Академический Документы
Профессиональный Документы
Культура Документы
Bo Thid
E LECTROMAGNETIC F IELD T HEORY
Bo Thid
Also available
Electromagnetic
Field Theory
Bo Thid
Swedish Institute of Space Physics
and
Department of Astronomy and Space Physics
Uppsala University, Sweden
AB
Contents
Preface
xi
1 Classical Electrodynamics
1.1 Electrostatics . . . . . . . . . . . . . . . . . . .
1.1.1 Coulombs law . . . . . . . . . . . . . .
1.1.2 The electrostatic field . . . . . . . . . . .
1.2 Magnetostatics . . . . . . . . . . . . . . . . . .
1.2.1 Ampres law . . . . . . . . . . . . . . .
1.2.2 The magnetostatic field . . . . . . . . . .
1.3 Electrodynamics . . . . . . . . . . . . . . . . .
1.3.1 Equation of continuity for electric charge
1.3.2 Maxwells displacement current . . . . .
1.3.3 Electromotive force . . . . . . . . . . . .
1.3.4 Faradays law of induction . . . . . . . .
1.3.5 Maxwells microscopic equations . . . .
1.3.6 Maxwells macroscopic equations . . . .
1.4 Electromagnetic Duality . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
1
2
2
5
5
6
8
8
9
10
11
14
14
15
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2 Electromagnetic Waves
2.1 The Wave Equations . . . . . . . . . . . . . . . .
2.1.1 The wave equation for E . . . . . . . . . .
2.1.2 The wave equation for B . . . . . . . . . .
2.1.3 The time-independent wave equation for E
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2.2
Plane Waves . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.1 Telegraphers equation . . . . . . . . . . . . . . . .
2.2.2 Waves in conductive media . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
23
24
24
24
25
26
28
29
30
ii
C ONTENTS
Electromagnetic Potentials
3.1 The Electrostatic Scalar Potential . . . . . . . . . . . . . . . .
3.2 The Magnetostatic Vector Potential . . . . . . . . . . . . . . .
3.3 The Electrodynamic Potentials . . . . . . . . . . . . . . . . .
3.3.1 Electrodynamic gauges . . . . . . . . . . . . . . . . .
Lorentz equations for the electrodynamic potentials . .
Gauge transformations . . . . . . . . . . . . . . . . .
3.3.2 Solution of the Lorentz equations for the electromagnetic potentials . . . . . . . . . . . . . . . . . . . . .
The retarded potentials . . . . . . . . . . . . . . . . .
Example 3.1 Electromagnetodynamic potentials . . . . .
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
35
35
36
36
38
38
39
.
.
.
.
40
44
44
45
Relativistic Electrodynamics
4.1 The Special Theory of Relativity . . . . . . . . . . . . . . . .
4.1.1 The Lorentz transformation . . . . . . . . . . . . . .
4.1.2 Lorentz space . . . . . . . . . . . . . . . . . . . . . .
Radius four-vector in contravariant and covariant form
Scalar product and norm . . . . . . . . . . . . . . . .
Metric tensor . . . . . . . . . . . . . . . . . . . . . .
Invariant line element and proper time . . . . . . . . .
Four-vector fields . . . . . . . . . . . . . . . . . . . .
The Lorentz transformation matrix . . . . . . . . . . .
The Lorentz group . . . . . . . . . . . . . . . . . . .
4.1.3 Minkowski space . . . . . . . . . . . . . . . . . . . .
4.2 Covariant Classical Mechanics . . . . . . . . . . . . . . . . .
4.3 Covariant Classical Electrodynamics . . . . . . . . . . . . . .
4.3.1 The four-potential . . . . . . . . . . . . . . . . . . .
4.3.2 The Linard-Wiechert potentials . . . . . . . . . . . .
4.3.3 The electromagnetic field tensor . . . . . . . . . . . .
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
47
47
48
49
50
50
51
52
54
54
54
55
57
59
59
60
62
65
.
.
.
.
67
67
67
68
70
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
iii
5.2
Other fields . . . . . . . . . . . . . . . . . . . . . . . . 82
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
6 Electromagnetic Fields and Matter
6.1 Electric Polarisation and Displacement . . . . . . . .
6.1.1 Electric multipole moments . . . . . . . . .
6.2 Magnetisation and the Magnetising Field . . . . . . .
6.3 Energy and Momentum . . . . . . . . . . . . . . . .
6.3.1 The energy theorem in Maxwells theory . .
6.3.2 The momentum theorem in Maxwells theory
Bibliography . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
85
85
85
88
90
90
91
93
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
95
97
99
101
104
104
104
106
.
.
.
.
.
.
.
.
.
.
.
.
.
107
107
108
111
114
114
118
120
121
122
123
125
127
132
.
.
.
.
.
.
.
.
.
.
.
.
.
.
iv
C ONTENTS
8.3.3
8.3.4
Vavilov-Cerenkov
radiation . . . . . . .
Bibliography . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
143
145
146
148
149
151
153
158
F Formulae
F.1 The Electromagnetic Field . . . . . . . . . . . . . . . . . . .
F.1.1 Maxwells equations . . . . . . . . . . . . . . . . . .
Constitutive relations . . . . . . . . . . . . . . . . . .
F.1.2 Fields and potentials . . . . . . . . . . . . . . . . . .
Vector and scalar potentials . . . . . . . . . . . . . .
Lorentz gauge condition in vacuum . . . . . . . . . .
F.1.3 Force and energy . . . . . . . . . . . . . . . . . . . .
Poyntings vector . . . . . . . . . . . . . . . . . . . .
Maxwells stress tensor . . . . . . . . . . . . . . . . .
F.2 Electromagnetic Radiation . . . . . . . . . . . . . . . . . . .
F.2.1 Relationship between the field vectors in a plane wave
F.2.2 The far fields from an extended source distribution . .
F.2.3 The far fields from an electric dipole . . . . . . . . . .
F.2.4 The far fields from a magnetic dipole . . . . . . . . .
F.2.5 The far fields from an electric quadrupole . . . . . . .
F.2.6 The fields from a point charge in arbitrary motion . . .
F.2.7 The fields from a point charge in uniform motion . . .
F.3 Special Relativity . . . . . . . . . . . . . . . . . . . . . . . .
F.3.1 Metric tensor . . . . . . . . . . . . . . . . . . . . . .
F.3.2 Covariant and contravariant four-vectors . . . . . . . .
F.3.3 Lorentz transformation of a four-vector . . . . . . . .
F.3.4 Invariant line element . . . . . . . . . . . . . . . . . .
F.3.5 Four-velocity . . . . . . . . . . . . . . . . . . . . . .
F.3.6 Four-momentum . . . . . . . . . . . . . . . . . . . .
F.3.7 Four-current density . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
159
159
159
159
159
159
160
160
160
160
160
160
160
160
161
161
161
161
162
162
162
162
162
162
162
163
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
F.3.8 Four-potential . . . . . . . . . . .
F.3.9 Field tensor . . . . . . . . . . . .
F.4 Vector Relations . . . . . . . . . . . . . .
F.4.1 Spherical polar coordinates . . . .
Base vectors . . . . . . . . . . .
Directed line element . . . . . . .
Solid angle element . . . . . . . .
Directed area element . . . . . .
Volume element . . . . . . . . .
F.4.2 Vector formulae . . . . . . . . . .
General vector algebraic identities
General vector analytic identities .
Special identities . . . . . . . . .
Integral relations . . . . . . . . .
Bibliography . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Appendices
163
163
163
163
163
164
164
164
164
164
164
165
165
166
166
159
M Mathematical Methods
M.1 Scalars, Vectors and Tensors
M.1.1 Vectors . . . . . . .
Radius vector . . . .
M.1.2 Fields . . . . . . . .
Scalar fields . . . . .
Vector fields . . . .
Tensor fields . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Example M.1 Tensors in 3D space .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
167
167
167
167
169
169
169
170
171
Dyadic product .
Vector product .
M.1.4 Vector analysis .
The del operator
.
.
.
.
Example M.6 The four-del operator in Lorentz space .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
178
179
179
179
180
vi
C ONTENTS
The divergence . . . . . . . . . . . . . . . . . .
Example M.8 Divergence in 3D . . . . . . . . . .
The Laplacian . . . . . . . . . . . . . . . . . . .
Example M.9 The Laplacian and the Dirac delta . .
The curl . . . . . . . . . . . . . . . . . . . . . .
Example M.10 The curl of a gradient . . . . . . . .
Example M.11 The divergence of a curl . . . . . .
M.2 Analytical Mechanics . . . . . . . . . . . . . . . . . . .
M.2.1 Lagranges equations . . . . . . . . . . . . . . .
M.2.2 Hamiltons equations . . . . . . . . . . . . . . .
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
182
182
182
182
183
183
184
185
185
185
186
List of Figures
1.1
1.2
1.3
Coulomb interaction . . . . . . . . . . . . . . . . . . . . . . . 3
Ampre interaction . . . . . . . . . . . . . . . . . . . . . . . . 6
Moving loop in a varying B field . . . . . . . . . . . . . . . . . 12
4.1
4.2
4.3
5.1
7.1
8.1
8.2
8.3
8.4
8.5
8.6
8.7
8.8
8.9
8.10
8.11
8.12
8.13
Linear antenna . . . . . . . . . . . . . . . . . . . . .
Electric dipole geometry . . . . . . . . . . . . . . . .
Loop antenna . . . . . . . . . . . . . . . . . . . . . .
Multipole radiation geometry . . . . . . . . . . . . . .
Electric dipole geometry . . . . . . . . . . . . . . . .
Radiation from a moving charge in vacuum . . . . . .
An accelerated charge in vacuum . . . . . . . . . . . .
Angular distribution of radiation during bremsstrahlung
Location of radiation during bremsstrahlung . . . . . .
Radiation from a charge in circular motion . . . . . . .
Synchrotron radiation lobe width . . . . . . . . . . . .
The perpendicular field of a moving charge . . . . . .
Vavilov-Cerenkov
cone . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
108
109
111
116
119
123
126
139
140
144
146
149
155
vii
Preface
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 R MAN , 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 undergraduate 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 classical 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.
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
xi
xii
P REFACE
B O T HID
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 electric 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. Clearly, this is in contradiction to electromagnetism on a truly microscopic scale, where charges and currents are known to
be spatially extended objects. However, the limiting processes used will yield
results which are correct on small as well as large macroscopic scales.
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 conservation 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 equationsthe Maxwell equations.
At the end of the chapter we study Diracs symmetrised form of Maxwells
equations by introducing (hypothetical) magnetic charges and magnetic currents into the theory. While not identified unambiguously in experiments yet,
magnetic charges and currents make the theory much more appealing for instance by allowing for duality transformations in a most natural way.
1.1 Electrostatics
The theory which describes physical phenomena related to the interaction between stationary electric charges or charge distributions in space is called electrostatics.1 For a long time electrostatics was considered an independent phys1 The
C LASSICAL E LECTRODYNAMICS
ical theory of its own, alongside other physical theories such as mechanics and
thermodynamics.
1
1
qq0
qq0 0
qq0 x x0
=
=
3
0
0
40 |x x |
40
40
|x x |
|x x0 |
(1.1)
where in the last step Equation (M.80) on page 181 was used. In SI units,
which we shall use throughout, the force F is measured in Newton (N), the
electric charges q and q0 in Coulomb (C) [= Ampre-seconds (As)], and the
length |x x0 | in metres (m). The constant 0 = 107 /(4c2 ) 8.8542 1012
Farad per metre (F/m) is the vacuum permittivity and c 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, electric charge in statcoulomb, and length in centimetres
(cm).
Estat lim
q0
F
q
(1.2)
1.1
E LECTROSTATICS
x x0
q0
x0
O
F IGURE 1.1: Coulombs law describes how a static electric charge q,
located at a point x relative to the origin O, experiences an electrostatic
force from a static electric charge q0 located at x0 .
Since the purpose of the limiting process is to assure that the test charge q does
not influence the field, the expression for Estat does not depend explicitly on q
but only on the charge q0 and the relative radius vector x x0 . 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 (1.1) and Equation (1.2) on the preceding page, and Formula (F.72)
on page 165, we find that the electrostatic field Estat at the field point x (also
known as the observation point), due to a field-producing electric charge q 0 at
the source point x0 , is given by
Estat (x) =
1
1
q0 0
q0
q0 x x 0
=
=
40 |x x0 |3
40
40
|x x0 |
|x x0 |
(1.3)
where in the last step Equation (M.80) on page 181 was used.
In the presence of several field producing discrete electric charges q0i , located at the points x0i , i = 1, 2, 3, . . . , respectively, in an otherwise empty space,
1 In
the preface to the first edition of the first volume of his book A Treatise on Electricity
and Magnetism, first published in 1873, James Clerk Maxwell describes this in the following,
almost poetic, manner [7]:
For instance, Faraday, in his minds eye, saw lines of force traversing all space
where the mathematicians saw centres of force attracting at a distance: Faraday
saw a medium 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.
C LASSICAL E LECTRODYNAMICS
the assumption of linearity of vacuum1 allows us to superimpose their individual E fields into a total E field
x x0i
1
Estat (x) =
q0i
(1.4)
x x0 3
40 i
i
If the discrete electric charges are small and numerous enough, we introduce the electric charge density located at x0 and write the total field as
1
x x0 3 0
1
1
d x =
(x0 )
(x0 )
d3x0
3
0
40 V 0
40 V 0
|x x0 |
|x x |
(1.5)
(x0 ) 3 0
1
=
d
x
40 V 0 |x x0 |
where we used Formula (F.72) on page 165 and the fact that (x0 ) does not depend on the unprimed coordinates on which operates. We emphasise that
Equation (1.5) above is valid for an arbitrary distribution of electric charges,
including discrete charges, in which case can be expressed in terms of one or
more Dirac delta distributions.
Since, according to formula Equation (M.90) on page 184, [(x)] 0
for any 3D 3 scalar field (x), we immediately find that in electrostatics
Estat (x) =
Estat (x) =
1
40
(x0 ) 3 0
d x =0
0
V 0 |x x |
(1.6)
(1.7)
1 In
1.2
M AGNETOSTATICS
1.2 Magnetostatics
While electrostatics deals with static electric charges, magnetostatics deals
with stationary electric currents, i.e., electric charges moving with constant
speeds, and the interaction between these currents. Let us discuss this theory
in some detail.
F(x) =
(1.8)
Here dl and dl0 are tangential line elements of the loops C and C 0 , respectively,
and, in SI units, 0 = 4 107 1.2566 106 H/m is the vacuum permeability. From the definition of 0 and 0 (in SI units) we observe that
107
1
(F/m) 4 107 (H/m) = 2 (s2 /m2 )
(1.9)
2
4c
c
which is a useful relation.
At first glance, Equation (1.8) above may appear unsymmetric in terms
of the loops and therefore to be a force law which is in contradiction with
Newtons third law. However, by applying the vector triple product bac-cab
Formula (F.54) on page 164, we can rewrite (1.8) as
0 0 =
1
0 JJ 0
dl
dl0
4 C C 0
|x x0 |
0 JJ 0
x x0
dl dl0
4 C C 0 |x x0 |3
F(x) =
(1.10)
Recognising the fact that 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
F(x) =
0 JJ 0
4
C C
x x0
dl dl0
3
0
|x x |
(1.11)
C LASSICAL E LECTRODYNAMICS
J
C
dl
x x0
dl0
x
C0
J0
x0
O
dBstat (x)
0 J 0 0 x x 0
dl
4
|x x0 |3
(1.12)
which expresses the small element dBstat (x) of the static magnetic field set
up at the field point x by a small line element dl0 of stationary current J 0 at
the source point x0 . The SI unit for the magnetic field, sometimes called the
magnetic flux density or magnetic induction, is Tesla (T).
If we generalise expression (1.12) to an integrated steady state electric
current density j(x), we obtain Biot-Savarts law:
x x0 3 0
0
0
j(x0 )
d x =
3
0
4 V 0
4
|x x |
j(x0 ) 3 0
0
=
dx
4
V 0 x x0
Bstat (x) =
V0
j(x0 )
1
d3x0
|x x0 |
(1.13)
1.2
M AGNETOSTATICS
where we used Formula (F.72) on page 165, Formula (F.60) on page 165, and
the fact that j(x0 ) does not depend on the unprimed coordinates on which
operates. Comparing Equation (1.5) on page 4 with Equation (1.13) on the
facing page, we see that there exists a close analogy between the expressions
for Estat and Bstat but that they differ in their vectorial characteristics. With this
definition of Bstat , Equation (1.8) on page 5 may we written
F(x) = J
C dl B
stat
(x)
(1.14)
0
4
j(x0 ) 3 0
d x =0
0
V0 x x
(1.15)
Bstat (x) =
1
d3x0
|x x0 |
(1.16)
In the first of the two integrals on the right hand side, we use the representation
of the Dirac delta function given in Formula (F.73) on page 165, and integrate
the second one by parts, by utilising Formula (F.59) on page 165 as follows:
1
d3x0
|x x0 |
V0
1
1
= x k 0 j(x0 ) 0
d3x0
0 j(x0 ) 0
d3x0
0 |
0|
0
0
x
x
|x
|x
V
V
k
1
1
= x k j(x0 ) 0
dS
0 j(x0 ) 0
d3x0
0
|x x0 |
S xk |x x |
V0
(1.17)
[j(x0 ) 0 ]0
Then we note that the first integral in the result, obtained by applying Gausss
theorem, vanishes when integrated over a large sphere far away from the localised source j(x0 ), and that the second integral vanishes because j = 0 for
C LASSICAL E LECTRODYNAMICS
stationary currents (no charge accumulation in space). The net result is simply
Bstat (x) = 0
V0
(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
(x)
0
stat
E (x) = 0
Estat (x) =
(1.19a)
(1.19b)
stat
(1.20a)
(x) = 0 j(x)
(1.20b)
Since there is nothing a priori which connects Estat directly with Bstat , we must
consider classical electrostatics and classical magnetostatics as two independent theories.
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 electric current is a transport
of electric charge. This fact manifests itself in the equation of continuity
and, as a consequence, in Maxwells 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 Faradays law of induction.
1.3
E LECTRODYNAMICS
(1.21)
which states that the time rate of change of electric charge (t, x) is balanced
by a divergence in the electric current density j(t, x).
V0
= 0 j(t, x) + 0
0 E(t, x)
t
0
4 t
V0
(t, x0 )0
1
d3x0
|x x0 |
(1.22)
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
40 t
V0
(t, x0 )0
1
d3x0 =
0
t
40
|x x |
= E(t, x)
t
V0
(t, x0 )
1
d3x0
|x x0 |
(1.23)
Later, we will need to consider this formal result further. The result is Maxwells
source equation for the B field
B(t, x) = 0 j(t, x) +
0 E(t, x)
t
(1.24)
where the last term 0 E(t, x)/t is the famous displacement current. This
term was introduced, in a stroke of genius, by Maxwell [6] in order to make
10
C LASSICAL E LECTRODYNAMICS
the right hand side of this equation divergence free when j(t, x) is assumed to
represent the density of the total electric current, which can be split up in ordinary conduction currents, polarisation currents and magnetisation currents.
The displacement current is an extra term which behaves like a current density
flowing in vacuum. As we shall see later, its existence has 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.
(1.25)
where is the electric conductivity (S/m). In the most general cases, for instance in an anisotropic conductor, is a tensor.
We can view Ohms law, Equation (1.25) above, as the first term in a Taylor
expansion of the law j[E(t, x)]. This general law incorporates non-linear effects
such as frequency mixing. Examples of media which are highly non-linear are
semiconductors 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 Ohms law with care. The conductivity is, in general, time-dependent
(temporal dispersive media) but then it is often the case that Equation (1.25) is
valid for each individual Fourier component of the field.
If the current is caused by an applied electric field E(t, x), this electric field
will exert work on the charges in the medium and, unless the medium is superconducting, there will be some energy loss. The rate at which this energy is
expended is 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 , Ohms law, Equation (1.25) above, takes the
form
j = (Estat + EEMF )
(1.26)
(Estat + EEMF ) dl
(1.27)
1.3
11
E LECTRODYNAMICS
CE
EMF
dl
(1.28)
E(t, x) dl =
d
dt
d
m (t, x)
dt
B(t, x) dS =
dS
B(t, x)
t
(1.29)
where m is the magnetic flux and S is the surface encircled by C 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(t, x) =
B(t, x)
t
(1.30)
which is valid for arbitrary variations in the fields and constitutes the Maxwell
equation which explicitly connects electricity with magnetism.
Any change of the magnetic flux m will induce an EMF. Let us therefore
consider the case, illustrated if Figure 1.3.4 on the following page, 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 wellknown operator formula
d
= +v
dt t
(1.31)
12
C LASSICAL E LECTRODYNAMICS
dS
B(x)
v
C
dl
B(x)
d
dt
B dS =
dS
B
(v )B dS
t S
(1.32)
During spatial differentiation v is to be considered as constant, and Equation (1.15) on page 7 holds also for time-varying fields:
B(t, x) = 0
(1.33)
(1.34)
1.3
13
E LECTRODYNAMICS
allowing us to rewrite Equation (1.32) on the facing page in the following way:
E(t, x) =
d
dt
EEMF dl =
S
B
dS
t
B dS
(1.35)
(B v) dS
EEMF dl =
B
dS
t
B
dS
t
(B v) dl
(1.36)
C
(EEMF v B) dl =
(1.37)
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) above
yields
(EEMF v B) =
B
t
(1.38)
(1.39)
(1.40)
(1.41)
Hence, we can conclude that for a stationary observer, the Maxwell equation
E =
B
t
(1.42)
14
C LASSICAL E LECTRODYNAMICS
(1.43a)
(1.43b)
(1.43c)
(1.43d)
In these equations (t, x) represents the total, possibly both time and space dependent, electric charge, i.e., free as well as induced (polarisation) charges,
and j(t, 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 Maxwells microscopic equations. Another name often used for them is the Maxwell-Lorentz equations. Together
with the appropriate constitutive relations, 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.
(1.44a)
H = H[t, x; E, B]
(1.44b)
1.4
15
E LECTROMAGNETIC D UALITY
Under certain conditions, for instance for very low field strengths, we may
assume that the response of a substance to the fields is linear so that
D = E
(1.45)
H = 1 B
(1.46)
i.e., that the derived fields are linearly proportional to the primary fields and
that the electric displacement (magnetising field) is only dependent on the electric (magnetic) field.
The field equations expressed in terms of the derived field quantities D and
H are
D = (t, x)
B
E =
t
B = 0
D
H =
+ j(t, x)
t
(1.47a)
(1.47b)
(1.47c)
(1.47d)
and are called Maxwells macroscopic equations. We will study them in more
detail in Chapter 6.
(1.48a)
(1.48b)
(1.48c)
(1.48d)
16
C LASSICAL E LECTRODYNAMICS
( B) 0 jm 0
t
(1.49)
where we used the fact that, according to Formula (M.94) on page 184, the
divergence of a curl always vanishes. Using (1.48c) to rewrite this relation, we
obtain the equation of continuity for magnetic monopoles
m
+ jm = 0
t
(1.50)
which has the same form as that for the electric monopoles (electric charges)
and currents, Equation (1.21) on page 9.
We notice that the new Equations (1.48) on the preceding page exhibit the
following symmetry (recall that 0 0 = 1/c2 ):
E cB
(1.51a)
cB E
e
(1.51b)
(1.51c)
c
e
cj j
(1.51d)
j cj
(1.51e)
(1.51f)
which is a particular case ( = /2) of the general duality transformation (depicted by the Hodge star operator)
?
E = E cos + cB sin
(1.52a)
c B = E sin + cB cos
(1.52b)
c = c cos + sin
(1.52c)
? m
(1.52d)
? e
= c sin + cos
? e
c j = cj cos + j sin
(1.52e)
? m
(1.52f)
j = cj sin + j cos
which leaves the symmetrised 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.
1.4
17
E LECTROMAGNETIC D UALITY
FARADAY S LAW AS
E XAMPLE 1.1
x
|
|x
1
0
m (x0 )
d3x0
=
4
|x x0 |
V0
Bstat (x) =
0
4
V
jm (x0 )
V
(x0 )
1
d3x0
|x x0 |
(1.53)
Estat ]
x x 0 3 0 0
dx =
4
|x x0 |3
Vj
0
(x0 )
1
d3x0
|x x0 |
(1.54)
Taking the curl of the latter and using the operator bac-cab rule, Formula (F.62) on
page 165, we find that
Estat (x) =
0
=
4
V0
jm (x0 ) 3 0
dx =
0
V0 x x
1
0
d3x0 +
[jm (x0 ) 0 ]0
0
4 V 0
|x x |
0
4
jm (x0 )2
1
d3x0
|x x0 |
(1.55)
Comparing with Equation (1.16) on page 7 for Estat and the evaluation of the integrals
there we obtain
Estat (x) = 0
Vj
(1.56)
We assume that Formula (1.54) is valid also for time-varying magnetic currents. Then,
with the use of the representation of the Dirac delta function, Equation (M.85) on
page 183, the equation of continuity for magnetic charge, Equation (1.50) on the facing page, and the assumption of the generalisation of Equation (1.53) above to timedependent magnetic charge distributions, we obtain, formally,
E(t, x) = 0
Vj
0
= 0 jm (t, x)
B(t, x)
t
0
4 t
V
0
(t, x0 )0
1
d3x0
|x x0 |
(1.57)
18
C LASSICAL E LECTRODYNAMICS
[cf. Equation (1.22) on page 9] which we recognise as Equation (1.48b) on page 15.
A transformation of this electromagnetodynamic result by rotating into the electric
realm of charge space, thereby letting jm tend to zero, yields the electrodynamic
Equation (1.48b) on page 15, i.e., the Faraday law in the ordinary Maxwell equations.
By postulating the indestructibility of an hypothetical magnetic charge, we have
thereby been able to replace Faradays experimental results on electromotive forces
and induction in loops as a foundation for the Maxwell equations by a more appealing
one.
E ND
E XAMPLE 1.2
OF EXAMPLE
1.1C
1
= (E cos + cB sin ) +
?E +
E sin + B cos
t
t
c
B
1 E
= 0 jm cos
cos + c0 je sin +
sin
t
c t
1 E
B
sin +
cos = 0 jm cos + c0 je sin
c t
t
= 0 (cje sin + jm cos ) = 0 ?jm
1
e
?B = ( E sin + B cos ) =
sin + 0 m cos
c
c0
= 0 ce sin + m cos = 0 ?m
?E = (E cos + cB sin ) =
(1.58)
(1.59)
(1.60)
1 ?E
1
1
?B 2
= ( E sin + B cos ) 2 (E cos + cB sin )
c t
c
c t
1 m
1 B
1 E
= 0 j sin +
cos + 0 je cos + 2
cos
c
c t
c t
(1.61)
1 E
1 B
2
cos
sin
c t
c t
1 m
= 0
j sin + je cos = 0 ?je
c
QED
E ND
OF EXAMPLE
1.2C
1.4
19
E LECTROMAGNETIC D UALITY
E XAMPLE 1.3
(1.62a)
(1.62b)
=
cos
cos
? m
= ce sin + ce tan cos = ce sin + ce sin = 0
1 e
1 e 2
? e
(j cos + je sin2 ) =
j
j = je cos + je tan sin =
cos
cos
? m
j = cje sin + cje tan cos = cje sin + cje sin = 0
? e
(1.63a)
(1.63b)
(1.63c)
(1.63d)
Hence, a fixed mixing angle, or, equivalently, a fixed ratio between the electric and
magnetic charges/currents, hides the magnetic monopole influence ( m and jm ) on
the dynamic equations.
We notice that the inverse of the transformation given by Equation (1.52) on page 16
yields
E = ?E cos c?B sin
(1.64)
(1.65)
Furthermore, from the expressions for the transformed charges and currents above, we
find that
?e
1 e
?E =
=
(1.66)
0
cos 0
and
?B = 0 ?m = 0
(1.67)
so that
E =
e
1 e
cos 0 =
cos 0
0
(1.68)
QED
E ND OF
EXAMPLE
1.3C
20
C LASSICAL E LECTRODYNAMICS
The invariance of Diracs symmetrised Maxwell equations under the similarity transformation 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. Such particles are referred to as dyons. 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.
E XAMPLE 1.4
T HE
and hence G
(1.69)
3,
(1.70)
is conserved. I.e.,
E 2 c2 B2 = Const
E B = Const
(1.71a)
(1.71b)
(1.72)
= E E c2 B B + ic(E B) + ic(B E)
= 0 + 0 + ic(E B) ic(E B) = 0
(1.73)
= E E + c2 B B ic(E B) + ic(B E)
= 0 + 0 ic(E B) ic(E B) = 2ic(E B)
(1.74)
E ND
OF EXAMPLE
1.4C
1.4
21
B IBLIOGRAPHY
D UALITY EXPRESSED IN
E XAMPLE 1.5
Expressed in the complex field vector, introduced in Example 1.4 on the facing page,
the duality transformation Equations (1.52) on page 16 become
?
(1.75)
while
?F 2 = ei G ei G = |F|2
(1.76)
G ?G = e2i G G
(1.77)
Furthermore, assuming that = (t, x), we see that the spatial and temporal differentiation of ?G leads to
?G
t ?G
= i(t )ei G + ei t G
(1.78a)
t
?G ?G = iei G + ei G
(1.78b)
?G ?G = iei G + ei G
(1.78c)
EXAMPLE
1.5C
Bibliography
[1] R. B ECKER, Electromagnetic Fields and Interactions, Dover Publications, Inc.,
New York, NY, 1982, ISBN 0-486-64290-9.
[2] W. G REINER, Classical Electrodynamics, Springer-Verlag, New York, Berlin,
Heidelberg, 1996, ISBN 0-387-94799-X.
[3] E. H ALLN, Electromagnetic Theory, Chapman & Hall, Ltd., London, 1962.
[4] J. D. JACKSON, Classical Electrodynamics, third ed., John Wiley & Sons, Inc.,
New York, NY . . . , 1999, ISBN 0-471-30932-X.
[5] L. D. L ANDAU AND E. M. L IFSHITZ, The Classical Theory of Fields, fourth revised English ed., vol. 2 of Course of Theoretical Physics, Pergamon Press, Ltd.,
Oxford . . . , 1975, ISBN 0-08-025072-6.
[6] J. C. M AXWELL, A dynamical theory of the electromagnetic field, Royal Society Transactions, 155 (1864).
22
C LASSICAL E LECTRODYNAMICS
2
Electromagnetic
Waves
(2.1a)
(2.1b)
(2.1c)
(2.1d)
and 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 (t, x) and current
distributions j(t, 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 these second order equations which, as we shall see are wave
equations, and then discuss the implications of them. We shall also show how
the B wave field can be easily calculated from the solution of the E wave
equation.
23
24
E LECTROMAGNETIC WAVES
( B) = 0 j + 0 E
t
t
t
(2.2)
(2.3)
(2.4)
and since EEMF = 0, Ohms law, Equation (1.26) on page 10, yields
j = E
(2.5)
E + 0 t E = 0
t
(2.6)
E 1 2 E
=0
t c2 t2
(2.7)
( E) = 0 E + 0 0 ( E) (2.8)
t
t
2.1
25
T HE WAVE E QUATIONS
which, with the use of Equation (F.67) on page 165 and Equation (2.1c) on
page 23 can be rewritten
( B) 2 B = 0
2
B
0 0 2 B
t
t
(2.9)
Using the fact that, according to (2.1c), B = 0 for any medium and rearranging, we can rewrite this equation as
B 1 2 B
=0
t c2 t2
2 B 0
(2.10)
This is the wave equation for the magnetic field. We notice that it is of exactly
the same form as the wave equation for the electric field, Equation (2.7) on the
facing page.
(2.11)
and insert this into Equation (2.7) on the preceding page. This yields
1 2
E0 (x)eit 2 2 E0 (x)eit
t
c t
1
= 2 E 0 (i)E0 (x)eit 2 (i)2 E0 (x)eit
c
1
2
2
= E 0 (i)E 2 (i) E =
c
2
E=0
= 2 E + 2 1 + i
c
0
2 E 0
(2.12)
2
c2
1+
i
E=0
(2.13)
2
E=0
c2
(2.14)
26
E LECTROMAGNETIC WAVES
(2.15)
(2.16)
k=
c
=
=
=
0 0 ck k
0
= R0
0 k
(2.17)
where in the last step we introduced the characteristic impedance for vacuum
R0 =
E XAMPLE 2.1
0
376.7
0
WAVE EQUATIONS IN
(2.18)
ELECTROMAGNETODYNAMICS
Derive the wave equation for the E field described by the electromagnetodynamic
equations (Diracs symmetrised Maxwell equations) [cf. Equations (1.48) on page 15]
e
(2.19a)
0
B
E =
0 jm
(2.19b)
t
m
B = 0
(2.19c)
E
+ 0 je
(2.19d)
B = 0 0
t
under the assumption of vanishing net electric and magnetic charge densities and in
the absence of electromotive and magnetomotive forces. Interpret this equation physically.
E =
Taking the curl of (2.19b) and using (2.19d), and assuming, for symmetry reasons, that
there exists a linear relation between the magnetic current density jm and the magnetic
2.1
27
T HE WAVE E QUATIONS
(2.20)
1/c2 ,
( E) = 0 jm
that
1 E
( B) = 0 m B
0 j e + 2
t
t
c t
= 0 m 0 e E +
2 E
1 E
E 1
0 e
2
c t
t c2 t2
(2.21)
= ( E) 2 E,
2 E 0 e +
m
c2
E 1 2 E
20 m e E = 0
t c2 t2
(2.22)
2 E + i0 e +
2
= 2 E +
c2
m
c2
E+
2
E 20 m e E
c2
1 0 m e
e + m /c2
E=0
+i
2
0
0
(2.23)
Realising that, according to Formula (2.18) on the preceding page, 0 /0 is the square
of the vacuum radiation resistance R0 , and rearranging a bit, we obtain the timeindependent wave equation in Diracs symmetrised electrodynamics
2 E +
2
c2
R20 m e
2
!"#
1+i
e + m /c2
R2
0 1 02 m e
*
%'&)(
E=0
(2.24)
From this equation we conclude that the existence of magnetic charges (magnetic
monopoles), and non-vanishing electric and magnetic conductivities would lead to
a shift in the effective wave number of the wave. Furthermore, even if the electric
conductivity vanishes, the imaginary term does not necessarily vanish and the wave
might therefore experience damping (or growth) according as m is positive (or negative) in
a perfect electric isolator. Finally, we note that in the particular case that
= R0 m e , the wave equation becomes a (time-independent) diffusion equation
2 E + i0 e +
m
c2
E=0
(2.25)
EXAMPLE
2.1C
28
E LECTROMAGNETIC WAVES
(2.26)
E
n
B
n
B
n
=0
=
(2.27a)
B
t
(2.27b)
=0
(2.27c)
= 0 j(t, x) + 0 0
E
E
= 0 E + 0 0
t
t
(2.27d)
= n
0 + 0 0
E
t
(2.28)
which simplifies to the first-order ordinary differential equation for the normal
component En of the electric field
dEn
+ En = 0
dt
0
(2.29)
(2.30)
This, together with (2.27a), shows that the longitudinal component of E, i.e.,
the component which is perpendicular to the plane surface is independent of
and has a time dependence which exhibits an exponential decay, with a decrement given by the relaxation time in the medium.
Scalar multiplying (2.27b) by n,
we similarly find that
0 = n
n
E
B
= n
t
(2.31)
or
n
B
=0
t
(2.32)
2.2
29
P LANE WAVES
From this, and (2.27c), we conclude that the only longitudinal component of
B must be constant in both time and space. In other words, the only non-static
solution must consist of transverse components.
=0
0
2
t c2 t2
(2.33)
=0
2 c2 t2
(2.34)
As is well known, each component of this equation has a solution which can
be written
Ei = f ( ct) + g( + ct),
i = 1, 2, 3
(2.35)
(2.36)
n = k
c
c
(2.37)
(2.38)
30
E LECTROMAGNETIC WAVES
E
= iB = ik n
E
(2.39)
k
1
1
n
E = k E = k E = 0 0 n
E
(2.40)
E
+
i
E
=
+ K2E = 0
0
0
0
2
2
(2.41)
where
K 2 = 0 0 2 1 + i
2
= 2
0
c
1+i
= k2 1 + i
0
0
(2.42)
where, in the last step, Equation (2.16) on page 26 was used to introduce the
wave number k. Taking the square root of this expression, we obtain
K=k
1+i
= + i
0
(2.43)
= (2 2 ) + 2i
0
(2.44)
or
2 = 2 k2
(2.45)
(2.46)
k2
20
2.2
31
P LANE WAVES
Squaring the latter and combining with the former, one obtains the second
order algebraic equation (in 2 )
k 4 2
420 2
2 (2 k2 ) =
(2.47)
+,,
,
=k
-
+,,
,
-
=k
1+
+1
(2.48a)
2
1+
(2.48b)
(2.49)
With the aid of Equation (2.40) on the facing page we can calculate the associated magnetic field, and find that it is given by
B=
1
1
1
K k E = ( k E)( + i) = ( k E) |A| ei
(2.50)
K = k 1+i
0
= 0 0 (1 + i)
. 1 i 0 /
=k i
0
= (1 + i)
20
1
2
k(1 + i)
0
2
20
(2.51)
From this analysis we conclude that when the wave impinges perpendicularly upon the medium, the fields are given, inside this medium, by
E0 = E0 exp
B0 = (1 + i)
0
exp i
2
0
( n
E0 )
2
0
t
2
0
(2.52a)
(2.52b)
32
E LECTROMAGNETIC WAVES
2
0
(2.53)
(2.55)
= F0 G0 + F0 G0 + F0 G0 e2it + F0 G0 e2it 4
43
(2.56)
1
= Re 1 F0 G0 + F0 G0 e2it 2
2
1
= Re 1 F0 eit G0 eit + F0 G0 e2it 2
2
1
= Re 1 F(t, x) G (t, x) + F0 G0 e2it 2
2
2.3
33
B IBLIOGRAPHY
Often in physics, we measure temporal averages (h i) of our physical observables. If so, we see that the average of the product of the two physical
quantities represented by F and G can be expressed as
1
1
hF Gi hRe {F} Re {G}i = Re 1 F G 2 = Re 1 F G 2
2
2
(2.57)
Bibliography
[1] J. D. JACKSON, Classical Electrodynamics, third ed., John Wiley & Sons, Inc.,
New York, NY . . . , 1999, ISBN 0-471-30932-X.
[2] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism,
second ed., Addison-Wesley Publishing Company, Inc., Reading, MA . . . , 1962,
ISBN 0-201-05702-6.
34
E LECTROMAGNETIC WAVES
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)
Taking the divergence of this and using Equation (1.7) on page 4, we obtain
Poissons equation
2 stat (x) = Estat (x) =
(x)
0
(3.2)
1
40
(x0 ) 3 0
d x +
|x x0 |
(3.3)
where the integration is taken over all source points x0 at which the charge
density (x0 ) is non-zero and is an arbitrary quantity which has a vanishing gradient. An example of such a quantity is a scalar constant. The scalar
function stat (x) in Equation (3.3) is called the electrostatic scalar potential.
35
36
E LECTROMAGNETIC P OTENTIALS
(3.4)
0
4
j(x0 ) 3 0
d x + a(x)
|x x0 |
(3.5)
where a(x) is an arbitrary vector field whose curl vanishes. From Equation (M.90) on page 184 we know that such a vector can always be written
as the gradient of a scalar field.
(3.6)
3.3
37
T HE E LECTRODYNAMIC P OTENTIALS
Inserting this expression into the other homogeneous Maxwell equation, Equation (1.30) on page 11, we obtain
E(t, x) =
(3.7)
A(t, x) = 0
t
(3.8)
A(t, x) = (t, x)
t
(3.9)
This means that in electrodynamics, E(t, x) can be calculated from the formula
E(t, x) = (t, x)
A(t, x)
t
(3.10)
and B(t, x) from Equation (3.6) on the preceding page. Hence, it is a matter
of taste whether we want to express the laws of electrodynamics in terms of
the potentials (t, x) and A(t, x), or in terms of the fields E(t, x) and B(t, x).
However, there exists an important difference between the two approaches: in
classical electrodynamics the only directly observable quantities are the fields
themselves (and quantities derived 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 (3.6) on the facing
page and Equation (3.10) above to calculate the fields or physical quantities
expressed in the fields.
Inserting (3.10) and (3.6) on the facing page into Maxwells equations (1.43)
on page 14 we obtain, after some simple algebra and the use of Equation (1.9)
on page 5, the general inhomogeneous wave equations
(t, x)
( A) =
t
0
2A
1
1
2 A 2 2 A + 2
= 0 j(t, x)
c t
c t
2 +
(3.11a)
(3.11b)
38
E LECTROMAGNETIC P OTENTIALS
1
(t, x)
2 A + 2
=
2
2
c t
t
c t
0
2
1 A
1
= 0 j(t, x)
2 A + A + 2
c2 t2
c t
(3.12a)
(3.12b)
1
=0
c2 t
(3.13)
because this condition simplifies the system of coupled equations Equation (3.12) above into the following set of uncoupled partial differential equa1 In
fact, the Dutch physicist Hendrik Antoon Lorentz, who in 1903 demonstrated the covariance of Maxwells equations, was not the original discoverer of this condition. It had been
discovered by the Danish physicist Ludwig Lorenz already in 1867 [4].
3.3
39
T HE E LECTRODYNAMIC P OTENTIALS
5
5
5
def
def
1 2
(t, x)
1 2
2 = 2 2 2 =
2
2
c t
c t
0
2
2
1
1 A
2 A = 2 2 2 A = 0 j(t, x)
2
2
c t
c t
(3.14a)
(3.14b)
Gauge transformations
We saw in Section 3.1 on page 35 and in Section 3.2 on page 36 that in electrostatics and magnetostatics we have a certain mathematical degree of freedom,
up to terms of vanishing gradients and curls, to pick suitable forms for the
potentials and still get the same physical result. In fact, the way the electromagnetic scalar potential (t, x) and the vector potential A(t, x) are related to
the physically observables gives leeway for similar manipulation of them
also in electrodynamics. If we transform (t, x) and A(t, x) simultaneously into
new ones 0 (t, x) and A0 (t, x) according to the mapping scheme
(t, x)
t
A(t, x) 7 A0 (t, x) = A(t, x) (t, x)
(t, x) 7 0 (t, x) = (t, x) +
(3.15a)
(3.15b)
+
=
t
t
t
t
t
B0 = A0 = A () = A
E0 = 0
(3.16a)
(3.16b)
where, once again Equation (M.90) on page 184 was used. Comparing these
expressions with (3.10) and (3.6) 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 Maxwells equations, invariant is called a gauge
transformation. 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.
40
E LECTROMAGNETIC P OTENTIALS
The potentials (t, x) and A(t, x) calculated from (3.11) on page 37, with an
arbitrary choice of A, can be further gauge transformed according to (3.15)
on the previous page. If, in particular, we choose A according to the Lorentz
condition, Equation (3.13) on page 38, and apply the gauge transformation
(3.15) on the resulting Lorentz potential equations (3.14) on the previous page,
these equations will be transformed into
1 2
2 +
2
2
c t
t
2
1 A
2 A
c2 t2
(t, x)
1 2
2 =
2
2
c t
0
2
1
2 = 0 j(t, x)
c2 t2
(3.17a)
(3.17b)
We notice that if we require that the gauge function (t, x) itself be restricted
to fulfil the wave equation
1 2
2 = 0
c2 t2
(3.18)
these transformed Lorentz equations will keep their original form. The set of
potentials which have been gauge transformed according to Equation (3.15)
on the preceding page with a gauge function (t, x) which is restricted to fulfil
Equation (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
A = 0.
The Coulomb gauge, defined by = 0, A = 0.
The temporal gauge, also known as the Hamilton gauge, defined by
= 0.
The axial gauge, defined by A3 = 0.
The process of choosing a particular gauge condition is referred to as gauge
fixing.
3.3
41
T HE E LECTRODYNAMIC P OTENTIALS
(one equation for and one equation for each of the three components of A).
Each of these four scalar equations is an inhomogeneous wave equation of the
following generic form:
(t, x) = f (t, x)
(3.19)
F 1 [ f (x)] f (t, x) =
def
f (x) eit d
(3.20a)
f (t, x) eit dt
(3.20b)
1
2
exists, and that the same is true for the generic potential component:
(t, x) =
(x) =
1
2
(x) eit d
(t, x) eit dt
(3.21a)
(3.21b)
Inserting the Fourier representations (3.20a) and (3.21a) into Equation (3.19)
above, and using the vacuum dispersion relation for electromagnetic waves
= ck
(3.22)
(3.23)
(3.24)
42
E LECTROMAGNETIC P OTENTIALS
and the solution of Equation (3.23) on the previous page which corresponds to
the frequency is given by the superposition
(x) =
(3.25)
(3.26)
Away from r = |x x0 | = 0, i.e., away from the source point x0 , this equation
takes the form
d2
(rG) + k2 (rG) = 0
dr2
(3.27)
ik|xx |
eikr
eikr
eik|xx |
e
+
C
+ C
C+
r
r
|x x0 |
|x x0 |
(3.28)
G( x x0 ) C +
1
1
+ C
,
0
|x x |
|x x0 |
x x0 0
(3.29)
The volume integrated Equation (3.24) on the preceding page can under this
assumption be approximated by
C+ + C 4
2
1
d3x0 + k2 C + + C 4
3
|x x0 |
0
3 0
1
d3x0
|x x0 |
(3.30)
( x x ) d x
3.3
43
T HE E LECTRODYNAMIC P OTENTIALS
In virtue of the fact that the volume element d3x0 in spherical polar coordinates
is proportional to |x x0 |2 , the second integral vanishes when |x x0 | 0. Furthermore, from Equation (M.85) on page 183, we find that the integrand in the
first integral can be written as 4(|x x0 |) and, hence, that
1
4
C+ + C =
(3.31)
Insertion of the general solution Equation (3.28) on the preceding page into
Equation (3.25) on the facing page gives
(x) = C +
f (x0 )
eik|xx | 3 0
d x + C
|x x0 |
f (x0 )
eik|xx | 3 0
dx
|x x0 |
(3.32)
(t, x) = C
f (x )
6
+ C
f (x0 )
6
|x x0 |
d d3x0
|
exp 7 i . t + k|xx
/98
(3.33)
|x x0 |
d d3x0
0
0
tret
= tret
(t, x x0 ) = t
k |x x0 |
|x x0 |
= t
c
0
k |x x0 |
|x x0 |
0
0
= t+
tadv = tadv (t, x x ) = t +
(3.34a)
(3.34b)
0 , x0 )
f (tret
d3x0 + C
|x x0 |
0 , x0 )
f (tadv
d3x0
|x x0 |
(3.35)
This is a solution to the generic inhomogeneous wave equation for the potential
components Equation (3.19) on page 41. We note that the solution at time t at
the field point x is dependent on the behaviour at other times t 0 of the source
at x0 and that both retarded and advanced t 0 are mathematically acceptable
solutions. However, if we assume that causality requires that the potential at
0 , x0 ), we must in
(t, x) is set up by the source at an earlier time, i.e., at (tret
44
E LECTROMAGNETIC P OTENTIALS
(t, x) =
(3.36a)
(3.36b)
Since these retarded potentials were obtained as solutions to the Lorentz equations (3.14) on page 39 they are valid in the Lorentz gauge but may be gauge
transformed according to the scheme described in subsection 3.3.1 on page 39.
As they stand, we shall use them frequently in the following.
E XAMPLE 3.1
E LECTROMAGNETODYNAMIC POTENTIALS
e
0
(3.37a)
E = 0 jm
B = 0
B
t
B = 0 j e + 0 0
(3.37b)
(3.37c)
E
t
(3.37d)
In this theory, one derives the inhomogeneous wave equations for the usual electric
scalar and vector potentials (e , Ae ) and their magnetic counterparts (m , Am ) by
assuming that the potentials are related to the fields in the following symmetrised
form:
e
A (t, x) Am
t
1
1
B = 2 m (t, x) 2 Am (t, x) + Ae
c
c t
E = e (t, x)
(3.38a)
(3.38b)
3.3
45
B IBLIOGRAPHY
e (t, x)
Ae =
t
0
m (t, x)
Am =
2 m +
t
0
1 2 Ae
1 e
2 e
e
A
+
A
+
c2 t2
c2 t
(3.39a)
(3.39b)
2 e +
1 2 Am
1 m
2 Am + A m + 2
2
2
c t
c t
= 0 je (t, x)
(3.39c)
= 0 jm (t, x)
(3.39d)
(3.40)
(3.41)
e (t, x)
1 2 e
2 e =
(3.42a)
2
2
c t
0
1 2 m
m (t, x)
2 m =
(3.42b)
2
2
c t
0
1 2 Ae
2 Ae = 0 je (t, x)
(3.42c)
c2 t2
1 2 Am
2 Am = 0 jm (t, x)
(3.42d)
c2 t2
exhibiting once again, the striking properties of Diracs symmetrised Maxwell theory.
E ND OF
EXAMPLE
3.1C
Bibliography
[1] L. D. FADEEV AND A. A. S LAVNOV, Gauge Fields: Introduction to Quantum
Theory, No. 50 in Frontiers in Physics: A Lecture Note and Reprint Series. Benjamin/Cummings Publishing Company, Inc., Reading, MA . . . , 1980, ISBN 08053-9016-2.
[2] M. G UIDRY, Gauge Field Theories: An Introduction with Applications, John
Wiley & Sons, Inc., New York, NY . . . , 1991, ISBN 0-471-63117-5.
[3] J. D. JACKSON, Classical Electrodynamics, third ed., John Wiley & Sons, Inc.,
New York, NY . . . , 1999, ISBN 0-471-30932-X.
[4] L. L ORENZ, Philosophical Magazine (1867), pp. 287301.
46
E LECTROMAGNETIC P OTENTIALS
4
Relativistic
Electrodynamics
The theory of relativity is not merely a scientific development of great importance in its own right. It is even more significant as the first stage of a radical
change in our basic concepts, which began in physics, and which is spreading
into other fields of science, and indeed, even into a great deal of thinking outside
of science. For as is well known, the modern trend is away from the notion of
sure absolute truth, (i.e., one which holds independently of all conditions, contexts, degrees, and types of approximation etc..) and toward the idea that a given
concept has significance only in relation to suitable broader forms of reference,
within which that concept can be given its full meaning.
47
48
R ELATIVISTIC E LECTRODYNAMICS
vt
y
y0
v
P(t, x, y, z)
P(t0 , x0 , y0 , z0 )
O
z
O0
x0
z0
F IGURE 4.1: Two inertial systems and 0 in relative motion with velocity v along the x = x0 axis. At time t = t0 = 0 the origin O0 of 0
coincided with the origin O of . At time t, the inertial system 0 has
been translated a distance vt along the x axis in . An event represented
by P(t, x, y, z) in is represented by P(t 0 , x0 , y0 , z0 ) in 0 .
Postulate 4.1 (Relativity principle; Poincar, 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 4.2 (Einstein, 1905). The velocity of light in empty space is independent of the motion of the source that emits the light.
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.
4.1
T HE S PECIAL T HEORY
OF
49
R ELATIVITY
where v = |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 and 0 ,
respectively, are connected by the following transformation:
ct0 = (ct x)
(4.3a)
x = (x vt)
(4.3b)
y =y
(4.3c)
z0 = z
(4.3d)
Taking the difference between the square of (4.3a) and the square of (4.3b) we
find that
c2 t02 x02 = 2 c2 t2 2xct + x2 2 x2 + 2xvt v2 t2 4
c2 t 2
v2
c2
2 2
= c t x2
1
v2
v2
2
x
1
c2
c2
(4.4)
From Equations (4.3) above we see that the y and z coordinates are unaffected
by the translational motion of the inertial system 0 along the x axis of system
. Using this fact, we find that we can generalise the result in Equation (4.4)
to
c2 t2 x2 y2 z2 = c2 t02 x02 y02 z02
(4.5)
which means that if a light wave is transmitted from the coinciding origins
O and O0 at time t = t0 = 0 it will arrive at an observer at (x, y, z) at time t
in and an observer at (x0 , y0 , z0 ) at time t0 in 0 in such a way that both observers conclude that the speed (spatial distance divided by time) of light in
vacuum is c. Hence, the speed of light in and 0 is the same. A linear coordinate transformation which has this property is called a (homogeneous) Lorentz
transformation.
50
R ELATIVISTIC E LECTRODYNAMICS
speed of light and t is time), and the remaining components are the components
of the ordinary 3 radius vector x defined in Equation (M.1) on page 168:
x = (x0 , x1 , x2 , x3 ) = (ct, x, y, z) (ct, x)
(4.6)
(4.7)
This summation process is an example of index contraction and is often referred to as index lowering.
(4.8)
4.1
T HE S PECIAL T HEORY
OF
51
R ELATIVITY
of the norm in a 4D Riemann space, then the explicit expression for the scalar
product of x with itself in this space must be
x x = c2 t 2 x 2 y2 z2
(4.9)
We notice that our space will have an indefinite norm which means that we deal
with a non-Euclidean space. We call the four-dimensional space (or spacetime) with this property Lorentz space and denote it ; 4 . A corresponding real,
linear 4D space with a positive definite norm which is conserved during ordinary rotations is a Euclidean vector space. We denote such a space 4 .
Metric tensor
By choosing the metric tensor in ;
g =
>
<=
=?
as
1
if = = 0
1 if = = i = j = 1, 2, 3
0
if =
6
(g ) =
@AAB
(4.10)
CEDD
1 0
0
0
0 1 0
0
0 0 1 0 F
0 0
0 1
(4.11)
i.e., a matrix with a main diagonal that has the sign sequence, or signature,
{+, , , }, the index lowering operation in our chosen flat 4D space becomes
nearly trivial:
x = g x = (ct, x)
(4.12)
CEDD
@AAB
x0
x1
x2 F
x3
CEDD
@AAB
1 0
0
0
0 1 0
0
0 0 1 0 F
0 0
0 1
@AAB
CEDD
x0
x1
x2 F
x3
CEDD
@AAB
x0
x1
x2 F
x3
(4.13)
Hence, if the metric tensor is defined according to expression (4.10) the covariant radius four-vector x is obtained from the contravariant radius four-vector
x simply by changing the sign of the last three components. These components are referred to as the space components; the zeroth component is referred
to as the time component.
52
R ELATIVISTIC E LECTRODYNAMICS
As we see, for this particular choice of metric, the scalar product of x with
itself becomes
x x = (ct, x) (ct, x) = c2 t2 x2 y2 z2
(4.14)
(4.15a)
g = g
g g =
g g =
g
g
=
=
(4.15b)
(4.15c)
(4.15d)
1
0
if =
if 6=
(4.16)
(4.17)
4.1
T HE S PECIAL T HEORY
OF
53
R ELATIVITY
+,,
1
1 2
c
H
dx 1
dt
dx 2
+
dt
dx 3
+
dt
I
dt
1
v2
2 + (v )2 + (v )2 dt = c 1
dt
(v
)
x
y
z
c2
c2
:
c
= c 1 2 dt = dt = c d
= c 1
(4.18)
where we introduced
d = dt/
(4.19)
Since d measures the time when no spatial changes are present, it is called the
proper time.
Expressing Equation (4.5) on page 49 in terms of the differential interval
ds and comparing with Equation (4.17) on the preceding page, we find that
ds2 = c2 dt2 dx2 dy2 dz2
(4.20)
(4.21)
(4.22)
(4.23)
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 ds is invariant under a
Lorentz transformation.
54
R ELATIVISTIC E LECTRODYNAMICS
Four-vector fields
Any quantity which relative to any coordinate system has a quadruple of real
numbers and transforms in the same way as the radius four-vector x does, is
called a four-vector. In analogy with the notation for the radius four-vector we
introduce the notation a = (a0 , a) for a general contravariant four-vector field
in ; 4 and find that the lowering of index rule, Equation (M.32) on page 174,
for such an arbitrary four-vector yields the dual covariant four-vector field
a (x ) = g a (x ) = (a0 (x ), a(x ))
(4.24)
The scalar product between this four-vector field and another one b (x ) is
g a (x )b (x ) = (a0 , a) (b0 , b) = a0 b0 a b
(4.25)
CEDD
@AAB
0 0
0 0
0
0
1 0F
0
0
0 1
(4.26)
the linear Lorentz transformation (4.3) on page 49, i.e., the coordinate transformation x x0 = x0 (x0 , x1 , x2 , x3 ), from one inertial system to another
inertial system 0 , can be written
x0 = x
(4.27)
(4.28)
1 + 2
1 + 1 2
(4.29)
4.1
T HE S PECIAL T HEORY
OF
55
R ELATIVITY
X0
X 00
x01
x1
F IGURE 4.2: Minkowski space can be considered an ordinary Euclidean
space where a Lorentz transformation from (x1 , X 0 = ict) to (x01 , X 00 =
ict0 ) corresponds to an ordinary rotation through an angle . This rotation
2
2
leaves the Euclidean distance x1 + X 0 = x2 c2 t2 invariant.
(4.30a)
X 1 = x1
(4.30b)
(4.30c)
(4.30d)
X =x
X =x
dS = ids
(4.30e)
56
R ELATIVISTIC E LECTRODYNAMICS
x0
= ct
x00
x0 = x 1
P0
ct
O=
O0
x01
x1 = x
F IGURE 4.3: Minkowski diagram depicting geometrically the transformation (4.34) from the unprimed system to the primed system. Here w
denotes the world line for an event and the line x0 = x1 x = ct the world
line for a light ray in vacuum. Note that the event P is simultaneous with
all points on the x1 axis (t = 0), including the origin O while the event
P0 , which is also simultaneous with all points on the x0 axis, including
O0 = O, to an observer at rest in the primed system, is not simultaneous
with O in the unprimed system but occurs there at time |P P0 | /c.
(4.31)
(4.32)
(4.33a)
(4.33b)
1 The
fact that our Riemannian space can be transformed in this way into an Euclidean one
means that it is, strictly speaking, a pseudo-Riemannian space.
4.2
57
(4.34a)
x = x cosh ct sinh
(4.34b)
(4.35a)
cosh =
(4.35b)
tanh =
(4.35c)
dx
= (c, v) =
d
B@ K
c
1
v2
c2
v
1
v2
c2
= (u0 , u)
(4.37)
which, when multiplied with the scalar invariant m0 yields the four-momentum
p = m0
dx
= m0 (c, v) =
d
B@ K
m0 c
1
v2
c2
m0 v
1
v2
c2
= (p0 , p)
(4.38)
58
R ELATIVISTIC E LECTRODYNAMICS
(4.39)
where
m = m0 =
m0
(4.40)
1 vc2
We can interpret this such that the Lorentz covariance implies that the masslike term in the ordinary 3D linear momentum is not invariant. A better way to
look at this is that p = mv = m0 v is the covariantly correct expression for the
kinetic three-momentum.
Multiplying the zeroth (time) component of the four-momentum p with
the scalar invariant c, we obtain
cp0 = m0 c2 =
m0 c2
v2
c2
= mc2
(4.41)
Since this component has the dimension of energy and is the result of a covariant description of the motion of a particle with its kinetic momentum described
by the spatial components of the four-momentum, Equation (4.38) on the preceding page, we interpret cp0 as the total energy E. Hence,
cp = (cp0 , cp) = (E, cp)
(4.42)
(m0 c2 )2
2
1 vc2
(4.43)
v2
1 2 = (m0 c2 )2
c
Since this is an invariant, this equation holds in any inertial frame, particularly
in the frame where p = 0 and there we have
E = m 0 c2
(4.44)
4.3
59
dx
= 0 u = 0 (c, v) = (c, v)
d
(4.45)
where we introduced
= 0
(4.46)
= = =
1 2
2
c2 t2
(4.47)
Since it has the characteristics of a four-scalar, the dAlembert operator is invariant and, hence, the homogeneous wave equation is Lorentz covariant.
,A
c
(4.48)
where is the scalar potential and A the vector potential, defined in Section 3.3
on page 36, we can write the inhomogeneous wave equations (Lorentz potential equations), Equations (3.14) on page 39, in the following compact (and
covariant) way:
A = 0 j
(4.49)
With the help of the above, we can formulate our electrodynamic equations
covariantly. For instance, the covariant form of the equation of continuity,
Equation (1.21) on page 9 is
j = 0
(4.50)
60
R ELATIVISTIC E LECTRODYNAMICS
and the Lorentz gauge condition, Equation (3.13) on page 38, can be written
A = 0
(4.51)
(4.52)
If only one dimension Lorentz contracts (for instance, due to relative motion along the x direction), a 3D spatial volume transforms according to
:
v2
1
dV = d3x = dV0 = dV0 1 2 = dV0 1 2
(4.53)
(4.54)
i.e., the charge in a given volume is conserved. We can therefore conclude that
the elementary charge is a universal constant.
A 4
=
0
,A
c
=
v=0
1
q0
,0
40 c |x x0 |0
(4.55)
where |x x0 |0 is the usual distance from the source point to the field point,
evaluated in the rest system (signified by the index 0).
Let us introduce the relative radius four-vector between the source point
and the field point:
R = x x0 = (c(t t0 ), x x0 )
(4.56)
2
(4.57)
4.3
61
We know that in vacuum the signal (field) from the charge q0 at x0 propagates to x with the speed of light c so that
x x0 = c(t t0 )
(4.58)
Inserting this into Equation (4.57) on the preceding page, we see that
R R = 0
(4.59)
R = ( x x 0 , x x 0 )
(4.60)
Now we want to find the correspondence to the rest system solution, Equation (4.55) on the preceding page, in an arbitrary inertial system. We note from
Equation (4.37) on page 57 that in the rest system
3u
B@ K
and
c
1
v2
c2
v
1
v2
c2
= (c, 0)
(4.61)
v=0
(R )0 = ( x x0 , x x0 )0 = ( x x0 0 , (x x0 )0 )
(4.62)
= (u )0 (R )0
= (c, 0) ( x x0 0 , (x x0 )0 ) = c x x0 0
(4.63)
q0
u
40 cu R
(4.64)
u R = (c, v) x x0 , (x x0 )4 = c x x0 v (x x0 )4
(4.65)
62
R ELATIVISTIC E LECTRODYNAMICS
= v
v
c
(4.66)
and introducing
s x x0
def
v (x x0 )
x x0 (x x0 )
c
(4.67)
we can write
u R = cs
(4.68)
and
u
=
cu R
1 v
,
cs c2 s
(4.69)
q0
40
1 v
,
=
cs c2 s
,A
c
(4.70)
where in the last step the definition of the four-potential, Equation (4.48) on
page 59, 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 v, the scalar and vector potentials are given by the expressions
1
q0 1
q0
=
0
40 s 40 |x x | (x x0 )
q0
v
q0 v
=
A(t, x) =
0
2
2
40 c s 40 c |x x | (x x0 )
(t, x) =
(4.71a)
(4.71b)
(4.72)
i, j 6= k
(4.73)
4.3
63
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 . For instance, the Maxwell equation
B
(4.74)
E =
t
can in this tensor notation be written
E j Ei
Bi j
=
(4.75)
xi x j
t
We know from Chapter 3 that the fields can be derived from the electromagnetic potentials in the following way:
B = A
(4.76a)
E =
(4.76b)
A
t
In component form, this can be written
A j Ai
= i A j j Ai
(4.77a)
xi x j
Ai
Ei = i
= i t Ai
(4.77b)
x
t
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,
Equation (4.48) on page 59:
Bi j =
A =
,A
c
(4.78)
Inspection of (4.78) and Equation (4.77) above makes it natural to define the
four-tensor
A A
F =
= A A
(4.79)
x x
This anti-symmetric (skew-symmetric), four-tensor of rank 2 is called the electromagnetic field tensor. In matrix representation, the contravariant field tensor can be written
3F
CEDD
@AAB
0
E x /c Ey /c Ez /c
E x /c
0
Bz
By
Ey /c
Bz
0
B x F
Ez /c By
Bx
0
(4.80)
64
R ELATIVISTIC E LECTRODYNAMICS
The covariant field tensor is obtained from the contravariant field tensor in the
usual manner by index contraction (index lowering):
F = g g F = A A
(4.81)
3 F
CEDD
@AAB
0
E x /c Ey /c Ez /c
E x /c
0
Bz
By
Ey /c Bz
0
B x F
Ez /c By
Bx
0
(4.82)
(4.83)
E x Ey Ez
+
+
x
y
z
(4.84)
or, equivalently,
E = 0 c2 =
(4.85)
which is the Maxwell source equation for the electric field, Equation (1.43a)
on page 14.
For = 1, Equation (4.84) yields
F 01 F 11 F 21 F 31
Bz By
1 E x
+0
+
= 0 j1 = 0 v x (4.86)
+ 1 + 2 + 3 = 2
0
x
x
x
x
c t
y
z
or, using 0 0 = 1/c2 ,
By Bz
E x
0 0
= 0 j x
z
y
t
(4.87)
4.3
65
B IBLIOGRAPHY
E
= 0 j(t, x)
t
(4.88)
which is the Maxwell source equation for the magnetic field, Equation (1.43d)
on page 14.
The two Maxwell field equations
E =
B
t
B = 0
(4.89)
(4.90)
(4.91)
Hence, Equation (4.83) on the facing page and Equation (4.91) above constitute Maxwells equations in four-dimensional formalism.
Bibliography
[1] J. A HARONI, The Special Theory of Relativity, second, revised ed., Dover Publications, Inc., New York, 1985, ISBN 0-486-64870-2.
[2] A. O. BARUT, Electrodynamics and Classical Theory of Fields and Particles,
Dover Publications, Inc., New York, NY, 1980, ISBN 0-486-64038-8.
[3] R. B ECKER, Electromagnetic Fields and Interactions, Dover Publications, Inc.,
New York, NY, 1982, ISBN 0-486-64290-9.
[4] D. B OHM, The Special Theory of Relativity, Routledge, New York, NY, 1996,
ISBN 0-415-14809-X.
[5] W. T. G RANDY, Introduction to Electrodynamics and Radiation, Academic
Press, New York and London, 1970, ISBN 0-12-295250-2.
[6] L. D. L ANDAU AND E. M. L IFSHITZ, The Classical Theory of Fields, fourth revised English ed., vol. 2 of Course of Theoretical Physics, Pergamon Press, Ltd.,
Oxford . . . , 1975, ISBN 0-08-025072-6.
[7] F. E. L OW, Classical Field Theory, John Wiley & Sons, Inc., New York, NY . . . ,
1997, ISBN 0-471-59551-9.
66
R ELATIVISTIC E LECTRODYNAMICS
[8] C. M LLER, The Theory of Relativity, second ed., Oxford University Press,
Glasgow . . . , 1972.
[9] H. M UIRHEAD, The Special Theory of Relativity, The Macmillan Press Ltd.,
London, Beccles and Colchester, 1973, ISBN 333-12845-1.
[10] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism,
second ed., Addison-Wesley Publishing Company, Inc., Reading, MA . . . , 1962,
ISBN 0-201-05702-6.
[11] J. J. S AKURAI, Advanced Quantum Mechanics, Addison-Wesley Publishing
Company, Inc., Reading, MA . . . , 1967, ISBN 0-201-06710-2.
[12] B. S PAIN, Tensor Calculus, third ed., Oliver and Boyd, Ltd., Edinburgh and
London, 1965, ISBN 05-001331-9.
[13] A. N. W HITEHEAD, Concept of Nature, Cambridge University Press, Cambridge . . . , 1920, ISBN 0-521-09245-0.
5
Electromagnetic
Fields and
Particles
67
68
E LECTROMAGNETIC F IELDS
AND
PARTICLES
Lagrange formalism
Let us no introduce a function L(4) which fulfils the variational principle
L(4) (x , u ) d = 0
(5.1)
where d is the proper time defined via Equation (4.18) on page 53, and the
endpoints are fixed. We shall show that L(4) acts as a kind of generalisation to
the common 3D Lagrangian.
We require that L(4) fulfils the following conditions:
1. The Lagrange function must be invariant. This implies that L(4) must be
a scalar.
2. The Lagrange function must yield linear equations of motion. This implies that L(4) must not contain higher than the second power of the fourvelocity u .
According to Formula (M.96) on page 185 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 m0 then in 3D the kinetic energy is
m0 v2 /2. This suggests that in 4D the Lagrangian for a free particle should be
1
free
L(4)
= m0 u u
2
(5.2)
For an interaction with the electromagnetic field we can introduce the interaction with the help of the four-potential given by Equation (4.78) on page 63 in
the following way
1
L(4) = m0 u u + qu A (x )
2
(5.3)
We call this the four-Lagrangian and shall now show how this function, together with the variation principle, Formula (5.1) above, yields covariant results which are physically correct.
The variation principle (5.1) with the 4D Lagrangian (5.3) inserted, leads
to
1
0
=
=
0
1
0
m0
u u + qu A / d
0 2
A
m0 (u u )
u + q A u + u x d
2 u
x
L(4) (x , u ) d =
m0 u u
(5.4)
+ q A u + u A x 4 d = 0
5.1
C HARGED PARTICLES
IN AN
69
E LECTROMAGNETIC F IELD
dx
d
(5.5)
which means that we can write the variation of u as a total derivative with
respect to :
u =
dx
d
=
x 4
d
d 3
(5.6)
Inserting this into the first two terms in the last integral in Equation (5.4) on
the preceding page, we obtain
1
L(4) (x , u ) d
m0 u
d
d
x 4 + qA
x 4 + qu A x d
d 3
d 3
(5.7)
Partial integration in the two first terms in the right hand member of (5.7)
gives
1
L(4) (x , u ) d
m0
du
dA
x q
x + qu A x d
d
d
(5.8)
where the integrated parts do not contribute since the variations at the endpoints vanish. A change of irrelevant summation index from to in the first
two terms of the right hand member of (5.8) yields, after moving the ensuing
common factor x outside the partenthesis, the following expression:
L(4) (x , u ) d
du
dA
m0
q
+ qu A x d
d
d
(5.9)
(5.10)
By inserting this expression (5.10) into the second term in right-hand member
of Equation (5.9) on the preceding page, and noting the common factor qu of
70
E LECTROMAGNETIC F IELDS
AND
PARTICLES
the resulting term and the last term, we obtain the final variational principle
expression
L(4) (x , u ) d
m0 du
+ qu A A 4
3
d
(5.11)
x d
Since, according to the variational principle, this expression shall vanish and
x is arbitrary between the fixed end points 0 and 1 , the expression inside
in the integrand in the right hand member of Equation (5.11) above must
vanish. In other words, we have found an equation of motion for a charged
particle in a prescribed electromagnetic field:
du
= qu A A 4
(5.12)
d
3
With the help of Equation (4.79) on page 63 we can express this equation in
terms of the electromagnetic field tensor in the following way:
m0
du
= qu F
(5.13)
d
This is the sought-for covariant equation of motion for a particle in an electromagnetic field. It is often referred to as the Minkowski equation. As the reader
can easily verify, the spatial part of this 4-vector equation is the covariant (relativistically correct) expression for the Newton-Lorentz force equation.
m0
Hamiltonian formalism
The usual Hamilton equations for a 3D space are given by Equation (M.101)
on page 185 in Appendix M. These six first-order partial differential equations
are
H dqi
=
(5.14a)
pi
dt
dpi
H
=
(5.14b)
qi
dt
where H(pi , qi , t) = pi q i L(qi , q i , t) is the ordinary 3D Hamiltonian, qi is a
generalised coordinate and pi is its canonically conjugate momentum.
We seek a similar set of equations in 4D space. To this end we introduce a
canonically conjugate four-momentum p in an analogous way as the ordinary
3D conjugate momentum:
p =
L(4)
u
(5.15)
5.1
C HARGED PARTICLES
IN AN
71
E LECTROMAGNETIC F IELD
(5.16)
With the help of these, the radius four-vector x , considered as the generalised
four-coordinate, and the invariant line element ds, defined in Equation (4.18)
on page 53, we introduce the following eight partial differential equations:
H(4) dx
=
(5.17a)
p
d
dp
H(4)
=
(5.17b)
x
d
which form the four-dimensional Hamilton equations.
Our strategy now is to use Equation (5.15) and Equations (5.17) above to
derive an explicit algebraic expression for the canonically conjugate momentum four-vector. According to Equation (4.42) on page 58, c times a fourmomentum has a zeroth (time) component which we can identify with the
total energy. Hence we require that the component p0 of the conjugate fourmomentum vector defined according to Equation (5.15) above be identical to
the ordinary 3D Hamiltonian H divided by c and hence that this cp0 solves the
Hamilton equations, Equations (5.14) on the facing page. This later consistency check is left as an exercise to the reader.
Using the definition of H(4) , Equation (5.16) above, and the expression for
L(4) , Equation (5.3) on page 68, we obtain
1
(5.18)
H(4) = p u L(4) = p u m0 u u qu A (x )
2
Furthermore, from the definition (5.15) of the canonically conjugate fourmomentum p , we see that
p =
L(4)
=
u
u
1
m0 u u + qu A (x ) = m0 u + qA
2
(5.19)
(5.20)
m0 c2
2
(5.21)
72
E LECTROMAGNETIC F IELDS
AND
PARTICLES
1
p qA 4
m0 3
(5.22)
H(4) =
1
1
p qA 4
p qA 4
m0 3
m0 3
3p
3p
qA 4
3 p qA
(5.23)
p 2qA p + q2 A A 4
That this four-Hamiltonian yields the correct covariant equation of motion can
be seen by inserting it into the four-dimensional Hamiltons equations (5.17)
and using the relation (5.22):
q
A
H(4)
= (p qA )
x
m0
x
q
A
= m0 u
m0
x
(5.24)
A
= qu
x
du
A
dp
= m0
q u
=
d
d
x
where in the last step Equation (5.19) on the preceding page was used. Rearranging terms, and using Equation (4.80) on page 63, we obtain
du
(5.25)
= qu A A 4 = qu F
d
3
which is identical to the covariant equation of motion Equation (5.13) on
page 70. We can then safely conclude that the Hamiltonian in question is
correct.
Recalling expression (4.48) on page 59 and representing the canonically
conjugate four-momentum as p = (p0 , p), we obtain the following scalar products:
m0
p p = (p0 )2 (p)2
1
A p = p0 (p A)
c
1
A A = 2 2 (A)2
c
(5.26a)
(5.26b)
(5.26c)
5.1
C HARGED PARTICLES
IN AN
73
E LECTROMAGNETIC F IELD
Inserting these explicit expressions into Equation (5.23) on the facing page,
and using the fact that for H(4) is equal to the scalar value m0 c2 /2, as derived
in Equation (5.21) on the preceding page, we obtain the equation
m0 c2
1
2
q2
=
(p0 )2 (p)2 qp0 + 2q(p A) + 2 2 q2 (A)2
2
2m0
c
c
(5.27)
2q 0
q2
p L (p)2 2qpMEN A + q2 (A)2O + 2 2 m20 c2 = 0
c
c
(5.28)
(pqA)2
(p qA)2 + m20 c2
(5.29)
(p qA)2 + m20 c2
(5.30)
(5.31)
Using the explicit expressions (Equation (5.30) on the preceding page) and
(Equation (5.31) on the previous page), we obtain the explicit expression for
the ordinary 3D Lagrange function
L = p v q c
(p qA)2 + m20 c2
(5.32)
m0 v
1
v2
c2
= mv
(5.33)
74
E LECTROMAGNETIC F IELDS
AND
PARTICLES
where the quantity mv is the usual kinetic momentum, we can rewrite this expression for the ordinary Lagrangian as follows:
2
L = qA v + mv q c m2 v2 + m20 c2
2
= mv q( A v) mc = q + qA v m0 c
v2
1 2
c
(5.34)
(5.35)
5.2
75
i1
m
k
a
i
m
k
a
i+1
k
a
k
a
F IGURE 5.1: A one-dimensional chain consisting of N discrete, identical mass points m, connected to their neighbours with identical, ideal
springs with spring constants k. The equilibrium distance between the
neighbouring mass points is a and i1 (t), i (t), i+1 (t) are the instantaneous deviations, along the x axis, of positions of the (i 1)th, ith, and
(i + 1)th mass point, respectively.
1 N
2
2
mi k(i+1 i )
2
i=1
(5.36)
L = aP
(5.37)
i=1
Here,
1 m 2
ka . i+1a i /
2 a i
(5.38)
is the so called linear Lagrange density. If we now let N and, at the same
time, let the springs become infinitesimally short according to the following
scheme:
a dx
m
dm
=
a
dx
ka Y
i+1 i
a
x
(5.39a)
linear mass density
(5.39b)
Youngs modulus
(5.39c)
(5.39d)
76
E LECTROMAGNETIC F IELDS
AND
PARTICLES
we obtain
L=
P
dx
(5.40)
where
PQ
, , , t = H
t x
2
t
x
I
(5.41)
Notice how we made a transition from a discrete description, in which the mass
points were identified by a discrete integer variable i = 1, 2, . . . , N, to a continuous description, where the infinitesimal mass points were instead identified by
a continuous real parameter x, namely their position along x .
A consequence of this transition is that the number of degrees of freedom
for the system went from the finite number N to infinity! Another consequence
is that P has now become dependent also on the partial derivative with respect
to x 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 (5.35) on the previous page yields:
L dt
=
=
R
PS
RUTV
,
dx dt
t x
P
P
+
+
dx dt
t
x
WX
. t /
. x /
(5.42)
=0
The last integral can be integrated by parts. This results in the expression
RUTV
B@
B@
F
WX
dx dt = 0
(5.43)
where the variation is arbitrary (and the endpoints fixed). This means that the
integrand itself must vanish. If we introduce the functional derivative
P
P
B@
(5.44)
5.2
77
B@
F
/
=0
(5.45)
Y
=
t2
x2
2
2
=0
Y t2 x2
(5.46)
i.e., the one-dimensional wave equation for compression waves which propa
gate with phase speed v = Y/ along the linear structure.
A generalisation of the above 1D results to a three-dimensional continuum
is straightforward. For this 3D case we get the variational principle
L dt =
=
=
Y
d3x dt
PS
d4x
x
R TV
B@
F
WX
(5.47)
4
d x
=0
where the variation is arbitrary and the endpoints are fixed. This means that
the integrand itself must vanish:
P
B@
=0
(5.48)
xi
B@
xi
(5.49)
78
E LECTROMAGNETIC F IELDS
B@
PARTICLES
AND
F
/
=0
(5.50)
(5.51)
Z
, ,
; t = PQ , , i
i
x
t
t x
(5.52)
If, as usual, we differentiate this expression and identify terms, we obtain the
following Hamilton density equations
t
Z
(5.53a)
(5.53b)
The Hamilton density functions are in many ways similar to the ordinary
Hamilton functions and lead to similar results.
tot
=P
mech
+P
inter
+P
field
(5.54)
where the mechanical part has to do with the particle motion (kinetic energy).
It is given by L(4) /V where L(4) is given by Equation (5.2) on page 68 and V is
5.2
79
the volume. Expressed in the rest mass density %0 , the mechanical Lagrange
density can be written
mech
1
= %0 u u
2
(5.55)
The P inter part which describes the interaction between the charged particles and the external electromagnetic field. A convenient expression for this
interaction Lagrange density is
inter
= j A
(5.56)
For the field part P 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
field
1
F F
40
(5.57)
tot
1
1
= %0 u u + j A +
F F
2
40
(5.58)
E XAMPLE 5.1
B2
0 E 2
0
(5.59)
i.e., the difference between the magnetic and electric field energy densities.
From Formula (4.80) on page 63 we recall that
F
""#
0
E x /c
E y /c
E z /c
E x /c E y /c E z /c
0
Bz
By
Bz
0
B x
By
Bx
0
*& (
(
(5.60)
F
""#
0
E x /c
E y /c
E z /c
E x /c
0
Bz
By
E y /c
Bz
0
Bx
E z /c
By
B x
0
*& (
(
(5.61)
where denotes the row number and the column number. Then, Einstein summation
80
E LECTROMAGNETIC F IELDS
AND
PARTICLES
(5.62)
1 2
B2
E
0 c 2 0
1
2
B2
0 E 2
0
(5.63)
QED
E ND
OF EXAMPLE
5.1C
EM
=P
inter
+P
field
= j A +
1
F F
40
(5.64)
inserted into the Euler-Lagrange equations, expression (5.48) on page 77, yields
two of Maxwells equations. To see this, we note from Equation (5.64) above
and the results in Example 5.1 that
P
EM
= j
(5.65)
5.2
81
Furthermore,
P EM
(
A )
1
F F 4
40
( A ) 3
1
=
( A A )( A A )
40
( A )
=
A A A A
40
( A )
(5.66)
A A + A A
=
1
A A A A 4
20
( A ) 3
'
But
A A 4 = A
A + A
A
( A ) 3
( A )
( A )
A + A
g g A
= A
( A )
( A )
= A
A + g g A
A (5.67)
( A )
( A )
= A
A + A
A
( A )
( A )
= 2 A
Similarly,
A A 4 = 2 A
( A ) 3
(5.68)
so that
P EM
(
A )
1
1
A A 4 = F
0 3
0
(5.69)
This means that the Euler-Lagrange equations, expression (5.48) on page 77,
for the Lagrangian density P EM and with A as the field quantity become
P
EM
P EM
(
A )
= j
1
F = 0
0
(5.70)
82
E LECTROMAGNETIC F IELDS
AND
PARTICLES
or
F = 0 j
(5.71)
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 following Lagrange density:
1
m2 2 4
23
(5.72)
m2 ) = 0
(5.73)
em|x|
|x|
(5.74)
which describes the Yukawa meson field for a scalar meson with mass m. With
=
1
c2 t
(5.75)
1 2 2
c + ()2 + m2 2
2
(5.76)
EM
=P
inter
+P
field
= j A +
1
F F + m2 A A
40
(5.77)
5.2
83
B IBLIOGRAPHY
(5.78)
Bibliography
[1] A. O. BARUT, Electrodynamics and Classical Theory of Fields and Particles,
Dover Publications, Inc., New York, NY, 1980, ISBN 0-486-64038-8.
[2] V. L. G INZBURG, Applications of Electrodynamics in Theoretical Physics
and Astrophysics, Revised third ed., Gordon and Breach Science Publishers,
New York, London, Paris, Montreux, Tokyo and Melbourne, 1989, ISBN 288124-719-9.
[3] H. G OLDSTEIN, Classical Mechanics, second ed., Addison-Wesley Publishing
Company, Inc., Reading, MA . . . , 1981, ISBN 0-201-02918-9.
[4] W. T. G RANDY, Introduction to Electrodynamics and Radiation, Academic
Press, New York and London, 1970, ISBN 0-12-295250-2.
[5] L. D. L ANDAU AND E. M. L IFSHITZ, The Classical Theory of Fields, fourth revised English ed., vol. 2 of Course of Theoretical Physics, Pergamon Press, Ltd.,
Oxford . . . , 1975, ISBN 0-08-025072-6.
[6] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism,
second ed., Addison-Wesley Publishing Company, Inc., Reading, MA . . . , 1962,
ISBN 0-201-05702-6.
[7] J. J. S AKURAI, Advanced Quantum Mechanics, Addison-Wesley Publishing
Company, Inc., Reading, MA . . . , 1967, ISBN 0-201-06710-2.
[8] D. E. S OPER, Classical Field Theory, John Wiley & Sons, Inc., New York, London, Sydney and Toronto, 1976, ISBN 0-471-81368-0.
84
E LECTROMAGNETIC F IELDS
AND
PARTICLES
6
Electromagnetic
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.
(x0 ) d3x0
(6.1)
where the is the charge density introduced in Equation (1.7) on page 4, the
electric dipole moment vector
p(x0 ) =
(6.2)
85
86
E LECTROMAGNETIC F IELDS
AND
M ATTER
(6.3)
q
1
1
(x x0 )i
+
pi
2
40 |x x0 | |x x0 |
|x x0 |
1
Qi j
+
|x x0 |3
3 (x x0 )i (x x0 ) j 1
i j + . . .
2 |x x0 | |x x0 | 2
(6.4)
x x0 3 0
1
d x =
0
3
|x x |
40
V
1
P(x0 ) 0
d3x0
|x x0 |
V
P(x0 )
p (x) =
P(x0 )
1
d3x0
|x x0 |
(6.5)
Using the expression Equation (M.83) on page 182 and applying the divergence theorem, we can rewrite this expression for the potential as follows:
1
40
1
=
40
p (x) =
P(x0 )
0 P(x0 ) 3 0
3 0
d
x
dx
|x x0 |
V
V |x x0 |
0
0
0
P(x ) n 2
P(x ) 3 0
d x
dx
0|
|x
S
V |x x0 |
0
(6.6)
6.1
E LECTRIC P OLARISATION
AND
87
D ISPLACEMENT
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 discontinuity in P n at the surface. Doing so, we find that the contribution
from the electric dipole moments to the potential is given by
p =
1
40
0 P(x0 ) 3 0
dx
|x x0 |
(6.7)
Comparing this expression with expression Equation (3.3) on page 35 for the
electrostatic potential from a static charge distribution , we see that P(x)
has the characteristics of a charge density and that, to the lowest order, the
effective charge density becomes (x) P(x), in which the second term is a
polarisation term.
The version of Equation (1.7) on page 4 where true and polarisation
charges are separated thus becomes
E =
(x) P(x)
0
(6.8)
Rewriting this equation, and at the same time introducing the electric displacement vector (C/m2 )
D = 0 E + P
(6.9)
we obtain
(0 E + P) = D = true (x)
(6.10)
where true is the true charge density in the medium. This is one of Maxwells
equations and is valid also for time varying fields. By introducing the notation pol = P for the polarised charge density in the medium, and
total = true + pol for the total charge density, we can write down the following alternative version of Maxwells equation (6.23a) on page 90
E =
total (x)
0
(6.11)
Often, for low enough field strengths |E|, the linear and isotropic relationship between P and E
P = 0 E
(6.12)
88
E LECTROMAGNETIC F IELDS
AND
M ATTER
(6.13)
where, approximately,
= 0 (1 + )
(6.14)
1
2
(6.15)
6.2
M AGNETISATION
AND THE
89
M AGNETISING F IELD
volume, M. Via the definition of the vector potential one can show that the
magnetisation current and the magnetisation is simply related:
jM = M
(6.16)
In a stationary medium we therefore have a total current which is (approximately) the sum of the three currents enumerated above:
jtotal = jtrue +
P
+M
t
(6.17)
P
+M
t
(6.18)
Moving the term M to the left hand side and introducing the magnetising
field (magnetic field intensity, Ampre-turn density) as
H=
B
M
0
(6.19)
and using the definition for D, Equation (6.9) on page 87, we can write this
incorrect equation in the following form
H = jtrue +
E
P
D
= jtrue +
0
t
t
t
(6.20)
(6.21)
where, approximately,
= 0 (1 + m )
(6.22)
90
E LECTROMAGNETIC F IELDS
AND
M ATTER
E =
(6.23a)
(6.23b)
B
t
(6.23c)
H = j(t, x) +
D
t
(6.23d)
and are called Maxwells macroscopic equations. These equations are convenient to use in certain simple cases. Together with the boundary conditions and
the constitutive relations, they describe uniquely (but only approximately!) the
properties of the electric and magnetic fields in matter.
(6.24)
Integration over the entire volume V and using Gausss theorem (the divergence theorem), we obtain
t
1
(H B + E D) d3x0 =
2
j E d3x0 +
(E H) nd
2x0
(6.25)
(6.26)
j E d3x0 =
j2 3 0
dx
j EEMF d3x0
(6.27)
6.3
E NERGY
AND
91
M OMENTUM
L
j EEMF d3x0 =
MEN
L
j2 3 0
dx +
MEN O t L
1
2x0
(E D + H B) d3x0 + (E H) nd
2
MEN
O L S M[N
O
Joule heat
Field energy
Radiated power
(6.28)
which is the energy theorem in Maxwells theory also known as Poyntings
theorem.
It is convenient to introduce the following quantities:
1
E D d3x0
2 V
1
H B d3x0
Um =
2 V
S = EH
Ue =
(6.29)
(6.30)
(6.31)
where U e is the electric field energy, U m is the magnetic field energy, both
measured in J, and S is the Poynting vector (power flux), measured in W/m2 .
(D B) + D
t
t
= E( D) B ( H)
(D B) D ( E) + H(
L MEN BO )
t
E + j B = ( D)E + H
(6.32)
=0
(D B)
t
92
E LECTROMAGNETIC F IELDS
AND
M ATTER
One verifies easily that the ith vector components of the two terms in
square brackets in the right hand member of (6.32) can be expressed as
[E( D) D ( E)]i =
1
D
E
E
D
+
2
xi
xi
x j
1
E i D j E D i j
2
(6.33)
and
[H( B) B ( H)]i =
1
B
H
H
B
+
2
xi
xi
x j
1
Hi B j B H i j
2
(6.34)
respectively.
Using these two expressions in the ith component of Equation (6.32) on
the previous page and re-shuffling terms, we get
D
B
E
H
+ H
+ (D B)i
D
B
xi
xi
xi
xi \
t
1
1
E i D j E D i j + Hi B j H B i j
2
2
(6.35)
(E + j B)i
=
x j
1
2
Introducing the electric volume force Fev via its ith component
(Fev )i = (E + j B)i
1
2
D
E
B
H
D
+ H
B
xi
xi
xi
xi 0
(6.36)
(6.37)
Fev + t (D B)
=
i
T i j
= ( T)i
x j
(6.38)
(6.39)
B = m 0 H = H
(6.40)
6.3
93
B IBLIOGRAPHY
(6.41)
i
where S is the Poynting vector defined in Equation (6.29) on page 91. Integration over the entire volume V yields
L
Fev d3x0 +
MEN
d
dt L
m 3 0
Sd x =
c2MEN
O
Field momentum
L
T n d2x0
MEN
(6.42)
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 Maxwells theory.
In vacuum (6.42) becomes
(E + v B) d3x0 +
1 d
c2 dt
S d3x0 =
V
T n d2x0
(6.43)
or
d mech d field
p
+ p
=
dt
dt
T n d2x0
(6.44)
Bibliography
[1] E. H ALLN, Electromagnetic Theory, Chapman & Hall, Ltd., London, 1962.
[2] J. D. JACKSON, Classical Electrodynamics, third ed., John Wiley & Sons, Inc.,
New York, NY . . . , 1999, ISBN 0-471-30932-X.
[3] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism,
second ed., Addison-Wesley Publishing Company, Inc., Reading, MA . . . , 1962,
ISBN 0-201-05702-6.
[4] J. A. S TRATTON, Electromagnetic Theory, McGraw-Hill Book Company, Inc.,
New York, NY and London, 1953, ISBN 07-062150-0.
94
E LECTROMAGNETIC F IELDS
AND
M ATTER
7
Electromagnetic
Fields from
Arbitrary Source
Distributions
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 inhomogeneous wave equation (3.19) on page 41 we presupposed the existence of a
Fourier transform pair (3.20a) on page 41 for the generic source term
(t, x) =
(x) =
1
2
(x) eit d
(7.1a)
(t, x) eit dt
(7.1b)
That such transform pairs exists is true for most physical variables which are
neither strictly monotonically increasing nor strictly monotonically decreasing
with time. For charge and current densities varying in time we can therefore,
without loss of generality, work with individual Fourier components w (x) and
jw (x), respectively. Strictly speaking, the existence of a single Fourier component assumes a monochromatic source (i.e., a source containing only one
single frequency component), which in turn requires that the electric and magnetic fields exist for infinitely long times. However, by taking the proper limits,
we may still use this approach even for sources and fields of finite duration.
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
(t, x) and current densities j(t, x), defined by the temporal Fourier transform
95
96
E LECTROMAGNETIC F IELDS
AND
M ATTER
pairs
(t, x) =
(x) eit d
1
2
(x) =
(7.2a)
(t, x) eit dt
(7.2b)
and
j(t, x) =
j (x) =
j (x) eit d
1
2
(7.3a)
j(t, x) eit dt
(7.3b)
under the assumption that only retarded potentials produce physically acceptable solutions.1
The temporal Fourier transform pair for the retarded vector potential can
then be written
(t, x) =
(x) eit d
1
(x) =
2
1
(t, x) e dt =
4
0
it
(7.4a)
0
eik|xx | 3 0
(x )
dx
|x x0 |
0
(7.4b)
where in the last step, we made use of the explicit expression for the temporal
Fourier transform of the generic potential component (x), Equation (3.32)
on page 43. Similarly, the following Fourier transform pair for the vector potential must exist:
A(t, x) =
1
A (x) =
2
A (x) eit d
(7.5a)
0
eik|xx | 3 0
0
j (x0 )
dx
A(t, x) e dt =
4
|x x0 |
it
(7.5b)
(7.6)
in order that all physical quantities be real. Similar transform pairs and requirements of real-valuedness exist for the fields themselves.
1 In
fact, John A. Wheeler and Richard P. Feynman derived in 1945 a fully self-consistent
electrodynamics using both the retarded and the advanced potentials [6]; See also [1].
7.1
97
T HE M AGNETIC F IELD
(7.7a)
i0 t
(7.7b)
(t, x) = 0 (x)ei0 t
(7.7c)
i0 t
(7.7d)
j(t, x) = j0 (x)e
A(t, x) = A0 (x)e
(7.8)
(7.9)
which provides a relation between (the Fourier transforms of) the vector and
scalar potentials.
Using the Fourier transformed version of Equation (7.8) above and Equation (7.5b) on the facing page, we obtain
0
B (x) = A (x) =
4
eik|xx | 3 0
j (x )
dx
|x x0 |
V
0
(7.10)
98
E LECTROMAGNETIC F IELDS
FROM
eik|xx |
d3x0
|x x0 |
V
0
x x0
j (x0 )
eik|xx | d3x0
3
0
|x x |
V
x x0 ik|xx0 |
1
3 0
+ j (x0 ) ik
e
0
|x x |
|x x0 | d x
V
0
4
0
=
4
B (x) =
j (x0 )
(7.11)
0 ik|xx | (x x0 )
j (x )e
d3x0
|x x0 |3
V
0
(ik)j(x0 )eik|xx | (x x0 ) 3 0
+
dx
|x x0 |2
V
0
=
4
From this expression for the magnetic field in the frequency () domain,
we obtain the total magnetic field in the temporal (t) domain by taking the
inverse Fourier transform (using the identity ik = i/c):
B(t, x) =
=
0
4
B (x) eit d
1
+
c
=
0
L 4
V
|x x0 |3
d3x0
|x x0 |2
0 , x0 ) (x x0 )
j(tret
0
d3x0 +
03
4c
|xM[N x |
Induction field
O L
d3x0
(7.12)
0 , x0 ) (x x0 )
j(tret
d3x0
2
0
|xM[
N x|
O
Radiation field
where
def
0
j(tret
, x0 )
j
t
(7.13)
0
t=tret
The first term, the induction field, dominates near the current source but falls
off rapidly with distance from it, is the electrodynamic version of the BiotSavart law in electrostatics, Formula (1.13) on page 6. The second term, the
radiation field or the far field, dominates at large distances and represents energy that is transported out to infinity. Note how the spatial derivatives ()
gave rise to a time derivative ()!
7.2
99
T HE E LECTRIC F IELD
eik|xx | 3 0 i0
1
(x0 )
dx +
=
40 V
4
|x x0 |
0
(x0 )eik|xx | (x x0 ) 3 0
1
dx
=
40 V
|x x0 |3
ik
eik|xx | 3 0
j (x )
dx
|x x0 |
V
0
(7.14)
dx
c |x x0 |
|x x0 |
(7.15)
(7.16)
Doing so in the last term of Equation (7.14) above, and also using the fact that
k = /c, we can rewrite this Equation as
1
E (x) =
40
1
c L
(x0 )eik|xx | (x x0 ) 3 0
dx
|x x0 |3
0
[0 j (x0 )](x x0 )
eik|xx | 3 0
0
ikj (x )
|x x0 | d xO
|x x0 |
MEN
(7.17)
I
The last vector-valued integral can be further rewritten in the following
way:
I =
=
[0 j (x0 )](x x0 )
eik|xx | 3 0
0
ikj
(x
)
|x x0 |
|x x0 | d x
0
eik|xx | 3 0
jm xl xl0
0
ik jl (x ) x l
0 |x x0 |
xm
|x x0 | d x
(7.18)
100
E LECTROMAGNETIC F IELDS
FROM
But, since
0
xm
xl xl0 ik|xx0 |
e
=
|x x0 |2
jm
jm xl xl0 ik|xx0 |
e
0
xm
|x x0 |2
xl xl0 ik|xx0 |
e
+ jm 0
xm |x x0 |2
(7.19)
we can rewrite I as
I =
jm
0
xm
xl xl0
eik|xx | 3 0
ik|xx0 |
+
ikj
dx
x
l
2
|x x0 |
|x x0 |
xl xl0
0
jm
x eik|xx | d3x0
2 l
0
|x x |
0
xm
(7.20)
ik
0
2
eik|xx |
+
j
j (x x0 ) (x x0 )eik|xx | d3x0
2
4
0
0
|x x |
|x x |
V]
j (x x0 ) (x x0 )
|x x0 |3
ik|xx0 |
eik|xx |
+ j
d3x0
|x x0 | ^
(7.21)
Using the triple product bac-cab Formula (F.54) on page 164 backwards, and
inserting the resulting expression for I into Equation (7.17) on the previous
page, we arrive at the following final expression for the Fourier transform of
the total E field:
0
1
eik|xx | 3 0 i0
E (x) =
(x0 )
dx +
40 V
4
|x x0 |
0|
0
ik|xx
0
1
(x )e 0 3(x x ) d3x0
=
40 V
|x x |
0
c V
|x x0 |3
1
+
c
eik|xx | 3 0
j (x )
dx
|x x0 |
V
0
(7.22)
Taking the inverse Fourier transform of Equation (7.22), once again using
the vacuum relation = kc, we find, at last, the expression in time domain for
7.3
101
T HE R ADIATION F IELDS
E (x) eit d
1
L 40
0 , x0 )(x x0 )
(tret
d3x0
|xMEN x0 |3
O
1
4
L 0c
0 , x0 ) (x x0 )](x x0 )
[j(tret
d3x0
4
0
M[|xN x |
O
(7.23)
Intermediate field
1
L 40 c
0 , x0 ) (x x0 )] (x x0 )
[j(tret
d3x0
0 |4
x
|x
MEN
O
Intermediate field
1
+
L 40 c2
0 , x0 ) (x x0 )] (x x0 )
[j(tret
d3x0
0 |3
x
|x
M[N
O
Radiation field
Here, the first term represents the retarded Coulomb field and the last term
represents 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 43 in Chapter 3.
102
E LECTROMAGNETIC F IELDS
FROM
Brad (t, x) =
Erad (t, x) =
it
Brad
d =
(x) e
it
Erad
d
(x) e
1
40 c2
where
def
0
j(tret
, x0 )
j
t
0
4c
0 , x0 ) (x x0 )
j(tret
d3x0 (7.24a)
2
0
|x x |
0 , x0 ) (x x0 )] (x x0 )
[j(tret
d3x0
|x x0 |3
(7.24b)
(7.25)
0
t=tret
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 Equation (7.24a) and
Equation (7.24b) above were included in Equation (7.11) on page 98 and Equation (7.22) on page 100, respectively and are explicitly given by
1 rad
B (t, x) eit dt
2
j (x0 ) (x x0 ) ik|xx0 | 3 0
k0
dx
= i
e
4 V
|x x0 |2
0
j (x0 ) k ik|xx0 | 3 0
e
= i
dx
4 V |x x0 |
1 rad
Erad
E (t, x) eit dt
(x) =
2
[j (x0 ) (x x0 )] (x x0 ) ik|xx0 | 3 0
k
dx
= i
e
40 c V
|x x0 |3
1
[j (x0 ) k] (x x0 ) ik|xx0 | 3 0
= i
dx
e
40 c V
|x x0 |2
Brad
(x) =
(7.26a)
(7.26b)
k x x0 k (x x0 ) k (x x0 ) k (x0 x0 )
k |x x0 | k (x0 x0 )
(7.27)
7.3
103
T HE R ADIATION F IELDS
S
dS = nd
2x
k
x x0
x x0
x0
x0 x 0
x0
V
O
F IGURE 7.1: Relation between the surface normal and the k vector for
radiation generated at source points x0 near the point x0 in the source
volume V. At distances much larger than the extent of V, the unit vector
n,
normal to the surface S which has its centre at x0 , and the unit vector
k of the radiation k vector from x0 are nearly coincident.
Recalling from Formula (F.48) and Formula (F.49) on page 164 that
dS = |x x0 |2 d = |x x0 |2 sin d d
and noting from Figure 7.1 that k and n are nearly parallel, we see that we can
approximate.
k dS
k n
=
dS d
2
|x x0 |
|x x0 |2
(7.28)
i
e
dx
e
40 c
|x x0 |2
V
0
1 eik|xx0 | (x x0 )
[j (x0 ) k] eik(x x0 ) d3x0
i
40 c |x x0 | |x x0 | V
(7.29b)
Brad
(x) i
104
E LECTROMAGNETIC F IELDS
FROM
hSi = hE Hi =
Using the far-field approximations (7.29a) and (7.29b) and the fact that 1/c =
(j k)e
ik(x0 x0 ) 3 0
dx
2
x x0
|x x0 |
(7.31)
dP
1
=
R0
2
d 32
(j k)e
ik(x0 x0 ) 3 0
dx
2
(7.32)
7.4
105
R ADIATED E NERGY
over the whole bandwidth. The total energy transmitted through a unit area is
the time integral of the Poynting vector:
S(t) dt =
=
(E H) dt
d
d0
(7.33)
(E H0 )ei(+ )t dt
If we carry out the temporal integration first and use the fact that
ei(+ )t dt = 2( + 0 )
(7.34)
S(t) dt = 2
= 2
2
0
2
=
0
= 2
= 2
(E H ) d
(E H ) d
(E H ) d +
(E H ) d +
(E H ) d
(E H ) d
(E H ) d
(7.35)
(E B + E B ) d
(E B + E B ) d
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 (7.26a) and (7.26b). The result, after integration over the area S of a large sphere which encloses the source, is
1
U=
4
0
0
S
j k ik|xx0 | 3 0 2
d x k ndS
d
e
|x x0 |
(7.36)
Inserting the approximations (7.27) and (7.28) into Equation (7.36) above
and also introducing
U=
U d
(7.37)
and recalling the definition (2.18) on page 26 for the vacuum resistance R 0 we
obtain
1
dU
d R0
d
4
(j k)e
ik(x0 x0 ) 3 0
dx
2
(7.38)
106
E LECTROMAGNETIC F IELDS
FROM
Bibliography
[1] F. H OYLE , S IR AND J. V. NARLIKAR, Lectures on Cosmology and Action at a
Distance Electrodynamics, World Scientific Publishing Co. Pte. Ltd, Singapore,
New Jersey, London and Hong Kong, 1996, ISBN 9810-02-2573-3(pbk).
[2] J. D. JACKSON, Classical Electrodynamics, third ed., John Wiley & Sons, Inc.,
New York, NY . . . , 1999, ISBN 0-471-30932-X.
[3] L. D. L ANDAU AND E. M. L IFSHITZ, The Classical Theory of Fields, fourth revised English ed., vol. 2 of Course of Theoretical Physics, Pergamon Press, Ltd.,
Oxford . . . , 1975, ISBN 0-08-025072-6.
[4] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism,
second ed., Addison-Wesley Publishing Company, Inc., Reading, MA . . . , 1962,
ISBN 0-201-05702-6.
[5] J. A. S TRATTON, Electromagnetic Theory, McGraw-Hill Book Company, Inc.,
New York, NY and London, 1953, ISBN 07-062150-0.
[6] J. A. W HEELER AND R. P. F EYNMAN, Interaction with the absorber as a mechanism for radiation, Reviews of Modern Physics, 17 (1945), pp. 157.
8
Electromagnetic
Radiation and
Radiating
Systems
In Chapter 3 we were able to derive general expressions for the scalar and vector potentials from which we then, in Chapter 7, calculated 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.
Thus, one can, at least in principle, calculate the radiated fields, Poynting
flux and energy for an arbitrary current density Fourier component and then
add these Fourier components together to construct the complete electromagnetic field at any time at any point in space. However, in practice, it is often
difficult to evaluate the source 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.
107
108
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
L2
L
2
F IGURE 8.1: A linear antenna used for transmission. The current in the
feeder and the antenna wire is set up by the EMF of the generator (the
transmitter). At the ends of the wire, the current is reflected back with a
180 phase shift to produce a antenna current in the form of a standing
wave.
sin[k(L/2 x30 )]
j0 (x0 ) = I0 (x10 )(x20 )
x 3
sin(kL/2)
(8.1)
8.1
R ADIATION
FROM
109
E XTENDED S OURCES
x3 = z
x
L
2
(x0 )
x2
x1
L2
F IGURE 8.2: We choose a spherical polar coordinate system (r = |x| , , )
and orient it so that the linear antenna axis (and thus the antenna current
density j ) is along the polar axis with the feed point at the origin.
V0
j0 k)e
=
dx
I02
2
2
sin[k(L/2 x30 )]
0
k sin eikx3 cos eikx0 cos dx30
I0
sin(kL/2)
L/2
L/2
ik(x0 x0 ) 3 0
4I02
L/2
0
2
(8.2)
Inserting this expression and d = 2 sin d into Formula (7.32) on page 104
110
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
and integrating over , we find that the total radiated power from the antenna
is
P(L) = R0 I02
1
4
sin d
(8.3)
kL0
12
L
R0 I02
(8.4)
P(L) P(L)
= 1 2 = R0
2
6
Ieff
2 I0
L
197
L
(8.5)
1
4
cos2
3 2 cos
sin
(8.6)
cos2
3 2 cos
sin
cos2 2 u4
3 du =
1 1 u2
cos2 . 2 u/ = 1 + cos(u)
2
1
1
1 + cos(u)
=
du
2 1 (1 + u)(1 u)
(8.7)
1 1 1 + cos(u)
1 1 1 + cos(u)
=
du +
du
4 1 (1 + u)
4 1 (1 u)
v
1 1 1 + cos(u)
du = 7 1 + u 8
=
2 1 (1 + u)
1 2 1 cos v
1
=
dv = [ + ln 2 Ci(2)]
2 0
v
2
1.22
d = [cos u] =
where in the last step the Euler-Mascheroni constant = 0.5772 . . . and the
cosine integral Ci(x) were introduced. Inserting this into the expression Equation (8.6) we obtain the value Rrad (/2) 73 .
8.1
R ADIATION
111
E XTENDED S OURCES
FROM
r
x3 = z = z 0
k
x2
z 0
j (x0 )
x1
x0
0
0
F IGURE 8.3: For the loop antenna the spherical coordinate system
(r, , ) describes the field point x (the radiation field) and the cylindrical
coordinate system (0 , 0 , z0 ) describes the source point x0 (the antenna
current).
i0 eik|x|
4 |x|
V0
eikx j k d3x0
(8.8)
In our case the generator produces a single frequncy and we feed the antenna
across a small gap where the loop crosses the positive x1 axis. The circumference of the loop is chosen to be exactly one wavelength = 2c/. This means
that the antenna current oscillates in the form of a sinusoidal standing current
112
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
(8.9)
For the spherical coordinate system of the field point, we recall from subsection F.4.1 on page 163 that the following relations between the base vectors
hold:
r = sin cos x 1 + sin sin x 2 + cos x 3
= cos cos x 1 + cos sin x 2 sin x 3
= sin x 1 + cos x 2
and
(8.10)
(8.11)
0 = sin 0 x 1 + cos 0 x 2
(8.12)
(8.13)
(8.14)
and
With these expressions inserted and d3x0 = 0 d0 d0 dz0 , the source integral
becomes
V0
eikx j k d3x0 = a
2
= I0 ak
+ I0 ak cos
2
0
(8.15)
8.1
R ADIATION
FROM
113
E XTENDED S OURCES
Utilising the periodicity of the integrands over the integration interval [0, 2],
introducing the auxiliary integration variable 00 = 0 , and utilising standard
trigonometric identities, the first integral in the RHS of (8.15) can be rewritten
2
00
0
2
= cos
= cos
=
00
0
2
00
1 1
+ cos 200 d00
2 2
(8.16)
2
1
00
eika sin cos d00
cos
2
0
2
00
1
eika sin cos cos 200 d00
+ cos
2
0
00
2
00
1
eika sin cos d00
sin
2
0
2
00
1
sin
eika sin cos cos 200 d00
2
0
(8.17)
in
ei cos cos n d =
in
2
ei cos cos n d
(8.18)
00
00
00
V0
eikx j k d3x0 = I + I
(8.20)
114
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
i0 eikr
I + I 4
(8.21)
4r 3
To obtain the desired physical magnetic field in the radiation (far) zone we
must Fourier transform back to t space and take the real part and evaluate it at
the retarded time:
Brad
(x) =
Brad (t, x) = Re
i0 e(ikrt )
4
3 I + I
4r
0
0
sin(kr t0 ) I + I 4
=
4r
3
I0 ak0
=
sin(kr t0 ) cos [J0 (ka sin ) J2 (ka sin)]
4r
+ cos sin [J0 (ka sin ) + J2 (ka sin )]
(8.22)
From this expression for the radiated B field, we can obtain the radiated E field
with the help of Maxwells equations.
(8.23)
= jtrue
t
(8.24a)
(8.24b)
8.2
115
M ULTIPOLE R ADIATION
and compare with Equation (8.23) on the preceding page, we see that (t, x)
satisfies this equation of continuity. Furthermore, if we compare with the electric polarisation [cf. Equation (6.9) on page 87], we see that the quantity is
related to the true charges in the same way as P is related to polarised charge.
The quantity is referred to as the polarisation vector since, formally, it treats
also the true (free) charges as polarisation charges.
We introduce a further potential e with the following property
e =
1 e
=A
c2 t
(8.25a)
(8.25b)
where and A are the electromagnetic scalar and vector potentials, respectively. As we see, e 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 potential and, as can be seen from (8.24) and (8.25), it satisfies
the inhomogeneous wave equation
e =
1 2 e
2 e =
2
2
c t
0
(8.26)
This equation is of the same type as Equation (3.19) on page 41, and has
therefore the retarded solution
e (t, x) =
1
40
0 , x0 )
(tret
d3x0
0|
0
x
|x
V
(8.27)
1
40
(x0 )eik|xx | 3 0
dx
|x x0 |
V0
(8.28)
(8.29)
we see that we can calculate the magnetic and electric fields, respectively, as
follows
1 C
c2 t
(8.30a)
E = C
(8.30b)
B=
116
E LECTROMAGNETIC R ADIATION
x x0
AND
x0
x x0
x
R ADIATING S YSTEMS
x0
V0
O
F IGURE 8.4: Geometry of a typical multipole radiation problem where
the field point x is located some distance away from the finite source volume V 0 centered around x0 . If k |x0 x0 | 1 k |x x0 |, then the radiation at x is well approximated by a few terms in the multipole expansion.
Clearly, the last equation is valid only outside the source volume, where
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 illustrated in Figure 8.4.
Under these assumptions, we can expand the Hertz vector, expression (8.27)
0 , x0 ) in the
on the previous page, due to the presence of non-vanishing (tret
vicinity of x0 , in a formal series. For this purpose we recall from potential
theory that
0
eik|(xx0 )(x x0 )|
eik|xx |
|x x0 | |(x x0 ) (x0 x0 )|
(8.31)
x0 )h(1)
n (k |x x0 |)
where
0
eik|xx |
is a Green function
|x x0 |
is the angle between x0 x0 and x x0 (see Figure 8.4 on page 116)
h(1)
n (k |x x0 |) is the spherical Hankel function of the first kind of order n
8.2
117
M ULTIPOLE R ADIATION
0 im( )
(1)m Pmn (cos )Pm
n (cos )e
Pn (cos ) =
(8.32)
m=n
where Pm
n is an associated Legendre polynomial and, in spherical polar coordinates,
x0 x0 = ( x0 x0 , 0 , 0 )
(8.33a)
x x0 = (|x x0 | , , )
(8.33b)
Inserting Equation (8.31) on the facing page, together with Equation (8.32)
above, into Equation (8.28) on page 115, we can in a formally exact way expand the Fourier component of the Hertz vector as
ik n
m
im
(2n + 1)(1)mh(1)
n (k |x x0 |) Pn (cos ) e
40 n=0
m=n
e =
V0
(x ) jn (k x
(8.34)
0 im0 3 0
x0 ) Pm
dx
n (cos ) e
We notice that there is no dependence on x x0 inside the integral; the integrand is only dependent on the relative source vector x0 x0 .
We are interested in the case where the field point is many wavelengths
away from the well-localised sources, i.e., when the following inequalities
k x0 x0 1 k |x x0 |
(8.35)
eik|xx0 |
k |x x0 |
(8.36)
and replace jn with the first term in its power series expansion:
jn (k x0 x0 )
2n n!
k x0 x0 4
(2n + 1)! 3
(8.37)
Inserting these expansions into Equation (8.34) above, we obtain the multipole
expansion of the Fourier component of the Hertz vector
e
e(n)
(8.38a)
n=0
118
E LECTROMAGNETIC R ADIATION
where
e(n) = (i)n
1 eik|xx0 | 2n n!
40 |x x0 | (2n)!
AND
R ADIATING S YSTEMS
V0
eik|xx0 |
40 |x x0 |
V0
(x0 ) d3x0 =
1 eik|xx0 |
p
40 |x x0 |
(8.39)
where p = V 0 (x0 ) d3x0 is the Fourier component of the electric dipole moment p(t, x0 ) = V 0 (x0 x0 )(t, x0 ) d3x0 [cf. Equation (6.2) on page 85 which
describes the static dipole moment]. If a spherical coordinate system is chosen
with its polar axis along p as in Figure 8.5 on the facing page, the components
of e(0) are
1 eik|xx0 |
p cos
40 |x x0 |
1 eik|xx0 |
def
e e(0) =
p sin
40 |x x0 |
def
er e(0) r =
(8.40a)
(8.40b)
def
e e(0) = 0
(8.40c)
Evaluating Formula (8.29) on page 115 for the help vector C, with the
spherically polar components (8.40) of e(0) inserted, we obtain
(0)
C = C,
=
1
40
1
eik|xx0 |
ik
|x x0 | p sin
|x x0 |
(8.41)
8.2
119
M ULTIPOLE R ADIATION
k
x3
Brad
Erad
x2
x1
F IGURE 8.5: If a spherical polar coordinate system (r, , ) is chosesn
such that the electric dipole moment p (and thus its Fourier transform p )
is located at the origin and directed along the polar axis, the calculations
are simplified.
Applying this to Equation (8.30) on page 115, we obtain directly the Fourier
components of the fields
0
1
eik|xx0 |
ik
4 |x x0 |
|x x0 | p sin
1
ik
x x0
1
2
cos
E =
2
40
|x x0 |
|x x0 |
|x x0 |
ik|xx0 |
1
ik
2
e
k
sin
+
|x x0 | p
|x x0 |2 |x x0 |
B = i
(8.42a)
(8.42b)
Keeping only those parts of the fields which dominate at large distances
(the radiation fields) and recalling that the wave vector k = k(x x0 )/ |x x0 |
where k = /c, we can now write down the Fourier components of the radiation
parts of the magnetic and electric fields from the dipole:
0 eik|xx0 |
0 eik|xx0 |
pw k sin =
(p k)
(8.43a)
4 |x x0 |
4 |x x0 |
1 eik|xx0 |
1 eik|xx0 |
p k2 sin =
[(p k) k]
Erad
=
40 |x x0 |
40 |x x0 |
(8.43b)
Brad
=
120
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
These fields constitute the electric dipole radiation, also known as E1 radiation.
eik|xx0 |
k x0 x0 (x0 ) cos d3x0
40 |x x0 | V 0
1 eik|xx0 |
= ik
[(x x0 ) (x0 x0 )] (x0 ) d3x0
40 |x x0 |2 V 0
e(1) = i
(8.44)
(8.45)
and introducing
i = xi x0,i
0i
(8.46a)
xi0 x0,i
(8.46b)
(8.47)
i.e., as the sum of two parts, the first being symmetric and the second antisymmetric in the indices i, j. We note that the antisymmetric part can be written
as
1
1
i , j 0i ,i 0j 4 = [, j (i 0i ) 0j (i ,i )]
2 3
2
1
= [ ( 0 ) 0 ( )] j
2
1
= 1 (x x0 ) [ (x0 x0 )] 2
2
(8.48)
j
The utilisation of Equations (8.24) on page 114, and the fact that we are
considering a single Fourier component,
(t, x) = eit
(8.49)
8.2
121
M ULTIPOLE R ADIATION
allow us to express in j as
= i
(8.50)
Hence, we can write the antisymmetric part of the integral in Formula (8.44)
on the preceding page as
1
(x x0 ) (x0 ) (x0 x0 ) d3x0
2
V0
1
= i (x x0 ) j (x0 ) (x0 x0 ) d3x0
2
V0
1
= i (x x0 ) m
(8.51)
1
2
V0
(8.52)
The final result is that the antisymmetric, magnetic dipole, part of e(1)
can be written
e,antisym
(1)
k eik|xx0 |
(x x0 ) m
40 |x x0 |2
(8.53)
In analogy with the electric dipole case, we insert this expression into Equation (8.29) on page 115 to evaluate C, with which Equations (8.30) on page 115
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
0 eik|xx0 |
(m k) k
4 |x x0 |
k eik|xx0 |
Erad
(x)
=
m k
40 c |x x0 |
Brad
(x) =
(8.54a)
(8.54b)
which are the fields of the magnetic dipole radiation (M1 radiation).
122
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
of the electric quadrupole tensor, which is defined in accordance with Equation (6.3) on page 86:
Q(t, x0 ) =
V0
(8.55)
Again we use this expression in Equation (8.29) on page 115 to calculate the
fields via Equations (8.30) on page 115. Tedious, but fairly straightforward
algebra (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
i0 eik|xx0 |
k Q 4 k
8 |x x0 | 3
i eik|xx0 |
Erad
k Q 4 k k
(x) =
80 |x x0 | 3
Brad
(x) =
(8.56a)
(8.56b)
(t, x) =
(8.57a)
(8.57b)
and consider a source region with such a limited spatial extent that the charges
and currents are well localised. Specifically, we consider a charge q0 , for instance an electron, which, classically, can be thought of as a localised, unstructured and rigid charge distribution with a small, finite radius. The part of this
charge distribution dq0 which we are considering is located in dV 0 = d3x0 in
the sphere in Figure 8.6 on the next page. Since we assume that the electron
(or any other other similar electric charge) is moving with a velocity v 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 (8.57) is
8.3
R ADIATION
FROM A
L OCALISED C HARGE
x(t)
IN
123
A RBITRARY M OTION
dr0
q0
dV 0
F IGURE 8.6: Signals which are observed at the field point x at time t
were generated at source points x0 (t0 ) on a sphere, centred on x and expanding, as time increases, with the velocity c outward from the centre.
The source charge element moves with an arbitrary velocity v and gives
rise to a source leakage out of the source volume dV 0 = d3x0 .
0 , x0 ) d3x0 and
0 , x0 ) d3x0 , respectively, because in the finite time
V (tret
V v(tret
interval during which the observed signal is generated, part of the charge distribution will leak out of the volume element d3x0 .
(x x0 ) v 0 0
dS dt
|x x0 |
(8.58)
where the last term represents the amount of source leakage due to the fact
that the charge distribution moves with velocity v(t 0 ). Since dt0 = dr0 /c and
124
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
dS 0 dr0 = d3x0 we can rewrite this expression for the net charge as
(x x0 ) v 3 0
dx
c |x x0 |
(x x0 ) v
0
= (tret
, x0 ) 1
d3x0
c |x x0 |
0
0
dq0 = (tret
, x0 ) d3x0 (tret
, x0 )
(8.59)
or
0
(tret
, x0 ) d3x0 =
dq0
(8.60)
)v
1 (xx
c|xx0 |
(8.61)
This is the expression to be used in the Formulae (8.57) on page 122 for the
retarded potentials. The result is (recall that j = v)
(t, x) =
1
40
A(t, x) =
0
4
dq0
|x x0 | (xxc0 )v
v dq0
0
|x x0 | (xxc )v
(8.62a)
(8.62b)
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 V dq0 = q0 to evaluate these expressions to
become
1
q0 1
q0
=
0
40 |x x0 | (xxc )v 40 s
v
q0 v v
q0
A(t, x) =
=
= (t, x)
0
40 c2 |x x0 | (xxc )v 40 c2 s c2
(t, x) =
where
s = s(t0 , x) = x x0 (t0 )
(x x0 (t0 )) v(t0 )
c
0 0
x x0 (t0 ) v(t0 )
= x x (t ) 1
|x x0 (t0 )| c
x x0 (t0 ) v(t0 )
= (x x0 (t0 ))
c
|x x0 (t0 )|
(8.63a)
(8.63b)
(8.64a)
(8.64b)
(8.64c)
8.3
R ADIATION
FROM A
L OCALISED C HARGE
IN
125
A RBITRARY M OTION
is the retarded relative distance. The potentials (8.63) are precisely the LinardWiechert potentials which we derived in Section 4.3.2 on page 60 by using a
covariant formalism.
It is important to realise that in the complicated derivation presented here,
the observer is in a coordinate system which has an absolute meaning and the
velocity v is that of the particle, whereas in the covariant derivation two frames
of equal standing were moving relative to each other with v. Expressed in the
four-potential, Equation (4.48) on page 59, the Linard-Wiechert potentials
become
A (x ) =
q0
40
1 v
= (, cA)
,
s cs
(8.65)
(8.66a)
A(t, x)
E(t, x) = (t, x)
t
(8.66b)
(8.67)
(in the interest of simplifying our notation, we drop the subscript ret on t 0
from now on). This means that we know the trajectory of the charge q0 , i.e., x0 ,
for all times up to the time t0 at which a signal was emitted in order to precisely
arrive at the field point x at time t. Because of the finite speed of propagation
of the fields, the trajectory at times later than t 0 is not (yet) known.
The retarded velocity and acceleration at time t 0 are given by
v(t0 ) =
dx0
dt0
a(t0 ) = v (t0 ) =
(8.68a)
dv d2 x0
=
dt0 dt0 2
(8.68b)
126
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
?
|x x0 |
v
c
q0
x0 (t0 )
v(t0 )
x0 (t)
x x0
x x0
x(t)
F IGURE 8.7: Signals which are observed at the field point x at time t
were generated at the source point x0 (t0 ). After time t0 the particle, which
moves with nonuniform velocity, has followed a yet unknown trajectory.
Extrapolating tangentially the trajectory from x0 (t0 ), based on the velocity v(t0 ), defines the virtual simultaneous coordinate x0 (t).
(8.69a)
d2
(x x0 (t0 )) = v(t0 )
dt0 2
(8.69b)
The retarded time t0 can, at least in principle, be calculated from the implicit
relation
t0 = t0 (t, x) = t
|x x0 (t0 )|
c
(8.70)
and we shall see later how this relation can be taken into account in the calculations.
According to Formulae (8.66) on the previous page the electric and magnetic fields are determined via differentiation of the retarded potentials at the
observation time t and at the observation point x. In these formulae the unprimed , i.e., the spatial derivative differentiation operator = x i /xi means
that we differentiate with respect to the coordinates x = (x1 , x2 , x3 ) while keeping t fixed, and the unprimed time derivative operator /t means that we differentiate with respect to t while keeping x fixed. But the Linard-Wiechert
8.3
R ADIATION
FROM A
L OCALISED C HARGE
IN
127
A RBITRARY M OTION
potentials and A, Equations (8.63) on page 124, are expressed in the charge
velocity v(t0 ) given by Equation (8.68a) on page 125 and the retarded relative
distance s(t0 , x) given by Equation (8.64) on page 124. This means that the
expressions for the potentials and A contain terms which are expressed explicitly in t0 , which in turn is expressed implicitly in t via Equation (8.70) on
the preceding page. Despite this complication it is possible, as we shall see
below, to determine the electric and magnetic fields and associated quantities
at the time of observation t. To this end, we need to investigate carefully the
action of differentiation on the potentials.
t0
x x0 (t0 ) =
x x0
|x x0 | t0
x x0 (t0 )4 =
(x x0 ) v(t0 )
(8.71)
|x x0 |
t0
t
t0
|x x0 |
= 1 0
t x c
t
0
0
0
(x x ) v(t ) t
= 1+
t
c |x x0 |
x
= 1
(8.72)
x
|x x0 |
|x x0 |
=
s
|x x0 | (x x0 ) v(t0 )/c
=
x
(8.73)
t
=
x
t0
t
t0
=
x
|x x0 |
s
t0
(8.74)
x
128
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
()t t0 = ()t
(8.75)
x x0
cs
(8.76)
which gives the following operator relation when ()t is acting on an arbitrary
function of t0 and x:
()t = ()t t0
t0
+ ()t0 =
x x0
cs
t0
+ ()t0
(8.77)
With the help of the rules (8.77) and (8.74) we are now able to replace t by t 0
in the operations which we need to perform. We find, for instance, that
1 q0
40 s
0
q
x x0 v(t0 ) x x0 s
t0
40 s2 |x x0 |
c
cs
x
0
0
0 q v(t )
A
A
=
t
t x t 4 s
0 0 0 0 s
q0
=
x x sv(t ) x x v(t ) 0
2
3
40 c s
t
()t =
(8.78a)
(8.78b)
x
Utilising these relations in the calculation of the E field from the LinardWiechert potentials, Equations (8.63) on page 124, we obtain
A(t, x)
t
q0
(x x0 (t0 )) |x x0 (t0 )| v(t0 )/c
=
40 s2 (t0 , x)
|x x0 (t0 )|
E(t, x) = (t, x)
s(t0 , x)
t0
|x x0 (t0 )| v (t0 )
c2
(8.79)
8.3
R ADIATION
FROM A
L OCALISED C HARGE
IN
129
A RBITRARY M OTION
Starting from expression (8.64a) on page 124 for the retarded relative distance
s(t0 , x), we see that we can evaluate (s/t 0 )x in the following way
s
t0
x x0 (x x c) v(t )
x
x
0
1 x x0 (t0 )4 0
t0
c
c
|x x0 |
(8.80)
where Equation (8.71) on page 127 and Equations (8.68) on page 125, respectively, were used. Hence, the electric field generated by an arbitrarily moving
charged particle at x0 (t0 ) is given by the expression
E(t, x) =
q0
|x x0 (t0 )| v(t0 )
0 0
(x
x
(t
))
d
40 s3 (t0 , x) L
MEcN
v2 (t0 )
c2 O
x x0 (t0 )
+
L c2
(x x0 (t0 ))
M[N
|x x0 (t0 )| v(t0 )
0
c
v (t )
O
Radiation field
(8.81)
The first part of the field, the velocity field, tends to the ordinary Coulomb field
when v 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.
From Figure 8.7 on page 126 we see that the position the charged particle
would have had if at t0 all external forces would have been switched off so that
the trajectory from then on would have been a straight line in the direction of
the tangent at x0 (t0 ) is x0 (t), the virtual simultaneous coordinate. During the
arbitrary motion, we interpret x x0 as the coordinate of the field point x relative to the virtual simultaneous coordinate x0 (t). Since the time it takes from a
signal to propagate (in the assumed vacuum) from x0 (t0 ) to x is |x x0 | /c, this
relative vector is given by
x x0 (t) = x x0 (t0 )
|x x0 (t0 )| v(t0 )
c
(8.82)
130
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
This allows us to rewrite Equation (8.81) on the previous page in the following
way
E(t, x) =
v2
(x x0 ) v
q0
(x
x
)
1
+ (x x0 )
0
3
2
40 s
c
c2
(8.83)
0
cs
t
0
0
0
A
xx
q
xx
v
=
0
2
2
0
40 c s |x x |
c |x x |
t x
A
x
(8.84)
where we made use of Equation (8.63) on page 124 and Formula (8.74) on
page 127. But, according to (8.78a),
q0
x x0
x x0
()
=
v
t
c |x x0 |
40 c2 s2 |x x0 |
(8.85)
so that
x x0
A
()t
c |x x0 |
t
0
xx
=
E(t, x)
c |x x0 |
B(t, x) =
(8.86)
The radiation part of the electric field is obtained from the acceleration
field in Formula (8.81) on the preceding page as
Erad (t, x) =
=
lim E(t, x)
|xx0 |
q0
(x x0 )
3
(x x0 )
40 c2 s
q0
=
(x x0 ) [(x x0 ) v ]
40 c2 s3
|x x0 | v
c
v
(8.87)
where in the last step we again used Formula (8.82) on the previous page.
Using this formula and Formula (8.86), the radiation part of the magnetic field
can be written
Brad (t, x) =
x x0
Erad (t, x)
c |x x0 |
(8.88)
8.3
R ADIATION
FROM A
L OCALISED C HARGE
IN
131
A RBITRARY M OTION
2
s2 (t0 , x) = x x0 2 x x0
(x x0 ) v
(x x0 ) v
+
c
c
(8.89)
(x x0 ) v
c
(x x0 ) v
+
c
02 2
|x x0 |2 v2
2 0 |x x | v
cos
+
sin2 0
c2
c2
|x x0 |2 v2
|x x0 |2 v2
2 0
2 0
=
(cos
+
sin
)
=
c2
c2
(8.90)
we find that
(x x0 ) v
c
(x x0 ) v
|x x0 |2 v2
c2
c
(8.91)
Furthermore, from Equation (8.82) on page 129, we obtain the following identity:
(x x0 ) v = (x x0 ) v
(8.92)
which, when inserted into Equation (8.91) above, yields the relation
(x x0 ) v
c
(x x0 ) v
|x x0 |2 v2
c2
c
(8.93)
Inserting the above into expression (8.89) for s2 , this expression becomes
2
s2 = x x 0 2 x x 0
= (x x0 )
(x x0 ) v |x x0 |2 v2
(x x0 ) v
+
2
c
c
c
|x x0 | v
c
(x x0 ) v
c
(x x0 (t)) v
= (x x0 (t))
c
(8.94)
(x x0 ) v
|x x0 |
c
132
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
(x x0 ) v
|x x0 |
c
= |x x0 | 1
=
|x x0 |
(8.95a)
v2 2
sin 0
c2
(8.95b)
v2
(x x0 ) v
1 2 +
c
c
(8.95c)
(x x0 ) v
v2
2
c
c
v2
vv
+ 2 (x x0 )
2
c
c
v
v
= 2 7 (x x0 ) + . (x x0 )/98
c
c
= 2 (x x0 ) 1
(8.96)
T HE
In the special case of uniform motion, the localised charge moves in a field-free, iso-
8.3
R ADIATION
FROM A
L OCALISED C HARGE
IN
133
A RBITRARY M OTION
lated space and we know that it will not be affected by any external forces. It will
therefore move uniformly in a straight line with the constant velocity v. This gives us
the possibility to extrapolate its position at the observation time, x0 (t), from its position at the retarded time, x0 (t0 ). Since the particle is not accelerated, v 0, the virtual
simultaneous coordinate x0 will be identical to the actual simultaneous coordinate of
the particle at time t, i.e., x0 (t) = x0 (t). As depicted in Figure 8.7 on page 126, the
angle between x x0 and v is 0 while then angle between x x0 and v is 0 .
We note that in the case of uniform velocity v, time and space derivatives are closely
related in the following way when they operate on functions of x(t) :
v
t
(8.97)
Hence, the E and B fields can be obtained from Formulae (8.66) on page 125, with the
potentials given by Equations (8.63) on page 124 as follows:
A
1 v
v
= 2
= 2
t
c t
c t
v v
vv
= +
= 1 2
c c
c
vv
= 2 1
c
v
v
v
B = A = 2 = 2 = 2
c
c
c
v
v
vv
v
v
= 2 2 1
= 2
c
c
c
c
c
v
= 2 E
c
E =
ef$
(8.98a)
(8.98b)
Here 1 = x i x i is the unit dyad and we used the fact that v v 0. What remains is just
to express in quantities evaluated at t and x.
From Equation (8.63a) on page 124 and Equation (8.96) on the facing page we find
that
1
q0
q0
=
s2
40
s
80 s3
q0
v
v
=
(x x0 ) +
(x x0 )
3
40 s
c
c
e
%g
(8.99)
When this expression for is inserted into Equation (8.98a) above, the following
result
134
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
vv
vv
q0
1
1 s2
c2
80 s3 c2
v
v
q0
(x x0 )
(x x0 ) +
=
40 s3
c
c
E(t, x) =
v v
vv v
v
(x x0 ) 2
(x x0 )
c c
c
c
c
% gi
q0
v v
v2
(x x0 ) (x x0 ) 2
(x x0 ) +
3
40 s
c c
c
v v
(x x0 )
c c
q0
40 s
(x x0 ) 1
3
(8.100)
v2
c2
follows. Of course, the same result also follows from Equation (8.83) on page 130
with v 0 inserted.
From Equation (8.100) above we conclude that E is directed along the vector from the
simultaneous coordinate x0 (t) to the field (observation) coordinate x(t). In a similar
way, the magnetic field can be calculated and one finds that
B(t, x) =
0 q 0
4s3
v2
c2
v (x x0 ) =
1
vE
c2
(8.101)
From these explicit formulae for the E and B fields we can discern the following cases:
1. v 0 E goes over into the Coulomb field ECoulomb
2. v 0 B goes over into the Biot-Savart field
3. v c E becomes dependent on 0
E XAMPLE 8.2
T HE
OF EXAMPLE
8.1C
Let us consider in more detail the treatment of the radiation from a uniformly moving
rigid charge distribution.
If we return to the original definition of the potentials and the inhomogeneous wave
equation, Formula (3.19) on page 41, for a generic potential component (t, x) and a
generic source component f (t, x),
(t, x) =
1 2
2 (t, x) = f (t, x)
c2 t2
(8.102)
8.3
R ADIATION
FROM A
L OCALISED C HARGE
IN
135
A RBITRARY M OTION
we find that under the assumption that v = v x 1 , this equation can be written
v2
c2
2 2 2
+
+
= f (x)
x21 x22 x23
(8.103)
(8.104a)
(8.104b)
(8.104c)
def
1 v2 /c2 1 , 2 , 3 ) f ()
(8.105)
in this space. This equation has the well-known Coulomb potential solution
1
f ( 0 ) 3 0
d
() =
4 V 0
(8.106)
(8.107)
2
v2
[(x2 x02 )2 + (x3 x03 )2 ]
(8.108)
2
c
Applying this to the explicit scalar and vector potential components, realising that for
a rigid charge distribution moving with velocity v the current is given by j = v, we
obtain
(x0 ) 3 0
1
dx
(8.109a)
(t, x) =
40 V s
1
v(x0 ) 3 0 v
A(t, x) =
d x = 2 (t, x)
(8.109b)
2
40 c V s
c
s = (x1 x01 )2 + 1
136
E LECTROMAGNETIC R ADIATION
R ADIATING S YSTEMS
AND
(8.111)
With the help of Equation (8.101) on page 134 and Equations (8.109) on the preceding
page, and the fact that t = v [cf.. Formula (8.97) on page 133], we can rewrite
expression (8.111) as
v
v
v
v
v
F = q E+v 2 E = q
(8.112)
c
c
c
c
c
Applying the bac-cab rule, Formula (F.54) on page 164, on the last term yields
%g
e)$
%g
v
v
v v2
v
=
c
c
c
c c2
which means that we can write
(8.113)
F = q
(8.114)
where
= 1
v2
c2
(8.115)
The scalar function is called the convection potential or the Heaviside potential.
When the rigid charge distribution is well localised so that we can use the potentials
(8.110) the convection potential becomes
= 1
v2
c2
q0
40 s
(8.116)
x1 x01
1 v2 /c2
(8.117)
OF EXAMPLE
8.2C
8.3
R ADIATION
FROM A
L OCALISED C HARGE
137
A RBITRARY M OTION
IN
s = x x0
(x x0 ) v
x x0 ,
c
vc
(8.118)
|x x0 | v
x x0 ,
c
vc
(8.119)
so that the radiation field Equation (8.87) on page 130 can be approximated by
Erad (t, x) =
q0
(x x0 ) [(x x0 ) v ],
3
2
0
40 c |x x |
vc
(8.120)
from which we obtain, with the use of Formula (8.86) on page 130, the magnetic field
Brad (t, x) =
q0
[v (x x0 )],
40 c3 |x x0 |2
vc
(8.121)
(8.122)
x x = x x0
(8.123a)
(8.123b)
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
S=
0
0 q0 2 (v)2
2 xx
sin
|x x0 |
162 c |x x0 |2
(8.124)
q0 2 v 2
0 q0 2 (v)2
=
6c
60 c3
(8.125)
where is the angle between v and x x0 . The total radiated power (integrated
over a closed spherical surface) becomes
P=
138
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
which is the Larmor formula for radiated power from an accelerated charge.
Note that here we are treating a charge with v c but otherwise totally unspecified motion while we compare with formulae derived for a stationary oscillating dipole. The electric and magnetic fields, Equation (8.120) on the
preceding page and Equation (8.121) on the previous 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 x0 (t0 ). The angular distribution is that which is frozen to the point from
which the energy is radiated.
8.3.3 Bremsstrahlung
An important special case of radiation is when the velocity v and the acceleration v are collinear (parallel or anti-parallel) so that v v = 0. This condition
(for an arbitrary magnitude of v) inserted into expression (8.87) on page 130
for the radiation field, yields
Erad (t, x) =
q0
(x x0 ) [(x x0 ) v ],
40 c2 s3
v k v
(8.126)
from which we obtain, with the use of Formula (8.86) on page 130, the magnetic field
Brad (t, x) =
q0 |x x0 |
[v (x x0 )],
40 c3 s3
v k v
(8.127)
The difference between this case and the previous case of v c is that the
approximate expression (8.118) on the preceding page for s is no longer valid;
we must instead use the correct expression (8.64) on page 124. The angular
distribution of the power flux (Poynting vector) therefore becomes
S=
sin2
0 q0 2 v 2
162 c |x x0 |2 1 v cos 4
c
x x0
|x x0 |
(8.128)
It is interesting to note that the magnitudes of the electric and magnetic fields
are the same whether v and v are parallel or anti-parallel.
We must be careful when we compute the energy (S integrated over time).
The Poynting vector is related to the time t when it is measured and to a fixed
surface in space. The radiated power into a solid angle element d, measured
relative to the particles retarded position, is given by the formula
dU rad ()
0 q0 2 v 2
sin2
d = S (x x0 ) x x0 d =
dt
162 c 1 v cos 4
c
3
d (8.129)
8.3
R ADIATION
FROM A
L OCALISED C HARGE
IN
139
A RBITRARY M OTION
v = 0.5c
v = 0.25c
v=0
v
F IGURE 8.8: Polar diagram of the energy loss angular distribution factor
sin2 /(1 v cos /c)5 during bremsstrahlung for particle speeds v = 0, v =
0.25c, and v = 0.5c.
On the other hand, the radiation loss due to radiation from the charge at retarded time t0 :
dU rad
dU rad
d
=
dt0
dt
t
t0
(8.130)
(8.131)
(8.132)
The angular factors of this expression, for three different particle speeds, are
plotted in Figure 8.8.
Comparing expression (8.129) on the preceding page with expression (8.132), we see that they differ by a factor 1 v cos /c which comes
from the extra factor s/ |x x0 | introduced in (8.131). Let us explain this in
geometrical terms.
During the interval (t0 , t0 + dt0 ) and within the solid angle element d the
particle radiates an energy [dU rad ()/dt0 ] dt0 d. As shown in 8.9 this energy
140
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
dS
dr
x
d
0
q
0
x02 vdt x01
x x02 + c dt0
is at time t located between two spheres, one outer with its origin in x01 (t0 ) and
one inner with its origin in x01 (t0 + dt0 ) = x01 (t0 ) + v dt and radius c[t (t0 + dt0 )] =
c(t t0 dt0 ).
From Figure 8.9 we see that the volume element subtending the solid angle
element
d =
dS
2
x x02
is
(8.133)
2
d3x = dS dr = x x02 d dr
(8.134)
Here, dr denotes the differential distance between the two spheres and can be
evaluated in the following way
0
x x20 v dt0 x x02
L x MExN 2 O
dr = x x02 + c dt0
v cos
=
x x02
x x02
dt0 =
(8.135)
cs
dt0
x x02
8.3
R ADIATION
L OCALISED C HARGE
FROM A
IN
141
A RBITRARY M OTION
where Formula (8.64) on page 124 was used in the last step. Hence, the volume
element under consideration is
s
dS c dt0
d3x = dS dr =
(8.136)
x x02
We see that the energy which is radiated per unit solid angle during the time
interval (t0 , t0 + dt0 ) is located in a volume element whose size is dependent.
This explains the difference between expression (8.129) on page 138 and expression (8.132) on page 139.
Let the radiated energy, integrated over , be denoted U rad . After tedious,
but relatively straightforward integration of Formula (8.132) on page 139, one
obtains
dU rad 0 q0 2 v 2
=
dt0
6c
. 1 /
v2
c2
v2
2 q0 2 v 2
1
3 40 c3
c2
(8.137)
If we know v(t0 ), we can integrate this expression over t 0 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 8.3
(8.138)
v dt
(8.139)
s x x0
(8.140)
(8.141)
From the general expression (8.86) on page 130 we conclude that E B and that
it suffices to consider E Erad . According to the bremsstrahlung expression for
Erad , Equation (8.126) on page 138,
E=
q0 sin
v (t0 t0 )
40 c2 |x x0 |
(8.142)
142
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
E
c
(8.143)
Fourier transforming expression (8.142) on the previous page for E is trivial, yielding
E =
q0 sin
v eit0
82 0 c2 |x x0 |
(8.144)
dU rad 0
dt =
dt0
1
0
S
EB dt0 d2x0 =
S
= 0 c
S E
1
0 c
B
0
dS dt0
S E
dt0 d2x0
(8.145)
dt0 d2x0
According to Parsevals identity [cf. Equation (7.35) on page 105] the following equality holds:
E 2 dt0 = 4
0
|E |2 d
(8.146)
S |E |2 d2x
(8.147)
q0 2 (v)2
163 0 c3
q0 2
30 c
S
0
v
c
sin2 2
d x d
|x x0 |2
d
2
0
sin2 sin d d
(8.148)
d
2
We see that the energy spectrum U rad is independent of frequency . This means that
if we integrate it over all frequencies [0, ], a divergent integral would result.
In reality, all spectra have finite widths, with an upper cutoff limit set by the quantum
8.3
R ADIATION
FROM A
L OCALISED C HARGE
IN
143
A RBITRARY M OTION
condition
1
(8.149)
h = m(v)2
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 h . If we adopt the picture that the total energy is quantised in terms of N
photons radiated during the process, we find that
U rad d
= dN
h
or, for an electron where q0 = |e|, where e is the elementary charge,
dN =
2
e2
40 h c 3
v
c
d
1 2
137 3
v
c
(8.150)
(8.151)
where we used the value of the fine structure constant = e2 /(40 h c) 1/137.
Even if the number of photons becomes infinite when 0, these photons have
negligible energies so that the total radiated energy is still finite.
E ND OF
EXAMPLE
8.3C
(8.152a)
0 0
x (t ) = a[ x 1 cos (t ) + x 2 sin (t )]
v(t0 ) = x 0 (t0 ) = a0 [ x 1 sin (t0 ) + x 2 cos (t0 )]
v = |v| = a0
a20
(8.152b)
(8.152c)
(8.152d)
(8.152e)
(8.152f)
x x0 = x x0 ( x 2 sin + x 3 cos )
(8.153)
144
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
x2
(t, x)
x x0
x
v
q0 0 0
(t , x )
(t0 )
x1
x3
F IGURE 8.10: Coordinate system for the radiation from a charged particle at x0 (t0 ) in circular motion with velocity v(t 0 ) along the tangent and
constant acceleration v (t0 ) toward the origin. The x1 x2 axes are chosen so
that the relative field point vector x x0 makes an angle with the x3 axis
which is normal to the plane of the orbital motion. The radius of the orbit
is a.
where is the angle between x x0 and the normal to the plane of the particle
orbit (see Figure 8.10). From the above expressions we obtain
(x x0 ) v = x x0 v sin cos
(x x0 ) v = x x0 v sin sin = x x0 v cos
(8.154a)
(8.154b)
where in the last step we simply used the definition of a scalar product and the
fact that the angle between v and x x0 is .
The power flux is given by the Poynting vector, which, with the help of
Formula (8.86) on page 130, can be written
S=
1
1
x x0
(E B) =
|E|2
0
c0
|x x0 |
(8.155)
(8.156)
8.3
R ADIATION
FROM A
L OCALISED C HARGE
IN
145
A RBITRARY M OTION
where the retarded distance s is given by expression (8.64) on page 124. With
the radiation part of the electric field, expression (8.87) on page 130, inserted,
and using (8.154a) and (8.154b) on the facing page, one finds, after some algebra, that
4
2
2
v
v
dU rad (, ) 0 q0 2 v 2 3 1 c sin cos . 1 c2 / sin sin
=
(8.157)
5
dt0
162 c
1 v sin cos 4
2
The angles and vary in time during the rotation, so that refers to a moving
coordinate system. But we can parametrise the solid angle d in the angle
and the (fixed) angle so that d = sin d d. Integration of Equation (8.157)
above over this d gives, after some cumbersome algebra, the angular integrated expression
dU rad 0 q0 2 v 2
=
dt0
6c
. 1 vc /
2
2
(8.158)
Cyclotron radiation
For a non-relativistic speed v c, Equation (8.157) above reduces to
dU rad (, ) 0 q0 2 v 2
(1 sin2 sin2 )
=
dt0
162 c
(8.159)
(8.160)
where is defined in Figure 8.10 on the preceding page. This means that we
can write
0 q0 2 v 2
dU rad () 0 q0 2 v 2
2
=
(1
cos
)
=
sin2
dt0
162 c
162 c
(8.161)
Consequently, a fixed observer near the orbit plane will observe cyclotron
radiation twice per revolution in the form of two equally broad pulses of radiation with alternating polarisation.
146
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
x2
(t, x)
x x0
v
q0 0 0
(t , x )
v
(t0 )
0
x1
x3
F IGURE 8.11: When the observation point is in the plane of the particle
orbit, i.e., = /2 the lobe width is given by .
Synchrotron radiation
When the particle is relativistic, v o c, the denominator in Equation (8.157)
on the preceding page becomes very small if sin cos 1, which defines the
forward direction of the particle motion ( /2, 0). Equation (8.157) on
the previous page then becomes
dU rad (/2, 0) 0 q0 2 v 2
1
=
0
2
dt
16 c 1 v 4
c
(8.162)
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 (8.161) on the preceding page and
Equation (8.162) 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 ( = /2), Equation (8.157) on the preceding page gives
4
2
v
v
dU rad (/2, ) 0 q0 2 v 2 3 1 c cos . 1 c2 / sin
=
5
dt0
162 c
1 v cos 4
2
(8.163)
8.3
R ADIATION
FROM A
L OCALISED C HARGE
IN
147
A RBITRARY M OTION
cos 0 =
sin 0 =
(8.164a)
1
v2
c2
(8.164b)
1
2
1 vc2
1,
v2
1,
c2
(8.165)
v2 1
=
c2
(8.166)
Hence, synchrotron radiation from ultra-relativistic charges is characterized by a radiation lobe width which is approximately
(8.167)
This angular interval is swept by the charge during the time interval
t0 =
(8.168)
(8.169)
in the direction toward the observer who therefore measures a pulse width of
length
l
vt0
v
v
v 1
= t0
= . 1 / t0 = . 1 /
. 1 /
c
c
c
c 0
c 0
v4
v4
2
1 c 1+ c 1
v
1
1 1
3v
= 3
1 2
= 3
0 L MEN c O 20 2 0
1+
L MEN cO
1/2
2
t = t0
148
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
(8.170)
As a general rule, the spectral width of a pulse of length t is o 1/t.
In the ultra-relativistic synchrotron case one can therefore expect frequency
components up to
1
= 23 0
(8.171)
t
A spectral analysis of the radiation pulse will exhibit Fourier components n 0
from n = 1 up to n 23 .
When N electrons are contributing to the radiation, we can discern between
three situations:
max
1. All electrons are very close to each other so that the individual phase
differences are negligible. The power will be multiplied by N 2 relative
to a single 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
(x x0 )
|x x0 | v
c
v
(8.172)
Integration over the solid angle gives the totally radiated power as
2
2
dU rad 0 q0 2 v 2 1 vc2 sin
=
3
dt0
6c
. 1 cv22 /
(8.173)
8.3
R ADIATION
L OCALISED C HARGE
FROM A
vt
IN
q0
149
A RBITRARY M OTION
v = v x
|x x0 |
B
E z
F IGURE 8.12: The perpendicular field of a charge q0 moving with velocity v = v x is E z .
Virtual photons
According to Formula (8.100) on page 134 and Figure 8.12,
E = Ez =
q0
40 s3
v2
(x x0 ) x 3
c2
(8.174)
b
q0
40 2 (vt)2 + b2 /2
(8.175)
3/2
This 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
energy. Fourier transforming, we obtain
E, =
1
2
E (t) eit dt =
q
42
0 bv
b
K1
v
b
v
(8.176)
Here, K1 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
E,
q
42
E, 0,
0 bv
b v
b v
(8.177a)
(8.177b)
150
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
2 3
E
d x = 0
bmax
bmin
2
E
vdt 2b db
(8.178)
where the volume integration is over the plane perpendicular to v. With the
use of Parsevals identity for Fourier transforms, Formula (7.35) on page 105,
we can rewrite this as
U=
q2
2
2 0 v
v/
bmin
bmax
U d = 40 v
2
E, d 2b db
(8.179)
bmin
db
d
b
q2
v
ln
2
2 0 v
bmin
(8.180)
2
c
d
ln
bmin
(8.181)
d
c
2
ln .
p1 p01 /
(8.182)
E1 cp1 cp01
m0 c2 E1 E10
(8.183)
8.3
R ADIATION
FROM A
L OCALISED C HARGE
IN
151
A RBITRARY M OTION
(8.184)
B = (t, x)H = m 0 H
(8.185)
E+ B = 0
t
B = 0
H D = j(t, x)
t
(8.186a)
(8.186b)
(8.186c)
(8.186d)
Assuming for simplicity that the electric permittivity and the magnetic
permeability , and hence the relative permittivity and the relative permeability m all have fixed values, independent on time and space, for each type
of material we consider, we can derive the general telegraphers equation [cf.
Equation (2.33) on page 29]
E
2 E
2 E
=0
2
t
t2
(8.187)
=0
2
t2
(8.188)
0 m 0
m
(8.189)
152
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
i = 1, 2, 3
(8.190)
(8.191)
k k k = kv =
v
v v
(8.192)
(8.193)
where now k is the wave vector in the medium given by Equation (8.192).
With these definitions, the vacuum formula for the associated magnetic field,
Equation (2.40) on page 30,
1
1
kE = kE
B = k E =
v
(8.194)
(8.195)
v =
8.3
R ADIATION
FROM A
L OCALISED C HARGE
IN
153
A RBITRARY M OTION
where, in the last step, we used Equation (8.192) on the facing page.
If the medium has a refractive index which, as is usually the case, dependent on frequency , we say that the medium is dispersive. Because in this
case also k() and (k), so that the group velocity
vg =
(8.196)
has a unique value for each frequency component, and is different from v .
Except in regions of anomalous dispersion, v is always smaller than c. In
a gas of free charges, such as a plasma, the refractive index is given by the
expression
2p
2
(8.197)
N q2
0 m
(8.198)
n2 () = 1
where
2p =
is the plasma frequency. Here m and N denote the mass and number density,
respectively, of charged particle species . In an inhomogeneous plasma, N =
N (x) so that the refractive index and also the phase and group velocities 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
is such that it coincides with p at some point in the medium, then at that
point v while vg 0 and the wave Fourier component at is reflected
there.
Vavilov-Cerenkov
radiation
As we saw in Subsection 8.1, a charge in uniform, rectilinear motion in vacuum does not give rise to any radiation; see in particular Equation (8.98a) on
page 133. Let us now consider a charge in uniform, rectilinear motion in a
medium with electric properties which are different from those of a (classical)
vacuum. Specifically, consider a medium where
= Const > 0
(8.199a)
= 0
(8.199b)
c
1
=
<c
n
0
(8.200)
154
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
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 |v|, which, of course, are below the phase speed
in vacuum, can experience that the particle speeds are higher than the phase
A(t, x) =
q0
1 q0
1
=
40 |x x0 | n (xx0 )v 40 s
c
1
q0 v
1 q0 v
=
40 c2 |x x0 | n (xx0 )v 40 c2 s
c
(8.201a)
(8.201b)
where now
(x x0 ) v
0
s = xx n
(8.202)
The need for the absolute value of the expression for s is obvious in the case
when v/c 1/n because then the second term can be larger than the first term;
if v/c 1/n 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 43]
0
0
tret
= tret
(t, x x0 ) = t
k |x x0 |
|x x0 | n
= t
c
0
k |x x0 |
|x x0 | n
0
0
tadv = tadv (t, x x ) = t +
= t+
(8.203a)
(8.203b)
so that the usual time interval t t 0 between the time measured at the point of
observation and the retarded time in a medium becomes
t t0 =
|x x0 | n
c
(8.204)
8.3
R ADIATION
FROM A
L OCALISED C HARGE
IN
A RBITRARY M OTION
q0
155
x(t)
x0 (t0 )
For v/c 1/n, the retarded distance s, and therefore the denominators in
Equations (8.201) on the preceding page vanish when
n(x x0 )
nv
v
= x x0
cos c = x x0
c
c
(8.205)
c
nv
(8.206)
In the direction defined by this angle c , the potentials become singular. During
the time interval t t0 given by expression (8.204) on the facing page, the field
exists within a sphere of radius |x x0 | around the particle while the particle
moves a distance
l = v(t t0 )
(8.207)
c
nv
(8.208)
156
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
The Vavilov-Cerenkov
cone is similar in nature to the Mach cone in acoustics.
In order to make some quantitative estimates of this radiation, we note that
we can describe the motion of each charged particle q0 as a current density:
j = q0 v (x0 vt0 ) = q0 v (x0 vt0 )(y0 )(z0 ) x 1
(8.209)
40 nc
(j k)e
ikx0 3 0
V
q0 2 n2
=
exp i
163 0 c3
dx
2
2
x0
kx0 cos
dx0 sin2 d
(8.212)
in 1934,
who was then a post-graduate student in S. I. Vavilovs research group at the Lebedev Institute
in Moscow. Vavilov wrote a manuscript with the experimental findings, put Cerenkov
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,
of the Cerenkov
effect [V. L. Ginzburg, private communication].
The first observation of this type of radiation was reported by Marie Curie in 1910, but she
never pursued the exploration of it [7].
8.3
R ADIATION
L OCALISED C HARGE
FROM A
IN
157
A RBITRARY M OTION
where is the angle between the direction of motion, x 01 , and the direction to
The integral in (8.212) is singular of a Dirac delta type. If
the observer, k.
we limit the spatial extent of the motion of the particle to the closed interval
[X, X] on the x0 axis we can evaluate the integral to obtain
4 X
q0 2 n2 sin2 sin2 1 nv
c cos v
3
d
2
43 0 c3
1 nvc cos 4 v
Urad d =
(8.213)
3
q0 2 n2 sin2 c
22 0 c3
sin2
7p. 1 + nvc /
7E. 1 + nvc / v 8
X
v
2
Urad () d
(8.214)
sin c
1
1
sin2
7E. 1 + nvc /
7p. 1 + nvc / v 8
X
v
2
d 1
=
c2
cX
n2 v2 n
sin2 x
dx
2
x
(8.215)
c2
cX
1 n2 v2
n
leading to the final approximate result for the total energy loss in the frequency
interval (, + d)
q0 2 X
U rad d =
20 c2
c2
d
n2 v2
(8.216)
c2
d
n2 ()v2
(8.217)
158
E LECTROMAGNETIC R ADIATION
AND
R ADIATING S YSTEMS
This result was derived under the assumption that v/c > 1/n(), i.e., under the
condition that the expression inside the parentheses in the right hand side is
positive. For all media it is true that n() 1 when , so there exist al
ways a highest frequency for which we can obtain Vavilov-Cerenkov
radiation
from a fast charge in a medium. Our derivation above for a fixed value of n is
valid for each individual Fourier component.
Bibliography
[1] R. B ECKER, Electromagnetic Fields and Interactions, Dover Publications, Inc.,
New York, NY, 1982, ISBN 0-486-64290-9.
[2] M. B ORN AND E. W OLF, Principles of Optics. Electromagnetic Theory of Propagation, Interference and Diffraction of Light, sixth ed., Pergamon Press, Oxford,. . . , 1980, ISBN 0-08-026481-6.
[3] V. L. G INZBURG, Applications of Electrodynamics in Theoretical Physics
and Astrophysics, Revised third ed., Gordon and Breach Science Publishers,
New York, London, Paris, Montreux, Tokyo and Melbourne, 1989, ISBN 288124-719-9.
[4] J. D. JACKSON, Classical Electrodynamics, third ed., John Wiley & Sons, Inc.,
New York, NY . . . , 1999, ISBN 0-471-30932-X.
[5] J. B. M ARION AND M. A. H EALD, Classical Electromagnetic Radiation, second ed., Academic Press, Inc. (London) Ltd., Orlando, . . . , 1980, ISBN 0-12472257-1.
[6] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism,
second ed., Addison-Wesley Publishing Company, Inc., Reading, MA . . . , 1962,
ISBN 0-201-05702-6.
[7] J. S CHWINGER , L. L. D E R AAD , J R ., K. A. M ILTON , AND W. T SAI, Classical
Electrodynamics, Perseus Books, Reading, MA, 1998, ISBN 0-7382-0056-5.
[8] J. A. S TRATTON, Electromagnetic Theory, McGraw-Hill Book Company, Inc.,
New York, NY and London, 1953, ISBN 07-062150-0.
[9] J. VANDERLINDE, Classical Electromagnetic Theory, John Wiley & Sons, Inc.,
New York, Chichester, Brisbane, Toronto, and Singapore, 1993, ISBN 0-47157269-1.
F
Formulae
(F.1)
B = 0
(F.2)
E =
(F.3)
B
t
H = j+ D
t
(F.4)
Constitutive relations
D = E
B
H=
(F.5)
j = E
(F.7)
P = 0 E
(F.8)
(F.6)
(F.9)
A
t
(F.10)
159
160
F ORMULAE
1
=0
c2 t
(F.11)
(F.12)
(F.13)
k E
c
(F.14)
Brad
(x) =
(F.15)
(F.16)
(F.17)
(F.18)
F.3
161
S PECIAL R ELATIVITY
Brad
(x) =
(F.19)
(F.20)
Brad
(x) =
(F.21)
(F.22)
x
)
1
+ (x x0 )
0
40 s3
c2
c2
E(t, x)
B(t, x) = (x x0 )
c|x x0 |
E(t, x) =
s = x x0 (x x0 )
v
c
x x0 = (x x0 ) |x x0 |
t0
t
=
x
|x x0 |
s
(F.23)
(F.24)
(F.25)
v
c
(F.26)
(F.27)
(x
x
)
0
40 s3
c2
v E(t, x)
B(t, x) =
c2
E(t, x) =
s=
|x x0 |2
(x x0 ) v
c
(F.28)
(F.29)
(F.30)
162
F ORMULAE
@AAB
CEDD
1 0
0
0
0 1 0
0
0 0 1 0 F
0 0
0 1
(F.31)
(F.32)
CEDD
0 0
0 0
= @A
AB 0 0 1 0F
0
0
0 1
1
= :
1 2
v
=
c
(F.33)
(F.34)
(F.35)
(F.36)
dt
= c d
(F.37)
F.3.5 Four-velocity
u =
dx
= (c, v)
d
(F.38)
F.3.6 Four-momentum
p = m0 cu =
E
,p
c
(F.39)
F.4
163
V ECTOR R ELATIONS
(F.40)
F.3.8 Four-potential
A =
,A
c
(F.41)
CEDD
@AAB
0
E x /c Ey /c Ez /c
E x /c
0
Bz
By
Ey /c
Bz
0
B x F
Ez /c By
Bx
0
(F.42)
x i
i=1
def
def
x i
xi
xi
(F.43)
,
,
=
x1 x2 x3
, ,
x y z
(F.44)
= sin x 1 + cos x 2
(F.45a)
(F.45b)
(F.45c)
164
F ORMULAE
(F.46a)
(F.46b)
x 3 = cos r sin
(F.46c)
(F.47)
(F.48)
(F.49)
Volume element
d3x = dV = drdS = r2 dr d
(F.50)
(F.51)
a (b c) = (a b) c
(F.53)
a b = b a = i jk a j bk x i
(F.52)
a (b c) = b(a c) c(a b)
(F.54)
(F.56)
a (b c) + b (c a) + c (a b) = 0
(a b) (c d) = (a b d)c (a b c)d
(F.55)
(F.57)
F.4
165
V ECTOR R ELATIONS
(F.58)
(a) = a + a
(F.59)
(a b) = b ( a) a ( b)
(F.61)
(a) = a a
(F.60)
(a b) = a( b) b( a) + (b )a (a )b
(F.62)
= 2
(F.64)
(a b) = a ( b) + b ( a) + (b )a + (a )b
= 0
( a) = 0
(F.63)
(F.65)
(F.66)
( a) = ( a) a
(F.67)
Special identities
In the following k is an arbitrary constant vector while x, as before, is the
radius vector and a an arbitrary vector field.
x = 3
(F.68)
(k x) = k
x
|x| =
|x|
1
x
= 3
|x|
|x|
x
1
= 2
= 4(x)
3
|x|
|x|
k
1
kx
= k
= 3
|x|
|x|
|x|
x
kx
=
if |x| 6= 0
k
|x|3
|x|3
k
1
= k2
= 4k(x)
2
|x|
|x|
(k a) = k( a) + k ( a) (k a)
(F.70)
x = 0
(F.69)
(F.71)
(F.72)
(F.73)
(F.74)
(F.75)
(F.76)
(F.77)
166
F ORMULAE
Integral relations
Let V(S ) be the volume bounded by the closed surface S (V). Denote the 3dimensional volume element by d3x( dV) and the surface element, directed
along the outward pointing surface normal unit vector n,
by dS( d2x n).
Then
( a) d3x =
() d3x =
dS a
dS
(F.78)
(F.79)
( a) d3x =
dS a
(F.80)
If S (C) is an open surface bounded by the contour C(S ), whose line element is dl, then
C dl =
C a dl =
dS
dS ( a)
(F.81)
(F.82)
Bibliography
[1] G. B. A RFKEN AND H. J. W EBER, Mathematical Methods for Physicists, fourth,
international ed., Academic Press, Inc., San Diego, CA . . . , 1995, ISBN 0-12059816-7.
[2] P. M. M ORSE AND H. F ESHBACH, Methods of Theoretical Physics, Part I.
McGraw-Hill Book Company, Inc., New York, NY . . . , 1953, ISBN 07-043316-8.
[3] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism,
second ed., Addison-Wesley Publishing Company, Inc., Reading, MA . . . , 1962,
ISBN 0-201-05702-6.
M
Mathematical
Methods
M.1 Scalars, Vectors and Tensors
Every physical observable can be described by a geometric object. We will
describe the observables in classical electrodynamics mathematically in terms
of scalars, pseudoscalars, vectors, pseudovectors, tensors or pseudotensors and
will not exploit differential forms to any significant degree.
A scalar describes a scalar quantity which 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 commonplace. The difference between a scalar, vector and tensor and a
pseudoscalar, pseudovector and a pseudotensor is that the latter behave differently under such coordinate transformations which cannot be reduced to pure
rotations.
Throughout we adopt the convention that Latin indices i, j, k, l, . . . 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 , , , , . . . ,
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
suitable representation is in terms of an ordered N-tuple, or row vector, of the
coordinates xN where N is the dimensionality of the space under consideration. The most basic vector is the radius vector which is the vector from the
origin to the point of interest. Its N-tuple representation simply enumerates the
coordinates which describe this point. In this sense, the radius vector from the
origin to a point is synonymous with the coordinates of the point itself.
167
168
M ATHEMATICAL M ETHODS
x = x i xi x i xi
def
(M.1)
i=1
xi (x1 , x2 , x3 ) (x, y, z)
(M.2)
x (x0 , x1 , x2 , x3 )
(M.3)
x (x0 , x1 , x2 , x3 )
(M.4)
x0
dx
x
(M.5)
M.1
S CALARS , V ECTORS
AND
169
T ENSORS
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 which enumerates the coordinates 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
by
def
(M.6)
(x0 , x1 , x2 , x3 ) (x )
(M.7)
which indicates that the four-scalar 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 the transformation rule, Equation (M.5) on the preceding
page, for the differential dx , the transformation rule for the differential operator /x under a transformation x x0 becomes
x
=
x0 x0 x
(M.8)
Vector fields
We can represent an arbitrary vector field a(x) in
as follows:
a(x) = x i ai (x)
(M.9)
(M.10)
170
M ATHEMATICAL M ETHODS
(M.11)
(M.12)
x0
y
x
(M.13)
x
y
x0
(M.14)
Tensor fields
We denote an arbitrary tensor field in 3 by A(x). This tensor field can be
represented in a number of ways, for instance in the following matrix form:
A(x) =
B@
Ai j (xk ) 4
def
(M.15)
where, in the last member, we have again used the more compact component
notation. Strictly speaking, the tensor field described here is a tensor of rank
two.
M.1
S CALARS , V ECTORS
AND
171
T ENSORS
0 if i 6= j
1 if i = j
(M.16)
B@
(i j ) =
1 0 0
0 1 0F
0 0 1
(M.17)
i jk =
>
=?
<=
1
0
1
(M.18)
(M.19)
In fact, tensors may have any rank n. In this picture a scalar is considered
to be a tensor of rank n = 0 and a vector a tensor of rank n = 1. Consequently,
the notation where a vector (tensor) is represented in its component form is
called the tensor notation. A tensor of rank n = 2 may be represented by a twodimensional array or matrix whereas higher rank tensors are best represented
in their component forms (tensor notation).
T ENSORS IN 3D SPACE
E XAMPLE M.1
(M.20)
Using (M.20) and Newtons second law, we find that the matter of mass m, which at a
172
M ATHEMATICAL M ETHODS
x3
d2x
x2
V
x1
F IGURE M.1: Terahedron-like volume element V containing matter.
given instant is located in V obeys the equation of motion
T n d2x cos 1 T x 1 d2x cos 2 T x 2 d2x cos 3 T x 3 d2x + Fext = ma
(M.21)
where Fext is the external force and a is the acceleration of the volume element. In
other words
m
Fext
T n = n1 T x 1 + n2 T x 2 + n3 T x 3 + 2 a
(M.22)
dx
m
m/d2x
T n = ni T x i ni T x i
(M.23)
i=1
From the above derivation it is clear that Equation (M.23) above is valid not only in
equilibrium but also when the matter in V is in motion.
Introducing the notation
T i j = T x i
(M.24)
for the jth component of the vector T x i , we can write Equation (M.23) on the facing
M.1
S CALARS , V ECTORS
AND
173
T ENSORS
T nj
= (T n ) j = ni T i j ni T i j
(M.25)
i=1
Using Equation (M.25) above, we find that the component of the vector T n in the
direction of an arbitrary unit vector m
is
T n m = T n m
= T nj
mj =
j=1
j=1
ni T i j n
(M.26)
m j ni T i j m j = n
T m
i=1
Hence, the jth component of the vector T x i , here denoted T i j , can be interpreted as the
i jth component of a tensor T. Note that T n m is independent of the particular coordinate
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 V of the body. If the external force density (force per unit volume) is
denoted by f and the velocity for a mass element dm is denoted by v, we obtain
d
dt
V v dm = V f d x + S Tn d x
(M.27)
V dt v j dm = V f j d x + S T nj d x = V f j d x + S ni Ti j d x
(M.28)
where, in the last step, Equation (M.25) was used. Setting dm = d 3x and using the
divergence theorem on the last term, we can rewrite the result as
V
d
v j d3x =
dt
V
f j d3x +
V
T i j 3
dx
xi
(M.29)
Since this formula is valid for any arbitrary volume, we must require that
T i j
d
vj fj
=0
dt
xi
(M.30)
or, equivalently
T i j
v j
+ v v j f j
=0
t
xi
(M.31)
Note that v j /t is the rate of change with time of the velocity component v j at a fixed
point x = (x1 , x1 , x3 ).
E ND OF
EXAMPLE
M.1C
174
M ATHEMATICAL M ETHODS
a (x ) g a (x )
(M.32)
This rule is often called lowering of index. The raising of index analogue of
the index lowering rule is:
def
a (x ) g a (x )
(M.33)
More generally, the following lowering and raising rules hold for arbitrary
rank n mixed tensor fields:
2 ...k1 k
1 2 ...k1
gk k A1k+1
k+2 ...n (x ) = Ak k+1 ...n (x )
(M.34)
1 2 ...k1 k
2 ...k1
gk k A1k k+1
...n (x ) = Ak+1 k+2 ...n (x )
(M.35)
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
A = g g A
(M.36)
i.e., the same rank 2 tensor in covariant form. This operation is also known as
a tensor contraction.
E XAMPLE M.2
FLAT
L ORENTZ SPACE
M.1
S CALARS , V ECTORS
AND
175
T ENSORS
metric tensor
1
if = = 0
1 if = = i = j = 1, 2, 3
0
if 6=
rs
g =
su
(M.37)
""#
1 0
0 1
0 0
0 0
(g ) =
0
0
0
0
1 0
0 1
*& (
(M.38)
i.e., a matrix with a main diagonal that has the sign sequence, or signature, {+, , , }
or
g =
t
rs
su
1 if = = 0
1
if = = i = j = 1, 2, 3
0
if =
6
(M.39)
(g ) =
""#
1 0
0 1
0 0
0 0
0
0
1
0
0
0
0
1
*&)(
(M.40)
a (a0 , a1 , a2 , a3 ) = (a0 , a)
(M.41)
According to the index lowering rule, Equation (M.32) on the preceding page, we
obtain the covariant version of this vector as
def
a (a0 , a1 , a2 , a3 ) = g a
(M.42)
a 0 = 1 a 0 + 0 a 1 + 0 a 2 + 0 a 3 = a0
0
(M.43)
1
(M.44)
a2 = 0 a0 + 0 a1 1 a2 + 0 a3 = a2
(M.45)
a1 = 0 a 1 a + 0 a + 0 a = a
0
a3 = 0 a + 0 a + 0 a + 1 a = a
(M.46)
or
a = (a0 , a1 , a2 , a3 ) = (a0 , a1 , a2 , a3 ) = (a0 , a)
(M.47)
176
M ATHEMATICAL M ETHODS
(M.48)
E ND OF
EXAMPLE
M.2C
(M.49)
(M.50)
where we made use of the index lowering and raising rules (M.32) and (M.33).
The result is a four-scalar, i.e., an invariant which is independent of in which
inertial system it is measured.
The quadratic differential form
ds2 = g dx dx = dx dx
(M.51)
i.e., the scalar product of the differential radius four-vector with itself, is an
invariant called the metric. It is also the square of the line element ds which is
the distance between neighbouring points with coordinates x and x + dx .
E XAMPLE M.3
def
def
A aR + iaI = aR a R + iaI a I A A
6 ),
expressed in terms of
(M.52)
def
(M.53)
(M.54)
Using this in Equation (M.52) above, we see that we can interpret this so that the
M.1
S CALARS , V ECTORS
AND
177
T ENSORS
aR
aR
aI
a R + i
aI
a R + i
aI
a I
a I
(M.55)
On the other hand, the definition of the scalar product in terms of the inner product of
complex vector with its own complex conjugate yields
def
(M.56)
aR
a2R + a2I
aR
a R + i
a2R + a2I
a2R + a2I
aI
a2R + a2I
a R + i
aI
a I
(M.57)
a2R + a2I
a2R + a2I
a I
E ND OF
METRIC IN
EXAMPLE
M.3C
L ORENTZ SPACE
E XAMPLE M.4
In 4 the metric tensor attains a simple form [see Equation (4.10) on page 51 for an
example] and, hence, the scalar product in Equation (M.50) on the preceding page can
be evaluated almost trivially and becomes
a b = (a0 , a) (b0 , b) = a0 b0 a b
The important scalar product of the
(M.58)
4.
The
(M.59)
4
(M.60)
E ND OF
EXAMPLE
M.4C
178
E XAMPLE M.5
M ATHEMATICAL M ETHODS
M ETRIC IN
GENERAL RELATIVITY
""#
(g ) =
e
0
0
0
0
e
0
0
0
0
r2
0
0
0
0
r2 sin2
*& (
(M.61)
where = (ct, r, , ) and = (ct, r, , ). In such a space, the metric takes the form
ds2 = c2 e (dt)2 e (dr)2 r2 (d)2 r2 sin2 (d)2
(M.62)
In general relativity the metric tensor is not given a priori but is determined by the
Einstein equations.
E ND OF
EXAMPLE
M.5C
Dyadic product
The dyadic product field A(x) a(x)b(x) with two juxtaposed vector fields
a(x) and b(x) is the outer product of a and b. Operating on this dyad from the
right and from the left with an inner product of an vector c one obtains
def
def
def
def
A c ab c a(b c)
(M.63a)
c A c ab (c a)b
(M.63b)
ab = x 1
x 2
x 3 4
B@
a1 b1 a1 b2 a1 b3
a1 b2 a2 b2 a2 b3 F
a1 b3 a3 b2 a3 b3
B@
x 1
x 2 F
x 3
(M.64)
which means that we can represent the tensor A(x) in matrix form as
3 Ai j (xk )
B@
a1 b1 a1 b2 a1 b3
a1 b2 a2 b2 a2 b3 F
a1 b3 a3 b2 a3 b3
(M.65)
which we identify with expression (M.15) on page 170, viz. a tensor in matrix
notation.
M.1
S CALARS , V ECTORS
AND
179
T ENSORS
Vector product
The vector product or cross product of two arbitrary 3D vectors a and b in
ordinary 3 space is the vector
c = a b = i jk a j bk x i
(M.66)
x i
def
xi
(M.67)
where x i is the ith unit vector in a Cartesian coordinate system. Since the
operator in itself has vectorial properties, we denote it with a boldface nabla.
In component notation we can write
i =
,
,
x1 x2 x3
(M.68)
180
M ATHEMATICAL M ETHODS
(M.69)
(M.70)
(M.71)
y0 = 0 x 4 y
(M.72)
and
respectively.
E XAMPLE M.6
T HE
In
L ORENTZ SPACE
1
, =
c t
1
,
c t
(M.73)
1
, =
c t
1
,
c t
(M.74)
1 2
2 =
c2 t2
(M.75)
, and sometimes defined with
E ND OF
EXAMPLE
M.6C
M.1
S CALARS , V ECTORS
AND
181
T ENSORS
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
a(x):
vector field
(M.76)
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.
In 4D, the four-gradient is a covariant vector, formed as a derivative of a
four-scalar field (x ), with the following component form:
(x ) =
(x )
x
(M.77)
RELATIVE DISTANCES IN
3D
E XAMPLE M.7
Very often electrodynamic quantities are dependent on the relative distance in 3 between two vectors x and x0 , i.e., on |x x0 |. In analogy with Equation (M.67) on
page 179, we can define the primed del operator in the following way:
0 = x i
= 0
x0i
(M.78)
Using this, the unprimed version, Equation (M.67) on page 179, and elementary
rules of differentiation, we obtain the following two very useful results:
|x x0 | = x i
|x x0 |
x x0
|x x0 |
=
=
i
xi
|x x0 |
x0i
= 0 |x x0 |
(M.79)
and
1
|x x0 |
x x0
= 0
|x x0 |3
1
|x x0 |
(M.80)
E ND OF
EXAMPLE
M.7C
182
M ATHEMATICAL M ETHODS
The divergence
We define the 3D divergence of a vector field in
as
ai (x)
= (x)
xi
(M.81)
E XAMPLE M.8
a (x )
x
(M.82)
D IVERGENCE IN 3D
For an arbitrary
0
a(x0 )
|x x0 |
0 a(x0 )
+ a(x0 ) 0
|x x0 |
1
|x x0 |
(M.83)
which demonstrates how the primed divergence, defined in terms of the primed
del operator in Equation (M.78) on the preceding page, works.
E ND OF
EXAMPLE
M.8C
The Laplacian
The 3D Laplace operator or Laplacian can be described as the divergence of
the gradient operator:
2 = = =
2
2
x i x j
= i j i j = 2i = 2 2
xi
x j
xi i=1 xi
(M.84)
The symbol 2 is sometimes read del squared. If, for a scalar field (x),
2 < 0 at some point in 3D space, it is a sign of concentration of at that
point.
M.1
S CALARS , V ECTORS
AND
183
T ENSORS
|x x0 |
= 2
E XAMPLE M.9
is
|x x0 |
= 4(x x0 )
(M.85)
EXAMPLE
M.9C
The curl
In 3 the curl of a vector field a(x), denoted a(x), is another
b(x) which can be defined in the following way:
a(x) = i jk x i j ak (x) = i jk x i
ak (x)
= b(x)
x j
vector field
(M.86)
where use was made of the Levi-Civita tensor, introduced in Equation (M.18)
on page 171.
The covariant 4D generalisation of the curl of a four-vector field a (x ) is
the antisymmetric four-tensor field
G (x ) = a (x ) a (x ) = G (x )
(M.87)
T HE CURL OF
E XAMPLE M.10
A GRADIENT
Using the definition of the 3 curl, Equation (M.86), and the gradient, Equation (M.76) on page 181, we see that
[(x)] = i jk x i j k (x)
(M.88)
(x) x i
x j xk
2
2
x2 x3 x3 x2
2
2
x3 x1 x1 x3
(x) x 2
2
2
x1 x2 x2 x1
(x) x 3
(x) x 1
(M.89)
184
M ATHEMATICAL M ETHODS
(M.90)
(M.91)
3.
E XAMPLE M.11
T HE
EXAMPLE
M.10C
DIVERGENCE OF A CURL
With the use of the definitions of the divergence (M.81) and the curl, Equation (M.86)
on the preceding page, we find that
[ a(x)] = i [ a(x)]i = i jk i j ak (x)
(M.92)
Using the definition for the Levi-Civita symbol, defined by Equation (M.18) on
page 171, we find that, due to the assumed well-behavedness of a(x),
i i jk j ak (x) =
=
i jk
ak
xi
x j
2
2
x2 x3 x3 x2
2
2
x3 x1 x1 x3
a2 (x)
2
2
x1 x2 x2 x1
a3 (x)
a1 (x)
(M.93)
0
i.e., that
[ a(x)] 0
(M.94)
a (x ) 6= 0
(M.95)
E ND OF
EXAMPLE
M.11C
M.2
185
A NALYTICAL M ECHANICS
Numerous vector algebra and vector analysis formulae are given in Chapter F. Those which are not found there can often be easily derived by using the
component 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].
dqi
,t = T V
dt
(M.96)
where qi is the generalised coordinate, T the kinetic energy and V the potential energy of a mechanical system, The Lagrangian satisfies the Lagrange
equations
L
L
=0
q i
qi
(M.97)
L
q i
(M.98)
(M.99)
(M.100)
After differentiating the left and right hand sides of this definition and setting
them equal we obtain
H
H
H
L
L
L
dpi +
dqi +
dt = q i dpi + pi dq i
dqi
dq i dt (M.101)
pi
qi
t
qi
q i
t
186
M ATHEMATICAL M ETHODS
(M.102a)
(M.102b)
Bibliography
[1] G. B. A RFKEN AND H. J. W EBER, Mathematical Methods for Physicists, fourth,
international ed., Academic Press, Inc., San Diego, CA . . . , 1995, ISBN 0-12059816-7.
[2] R. A. D EAN, Elements of Abstract Algebra, John Wiley & Sons, Inc.,
New York, NY . . . , 1967, ISBN 0-471-20452-8.
[3] A. A. E VETT, Permutation symbol approach to elementary vector analysis,
American Journal of Physics, 34 (1965), pp. 503507.
[4] P. M. M ORSE AND H. F ESHBACH, Methods of Theoretical Physics, Part I.
McGraw-Hill Book Company, Inc., New York, NY . . . , 1953, ISBN 07-043316-8.
[5] B. S PAIN, Tensor Calculus, third ed., Oliver and Boyd, Ltd., Edinburgh and London, 1965, ISBN 05-001331-9.
[6] W. E. T HIRRING, Classical Mathematical Physics, Springer-Verlag, New York,
Vienna, 1997, ISBN 0-387-94843-0.
Index
acceleration field, 129
advanced time, 43
Ampres law, 5
Ampre-turn density, 89
anisotropic, 152
anomalous dispersion, 153
antisymmetric tensor, 63
associated Legendre polynomial, 117
associative, 55
axial gauge, 40
axial vector, 63, 179
contravariant vector, 50
convection potential, 136
convective derivative, 11
cosine integral, 110
Coulomb gauge, 40
Coulombs law, 2
covariant, 48
covariant component form, 168
covariant field tensor, 64
covariant four-tensor field, 174
covariant four-vector, 170
covariant four-vector field, 54
covariant vector, 50
cross product, 179
curl, 183
cutoff, 142
cyclotron radiation, 145, 148
187
188
M ATHEMATICAL M ETHODS
electric conductivity, 10
field Lagrange density, 79
electric current density, 6
field point, 3
electric dipole moment, 118
field quantum, 143
electric dipole moment vector, 85
fine structure constant, 143, 150
electric dipole radiation, 120
four-current, 59
electric displacement, 14
four-del operator, 180
electric displacement vector, 87
four-dimensional Hamilton equations,
electric field, 2
71
electric field energy, 91
four-dimensional vector space, 50
electric monopole moment, 85
four-divergence, 182
electric permittivity, 151
four-gradient, 181
electric polarisation, 86
four-Hamiltonian, 71
electric quadrupole moment tensor, 86
four-Lagrangian, 68
electric quadrupole radiation, 122
four-momentum, 57
electric quadrupole tensor, 122
four-potential, 59
electric susceptibility, 87
four-scalar, 169
electric volume force, 92
four-tensor fields, 174
electrodynamic potentials, 36
four-vector, 54, 170
electromagnetic field tensor, 63
four-velocity, 57
electromagnetic scalar potential, 37
Fourier component, 25
electromagnetic vector potential, 36
Fourier transform, 41
electromagnetodynamic equations, 16
functional derivative, 76
electromagnetodynamics, 16
fundamental tensor, 50, 168, 174
electromotive force (EMF), 10
Galileos law, 47
electrostatic scalar potential, 35
gauge fixing, 40
electrostatics, 1
gauge function, 39
energy theorem in Maxwells theory,
gauge invariant, 39
91
gauge transformation, 39
equation of continuity, 9, 59
equation of continuity for magnetic monopoles, Gausss law, 4
general inhomogeneous wave equations,
16
37
equations of classical electrostatics, 8
generalised coordinate, 70, 185
equations of classical magnetostatics,
generalised four-coordinate, 71
8
Gibbs notation, 179
Euclidean space, 56
gradient, 181
Euclidean vector space, 51
Green function, 42, 116
Euler-Lagrange equation, 77
group theory, 55
Euler-Lagrange equations, 77
group velocity, 153
Euler-Mascheroni constant, 110
event, 55
Hamilton density, 78
Hamilton density equations, 78
far field, 98
far zone, 101
Hamilton equations, 70, 186
Faradays law, 11
Hamilton function, 185
field, 169
Hamilton gauge, 40
189
Hamiltonian, 185
Heaviside potential, 136
Helmholtz theorem, 38
help vector, 115
Hertz method, 114
Hertz vector, 115
Hodge star operator, 16
homogeneous wave equation, 24
Huygens principle, 41
identity element, 55
in a medium, 154
incoherent radiation, 148
indefinite norm, 51
index contraction, 50
index lowering, 50
induction field, 98
inertial reference frame, 47
inertial system, 47
inhomogeneous Helmholtz equation, 41
inhomogeneous time-independent wave
equation, 41
inhomogeneous wave equation, 41
inner product, 176
instantaneous, 138
interaction Lagrange density, 79
intermediate field, 101
invariant, 169
invariant line element, 53
inverse element, 55
irrotational, 4, 183
Kelvin function, 149
kinetic energy, 75, 185
kinetic momentum, 74
Kronecker delta, 171
Lagrange density, 75
Lagrange equations, 185
Lagrange function, 74, 185
Lagrangian, 74, 185
Laplace operator, 182
Laplacian, 182
Larmor formula for radiated power, 138
law of inertia, 47
190
M ATHEMATICAL M ETHODS
positive definite, 56
positive definite norm, 51
potential energy, 75, 185
potential theory, 116
power flux, 91
Poynting vector, 91
Poyntings theorem, 91
Proca Lagrangian, 82
propagator, 42
proper time, 53
pseudo-Riemannian space, 56
pseudoscalar, 167
pseudoscalars, 179
pseudotensor, 167
pseudotensors, 179
pseudovector, 63, 167, 179
quadratic differential form, 52, 176
quantum mechanical nonlinearity, 4
radiation field, 98, 101, 129
radiation fields, 101
radiation gauge, 40
radiation resistance, 110
radius four-vector, 50
radius vector, 167
raising of index, 174
rank, 170
rapidity, 57
refractive index, 152
relative electric permittivity, 92
relative magnetic permeability, 92
relative permeability, 151
relative permittivity, 151
Relativity principle, 48
relaxation time, 25
rest mass density, 79
retarded Coulomb field, 101
retarded potentials, 44
retarded relative distance, 125
retarded time, 43
Riemannian metric, 52
Riemannian space, 50, 168
row vector, 167
scalar, 167, 182
191
Vavilov-Cerenkov
radiation, 154, 156
vector, 167
vector product, 179
velocity field, 129
virtual simultaneous coordinate, 126,
129
wave equations, 23
wave vector, 29, 152
world line, 55
Youngs modulus, 75
Yukawa meson field, 82
Companion volume to
E LECTROMAGNETIC
F IELD T HEORY
Exercises
Please note that this is a
VERY preliminary draft!
Tobia Carozzi Anders Eriksson
Bengt Lundborg Bo Thid
Mattias Waldenvik
Department of Space and Plasma Physics
Uppsala University
and
Swedish Institute of Space Physics
Uppsala Division
Sweden
Ipsum
C ONTENTS
Preface
vii
1 Maxwells Equations
1.1
1.2
1.3
Coverage . . . . . . . . . . . . . . . . . . . . . .
Formulae used . . . . . . . . . . . . . . . . . . . .
Solved examples . . . . . . . . . . . . . . . . . .
Example 1.1 Macroscopic Maxwell equations .
Solution . . . . . . . . . . . . . . . . . . .
1
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Solution . . . . . . . . . . . . . . . . . . . . . .
Example 1.3 The charge continuity equation . . . .
Solution . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Coverage . . . . . . . . . . . . . . . . . . . . . .
Formulae used . . . . . . . . . . . . . . . . . . . .
Solved examples . . . . . . . . . . . . . . . . . .
Example 2.1 The Aharonov-Bohm effect . . .
Solution . . . . . . . . . . . . . . . . . . .
Example 2.2 Invent your own gauge . . . . .
Solution . . . . . . . . . . . . . . . . . . .
9
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Solution . . . . . . . . . . . . . . . . . . . . . . .
Example 2.4 Simple dispersion relation . . . . . . . .
Solution . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3 Relativistic Electrodynamics
3.1
Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
1
1
1
2
4
4
5
5
9
9
9
9
10
11
11
13
13
15
15
17
17
i
ii
3.2
3.3
Formulae used . . . . . . . . . . . . . . . . . . . . . .
Solved examples . . . . . . . . . . . . . . . . . . . .
Example 3.1 Covariance of Maxwells equations .
Solution . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
17
18
18
18
Example 3.2 Invariant quantities constructed from the field tensor 20
Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Example 3.3 Covariant formulation of common electrodynamics formulas . . . . . . . . . . . . . . . . . . . .
Solution . . . . . . . . . . . . . . . . . . . . . . . . . . .
Example 3.4 Fields from uniformly moving charge via Lorentz
transformation . . . . . . . . . . . . . . . . . . .
Solution . . . . . . . . . . . . . . . . . . . . . . . . . . .
Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Formulae used . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Solved examples . . . . . . . . . . . . . . . . . . . . . . . . . .
Example 4.1 Canonical quantities for a particle in an EM field .
Solution . . . . . . . . . . . . . . . . . . . . . . . . . . .
Example 4.2 Gauge invariance of the Lagrangian density . . .
Solution . . . . . . . . . . . . . . . . . . . . . . . . . . .
Coverage . . . . . . . . . . . . . . . . . . .
Formulae used . . . . . . . . . . . . . . . . .
Solved examples . . . . . . . . . . . . . . .
Example 5.1 EM quantities potpourri . .
Solution . . . . . . . . . . . . . . . .
Example 5.2 Classical electron radius .
Solution . . . . . . . . . . . . . . . .
Example 5.3 Solar sailing . . . . . . .
Solution . . . . . . . . . . . . . . . .
27
27
28
28
28
29
29
31
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Example 5.4 Magnetic pressure on the earth .
Solution . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
23
23
27
21
21
Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Formulae used . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Solved examples . . . . . . . . . . . . . . . . . . . . . . . . . .
31
31
32
32
32
35
35
37
37
39
39
41
41
41
42
Example 6.1 Instantaneous current in an infinitely long conductor 42
Draft version released 15th November 2000 at 20:39
iii
Solution . . . . . . . . . . . . . . . . .
Example 6.2 Multiple half-wave antenna .
Solution . . . . . . . . . . . . . . . . .
Example 6.3 Travelling wave antenna . . .
Solution . . . . . . . . . . . . . . . . .
Example 6.4 Microwave link design . . .
Solution . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
7 Multipole Radiation
7.1
7.2
7.3
Coverage . . . . . . . . . . . . . . . . . . .
Formulae used . . . . . . . . . . . . . . . . .
Solved examples . . . . . . . . . . . . . . .
Example 7.1 Rotating Electric Dipole .
Solution . . . . . . . . . . . . . . . .
Example 7.2 Rotating multipole . . . .
Solution . . . . . . . . . . . . . . . .
Example 7.3 Atomic radiation . . . . .
Solution . . . . . . . . . . . . . . . .
Example 7.4 Classical Positronium . . .
Solution . . . . . . . . . . . . . . . .
53
.
.
.
.
.
.
.
.
.
.
.
Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Formulae used . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Solved examples . . . . . . . . . . . . . . . . . . . . . . . . . .
Solution . . . . . . . . . . . . . . . . . . . . . . . . . . .
Example 8.2 Synchrotron radiation perpendicular to the acceleration . . . . . . . . . . . . . . . . . . . . . .
Solution . . . . . . . . . . . . . . . . .
Example 8.3 The Larmor formula . . . .
Solution . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
53
53
54
54
54
56
56
58
58
59
59
63
9.1
9.2
9.3
42
47
47
50
50
51
51
63
63
64
64
64
66
66
67
67
69
69
71
71
71
72
Solution . .
Example 9.2
Solution . .
Example 9.3
Solution . .
Example 9.4
Solution . .
72
. . . . . . . . . . . . . . . . . . . . . . . . . 72
Radiative reaction force from conservation of energy 74
. . . . . . . . . . . . . . . . . . . . . . . . . 74
Radiation and particle energy in a synchrotron . .
77
. . . . . . . . . . . . . . . . . . . . . . . . . 77
Radiation loss of an accelerated charged particle .
79
. . . . . . . . . . . . . . . . . . . . . . . . . 79
iv
L IST
6.1
6.2
6.3
43
44
47
9.1
74
OF
F IGURES
vi
P REFACE
B. T.
vii
viii
P REFACE
L ESSON 1
Maxwells Equations
1.1
Coverage
In this lesson we examine Maxwells equations, the cornerstone of electrodynamics. We start by practising our math skill, refreshing our knowledge of vector
analysis in vector form and in component form.
1.2
Formulae used
E 0
B 0
E
B
1.3
B
0 j
(1.1a)
(1.1b)
(1.1c)
1
E
c2 t
(1.1d)
Solved examples
E XAMPLE 1.1
E 0
B 0
B
E
0 j
(1.2a)
(1.2b)
(1.2c)
1
E
c2 t
(1.2d)
sometimes known as the microscopic Maxwell equations or the Maxwell-Lorentz equations. In the presence of a medium, these equations are still true, but it may sometimes
be convenient to separate the sources of the fields (the charge and current densities) into
an induced part, due to the response of the medium to the electromagnetic fields, and an
extraneous, due to free charges and currents not caused by the material properties. One
then writes
j
jind jext
ind ext
(1.3)
(1.4)
The electric and magnetic properties of the material are often described by the electric
polarisation P (SI unit: C/m2 ) and the magnetisation M (SI unit: A/m). In terms of these,
the induced sources are described by
jind
ind
P t
P
(1.5)
(1.6)
To fully describe a certain situation, one also needs constitutive relations telling how P
and M depends on E and B. These are generally empirical relations, different for different
media.
Show that by introducing the fields
0 E P
B 0 M
D
H
(1.7)
(1.8)
the two Maxwell equations containing source terms (1.2a) and (??) reduce to
ext
jext
(1.9)
D
t
(1.10)
(1.11)
Solution
If we insert
j
jind jext
ind ext
(1.12)
(1.13)
and
1.3. S OLVED
jind
ind
EXAMPLES
M
P
t
P
(1.14)
(1.15)
into
0 j
0
1
E
c2 t
(1.16)
(1.17)
(1.18)
we get
0 jext
P
t
1
E
P
ext
B
M
1
E
c2 t
(1.19)
(1.20)
B
M!"
0
0 E P #
jext
P 0 E
t
(1.21)
ext
(1.22)
0 E P
(1.23)
B
M
0
(1.24)
H jext
D
t
D ext
(1.25)
(1.26)
QED $
The reason these equations are known as macroscopic are that the material properties
described by P and M generally are average quantities, not considering the atomic properties of matter. Thus E and D get the character of averages, not including details around
single atoms etc. However, there is nothing in principle preventing us from using largescale averages of E and B, or even to use atomic-scale calculated D and H although this is
a rather useless procedure, so the nomenclature microscopic/macroscopic is somewhat
misleading. The inherent difference lies in how a material is treated, not in the spatial
scales.
E ND
OF EXAMPLE
1.1 %
E XAMPLE 1.2
Solution
Maxwells equations in vector form are written:
E 0
B 0
E
B
0 j
(1.27)
(1.28)
(1.29)
1
E
c2 t
(1.30)
In these equations, E, B, and j are vectors, while is a scalar. Even though all the equations
contain vectors, only the latter pair are true vector equations in the sense that the equations
themselves have several components.
When going to component notation, all scalar quantities are of course left as they are.
Vector quantities, for example E, can always be expanded as E 3j & 1 E j x j E j x j ,
where the last step assumes Einsteins summation convention: if an index appears twice in
the same term, it is to be summed over. Such an index is called a summation index. Indices
which only appear once are known as free indices, and are not to be summed over. What
symbol is used for a summation index is immaterial: it is always true that a i bi ak bk ,
since both these expressions mean a1 b1 a2 b2 a3 b3 a b. On the other hand, the
expression ai ak is in general not true or even meaningful, unless i k or if a is the null
vector.
The three E j are the components of the vector E in the coordinate system set by the three
unit vectors x j . The E j are real numbers, while the x j are vectors, i.e. geometrical objects.
Remember that though they are real numbers, the E j are not scalars.
Vector equations are transformed into component form by forming the scalar product of
both sides with the same unit vector. Let us go into ridiculous detail in a very simple case:
G x k
G j x j x k
G j jk
Gk
H
H x k
Hi x i x k
Hi ik
Hk
(1.31)
(1.32)
(1.33)
(1.34)
(1.35)
This is of course unnecessarily tedious algebra for an obvious result, but by using this
careful procedure, we are certain to get the correct answer: the free index in the resulting
equation necessarily comes out the same on both sides. Even if one does not follow this
complicated way always, one should to some extent at least think in those terms.
Nabla operations are translated into component form as follows:
1.3. S OLVED
EXAMPLES
'
(
xi
V
'
V (
xi i
'
V i jk x i
V (
xj k
x i
xi
Vi
xi
V
i jk k
xj
(1.36)
(1.37)
(1.38)
0 j
)'
0
0 )'
*
B
)'
t
1 E
c2 t
)'
Ei
xi
0
Bi
0
xi
E
i jk k
B
xj
t i
i jk
Bk
xj
0 ji
(1.39)
(1.40)
(1.41)
1 Ei
c2 t
(1.42)
E ND
OF EXAMPLE
1.2 %
E XAMPLE 1.3
Derive the continuity equation for charge density from Maxwells equations using (a)
vector notation and (b) component notation. Compare the usefulness of the two systems of
notations. Also, discuss the physical meaning of the charge continuity equation.
Solution
Vector notation In vector notation, a derivation of the continuity equation for charge
looks like this:
Compute
1. Apply
t E
in two ways:
to Gausss law:
E +
t
0 t
Draft version released 15th November 2000 at 20:39
(1.43)
1
B + 0
j
E
c2
t
Use
(1.44)
,
+- 0 and 0 0 c2 1:
.
1
j
0
(1.45)
j 0/
t
(1.46)
x i t Ei
1. Take
in two ways:
of Gausss law:
Ei
t xi
0 t
(1.47)
Bk
xi 0 i jk x j 1
0
('
j
xi i
1 Ei
c2 x i t
- 0 is valid also if Ai
(1.48)
xi ,
Ei
1
j
t xi
0 x i i
and that 0 0 c2
1:
(1.49)
t
ji
0/
xi
Draft version released 15th November 2000 at 20:39
(1.50)
1.3. S OLVED
EXAMPLES
Comparing the two notation systems We notice a few points in the derivations
above:
2 In the vector notation system, we sometimes need to keep some vector formulas in
memory or to consult a math handbook, while with the component system you need
only the definitions of i jk and i j .
2 Although not seen here, the component system of notation is more explicit (read
unambiguous) when dealing with tensors of higher rank, for which vector notation
becomes cumbersome.
j 0
(1.51)
t
is known as a continuity equation. Why? Well, integrate the continuity equation over some
volume V bounded by the surface S. By using Gausss theorem, we find that
3
dQ
43
d3x 53
j d x 63 j dS
(1.52)
dt
t
V
V
S
which says that the change in the total charge in the volume is due to the net inflow of
electric current through the boundary surface S. Hence, the continuity equation is the field
theory formulation of the physical law of charge conservation.
E ND
OF EXAMPLE
1.3 %
L ESSON 2
Electromagnetic
Potentials and Waves
2.1
Coverage
Here we study the vector and scalar potentials A and and the concept of gauge
transformation.
One of the most important physical manifestation of Maxwells equations is
the EM wave. Seen as wave equations, the Maxwell equations can be reduced
to algebraic equations via the Fourier transform and the physics is contained in
so-called dispersion relations which set the kinematic restrictions on the fields.
2.2
Formulae used
E
7
B
2.3
A
t
Solved examples
E XAMPLE 2.1
8 r0 9 9 z : Bz
; r0 9 9 z :
0
(2.1)
(2.2)
10
AND
WAVES
Solution
A interesting question in electrodynamics is whether the EM potentials and A are more
than mathematical tools, and alternatives to the Maxwell equations, from which we can
derive the EM fields. Could it be that the potentials and not Maxwells equations are more
fundamental? Although the ultimate answer to these questions is somewhat metaphysical,
it is exactly these questions that make the Aharonov-Bohm effect. Before we discuss this
effect let us calculate the vector field from the given magnetic field.
The equations connecting the potentials with the fields are
A
t
E
B
(2.3)
(2.4)
In this problem we see that we have no boundary conditions for the potentials. Also, let us
use the gauge 0.
This problem naturally divides into two parts: the part within the magnetic field and the
part outside the magnetic field. Let us start with the interior part:
1
r
1 Az
r
Ar
z
rA
r
A
t
A
z
Az
r
Ar
!
0
(2.5a)
0
(2.5b)
0
(2.5c)
B
(2.5d)
Br
2
(2.7)
Moving to the outer problem, we see that the only difference compared with the inner
problem is that B 0 so that we must consider
1 rA
r r
0
(2.8)
2.3. S OLVED
11
EXAMPLES
Br02
2r
0!
<
(2.10)
Now in electrodynamics (read: in this course) the only measurable quantities are the fields.
So the situation above, where we have a region in which the magnetic field is zero but
the potential is non-zero has no measurable consequence in classical electrodynamics. In
quantum mechanics however, the Aharonov-Bohm effect shows that this situation does
have a measurable consequence. Namely, when letting charged particles go around this
magnetic field (the particles are do not enter the magnetic field region because of a impenetrable wall) the energy spectrum of the particles after passing the cylinder will have
changed, even though there is no magnetic field along their path. The interpretation is that
the potential is a more fundamental quantity than the field.
E ND
OF EXAMPLE
2.1 %
E XAMPLE 2.2
Name some common gauge conditions and discuss the usefulness of each one. Then invent
your own gauge and verify that it is indeed a gauge condition.
Solution
Background The Maxwell equations that do not contain source terms can be solved
by using the vector potential A and the scalar potential , defined through the relations
B
E
(2.11)
A
t
(2.12)
Assuming linear, isotropic and homogeneous media, we can use the constitutive relations
D E H B , and j E j= (where j= is the free current density arising from
other sources than conductivity) and definitions of the scalar and vector potentials in the
remaining two Maxwell equations and find that
2
2 A
A
t
(
2A
A !> j=
A 2
t
t
t
Draft version released 15th November 2000 at 20:39
(2.13)
(2.14)
12
AND
WAVES
These equations are used to determine A and from the source terms. And once we have
found A and it is straight forward to derive the E and B fields from (2.11) and (2.12).
The definitions of the scalar and vector potentials are not enough to make A and unique,
i.e. , if one is given A and then (2.11) and (2.12) determine B and E, but if one is given
B and E there many ways of choosing A and . This can be seen through the fact that A
and can be transformed according to
A=
A
=
(2.15)
(2.16)
where is an arbitrary scalar field, but the B and E fields do not change. This kind of
transformation is called a gauge transformation and the fact that gauge transformations do
not affect the physically observable fields is known as gauge invariance.
Gauge conditions The ambiguity in the definitions of A and can be used to introduce
a gauge condition. In other words, since the definitions (2.11) and (2.12) do not completely
define A and we are free to add certain conditions. Some common gauge conditions are
Coulomb gauge
Lorentz gauge
Temporal gauge
A t
A 0
0
0
The Coulomb gauge is most useful when dealing with static fields. Using
(2.13) and (2.14, for static fields, reduces to
2 A
j
2
0 then
(2.17)
(2.18)
The Lorentz gauge is the most commonly used gauge for time-varying fields. In this case
(2.13) and (2.14) reduce to
2
2 !
t
t
2
2
2 ! A
j
t
t
2
(2.19)
(2.20)
So the Lorentz transform decouples (2.13) and (2.14) and puts and A on equal footing.
Furthermore, the resulting equations are manifestly covariant.
In the temporal gauge one discards the scalar potential by setting
(2.13) and (2.14) reduce to
1 2A
c2 t 2
6
j
0. In this gauge
(2.21)
2.3. S OLVED
13
EXAMPLES
Thus the single vector A describes both E and B in the temporal gauge.
How to invent your own gauge Gauges other than Coulomb, Lorentz and the temporal mentioned above are rarely used in introductory literature in Electrodynamics, but
it is instructive to consider what constitutes a gauge condition by constructing ones own
gauge.
Of course, a gauge condition is at least a scalar equation containing at least one of the
components of A or . Once you have an equation that you think might be a gauge, it must
be verified that it is a gauge. To verify that a condition is a gauge condition it is sufficient
to show that any given set of A and can be made to satisfy your condition. This is done
through gauge transformations. So given a A and a which satisfy the physical conditions
through (2.13) and (2.14) we try to see if it is possible (at least in principle) to find a gauge
transformation to some new potential A= and = , which satisfy your condition.
E ND
OF EXAMPLE
2.2 %
E XAMPLE 2.3
Fourier transform Maxwells Equation. Use the Fourier version of Maxwells equations
to investigate the possibility of waves that do not propagate energy; such waves are called
static waves.
Solution
Maxwells equations contain only linear operators in time and space. This makes it easy to
Fourier transform them. By transforming them we get simple algebraic equations instead of
differential equations. Furthermore, the Fourier transformed Maxwell equations are useful
when working with waves or time-varying fields, especially since the response function,
i.e. the dielectric function, is in many case more fundamentally described as a function of
angular frequency than length x.
To perform this derivation we need formulas on how to translate the operators
and t in Maxwells equations.
? "
,
3 A
dt ei t f t
(2.22)
3 A
d3x e
A ik B x
A x
(2.23)
3 A
dt d3x e
A iC kB xA
D F t x
9
(2.24)
14
3 A
'
E3 A
AND
WAVES
E we get
dt d3x e
A iC kB xA
D E
dt d3x e
A iC kB xA
D Ei t 9 x
xi
A
A
3 dt F Ei t 9 x e i C k B x t DHG
3 dt d3x iki Ei t 9 x O
i
xI
J
MK L
N
& 0
iki 3 dt d3x Ei t 9 x
iki Ei 9 k + ik E 9 k
'
(2.25)
E we get
D E t
9
E t x
3 dt d3x e A i C k B x A t D i jk j 9 ei
3 dt d3x e A i C k B x A
A
A
3 dt P Ei /Q/M/ e i C k B x
J
KML
& 0
ii jk k j Ek 9 k ei
xk
A
A
DSR 5
3 dt d3x TU ii jk k j Ek t 9 x e i C k B x
N
De
V
(2.26)
where we have once again used partial integration. One may proceed analogously for
tE t 9 x . Trivially, one gets similar equations for the transformation of the D, H and B
fields. Thus we have found that
V t x #' ik V
k
9
9
V t 9 x #' ik V 9 k
V t 9 x
' i V 9 k
t
(2.27a)
(2.27b)
(2.27c)
where V t 9 x is an arbitrary field and ' denotes here Fourier transform. These transformation rules are easy to remember by saying that roughly the Fourier transform of is
ik and the Fourier transform of t is i .
Now we can use (2.27a), (2.27b) and (2.27c) on Maxwells equations. We then get, after
some simple trimming
E 9
H
9
D
9
B 9
k
ik
k
k
k : B 9 k
k
k : j 9 k W i D
9
k : i 9 k
k : 0
Draft version released 15th November 2000 at 20:39
(2.28a)
(2.28b)
(2.28c)
(2.28d)
2.3. S OLVED
15
EXAMPLES
where we have dropped the notation. These are the Fourier versions of Maxwells equation.
As an example of the use of the Fourier transformed Maxwells equations let us derive
static waves. Static waves are one possible oscillation mode for the E and H fields. Lets
say that we have a mode such that the E E field is oscillating at 0 and
<
that it has a k k 0 which is parallel to the electric field, so k X E . From (2.28a)
<
this implies that
B k
B
0
0
(2.29)
(2.30)
So, we see that S E H 0 trivially. The lesson here is that you can have time-varying
fields that do not transmit energy! These waves are also called longitudinal waves for
obvious reasons.
E ND
OF EXAMPLE
2.3 %
E XAMPLE 2.4
Solution
A dispersion relation is a relation between and k, usually something like
D 9 k +
(2.31)
From this one can solve for which is then a function of k1 9 k2 9 k3 . For isotropic media
then will be a function of Y k YU k only. A dispersion relation determines what modes (i.e.
what combinations of k and ) are possible. The dispersion relation is derivable in principal once one has explicit knowledge of the dielectricity function (or response function)
for the medium in question.
The two vector equations in Maxwells equations are
(2.32)
(2.33)
B
t
D
j
t
16
k
k
B
(2.34)
E
(2.35)
k E + k B
k k2 E 2 E
J kKML EN
.
&
WAVES
B
AND
(2.36)
(2.37)
0 Progressive wave!
2
2
k E 0
(2.38)
2 0
(2.39)
k2
2
"Z [
(2.40)
k
"Z ku
(2.41)
1
vg uk + u
k
k
so in this simple case the group velocity is the same as the phase velocity.
(2.42)
For the case of a conducting medium, in which j E, the two vector equations applied
on a wave which at first resembles the progressive wave we used above gives
k
k
B
B
(2.43)
i E E
(2.44)
k2 E i E 2 E
. T 2 i u2 k 2 V E 0
\
(2.45)
(2.46)
So that
i
1
Z
4u2 k2
2
2]
and the group velocity is
2
2
(2.47)
2u2 k
"Z ^
k
4u2 k2 2 2
If 0 then we are back again to the previous problem as can be verified.
E ND
(2.48)
OF EXAMPLE
2.4 %
L ESSON 3
Relativistic
Electrodynamics
3.1
Coverage
3.2
Formulae used
0
0
bc
0
1
0
0
dfe
0
e
0
0
1
0 g
0
(3.1)
F
bc
`b
0
E1
E2
E3
E1
0
cB3
cB2
E3
0
cB1
cB2
cB1 g
0
dfe
E2
cB3
e
(3.2)
17
18
3.3
E XAMPLE 3.1
Solved examples
Solution
The dAlembert operator
1 2
2
(3.3)
c2 t 2
is a fundamental operator in electrodynamics. The dynamics of EM fields is completely
described using the dAlembertian. Galilean transformations, even though closest to our
intuitive picture of the fabric of space-time, does not leave the dAlembertian invariant. A
Galilean transformation is simply
2
k
jjl
ijj
x1= x1
x2= x2
x3= x3
t =m t
(3.4)
vt
where the origin of the primed system is moving relative the unprimed along the 3-direction
with velocity v. Now we introduce this transformation by expanding each differential in
the unprimed coordinate system in terms of the differential in the primed system by using
the chain rule of derivation, i.e. we evaluate x x= x x= , so
x1
x2
x3
t=
x1 t =
x1=
x1=
x1 x1=
t=
x2 t =
x2=
x1=
x1 x1=
t=
x3 t =
x3=
x1=
x3 x1=
x2=
x1 x2=
x3=
x1 x3=
(3.5)
(3.6)
x2=
x2 x2=
x3=
x2 x3=
(3.7)
(3.8)
x2=
x3 x2=
x3=
x3 x3=
(3.9)
(3.10)
3.3. S OLVED
19
EXAMPLES
and so
2
x21
2
x12 ;
2
x22
2
x22 ;
2
x23
2
x32
(3.11)
but
t=
x1=
t t=
t x1=
v
t=
x3=
t
x2=
t x2=
x3=
t x3=
(3.12)
(3.13)
and so
2
t2
2
(3.14)
t = v x= t = v x= + t = 2 v x= 2 2v t = x=
3
3
3
3
where we have used the fact that the operators t = and x3= commute. Thus we have
found that
'
h 2=
= 2
1 2
c2 t = 2
v2 2
c2 x3= 2
2v 2
c2 t = x3=
0
(3.15)
Which obviously does not have the same form as the dAlembertian in the unprimed system!
Let us do the same calculations for the case of a Lorentz transformation; more specifically
we consider a boost along the 3 axis which is given by
x n
L n x
(3.16)
where
L n
poqqr
0
0
0
1
0
0
0
0
1
0
sMt
t
0
0
(3.17)
(remember that runs over 0, 1, 2, 3). Since and depend on v The 4-gradient,
- x transforms as
x n
x n
T L n x V n
x
L n
L n n L n n
x
1
(3.18)
so
g L n L n n n
2 2 2 02 12 22 2 2 2 32 g n n n n
(3.19)
In other words we have found that the dAlembertian is invariant under Lorentz boosts.
E ND
OF EXAMPLE
3.1 %
20
E XAMPLE 3.2
constructed solely with the field tensor and its dual tensor, are invariant under Lorentz
transformations? Having found these quantities you should be able to answer the questions:
2 can a purely electric field in one inertial system be seen as a purely magnetic field
in another?
2 and, can a progressive wave be seen as a purely electric or a purely magnetic field
in an inertial system?
Solution
The dual tensor of F is given by
F
2
g g F
2
where
qp
q o
E1
0
E1
E2
E3
E2
0
cB3
cB2
cB2
cB1 u
0
cB1
qp
q o
r
0
1
0
0
0
0
1
0
0
0
0
1
(3.21)
and
1
0
0
0
s tt
E3
cB3
(3.20)
s tt
u
(3.22)
We determine first the field tensor with two covariant indices through the formula
g g F poqqr
so
qp
q o
0
cB1
cB2
cB3
0
E1
E2
E3
E1
0
cB3
cB2
E2
cB3
0
cB1
E3
cB2
cB1
0
sMtt
u
(3.23)
st
cB1 cB2 cB3 t
E2
0
E1
0
E1 u
E3
E1
0
E2
'
(3.24)
cB and B
'w E c.
3.3. S OLVED
21
EXAMPLES
F and F F .
4cE B
(3.25)
2 c 2 B2 E 2
This means that E B and E 2 c2 B2 are Lorentz invariant scalars.
F F
(3.26)
Relation of EM fields in different inertial systems Now that we know that E B and
E c B are Lorentz invariant scalars, let see what they say about EM fields in different
inertial systems. Let us say that X - E B and Y - E 2 c2 B2 . All inertial systems must
have the same value for X and Y . A purely electric field in one inertial system means that
B 0, so X 0 and Y x 0. A purely magnetic field would mean that E =y 0, so X 0
but Y z 0. In other words it does not seem that a purely electric field can be a purely
magnetic field in any inertial system.
2
2 2
OF EXAMPLE
3.2 %
x
u
p
k
J
A
F
ct 9 x ,
c 9 v ,
E c 9 p ,
c 9 k ,
c 9 j ,
c 9 A ,
v F c 9 F
E XAMPLE 3.3
22
4-gradient
- x
ct 9 x
as our operator building block and also the second rank 4-tensor
field tensor
qp
q o
r
0
E1
E2
E3
E1
0
cB3
cB2
E2
cB3
0
cB1
s tt
E3
cB2
cB1 u
0
Observe that we use indices which run 0, 1, 2, 3 where the 0-component is time-like
component, there is also the system where indices run 1, 2, 3, 4 and the 4-component
is the time-like component. Beware!
A sufficient condition to formulate covariant electrodynamic formulas is that we make
our formulas by combine the above 4-vectors. To make sure we have a covariant form
we take outer product (i.e. simply combine the tensors so that all the indices are free)
and then perform zero or more contractions, i.e. equate two indices and sum over this
index (notationally this means we create a repeated index). In the notation we use here
contractions must be between a contravariant (upper) index and a covariant (lower) index.
One can always raise or lower a index by including a metric tensor g . On top of this
sufficient condition, we will need to use our knowledge of the formulas we will try to
make covariant, to accomplish our goal.
The continuity equation We know that the continuity equation is a differential equation which includes the charge density and the current density and that it is a scalar equation. This leads us to calculate the contraction of the outer product between the 4-gradient
and the 4-current J
J 0
(3.27)
This is covariant version of the continuity equation, thus in space-time the continuity equation is simply stated as the 4-current density is divergence-free!
Lorentz force We know that the left hand side of Lorentz force equation is a 3-force.
Obviously we should use the covariant 4-vector force instead. And on the right hand side
of th Lorentz equation is a 3-vector quantity involving charge density and current density
and the E and B fields. The EM fields are of course contained in the field tensor F . To
get a vector quantity from F and J we contract these so our guess is
F
F J
(3.28)
F J
(3.29)
3.3. S OLVED
23
EXAMPLES
0
(3.30)
E ND
OF EXAMPLE
3.3 %
qq o
A A p
r
0
Ex
Ey
Ez
Ex
0
cBz
cBy
Ey
cBz
0
cBx
s tt
Ez
cBy
cBx u
0
(3.31)
v
E c2
are obtained via a Lorentz transformation of the corresponding fields in the rest system of
the charge.
Solution
We wish to transform the EM fields. The EM fields in a covariant formulation of electrodynamics is given by the electromagnetic field tensor
F
oqqr
cB3
0
cB3
cB2
E1
0
cB1
E2
cB2
cB1
0
E3
E1
E2
E3
0
s tt
u
(3.32)
L L F
(3.33)
~LF0 L
(3.34)
where
L
qp
q o
0
0
0
1
0
0
0
0
1
0
st
t
0
0
(3.35)
E XAMPLE 3.4
24
1
where
2 and v c where v v1 x.
The fields in the rest system S0 are
q
x0 x y0 y z0 z
^
40 x0 2 y0 2 z0 2 3
E0
B0
0
(3.36)
(3.37)
so that
F
poqqr
0
0
0
E10
0
0
0
E20
E10
E20
E30
0
0
0
0
E30
sMtt
u
(3.38)
~L qorq
~
LFL
0
0
0
E10
qw
q o
0
0
qw
qr o
0
1
0
0
0
0
0
E20
0
0
1
0
E10
E20
0
0
0
E30
st
t
0
0
E10
E30
2
1 2 E10
E10
E20
E30
0
s tt
0
0
qoq
u
r
E10
E0
oqqr E20
3
E10
E20
E30
0
0
E20
0
0
E30
0
1
0
0
0
0
0
E20
0
0
1
0
st
t
0
0
0
0
0
E30
u
st
E10 t
E20
E30 u
E10
(3.40)
st
t
1 2 E10
0
E2
u
E30
2 0
2 0
E1 E1
2
(3.39)
(3.41)
E1
E2
E3
E10
B1
0
(3.45)
B2
(3.46)
(3.42)
(3.43)
(3.44)
E20
E30
0
E
c 3
0
B3
E
c 2
oqqr
where s
x0
y0
z0
ct 0
vt
sMtt
u L oqrq
(3.47)
x
y
z
ct
sMtt
u poqqr
sMt
t
x s
y
u
z
ct x
(3.48)
3.3. S OLVED
q O x s x y y zz
^
40 2 x s 2 y2 z2 3
(3.49)
s x 2 y 2 z2
R20
s
25
EXAMPLES
x R0 sin
q
40 2
2
x s
y2
2
+ R0 cos
2
R0
x s
z2
(3.50)
y2 z2
2
(3.51)
(3.52)
R20 y2 z2
y2
z2
(3.53)
R20 y2 z2 2 1 ~
R20 1 2 sin2
(3.54)
2
1
(3.55)
R0
q
1 2
^
3
40
3
R0 1 2 sin2
(3.56)
E ND
OF EXAMPLE
3.4 %
26
L ESSON 4
Lagrangian and
Hamiltonian
Electrodynamics
4.1
Coverage
4.2
Formulae used
mc2
q qv A
(4.1)
A useful Lagrangian density for EM field and its interaction with charged particles
is given by
1
0 c2 B2
E 2
A j
2
(4.2)
27
28
4.3
L ESSON 4. L AGRANGIAN
E XAMPLE 4.1
AND
H AMILTONIAN E LECTRODYNAMICS
Solved examples
Solution
We know from analytical mechanics that the canonical momentum P is found through
L
Pi
(4.3)
vi
so with
mc2
q qv A
(4.4)
L
we find that
L
vi
vv
mc2 1 i 2 i q qv j A j !
vi
c
]
v
i
qAi
mc2 |
v v
c2 1 cj 2 j
Pi
m vi qAi
P p qA
.
(4.5)
(4.6)
d
q qv j A j
q qv j A j
xi
dt vi
v A
dA
q j j q i
q
xi
xi
dt
dA
Q q q v A q
dt
Qi
(4.7)
f
6
f
A
A q
q q v A q v
q v A
t
5
A
A q
q qv
t
qE qv B
Draft version released 15th November 2000 at 20:39
(4.8)
(4.9)
4.3. S OLVED
29
EXAMPLES
OF EXAMPLE
4.1 %
E XAMPLE 4.2
1
c 2 B2
2 0
*
E 2 W A j
(4.10)
A=
=
(4.11)
(4.12)
Solution
Let us insert the gauge transformation relations into the Lagrangian density. Remembering
that E and B are invariant under gauge transformations we find that
'
E 2 A=
j =
(4.13)
t
1
c2 B2 E 2 A= j = j
(4.14)
2 0
t
1
(4.15)
c2 B2 E 2 A= j =
j
j
2 0
t
t
1
0 c2 B2 E 2 A= j =
j
j
(4.16)
2
t
t
(4.17)
0
j t
=
1
c 2 B2
2 0
30
L ESSON 4. L AGRANGIAN
AND
H AMILTONIAN E LECTRODYNAMICS
where
we have used the continuity equation. We see that
as , but what about the other two terms?
Some thought reveals that the neccessary condition for a Lagrangian to be physically acceptable is not the Lagrangian itself is invariant but rather that the variation of the action
integral S p
d3x dt is invariant. So now we would like to check that the gauge
transformations indeed do not affect any the variation of the action. Now it is possible
to proceed in two different ways to do this: one is simply carry out the integration in the
definition of the action integral and check that its variation is zero, or two, remembering
that the variation of the action is equivalent to the Euler-Lagrange equations, one could
plug in the Lagrangian density (4.17) into Euler-Lagrange equations to check the resulting
equations differ from the Maxwell equations.
Let us use the first alternative. Since the action is linear in
Sext
33
(
j
!
t
it is sufficient to examine
d3x dt
33 j dS dt 3 d3x
33 j dS dt 3 d3x
t1
t0
t1
33
t
j!
d3x dt
(4.18)
t0
where we have used the continuity equation. Furthermore, if we assume no flux source/sink
at infinity then we can write
Sext
3 d3x
t1
(4.19)
t0
Now when taking the variation, we realize that we must hold t0 and t1 the end point of the
particle path fixed, and thus
Sext 0
(4.20)
As one would expect, gauge transformations of the potentials do not effect the physics of
the problem!
E ND
OF EXAMPLE
4.2 %
L ESSON 5
Electromagnetic
Energy, Momentum
and Stress
5.1
Coverage
Here we study the force, energy, momentum and stress in an electromagnetic field.
5.2
Formulae used
Poyntings vector
S
(5.1)
Ei D j Hi B j
1
E D Hk Bk
2 ij k k
(5.2)
31
32
AND
S TRESS
Table 5.1. The following table gives the relevant quantities. The field vectors in
this table are assumed to be real. If the given fields are complex, use the real part
in the formulas.
Name
Energy density
Intensity
Momentum density
Stress
5.3
E XAMPLE 5.1
Symbol
Uv
S
PEM
T
Formula
D 12 H B
E H
E H
Hi B j 12 i j Ek Dk
1
2E
Ei D j
Hk Bk
SI unit
J/m3
W/m2
kg/s m2
Pa
Solved examples
EM QUANTITIES POTPOURRI
Determine the instantaneous values of the energy density, momentum density, intensity
and stress associated with the electromagnetic fields for the following cases:
A
E E e ei C k 1 x 1
H k0 k2 E
a)
Br
B
for k1
, k2 k3 0.
i
0 m cos
2 r 3
0 m sin
4 r 3
E and B H.
Solution
Background
(a) Progressive wave This case is an example of a progressive or propagating wave.
Since E and H are complex we must first take their real parts:
Re E
E0 cos k1 x1 t e2
Re H
k1
E cos k1 x1 t e3
0
(5.3)
(5.4)
5.3. S OLVED
2
k12
E 2 cos2 k1 x1 t ~
E0 cos2 k1 x1 t
2
2 2 0
[
E02 cos2 x1 t
Uv
33
EXAMPLES
(5.5)
(5.6)
k
E0 cos k1 x1 t e2 1 E0 cos k1 x1 t e3 +
[
2
E0 cos2 x1 t e2 e3
]
[
2
E0 cos2 x1 t e1
]
(5.7)
(5.8)
(5.9)
The momentum is
PEM
[
[
S E02 cos2 x1 t
(5.10)
The stress is
T11
T21
T22
T32
T33
2 2
E
H
2 2
2 3
k12 2
* E02 cos2 k1 x1 t
E cos2 k1 x1 t
2
2 0
* E02 cos2 k1 x1 t
T31 T12 0
2
E22
H
2
2 3
k12 2
E cos2 k1 x1 t
E02 cos2 k1 x1 t
2
2 0
T13 T23 0
2 2
H
E T22 0
2 3 2 2
*
(5.11)
(5.12)
(5.13)
(5.14)
(5.15)
(b) Evanescent wave This case is a example of an evanescent wave. We take real part
of the fields keeping in mind the fact that k1
A
Re E E0 e
x1
i is imaginary:
sin t e2
A
e1 Im E +
e1
E0 e
E0 A x 1
e
cos t e3
Re H
(5.16)
x1
cos t e2
(5.17)
34
AND
2 A 2 x 1 2
2 E02 A 2 x1
e
sin t
cos2 t
E0 e
2
2 2
A
E02 e 2 x1
2
sin2 t
cos2 t !
2
2
Uv
S TRESS
(5.18)
(5.19)
The intensity is
E02 A
e
2 x 1
(5.20)
The momentum is
PEM
S
E02 A
e
2 x 1
(5.21)
The stress is
2 2
E
H
2 2
2 3
T31 T12 0
T11
T21
T32
2 2
E
H
2 2
2 3
T13 T23 0
T22
T33
2
H
2 3
E02 e
2 x 1
sin2 t
2
cos2 t !
2
(5.22)
(5.23)
E02 e 2 x1
sin2 t
2
E T22
2 2
cos2 t !
2
2
(5.24)
(5.25)
E02 e 2 x1
sin2 t
cos2 t !
2
2
(5.26)
(c) Magnetic dipole This case is a magnetic dipole. The fields are real and in spherical
coordinates.
The energy density is
1
B2
2 0 r
Uv
B2 +
1
2 0
02 m2 4 cos2
16 2
r6
02 m2 sin2
!
16 2 r6
0 m
1
4 cos2 sin2 6
32 2
r
0 m2
1 3 cos2
32 2 r6
2
The intensity is S
components are
(5.27)
5.3. S OLVED
1 2
B
0 r
Trr
35
EXAMPLES
Uv
0 m2 cos2
4 2 r6
0 m2
1 3 cos2
32 2 r6
T r
Tr
0 m2
8 cos2 1 3 cos2
32 2 r6
m2
0 2 6 1 5 cos2
32 r
0 m2 sin cos
1
Br B
0
8
r6
0
T r
T r
1 2
B
0
Uv
0 m2 sin2
16 2 r6
(5.28)
(5.29)
(5.30)
(5.31)
0 m2
1 3 cos2
32 2 r6
0 m2
2 sin2 1 3 cos2
32 2 r6
m2
0 2 6 1 3 cos2 2 sin2
32 r
0
0
0
0
T
Tr
T
T
(5.32)
(5.33)
(5.34)
(5.35)
(5.36)
E ND
OF EXAMPLE
5.1 %
Solution
One the failures of Maxwells equations or classical electrodynamics is on question of
mass of point particles. From relativity one can show that fundamental particles should be
point-like. However, if one calculates the electromagnetic mass from Maxwells equation
one gets an infinite result due to the singularity in Gauss law
E 0 . This points
to the fact that Maxwells equations has a minimum length scale validity where quantum
mechanics takes over.
We will calculate this problem as follows: determine the electric field in all of space and
then integrate the formula for the energy density of the electric field. This integral will
contain the radius of the electron since it partitions the integration. We then relate this field
energy to the mass of the electron, which is a known quantity. We use this relation to solve
E XAMPLE 5.2
36
AND
S TRESS
E r
8 re + 0
(5.38)
e
4 3 re3 0
(5.39)
(5.40)
We can easily verify that (5.40) is indeed a solution to (5.39) and furthermore it is continuous with (5.37), thus making the solution unique.
We may now determine the energy density of the electric field due to the electron. It is
simply
1
E2
2 0
Uv r
Uv
(5.41)
; re +
Uv r 8 re +
1
e
32 2 0 r4
2
e
r2
32 2 0 re6
(5.42)
3
all space
4
3
re
0
8 re Uv r ; re d3x
Uv r
e2
r2 r2 dr d 3
32 2 0 re6
e2
r5
80 re6 5
re
e2
80
re
re
e2 1 2
r dr d
32 2 0 r4
3e2 1
200 re
(5.43)
Finally, we relate the total electric field energy to the rest mass of the electron and solve
for the electron radius,
m e c2
3e 1
200 re
re
(5.44)
me c2
3
e2
5 40 me c2
Draft version released 15th November 2000 at 20:39
(5.45)
5.3. S OLVED
37
EXAMPLES
This last result can be compared with the de facto classical electron radius which is defined
as
re
e2
40 me c2
(5.46)
OF EXAMPLE
5.2 %
S OLAR SAILING
E XAMPLE 5.3
Investigate the feasibility of sailing in our solar system using the solar wind. One technical
proposal uses kapton, 2 mm in thickness, with a 0.1 mm thick aluminium coating as material for a sail. It would weigh 1 g/m2 . At 1 AU (= astronomical unit distance between
sun and earth) the intensity of the electromagnetic radiation from the sun is of the order
1 / 4 103 W/m2 . If we assume the sail to be a perfect reflector, what would the acceleration
be for different incidence angles of the suns EM radiation on the sail?
Solution
In this problem we will use the principle of momentum conservation. If the continuity
equation for electromagnetic momentum PEM t
T Pmech t is integrated over
all space the stress term disappears and what is left says that a change in electromagnetic
momentum is balanced by mechanical force.
Imagine a localised pulse of a progressive electromagnetic wave incident on a plane. That
the wave is progressive means simply that it is purely radiation (more about progressive
waves in the next lesson). Let us characterize this wave by a Poynting vector S of duration
t. It travels in space with velocity c, it lights up the area A on the surface on the sail,
and, furthermore, S makes an angle 180 with the normal of the sail surface. The
momentum carried by such a wave is
2
Pbefore
EM V + S c A c t cos
pbefore
EM
(5.47)
pbefore
n
EM
Y SY
c
A t cos2
(5.48)
After hitting the sail, the pulse will be characterized with all the same quantities as before
the impact, except that the component along the surface normal will be opposite in sign.
This is because the sail is assumed to be perfectly reflecting. Thus,
pafter
n
EM ^
n
pbefore
EM
^
Y SY
c
A t cos2
(5.49)
Pmech F so for
38
pafter
EM
pbefore
EM
2
Y SY
A cos2
t
c
The other force components are zero by symmetry.
F ^
n
n
F ^
A
2
n
^
AND
S TRESS
(5.50)
Y SY
c
cos2
(5.51)
We now have the pressure exerted by a pulse charcterized by S incident on a surface with
an angle . The solar wind can be seen as a multitude of such pulses radiating radially
outwards from the sun. At 1 AU, the solar constant is 1.3 kW m2 . The solar constant is
the intensity of EM radiation or in other words, the magnitude of the Poynting vector Y S Y .
From equation (5.51), we find that
2
1 / 3 03 W/m2
cos2
3 108 m/s
(5.52)
Newtons second law gives the acceleration a of the sail (which, as we recall, weighs
A
3 10 3 kg/m2 ) due to the solar wind can at most be
a
0 / 9 10 5 N/m2
3 10 A 3 kg/m2
3 mm/s2
(5.53)
OF EXAMPLE
5.3 %
5.3. S OLVED
39
EXAMPLES
E XAMPLE 5.4
Determine the magnetic pressure due to the Earths magnetic field at the magnetic
A
poles (take B 6 10 5 T) and compare this with the Earths atmospheric pressure
5
(1 atm 1 / 01 10 Pa). Now assume that the magnetic dipole moment is proportional to
the angular velocity of the earth, how much faster would the angular velocity need to be
for the magnetic pressure to be comparable to the atmospheric pressure?
Solution
In this exercise we see that EMF can exert pressure not only via radiation pressure, but also
through static fields.
The static magnetic pressure is quantified in the Maxwellian stress tensor, which for this
case is
T
B2r 0
0
0
0 0
0 0
0 0
Uv
0 Uv
0
0
Uv
(5.54)
So by taking the inner product of T and the unit vector pointing in the direction the north
pole, namely 1 0 0 , we find the pressure in the radial direction to be
B2r
2 0
A 52
A
6 10
2 1 / 3 10
1 / 3 10 A 6
1 / 4 10 A 3 Pa
(5.55)
mB
Tperiod
(5.56)
so that
1
0
P0
P1
(5.57)
where we denote the current values of the pressure and rotational period time,
respectively and the hypothetical values 1 and T1 . Solving for T1 we arrive at
.
So with
T1
T1
T0
and T0
(5.58)
24 3600
105
10 s
(5.59)
E ND
OF EXAMPLE
5.4 %
40
L ESSON 6
Radiation from
Extended Sources
6.1
Coverage
We will study the important general problem of how to calculate the EM fields
induced by spatially extended, time-varying sources. This problem is solved in
different ways depending on the explicit form of the source distribution. For truly
extended bodies with non-monochromatic time dependence, we use the general
expressions for the retarded potentials. And for monochromatic, one-dimensional
current distributions, which we will call antennas here, we use the formulas given
below.
6.2
Formulae used
0
4
t x
1
40
j x t
d3x
x
x
x t
d3x
x
x
4
x
x
E
ik sin eik x x_
40 c x
x
41
42
L ESSON 6. R ADIATION
S
k2 sin2 0 c
32 2 x
x 2
FROM
E XTENDED S OURCES
where
6.3
E XAMPLE 6.1
j z e
ikz_ cos
dz
Solved examples
0 . Determine B, E and S
Solution
This problem belongs to the most general category of problems of the kind where given
a source distribution one wants to find the EM fields. This is because the source is not
monochromatic, so it is not an antenna, and furthermore it is an extended distribution, so
multipole expansion analysis is not possible. So we must use the most general formula for
calculating fields induced by time-varying sources, which in the Lorentz gauge take the
form
j t = 9 x=H
Y x x= Y
A t 9 x @
0
3
4
d3x=
t 9 x +
1
40
t = x=
3 d3x= 9
Y x x= Y
Draft version released 15th November 2000 at 20:39
(6.1)
(6.2)
6.3. S OLVED
43
EXAMPLES
where the source time t = is to be replaced by t =y t ,Y x x=Y c so t = is seen as a function
of t, x and x= . The function t Y x x= Y c is known as the retarded time tret
= , an expression
which is meaningful only relative the field point t 9 x .
The algorithm of solution now that we have decided to use (6.1) is to find the explicit
form for j t = 9 x=H and t = 9 x=H , perform the integrations to obtain the potentials A t 9 x and
t 9 x and, finally, derive the E and B fields from the potentials in the normal fashion and
the Poynting vector from the fields.
Let us find the explicit expression for the current density j x = 9 t =H , which is illustrated in
Fig. 6.1. In many problems, the expressions for the sources consist of a time-dependent
part times a space-dependent part. This is one such case. The switching on at t =) 0 can
be written as a step function t = . And if we orient the wire along the z axis, the space
dependent part can be written x=H y=H J z .
On the other hand, the charge distribution t 9 x + 0 as given in the problem formulation.
This can be seen as and charges flowing in opposite directions such as to keep the
total charge density t 9 x + 0 but this could still have a total current density j t 9 x 0.
<
x= y= J t = z
(6.3)
t = 9 x= + 0
(6.4)
We insert (6.3) into (6.1), remembering to replace the source time t = with t
and perform the integration. The integrations over x = and y= are trivial:
J
t Y x z= z Y c
A t 9 x + z 0 3
dz=
Y x z= z Y
4 A
Y x x=Y c,
(6.5)
Figure 6.1. The current density distribution j is along the z axis and is turned on
at t = 0. We use cylindrical coordinates.
For the remaining z= integration we use cylindrical coordinates (see Figure BLP1.2cyl) so
we can write
44
L ESSON 6. R ADIATION
x r r
Y x x= Y
[
r2
FROM
E XTENDED S OURCES
(6.6)
(6.7)
= /
z2
The step function in the integrand is zero when its argument is less than zero. This means
that when integrating over z= , only those z= contribute which satisfy
>Y r r z= z Y c ; 0 .
ct
r2
z= 2
(6.8)
; 09
(6.9)
as
z= 2
r[ 2 8 c2t 2
z= 8Y c2t 2 r2 Y
.
or
c2 t 2
(6.10)
(6.11)
r 2 8 z= 8
r2 - a
c2 t 2
(6.12)
P
r
r c
" 2 r2 c
2 2 r2 2
Figure 6.2. This series of snapshots shows what the part of the current is seen at
the field point P at different field times t.
So we have now that
J a
A z 0
4 a
r2
1 " 1
z2
J
z 0 ln
4
1
dz
r2
c2 t 2
r2
c2 t 2
(6.13)
6.3. S OLVED
45
EXAMPLES
0, since t = 9 x=H+
0.
Now that we have the vector potential we can derive the E and B fields. The B field is
B
(6.14)
and since A only has its z -component different from zero in the cylindrical system, we
have
Az
(6.15)
B
r
Az
r .
Introducing f
f
r 1
2 2
r
ct f
(6.16)
0 J
1
ln
4 r
1
0 J 2
4 r f
Az
r
2 r
f
f
!!
(6.17)
0 J
1
(6.18)
r2
c2 t 2
0 J 2 r
A
t
1
J
z 0
ln
4 t
1
f
!!
f
(6.19)
But
f
t
1 2r2
2 f c2 t 3
1 r2
f c2 t 3
(6.20)
and so
1
ln
t
1
1
1
f
!!4
f
f 2 t
f 1 f
2
ft
(6.21)
and
E
0 J
| 1
2 t 1
z
Notice that as t
'
(6.22)
r2
c2 t 2
then E
'
0!
46
L ESSON 6. R ADIATION
1
E B
0
0
J2
r
2
2 2 rt 1 cr2t 2
1
J
| 0
0 2 r 1
r2
c2 t 2
2 t
0 J
1
r2
c2 t 2
FROM
E XTENDED S OURCES
z
(6.23)
(6.24)
From this expression we can for example calculate the radiated power per unit length by
integrating over a cylindrical surface C enclosing the wire:
S dC
0
J2
2 t 1 cr22t 2
(6.25)
We see that an infinite power is transmitted starting at t 0 and r 0 which travels out
to infinite r. So in practice it is impossible to produce this physical setup. This is due to
the quick turn on. In the physical world only gradual turn ons are possible. This many
be seen as a consequence of what is known as Gibbs phenomenom.
In the above we tacitly employed the retarded potential without discussing the possibility of using the advanced potential. Let us see what happens if we apply the advanced
potential to this problem. The only thing that changes from the outset is that the source
time t = is replaced by the advanced time ta - t Y x x=Y c instead of the retarded time
tr - t Y x x=Y c. With ta as the argument to the step function, the contribution to the
integral comes only from those z= which satisfy
"Y r r z= z Y c ; 0 . t 8Y r r z= z Y c
(6.26)
If we assume that
8 09
(6.27)
.
.
z = 2 r 2 ; c2 t 2
[
z= ;>Z
c2 t 2 r 2
or
z=
8"
c2 t 2
r2 9
(6.28)
(6.29)
c2 t 2
r 2 8 z=
(6.30)
One can proceed further and calculate the resulting integral. But what is interesting is that
now we see that the relation (6.9) seems to say that we have no information about what
happened before turn on, while the relation (6.27) says we have no information about what
happened after turn on. Physics seems to be conspiring on us in such a way that we cannot
compare the advanced and the retarded potential at the same time!
E ND
OF EXAMPLE
6.1 %
6.3. S OLVED
47
EXAMPLES
tribution in the form of a standing wave with current nodes at the its endpoints. Determine
the angular distribution of the radiated electromagnetic power from the antenna.
Hint: It can me convenient to treat even m and odd m separately.
Solution
One realises that the setup in this problem is an antenna, since we have a monochromatic
source and the current is an extended one-dimensional distribution. Thus, we may use the
antenna formulae, but since only the radiated effect is ask for all we need is
S +
k2 sin2 0 c
Y
32 2 r2
Y 2 r
(6.31)
where
l
43 A j z= e
l
ikz cos
dz=
(6.32)
l
a) even .
z=
sin kz=
l
b) odd
z=
cos kz=
Figure 6.3. Depending on whether the length of the wire is an even (as seen in
a) or odd (as seen in b) multiple of half the wave length, the current distribution
is sin kz= or cos kz= .
We perform both integrations over z= from l to l which is the total length L, which is
a multiple m of 2. These facts give us a relation between the integration limits and k,
namely
m
L 2l m 2 . l
/
(6.33)
2 k
Draft version released 15th November 2000 at 20:39
E XAMPLE 6.2
48
L ESSON 6. R ADIATION
FROM
E XTENDED S OURCES
(6.34)
and so
l
u kz=
3 A J0 cos kz= e A ikzn cos dz= dz= 1 k du
l
m 2
J
A
A
03 A
eiu e iu e iu cos du
2k m 2
m 2
A
A A
J
eiu C 1 cos D
e iu C 1 cos D
0
2k i 1 cos
i 1 cos A m 2
m
m
1 cos sin
1 cos
2
2
sin
(6.36)
Let us consider the case when m is even so the current distribution is written
j z= + J0 sin kz=
(6.37)
We remember the relation (6.33) which is still valid but now m is an even number. So we
have that
6.3. S OLVED
2k
A
J0 3 m e
ikz cos
2k
m 2
J0
2ik
J0
2ik
J0 ei C 1
2ik
3 A
3 A
m 2
m 2
m 2
J0
ik sin2
J0
ik sin2
49
EXAMPLES
iu cos
A
e C
u kz=
dz= 1 du
u
A
eiu e iu du
iu 1 cos
D e A iu C 1}
eA iC 1A
i 1 cos
cos
m
2
cos
ei C 1
e A i C 1}
i 1 cos
cos
m R
cos
2
2J0
cos T
ik sin2
2J0
sin T
i
k sin2
.Y Y 2
m
2
D du
D
m
2
cos
D m2
m 2
m 2
m R
1 cos sin P 1
1 cos sin P 1 cos
2
m R
m R
T sin P 1 cos
sin P 1 cos
V
cos T sin P 1
cos
m R
cos
2
m
m
sin T
cos V
2 V
2
m
cos V
2
m R
2 V~
m
m
cos sin T
cos T
cos V+
2 V
2
sin P 1 cos
(6.38)
4J02
m
sin2 T
cos V
4
2
2
k sin
(6.39)
(6.40)
E ND
OF EXAMPLE
6.2 %
50
E XAMPLE 6.3
L ESSON 6. R ADIATION
FROM
E XTENDED S OURCES
Solution
We need the formula
where
k2 sin2
32 2 r2
0
Y
0
L 2
43 A dz= e A
L 2
ik cos z
Y 2 r
(6.41)
n j z=
(6.42)
L 2
J0 3 A dz= eik C 1 A cos D zn
L 2
L 2
A
eik C 1 cos D zn
1 i
A
e e
J0
sin
ik 1 cos A L 2
2i
kL
2
sin
1 cos !
J0
k 1 cos
2
4J02
k2 1 cos
Y Y2
sin2
kL
2
(6.43)
1 cos !
(6.44)
Finally we have
J02 sin2
8 2 r2
0 sin2 kL
2 1 cos
r
0
1 cos 2
(6.45)
E ND
OF EXAMPLE
6.3 %
6.3. S OLVED
51
EXAMPLES
E XAMPLE 6.4
Microwave links are based on direct waves, i.e., propagation along the line of sight between
the transmitter and receiver antennas. Reflections from the ground or a water surface may
in unfavourable cases distort the transmission. Study this phenomenon using the following
simple model:
The transmitter antenna T is a horizontal half-wave dipole placed a distance h 1 above the
ground level. The receiving antenna R is in the main lobe from T, at a horizontal distance
D from T, and at height h2 . The signal at R is considered to be composed of the direct wave
and a wave reflected from the ground. The reflection is assumed to cause a phase shift
in the wave, but no loss of power. The ground is considered flat over the distance D, and
h1 9 h2 D.
1. Calculate the electric field E (magnitude and direction) in at R if the transmitter
antenna is fed with a current j.
2. Discuss the meaning of the result.
3. Simplify the result for the case h1 h2
D.
r
1
h2
r2
h1
h1
Solution
The Fourier transform of the E field in the far zone (radiation field) from a half-wave dipole
antenna, i.e., a linearly extended current distribution with length 2 in the z direction,
E rad r +
i eikr
r
40 c r
4
dz= eik B xn j k
3 A
4
Since k k r and we study the radiation in the maximum direction, i.e., perpendicular to
j z=~ J cos kz= z so that k x=m 0, this expressions simplifies to
Draft version released 15th November 2000 at 20:39
52
L ESSON 6. R ADIATION
FROM
E XTENDED S OURCES
4
3 A
dz= J cos kz= M r z r
4
ikr
ik e J
iJ eikr
z
J sin kzKML = A 4 N 4 z
40 c r k
20 c r
ik eikr
40 c r
E rad r ~
2
A superposition of the direct and reflected contributions (with the distance from the transmitter T to the receiver R equal to r1 and r2 , respectively), with due regard to the phase
shift (corresponding to a change of sign in the current), gives the Fourier transform of
the total E field at the far zone point R :
Etot rad r @
Since h1 9 h2
refl
Edir
rad r E rad r +
iJ
20 c
eikr1
r1
eikr2
! z
r2
(6.46)
h h 2
h2 h2
h h
r
h1 h2 2 D 1 2 D 1 2 1 2 r
2D
2D
D
2
2
2
2
^
h
h
h
h
h
h
r
2
D 1 2 1 2 r
r2
D2 h1 h2 2 D 1
2D
2D
D
2
r1
D2
where
r
D
h21
h22
D
(6.47)
2D
and
2h1 h2
(6.48)
D
is the difference in path distance. Insertion of (6.47) and (6.48) into (6.46), with
k c 2 , gives the Fourier transform of the field at R
r
Etot
rad r +
iJ eikr A
Te
20 c D
eikr 2 z
V
ikr 2
2 h1 h2
J eikr
sin
z
0 c D
D
Re Etot rad e
i t
J
2 h1 h2
2
sin
cos
0 cD
D
D
h21
h22
2D
! t z
n
E ND
OF EXAMPLE
6.4 %
L ESSON 7
Multipole Radiation
7.1
Coverage
We look at electric dipole, magnetic dipole and electric quadrupole radiation. Multipole radiation analysis is important since it simplifies the calculation of radiation
fields from time-varying field and since EM multipoles exist in many fields of
physics such as astrophysics, plasma physics, atomic physics, and nuclear physics.
7.2
Formulae used
0 eik x
p 1 k
4 x
E rad x
1 eik x
p 1 k
k
40 x
E rad x
0 eik x
m k k
4 x
k eik x
m k
40 c x
53
54
B rad
i 0 eik x
k Q k
8 x
E rad
i eik x
k Q k k
80 x
7.3
E XAMPLE 7.1
Solved examples
(a) Determine the time-dependent electromagnetic fields at large distances from the dipole.
(b) Determine the radiated average power angular distribution and the total radiated
power.
Solution
(a) We can write the time-varying dipole momentum relative the location of the dipole
as
p t +
p0 cos t x
sin t y
(7.1)
which represents a constant dipole moment p0 times unit vector rotating with angular frequency . This can also be rewritten in complex form
p t +
Re p0 e
p t +
A
p0 e
or
i t
i t
ip0 e A
ip0 e A
i t
i t
y
(7.2)
c/ c/
(7.3)
where c.c., stands for the complex conjugate of the term opposite the sign. In what
follows we use the convention that we write the dipole expression as a complex quantity
but we drop the c.c. term, which is commonplace when discussing harmonic oscillation. It
is easy to identify the Fourier component of the dipole moment in this case
p t +
p0 x
ip0 y /
(7.4)
We notice, in this complex space variable space, that the y component has the phase factor
i ei 2 relative the x component, which is due the circular rotation.
We would like to express (7.4) in spherical components rather than Cartesian components
since the expressions for dipole fields in spherical components are simpler, so we transform
7.3. S OLVED
55
EXAMPLES
(7.5)
(7.6)
(7.7)
(7.8)
(7.9)
(7.10)
Now that we have the Fourier component of the dipole moment expressed in spherical
components we insert this into the dipole radiation fomulae:
0 eikr
p k
4 r
1 eikr
E
p k
k
40 r
(7.11)
(7.12)
First we calculate
p
k p0 ke sin cos i
1
0
0
p0 kei i cos
i
(7.13)
(7.14)
p
k
k p0 ke 0 i cos
1 0
0
2 i
p0 k e cos i
i
(7.15)
(7.16)
so finally we can write the field in space and time coordinates (remember:
A
B t 9 x + Re B x e i t ),
0
ei C kr t D
Re
p0 kei i cos
4
r
p k
}
0 0 Re ei C kr A t D ei i cos
4 r
B t 9 x :
(7.17)
(7.18)
0 2 p0
sin kr t cos cos kr t
4 cr
2 p0
cos cos kr t sin kr t
40 c2 r
Draft version released 15th November 2000 at 20:39
(7.19)
(7.20)
56
1
2 0 E
1
1 0 1 1
E B
2 0
20 4 40 r2
k p k p
k
2
2
32 0 r
Y p k Y 2 r
32 2 0 r2
k
O p
p
k k
k
p
p
k +
k O~
(7.21)
(7.22)
(7.23)
0 p20 4
1 cos2 r
32 2 cr2
(7.24)
3
Sr r2 d
(7.25)
0 p20 4
2 3
d sin 1 cos2 +
2
32 c
0
1
0 p20 4
2
2 3
2
A 1 dx 1 x ~
32 c
0 p20 4
6 c
(7.26)
(7.27)
(7.28)
E ND
E XAMPLE 7.2
OF EXAMPLE
7.1 %
ROTATING MULTIPOLE
Two point charges of equal charge q are located in the xy plane at either end of the diameter
of a circle of radius a. The particles rotate with a constant angular speed in the plane of
the circle.
Determine
Solution
7.3. S OLVED
r1=
r2=
v1=
v2=
p
57
EXAMPLES
a cos 0t x sin 0t y
* r1=
0 a sin t x cos 0t y
* v1=
(7.29)
(7.30)
(7.31)
(7.32)
qn rn= 0
(7.33)
1
rn=
2
n
qvn=
1
2
1
r=
2 1
3 A
qv1=
r1=
q
v1= O+ q0 a2 z
z q0 a2
Qi j
qn xin= x= jn
(7.36)
Qxx
Qyy
Qxy
Qiz
(7.37)
(7.38)
(7.39)
(7.40)
Qxx
Qyy
Qxy
qa2 1 e A i20t
A
qa2 1 e i20t
Qyx qa2 ie A i20t
(7.41)
(7.42)
(7.43)
e i t d
q0 a2 z
(7.34)
(7.35)
Fourier transform
Qxx
1
3 A 1 1 2ei20t 1 2e A
2
1
qa2
W 20 O
2
qa2
Qyy
Qxy
1
3 A qa2 1 cos 20t ei t dt
2
qa2 1 2 20 1 2
qa2
1
2
3 A
qa2 ro
i20 t
(7.44)
20 O
(7.45)
ei t dt
(7.46)
(7.47)
A i C 2 0 A D t
dt qa2 i 20
1 0 0
0 1 0
0 0 0
s
u qa2 20 r
o
(7.48)
1
0
0
0 0
1 0
0 0
s
(7.49)
58
i0 eikr
k Q
k
8 r
i eikr
k Q
k
80 r O
1
E
2 0
(7.50)
(7.51)
(7.52)
E ND
E XAMPLE 7.3
OF EXAMPLE
7.2 %
ATOMIC RADIATION
A transition in an atom is described as quantum matrix element of a radiation operator
between the un-normalised eigenstates
f
e
re A r exp i E 1 t cos
e A r exp i E 0 t
(7.53)
c f f ce e
(7.54)
where cf and ce can be viewed as given constants, chosen such that becomes normalised.
According to semiclassical theory, one can interpret the magnitude squared of the wave
function as a particle density function. Determine, according to this semiclassical interpretation, the power emitted by the atom via the dipole radiation which appears due to the
transition between the two states.
The power from an electric dipole is given by
P
0 4 Y p Y 2
12 c
(7.55)
Solution
The charge density is
q
q TY c f f Y 2 c f ce e f c f ce e f "Y ce e Y 2 V
(7.56)
(7.57)
Only the two last terms contribute because they are non-static. Via inspection we find the
Fourier components for E1 E0 OU
(7.59)
7.3. S OLVED
59
EXAMPLES
3
qc f ce r cos exp
2r r r r2 dr d
qc f ce 3 r4 exp 2r dr ! 3
1
2
qc f ce 4! 3
2
0
12qc f ce 0 9 0 9 3
0
2
0
12qc f ce 0 9 0 9 2 3
r d3x
2
0
(7.60)
cos sin r d d
(7.61)
3
0
1 2 sin
(7.62)
(7.63)
(7.64)
(7.65)
(7.66)
(7.67)
(7.68)
and a similar integral for the complex conjugate term. The power from the electric dipole
radiation due to the transition between the two states is given by
P
0 4 Y p Y 2
12 c
16 q2 0 E1
3c 4
E0 4
2
c f ce /
(7.69)
E ND
OF EXAMPLE
7.3 %
C LASSICAL P OSITRONIUM
E XAMPLE 7.4
Calculate the radiation from a positron-electron head-on collision and subsequent annihilation, classically by assuming that the particles travel at a constant velocity v 0 c up until
the time the annihilate.
Solution
Background Consider a system of localised charges in motion. If we assume that we
are observing at a distance Y x Ym R much greater than extension Y x = Y of the charge system
and further that v c, it can be argued that the source time t = is approximately t R c
(where R Y x Y ) instead of t ,Y x x=Y c. This is because, due to the first assumption
x
(7.70)
Y x x= Y R x=
R
Since the timescale for the system is of the order T Y x =Y v and since v c we have that
Y x=Y
c
(7.71)
60
t R c
(7.72)
R R
1
p
40 rc2 O
(7.77)
R
0
p
4 rc2
R
It can be shown that the angular and spectral distribution of energy is
B
d
d d
O p
R R
2 c3
(7.78)
(7.79)
2qe v0t z
t
t
0
z 0
; 0
2p
2qe v0 t z
t2
For the spectral and angular distributions of the radiation we have
(7.80)
(7.81)
(7.82)
d
R R 2
p
(7.83)
d d
2 c3
where subscript denotes the Fourier transform of the dipole. Now the Fourier transform
of the dipole is simply p 2qv0 z 2 1 2 , so that
Draft version released 15th November 2000 at 20:39
7.3. S OLVED
61
EXAMPLES
d
q2e v20 z R 2 2 c3
(7.84)
d d
In the final result we notice that there is no dependence on so the spectral density is
white noise.
E ND
OF EXAMPLE
7.4 %
62
L ESSON 8
Radiation from
Moving Point
Charges
8.1
Coverage
In our series on deriving fields from given sources we have come to the most fundamental case: the moving point charge. The fields are derived from the LinardWiechert potentials. In what follows we will assume that the motion x t is known
in advance. We find that accelerating charges radiate. We also look other mechanisms for a point charge to radiate such as Cerenkov emission.
8.2
Formulae used
According to the Formulae (F.2326), the fields from a charge in arbitrary motion
are given by
E t x
q
Rv 1
40 s3
B t x
x
x
s
x
x
v2
c2
E t x
c x
x
x
x
v
c
Rv
x
x
c2
(8.1)
(8.2)
(8.3)
63
64
L ESSON 8. R ADIATION
FROM
x
x vc
x
x
Rv
t
t
x
x
x
s
(8.4)
(8.5)
Alternative formulae
x
x
v
c
Rv
x
x0
tt
E XAMPLE 8.1
r
r
s
8.3
(8.6)
x
s
r
r
r
v
c
(8.7)
rv
r
r cos
r0
(8.8)
r 1
cos
(8.9)
1
cos
(8.10)
Solved examples
Solution
In general the fields due to a single point charge may be written as
E
B
Ev Erad
Bv Brad
(8.11)
(8.12)
8.3. S OLVED
65
EXAMPLES
Ev and Bv are known as the velocity fields and Erad and Brad are known as acceleration
fields. These can be derived from the Linard-Wiechert potentials and result in
q
r r 1 2
40 s3
q
T r P r r R
3
40 cs
Ev
Erad
(8.13)
(8.14)
and
Bv
Brad
q
r 1 2
3
40 cs
q
r Pr
r
2
3
40 c s
(8.15)
R
r
(8.16)
- v c and
s r r
(8.17)
(8.18)
where
The big outer square brackets P R indicate that one should evaluate their content at the
retarded time t = t Y x x= t =HQY c, where x= t =H is the given motion of the charged
particle. It is this that makes the equations for the fields difficult to evaluate in general. It
may not be difficult to get an expression for x = t =H , but then to solve the equation for the
retarded time t = t Y x x= t = QY c for t = t = t 9 x . In this case it is not necessary to
perform this transformation of variables since we are not interested in the time evolution,
so we drop the brackets.
It is easy to verify that for uniform motion of the charge q, or in other words - 0, that
Erad
c
so the Poynting vector in this case is
S
1
E
0 c
E +
0
E E
0
(8.19)
|
r r +
r02 r0 2
(8.20)
q
1
1 2 3
40
r0 1 2 sin2
r
3 2 0
(8.21)
where is the angle between r0 and . Inserting this relation for E into the relation for S
we obtain
66
L ESSON 8. R ADIATION
S
So
and
(8.22)
r0
r0 sin
(8.23)
r0
0 q 2
1
1 2 2 6
r0 r0
2
2
2
2
3
0 16 0
r0 1 sin
FROM
(8.24)
So finally
S
q2
v 1 2 2 sin
2
16 0 r04 1 2 sin2 3
(8.25)
And now when we integrate the Poynting vector over a spherical surface A with radius R
which encloses the moving charge which results in
P
S dA S r d 0
(8.26)
since the Poynting vector is not radial, so a charge in uniform motion in vacuo, does not
radiate energy.
E ND
E XAMPLE 8.2
OF EXAMPLE
8.1 %
Solution
We consider only the formulas for the radiation fields, for which the denominator is the
cube of the retarded relative distance
r r r r cos r 1 cos
Now, we have that v r 0, so
r rv v : rv J rKML vN v r rv
& 0
v r r r
v r2 r2 cos
v r2 1 cos
v rs
s
(8.27)
(8.28a)
(8.28b)
(8.28c)
(8.28d)
8.3. S OLVED
67
EXAMPLES
0 q
q v r
v rs 0 2
4 s3
4 s
The Poynting vector is given by
Erad
1
Y E Y 2 r
0 c
v2 r
02 q2
2
2
16 0 c r 1 cos
(8.29)
(8.30)
Remember that the Poynting vector represents energy flux per unit time at the field point
at the time t. The charges energy loss must be related to the time t = , i.e., the time when the
energy was emitted!
U
Sr r 2
t
(8.31)
but
U
t=
q2
t U
v2
r2
1 cos 0 2 2
t= t
16 c r 1 cos 4
0 q2 v2
1
16 2 c 1 cos 3
(8.32)
E ND
OF EXAMPLE
8.2 %
T HE L ARMOR FORMULA
E XAMPLE 8.3
Derive the Larmor formula by calculating the radiated power of an accelerating point
charge due to electric dipole emission. Apply the Larmor formula to linear harmonic motion and circular motion.
Solution
The Larmor formula is a very useful equation for deriving the power of emission from
non-relativistic accelerating charged particles. It can be derived from the radiation fields
of a non-relativistic (v c) charged particle, but it can also be seen as an electric dipole
relative a co-moving coordinate system. We shall investigate the latter.
First derive the power emitted by an electric dipole. Using a time domain (non-Fourier)
version of the dipole fields:
1
p
40 rc2
r
0
p r
4 rc
we find the Poynting vector to be
B
(8.33)
(8.34)
68
L ESSON 8. R ADIATION
1
E B
0
1
16 2 0 r2 c3
1
2
16 0 r2 c3
1
2
16 0 r2 c3
p
T Y p
r
r Y 2 r
Y p
p
FROM
r +
r p r p r V
r Y 2 r
(8.35)
S r r2 sin d d
1
16 2 0 c3
1
16 2 0 c3
Y p Y 2
80
c3
3
3
0
2
0
Y p
r Y 2 sin d d
Y p Y 2 sin3 d d
sin3 d
Y p Y 2 4
80 c3 3
Y p Y 2
60 c3
(8.36)
p t +
qx t
(8.37)
P t +
q2 Y x Y 2
60 c3
q2 Y a Y 2
60 c3
(8.38)
x t .
x0 cos 0t
2
9 a t + x t + 0 x0 cos 0t /
q2 04 x20 cos2 0t
P t +
6 c3
(8.39)
(8.40)
q2 04 x20
120 c3
(8.41)
Circular motion
x t +
R0 cos 0t
9 y t + R0 sin 0t 9
Draft version released 15th November 2000 at 20:39
(8.42)
8.3. S OLVED
69
EXAMPLES
q2 04 R20
60 c3
P t ~
(8.43)
Validity of the Larmor formula The Larmor formula although not covariant in form
can indeed be extended such as to be valid for all inertial frames.
One point that should definitely be raised is power radiated for more than one accelerating
charge. It is not so simple that one may assume that power is proportional to the number of
sources N. What must be understood is whether or not the sources are radiating coherently
or incoherently. For example, consider the above case of circular motion. If we had a
large collection of particles such as the case with in a particle storage ring or in circular
wire, the radiation is not automatically proportional to N. If they are bunched the power is
proportional to N 2 , this is coherent radiation. If the charges are distributed homogeneously
the radiated power is 0. And if the charges are distributed evenly but fluctuate thermally
then power is proportional to N, this is incoherent radiation.
E ND
OF EXAMPLE
8.3 %
E XAMPLE 8.4
Solution
The motion of the charge is given by x = t =H+
x x= 2 y2 x v t = t 2 y2
x2 v2 t t = 2 2vx t t = y2
r2
but r
c
n
r2
v t = t so that
(8.44)
t= 2 0
(8.45)
v2 t t = 2 2vx t t = y2
v2
c2
!
n2
c2
t
n2
t= 2 0
2
2
2
t t = 2vx t t = y x 0
(8.46)
(8.47)
This is a quadratic equation in t t = . For a fixed t we would in other words have two
values of t = . It remains to show that the result is physically reasonable.
Draft version released 15th November 2000 at 20:39
70
L ESSON 8. R ADIATION
2vx Z
t t = #
vx Z
t t=
4v2 x2 4 x2 y2 v2
2 v 2 c2 n 2
FROM
c2 n 2
x2 y2 v2 y2
2
v2 nc2
c2
n2
(8.48)
c2 2
x y2 #; v2 y2
n2
c2
y2
;
n 2 v2
x2 y2
2
sin c ; sin2
8 c
.
.
.
.
(8.49)
(8.50)
(8.51)
(8.52)
t t=
v2
c2
!
n2
Z
]
c2 2
x
n2
Z vx
]
y2 v2 y2 vx
c2 x2 y2
n 2 v2
x2
y2
x2
vx
vx Z sin2 c 2 tan2 1
cos
0 ]
1
1 cos2 c 1 cos2 1
vx Z
cos2
0 ]
1
cos2
c
1
vx Z 1
cos2
0 ]
1
t t=
J
v2
c2
(8.53)
cos2
c
2 ! vx Z 1
1
2
n
cos
]
0J
1N
MK L
N
KML
(8.54)
E ND
OF EXAMPLE
8.4 %
L ESSON 9
Radiation from
Accelerated Particles
9.1
Coverage
The generation of EM fields via Linard-Wiechert potentials are considered simultaneously with the Lorentz force to give a self-consistent treatment of radiation
problems. In the previous lesson we solve radiation problems from given expressions for the motion. Now we consider how charges actually move in the EM fields
and thus present the Lorentz force. We also discuss the effect of radiation on the
motion of the radiating body itself known as radiative reaction
9.2
Formulae used
q E v
(9.1)
71
72
L ESSON 9. R ADIATION
9.3
E XAMPLE 9.1
FROM
ACCELERATED PARTICLES
Solved examples
Solution
Background on equations of motion for charged particles As we know, the motion of a charged particle in electric and magnetic fields is given by the Lorentz force
F
q E v
B
(9.2)
This can be seen as a definition of the E and B fields and also the fundamental equation
for measuring the fields. The E field gives the force parallel to direction of motion and B
gives the force perpendicular to the direction of motion. In this certain sense the Lorentz
force is trivial: it is simply a definition of the EM fields.
The equation for the Lorentz force is relativistically correct as it stands, as long as one
interpretes
d m v
(9.3)
F
dt
^
where m is the mass of the particle, - 1 1 v2 c2 , v is the three-velocity, and t is the
time. The fact that the Lorentz force in this form
d m v
q E v B
(9.4)
dt
is Lorentz invariant is not immediately clear but can easily be shown. On the other hand
this equation is difficult to solve for v t because contains v.
In many cases one has conditions which are non-relativistic and under such conditions it
is possible to simplify (9.3) and thus also (9.4). One simply uses the fact that as v c ' 0
then ' 1. In this case F mdv dt, (i.e. the Newtonian force definition), so the Lorentz
force becomes
q
dv
E v B
(9.5)
dt
m
This equation is the equation of motion for non-relativistic charge particles in EM fields.
The motion As is well known, charged particles are uniformly accelerated in a static
and homogenous electric field, and in a static and homogenous magnetic field the charged
particles perform circular motion. So what happens in a combined electric and magnetic
fields?
The equations of motion are given in (9.5) The first step in solving this equation is to
separate the motion into motion parallel with the B field which we will denote with v and
9.3. S OLVED
dv
73
EXAMPLES
v y v and thus
dv
q
q
E v
B
E y
dt
dt
m
m
and since the X and { are mutually orthogonal the equation separates into
q
E
dt
m
and for the perpendicular direction
dv
dt
(9.6)
(9.7)
q
E v
B
(9.8)
m
dv
q
q
v
E
B
(9.9)
.
dt
m
m
This is a first order linear ordinary differential equation in v . It is also inhomogeneous
which means that we have both a solution to the homogenous equation and solution to the
inhomogeneous equation. Let us call the solution to the inhomogeneous equation D and
the solution to the homogenous equation R . For the inhomogeneous solution, we notice
that neither the inhomogeneity nor the coefficient of the zeroth order term depends on t so
the solution D itself cannot be time dependent. With this assumption, and that is left is
D B E . This is easily solved, by taking the cross product of this equation with B
we find
D B B E B
2
J DKML BN B B D E B
& 0
E
B
D
2
.
.
(9.10)
(9.11)
(9.12)
q
r
m
(9.13)
on the other hand, can be shown to have the form r c e i t where c is a constant
vector perpendicular to B and satisfies c ic4 0 (where the denotes scalar product
defined as the inner product of the vectors), so for example c 1 9 i 9 0 if B is in the 3
direction.
Interpretation of the Motion Having derived the solutions for the motion of the
charged particle we are now in a position to describe it in words. First of all we see that the
motion consists of three separate parts. First we have the motion along the B field which
simply is not effected by the B field, so that part of the E field which is along the B field
accelerates the charge as if there were no B field. The motion perpendicular to the B field
can further be separated into two parts. One part, which here we have denoted r , represDraft version released 15th November 2000 at 20:39
74
L ESSON 9. R ADIATION
FROM
ACCELERATED PARTICLES
y
x
Figure 9.1. Motion of a charge in an electric and a magnetic field. Here the
electric field is along the y axis and the magnetic field is along the z axis. There
is a background velocity in the z axis for reasons of clarity.
ents so called gyro-harmonic rotation, i.e. the particle moves in a circular orbit, with its
axis of rotation parallel with the B field with period of rotation 2 c where c qB m is
known as the cyclotron frequency. This gyro-harmonic motion is the complete solution to
our problem if the E field was not considered. The second part of the motion in the perpendicular plane however, which we here denoted D involves both the B field and the E field.
It is known as the drift velocity since this is the velocity of the centre of the gyro-harmonic
motion.
The motion is illustrated in Figure 9.1.
E ND
E XAMPLE 9.2
OF EXAMPLE
9.1 %
Solution
Background Among other things, Maxwells introduction of the electric and magnetic
fields was a mathematical technique to divide the work of solving the motion of charge
particles in EM fields. Instead of action-at-a-distance, the fields naturally divide the probDraft version released 15th November 2000 at 20:39
9.3. S OLVED
75
EXAMPLES
lem into two parts: the generation of fields from moving charged particles and the motion
of charged particles in EM fields. In a certain sense this can be seen as a division into, on
the one side a cause and on the other side, an effect of EM interaction, but the formulas
governing these processes are not very symmetric with respect to each other in classical
electrodynamics: the generation of fields from moving point charges is determined via the
Linard-Wiechert potentials and the motion of point charged particles is determined by the
Lorentz force.
Despite the lack of symmetry, this division seems successful at describing EM interaction
except for one thing: it fails to describe the self-interaction of charged point particles. We
know that the motion of charged particles is governed by the Lorentz force, but at the same
time we know that acceleration of charged particles causes energy emission. These two
facts have been, until now, treated separately, but taken together we realize something is
missing. This can be seen for example in the case of a single charge under the influence
of a mechanical force in a region of space with no EM fields except for the field from the
charge itself. From the Linard-Wiechert potentials we know that the mechanical force
will cause the charge to radiate and if energy is to be conserved the emission must take
its energy from the kinetic energy. Since the Lorentz force is zero and there are no other
electromagnetic interactions that we know of, as of yet, we have no way of accounting for
this radiative friction. This, as of yet not mentioned force, is known as radiative reaction
or the radiative damping force.
One question that comes to mind, after the above discussion, is how can so many problems be described by classical electrodynamics without considering radiative reaction?
Obviously, it should have a negligible effect in most cases but what are the limiting conditions? Certainly, these should be determined by considering the conditions under which
the energy emitted is of the same order as the kinetic energy of the charge. If we consider
non-relativistic motion, the energy emitted by a charge accelerating at the order of a, under
a period of duration of order T , is given by the Larmor formula and is of the order of
q2 a2 T
60 c3
Erad
(9.14)
On the other hand, the acceleration bestows the charge with kinetic energy on the order of
Ekin
m aT
2
(9.15)
m aT
2
(9.16)
or
T
q2
30 c3
(9.17)
76
L ESSON 9. R ADIATION
FROM
ACCELERATED PARTICLES
and if we define the characteristic time as - q2 30 c3 we can say that the effects of the
.
radiative reaction are negligible in measurements made over timescales on order of T
Accounting for radiative reaction Having demonstrated the need for a force which
accounts for radiative effects of accelerating charges we set out to determine its form. From
the conservation of energy it is clear that the force we are looking for, which we will denote
Frad , must satisfy
3 Frad v dt 3
q2 v v
dt
60 c3
(9.18)
where have integrated the Larmor formula over time on the right hand side. Partial integration yields
3 Frad v dt 3
q2 v v
dt
60 c3
q2 v v
60 c3
(9.19)
Frad
q2 v
!
60 c3
v dt
0
(9.20)
and so
Frad
q2 v
60 c3
m v
(9.21)
We now can correct the equation of motion to include the radiative reaction
m v v ~
Fext
(9.22)
OF EXAMPLE
9.2 %
9.3. S OLVED
77
EXAMPLES
E XAMPLE 9.3
An electron moves in a circular orbit in a synchrotron under the action of the homogeneous
magnetic field B0 .
(a) Calculate the energy which is lost in electromagnetic radiation per revolution if the
speed v of the electron is assumed to be constant.
(b) At which particle energy does the radiated energy per revolution become equal to the
particles total energy, if B0
Hint:
1 / 5 T?
d m0 v
dt =
dU
dt =
F9
0 q2 v 2 4
6 c
Solution
(a) Since we wish to consider synchrotron motion we use the relativistically correct
equation of motion
dp
qv B
dt =
where the right hand side is the Lorentz force with E
momentum 4-vector so
dp
d
d
m v + m0 v m0 v
dt =
dt = 0
dt =
where
d
dt =
d
dt =
But v v
v 2 c2 1
1 3d
3
v v + 2 v v
2
2c
dt
1 3
d v2
!
dt = c2
(9.23)
(9.25)
(9.26)
d
dt
0 and thus
dp
(9.27)
m0 v
dt =
This means that the power necessary to keep the particle in a circular orbit, disregarding
radiation losses, is
v
dp
2
m0 v v 2 v2 v v !> 0
dt =
c
(9.28)
which comes as no surprise. However, the circular motion does emit EM radiation. To find
the expression for the power loss due to radiation we need an expression for the accelera-
78
L ESSON 9. R ADIATION
FROM
ACCELERATED PARTICLES
qv
(9.29)
The acceleration is found by taking the norm of this last equation and since v
have is the scalar equation
m0 Y v YY q Y vB0
X B all we
(9.30)
Y q Y B0
m0
(9.31)
The factor m 0 - c is known as the synchrotron angular frequency and is the relativistic
0
value of the angular frequency for gyro-harmonic motion or cyclotron angular frequency.
qB
Now we can insert this expression into the relativistic generalisation of the Larmor formula
q2 v2 4
q2 2 4
dU
0
0 c v2
dt =
6 c
6 c
To find the energy per revolution, we need the period of revolution which is T
so
dU
2 0 q2 c2 4 2
v
dt =
c 6 c
0 Y q Y 3 3 B0 v 2
3cm0
T
Urev
0 q2 c2 4 2
v
6 c
(9.32)
2 c ,
(9.33)
(9.34)
(b) The task here is to equate the total energy and the radiated energy and solve for
velocity and then see what total energy that velocity is associated with.
The total energy is E m0 c2 and the radiated energy is Urev
which when equated gives
c2 1
v2 2
0 q 3 B0
3cm20
c2
v2
1
v2
c2
(9.35)
0 q 3 B0
3cm20
1
1
0 Y q Y 3 3 v2 B0 3cm0 ,
(9.36)
(9.37)
0 q3 B0
3cm20
1
2cm20
0 Y q Y 3 B0
(9.38)
Thus the particle energy for which the radiated energy is equal to the total particle energy
is
9.3. S OLVED
79
EXAMPLES
m 0 c2 1
2cm20
0 Y q Y 3 B0
0 / 16 TeV
(9.39)
This is obviously the upper limit for which radiation effect can be neglected in the treatment
of synchrotron motion of the particle.
E ND
OF EXAMPLE
9.3 %
E XAMPLE 9.4
Solution
The relativistic equation of motion is
dp
dt =
d
m v +
dt = 0
2
v v v !, qE0
2
c
m0 v
(9.40)
X v X E0 then
Since v
m0 v 1 2 2 +
m0 v
.
m0 3 v
2
!
1 2
m0 v 1
1
!, m0 v 2
1 2
Y q Y E0
(9.41)
Y q Y E0
(9.42)
Y q Y E0
m0 3
(9.43)
Having found v we try to derive the radiation field generated by this motion; we have that
0 q
r rv v
4 s3
0 q
1
r
4 r3 1 cos 3
Erad
PT r r V
c
0 q
1
r
3
3
4 r 1 cos
ro r
0 q
4 r3 1
0 q
4 r3 1
0 qv
4 r 1
1
cos
1
cos
sin
cos
r
v
v R
r u
Jv KML vN
c
&
v
r2 v sin
(9.44)
80
L ESSON 9. R ADIATION
FROM
ACCELERATED PARTICLES
where have used the particular geometry of the vectors involved. Then we may determine
the Poynting vector, which is
S
1
2
Erad r
0 c
Sr r
(9.45)
The radiated energy per unit area per unit time corresponding to this is
U
Sr r 2
t
rad 2
r2
0 c
(9.46)
but the energy radiated per unit area per unit time at the source point is
U
t=
U t
t t=
U s
t r
rs
2
Erad
0 c
cos 02 q2 v2
r2 1
0 q2 v2
sin2
2
16 c 1 cos
0 c
16 2
sin2
r2 1 cos
(9.47)
0 q2 v2
sin3
3
2
d
5
16 2 c
0 1 cos
q2 v2 1 1 x2
x cos
2 0 2 3
dx
dx
d sin
8 c A 1 1 x 5
q2 v2 1
1
x2
2 0 2 3 A
! dx
8 c
1
1 x 5 1 x 5
q2 v2 4
0 q2 v2 4 6
0 q4 E02
1
2 0 2
2
8 c 3 1 2 3
8 2 c 3
6 m20 c
U
t=
U
d
t=
We compare this expression for the radiated energy with the total energy E
dE
dt =
m0 c2
d
dt =
(9.48)
m0 c2 , so
(9.49)
m0 v
Y q Y E0
Y q Y E0
(9.50)
(9.51)
Hence
9.3. S OLVED
m0 v
81
EXAMPLES
1
Y q Y E0 1 2 !
J KML N
d
dt =
(9.52)
v2
c2
d
dt =
Y q Y E0 v
(9.53)
m0 c2
and
dE
dt =
m0 c2
Finally, for E0
dU
dt =
Y q Y E0 v
m0 c2
Y q Y E0 v
(9.54)
dE
dt =
0 Y e Y 3 E0
6 cm20 v
E0 4 10 7 1 / 6 10 19 3
v 6 2 / 998 108 9 / 11 10 A 31
1/ 1
10
c
(9.55)
2 / 998
E ND
12 E0
10 8 m/s)
OF EXAMPLE
9.4 %
82
L ESSON 9. R ADIATION
FROM
ACCELERATED PARTICLES