Electromagnetic Field Theory (Ebook)

Bo Thidé
E LECTROMAGNETIC F IELD T HEORY
Released 14th December 1998 at 13:56
Downloaded from http://www.plasma.uu.se/CED/Book

Also available
E LECTROMAGNETIC F IELD T HEORY
E XERCISES
by
Tobia Carozzi, Anders Eriksson, Bengt Lundborg,
Bo Thidé and Mattias Waldenvik
E LECTROMAGNETIC
F IELD T HEORY
Bo Thidé
Department of Space and Plasma Physics
Uppsala University
and
Swedish Institute of Space Physics
Uppsala Division
Sweden
Σ
Ipsum
This book was typeset in LATEX 2
on an HP9000/700 series workstation
and printed on an HP LaserJet 5000GN printer.

Copyright c 1997, 1998 by
Bo Thidé
Uppsala, Sweden
All rights reserved.
Electromagnetic Field Theory

ISBN X-XXX-XXXXX-X
C ONTENTS
Preface xi
1 Classical Electrodynamics 1
1.1 Electrostatics . . . . . . . . . . . . . . . . . . . .1 . . . . . . . .
1.1.1 Coulomb’s law . . . . . . . . . . . . . . .2 . . . . . . . .
1.1.2 The electrostatic field . . . . . . . . . . . .3 . . . . . . . .
1.2 Magnetostatics . . . . . . . . . . . . . . . . . . .5 . . . . . . . .
1.2.1 Ampère’s law . . . . . . . . . . . . . . . .5 . . . . . . . .
1.2.2 The magnetostatic field . . . . . . . . . . .6 . . . . . . . .
1.3 Electrodynamics . . . . . . . . . . . . . . . . . .8 . . . . . . . .
1.3.1 Equation of continuity . . . . . . . . . . .9 . . . . . . . .
1.3.2 Maxwell’s displacement current . . . . . .9 . . . . . . . .
1.3.3 Electromotive force . . . . . . . . . . . . .
10 . . . . . . . .
1.3.4 Faraday’s law of induction . . . . . . . . .
11 . . . . . . . .
1.3.5 Maxwell’s microscopic equations . . . . .
13 . . . . . . . .
1.3.6 Maxwell’s macroscopic equations . . . . .
14 . . . . . . . .
1.4 Electromagnetodynamics . . . . . . . . . . . . . .
15 . . . . . . . .
Example 1.1 Invariance of the electromagnetodynamic equations 16
Example 1.2 Maxwell from Dirac-Maxwell equations for a fixed
mixing angle . . . . . . . . . . . . . . . . . . . 17
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2 Electromagnetic Waves 21
2.1 The wave equation . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2 Plane waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2.1 Telegrapher’s equation . . . . . . . . . . . . . . . . . . . 25
2.2.2 Waves in conductive media . . . . . . . . . . . . . . . . . 26
2.3 Observables and averages . . . . . . . . . . . . . . . . . . . . . . 27
i
ii
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3 Electromagnetic Potentials 31
3.1 The electrostatic scalar potential . . . . . . . . . . . . . . . . . . 31
3.2 The magnetostatic vector potential . . . . . . . . . . . . . . . . . 32
3.3 The electromagnetic scalar and vector potentials . . . . . . . . . . 32
3.3.1 Electromagnetic gauges . . . . . . . . . . . . . . . . . . 34
Lorentz equations for the electromagnetic potentials . . . 34
Gauge transformations . . . . . . . . . . . . . . . . . . . 35
3.3.2 Solution of the Lorentz equations for the electromagnetic
potentials . . . . . . . . . . . . . . . . . . . . . . . . . . 36
The retarded potentials . . . . . . . . . . . . . . . . . . . 39
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4 The Electromagnetic Fields 43

4.1 The magnetic field . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.2 The electric field . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5 Relativistic Electrodynamics 53
5.1 The special theory of relativity . . . . . . . . . . . . . . . . . . . 53
5.1.1 The Lorentz transformation . . . . . . . . . . . . . . . . 54
5.1.2 Lorentz space . . . . . . . . . . . . . . . . . . . . . . . . 55
Metric tensor . . . . . . . . . . . . . . . . . . . . . . . . 56
Radius four-vector in contravariant and covariant form . . 56
Scalar product and norm . . . . . . . . . . . . . . . . . . 57
Invariant line element and proper time . . . . . . . . . . . 57
Four-vector fields . . . . . . . . . . . . . . . . . . . . . . 58
The Lorentz transformation matrix . . . . . . . . . . . . . 59
The Lorentz group . . . . . . . . . . . . . . . . . . . . . 59
5.1.3 Minkowski space . . . . . . . . . . . . . . . . . . . . . . 60
5.2 Covariant classical mechanics . . . . . . . . . . . . . . . . . . . 62
5.3 Covariant classical electrodynamics . . . . . . . . . . . . . . . . 63
5.3.1 The four-potential . . . . . . . . . . . . . . . . . . . . . 64
5.3.2 The Liénard-Wiechert potentials . . . . . . . . . . . . . . 65
5.3.3 The electromagnetic field tensor . . . . . . . . . . . . . . 67
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
iii
6 Interactions of Fields and Particles 73

6.1 Charged Particles in an Electromagnetic Field . . . . . . . . . . 73 .
6.1.1 Covariant equations of motion . . . . . . . . . . . . . . 73 .
Lagrange formalism . . . . . . . . . . . . . . . . . . . 73 .
Hamiltonian formalism . . . . . . . . . . . . . . . . . . 76 .
6.2 Covariant Field Theory . . . . . . . . . . . . . . . . . . . . . . 80 .
6.2.1 Lagrange-Hamilton formalism for fields and interactions 80 .
The electromagnetic field . . . . . . . . . . . . . . . . . 84 .
Example 6.1 Field energy difference expressed in the field tensor 84
Other fields . . . . . . . . . . . . . . . . . . . . . . . . . 88
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7 Interactions of Fields and Matter 91

7.1 Electric polarisation and the electric displacement vector . . . . . 91
7.1.1 Electric multipole moments . . . . . . . . . . . . . . . . 91
7.2 Magnetisation and the magnetising field . . . . . . . . . . . . . . 94
7.3 Energy and momentum . . . . . . . . . . . . . . . . . . . . . . . 95
7.3.1 The energy theorem in Maxwell’s theory . . . . . . . . . 96
7.3.2 The momentum theorem in Maxwell’s theory . . . . . . . 97
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
8 Electromagnetic Radiation 103

8.1 The radiation fields . . . . . . . . . . . . . . . . . . . . . . . . . 103
8.2 Radiated energy . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
8.2.1 Monochromatic signals . . . . . . . . . . . . . . . . . . . 106
8.2.2 Finite bandwidth signals . . . . . . . . . . . . . . . . . . 106
8.3 Radiating systems . . . . . . . . . . . . . . . . . . . . . . . . . . 108
8.3.1 Simple geometries . . . . . . . . . . . . . . . . . . . . . 108
Linear antenna . . . . . . . . . . . . . . . . . . . . . . . 108
8.3.2 Multipole expansion . . . . . . . . . . . . . . . . . . . . 110
The Hertz potential . . . . . . . . . . . . . . . . . . . . . 110
Electric dipole radiation . . . . . . . . . . . . . . . . . . 113
Magnetic dipole radiation . . . . . . . . . . . . . . . . . 114
Electric quadrupole radiation . . . . . . . . . . . . . . . . 116
8.3.3 Radiation from charges moving in vacuum . . . . . . . . 117
Uniformly moving charges . . . . . . . . . . . . . . . . . 119
Accelerated charges . . . . . . . . . . . . . . . . . . . . 122
Radiation for small velocities . . . . . . . . . . . . . . . 126
Bremsstrahlung . . . . . . . . . . . . . . . . . . . . . . . 127
iv
Example 8.1 Bremsstrahlung at low speeds and short accelera-

tion times . . . . . . . . . . . . . . . . . . . . . 130
Cyclotron and synchrotron radiation . . . . . . . . . . . . 132
Radiation in the general case . . . . . . . . . . . . . . . . 137
The convection potential and the convection force . . . . . 138
Virtual photons . . . . . . . . . . . . . . . . . . . . . . . 141
8.3.4 Radiation from charges moving in matter . . . . . . . . . 143
Vavilov-Čerenkov radiation . . . . . . . . . . . . . . . . 145
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
F Formulae 153
F.1 The Electromagnetic Field . . . . . . . . . . . . . . . . . . . . . 153
F.1.1 Maxwell’s equations . . . . . . . . . . . . . . . . . . . . 153
Constitutive relations . . . . . . . . . . . . . . . . . . . . 153
F.1.2 Fields and potentials . . . . . . . . . . . . . . . . . . . . 154
Vector and scalar potentials . . . . . . . . . . . . . . . . 154
Lorentz’ gauge condition in vacuum . . . . . . . . . . . . 154
F.1.3 Force and energy . . . . . . . . . . . . . . . . . . . . . . 154
Poynting’s vector . . . . . . . . . . . . . . . . . . . . . . 154
Maxwell’s stress tensor . . . . . . . . . . . . . . . . . . . 154
F.2 Electromagnetic Radiation . . . . . . . . . . . . . . . . . . . . . 154
F.2.1 Relationship between the field vectors in a plane wave . . 154
F.2.2 The far fields from an extended source distribution . . . . 154
F.2.3 The far fields from an electric dipole . . . . . . . . . . . . 155
F.2.4 The far fields from a magnetic dipole . . . . . . . . . . . 155
F.2.5 The far fields from an electric quadrupole . . . . . . . . . 155
F.2.6 The fields from a point charge in arbitrary motion . . . . . 155
F.2.7 The fields from a point charge in uniform motion . . . . . 156
F.3 Special Relativity . . . . . . . . . . . . . . . . . . . . . . . . . . 156
F.3.1 Metric tensor . . . . . . . . . . . . . . . . . . . . . . . . 156
F.3.2 Covariant and contravariant four-vectors . . . . . . . . . . 156
F.3.3 Lorentz transformation of a four-vector . . . . . . . . . . 156
F.3.4 Invariant line element . . . . . . . . . . . . . . . . . . . . 157
F.3.5 Four-velocity . . . . . . . . . . . . . . . . . . . . . . . . 157
F.3.6 Four-momentum . . . . . . . . . . . . . . . . . . . . . . 157
F.3.7 Four-current density . . . . . . . . . . . . . . . . . . . . 157
F.3.8 Four-potential . . . . . . . . . . . . . . . . . . . . . . . . 157
F.3.9 Field tensor . . . . . . . . . . . . . . . . . . . . . . . . . 157
F.4 Vector Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
F.4.1 Spherical polar coordinates . . . . . . . . . . . . . . . . . 158
v
Base vectors . . . . . . . . . . . . . . . . . . . . . . . . 158

Directed line element . . . . . . . . . . . . . . . . . . . . 158
Solid angle element . . . . . . . . . . . . . . . . . . . . . 158
Directed area element . . . . . . . . . . . . . . . . . . . 158
Volume element . . . . . . . . . . . . . . . . . . . . . . 158
F.4.2 Vector formulae . . . . . . . . . . . . . . . . . . . . . . . 159
General relations . . . . . . . . . . . . . . . . . . . . . . 159
Special relations . . . . . . . . . . . . . . . . . . . . . . 160
Integral relations . . . . . . . . . . . . . . . . . . . . . . 160
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
M Mathematical Methods 165

M.1 Scalars, Vectors and Tensors . .
. . . . . . . . . . . . . . . . . . 165
M.1.1 Vectors . . . . . . . . .
. . . . . . . . . . . . . . . . . . 165
Radius vector . . . . . .
. . . . . . . . . . . . . . . . . . 165
M.1.2 Fields . . . . . . . . . .
. . . . . . . . . . . . . . . . . . 167
Scalar fields . . . . . . .
. . . . . . . . . . . . . . . . . . 167
Vector fields . . . . . .
. . . . . . . . . . . . . . . . . . 167
Tensor fields . . . . . .
. . . . . . . . . . . . . . . . . . 168
Example M.1 Tensors in 3D space . . . . . . . . . . . . . . 170
M.1.3 Vector algebra . . . . . . . . . . . . . . . . . . . . . . . 173
Scalar product . . . . . . . . . . . . . . . . . . . . . . . 173
Example M.2 Scalar product, norm and metric in Lorentz space 173
Example M.3 Metric in general relativity . . . . . . . . . . . 173
Dyadic product . . . . . . . . . .
. . . . . . . . . . . . . 174
Vector product . . . . . . . . . .
. . . . . . . . . . . . . 174
M.1.4 Vector analysis . . . . . . . . . .
. . . . . . . . . . . . . 175
The del operator . . . . . . . . .
. . . . . . . . . . . . . 175
Example M.4 The four-del operator in Lorentz space . . . . . 176
The gradient . . . . . . . . . . . . . . . . . . . . . . . . 176
Example M.5 Gradients of scalar functions of relative distances
in 3D . . . . . . . . . . . . . . . . . . . . . . . 176
The divergence . . . . . . . . . . . . . . . . . . . . . . . 177
Example M.6 Divergence in 3D . . . . . . . . . . . . . . . 177
The Laplacian . . . . . . . . . . . . . . . . . . . . . . . . 178
Example M.7 The Laplacian and the Dirac delta . . . . . . . 178
The curl . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
Example M.8 The curl of a gradient . . . . . . . . . . . . . 178
Example M.9 The divergence of a curl . . . . . . . . . . . . 179
M.2 Analytical Mechanics . . . . . . . . . . . . . . . . . . . . . . . . 180
M.2.1 Lagrange’s equations . . . . . . . . . . . . . . . . . . . . 180
M.2.2 Hamilton’s equations . . . . . . . . . . . . . . . . . . . . 181
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
vi
L IST OF F IGURES
1.1 Coulomb interaction . . . . . . . . . . . . . . . . . . . . . . . . 2

1.2 Ampère interaction . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Moving loop in a varying B field . . . . . . . . . . . . . . . . . . 12
5.1 Relative motion of two inertial systems . . . . . . . . . . . . . . 54

5.2 Rotation in a 2D Euclidean space . . . . . . . . . . . . . . . . . . 60
5.3 Minkowski diagram . . . . . . . . . . . . . . . . . . . . . . . . . 61
6.1 Linear one-dimensional mass chain . . . . . . . . . . . . . . . . . 80
8.1 Radiation in the far zone . . . . . . . . . . . . . . . . . . . . . . 105

8.2 Radiation from a moving charge in vacuum . . . . . . . . . . . . 117
8.3 A uniformly moving charge in vacuum . . . . . . . . . . . . . . . 120
8.4 An accelerated charge in vacuum . . . . . . . . . . . . . . . . . . 125
8.5 Angular distribution of radiation during bremsstrahlung . . . . . . 128
8.6 Location of radiation during bremsstrahlung . . . . . . . . . . . . 129
8.7 Radiation from a charge in circular motion . . . . . . . . . . . . . 133
8.8 Synchrotron radiation lobe width . . . . . . . . . . . . . . . . . . 136
8.9 The perpendicular field of a moving charge. . . . . . . . . . . . . 141
8.10 Vavilov-Čerenkov cone . . . . . . . . . . . . . . . . . . . . . . . 147
M.1 Tetrahedron-like volume element of matter . . . . . . . . . . . . . 171
vii
To the memory of
L EV M IKHAILOVICH E RUKHIMOV
dear friend, remarkable physicist
and a truly great human
P REFACE
This book is the result of a twenty-five year long love affair. In 1972, I took my
first advanced course in electrodynamics at the Theoretical Physics department,
Uppsala University. Shortly thereafter, I joined the research group there and took
on the task of helping my supervisor, professor P ER -O LOF F RÖMAN, with the
preparation of a new version of his lecture notes on Electricity Theory. These two
things opened up my eyes for the beauty and intricacy of electrodynamics, already
at the classical level, and I fell in love with it.
Ever since that time, I have off and on had reason to return to electrodynamics,
both in my studies, research and teaching, and the current book is the result of
my own teaching of a course in advanced electrodynamics at Uppsala University
some twenty odd years after I experienced the first encounter with this subject.
The book is the outgrowth of the lecture notes that I prepared for the four-credit
course Electrodynamics that was introduced in the Uppsala University curriculum
in 1992, to become the five-credit course Classical Electrodynamics in 1997. To
some extent, parts of these notes were based on lecture notes prepared, in Swedish,
by B ENGT L UNDBORG who created, developed and taught the earlier, two-credit
course Electromagnetic Radiation at our faculty.
Intended primarily as a textbook for physics students at the advanced under-
graduate or beginning graduate level, I hope the book may be useful for research
workers too. It provides a thorough treatment of the theory of electrodynamics,
mainly from a classical field theoretical point of view, and includes such things
as electrostatics and magnetostatics and their unification into electrodynamics, the
electromagnetic potentials, gauge transformations, covariant formulation of clas-
sical electrodynamics, force, momentum and energy of the electromagnetic field,
radiation and scattering phenomena, electromagnetic waves and their propagation
in vacuum and in media, and covariant Lagrangian/Hamiltonian field theoretical
methods for electromagnetic fields, particles and interactions. The aim has been to
write a book that can serve both as an advanced text in Classical Electrodynamics
and as a preparation for studies in Quantum Electrodynamics and related subjects.
xi
xii P REFACE
In an attempt to encourage participation by other scientists and students in the

authoring of this book, and to ensure its quality and scope to make it useful in
higher university education anywhere in the world, it was produced within a World-
Wide Web (WWW) project. This turned out to be a rather successful move. By
making an electronic version of the book freely down-loadable on the net, I have
not only received comments on it from fellow Internet physicists around the world,
but know, from WWW ‘hit’ statistics that at the time of writing this, the book
serves as a frequently used Internet resource. This way it is my hope that it will
be particularly useful for students and researchers working under financial or other
circumstances that make it difficult to procure a printed copy of the book.
I am grateful not only to Per-Olof Fröman and Bengt Lundborg for providing the
inspiration for my writing this book, but also to C HRISTER WAHLBERG at Uppsala
University for interesting discussions on electrodynamics in general and on this
book in particular, and to my former graduate students M ATTIAS WALDENVIK
and T OBIA C AROZZI as well as A NDERS E RIKSSON, all at the Swedish Institute
of Space Physics, Uppsala Division, and who have participated in the teaching and
commented on the material covered in the course and in this book. Thanks are also
due to my long-term space physics colleague H ELMUT KOPKA of the Max-Planck-
Institut für Aeronomie, Lindau, Germany, who not only taught me about the prac-
tical aspects of the of high-power radio wave transmitters and transmission lines,
but also about the more delicate aspects of typesetting a book in TEX and LATEX.
I am particularly indebted to Academician professor V ITALIY L. G INZBURG for
his many fascinating and very elucidating lectures, comments and historical foot-
notes on electromagnetic radiation while cruising on the Volga river during our
joint Russian-Swedish summer schools.
Finally, I would like to thank all students and Internet users who have down-
loaded and commented on the book during its life on the World-Wide Web.
I dedicate this book to my son M ATTIAS, my daughter K AROLINA, my
high-school physics teacher, S TAFFAN RÖSBY, and to my fellow members of the
C APELLA P EDAGOGICA U PSALIENSIS.
Uppsala, Sweden B O T HIDÉ

December, 1998
C HAPTER 1
Classical
Electrodynamics
Classical electrodynamics deals with electric and magnetic fields and interactions
caused by macroscopic distributions of electric charges and currents. This means
that the concepts of localised charges and currents assume the validity of certain
mathematical limiting processes in which it is considered possible for the charge
and current distributions to be localised in infinitesimally small volumes of space.
This is in obvious contradiction to electromagnetism on a microscopic scale, where
charges and currents are known to be spatially extended objects. However, the
limiting processes yield results which are correct on a macroscopic scale.
In this Chapter we start with the force interactions in classical electrostatics and
classical magnetostatics and introduce the static electric and magnetic fields and
find two uncoupled systems of equations for them. Then we see how the conser-
vation of electric charge and its relation to electric current leads to the dynamic
connection between electricity and magnetism and how the two can be unified in
one theory, classical electrodynamics, described by one system of coupled dynamic
field equations.
1.1 Electrostatics
The theory that describes physical phenomena related to the interaction between
stationary electric charges or charge distributions in space is called electrostatics.
1
2 C HAPTER 1. C LASSICAL E LECTRODYNAMICS
x x
x

x
Figure 1.1. Coulomb’s law describes

how a static electric charge , located at a
point x relative to the origin , experiences an electrostatic force from a static
electric charge located at x .
1.1.1 Coulomb’s law

It has been found experimentally that in classical electrostatics the interaction
between two stationary electrically charged bodies can be described in terms of
a mechanical force. Let us consider the simple case depicted in Figure 1.1 where
F denotes the force acting on a charged particle with charge located at x, due to
the presence of a charge located at x . According to Coulomb’s law this force is,
in vacuum, given by the expression
x x
F x

4 0 x x 3
1
(1.1)
4 0 x x
where we have used results from Example M.5 on page 177. In SI units, which
we shall use throughout, the force F is measured
in Newton (N), the charges
and in Coulomb (C), and the length x x in metres (m). The constant
0 107 4 2
8 8542 10 12 Farad per metre (F/m) is the vacuum permit-
tivity and 2 9979 108 m/s is the speed of light in vacuum. In CGS units
0 1 4
and the force is measured in dyne, the charge in statcoulomb, and
length in centimetres (cm).
1.1. E LECTROSTATICS 3
1.1.2 The electrostatic field

Instead of describing the electrostatic interaction in terms of a “force action at a
distance,” it turns out that it is often more convenient to introduce the concept of
a field and to describe the electrostatic interaction in terms of a static vectorial
electric field Estat defined by the limiting process

def F
Estat ! lim0 (1.2)
where F is the electrostatic force, as defined in Equation (1.1) on the facing page,
from a net charge on the test particle with a small electric net charge . Since the
purpose of the limiting process is to assure that the test charge does not influence
the field, the expression for Estat does not depend explicitly on but only on the
charge and the relative radius vector x x . This means that we can say that
any net electric charge produces an electric field in the space that surrounds it,
regardless of the existence of a second charge anywhere in this space.1
Using formulae (1.1) and (1.2), we find that the electrostatic field Estat at the
field point x (also known as the observation point), due to a field-producing charge
at the source point x , is given by
x x
Estat x

4 x x 3
0
1
" (1.3)
4 0 x x
In the presence of several field producing discrete charges # , at x# , $% 1 & 2 & 3 &''' ,
respectively, the assumption of linearity of vacuum2 allows us to superimpose their
individual E fields into a total E field
1
In the preface to the first edition of the first volume of his book A Treatise on Electricity and Mag-
netism, first published in 1873, James Clerk Maxwell describes this in the following, almost poetic,
manner [5]:
“For instance, Faraday, in his mind’s eye, saw lines of force traversing all space where
the mathematicians saw centres of force attracting at a distance: Faraday saw a me-
dium where they saw nothing but distance: Faraday sought the seat of the phenomena
in real actions going on in the medium, they were satisfied that they had found it in a
power of action at a distance impressed on the electric fluids.”
2
In fact, vacuum exhibits a quantum mechanical nonlinearity due to vacuum polarisation effects
manifesting themselves in the momentary creation and annihilation of electron-positron pairs, but
classically this nonlinearity is negligible.
# x x#
Estat x
)( (1.4)
# 4 0 x x# 3
If the discrete charges are small and numerous enough, we introduce the charge
density * located at x and write the total field as
1 x x 3.
Estat x
* x
3 -
4 0 + , x x
1 1
* x
- 3. (1.5)
4 0 + , x x
where, in the last step, we used formula Equation (M.60) on page 177. We em-
phasise that Equation (1.5) above is valid for an arbitrary distribution of charges,
including discrete charges, in which case * can be expressed in terms of one or
more Dirac delta functions.
10
Since, according to formula Equation (M.70) on page 179, / x
32 0
0
for any 3D 4 3 scalar field x
, we immediately find that in electrostatics
1 1
Estat x
* x
65 7 - 3.
4 0 + , x x
0 (1.6)
I.e., Estat is an irrotational field.

Taking the divergence of the general Estat expression for an arbitrary charge dis-
tribution, Equation (1.5), and using the representation of the Dirac delta function,
Equation (M.65) on page 178, we find that
98 98 1 x x 3.
Estat x
* x
3 -
4 0 + , x x
1 98:
1
" * x
- 3.
4 0 + , x x
1 1
" * x
; 2 - 3.
4 0 + , x x
1
* x
< x x
- 3.
0+ ,
* x
(1.7)
0
which is Gauss’s law in differential form.
1.2. M AGNETOSTATICS 5
1.2 Magnetostatics
While electrostatics deals with static charges, magnetostatics deals with stationary
currents, i.e., charges moving with constant speeds, and the interaction between
these currents.
1.2.1 Ampère’s law

Experiments on the interaction between two small current loops have shown that
they interact via a mechanical force, much the same way that charges interact. Let
F denote such a force acting on a small loop = carrying a current > located at x,
due to the presence of a small loop =? carrying a current > located at x . According
to Ampère’s law this force is, in vacuum, given by the expression
0 >%> l x x
F x
@ l -
4 A B A BDC - x x 3
0 >%> 1
" @ l 65 - l 7
4 A B A BDC - x x
(1.8)
Here - l and - l are tangential line elements of the loops = and = , respectively, and,
in SI units, 0 4E 10 7 1 2566 10 6 H/m is the vacuum permeability.
@
From the definition of 0 and 0 (in SI units) we observe that
@
107 1 2 2
0 0 (F/m) 4F 10 7 (H/m) (s /m ) (1.9)
@ 4 2 2
which is a useful relation.
At first glance, Equation (1.8) above appears to be unsymmetric in terms of the
loops and therefore to be a force law that is in contradiction with Newton’s third
law. However, by applying the vector triple product “bac-cab” Formula (F.56) on
page 159, we can rewrite (1.8) in the following way
0 >%> 8 1
F x
" @ 5- l 7 - l
4 A B A BDC x x
0 >G> x x 8
@ 3 - l - l (1.10)
4 A B A BDC x x
Recognising the fact the integrand in the first integral is an exact differential so that
this integral vanishes, we can rewrite the force expression, Equation (1.8) above,
in the following symmetric way
0 >G> x x 8
F x
H @ l - l (1.11)
4 A B A BDC x x 3 I
This clearly exhibits the expected symmetry in terms of loops = and = .
KL J
l
x x J
l
x
K

x L

K
Figure 1.2. Ampère’s law describes how a small loop , carrying a static electric
J
current through its tangential line K element l located at x, experiences a mag-
L force from a smallJ loop , carrying a static electric current through
netostatic
the tangential line element l located at x . The loops can have arbitrary
L shapes
as long as they are simple and closed.
1.2.2 The magnetostatic field

In analogy with the electrostatic case, we may attribute the magnetostatic interac-
tion to a vectorial magnetic field field Bstat . I turns out that Bstat can be defined
through

def 0 > x x
- Bstat x
@
4 - l
x x
3 (1.12)
which expresses the small element - Bstat x

of the static magnetic field set up at
the field point x by a small line element - l of stationary current > at the source
point x . The SI unit for the magnetic field, sometimes called the magnetic flux
density or magnetic induction, is Tesla (T).
Generalising expression (1.12) to an integrated steady state current distribution
j x
, we obtain Biot-Savart’s law:
0 x x 3.
Bstat x
@ j x
3 -
4 + , x x
1
@
0
j x
- 3. (1.13)
4 + , x x
Comparing Equation (1.5) on page 4 with Equation (1.13), we see that there exists
a close analogy between the expressions for Estat and Bstat but that they differ in
1.2. M AGNETOSTATICS 7
their vectorial characteristics. With this definition of Bstat , Equation (1.8) on page 5
may we written
F x
"> l Bstat x
(1.14)
A B -
In order to study the properties of Bstat , we investigate its divergence and curl.
Taking the divergence of both sides of Equation (1.13) on the facing page, we
obtain
98 0 98 1
Bstat x
@ j x
- 3.
4 + , x x
8 1
@
0
j x
5 7 - 3.
4 + , x x
0 (1.15)
10
where Equation (M.70) on page 179 for / x
M2 was used.
Applying the operator “bac-cab” rule, Formula (F.69) on page 159, the curl of
Equation (1.13) on the facing page can be written
1
Bstat x
" @
0
j x
- 3.
4 + , x x
1
" @
0
j x
; 2 - 3.
4 + , x x
N 8 1
@
0
/ j x
2 - 3. (1.16)
4 + , x x
If, in the first of the two integrals on the right hand side, we use the representation
of the Dirac delta function Equation (M.65) on page 178, and integrate the second
one by parts, by utilising Formula (F.61) on page 159 as follows:
8: 1 3.
/ j x
2 -
+ , x x
8QP 1
xO j x
5FR . 7TS - 3.
+ , O x x
R
8 1
j x
V - 3.
+ , U x x
1 8
xO j x
R . - S
+ W O x x
R
8 1
j x
V - 3. 0 (1.17)
+ , U x x
Here the first integral, obtained by applying Gauss’s theorem, vanishes when integ-
rated over a large sphere far away from the localised source j x
, and the second
98
one vanishes because j 0 for stationary currents (no charge accumulation in
space). The net result is simply

Bstat x
0 j x
< x x
- 3. 0j x
(1.18)
@ + , @
1.3 Electrodynamics
As we saw in the previous sections, the laws of electrostatics and magnetostatics
can be summarised in two pairs of time-independent, uncoupled vector differential
equations, namely the equations of classical electrostatics
98 * x
Estat x
(1.19a)
0

Estat x
0 (1.19b)
and the equations of classical magnetostatics

98
Bstat x
0 (1.20a)

Bstat x
0j x
(1.20b)
@
Since there is nothing a priori that connects Estat directly with Bstat , we must con-
sider classical electrostatics and classical magnetostatics as two independent the-
ories.
However, when we include time-dependence, these theories are unified into one
theory: classical electrodynamics. This unification of the theories of electricity
and magnetism is motivated by two empirically established facts:
1. Electric charge is a conserved quantity and current is a transport of electric

charge. This fact manifests itself in the equation of continuity and, as a
consequence, in Maxwell’s displacement current.
2. A change in the magnetic flux through a loop will induce an EMF electric
field in the loop. This is the celebrated Faraday’s law of induction.
1.3. E LECTRODYNAMICS 9
1.3.1 Equation of continuity

Let j denote the electric current density (A/m2 ). In the simplest case it can be
defined as j v* where v is the velocity of the charge density. In general, j has
to be defined in statistical mechanical terms as j XY& x
Z [\ [ ] v ^ [ X_& x & v
- 3̀
0
where ^ [ X_& x & v
is the (normalised) distribution function for particle species
with electrical charge [ .
The electric charge conservation law can be formulated in the equation of con-
tinuity
* XY& x
N 98
R j X_& x
0 (1.21)
X
R
which states that the time rate of change of electric charge * XY& x
is balanced by a
divergence in the electric current density j X_& x
.
1.3.2 Maxwell’s displacement current

We recall from the derivation of Equation (1.18) on the facing page that we used
a8
the fact that in magnetostatics j x
0. In the case of non-stationary sources
and fields, we must, in accordance with the continuity Equation (1.21) above, set
b8
j X_& x
c * X_& x
X . Doing so, and formally repeating the steps in the
R R
derivation of Equation (1.18) on the facing page, we would obtain the formal result

B X_& x
0 j X_& x
< x x
- 3.
@ + ,
N 0 1 3.
@ R * XY& x
-
4 X + , x x
R
N
0 j X_& x
0 R 0 E XY& x
(1.22)
@ @ X
R
where, in the last step, we have assumed that a generalisation of Equation (1.5) on
page 4 to time-varying fields allows us to make the identification
1 1
R * XY& x
- 3.
4 0 X + , x x
R
1 1

R * YX & x
- 3.
4 0 X + , x x
R
dR E XY& x
(1.23)
X
R
Later, we will need to consider this formal result further. The result is Maxwell’s
source equation for the B field
N
B X_& x
0 j X_& x
R 0 E XY& x
(1.24)
@ X
R

where the last term 0 E X_& x
X is the famous displacement current. This term
R R
was introduced, in a stroke of genius, by Maxwell in order to make the right hand
side of this equation divergence free when j X_& x
is assumed to represent the dens-
ity of the total electric current, which can be split up in “ordinary” conduction
currents, polarisation currents and magnetisation currents. The displacement cur-
rent is an extra term which behaves like a current density which flows in vacuum
and, as we shall see later, its existence has very far-reaching physical consequences
as it predicts the existence of electromagnetic radiation that can carry energy and
momentum over very long distances, even in vacuum.
1.3.3 Electromotive force

If an electric field E XY& x
, is applied to a conducting medium, a current density
j XY& x
will be produced in this medium. There exist also hydrodynamical and
chemical processes which can create currents. Under certain physical conditions,
and for certain materials, one can sometimes assume a linear relationship between
the current density j and E, called Ohm’s law:
j XY& x
fe E XY& x
(1.25)
where e is the electric conductivity (S/m). In the most general cases, for instance
in an anisotropic conductor, e is a tensor.
We can view Ohm’s law, Equation (1.25), as the first term in a Taylor expan-
sion of the law j / E XY& x
32 . This general law incorporates non-linear effects such
as frequency mixing. Examples of media which are highly non-linear are semi-
conductors and plasma. We draw the attention to the fact that even in cases when
the linear relation between E and j is a good approximation, we still have to use
Ohm’s law with care. The conductivity e is, in general, time-dependent (temporal
dispersive media) but then it is often the case that Equation (1.25) above is valid
for each individual Fourier component of the field. We shall not, however, dwell
upon such complicated cases here.
If the current is caused by an applied electric field E XY& x
, this electric field will
exert work on the charges in the medium and, unless the medium is superconduct-
ing, there will be some energy loss. The rate at which this energy is expended is
8
j E per unit volume. If E is irrotational (conservative), j will decay away with
time. Stationary currents therefore require that an electric field which corresponds
to an electromotive force (EMF) is present. In the presence of such a field EEMF ,

Ohm’s law, Equation (1.25) on the facing page, takes the form
N
j fe Estat EEMF
(1.26)
The electromotive force is defined as
g N 8
Estat EEMF
- l (1.27)
A B
where - l is a tangential line element of the closed loop = .
1.3.4 Faraday’s law of induction

In Subsection 1.1.2 we derived the differential equations for the electrostatic field.

In particular, we derived Equation (1.6) on page 4 which states that Estat x
0
and thus that Estat is a conservative field (it can be expressed as a gradient of a scalar
field). This implies that the closed line integral of Estat in Equation (1.27) vanishes
and that this equation becomes
g 8
EEMF - l (1.28)
A B
It was found experimentally that a nonconservative EMF field is produced in a
closed circuit = if the magnetic flux through this circuit varies with time. This is
formulated in Faraday’s law which, in Maxwell’s generalised form, reads
g 8
X_& x
E X_& x
- l
A B
- Φm X_& x

- X
8
- B X_& x
- S
X
- + W
8
- S R B X_& x
(1.29)
+ W X
R
where Φm is the magnetic flux and h is the surface encircled by = which can
be interpreted as a generic stationary “loop” and not necessarily as a conducting
circuit. Application of Stokes’ theorem on this integral equation, transforms it into
the differential equation

E X_& x
fR B X_& x
(1.30)
X
R
which is valid for arbitrary variations in the fields and constitutes the Maxwell
equation which explicitly connects electricity with magnetism.
k
v S
B i xj
v
l
k
l
B i xj
Figure 1.3. A loop m which moves with velocity v in a spatially varying mag-
netic field B n x o will sense a varying magnetic flux during the motion.
An EMF is induced by all changes of the magnetic flux Φm . Let us therefore

consider the case, illustrated if Figure 1.3, that the “loop” is moved in such a way
that it links a magnetic field which varies during the movement. The convective
derivative is evaluated according to the well-known operator formula
p
prqtscu qwv v xy (1.31)
u
which follows immediately from the rules of differentiation of an arbitrary differ-
q_| q }~}
entiable function z { x { . Applying this rule to Faraday’s law, Equation (1.29)
on the previous page, we obtain
p
qY| } p
{ x s prq B x S
p } p
s S x u q B { v xy B x S (1.32)
u
In spatial differentiation v is to be considered as constant, and Equation (1.15)
on page 7 holds also for time-varying fields:
q_| }
yx B { x s 0 (1.33)
(it is one of Maxwell’s equations) so that

8:
B v
v
B (1.34)
allowing us to rewrite Equation (1.32) on the preceding page in the following way:
g 8
X_& x
EEMF - l
A B
8
- B - S
- X + W
8 8
R B - S B v
- S (1.35)
+ W X + W
R
With Stokes’ theorem applied to the last integral, we finally get
g 8
X_& x
EEMF - l
A B
8 8
R B - S B v
- l (1.36)
+ W X A B
R
or, rearranging the terms,
8 8
EEMF v B
- l " R B - S (1.37)
+ B + W X
R
where EEMF is the field which is induced in the “loop,” i.e., in the moving system.
The use of Stokes’ theorem “backwards” on Equation (1.37) yields

EEMF v B
" R B (1.38)
X
R
In the fixed system, an observer measures the electric field
E EEMF v B (1.39)
Hence, a moving observer measures the following Lorentz force on a charge
N
EEMF f E v B
(1.40)
corresponding to an “effective” electric field in the “loop” (moving observer)
N
EEMF E v B
(1.41)
Hence, we can conclude that for a stationary observer, the Maxwell equation

E R B (1.42)
X
R
is indeed valid even if the “loop” is moving.
1.3.5 Maxwell’s microscopic equations

We are now able to collect the results from the above considerations and formulate
the equations of classical electrodynamics valid for arbitrary variations in time and
space of the coupled electric and magnetic fields E XY& x
and B X_& x
. The equations
are
98 * _X & x
E (1.43a)
0
N B
E R 0 (1.43b)
9R 8 X
B 0 (1.43c)
E
B 0 0 R 0 j X_& x
(1.43d)
@ X @
R
In these equations * X_& x
represents the total, possibly both time and space depend-
ent, electric charge, i.e., free as well as induced (polarisation) charges, and j X_& x
represents the total, possibly both time and space dependent, electric current, i.e.,
conduction currents (motion of free charges) as well as all atomistic (polarisation,
magnetisation) currents. As they stand, the equations therefore incorporate the
classical interaction between all electric charges and currents in the system and
are called Maxwell’s microscopic equations. Another name often used for them
is the Maxwell-Lorentz equations. Together with the appropriate constitutive re-
lations, which relate * and j to the fields, and the initial and boundary conditions
pertinent to the physical situation at hand, they form a system of well-posed partial
differential equations which completely determine E and B.
1.3.6 Maxwell’s macroscopic equations

The microscopic field equations (1.43) provide a correct classical picture for ar-
bitrary field and source distributions, including both microscopic and macroscopic
scales. However, for macroscopic substances it is sometimes convenient to intro-
duce new derived fields which represent the electric and magnetic fields in which,
in an average sence, the material properties of the substances are already included.
These fields are the electric displacement D and the magnetising field H. In the
most general case, these derived fields are complicated nonlocal, nonlinear func-
tionals of the primary fields E and B:
D D / X_& x; E & B2 (1.44a)
H H / X_& x; E & B2 (1.44b)
Under certain conditions, for instance for very low field strenghts, we may assume
that the response of a substance is linear so that
D _X & x
E (1.45)
H X_& x
B
1
(1.46)
@
i.e., that the derived fields are linearly proportional to the primary fields and that
the elecric displacement (magnetising field) is only dependent on the electric (mag-
1.4. E LECTROMAGNETODYNAMICS 15
netic) field.
The field equations expressed in terms of the derived field quantities D and H
are
98
D * X_& x
(1.47a)
N B
E R 0 (1.47b)
9R 8 X
B 0 (1.47c)
D
H R j XY& x
(1.47d)
X
R
and are called Maxwell’s macroscopic equations.
1.4 Electromagnetodynamics
If we look more closely at the microscopic Maxwell equations (1.48), we see that
they exhibit some, albeit not a complete, symmetry. Let us for explicitness de-
note the electric charge density ** XY& x
by * e and the electric current density
j j X_& x
by je . We further make the ad hoc assumption that there exist mag-
netic monopoles represented by a magnetic charge density, denoted * m * m XY& x
,
and a magnetic current density, denoted jm jm X_& x
. With these new quantities
included in the theory, the Maxwell equations can be written
98 * e
E (1.48a)
0
N B
E R " 0 jm (1.48b)
9R 8 X @
B 0* m (1.48c)
@
1 E
B R 0 je (1.48d)
2 X @
R
We shall call these equations the Dirac-Maxwell equations or the electromagneto-
dynamic equations
Taking the divergence of (1.48b), we find that
98 98 8
E
R B
0 jm 0 (1.49)
X @
R
where we used the fact that, according to Formula (M.74) on page 180, the diver-
gence of a curl always vanishes. Using (1.48c) to rewrite this relation, we obtain
the equation of continuity for magnetic monopoles
* m N 8
R jm 0 (1.50)
X
R
which has the same form as that for the electric charges and currents, Equation (1.21)
on page 9.
We notice that the new Equations (1.48) on the previous page exhibit the follow-
ing symmetry (recall that 0 0 1 2 ):
@
E B (1.51a)
1
B E (1.51b)

1
* e * m (1.51c)

* m '* e (1.51d)
1
je jm (1.51e)

jm je (1.51f)
which is a partiular case ( f 2) of the general duality transformation

N
E E cos B sin (1.52a)
1 N
B " E sin B cos (1.52b)

N 1
* e * e cos * m sin (1.52c)

N
* m "w* e sin * m cos (1.52d)
N 1
je je cos jm sin (1.52e)

N
jm "w je sin jm cos (1.52f)
which leaves the Dirac-Maxwell equations, and hence the physics they describe
(often referred to as electromagnetodynamics), invariant. Since E and je are (true
or polar) vectors, B a pseudovector (axial vector), * e a (true) scalar, then * m and
, which behaves as a mixing angle in a two-dimensional “charge space,” must be
pseudoscalars and jm a pseudovector.

E XAMPLE 1.1 I NVARIANCE OF THE ELECTROMAGNETODYNAMIC EQUATIONS
Show that the symmetric, electromagnetodynamic form of Maxwell’s equations (the Dirac-
Maxwell equations), Equations (1.48) on the preceding page are invariant under the duality
transformation (1.52).
Explicit application of the transformation yields
1.4. E LECTROMAGNETODYNAMICS 17
H' 6:
E E cos B sin cos sin
e
0
0 m

1 1
cos sin ¢¡F e
(1.53)

e m
0 0

¤£ B ¤£© 1
E ¦¥ E cos ª B sin Q«¥ E sin B cos ¡

¥¨§ ¥¬§
B 1 E
¥ cos 0 jm cos ¥ sin 0 je sin

¥¨§ 1 E B ¥¨§
¥ sin ¥ cos w 0 jm cos ª 0 je sin

0 ¥¨ j§ e sin j¥¨m§ cos 0 jm (1.54)
and analogously for the other two Dirac-Maxwell equations. QED
E ND OF EXAMPLE 1.1 ®
The invariance of the Dirac-Maxwell equations under the similarity transforma-

tion means that the amount of magnetic monopole density * m is irrelevant for the
physics as long as the ratio * m * e tan is kept constant. So whether we assume
that the particles are only electrically charged or have also a magnetic charge with
a given, fixed ratio between the two types of charges is a matter of convention, as
long as we assume that this fraction is the same for all particles. By varying the
mixing angle we can change the fraction of magnetic monopoles at will without
changing the laws of electrodynamics. For 0 we recover the usual Maxwell
electrodynamics as we know it.

M AXWELL FROM D IRAC -M AXWELL EQUATIONS FOR A FIXED MIXING ANGLE E XAMPLE 1.2
Show that for a fixed mixing angle such that
F e tan (1.55a)

m

jm F je tan (1.55b)
the Dirac-Maxwell equations reduce to the usual Maxwell equations.
Explicit application of the fixed mixing angle conditions on the duality transformation
(1.52) on page 16 yields
1 1
cos sin ¯ cos e tan sin
e
e

m
e

1 1
e cos
2
sin2 (1.56a)
cos e cos
e

e sin ª e tan cos ¯ e sin ª sin w 0 (1.56b)
m
e
1 1
je je cos je tan sin w je cos2 je sin2 je (1.56c)
cos cos
jm je sin ª je tan cos w je sin je sin w 0 (1.56d)
Hence, a fixed mixing angle, or, equvalently, a fixed ratio between the electric and magnetic
charges/currents, “hides” the magnetic monopole influence ( m and jm ) on the dynamic

equations. Furthermore, we notice that
H' H 6 6
E E cos B sin E cos sin
H 6
0 m
E cos 2 tan sin w E cos e

tan sin

0 e
0
H sin2 1
E cos e
e
(1.57)
0 cos cos 0
or
H 1
E e
1 sin2 e
(1.58)
cos2 0 0
and so on for the other equations. QED
E ND OF EXAMPLE 1.2 ®
18
B IBLIOGRAPHY 1
[1] Richard Becker. Electromagnetic Fields and Interactions. Dover Publications, Inc.,
New York, NY, 1982. ISBN 0-486-64290-9.
[2] Erik Hallén. Electromagnetic Theory. Chapman & Hall, Ltd., London, 1962.
[3] John D. Jackson. Classical Electrodynamics. Wiley & Sons, Inc., New York, NY . . . ,
second edition, 1975. ISBN 0-471-43132-X.
[4] Lev Davidovich Landau and Evgeniy Mikhailovich Lifshitz. The Classical Theory of
Fields, volume 2 of Course of Theoretical Physics. Pergamon Press, Ltd., Oxford . . . ,
fourth revised English edition, 1975. ISBN 0-08-025072-6.
[5] James Clerk Maxwell. A Treatise on Electricity and Magnetism, volume 1. Dover
Publications, Inc., New York, NY, third edition, 1954. ISBN 0-486-60636-8.
[6] David Blair Melrose and R. C. McPhedran. Electromagnetic Processes in Dispersive
Media. Cambridge University Press, Cambridge . . . , 1991. ISBN 0-521-41025-8.
[7] Wolfgang K. H. Panofsky and Melba Phillips. Classical Electricity and Magnetism.
Addison-Wesley Publishing Company, Inc., Reading, MA . . . , third edition, 1962.
ISBN 0-201-05702-6.
[8] Julius Adams Stratton. Electromagnetic Theory. McGraw-Hill Book Company, Inc.,
New York, NY and London, 1953. ISBN 07-062150-0.
[9] Jack Vanderlinde. Classical Electromagnetic Theory. John Wiley & Sons, Inc., New
York, Chichester, Brisbane, Toronto, and Singapore, 1993. ISBN 0-471-57269-1.
19
20
C HAPTER 2
Electromagnetic
Waves
Maxwell’s microscopic equations (1.43) on page 14, which are usually written in
the following form
98 * _X & x
E (2.1a)
0

E" R B (2.1b)
98 X
B 0 R (2.1c)
N
B 0j X_& x
0 0 R E (2.1d)
@ @ X
R
can be viewed as an axiomatic basis for classical electrodynamics. In particular,
these equations are well suited for calculating the electric and magnetic fields E
and B from given, prescribed charge distributions * XY& x
and current distributions
j X_& x
of arbitrary time- and space-dependent form.
However, as is well known from the theory of differential equations, these four
first order, coupled partial differential vector equations can be rewritten as two un-
coupled, second order partial equations, one for E and one for B. We shall derive
the second order equation for E, which, as we shall see is a homogeneous wave
equation, and then discuss the implications of this equation. We shall also show
how the B field can be easily calculated from the solution of the E equation.
2.1 The wave equation

Let us consider a volume with no net charge, *F 0, and no electromotive force
EEMF 0. Taking the curl of (2.1b) and using (2.1d), we obtain
21
22 C HAPTER 2. E LECTROMAGNETIC WAVES

E
R B
X
R N
0 R j 0 R E (2.2)
@ X X
R R
According to the operator triple product “bac-cab” rule Equation (F.69) on page 159
98
E
E
; 2E (2.3)
Furthermore, since *° 0, Equation (2.1a) on the previous page yields
98
E 0 (2.4)
and since EEMF 0, Ohm’s law, Equation (1.26) on page 11, yields
j fe E (2.5)
we find that Equation (2.2) can be rewritten
N
; 2
E 0 R e E 0 R E 0 (2.6)
@ X X
R R
or, also using Equation (1.9) on page 5,
1 2
; 2
E E 0 0 eR E (2.7) R
X 2 X2 @
R R
which is the homogeneous wave equation for E.
We look for a solution in the form of a time-harmonic wave, and make therefore
the following Fourier component Ansatz
E E0 x
± Q² ³¬´ (2.8)
Inserting this into Equation (2.7), we obtain
1 2
; 2
E e0 R E0 x
± Q² ³¨´ R E0 x
± Q² ³¬´
@ X 2 X2
R R 1
; 2E 0 e $¶µ
E0 x
± Q² ³¬´ ·¸µ
2 E0 x
± Q² ³¨´
@ 2
1
; 2E 0 e ·¸µ
E w·¹µ
2 E 0 (2.9)
@ 2
which we can rewrite as
N µ 2 N e N µ 2 N ·
; 2
E 1 · E ; 2
E 1 º µ E 0 (2.10)
2 0µ 2
The quantity º ) 0 e is called the relaxation time of the medium in question. In

the limit of long º , Equation (2.10) tends to
2.2. P LANE WAVES 23
N µ 2
; 2
E E 0 (2.11)
2
which is a time-independent wave equation for E, representing weakly damped

propagating waves. In the short º limit we have instead
N
; 2
E ·¸µ 0e E 0 (2.12)
@
which is a time-independent diffusion equation for E.
For most metals º©» 10 14 s, which means that the diffusion picture is good for
all frequencies lower than optical frequencies. Hence, in metallic conductors, the
propagation term 2 E 2 X 2 is negligible even for VHF, UHF, and SHF signals.
R R
Alternatively, we may say that the displacement current 0 E X is negligible rel-
R R
ative to the conduction current j e E.
If we introduce the vacuum wave number
¼ µ
(2.13)

we can write, using the fact that ½ 1 ¿¾ 0 0 according to Equation (1.9) on
@
page 5,
1 1 0
ÀrÁÂ Ä Ã ÁÂÅÃÄ Æ¢Ç ÂÈÃ ÇÊÉ ËÄ ÂÃ ÇÍÌ 0 (2.14)
0 0 0
where in the last step we introduced the characteristic impedance for vacuum
0
Ì 0 ÂÎÉ ËÄ 376 Ð 7 Ω (2.15)
0 Ï
2.2 Plane waves

Consider now the case where all fields depend only on the distance Ñ to a given
plane with unit normal Ò ˆ . Then the del operator becomes
Ó ÒH
Â ˆ Ô (2.16)
Ô Ñ
and Maxwell’s equations attain the form

ÒÖˆ Õ Ô
E
0 (2.17a)
Â
Ô Ñ
Ò)ˆ × Ô
E Ô B (2.17b)
Â"Ø
Ô Ñ ÔÚÙ
ÒÖ
ˆ ÕÚÔ
B
0 (2.17c)
Â
Ô Ñ
Ò)
ˆ ×ÛÔ
B Ä Ô E Eß Ä Ô E
Â 0j Ü ÙYÝ x Þ ß 0 0 Â 0 0 0 (2.17d)
Ô Ñ Ë Ë ÔàÙ Ë Ã Ë ÔÚÙ
Scalar multiplying (2.17d) by Ò ˆ , we find that
0Â ÒÖ
ˆ Õ á Ò)
ˆ ×¦Ô
B ÒÖ
ˆ Õ á ß Ä Ô E (2.18)
Â 0 0 0
Ô Ñ â Ë Ã Ë Ô Ù â
à
which simplifies to the first-order ordinary differential equation for the normal
component ãä of the electric field
- ã ä ß ãä Â 0 (2.19)
ÃÄ
- Ù 0
with the solution
ãä Â ãwä 0 ¨
å æQç¢èêéë
0
Â ãä 0 ¨
å æìèêéMí (2.20)
This, together with (2.17a), shows that the longitudinal component of E, i.e., the
component which is perpendicular to the plane surface is independent of Ñ and has
a time dependence which exhibits an exponential decay, with a decrement given by
the relaxation time À in the medium.
Scalar multiplying (2.17b) by Ò ˆ , we similarly find that
0Â ÒÖ
ˆ Õ á Ò)
ˆ ×¦Ô
E ÒÖÕQÔ B
ÂØ ˆ (2.21)
Ô Ñ â ÔàÙ
or
ÒÖ
ˆ Õ Ô
B
0 (2.22)
Â
ÔàÙ
From this, and (2.17c), we conclude that the only longitudinal component of B
must be constant in both time and space. In other words, the only non-static solu-
tion must consist of transverse components.
2.2. P LANE WAVES 25
2.2.1 Telegrapher’s equation

In analogy with Equation (2.7) on page 22, we can easily derive the equation
2 1 Ô 2E
Ô E Ô E
2 Ø 0 Ø Æ 2 2 Â 0 (2.23)
Ô Ñ Ë Ã ÔàÙ ÔÚÙ
This equation, which describes the propagation of plane waves in a conducting
medium, is called the telegrapher’s equation. If the medium is an insulator so that
Â 0, then the equation takes the form of the one-dimensional wave equation
Ã 2
Ô E 1 Ô 2E
Ø Æ 2 Â
0 (2.24)
Ô Ñ
2 2
ÔÚÙ
As is well known, each component of this equation has a solution which can be
written
ã î Â)ï Ü Ñ Ø Æ Ù Þ ßð Ü ÑÊß Æ Ù Þ Ý ñ Â 1 Ý 2 Ý 3 (2.25)
where ï and ð are arbitrary (non-pathological) functions of their respective argu-
ments. This general solution represents perturbations which propagate along Ñ ,
where the ï perturbation propagates in the positive Ñ direction and the ð perturba-
tion propagates in the negative Ñ direction.
If we assume that E is time-harmonic, i.e., can be represented by a Fourier com-
ponent proportional to exp Øwò Á Ù , the solution of Equation (2.24) becomes
E Â E0 å¨æQóõô÷ö¨èêøùYúüû (2.26)
By introducing the wave vector
Á Á ý
k Â Ç Òˆ Â Æ Òˆ Â Æ ˆ (2.27)
this solution can be written as
E Â E0 å¢óõô k þ x æìö¬èÿû (2.28)
Let us consider the minus sign in the exponent in Equation (2.26) above. This
corresponds to a wave which propagates in the direction of increasing Ñ . Inserting
this solution into Equation (2.17b) on the facing page, gives
Ò)
ˆ ×ÛÔ
E Á Ç Ò ×
Ñ Âfò B Âò ˆ E (2.29)
Ô
or, solving for B,
B Â Á Ò)
Ç 1 1ý
ˆ × EÂ Á k× EÂ Æ ˆ × EÂ
Ä Ò)× E
0 0 ˆ (2.30)
Ë
Hence, to each transverse component of E, there exists an associated magnetic field
given by Equation (2.30) above. If E and/or B has a direction in space which is
constant in time, we have a plane polarised wave (or linearly polarised wave).
2.2.2 Waves in conductive media

Assuming that our medium has a finite conductivity , and making the time-
Ã
harmonic wave Ansatz in Equation (2.23) on the preceding page, we find that the
time-independent telegrapher’s equation can be written

2 2
Ô Eß Ä Á 2
Eß ò Á E Ô Eß 2
EÂ 0 (2.31)
0 0 0 Â
Ô Ñ Ô Ñ
2 2
Ë Ë Ã
where
Á

2
2 Ä Á 2 á 1ß á 1ß Ç 2 á 1ß (2.32)
Â 0 0 òÄ Ã Á â Â Æ 2 òÄ Ã Á â Â òÄ Ã Á â
Ë 0 0 0
where, in the last step, Equation (2.13) on page 23 was used to introduce the wave
number Ç . Taking the square root of this expression, we obtain
Â
Ç
É 1ß ò Ä Ã ÁÂ
0
ß ò (2.33)

where
É 1 ß
Ç
2
ë ö
ç ß 1
FÂ 0
(2.34a)
2
É
ë ç ö
Ç
2
1ß Ø 1
1Â 2
0
(2.34b)
Hence, the solution of the time-independent telegrapher’s equation, Equation (2.31),

can be written
EÂ E0 å¨æ ú'å¢óõô¬ú_æìö¬èêû (2.35)
With the aid of Equation (2.30) on the previous page we can calculate the associ-
ated magnetic field, and find that it is given by
BÂ Á
1
ýˆ ×
EÂ Á
1 ýˆ ×
Ü E ÞÜ ß ò Þ Â Á
1 ýˆ ×
Ü EÞ å¢ó (2.36)
where we have, in the last step, rewritten ß ò in the amplitude-phase form

exp ò . From the above, we immediately see that E is damped and that E and

B in the wave are out of phase.
In the case that Ä 0 Á , we can approximate as follows:
Ã
Ç á 1ß
1
2
ò Ä Ã Á
1 Ø1ò Ä
Ç 0
Á 1
2
Ç Ü1ß
Â òÄ Ã Á â Â
Á Ï
òÞ É 2Ä Ã 0 Á

0 0
Ã
Ä Á Ü1ß 0
Â 0 0 òÞ É 2Ä Ã ÁÂ
Ü1ß òÞ É Ë 2Ã (2.37)
Ë 0
From this analysis we conclude that when the wave impinges perpendicularly
upon the medium, the fields are given, inside this medium, by
Á Á
E Â E0 exp Ø É ! Ë 2Ã
0
#"
Ñ exp ò á É ! Ë 2Ã
0
Ñ Ø Á Ù
â
" (2.38a)
Ü1ß òÞ É
0 Ò ×
B Â Ë 2Á Ã Ü ˆ E Þ (2.38b)
Hence, both fields fall off by a factor 1 å at a distance $

% Â
2
Á (2.39)
%
0
Ë Ã
This distance is called the skin depth.
2.3 Observables and averages

In the above we have used complex notation quite extensively. This is for mathem-
atical convenience only. For instance, in this notation differentiations are almost

trivial to perform. However, every physical measurable quantity is always real val-
ued. I.e., “Ephysical Â Re Emathematical .” It is particularly important to remember
'&)(
this when one works with products of physical quantities.
Generally speaking, we tend to measure temporal averages Ü !Þ of our physical

observables. If we have two physical vectors F and G which both are time har-
monic, i.e., can be represented by Fourier components proportional to exp Øò Á Ù ,
it is easy to show that the average of the product of the two physical quantities

represented by F and G can be expressed as
& Re F Õ Re G ( Â 12 Re F Õ G* Â
1
2
Re F Õ G * (2.40)
where * denotes complex conjugate.
27
B IBLIOGRAPHY 2
ISBN 0-201-05702-6.
29
30
C HAPTER 3
Electromagnetic
Potentials
Instead of expressing the laws of electrodynamics in terms of electric and magnetic
fields, it turns out that it is often more convenient to express the theory in terms
of potentials. In this Chapter we will introduce and study the properties of such
potentials.
3.1 The electrostatic scalar potential

As we saw in Equation (1.6) on page 4, the electrostatic field Estat Ü x Þ is irrotational.
Hence, it may be expressed in terms of the gradient of a scalar field. If we denote
,+
this scalar field by Ø stat Ü x Þ , we get
Estat Ü x Þ ÂØ
Ó
stat Ü x Þ + (3.1)
-
Taking the divergence of this and using Equation (1.7) on page 4, we obtain Pois-
sons’ equation
2
+ stat Ü x Þ ÂØ
Ó Õ E Ü xÞ
stat Â Ø
/. Ä Ü xÞ
0
(3.2)
A comparison with the definition of Estat , namely Equation (1.5) on page 4, after
Ó
the has been moved out of the integral, shows that this equation has the solution
+ stat Ü
0 1 2 x. Ø x - ¨ß
xÞ Â
1ÜxÞ
4 Ä 0
. 3
(3.3)
where the integration is taken over all source points x at which the charge density
Þ is non-zero and is an arbitrary quantity which has a vanishing gradient.
. Ü x example
An of such a quantity is a scalar constant. The scalar function + Ü x Þ in
stat
Equation (3.3) above is called the electrostatic scalar potential.
31
32 C HAPTER 3. E LECTROMAGNETIC P OTENTIALS
3.2 The magnetostatic vector potential
3
Consider the equations of magnetostatics (1.20) on page 8. From Equation (M.74)
Ó Õ Ó ×
on page 180 we know that any 3D vector a has the property that Ü aÞ 0
and in the derivation of Equation (1.15) on page 7 in magnetostatics we found that
Ó Õ
Bstat Ü x Þ Â 0. We therefore realise that we can always write
Ó ×
Bstat Ü x Þ Â Astat Ü x Þ (3.4)
where Astat Ü x Þ is called the magnetostatic vector potential.
We saw above that the electrostatic potential (as any scalar potential) is not
unique: we may, without changing the physics, add to it a quantity whose spa-
tial gradient vanishes. A similar arbitrariness is true also for the magnetostatic
vector potential.
In the magnetostatic case, we may start from Biot-Savart’s law as expressed by
Ó
Equation (1.13) on page 6 and “move the out of the integral:”
jÜ x Þ ×
xØ x 3.
0 12 x Ø x -
0
Bstat Ü x Þ Â
4Ë 3
Ó á
jÜ x Þ ×
1 3.
ÂØ 4Ë 0 x Ø x â
12
0
-
Ó × jÜ x Þ .
4Ë 0 1 2 x Ø x -
0 3
Â (3.5)
An identification of terms allows us to define the static vector potential as

jÜ x Þ 3.
0 1 2 x Ø x -
0
Astat Ü x Þ Â ¨ß a Ü x Þ (3.6)
4Ë
where a Ü x Þ is an arbitrary vector field whose curl vanishes. From Equation (M.70)
on page 179 we know that such a vector can always be written as the gradient of a
scalar field.
3.3 The electromagnetic scalar and vector

potentials
Let us now generalise the static analysis above to the electrodynamic case, i.e., the
.
case with temporal and spatial dependent sources Ü ÙYÝ x Þ and j Ü ÙYÝ x Þ , and corres-
ponding fields E Ü Ù_Ý x Þ and B Ü ÙYÝ x Þ , as described by Maxwell’s equations (1.43) on
page 14. In other words, let us study the electromagnetic potentials Ü Ù_Ý x Þ and
A Ü ÙYÝ x Þ .
+
3.3. T HE ELECTROMAGNETIC SCALAR AND VECTOR POTENTIALS 33
From Equation (1.43c) on page 14 we note that also in electrodynamics the

Ó
homogeneous equation Õ B Ü Ù_Ý x Þ Â 0 remains valid. Because of this divergence-
free nature of the time- and space-dependent magnetic field, we can express it as
the curl of an electromagnetic vector potential:
Ó × A Ü xÞ
B Ü Ù_Ý x Þ Â YÙ Ý (3.7)
Inserting this expression into the other homogeneous Maxwell equation, Equa-
tion (1.30) on page 11, we obtain
Ó × E Ü xÞ
Ù_Ý ÂØ
Ô /Ó
Ô Ù
à
× A Ü xÞ
ÙYÝ ÂØ
Ó
54 ×«Ô A Ü x Þ
ÔàÙ
Ù_Ý (3.8)
or, rearranging the terms,
Ó × á E Ü xÞ ß Ô A Ü xÞ
Ù_Ý ÙYÝ â Â 0 (3.9)
Ô Ù
à
As before we utilise the vanishing curl of a vector expression to write this vector
expression as the gradient of a scalar function. If, in analogy with the electrostatic
case, we introduce the electromagnetic scalar potential function Ø Ü ÙYÝ x Þ , Equa-
tion (3.9) becomes equivalent to
6+
E Ü Ù_Ý x Þ ß Ô A Ü xÞ
Ô Ù
à
YÙ Ý Â Ø
H
Ó
+ Ü ÙYÝ x Þ (3.10)
This means that in electrodynamics, E Ü Ù_Ý x Þ can be calculated from the formula
E Ü Ù_Ý x Þ
Â Ø
Ó
+ Ü Ù_Ý x Þ Ø Ô A Ü xÞ
Ô Ù
à
YÙ Ý (3.11)
and B Ü Ù_Ý x Þ from Equation (3.7) above. Hence, it is a matter of taste whether we
want to express the laws of electrodynamics in terms of the potentials Ü Ù_Ý x Þ and +
A Ü Ù_Ý x Þ , or in terms of the fields E Ü Ù_Ý x Þ and B Ü Ù_Ý x Þ . However, there exists an im-
portant difference between the two approaches: in classical electrodynamics the
only directly observable quantities are the fields themselves (and quantities de-
rived from them) and not the potentials. On the other hand, the treatment becomes
significantly simpler if we use the potentials in our calculations and then, at the
final stage, use Equation (3.7) and Equation (3.11) above to calculate the fields or
physical quantities expressed in the fields.
Inserting (3.11) and (3.7) into Maxwell’s equations (1.43) on page 14 we obtain,
-
after some simple algebra and the use of Equation (1.9) on page 5,
- + Ô Ü Ó Õ AÞ
/.
Ü YÙ Ý x Þ
2
ß Â Ø
Ä (3.12a)
Ô Ù
Ú 0
Ó á Ó Õ
+â
2
A Ø Æ 2 Ô 2A Ø
1 1 Ô
2
Aß Æ 2 ÂØ 0j Ü Ù_Ý x Þ (3.12b)
ÔàÙ ÔàÙ Ë
These two second order, coupled, partial differential equations, representing in all
+
four scalar equations (one for and one each for the . 1 Ý . 2 , and . 3 components
of A) are completely equivalent to the formulation of electrodynamics in terms of
Maxwell’s equations, which represent six scalar first-order, coupled, partial differ-
ential equations (one equation for each of the components of E and B).
3.3.1 Electromagnetic gauges

Lorentz equations for the electromagnetic potentials
As they stand, Equations (3.12) look complicated and may seem to be of lim-
ited use. However, if we write Equation (3.7) on the preceding page in the form
Ó × Ó ×
A Ü Ù_Ý x Þ Â B Ü ÙYÝ x Þ we can consider this as a specification of A. But we
know from Helmholtz’ theorem that in order to determine the (spatial behaviour)
Ó Õ
of A completely, we must also specify A. Since this divergence does not enter
Ó
the derivation above, we are free to choose Õ A in whatever way we like and still
obtain the same physical results! This is somewhat analogous to the freedom of
adding an arbitrary scalar constant (whose grad vanishes) to the potential energy in
classical mechanics and still get the same force.
Ó
With a judicious choice of Õ A, the calculations can be simplified considerably.
Lorentz introduced
Ó 1 Ô
2
Õ Aß
ÔÚÙ
Â 0 Æ + (3.13)
which is called the Lorentz gauge condition, because this choice simplifies the sys-
tem of coupled equations (3.12) above into the following set of uncoupled partial
differential equations: -
7 + 3 2 def
Æ
1 Ô 2
+ - + Â8. Ü ÄÙYÝ xÞ
2
(3.14a)
2 2 Ø
ÔàÙ
7
0
2
A 3
def 1 Ô 2
Æ 2 2A Ø
2
AÂ 0j Ü ÙYÝ x Þ (3.14b)
7 ÔàÙ Ë
where 2 is the d’Alembert operator discussed in Example M.4 on page 176. We
shall call (3.14) the Lorentz equations for the electromagnetic potentials.
Gauge transformations
We saw in Section 3.1 on page 31 and in Section 3.2 on page 32 that in electro-
statics and magnetostatics we have a certain mathematical degree of freedom, up
to terms of vanishing gradients and divergences, to pick suitable forms for the po-
tentials and still get the same physical result. In fact, the way the electromagnetic
+
scalar potential Ü Ù_Ý x Þ and the vector potential A Ü Ù_Ý x Þ are related to the physically
observables gives leeway for similar “manipulation” of them also in electrodynam-
+
ics. If we transform Ü Ù_Ý x Þ and A Ü ÙYÝ x Þ simultaneously into new ones Ü Ù_Ý x Þ and
A Ü Ù_Ý x Þ according to the scheme
+
+ Ü Ù_Ý x Þ Â9+ Ü ÙYÝ x Þ ßÔ Γ Ü xÞ
ÔÓ Ù
Ú
Ù_Ý (3.15a)
A Ü Ù_Ý x Þ Â A Ü Ù_Ý x Þ Ø Γ Ü Ù_Ý x Þ (3.15b)
where Γ Ü Ù_Ý x Þ is an arbitrary, differentiable scalar function called the gauge func-
tion, and insert the transformed potentials into Equation (3.11) on page 33 for the
electric field and into Equation (3.7) on page 33 for the magnetic field, we obtain
the transformed fields
E Â"Ø
Ó
Ø
Ô A
ÔàÙ
+ ÂØ
Ó
Ø
Ô Ó Γ
ÔàÙ
Ø
ÔÚÙ
+
Ô Aß Ô Ó Γ
ÔàÙ
Â"Ø
Ó ×
Ó
Ø
Ô A
ÔàÙ Ó
+ Ó Ó Ó
(3.16a)
B Â A Â × A × ΓÂ × A (3.16b)
Ø
where, once again Equation (M.70) on page 179 was used. Comparing these ex-
pressions with (3.11) and (3.7) we see that the fields are unaffected by the gauge
transformation (3.15). A transformation of the potentials and A which leaves
the fields, and hence Maxwell’s equations, invariant is called a gauge transforma-
+
tion. A physical law which does not change under a gauge transformation is said
to be gauge invariant. By definition, the fields themselves are, of course, gauge
invariant.
with an arbitrary choice of

Ó Õ +
The potentials Ü ÙYÝ x Þ and A Ü ÙYÝ x Þ calculated from (3.12) on the facing page,
A, can be further gauge transformed according to
Ó
(3.15) above. If, in particular, we choose Õ A according to the Lorentz condition,
Equation (3.13) on the facing page, and apply the gauge transformation (3.15) on
- -
the resulting Lorentz equations (3.14) on the preceding page, these equations will
be transformed into
+ - + . Ü ÄÙ_Ý xÞ
2 2
1 Ô Ô á 1 Ô Γ
- Γ
2 2
Æ 2 ©Ø Ø Æ 2 2 Ø Â (3.17a)
2
ÔàÙ Ô Ù
à ÔÚÙ â 0
1 Ô 2 Ó á 1 Ô 2
Æ 2 2Γ Ø Γ 0j Ü Ù_Ý x Þ
2 2
Æ 2 2A Ø AØ
â Â (3.17b)
ÔÚÙ ÔàÙ Ë
We notice that if we require that the gauge function Γ Ü _Ù Ý x Þ itself be restricted to

fulfil the wave equation -
1 Ô 2
2
ΓÆ
2 Ø
2
ΓÂ 0 (3.18)
ÔàÙ
these transformed Lorentz equations will keep their original form. The set of po-
tentials which have been gauge transformed according to Equation (3.15) on the
preceding page with a gauge function Γ Ü ÙYÝ x Þ which is restricted to fulfil Equa-
tion (3.18) above, i.e., those gauge transformed potentials for which the Lorentz
equations (3.14) are invariant, comprises the Lorentz gauge.
Other useful gauges are
: The radiation gauge, also known as the transverse gauge, defined by
Ó Õ
A Â 0.
: The Coulomb gauge, defined by Â 0,
Ó Õ
A Â 0. +
: The temporal gauge, also known as the Hamilton gauge, defined by + Â 0.
: The axial gauge, defined by 3 Â 0.
The process of choosing a particular gauge condition is referred to as gauge fixing.
3.3.2 Solution of the Lorentz equations for the electromag-

netic potentials
As we see, the Lorentz equations (3.14) on page 34 for Ü Ù_Ý x Þ and A Ü ÙYÝ x Þ represent +
a set of uncoupled equations involving four scalar equations (one equation for
and one equation for each of the three components of A). Each of these four scalar
+
7
equations is an inhomogeneous wave equation of the following generic form:
2
Ψ Ü ÙYÝ x Þ )
Â ï Ü ÙYÝ x Þ (3.19)
+
where Ψ is a shorthand for either or one of the vector components of A, and ï is
the pertinent generic source component.
We assume that our sources are well-behaved enough in time Ù so that the Fourier
transform pair for the generic source function
; 54 3 def
Ü x Þ å¨æQó ö¬è - Á
/ï Ü xÞ ï Ü ÙYÝ x Þ Â
1 æ< ïö
1
æ (3.20a)
ö
; / ï Ü Ù_Ý xÞ54 3 1
<
20 1 <
def
ï Ü xÞ Â ï Ü ÙYÝ x Þ ¢å ó ö¬è - Ù (3.20b)
ö
æ
exists, and that the same is true for the generic potential component: <
Á
1 æ < Ψö Ü xÞ å¨æQó ö¨è -
Ψ Ü ÙYÝ x Þ Â (3.21a)
0 < 1 æ < Ψ Ü ÙYÝ xÞ å¢ó ö¬è - Ù

1
Ψ Ü xÞ Â (3.21b)
ö 2
<
Inserting the Fourier representations (3.20a) and (3.21a) into Equation (3.19) on
the preceding page, and using the vacuum dispersion relation for electromagnetic
waves
Á Æ:Ç (3.22)
Â
-
the generic 3D inhomogeneous wave equation Equation (3.19) on the facing page
turns into
Ψ Ü x Þ ß Ç 2 Ψ Ü x Þ ÂHØÊï Ü x Þ
2
(3.23)
ö ö ö
which is a 3D inhomogeneous time-independent wave equation, often called the
3D inhomogeneous Helmholtz equation.
As postulated by Huygen’s principle, each point on a wave front acts as a point
source for spherical waves which form a new wave from a superposition of the
individual waves from each of the point sources on the old wave front. The solution
-
of (3.23) can therefore be expressed as a superposition of solutions of an equation
where the source term has been replaced by a point source:
2 = Ü xÝ x Þ ß Ç 2 = Ü x Ý x Þ Â"Ø
% ÜxØ xÞ (3.24)
and the solution of Equation (3.23) above which corresponds to the frequency Á is
given by the superposition
= Ü x Ý x Þ - 3.
Ψ Â
ö 1 ï
ö
ÜxÞ (3.25)
=
The function Ü x Ý x Þ is called the Green’s function or the propagator.
%
In Equation (3.24), the Dirac generalised function Ü x Ø x Þ , which represents
>
the point source, depends only on x Ø x and there is no angular dependence in
>
the equation. Hence, the solution can only be dependent on Â x Ø x . If we
interpret as the radial coordinate in a spherically polar coordinate system, the
“spherically symmetric” Ü Þ is given by the solution of = >
>- >
= >= > % Ü> Þ
2
- Ü Þ ß Ç 2
Ü Þ ÂØ (3.26)
2
Away from > Â x Ø x Â 0, i.e., away from the source point x , this equation
takes the form
= >=
>- Ü >
2
- Þ ß Ç 2
Ü Þ Â 0 (3.27)
2
with the well-known general solution

= Â@?BA å ó ùDC ß ? æ å æQó ùDC 3 ?BA å ó÷ùFE æ G E ß ? æ å æQó÷ùFE æ G E x x x x
> > x Ø x x Ø x (3.28)
where ? ø are constants.

In order to evaluate the constants ? ø , we insert the general solution Equa-
a small volume around > Â x Ø x Â 0. Since

tion (3.28) above into Equation (3.24) on the preceding page and integrate over
= Ü HH x Ø xIHH Þ J ?BA 1 ß ? æ 1 Ý HH x Ø xKHHML 0

x Ø x x Ø x (3.29)
N N
Equation (3.24) on the previous page can under this assumption be approximated
by -
?BA ß ? æO 1 á x Ø 1 x â - . ß Ç ?BA ß ? æO 1 x Ø 1 x - .
2 3 2 3
% Ü H x Ø xPH Þ .
1 H H-
3
Â"Ø (3.30)
In virtue of the fact that the volume element - . in spherical polar coordinates is 3
proportional to x Ø x , the second integral vanishes when x Ø x

L 0. Further-
2
%
integral can be written as Ø 40 Ü' x Ø x Þ and, hence, that
more, from Equation (M.65) on page 178, we find that the integrand in the first
?QA ß ? æ Â 410 (3.31)

Insertion of the general solution Equation (3.28) into Equation (3.25) on the
previous page gives
R?QA 1 å ó÷ù æ 3. FE G E
x x
å æQó ù æ 3.
? 1 FE G E
x x
Ψ Ü xÞ Â
ö
ï Ü x Þ x x - ¨ß
ö Ø
æ ï ÜxÞ x x -
ö Ø
(3.32)

The Fourier transform to ordinary Ù domain of this is obtained by inserting the
above expression for Ψ Ü x Þ into Equation (3.21a) on the preceding page:
FE G E U T
ö
exp Øò Á Ù Ø ù x æ x S
@?BA /1 1
Ψ Ü Ù_Ý x Þ Â ï ÜxÞ
ö x x
ö -
Á 3.
-
FE G E UT
Ø
exp Øwò Á Ù ß ù x æ x S
? 1/1
ß æ ï ÜxÞ
ö xØ x
ö - -
Á 3.

(3.33)

$ $
If we introduce the retarded time Ù ret and the advanced time Ù adv in the following
way [using the fact that in vacuum Ç Á Â 1 Æ , according to Equation (3.22) on
page 37]:
Ù ret Â HH KHH
Ù ret Ü Ù_Ý x Ø x Þ Â Ù Ø
Ç x x
Á

Ø
Â
Ù Ø
x Ø Æ x (3.34a)
Ü Ù_Ý H x Ø xKH Þ
Ç x x x Ø Æ x
Ù adv Â Ù adv H H Â Ù ß Á
Ø
Â Ù ß (3.34b)
and use Equation (3.20a) on page 36, we obtain
Ψ Ü ÙYÝ x Þ Â @?BA 1 ï xÜ Ù Ø Ý xx Þ
ret
-
3.
¨ß ? æ 1 ï Ü xÙ Ø xÝ x Þ
adv
-
3.
(3.35)
This is a solution of the generic inhomogeneous wave equation for the potential
components Equation (3.19) on page 36. We note that the solution at time Ù at
the field point x is dependent on the behaviour at other times Ù of the source at
x and that both retarded and advanced Ù are mathematically acceptable solutions.
However, if we assume that causality requires that the potential at Ü ÙYÝ x Þ is set up by
the source at an earlier time, i.e., at Ü Ù ret Ý x Þ , we must in Equation (3.35) set æ Â 0 ?$ 0
and therefore, according to Equation (3.31) on the preceding page, Â 1 rÜ 4 Þ . ?A
The retarded potentials
From the above discussion on the solution of the inhomogeneous wave equation
we conclude that under the assumption of causality the electromagnetic potentials
in vacuum can be written
+ Ü Ù_Ý x Þ Â
1
4 Ä . ÜÙ ÝxÞ
0 1 x Ø x -
0
ret 3.
(3.36a)
jÜ Ù Ý x Þ .
4Ë 0 1 x Ø x -
0 ret
A Ü Ù_Ý x Þ Â 3
(3.36b)
Since these retarded potentials were obtained as solutions to the Lorentz equations
(3.14) on page 34 they are valid in the Lorentz gauge but may be gauge transformed
according to the scheme described in Subsection 3.3.1 on page 35. As they stand,
we shall use them frequently in the following.
39
B IBLIOGRAPHY 3
[1] L. D. Fadeev and A. A. Slavnov. Gauge Fields: Introduction to Quantum The-

ory. Number 50 in Frontiers in Physics: A Lecture Note and Reprint Series. Ben-
jamin/Cummings Publishing Company, Inc., Reading, MA . . . , 1980. ISBN 0-8053-
9016-2.
[2] Mike Guidry. Gauge Field Theories: An Introduction with Applications.
Wiley & Sons, Inc., New York, NY . . . , 1991. ISBN 0-471-63117-5.
ISBN 0-201-05702-6.
41
42
C HAPTER 4
The Electromagnetic
Fields
While, in principle, the electric and magnetic fields can be calculated from the
Maxwell equations in Chapter 1, or even from the wave equations in Chapter 2, it
is often physically more lucid to calculate them from the electromagnetic potentials
derived in Chapter 3. In this Chapter we will derive the electric and magnetic fields
from the potentials.
We recall that in order to find the solution (3.35) for the generic inhomogen-
eous wave equation (3.19) on page 36 we presupposed the existence of a Fourier
transform pair (3.20a) on page 36 for the generic source term
Á
ï Ü Ù_Ý x Þ Â
1 æ< ï
ö
Ü xÞ ¨
å æQó ö¬è - (4.1a)
ï
ö
Ü xÞ Â
1
<
20 1 < ï Ü Ù_Ý x Þ ¢å ó ö¨è - Ù (4.1b)
æ
and for the generic potential component <
V V Ü x Þ å¨æQó ö¨è - Á
1 æ<
Ü ÙYÝ x Þ Â
ö
(4.2a)
V Ü xÞ Â 1 < V Ü Ù_Ý xÞ å¢ó ö¬è Ù

20 1 < - (4.2b)
ö
æ
That such transform pairs< exists is true for most physical variables which are not
strictly monotonically increasing and decreasing with time. For charges and cur-
rents varying in time we can therefore, without loss of generality, work with in-
dividual Fourier components. Strictly speaking, the existence of a single Fourier
component assumes a monochromatic source, which in turn requires that the elec-
tric and magnetic fields exist for infinitely long times. However, by taking the
proper limits, we can still use this approach even for sources and fields of finite
duration.
43
44 C HAPTER 4. T HE E LECTROMAGNETIC F IELDS
This is the method we shall utilise in this Chapter in order to derive the electric
and magnetic fields in vacuum from arbitrary given charge densities Ü Ù_Ý x Þ and
current densities j Ü Ù_Ý x Þ , defined by the Fourier transform pairs .
Á
. Ü ÙYÝ xÞ Â 1 æ < . ö Ü xÞ å¨æQó ö¨è - (4.3a)
.ö Ü xÞ Â
1
< Ü xÞ
20 1 < . ÙYÝ å¢ó ö¬è - Ù
(4.3b)
æ
and <
j Ü x Þ å¨æQó ö¬è - Á
j Ü ÙYÝ x Þ Â
1 æ< ö (4.4a)
j Ü xÞ Â
ö
1
< j Ü Ù_Ý xÞ å¢ó ö¨è - Ù
20 1 <
(4.4b)
æ
under the assumption < that only retarded potentials produce physically acceptable
solutions. 1
The Fourier transform pair for the retarded vector potential can then be written
+ 1 æ < + ö Ü xÞ å¨æQó ö¬è -

Ü Ù_Ý x Þ Â Á (4.5a)
< G
å ó÷ùFE æ E .
+ ö Â 20 1 < + å¢ó ö¬è - Â 40 Ä 1 . ö x Ø x -
1 1 x x
Ü x Þ Ü _
Ù Ý x Þ Ù Ü x Þ (4.5b) 3
0
æ
where in the last step, we< made use of theV explicit
form of the generic potential component
expression for the Fourier trans-
Ü x Þ , Equation (3.32) on page 38. Sim-
ilarly, the following Fourier transform pair ö for the vector potential must exist:
A Ü x Þ å¨æQó ö¨è - Á
A Ü ÙYÝ x Þ Â
1 æ< ö
(4.6a)
1
< FE G E
x x
j Ü x Þ å ó÷ù æ 3.
20 1 < 0 1
0
A Ü xÞ Â A Ü ÙYÝ x Þ å¢ó ö¬è - Ù Â (4.6b)
ö 4Ë ö xØ x -
æ
Clearly, we must require that <
A Â A Ý
ö æìö
* ö
Â
æìö
+ W+ * (4.7)
in order that all physical quantities be real. Similar transform pairs and require-
ments of real valuedness exist for the fields themselves.
In the limit that the sources can be considered as being monochromatic, we have
1
In fact, John A. Wheeler and Richard P. Feynman derived in 1945 a fully self-consistent electro-
dynamics using both the retarded and the advanced potentials [6]; See also [1].
4.1. T HE MAGNETIC FIELD 45
the much simpler expressions
. j ÜÜ Ù_ÙYÝÝ xxÞÞ ÂÂ j. Ü Ü xxÞ Þ å¨å¨æQæQó óö¬ö¬è è

0
0 (4.8a)
(4.8b)
+ Ü Ù_Ý xÞ ÂW+ Ü xÞ å¨æQó ö¨è 0 (4.8c)
A Ü Ù_Ý x Þ Â A0 Ü x Þ ¨
å æQó ö¬è (4.8d)
where again the real-valuedness of all these quantities is implied. As discussed

above, we can safely assume that all formulas derived for a general Fourier rep-
resentation of the source (general distribution of frequencies in the source) are
valid for these simple limiting cases. We note that in this context, we can make
Á Á Þ,j Á Á Þ etc., and that we % %
the formal identification
ö
Â 0 Ü Ø
ö
Â j0 Ü Ø
. .
therefore often identify
ö .
with 0 , j with j0 and so on.
ö .
4.1 The magnetic field

Let us now compute the magnetic field from the vector potential, defined by Equa-
tion (4.6a) and Equation (4.6b) on the preceding page, and Formula (3.7) on page 33:
Ó × A Ü xÞ
B Ü Ù_Ý x Þ Â ÙYÝ (4.9)
The calculations are much simplified if we work in Á space and, at the final
stage, Fourier transform back to ordinary Ù space. We are working in the Lorentz
gauge and note that in Á space the Lorentz condition, Equation (3.13) on page 34,
takes the form
Ç
Ó Õ A
ö
Ø ò Æ
F +ö Â 0 (4.10)
which provides a relation between (the Fourier transforms of) the vector and scalar
potentials.
Using the Fourier transformed version of Equation (4.9) and Equation (4.6b) on
the preceding page, we obtain
Ó × A Ü xÞ Ó × FE G E
x x
j Ü x Þ å ó÷ù æ 3.
0 12 X
0
B Ü xÞ Â Â (4.11)
ö ö 4Ë ö xØ x
Using Formula (F.62) on page 159, we can rewrite this as

j ÜxÞ × ZY Ó\[ å ó÷ùFE æ G E . x x
0 12 x Ø x ]B^_X
0
B Ü x Þ ÂØ Ë 3
ö 4 ö
ÂØ 4Ë 0
Y j Ü x Þ × [ Ø x Ø x å¢ó ùFE æ G E .
12 ö x Ø x ] X
0 x x 3
3
j ÜxÞ × á òÇ
xØ x
F E GE 1 .
ß
12 ö x Ø x â x Ø x X ^
x x 3
¢
å ó ù æ
Y j Ü x Þ å ó÷ùFE æ G E × Ü x Ø x Þ .
x x
Â 4Ë 0 X
1 2 ö x Ø x
0 3
3
Ü Øò Ç Þ j Ü x Þ å ó÷ùFE æ G E × Ü x Ø x Þ . x x
ß
12 x Ø x X ^
3
ö 2
(4.12)
From this expression for the magnetic field in the frequency (Á ) domain, we
$
obtain the magnetic field in the temporal (Ù ) domain by taking the inverse Fourier
transform (using the identity Øwò Ç ÂØwò Á Æ ):
B Ü x Þ å¨æQó ö¨è Á
B Ü ÙYÝ x Þ Â
ö 1 < X
< Sb FE G E X T
æ
j Ü x Þ å óõôõù x æ x æìö¬èÿû Á × Ü x Ø x Þ
3.
a0 ` 1 2
ö
X
0
Â 4Ë 3
xØ x
Sb FE G E X T
Ü Øwò Á Þ j Ü x Þ å ó¹ôõù x æ x æìö¬èÿû Á × Üx xÞ
X Kc
1 Ø 3.
ö
ß Æ
12 xØ x 2

j Ü Ù ret Ý x Þ × Ü x Ø x Þ 3.
d0 12 ef X g
0
Â 4Ë
xØ x 3
Induction field
j̇ Ü Ù ret Ý xÞ × ÜxØ xÞ 3.
d0 12 xe f Ø x X g
0
ß (4.13)
4Ë Æ 2
Radiation field
where
á Ô j 3
def
ih G
j̇ Ü Ù ret Ý xÞ (4.14)
ÔàÙ â
è rè ret
The first term, the induction field, dominates near the current source but falls off
rapidly with distance from it, while the second term, the radiation field or the far
field, dominates at large distances and represents energy that is transported out to
Ó
infinity. Note how the spatial derivatives ( ) gave rise to a time derivative (˙)!
4.2. T HE ELECTRIC FIELD 47
4.2 The electric field

To calculate the electric field, we use the Fourier transformed version of For-
mula (3.11) on page 33, inserting Equations (4.5b) and (4.6b) as the explicit ex-
pressions for the Fourier transforms of and A: +
E Ü x Þ ÂØ
Ó
+
Ü xÞ ß ò Á A Ü xÞ
ö ö
1 Ó
ö
å ó÷ù æ FE G E
x x
3. ò 0
Á
j Ü x Þ å ó÷ù æ
x x
3.FE G E
ÂØ 4 Ä 0 12 . Xö
Ü x Þ
x Ø x
¨ß Ë 4 ö 0 12 x Ø x X
Ü x Þ å ó÷ùFE æ G E Ü x Ø x Þ .
0
1 Y
.
x x
40 12 x Ø x X
3
Â Ä ö 3
0
j Ü x Þ å ó ùjE æ G E .
1 2 . x Ø x Ø ö Æ â x Ø x X ^ (4.15)
Ç á Ü x Þ'Ü x Ø x Þ x x
3
Øò ö
Using the Fourier transform of the continuity equation (1.21) on page 9

Ó Õ j ÜxÞ Á
ö
Ø1ò
ö .
ÜxÞ Â 0 (4.16)
we see that we can express

Ó
ö .
in terms of j as follows
ö
Ü x Þ ÂHØ Á ò Õ
. ö
j ÜxÞ
ö
(4.17)
$
Doing so in the last term of Equation (4.15) above, and also using the fact that
Ç Á Æ , we can rewrite this Equation as
Â
Y FE G E
0 12 .ö
1 Ü x Þ å ÷ó ù x æ x Ü x Ø x Þ 3.
E Ü xÞ Â
ö 4 Ä 0
xØ x 3 X
1 á
Ó Õ j Ü x Þ Ü x x Þ
Ø Ç jE G E .
x x
å óù æ
d1 2
Ø ò j ÜxÞâ
ef X g^
3
Ø Æ ö F
xØ x ö xØ x
I
ö (4.18)
The last integral can be further rewritten in the following way:
á
Ó Õ j Ü x ÞÜ x x Þ
Ø Ç j Üx Þ å óù æ
x x
3. jE G E
I Â
12 ö
Øò â xØ x X
ö
n n
ö
o n jE G E
xØ x
k . .
k n
1 2 ml l
x x
á Ô Ø Ç Ü x Þ ˆ àå ó ù æ 3.
Â
Ô
ö.
xØ x F
Ø ò
ö â xØ x X (4.19)
But, since
[k nØ .n .
G k . n . n
G
.
Ô
öpl x Ø x
å¢ó ùFE æ E ]
Â
x x
. p
ö l â x Ø x å¢ó÷ùFE æ E
á Ô Ø x x
l l
2 2
Ô Ô
[ . n . n
ß k Ô Ø G (4.20)
å¢ó÷ùFE æ E ]
öpl Ô xØ x
x x
.
l
2
we can rewrite I as
Y k . Ô [ . n Ø . n o ˆ n å¢ó ùFE æ G E ß ò Ç j å ó÷ùFE æ G E .

ö
x x
1 2 öpl Ô l x Ø x ] ö x Ø x ^ X
x x 3
I Â"Ø 2
ö
[ . n . n
Ô
k Ø o ˆ n å¢ó ùFE æ G E .
ß
1 2 Ô l öpl x Ø x ] X
x x 3
. 2
(4.21)
where, according to Gauss’s theorem, the last term vanishes if j is assumed to be

ö the derivative in
limited and tends to zero at large distances. Further evaluation of
the first term makes it possible to write
[ FE G E r FE æ G E ] X
q
x x
j Õ Ü x Ø x Þ Ü x Ø x Þ å¢ó÷ù
å óù æ 2 3.
12 ß x x
I "Â Ø Ø j 2 4
ö ö xØ x xØ x ö
Ç [ s4
/ j Õ Ü x Ø x Þ ¿Ü x Ø x Þ
FE G E
x x
å óù æ FE G E 3.
12 ß
] X
x x
Øò Ø ö å ¢ ÷
ó ù æ j (4.22)
xØ x 3 ö xØ x
Using the triple product “bac-cab” Formula (F.56) on page 159 backwards, and
inserting the resulting expression for I into Equation (4.18) on the previous page,
we arrive at the following final expression ö for the Fourier transform of the total
jE G E GE
E-field:
1 Ó å óù æ
x x
3. ò 0 jE
Á x x
å óù æ 3.
E Ü x Þ ÂØ
ö 4 Ä 0 0 12 . X
ö
Ü x Þ
xØ x 0 12 ¨ß
Ë4 j
ö
Ü x Þ
xØ x X
1 Y . jE G E
Ü x Þ å ó ù x æ x Ü x Ø x Þ 3.
Â 4 Ä ö
0 12 xØ x 3X
FE G E
0
1 s4
/ j Ü x Þ å ó÷ù x æ x Õ Ü x Ø x Þ MÜ x Ø x Þ 3.
ß Æ
12ö
X
xØ x 4
1 FE G E s4
/ j Ü x Þ å ó÷ù x æ x × Ü x Ø x Þ × Ü x Ø x Þ 3.
ß Æ
12 ö
X
xØ x 4
ò
Ç
FE G E s4
/ j Ü x Þ å ó÷ù x æ x × Ü x Ø x Þ × Ü x Ø x Þ 3.
Ø Æ
12 ö
xØ x 3 X (4.23)
^
t
Taking the inverse Fourier transform of Equation (4.23), once again using the
vacuum relation Ç Æ , we find, at last, the expression in time domain for the
Â
total electric field:
Á
E Ü Ù_Ý x Þ Â
1 < æ
X
E Ü xÞ ¨
ö
å æQó ö¬è
<
d 0 1 2 . e f
1 Ü Ù ret Ý x Þ'Ü x Ø x Þ 3.
Â 4 0 Ä
xØ x 3 X g
Retarded Coulomb field
1 / jÜ Ù ret Ý x Þ Õ Ü x Ø x Þ MÜ x Ø x Þ s4 3.
d0 12 e xf Ø x X g
ß
4 Ä 0Æ 4
Intermediate field
1 / jÜ Ù ret Ý xÞ × ÜxØ xÞ × ÜxØ xÞ 54 3.
d0 12 e f X g
ß
4 Ä 0Æ xØ x 4
Intermediate field
1 / j̇ Ü Ù ret Ý xÞ × ÜxØ xÞ × ÜxØ xÞ s4 3.
d0 12 e xf Ø x X g
ß Ä
4 0Æ 2 3
Radiation field
(4.24)
Here, the first term represents the retarded Coulomb field and the last term repres-
ents the radiation field which carries energy over very large distances. The other
two terms represent an intermediate field which contributes only in the near zone
and must be taken into account there.
With this we have achieved our goal of finding closed-form analytic expressions
for the electric and magnetic fields when the sources of the fields are completely
arbitrary, prescribed distributions of charges and currents. The only assumption
made is that the advanced potentials have been discarded; recall the discussion
following Equation (3.35) on page 39 in Chapter 3.
49
B IBLIOGRAPHY 4
[1] Sir Fred Hoyle and Jayant V. Narlikar. Lectures on Cosmology and Action at a Dis-
tance Electrodynamics. World Scientific Publishing Co. Pte. Ltd, Singapore, New
Jersey, London and Hong Kong, 1996. ISBN 9810-02-2573-3(pbk).
ISBN 0-201-05702-6.
[5] Julius Adams Stratton. Electromagnetic Theory. McGraw-Hill Book Company, Inc.,
New York, NY and London, 1953. ISBN 07-062150-0.
[6] John Archibald Wheeler and Richard Phillips Feynman. Interaction with the absorber
as a mechanism for radiation. Reviews of Modern Physics, 17, 1945.
51
52
C HAPTER 5
Relativistic
Electrodynamics
We saw in Chapter 3 how the derivation of the electrodynamic potentials led, in a
$
most natural way, to the introduction of a characteristic, finite speed of propagation
that equals the speed of light Æ Â 1 Ä 0 0 and which can be considered as a con-
stant of nature. To take this finite speed ofË propagation of information into account,
and to ensure that our laws of physics be independent of any specific coordinate
frame, requires a treatment of electrodynamics in a relativistically covariant (co-
ordinate independent) form. This is the object of the current chapter.
5.1 The special theory of relativity

An inertial system, or inertial reference frame, is a system of reference, or rigid
coordinate system, in which the law of inertia (Galileo’s law, Newton’s first law)
holds. In other words, an inertial system is a system in which free bodies move
uniformly and do not experience any acceleration. The special theory of relativity
describes how physical processes are interrelated when observed in different iner-
tial systems in uniform, rectilinear motion relative to each other and is based on
two postulates:
Postulate 1 (Relativity principle; Poincaré, 1905) All laws of physics (except the
laws of gravitation) are independent of the uniform translational motion of the
system on which they operate.
Postulate 2 (Einstein, 1905) The velocity of light is independent of the motion of
the source.
A consequence of the first postulate is that all geometrical objects (vectors,
tensors) in an equation describing a physical process must transform in a covariant
manner, i.e., in the same way.
53
54 C HAPTER 5. R ELATIVISTIC E LECTRODYNAMICS
uDv ww xx
| |

vzvyiyi{j{y}|~y|'yi yi P
Σ Σ v
{ {

Figure 5.1. Two inertial systems Σ and Σ in relative motion with velocity v

of Σ coincided with the

along the axis. At time 0 the origin
origin of Σ. At time , the inertial system Σ has been translated a distance

along the axis in Σ. An event represented by Ps K
in Σ is represented
. by K K
in Σ .
5.1.1 The Lorentz transformation
Let us consider two three-dimensional inertial systems Σ and Σ in vacuum which

are in rectilinear motion relative to each other in such a way that Σ moves with
constant velocity v along the . axis of the Σ system. The times and the spatial co-
P D P D
ordinates as measured in the two systems are Ù and Ü . Ý Ý Þ , and Ù and Ü . Ý Ý Þ ,
respectively. At time Ù Â Ù Â 0 the origins and
and the . and . axes of the
two inertial systems coincide and at a later time Ù they have the relative location as
depicted in Figure 5.1.
For convenience, let us introduce the two quantities
1ÂÆ (5.1)
Â 1 1Ø 2
(5.2)
where Â v . In the following, we shall make frequent use of these shorthand

notations.
As shown by Einstein, the two postulates of special relativity require that the
spatial coordinates and times as measured by an observer in Σ and Σ , respectively,
are connected by the following transformation:
5.1. T HE SPECIAL THEORY OF RELATIVITY 55
Æ
Ù Â ÜÆ Ù Ø . Þ (5.3a)
Â
. Ü. Ø
Â
ÙÞ
(5.3b)
(5.3c)
Â (5.3d)
N
Taking the difference between the square of (5.3a) and the square of (5.3b) we find
that
Æ 2 2
Ù Ø
. 2

Â
2 Æ 2 2
Ù Ø 2
. Æ
Ù ß
. 2 2
Ø
. 2
ß 2.
Ù Ø
ÙO
2 2
Â
1 Æ 2 2
Ù
á 1
Ø Æ
2
Ø
. 2 á 1
Ø Æ â
2
Æ
2 2â 2
1Ø 2
Æ 2 2 . 2
Â Ù Ø (5.4)
From Equation (5.3) on the facing page we see that the and coordinates are
unaffected by the translational motion of the inertial system Σ along the . axis of

system Σ. Using this fact, we find that we can generalise the result in Equation (5.4)
above to
Æ 2 2
Ù Ø
. 2
Ø 2
Ø 2
Â
Æ 2 2
Ù Ø
. 2
Ø 2
Ø 2
(5.5)

which means that if a light wave is transmitted from the coinciding origins and
at time Ù Â Ù Â 0 it will arrive at an observer at Ü . Ý Ý Þ at time Ù in Σ and an P D
P D
observer at Ü . Ý Ý Þ at time Ù in Σ in such a way that both observers conclude
that the speed (spatial distance divided by time) of light in vacuum is Æ . Hence, the
speed of light in Σ and Σ is the same. A linear coordinate transformation which
has this property is called a (homogeneous) Lorentz transformation.
5.1.2 Lorentz space

Let us introduce an ordered quadruple of real numbers, enumerated with the help
of upper indices Â 0 Ý 1 Ý 2 Ý 3, where the zeroth component is Æ Ù (Æ is the speed
Ë
of light and Ù is time),
¡
and the remaining components are the components of the
ordinary 3 radius vector x defined in Equation (M.1) on page 166:
. ¢ Â Ü. 0
Ý
. 1
Ý
. 2
Ý
. 3
Þ Â Ü Æ Ù_Ý . Ý P ÝD Þ 3 Ü Æ Ù_Ý x Þ (5.6)
We then interpret this quadruple . as (the th component of) a radius four-vector ¢

Ë
in a real, linear, four-dimensional vector space.
Metric tensor
We want our space to be a Riemannian space, i.e., a space where a distance and a
scalar product are defined. We therefore need to define in this space a metric tensor,
also known as the fundamental tensor, which we shall denote ð and choose as ¢£
(in matrix notation):
0 1 0 0
§ ¨¨
Üð ¢£
Þ Â ¤¥¥¦
0
0
0
0
Ø 1
0
0
Ø 1
© (5.7)
Ø 1 0 0 0
i.e., with a main diagonal with sign sequence, or signature, ß Ý Ø Ý Ø Ý Ø . 1
Radius four-vector in contravariant and covariant form

¢
The radius four-vector . Â Ü . 0 Ý . 1 Ý . 2 Ý . 3 Þ Â Ü Æ Ù_Ý x Þ , as defined in Equation (5.6)
on the previous page, is, by definition, the prototype of a contravariant vector (or,
more accurately, a vector in contravariant component form). The corresponding
¢
covariant vector . is obtained as (the upper index in . is summed over and is ¢
therefore a dummy index and may be replaced by another Ë dummy index ): ª
.
¢ Â ð
.
¢ £ «£ (5.8)
This process is an example of contraction and is often called the “lowering” of
index.
Index lowering of the contravariant radius four-vector . amounts to multiplying ¢
¢
the column vector representation of . from the left by the matrix representation,
Equation (5.7), of ð ¢£
to obtain the column vector representation of . . The simple ¢
¢£
diagonal form of ð , Equation (5.7) above, means that the index lowering opera-
§ ¨¨
tion in our chosen flat 4D space is nearly trivial:
.
0 1 0 0 0 . 0 . 0§ ¨¨ § ¨¨ § ¨¨
. . . 1
¤¥¥¦ © ¤¥¥¦ © ¥¦¥¤ © ¤¥¥¦ ©
0 Ø 1 0 0 1
1 Ø
. Â 0 0 Ø 1 0 . 2 Â . 2 (5.9)
2 Ø
. 0 0 0 Ø 1 . 3 . 3
3 Ø
which we can describe as
¢.
Â ð ¢ £ «£
. Æ
Â Ü ÙYÝ Ø x Þ (5.10)
¢
i.e., the covariant radius four-vector
.
.
¢
is obtained from the contravariant radius
four-vector simply by changing the sign of the last three components. These
1
Without changing the physics, one can alternatively choose a signature . The latter ¬~¯® ° ® ° ® ° ±
has the advantage that the transition from 3D to 4D becomes smooth, while it will introduce some
annoying minus signs in the theory. In current physics literature, the signature
to be the most commonly used one.
seems ¬ ° ®I¯®I²®I³±
components are referred to as the space components; the zeroth component is re-
ferred to as the time component.
Scalar product and norm

¢
Taking the scalar product of . with itself, we get, by definition,
¢ £ U£ ¢ ¢ ¢
ð . :. Â . . Â Ü . 0 Ý . 1 Ý . 2 Ý . 3 Þ Õ Ü . 0 Ý . 1 Ý . 2 Ý . 3 Þ Â Ü Æ ÙYÝ Ø x Þ Õ Ü Æ ÙYÝ x Þ
Æ . Õ Æ .
P D
Â Ü ÙYÝ Ý àÝ Þ Ü ÙYÝ Ø Ý Ø Ý Ø Þ Â Æ 2 2 . 2
Ù Ø Ø
2

Ø
2
(5.11)
which acts as a norm or distance in our 4D space. We see, by comparing Equa-
tion (5.11) and Equation (5.5) on page 55, that this norm is conserved (invariant)
during a Lorentz transformation. We notice further from Equation (5.11) that our
space has an indefinite norm which means that we deal with a non-Euclidean space.
´
The four-dimensional space (or space-time) with these properties is called Lorentz
space and is denoted 4 . The corresponding real, linear 4D space with a positive
¡
definite norm which is conserved during ordinary rotations is a Euclidean vector
space which we denote 4 .
´
The 4 metric tensor Equation (5.7) on the preceding page has a number of
interesting properties: Firstly, we see that this tensor has a trace Tr Ü ð Þ Â Ø 2 ¢£
¡
whereas in 4 , as in any vector space with definite norm, the trace equals the space
dimensionality. Secondly, we find, after trivial algebra, that the following relations
between the contravariant, covariant and mixed forms of the metric tensor hold:
ð ¢¢ ££ Â ð £ ¢ (5.12a)
Â ð ¢ £
ð
ð £ µ ð µ¶¢ Â ð £¢ Â
% £¢ (5.12b)
ð £ µ ð µ¶¢ Â ð ¢ £ Â
% ¢£ (5.12c)
% ¢
(5.12d)
Here we have introduced the 4D version of the Kronecker delta £ , a mixed four-
tensor of rank 2 which fulfils
% £¢ % ¢ £ !
Â Â
0 if Â
¸· ªª (5.13)
1 if Ë Â
Ë
Invariant line element and proper time
between the two points . and . ß ¢ ¢
. in 4 can be ¢ ´
The differential distance
X¹ X
¢ £ X «£ X ¢ X ¢ X ¢ X
calculated from the Riemannian metric, given by the quadratic differential form
. . . . . 0Þ 2 Ü . 1Þ 2 Ü . 2Þ 2 Ü . 3Þ 2
X¹ Â ð Â Ü X X X
2
Â Ø Ø Ø
X X X X
Æ 2 2 . 2 2 2
Â Ù Ø Ø Ø (5.14)
where the metric tensor is as in Equation (5.7) on the facing page. The square root
of this expression is the invariant line element

Æ 1 YáX. 2
á X Ù â
2
X Ù â ^
á
2
X¹ Â X Ù 1Ø Æ 2
XÙâ
ß
X
ß
X
Æ
Æ X Ù X Ù
2
Æ É 1 Æ 1 Æ À
XÙÂ X
2 (5.15)
Â Ø 2 Â Ø Â
where we introduced
X Ù$
À
X Â (5.16)
Since À measures the time when no spatial changes are present, it is called the
proper time.
X
Expressing Equation (5.5) on page 55 in terms of the differential interval and
comparing with Equation (5.14) on the preceding page, we find that
X¹
X X
Æ .
X¹ XÙ Ø X
2 2 2 2 2 2
Â Ø Ø (5.17)
is invariant during a Lorentz transformation. Conversely, we may say that every co-
ordinate transformation which preserve this differential interval is a Lorentz trans-
formation.
If in some inertial system
º
X X
. Æ
X ß ß
XÙ
2 2 2 2 2
(5.18)
X¹ is a time-like interval, but if

»
X X
. Æ
X ß ß
XÙ
2 2 2 2 2
(5.19)
X¹ is a space-like interval, whereas
X X
. Æ
X ß ß
XÙ
2 2 2 2 2
Â (5.20)
is a light-like interval; we may also say that in this case we are on the light cone.
A vector which has a light-like interval is called a null vector. The time-like,
space-like or light-like aspects of an interval
formation.
is invariant under a Lorentz trans-
X¹
Four-vector fields
Any quantity that relative to any coordinate system has a quadruple of real numbers
and which transform in the same way as the radius four-vector . , is called a ¢
¼¢ ¼ ´
four-vector. I analogy with the notation for the radius four-vector we introduce
Â Ü Ý a Þ for a general contravariant four-vector field in
the notation 0 4 and
find that the “lowering of index” rule, Equation (M.20) on page 169, for such an
arbitrary four-vector yields the dual covariant four-vector field
¼¢ µ ¢£¼£ µ ¼ µ
Ü. Þ Â ð Ü . Þ Â Ü 0 Ü . Þ Ý Ø a Ü . Þ!Þ µ (5.21)
The scalar product between this four-vector field and another one Ü . Þ is ½¢ µ
¢£¼ £ µ ½¢ µ ¼
ð ½
Ü . Þ Ü . Þ Â Ü 0 Ý Ø aÞ Õ Ü 0 Ý bÞ Â 0 0 Õ
Ø a b ¼ ½ (5.22)
µ
.
µ
which is a scalar field, i.e., an invariant scalar quantity Ü Þ which depends on
time and space, as described by . Â Ü Æ Ù_Ý . Ý Ý Þ . P D
The Lorentz transformation matrix
Introducing the transformation matrix
§ ¨¨
¾Ø ³ 0 0
¢£
ÜΛ Þ Â ¥¤¦¥ 0 0
Ø ³
¾ 0 0
© (5.23)
1 0
0 0 0 1
¢L ¢ ¢
the linear Lorentz transformation (5.3) on page 55, i.e., the coordinate transforma-
tion . .
Â
. Ü . 0 . 1 . 2 . 3 Þ , from one inertial system Σ to another inertial
Ý Ý Ý
system Σ , can be written
.¢ ¢ £ «£
Â Λ
. (5.24)
The inverse transform then takes the form
U£ .
Â ÜΛ Þ æ
1. ¢£ ¢ (5.25)
The Lorentz group

It is easy to show, by means of direct algebra, that two successive Lorentz trans-
formations of the type in Equation (5.25) above, and defined by the speed para-

meters 1 and 2 , respectively, correspond to a single transformation with speed
parameter
FÂ 1 ß ß
1 2
1 2
(5.26)
This means that the nonempty set of Lorentz transformations constitute a closed
algebraic structure with a binary operation which is associative. Furthermore,
one can show that this set possesses at least one identity element and at least one
inverse element. In other words, this set of Lorentz transformations constitute a
mathematical group. However tempting, we shall not make any further use of
group theory.
À
À Á 0
0
Â
¿Á 1
Â
¿ 1
ÃiÄ Å Æ ÇÉÈÊPËKÌ ÃiÄMÍ Å Æ Í ÇÉÈÊIËKÌ

Figure 5.2. Minkowski space can be considered an ordinary Euclidean space
ÎÒ Ê Ë
where a Lorentz transformation from 1 0 to 1 0
cor-
Ð
iÏ Ä ÏÑ Æ Ð Ç
responds to an ordinary rotation through an angle . This rotation leaves the
Euclidean distance 1
2 0 2 2 2 2
invariant. Ä
5.1.3 Minkowski space
ÓÕÔ×Ö Ø Ó Ù Ó Ù Ó Ù Ó Ú
ÛÜÖÝÓ Þ ß
Specifying a point 0 1 2 3 in 4D space-time is a way of saying
Ø Ó ÙPà«ÙDá Ú Ö Ø Ó Ù Ó Ù Ó Ú
that “something takes place at a certain time 0 and at a certain place
1 2 3 .” Such a point is therefore called an event. The traject-
Ó ÖÓ
ory for an event as a function of time and space is called a world line. For instance,
the world line for a light ray which propagates in vacuum is the trajectory 0 1.
â äÖ ã}Ó Öã ß Û
If we introduce
åæÜç ã å è
0 0
(5.27a)
ç ê ë 1, we can rewrite Equation (5.14) on page 57 as

(5.27b)
where é
åæ ç ì å â í î ì åðï í î ì åðï í î ì åðï í
2 0 2 1 2 2 2 3 2
(5.28)
ñ
i.e., as a 4D differential form which is positive definite just as is ordinary 3D Euc-
lidean space 3 . We shall call the 4D Euclidean space constructed in this way the
Minkowski space 4 . ò
ó ï
As before, it suffices to consider the simplified case where the relative motion
between Σ and Σ is along the axes. Then
åæ ç ì å â í î ì åðï í
2 0 2 1 2
â ï
(5.29)
and we consider and 0 1
â ï plane
as orthogonal axes in an Euclidean space. As in all
Euclidean spaces, every interval is invariant under a rotation of the 0 1
ü
ô õ9öø÷
0
ôù 0 ô õ ô
0 1
û
ýù ôù 1
ú õ úù û öø÷
ý ô õ ô
1
þ
Figure 5.3. Minkowski diagram depicting geometrically the transformation
ÿ ÿ ÿ
(5.31) from the unprimed system to the primed system. Here denotes the
ÿ
world line for an event and the line 0 1

the world line for a light

ray in vacuum. Note that the event is simultaneous with all points on the 1

ÿ
axis ( 0), including the origin while the event , which is also simultan-

eous with all points on the axis, including , to an observer at rest in the

primed system, is not simultaneous with in the unprimed system but occurs
there at time .
â óïó:
0 1
â ó ç ë ï sin î â cos
through an angle into
ï ó çäï cos î â sin

0 1 0
(5.30a)
1 1 0
(5.30b)
See Figure 5.2.
If we introduce the angle
ç ë é , often called the rapidity or the Lorentz boost
parameter, and transform back to the original space and time variables by using
ó ç ë ï
Equation (5.27b) on the preceding page backwards, we obtain
ï ó çäï ë î cosh
sinh
sinh (5.31a)
cosh (5.31b)
ç
which are identical to the transformation equations (5.3) on page 54 if we let
sinh ç
(5.32a)
cosh
tanh
ç (5.32b)
(5.32c)
It is therefore possible to envisage the Lorentz transformation as an “ordinary”

rotation in the 4D Euclidean space 4 . Equation (5.26) on page 59 for successive
Lorentz transformation then corresponds to the tanh addition formula
1tanh tanh
tanh
1 2
tanh 1
2
tanh 1 2
(5.33)
The use of "!# and , which leads to the interpretation of the Lorentz trans-
4
formation as an “ordinary” rotation, may, at best, be illustrative, but is not very

physical. Besides, if we leave the flat $ space and enter the curved space of
4
general relativity, the “"!# ” trick will turn out to be an impasse. Let us therefore
immediately return to $ where all components are real valued.
4
5.2 Covariant classical mechanics

The measure %& of the differential “distance” in $ 4 allows us to define the four-
velocity
( , 89
' ( % )
% & +* 1 - /! . 01 2 1 5 - 2
u 1 u
(5.34)
3 46 ! 1 7 3 46 2
2
2
2
which, when multiplied with the scalar invariant : ! yields the four-momentum 0
2
( 89
;<( : ! % ) : ! : !
% & +* : ! !=- u 01 2 1 3546 - 2 1 3746
u 2
2 0 0
0 0 (5.35)
2 2
2 2
From this we see that we can write
p : u (5.36)
where
: >* : 2 : 0
(5.37)
3746
0
2
1 2
i.e., that Lorentz covariance implies that the mass-like term in the ordinary 3D
linear momentum is not invariant.
5.3. C OVARIANT CLASSICAL ELECTRODYNAMICS 63
The zeroth (time) component of the four-momentum ;( is given by
; >* : ! 2 : ! :?!
0 2 0
2
2
(5.38)
1 374 6
0
2
2
We interpret this as the energy @ , i.e., we can write

;( @A-B! p (5.39)
Scalar multiplying this four-momentum with itself, we obtain

; ( ;( ð ( C ; C ;<( ; 3
0 2
; 3 ; 3 ;
1 2 2 2 3 2
@A-D3E! p F @A-B! p @ 3G! H p H : !

2 2 2
0
2 2
(5.40)
Since this is an invariant, this equation holds in any inertial frame, particularly in

the frame where p 0 where we thus have
@ : ! 0
2
(5.41)
which is probably the most well-known physics formula ever.
5.3 Covariant classical electrodynamics

In the rest inertial system the charge density is I 0. The four-vector (in contravariant
component form)
J ( I % ) ( , I -BI v
%&
0
K! . (5.42)
where we introduced
I >* I 0 (5.43)
is called the four-current.

As is shown in Example M.4 on page 176, the d’Alembert operator is the scalar
product of the four-del with itself:
L NM ( M ( NM ( M ( 1 M 3 O 2
! M#
2 2
2 2
(5.44)
Since it has the characteristics of a four-scalar, the d’Alembert operator is invariant

and, hence, the homogeneous wave equation is Lorentz covariant.
5.3.1 The four-potential

If we introduce the four-potential
P ( QR-B! A (5.45)
where Q is the scalar potential and A the vector potential, defined in Section 3.3
on page 32, we can write the inhomogeneous wave equations (Lorentz equations)
Equation (3.14) on page 34 in the following compact (and covariant) way:
L P ( JS (
2
(5.46)
0
With the help of the above, we can formulate our electrodynamic equations co-
variantly. For instance, the covariant form of the equation of continuity, Equa-
tion (1.21) on page 9 is
J U
( T M J(
M( M)( 0 (5.47)
and the Lorentz gauge condition, Equation (3.13) on page 34, can be written
P ( T M P(
M( M)( 0 (5.48)
The gauge transformations (3.15) on page 35 in covariant form are simply
PWV ( P ( 3 M ( / ! Γ ) C X P ( 3 M ! Γ ) C
M)( (5.49)
If only one dimension Lorentz contracts (for instance, due to relative motion
)
along the direction), a 3D spatial volume transforms according to
%Y % ) *1 %Y %Y Z 1 \3 [ %Y ] 1 _

3 ^!
2
3 2
0 0 0 2
(5.50)
then from Equation (5.43) on the preceding page we see that
I`%Y I %Y 0 0 (5.51)
i.e., the charge in a given volume is conserved. We can therfore conclude that the
electron charge is a universal constant.
5.3.2 The Liénard-Wiechert potentials

Let us now solve the Lorentz equation (the inhomogeneous wave equation) (3.14)
V
defined by the radius four-vector ) V ( ) V !Db V - ( ) V - ) V - )
on page 34 in vacuum for the case of a well-localised charge at a source point
0 1 2 3 . The field (obser- V a
vation) point is denoted by the radius four-vector 0 ) )
1 2 3 . !Dbc- ) - ) - )
V
In the rest system we know that the solution is simply
P ( d 4 e a S H x 3 1 xV H - 0f
0 (5.52)
V
where H x 3 x H is the usual distance from the source point to the field point, eval-
0 0
0
uated in the rest system (signified by the index “0”).
Let us introduce the relative radius four-vector between the source point and the
g
field point:
( ) ( 3 ) V ( ! bR3Gb V - x 3 xV (5.53)
Scalarg multiplying
g this relative four-vector with itself, we obtain
( ( ! bh3ib V - x 3 xV F ! bD3jb V -D3 x 3 xV B ! bh3ib V 3lk x 3 xV k (5.54)
V kV k
2 2 2
We know that in vacuum the signal (field) from the charge a at ) ( propagates
to ) ( with the speed of light ! so that
kk x 3 xV kk ! bm3Gb V (5.55)
g this
Inserting g into Equation (5.54), we see that
( ( 0 (5.56)
or thatg Equation (5.53) above can be written
( k x 3 xV k - x 3 xV
k k (5.57)
Now we want to find the correspondence to the rest system solution, Equa-
tion (5.52), in an arbitrary inertial system. We note from Equation (5.34) on
89
page 62 that in the rest system
'( 01 2 1
- 2 u 1 - 0 (5.58)
1 3N4 6 ! 1 7 3 46 n o p o q
0 2 2
2 2
4 u 0 0
and g
( kk x 3 xV kk - x 3 gxV kk x 3 xV kk - x 3 xV (5.59)
Like all scalar products, ' ( ( is invariant, so we can evaluate it in any inertial
0 0 0 0
g g g
system. If we evaluate it in the rest system the result is:
'( ( '( ( '( (
1 - 0 F kk x 3 xV kk - x 3 xV kk x 3 xV kk
0 0 0
0 0 0
(5.60)
We therefore see that the expression
P ( a VS ' ' g (
4e 0
C Cg g (5.61)
subject to the condition ( ( 0 has the proper transformation properties (proper

tensor form) and reduces, in the rest system, to the solution Equation (5.52) on the
preceding page. It is therefore the correct solution, valid in any inertial system.
According to Equation (5.34) on page 62 and Equation (5.57) on the preceding
page
g ,
' C C r* 1 - u!/. F kk x 3 xV kk -D3 x 3 xV s
V V
x3 x F u
r* d kk x 3 x kk 3 ! f (5.62)
Introducing
V
& kk x 3 xV kk 3 x 3 !x F u (5.63)
g
we can write
' C C >* & (5.64)
and
' g(
' C C d 1& - ! u& f (5.65)
from which we see that the solution (5.61) can be written
P ( ) t a VS d 1 - u f QR-B! A
4e & !& (5.66)
0
where in the last step the definition of the four-potential, Equation (5.45) on page 64,
was used. Writing the solution in the ordinary 3D-way, we conclude that for a very
localised charge volume, moving relative an observer with a velocity u, the scalar
V 1
and vector potentials are given by the expressions
a a V
Q b- x 4e S & 4e S H x 3 xV H 3 n u 6 vwqyx
1
(5.67a)
Vu 1 Vu
x x u
0 0
a a
A b- x
4e S ! & 4e S ! H x 3 xV H 3 n u 6 v qzx
1
2 2 x x u
(5.67b)
0 0
These potentials are called the Liénard-Wiechert potentials.
5.3.3 The electromagnetic field tensor

Consider a vectorial (cross) product c between two ordinary vectors a and b:
c a{ b7 | } ~ch } ~ ˆ
3 ˆ
2 3 3 2 1
3 ˆ
3 1 1 3 2
3 ˆ
1 2 2 1 3 (5.68)
We notice that the th component of the vector c can be represented as
! + } ~ 3 ~ } ! } ~ 3E! ~ } - - J (5.69)
In other words, the pseudovector c a { b can be considered as an antisymmetric
tensor of rank two!
The same is true for the curl operator 7{ . For instance, the Maxwell equation
{ E 3 M M Bb (5.70)
can in this tensor notation be written
M @ }~ 3 M @ ~ } 3 M } ~
M) M) Mb (5.71)
We know from Chapter 3 that the fields can be derived from the electromagnetic
potentials in the following way:
B { A (5.72a)
E 3jQ3
M P
Mb (5.72b)
P~ MP} P
In component form, this can be written
} ~ M ) } 3 M ) P ~ NM } ~ 3 M ~ P }
M (5.73a)
@ } 3 M M ) Q } 3 M M b } 3 M } Q3 M P } (5.73b)

From this, we notice the clear difference between the axial vector (pseudovector)
B and the polar vector (“ordinary vector”) E.
Our goal is to express the electric and magnetic fields in a tensor form where
the components are functions of the covariant form of the four-potential, Equa-
tion (5.45) on page 64:
P ( QR-D3E! A (5.74)
Inspection of (5.74) and Equation (5.73) on the previous page makes it natural to
(C MP C 3 MP ( N
define the covariant four-tensor
M ( P C 3 MCP (
(
M) M) C (5.75)
This anti-symmetric (skew-symmetric), covariant four-tensor of rank 2 is called the

electromagnetic field tensor. In matrix representation, the covariant field tensor
@ @ @W 89
can be written
( C NM ( P C 3 M C P ( 0 3E@ 0 E3 ! !

0
1 E3 @ ! 0 3
! (5.76)
3W@E 3
! ! 0
3W@ 3W@ 3W@ 89
The matrix representation for the contravariant field tensor is
( C NM ( P C 3 M C P ( 0 @ 0 3
! !
0
1 @ ! 0 3E! (5.77)
@E 3E! ! 0
curl of the four-potential vector ( .
P
It is perhaps interesting to note that the field tensor is a sort of four-dimensional
One can show that the two Maxwell source equations
F E SI (5.78)
{ B d j S M M Eb f d I u S M M Eb f
0
0 0 0 0 (5.79)
M C C ( JS (
correspond to
M) 0
(5.80)
and that the two Maxwell “field” equations
{ E 3 M M Bb (5.81)
F B 0 (5.82)
correspond to
68
M ( C M C( t M tC (
M)t M) M) 0 (5.83)
Hence, Equation (5.80) on the facing page and Equation (5.83) above constitute
Maxwell’s equations in four-dimensional formalism.
B IBLIOGRAPHY 5
[1] J. Aharoni. The Special Theory of Relativity. Dover Publications, Inc., New York,
second, revised edition, 1985. ISBN 0-486-64870-2.
[2] Asim O. Barut. Dynamics and Classical Theory of Fields and Particles. Dover Pub-
lications, Inc., New York, NY, 1980. ISBN 0-486-64038-8.
[3] Walter T. Grandy. Introduction to Electrodynamics and Radiation. Academic Press,
New York and London, 1970. ISBN 0-12-295250-2.
[5] C. Møller. The Theory of Relativity. Oxford University Press, Glasgow . . . , second
edition, 1972.
ISBN 0-201-05702-6.
[7] J. J. Sakurai. Advanced Quantum Mechanics. Addison-Wesley Publishing Com-
pany, Inc., Reading, MA . . . , 1967. ISBN 0-201-06710-2.
[8] Barry Spain. Tensor Calculus. Oliver and Boyd, Ltd., Edinburgh and London, third
edition, 1965. ISBN 05-001331-9.
71
72
C HAPTER 6
Interactions of Fields
and Particles
In this Chapter we study the interaction between electric and magnetic fields and
electrically charged particles. The analysis is based Lagrangian and Hamiltonian
methods, is fully covariant, and yields results which are relativistically correct.
6.1 Charged Particles in an Electromagnetic Field

We first establish a relativistic correct theory describing the motion of charged
particles in prescribed electric and magnetic fields. From these equations we may
then calculate the charged particle dynamics in the most general case.
6.1.1 Covariant equations of motion

We can obtain an equation of motion if we find a Lagrange function in 4D for our
problem and then apply a variational principle or if we find a Hamiltonian in 4D
and solve the corresponding Hamilton’s equations. We shall do both.
Lagrange formalism
Call the 4D Lagrange function nq and assume that it fulfils the variational prin-

4
ciple
n ( '(
q) - % &
1
4 0 (6.1)
0
%&
where is the invariant line element given by Equation (5.15) on page 58, and the
nq
endpoints are fixed.
We must require that 4 fulfils the following conditions:
73
74 C HAPTER 6. I NTERACTIONS OF F IELDS AND PARTICLES
1. The Lagrange function must be invariant. This implies that nq

4 must be a
scalar.
n q must not con-

2. The Lagrange function must be linear. This implies that
tain higher than the second power of the four-velocity . '( 4
According to Formula (M.76) on page 180 the ordinary 3D Lagrangian is the
: ^
difference between the kinetic and potential energies. A free particle has only
:
kinetic energy. If the particle mass is 0 then in 3D the kinetic energy is 0 2 2.
This suggests that in 4D the Lagrangian for a free particle should be
n q :?2! ' (' (

2
free
4 (6.2)
For an interaction with the electromagnetic field we can introduce the interaction
with the help of the four-potential given by Equation (5.74) on page 68 in the
following way
n q :?2! ' (' ( ¡ ' ( P ( ) t

2
4 (6.3)
We call this the four-Lagrangian.

The variation principle (6.1) with the 4D Lagrangian (6.3) inserted, leads to
n ( '( : ! '(`' '( P
q ) - % & d 2 ( ¡ ( f % &
1 1 2
0
4
0
: ! M '(' '( P
0
M P( C

' (
'
2 M ' ( ( ¡ d ( M ) C ) f £ % &
1
0
2
(
¢ ' '(
0
P
' (
' ( M P( C
:?! ( ¤ d ( M ) C ) f £ % & 0
1
2
(6.4)
0 ¢
According to Equation (5.34) on page 62, the four-velocity is
'( % ) (
%& (6.5)
which means that we can write the variation of ' (

to & :
as a total derivative with respect
'( d % ) ( f % ) (
%& %& (6.6)
Inserting this into the first two terms in the last integral in Equation (6.4), we obtain
6.1. C HARGED PARTICLES IN AN E LECTROMAGNETIC F IELD 75

n ( '(
q ) - % &
1
4

d :?! ' ( % % & ) ( ¤ P ( % % & ) ( ¡ '( M P C( ) C f % &
0
M)
2
(6.7)
0

Partial integration in the two first terms in the right hand member in (6.7) gives
n ( '(
q ) - % &
1
4
0
% ' ( ( % P ( ( '( M P C( ) C f % &
d 3E: ! % & ) 3 % & ) ¡
1
M)
2
0 (6.8)
0
where the integrated parts do not contribute since the variations at the endpoints

vanishes. Change of irrelevant summation index from to in the first two terms ¥

in the right hand member in (6.8) yields, after breaking out a common factor
n ( '(
q ) - % &
1
4
0
% ' C % P C '( PM ( C
d 3E: ! % & 3 % & ¡ M)Cf ) %&
1
2
0 (6.9)
0
%P C%&
Applying well-known rules of differentiation and the expression (5.34) for the
four-velocity, we can reform the expression for as follows:
% P C M P C % ) ( M P C '(
%& M)( %& M)( (6.10)
'(
By inserting this expression (6.10) into the right-hand member of Equation (6.9)
above, and moving out a common factor , we obtain the final variational prin-

ciple expression
n ( '(
q ) - % &
1
4
0
% ' C '( M P ( M P C
3E: ! % & d M ) C 3 M ) ( f £ ) C % &
1
2
(6.11)
¢
0
0
)C
)( )(
Since, according to the variational principle, this expression shall vanish and
is arbitrary along the world-line from 0 tp 1 , the expression inside in the ¦X
integrand in the right hand member of (6.11) must vanish. In other words, we have
found an equation of motion:
% ' C '( M P ( M P C
: ! % & + d M ) C 3 M ) ( f
0
2
(6.12)
With the help of Equation (5.77) on page 68 we can express this equation in terms
of the electromagnetic field tensor in the following way:
'
: ! % % & C + '( C (
0
2
(6.13)
This is the final equation of motion for a particle in an electromagnetic field. It is

often referred to as the Minkowski equation.
Hamiltonian formalism
The usual Hamilton equations for a 3D space are given by Equation (M.81) on
page 181 in Chapter M. These six first-order partial differential equations are
M;§ % }
M } %¨b ; (6.14a)
M§ 3 % }
M } %¨b (6.14b)
where § ; } - } -Bb ; ˙ 3¡ } - ˙} -Bb is the ordinary 3D Hamiltonian, } is a gen-

eralised coordinate and ; } is its canonically conjugate momentum.
velocity ' ( , as given by Equation (5.34) on page 62, and the four-momentum ; ( ,
We seek a similar set of equations in 4D space. To this end we utilise the four-
as given by Equation (5.35) on page 62, to introduce the four-Hamiltonian
§ n q ;<(' ( 3 n q
4 4 (6.15)
where ; ( is considered as the canonically conjugate four-momentum
;<( M ' n q
4
M ( (6.16)
where n q is as in Equation (6.3) on page 74. With the help of these, the radius
four-vector ) ( , considered as the generalise four-coordinate, and the invariant line
4
element % & , defined in Equation (5.15) on page 58, we introduce the following
eight partial differential equations:
M§ ; n q % ) (
M n( % & ;
4
(6.17a)
M§ q 3 % (
M)( %&
4
(6.17b)
which form the four-dimensional Hamilton equations.

;(
We note that by solving (6.17) we obtain which, according to Equation (5.39)
;
on page 63, has a zeroth (time) component which we identify with the total energy.
Hence we must require that this component 0 solves the ordinary 3D Hamilton
§ nq
equations (6.14).
n q
Using to the definition of 4 , Equation (6.15) on the preceding page, and the
expression for 4 , Equation (6.3) on page 74, we obtain
§ n q ;<(`' ( 3 n q ; (' ( 3 : 2! ' (' ( 3 '( P ( ) C

2
0
4 4 (6.18)
Furthermore, from the definition (6.16) of the conjugate four-momentum ; ( , we

see that
;( M ' n q M' d : ! ' (' ( ¡ '( P ( ) C f

4 0
2
M ( ' ( M P ( ( 2
: ! ¡0
2
(6.19)
Inserting this into (6.18), we obtain
§ n q : ! ' (' ( P ( ' ( 3 : 2! ' (' ( 3 '( P ( ) C

2
2 0
4 0
12 : ! '(' (
0
2
(6.20)
Since the four-velocity scalar-multiplied by itself is ' ( ' ( 1, we clearly see

from Equation (6.20) above that § n q is a scalar invariant, whose value is simply
4
§ n q 12 : !
4 0
2
(6.21)
However, at the same time (6.19) provides the algebraic relationship
'( 1 <; ( 3 P (
: ! 2
(6.22)
' ( , one gets

0
and if this is used in (6.20) to eliminate

§ n q : 2! d : 1! ;<( 3 P ( : 1! ; ( 3 P ( f
2
0
4 2 2
P ( ; ( 3 P (
0 0
2: 1 ! 2
;<( 3
© ;<(; ( 3 2 P (h; ( ¡ P ( P (`ª
0
2: 1 ! 0
2
2
(6.23)
That this four-Hamiltonian yields the correct covariant equation of motion can be
seen by inserting it into the four-dimensional Hamilton’s equations (6.17) and using
M§ n q 3 ; C 3 P C M P C
the relation (6.22):
M)( : ! M)(
4
P 2
3 : ! : ! ' C MM ) ( C
0
2
PC 2 0
' C M
0
3
M)(
; '
3 % % & ( 3
:?! % % & ( 3 M P C( ' C
M)
2
(6.24)
where in the last step Equation (6.19) on the preceding page was used. Rearranging
'( 'C MP C MP(

terms, and using Equation (5.77) on page 68, we obtain
: ! % & + d M ) ( 3 M ) C f + ' C ( C

%
0
2
(6.25)
which is identical to the covariant equation of motion Equation (6.13) on page 76.
§ nq : !
We can then safely conclude that the Hamiltonian in question is correct.
Setting expression (6.23) above for 4 equal to the scalar value 0 2 2, as
derived above, and using the fact that
;<( ; B- ! p
P(
0
;<(; ; QR-B! A
P (; (( ; 3! p
0 2 2 2
P (P ( Q 3G! p F A
0 2
Q 3G! A
2 2 2
we obtain the equation

: ! 1 ; 3G! p 3 2 Q ; 2 ! p F A ¤ Q 3 ! A
2
2: !
0 0 2 2 2 0 2 2 2 2 2 2
2 2
0
(6.27)
which is a second order algebraic equation in ; 0
; ; 3 2 Q ; 3G; ! p « 2 ! p F A 3 ! A ¬ ¤ Q 3 : !
0 2 0 2 2 2 2 2 2 2 2 2
0
2 2
3 2 Q 3! p 3 2 p F A ¡ A ¤ Q G3 : !

0 2 0 2 2 2 2 2 2 2 4
; 3 2 Q ; 3 ! p 3 A Q 3G: ! 0
0
0 2 0 2 2 2 2 2 4
(6.28)
0
with the solution
; + Q®¡! 2 p 3 A : !
0 2 2 2
0 (6.29)
Since the fourth component (time component) 0 of the four-momentum is ; ;(

the total energy, the positive solution in (6.29) must be identified with the ordinary
§
Hamilton function . This means that

T ; 2
§ + Q ! p 3 A : !
0 2 2 2
0 (6.30)
is the ordinary 3D Hamilton function for a charged particle moving in scalar and
vector potentials associated with prescribed electric and magnetic fields.
§
nq § nq
The ordinary Lagrange and Hamilton functions and are related to each other
by the 3D transformation [cf. the 4D transformation (6.15) between 4 and 4 ]
pF u3 § (6.31)
Furthermore, we make the identification
p 3 A 2 : 0u
: u (6.32)
1 37¯ 6 2
2
Together with (6.30) and (6.31), gives the ordinary 3D Lagrange function
'
3 Q ¤ A F u 3G: ! ] 1 3 !
2
2
0 2
(6.33)
for a charged particle moving in scalar and vector potentials associated with pre-
scribed electric and magnetic fields.
±³²µ´ 1
±³² ±³² ¶ 1
¸ · ¸ · ¸ · ¸ · ¸ ¹
° ° ° °
º
»
Figure 6.1. A one-dimensional chain consisting of discrete, identical mass
¼
points , connected to their neighbours with identical, ideal springs with spring
½ and ¾ ¿ À ÁÃÂÅÄ h¾ ¿ ÃÁ ÂÅÄ h¾ ¿ Æ ÃÁ ÂÅÄ
constants . The equilibrium distance between the neighbouring mass points is
, , are the instantaneous deviations, along the axis, Ç
ÁyÈ É Ä È ÁyÈ¨Ê Ä
1 1
of positions of the 1 th, th, and 1 th mass point, respectively.
6.2 Covariant Field Theory

So far, we have considered two classes of problems. Either we have calculated the
fields from given, prescribed distributions of charges and currents, or we have de-
rived the equations of motion for charged particles in given, prescribed fields. Let
us now put the fields and the particles on an equal footing and present a theoretical
description which treats the fields, the particles, and their interactions in a unified
way. This involves transition to a field picture with an infinite number of degrees
of freedom. We shall first consider a simple mechanical problem whose solution
is well known. Then, drawing inferences from this model problem, we apply a
similar view on the electromagnetic problem.
6.2.1 Lagrange-Hamilton formalism for fields and interac-

tions
Consider Ë identical mass points, each with mass and connected to its neigh- :
bour along a one-dimensional straight line, which we choose to be the axis, by )

identical ideal springs with spring constants . At equilibrium the mass points are

at rest, distributed evenly with a distance to their two nearest neighbours. After

perturbation, the motion of mass point will be a one-dimensional oscillatory mo-

Ì }b
tion along ˆ . Let us denote the magnitude of the deviation for mass point from
its equilibrium position by ˆ.
The solution to this mechanical problem can be obtained if we can find a Lag-

rangian (Lagrange function) which satisfies the variational equation
6.2. C OVARIANT F IELD T HEORY 81

Ì } - Ì ˙} -Bb %¨b 0 (6.34)
According to Equation (M.76) on page 180, the Lagrangian is where >Í 3¡Y
Í Y
denotes the kinetic energy and the potential energy of a classical mechanical
system with conservative forces. In our case the Lagrangian is
Ï « : Ì ˙ 3 Ì 3GÌ ¬
1
} }Ð }
} Îo
2 2
1 (6.35)
2
1
Let us write the Lagrangian, as given by Equation (6.35), in the following way:
Ï Î `Ñ } (6.36)
}o 1
Here,
Ñ } 1 : Ì ˙} 3¡ d Ì } Ð 3GÌ } f Ó 2
Ò
1

2
(6.37)
2
is the so called linear Lagrange density. If we now let and, at the Ë Ô Õ

same time, let the springs become infinitesimally short according to the follow-
ing scheme:
ÔÖ% )
: Ô
(6.38a)
ÔØ×
(linear mass density) (6.38b)
(Young’s modulus) (6.38c)
Ì } Ð 3GÌ } Ô Ù Ú
Ù
1
(6.38d)
we obtain
Ñ %) (6.39)
where
Ó
Ñ d Ì- M M Ì b - MM )Ì B- b f Ò d M M Ì b f 3× d MM )Ì f
2 2
1
(6.40)
2
Notice how we made a transition from a discrete description, in which the mass
points were identified by a discrete integer variable 1 2 - -DÛDÛDÛ-sË
, to a continu-
ous description, where the infinitesimal mass points were instead identified by a
)
continuous real parameter , namely their position along ˆ .
A consequence of this transition is that the number of degrees of freedom for

the system went from the finite number Ë
to infinity! Another consequence is
that Ñ has now become dependent also on the partial derivative with respect to )
Ì
of the “field coordinate” . But, as we shall see, the transition is well worth the
price because it allows us to treat all fields, be it classical scalar or vectorial fields,
or wave functions, spinors and other fields that appear in quantum physics, on an
equal footing.

Under the assumption of time independence and fixed endpoints, the variation
principle (6.34) on the previous page yields:
Ü b
ÝÞÑ d Ì - M M Ì b - M M )Ì f % ) %¨b
M Ñ Ì
MÌ
M, Ñ d M M Ì b f M, Ñ Ù Ú d MM )Ì f ßà % ) %¨b
M ÙÙ Ú . M Ù.
0 (6.41)
ÝÞ 89
The last integral can be integrated by parts. This results in the expression
89
M<Ñ 3 M 01 M, Ñ M MÑ
M Ì M b M Ù Ú . 3 M ) 01 M , Ù Ú . ßà Ìá% ) %¨b 0 (6.42)
Ù Ù
where the variation is arbitrary (and the endpoints fixed). This means that the
itself must vanish. If we introduce the functional derivative
integrand
89
Ñ MÑ 3 M 01 M, Ñ
Ì M Ì M ) M Ù Ú (6.43)
Ù .
we can express this as
89
Ñ 3 M 01 M, Ñ
Ì M b M Ù Ú 0 (6.44)
Ù .
which is the one-dimensional Euler-Lagrange equation.
Inserting the linear mass point chain Lagrangian density, Equation (6.40) on the
preceding page, into Equation (6.44) above, we obtain the equation of motion for
our one-dimensional linear mechanical structure. It is:
M M b âÌ 3× M M ) Ì d × M M b 3 M M ) f Ì 0
2 2 2 2
2 2 2 2
(6.45)
Z × along the linear structure.

i.e., the one-dimensional wave equation for compression waves which propagate
with phase speed
ä
^ ã
this 3D case we get the variational principle
A generalisation of the above 1D results to a three-dimensional continuum is
straightforward. For
å Ü b ÝÞ Ñ % )WÜ b
3
Ñ d Ì - M M ) Ì ( f % )
4
89
M<Ñ 3 M ( 01 ,MÑ
M Ì M ) M Ù Ú . ßà Ìá% ) 0
4
(6.46)
Ù
where the variation Ì is arbitrary and the endpoints are fixed. This means that the
89
integrand itself must vanish:
MÑ 3 M ( 01 ,MÑ
M Ì M ) M ÙÚ . 0 (6.47)
Ù

This constitutes the three-dimensional Euler-Lagrange equations.

Introducing the three-dimensional functional derivative
Ñ MÑ 3 M } æ MÑ
Ì M Ì M ) M Ù Ú ç è (6.48)
Ù
we can express this as
89
Ñ 3 M 01 M<, Ñ
Ì M b M Ù Ú 0 (6.49)
Ù .
I analogy with particle mechanics (finite number of degrees of freedom), we may
introduce the canonically conjugate momentum density
e ) ( e bc- x M , Ñ
M ÙÙ Ú .
(6.50)
and define the Hamilton density

é d e-BÌ- M Ì ; b f e M Ì 3 Ñ d Ì- M Ì - M Ì f
M)} Mb Mb M)} (6.51)
If, as usual, we differentiate this expression and identify terms, we obtain the fol-
M é M Ì
lowing Hamilton density equations
Mée M b e (6.52a)
3 M
Ì Mb (6.52b)
The Hamilton density functions are in many ways similar to the ordinary Hamilton
functions and lead to similar results.
The electromagnetic field

Above, when we described the mechanical field, we used a scalar field x. Ì b-
If we want to describe the electromagnetic field in terms of a Lagrange density
Ñ P ( ) t
and Euler-Lagrange equations, it comes natural to express in terms of theÑ
four-potential .
The entire system of particles and fields consists of a mechanical part, a field
part and an interaction part. We therefore assume that the total Lagrange density
Ñ tot for this system can be expressed as
Ñ +Ñ
tot
Ñ Ñ
mech inter field
(6.53)
is given by n q Y where n q is given by Equation (6.2) on page 74 and Y is the

where the mechanical part has to do with the particle motion (kinetic energy). It
volume. Expressed in the rest mass density I , the mechanical Lagrange density
4 4
0
m
can be written
Ñ 12 I ! '(`' (
mech 0 2
m (6.54)
The Ñ inter
part which describes the interaction between the charged particles
and the external electromagnetic field. A convenient expression for this interaction
J(P (
Lagrange density is
Ñ inter
(6.55)
For the field part Ñ field
we choose the difference between magnetic and electric
energy density (in analogy with the difference between kinetic and potential energy
in a mechanical field). Using the field tensor, we express this field Lagrange density
as
Ñ field 1
4
0 S (C (C (6.56)
so that the total Lagrangian density can be written
Ñ tot
12 I ! ' (' ( J ( P ( 14 S ( C ( C
0 2
m 0 (6.57)
êF IELD ENERGY DIFFERENCE EXPRESSED IN THE FIELD TENSOR E XAMPLE 6.1

Show, by explicit calculation, that
ë
ìíDë ìîí
ï 2 zÁ ð ñ É ò Ä
2 2 2
(6.58)
From Formula (5.77) on page 68 we recall that
É ò ü û É
ò ý É ö
ò þ ÿ
ë
ìí ï ó =ì ôõí É ó hí ôöì ø
0
ñ ñ
ï ÷ùú òüòýû ð ñ 0 þ É ð 0 þ É ð ð ñ ý û (6.59)
òöþ É ð ñ ý ð ñ û 0
and from Formula (5.76) on page 68 that
ü
ò û ò ý ö
ò þ ÿ
ñ
ë ìíEïóîì ô í É äó í ô ì ï ùú÷ É òüû ñ 0 É ð þ ð ñ ý
0
ñ
É òý ð ñ þ ñ 0 É ð 0 û (6.60)
É òþ É ð ý ð û
where denotes the row number and the column number. Then, Einstein summation and
direct substitution yields
ë
ìíDë ìîíEï ë ë Ê ë ë Ê ë ë Ê ë ë
Ê ëë ëë Ê ëë ëë Ê ëë ëë Ê ë ë
00 01 02 03
00 01 02 03
ë ë
10 11 12 13
Ê ë ë Ê ë ë Ê ë ë Ê
10 11 12 13
ë ë
20 21 22 23
Ê Ê Ê Ê
20 21 22 23
30 31 32 33
ï 0É ò û É ò ý É ò þ
30 31 32 33
2 2 2
É ò û Ê 0Ê ð ñ þ Ê ð ñ ý
2 2 2 2 2
2
É ò ý Ê ð ñ þ Ê 0Ê ð ñ û
2 2 2 2
2
É òþÊ ðñýÊ ðñûÊ 0
2 2 2 2
ï É 2ò û É 2ò ý É 2ò þ Ê ð ñ û Ê ð
2 2 2 2 2 2 ñýÊ ðñþ
2 2 2
ï É 2ò Ê 2ð ñ ï 2 Ázð ñ É ò Ä
2 2 2 2 2 2
(6.61)
QED
E ND OF EXAMPLE 6.1
Ñ PC
Using tot in the 3D Euler-Lagrange equations, Equation (6.47) on page 83
Ì
(with replaced by ), we can derive the dynamics for the whole system. For
instance, the electromagnetic part of the Lagrangian density
Ñ EM
>Ñ inter
Ñ field
J C P C 14 S ( C ( C 0 (6.62)
inserted into the Euler-Lagrange equations, expression (6.47) on page 83, yields
two of Maxwell’s equations. To see this, we note from Equation (6.62) above and
the results in Example 6.1 that

M<Ñ P J C
EM
M C (6.63)
Furthermore,
S ,
M ( M <M MÑ ( P C £ 4 M ( M M (M P C t t . £
EM
0
¢S ¢ P
4 M (
M M ( P C M t 3 M P t M t P 3 M P t
0 M
S
4 M ( M M (M P C M t P M t P 3 M t P M P t
0
¢
3 M P t M t P ¤M P t M P t £
S , P P
2M ( M M ( C M t M t 3 M t P M P t . £
0 M P
¢ (6.64)
, P P
But
MP M t M N
M t P MP M P
M M ( C t . M M ( C t
NM t M M (M P C M t P
P
NM t P M M (M P C M t P
NM t P M M (M P C ð¨t M ð P
ð¨t ð M t P M M (M P C M P
NM P M M (M P C M P
2M ( P C (6.65)
Similarly,
MP , P P P
M t M 2M C (
M M ( C t. (6.66)
(C
so that

M Ñ P £
M ( M M ( C M ( M
EM
S ( P C 3 C P (
M M)( S M (6.67)
¢
0 0
the Lagrangian density Ñ EM P

This means that the Euler-Lagrange equations, expression (6.47) on page 83, for
and with C as the field quantity become
MÑ P 3 M ( MÑ P £ J C 3 S M ( ( C 0
EM EM
M C M¢ M ( C 0
M) (6.68)
M ( ( C JS C
or
M) 0
(6.69)
Explicitly, setting ¥ 0 in this covariant equation and using the matrix rep-

electromagnetic field tensor ( C , we obtain
resentation Formula (5.77) on page 68 for the covariant component form of the
M M M M 0 M @ M @ M @E
00 10 20 30
M)
0
M) 1
M) 2
M)
3
M ) I M M
F E S
0
(6.70)
which is the Maxwell source equation for the electric field, Equation (1.43a) on
page 14. For ¥
1 we get
M M M M 3 1 M @ 0 ! M ! M
01 11 21 31
M) 0
M) M)
1
M ) I '! M b
2 3
M M
S (6.71)
J
or, using S 1 ! and identifying I ' ,
0
2
0 0
MK 3 M 3 S M @ J
M M Mb 0 0 0 (6.72)
and similarly for ¥ 2 - 3. In summary, in three-vector form, we can write the

result as
{ B 3 S M M Eb j b- x
0 0 0 (6.73)
which is the Maxwell source equation for the magnetic field, Equation (1.43d) on
page 14.
Other fields
In general, the dynamic equations for most any fields, and not only electromagnetic
ones, can be derived from a Lagrangian density together with a variational principle
(the Euler-Lagrange equations). Both linear and non-linear fields are studied with
this technique. As a simple example, consider a real, scalar field which has the Ì
following Lagrange density:
Ñ 1
2 © M ( Ì M ( Ìâ3G: Ì ª 2 2 1
2
d M M ) Ì ( M M ) Ì ( 3G: Ì f
2 2
(6.74)
Insertion into the 1D Euler-Lagrange equation, Equation (6.44) on page 82, yields
the dynamic equation
L 3G: Ì
2 2
0 (6.75)
with the solution
u"#%$
Ì n x u"!yq H x H
x$
kx
(6.76)
which describes the Yukawa meson field for a scalar meson with mass : . With
e ! 1 MM Ìb2
(6.77)
we obtain the Hamilton density

é 1 ! e Ì : Ì
2 2 2 2 2
(6.78)
2
which is positive definite.
Another Lagrangian density which has attracted quite some interest is the Proca
Lagrangian
Ñ EM
>Ñ inter
Ñ field
J C P C 14 S ( C ( C : P ( P (
0
2
(6.79)
M ( ( C : P C JS C
which leads to the dynamic equation
M)
2
(6.80)
0
This equation describes an electromagnetic field with a mass, or, in other words,
massive photons. If massive photons would exist, large-scale magnetic fields, in-
cluding those of the earth and galactic spiral arms, would be significantly modified
to yield measurable discrepances from their usual form. Space experiments of this
kind onboard satellites have led to stringent upper bounds on the photon mass. If
the photon really has a mass, it will have an impact on electrodynamics as well as
on cosmology and astrophysics.
88
B IBLIOGRAPHY 6
[1] Asim O. Barut. Dynamics and Classical Theory of Fields and Particles. Dover Pub-
lications, Inc., New York, NY, 1980. ISBN 0-486-64038-8.
[2] Herbert Goldstein. Classical Mechanics. Addison-Wesley Publishing Company, Inc.,
Reading, MA . . . , second edition, 1981. ISBN 0-201-02918-9.
[3] Walter T. Grandy. Introduction to Electrodynamics and Radiation. Academic Press,
New York and London, 1970. ISBN 0-12-295250-2.
ISBN 0-201-05702-6.
[6] J. J. Sakurai. Advanced Quantum Mechanics. Addison-Wesley Publishing Com-
pany, Inc., Reading, MA . . . , 1967. ISBN 0-201-06710-2.
89
90
C HAPTER 7
Interactions of Fields
and Matter
The microscopic Maxwell equations (1.43) derived in Chapter 1 are valid on all
scales where a classical description is good. However, when macroscopic matter is
present, it is sometimes convenient to use the corresponding macroscopic Maxwell
equations (in a statistical sense) in which auxiliary, derived fields are introduced in
order to incorporate effects of macroscopic matter when this is immersed fully or
partially in an electromagnetic field.
7.1 Electric polarisation and the electric

displacement vector
7.1.1 Electric multipole moments
The electrostatic properties of a spatial volume containing electric charges and
located near a point x0 can be characterized in terms of the total charge or electric
monopole moment
&' ( ) * + 3/
x, - . , (7.1)
where the * is the charge density introduced in Equation (1.7) on page 4), the
electric dipole moment vector
' ( ) + * + 3/
p x,0 x0 - x, - . , (7.2)
'
with components 1 2 , 3 1 4 2 4 3, the electric quadrupole moment tensor
( ) +
Q' x,0 x0 - + x,50 x0 - * +
x, - .
3/
, (7.3)
'
with components 6 2 7 4384:9 1 4 2 4 3, and higher order electric moments.
91
92 C HAPTER 7. I NTERACTIONS OF F IELDS AND M ATTER
In particular, the electrostatic potential Equation (3.3) on page 31 from a charge

distribution located near x0 can be Taylor expanded in the following way:
& +
; + ' 1 1 x0 x0 - 2
stat x- @ @ A @ @ 21 2 @ @ A
4<>= 0 ? x 0 x0 x 0 x0 x 0 x0
1 3 + x 0 x0 -C2 + x 0 x0 - 7 1
@ @36 2 7 B @ @ @ @ 0 27 E AGFFFIH
x 0 x0 2 x 0 x0 x 0 x0 2D
(7.4)
where Einstein’s summation convention over 3 and 9 is implied. As can be seen

from this expression, only the first few terms are important if the field point (ob-
servation point) is far away from x0 .
For a normal medium, the major contributions to the electrostatic interactions
come from the net charge and the lowest order electric multipole moments induced
by the polarisation due to an applied electric field. Particularly important is the
dipole moment. Let P denote the electric dipole moment density (electric dipole
moment per unit volume; unit: C/m2 ), also known as the electric polarisation, in
some medium. In analogy with the second term in the expansion Equation (7.4) on
page 92, the electric potential from this volume distribution P + x, - of electric dipole
moments p at the source point x, can be written
; + ' 1 ( ) x 0 x,
p x- P + x, - J @ @ .
3/
,
4<>= 0 x 0 x, 3
' 1 ( ) 1
0 P + x, - JML B @ @E .
3/
,
4<K= 0 x 0 x,
' 1 ( ) 1
P + x, - JMLN, B @ @E .
3/
, (7.5)
4<>= 0 x 0 x,
Using the expression Equation (M.63) on page 178 and applying the divergence
theorem, we can rewrite this expression for the potential as follows:
; + ' 1 ( ) P + x, - 3/ ( ) L , JP + x, - 3/
p x- LN,J B @ @E . ,0 @ @ . ,QP
4<>= 0 O x 0 x, x 0 x,
' 1 P + x, - J n̂ 2/ ( ) L , JP + x, - 3/
@ @ . 0 @ @ . ,P
4<>= 0 O R S x 0 x, x 0 x,
(7.6)
where the first term, which describes the effects of the induced, non-cancelling
dipole moment on the surface of the volume, can be neglected, unless there is a
7.1. E LECTRIC POLARISATION AND THE ELECTRIC DISPLACEMENT VECTOR 93
discontinuity in P J T ˆ at the surface. Doing so, we find that the contribution from
the electric dipole moments to the potential is given by
; ' 1 ( ) 0UL , J P + x, - 3/
p @ @ . , (7.7)
4<K= 0 x0 x,
Comparing this expression with expression Equation (3.3) on page 31 for the elec-
trostatic potential from a static charge distribution * , we see that 0VLWJ P + x - has the
characteristics of a charge density and that, to the lowest order, the effective charge
density becomes * + / - 0XLYJ P + / - , in which the second term is a polarisation term.
The version of equation Equation (1.7) on page 4 where “true” and polarisation
charges are separated thus becomes
* +
' x- 0[LZJ P + x-
LZJ E (7.8)
= 0
Rewriting this equation, and at the same time introducing the electric displacement
vector (C/m2 )
'
D = 0E A P (7.9)
we obtain
+ ' 'G* +
LZJ = 0E A P- LZJ D true x- (7.10)
where * true is the “true” charge density in the medium. This is one of Max-
well’s equations and is valid also for time varying fields. By introducing the
notation * pol ' 0VL\J P for the “polarised” charge density in the medium, and
* 'G* *
total true A pol for the “total” charge density, we can write down the following
alternative version of Maxwell’s equation (7.23a) on page 95
* +
' total x-
LZJ E (7.11)
= 0
@ @
Often, for low enough field strengths E , the linear and isotopic relationship
between P and E
'
P = 0] E (7.12)
is a good approximation. The quantity ] is the electric susceptibility which is ma-

terial dependent. For electromagnetically anisotropic media such as a magnetised
plasma or a birefringent crystal, the susceptibility is a tensor. In general, the rela-
tionship is not of a simple linear form as in Equation (7.12) but non-linear terms
are important. In such a situation the principle of superposition is no longer valid
and non-linear effects such as frequency conversion and mixing can be expected.
Inserting the approximation (7.12) into Equation (7.9) on the previous page, we
can write the latter
'
D = E (7.13)
where, approximately,
' +
= = 0 1A ] - (7.14)
7.2 Magnetisation and the magnetising field

An analysis of the properties of stationary magnetic media and the associated cur-
rents shows that three such types of currents exist:
1. In analogy with “true” charges for the electric case, we may have “true”
currents jtrue , i.e., a physical transport of true charges.
2. In analogy with electric polarisation P there may be a form of charge trans-

port associated with the changes of the polarisation with time. We call such
currents induced by an external field polarisation currents. We identify them
with ^ P _ ^` .
3. There may also be intrinsic currents of a microscopic, often atomic, nature

that are inaccessible to direct observation, but which may produce net effects
at discontinuities and boundaries. We shall call such currents magnetisation
currents and denote them jm .
No magnetic monopoles have been observed yet. So there is no correspondence
in the magnetic case to the electric monopole moment (7.1). The lowest order mag-
netic moment, corresponding to the electric dipole moment (7.2), is the magnetic
dipole moment
' 1 ( ) +
m x,50 x0 - a j + x, - .
3/
, (7.15)
2
For a distribution of magnetic dipole moments in a volume, we may describe this
volume in terms of the magnetisation, or magnetic dipole moment per unit volume,
M. Via the definition of the vector potential one can show that the magnetisation
current and the magnetisation is simply related:
'
jm Lba M (7.16)
In a stationary medium we therefore have a total current which is (approxim-
ately) the sum of the three currents enumerated above:
7.3. E NERGY AND MOMENTUM 95
' ^ P
jtotal jtrue A A Lba M (7.17)
^`
We then obtain the Maxwell equation
' d ^ P
Lca B 0 B jtrue A A Lba ME (7.18)
^`
Moving the term Lea M to the left hand side and introducing the magnetising field
(magnetic field intensity, Ampère-turn density) as
B
H' d 0 M (7.19)
0
and using the definition for D, Equation (7.9) on page 93, we can write this Max-
well equation in the following form
' ^ D
Lca H jtrue A (7.20)
^`
We may, in analogy with the electric case, introduce a magnetic susceptibility
for the medium. Denoting it ] m , we can write
B
H' d (7.21)
where, approximately,
d ' d +
0 1A ] m- (7.22)
7.3 Energy and momentum

As mentioned in Chapter 1, Maxwell’s equations expressed in terms of the derived
field quantities D and H can be written
'f* +
LZJ D `4 x- (7.23a)
'
LZJ B 0 (7.23b)
' ^ B
Lca E 0 (7.23c)
^`
' ^
Lca H j+ `g4 x- A D (7.23d)
^`
and are called Maxwell’s macroscopic equations. These equations are convenient
to use in certain simple cases. Together with the boundary conditions and the con-
stitutive relations, they describe uniquely (but only approximately!) the properties
of the electric and magnetic fields in matter. We shall use them in the following
considerations on the energy and momentum of the electromagnetic field and its
interaction with matter.
7.3.1 The energy theorem in Maxwell’s theory

Scalar multiplying (7.23c) by H, (7.23d) by E and subtracting, we obtain
+
HJ Lba E - 0 E J + Lca H - ' LZJ
E a H- +
' ^ B ^ D ' 1 ^ +
0 HJ 0 EJ j0 EJ 0 H J B A E J D- 0 j J E (7.24)
^` ^` 2 ^`
Integration over the entire volume h and using Gauss’s theorem (the divergence
theorem), we obtain
^ ( ) 1+
0 H J B A E J D - . 3/ , ' ( ) j J E . 3/ , A ( + E a H - Ji. S, (7.25)
^` 2 S
But, according to Ohm’s law in the presence of an electromotive force field,
Equation (1.26) on page 11:
'Gj +
j E A EEMF - (7.26)
which means that
( ) 3/ ' ( ) 9 2 3/ ( ) 3/
j J E. , j . , 0
j J EEMF . , (7.27)
Inserting this into Equation (7.25)

( ) 3/ ' ( ) 9 2 3/
j J EEMF . , j . ,
k l m n k l m n
Applied electric power Joule heat
^ ( ) 1+ 3/
A E J D A H J B- . ,
^` k 2 l m n
Field energy
A ( +
E a H- Ji. S, (7.28)
k S l m n
Radiated power
which is the energy theorem in Maxwell’s theory also known as Poynting’s theorem.
It is convenient to introduce the following quantities:
o ' 1 ( )
e E J D . 3/ , (7.29)
2
o ' 1 ( )
m H J B . 3/ , (7.30)
2
S' Ea H (7.31)
o o
where e is the electric field energy, m is the magnetic field energy, both measured
in J, and S is the Poynting vector (power flux), measured in W/m2 .
7.3. E NERGY AND MOMENTUM 97
7.3.2 The momentum theorem in Maxwell’s theory
Let us now investigate the momentum balance (force actions) in the case that a
field interacts with matter in a non-relativistic way. For this purpose we consider
the force density given by the Lorentz force per unit volume * E A j a B. Using
Maxwell’s equations (7.23) and symmetrising, we obtain
* ' + ^ D
EA j a B LZJ D- E A B Lba H0 E a B
^`
' ^ D
E+ LZJ D- A +
Lba H- a B0 a B
^`
'
E+ LZJ D- 0 B a
+
Lca H-
^ + ^ B
0 D a B- A D a
^ `
^`
'
E + LZJ D- 0 B a
+
Lca H-
^ + +
0 D a B- 0 D a Lca E- A H + k LZl mJ B
n -
^` p 0
' q + + A q + +
E LZJ D- 0 D a Lca E -:r H LZJ B- 0 B a Lca H -Cr
^ +
0 D a B- (7.32)
^`
One verifies easily that the 3 th vector components of the two terms in square
brackets in the right hand member of (7.32) can be expressed as
q
E+ LZJ D- 0 D a + Lba E -Crs2
' 1 ^ D ^ E ^ 1
B EJ / 0 DJ / E A / B t 2 u 7 0 EJ D 27 E (7.33)
2 ^ 2 ^ 2 ^ 7 2 D
and
q
H+ LZJ B- 0 B a + Lca H -Cr 2
' 1 ^ B ^ H ^ 1
B HJ / 0 BJ / E A / B v 2 wx7 0 BJ H 27 E (7.34)
2 ^ 2 ^ 2 ^ 7 2 D
respectively.
Using these two expressions in the 3 th component of Equation (7.32) above and
re-shuffling terms, we get
+* 1 ^ D ^ E
EA j a B- 2 0 B EJ / 0 DJ / E
2 ? ^ 2 ^ 2
^ B ^ H ^ +
A B HJ / 0 BJ / E H A D a B- 2
^ 2 ^ 2 ^`
' ^ 1 1
/ B t 2 u 7 0 EJ D 27 A v 2 wx7 0 HJ B 27 E (7.35)
^ 7 2 D 2 D
Introducing the electric volume force Fev via its 3 th component

+ ' +*
Fev - 2 EA j a B- 2
1 ^ D ^ E ^ B ^ H
0 B E J / 0 DJ / E A B HJ / 0 BJ / E H (7.36)
2 ? ^ 2 ^ 2 ^ 2 ^ 2
and the Maxwell stress tensor

y ' 1 1
27 t 2 u 7 0 EJ D 27 A v 2 wx7 0 HJ B 27 (7.37)
2 D 2 D
we finally obtain the force equation
y
^ + ' ^ 27
Fev A D a B- P / (7.38)
O ^` 2 ^ 7
If we introduce the relative electric permittivity z and the relative magnetic per-
meability z m as
' '
D z = 0E = E (7.39)
' d ' d
B z m 0H H (7.40)
we can rewrite (7.38) as
y
^ 27 ' zz m ^ S
/ B Fev A { 2 E (7.41)
^ 7 ^` 2
where S is the Poynting vector defined in Equation (7.31) on page 96. Integration
over the entire volume h yields
( ) 3/ . ( ) zz m 3/ ' ( 2/
Fev . , A { 2 S. , T| ˆ . , (7.42)
k l m n .` k l m n k S l m n
Force on the matter Field momentum Maxwell stress
which expresses the balance between the force on the matter, the rate of change
of the electromagnetic field momentum and the Maxwell stress. This equation is
called the momentum theorem in Maxwell’s theory.
In vacuum (7.42) becomes
( ) * + 3/ 1 . ( ) 3/ ' ( 2/
EA v a B- . , A { 2 S. , T| ˆ . , (7.43)
.` S
or
. . ' ( 2/
pmech A pfield T| ˆ . , (7.44)
.` .` S
99
B IBLIOGRAPHY 7
ISBN 0-201-05702-6.
101
102
C HAPTER 8
Electromagnetic
Radiation
In this chapter we will develop the theory of electromagnetic radiation, and there-
fore study electric and magnetic fields which are capable of carrying energy and
momentum over large distances. In Chapter 3 we were able to derive general ex-
pressions for the scalar and vector potentials from which we then, in Chapter 4 cal-
culated the total electric and magnetic fields from arbitrary distributions of charge
and current sources. The only limitation in the calculation of the fields was that the
advanced potentials were discarded.
We shall now study these fields further under the assumption that the observer is
located in the far zone, i.e., very far away from the source region(s). We therefore
study the radiation fields which are dominating in this zone.
8.1 The radiation fields

From Equation (4.13) on page 46 and Equation (4.24) on page 49, which give the
total electric and magnetic fields, we obtain the radiation fields
~
Brad + `4 x- ' ( }
Brad + x - .
~
d }
' 0 ( ) j̇ + ` ret
, 4 x, - a
+
x 0 x, - 3/
{ @ @2 . , (8.1)
4< x 0 x,
( ~
Erad + `4 x- ' }
Erad + x - .
~
} q + + +
' 1 ( ) j̇ ` ret
, 4 x, - a x 0 x, -Cr a x 0 x, - 3/
{ 2 @ @3 . , (8.2)
4<K= 0 x 0 x,
where
103
104 C HAPTER 8. E LECTROMAGNETIC R ADIATION
^ j
j̇ +
def
`Cret
, 4 x, - B E (8.3)
^` p ret

Instead of studying the fields in the time domain, we can often make a spectrum
analysis into the frequency domain and study each Fourier component separately.
A superposition of all these components and a transformation back to the time
domain will then yield the complete solution.
The Fourier representation of the radiation fields (8.1) (8.2) were included in
Equation (4.12) on page 46 and Equation (4.23) on page 48, respectively and are
explicitly given by
1
Brad + x - ' ( }
Brad + `4 x- .`
2< ~
d }
' 0 ( ) j + x , - a
+
x 0 x, - x ~ x 3/
0 @ @2 . ,
4< x 0 x,
d +
' 0 ( ) j x, - a k Q x ~ x 3/
0 @ @ . , (8.4)
4< x 0 x,
1 (
Erad + x - ' }
Erad + `4 x- .`
2< ~
} q
' ( ) j + x , - a
+
x 0 x, -:r a
+
x 0 x, - Q x ~ x 3/
0 { @ @3 . ,
4<>= 0 x 0 x,
q
' 1 ( ) j + x , - a kr a + x 0 x , - QM x ~ x 3/
0 { @ @2 . , (8.5)
4<>= 0 x 0 x,
If the source is located@ inside a volume h near x0 and has such a limited spatial
@ @ @
extent that max x, 0 x0 @ x 0 x@, , and the@ integration
@
surface , centered on x0 ,
has a large enough radius x 0 x0 max x, 0 x0 , we see from Figure 8.1 that
we can approximate
+ @ @
x 0 x, k J x 0 x, - x 0 x0 0 k J + x,0 x0 - (8.6)

and
T
ˆ J ˆ
@ @ 2 .X. Ω (8.7)
x 0 x0
' @ @2
where . x 0 x0 . Ω, was used.
8.1. T HE RADIATION FIELDS 105
ˆ 2
S
x x
¡ˆ
x x0 x
x x0
x x0 ¤
Figure 8.1. Relation between the surface normal and the k vector for radiation
generated at source points x¥ near the point x0 in the source volume ¦ . At dis-
tances
¨
much larger than the extent of ¦ , the unit vector § ˆ , normal to the surface
©
which has its centre at x0 , and the unit vector ˆ of the radiation k vector from
x¥ are nearly coincident.
Within approximation (8.6) the expressions (8.4) and (8.5) for the radiation fields
can be approximated as
d
QM x ~ x0 ( ) j + x , - a k ~ k ¬® x ~ x0 ¯
Brad + x - 3/
0
ª0« @ @ . ,
4< x 0 x,
d QM x ~ x0
0 ~ k ¬® x ~ x0 ¯
ª0« @ @ ( ) q j + x, - a kr .
3/
, (8.8)
4< x 0 x0
q
1 Q x ~ x0 ( ) j + x, - a kr a
+
x 0 x, - ~ k ¬® x ~ x0 ¯ 3/
Erad + x - ª0« { @ @2 . ,
4<K= 0 x 0 x,
x ~ x0 +
1 x0 x0 - ( ) q ~ k ¬® x ~ x0 ¯
f { @ @ @ @ a j + x, - a kr .
3/
,
4<K= 0 x 0 x0 x 0 x0
(8.9)
@ @ @ @
I.e., if max x, 0 x0 x 0 x, , then the fields can be approximated as spherical
waves multiplied by dimensional and angular factors, with integrals over points in
the source volume only.
8.2 Radiated energy

Let us consider the energy that carried in the radiation fields Brad , (8.1), and Erad ,
Equation (8.2) on page 103. We have to treat signals with limited lifetime and
hence finite frequency bandwidth differently from monochromatic signals.
8.2.1 Monochromatic signals

If the source is strictly monochromatic, we can obtain the temporal average of the
radiated power ° directly, simply by averaging over one period so that
± ' ± ' 1 ' 1 ~ + ~

S² E a H² d Re ³ E a B´ µ Re ¶ E a B - ´ ·
2 0 2d 0
' 1 ~ ' 1
Re ¶ E a B ´ · Re ³ E a B ´ µ (8.10)
2d 0 2d 0
Using the far-field approximations (8.8) and (8.9) and the fact that 1 _ {¸' ¹ d
= 0 0
and º 0 '¼» d 0 _ = 0 according to the definition (2.15) on page 23, we obtain
2
± ' 1 1 ( ) q½+ ~ k ¬Q x ~ x0 ¯ 3/ x 0 x0
S² º @2 j a k -:r . , @ @ (8.11)
32< 2 0 @ x 0 x0
x 0 x0

or, making use of (8.7) on page 104,
2
.° ' 1 ( ) q½+ ~ k ¬Q x ~ x0 ¯ 3/
º j a k -:r . , (8.12)
. Ω 32< 2 0

which is the radiated power per unit solid angle.
8.2.2 Finite bandwidth signals

A signal with finite pulse width in time (` ) domain has a certain spread in fre-
quency ( ) domain. To calculate the total radiated energy we need to integrate
over the whole bandwith. The total energy transmitted through a unit area is the
8.2. R ADIATED ENERGY 107
time integral of the Poynting vector:

( } (
S+ ` - .`
' } +
E a H- .`
~ ~
} }
' ( } ( } ( } + ~ ¿ ¯
. .¾, E a H - . `
(8.13)
~ ~ ~
} } }
If we carry out the temporal integration first and use the fact that
( } ~ ¿ ¯ ' +
. `
2< A ¾, - (8.14)
~ D
}
Equation (8.13) can be written [cf. Parseval’s identity]
( }
S+ ` - .`
'
2< ( } +
E a H~ - .
~ ~
} } 0
' ( } + A ( +
2< B E a H~ - . E a H~ - . E
0 ~
}
' ( } + ( } +
2< B E a H ~ - .À0 E~ a H - . E
0 0
' 2< ( } +
d E a B~ A E~ a B - .
0 0
' 2< ( } +
d E a B ´ A E ´ a B - . (8.15)
0 0
where the last step follows from the real-valuedness of E and B . We insert the
Fourier transforms of the field components which dominate at large distances, i.e.,
the radiation fields (8.4) and (8.5).
The result, after integration over the area of a large sphere which encloses the
source, is
d 2
o ' 1 0 ( ( } ( ) j a k Q x ~ x 3/ T
@ @ . , ˆ J ˆ .Ã. (8.16)
4<ÂÁ = 0 S 0 x 0 x,

(8.6) and (8.7) into Equation
Inserting the approximations (8.16) above and also
introducing
o ' ( } o
. (8.17)
0
and recalling the definition (2.15) on page 23 for the vacuum resistance º 0 we
obtain
o 2
. 1 ~ ~
º 0
( ) +
j a k - k ¬® x x0 ¯ . 3/ , .
.Ä (8.18)
. Ω 4<

which, at large distances, is a good approximation to the energy that is radiated
per unit solid angle . Ω in a frequency band . . It is important to notice that
Formula (8.18) includes only source coordinates. This means that the amount of
energy that is being radiated is independent on the distance to the source (as long
as it is large).
8.3 Radiating systems

As shown above, one can, at least in principle, calculate the radiated fields, Poynt-
ing flux and energy for an arbitrary current density Fourier component. However, in
practice, it is often difficult to evaluate the integrals unless the current has a simple
distribution in space. In the general case, one has to resort to approximations. We
shall consider both these situations.
8.3.1 Simple geometries

Certain radiation systems have a geometry which is one-dimensional, symmetric
or in any other way simple enough that a direct calculation of the radiated fields
and energy is possible. This is for instance the case when the current flows in one
direction in space only and is limited in extent. An example of this is a linear
antenna.
Linear antenna
Let us apply Equation (8.12) on page 106 for calculating the power from a linear,
transmitting antenna, fed across a small gap at its centre with a monochromatic
source. The antenna is a straight, thin conductor of length Å which carries a one-
dimensional time-varying current so that it produces electromagnetic radiation.
We assume that the conductor resistance and the energy loss due to the electro-
magnetic radiation are negligible. Since we can assume that the antenna wire is
infinitely thin, the current vanishes at the end point. The current therefore forms a
standing wave with wave number ' _ { and can be written

@/ @
+ 'fÆ + / +/ sin q +
Å _ 20 3, -:r Ç
j0 x , - 0 1, - 2, - + ˆ3 (8.19)
D D sin Å _ 2-

where the amplitude Æ 0 is constant. In order to evaluate Formula (8.12) on page 106
8.3. R ADIATING SYSTEMS 109
with the explicit monochromatic current (8.19) inserted, we need the expression
2
~ ~

( ) +
j0 a k - k ¬Q x x0 ¯ . 3/ ,

2 q + @/ @ 2
' sin Å _ 20 3, -:r ~Q Ë ~Q Ë 0
(ÄÈ É Æ
sin Ê 3ÎÌ ÍÏÑÐ ÑÌ ÍÏÑÐ . / ,
0 + 3
2 sin Å _ 2-
È É
2 sin2 Ê 2 2
'GÆ 2 Q Ë cos Ð
2
2 ( È É q + / + / /
0 2+
0
sin Å _ 20 3, -:r cos 3, cos Ê - . 3,
sin Å _ 2- 0

q½+ + 2
' Æ 2 cos Å _ 2 - cos Ê r 0 cos Å _ 2-
4 0 B E (8.20)
sin Ê sin + Å _ 2-

inserting this expression and . Ω ' 2< sin Ê . Ê into Formula (8.12) on page 106
and integrating over Ê , we find that the total radiated power from the antenna is
+ ' Æ 2 1 ( Ò cos q½+ Å _ 2 - cos Ê r 0 cos + Å _ 2- 2

° Å - º 0 0 B E sin Ê . Ê (8.21)
4< 0 sin Ê sin + Å _ 2-

One can show that
+ ' < ÅÔ 2
lim ° Å - B E º 0Õ 0 (8.22)
0 12
È Ó
Ô
where is the vacuum wavelength.
The quantity
+ + 2 2
+ ' ° Å - ' ° Å - ' < ÅÔ ÅÔ
º rad Å - Æ 2 1Æ 2
º 0 B E 197 B E Ω (8.23)
eff 2 0
6
Ô
is called the radiation resistance. For the technologically important case of a half-
' '
wave antenna, i.e., for Å _ 2 or Å < , Formula (8.21) reduces to

Ô Ö
+ ' Æ 2 1 ( Ò cos2 Ò2 cos Ê ×
° _ 2- º 0 0 . Ê (8.24)
4< 0 sin Ê
The integral in (8.24) can be evaluated numerically. It can also be evaluated

analytically as follows:
cos2 Ö cos Ê × 1 cos2 Ö Ò 2 Ú×
( Ò Ò2 ' Ø ' ( '
. Ê cos Ê ÙÛÚ Ü .Ú
0 sin Ê ~ 1 1 0XÚ 2
Ý
Þ < ' 1A cos + <>Ú -
cos2 Úß
2 2 à
' 1 (
1
1 A cos + <KÚ -
+ A .Ú
2 ~ 1 1 Ú - + 1 0XÚ -
' 1 (
1
1 A cos + <>Ú - 1 (
1
1 A cos + <KÚ -
+ A .Ú A + .Ú
4 ~ 1 1 Ú - 4 ~ 1 1 0XÚ -
' 1 (
1
1 A cos + <>Ú - 'âá
+ A .Ú 1A ÚãÙ ä
2 ~ 1 1 Ú - <xå
' 1 ( 2Ò 1 0 cos ' 1qæ

ä . A ln 2<ã0 Ci + 2< -Cr
2 0 ä 2
ä
1 F 22 (8.25)
where in the last step the Euler-Mascheroni constant æç' 0 F 5772 FFF and the cosine
integral Ci + / - were introduced. Inserting this into
Ô the expression Equation (8.24)
+
on the previous page we obtain the value º rad _ 2 - 73 Ω.
8.3.2 Multipole expansion

In the general case, and when we are interested in evaluating the radiation far from
the source volume, we can introduce an approximation which leads to a multi-
pole expansion where individual terms can be evaluated analytically. We shall use
Hertz’ method to obtain this expansion.
The Hertz potential

Let us consider the continuity equation, which, according to expression (1.21) on
page 9, can be written
* +
^ `4 x-
A LZJ j+ `4 x- '
0 (8.26)
^`
If we introduce a vector field p + `4 x - such that
' *
LZJ p 0 true (8.27)
^ p '
jtrue (8.28)
^`
and compare with Equation (8.26) above, we see that p + `g4 x - satisfies this continu-
ity equation. Furthermore, if we compare with the electric polarisation [cf. Equa-
tion (7.9) on page 93], we see that the quantity p is related to the “true” charges in
the same way as P is related to polarised charge. Therefore, p is referred to as the
polarisation vector.
We introduce a further potential Z with the following property
' ;
LZJ Z 0 (8.29a)
1 ^ Z '
{ 2 A (8.29b)
^ `

;
where and A are the electromagnetic scalar and vector potentials, respectively.
As we see Z acts as a “super-potential” in the sense that it is a potential from which
we can obtain other potentials. It is called the Hertz’ vector or polarisation poten-
tial and, as can be seen from (8.28) and (8.29b), it satisfies the inhomogeneous
è
wave equation
2 ' 1 ^ 2 2 ' p
Z { 2 Z0 é Z (8.30)
^` 2 = 0
This equation is of the same type as Equation (3.19) on page 36, and has there-
fore the retarded solution
1 ( p + ` ret
, 4 x, -
Z+ `4 x- ' @ @ .
3/
, (8.31)
4<>= 0 x 0 x,
with Fourier components
QM ~ x
1 ( p + x , - x
Z + x - ' @ @ .
3/
, (8.32)
4<K= 0 x 0 x,
If we introduce the help vector C such that
'
C Lca Z (8.33)
we see that we can calculate the magnetic and electric fields, respectively, as fol-
lows
' 1 ^ C
B { 2 (8.34a)
^ `

'
E Lca C (8.34b)
where the last equation is valid only outside the source volume, where LZJ E ' 0.
Since we are mainly interested in the fields in the far zone, a long distance from
the source region, this is no essential limitation.
Assume that the source region is a limited volume around some central point
x0 far away from the field (observation) point x. Under these assumptions, we
can expand expression (8.31) on the preceding page the Hertz’ vector, due to the
presence of non-vanishing p + ` , 4 x, - in the vicinity of x0 , in a formal series. For this
purpose we recall from potential theory that
Q x ~ x ê }
@ @
@ @ '
+
2ì A 1- ° ë
+
cos Θ - 9 ë
+
x,0 x0 1 +
- í ë ¯ x 0 x0 - (8.35)
x 0 x, ë p
0
where
Θ is the angle between x,0 x0 and x 0 x0
+
° ë cos Θ - is the Legendre polynomial of order ì
9 ë + x,0 x0 - is the spherical Bessel function of the first kind of order ì
@
ë 1 ¯ + @
í x0 x0 - is the spherical Hankel function of the first kind of order ì

According to the addition theorem for Legendre polynomials, we can write
ë
ê î î ~ î î ðï ~ ï ¯
+ ' + + +
° ë cos Θ - 0 1- ° ë cos Ê - ° ë cos Ê ,- (8.36)
î p ~ ë
î
where ° ë is an associated Legendre polynomial and, in spherical polar coordin-
ates,
' + ;
x,0 x0 x,0 x0 4 Ê ,ñ4 , - (8.37a)
' + @ ; @
x 0 x0 x0 x0 4 Ê 4 - (8.37b)
Inserting Equation (8.35) above, together with Equation (8.36), into Equation (8.32)
on the previous page, we can in a formally exact way expand the Fourier compon-
ent of the Hertz’ vector as
ë
ê } ê î @ @ î î ï
' + 1 + +
Z 0 1- í ë ¯ x 0 x0 - ° ë cos Ê -
4<K= 0 ë p î p ~ ë
0
( ) ~ î
a p + x , - 9 ë
+
x,0 x0 - ° ë
+
cos Ê ,- .
3/
, (8.38)

We notice that there is no dependence on x 0 x0 inside the integral; the integrand
is only dependent on the relative source vector x, 0 x0 .
We are interested in the case where the field point is many wavelengths away
from the sources, i.e., when the following inequalities
@ @
x,0 x0 1 x 0 x0 (8.39)

1¯
hold. Then we may to a good approximation replace í ë with the first term in its
asymptotic expansion:
x ~ x0
1 + ë
í ë ¯ 0 - @ @ (8.40)
x 0 x0
and replace 9 ë with the first term in its power series expansion:
ë
2 ì ! ë
+ Ö
9 ë x,0 x0 - + x,0 x0 × (8.41)
2ì A 1- !

Inserting these expansions into Equation (8.38) on the facing page, we obtain the
multipole expansion of the Fourier component of the Hertz’ vector
ê } ë
Z Z ¯ (8.42a)
ë p 0
where
Q x ~ x0 ë
ë ë 1 2 ì ! ( ) ë
Z ¯ ' +
0« - @ @ p + x , - +
x,0 x0 - ° ë
+
cos Θ - .
3/
,
4<>= 0 x 0 x0 + 2ì - !

(8.42b)
This expression is approximately correct only if certain care is exercised; if many

terms are needed for an accurate result, the expansions of the spherical Hankel and
Bessel functions used above may not be consistent and must be replaced by more
accurate expressions. Taking the inverse Fourier transform of Zò will yield the
Hertz’ vector in time domain, which inserted into Equation (8.33) on page 111 will
yield C. The resulting expression is then in turn to be inserted into Equation (8.34)
on page 111 in order to obtain the radiation fields.
For a linear source distribution along the polar axis, Θ ' Ê in expression (8.42b)
above, and ° ë + cos Ê - gives the angular distribution of the radiation. In the general
case, however, the angular distribution must be computed with the help of For-
mula (8.36) on the facing page. Let us now study the lowest order contributions to
the expansion of Hertz’ vector.
Electric dipole radiation

'
Choosing ì 0 in expression (8.42b) above, we obtain
Q x ~ x0 x ~ x0
' 1
Z 0 ¯ @ @ ( ) p + x, - .
3/
,
' @ @ p 0 ¯ (8.43)
4<K= 0 x 0 x0 4<K= 0 x 0 x0
0
where p ¯ ' p"ó 1 is the Fourier component of the electric dipole moment; cf.
Equation (7.2) on page 91 which describes the static dipole moment. If a spherical
0
coordinate system is chosen with its polar axis along pó 1 , the components of Z ¯
are ôöõ ô
Q x ~ x0
' 0¯ ' 1
ô ô cos Ê @ 1 ó 1 cos Ê @ (8.44a)
4<>= 0 x 0 x0
Q ~
ô ' 0¯ ' 1 x x0
0 sin Ê 0 @ @ 1 ó 1 sin Ê (8.44b)
Ð 4<K= 0 x 0 x0
'
ï 0 (8.44c)
Evaluating Formula (8.33) on page 111 for the help vector C, with the spheric-
0
ally polar components (8.44) of Z ¯ inserted, we obtain
Q x ~ x0
0¯ ù 1 1
C 'ø÷
"ó ï ˆ '
B @ @ 0X E @ @ 1 ó 1 sin Ê ù ˆ (8.45)
4<K= 0 x 0 x0 x 0 x0
Applying this to Equation (8.34) on page 111, we obtain directly the Fourier com-
ponents of the fields:
d Q x ~ x0
1
' @ 1 ó 1 sin Ê ù
0
B 0« B @ @ 0À E @ ˆ (8.46)
4< x 0 x0 x 0 x0
' 1 1 x 0 x0
E 2 ú @ @2 0 @ @û cos Ê @ @ (8.47)
4<K= 0 ? x 0 x0 x 0 x0 x 0 x0
Q x ~ x0
1 ü
A ú @ @2 0 @ @ 0 2û
sin Ê ˆ H @ @ 1 "ó 1 (8.48)
x 0 x0 x 0 x0 x 0 x0
Keeping only the parts which dominate at large @

distances
@
(radiation field) and
recalling that the wave vector k ' +
x 0 x0 -_ x 0 x0 where ' _ { , we can

now write down the Fourier components of the radiation parts of the magnetic and
electric fields from the dipole:
Q x ~ x0
d
' 0
Brad 0 @ + pó 1 a
@ k- (8.49a)
4 < x 0 x0
Q ~
' 1 x x0 ½q +
Erad 0 @ @ pó 1 a k- a kr (8.49b)
4>
< = 0 x 0 x0
These fields constitute the electric dipole radiation, also known as E1 radiation.
Magnetic dipole radiation

The next term in the expression (8.42b) on the previous page for the expansion of
the Fourier transform of the Hertz’ vector is for ì ' 1:
Q x ~ x0

Z 0«
1¯ ' @ @ ( ) x,0 x0 p + x, - cos Θ . 3/ , (8.50)
4<>= 0 x 0 x0
Q x ~ x0
' 1 ( ) q½+
0« @ @ x 0 x0 - J + x,0 x0 -:r p + x, - . 3/ , (8.51)
4<K= 0 x 0 x0 2
Here, the term q½+ x 0 x0 - J + x, 0 x0 -:r p + x, - can be rewritten

q½+ +
x 0 x0 - J x,0 x0 -:r p + x, - ' + /
2 0
/
0ó 2 -
+/
,50
/
0ó 2 - p + x , - (8.52)
ý
and introducing
ý 2 ' /
2 0
/
0ó 2 (8.53a)
' / /
2, 2, 0 0ó 2 (8.53b)
1¯
the 9 th component of the integrand in Z ý
can be broken
ý
upý into
q½+ 1 Ö +
³ x 0 x0 - J x,0 x0 -:r p + x, -gµ
2 1 ó 7 2 , A 1 ó 2 7 , × 2
'
(8.54a)
2ý ý ý
1 Ö
A 2 1 ó 7 2 , 0þ1 ó 2 7 , × (8.54b)
2
i.e., as the sum two parts, the first being symmetric and the second antisymmetric
in the indices 384:9 . We note that the antisymmetric part can be written as
ý ý ý ý ý ý ý
1 Ö ' 1 + +
2 1 ó 7 2 , 0ç1 ó 2 7 , × 1 ó 7 2 2, - 0 7, 2 1 ó 7 -
2 2
' 1q
p + ÿ J ÿ - 0 ÿ + ÿ J p -:r 7
2
' 1 + q +
¶ x 0 x0 - a p a x,0 x0 -:r · 7 (8.55)
2
From Equation (8.28) on page 110 and the fact that we are considering a single
Fourier component
~
p+ `4 x- '
p (8.56)
which allows us to express p in j as
jò
p ' (8.57)

we can write the antisymmetric part of the integral in Formula (8.51) above as
1+
x 0 x0 - a ( ) p + x, - a + x,50 x0 - h ,
2
1 +
'
x 0 x0 - a ( ) j + x, - a + x,0 x0 - . 3/ ,
2
' 1+
0 x 0 x0 - a m ò (8.58)

where we introduced the Fourier transform of the magnetic dipole moment
' 1 ( ) +
mò x,0 x0 - a j + x , - .
3/
, (8.59)
2
1¯
The final result is that the antisymmetric, magnetic dipole, part of Z can be
written
x ~ x0
ó antisym ' +
Z 1 ¯ 0 @ @2 x 0 x0 - a m (8.60)
4<K= 0 x 0 x0
In analogy with the electric dipole case, we insert this expression into Equation (8.33)
on page 111 to evaluate C, with which Equations (8.34) on page 111 then gives the
B and E fields. Discarding, as before, all terms belonging to the near fields and
transition fields and keeping only the terms that dominate at large distances, we
obtain
QM x ~ x0
d
0
Brad + x - '
0 @ + @ m a k- a k (8.61a)
4 < x 0 x0
x ~ x0

Erad + x - '
{ @ @ m a k (8.61b)
4K
< = 0 x 0 x0
which are the fields of the magnetic dipole radiation (M1 radiation).
Electric quadrupole radiation

1 ó sym
The symmetric part Z ¯ of the ì ' 1 contribution in the Equation (8.42b) on
page 113 for the expansion of the Hertz’ vector can be expressed in terms of the
electric quadrupole tensor, which is defined in accordance with Equation (7.3) on
page 91:
( ) +
Q+ ` - ' x,0 x0 - + x,50 x0 - * +
`g4 x, - .
3/
, (8.62)
Again we use this expression in Equation (8.33) on page 111 to calculate the
fields via Equations (8.34) on page 111. Tedious, but fairly straightforward al-
gebra (which we will not present here), yields the resulting fields. The radiation
components of the fields in the far field zone (wave zone) are given by
d x ~ x0
+ ' 0
Brad x - @ @ + k J Q - a k (8.63a)
8< x 0 x0
Q x ~ x0

Erad + x - ' @ @ sq + k J Q - a kr a k (8.63b)
8K
< = 0 x 0 x0
This type of radiation is called electric quadrupole radiation or E2 radiation.


x
u
x x S

x
Figure 8.2. Signals which are observed at the field point x were generated at
source points x on a sphere, centred on x and expanding, as time increases,
with the velocity c outward from the center. The source charge element moves
with an arbitrary velocity u and gives rise to a source “leakage” out of the source
volume 3 .
8.3.3 Radiation from charges moving in vacuum

The derivation of the radiation fields for the case of the source moving relative
to the observer is considerably more complicated than the stationary cases stud-
ied above. In order to handle this non-stationary situation, we use the retarded
potentials (3.36) on page 39 in Chapter 3
) * + -! , " , ,
!#" 1
. ret x, . $ 0 31
x$ % (8.64a)
4&(' 0 x/ x
) * !, " ,
!#" j ret x $ 31 ,
A x$ % 2
0
. ,. 0 (8.64b)
4& x/ x
and consider a source region with such limited extent that the charges and currents
,
are well localised. Specifically, we consider a charge 3 , for instance an electron,
with unspecified but limited spatial extent. Classically, an electron is a localised,
unstructured charge distribution with a small, finite radius. The part of this “charge
, , ,
distribution” 0 3 which we are considering is located in 054 % 0 31 in the sphere in
Figure 8.2. Since we assume that the electron (or any other other charge) is moving
with a velocity u whose direction is arbitrary and whose magnitude can be almost
comparable to the speed of light, we cannot say that the charge and current to
) )
be used in (8.64) is 6 * + ` ret , 4 x, - .
3/ and 6
, u* + ` ret
, 4 x, - .
3/ , respectively, because
,
during the finite time interval that the observed signal is generated, part of the
charge distribution will “leak” out of the volume element . 3/ , .
The charge distribution which@ contributes @
to the field at + `4 x - is located at + ` , 4 x, -
on a sphere with radius 7 ' x 0 x, ' { + `V0` , - . The radius interval of this
sphere from which radiation is received at the field point x during the time interval
`4I` A .` - is 7 47 A .87 - and the net amount of charge in this radial interval is
+ +
+
x 0 x, - J u
.
&
,
'G* +
`Cret
, 4 x, - .¸. %0 87 * +
` ret
, 4 x, - @ @ .Ã.` (8.65)
x 0 x,
where the last term represents the amount of “source leakage” due to the fact that
the charge distribution moves with velocity u. Since .` ' .87 _ { and .¸.87 ' . 3/ ,
we can rewrite this expression for the net charge as
+
& 'f* + 3/ * + x 0 x, - J u 3/
. , ` ret
, 4 x, - . ,0 ` ret
, 4 x, - { @ @ . ,
x 0 x,
+
'f* + x 0 x, - J u 3/
` ret
, 4 x, - B 10 { @ @ E . , (8.66)
x 0 x,
or
&
* + 3/ ' . ,
`Cret
, 4 x, - . , (8.67)
x ~ x ¯ ¬ u
10 ~
Ì x x
which leads to the expression

* + &
` , 4 x, - . ,
@ ret @ . 3/ , ' @ @ (8.68)
x 0 x, x ~ x ¯ ¬ u
x 0 x, 0
Ì
This is the expression to be used in the expressions (8.64) for the retarded poten-
tials. The result is
&
; + ' 1 ( . ,
`g4 x- @ @ (8.69a)
4<>= x ~ x ¯ ¬ u
0 x 0 x, 0
Ì
d &
u. ,
A+ ' 0 (
`g4 x- @ @ (8.69b)
4< x ~ x ¯ ¬ u
x 0 x, 0
Ì
For a sufficiently small and well localised charge distribution we can, assuming
that the integrands do not change sign in the integration volume, use the mean
)
value theorem and the fact that 6 .
&
,
'G&
, , evaluate these expressions to become
& &
; + ' 1 , ' 1 ,
`4 x-
4<K=
@ @ x ~ x ¯ ¬ u 4<K= 9 (8.70a)
0 x 0 x, 0 0
Ì
& &
1 , u 1 ,u
A+ `4 x- '
{ 2 @ @ x ~ x ¯ ¬ u
'
{ 2 9 (8.70b)
4<K= 0 x 0 x, 0 4<K= 0
Ì
where
+
9 '
x 0 x, 0
x 0 x, - J u
(8.71)
{

is the retarded relative distance. These potentials are precisely the Liénard-Wiechert
potentials which we derived in Equation (8.70) above using a covariant formalism.
It is important to realise that in the complicated derivation presented here the ob-
server is in a coordinate system which has an “absolute” meaning and the velocity
u is that of the particle, whereas in the covariant derivation two frames of equal
standing were moving relative to each other with u. Expressed in the four-potential,
Equation (5.45) on page 64, the Liénard-Wiechert potentials become
:<; + / = '
&
, 1 u ' + ; {
-
4<>= 0
B 9 4 { 9 E 4 A- (8.72)
The Liénard-Wiechert potentials are applicable in all problems where a spa-

tially localised charge emits electromagnetic radiation, and we shall now study
such emission problems.
Uniformly moving charges

Generally speaking, the Liénard-Wiechert potentials (8.70), which give rise to ra-
diation observed at the field point + `g4 x - , can only be determined as functions of
the retarded quantities at + ` ret
, 4 x, - ; there is usually no way that an observer at time
` can obtain any knowledge about what has happened to the charge after ` ret , . The
charge may even have ceased to exist! However, if the charge moves in a field-free,
isolated space we know that it will not be affected by any external forces and will
therefore move uniformly in a straight line. This gives us the possibility to estimate
(but not to observe!) its position at time ` . To compute the fields from a uniformly
moving charge is therefore relatively easy.
Assume that we have a charged particle which moves in vacuum with uniform
velocity u. At time ` , , when the particle passes point x, , it emits a field signal
G x H xD G ? @BA x C
I u 0 u
? @EDFA xD C >
x H xD
? @BA x C
Figure 8.3. Signals which were generated at the source point JK ¥L x¥NM are ob-
served at the field point JK L x M . The particle, which moves with constant, uniform
velocity u, has then reached the “simultaneous” coordinate x0 .
{
which propagates with speed and arrives at the observation point at time ` . In
other words,
' + {
x 0 x, ` 0X` , - (8.73)

Under the same time interval `«0` , , the particle has moved in a straight line a
distance
@ @
+ ' x 0 x, Ú
`ö0X` , - Ú { (8.74)
to the “simultaneous” position + `4 x0 - . The radius vector of this computed point
relative to the field point + `g4 x - is
@ @
' x 0 x, u
x 0 x0 x 0 x,0 { (8.75)
According to Equation (8.71) on the previous page, the square of the retarded
relative distance 9 is
+ + 2
9 2 '
x 0 x,
2
0 2 x 0 x,
x 0 x, - J u
A B
x 0 x, - J u
E (8.76)
{ {

Combining this with the identity
2 2 ' q½+ q½+
x 0 x, Ú x 0 x, - J ur 2 A x 0 x, - a ur 2 (8.77)

and, from Formula (8.75),
+ ' +
x 0 x, - a u x 0 x0 - a u (8.78)
we can express 9 in terms of the “simultaneous” coordinate x0 :
+
9
2
sin2 O
' @ @2 x 0 x0 - a u ' @ @ Ú 2
x 0 x0 0 B { E x 0 x0 10 { 2 (8.79)
Á
where O is the angle between u and x 0 x0 . Consequently, in the special case

of a uniform motion, the retarded distance 9 in the Liénard-Wiechert potentials
(8.70) can be expressed in terms of the “simultaneous” coordinate, viz., the point
at which the particle has arrived at the time ` when we obtain the first knowledge
of its existence at the source point x, at the retarded time ` , .
The electric field is calculated from the potentials in the usual way:
B+ `4 x- '
Lba A+ `4 x- (8.80a)
; + ^ A+ `g4 x-
E+ `4 x- '
0VL `4 x- 0 (8.80b)
^`
Because of the uniformity of the velocity u we can transform temporal differences
into spatial differences, and vice versa. A stationary observer at the observation
(field) point x measures at time ` A .` a field which, at time ` existed at a point a
distance 0 u .` away from x so that
A+ ` A .`4 x- 0 A+ `4 x- '
A+ `4 x 0 u .` - 0 A+ `g4 x- (8.81)
This means that partial time derivatives can be replaced by partial space derivatives:
^
A Ù 0 u JML A (8.82)
^`
where we notice the dyadic product.
The insertion of the Liénard-Wiechert potentials (8.70) and the relation (8.82)
into Formula (8.80b) above gives
&
' , Þ uu 9
9
2
E { 2 3
10 JQßL (8.83)
8<K= 0
With 9 given by Formula (8.79), this can be rewritten as
&
' , Ú 2 +
E
4<>= 0
9 3
B 10 { 2E x 0 x0 -
& 1 0QP
2
' , 2 +
Ì x 0 x0 - (8.84)
O
3
4<>= 0 @ @3 Þ
x 0 x0 10 PÌ 2
2 sin 2
ß
2
We see that E is directed along the vector from the “simultaneous” source point
+ +
`4 x0 - to the field (observation) point `4 x - . In a similar way, the magnetic field
can be calculated and one finds that
d &
0 , Ú 2 1
B' 9 3
B 10 { E u a
2
+
x 0 x0 - ' { 2 u a E (8.85)
4<K= 0
From these explicit formulas for the E and B fields we see that
1. ÚãÙ 0 R E goes over into the Coulomb field ECoulomb
2. ÚãÙ 0 R B goes over into the Biot-Savart field
3. ÚãÙ
{
R E becomes dependent on O
4. ÚãÙ
{
4 sin O 0 R E Ù
+
1 0XÚ 2_ { 2- ECoulomb
5. ÚãÙ
{
4 sin O 1 R E Ù
+
1 0XÚ 2 { 2 ~ 1 2 ECoulomb
_ - É
Accelerated charges
Consider a point charge & , and assume that its trajectory is known as a function of
retarded time
' +
x, x, ` ,- (8.86)
This means that we know the source point x, at which a signal is emitted at time ` ,
in order to arrive at the field point x at time ` . We obtain the retarded velocity and
acceleration from
. x,
u+ `C, -
'
(8.87a)
.` ,
. u . 2 x,
a+ ` ,-
'
u̇ + ` ,-
' '
(8.87b)
.` , .` , 2
If we choose the field point as fixed, (8.87) yields for the relative vector x 0 x, :
. + '
x 0 x, - 0 u+ `C, - (8.88a)
.` ,
. 2 + '
x 0 x, - 0 u̇ + `C, - (8.88b)
.` , 2
The retarded time ` , can, at least in principle, be obtained from the implicit relation
@ + + @
' x ` - 0 x, ` ,-
` 0X` , { (8.89)
The fields are determined, as usual, from the potentials, Formulae (8.80) on the
previous page. In these formulae the unprimed L , i.e., the spatial derivative differ-
entiation operator
' ^ Ç
L / ˆ2 (8.90)
^ 2
means that we differentiate with respect to the coordinates of x when we keep `

fixed, and the unprimed time derivative operator ^ _ ^` that we differentiate with
respect to ` while keeping x fixed. But the Liénard-Wiechert potentials, Equa-

tions (8.70) on page 119, are expressed in the u given by Equation (8.87b) on the
preceding page and the retarded relative distance 9 given by Equation (8.71) on
;
page 119. This means that our potentials and A are expressed in ` , ! Since we can
only know the fate of our sources at ` , and earlier, we therefore need to express all
our operators in ` , .
We introduce the convention that an index x or ` on the operators means that
they should be applied at constant x and ` , respectively. With this convention, we
find that
@ @ +
^ x 0 x, ' x 0 x, ^ + ' x 0 x, - J u
B E { { @ @ J B E x 0 x, - 0 { @ @ (8.91)
^` , x x 0 x, ^` , x x 0 x,
+
Furthermore, by applying the operator ^ _ ^` - x to Equation (8.89) on the preceding
page we find that
@ @ @ @
^` , ' ^ x 0 x, ' ^ x 0 x, ^` ,
10 B E B E { B E { P B E
^` x ^` x O ^` , x ^` x
+
' x 0 x, - J u ^` ,
0 { @ @ B E (8.92)
x 0 x, ^` x
which we can rearrange into

@ @ @ @
^` , ' x 0 x, ' x 0 x,
B
^`
E @
x0
@
x, 0 + x 0 x, - J u_ { 9 (8.93)
x
Here 9 is the retarded relative distance given by Equation (8.71) on page 119. Mak-
ing use of this, we obtain the following operator relation
@ @
^ ' ^` , ^ ' x 0 x, ^
B
^`
E B
^`
E B
^` ,
E 9 B
^` ,
E (8.94)
x x x x
By applying { +
L - to Equation (8.89) on the preceding page we obtain
{ + ' + 'x 0 x, + +
0 L - ` , L - x 0 x, @ @ J L - x 0 x, -
x 0 x,
+
' x 0 x, x 0 x, - J u +
@ @ 0 @ @ L - `C, (8.95)
x 0 x, x 0 x,
from which we see that
+ ' x 0 x,
L - ` , 0 { 9 (8.96)
This gives the following operator relation when + L - is acting on an arbitrary func-
tion of / and ` , :
^ x 0 x, ^
+
L -
' +
L - A TS + L - ` , VU B E
' +
L - 0 { 9 B E (8.97)
^` , x ^` , x
With the help of the rules (8.97) and (8.94) we are able to replace ` by ` , in the
operations which we need to perform. We find, for instance, that
&
; + ; ' 1 ,
L L - L
4<>= 0
B 9 E
&
' , x 0 x, u x 0 x, ^ 9
0
4<>= 0
9 2
O
@
x 0 x,
@ 0 { 0 { 9 B
^` ,
E P (8.98a)
x
d &
^ ^ ' ^ 0 ,u
^`
A B
^`
E A
^`
B
4< 9 E
x
&
' , 9 ^ 9
4<K= 0
{ 2 3 9 O
x 0 x,

u̇ 0

x 0 x, u

B
^` ,
E P (8.98b)
x
Applying this to the calculation of the E field with the use of the Liénard-Wiechert
potentials, Equations (8.70) on page 119, we obtain
; ^ A
E+ `g4 x- '
0UL 0
^`
& + @ @ {
' , x 0 x, - 0 x 0 x, u _
4<>= 0
9 2
?
@
x 0 x,
@
+ @ @ { @ @
x 0 x, - 0 x 0 x, u _ ^ 9 x 0 x, u̇
0 { 9 B
^` ,
E 0 { 2 H (8.99)
x
In analogy with the uniform motion, we can introduce certain useful quantities
as depicted in Figure 8.4. During the arbitrary motion, we interpret x 0 x0 as the
radius vector of the field point relative to the virtual simultaneous position x0 . This
is the position the charged particle would have had if at ` , all external forces would
have been switched off so that the trajectory would have been a straight line in the
direction of the tangent at x, . The particle would, from that point onwards, have
continued with the constant velocity u + ` , - from x, to x0 .
In order to simplify expression (8.99) above further, we use the fact that
+ +
B
^
E
9 ' Ú 2 x 0 x, - J u
@ @
x 0 x, - J u̇
{ 0 0 { (8.100)
^` , x x 0 x,
and find that we can write the electric field from an arbitrarily moving particle at
+
` , 4 x, - is given by the expression
\ x ] x[ \ WXY x Z WX [ Z
^ u 0 u
WXE[FY x[ Z
x ] x[
WXY x Z
Figure 8.4. Signals which are observed at the field point JK L x M were generated at
the source point JKÑ¥ L x¥ M . After time KÑ¥ the particle, which moves with nonuniform
velocity, has followed a yet unknown trajectory. Extrapolating the trajectory
from (K ¥L x¥M , based on the velocity u JK ¥M , defines the “virtual simultaneous radius
vector” x0 .
& @ @
+ ' , + x 0 x, u Ú 2
E `4 x-
4<K= 9 3_ B x 0 x, - 0 { E B 10 { 2E
0 k l m n
Coulomb field when `ba
@
0
@ c
x 0 x, + x 0 x, u
A { 2 a B x 0 x, - 0 { E a u̇P (8.101)
k O l m n
Radiation field
The first part of the field, the velocity field tends to the ordinary Coulomb field
when Ú Ù 0 and does not contribute to the radiation. The second part of the field,
the acceleration field, is radiated into the far zone and is therefore also called the
radiation field.
In a similar way we can compute the magnetic field:
x 0 x, ^
B+ `4 x- '
Lba A +
L - a A ' +
L - a A0 { 9 a B E A
^` , x
&
' , x 0 x, x 0 x, ^
0 @
4<>= 0 { 2 2 x 0 9 x,
@ a u0 @
x 0 x,
@ a B
^`
E A (8.102)
x
where we made use of Equation (8.70) on page 119 and Formula (8.94) on page 123.
But, according to (8.98a),
&
x 0 x, + ; ' , x 0 x,
@
x 0 x,
@ { a L - @
4<K= 0 { 2 2 x 0 9 x,
@ a u (8.103)
so that
x0 x, ; ^
B+ `g4 x- ' @ @ { a 0
+
L - 0 B E AP
x0 x, O ^` x
' x0 x,
@ @ { a E+ `4 x- (8.104)
x0 x,
The radiation part of the electric field is obtained from the acceleration field in
Formula (8.101) on the preceding page as
Erad + `4 x- '
lim E+ `4 x-
x ~ x
& Ó } @ @
, x 0 x, u
' +
9 x 0 x, - ed
a
+
x 0 x, - 0 P a u̇ f
4<>= 0 { 2 3 O
{
&
' , + q½+
4<>= 0 { 2 3 9 x 0 x, - a x 0 x0 - a u̇r (8.105)
where in the last step we used Formula (8.75) on page 120.
Radiation for small velocities

{
When the charge moves at low speeds Ú _ 1 and Formula (8.71) on page 119
simplifies to
+
9 '
x 0 x, 0
x 0 x, - J u
x 0 x, 4 Ú
{
(8.106)
{

and Formula (8.75) on page 120
@ @
' x 0 x, u {
x 0 x0 x 0 x,0 { x 0 x, 4 Ú (8.107)
so that the radiation field Equation (8.105) can be approximated by

&
,
Erad + `4 x- '
{ 2 @ @3
+
x 0 x, - a
q½+
x 0 x, - a u̇r 4 Ú
{
(8.108)
4<K= 0 x 0 x,
from which we obtain, with the use of Formula (8.104) above, the magnetic field
&
,
Brad + `4 x- '
@ @2
q
u̇ a
+
x 0 x, -Cr 4 Ú
{
(8.109)
4< { 3 x 0 x,
It is interesting to note the close correspondence which exists between the non-
relativistic fields (8.108) and (8.109) and the electric dipole field Equations (8.49)
on page 114 if we introduce
'G& +
p1 , x, C` , - (8.110)
and at the same time make the transitions
& ' 2
, u̇ p̈1 Ù 0 pó 1 (8.111a)
'
x 0 x, x 0 x0 (8.111b)
The power flux in the far zone is described by the Poynting vector as a function
of Erad and Brad . We use the close correspondence with the dipole case to find that
it becomes
& 2+d 2
' 0 , u̇ - x 0 x,
S @ @ sin2 Ê @ @ (8.112)
16< 2 { x 0 x, 2 x 0 x,
where Ê is the angle between u̇ and x 0 x0 . The total radiated power (integrated
over a closed spherical surface) becomes
d & 2+ 2 & 2 2
' 0 , u̇ - ' , Ú˙
° { (8.113)
6< 6<K= 0 { 3
which is the Larmor formula for radiated power from an accelerated charge. Note
{
that here we are treating a charge with Ú but otherwise totally unspecified mo-
tion while we compare with formulae derived for a stationary oscillating dipole.
The electric and magnetic fields, Equation (8.108) on the preceding page and Equa-
tion (8.109) on the facing page, respectively, and the expressions for the Poynting
flux and power derived from them, are here instantaneous values, dependent on
the instantaneous position of the charge at x, + ` , - . The angular distribution is that
which is “frozen” to the point from which the energy is radiated.
Bremsstrahlung
An important special case of radiation is when the velocity u and the acceleration
u̇ are collinear (parallel or antiparallel) so that u a u̇ ' 0. This condition (for an
arbitrary magnitude of u) inserted into expression (8.105) on the preceding page
for the radiation field, yields
&
,
Erad + `4 x- '
9
{ 2 3
+
x 0 x, - a
q½+
x 0 x, - a u̇r 4 u g u̇ (8.114)
4<>= 0
from which we obtain, with the use of Formula (8.104) on the facing page, the
magnetic field
& @ @
, x 0 x,
Brad +
`4 x- '
{ 3 3 9
q
u̇ a
+
x 0 x, -:r 4 u g u̇ (8.115)
4<K= 0
{
The difference between this case and the previous case of Ú is that the ap-
9
proximate expression (8.106) on the preceding page for is not valid; we must
h5i 0 j 5k
h5i h5i 0 j 25k

0
u
Figure 8.5. Polar diagram of the energy loss angular distribution factor
sin2 lnm J 1 É ` cos lnmpo M 5 during bremsstrahlung for particle speeds `rq 0,
`sq 0 t 25o , and `bq 0 t 5o .
instead use the correct expression (8.71) on page 119. The angular distribution of
the power flux (Poynting vector) therefore becomes
2z ˙ 2 sin2 x ~ xx
Su
y
w
v 2| } x 0

16{ x ~ xx } 2 1 Q
~ cos 6 }x ~ xx } (8.116)
It is interesting to note that the magnitudes of the electric and magnetic fields are
the same whether u and u̇ are parallel or antiparallel.
We must be careful when we compute the energy (S integrated over time). The
Poynting vector is related to the time when it is measured and to a fixed surface
in space. The radiated power into a solid angle element Ω, measured relative to
the particle’s retarded position, is given by the formula
2z ˙ 2
rad Ω u S x ~ xx x ~ xx Ω u v 0 nw x 2 |
sin2
Ω
8 16{ ~ cos 6
1Q
(8.117)
On the other hand, the radiation loss due to radiation from the charge at retarded
time x :
rad x
Ω u Ω rad
(8.118)
8 x x 8

Using Formula (8.94) on page 123, we obtain

S
Ω
x2 #

u
x
<

x2
x1
Figure 8.6. Location of radiation between two spheres as the charge moves with
velocity u from x1 to x2 during the time interval y¡ ¢£ .
¤¥ ¤¥ §® ¤
¤ ¤
¤8¦-rad
§ Ω ¨ ¤8¦rad ª © § ª Ω ¨ S¬ x« x Ω (8.119)
x« x ©
Inserting Equation (8.116) on the facing page for S into (8.119), we obtain the
explicit expression for the energy loss due to radiation evaluated at the retarded
time
¤¥ ® § 2³ 2
rad ¯ ¤ 0² ˙ sin2 ¯ ¤
¤8¦°§ Ω¨ ± µ ¶ Ω (8.120)
16´ 2
1 «Q· ¸ cos ¯ ¹
5
The angular factors of this expresssion, for three different particle speeds, is plotted
in Figure 8.5.
Comparing expression (8.117) on the preceding page with expression (8.120),
³ µ
§ ª they differ by a factor 1 « cos ¯ º which comes from the extra factor
we ª see that
º x « x introduced in (8.119). Let us explain this in geometrical terms.
© During the interval ¦ §B» ¦ §¼ ¤8¦ § ® and within the solid angle element ¤ Ω the particle
¤¥ ® ¤8¦ § ¾ ¤8¦ §
¦§
radiates an energy ½ rad ¯ º . As shown in Figure 8.6 § ® energy is at time
§ ¦ this
§ ¦ §¼ two
located between
¤8¦ § ® spheres,§ ¦ § ®one
¼ outer ¦ in¦ §x¼ 1 ¤8¦ § ® and
¤8¦ with its origin ¾ one ¦§®
¦ § ¤8with
¦ inner
its origin in x1 ¨ x1 u and radius µ ½ « ¨ µ « « .
From Figure 8.6 we see that the volume element subtending the solid angle ele-
ment
¤5¿
¤
Ω ¨ ª § ª2 (8.121)
x « x2
is
3À Á 5ÂÃ8Ä Á x ~ x2x 2 Ω 8Ä (8.122)

Here, 8Ä denotes the differential distance between the two spheres and can be eval-
uated in the following way
8Ä Á x ~ | 8 x ~ x ~ x2x u 8 x ~ x ~

x2x ÆÅ x2x
Ç } x ~ È xÉ 2x } Ê
z cos
Á | ~ x ~ x2x u 8 x Á |Ë
} x ~ x2x } } x ~ x2x } 8 x (8.123)
where Formula (8.71) on page 119 was used in the last step. Hence, the volume
element under consideration is
Ë
3À Á 5ÂÃ8Ä Á } Â | 8 x
x ~ x2x } Ì
(8.124)
We see that the energy which is radiated per unit solid angle during the time interval
xÍ x Å 8 x is located in a volume element whose size is dependent. This explains
the difference between expression (8.117) on page 128 and expression (8.120) on
the previous page.
Let the radiated energy, integrated over Ω, be denoted ˜ rad . After tedious, but
relatively straightforward integration of Formula (8.120) on the preceding page,
one obtains
˜ rad Á 0 wyx 2 z ˙ 2 1
v | Î (8.125)
8 x 6{ 2 3
1 ~ 2 Ï
If we know u x , we can integrate this expression over x and obtain the total
energy radiated during the acceleration or deceleration of the particle. This way
we obtain a classical picture of bremsstrahlung (braking radiation). Often, an
atomistic treatment is required for an acceptable result.
E XAMPLE 8.1
Ð B REMSSTRAHLUNG AT LOW SPEEDS AND SHORT ACCELERATION TIMES
Calculate the bremsstrahlung when a charged particle, moving at a non-relativistic speed,
is accelerated or decelerated during an infinitely short time interval.
We approximate the velocity change at time ÑÒ8ÓÔÑ 0 by a delta function:
u̇ ÕÑ ÒÖ Ó ∆u ×ØÕÑ Ò Ù Ñ 0Ö (8.126)
which means that
∆u ÓÛÚÝÜ u̇ ßØÑ (8.127)

Þ
Ü
Also, we assume à mpoâá 1 so that, according to Formula (8.71) on page 119,

ãåäçæ x Ù xÒ æ (8.128)
and, according to Formula (8.75) on page 120,
x Ù x0 ä x Ù xÒ (8.129)
From the general expression (8.104) on page 126 we conclude that E è B and that it
suffices to consider éëê æ Erad æ . According to the “bremsstrahlung expression” for Erad ,
Equation (8.114) on page 127,
Ò sin î
éìÓ 4ñ
í
ï ð 0 2 æx Ù
o xÒ æ ∆àå×ØÕÑ Ò Ù Ñ 0 Ö (8.130)
In this simple case òóê æ Brad æ is given by
òóÓ é o (8.131)
Fourier transforming expression (8.130) above for é is trivial, yielding
éõôöÓ í Ò sin î ôyú

8ï ð 0 o 2 æ x Ù
2 xÒ æ ∆àø÷ù 0
(8.132)
We note that the magnitude of this Fourier component is independent of û . This is a

consequence of the infinitely short “impulsive step” ×ØÕÑÒ Ù Ñ 0 Ö in the time domain which
produces an infinite spectrum in the frequency domain.
The total radiation energy is given by the expression
ü ˜ Ó Ú ß ü ˜ rad ß¢Ñ Ò Ó Ú Ü Ú¢ýöþ

Øß Ñ Ò E ÿ B
ß S Øß Ñ Ò
rad
Þ 0
Ú ýåÚ Ü é ò¡ß¢Ñ Ò ß 2 Ó
Ü
o Ú¢ýåÚ Þ Ü é 2 Øß Ñ Ò ß
1 1
Ó 2
0 Þ 0
Ó ð 0 o Ú ý Ú Ü é 2 ßØÑ Ò ß 2
Ü Ü
(8.133)
Þ
Ü
According to Parseval’s identity [cf. Equation (8.15) on page 107] the following equality
holds:
ÚÝÜ é 2 Øß Ñ Ò Ó 4ï<ÚÝÜ æ éâô æ 2 Øß û (8.134)

Þ 0
Ü
which means that the radiated energy in the frequency interval Õûû ßØû Ö is

ü˜
rad ßØû Ó 4ï5ð 0
o þ Ú ý æõ
é ô æ2 ß
2
ßØû (8.135)
For our infinite spectrum, Equation (8.132) on the previous page, we obtain
í Ò ï Õ 3∆ð à o 3 Ú ý æ sin î æ 2 ß 2 ßØû

ü˜ Ö2
ßØû
2 2
Ó 16
x Ù xÒ
rad
Ö
í Ò ï Õ 3∆ð à o 3 Ú ß Ú sin2 î sin î ß î ßØû
2 2
Ó 16
0 0
0
Ó 3ïñí ðÒ o þ ∆o à ßØû
2 2
(8.136)
0 2ï
ü
We see that the energy spectrum ˜ rad is independent of frequency û . This means that if
we integrate it over all frequencies û 0 Ö , a divergent integral would result.

In reality, all spectra have finite widths, with an upper cutoff limit set by the quantum
condition
û Ó 1
Õ ∆à Ö 2

(8.137)
2
which expresses that the highest possible frequency in the spectrum is that for which all
kinetic energy difference has gone into one single field quantum (photon) with energy û .

If we adopt the picture that the total energy is quantised in terms of ô photons radiated

during the process, we find that
ü˜
rad ßØû Ó ß (8.138)
û
or, for an electron where Ò Ó Ù æ ÷ æ , where ÷ is the elementary charge,
í
2 Øß û 1 2 ∆à 2 ßØû
÷
ß Ó 4ï5ð o 32ï þ ∆o à ä
2
0 û 137 3ï
þ o û (8.139)
where we used the value of the fine structure constant ÷ 2 m Õ 4ïñð 0 o Ö ä 1 m 137.
Even if the number of photons becomes infinite when û 0, these photons have negligible
energies so that the total radiated energy is finite.
E ND OF EXAMPLE 8.1
Cyclotron and synchrotron radiation

Formula (8.104) and Formula (8.105) on page 126 for the magnetic field and the
radiation part of the electric field are general, valid for any kind of motion of the
localised charge. A very important special case is circular motion, i.e., the case

u u̇.

With the charged particle orbiting in the À plane as in Figure 8.7, an orbit radius
"%$ - x( x . x&
x
u
",$ &%- x& (

* + !#"%$'&)(
u̇
0
Figure 8.7. Coordinate system for a particle in circular motion with velocity
u ÕÑÒ Ö along the tangent and constant acceleration u̇ ÕÑÒ Ö toward the origin at the
source point ÕÑ Ò xÒ Ö . The /
0 1
axes are chosen so that the relative field point vector
x Ù xÒ makes an angle with the axis which is normal to the plan of the orbital
motion. The radius of the orbit is . 2
3 , and an angular frequency 4 0, we obtain
5 x Á 4 x (8.140a)
3 6 7 ˆ cos 5
x Å 8 ˆ sin 5 x :9
0
xx x Á (8.140b)
u x Á ẋx x Á 3 4 0 6 ~ 7 ˆ sin 5 x Å 8 ˆ cos 5 x ; 9 (8.140c)
z Á } u} Á 3 4 0 (8.140d)
u̇ x Á ẍx x Á ~ 3 4 02 6 7 ˆ cos 5 x Å 8 ˆ sin 5 x ; 9 (8.140e)
z˙ Á } u̇ } Á 3 4 02 (8.140f)
We also see that we can express the relative vector x ~ xx as
x ~ xx Á x ~ xx 8 ˆ sin < Å = ˆ cos < (8.141)
where < is the angle between x ~ xx and the normal plane of the particle orbit.
From these expressions we obtain
x~ Á x ~ xx z sin < cos 5

xx u (8.142a)
x ~ xx u̇ Á ~ x ~ xx z ˙ sin < sin 5 (8.142b)
The power flux is given by the Poynting vector, which, with the help of For-
mula (8.104) on page 126, can be written
x ~ xx
Á E> B Á } E} 2
1 1
S | }x ~xx } (8.143)
v 0
v 0
Inserting this into Equation (8.119) on page 129, we obtain
rad < Í 5 Á }x ~ xx }Ë

} E} 2
8 x | (8.144)
v 0
where the retarded distance Ë is given by expression (8.71) on page 119. With the
radiation part of the electric field, expression (8.105) on page 126, inserted, and
using (8.142a) and (8.142b) above, one obtains, after some algebra, that
Î
5 5
< Í5 < ~ 2 Ï sin2 <
z 1 ~Q sin cos 2~ 2
sin2
Á v 0w x 2 ˙2 1
cos 5
rad
16{ 2 |

8 x 1~ sin < 5
(8.145)
5
The angles and vary in time during the rotation, so that refers to a moving
coordinate system. But we can parametrise the solid angle Ω in the angle and 5
<
the (fixed) angle so that Ω Á sin . Integration of Equation (8.145) over < < 5
this Ω, gives after some cumbersome algebra the angular integrated expression
˜ rad Á 0 w x 2 z ˙ 2 1
v | Î (8.146)
8 x 6{
1Q
~ 22 Ï
2
In Equation (8.145) above, two limits are particularly interesting:
1. z ?A
| @ 1 which corresponds to cyclotron radiation.
2. z ?| B 1 which corresponds to synchrotron radiation.

Cyclotron radiation For a non-relativistic speed z @ | , Equation (8.145) on the

facing page reduces to
< Í 5 Á v 0 wyx 2 z ˙ 2 1 ~ < 5

rad
sin2 sin2
16{ 2 |
(8.147)
8 x
But, according to Equation (8.142b) on the preceding page
sin2 < sin2 5 Á cos2 (8.148)
where is defined in Figure 8.7. This means that we can write
Á v 0 yw x 2 z ˙ 2 1 ~ cos2
2z 2
Á v 0 yw x ˙ sin2

rad
16{ 2 | 16{ 2 |
(8.149)
8 x
Consequently, a fixed observer near the orbit plane will observe cyclotron radi-
ation twice per revolution in the form of two equally broad pulses of radiation with
alternating polarisation.
Synchrotron radiation When the particle is relativistic, z | , the denominator B

< 5 C
D
< C ? 5EC
in Equation (8.145) on the facing page becomes very small if sin cos 1,
which defines the forward direction of the particle motion ( { 2Í 0).
Equation (8.145) on the preceding page then becomes
rad { ? 2 Í 0 Á v 0 w x 2 z ˙ 2 1
16{ 2 |
(8.150)
8 x ~
1Q
3
which means that an observer near the orbit plane sees a very strong pulse followed,
half an orbit period later, by a much weaker pulse.
The two cases represented by Equation (8.149) and Equation (8.150) above are
very important results since they can be used to determine the characteristics of the
particle motion both in particle accelerators and in astrophysical objects where a
direct measurement of particle velocities are impossible.
< ?
In the orbit plane ( Á { 2), Equation (8.145) on the facing page gives
Î
? 2Í 5 z

1~ 5 2~ 1~ 2 Ï
2
sin2 5
{ Á v 0 yw x 2 ˙2 cos
5
rad
8 x 16{ 2 | 1 ~Q cos 5
(8.151)
which vanishes for angles 5 0 which fulfil

"%$ - x( y
x . x&
∆

u
∆
"%$ &%- x& (
+ !#u̇"%$F&)(
0
0
Figure 8.8. When the observation point is in the plane of the particle orbit, i.e.,
ÓÔï m 2 the lobe width is given by ∆î .
z
cos 5 0
Á
| (8.152a)
z
sin 5 H
Á G 0 1~ |
2
2
(8.152b)
5
Hence, the angle 0 is a measure of the synchrotron radiation lobe width ∆ ; see
Figure 8.8. For ultra-relativistic particles, defined by
z
I Á J 1
1Í G 1~ |
2
@ 1Í (8.153)
K
2
1~ 2
2
one can approximate

z
5 C
0 sin 5 E
Á G 0 1~ |
2
2
Á
I
1
(8.154)
Hence, synchrotron radiation from ultra-relativistic charges is characterized by

a radiation lobe width which is approximately
∆ C I1 (8.155)
This angular interval is swept by the charge during the time interval
∆
∆ x Á 4 (8.156)
0
during which the particle moves a length interval
∆
L
∆ Á z ∆ x Á z
4 0
(8.157)
in the direction toward the observer who therefore measures a pulse width of length
∆ Á ∆ x ~
∆ L Á ∆ x ~
z ∆ x Î
Á 1~ Ï
z
∆ x Á
Î z
~ | Ï
∆
|

|

| 1
4 0
Î z z 2
C 1~ | Ï I4
1 Á 1 ~ 1z Å
I4 C
1
1~ | 2 I4
1
Ç ÈÉ Ê
Ç ÅÈ É | Ê
2
1? I
0 1 0 0
2 C 2
Á
I 4
1 1
(8.158)
2 3 0
As a general rule, the spectral width of a pulse of length ∆ is ∆ 1 ∆ . In the 4 B ?
ultra-relativistic synchrotron case one can therefore expect frequency components
up to
4 1 Á C
2 3 0 I 4 (8.159)
∆
max
M4
M C I
A spectral analysis of the radiation pulse will exhibit Fourier components 0 from
M Á 1 up to 2 3.
When N
electrons are contributing to the radiation, we can discern between
three situations:
1. All electrons are very close to each other so that the individual phase differ-
ences are negligible. The power will be multiplied by 2 relative to a single N
electron and we talk about coherent radiation.
2. The electrons are perfectly evenly distributed in the orbit. This is the case,
for instance, for electrons in a circular current in a conductor. In this case the
radiation fields cancel completely and no far fields are generated.
3. The electrons are unevenly distributed in the orbit. This happens for an open
ring current which is subject to fluctuations of order as for all open O N
systems. As a result we get incoherent radiation. Examples of this can be
found both in earthly laboratories and under cosmic conditions.
Radiation in the general case

We recall that the general expression for the radiation E field from a moving charge
concentration is given by expression (8.105) on page 126. This expression in Equa-
tion (8.144) on page 134 yields the general formula

Á v 0 yw x 2 } x ~ xx }
>RQ
}x ~ xx } u
S > u̇ T
2
P x~ xx x~ xx ~
rad
8 x 16{ 2 | Ë 5 |
(8.160)
Integration over the solid angle Ω gives the totally radiated power as
˜ rad Á 0 w x 2 z ˙ 2 1 Q ~ 2 sin2 U
2
v | Î (8.161)
8 x 6{
1Q ~ 22 Ï
3
U
where is the angle between u and u̇.
V U
In the limit u u̇, sin Á 0, which corresponds to bremsstrahlung. For u u̇,
U
sin Á 1, which corresponds to cyclotron radiation or synchrotron radiation.
The convection potential and the convection force

Let us consider in more detail the treatment of the radiation from a uniformly mov-
ing rigid charge distribution. We recall from Equation (8.82) on page 121 that
in this particular physical case of uniform velocity u, time and space derivatives
are closely related in the following way when they operate on a scalar function of
Í x :
XW
~ u ZY (8.162)

If we return to the original definition of the potentials and the inhomogeneous wave
equation, Formula (3.19) on page 36, for a generic potential component Ψ Í x and
a generic source component Í x , [
\ Í x Á | 2 2 ~ ] 2 Ψ Í x Á [ Í x
2
Ψ
1 2
(8.163)

we find that under the assumption that u Á z 7 ˆ 1 , this equation can be written
z 2 2Ψ
1~ | 2
2Ψ 2Ψ
Á ^~ [ x
À 21 Å À 22 Å À 23
(8.164)

i.e., in a time-independent form. Transforming

_1Á ` À 1
1 ~ z 2? | 2
(8.165a)
_2Á À 2
_3Á À 3 (8.165b)
(8.165c)
the time-independent equation (8.164) reduces to an ordinary Poisson’s equation
a] Ψ _ Á ~A[ J
2
1~ z 2 ?| _ Í_ Í_
2
1 2 3
(8.166)
which has as the well-known Coulomb potential solution
Ψ b Á 1 [ b x _ x
c d b ~ b x
3
(8.167)
4{
After inverse transformation back to the original coordinates, this becomes
Ψ x Á 1 [ xx 3À x
4{ cd Ë (8.168)
where, in the denominator,
Ë Á À 1 ~ À 1x 2 Å 1 ~ z 2 ? | 6À
2
2 ~ À 2x 2 Å À 3 ~ À 3x 2 9
1
2
(8.169)
Applying this to the explicit scalar and vector potential components, realising that
e
for a rigid charge distribution moving with velocity u, the current is given by
jÁ e u we obtain
f Í x Á
1 xx 3À x e
4{hg 0 c d Ë (8.170a)
ue xx 3À Á u f
A Í x Á
1
4{hg 0 | 2 c d Ë x | 2 Íx (8.170b)
For a localised charge where i e 3À x Á wnx , these expressions reduce to
f Í x Á
wnx
4{hg Ë 0
(8.171a)
A Í x Á wyx u
4{hg | Ë
2
(8.171b)
0
which we recognise as the Liénard-Wiechert potentials; cf. Equations (8.70) on

page 119. We notice, however, that the derivation here, based on a mathematical
technique which in fact is a Lorentz transformation, is of more general validity
than the one leading to Equations (8.70) on page 119.
Let us now consider the action of the fields produced from a moving, rigid charge
distribution represented by wnx moving with velocity u, on a charged particle w , also
moving with velocity u. This force is given by the Lorentz force
F Á w EÅ u > B (8.172)
The E and B fields are obtained from Formulae (8.80) on page 121 and the poten-
tials given by Equations (8.170) in the following way, making use of Formula (8.162)
f
on page 138:
Á ~^Y f ~ A Á ~^Y f ~ 1 u Á ~^Y f Å u u ZY f
E | 2 |2
Î u f u f
| j Y Ï | ~kY
Á (8.173a)
Î uf
B Á Yl> A Á Yl> Á Y f > u Á ~ u >mY f
|2 Ï |2 |2
Á u >on u Z Y f Ï u ~kY fqp Á u > E
Î
|2 | | |2 (8.173b)
where we used the fact that u u 0. > r

This means that we can rewrite expression (8.172) on the preceding page as
> p ZY f kY f >mY f Ï p
Î u Î u Î u
FÁ w EÅ u |2 E Ï Á w | n
u
Ï | ~ ~
u
>
| | n >
(8.174)
Applying the “bac-cab” rule, Formula (F.56) on page 159, on the last term yields
>Y f f Y f
Î Î z
> Ï Á | ZY
u u u u 2
| | Ï | ~ | 2
(8.175)
which means that we can write

F Á ~ w Y U (8.176)
where
U Á 1~
z 2 f
| 2 (8.177)
U
The scalar function is called the convection potential or the Heaviside poten-
tial. When the rigid charge distribution is well localised so that we can use the
potentials (8.171) the convection potential becomes
z
U Á 1~
2
wyx
| 2 4{ hg 0
Ë (8.178)
The convection potential from a point charge is constant on flattened ellipsoids of

revolution, defined through Equation (8.169) on the preceding page as
s À 1~ À x 2
` 1 ~ z 2 ? | 2 t Å À 2 ~ À 2x 2 Å À 3 ~ À 3x 2
1
Á I 2 À 1 ~ À x 2 Å À 2 ~ À x 2 Å À 3 ~ À x 2 Á Const (8.179)
1 2 3
U
These Heaviside ellipsoids are equipotential surfaces, and since the force is propor-
tional to the gradient of , which means it is perpendicular to the ellipsoid surface,
v{
y
uwv x
z
u ˆ
|~} ˆ
Figure 8.9. The perpendicular field of a moving charge.
the force between the two charges is in general not directed along the line which
connects the charges. A consequence of this is that a system consisting of two
comoving charges connected with a rigid bar, will experience a torque. This is the
idea behind the Trouton-Noble experiment, aimed at measuring the absolute speed
of the earth or the galaxy. The negative outcome of this experiment is explained
by the special theory of relativity which postulates that mechanical laws follow the
same rules as electromagnetic laws, so that a compensating torque appears due to
mechanical stresses within the charge-bar system.
Virtual photons
According to Formula (8.84) on page 121 and Figure 8.9
Á Á wx 1~
z 2
x0 =ˆ
4{hg 0
Ë 3 | 2 x~
Á wx
4{hg I 6 z 0 2 2Å 2 ?I 9 2 3 2
(8.180)
which represents a contracted field, approaching the field of a plane wave. The
passage of this field “pulse” corresponds to a frequency distribution of the field
8

energy. Fourier transforming, we obtain
Á 1
2{ c
Á
4{ 2
w
g 0
z Q zq 4 I 1
4
zq I S
(8.181)

Here, 1 is the Kelvin function (Bessel function of the second kind with imaginary
argument) which behaves in such a way for small and large arguments that
w 4 @ z I
4{ g z Í (8.182a)
0 4 zI
Í
2
0
(8.182b)
K length is of the order ? zqI .
showing that the “pulse”
the electric and magnetic fields,
Due to the equipartition of the field energy into
the total field energy can be written
Á g
ÀxÁ g z 8 2{ max
cd c c
2 3 2
0 0 (8.183)
min
where the volume integration is over the plane perpendicular to u. With the use
of Parseval’s identity for Fourier transforms, Formula (8.15) on page 107, we can
rewrite this as
Á g z c c } # } 4 2{
4 Á max
Ô 4{
c
2

0
0

min 0
C w z
4
2
2{ g c c
2
(8.184)
0 0

min
C w ln zI
from which we conclude that
2
2{ g z 2 0 4 min
(8.185)
where an explicit value of can be calculated in quantum theory only.

min

As in the case of bremsstrahlung, it is intriguing to quantise the energy into
|I
photons [cf. Equation (8.138) on page 132]. Then we find that
NXö 4 C 2{ < ln 4 4 4 (8.186)
where < Á ? 4{g | C 1 ? 137 is the fine structure constant.

min
2
0
Let us consider the interaction of two electrons, 1 and 2. The result of this
interaction is that they change their linear momenta from p1 to p1x and p2 to p2x ,
respectively. Heisenberg’s uncertainty principle gives min } p1 ~ p1x } so that ?
I
the number of photons exchanged in the process is of the order
N 4 C {

2 < Î | 4
p1 ~ px Ï
4
ln 1
4 (8.187)

Since this change in momentum corresponds to a change in energy 4 Á ~
Á I | , we see that 1 x
1
and 1 0
2<
} | p ~ | px } 4
C
NXö 4 { ln | ~ x 4 1
2
1 1
(8.188)
0 1 1
a formula which gives a reasonable account of electron- and photon-induced pro-
cesses.
8.3.4 Radiation from charges moving in matter

When electromagnetic radiation is propagating through matter, new phenomena
may appear which are (at least classically) not present in vacuum. As mentioned
earlier, one can under certain simplifying assumptions include, to some extent, the
influence from matter on the electromagnetic fields by introducing new, derived
field quantities D and H according to
D Á g Í x E Á g 0E (8.189)
BÁ
v Í x H Á m 0H
v
(8.190)
Expressed in terms of these derived field quantities, the Maxwell equations, often
called macroscopic Maxwell equations, take the form
Y D Á e Í x (8.191a)
Y> EÅ

B Á 0 (8.191b)
Y B Á 0 (8.191c)
Yl> H~

D Á j Í x (8.191d)
Assuming for simplicity that the electric permittivity and the magnetic per-
g
meability , and hence the relative permittivity and the relative permeability m
v
all have fixed values, independent on time and space, for each type of material we
consider, we can derive the general telegrapher’s equation [cf. Equation (2.23) on
page 25]
~Á 0 ~g
2E E 2E
(8.192)
v
2 v 2
(1D) wave
describing propagation
in a material medium.
In Chapter 2 we concluded that the existence of a finite conductivity, manifesting
itself in a collisional interaction between the charge carriers, causes the waves
to decay exponentially with time and space. Let us therefore assume that in our

medium Á 0 so that the wave equation simplifies to
Á 0~g
2E 2E
(8.193)
v 2
2

If we introduce phase velocity in the medium as
the
¡ Á 1 Á 1 Á ¢|
O g v O g 0 mv 0 O m (8.194)
where according to Equation (1.9) on page 5 | Á 1 ? O g

0 0 is the speed of light,
v
i.e., the phase speed of electromagnetic waves in vacuum, then the general solution
to each component of Equation (8.193)
£Á [ ~ ¡ ¥
Å ¤ Å ¡ Í ¦ Á 1Í 2Í 3 (8.195)

The ratio of the phase speed in vacuum and in the medium
| Á
¡ O ¢ m
Á | O gv r M def
(8.196)
is called the refractive index of the medium. In general is a function of both

M
time and space as are the quantities , , , and m themselves. If, in addition,
v
g
the medium is anisotropic or birefringent, all these quantities are rank-two tensor
fields. Under our simplifying assumptions, in each medium we consider Á Const M
for each frequency component of the fields.
Associated with the phase speed of a medium for a wave of a given frequency 4
we have a wave vector, defined as
r¨§ © ˆ Á § ª ˆ ¡ Á 4 v¡
k
def
¡ ¡ (8.197)
As in the vacuum case discussed in Chapter 2, assuming that E is time-harmonic,

i.e., can be represented by a Fourier component proportional to exp ~ , the « ¬ 4
solution of Equation (8.193) can be written

%
® ¯ ±°
EÁ E 0
kx
(8.198)
where now k is the wave vector in the medium given by Equation (8.197) above.
With these definitions, the vacuum formula for the associated magnetic field, Equa-
tion (2.30) on page 25,
B Á O g v ©ˆ > E Á 1 ©
¡ ˆ> E Á
4
1
k > E (8.199)
is valid also in a material medium (assuming, as mentioned, that has a fixed

³² M
constant scalar value). A consequence of a Á 1 is that the electric field will, in
general, have a longitudinal component.
It is important to notice that depending on the electric and magnetic properties
of a medium, and, hence, on the value of the refractive index , the phase speed in M
the medium can be smaller or larger than the speed of light:
~¡ Á | Á 4
M § (8.200)
where, in the last step, we used Equation (8.197) above.

If the medium has a refractive index which, as is usually the case, dependent on
4
frequency , we say that the medium is dispersive. Because in this case also k 4
4
and k), so that the group velocity
Á 4
g
§ (8.201)
has a unique value for each frequency component, and is different from . Except ¡
in regions of anomalous dispersion, is always smaller than | . In a gas of free ¡
charges, such as a plasma, the refractive index is given by the expression
M ´ Á 1~
4 2
4
2 p
2
(8.202)
where · ·
4 Á¶µ· Ng w ·
2
2
(8.203)
· ·
p
0
is the plasma frequency. Here and denote the mass and number dens- N
· ·
ity, respectively, of charged particle species . In an inhomogeneous plasma,
N Á N
x so that the refractive index and also the phase and group velocit-
ies are space dependent. As can be easily seen, for each given frequency, the phase
and group velocities in a plasma are different from each other. If the frequency 4
¡
is such that it coincides with p at some point in the medium, then at that point 4
Wl¸
while g
W
0 and the wave Fourier component at is reflected there. 4
Vavilov-Čerenkov radiation
As we saw in Subsection 8.3.3, a charge in uniform, rectilinear motion in vacuum
does not give rise to any radiation; see in particular Equation (8.84) on page 121.
Let us now consider a charge in uniform, rectilinear motion in a medium with elec-
tric properties which are different from those of a (classical) vacuum. Specifically,
consider a medium where
g Á Const ¹³g 0 (8.204a)

Á (8.204b)
v v 0
This implies that in this medium the phase speed is
¡ Á | Á 1 |
M O gv º 0
(8.205)
Hence, in this particular medium, the speed of propagation of (the phase planes of)
electromagnetic waves is less than the speed of light in vacuum, which we know is
an absolute limit for the motion of anything, including particles. A medium of this
kind has the interesting property that particles, entering into the medium at high
speeds } u } , which, of course, are below the phase speed in vacuum, can experience
that the particle speeds are higher than the phase speed in the medium. This is the
basis for the Vavilov-Čerenkov radiation that we shall now study.
If we recall the general derivation, in the vacuum case, of the retarded (and
advanced) potentials in Chapter 3 and the Liénard-Wiechert potentials, Equa-
tions (8.70) on page 119, we realise that we obtain the latter in the medium by
a simple formal replacement | | in the expression (8.71) on page 119 for Ë . ?M
Hence, the Liénard-Wiechert potentials in a medium characterized by a refractive
W
M
index , are
f
®M » ° ¯
wyx wyx
Í x Á Á
1 1
4{g 0
}
x ~ xx }~ x x u

4{g 0
Ë (8.206a)
® »°¯
A Í x Á
1 wyx u Á 1 wyx u
4{g 0
| 2 }
x ~ xx ~ M
} x x u

4{ hg 0
| 2 Ë (8.206b)
where now
x~ xx u
Ë Á x ~

xx ~ M | (8.207)
The need for the absolute value of the expression for Ë is obvious in the case when
? ½¼ ? M
z |
? ³@ ? M
z |
1 because then the second term can be larger than the first term; if
1 we recover the well-known vacuum case but with modified phase
speed. We also note that the retarded and advanced times in the medium are [cf.
Equation (3.34) on page 39]
rx et Á rx et Í x ~ xx Á ~ § }x ~ } Á
xx
~
} x ~ xx } M
4 | (8.208a)
x Á x Í x ~ xx Á Å § } x ~ xx } Á
Å
} x ~ xx } M
adv adv
4 | (8.208b)
so that the usual time interval ~ó x between the time measured at the point of
observation and the retarded time in a medium becomes
} x ~ xx } M
~ x Á | (8.209)
For z ? X
| ¼ 1 ? M , the retarded distance Ë , and therfore the denominators in Equa-
tions (8.206) vanish when
¾c ¿c u
Figure 8.10. Instantaneous picture of the expanding field spheres from a point
charge moving with constant speed à mpo 1 m in a medium where 1. This À Á ÁÀ
generates a Vavilov-Čerenkov shock wave in the form of a cone.
M x ~ xx | Á x ~
u
xx M z
cos c Á x ~ xx
| (8.210)
or, equivalently, when

Á |
cos c
M z (8.211)
In the direction defined by this angle c , the potentials become singular. During
the time interval ~ x given by expression (8.209) on the preceding page, the field
exists within a sphere of radius } x ~ xx } around the particle while the particle moves
a distance
L Á z ~ x (8.212)
along the direction of u.

In the direction c where the potentials are singular, all field spheres are tangent
to a straight cone with its apex at the instantaneous position of the particle and with
the apex half angle c defined according to<
|
sin c < Á cos c Á
M z (8.213)
speed | ?M
This cone of potential singularities and field sphere circumferences propagates with
in the form of a shock front, called Vavilov-Čerenkov radiation.1 The
Vavilov-Čerenkov cone is similar in nature to the Mach cone in acoustics.
1
The first observation of this radiation was made by P. A. Čerenkov in 1934, who was then a post-
graduate student in S. I. Vavilov’s research group at the Lebedev Institute in Moscow. Vavilov wrote
In order to make some quantitative estimates of this radiation, we note that we

can describe the motion of each charged particle wnx as a current density:
j Á w x u Â xx ~ u x Á w xz Â À x ~ z x Â x Â Ã x 7ˆ (8.214)
which has the trivial Fourier transform
j
Á wyx Ä » Â x Â Ã x 7 ˆ (8.215)
2{
This Fourier component can be used in the formulae derived for a linear current in
Subsection 8.3.1 if only we make the replacements
g W g Á M g 2
§ W M |4
0 0 (8.216a)
(8.216b)

In this manner, using j from Equation (8.215), the resulting Fourier transforms of
the Vavilov-Čerenkov magnetic and electric radiation fields can be calculated from
the expressions (8.4) and (8.5) on page 104, respectively.
The total energy content is then obtained from Equation (8.15) on page 107
(integrated over a closed sphere at large distances). For a Fourier component one
obtains [cf. Equation (8.18) on page 107]
Å ΩC
1

> k ¯ » À x Ω 2
4{hg M | c d
kx 3
j

rad
0
Á wyx M 4 4 x ~k§ À x cos À x sin Ω

À
±
Q ¬
2
2 2
16{ g | c z S
2
exp
3
0
3

(8.217)
where is the angle between the direction of motion, 7 ˆ x , and the direction to the
©
the spatial extent of the motion of the particle to the closed interval 6 ~Æ Í Æ 9 on the
observer, ˆ . The integral in (8.217) is singular of a “Dirac delta type.” If we limit
a manuscript with the experimental findings, put Čerenkov as the author, and submitted it to Nature.
In the manuscript, Vavilov explained the results in terms of radioactive particles creating Compton
electrons which gave rise to the radiation (which was the correct interpretation), but the paper was
rejected. The paper was then sent to Physical Review and was, after some controversy with the
American editors who claimed the results to be wrong, eventually published in 1937. In the same
year, I .E. Tamm and I. M. Frank published the theory for the effect (“the singing electron”). In fact,
predictions of a similar effect had been made as early as 1888 by Heaviside, and by Sommerfeld in
his 1904 paper “Radiating body moving with velocity of light”. On May 8, 1937, Sommerfeld sent a
letter to Tamm via Austria, saying that he was surprised that his old 1904 ideas were now becoming
interesting. Tamm, Frank and Čerenkov received the Nobel Prize in 1958 “for the discovery and the
interpretation of the Čerenkov effect” [V. L. Ginzburg, private communication].
Ç È axis we can evaluate the integral to obtain
É Å#Ê Ω ËÍÌ È M 4 sin Î sin ÑÓÒ 1 ÔÖÕZØ × cos Î Ù Ú ×ÝÛ Ü Ê Ω
2 2 2 2
rad
4Ï g Ð ÑÓÒ 1 Ô½ÕZØ × cos Î ÙÛ
3 3
(8.218)
2
0
×ÞÜ
which has a maximum in the direction Î c as expected. The magnitude of this
maximum grows and its width narrows as Æ ß
¸ . The integration of (8.218)
over Ω therefore picks up the main contributions from Î à Î c . Consequently, we
can set sin Î à sin Î c and the result of the integration is
2 2
É ˜ á Ë 2Ï â ã É á sin Î Ê Î Ë 2Ï â ã É áÊ ä Ô cos Î å

rad
0
rad æ çêè é 0
rad
ïñð ÚÛ
à Ì È ë ì í sin Î c â sin 1 ò ÕZ×ôØó õ ×Ýö Ê ê
2 1 2
2 2
2Ï Ð î ï÷ð 1 ò ÕZ×ôØ ó õ Û
2 3
(8.219)
2
×ö
0 1
ê
The integral (8.219) is strongly in peaked near ËøÔ Ð ù ä ë ú å , or, equivalently,
near cos Î c Ë Ð ù ä ë ú å so we can extend the integration limits to ûXü without
introducing too much error. Via yet another variable substitution we can therefore
approximate
ï÷ð 1 ò ÕZ×ôØÝó õ Ú Û ê Ç Ç
cos Î c â
1 sin2
× ö Ê à ý 1 Ô ë Ð ú þ Ð÷ì ÿ ë â
Ý 2
Ç Ê
sin2
î î
2
sin
ïñð 1 ò ÕZ×ôØ ó õ Û 2 2 2 2
×ö
1
Ë Ð÷ì ÿ ë Ï ý 1 Ô ë Ð ú þ
2
2 2
(8.220)
ä ì Óì ò Ê ì å
leading to the final approximate result for the total energy loss in the frequency
interval
É ˜ á#Ê ì Ë Ì È ÿ ý 1 Ô Ð ì Ê ì
2 2
rad
2Ï í Ð ë ú þ 0
2 2 2
(8.221)
As mentioned earlier, the refractive index is usually frequency dependent. Real-

ising this, we find that the radiation energy per frequency unit and per unit length
É #á Ê ì
is
Èì ý Ð ì
˜ rad
Ë Ï í Ð Ô ë äì åú þ Ê
Ì
2 2
ÿ 2 4 0 2
1 2 2
(8.222)
This result was derived under the assumption that 1 , i.e., under the ú ù Ð ù ë äì å
ive. For all media it is true that 1 when ë äì å ß
, so there exist always aì ß ü
condition that the expression inside the parentheses in the right hand side is posit-
149
ë
highest frequency for which we can obtain Vavilov-Čerenkov radiation from a fast
charge in a medium. Our derivation above for a fixed value of is valid for each
individual Fourier component.
B IBLIOGRAPHY 8
[1] Richard Becker. Electromagnetic Fields and Interactions. Dover Publications, Inc.,
New York, NY, 1982. ISBN 0-486-64290-9.
[2] Vitaliy Lazarevich Ginzburg. Applications of Electrodynamics in Theoretical Physics
and Astrophysics. Gordon and Breach Science Publishers, New York, London, Paris,
Montreux, Tokyo and Melbourne, Revised third edition, 1989. ISBN 2-88124-719-9.
[4] Jerry B. Marion and Mark A. Heald. Classical Electromagnetic Radiation. Academic
Press, Inc. (London) Ltd., Orlando, . . . , second edition, 1980. ISBN 0-12-472257-1.
ISBN 0-201-05702-6.
[6] Jack Vanderlinde. Classical Electromagnetic Theory. John Wiley & Sons, Inc., New
York, Chichester, Brisbane, Toronto, and Singapore, 1993. ISBN 0-471-57269-1.
151
152
A PPENDIX F
Formulae
F.1 The Electromagnetic Field
F.1.1 Maxwell’s equations
D Ë (F.1)

B Ë 0

E ËoÔ
(F.2)
B (F.3)

H Ë jò D (F.4)
Constitutive relations
D Ë íE (F.5)
H Ë
B
(F.6)
j Ë E (F.7)
PË í 0 E (F.8)
153
154 A PPENDIX F. F ORMULAE
F.1.2 Fields and potentials

Vector and scalar potentials
Ë
A
B

(F.9)
E Ë Ô Ô A (F.10)
Lorentz’ gauge condition in vacuum

A ò
1

Ð 2
0 (F.11)
F.1.3 Force and energy

Poynting’s vector
S
E
H (F.12)
Maxwell’s stress tensor

ò Ô 1 "! ! #
ä ò ! ! å (F.13)
2
F.2 Electromagnetic Radiation

F.2.1 Relationship between the field vectors in a plane wave
$ˆ E
B
Ð (F.14)
F.2.2 !-, from

The far fields , an extended source distribution
0
(Ô ' )+*
Û ä å . . â / 021 3 3 4 ) î * k 5 x jÛ k
x
B&% rad x
Ï
0
(F.15)
4 x
!-, , 0
)+* x 6 î
Û ä å
E&% rad x

4 Ïí Ð
'
0
. . ˆ â/ 01
x
3 3 4 ) * k 5 x
jÛ k (F.16)
F.2. E LECTROMAGNETIC R ADIATION 155
F.2.3 The far fields

from an electric dipole
!-, ,
ì )+*
Û ä å Ô . . pÛ&% 1 k
x
B7% rad x
4Ï
0
(F.17)
x
!-, ,
1 )+* x
Û ä å Ô
E7% rad x í
4Ï 0 x
. . ä pÛ&% 1 k å k (F.18)
F.2.4 The far fields

from a magnetic dipole
!-, ,
)+*
Û ä å Ô 4Ï . . ä mÛ å
x
B7% rad x
0
k k (F.19)
x
!-, ,
)+* x
EÛ7% ä x å
8 . . mÛ
4Ï í Ð
rad k (F.20)
0 x
F.2.5 The far fields from an electric quadrupole

!-, ,
ì )*
Û ä å ' . . ä k QÛ å k
x
B7% rad x
8Ï
0
(F.21)
x
!-, ,
' ) * . x.:9 k Q å k;
Û ä å
E7% rad x
8Ï í x
ä Û 0
k (F.22)
F.2.6 The fields from a point charge in arbitrary motion

> ú
E xå R? ý 1 Ô 4å
< R?
ä Ð þ ò äxÔ x
2 u̇
4Ï í Ð @ (F.23)
0 = 3 2 2
4x å E. ä x å .

B ä xå ä x Ô Ð x Ô x4 (F.24)
=
. x Ô x4 . Ô ä x Ô x4 å uÐ (F.25)
R?

ä x Ô x4 å Ô . x Ô x4 . uÐ (F.26)
4 . x Ô x4
.
ý þ (F.27)
x =
F.2.7 The fields from a point charge in uniform motion

ú
ä xå ý
Ô < 2
4Ï í = Ð þ
E 1 3 2
R0 (F.28)
0
u E ä xå

Bä x å
Ð 2
(F.29)
.R . Ô ý R u 2
= Ð þ
2 0
0 (F.30)
R
xÔ x (F.31)
0 0
F.3 Special Relativity

F.3.1 Metric tensor
H II
1 0 0 0
ACB D EFGF 0 Ô 1 0 0
0 0 Ô 1 0 J
(F.32)
0 0 0 Ô 1
F.3.2 Covariant and contravariant four-vectors

LK B CA B D K D (F.33)
F.3.3 Lorentz transformation of a four-vector

3 4B ΛD
B 3D
(F.34)
H II
B EF
Ô 0 0
ΛD GF Ô M0 M&N 0 0
0J
(F.35)
M&N
0
M00 1
0 1
O 1
1 Ô 2
(F.36)
M ú
N
Ð (F.37)
N
F.3. S PECIAL R ELATIVITY 157
F.3.4 Invariant line element

Q
P 1 Q
P 1SR
1= (F.38)
F.3.5 Four-velocity
B
ú B d3 ð uP õ (F.39)
d=
M M
F.3.6 Four-momentum
T B QU 0 P 2 ú B ä P p å (F.40)
F.3.7 Four-current density

VB ú B ð uP õ (F.41)
0
F.3.8 Four-potential
W B P Aå
ä (F.42)
F.3.9 Field tensor

Z H II
D Y W B EF 0Z Ô Ô P [ \ Ô P ][ \
X B D W Ô 3D GF [ P 0 \ Ô P Z
3B Ô J (F.43)
\ Ô P [ P Z
0
0
F.4 Vector Relations

å
Let x be the radius vector (coordinate
ä å
3 1 3 2 3 3 ^ 3 `_Yba and let . x . denote the magnitude (“length”) of x. Let fur-
vector), from the origin to the point
ä å ä åä
ther c x x ededed be arbitrary scalar fields and a x b x c x d x ededed arbitrary ä å ä å ä å ä å
vector fields.
N
The differential vector operator is in Cartesian coordinates given by
^ f
^ 6 ˆ 3 def
^ h
i
3
def
g (F.44)
1
6
where ˆ , j
1 2 3 is the j th unit vector and 6 ˆ 1 ^ 6 ˆ , 6 ˆ 2 ^ k ˆ , and 6 ˆ 3 ^ l ˆ . In
component (tensor) notation
can be written

m ý 3 3 3 þ
ý 3 _ a þ (F.45)
1 2 3
F.4.1 Spherical polar coordinates

Base vectors
nˆ sin o cos p 6 ˆ ò sin o sin p k ˆ ò cos o l ˆ (F.46)
qˆ cos o cos p 6 ˆ ò cos o sin p ks
ˆ r sin o l ˆ (F.47)
tˆ r sin p 6 ˆ ò cos p k ˆ (F.48)
Directed line element

1 3 6 ˆ 1 l S1 u n ˆ òuv1 o q ˆ ò#u 2
sin o 1p t ˆ (F.49)
Solid angle element

1 Ω
sin o 1 o 1p (F.50)
Directed area element

1 23 w ˆ 1 S 1 x n ˆ u 2 1 Ωn ˆ (F.51)
Volume element
1 33 1y 1SuC1x u 2 S1 uv1 Ω (F.52)
F.4. V ECTOR R ELATIONS 159
F.4.2 Vector formulae

General relations
a b
b a
ez {|"Qz7{ cos o (F.53)
a
b
r b
a
} !2z {~! 6 ˆ (F.54)
a ä b cå ä a bå c
(F.55)
a ä b cå
b a cå r c a bå
ä ä (F.56)
a

ä b cå ò b ä c aå ò c ä a bå 0 (F.57)
ä a bå ä c då a 9 b ä c då ; ä a cå ä b då r ä a då ä b cå (F.58)
ä a bå ä c då ä a b då c r ä a b cå d (F.59)
c å c ò c
ä (F.60)
N c a å a - N c N òc a

ä (F.61)

c a å c
a r a s c
ä (F.62)
a b å b
aå r a å
ä ä ä b (F.63)

ä a bå a ä bå r b ä aå ò ä b - å a r ä a - å b (F.64)

ä a bå a ä bå ò b ä
aå ò ä b - å a ò ä a - å b (F.65)
L c m 2 c (F.66)

s c 0 (F.67)

aå
ä 0 (F.68)

ä a å aå r m
ä 2
a (F.69)
Special relations
In the following k is an arbitrary constant vector.
x 3 (F.70)

x
0 (F.71)
. x. . x. (F.72)
x
ý . 1. þ r . x.
3
(F.73)
x x
ý . x.
x þ 3
r m 2 ý . 1. þ
x
4 ä xå (F.74)
>
ý . k. r k. . x
ý . 1. þ
x þ k @ x3x
(F.75)
>

k
ý . x. 3 þ @ r ý k. . 3x þ if . x . 0 (F.76)
x x
m ý . k. þ
2
k
m . 1 . þ r 4 k ä x å
2 ý (F.77)
x x

k a å k aå ò k
ä ä ä aå r ä k aå (F.78)
Integral relations
ä å
Let y x be the volume bounded by the closed surface x y . Denote the 3-
å ä å
dimensional volume element by 1 33 ^ d y and the surface element, directed along ä
the outward pointing surface normal unit vector w ˆ , by 1 S ^ d23 w ˆ . ä å
â / ä aå 1 3 1 S a
3
(F.79)
â / ä c å 1 3 1 Sc
3
(F.80)
â / ä aå 1 3 1 S a 3
(F.81)
If x ä å is an open surface bounded by the contour ä x å , whose line element is

1 l, then
160
F.4. V ECTOR R ELATIONS 161
c1 l 1 S s c (F.82)
a
1 l 1 S

ä a å (F.83)
B IBLIOGRAPHY F
[1] George B. Arfken and Hans J. Weber. Mathematical Methods for Physicists. Academic
Press, Inc., San Diego, CA . . . , fourth, international edition, 1995. ISBN 0-12-059816-
7.
[2] Philip M. Morse and Herman Feshbach. Methods of Theoretical Physics. Part I.
McGraw-Hill Book Company, Inc., New York, NY . . . , 1953. ISBN 07-043316-8.
ISBN 0-201-05702-6.
163
164
A PPENDIX M
Mathematical
Methods
M.1 Scalars, Vectors and Tensors
Every physical observable can be described by a geometric object. We will de-
scribe the observables in classical electrodynamics mathematically in terms of scal-
ars, pseudoscalars, vectors, pseudovectors, tensors or pseudotensors and will not
exploit differential forms to any significant degree.
A scalar describes a scalar quantity that may or may not be constant in time
and/or space. A vector describes some kind of physical motion due to vection
and a tensor describes the motion or deformation due to some form of tension.
However, generalisations to more abstract notions of these quantities are common-
place. The difference between a scalar, vector and tensor and a pseudoscalar,
pseudovector and a pseudotensor is that the latter behave differently under such
V
coordinate transformations which cannot be reduced to pure rotations.
Throughout we adopt the convention that Latin indices j 8 ededed run over the
range 1 2 3 to denote vector or tensor components in the real
Euclidean three-
dimensional (3D) configuration space 3 , and Greek indices ededed , which
are used in four-dimensional (4D) space, run over the range 0 1 2 3.
M.1.1 Vectors
Radius vector
Any vector can be represented mathematically in several different ways. One suit-
able representation is in terms of an ordered -tuple, or row vector, of the co-
ordinates 3 where is the dimensionality of the space under consideration. The
most basic vector is radius vector which is the vector from the origin to the point
of interest. Its -tuple representation simply enumerates the coordinates which
165
166 A PPENDIX M. M ATHEMATICAL M ETHODS
describe this point. In this sense, the radius vector from the origin to a point is
3 and the radius vector can be represented by
synonymous with the coordinates of the point itself.
In the 3D space 3 , we have
å 1 2 3. The coordinates 3 are scalar
the triplet 3 1 3 2 3 3 of coordinates 3 , j
ä
quantities that describe the position along the unit base vectors 6 ˆ which span 3 .
Therefore a representation of the radius vector in 3 is
f 3
6 ˆ 3 def
^ 6ˆ 3
x g (M.1)
1
where we have introduced Einstein’s summation convention (EΣ) which states that
a repeated index in a term implies summation over the range of the index in ques-
tion. Whenever possible and convenient we shall in the following always assume
EΣ and suppress explicit summation in our formulae. Typographically, we repres-
ent a 3D vector by a boldface letter or symbol in a Roman font.
Alternatively, we may describe the radius vector in component notation as fol-
lows:
3 def
^ ä 3 1 3 2 3 3
å ^ ä3 `_Yba å (M.2)
This component notation is particularly useful in 4D space where we can repres-
ent the radius vector either in its contravariant component form
3 B def
^ ä 3 0 3 1 3 2 3 3 å (M.3)
or its covariant component form
3 B def
^ ä 3 0 3 1 3 2 3 3
å (M.4)
The relation between the covariant and contravariant forms is determined by the
metric tensor (also known as the fundamental tensor) whose actual form is dictated
by the physics. The dual representation of vectors in contravariant and covariant
forms is most convenient when we work in a non-Euclidean vector space with
an indefinite metric. An example is Lorentz space 4 which is a 4D Riemannian
space. 4 is often utilised to formulate the special
B theory B of3 4relativity.
B 3 0 3 1 3 2 3 3
We note that for a change of coordinates 3 3 4 ß ä , due å
B
to a transformation from a system Σ to another system Σ 4 , the differential radius
vector 1 3 transforms
B as

31 4 B 3 3 4 D 1 3 D (M.5)
3 4 B
D
which follows trivially from the rules of differentiation of considered as func-
tions of four variables 3 .
M.1. S CALARS , V ECTORS AND T ENSORS 167
M.1.2 Fields
A field is a physical entity which depends on one or more continuous parameters.
Such a parameter can be viewed as a “continuous index” that enumerates the “co-
ordinates” of the field. In particular, in a field which depends on the usual radius
vector x of 3 , each point in this space can be considered as one degree of freedom
so that a field is a representation of a physical entity which has an infinite number
of degrees of freedom.
Scalar fields
We denote an arbitrary scalar field in 3 by
c ä xå c 3 3 3
ä 1 2 3
å ^ c ä3 å
def
(M.6)
This field describes how the scalar quantity c varies continuously in 3D 3 space.
In 4D, a four-scalar field is denoted
c ä 3 0 3 1 3 2 3 3 å ^ c ä3 B
def å (M.7)
which indicates that the four-scalar c depends on all four coordinates spanning
this space. Since a four-scalar has the same value at a given point regardless of
coordinate system, it is also called an invariant.
Analogous to B the transformation rule, Equation (M.5) on the facing page,
3 B for the
differential 1 3 , the
B 3 4B
transformation rule for the differential operator under ù
a transformation 3 ß
becomes
3D
3 4B 3 4B 3 D (M.8)
which, again, follows trivially from the rules of differentiation.
Vector fields
We can represent an arbitrary vector field a x in ä å 3 as follows:
ax ä å 6 ˆ z xå
ä (M.9)
In component notation this same vector can be represented as

z
ä xå ä z ä xå z ä xå z ä xåå
1 2 3
Qz 3
ä å (M.10)
In 4D, an arbitrary four-vector field in contravariant component form can be

represented as
z B 3 Då z 03 Dåz 13 Dåz 23 Dåz 33 D åå

ä ä ä ä ä ä (M.11)
or, in covariant component form, as

z B 3 Då z 3 Dåz 3 Dåz 3 Dåz 3 D åå
ä ä 0ä 1ä 2ä 3ä (M.12)
D3 zB z
where is the radius four-vector. Again, the relation between and B is de-
termined by the metric of the physical 4D system under consideration.
Whether an arbitrary -tuple fulfils the requirement of being an ( -dimen-
_ B properties
sional) contravariant vector or not, depends on its transformation
_ `
_ `
_
during
2 `_ 3 å
a change of coordinates. For instance, in 4D an assemblage 0 1
ä
constitutes a contravariant four-vector (or the contravariant components of a four-B
B
vector) if and only if, during a transformation from a system Σ with coordinates 3
to a system Σ4 with coordinates 3 4 , it transforms to the new system according to
the rule
B 3 4B D
_4 3 D _ (M.13)
B
i.e., in the same way as the differential coordinate element 1 3 transforms accord-
ing to Equation (M.5) on page 166.
The analogous requirement for a covariant four-vector is that it transforms, dur-
ing the change from Σ to Σ4 , according to the rule
3 D
_ B4 3 B _ D
4 (M.14)
3B
i.e., in the same way as the differential operator ù transforms according to
Equation (M.8) on the preceding page.
Tensor fields
ä å
We denote an arbitrary tensor field in 3 by A x . This tensor field can be repres-
ented in a number of ways, for instance in the following matrix form:
H
W
ä xåå W
ä xåå W
ä xåå J
ä å GE W W W ^ W ä 3 ! å
11 12 13
A x
W 21 ä xå W 22 ä xå W 23 ä xå def
(M.15)
31 ä x 32 ä x 33 äx
where, in the last member, we have again used the more compact component nota-
tion. Strictly speaking, the tensor field described here is a tensor of rank two.
particularly simple rank-two tensor in 3

A, with the following properties:
is the 3D Kronecker delta symbol
0 if j VV
1 if j
(M.16)
The 3D Kronecker delta has the following matrix representation

H
ä å GE 1 0 0
0 1 0J (M.17)
0 0 1
Another common and useful tensor is the fully antisymmetric tensor of rank 3,
also known as the Levi-Civita tensor
V
1 if j 8 is an even permutation of 1,2,3
} ! V
0 if at least two of j 8 are equal
r V
1 if j 8 is an odd permutation of 1,2,3
(M.18)
with the following further property

} !-} b ! r ` ! (M.19)
ë
ë
0 and a vector a tensor of rank 1. Consequently, the ë
In fact, tensors may have any rank . In this picture a scalar is considered to be
a tensor of rank
ë
2 may be represented by a two-dimensional
notation where a vector (tensor) is represented in its component form is called the
tensor notation. A tensor of rank
array or matrix whereas higher rank tensors are best represented in their component
forms (tensor notation).
In 4D, we have three forms of four-tensor fields of rank . We speak of: ë
W B B B ¡ 3 D å ,
A contravariant four-tensor field, denoted 1 2
ä
or a covariant four-tensor field, denoted W B B B ¡ ä 3 D å ,
or a mixed four-tensor field, denoted W BBC¢ B £ BCB ¡¢ ä 3 D å .
1 2
1 2
1
The 4D metric tensor (fundamental tensor) mentioned above is a particularly

important four-tensor of rank 2. In covariant component form we shall denote it
ACB D . This metric
four-vector
z B andtensor z
determines the relation between an arbitrary contravariant
its covariant counterpart B according to the following rule:
z B 3 ¤ å def D å
ä ^ A B D z ä3 ¤ (M.20)
This rule is sometimes called the “lowering of index.”

BD
From the above we see that ACB D “lowers” one index and that A “raises” one in-
dex of the tensor on which it operates. In particular, the “raising of index” analogue
of the “lowering of index” rule, Equation (M.20) on the previous page, is:
z B 3 ¤ å def BD å
ä ^ A z D ä3 ¤ (M.21)
ë
More generally, the following “lowering” and “raising” rules hold for arbitrary
rank mixed tensor fields:
BB ¢ B £ B ¢ BC£ ¢¥ B BC¡ ¢ 3 ¤ W BD ¢ BB ¢ £ BC¢bB ¥ ¡ 3 ¤
A D ¢BC¢ W 1 2 1
ä å 1 2
ä å
1
(M.22)
BCB ¢bB BC¢ £ B ¢ ¥B ¡ ä 3 ¤ å W BCB ¢ B £ BC ¢ B £ ¢¥ B D ¡¢ ä 3 ¤ å
1 2 1
A D¢B¢W 1 2
1
1 1 2
1 2
1
(M.23)
Successive lowering and raising of more than one index is achieved by a repeated
application of this rule. For example, a dual application of the lowering operation
on a rank 2 tensor in contravariant form yields
W B D A B A§¦ D W ¤ ¦
¤ (M.24)
i.e., the same rank 2 tensor in covariant form. This operation is also known as a
tensor contraction.
E XAMPLE M.1
¨ T ENSORS IN 3D SPACE
Consider the tetrahedron-like volume element © indicated in Figure M.1 on the facing page
analysis. Let ª S «ª 2¬ ˆ in Figure M.1 on the next page be a directed surface element of
of a solid, fluid, or gaseous body, whose atomistic structure is irrelevant for the present
this volume element and let the vector T® ˆ ª 2¬ be the force that matter, lying on the side of
ª 2¬ toward which the unit normal vector ˆ points, acts on matter that lies on the opposite
side of ª 2¬ . This force concept is meaningful only if the forces are short-range enough that
they can be assumed to act only in the surface proper. According to Newton’s third law,
this surface force fulfils
T¯ ® ˆ « ° T® ˆ (M.25)
Using (M.25) and Newton’s second law, we find that the matter of mass ± , which at a
given instant is located in © obeys the equation of motion
T® ˆ ª 2¬ ° T² ˆ 1 ª 2 ¬ ° T² ˆ 2 ª 2 ¬ ° T² ˆ 3 ª 2 ¬³ Fext « ± a (M.26)
where Fext is the external force and a is the acceleration of the volume element. In other
words
T® ˆ « ´ ² ³ ´ ² ³ ´ ² ³ ± 2
ª ¬ µ a° ± ¶
Fext
1T ˆ 1 2T ˆ 2 3T ˆ 3 (M.27)
· 3
ºˆ
¹· 2
· 2
· 1
Figure M.1. Terahedron-like volume element » containing matter.
Since both a and Fext ¼ ½ remain finite whereas ½ ¼e¾ 2Á

¿ À 0 as » À 0, one finds that in
this limit
Æ Æ
ÅÆ Ç3
TÂÄ
ˆ Ã T ÉLˆ ÊÌË TÉCˆ Ê (M.28)
1 È È
From the above derivation it is clear that Equation (M.28) is valid not only in equilibrium
but also when the matter in » is in motion.
Introducing the notation
ÆÎ Î
Í ÃÐÏ TÉ ˆ ÊÒÑ (M.29)
for the Ó th component of the vector TÉ ˆ Ê , we can write Equation (M.28) above in component
form as follows
Î Î Æ ÔÆ Î Æ Æ Î
Í Â ÃÐÏ TÂ Ñ Ã Å Æ Ç 3 Í Ë Í (M.30)
ˆ ˆ
1 È È
Using Equation (M.30), we find that the component of the vector T Â ˆ in the direction of an
arbitrary unit vector Õ ˆ is

Ö ® × « T ® ˆÚ Ø Õ Úˆ
ˆ ˆ Ý ÝÚÞ Úàß Ý ÝÚ Ú
ÙÚ Û 3 Ö ® ÙÚ Û3 Ù Ý Û 3 Ö
« ˆ ± « ´ ± ´ Ö ± « ˆ Ø T Ø Õ ˆ (M.31)
1 Ü
1 1
ÝÚ
Hence, the á th component of the vector T ²Lˆ â , here denoted
Ö
ã á th component of a tensor T. Note that Ö ® × is independent of, can the
be interpreted as the
particular coordinate
ˆ ˆ
system used in the derivation.
We shall now show how one can use the momentum law (force equation) to derive the
equation of motion for an arbitrary element of mass in the body. To this end we consider
a part © of the body. If the external force density (force per unit volume) is denoted by f
and the velocity for a mass element ª ± is denoted by v, we obtain
ª 3
¬ ³ T® ˆ ª 2¬
ª§äæå ç v ª ± « å ç f ª åéè (M.32)
The á th component of this equation can be written

Ú Ú Ú Ú Ý ÝÔÚ
ª ª± « 3
¬ ³ Ö ® ¬ ¬ ³ ´ Ö ª 2¬

åç ë ª åìè ˆ ª « å ç ë ª
2 3
å ç §ª ä2ê åìè (M.33)
where, in the last step, Equation (M.30) on the previous page was used. Setting ª ± «îíæª 3¬
and using the divergence theorem on the lastÝÚ term, we can rewrite the result as
Ú Ú Ö Ý
í ª ª 3¬
« ª 3
¬ ³ 3¬
å ç ìª ä2ê åç ë åç ï ¬ ª (M.34)
ï
Since this formula is validÝÚ for any arbitrary volume, we must require that
Ú Ú Ö Ý
í ªìª ä ê ° ë ° ï ¬ « 0 (M.35)
ï
or, equivalently Ú Ú Ú Ö ÝÝ Ú
í ï ê ä ³ í v Øeð ê ° ë ° ï ¬ « 0 (M.36)
Úòñ Ú
ï ï
Note that
ê ä is the rate of change with time of the velocity component ê
point x «ôï ó ¬ 1 õ ¬ ï 1 õ ¬ 3 ö .
at a fixed
E ND OF EXAMPLE M.1 ÷
M.1.3 Vector algebra

Scalar product
The scalar product (dot product, inner product) of two arbitrary 3D vectors a and
b in ordinary 3 space is the scalar number
aø b
6 ˆ z ø 6 ˆ 2{ù" 6 ˆ ø 6 ˆ ez {ù"
ez {ù"úz { (M.37)

where we used the fact that the scalar product 6 ˆ ø 6 ˆ is a representation of the
Kronecker delta
defined in Equation (M.16) on page 169. In Russian literature,
the scalar product is often denoted û ab ü .
In 4D space we define the scalar product of two arbitrary four-vectors
zþý {ÿý and
z ý { ý ý z { ý z { þý z ý {
in the following way
(M.38)
where we made use of the index “lowering” and “raising” rules (M.20) and (M.21).
The result is a four-scalar, i.e., an invariant which is independent on in which iner-
tial system it is measured.
2 ý ý ý ý
The quadratic differential form
(M.39)
i.e., the scalar product of the differential radius four-vector with itself, is an in-

ý and ý which
variant called the metric. It is also the square of the line element
distance between neighbouring points with coordinates

ý . is the
¨ S CALAR PRODUCT, NORM AND METRIC IN L ORENTZ SPACE E XAMPLE M.2

In 4 the metric tensor attains a simple form [see Equation (5.7) on page 56 for an example]
and, hence, the scalar product in Equation (M.38) can be evaluated almost trivially and
becomes
õ a õ b a b
0
0
0
0
(M.40)
The important scalar product of the radius four-vector with itself becomes
4
õ x õ x ! õ x " õ x

! $#
0

0
2 1 2 2 2 3 2 2
(M.41)
which is the indefinite, real norm of . The metric is the quadratic differential form
4 4
% # % % & % % % %
2 2 2 1 2 2 2 3 2
(M.42)
E ND OF EXAMPLE M.2 ÷
E XAMPLE M.3
' M ETRIC IN GENERAL RELATIVITY
In the general theory of relativity, several important problems are treated in a 4D spherical
polar coordinate system where the radius four-vector can be given as and !"( ) (!*("+
the metric tensor is
/0 243 798
, .- 1 2 50 6 0
)
0
*:
0 0 0
0 0 2
)0
(M.43)
0 0 0 2
sin2
""( ) (<*=("+ !"( ) (!*("+
where ; ; and > > . In such a space, the metric takes the form
% # & 2 3 % 2 6 % )
2 2 2 2 2
) * )
% sin %
2 2 2
* +
2
(M.44)
In general relativity the metric tensor is not given a priori but is determined by the Einstein
equations.
E ND OF EXAMPLE M.3 ?
Dyadic product
A tensor A û x ü can sometimes be represented in the dyadic form A û x ü a û xü b û xü . @
The dyadic notation with two juxtaposed vectors a and b is interpreted as an outer
product and this dyad is operated on by another vector c “from the right” and “from
the left” with a scalar (inner) product in the following two ways:
Aø c
def
@ ab ø c @
def
a û b ø cü (M.45a)
cø A
def
@ c ø ab @ û c ø aü b
def
(M.45b)
thus producing new vectors, proportional to a and b. In mathematics, a dyadic

product is often called tensor product and is frequently denoted a b. A
Vector product
B
The vector product or cross product of two arbitrary 3D vectors a and b in ordinary
3 space is the vector
c a C b
ED F GIHKJ GML H N ˆ F (M.46)
Here
D F GIH
is the Levi-Civita tensor defined in Equation (M.18) on page 169. Some-
times the vector product of a and b is denoted a b or, particularly in the Russian O
literature, ab . P Q
A spatial reversal of the coordinate system û 1 2 3 ü û 1 2 3ü R R R T ST ST

S S
ate system a a and b
R R
changes sign of the components of the vectors a and b so that in the new coordin-
T T
b, which is to say that the direction of an ordinary
vector is not dependent on the choice of directions of the coordinate axes. On the
other hand, as is seen from Equation (M.46) on the facing page, the cross product
vector c does not change sign. Therefore a (or b) is an example of a “true” vector,
or polar vector, whereas c is an example of an axial vector, or pseudovector.
A prototype for a pseudovector is the angular momentum vector and hence the
attribute “axial.” Pseudovectors transform as ordinary vectors under translations
and proper rotations, but reverse their sign relative to ordinary vectors for any co-
ordinate change involving reflection. Tensors (of any rank) that transform analog-
ously to pseudovectors are called pseudotensors. Scalars are tensors of rank zero,
being the pseudoscalar ˆ ø û ˆ NF NG C NH

and zero-rank pseudotensors are therefore also called pseudoscalars, an example
ˆ ü . This triple product is a representation of the
UVXWcomponent of the Levi-Civita tensor
D F GIH
which is a rank three pseudotensor.
M.1.4 Vector analysis

The del operator
In
B 3 the del operator is a differential vector operator, denoted in Gibb’s notation
by Y and defined as
Y def@ N ˆ F Z Z F def@\[ (M.47)

F U
where N ˆ is the th unit vector in a Cartesian coordinate system. Since the operator
in itself has vectorial properties, we denote it with a boldface nabla. In “compon-
ent” notation we can write
ZF ] Z Z Z
Z 1S Z 2S Z 3^ (M.48)
In 4D, the contravariant component representation of the four-del operator is

defined by
Z ý ] Z Z S Z Z S Z Z S Z Z
3^
(M.49)
0 1 2
whereas the covariant four-del operator is
Zý ] Z Z Z Z
Z 0 S Z 1 S Z 2 S Z 3^ (M.50)
We can use this four-del operator to express the transformation properties (M.13)
and (M.14) on page 168 as

_ R ýa` Z R ýb _ (M.51)
` Z ýR b _
and
_ ýR (M.52)
respectively.
E XAMPLE M.4
' T HE FOUR - DEL OPERATOR IN L ORENTZ SPACE

In 4 the contravariant form of the four-del operator can be represented as
c dfe 1 c c ( gih fe 1 c c ( jkh (M.53)
and the covariant form as

c fe 1 c c ( ig h le 1 c c ( k
j h (M.54)
Taking the scalar product of these two, one obtains
c c 1 c c m $n
2
2 2
(M.55)
n
2 2
which is the d’Alembert operator, sometimes denoted , and sometimes defined with an
opposite sign convention.
With the help of the del operator we can define the gradient, divergence and curl
of a tensor (in the generalised sense).
The gradient
The gradient of an
B 3 scalar field o û xü , denoted Ypo û ü , is an B 3 vector field a û x ü :
Ypo û xü [qo û xü N ˆ F Z F o û xü a û xü (M.56)
From this we see that the boldface notation for the nabla and del operators is very
handy as it elucidates the 3D vectorial property of the gradient.
o
ý
In 4D, the four-gradient is a covariant vector, formed as a derivative of a four-
scalar field û ü , with the following component form:

Zýo û ü Z o û ý ü
Z (M.57)
' G RADIENTS OF SCALAR FUNCTIONS OF RELATIVE DISTANCES IN 3D E XAMPLE M.5

Very often electrodynamic quantities are dependent on the relative distance in between r 3
s t sut
two vectors x and x , i.e., on x x . In analogy with Equation (M.47) on page 175, we
can define the “primed” del operator in the following way:
vw c
j s ˆ c w
s gs (M.58)
Using this, the “unprimed” version, Equation (M.47) on page 175, and elementary rules of
differentiation, we obtain the following two very useful results:
v w c t x xsxt
j t x x s t ˆ cw
x xs
t x xs t
v w c t x xsut
ˆ c ws

j s t x xs t (M.59)
e e
j t x 1 xs t h t xx xxs s t j s t x 1 xs t h
3
(M.60)
The divergence
We define the 3D divergence of a vector field in
B 3 as
J F û xü
Yø a û xü [ ø N ˆ G J G û xü y F G Z F J G û xü
Z Z F J F û xü
Z F o û xü (M.61)
B
which, as indicated by the notation o û x ü , is a scalar field in 3 . We may think of
the divergence as a scalar product between a vectorial operator and a vector. As
is the case for any scalar product, the result of a divergence operation is a scalar.
J
ý is the following four-scalar:
Again we see that the boldface notation for the 3D del operator is very convenient.
The four-divergence of a four-vector

Z ý Jý û ü

Z ýJ ý û ü Z J ý û ü
Z ý (M.62)
E XAMPLE M.6
' D IVERGENCE IN 3D
For an arbitrary r 3
sz
vector field a x , the following relation holds:
e e
j s t xa xs{x s t h j t xs axxs s|t& } a xs j s t x 1 xs t h (M.63)
which demonstrates how the “primed” divergence, defined in terms of the “primed” del
operator in Equation (M.58) on the preceding page, works.
The Laplacian
The 3D Laplace operator or Laplacian can be described as the divergence of the
gradient operator:
Z F N ˆ F ø N ˆ GZ G y F G Z F Z G 3 2
~ Yø.Y Z F2 Z Z 2F @ F Z Z 2F
2
∆
Z Z
2
1
(M.64)
The symbol
~ 2 is sometimes read del squared. If, for a scalar field
o û xü , ~ 2
o 0
at some point in 3D space, it is a sign of concentration of o at that point.
E XAMPLE M.7
' THE LAPLACIAN AND THE DIRAC DELTA
A very useful formula in 3D r is 3
e e
j j t x 1 xs t h m t x 1 xs t h 4ï x xs
2
(M.65)

where x xsz is the 3D Dirac delta “function.”
E M.7 ? ND OF EXAMPLE
The curl
B
In 3 the curl of a vector field a û x ü , denoted Y C a û xü , is another B
3 vector field
b û x ü which can be defined in the following way:
D F I
G H F G
J H D F I
G H F Z JH û xü
N ˆ Z N
Y C a û xü
û xü ˆ
Z G b û xü (M.66)
where use was made of the Levi-Civita tensor, introduced in Equation (M.18) on
page 169.
ý
The covariant 4D generalisation of the curl of a four-vector field û ü is the
J
antisymmetric four-tensor field
ý û ü Z ý J û ü T Z J ý û ü T ý û ü (M.67)
A vector with vanishing curl is said to be irrotational.
' T HE CURL OF A GRADIENT E XAMPLE M.8

Using the definition of the r 3
curl, Equation (M.66) on the facing page, and the gradient,
Equation (M.56) on page 176, we see that

j jk x w ! v ˆ w c c x (M.68)

which, due to the assumed well-behavedness of x , vanishes:
w! v w c c w ! c c
ˆ x cc x v ˆ w
e c c
c c c c h x v ˆ
2 2
1
e c c
2 3 3 2
} ccc cc h x v ˆ

2 2
2
e
3 1
c
1 3

c c c c h x v ˆ
2 2
0 } 1 2 2 1
3
(M.69)
We thus find that

j jk x 0 (M.70)

for any arbitrary, well-behaved r 3
scalar field x .
In 4D we note that for any well-behaved four-scalar field 3
c c - c - c 3
0 (M.71)
so that the four-curl of a four-gradient vanishes just as does a curl of a gradient in r 3

.
Hence, a gradient is always irrotational.
' T HE DIVERGENCE OF A CURL E XAMPLE M.9

With the use of the definitions of the divergence (M.61) and the curl, Equation (M.66) on
the facing page, we find that
cw
j j ax ja a x w w ! c w c x (M.72)
Using the definition for the Levi-Civita symbol, defined by Equation (M.18) on page 169,

we find that, due to the assumed well-behavedness of a ,
c w w! c x c c w! c
w c

e c c
c c
2
c c h
2
x
1
e c c
2 3 3 2
} ccc c
c h x
2 2
2
e
3 1
c1 3

c c cc h x
2 2
0 } 1 2 2 1
3
(M.73)
i.e., that

j j ax 0 (M.74)

for any arbitrary, well-behaved r 3
vector field a x .
In 4D, the four-divergence of the four-curl is not zero, for
c M- . - c c - - 3 n 3
2
0 (M.75)
Numerous vector algebra and vector analysis formulae are given in Chapter F.
Those that are not found there can often be easily derived by using the compon-
ent forms of the vectors and tensors, together with the Kronecker and Levi-Civita
tensors and their generalisations to higher ranks. A short but very useful reference
in this respect is the article by A. Evett [3].
M.2 Analytical Mechanics

M.2.1 Lagrange’s equations
As is well known from elementary analytical mechanics, the Lagrange function or

Lagrangian is given by
F F ] F F ¢¡
û S ˙ S" ü S S" T¤£
^
(M.76)
F ¡
where is the generalised coordinate, the kinetic energy and £ the potential
energy of a mechanical system, The Lagrangian satisfies the Lagrange equations
Z ] Z F T Z F
Z Z ˙ ^ Z 0 (M.77)
define the to the generalised coordinate

F
¥ F We
according to
canonically conjugate momentum
¥ F Z F
Z ˙ (M.78)
and note from Equation (M.77) that

Z F ¥ ˙F
Z (M.79)
M.2.2 Hamilton’s equations

From , the Hamiltonian (Hamilton function) ¦ can be defined via the Legendre
transformation
¦ û ¥ F S F "S ü ¥ F ˙F T û F S ˙F "S ü (M.80)
After differentiating the left and right hand sides of this definition and setting them
equal we obtain
Z ¦ F ¥ F
Z ¦ F F
Z ¦ ˙F ¥ F
¥ F ˙F T Z F F T Z F ˙F T Z
Z¥ Z Z Z Z ˙ Z
(M.81)
F
According to the definition of ¥ , Equation (M.78) above, the second and fourth
terms on the right hand side cancel. Furthermore, noting that according to Equa-
T ¥ F F
tion (M.79) the third term on the right hand side of Equation (M.81) above is equal
to ˙ and identifying terms, we obtain the Hamilton equations:
Z¦F F
Z¥ ˙F (M.82a)
F
Z¦F ¥T ˙F T ¥
Z
(M.82b)
181
B IBLIOGRAPHY M
[1] George B. Arfken and Hans J. Weber. Mathematical Methods for Physicists. Academic
Press, Inc., San Diego, CA . . . , fourth, international edition, 1995. ISBN 0-12-059816-
7.
[2] R. A. Dean. Elements of Abstract Algebra. Wiley & Sons, Inc., New York, NY . . . ,
1967. ISBN 0-471-20452-8.
[3] Arthur A. Evett. Permutation symbol approach to elementary vector analysis. Amer-
ican Journal of Physics, 34, 1965.
[4] Philip M. Morse and Herman Feshbach. Methods of Theoretical Physics. Part I.
McGraw-Hill Book Company, Inc., New York, NY . . . , 1953. ISBN 07-043316-8.
[5] Barry Spain. Tensor Calculus. Oliver and Boyd, Ltd., Edinburgh and London, third
edition, 1965. ISBN 05-001331-9.
183
I NDEX
acceleration field, 125 component notation, 166

advanced time, 38 concentration, 178
Ampère’s law, 5 conservative field, 11
Ampère-turn density, 95 conservative forces, 81
anisotropic, 144 constitutive relations, 14
anomalous dispersion, 145 contraction, 56
antisymmetric tensor, 67 contravariant component form, 56, 166
associated Legendre polynomial, 112 contravariant field tensor, 68
associative, 59
contravariant four-tensor field, 169
axial gauge, 36
contravariant four-vector, 168
axial vector, 67, 175
contravariant four-vector field, 58
Biot-Savart’s law, 6 contravariant vector, 56
birefringent, 144 convection potential, 140
braking radiation, 130 convective derivative, 12
bremsstrahlung, 130, 138 cosine integral, 110
Coulomb gauge, 36
canonically conjugate four-momentum,
Coulomb’s law, 2
76
covariant, 53
canonically conjugate momentum, 76,
covariant component form, 166
181
covariant field tensor, 68
canonically conjugate momentum dens-
ity, 83 covariant four-tensor field, 169
characteristic impedance, 23 covariant four-vector, 168
charge density, 4 covariant four-vector field, 59
classical electrodynamics, 8 covariant vector, 56
closed algebraic structure, 59 cross product, 174
coherent radiation, 137 curl, 178
collisional interaction, 143 cutoff, 132
complex notation, 27 cyclotron radiation, 134, 138
185
186 I NDEX
d’Alembert operator, 34, 63, 176 electromagnetodynamic equations, 15

del operator, 175 electromagnetodynamics, 16
del squared, 178 electromotive force (EMF), 11
differential distance, 57 electrostatic scalar potential, 31
differential vector operator, 175 electrostatics, 1
Dirac delta, 178 energy theorem in Maxwell’s theory,
Dirac-Maxwell equations, 15 96
dispersive, 145 equation of continuity, 9, 64
displacement current, 10 equation of continuity for magnetic
divergence, 177 monopoles, 15
dot product, 173 equations of classical electrostatics, 8
duality transformation, 16 equations of classical magnetostatics,
dummy index, 56 8
dyadic form, 174 Euclidean space, 60
Euclidean vector space, 57
E1 radiation, 114 Euler-Lagrange equation, 82
E2 radiation, 116 Euler-Lagrange equations, 83
Einstein equations, 174 Euler-Mascheroni constant, 110
Einstein’s summation convention, 166 event, 60
electric charge conservation law, 9
electric conductivity, 10 far field, 46
electric dipole moment, 113 far zone, 103
electric dipole moment vector, 91 Faraday’s law, 11
electric dipole radiation, 114 field, 167
electric displacement, 14 field Lagrange density, 84
electric displacement vector, 93 field point, 3
electric field, 3 field quantum, 132
electric field energy, 96 fine structure constant, 132, 142
electric monopole moment, 91 four-current, 63
electric permittivity, 143 four-del operator, 175
electric polarisation, 92 four-dimensional Hamilton equations,
electric quadrupole moment tensor, 91 77
electric quadrupole radiation, 116 four-dimensional vector space, 55
electric quadrupole tensor, 116 four-divergence, 177
electric susceptibility, 93 four-gradient, 176
electric volume force, 98 four-Hamiltonian, 76
electromagnetic field tensor, 68 four-Lagrangian, 74
electromagnetic potentials, 32 four-momentum, 62
electromagnetic scalar potential, 33 four-potential, 64
electromagnetic vector potential, 33 four-scalar, 167
I NDEX 187
four-tensor fields, 169 inertial reference frame, 53

four-vector, 58, 167 inertial system, 53
four-velocity, 62 inhomogeneous Helmholtz equation,
Fourier component, 22 37
Fourier transform, 36 inhomogeneous time-independent wave
functional derivative, 82 equation, 37
fundamental tensor, 56, 166, 169 inhomogeneous wave equation, 36
inner product, 173
Galileo’s law, 53 instantaneous, 127
gauge fixing, 36 interaction Lagrange density, 84
gauge function, 35 intermediate field, 49
gauge invariant, 35 invariant, 167
gauge transformation, 35 invariant line element, 57
Gauss’s law, 4 inverse element, 59
generalise four-coordinate, 76 irrotational, 4, 178
generalised coordinate, 76, 180
Gibb’s notation, 175 Kelvin function, 141
gradient, 176 kinetic energy, 81, 180
Green’s function, 37 Kronecker delta, 169
group theory, 59
group velocity, 145 Lagrange density, 81
Lagrange equations, 180
Hamilton density, 83
Lagrange function, 80, 180
Hamilton density equations, 84
Lagrangian, 80, 180
Hamilton equations, 76, 181
Laplace operator, 178
Hamilton function, 181
Hamilton gauge, 36 Laplacian, 178
Hamiltonian, 181 Larmor formula for radiated power,
Heaviside potential, 140 127
Helmholtz’ theorem, 34 law of inertia, 53
help vector, 111 Legendre polynomial, 112
Hertz’ method, 110 Legendre transformation, 181
Hertz’ vector, 111 Levi-Civita tensor, 169
homogeneous wave equation, 21, 22 Liénard-Wiechert potentials, 67, 119,
Huygen’s principle, 37 139
light cone, 58
identity element, 59 light-like interval, 58
in a medium, 146 line element, 173
incoherent radiation, 137 linear mass density, 81
indefinite norm, 57 linearly polarised wave, 25
induction field, 46 longitudinal component, 24
188 I NDEX
Lorentz boost parameter, 61 Minkowski equation, 76

Lorentz equations, 34 Minkowski space, 60
Lorentz force, 13, 97, 139 mixed four-tensor field, 169
Lorentz gauge, 36 mixing angle, 16
Lorentz gauge condition, 34, 64 momentum theorem in Maxwell’s the-
Lorentz space, 57, 166 ory, 98
Lorentz transformation, 55, 139 multipole expansion, 110, 113
M1 radiation, 116 near zone, 49

Mach cone, 147 Newton’s first law, 53
macroscopic Maxwell equations, 143 non-Euclidean space, 57
magnetic charge density, 15 non-linear effects, 10
magnetic current density, 15 norm, 57, 173
magnetic dipole moment, 94, 115 null vector, 58
magnetic dipole radiation, 116
observation point, 3
magnetic field, 6
Ohm’s law, 10
magnetic field energy, 96
one-dimensional wave equation, 25
magnetic field intensity, 95
outer product, 174
magnetic flux, 11
magnetic flux density, 6 Parseval’s identity, 107, 131, 142
magnetic induction, 6 phase velocity, 143
magnetic monopoles, 15 photon, 132
magnetic permeability, 143 physical measurable, 27
magnetic susceptibility, 95 plane polarised wave, 25
magnetisation, 94 plasma, 145
magnetisation currents, 94 plasma frequency, 145
magnetising field, 14, 95 Poisson’s equation, 138
magnetostatic vector potential, 32 Poissons’ equation, 31
magnetostatics, 5 polar vector, 67, 175
massive photons, 88 polarisation charges, 93
mathematical group, 59 polarisation currents, 94
matrix form, 168 polarisation potential, 111
Maxwell stress tensor, 98 polarisation vector, 111
Maxwell’s macroscopic equations, 15, positive definite, 60
95 positive definite norm, 57
Maxwell’s microscopic equations, 14 potential energy, 81, 180
Maxwell-Lorentz equations, 14 potential theory, 112
mechanical Lagrange density, 84 power flux, 96
metric, 166, 173 Poynting vector, 96
metric tensor, 56, 166, 169 Poynting’s theorem, 96
I NDEX 189
Proca Lagrangian, 88 skew-symmetric, 68

propagator, 37 skin depth, 27
proper time, 58 source point, 3
pseudoscalar, 165 space components, 57
pseudoscalars, 175 space-like interval, 58
pseudotensor, 165 space-time, 57
pseudotensors, 175 special theory of relativity, 53
pseudovector, 67, 165, 175 spherical Bessel function of the first
kind, 112
quadratic differential form, 57, 173 spherical Hankel function of the first
quantum mechanical nonlinearity, 3 kind, 112
radiation field, 46, 49, 125 spherical waves, 105
radiation fields, 103 super-potential, 111
radiation gauge, 36 synchrotron radiation, 134, 138
radiation resistance, 109 synchrotron radiation lobe width, 136
radius four-vector, 55
radius vector, 165 telegrapher’s equation, 25, 143
rank, 168 temporal dispersive media, 10
rapidity, 61 temporal gauge, 36
refractive index, 144 tensor, 165
relative electric permittivity, 98 tensor contraction, 170
relative magnetic permeability, 98 tensor field, 168
relative permeability, 143 tensor notation, 169
relative permittivity, 143 tensor product, 174
Relativity principle, 53 three-dimensional functional derivat-
relaxation time, 22 ive, 83
rest mass density, 84 time component, 57
retarded Coulomb field, 49 time-harmonic wave, 22
retarded potentials, 39 time-independent diffusion equation,
retarded relative distance, 119 23
retarded time, 38 time-independent telegrapher’s equa-
Riemannian metric, 57 tion, 26
Riemannian space, 56, 166 time-independent wave equation, 23
row vector, 165 time-like interval, 58
total charge, 91
scalar, 165, 177 transverse components, 24
scalar field, 59, 167 transverse gauge, 36
scalar product, 173
shock front, 147 vacuum permeability, 5
signature, 56 vacuum permittivity, 2
190 I NDEX
vacuum polarisation effects, 3

vacuum wave number, 23
Vavilov-Čerenkov radiation, 146, 147
vector, 165
vector product, 174
velocity field, 125
virtual simultaneous radius vector, 125
wave vector, 25, 144

world line, 60
Young’s modulus, 81
Yukawa meson field, 88

Electromagnetic Field Theory (Ebook)

Enviado por

Dados do documento

Descrição original:

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Electromagnetic Field Theory (Ebook)

Enviado por

Direitos autorais:

Formatos disponíveis

Bo Thidé

E LECTROMAGNETIC F IELD T HEORY

Released 14th December 1998 at 13:56

Downloaded from http://www.plasma.uu.se/CED/Book

Electromagnetic Field Theory

4 The Electromagnetic Fields 43

6 Interactions of Fields and Particles 73

7 Interactions of Fields and Matter 91

8 Electromagnetic Radiation 103

Example 8.1 Bremsstrahlung at low speeds and short accelera-

Base vectors . . . . . . . . . . . . . . . . . . . . . . . . 158

M Mathematical Methods 165

1.1 Coulomb interaction . . . . . . . . . . . . . . . . . . . . . . . . 2

5.1 Relative motion of two inertial systems . . . . . . . . . . . . . . 54

6.1 Linear one-dimensional mass chain . . . . . . . . . . . . . . . . . 80

8.1 Radiation in the far zone . . . . . . . . . . . . . . . . . . . . . . 105

M.1 Tetrahedron-like volume element of matter . . . . . . . . . . . . . 171

In an attempt to encourage participation by other scientists and students in the

Uppsala, Sweden B O T HIDÉ

Figure 1.1. Coulomb’s law describes

1.1.1 Coulomb’s law

1.1.2 The electrostatic field

I.e., Estat is an irrotational field.

1.2.1 Ampère’s law

1.2.2 The magnetostatic field

which expresses the small element - Bstat x

and the equations of classical magnetostatics

1. Electric charge is a conserved quantity and current is a transport of electric

1.3.1 Equation of continuity

1.3.2 Maxwell’s displacement current

1.3.3 Electromotive force

to an electromotive force (EMF) is present. In the presence of such a field EEMF ,

1.3.4 Faraday’s law of induction

An EMF is induced by all changes of the magnetic flux Φm . Let us therefore

(it is one of Maxwell’s equations) so that

1.3.5 Maxwell’s microscopic equations

1.3.6 Maxwell’s macroscopic equations

which is a partiular case (  f  2) of the general duality transformation

and analogously for the other two Dirac-Maxwell equations. QED ­

The invariance of the Dirac-Maxwell equations under the similarity transforma-

F e tan  (1.55a)

the Dirac-Maxwell equations reduce to the usual Maxwell equations.

 E cos  2  tan  sin w E cos   e

and so on for the other equations. QED ­

2.1 The wave equation

The quantity º ) 0  e is called the relaxation time of the medium in question. In

which is a time-independent wave equation for E, representing weakly damped

2.2 Plane waves

and Maxwell’s equations attain the form

with the solution

2.2.1 Telegrapher’s equation

2.2.2 Waves in conductive media

Hence, the solution of the time-independent telegrapher’s equation, Equation (2.31),

EÂ E0 å¨æ  ú'å¢óõô¬ú_æìö¬èêû (2.35)

where we have, in the last step, rewritten ß ò in the amplitude-phase form

Hence, both fields fall off by a factor 1 å at a distance $

2.3 Observables and averages

where * denotes complex conjugate.

3.1 The electrostatic scalar potential

3.2 The magnetostatic vector potential

An identification of terms allows us to define the static vector potential as

3.3 The electromagnetic scalar and vector

which is a partiular case ( f 2) of the general duality transformation

and analogously for the other two Dirac-Maxwell equations. QED

F e tan (1.55a)

E cos 2 tan sin w E cos e

and so on for the other equations. QED

The quantity º ) 0 e is called the relaxation time of the medium in question. In

EÂ E0 å¨æ ú'å¢óõô¬ú_æìö¬èêû (2.35)

> > x Ø x x Ø x (3.28)

a small volume around > Â x Ø x Â 0. Since

proportional to x Ø x , the second integral vanishes when x Ø x

where Â v . In the following, we shall make frequent use of these shorthand

ÃiÄ Å Æ ÇÉÈÊPËKÌ ÃiÄMÍ Å Æ Í ÇÉÈÊIËKÌ

ï ó çäï cos î â sin

@A-D3E! p F @A-B! p @ 3G! H p H : !

%Y % ) *1 %Y %Y Z 1 \3 [ %Y ] 1 _

I`%Y I %Y 0 0 (5.51)

subject to the condition ( ( 0 has the proper transformation properties (proper

@ } 3 M M ) Q } 3 M M b } 3 M } Q3 M P } (5.73b)

( C NM ( P C 3 M C P ( 0 3E@ 0 E3 ! !