φ(x

1

)φ(x

2

)φ(x

3

)φ(x

4

) :

d

4

y φ

4

(y) :

0

= −

iλ

4!

d

4

y 4! G

F

(x

1

−y)G

F

(x

2

−y)G

F

(x

3

−y)G

F

(x

4

−y), (5.29)

where we have taken into account that contractions involving two ﬁelds inside : : vanish.

The factor 4! inside the integrand is a combinatorial factor: it is equal to the number of

permutations which must be summed over according to Wick’s theorem and cancels the

4! in the denominator of the interaction Lagrangian. Graphically this contribution is

represented by

x

3

x

4

x

1

x

2

y

−iλ

d

4

y

where the integration over y denotes the sum over all possible locations of the interaction

point y. Without normal ordering we would have encountered the following contributions

for r = 1:

x

3

x

4

x

1

x

2

+

x

3

x

4

x

1

x

2

+. . .

Such contributions are corrections to the vacuum and are cancelled by the denomina-

tor. This demonstrates how normal ordering simpliﬁes the calculation by automatically

subtracting terms which do not contribute to the actual scattering process.

To summarise, the ﬁnal answer for the scattering amplitude to O(λ) is given by

eq. (5.29).

5.4 Graphical representation of the Wick expansion: Feynman rules

We have already encountered the graphical representation of the expansion of Green’s

functions in perturbation theory after applying Wick’s theorem. It is possible to formulate

a simple set of rules which allow to draw the graphs directly without using Wick’s theorem

and to write down the corresponding algebraic expressions.

We again consider a neutral scalar ﬁeld whose Lagrangian is

L =

1

2

∂

µ

φ∂

µ

φ −

1

2

m

2

φ

2

−

λ

4!

φ

4

. (5.30)

Suppose now that we want to compute the O(λ

m

) contribution to the n-point Green’s

function G

n

(x

1

, . . . , x

n

). This is achieved by going through the following steps:

(1) Draw all distinct diagrams with n external lines and m 4-fold vertices:

40

• Draw n dots and label them x

1

, . . . , x

n

(external points)

• Draw m dots and label them y

1

, . . . , y

m

(vertices)

• Join the dots according to the following rules:

– only one line emanates from each x

i

– exactly four lines run into each y

j

– the resulting diagram must be connected, i.e. there must be a continuous

path between any two points.

(2) Assign a factor −

iλ

4!

d

4

y

i

to the vertex at y

i

(3) Assign a factor G

F

(x

i

−y

j

) to the line joining x

i

and y

j

(4) Multiply by the number of contractions ( from the Wick expansion which lead to

the same diagram.

These are the Feynman rules for scalar ﬁeld theory in position space.

Let us look at an example, namely the 2-point function. According to the Feynman

rules the contributions up to order λ

2

are as follows:

O(1):

x

1

x

2

= G

F

(x

1

−x

2

)

O(λ):

x

1

x

2

y

“tadpole diagram”;

(cancelled by normal ordering)

O(λ

2

):

x

1

x

2

y

1

y

2

“tadpole diagram”;

(cancelled by normal ordering)

O(λ

2

):

x

1

x

2

y

1

y

2

= (

−

iλ

4!

2

d

4

y

1

d

4

y

2

G

F

(x

1

−y

1

) [G

F

(y

1

−y

2

)]

3

G

F

(y

2

−x

2

)

The combinatorial factor for this contribution is worked out as ( = 4 4!. Note that

the same graph, but with the positions of y

1

and y

2

interchanged is topologically distinct.

Numerically it has the same value as the above graph, and so the corresponding expression

has to be multiplied by a factor 2.

Another contribution at order λ

2

is

41

O(λ

2

):

x

1

x

2

y

1

y

2

vacuum contribution;

not connected

This contribution must be discarded, since not all of the points are connected via a

continuous line.

Let us end this discussion with a small remark on the tadpole diagrams encountered

above. These contributions to the 2-point function are cancelled if the interaction term

is normal-ordered. However, unlike the case of the 4-point function, the corresponding

diagrams satisfy the Feynman rules listed above. In particular, the diagrams are connected

and are not simply vacuum contributions. They must hence be included in the expression

for the 2-point function.

5.5 Feynman rules in momentum space

It is often simpler to work in momentum space, and hence we will discuss the derivation

of Feynman rules in this case. If one works in momentum space, the Green’s functions

are related to those in position space by a Fourier transform

G

n

(x

1

, . . . , x

n

) =

d

4

p

1

(2π)

4

d

4

p

n

(2π)

4

e

ip

1

·x

1

+...+ipn·xn ¯

G

n

(p

1

, . . . , p

n

). (5.31)

The Feynman rules then serve to compute the Green’s function

¯

G

n

(p

1

, . . . , p

n

) order by

order in the coupling.

In every scattering process the overall momentum must be conserved, and hence

n

¸

i=1

p

i

= 0. (5.32)

This can be incorporated into the deﬁnition of the momentum space Green’s function one

is trying to compute:

¯

G

n

(p

1

, . . . , p

n

) = (2π)

4

δ

4

n

¸

i=1

p

i

(

n

(p

1

, . . . , p

n

). (5.33)

Here we won’t be concerned with the exact derivation of the momentum space Feynman

rules, but only list them as a recipe.

Feynman rules (momentum space)

(1) Draw all distinct diagrams with n external lines and m 4-fold vertices:

• Assign momenta p

1

, . . . , p

n

to the external lines

• Assign momenta k

j

to the internal lines

42

(2) Assign to each external line a factor

i

p

2

k

−m

2

+ i

(3) Assign to each internal line a factor

d

4

k

j

(2π)

4

i

k

2

j

−m

2

+ i

(4) Each vertex contributes a factor

−

iλ

4!

(2π)

4

δ

4

¸

momenta

,

(the delta function ensures that momentum is conserved at each vertex).

(5) Multiply by the combinatorial factor (, which is the number of contractions leading

to the same momentum space diagram (note that ( may be diﬀerent from the

combinatorial factor for the same diagram considered in position space!)

5.6 S-matrix and truncated Green’s functions

The ﬁnal topic in these lectures is the derivation of a simple relation between the S-

matrix element and a particular momentum space Green’s function, which has its external

legs amputated: the so-called truncated Green’s function. This further simpliﬁes the

calculation of scattering amplitudes using Feynman rules.

Let us return to the LSZ formalism and consider the scattering of m initial particles

(momenta p

1

, . . . , p

m

) into n ﬁnal particles with momenta k

1

, . . . , k

n

. The LSZ formula

tells us that the S-matrix element is given by

k

1

, . . . , k

n

; out

p

1

, . . . , p

m

; in

= (i)

n+m

m

¸

i=1

d

4

x

i

n

¸

j=1

d

4

y

j

exp

−i

m

¸

i=1

p

i

x

i

+ i

n

¸

j=1

k

j

y

j

¸

m

¸

i=1

x

i

+ m

2

n

¸

j=1

y

j

+ m

2

G

n+m

(x

1

, . . . , x

m

, y

1

, . . . , y

n

). (5.34)

Let us have a closer look at G

n+m

(x

1

, . . . , x

m

, y

1

, . . . , y

n

). As shown in Fig. 4 it can

be split into Feynman propagators, which connect the external points to the vertices at

z

1

, . . . , z

n+m

, and a remaining Green’s function G

n+m

, according to

G

n+m

=

d

4

z

1

d

4

z

n+m

G

F

(x

1

−z

1

) G

F

(y

n

−z

n+m

) G

n+m

(z

1

, . . . , z

n+m

), (5.35)

where, perhaps for obvious reasons, G

n+m

is called the truncated Green’s function.

Putting eq. (5.35) back into the LSZ expression for the S-matrix element, and using

that

x

i

+ m

2

G

F

(x

i

−z

i

) = −iδ

4

(x

i

−z

i

) (5.36)

43

x

1

x

2

x

3

G

z

1

z

2

z

3

Figure 4: The construction of the truncated Green’s function in position space.

one obtains

k

1

, . . . , k

n

; out

p

1

, . . . , p

m

; in

= (i)

n+m

m

¸

i=1

d

4

x

i

n

¸

j=1

d

4

y

j

exp

−i

m

¸

i=1

p

i

x

i

+ i

n

¸

j=1

k

j

y

j

¸

(5.37)

(−i)

n+m

d

4

z

1

d

4

z

n+m

δ

4

(x

1

−z

1

) δ

4

(y

n

−z

n+m

) G

n+m

(z

1

, . . . , z

n+m

).

After performing all the integrations over the z

k

’s, the ﬁnal relation becomes

k

1

, . . . , k

n

; out

p

1

, . . . , p

m

; in

=

m

¸

i=1

d

4

x

i

n

¸

j=1

d

4

y

j

exp

−i

m

¸

i=1

p

i

x

i

+ i

n

¸

j=1

k

j

y

j

¸

G

n+m

(x

1

, . . . , x

m

, y

1

, . . . , y

n

)

≡ (

n+m

(p

1

, . . . , p

m

, k

1

, . . . , k

n

), (5.38)

where (

n+m

is the truncated n+m-point function in momentum space. This result shows

that the scattering matrix element is directly given by the truncated Green’s function

in momentum space. The latter can be obtained using the Feynman rules without the

expression for the external legs.

Problems

5.1 Verify that

: φ(x

1

)φ(x

2

) : = : φ(x

2

)φ(x

1

) :

Hint: write φ = φ

+

+φ

−

, where φ

+

and φ

−

are creation and annihilation components

of φ.

5.2 Verify that

G

F

(x −y) = i

d

4

p

(2π)

4

e

ip·(x−y)

p

2

−m

2

+ i

is a Green’s function of (∂

µ

∂

µ

+ m

2

) as →0 (where ∂

µ

≡ ∂/∂x

µ

).

5.3 Find the expressions corresponding to the following momentum space Feynman di-

agrams

44

Integrate out all the δ-functions but do not perform the remaining integrals.

6 Concluding remarks

Although we have missed out on many important topics in Quantum Field Theory, we got

to the point where we established contact between the underlying formalism of Quantum

Field theory and the Feynman rules, which are widely used. The main concepts of the

formulation were discussed: we introduced ﬁeld operators, multi-particle states that live

in Fock spaces, creation and annihilation operators, as well as tools like normal ordering,

all of which served to overcome the problems we encountered when we na¨ıvely tried to

write down a relativistic quantum theory. The basic ingredients to formulate a quantum

theory for a real scalar ﬁeld can be used to quantise the electron and photon ﬁelds. This

is discussed in the lectures by Robert Thorne at this school. Renormalisation is a topic

which is not so easily discussed in a relatively short period of time, and hence I refer the

reader to standard textbooks on Quantum Field Theory, which are listed below. The

same applies to the method of quantisation via path integrals.

Acknowledgements

I would like to thank Julia Sedgbeer for the invitation and for all her eﬀorts which made

the school a very pleasant experience. I would also like to thank Margaret Evans for

her most friendly and eﬃcient way of organising such an event. Many thanks go to my

fellow lecturers, the tutors and all the students, whose persistent questions lead to many

clariﬁcations and improvements.

References

[1] F. Mandl and G. Shaw, Quantum Field Theory, Wiley 1984.

[2] C. Itzykson and J.B. Zuber, Quantum Field Theory, McGraw-Hill 1987.

[3] M.E. Peskin and D.V. Schroeder, An Introduction to Quantum Field Theory, Addi-

son Wesley 1995

[4] S. Weinberg, The Quantum Theory of Fields, Vol. 1, CUP 1995

45

A Notation and conventions

4-vectors:

x

µ

= (x

0

, x) = (t, x)

x

µ

= g

µν

x

ν

= (x

0

, −x) = (t, −x)

Metric tensor: g

µν

= g

µν

=

¸

¸

¸

1 0 0 0

0 −1 0 0

0 0 −1 0

0 0 0 −1

Scalar product:

x

µ

x

µ

= x

0

x

0

+ x

1

x

1

+ x

2

x

2

+ x

3

x

3

= t

2

−x

2

Gradient operators:

∂

µ

≡

∂

∂x

µ

=

∂

∂t

, −∇

∂

µ

≡

∂

∂x

µ

=

∂

∂t

, ∇

d’Alembertian: ∂

µ

∂

µ

=

∂

2

∂t

2

−∇

2

≡

Momentum operator:

ˆ p

µ

= i∂

µ

=

i

∂

∂t

, −i∇

=

ˆ

E, ˆ p

(as it should be)

δ-functions:

d

3

p f(p) δ

3

(p −q) = f(q)

d

3

xe

−ip·x

= (2π)

3

δ

3

(p)

d

3

p

(2π)

3

e

−ip·x

= δ

3

(x)

(similarly in four dimensions)

Note:

δ(x

2

−x

2

0

) = δ¦(x −x

0

)(x + x

0

)¦

=

1

2x

¦δ(x −x

0

) + δ(x + x

0

)¦

46

Contents

0 Prologue 1 Ingredients of Quantum Field Theory 1.1 Basic concepts of Quantum Mechanics . . . . . . . . . 1.2 Special Relativity . . . . . . . . . . . . . . . . . . . . . 1.3 Relativistic wave equation: The Klein-Gordon equation Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Quantisation of the free scalar ﬁeld 2.1 Heisenberg picture . . . . . . . . . . . . . . . . . . 2.2 Plane wave solutions of the Klein-Gordon equation 2.3 Quantisation of the real Klein-Gordon ﬁeld . . . . . 2.4 Causality and commutation relations . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Lagrangian formalism 3.1 Classical Mechanics . . . . . . . . . . . . 3.2 Classical ﬁeld theory . . . . . . . . . . . 3.3 Quantum ﬁeld theory . . . . . . . . . . . 3.4 Summary: canonical quantisation for real Problems . . . . . . . . . . . . . . . . . . . . 4 Interacting scalar ﬁelds 4.1 The S-matrix . . . . . . . . . . . . . . 4.2 More on time evolution: Dirac picture 4.3 S-matrix and Green’s functions . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 3 3 5 6 7

8 . 8 . 9 . 10 . 12 . 16 . . . . . . . . . . . . . . . . 17 17 19 22 24 25 26 27 28 31 34 34 36 37 38 40 42 43 44 45 45 46

. . . . . . . . . . . . . . . . . . . . . scalar ﬁelds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5 Perturbation Theory 5.1 Wick’s Theorem . . . . . . . . . . . . . . . . . 5.2 The Feynman propagator . . . . . . . . . . . . . 5.3 Two-particle scattering to O(λ) . . . . . . . . . 5.4 Graphical representation of the Wick expansion: 5.5 Feynman rules in momentum space . . . . . . . 5.6 S-matrix and truncated Green’s functions . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . 6 Concluding remarks Acknowledgements A Notation and conventions

. . . . . . . . . . . . . . . . . . . . . . . . . . . Feynman rules . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

When I became a student of Pomeranchuk in 1950 I heard from him a kind of joke that the Book of Physics had two volumes: vol.1 is “Pumps and Manometers”, vol.2 is “Quantum Field Theory” Lev Okun

0

Prologue

The development of Quantum Field Theory is surely one of the most important achievements in modern physics. Presently, all observational evidence points to the fact that Quantum Field Theory (QFT) provides a good description of elementary particles for a wide range of energies, leading up to the Planck mass, E < MPlanck 1019 GeV. His∼ torically, Quantum Electrodynamics (QED) emerged as the prototype of modern QFT’s. It was developed in the late 1940s and early 1950s chieﬂy by Feynman, Schwinger and Tomonaga, and is perhaps the most successful theory in physics: the anomalous magnetic dipole moment of the electron is predicted by QED with an accuracy of one part in 1010 . The scope of these lectures is to provide an introduction to the formalism of Quantum Field Theory. This is best explained by restricting the discussion to the quantum theory of scalar ﬁelds. Furthermore, I shall use the Lagrangian formalism and canonical quantisation, thus leaving aside the quantisation approach via path integrals. Since the main motivation for these lectures is the discussion of the underlying formalism leading to the derivation of Feynman rules, the canonical approach is totally adequate. The physically relevant theories of QED, QCD and the electroweak model are covered in the lectures by Nick Evans, Michael Kr¨mer and Adrian Signer. a The outline of these lecture notes is as follows: in section 1 we shall review the two main ingredients of QFT, namely Quantum Mechanics and Special Relativity. Section 2 discusses the relativistic wave equation for free scalar ﬁelds, i.e. the Klein-Gordon equation. We shall derive a consistent quantum interpretation of its solutions in terms of the Klein-Gordon ﬁeld. In section 3 the Lagrangian formalism will be discussed, which is based on the Principle of least Action and provides a general framework to impose the canonical quantisation rules. The more interesting case of interacting scalar ﬁelds is presented in section 4: we shall introduce the S-matrix and examine its relation with the Green’s functions of the theory. Finally, in section 5 the general method of perturbation theory is presented, which serves to compute the Green functions in terms of a power series in the coupling constant. Here, Wick’s Theorem is of central importance in order to understand the derivation of Feynman rules.

1

1.1

**Ingredients of Quantum Field Theory
**

Basic concepts of Quantum Mechanics

Let us recall that physical states in Quantum Mechanics are described by vectors |ψ in a Hilbert space H. On the other hand, physical observables are described by hermitian

3

operators acting on the state vectors |ψ . Common examples for operators are position operator : x ˆ momentum operator : p = −i ˆ

2 2 p2 ˆ + V (ˆ) = − x + V (ˆ). x (1.1) 2m 2m The required hermiticity of operators implies that these have real eigenvalues, so that their expectation values can be interpreted as the result of a measurement. The deﬁnition of the Hamiltonian follows the expression for the total energy in classical mechanics, i.e. p2 + V (x). T +V = (1.2) 2m If the potential energy V vanishes, then the total energy E is given by

ˆ Hamiltonian : H =

p2 (1.3) 2m which is the classical, non-relativistic energy-momentum relation. In Quantum Mechanics, the operators must satisfy commutation relations, for instance, E= x, p = i ˆ ˆ [ˆ, x] = p, p = 0. x ˆ ˆ ˆ (1.4) (1.5)

Equation (1.4) expresses the well-known fact that precise and independent measurements of x and p cannot be made. To summarise the above, one may say that quantisation is achieved by replacing physical observables by the corresponding hermitian operators and imposing suitable commutation relations. Quantum Mechanics can be formulated in terms of diﬀerent so-called quantum pictures, which treat the time dependence of state vectors and operators in diﬀerent ways. In the Schr¨dinger picture state vectors are time-dependent, whereas operators describo ing observables are time-independent. The state vector |ψ can be represented by the o wavefunction Ψ(x, t), which satisﬁes the Schr¨dinger equation: ∂ ˆ Ψ(x, t) = HΨ(x, t) (1.6) ∂t ˆ The expectation value of some hermitian operator O at a given time t is then deﬁned as i ˆ O t= ˆ d3 x Ψ∗ (x, t)OΨ(x, t), (1.7)

and the normalisation of the wavefunction is given by d3 x Ψ∗ (x, t)Ψ(x, t) = 1 t . (1.8)

Since Ψ∗ Ψ is positive, it is natural to interpret it as the probability density for ﬁnding a particle at position x. Furthermore one can derive a conserved current j, as well as a continuity equation by considering Ψ∗ × (Schr.Eq.) − Ψ × (Schr.Eq.) ∗ . 4 (1.9)

The continuity equation reads ∂ ρ = − ·j ∂t where the density ρ and the current j are given by ρ = Ψ∗ Ψ j= 2im (Ψ∗ Ψ − ( Ψ∗ )Ψ) (positive). i. cannot aﬀect each other. we conclude that this probability must be conserved: particles cannot be created or destroyed in our theory. The quantity Ψ(x. (1. In the time interval (x0 − y 0 ) the light travels a distance |x − y|. where x0 = ctx is the time coordinate (our conventions for 4-vectors. 1. let us ﬁrst recall some basic facts about special relativity.2 Special Relativity Before we try to extend the formalism of Quantum Mechanics to relativistic systems.e. Using eq. two events which are characterised by space-time points xµ and y µ are said to be causal if the distance (x − y)2 is time-like.e.13) where in the last line we have used Gauss’s theorem. can be rewritten and we obtain ∂ (1. For instance. ∂t S In other words. By contrast. Consider a ﬁnite volume V with boundary S. (x − y)2 > 0. provided that j = 0 everywhere at the boundary S. (real). (1.12) Now that we have derived the continuity equation let us discuss the probability interpretation of Quantum Mechanics in more detail.15) This deﬁnes the light cone about y. 5 . etc. Since 1 t represents the total probability for ﬁnding the particle anywhere inside the volume V . In the forward light cone (FLC) the concepts of “before” and “after” make sense. The integrated continuity equation is ∂ρ 3 dx = − ∂t = − S · j d3 x V V j · dS (1. i. the metric tensor. Let xµ = (x0 . are listed in Appendix A). This has important consequences about the way in which diﬀerent events can aﬀect each other.8) the lhs. since the point x is not contained inside the light cone about y. Suppose that someone sends out a light signal at time y 0 and position y. (1.11) (1.14) 1 t = − j · dS = 0.10) (1. x) denote a 4-vector. One important postulate of special relativity states that no signal and no interaction can travel faster than the speed of light. (x − y)2 < 0. t) is interpreted as a one-particle wave function. so that (x − y)2 = (x0 − y 0 )2 (x0 − y 0 )2 − (x − y)2 ≡ (x − y)2 = 0. Non-relativistic Quantum Mechanics thus provides a consistent formalism to describe a single particle. we ﬁnd that the time derivative of 1 t vanishes. two events characterised by a space-like separation.

19) . time-like (x − y)2 = 0. In non-relativistic Quantum Mechanics the commutation relations among operators indicate whether precise and independent measurements of the corresponding observables can be made. 1. space-like space Figure 1: The light cone about y. This can be rewritten as 2 ⇒ 2 ∂2 − 2 φ(x. t) = 0 ∂t2 + m2 φ(x. t). (1.18) which is the Klein-Gordon equation (KGE) for a free particle. As in the non-relativistic case one can derive a continuity equation: φ∗ × (KGE) − φ × (KGE)∗ .time (x − y)2 > 0. (1. then a measurement of one observables aﬀects that of the other.17) where φ(x. 6 (1. t) + m2 φ(x. If the commutator does not vanish. light-like y (x − y)2 < 0. which gives − 2 ∂2 φ(x. We may try to quantise the theory by replacing observables by the corresponding hermitian operators. t) + m2 φ(x. t) = 0. Events occurring at points x and y are said to be time-like (space-like) if x is inside (outside) the light cone about y.16) where we have set c = 1. We will return to this important issue in section 2. From the above it is then clear that the issue of causality must be incorporated into the commutation relations of the relativistic version of our quantum theory: whether or not independent and precise measurements of two observables can be made depends also on the separation of the 4-vectors characterising the points at which these measurements occur. t) = − ∂t2 2 2 φ(x.3 Relativistic wave equation: The Klein-Gordon equation In analogy to the non-relativistic case we shall now derive a relativistically invariant wave equation. Our starting point is the relativistic energy-momentum relation E 2 = p2 + m2 (1. t) is the wavefunction.

the quantity ρ. Verify that the continuity equation can be written in manifestly covariant form. Therefore we will have to abandon the interpretation which was so successful in the nonrelativistic case.The density and current in this case are given by ∂ ∂ ρ = i φ∗ φ − ( φ∗ )φ ∂t ∂t ∗ j = −i (φ φ − ( φ∗ )φ) (real) (real) (1. 7 . So far we have assumed that the number of particles should be constant in the relativistic theory. V φ∗ φ d3 x may vary over time due to the change in the number of particles. t). Set c = 1 and verify that the continuity equation can be written as ∂µ j µ = 0 where jµ = i (φ∗ (∂µ φ) − (∂µ φ∗ ) φ) . 2. As we shall see. o 2 2 p2 + m2 (1.20) (1. though real. There are states of arbitrarily low energies: E=± so that the system has no ground state. the problem of negative energies remains unsolved. Although V ρ d3 x is again independent of time. t) ∂t show that the probability density ρ = Ψ∗ Ψ satisﬁes the continuity equation eq.2 Derive the corresponding continuity equation for the Klein-Gordon equation using a similar procedure as in Problem 1.e.22) − 2m + V (x) Ψ(x.21) However. 3. with the current j given by eq.1 Starting from the Schr¨dinger equation for the wavefunction Ψ(x. we run into a number of problems: 1. If we drop this requirement and introduce the concept of particle creation or annihilation. this is ultimately possible if we re-interpret φ as a ﬁeld operator which can destroy or create particles. j). it cannot be shown that V φ∗ φ d3 x is independent of time. (1. is not positive deﬁnite. ∂µ j µ = 0.12). then the interpretation of φ as a one-particle wavefunction does not have to be upheld. if we try to take over the probabilistic interpretation encountered in the nonrelativistic case. j µ = (cρ. since it does not satisfy a continuity equation. i. (1. 1. Problems 1. but then one ﬁnds that φ∗ φ is not conserved.10). which means that ρ cannot be interpreted as a probability density. One might try to interpret φ∗ φ as a probability density instead. In other words. t) = i ∂ Ψ(x. As a consequence. Although this apparently overcomes the problem that a consistent relativistic theory for one particle cannot be formulated.1.

o Let us consider the expectation value of an observable A. t)AS Ψ(x.1) ∂t By contrast. S|AS |t. The state vectors |H of the Heisenberg picture are time-independent but may carry a time label. S = H|t. in which operators and states are labelled by “H”. t). From now on we will always set = c = 1.3) ˆ (2. S .2) we obtain ˆ ˆ t. operators have no explicit time dependence in the Schr¨dinger picture. (2. however. ˆ ˆ (2. It is possible to trade oﬀ the time dependence between the state o vectors and the operators.e. ˆ t.7) ˆ ˆ The above relations between AS and AH (t) and the respective state vectors deﬁne the Heisenberg picture. In order to indicate this we shall use the notation |t. has to remain unchanged. ﬁnally arriving at a consistent way to quantise it. As mentioned earlier.6) |H ≡ |t0 . Then we have ˆ ˆ t. S . the physical states of a system are represented by vectors in a Hilbert space. i. S = t0 . S . S = e−iH(t−t0 ) |t0 .5) (2. S|eiH(t−t0 ) AS e−iH(t−t0 ) |t0 . S|AS |t. S = H|AH (t)|H (2. S|AS |t. The o formal solution of eq.2 Quantisation of the free scalar ﬁeld We are now going to discuss the Klein-Gordon equation for free. where o the state vectors have an explicit time dependence. S ≡ ˆ d3 x Ψ∗ (x.e.2) ˆ where the subscript “S” on AS denotes that the corresponding operator is deﬁned in the Schr¨dinger picture. We have already encountered the Schr¨dinger picture. non-interacting particles in more detail. i. S . S for a Schr¨dinger state at time t which satisﬁes the Schr¨dinger o o equation: ∂ ˆ i |t. The expectation value.4) ˆ We can thus deﬁne the time-dependent operator AH (t) and the time-independent state vector |H through ˆ ˆ ˆ ˆ AH (t) = eiH(t−t0 ) AS e−iH(t−t0 ) (2. which 8 . 2.1) in a time interval from t0 to t is given by |t. and derive a result which will prove useful for the later analysis. (2. (2. since both are merely mathematical devices to describe physical quantities.1 Heisenberg picture Before we are ready to discuss solutions of the Klein-Gordon equation we have to look at the equations of motion in more detail. which is easily veriﬁed by inserting it into the Schr¨dinger equation. (2. H = |t0 . Using the solution eq.

14) These are called the generalised Heisenberg equations of motion. Using eq. and by using the 9 . (2. H + ixk |x. Here. H − xk pj |x. t) . t) = i H.6)).12) then we can rewrite the equations of motion in covariant form ∂ ˆ AH (x) = i pµ . H : ˆ ˆ [ˆj . H ) + ixk |x. ∂t (2. 2. ∂xj (2. t) = i pj . any Heisenberg operator satisﬁes the following equations of motion: ∂ ˆ ˆ ˆ AH (x. ˆ ˆ ∂xj If. we use the 4-vectors ∂ = ∂xµ ∂ . t) ∂t ∂ ˆ AH (x. AH (x. H .10) To summarise. [ˆj . What we have recovered is the uncertainty relation. AH (x. t). State vectors in both pictures can depend on a spatial coordinate x. k refer to the components of x and p). characterised by φ∗ (x) = φ(x) and interpret φ(x) as a classical. xk ] |x.9) (The indices j. H and compute p. H = −i ∂xj ∂xj ∂xj = −iδjk |x. x |x. real ﬁeld.− ∂t ⇒ ˆ ˆ pµ = H.simply means that these vectors are identiﬁed with the corresponding Schr¨dinger states o at that particular time (see eq.2 Plane wave solutions of the Klein-Gordon equation In covariant form the Klein-Gordon equation reads + m2 φ(x) = 0. H − ixk |x. Let us consider a Heisenberg state |x.11) (2.8) which is the Heisenberg equation of motion.15) Let us consider real solutions. (2. AH (x) . (2. xk ] = −iδjk . “classical” simply means “not quantised”. AH (t) . H p ˆ = pj xk |x. AH (x. ˆ ˆ ∂xµ x ≡ xµ = (t.5) it is easy to derive ∂ ˆ ˆ ˆ AH (t) = i H. H ˆˆ ˆ ˆ ∂ ∂ (ˆk |x. H x = −i ∂xj ∂xj ∂xk ∂ ∂ |x. in accordance with the conventions in Appendix A.13) (2. x). This result can easily be generalised to Heisenberg p ˆ ˆH which have an explicit dependence on x: operators A pj . (2. t) = −i ˆ ˆ ∂ ˆ AH (x. p ˆ (2. (2.

(2. This illustrates the close relationship of a relativistically invariant theory like classical electromagnetism with the Klein-Gordon equation and motivates the interpretation of φ as a ﬁeld variable. i = 1.3 Quantisation of the real Klein-Gordon ﬁeld Starting from the classical solution to the KGE in eq. It is well known that in Maxwell’s theory the scalar and vector potentials (denoted by Φ and V .17) These wave equations are identical to the Klein-Gordon equation.18) The Klein-Gordon equation is satisﬁed if (k 0 )2 − k 2 = m2 so that k0 = ± k 2 + m2 .e. From the general solution one easily reads oﬀ that φ is real. i = 1. 2.19) If we choose the positive branch of the square root then we can deﬁne the energy as E(k) = k 2 + m2 > 0 (2. (2. To ﬁnd the solutions of the Klein-Gordon equation let us consider an ansatz of plane waves 0 φ(x) ∝ ei(k t−k·x) (2. i. Vi (x) = 0.20) and obtain two types of solutions which read φ+ (x) ∝ ei(E(k)t−k·x) . i. as well as the imposition of suitable commutation relations among these operators. Bi (x) = 0. 2.e. (2.22) where α(k) is an arbitrary complex coeﬃcient.23) (2. respectively) satisfy wave equations of the type Φ(x) = 0. Using E(k)t − k · x = k µ kµ = kµ k µ = k · x this solution reads φ(x) = d3 k eik·x α∗ (k) + e−ik·x α(k) . Recall that quantisation is achieved by replacing observables by the corresponding operators. (2.concept of “ﬁelds” we make the analogy with electromagnetism. φ− (x) ∝ e−i(E(k)t−k·x) . except for the mass term. In electromagnetism one can indeed express one component of 10 . which is absent in the Maxwell theory.23) we will now derive a quantum interpretation of φ. 3. 3. Ei (x) = 0. 2. φ = φ∗ .16) and the same is true for the components of the electric and magnetic ﬁelds.21) The general solution is a superposition of φ+ and φ− . (2. (2π)3 2E(k) (2.

φ is a real solution of the ˆ ˆ KGE if the operator φ is hermitian and satisﬁes the KGE as well.32) Furthermore. eq.e. a† (k)] + e−ik·x i[ˆµ .27) ∂xµ ˆ From the classical solution φ(x) of eq. by the expectation value of a ﬁeld operator. 0|0 = 1.31) ˆ [ˆµ . which is a classical observable.25) Historically this is how a ﬁeld theory was quantised for the ﬁrst time. Ei (x) = classical observable ˆ |Ei (x)| . and it we normalise it to one. value of ﬁeld operator (2.14)): ∂ ˆ φ(x) = i pµ .f. a(k)] = −k µ a(k) p ˆ These relation will help later us to specify the state vectors in more detail. the vacuum has zero energy and momentum. φ is a Heisenberg operator and must therefore satisfy the generalised Heisenberg equation (c. Here we have used empty bra and ket symbols to indicate that the states have to be speciﬁed in more detail – this is discussed later.29) By comparing the integrands we ﬁnd p ˆ [ˆµ . which implies that the momentum operator pµ has a zero eigenvalue when acting on |0 : ˆ pµ |0 = 0.28) Here a† (k) and a(k) are operators. Let us now consider the vacuum state. a† (k)] = k µ a† (k) ˆ (2. This state is denoted by |0 .26) ˆ ˆ The condition φ = φ∗ is obtained if φ† = φ.the electric ﬁeld. which gives expression for φ ∂ ˆ φ(x) = ∂xµ = d3 k (2π)3 2E(k) d3 k (2π)3 2E(k) a eik·x (ik µ )ˆ† (k) + e−ik·x (−ik µ )ˆ(k) a eik·x i[ˆµ . If we follow the same approach for the real Klein-Gordon ﬁeld. i. ˆ then we have to replace the ﬁeld φ by the expectation value of the ﬁeld operator φ: ˆ φ(x) = |φ(x)| (2. (2. which does not contain any particles.30) (2. ˆ 3 2E(k) (2π) (2.e. φ(x) . (2. i. whose properties can be worked out by inserting the ˆ ˆ ˆ into the generalised Heisenberg equation. (2.23) we can deduce the following form of φ(x) ˆ φ(x) = d3 k ˆ eik·xa† (k) + e−ik·xa(k) .24) (2. In other words. Furthermore. ˆ 11 (2. exp. Ei (x). ˆ ˆ (2.33) . a(k)] . p ˆ p ˆ (2.

41) In this case x is outside the light cone about x and vice versa. this means that a(k)|0 is a state with negative energy. and E(k) = k 2 + m2 . (2.30) we ﬁnd that p ˆ [ˆµ . and. ˆ ˆ ˆ (2. ˆ (2. so that we deﬁne a(k)|0 = 0. Such ˆ states are clearly unphysical. k). 2. a† (k)]|0 .40) The states which appear in our quantised version of the Klein-Gordon equation are thus multi-particle states. ˆ (2. Similarly we ﬁnd that ˆ pµ a(k)|0 = −k µ a(k)|0 . after repeating ˆ a this procedure n times one obtains ˆ a† (k n ) · · · a† (k 1 )|0 ˆ n-particle state (2. (2. Using eq. cannot aﬀect each 12 .35) However. such that (x − x )2 < 0 (space-like). that pµ a† (k 2 )ˆ† (k 1 )|0 ˆˆ a p a(k 2 )ˆ (k 1 )|0 ˆ ˆ a µ † µ µ = (k1 + k2 )ˆ† (k 2 )ˆ† (k 1 )|0 a a (2. a† (k)]|0 p ˆ ⇒ pµ a† (k)|0 ˆ ˆ = pµ a† (k)|0 − a† (k)ˆµ |0 ˆˆ ˆ p = k µ a† (k)|0 ˆ = k µ a† (k)|0 . (2. which is central in Special Relativity. one can show by applying eqs. measurements of some observable).38) = µ (k1 − µ k2 )ˆ(k 2 )ˆ† (k 1 )|0 a a It is thus natural to interpret a† (k 2 )ˆ† (k 1 )|0 as a two-particle state. x and x . Consider two 4-vectors.31). a† (k)|0 is an eigenstate of pµ with eigenvalue k µ = (E(k).Now we consider [ˆµ .37) .30) and (2. Events occurring at x and x respectively (for instance. where k µ is ˆ ˆ the 4-momentum of a relativistic particle with mass m. We still have to look at the normalisation of these states in detail.4 Causality and commutation relations In section 1 it was mentioned that commutation relations of operators in the relativistic case must be imposed in accordance with the principle of causality. respectively decrease the number of particles in a ˆ given state by one. We call a† (k) : creation operator ˆ a(k) : annihilation operator.39) ˆ The operators a† (k) and a(k) increase. −k 0 = −E(k) < 0. ˆ and also introduce the commutation rules for our ﬁeld operator φ.36) (2. (2.34) In other words. ˆ Furthermore. Therefore it makes sense to interpret a† (k)|0 as a one-particle state.

φ(y.42) with explicit space and and time coordinates. t ) = 0.49) If we take the result of eq. ∂t ˆ These expressions are referred to as the equal time commutators of the ﬁeld operator φ. t) reads ˆ d3 x eip·x φ(x. ˆ 3 2E(k) (2π) (2. Before we can use them in order to deduce the commutators for creation and annihilation ˆ operators. φ(x . ˆ ˆ (2. by letting k → −k in the term proportional to a(k) in eq. eiE2 t a† (p2 ) + e−iE2 t a(−p2 ) ˆ ˆ ˆ ˆ = 2E1 2E2 d3 x ˆ ˆ d3 y eip1 ·x eip2 ·y φ(x. It also vanishes for t = t. ˆ ˆ e (2π)3 2 (2. t). t) ≡ φ(x. we rewrite φ. φ(x ) = 0 for (x − x )2 < 0.42) is an important relation in order to derive the commutation rules for the creation and annihilation operators a and a† .42) This condition is sometimes called micro-causality.28). ˆ 2E (2.44) φ(x. t) = d3 k (2π)3 2E(k) ˆ ˆ d3 x ei(p−k)·x eiEt a† (k) + e−iEt a(−k) (2π)3 δ 3 (p−k) = 1 ˆ eiEt a† (p) + e−iEt a(−p) .50) .43) which is just eq. φ(x . (2. Equation (2. (2. the commutator vanishes in a ﬁnite interval |t − t|. t) yields −2i d3 x eip·x ∂ ˆ φ(x. φ(x .45) ˆ The Fourier transform of φ(x. |t − t| < |x − x | = 0. t) = ∂t d3 k i −ik·x iEt † e a (k) − e−iEt a(−k) . t) = Similarly one obtains ∂ ˆ ˆ ∂0 φ(x. ˆ ˆ φ(x. (2. t) = 0. Our starting point is ˆ ˆ ˆ ˆ φ(x. t).48) we ﬁnd eiE1 t a† (p1 ) + e−iE1 t a(−p1 ). (2. t) . t) = eiEt a† (p) − e−iEt a(−p) . t). (2. x = x . t). t) = 0.47) where E = E(p) in the last line. 13 (2. ˆ ˆ ∂t (2. hence ∂ ˆ ˆ ˆ ˆ (2. As long as |t − t| < |x − x |.46) d3 k ˆ e−ik·x eiEt a† (k) + e−iEt a(−k) .48) ˆ and a similar procedure applied to the Fourier transform of ∂0 φ(x. φ(x. It is then clear that the ﬁeld operators φ(x) and φ(x ) must commute ˆ ˆ φ(x).ˆ ˆ other – they are independent. Thus we ﬁnd 2E ˆ d3 x eip·x φ(x. t) = eiEt a† (p) + e−iEt a(−p) .

ˆ a ˆ a (2.51) which tells us the important result that both creation and annihilation operators commute among themselves! Let us discuss the consequences by considering two creation operators. However.and far right-hand sides refer to particle one and two. The fact that we have considered real solutions to the KGE means that we consider particles without charge. We see that. in analogy with Quantum Mechanics. a† (p1 ) and ˆ † a (p2 ). a(−p2 ) = 0. t) + iE(q) ∂0 φ(x. so far we have not imposed any commutation rules for the case x = y. t). a† (p2 ) ˆ ˆ ˆ ˆ ˆ ˆ + ei(E1 −E2 )t a† (p1 ).52) (2. p2 = |p2 . a† (p2 ) = a(−p1 ).e.49) ˆ ˆ † for a and a then we ﬁnd ˆ ˆ ˆ a(p).56) × E(p) δ 3 (x − y) + E(q) δ 3 (y − x) = ei(E(p)−E(q))t d3 x e−i(p−q)·x (E(p) + E(q)). Let us try ˆ ˆ φ(x. We then obtain x ˆ a(p). a(−p2 ) = 0. We still have to derive the commutator of a(p). If we solve eq. We have thus derived that our particles must be bosons ! The quantised Klein-Gordon theory describes relativistic particles which are bosons. ˆ ˆ This relation is valid for all times t. The equal time commutator tells us that ˆ ˆ φ(x.53) where the positions of the 3-momenta on the far left.54) ˆ ˆ ˆ ˆ × −iE(p) φ(x.57) 14 . The fact that creation operators commute implies that ˆ |p1 . must vanish. t) = 0 for x = y. a† (q) . i. and we deduce that a† (p1 ). t) = iδ 3 (x − y). (2. t). the state |p1 . a(−p2 ) + e−i(E1 +E2 )t a(−p1 ). (2π)3 δ 3 (p−q) (2. t). φ(y. An example of a neutral particle with spin zero is the neutral pion. p2 ≡ a† (p1 )ˆ† (p2 )|0 = a† (p2 )ˆ† (p1 )|0 ≡ |p2 . which gives ei(E1 +E2 )t a† (p1 ).55) as required by causality. (2.48) and (2. ∂0 φ(y. since |p1 . π 0 . t) This expression can be simpliﬁed using the commutators for the ﬁeld operators. respectively. a† (p2 ) + e−i(E1 −E2 )t a(−p1 ). where [ˆj . which have integer spin. a† (q) ˆ ˆ = ei(E(p)−E(q))t d3 x d3 y e−ip·x eiq·y (2. p1 .The equal time commutator tells us that the rhs. a† (q) = ei(E(p)−E(q))t ˆ d3 x d3 y e−ip·x eiq·y . The formalism that we have developed so far can be used to describe free neutral pions. p2 is symmetric if we exchange the two particles. (2. p1 . pk ] = iδjk . t). ∂0 φ(y. ˆ ˆ ˆ ˆ (2. ∂0 φ(y. acting on the vacuum |0 .

ˆ (2. a(k 2 )] = 0 a ˆ Relation (2. ˆ ˆ φ(x.62) which ensures that the resulting theory obeys causality. d3 k k |k = (2π)3 2E(k ) d3 k E(k) 3 δ (k − k) = 1. A consistent quantisation of the free Klein-Gordon ﬁeld has been achieved by going through the following steps: • Abandon the interpretation of φ(x. (2. ˆ |k = a† (k)|0 .60) ˆ |k = a† (k )|0 . t) = 0.e. i. 15 (2. ˆ ˆ • Impose the equal time commutators ˆ ˆ φ(x. t). • This implies the following commutation rules for a† .64) (2.66) a(k)|0 = 0 : ˆ . i. ˆ • Expand φ in terms of creation and annihilation operators a† and a. φ(y. a (k 2 ) ˆ ˆ † = 0 = (2π) 2E(k 1 )δ (k 1 − k 2 ) 3 3 (2. (2. let us make a brief summary of what we have actually done. a† (k) |0 = (2π)3 2E(k)δ 3 (k − k) > 0. 0|0 = 1 one-meson state. t). The state vectors are interpreted as multi-boson (or multi-meson) states: |0 a (k)|0 = |k ˆ † : : vacuum. a† (k 2 ) ˆ ˆ a(k 1 ). t) as a one-particle wave function and think of it as a ﬁeld variable. Then we have k |k = = a 0|ˆ(k )ˆ† (k)|0 a 0| a(k ).59) ˆ ˆ We see that the way we have set up the commutation rules for φ and ∂0 φ in eq. To conclude this chapter.58) Armed with this result we can now look at the normalisation of our multi-boson states. (2.Our commutation rule for a and a† thus reads ˆ ˆ ˆ a(p). ˆ • Replace the classical ﬁeld φ by a ﬁeld operator φ.63) (2.63) implies that our particles must be bosons. Let |k and |k be two one-meson states.65) [ˆ(k 1 ). t) = iδ 3 (x − y).56) ensures that our multi-meson states have positive norm. ∂0 φ(y. ˆ ˆ (2. E(k ) (2.e. positive energy exclude states with negative energy. a: ˆ ˆ a† (k 1 ). a† (q) = (2π)3 2E(p)δ 3 (p − q).61) and thus our states form indeed a proper Hilbert space.

8). AH (x. a† (k 1 ) · · · a† (k n )|0 = |k 1 . a† (k) = k µ a† (k).6 Starting from the expression for ∂0 φ(x.3 Given the relativistic invariance of the measure d4 k.10). t) = d3 k i −ik·x iEt † e ˆ e a (k) − e−iEt a(−k) ˆ (2π)3 2 pµ .65) implies that one-meson states must have positive norm.37) and (2. eq. for a Heisenberg operator ˆ AH (x. .1 Let AH (t) and AS be operators in the Heisenberg and Schr¨dinger pictures. 2. a(k) = −k µ a(k) ˆ ˆ ˆ 16 . where |x. (2. k n . ˆ ˆ (2.2 Derive the generalised uncertainty relation. eq. t). 2.g. relation (2. e. 2. t): ˆ ∂0 φ(x. (2. Hint: Consider pj . ˆ ˆ ˆ to derive eqs. t) |x.23) is indeed a solution of the Klein-Gordon equation. respeco tively. Hint: Start from the expression d4 k δ(k 2 − m2 ) θ(k0 ) 3 (2π) and use 1 (δ(x − x0 ) + δ(x + x0 )) 2x What is the signiﬁcance of the δ and θ functions above? δ(x2 − x2 ) = 0 2. Using the relation ˆ ˆ ˆ ˆ AH (t) = eiH(t−t0 ) AS e−iH(t−t0 ) derive the Heisenberg equation of motion. provided that E(k) = k 2 + m2 . Problems ˆ ˆ 2.67) The resulting Hilbert space of multi-particle states is called a Fock space. .Furthermore. H is an arbitrary Heisenˆ ˆ berg state. .4 Verify that eq. . ˆ 2. (2.5 Use the commutator relations pµ . H . (2. show that the integration measure d3 k (2π)3 2E(k) is Lorentz-invariant.38). The generalisation to n-meson states is achieved by repeated applications of creation operators.

x). be derived from the Principle of least Action. a† (q)] = ei(E(p)−E(q))t a ˆ d3 x d3 y e−ip·x+iq·y . ˙ 17 (3. t). ˆ eiEt a† (k) + e−iEt a(−k) = 2E(k) ˆ ˆ d3 x eik·x φ(x. This well-known result is Newton’s 2nd law. Use this result together with eq. t) 3 Lagrangian formalism We could have derived all results in the previous section in a much more axiomatic fashion. The Lagrangian formalism provides a generalisation such that the equation of motion can. (2. respectively. 3. The Lagrangian formalism is used in all modern theories of matter. a† (q)] ˆ reads [ˆ(p).3) . ˆ a ˆ to solve for a† (k) and a(k).e. t). t). (3.48) in the lecture notes. and its extensions. The Lagrangian for a single classical particle is deﬁned as ˙ L = T − V = 1 mx2 − V (x) 2 = L(x. LGSW (3. before we move on to classical ﬁeld theory and then quantise it. LQCD . i. which is ∂V m¨ = F ≡ − x .1) for Quantum Electrodynamics. φ(y. ˆ ˆ ˆ ˆ × i E(q) ∂0 φ(x. A typical problem is then to ﬁnd the trajectory x(t) of that particle by integrating its equation of motion. In this section we will introduce the Lagrangian formalism in the context of classical mechanics. for instance LQED .invert the Fourier transform to obtain eiEt a† (k) − e−iEt a(−k) = −2i ˆ ˆ ˆ d3 x eik·x ∂0 φ(x. Verify that the expression for the commutator [ˆ(p). t). t) − E(p) φ(x. Quantum Chromodynamics and the electroweak (GlashowSalam-Weinberg) theory. most notably in the Standard Model of Elementary Particle Physics.2) ∂x where V is the potential.1 Classical Mechanics In Classical Mechanics one considers point particles of mass m at some position x. The starting point for any theoretical treatment of fundamental interactions is the relevant Lagrangian. ∂0 φ(y. by starting from the Lagrangian of the theory. in fact.

∂x dt ∂ x ˙ If we insert eq. because of the boundary conditions.8) must vanish for arbitrary variations δx. and it is a legitimate question to ask what we have gained. Let us minimise S in the standard fashion by considering a variation in the path x(t): x(t) −→ x (t) = x(t) + δx(t). (3. x (t1 ) = x(t1 ) x (t2 ) = x(t2 ) ⇒ δx(t1 ) = δx(t2 ) = 0. its trajectory) must be such that the action S is a minimum. In other words. δx(t) x.e. In this example the Lagrangian formalism and Newton’s law are equivalent. The Principle of least Action then states that the motion of the particle (i.e. and since the remaining integral in eq.4) is called the action. x) dt ˙ t1 (3. then δS = 0. (3.7) The variation δS is obtained from the Taylor expansion of the integrand about (x. this is only possible if the integrand itself vanishes. i. Thus we have derived the Euler-Lagrange equation ∂L d ∂L − = 0.3) into the Euler-Lagrange equation we obtain ∂L ∂x d ∂L dt ∂ x ˙ = − = ∂V (x) =F ∂x ∂V ∂x (3. x) ˙ t2 δS = t1 ∂L ∂L δx + δx ˙ ∂x ∂x ˙ t2 t2 dt δx dt (3.The integral S= t2 L(x. The key advantage of the Lagrangian 18 .6) The variation of S can then be worked out in detail by considering t2 S + δS = t1 L(x + δx. dt (3. If S is a minimum.8) = ∂L δx ∂x ˙ + t1 t1 ∂L d ∂L − ∂x dt ∂ x ˙ The ﬁrst term in the last line vanishes. the requirement that S be a minimum singles out one particular trajectory x(t).10) ⇒ m¨ = F = − x So we ﬁnd that the equation of motion (here: Newton’s 2nd law) is reproduced by the Principle of least Action. (3. (3. ˙ ˙ δx = ˙ d δx.9) d mx = m¨ ˙ x dt (Newton’s law). (3. x + δ x) dt.5) As a boundary condition we require that the points at the boundaries stay ﬁxed.

We can now have a look at conservation laws. ˙ 19 . p) = + dt ∂x dt ∂p dt = −px + xp = 0. and the Lagrangian is given by L(φ. (3. ˙ ∂x ˙ (3. x) −→ L(φ. o The “dictionary” which provides the relation to ﬁeld theory reads as follows: Classical Mechanics: Classical ﬁeld theory: t −→ xµ x(t) −→ φ(x) (ﬁeld conﬁguration) L(x. It is now easy to derive Hamilton’s equations (see exercise ) ∂H ∂H = −p.2 Classical ﬁeld theory (3. ˙ (3.16) where L is usually referred to as the Lagrangian density.12) The Hamiltonian H(x. t) = φ(x).15) The action is deﬁned as S= d4 x L(φ. We will return to this point below when we discuss classical ﬁeld theory. (3.13) ∂x ∂p The Euler-Lagrange equations and Hamilton’s equations provide an entirely equivalent description of the system. x) ˙ ˙ 1 = mx2 − 2 mx2 + V (x) ˙ ˙ ∂L = mx.11) = 1 mx2 ˙ 2 + V (x) = T + V. ∂µ φ) (Lagrangian density) ˙ (3. it is a function of the position variable x and the conjugate momentum1 p. ˙˙ ˙ ˙ and hence we conclude that the total energy T + V is conserved. p) is the total energy of the system. 3. (3. To this end we deﬁne the conjugate momentum p by p≡ and the Hamiltonian H via H(x. ∂µ φ). p) ≡ px − L(x. Assuming that the Hamiltonian has no explicit time dependence we ﬁnd ∂H dx ∂H dp d H(x.formalism is that it helps us to study the implications of symmetries for a given theory.17) It should be noted that the conjugate momentum is in general not equal to mx. ∂µ φ) = 1 d3 x L(φ.14) We can extend the Lagrangian formalism for a classical point particle to ﬁeld theory. ˙ = x. We shall now establish the relation of the Principle of least Action with the Hamiltonian formalism. ∂µ φ). The rˆle of the particle’s trajectory x(t) is now played by the classical ﬁeld φ(x.

∂φ ∂(∂µ φ) where in the second term a summation over the Lorentz index µ is implied. In a nutshell. (3.21) (3.23) The Euler-Lagrange equation then implies ( + m2 )φ(x) = 0. (3.20) ∂L = ∂µ ∂ µ φ = ∂(∂µ φ) (3. pµ and the conjugate momentum π are not the same.18) The variation of the action becomes δS = = =0 d4 x ∂L ∂L δφ + δ(∂µ φ) ∂φ ∂(∂µ φ) ∂L ∂L ∂L + d4 x δφ − ∂µ ∂(∂µ φ) ∂φ ∂(∂µ φ) δφ.24) and one recovers the Klein-Gordon equation from a Lagrangian via the Euler-Lagrange equation. Noether’s theorem says that the invariance of a Lagrangian under a symmetry transformation implies the existence of a conserved 20 . δ∂µ φ = ∂µ δφ. Let us now consider the Lagrangian L = 1 ∂ µ φ∂µ φ − 1 m2 φ2 2 2 The functional derivatives yield ∂L = −m2 φ. ∂φ so that ∂µ ∂L = ∂µφ ∂(∂µ φ) φ. (3. let us consider variations of the ﬁeld and its derivative according to φ → φ + δφ. This yields the Euler-Lagrange equations for a classical ﬁeld theory: ∂L ∂L − ∂µ = 0. (3. ∂µ φ → ∂µ φ + δ∂µ φ.25) = π(x) ≡ ˙ ∂(∂0 φ(x)) ∂ φ(x) Note that the momentum variables p.19) at boundaries As in the previous subsection one can argue that the integrand itself must vanish if δS = 0. ∂µ φ) = ∂0 φ(x). In classical mechanics we have shown that the total energy is conserved. ∂µ φ) ∂L(φ.In order to derive the Euler-Lagrange equations in this case. then we ﬁrst have to have a look at what is known as Noether’s Theorem. The word “momentum” is used only as a semantic analogy to classical mechanics. (3.22) (3. If we want to do the same for classical ﬁeld theory. In analogy to classical mechanics we can deﬁne a conjugate momentum π through ∂L(φ.

29) in the second line. which is related to Θ0j .31) d3 x ∂ 0 Θ00 d3 x ∂ j Θj0 = V S = dS · Θ0 = 0. We have thus established the Lagrange-Hamilton formalism for classical ﬁeld theory: we derived the equation of motion (Euler-Lagrange equation) from the Lagrangian and introduced the conjugate momentum. (3.14). the conservation of 3-momentum p is associated with translational invariance of the Lagrangian. 21 . We then deﬁned the Hamiltonian (density) and considered conservation laws by studying the energy-momentum tensor Θµν . the transformation x → x + a. (3.32) where we have used eq.28) states that the energy-momentum tensor is conserved for every component ν: ∂ µ Θµν ≡ ∂ 0 Θ0ν − ∂ j Θjν = 0. ∂µ φ).quantity. provided that there is no energy ﬂow through the surface S which encloses the volume V . (3. For instance. The Hamiltonian density is a conserved quantity. (3. (3. is conserved as well. (3. (3. a : constant 3-vector. The conservation of energy can now be shown by considering ∂ ∂t d3 x Θ00 = V V (3.e. (3.26) while the conservation of energy comes from the invariance of the Lagrangian under time translations t → t + τ. Θµν . i. Equation (3.28) where the expression in curly brackets is called the energy-momentum tensor. In a similar manner one can show that 3-momentum pj . φ) ≡ Θ00 = π(x)(∂0 φ(x)) − L(φ.29) Let us examine Θ00 in more detail: Θ00 = ∂L ∂0 φ − g00 L ∂(∂ 0 φ) = π(x)(∂0 φ(x)) − L. and so we can deﬁne the Hamiltonian density H as H(π. τ : constant time interval.27) For our classical ﬁeld theory one can use Noether’s theorem to derive the following relation ∂µ ∂L ∂ν φ − gµν L ∂(∂ µ φ) = 0.30) The last line is reminiscent of the deﬁnition of the Hamiltonian in eq.

Thus we obtain ˆ 0|H|0 = δ 3 (0) 1 2 d3 p E(p) = ∞. but it is quite easy to work around it. π (y.33) (3. π (y. Let us make this ˆ more explicit by considering the modiﬁed Hamiltonian H ˆ H 1 = 2 = 1 2 d3 p ˆ a ˆ a a a ˆ a E(p) a† (p)ˆ(p) + a(p)ˆ† (p) − 0|ˆ† (p)ˆ(p) + a(p)ˆ† (p)|0 3 2E(p) (2π) d3 p E(p) 2ˆ† (p)ˆ(p) + a(p). whereas the ﬁrst can be replaced by the commutator.40) . (3. t) = iδ 3 (x − y).(3. a† (p) |0 a a ˆ ˆ ˆ ˆ (2π)3 2E(p) 22 .39) which means that the energy of the ground state is inﬁnite! This result seems rather paradoxical. the operator of the total energy. t). = ˆ π (x) ≡ ∂0 φ(x) ˆ (3. ˆ ˆ ˆ Both operators φ and π can be expanded in terms of creation and annihilation operators ˆ † a and a. ˆ a ˆ ˆ ˆ a (3. a† (p) − 0| a(p). We can simply redeﬁne the vacuum energy by setting it to zero. i. π to ﬁeld operators ˆ φ(x) and π (x).34) π (x.3.37) The ﬁrst term in curly brackets vanishes. The second can ˆ be rewritten as a(p)ˆ† (p)|0 = a(p). and by imposing the (equal time) commutators ˆ ˆ ˆ φ(x. i. t). a† (p) + a† (p)ˆ(p) |0 .e. the ground state. t) = 0. The ˆ ˆ starting point is the expression for the Hamiltonian density ˆ ˆ ˆ H = π (x)(∂0 φ(x)) − L(φ. π ). t) ˆ ˆ φ(x. t). φ(y. 3 2E(p) (2π) (3. (3.38) It is now the second term which vanishes.36) Let us now work out the energies of some of our multi-particle states. What we are actually interested in is the energy of multi-particle states relative to the vacuum. thereby subtracting the ˆ vacuum contribution from expectation values of the Hamiltonian H. In this case it does not really matter what the absolute value of the ground state energy is.e.3 Quantum ﬁeld theory We shall now quantise the theory by promoting the classical ﬁelds φ. Let’s work out the Hamiltonian. since a annihilates the vacuum. Starting with the vacuum as the ground state we get ˆ 0|H|0 = 1 2 d3 p E(p) (2π)3 2E(p) 0|ˆ† (p)ˆ(p)|0 + 0|ˆ(p)ˆ† (p)|0 a a a a . ˆ In terms of a† and a one obtains ˆ ˆ ˆ H≡ d3 x H = 1 4 1 = 2 d3 p a ˆ a a† (p)ˆ(p) + a(p)ˆ† (p) ˆ (2π)3 d3 p ˆ a ˆ a E(p) a† (p)ˆ(p) + a(p)ˆ† (p) .35) (3.

44) It is important to keep in mind that a and a† always commute inside : · · · :. Let us see what 3p happens when the operator (2π)d 2E(p) a† (p)ˆ(p) acts on |k : ˆ a 3 d3 p a a† (p)ˆ(p) |k ˆ (2π)3 2E(p) = = = d3 p a ˆ a† (p)ˆ(p) a† (k) |0 ˆ (2π)3 2E(p) d3 p a† (p) ˆ 3 2E(p) (2π) a(p). In particular we now ﬁnd that ˆ 0|H |0 = 0. (3. In particular.e. This is true ˆ ˆ † for an arbitrary string of a and a .46) ˆ and.47) d3 p ˆ a ˆ a E(p) a† (p)ˆ(p) + a(p)ˆ† (p) : (2π)3 2E(p) d3 p ˆ a E(p) a† (p)ˆ(p). which appears in its deﬁnition. |k . 0| : H : |0 = 0. 3 2E(p) (2π) (3. The normal ordered Hamiltonian thus produces a sensible result for the vacuum energy. (3. Consider ˆˆ a one-meson state. then it is easy to see that Ψ |H vac |Ψ = 0. of a particle with mass m and momentum k. we shall be interested in the combination a† a. A simple way to remove the vacuum contribution is to introduce normal ordering. a† (k) + a† (k)ˆ(p) |0 ˆ ˆ ˆ a (3. ˆˆ ˆˆ i.Here the subtraction of the vacuum energy is shown explicitly. For instance : 1 2 (3. in particular. Normal ordering means that all annihilation operators appear to the right of any creation operator. a† (p) |0 ˆ ˆ ˆ ˆ . a† (p) − 0| a(p). ˆ 23 . (3. we ﬁnd that ˆ ˆ Ψ | : H : |Ψ = Ψ |H |Ψ . and we can rewrite is as ˆ H = 1 2 + 1 = 2 1 2 d3 p E(p)ˆ† (p)ˆ(p) a a (2π)3 2E(p) d3 p E(p) (2π)3 2E(p) a(p). With this deﬁnition we can write the normal-ordered ˆ ˆ Hamiltonian as 1 ˆ :H: = : 2 = Hence.43) a† (p)ˆ(p) + a(p)ˆ† (p) : = a† (p)ˆ(p).41) d3 p ˆ ˆ a E(p) a† (p)ˆ(p) + H vac 3 2E(p) (2π) ˆ The operator H vac ensures that the vacuum energy is properly subtracted: if |Ψ and |Ψ ˆ denote arbitrary Fock states. The notation is : aa† : = a† a.42) as it should be. ˆ a ˆ a ˆ a (3. Let us now look at the interpretation of the Hamiltonian in more detail. the normal-ordered operators are enclosed within colons.45) ˆ d3 p a† (p) δ 3 (p − k)|0 = a† (k)|0 = |k .

. Start from the classical Lagrangian density L(φ. k . This is ﬁnally the solution of the negative energy problem encountered earlier.49) returns the number of mesons in a given Fock state. .50) Note that the normal-ordered Hamiltonian diﬀers from the number operator just by a factor of E(p). One can easily generalise this to a system of n mesons of equal mass but unequal momenta k 1 . . .55) . 2 2 φ(x) : classical ﬁeld (3. ∂µ φ) = 1 (∂µ φ∂ µ φ) − 1 m2 φ2 . . . . (3. ˆ a (2π)3 2E(p) At this point the pattern becomes clear: the operator d3 p a a† (p)ˆ(p) ˆ 3 2E(p) (2π) (3. . If we work out the energy of an n-meson state we get ˆ : H : |k. (2π)3 2E(p) (3. . k = 2 |k. i. 3. . k n .48) (3. ∂µ φ) π(x) = 24 (3. ∂(∂0 φ(x)) H = π(x)(∂0 φ(x)) − L(φ. . (3. This is easily generalised to n-meson states.54) 3. + E(k n )} |k 1 . . k = nE(k) |k. .e.53) 2. one can derive d3 p a a† (p)ˆ(p) |k. . . Let us now a ˆ 3 look at a two meson state |k. . k . ˆ (2π)3 2E(p) n momenta n momenta (3. . k n = {E(k 1 ) + . . . and since E(k) > 0. . k = n |k. . ∂φ ∂(∂µ φ) ⇒ + m2 φ(x) = 0. k = d3 p a a E(p)ˆ† (p)ˆ(p) |k. Since the number operator has strictly non-negative eigenvalues. . . .p We ﬁnd that |k is an eigenstate of (2π)d 2E(p) a† (p)ˆ(p) with eigenvalue 1. .4 Summary: canonical quantisation for real scalar ﬁelds 1. k . .51) where E(k) = k 2 + m2 > 0. . . . k . Using similar manipulations as above we ﬁnd that 3 d3 p a† (p)ˆ(p) |k. . . the energy of any Fock state cannot be negative. k n : ˆ : H : |k 1 .52) ˆ We conclude that the Hamiltonian : H : returns the total energy of an n-meson state. . and is therefore referred to as the number operator. Deﬁne a conjugate momentum π(x) and the Hamiltonian density by ∂L = ∂0 φ(x). . This implies that the energy of the n-meson state is positive. The Principle of least Action implies the equation of motion ∂L ∂L − ∂µ = 0. .

ˆ a (2π)3 (3. let us note that the term “canonical quantisation” refers to the imposition of commutation relations among ﬁeld operators. 2 Use this result to express the Hamiltonian 1 ˆ H= 2 d3 x ˆ ˆ ˆ ∂0 φ)2 + ( φ)2 + m2 φ2 of the quantised theory in terms of creation and annihilation operators and show that it is given by 1 ˆ H= 2 3. 5. (3. π (y. The energy of a given Fock state is given by the normal-ordered Hamiltonian.2 Show that the Hamiltonian density H for a free scalar ﬁeld is given by H= 1 (∂0 φ)2 + ( φ)2 + m2 φ2 . More details can be found in standard textbooks on Quantum Field Theory. π (y. H(x.4.3 Verify eq. which is actually more ﬂexible. t) = iδ 3 (x − y) ˆ ˆ φ(x. p) ≡ px − L(x. ˙ =x ˙ ∂x ∂p 3. which are used to generate ˆ ˆ ˆ a basis of Fock states. is based on the path integral formulation. d3 p E(p) a† (p)ˆ(p) + a(p)ˆ† (p) ˆ a ˆ a (2π)3 2E(p) 25 . Quantise the theory by regarding φ(x). t). 1 ˆ :H:= 2 d3 p † a (p)ˆ(p). t). ˆ ˆ (3. t) = π (x. x). π(x) as ﬁeld operators with equal time commutators ˆ ˆ φ(x. This approach is widely used for the quantisation of non-Abelian gauge theories. φ(y. As a ﬁnal remark in this section.57) acting on that state. Normal ordering ensures that the vacuum energy is subtracted and that the total energy of a given Fock state is non-negative.56) ˆ The operators φ and π are expanded in terms of a† and a. derive ˙ ˙ Hamilton’s equations ∂H ∂H = −p.1 Starting from the deﬁnition of the Hamiltonian. but a detailed discussion is beyond the scope of these lectures. Problems 3. t).50). t) = 0. An alternative method of quantisation.

We shall leave the details of Lint unspeciﬁed for the moment. .3) which follows from its deﬁnition. thereby producing a number of particles in the ﬁnal state. The Hamiltonian density of the interaction is related to Lint simply by Hint = −Lint . The full Lagrangian L is given by L = L0 + Lint where 1 L 0 = 2 ∂ µ φ ∂ µ φ − 1 m2 φ2 2 (4.4 Using the expressions for φ and π in terms of a and a† . p1 p2 Figure 2: k1 k2 kn Scattering of two initial particles with momenta p1 and p2 into n particles with momenta k 1 . This is schematically shown in Fig. . What we will be concerned with mostly are scattering processes. (4. . 26 . Our task is to ﬁnd a description of such a scattering process in terms of the underlying quantum ﬁeld theory.ˆ 3. in which two initial particles with momenta p1 and p2 scatter. π (x ) = ˆ 2 d3 p eip·(x−x ) + e−ip·(x−x ) 3 (2π) Show that for t = t one recovers the equal time commutator ˆ φ(x. It is then convenient to drop the “hats” on the operators that we have considered up to now So far we have only discussed free ﬁelds without any interaction between them. t) = iδ 3 (x − x ) ˆ 4 Interacting scalar ﬁelds From now on we shall always discuss quantised real scalar ﬁelds. t).2) is the free Lagrangian density discussed before. show that the unequal time ˆ ˆ ˆ ˆ commutator φ(x). . k n in the ﬁnal state. characterised by momenta k 1 . 2. π (x ) is given by ˆ i ˆ φ(x). π (x . . kn . As this does not make for a very interesting theory. . let us now add an interaction Lagrangian Lint .1) (4. . .

13) .6) Operators describing free ﬁelds can be expressed as a superposition of plane waves (see eq. i. the a† ’s and a† ’s generate two diﬀerent bases of the Fock ˆin space. in = |0. the incoming particles evolve independently and freely.9) we have assumed that there is a stable and unique vacuum state: |0 = |0. k n .11) Here |in and |out denote generic “in” and “out” states. Since the physics that we want to describe must be independent of the choice of basis.5) The ﬁelds φin .e.8) (4. = |k 1 . and the corresponding operator is t→+∞ lim φ(x) = φout (x). They both satisfy the free Klein-Gordon equation. (4. Long before t. and therefore we ˆ in ˆout have labelled the Fock states accordingly. We can relate the two bases by introducing a unitary operator S such that φin (x) = S φout (x) S † |in = S |out .12) (4. (4.10) ˆout Mathematically speaking. (4. . Note that the operators a† and a also ˆ ˆ carry subscripts “in” and “out”. . out . In eqs. The scattering (interaction) process takes place during a short interval around some particular time t with −∞ t ∞. φout are the asymptotic limits of the Heisenberg operator φ. |out = S † |in . (4. (4. (4. (4. in . ( + m2 )φin (x) = 0. .4) Long after the collision the particles in the ﬁnal state evolve again like in the free theory.1 The S-matrix The timescales over which interactions happen are extremely short.4.28)).7) with an entirely analogous expression for φout (x). For instance a† (p1 ) a† (p2 )|0 ˆin ˆin a† (k 1 ) · · · a† (k n )|0 ˆout ˆout = |p1 .8) and (4. ˆin (4.9) We must now distinguish between Fock states generated by a† and a† . They are described by a ﬁeld operator φin deﬁned through t→−∞ lim φ(x) = φin (x). Thus. 27 S † S = 1. We can now use the creation operators a† and a† to build up Fock states from the ˆin ˆout vacuum. out . expectation values expressed in terms of “in” and “out” operators and states must satisfy in| φin (x) |in = out| φout (x) |out . . (4. for φin we have φin (x) = d3 k (2π)3 2E(k) ˆ eik·xa† (k) + e−ik·xain (k) . p2 . ( + m2 )φout (x) = 0. (2.

e. i. t) = φI (x. H = H0 + Hint . What we are ultimately interested in are transition amplitudes between an initial state i of. in ≡ Sﬁ .16) Note that this relation involves the full Hamiltonian H = H0 + Hint in the interacting theory. S = 1. The transition amplitude is then given by f. where one tries to infer the nature of the interaction by studying the products of the scattering of particles that have been collided with known energies. All information about the scattering is thus encoded in the Smatrix. t) = eiH0 t φS e−iH0 t . which is a measurable quantity is then proportional to |Sﬁ |2 . By contrast. and in the Schr¨dinger picture.e. Hint = 0. (4. 4. before we try to derive the relation between S and Hint we have to take a slight detour. in| S |i. for Lint = 0 the distinction between φin and φout is not necessary. We have already found solutions to the Klein-Gordon equation in the free theory. t) which we have encountered are Heisenberg ﬁelds and thus time-dependent.14) By comparing “in” with “out” states one can extract information about the interaction – this is the very essence of detector experiments. say. and then the relation between diﬀerent bases of the Fock space becomes trivial.15) The S-matrix element Sﬁ therefore describes the transition amplitude for the scattering process in question. The relation 28 . Nevertheless. in the absence of any interaction. As we will see below. It now turns out to be useful to introduce a new quantum picture for the interacting theory. out = f. t) and π(x. This is the so-called Dirac or Interaction picture. The relation between ﬁelds in the Interaction picture. However. which must therefore be closely related to the interaction Hamiltonian density Hint . out| i. and so we know how to handle time evolution in this case. The scattering cross section. and a ﬁnal state f .e.2 More on time evolution: Dirac picture The operators φ(x.17) At t = −∞ the interaction vanishes. and hence the ﬁelds in the Interaction and Heisenberg pictures are identical. in which the time dependence is governed by H0 only. ˆin ˆout ain = S aout S † . as one would expect. p2 . Note that the plane wave solutions of φin and φout also imply that a† = S a † S † . t) for t → −∞. ˆ ˆ (4. They can thus be identiﬁed. i. φS . For instance. (4. state vectors in the Heisenberg picture can carry a time label. i. whereas the ﬁnal one must be an “out” state. for instance n particles of unequal momenta. in = f. is given o by φI (x. two particles of momenta p1 . the “in”-states of the previous subsection are deﬁned at t = −∞. t) = eiHt φS e−iHt .S is called the S-matrix or S-operator. this information is contained in the elements of the S-matrix. Obviously the initial state is characterised by an “in” state. φI . The state vectors are time-independent in the sense that they do not satisfy a non-trivial equation of motion. (4. φH (x. out| S |i. The relation of the Heisenberg operator φH (x) with its counterpart φS in the Schr¨dinger picture is given by o φH (x.

i. We regard t0 merely as a time label and ﬁx it at −∞. t) contains the information about the interaction.e. t) = φH (x.21) and (4. (4. they always act in the basis of “in” vectors. 29 .12) we then read oﬀ that t→∞ (4. An important boundary condition for U (t) is t→−∞ lim U (t) = 1.24) lim U (t) = S.22) yield φin = U (∞) φout U † (∞).21) Furthermore.20) (4. measured relative to t0 = −∞.22) The relation between φI and φH at any time t is given by φI (x. t).e. t) ≡ φH (x.19) The ﬁeld φH (x. −∞ < t < ∞. we can now make the following identiﬁcations: t → −∞ : φin (x. (4. such that φin (x. which means that in this limit there is no distinction between Heisenberg and Dirac ﬁelds. t) = U (t) φH (x. (4. t) = eiHt φS e−iHt = eiHt e−iH0 t eiH0 t φS e−iH0 t eiH0 t eiH0 t e−iHt φI (x. t) = φI (x. t).18) where we have introduced the unitary operator U (t) U (t) = eiH0 t e−iHt . which tells us how operators and state vectors deviate from the free theory at time t. where the interaction vanishes. t0 ). we write U (t) instead of U (t. (4. (4. U † U = 1. From the deﬁnition of the S-matrix.between φH and φI can be worked out easily: φH (x.26) then simply states that U becomes unity as t → t0 . since the ﬁelds φI evolve over time with the free Hamiltonian H0 . i. (4. t) = φI (x.23) (4. since it evolves over time with the full Hamiltonian. t) U (t).25) We have thus derived a formal expression for the S-matrix in terms of the operator U (t). t). long before the interaction process. t → +∞ : φout (x. Equation (4. As t → ∞ the identiﬁcations of eqs. eq.26) What we mean here is the following: the operator U actually describes the evolution relative to some initial time t0 . (4. which we will normally suppress. In order to describe the “in” and “out” ﬁeld operators.t) = U −1 (t) φI (x. t) U −1 (t). (4.

dt Hint (t) = eiH0 t Hint e−iH0 t . The time-ordered product of two ﬁelds φ(t1 ) and φ(t2 ) is deﬁned as T {φ(t1 ) φ(t2 )} = φ(t1 )φ(t2 ) t1 > t2 φ(t2 )φ(t1 ) t1 < t2 ≡ θ(t1 − t2 ) φ(t1 )φ(t2 ) + θ(t2 − t1 ) φ(t2 )φ(t1 ). such that the resulting expressions still obey causality. . (4.29) where t2 < t1 < t. and depends on the ﬁelds φin . This procedure can be iterated further.Using the deﬁnition of U (t). (4. Let us now solve the equation of motion for U (t) with the boundary condition lim U (t) = 1. i=1 −∞ (4. πin in the “in” basis.31) where θ denotes the step function. (4. were it not for the fact that the upper integration bounds were all diﬀerent. (4. −∞ ⇒ U (t) = 1 − i (4. (4. Integrating eq. and that the ordering tn < tn−1 < . but we can substitute our new expression for U (t) into the integrand. . it is an easy exercise to derive the equation of motion for U (t): i d U (t) = Hint (t) U (t).28) The rhs. so that the nth term in the sum is t t1 tn−1 (−i)n −∞ dt1 −∞ dt2 · · · −∞ dtn Hint (t1 ) Hint (t2 ) · · · Hint (tn ). eq.27) gives t→−∞ t −∞ d U (t1 ) dt1 = −i dt1 t Hint (t1 ) U (t1 ) dt1 −∞ t U (t) − U (−∞) = −i −∞ t Hint (t1 ) U (t1 ) dt1 Hint (t1 ) U (t1 ) dt1 . thereby ensuring the causality of the theory.27) The time-dependent operator Hint (t) is deﬁned in the interaction picture. (4. (4. Time ordering is an important issue.30) This iterative solution could be written in much more compact form. since one has to ensure that the interaction Hamiltonians act at the proper time. which gives t t1 U (t) = 1 − i −∞ t Hint (t1 ) 1 − i −∞ t Hint (t2 ) U (t2 ) dt2 dt1 t1 = 1−i −∞ Hint (t1 )dt1 − −∞ dt1 Hint (t1 ) −∞ dt2 Hint (t2 ) U (t2 ).30) we obtain (−i)n n! n t dti T {Hint (t1 ) Hint (t2 ) · · · Hint (tn )} . By introducing the time-ordered product of operators. one can use a compact notation. < t1 < t had to be obeyed.19). Using time ordering for the nth term of eq. The generalisation to products of n operators is obvious.32) 30 . still depends on U .

As we shall see. out φ(x1 ) p2 . k 2 .3 S-matrix and Green’s functions The S-matrix. k n . out φin (x1 ) p2 . out a† (p1 ) p2 . k n . . . diﬀerentiable function f (t). On the other hand. The following identities will prove useful a† (p) = i ˆ ≡ −i a(p) = −i ˆ ≡ i d3 x ∂0 e−iq·x φ(x) − e−iq·x (∂0 φ(x)) ←→ d3 x e−iq·x ∂0 φ(x). We can now rewrite limt1 →−∞ using the following identity. d3 x ∂0 eiq·x φ(x) − eiq·x (∂0 φ(x)) ←→ (4. k n . p2 . . in .e. in . . Our task is now to in express a† in terms of φin . . (4.and since this looks like the nth term in the series expansion of an exponential. (4. (4.38) where in the last line we have used eq. which can be computed in various ways. and repeat this procedure for all other momenta labelling in our Fock states. k 2 . 4. the non-trivial case is described by the T -matrix. which relates the “in” and “out” ﬁelds before and after the scattering process. . in k 1 . this will relate the S-matrix element Sﬁ to the vacuum expectation value of ﬁelds. out p1 . . Let us consider again the scattering process depicted in Fig. (4. a Green function. . . . The fact that S contains the unit operator means that also the case where none of the particles scatter is encoded in S. can be written as S = 1 + iT. . . we can ﬁnally rewrite the solution for U (t) in compact form as t U (t) = T exp −i −∞ Hint (t ) dt .35) where a† is the creation operator pertaining to the “in” ﬁeld φin . (4.37) The S-matrix element can then be rewritten as Sﬁ = −i d3 x1 e−ip1 ·x1 ∂0 t1 →−∞ ←→ k 1 . . The S-matrix element in this case is Sﬁ = = k 1 .33) where the “T ” in front ensures the correct time ordering. k n . which holds for an arbitrary. .39) 31 . .36) d3 x eiq·x ∂0 φ(x).34) where T is commonly called the T -matrix. in (4. and this is what we are interested in. . dt (4.4) to replace φin by φ. whose limit t → ±∞ exists: +∞ t→−∞ lim f (t) = lim f (t) − t→+∞ −∞ df dt. 2. in ←→ = −i lim d3 x1 e−ip1 ·x1 ∂0 k 1 . i.

.45) × 0. . . The above expression is known as the Lehmann-Symanzik-Zimmermann (LSZ) reduction formula.The S-matrix element then reads Sﬁ = −i lim +i −∞ t1 →+∞ +∞ d3 x1 e−ip1 ·x1 ∂0 dt1 ∂ ∂t1 ←→ k 1 . out φ(x1 ) p2 .44) d4 x1 e−ip1 ·x1 x1 + m2 k 1 . out φ(x1 ) p2 . kn . . . out T {φ(y1 ) · · · φ(yn )φ(x1 )φ(x2 )} 0. .42) The time derivatives in the integrand can be worked out: ∂0 ∂0 e−ip1 ·x1 φ(x1 ) − e−ip1 ·x1 (∂0 φ(x1 )) 2 2 2 e−ip1 ·x1 φ(x1 ) − e−ip1 ·x1 ∂0 φ(x1 ) 2 + m2 e−ip1 ·x1 φ(x1 ) + e−ip1 ·x1 ∂0 φ(x1 ) . k n . k n . .(4. k n . The VEV in the LSZ formula for the scattering of two initial particles into n particles in the ﬁnal state is called the (n + 2)-point Green’s function: Gn+2 (y1 . . k n . . . out T {φ(y1 ) · · · φ(yn )φ(x1 )φ(x2 )} 0. . in (4. in .41) out This is non-zero only if p1 is equal to one of k 1 . . . in . . . in . . (4. . .46) . . . . as well as the n momenta of the “out” state. . k n .34). (4. The ﬁnal expression for Sﬁ is Sﬁ = (i)n+2 × x1 d 4 x1 + m2 d 4 x2 x2 d4 y1 · · · y1 d4 yn e(−ip1 ·x1 −ip2 ·x2 +ik1 ·y1 +···+kn ·yn ) + m2 · · · yn + m2 + m2 (4. out φ(x1 ) p2 . It relates the formal deﬁnition of the scattering amplitude to a VEV of time-ordered ﬁelds. . x2 ) = 0. k n .40) The ﬁrst term in this expression involves limt1 →+∞ φ = φout . in . . out ∂0 − + m2 φ(x1 ) p2 . x1 . We are then left with the following expression for Sﬁ : Sﬁ = −i d4 x1 k 1 . out a† (p1 ) p2 . and hence the ﬁrst term does not contribute to the T -matrix of eq. out ∂0 ∂0 e−ip1 ·x1 φ(x1 ) − e−ip1 ·x1 (∂0 φ(x1 )) p2 . yn . . . means that the particle with momentum p1 does not scatter. in . where the time-ordering inside the vacuum expectation value (VEV) ensures that causality is obeyed. . . which gives rise to a contribution ∝ k 1 .43) 2 −ip1 ·x1 = p2 e−ip1 ·x1 . . one of which has been reduced to a one-particle state. . For the S-matrix element one obtains 1 2 2 d4 x1 e−ip1 ·x1 k 1 . however. = − E(p1 ) =− − where we have used that − Sﬁ = i = i e (4. (4. What we have obtained after this rather lengthy step of algebra is an expression in which the ﬁeld operator is sandwiched between Fock states. . in . We can now successively eliminate all momentum variables from the Fock states. 32 (4. in ←→ d3 x1 e−ip1 ·x1 ∂0 k 1 . . y2 . by repeating the procedure for the momentum p2 . This.

−t) = S. xn ) = 0 |T {φ(x1 ) · · · φ(xn )}| 0 .51) One can easily convince oneself that U (t. . i. which we know how to handle.25). . which can be measured in an experiment. The relation between U (t) and the S-matrix eq. The LSZ formula then provides the link to the scattering amplitude Sﬁ . we ﬁrst have to isolate the dependence of the ﬁelds on the interaction Hamiltonian. as well as the boundary condition eq. so that we can forget about writing T (· · ·) everywhere.26) tell us that t→∞ 2 (4.48) into the deﬁnition of Gn one obtains Gn = 0|U −1 (t1 )φin (t1 )U (t1 ) U −1 (t2 )φin (t2 )U (t2 ) · · · × U −1 (tn )φin (tn )U (tn )|0 . (4. U (t.48) We now assume that the ﬁelds are properly time-ordered. An important observation at this point is that it involves pairs of U and its inverse. t1 ) provides the net time evolution from t1 to t. We can now write Gn as Gn = 0 U −1 (t) T φin (t1 ) · · · φin (tn ) U (t. t2 ) · · · U (tn . (4. t1 ). (4. −t) Let us now take t → ∞. (4. −t) U (−t) 0 . (4.In the next section we shall discuss a method how to compute the Green’s function of scalar ﬁeld theory. t1 ) U (t1 . the φ’s are not the φin ’s. This will allow for an explicit calculation of Gn+2 . (4. .e. whose time evolution is governed by the full Hamiltonian H0 + Hint . In order to be useful for actual computations. Before we can tackle the actual computation of Green’s function. and |Sﬁ |2 ﬁnally yields the cross section. (4. . Now we introduce another time label t such that t function we now obtain t1 and −t (4.52) lim U (−t) = 1. t→∞ lim U (t. > tn . for instance U (t)U −1 (t1 ) ≡ U (t. For the n-point Gn = 0 U −1 (t) U (t)U −1 (t1 )φin (t1 )U (t1 ) U −1 (t2 )φin (t2 )U (t2 ) · · · × U −1 (tn )φin (tn )U (tn )U −1 (−t) U (−t) 0 . .47) The ﬁelds φ which appear in this expression are Heisenberg ﬁelds. t1 > t2 > . . Let us consider the n-point Green’s function Gn (x1 . Let us recall the relation between the Heisenberg ﬁelds φ(t) and the “in”-ﬁelds2 φ(t) = U −1 (t) φin (t) U (t).49) t1 . After inserting eq. (4. 33 . In particular.53) Here and in the following we suppress the spatial argument of the ﬁelds for the sake of brevity. we must take a further step.50) The expression in curly braces is now time-ordered by construction.

At the end we will derive a set of rules.56) (4.2 Derive eqs.36) and (4. . 5 Perturbation Theory In this section we are going to calculate the Green’s functions of scalar quantum ﬁeld theory explicitly. xn ) = The S-matrix is given by +∞ 0| T {φin (x1 ) · · · φin (xn ) S} |0 . . (4. (4. eq.which can be inserted into the above expression. 0|S|0 (4. Hint = Hint (φin . . In a paper by Gell-Mann and Low it was argued that the time evolution operator must leave the vacuum invariant (up to a phase). 0|U (∞)|0 (4. This completes the derivation of a relation between the general deﬁnition of the scattering amplitude Sﬁ and the VEV of time-ordered “in”-ﬁelds. We will specify the interaction Lagrangian in detail and use an approximation known as perturbation theory. which justiﬁes the ansatz 0|U −1 (∞) = K 0|.58) S = T exp −i −∞ Hint (t) dt . (4. πin ). 0|S|0 (4. We still have to work out the meaning of 0|U −1 (∞) in the expression for Gn . . The link between the scattering amplitude and the underlying ﬁeld theory is provided by the n-point Green’s function. Multiplying this relation with |0 from the right gives 0|U −1 (∞)|0 = K 0|0 = K.59) and thus we have ﬁnally succeeded in expressing the n-point Green’s function exclusively in terms of the “in”-ﬁelds.1 Using the deﬁnition U (t) = eiH0 t e−iHt derive the equation of motion for U (t).54) with K being the phase. Problems 4.27). (4. 4.57) After inserting all these relations into the expression for Gn we obtain Gn (x1 . Gell-Mann and Low showed that 0|U −1 (∞)|0 = which implies K= 1 .55) 1 . which 34 .37). Furthermore.

Secondly. This only makes sense if λ is suﬃciently small. For our theory of a real scalar ﬁeld in four space-time dimensions. 2 2 4! (5. . These are the famous Feynman rules. 3 This is despite the subtraction of the vacuum energy discussed in section 3. our interacting theory should be renormalisable. We start by making a deﬁnite choice for the interaction Lagrangian Lint . and can be easily generalised to apply to other. the second issue will not be addressed in these lectures. In other words. the procedure of expanding Green’s functions in powers of the coupling is referred to as perturbation theory. The full Lagrangian of the theory then reads L = L0 + Lint = 1 λ 1 (∂µ φ)2 − m2 φ2 − φ4 . The requirement of renormalisability arises because if one computes quantities like the energy or charge of a particle. one typically obtains a divergent result 3 . Thus we choose λ Lint = − φ4 (x). . it turns out that the only interaction term which leads to a renormalisable theory must be quartic in the ﬁelds. one has to obey some basic principles: ﬁrstly. Although one may think of many diﬀerent expressions for Lint . . Despite being of great importance.3) d4 x φ4 (x) . xn ) ∞ Hint = λ 4! λ 4! d3 x φ4 (x. the interaction Lagrangian must act as a small perturbation on the system. say. in = r=0 iλ − 4! r 1 r! r 0 T iλ − 4! r φin (x1 ) · · · φin (xn ) 1 r! 0 T d 4 d 4 y φ4 (y) in r 0 .4) ∞ y φ4 (y) in 0 r=0 In order to evaluate this expression we must expand Gn in powers of the coupling λ. the electric charge describing the strength of the interaction between photons and electrons. S = T exp −i The n-point Green’s function is Gn (x1 . (5. t) in (5. in which these divergences can be removed by suitable redeﬁnitions of the ﬁelds and the parameters (masses and coupling constants). .2) and the explicit expressions for the interaction Hamiltonian and the S-matrix are Hint = −Lint . 35 .1) 4! where the coupling constant λ describes the strength of the interaction between the scalar ﬁelds. (5. Lint must be chosen such that the potential it generates is bounded from below – otherwise the system has no ground state. There are classes of quantum ﬁeld theories. As a consequence. more complicated ﬁeld theories. called renormalisable. much like.represent a systematic prescription for the calculation of Green’s functions.

i.5. For the case of three ﬁelds. however. (5.4) involves the time-ordered product over at least n ﬁelds. This is known as Wick’s theorem. The timeordered product T {φ(x1 )φ(x2 )} is given by T {φ(x1 )φ(x2 )} = φ(x1 )φ(x2 )θ(t1 − t2 ) + φ(x2 )φ(x1 )θ(t2 − t1 ) = : φ(x1 )φ(x2 ) : θ(t1 − t2 ) + θ(t2 − t1 ) (5. 0|T {φin (x1 ) · · · φin (xn )} |0 in terms of VEV’s involving two ﬁelds only.. +. i. + : φ(x1 ) · · · φ(xi ) · · · φ(xj ) · · · φ(xk ) · · · φ(xl ) · · · φ(xn ) : × 0|T {φ(xi )φ(xj )}|0 0|T {φ(xk )φ(xl )}|0 + perms. (5. Here we have used the important observation that : φ(x1 )φ(x2 ) : = : φ(x2 )φ(x1 ) :.9) At this point the general pattern becomes clear: any time-ordered product of ﬁelds is equal to its normal-ordered version plus terms in which pairs of ﬁelds are removed from the normal-ordered product and sandwiched between the vacuum to form 2-point functions..5) We are now going to combine normal-ordered products with time ordering.e. (5.6) + 0|φ(x1 )φ(x2 )θ(t1 − t2 ) + φ(x2 )φ(x1 )θ(t2 − t1 )|0 .1 Wick’s Theorem The n-point Green’s function in eq. Let us for the moment ignore the subscript “in” and return to the deﬁnition of normal-ordered ﬁelds.+ + 0|T {φ(x1 )φ(x2 )}|0 0|T {φ(x3)φ(x4 )}|0 · · · 0|T {φ(xn−1 )φ(xn )}|0 + perms.8) (5. 4 (5.10) The reverse is. φ(x1 )φ(x2 ) = : φ(x1 )φ(x2 ) : + 0|φ(x1 )φ(x2 )|0 . There is a method to express VEV’s of n ﬁelds. Then one sums over all permutations.4 Equation (5.6) is Wick’s theorem for the case of two ﬁelds: T {φ(x1 )φ(x2 )} = : φ(x1 )φ(x2 ) : + 0|T {φ(x1 )φ(x2 )} |0 .. Wick’s theorem yields T {φ(x1 )φ(x2 )φ(x3 )} = : φ(x1 )φ(x2 )φ(x3 ) : + : φ(x1 ) : 0|T {φ(x2 )φ(x3 )}|0 + : φ(x2 ) : 0|T {φ(x1 )φ(x3 )}|0 + : φ(x3 ) : 0|T {φ(x1 )φ(x2 )}|0 (5.7) which means that normal-ordered products of ﬁelds are automatically time-ordered.e. Without proof we give the expression for the general case of n ﬁelds (n even): T {φ(x1 ) · · · φ(xn )} = : φ(x1 ) · · · φ(xn ) : + : φ(x1 ) · · · φ(xi ) · · · φ(xj ) · · · φ(xn ) : 0|T {φ(xi )φ(xj )}|0 + perms. The normal-ordered product : φ(x1 )φ(x2 ) : diﬀers from φ(x1 )φ(x2 ) by the vacuum expectation value. not true! 36 .

If we assume tx > ty then ˆ ˆ G2 (x. y) = 0|T {φin (x)φin (y)}|0 . where two of the ﬁelds are also normalordered: 0|T {φ(x1 )φ(x2 ) : φ(x3 )φ(x4 ) :}|0 = φ (x1 )φ(x2 ) : φ (x3 )φ(x4 ) : +φ (x1 )φ (x2 ) : φ(x3 )φ(x4 ) : +φ(x1 )φ (x2 ) : φ(x3 )φ(x4 ) : . though.e. Such contractions contribute only to the vacuum. (5.The symbol φ(xi ) indicates that φ(xi ) has been removed from the normal-ordered product.12) The second example is again a 4-point function. y) can be written as G2 (x. 5. The only surviving contribution is that in which all ﬁelds have been paired or “contracted”. the contraction of : φ(x3 )φ(x4 ) : vanishes by construction.13) (5. Let us now go back to 0|T {φ(x1 ) · · · φ(xn )}|0 . +φ (x1 )φ (x2 )φ(x3 )φ(x4 ) + φ(x1 )φ (x2 )φ(x3 )φ(x4 ) (5. Wick’s theorem can now be rephrased as 0|T {φ(x1 ) · · · φ(xn )}|0 = sum of all possible contractions of n ﬁelds.10) survives: by deﬁnition the VEV of a normal-ordered product of ﬁelds vanishes. If we insert Wick’s theorem. as we shall see explicitly below. contractions which only involve ﬁelds inside a normal-ordered product vanish. then we ﬁnd that only the contribution in the last line of eq. Let us look at a few examples. Normal ordering can therefore simplify the calculation of Green’s functions quite considerably. Let us have a closer look at G2 (x. so only the last two terms survive! As a general rule. and it is precisely the last line of Wick’s theorem in which no normal-ordered products are left.11) i. (5. y) = d3 p d 3 q (2π)6 4E(p)E(q) × 0 a† (p) eip·x + a(p) e−ip·x ˆ ˆ = a† (q) eiq·y + a(q) e−iq·y ˆ ˆ 0 (5.2 The Feynman propagator Using Wick’s Theorem one can relate any n-point Green’s functions to an expression involving only 2-point functions.15) We can now insert the solution for φ in terms of a and a† . The ﬁrst is the 4-point function 0|T {φ(x1 )φ(x2 )φ(x3 )φ(x4 )}|0 = φ (x1 )φ(x2 )φ (x3 )φ(x4 ) . (5. the pair of ﬁelds which is contracted is joined by the braces. Sometimes a contraction is represented by the notation: φ (xi )φ(xj ) ≡ 0|T {φ(xi )φ(xj )}|0 . (5. ˆ a (2π)6 4E(p)E(q) 37 .14) In this example.16) d3 p d 3 q e−ip·x+iq·y 0 a(p)ˆ† (q) 0 .

a† (q) 0 ˆ ˆ 6 4E(p)E(q) (2π) d3 p e−ip·(x−y) .21) x+m and describes the propagation of a meson between the space-time points x and y. Equation (5.p1 p2 k1 k2 Figure 3: Scattering of two initial particles with momenta p and p into 2 particles with 1 2 momenta k 1 and k 2 . p2 . in . (2π)3 2E(p) (5. y) actually depends only on the diﬀerence (x − y). (5. (2π)3 2E(p) (5. using contour integration one can show that this expression can be rewritten as a 4-dimensional integral G2 (x. in|S|p1 . (2π)4 p2 − m2 + i (5. k 2 . 3. We can now replace a(p)ˆ † (q) by its commutator: ˆ a G2 (x.18) Furthermore. out|p1 . reads G2 (x.22) . it satisﬁes 2 GF (x − y) = −iδ 4 (x − y). 5. y) = d3 p e−ip·(x−y) θ(tx − ty ) + eip·(x−y) θ(ty − tx ) . in k 1 . as shown in Fig.3 Two-particle scattering to O(λ) Let us now consider a scattering process in which two incoming particles with momenta p1 and p2 scatter into two outgoing ones with momenta k 1 and k 2 .19) where is a small parameter which ensures that G2 does not develop a pole. y) = i d4 p eip·(x−y) .17) and the general result.19) is called the Feynman propagator GF (x − y): GF (x − y) ≡ 0|T {φ(x)φ(y)}|0 = i eip·(x−y) d4 p . k 2 . i. after restoring time-ordering. y) = = d3 p d 3 q e−ip·x+iq·y 0 a(p).20) The Feynman propagator is a Green’s function of the Klein-Gordon equation. 38 (5. The S-matrix element in this case is Sﬁ = = k 1 . (2π)4 p2 − m2 + i (5.e. p2 . This calculation has established that G2 (x. This shows that G2 can be interpreted as the amplitude for a meson which is created at y and destroyed again at point x.

Therefore denominator = 1 to order λ. Let us work out G4 in powers of λ using Wick’s theorem. hence r=1: denominator = 0.23) Suppressing the subscripts “in” from now on. i.25) If r = 1. . . 39 . so r = 0 in the numerator does not describe scattering and is hence not a contribution to the T -matrix. then the expression in the denominator only involves ﬁelds which are normalordered.28) But this is the same answer as if we had set λ = 0. S = T exp −i λ 4! d4 x : φ4 (x) : in (5. is non-zero.26) The contribution for r = 2. we note that for r = 0 one ﬁnds r=0: denominator = 1. we start with r = 0 and apply Wick’s theorem. the expression we have to evaluate order by order in λ is Gn (x1 . . The LSZ formula eq.1 we conclude that these contributions must vanish. To make life simpler. we shall introduce normal ordering into the deﬁnition of S. (5. Following the discussion at the end of section 5.24) iλ 4! r − = r=0 1 r! ∞ r 0 T iλ − 4! φ(x1 )φ(x2 )φ(x3 )φ(x4 ) r d4 y : φ4 (y) : r 4 0 . which gives r=0: 0|T {φ(x1 )φ(x2 )φ(x3 )φ(x4 )}|0 = GF (x1 − x2 ) GF (x3 − x4 ) + GF (x1 − x3 ) GF (x2 − x4 ) + GF (x1 − x4 ) GF (x2 − x3 ). But then the case of r = 2 corresponds already to O(λ2 ).and S = 1 + iT . xn ) ∞ (5. which can be graphically represented as x1 x3 + x2 x4 x2 x4 x1 x3 + x2 x4 x1 x3 (5. (5.e.27) Turning now to the numerator. however. which is higher than the order which we are working to. (5. (4.45) tells us that we must compute G4 in order to obtain Sﬁ . . r=0 1 r! 0 T d y : φ (y) : 4 0 Starting with the denominator.

. . Such contributions are corrections to the vacuum and are cancelled by the denominator. . the ﬁnal answer for the scattering amplitude to O(λ) is given by eq.29) d4 y 4! GF (x1 − y)GF (x2 − y)GF (x3 − y)GF (x4 − y)..29).For r = 1 in the numerator we have to evaluate r=1: iλ 4! iλ = − 4! − 0 T φ(x1 )φ(x2 )φ(x3 )φ(x4 ) : d4 y φ4 (y) : 0 (5.4 Graphical representation of the Wick expansion: Feynman rules We have already encountered the graphical representation of the expansion of Green’s functions in perturbation theory after applying Wick’s theorem. To summarise. Without normal ordering we would have encountered the following contributions for r = 1: x1 x3 + x2 x4 x2 x4 x1 x3 +. . The factor 4! inside the integrand is a combinatorial factor: it is equal to the number of permutations which must be summed over according to Wick’s theorem and cancels the 4! in the denominator of the interaction Lagrangian. This demonstrates how normal ordering simpliﬁes the calculation by automatically subtracting terms which do not contribute to the actual scattering process. It is possible to formulate a simple set of rules which allow to draw the graphs directly without using Wick’s theorem and to write down the corresponding algebraic expressions. xn ). We again consider a neutral scalar ﬁeld whose Lagrangian is 1 1 λ L = ∂µ φ∂ µ φ − m2 φ2 − φ4 . This is achieved by going through the following steps: (1) Draw all distinct diagrams with n external lines and m 4-fold vertices: 40 . where we have taken into account that contractions involving two ﬁelds inside : · · · : vanish. 5.30) Suppose now that we want to compute the O(λm ) contribution to the n-point Green’s function Gn (x1 . (5. 2 2 4! (5.. Graphically this contribution is represented by x1 −iλ d4 y x2 y x4 x3 where the integration over y denotes the sum over all possible locations of the interaction point y.

. Numerically it has the same value as the above graph. namely the 2-point function. ym (vertices) • Join the dots according to the following rules: – only one line emanates from each xi – exactly four lines run into each yj – the resulting diagram must be connected. there must be a continuous path between any two points. According to the Feynman rules the contributions up to order λ2 are as follows: O(1): x1 x2 = GF (x1 − x2 ) O(λ): x1 y y2 x1 O(λ2 ): x1 y1 y1 =C − iλ 4! x2 y 2 x2 2 x2 “tadpole diagram”. but with the positions of y1 and y2 interchanged is topologically distinct. Note that the same graph. These are the Feynman rules for scalar ﬁeld theory in position space. . . (cancelled by normal ordering) d4 y1 d4 y2 GF (x1 − y1 ) [GF (y1 − y2 )]3 GF (y2 − x2 ) The combinatorial factor for this contribution is worked out as C = 4 · 4!. i.e. . . (cancelled by normal ordering) O(λ2 ): “tadpole diagram”. . Let us look at an example. (2) Assign a factor − iλ 4! d4 yi to the vertex at yi (3) Assign a factor GF (xi − yj ) to the line joining xi and yj (4) Multiply by the number of contractions C from the Wick expansion which lead to the same diagram. Another contribution at order λ2 is 41 . . xn (external points) • Draw m dots and label them y1 .• Draw n dots and label them x1 . and so the corresponding expression has to be multiplied by a factor 2. .

but only list them as a recipe. pn ). (5. .+ipn ·xn e Gn (p1 . However. not connected This contribution must be discarded.32) This can be incorporated into the deﬁnition of the momentum space Green’s function one is trying to compute: n Gn (p1 . the diagrams are connected and are not simply vacuum contributions. . . since not all of the points are connected via a continuous line. . .31) The Feynman rules then serve to compute the Green’s function Gn (p1 ..5 Feynman rules in momentum space It is often simpler to work in momentum space. They must hence be included in the expression for the 2-point function. If one works in momentum space. unlike the case of the 4-point function. . i=1 (5. the Green’s functions are related to those in position space by a Fourier transform Gn (x1 . pn ). xn ) = d4 p1 ··· (2π)4 d4 pn ip1 ·x1 +. . Feynman rules (momentum space) (1) Draw all distinct diagrams with n external lines and m 4-fold vertices: • Assign momenta p1 . .O(λ2 ): x1 y1 y2 x2 vacuum contribution. . the corresponding diagrams satisfy the Feynman rules listed above. . . These contributions to the 2-point function are cancelled if the interaction term is normal-ordered. . .33) Here we won’t be concerned with the exact derivation of the momentum space Feynman rules. pn ) order by order in the coupling. and hence we will discuss the derivation of Feynman rules in this case. pn to the external lines • Assign momenta kj to the internal lines 42 . . pn ) = (2π) δ 4 4 i=1 pi Gn (p1 . . . . and hence n pi = 0. Let us end this discussion with a small remark on the tadpole diagrams encountered above. 5. . .. . . . In particular. (2π)4 (5. . . In every scattering process the overall momentum must be conserved.

. which connect the external points to the vertices at z1 . 4 it can be split into Feynman propagators. . y1 . yn ). zn+m ). . out p1 . This further simpliﬁes the calculation of scattering amplitudes using Feynman rules. and using that 2 GF (xi − zi ) = −iδ 4 (xi − zi ) (5. xm . . . pm . . . . . according to Gn+m = d4 z1 · · · d4 zn+m GF (x1 − z1 ) · · · GF (yn − zn+m ) Gn+m (z1 .34) × i=1 + m2 j=1 yj + m2 Gn+m (x1 . . . . . . . . . (5. . . . Gn+m is called the truncated Green’s function. Let us return to the LSZ formalism and consider the scattering of m initial particles (momenta p1 . . pm ) into n ﬁnal particles with momenta k 1 . . i − m2 + i (the delta function ensures that momentum is conserved at each vertex). which is the number of contractions leading to the same momentum space diagram (note that C may be diﬀerent from the combinatorial factor for the same diagram considered in position space!) 5.(2) Assign to each external line a factor p2 k (3) Assign to each internal line a factor d4 kj i 4 k 2 − m2 + i (2π) j (4) Each vertex contributes a factor − iλ (2π)4 δ 4 4! momenta . in m n m n = (i)n+m i=1 m d 4 xi j=1 n xi d4 yj exp −i i=1 p i · xi + i j=1 kj · yj (5. As shown in Fig. . zn+m . . perhaps for obvious reasons. which has its external legs amputated: the so-called truncated Green’s function. .6 S-matrix and truncated Green’s functions The ﬁnal topic in these lectures is the derivation of a simple relation between the Smatrix element and a particular momentum space Green’s function. The LSZ formula tells us that the S-matrix element is given by k 1 .35) where.36) xi + m 43 . . . . . . (5) Multiply by the combinatorial factor C. . yn ). . . . k n . kn . . and a remaining Green’s function Gn+m . . y1 . xm . . . (5.35) back into the LSZ expression for the S-matrix element. . . Let us have a closer look at Gn+m (x1 . Putting eq.

k n . in m n m n = (i)n+m i=1 d 4 xi j=1 d4 yj exp −i i=1 p i · xi + i j=1 kj · yj (5. .38) where G n+m is the truncated n + m-point function in momentum space. . kn ).37) ×(−i)n+m d4 z1 · · · d4 zn+m δ 4 (x1 − z1 ) · · · δ 4 (yn − zn+m ) Gn+m (z1 . in m n m n = i=1 d xi j=1 4 d yj exp −i i=1 4 p i · xi + i j=1 kj · yj × Gn+m (x1 . out p1 . . . . 5. After performing all the integrations over the zk ’s. yn ) ≡ G n+m (p1 . one obtains k 1 . . . . the ﬁnal relation becomes k 1 . pm . out p1 . . . . . k1 .3 Find the expressions corresponding to the following momentum space Feynman diagrams 44 . . .2 Verify that GF (x − y) = i is a Green’s function of (∂ µ ∂µ + m2 ) as eip·(x−y) d4 p (2π)4 p2 − m2 + i → 0 (where ∂µ ≡ ∂/∂xµ ). 5. . . where φ+ and φ− are creation and annihilation components of φ. . . . .1 Verify that : φ(x1 )φ(x2 ) : = : φ(x2 )φ(x1 ) : Hint: write φ = φ+ +φ− . . xm . pm . pm . . The latter can be obtained using the Feynman rules without the expression for the external legs. k n . . . . . . . . This result shows that the scattering matrix element is directly given by the truncated Green’s function in momentum space. (5. . .x1 x2 x3 Figure 4: z1 z2 G z3 The construction of the truncated Green’s function in position space. y1 . . . zn+m ). . . Problems 5. .

Zuber. Acknowledgements I would like to thank Julia Sedgbeer for the invitation and for all her eﬀorts which made the school a very pleasant experience. multi-particle states that live in Fock spaces. This is discussed in the lectures by Robert Thorne at this school. whose persistent questions lead to many clariﬁcations and improvements. References [1] F.Integrate out all the δ-functions but do not perform the remaining integrals. Many thanks go to my fellow lecturers. Itzykson and J. 6 Concluding remarks Although we have missed out on many important topics in Quantum Field Theory. all of which served to overcome the problems we encountered when we na¨ ıvely tried to write down a relativistic quantum theory. The basic ingredients to formulate a quantum theory for a real scalar ﬁeld can be used to quantise the electron and photon ﬁelds. Weinberg. Shaw. CUP 1995 45 .E. [2] C. The Quantum Theory of Fields. creation and annihilation operators. the tutors and all the students. Quantum Field Theory. An Introduction to Quantum Field Theory. which are listed below. 1. Peskin and D. as well as tools like normal ordering. Vol. we got to the point where we established contact between the underlying formalism of Quantum Field theory and the Feynman rules. Addison Wesley 1995 [4] S. The same applies to the method of quantisation via path integrals. Wiley 1984. McGraw-Hill 1987.V. Schroeder. The main concepts of the formulation were discussed: we introduced ﬁeld operators. Mandl and G. Quantum Field Theory.B. and hence I refer the reader to standard textbooks on Quantum Field Theory. Renormalisation is a topic which is not so easily discussed in a relatively short period of time. which are widely used. [3] M. I would also like to thank Margaret Evans for her most friendly and eﬃcient way of organising such an event.

−x) = (t. ∂t ∂2 − ∂t2 2 d’Alembertian: ∂ µ ∂µ = Momentum operator: pµ = i ∂ µ = ˆ δ-functions: i ∂ .− ∂t ∂ . x) xµ = gµν xν = (x0 . x) = (t.A Notation and conventions 4-vectors: xµ = (x0 . −x) 1 0 0 0 0 −1 0 0 Metric tensor: gµν = g µν = 0 0 −1 0 0 0 0 −1 Scalar product: xµ xµ = x 0 x0 + x 1 x1 + x 2 x2 + x 3 x3 = t 2 − x2 Gradient operators: ∂µ ≡ ∂ = ∂xµ ∂ ∂µ ≡ µ = ∂x ∂ . −i ∂t ≡ ˆ ˆ = E. p (as it should be) d3 p f (p) δ 3 (p − q) = f (q) d3 x e−ip·x = (2π)3 δ 3 (p) d3 p −ip·x e = δ 3 (x) (2π)3 (similarly in four dimensions) Note: δ(x2 − x2 ) = δ{(x − x0 )(x + x0 )} 0 1 {δ(x − x0 ) + δ(x + x0 )} = 2x 46 .

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue listening from where you left off, or restart the preview.

scribd