AN INTRODUCTION TO QUANTUM FIELD THEORY
Erik Verlinde
Theory Division, CERN, CH-1211 Geneva 23, Switzerland
Abstract
The basic concepts and ideas underlying quantum eld theory
are presented, together with a brief discussion of Lagrangian eld
theory, symmetries and gauge invariance. The use of the Feyn-
man rules to calculate scattering amplitudes and cross sections is
illustrated in the context of quantum electro dynamics.
1. INTRODUCTION
Quantum eld theory provides a successful theoretical framework for describing
elementary particles and their interactions. It combines the theory of sp ecial relativity
and quantum mechanics and leads to a set of rules that allows one to compute physical
quantities that can b e compared with high energy exp eriments. The physical quantityof
most interest in high energy exp eriments is the cross-section . It determines directly the
number of events that an exp erimentalist will b e able to see, and also is in a direct way
related to the quantum mechanical probabilities that can b e computed with the help of
eld theory.
In these notes I will present the basic concepts and ideas underlying quantum eld
theory. Starting from the quantum mechanics of relativistic particles I will intro duce
quantum eld op erators and explain how these are used to describ e the interactions among
elementary particles. These notes also contain a brief discussion of Lagrangian eld theory,
symmetries and gauge invariance. Finally, after presenting a heuristic derivation of the
Feynman rules, I'll illustrate in the context of quantum electro dynamics how to calculate
the amplitude that determines the cross section for a simple scattering pro cess. But b efore
we turn to the quantum mechanics, let us rst discuss the meaning of the cross section in
more familiar classical terms.
1.1 From Events to Cross Sections
For concreteness, let us consider an exp eriment with two colliding particle b eams in
which one b eam consists of, say, electrons and the other b eam of p ositrons. An exp erimen-
talist can now collect and count the events in which an electron and a p ositron collide and
pro duce several other particles. The number of events N that o ccur in a certain amount
of time dep ends on various exp erimental details, in particular the densities and of
1 2
the particles inside the b eams and of course the particle velo cities ~v and ~v .
1 2
To nd out how these parameters a ect the number of events, let us rst consider
the probability that an event o ccurs inside a a small volume V and within a time
interval t. In this case we can assume that the densities and velo cities are constant.
Now imagine that one of the particles, say the electron, has a nite cross-section as seen
from the p ersp ective of the other particle, the p ositron. After a time t each electron
has moved with resp ect to the p ositrons over a distance j~v ~v jt, where ~v ~v is the
1 2 1 2
relativevelo city, and in this way it has swept out a part of space with a volume equal to
j~v ~v jt. The probability that in this time the electron has collided with a p ositron
1 2
is equal to the numb er of p ositrons inside this volume, whichis j~v ~v jt. Finally,
2 1 2
to obtain the total number N of collisions wehavetomultiply this with the number
of electrons inside the volume V ; this number is V . In this waywe nd that the
1
number of events that o ccur in V and t is
N = j~v ~v jV t (1)
1 2 1 2
In a realistic situation the densities (~x; t) of the particles and their velo cities are not
i
constant in space and time. Instead of the velo city ~v it is also more appropriate to use
~
the current density j that describ es the ux of the particles p er unit of area. It is de ned
~
by j = ~v where ~v is the lo cal velo city at a certain p oint in space and time. By rep eating
the ab ove argumentation we can express the number of events in an in nitesimal volume
3
~ ~
dV and time interval dt as dN = j j j jd ~xdt. Then byintegrating of space and
2 1 1 2
time we obtain the following expression for the total number of events N
Z
N = dt L(t); (2)
1)
where
Z
3
~ ~
L(t)= d xj (~x; t)j (~x; t) (~x; t)j (~x; t)j: (3)
1 2 2 1
All exp erimental details that in uence the number of events are combined in the time
integral of L(t), which is called the integrated luminosity. The integrated luminositydoes
not dep end on the kind of pro cess that one is interested in, and so it can b e measured by
lo oking at a reference pro cess for which the cross section is accurately known. With this
knowledge one can then determine the cross sections for all other pro cesses. These cross-
sections are indep endent of the exp erimental details, and thus can b e used to compare
the data that come from di erent exp eriments. They are also the relevant quantities that
one wishes to compute using quantum eld theory.
1.2 Quantum Mechanical Probabilities
Just to give an idea of how the cross section is related to a quantum mechanical
amplitude, let us assume that b efore the collision the densities and velo cities of the par-
ticles inside the b eams are accurately known. In terms of quantum mechanics this means
1) A more careful analysis shows that the integrand has an additional relativistic con-
1
~ ~ ~ ~
tribution jj j j. This term can b e dropp ed in a situation where j and j are
1 2 1 2
c
(approximately) parallel.
that one has sp eci ed the initial quantum state j i of the colliding particles. Similarly,
i
by measuring the energies and directions of the particles that are b eing pro duced by the
collisions one selects a certain nal wave-function j i that describ es the state of the
f
out-going particles. According to the rules of quantum mechanics the probablity W for
this event to happ en is given by the absolute valued squared of the overlap b etween the
nal state and the initial state
2
W = jh j ij
i!f i f
Quantum eld theory provides the to ols to compute the amplitudes h j i, and hence
i f
determine the quantum mechanical probabilities W . In fact, the quantity that is most
directly computed is the probability W that a collision takes place p er unit of volume
and time for a situation with constant particle densities and .For this situation W
1 2
may b e identi ed with the ratio N=V t, which according to equation (1) gives the
cross section times the pro duct of the densities and relativevelo city. All this will b e
explained in more detail in the following sections.
2. RELATIVITY AND QUANTUM MECHANICS
The elementary particles that are used in high energy exp eriments havevelo cities
close to the sp eed of light, and should therefore b e describ ed by the theory of sp ecial
relativity.For relativistic particles it is convenient to sp ecify their motion, instead by
their velo city ~v , in terms of their momentum
m~v
q
(4) ~p =
~v
2
1 ( )
c
An imp ortant reason for this is that during collision the total momentum of the particles
will b e conserved. The total energy of the particles will also b e conserved. In sp ecial
relativity the energy of a particle with momentum ~p is given by
q
2 2 2 4
E = ~p c + m c (5)
2
For a particle at rest this b ecomes Einstein's famous relation E = mc .
2.1 Relativistic WaveFunctions
In quantum mechanics the energy E b ecomes identi ed with the Hamiltonian op er-
ator H . The time evolution of the wave function for a single particle is, as usual, governed
by the Schrodinger equation in which the Hamiltonian H acts as a di erential op erator.
The form of this di erential op erator can b e read of from (5) by replacing the momentum
~
~
~p by the op erator p = ih r. This gives
ih@ (~x;t)= H (~x; t) (6)
t
with
q
2
2 2 4
~
H = h r + m c : (7)
An imp ortant di erence with the non-relativistic case is, however, that the Hamiltonian
involves a square ro ot. Still the Schrodinger equation can easily b e solved by expanding
the wave-function in terms of plane waves
~ ~
ik~x i! (k )t
(~x;t)=e :
~ k
~
These plane waves are the eigenfunctions of the momentum op erators p with eigenvalue
~ ~
p~ =hk and are also energy eigenstates with eigenvalue E =h! (k ). The most general
2)
solution to the Schrodinger equation can now b e written as
Z
3
~
d k 1
~ ~
ik ~x i! (k )t
~
(~x;t)= e (k ); (8)
3
~
(2 )
2! (k )
where
q
2 2
~ ~
! (k )= k +m : (9)
To simplify the formulas we put here and in the followingh = c = 1. When the momentum
~ ~
p~ is well sp eci ed the wave-function (k ) will b e strongly p eaked near k = ~p. The p osition
of the particle will in this case not b e very precisely determined due to the uncertainty
principle. So the wave function gives us only a probability (~x; t) for nding a particle at
a p osition ~x at a certain time t. Because probabilities have to b e conserved, there must
~
also exist a probability current density j (~x; t) that ob eys the conservation law
~
~
@ (~x; t)= r j(~x; t) (10)
t
For the relativistic wave function the probability density and current density are given by
the expressions
i
= ( @ @ )
t t
2m
i
~ ~
~
j = ( r r ) (11)
2m
~
These densities can in fact b e identi ed with the particle densities and j inside the
~
! k
~
b eam. Notice that for a plane wave these densities are given by = and j = .To
~
k
m m
verify that these expressions indeed satisfy the conservation law (10) one has to use the
`square' of the Schrodinger equation (6)
2 2 2
~
(@ r + m ) (~x;t)=0 (12)
t
This equation is called the Klein-Gordon equation. We note, however, that the Klein-
Gordon equation allows more general solutions than the original Schrodinger equation.
~ ~
Namely,ifwechange the sign of the time frequency ! (k ) ! !(k) in the expression (8)
it would still b e a solution of (12). These negative frequency solutions are not physically
acceptable as wave functions b ecause they would corresp ond to particles with negative
energy.
2.2 The Dirac Equation
For particles with spin the wave function not only carries information ab out the
momentum and energy but also sp ecify the p olarization of the particle. An electron has
spin 1=2, and so its wave function has two indep endent comp onents (~x;t), for spin up,
"
~ ~
2) Here the normalization factor 2! (k )ischosen so that the function (k ) has a relativistic
invariant meaning.
and and (~x; t), for spin down. Both comp onents satisfy the Schrodinger equation (6)
#
~ ~
and through a similar expansion as (8) lead to two indep endent functions (k ) and (k )
" #
that can b e combined in a two comp onent "spinor"
!
"
~ ~
(k )= (k) (13)
#
In the rest frame of the electron the direction of the spin is measured by the Pauli matrices
2
. These are 2 2 matrices satisfying = 1 and = = i , etc.
i 1 2 2 1 3
i
1
Dirac discovered a relativistic wave equation for a spin particle in terms of a
2
4-comp onentwave-function .Itisconvenient to rst decomp ose this four comp onent
a
spinor in twotwo-comp onent spinors and .
+
!
+
=