Classical Electromagnetism by R-Fitzpatrick

Classical Electromagnetism:
An intermediate level course

Richard Fitzpatrick
Professor of Physics
The University of Texas at Austin
Contents
1 Introduction
1.1 Intended audience
1.2 Major sources . . .
1.3 Preface . . . . . .
1.4 Outline of course .
1.5 Acknowledgements
.
.
.
.
.
7
7
7
8
9
10
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
11
11
11
14
16
18
21
24
25
26
27
30
31
33
34
35
40
44
46
50
3 Time-independent Maxwell equations

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2 Coulombs law . . . . . . . . . . . . . . . . . . . . . . . . . . . .
53
53
53
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2 Vectors
2.1 Introduction . . . . . . .
2.2 Vector algebra . . . . . .
2.3 Vector areas . . . . . . . .
2.4 The scalar product . . . .
2.5 The vector product . . . .
2.6 Rotation . . . . . . . . . .
2.7 The scalar triple product .
2.8 The vector triple product
2.9 Vector calculus . . . . . .
2.10 Line integrals . . . . . . .
2.11 Vector line integrals . . .
2.12 Surface integrals . . . . .
2.13 Vector surface integrals .
2.14 Volume integrals . . . . .
2.15 Gradient . . . . . . . . . .
2.16 Divergence . . . . . . . .
2.17 The Laplacian . . . . . . .
2.18 Curl . . . . . . . . . . . .
2.19 Summary . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3.13
3.14
The electric scalar potential . . .

Gauss law . . . . . . . . . . . . .
Poissons equation . . . . . . . .
Ampères experiments . . . . . .
The Lorentz force . . . . . . . . .
Ampères law . . . . . . . . . . .
Magnetic monopoles? . . . . . .
Ampères circuital law . . . . . .
Helmholtzs theorem . . . . . . .
The magnetic vector potential . .
The Biot-Savart law . . . . . . .
Electrostatics and magnetostatics
.
.
.
.
.
.
.
.
.
.
.
.
4 Time-dependent Maxwells equations

4.1 Introduction . . . . . . . . . . . .
4.2 Faradays law . . . . . . . . . . . .
4.3 Electric scalar potential? . . . . . .
4.4 Gauge transformations . . . . . . .
4.5 The displacement current . . . . .
4.6 Potential formulation . . . . . . . .
4.7 Electromagnetic waves . . . . . . .
4.8 Greens functions . . . . . . . . . .
4.9 Retarded potentials . . . . . . . . .
4.10 Advanced potentials? . . . . . . .
4.11 Retarded fields . . . . . . . . . . .
4.12 Summary . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5 Electrostatics
5.1 Introduction . . . . . . . . . . . . . . .
5.2 Electrostatic energy . . . . . . . . . . .
5.3 Ohms law . . . . . . . . . . . . . . . . .
5.4 Conductors . . . . . . . . . . . . . . . .
5.5 Boundary conditions on the electric field
5.6 Capacitors . . . . . . . . . . . . . . . . .
5.7 Poissons equation . . . . . . . . . . . .
3
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
59
62
68
70
74
78
79
82
88
94
97
99
.
.
.
.
.
.
.
.
.
.
.
.
104
104
104
108
110
113
119
120
128
133
140
143
147
.
.
.
.
.
.
.
150
150
150
156
157
164
165
170
5.8
5.9
5.10
5.11
5.12
The uniqueness theorem . . . . . . . . . . . . .

One-dimensional solution of Poissons equation
The method of images . . . . . . . . . . . . . .
Complex analysis . . . . . . . . . . . . . . . . .
Separation of variables . . . . . . . . . . . . . .
6 Dielectric and magnetic media

6.1 Introduction . . . . . . . . . . . . . . . . . .
6.2 Polarization . . . . . . . . . . . . . . . . . . .
6.3 Boundary conditions for E and D . . . . . . .
6.4 Boundary value problems with dielectrics . .
6.5 Energy density within a dielectric medium . .
6.6 Magnetization . . . . . . . . . . . . . . . . .
6.7 Magnetic susceptibility and permeability . . .
6.8 Ferromagnetism . . . . . . . . . . . . . . . .
6.9 Boundary conditions for B and H . . . . . . .
6.10 Boundary value problems with ferromagnets
6.11 Magnetic energy . . . . . . . . . . . . . . . .
7 Magnetic induction
7.1 Introduction . . . . . . . .
7.2 Inductance . . . . . . . . .
7.3 Self-inductance . . . . . . .
7.4 Mutual inductance . . . . .
7.5 Magnetic energy . . . . . .
7.6 Alternating current circuits
7.7 Transmission lines . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
8 Electromagnetic energy and momentum

8.1 Introduction . . . . . . . . . . . . .
8.2 Energy conservation . . . . . . . . .
8.3 Electromagnetic momentum . . . . .
8.4 Momentum conservation . . . . . .
9 Electromagnetic radiation
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
171
176
178
183
188
.
.
.
.
.
.
.
.
.
.
.
194
194
194
197
198
203
204
206
207
210
211
213
.
.
.
.
.
.
.
216
216
216
218
222
225
231
234
.
.
.
.
242
242
242
246
250
253
9.1
9.2
9.3
9.4
9.5
9.6
9.7
9.8
9.9
9.10
9.11
9.12
9.13
Introduction . . . . . . . . . . . . . . . .
The Hertzian dipole . . . . . . . . . . . .
Electric dipole radiation . . . . . . . . . .
Thompson scattering . . . . . . . . . . . .
Rayleigh scattering . . . . . . . . . . . . .
Propagation in a dielectric medium . . . .
Dielectric constant of a gaseous medium .
Dielectric constant of a plasma . . . . . .
Faraday rotation . . . . . . . . . . . . . .
Propagation in a conductor . . . . . . . .
Dielectric constant of a collisional plasma
Reflection at a dielectric boundary . . . .
Wave-guides . . . . . . . . . . . . . . . .
10 Relativity and electromagnetism

10.1 Introdunction . . . . . . . . . . . . . .
10.2 The relativity principle . . . . . . . . .
10.3 The Lorentz transformation . . . . . .
10.4 Transformation of velocities . . . . . .
10.5 Tensors . . . . . . . . . . . . . . . . .
10.6 The physical significance of tensors . .
10.7 Space-time . . . . . . . . . . . . . . .
10.8 Proper time . . . . . . . . . . . . . . .
10.9 4-velocity and 4-acceleration . . . . .
10.10 The current density 4-vector . . . . . .
10.11 The potential 4-vector . . . . . . . . .
10.12 Gauge invariance . . . . . . . . . . . .
10.13 Retarded potentials . . . . . . . . . . .
10.14 Tensors and pseudo-tensors . . . . . .
10.15 The electromagnetic field tensor . . .
10.16 The dual electromagnetic field tensor .
10.17 Transformation of fields . . . . . . . .
10.18 Potential due to a moving charge . . .
10.19 Fields due to a moving charge . . . . .
5
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
253
253
259
260
263
266
267
268
273
276
279
281
292
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
298
298
298
300
304
307
313
314
319
321
321
323
324
325
327
331
334
337
337
339
10.20 Relativistic particle dynamics . . .

10.21 The force on a moving charge . . .
10.22 The electromagnetic energy tensor
10.23 Accelerated charges . . . . . . . .
10.24 The Larmor formula . . . . . . . .
10.25 Radiation losses . . . . . . . . . .
10.26 Angular distribution of radiation .
10.27 Synchrotron radiation . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
341
343
345
348
353
357
359
360
1 INTRODUCTION
Introduction
1.1 Intended audience

These lecture notes outline a single semester course intended for upper division
undergraduates.
1.2 Major sources

The textbooks which I have consulted most frequently whilst developing course
material are:
Classical electricity and magnetism: W.K.H. Panofsky, and M. Phillips, 2nd edition
(Addison-Wesley, Reading MA, 1962).
The Feynman lectures on physics: R.P. Feynman, R.B. Leighton, and M. Sands, Vol.
II (Addison-Wesley, Reading MA, 1964).
Special relativity: W. Rindler (Oliver & Boyd, Edinburgh & London UK, 1966).
Electromagnetic fields and waves: P. Lorrain, and D.R. Corson, 3rd edition (W.H. Freeman & Co., San Francisco CA, 1970).
Electromagnetism: I.S. Grant, and W.R. Phillips (John Wiley & Sons, Chichester
UK, 1975).
Foundations of electromagnetic theory: J.R. Reitz, F.J. Milford, and R.W. Christy,
3rd edition (Addison-Wesley, Reading MA, 1980).
The classical theory of fields: E.M. Lifshitz, and L.D. Landau, 4th edition [ButterworthHeinemann, Oxford UK, 1980].
Introduction to electrodynamics: D.J. Griffiths, 2nd edition (Prentice Hall, Englewood Cliffs NJ, 1989).
1 INTRODUCTION
1.3 Preface
Classical electromagnetic radiation: M.A. Heald, and J.B. Marion, 3rd edition (Saunders College Publishing, Fort Worth TX, 1995).
Classical electrodynamics: W. Greiner (Springer-Verlag, New York NY, 1998).
In addition, the section on vectors is largely based on my undergraduate lecture
notes taken from a course given by Dr. Stephen Gull at the University of Cambridge.
1.3 Preface
The main topic of this course is Maxwells equations. These are a set of eight
first-order partial differential equations which constitute a complete description
of electric and magnetic phenomena. To be more exact, Maxwells equations
constitute a complete description of the behaviour of electric and magnetic fields.
Students entering this course should be quite familiar with the concepts of electric
and magnetic fields. Nevertheless, few can answer the following important question: do electric and magnetic fields have a real physical existence, or are they
merely theoretical constructs which we use to calculate the electric and magnetic
forces exerted by charged particles on one another? As we shall see, the process
of formulating an answer to this question enables us to come to a better understanding of the nature of electric and magnetic fields, and the reasons why it is
necessary to use such concepts in order to fully describe electric and magnetic
phenomena.
At any given point in space, an electric or magnetic field possesses two properties, a magnitude and a direction. In general, these properties vary (continuously)
from point to point. It is conventional to represent such a field in terms of its
components measured with respect to some conveniently chosen set of Cartesian
axes (i.e., the conventional x-, y-, and z-axes). Of course, the orientation of these
axes is arbitrary. In other words, different observers may well choose different
coordinate axes to describe the same field. Consequently, electric and magnetic
fields may have different components according to different observers. We can
see that any description of electric and magnetic fields is going to depend on
8
1 INTRODUCTION
1.4 Outline of course
two seperate things. Firstly, the nature of the fields themselves, and, secondly,
our arbitrary choice of the coordinate axes with respect to which we measure
these fields. Likewise, Maxwells equationsthe equations which describe the
behaviour of electric and magnetic fieldsdepend on two separate things. Firstly,
the fundamental laws of physics which govern the behaviour of electric and magnetic fields, and, secondly, our arbitrary choice of coordinate axes. It would be
helpful if we could easily distinguish those elements of Maxwells equations which
depend on physics from those which only depend on coordinates. In fact, we can
achieve this by using what mathematicians call vector field theory. This theory
enables us to write Maxwells equations in a manner which is completely independent of our choice of coordinate axes. As an added bonus, Maxwells equations
look a lot simpler when written in a coordinate-free manner. In fact, instead of
eight first-order partial differential equations, we only require four such equations
within the context of vector field theory.
1.4 Outline of course

This course is organized as follows. Section 2 consists of a brief review of those
elements of vector field theory which are relevent to Maxwells equations. In
Sect. 3, we derive the time-independent version of Maxwells equations. In
Sect. 4, we generalize to the full time-dependent set of Maxwell equations. Section 5 discusses the application of Maxwells equations to electrostatics. In Sect. 6,
we incorporate dielectric and magnetic media into Maxwells equations. Section 7 investigates the application of Maxwells equations to magnetic induction.
In Sect. 8, we examine how Maxwells equations conserve electromagnetic energy
and momentum. In Sect. 9, we employ Maxwells equations to investigate electromagnetic waves. We conclude, in Sect. 10, with a discussion of the relativistic
formulation of Maxwells equations.
1 INTRODUCTION
My thanks to Prof. Wang-Jung Yoon [Chonnam National University, Republic of
Korea (South)] for pointing out many typographical errors appearing in earlier
editions of this work.
10
2 VECTORS
Vectors
2.1 Introduction
In this section, we shall give a brief outline of those aspects of vector algebra, vector calculus, and vector field theory which are needed to derive and understand
Maxwells equations.
This section is largely based on my undergraduate lecture notes from a course
given by Dr. Stephen Gull at the University of Cambridge.
2.2 Vector algebra

Q
P
Figure 1:
In applied mathematics, physical quantities are (predominately) represented

by two distinct classes of objects. Some quantities, denoted scalars, are represented by real numbers. Others, denoted vectors, are represented by directed line
elements in space: e.g., PQ (see Fig. 1). Note that line elements (and, therefore, vectors) are movable, and do not carry intrinsic position information. In
fact, vectors just possess a magnitude and a direction, whereas scalars possess
a magnitude but no direction. By convention, vector quantities are denoted by
bold-faced characters (e.g., a) in typeset documents, and by underlined characters (e.g., a) in long-hand. Vectors can be added together, but the same units
must be used, just like in scalar addition. Vector addition can be represented

using a parallelogram: PR = PQ + QR (see Fig. 2). Suppose that a PQ SR,
11
2 VECTORS
2.2 Vector algebra

R
b
Q
S
a
Figure 2:
b QR PS, and c PR. It is clear from Fig. 2 that vector addition is commutative: i.e., a + b = b + a. It can also be shown that the associative law holds: i.e.,
a + (b + c) = (a + b) + c.
There are two approaches to vector analysis. The geometric approach is based
on line elements in space. The coordinate approach assumes that space is defined
by Cartesian coordinates, and uses these to characterize vectors. In physics, we
generally adopt the second approach, because it is far more convenient.
In the coordinate approach, a vector is denoted as the row matrix of its components along each of the Cartesian axes (the x-, y-, and z-axes, say):
a (ax , ay , az ).
(2.1)
Here, ax is the x-coordinate of the head of the vector minus the x-coordinate of
its tail. If a (ax , ay , az ) and b (bx , by , bz ) then vector addition is defined
a + b (ax + bx , ay + by , az + bz ).
(2.2)
If a is a vector and n is a scalar then the product of a scalar and a vector is defined
n a (n ax , n ay , n az ).
(2.3)
It is clear that vector algebra is distributive with respect to scalar multiplication:

i.e., n (a + b) = n a + n b.
12
2 VECTORS
2.2 Vector algebra

y
Figure 3:
Unit vectors can be defined in the x-, y-, and z-directions as ex (1, 0, 0),
ey (0, 1, 0), and ez (0, 0, 1). Any vector can be written in terms of these unit
vectors:
a = a x ex + a y ey + a z ez .
(2.4)
In mathematical terminology, three vectors used in this manner form a basis of
the vector space. If the three vectors are mutually perpendicular then they are
termed orthogonal basis vectors. However, any set of three non-coplanar vectors
can be used as basis vectors.
Examples of vectors in physics are displacements from an origin,
r = (x, y, z),
(2.5)
and velocities,
v=
r(t + t) r(t)
dr
= lim
.
dt t0
t
(2.6)
Suppose that we transform to a new orthogonal basis, the x 0 -, y 0 -, and z 0 -axes,

which are related to the x-, y-, and z-axes via a rotation through an angle
around the z-axis (see Fig. 3). In the new basis, the coordinates of the general
displacement r from the origin are (x 0 , y 0 , z 0 ). These coordinates are related to
the previous coordinates via the transformation:
x 0 = x cos + y sin ,
(2.7)
y 0 = x sin + y cos ,
(2.8)
z 0 = z.
(2.9)
13
2 VECTORS
2.3 Vector areas
We do not need to change our notation for the displacement in the new basis.
It is still denoted r. The reason for this is that the magnitude and direction of r
are independent of the choice of basis vectors. The coordinates of r do depend on
the choice of basis vectors. However, they must depend in a very specific manner
[i.e., Eqs. (2.7)(2.9)] which preserves the magnitude and direction of r.
Since any vector can be represented as a displacement from an origin (this is
just a special case of a directed line element), it follows that the components of
a general vector a must transform in an analogous manner to Eqs. (2.7)(2.9).
Thus,
ax 0 = ax cos + ay sin ,
(2.10)
ay 0 = ax sin + ay cos ,
(2.11)
az 0 = a z ,
(2.12)
with similar transformation rules for rotation about the y- and z-axes. In the coordinate approach, Eqs. (2.10)(2.12) constitute the definition of a vector. The
three quantities (ax , ay , az ) are the components of a vector provided that they
transform under rotation like Eqs. (2.10)(2.12). Conversely, (ax , ay , az ) cannot
be the components of a vector if they do not transform like Eqs. (2.10)(2.12).
Scalar quantities are invariant under transformation. Thus, the individual components of a vector (ax , say) are real numbers, but they are not scalars. Displacement vectors, and all vectors derived from displacements, automatically satisfy
Eqs. (2.10)(2.12). There are, however, other physical quantities which have
both magnitude and direction, but which are not obviously related to displacements. We need to check carefully to see whether these quantities are vectors.
2.3 Vector areas

Suppose that we have planar surface of scalar area S. We can define a vector
area S whose magnitude is S, and whose direction is perpendicular to the plane,
in the sense determined by the right-hand grip rule on the rim (see Fig. 4). This
quantity clearly possesses both magnitude and direction. But is it a true vector?
We know that if the normal to the surface makes an angle x with the x-axis then
14
2 VECTORS
2.3 Vector areas
Figure 4:
the area seen looking along the x-direction is S cos x . This is the x-component
of S. Similarly, if the normal makes an angle y with the y-axis then the area
seen looking along the y-direction is S cos y . This is the y-component of S. If
we limit ourselves to a surface whose normal is perpendicular to the z-direction
then x = /2 y = . It follows that S = S (cos , sin , 0). If we rotate the
basis about the z-axis by degrees, which is equivalent to rotating the normal to
the surface about the z-axis by degrees, then
Sx 0 = S cos ( ) = S cos cos + S sin sin = Sx cos + Sy sin , (2.13)
which is the correct transformation rule for the x-component of a vector. The
other components transform correctly as well. This proves that a vector area is a
true vector.
According to the vector addition theorem, the projected area of two plane
surfaces, joined together at a line, looking along the x-direction (say) is the xcomponent of the resultant of the vector areas of the two surfaces. Likewise, for
many joined-up plane areas, the projected area in the x-direction, which is the
same as the projected area of the rim in the x-direction, is the x-component of
the resultant of all the vector areas:
X
S=
Si .
(2.14)
i
If we approach a limit, by letting the number of plane facets increase, and their
areas reduce, then we obtain a continuous surface denoted by the resultant vector
15
2 VECTORS
2.4 The scalar product
area:
S=
Si .
(2.15)
It is clear that the projected area of the rim in the x-direction is just S x . Note that
the rim of the surface determines the vector area rather than the nature of the
surface. So, two different surfaces sharing the same rim both possess the same
vector area.
In conclusion, a loop (not all in one plane) has a vector area S which is the
resultant of the vector areas of any surface ending on the loop. The components
of S are the projected areas of the loop in the directions of the basis vectors. As a
corollary, a closed surface has S = 0, since it does not possess a rim.

A scalar quantity is invariant under all possible rotational transformations. The
individual components of a vector are not scalars because they change under
transformation. Can we form a scalar out of some combination of the components of one, or more, vectors? Suppose that we were to define the ampersand
product,
a & b = ax by + ay bz + az bx = scalar number,
(2.16)
for general vectors a and b. Is a & b invariant under transformation, as must
be the case if it is a scalar number? Let us consider an example. Suppose that
a = (1, 0, 0) and b = (0, 1, 0). It is easily seen that a & b = 1. Letus now
rotate
the basis through

45
about the z-axis. In the new basis, a = (1/ 2, 1/ 2, 0)

and b = (1/ 2, 1/ 2, 0), giving a & b = 1/2. Clearly, a & b is not invariant under
rotational transformation, so the above definition is a bad one.
Consider, now, the dot product or scalar product:
a b = ax bx + ay by + az bz = scalar number.
(2.17)
Let us rotate the basis though degrees about the z-axis. According to Eqs. (2.10)
(2.12), in the new basis a b takes the form
a b = (ax cos + ay sin ) (bx cos + by sin )
16
2 VECTORS
+(ax sin + ay cos ) (bx sin + by cos ) + az bz
(2.18)
= a x bx + a y by + a z bz .
Thus, a b is invariant under rotation about the z-axis. It can easily be shown that
it is also invariant under rotation about the x- and y-axes. Clearly, a b is a true
scalar, so the above definition is a good one. Incidentally, a b is the only simple
combination of the components of two vectors which transforms like a scalar. It
is easily shown that the dot product is commutative and distributive:
a b = b a,
a (b + c) = a b + a c.
(2.19)
The associative property is meaningless for the dot product, because we cannot
have (a b) c, since a b is scalar.
We have shown that the dot product a b is coordinate independent. But what
is the physical significance of this? Consider the special case where a = b. Clearly,
a b = ax2 + ay2 + az2 = Length (OP)2 ,
(2.20)
if a is the position vector of P relative to the origin O. So, the invariance of a a

is equivalent to the invariance of the length, or magnitude, of vector a under
transformation. The length of vector a is usually denoted |a| (the modulus of
a) or sometimes just a, so
a a = |a|2 = a2 .
(2.21)
B
ba
a
Figure 5:
17
2 VECTORS
2.5 The vector product
Let us now investigate the general case. The length squared of AB (see Fig. 5)
is
(b a) (b a) = |a|2 + |b|2 2 a b.
(2.22)
(AB)2 = (OA)2 + (OB)2 2 (OA) (OB) cos ,
(2.23)
However, according to the cosine rule of trigonometry,
where (AB) denotes the length of side AB. It follows that

a b = |a| |b| cos .
(2.24)
Clearly, the invariance of ab under transformation is equivalent to the invariance

of the angle subtended between the two vectors. Note that if a b = 0 then either
|a| = 0, |b| = 0, or the vectors a and b are perpendicular. The angle subtended
between two vectors can easily be obtained from the dot product:
cos =
ab
.
|a| |b|
(2.25)
The work W performed by a constant force F moving an object through a

displacement r is the product of the magnitude of F times the displacement in
the direction of F. If the angle subtended between F and r is then
W = |F| (|r| cos ) = F r.
(2.26)
The rate of flow of liquid of constant velocity v through a loop of vector area S
is the product of the magnitude of the area times the component of the velocity
perpendicular to the loop. Thus,
Rate of flow = v S.
(2.27)

We have discovered how to construct a scalar from the components of two general vectors a and b. Can we also construct a vector which is not just a linear
combination of a and b? Consider the following definition:
a x b = (ax bx , ay by , az bz ).
18
(2.28)
2 VECTORS
Is a x b a proper vector? Suppose that a = (1, 0, 0), b = (0, 1, 0). Clearly,
axb =
0. However,
if we rotate
thebasis through 45 about the z-axis then

a = (1/ 2, 1/ 2, 0), b = (1/ 2, 1/ 2, 0), and a x b = (1/2, 1/2, 0). Thus,
a x b does not transform like a vector, because its magnitude depends on the
choice of axes. So, above definition is a bad one.
Consider, now, the cross product or vector product:
a b = (ay bz az by , az bx ax bz , ax by ay bx ) = c.
(2.29)
Does this rather unlikely combination transform like a vector? Let us try rotating
the basis through degrees about the z-axis using Eqs. (2.10)(2.12). In the new
basis,
cx 0 = (ax sin + ay cos ) bz az (bx sin + by cos )
= (ay bz az by ) cos + (az bx ax bz ) sin
= cx cos + cy sin .
(2.30)
Thus, the x-component of a b transforms correctly. It can easily be shown that

the other components transform correctly as well, and that all components also
transform correctly under rotation about the y- and z-axes. Thus, ab is a proper
vector. Incidentally, a b is the only simple combination of the components of
two vectors which transforms like a vector (which is non-coplanar with a and b).
The cross product is anticommutative,
a b = b a,
(2.31)
a (b + c) = a b + a c,
(2.32)
a (b c) 6= (a b) c.
(2.33)
distributive,
but is not associative:
The cross product transforms like a vector, which means that it must have a
well-defined direction and magnitude. We can show that a b is perpendicular
to both a and b. Consider a a b. If this is zero then the cross product must be
19
2 VECTORS

thumb
a b
middle finger
b
index finger
Figure 6:
perpendicular to a. Now
a a b = ax (ay bz az by ) + ay (az bx ax bz ) + az (ax by ay bx )
(2.34)
= 0.
Therefore, a b is perpendicular to a. Likewise, it can be demonstrated that a b

is perpendicular to b. The vectors a, b, and a b form a right-handed set, like the
unit vectors ex , ey , and ez . In fact, ex ey = ez . This defines a unique direction
for a b, which is obtained from the right-hand rule (see Fig. 6).
Let us now evaluate the magnitude of a b. We have
(a b)2 = (ay bz az by )2 + (az bx ax bz )2 + (ax bz ay bx )2
= (ax2 + ay2 + az2 ) (bx2 + by2 + bz2 ) (ax bx + ay by + az bz )2
= |a|2 |b|2 (a b)2
= |a|2 |b|2 |a|2 |b|2 cos2 = |a|2 |b|2 sin2 .
(2.35)
Thus,
|a b| = |a| |b| sin .
(2.36)
Clearly, a a = 0 for any vector, since is always zero in this case. Also, if
a b = 0 then either |a| = 0, |b| = 0, or b is parallel (or antiparallel) to a.
Consider the parallelogram defined by vectors a and b (see Fig. 7). The scalar
area is a b sin . The vector area has the magnitude of the scalar area, and is
20
2 VECTORS
2.6 Rotation
a
Figure 7:
normal to the plane of the parallelogram, which means that it is perpendicular to

both a and b. Clearly, the vector area is given by
S = a b,
(2.37)
with the sense obtained from the right-hand grip rule by rotating a on to b.
Suppose that a force F is applied at position r (see Fig. 8). The moment, or
torque, about the origin O is the product of the magnitude of the force and the
length of the lever arm OQ. Thus, the magnitude of the moment is |F| |r| sin .
The direction of the moment is conventionally the direction of the axis through
O about which the force tries to rotate objects, in the sense determined by the
right-hand grip rule. It follows that the vector moment is given by
M = r F.
(2.38)
2.6 Rotation
Let us try to define a rotation vector whose magnitude is the angle of the rotation, , and whose direction is the axis of the rotation, in the sense determined
by the right-hand grip rule. Is this a good vector? The short answer is, no. The
problem is that the addition of rotations is not commutative, whereas vector addition is commuative. Figure 9 shows the effect of applying two successive 90
rotations, one about x-axis, and the other about the z-axis, to a six-sided die. In
the left-hand case, the z-rotation is applied before the x-rotation, and vice versa
21
2 VECTORS
2.6 Rotation
P
r
r sin
Figure 8:
in the right-hand case. It can be seen that the die ends up in two completely
different states. Clearly, the z-rotation plus the x-rotation does not equal the xrotation plus the z-rotation. This non-commuting algebra cannot be represented
by vectors. So, although rotations have a well-defined magnitude and direction,
they are not vector quantities.
But, this is not quite the end of the story. Suppose that we take a general
vector a and rotate it about the z-axis by a small angle z . This is equivalent to
rotating the basis about the z-axis by z . According to Eqs. (2.10)(2.12), we
have
a 0 ' a + z ez a,
(2.39)
where use has been made of the small angle expansions sin ' and cos ' 1.
The above equation can easily be generalized to allow small rotations about the
x- and y-axes by x and y , respectively. We find that
a 0 ' a + a,
(2.40)
= x ex + y ey + z ez .
(2.41)
where
Clearly, we can define a rotation vector , but it only works for small angle
rotations (i.e., sufficiently small that the small angle expansions of sine and cosine
are good). According to the above equation, a small z-rotation plus a small xrotation is (approximately) equal to the two rotations applied in the opposite
22
2 VECTORS
2.6 Rotation
z
x
y
z-axis
x-axis
x-axis
z-axis
Figure 9:
23
2 VECTORS
2.7 The scalar triple product
c
b
Figure 10:
order. The fact that infinitesimal rotation is a vector implies that angular velocity,
,
t0 t
= lim
(2.42)
must be a vector as well. Also, if a 0 is interpreted as a(t+t) in the above equation

then it is clear that the equation of motion of a vector precessing about the origin
with angular velocity is
da
= a.
(2.43)
dt
2.7 The scalar triple product
Consider three vectors a, b, and c. The scalar triple product is defined a b c.
Now, b c is the vector area of the parallelogram defined by b and c. So, a b c
is the scalar area of this parallelogram times the component of a in the direction
of its normal. It follows that a b c is the volume of the parallelepiped defined
by vectors a, b, and c (see Fig. 10). This volume is independent of how the triple
product is formed from a, b, and c, except that
a b c = a c b.
(2.44)
So, the volume is positive if a, b, and c form a right-handed set (i.e., if a lies
above the plane of b and c, in the sense determined from the right-hand grip
rule by rotating b onto c) and negative if they form a left-handed set. The triple
24
2 VECTORS
product is unchanged if the dot and cross product operators are interchanged:
a b c = a b c.
(2.45)
The triple product is also invariant under any cyclic permutation of a, b, and c,
a b c = b c a = c a b,
(2.46)
but any anti-cyclic permutation causes it to change sign,

a b c = b a c.
(2.47)
The scalar triple product is zero if any two of a, b, and c are parallel, or if a, b,
and c are co-planar.
If a, b, and c are non-coplanar, then any vector r can be written in terms of
them:
r = a + b + c.
(2.48)
Forming the dot product of this equation with b c, we then obtain
r b c = a b c,
(2.49)
so
rbc
.
(2.50)
abc
Analogous expressions can be written for and . The parameters , , and
are uniquely determined provided a b c 6= 0: i.e., provided that the three basis
vectors are not co-planar.
=

For three vectors a, b, and c, the vector triple product is defined a (b c).
The brackets are important because a (b c) 6= (a b) c. In fact, it can be
demonstrated that
a (b c) (a c) b (a b) c
(2.51)
and
(a b) c (a c) b (b c) a.
25
(2.52)
2 VECTORS
2.9 Vector calculus
Let us try to prove the first of the above theorems. The left-hand side and
the right-hand side are both proper vectors, so if we can prove this result in
one particular coordinate system then it must be true in general. Let us take
convenient axes such that the x-axis lies along b, and c lies in the x-y plane. It
follows that b = (bx , 0, 0), c = (cx , cy , 0), and a = (ax , ay , az ). The vector b c
is directed along the z-axis: b c = (0, 0, bx cy ). It follows that a (b c) lies
in the x-y plane: a (b c) = (ay bx cy , ax bx cy , 0). This is the left-hand side
of Eq. (2.51) in our convenient axes. To evaluate the right-hand side, we need
a c = ax cx + ay cy and a b = ax bx . It follows that the right-hand side is
RHS = ( [ax cx + ay cy ] bx , 0, 0) (ax bx cx , ax bx cy , 0)
= (ay cy bx , ax bx cy , 0) = LHS,
(2.53)
which proves the theorem.
2.9 Vector calculus

Suppose that vector a varies with time, so that a = a(t). The time derivative of
the vector is defined
da
a(t
+
t)
a(t)
.
= lim
(2.54)
t0
dt
t
When written out in component form this becomes
da
dax day daz
=
,
,
.
dt
dt dt dt
!
(2.55)
Suppose that a is, in fact, the product of a scalar (t) and another vector b(t).
What now is the time derivative of a? We have
dax
d
d
dbx
= ( bx ) =
bx +
,
dt
dt
dt
dt
(2.56)
da d
db
=
b+
.
dt
dt
dt
(2.57)
which implies that
26
2 VECTORS
2.10 Line integrals

Q
P
x
Figure 11:
It is easily demonstrated that

da
db
d
(a b) =
b+a
.
dt
dt
dt
(2.58)
da
db
d
(a b) =
b+a
.
dt
dt
dt
(2.59)
Likewise,
It can be seen that the laws of vector differentiation are analogous to those in
conventional calculus.
2.10 Line integrals

Consider a two-dimensional function f(x, y) which is defined for all x and y.
What is meant by the integral of f along a given curve from P to Q in the x-y
plane? We first draw out f as a function of length l along the path (see Fig. 11).
The integral is then simply given by
ZQ
f(x, y) dl = Area under the curve.
(2.60)
P
As an example of this, consider the integral of f(x, y) = x y between P and

Q along the two routes indicated in Fig. 12. Along route 1 we have x = y, so
27
2 VECTORS
2.10 Line integrals

y
Q = (1, 1)
2
2
P = (0, 0)
Figure 12:
dl =
2 dx. Thus,
ZQ
x y dl =
Z1
The integration along route 2 gives

ZQ
Z1

x y dl =
x y dx
0
= 0+
Z1
0
2
2 dx =
.
3

y=0
(2.61)
Z1
1
y dy = .
2

x y dy

0
x=1
(2.62)
Note that the integral depends on the route taken between the initial and final
points.
The most common type of line integral is that where the contributions from
dx and dy are evaluated separately, rather that through the path length dl:
ZQ
[f(x, y) dx + g(x, y) dy] .
(2.63)
P
As an example of this, consider the integral

ZQ
h
i
y3 dx + x dy
(2.64)
along the two routes indicated in Fig. 13. Along route 1 we have x = y + 1 and
28
2 VECTORS
2.10 Line integrals

y
Q = (2, 1)
2
2
P = (1, 0)
Figure 13:
dx = dy, so
ZQ
Along route 2,
ZQ
Z1
0
i
7
y3 dy + (y + 1) dy = .
4
Z2

3
y dx

1
y=0
Z1

x dy

0
x=2
= 2.
(2.65)
(2.66)
Again, the integral depends on the path of integration.

Suppose that we have a line integral which does not depend on the path of
integration. It follows that
ZQ
(f dx + g dy) = F(Q) F(P)
(2.67)
P
for some function F. Given F(P) for one point P in the x-y plane, then
ZQ
(f dx + g dy)
F(Q) = F(P) +
(2.68)
defines F(Q) for all other points in the plane. We can then draw a contour map of
F(x, y). The line integral between points P and Q is simply the change in height
in the contour map between these two points:
ZQ
ZQ
(f dx + g dy) =
dF(x, y) = F(Q) F(P).
(2.69)
P
29
2 VECTORS
2.11 Vector line integrals
Thus,
dF(x, y) = f(x, y) dx + g(x, y) dy.
For instance, if F = x y3 then dF = y3 dx + 3 x y2 dy and
ZQ

h
iQ
y3 dx + 3 x y2 dy = x y3 P
(2.70)
(2.71)
is independent of the path of integration.

It is clear that there are two distinct types of line integral. Those which depend
only on their endpoints and not on the path of integration, and those which
depend both on their endpoints and the integration path. Later on, we shall
learn how to distinguish between these two types.
2.11 Vector line integrals

A vector field is defined as a set of vectors associated with each point in space.
For instance, the velocity v(r) in a moving liquid (e.g., a whirlpool) constitutes a
vector field. By analogy, a scalar field is a set of scalars associated with each point
in space. An example of a scalar field is the temperature distribution T (r) in a
furnace.
Consider a general vector field A(r). Let dl = (dx, dy, dz) be the vector element of line length. Vector line integrals often arise as
ZQ
ZQ
A dl = (Ax dx + Ay dy + Az dz).
(2.72)
P
For instance, if A is a force then the line integral is the work done in going from
P to Q.
As an example, consider the work done in a repulsive, inverse-square, central
field, F = r/|r3 |. The element of work done is dW = F dl. Take P = (, 0, 0)
and Q = (a, 0, 0). Route 1 is along the x-axis, so
Za
!
" #a
1
1
1
= .
(2.73)
2 dx =
W=
x
x a
30
2 VECTORS
2.12 Surface integrals
The second route is, firstly, around a large circle (r = constant) to the point (a,
, 0), and then parallel to the y-axis. In the first, part no work is done, since F
is perpendicular to dl. In the second part,
0
Z0
y dy
1
1
=
=
.
(2.74)
W=
2 + y2 )3/2
2 + a2 )1/2
(a
(y
a
In this case, the integral is independent of the path. However, not all vector line
integrals are path independent.

Let us take a surface S, which is not necessarily co-planar, and divide in up into
(scalar) elements Si . Then
ZZ
X
f(x, y, z) dS = lim
f(x, y, z) Si
(2.75)
Si 0
is a surface integral. For instance, the volume of water in a lake of depth D(x, y)
is
ZZ
V=
D(x, y) dS.
(2.76)
To evaluate this integral we must split the calculation into two ordinary integrals.
The volume in the strip shown in Fig. 14 is
Z
x2
D(x, y) dx dy.
(2.77)
x1
Note that the limits x1 and x2 depend on y. The total volume is the sum over all
strips:
Z
ZZ
Z y2
x2 (y)
D(x, y) dx dy.
(2.78)
D(x, y) dx
dy
V=
y1
x1 (y)
Of course, the integral can be evaluated by taking the strips the other way around:
Z y2 (x)
Z x2
dx
D(x, y) dy.
(2.79)
V=
x1
y1 (x)
31
2 VECTORS

y
y
dy
y
x1
x2
Figure 14:
Interchanging the order of integration is a very powerful and useful trick. But
great care must be taken when evaluating the limits.
As an example, consider
ZZ
x2 y dx dy,
(2.80)
where S is shown in Fig. 15. Suppose that we evaluate the x integral first:
Z
1y
3 1y
y
x
dy
x2 y dx = y dy
= (1 y)3 dy.
(2.81)
3 0
3
0
Let us now evaluate the y integral:
Z1
4
y
1
y
dy =
y2 + y3
.
3
3
60
0
(2.82)
We can also evaluate the integral by interchanging the order of integration:

Z1
Z 1x
Z1 2
1
x
(1 x)2 dx = .
(2.83)
x2 dx
y dy =
60
0
0
0 2
In some cases, a surface integral is just the product of two separate integrals.
For instance,
ZZ
x2 y dx dy
(2.84)
S
32
2 VECTORS
2.13 Vector surface integrals

y
(0, 1)
1y=x
(0, 0)
(1, 0)
Figure 15:
where S is a unit square. This integral can be written

Z
Z
Z1 Z1
1
1
11 1
2
2
dx x y dy = x dx y dy =
= ,
32 6
0
0
0
0
(2.85)
since the limits are both independent of the other variable.

In general, when interchanging the order of integration, the most important
part of the whole problem is getting the limits of integration right. The only
foolproof way of doing this is to draw a diagram.
2.13 Vector surface integrals

Surface integrals often occur during vector analysis. For instance, the rate of flow
of a liquid of velocity v through an infinitesimal surface of vector area dS is v dS.
The net rate of flow through a surface S made up of lots of infinitesimal surfaces
is
ZZ
X

v dS = lim
v cos dS ,
(2.86)
S
dS0
where is the angle subtended between the normal to the surface and the flow
velocity.
Analogously to line integrals, most surface integrals depend both on the surface and the rim. But some (very important) integrals depend only on the rim,
33
2 VECTORS
2.14 Volume integrals
and not on the nature of the surface which spans it. As an example of this, consider incompressible fluid flow between two surfaces S1 and S2 which end on the
same rim. The volume between the surfaces is constant, so what goes in must
come out, and
ZZ
ZZ
v dS.
(2.87)
v dS =
S2
S1
It follows that
ZZ
v dS
(2.88)
depends only on the rim, and not on the form of surfaces S1 and S2 .
2.14 Volume integrals
A volume integral takes the form
ZZZ
(2.89)
f(x, y, z) dV,
V
where V is some volume, and dV = dx dy dz is a small volume element. The

volume element is sometimes written d3 r, or even d. As an example of a volume
integral, let us evaluate the centre of gravity of a solid hemisphere of radius a
(centered on the origin). The height of the centre of gravity is given by
ZZZ
, ZZZ
z=
z dV
dV.
(2.90)
The bottom integral is simply the volume of the hemisphere, which is 2 a 3 /3.
The top integral is most easily evaluated in spherical polar coordinates, for which
z = r cos and dV = r2 sin dr d d. Thus,
ZZZ
Z a Z /2
Z 2
z dV =
dr
d
d r cos r2 sin
0
giving
Za
r dr
Z /2
sin cos d
Z 2
0
a4 3
3a
z=
=
.
3
4 2 a
8
34
a4
d =
,
4
(2.91)
(2.92)
2 VECTORS
2.15 Gradient
2.15 Gradient
A one-dimensional function f(x) has a gradient df/dx which is defined as the
slope of the tangent to the curve at x. We wish to extend this idea to cover scalar
fields in two and three dimensions.
y
contours of h(x, y)
Figure 16:
Consider a two-dimensional scalar field h(x, y), which is (say) the height of
a hill. Let dl = (dx, dy) be an element of horizontal distance. Consider dh/dl,
where dh is the change in height after moving an infinitesimal distance dl. This
quantity is somewhat like the one-dimensional gradient, except that dh depends
on the direction of dl, as well as its magnitude. In the immediate vicinity of some
point P, the slope reduces to an inclined plane (see Fig. 16). The largest value of
dh/dl is straight up the slope. For any other direction
dh
dh
=
dl
dl
cos .
(2.93)
max
Let us define a two-dimensional vector, grad h, called the gradient of h, whose

magnitude is (dh/dl)max , and whose direction is the direction up the steepest
slope. Because of the cos property, the component of grad h in any direction
equals dh/dl for that direction. [The argument, here, is analogous to that used
for vector areas in Sect. 2.3. See, in particular, Eq. (2.13). ]
The component of dh/dl in the x-direction can be obtained by plotting out the
profile of h at constant y, and then finding the slope of the tangent to the curve
at given x. This quantity is known as the partial derivative of h with respect to x
35
2 VECTORS
2.15 Gradient
at constant y, and is denoted (h/x)y . Likewise, the gradient of the profile at

constant x is written (h/y)x . Note that the subscripts denoting constant-x and
constant-y are usually omitted, unless there is any ambiguity. If follows that in
component form
!
h h
,
.
(2.94)
grad h =
x y
Now, the equation of the tangent plane at P = (x0 , y0 ) is
hT (x, y) = h(x0 , y0 ) + (x x0 ) + (y y0 ).
(2.95)
This has the same local gradients as h(x, y), so

=
h
,
x
h
,
y
(2.96)
by differentiation of the above. For small dx = x x0 and dy = y y0 , the

function h is coincident with the tangent plane. We have
dh =
h
h
dx +
dy,
x
y
(2.97)
but grad h = (h/x, h/y) and dl = (dx, dy), so

dh = grad h dl.
(2.98)
Incidentally, the above equation demonstrates that grad h is a proper vector, since
the left-hand side is a scalar, and, according to the properties of the dot product, the right-hand side is also a scalar, provided that dl and grad h are both
proper vectors (dl is an obvious vector, because it is directly derived from displacements).
Consider, now, a three-dimensional temperature distribution T (x, y, z) in (say)
a reaction vessel. Let us define grad T , as before, as a vector whose magnitude is
(dT/dl)max , and whose direction is the direction of the maximum gradient. This
vector is written in component form
T T T
grad T =
,
,
.
x y z
!
36
(2.99)
2 VECTORS
2.15 Gradient
Here, T/x (T/x)y,z is the gradient of the one-dimensional temperature profile at constant y and z. The change in T in going from point P to a neighbouring
point offset by dl = (dx, dy, dz) is
dT =
T
T
T
dx +
dy +
dz.
x
y
z
(2.100)
dT = grad T dl.
(2.101)
In vector form, this becomes
Suppose that dT = 0 for some dl. It follows that

dT = grad T dl = 0.
(2.102)
So, dl is perpendicular to grad T . Since dT = 0 along so-called isotherms

(i.e., contours of the temperature), we conclude that the isotherms (contours)
are everywhere perpendicular to grad T (see Fig. 17).
T = constant
dl
grad T
isotherms
Figure 17:
It is, of course, possible to integrate dT . The line integral from point P to point
Q is written
Z
Z
Q
dT =
grad T dl = T (Q) T (P).
(2.103)
This integral is clearly independent of the path taken between P and Q, so

R
Q
P grad T dl must be path independent.
37
2 VECTORS
2.15 Gradient
RQ
In general, P A dl depends on path, but for some special vector fields the
integral is path independent. Such fields are called conservative fields. It can be
shown that if A is a conservative field then A = grad for some scalar field .
The proof of this is straightforward. Keeping P fixed we have
ZQ
A dl = V(Q),
(2.104)
P
where V(Q) is a well-defined function, due to the path independent nature of the
line integral. Consider moving the position of the end point by an infinitesimal
amount dx in the x-direction. We have
Z Q+dx
V(Q + dx) = V(Q) +
A dl = V(Q) + Ax dx.
(2.105)
Q
Hence,
V
= Ax ,
x
with analogous relations for the other components of A. It follows that
(2.106)
(2.107)
A = grad V.
In physics, the force

R due to gravity is a good example of a conservative field.
If A is a force, then A dl is the work done in traversing some path. If A is
conservative then
I
A dl = 0,
(2.108)
H
where corresponds to the line integral around some closed loop. The fact that
zero net work is done in going around a closed loop is equivalent to the conservation of energy (this is why conservative fields are called conservative). A
good example of a non-conservative field is the force due to friction.
Clearly, a
H
frictional system loses energy in going around a closed cycle, so A dl 6= 0.
It is useful to define the vector operator

,
,
,
x y z
!
38
(2.109)
2 VECTORS
2.15 Gradient
which is usually called the grad or del operator. This operator acts on everything
to its right in a expression, until the end of the expression or a closing bracket is
reached. For instance,
f f f
,
,
grad f = f =
.
x y z
!
(2.110)
For two scalar fields and ,

(2.111)
grad ( ) = grad + grad

can be written more succinctly as
( ) = + .
(2.112)
Suppose that we rotate the basis about the z-axis by degrees. By analogy
with Eqs. (2.7)(2.9), the old coordinates (x, y, z) are related to the new ones
(x 0 , y 0 , z 0 ) via
Now,
=
x 0
x 0
y 0 ,z 0
x = x 0 cos y 0 sin ,
(2.113)
y = x 0 sin + y 0 cos ,
(2.114)
z = z 0.
(2.115)
+
x
x 0
y 0 ,z 0
+
y
x 0
y 0 ,z 0
,
z
(2.116)
giving
=
cos
+
sin
,
x 0
x
y
(2.117)
x 0 = cos x + sin y .
(2.118)
and
It can be seen that the differential operator transforms like a proper vector,
according to Eqs. (2.10)(2.12). This is another proof that f is a good vector.
39
2 VECTORS
2.16 Divergence
2.16 Divergence
H
Let us start with a vector field A. Consider S A dS over some closed surface S,
where dS denotes an outward pointing surface element. This surface integral
is
H
usually called the flux of A out of S. If A is the velocity of some fluid, then S AdS
is the rate of flow of material out of S.
If A is constant in space then it is easily demonstrated that the net flux out of
S is zero,
I
I
A dS = A dS = A S = 0,
(2.119)
since the vector area S of a closed surface is zero.
y + dy
z + dz
z
z
x + dx
y
x
Figure 18:
Suppose, now, that A is not uniform in space. Consider aHvery small rectangular volume over which A hardly varies. The contribution to A dS from the two
faces normal to the x-axis is
Ax (x + dx) dy dz Ax (x) dy dz =
Ax
Ax
dx dy dz =
dV,
x
x
(2.120)
where dV = dx dy dz is the volume element (see Fig. 18). There are analogous
contributions from the sides normal to the y- and z-axes, so the total of all the
contributions is
I
!
Ax Ay Az
A dS =
+
+
dV.
(2.121)
x
y
z
40
2 VECTORS
2.16 Divergence
The divergence of a vector field is defined

div A = A =
Ax Ay Az
+
+
.
x
y
z
(2.122)
Divergence is a good scalar (i.e., it is coordinate independent), since it is the dot

product of the vector operator with A. The formal definition of div A is
H
A dS
.
(2.123)
div A = lim
dV0
dV
This definition is independent of the shape of the infinitesimal volume element.
S
interior contributions cancel
.
Figure 19:
One of the most important results in vector field theory is the so-called divergence theorem or Gauss theorem. This states that for any volume V surrounded
by a closed surface S,
I
Z
S
A dS =
div A dV,
(2.124)
where dS is an outward pointing volume element. The proof is very straightforR

ward. We divide up the volume into lots of very small cubes, and sum A dS
over all of the surfaces. The contributions from the interior surfaces cancel out,
leaving just the contribution from the outer surface (see Fig. 19). We can use
Eq. (2.121)
R for each cube individually. This tells us that the summation is equivalent to div A dV over the whole volume. Thus, the integral of A dS over
41
2 VECTORS
2.16 Divergence
the outer surface is equal to the integral of div A over the whole volume, which
proves the divergence theorem.
Now, for a vector field with div A = 0,
I
A dS = 0
(2.125)
for any closed surface S. So, for two surfaces on the same rim (see Fig. 20),
Z
Z
A dS.
(2.126)
A dS =
S2
S1
Thus, if div A = 0 then the surface integral depends on the rim but not the nature
of the surface which spans it. On the other hand, if div A 6= 0 then the integral
depends on both the rim and the surface.
S2
Rim
S1
Figure 20:
H Consider an incompressible fluid whose velocity field is v. It is clear that

v dS = 0 for any closed surface, since what flows intoR the surface must flow
out again. Thus, according to the divergence theorem, div v dV = 0 for any
volume. The only way in which this is possible is if div v is everywhere zero.
Thus, the velocity components of an incompressible fluid satisfy the following
differential relation:
vx vy vz
+
+
= 0.
(2.127)
x
y
z
Consider,
H now, a compressible fluid of density and velocity v. The surface
integral S v dS is the net rate of mass flow out of the closed surface S. This
42
2 VECTORS
2.16 Divergence
must be equal to the rate of

R decrease of mass inside the volume V enclosed by S,
which is written (/t)( V dV). Thus,
I
Z
!
v dS =
dV
(2.128)
t V
S
for any volume. It follows from the divergence theorem that
div ( v) =
.
t
(2.129)
This is called the equation of continuity of the fluid, since it ensures that fluid
is neither created nor destroyed as it flows from place to place. If is constant
then the equation of continuity reduces to the previous incompressible result,
div v = 0.
2
Figure 21:
It is sometimes helpful to represent a vector field A by lines of force or fieldlines. The direction of a line of force at any point is the same as the direction of A.
The density of lines (i.e., the number of lines crossing a unit surface perpendicular
to A) is equal to |A|. For instance, in Fig. 21, |A| is larger at point 1 than at point
2. The number of lines crossing a surface element dS is AdS. So, the net number
of lines leaving a closed surface is
I
Z
A dS = div A dV.
(2.130)
S
If div A = 0 then there is no net flux of lines out of any surface. Such a field is
called a solenoidal vector field. The simplest example of a solenoidal vector field
is one in which the lines of force all form closed loops.
43
2 VECTORS
2.17 The Laplacian
2.17 The Laplacian

So far we have encountered

grad =
,
,
,
x y z
!
(2.131)
which is a vector field formed from a scalar field, and

div A =
Ax Ay Az
+
+
,
x
y
z
(2.132)
which is a scalar field formed from a vector field. There are two ways in which
we can combine grad and div. We can either form the vector field grad (div A)
or the scalar field div (grad ). The former is not particularly interesting, but
the scalar field div (grad ) turns up in a great many physics problems, and is,
therefore, worthy of discussion.
Let us introduce the heat flow vector h, which is the rate of flow of heat energy per unit area across a surface perpendicular to the direction of h. In many
substances, heat flows directly down the temperature gradient, so that we can
write
h = grad T,
(2.133)
H
where is the thermal conductivity. The net rate of heat flow S h dS out of
some closed surface S must be equal to the rate of decrease of heat energy in the
volume V enclosed by S. Thus, we can write
I
Z
!
h dS =
c T dV ,
(2.134)
t
S
where c is the specific heat. It follows from the divergence theorem that
div h = c
T
.
t
(2.135)
Taking the divergence of both sides of Eq. (2.133), and making use of Eq. (2.135),
we obtain
T
div ( grad T ) = c
,
(2.136)
t
44
2 VECTORS
2.17 The Laplacian
or
T
.
t
If is constant then the above equation can be written
( T ) = c
div (grad T ) =
(2.137)
c T
.
t
(2.138)
The scalar field div (grad T ) takes the form

T
T
T
+
+
div (grad T ) =
x x
y y
z z
!
2 T 2 T 2 T
+ 2 + 2 2 T.
2
x
y
z
(2.139)
Here, the scalar differential operator

2
2
2
2+ 2+ 2
x
y
z
2
(2.140)
is called the Laplacian. The Laplacian is a good scalar operator (i.e., it is coordinate independent) because it is formed from a combination of div (another good
scalar operator) and grad (a good vector operator).
What is the physical significance of the Laplacian? In one dimension, 2 T
reduces to 2 T/x2 . Now, 2 T/x2 is positive if T (x) is concave (from above) and
negative if it is convex. So, if T is less than the average of T in its surroundings
then 2 T is positive, and vice versa.
In two dimensions,
2 T 2 T
T = 2 + 2.
(2.141)
x
y
Consider a local minimum of the temperature. At the minimum, the slope of T
increases in all directions, so 2 T is positive. Likewise, 2 T is negative at a local
maximum. Consider, now, a steep-sided valley in T . Suppose that the bottom of
the valley runs parallel to the x-axis. At the bottom of the valley 2 T/y2 is large
and positive, whereas 2 T/x2 is small and may even be negative. Thus, 2 T is
positive, and this is associated with T being less than the average local value.
2
45
2 VECTORS
2.18 Curl
Let us now return to the heat conduction problem:

2 T =
c T
.
t
(2.142)
It is clear that if 2 T is positive then T is locally less than the average value, so
T/t > 0: i.e., the region heats up. Likewise, if 2 T is negative then T is locally
greater than the average value, and heat flows out of the region: i.e., T/t < 0.
Thus, the above heat conduction equation makes physical sense.
2.18 Curl
Consider a vector field A, and
H a loop which lies in one plane. The integral of A
around this loop is written A dl, whereHdl is a line element of the loop. If A is
a conservative field then A
H = grad and A dl = 0 for all loops. In general, for
a non-conservative field, A dl 6= 0.
H
For a small loop we expect A dl to be
H proportional to the area of the loop.
Moreover, for a fixed area loop we expect A dl to depend on theHorientation of
the loop. One particular orientation will give the maximum value: A dl = Imax .
If the loop subtends an angle with this optimum orientation then we expect
I = Imax cos . Let us introduce the vector field curl A whose magnitude is
H
A dl
|curl A| = lim
(2.143)
dS0
dS
for the orientation giving Imax . Here, dS is the area of the loop. The direction
of curl A is perpendicular to the plane of the loop, when it is in the orientation
giving Imax , with the sense given by the right-hand grip rule.
Let us Hnow express curl A in terms of the components of A. First, we shall
evaluate A dl around a small rectangle in the y-z plane (see Fig. 22). The
contribution from sides 1 and 3 is
Az (y + dy) dz Az (y) dz =
46
Az
dy dz.
y
(2.144)
2 VECTORS
2.18 Curl
z
4
z + dz
1
z
3
y
y + dy
y
Figure 22:
The contribution from sides 2 and 4 is

Ay (z + dz) dy + Ay (z) dy =
Ay
dy dz.
y
So, the total of all contributions gives

I
!
Az Ay
A dl =
dS,
y
z
(2.145)
(2.146)
where dS = dy dz is the area of the loop.
Consider a non-rectangular (but still small)

H loop in the y-z plane. We can
divide it into rectangular elements, and form A dl over all the resultant loops.
The interior contributions cancel, so we are just left with the contribution from
the outer loop. Also, the area of the outer loop is the sum of all the areas of the
inner loops. We conclude that
I
!
Az Ay
A dl =
dSx
(2.147)
y
z
is valid for a small loop dS = (dSx , 0, 0) of any shape in the y-z plane. Likewise,
we can show that if the loop is in the x-z plane then dS = (0, dSy , 0) and
I
!
Ax Az
A dl =
dSy .
(2.148)
z
x
Finally, if the loop is in the x-y plane then dS = (0, 0, dSz ) and
I
!
Ay Ax
A dl =
dSz .
x
y
47
(2.149)
2 VECTORS
2.18 Curl
z
dS
2
3
x
Figure 23:
Imagine an arbitrary loop of vector area dS = (dSx , dSy , dSz ). We can construct this out of three loops in the x-, y-, and z-directions, as indicated in Fig. 23.
If we form the line integral around all three loops then the interior contributions
cancel, and we are left with the line integral around the original loop. Thus,
I
I
I
I
A dl = A dl1 + A dl2 + A dl3 ,
(2.150)
giving
where
A dl = curl A dS = |curl A| |dS| cos ,
(2.151)
Az Ay Ax Az Ay Ax
curl A =
.
y
z
z
x x
y
(2.152)
curl A = A.
(2.153)
Note that
This demonstrates that curl A is a good vector field, since it is the cross product
of the operator (a good vector operator) and the vector field A.
Consider a solid body rotating about the z-axis. The angular velocity is given
by = (0, 0, ), so the rotation velocity at position r is
v=r
48
(2.154)
2 VECTORS
2.18 Curl
[see Eq. (2.43) ]. Let us evaluateH curl v on the axis of rotation. The x-component
is proportional to the integral v dl around a loop in the y-z plane. This
H is
plainly zero. Likewise, the y-component is also zero. The z-component is v
dl/dS around some loop in the x-y plane. Consider a circular loop. We have
H
v dl = 2 r r with dS = r2 . Here, r is the radial distance from the rotation
axis. It follows that (curl v)z = 2 , which is independent of r. So, on the axis,
curl v = (0 , 0 , 2 ). Off the axis, at position r0 , we can write
v = (r r0 ) + r0 .
(2.155)
The first part has the same curl as the velocity field on the axis, and the second
part has zero curl, since it is constant. Thus, curl v = (0, 0, 2 ) everywhere in
the body. This allows us to form a physical picture of curl A. If we imagine A as
the velocity field of some fluid, then curl A at any given point is equal to twice
the local angular rotation velocity: i.e., 2 . Hence, a vector field with curl A = 0
everywhere is said to be irrotational.
Another important result of vector field theory is the curl theorem or Stokes
theorem,
I
Z
A dl = curl A dS,
(2.156)
C
for some (non-planar) surface S bounded by a rim C. This theorem can easily be
proved by splitting the loop up into many small rectangular loops, and forming
the integral around all of the resultant loops. All of the contributions from the
interior loops cancel, leaving just the contribution from the outer rim. Making
use of Eq. (2.151) for each of the small loops, we can see that the contribution
from all of the loops is also equal to the integral of curl A dS across the whole
surface. This proves the theorem.
One immediate consequence of Stokes theorem is that curl A is incompressible. Consider two surfaces,
R S1 and S2 , which share the same rim. It is clear
from StokesHtheorem that curl A dS is the same for both surfaces. Thus, it
follows that curl A dS =
H 0 for any closed
R surface. However, we have from the
divergence theorem that curl A dS = div (curl A) dV = 0 for any volume.
Hence,
div (curl A) 0.
(2.157)
49
2 VECTORS
2.19 Summary
So, curl A is a solenoidal field.

H
We have seen that for a conservative field A dl = 0 for any loop. This isHentirely equivalent to A = grad . However, the magnitude of curl A is lim dS0 A
dl/dS for some particular loop. It is clear then that curl A = 0 for a conservative
field. In other words,
curl (grad ) 0.
(2.158)
Thus, a conservative field is also an irrotational one.
Finally, it can be shown that
curl (curl A) = grad (div A) 2 A,
(2.159)
2 A = (2 Ax , 2 Ay , 2 Az ).
(2.160)
where
It should be emphasized, however, that the above result is only valid in Cartesian
coordinates.
2.19 Summary
Vector addition:
a + b (ax + bx , ay + by , az + bz )
Scalar multiplication:
n a (n ax , n ay , n az )
Scalar product:
a b = a x bx + a y by + a z bz
Vector product:
a b = (ay bz az by , az bx ax bz , ax by ay bx )
Scalar triple product:
a b c = a b c = b c a = b a c
50
2 VECTORS
2.19 Summary
Vector triple product:

a (b c) = (a c) b (a b) c
(a b) c = (a c) b (b c) a
Gradient:

,
,
grad =
x y z
Divergence:
div A =
Curl:
Gauss theorem:
Stokes theorem:
Ax Ay Az
+
+
x
y
z
Az Ay Ax Az Ay Ax
curl A =
y
z
z
x x
y
I
I
A dS =
A dl =
div A dV
V
curl A dS
Del operator:

=
,
,
x y z
grad =
div A = A
curl A = A
Vector identities:
2 2 2
+ 2
= 2 = 2 +
x
y2
z
A=0
= 0
2 A = ( A) A
51
2 VECTORS
2.19 Summary
Other vector identities:

( ) = +
( A) = A + A
( A) = A + A
(A B) = B A A B
(A B) = A ( B) B ( A) + (B )A (A )B
(A B) = A ( B) + B ( A) + (A )B + (B )A
Cylindrical polar coordinates:
x = r cos ,
y = r sin , z = z, dV = r dr d dz
!
f 1 f f
,
,
f =
r r z
1 (rAr ) 1 A Az
A=
+
+
r
r
r
A
A
A
1
1
1
A
(rA
)
A
r
z
z
A=
r
z z
r r r
r
!
1
f
1 2 f 2 f
2 f =
r
+ 2 2+ 2
r r r
r
z
Spherical polar coordinates:
x = r sin cos , y = r sin sin , z = r cos , dV = r2 sin dr d d
!
1 f
f 1 f
,
,
f =
r r r sin
1
1 A
1
(sin A ) +
A = 2 (r2 Ar ) +
r r
r sin
r sin
1 (sin A )
1 A
( A)r =
r sin
r sin
1 Ar 1 (rA )
( A) =
r sin
r r
1 (rA ) 1 Ar
( A)z =
r r
r
!
!
2 f
1 2 f
1
f
1
2
f= 2
r
+ 2
sin
+ 2
r r
r
r sin
r sin2 2
52
3 TIME-INDEPENDENT MAXWELL EQUATIONS
Time-independent Maxwell equations
3.1 Introduction
In this section, we shall take the familiar force laws of electrostatics and magnetostatics, and recast them as vector field equations.
3.2 Coulombs law

Between 1785 and 1787, the French physicist Charles Augustine de Coulomb
performed a series of experiments involving electric charges, and eventually established what is nowadays known as Coulombs law. According to this law, the
force acting between two electric charges is radial, inverse-square, and proportional to the product of the charges. Two like charges repel one another, whereas
two unlike charges attract. Suppose that two charges, q1 and q2 , are located at
position vectors r1 and r2 . The electrical force acting on the second charge is
written
q1 q2 r 2 r 1
(3.1)
f2 =
4 0 |r2 r1 |3
in vector notation (see Fig. 24). An equal and opposite force acts on the first
charge, in accordance with Newtons third law of motion. The SI unit of electric
charge is the coulomb (C). The magnitude of the charge on an electron is 1.6022
1019 C. The universal constant 0 is called the permittivity of free space, and takes
the value
0 = 8.8542 1012 C 2 N1 m2 .
(3.2)
Coulombs law has the same mathematical form as Newtons law of gravity.
Suppose that two masses, m1 and m2 , are located at position vectors r1 and r2 .
The gravitational force acting on the second mass is written
f2 = G m1 m2
53
r2 r 1
|r2 r1 |3
(3.3)

f
q1
3.2 Coulombs law
r2 r1
q2
f
r2
r1
Figure 24:
in vector notation. The gravitational constant G takes the value

G = 6.6726 1011 N m2 kg2 .
(3.4)
Coulombs law and Newtons law are both inverse-square force laws: i.e.
|f2 |
1
.
|r2 r1 |2
(3.5)
However, they differ in two crucial respects. Firstly, the force due to gravity is
always attractive (there is no such thing as a negative mass). Secondly, the magnitudes of the two forces are vastly different. Consider the ratio of the electrical
and gravitational forces acting on two particles. This ratio is a constant, independent of the relative positions of the particles, and is given by
q1 q2
1
|felectrical |
=
.
|fgravitational | m1 m2 4 0 G
(3.6)
For electrons, the charge to mass ratio is q/m = 1.759 1011 C kg1 , so
|felectrical |
|fgravitational |
= 4.17 1042 .
(3.7)
This is a colossal number! Suppose we are studying a physics problem involving

the motion of particles in a box under the action of two forces with the same
range, but differing in magnitude by a factor 1042 . It would seem a plausible
approximation (to say the least) to start the investgation by neglecting the weaker
54
3.2 Coulombs law
force. Applying this reasoning to the motion of particles in the Universe, we

would expect the Universe to be governed entirely by electrical forces. However,
this is not the case. The force which holds us to the surface of the Earth, and
prevents us from floating off into space, is gravity. The force which causes the
Earth to orbit the Sun is also gravity. In fact, on astronomical length-scales gravity
is the dominant force, and electrical forces are largely irrelevant. The key to
understanding this paradox is that there are both positive and negative electric
charges, whereas there are only positive gravitational charges. This means that
gravitational forces are always cumulative, whereas electrical forces can cancel
one another out. Suppose, for the sake of argument, that the Universe starts out
with randomly distributed electric charges. Initially, we expect electrical forces to
completely dominate gravity. These forces try to make every positive charge get
as far away as possible from the other positive charges, and as close as possible
to the other negative charges. After a while, we expect the positive and negative
charges to form close pairs. Just how close is determined by quantum mechanics,
but, in general, it is pretty close: i.e., about 1010 m. The electrical forces due to
the charges in each pair effectively cancel one another out on length-scales much
larger than the mutual spacing of the pair. It is only possible for gravity to be the
dominant long-range force if the number of positive charges in the Universe is
almost equal to the number of negative charges. In this situation, every positive
charge can find a negative charge to team up with, and there are virtually no
charges left over. In order for the cancellation of long-range electrical forces to be
effective, the relative difference in the number of positive and negative charges in
the Universe must be incredibly small. In fact, positive and negative charges have
to cancel each other out to such accuracy that most physicists believe that the net
charge of the universe is exactly zero. But, it is not enough for the Universe to
start out with zero charge. Suppose there were some elementary particle process
which did not conserve electric charge. Even if this were to go on at a very low
rate, it would not take long before the fine balance between positive and negative
charges in the Universe was wrecked. So, it is important that electric charge is
a conserved quantity (i.e., the net charge of the Universe can neither increase or
decrease). As far as we know, this is the case. To date, no elementary particle
reactions have been discovered which create or destroy net electric charge.
55
3.2 Coulombs law
In summary, there are two long-range forces in the Universe, electromagnetism

and gravity. The former is enormously stronger than the latter, but is usually hidden away inside neutral atoms. The fine balance of forces due to negative and
positive electric charges starts to break down on atomic scales. In fact, interatomic and intermolecular forces are all electrical in nature. So, electrical forces
are basically what prevent us from falling though the floor. But, this is electromagnetism on the microscopic or atomic scalewhat is usually termed quantum
electromagnetism. This course is about classical electromagnetism. That is, electromagnetism on length-scales much larger than the atomic scale. Classical electromagnetism generally describes phenomena in which some sort of violence
is done to matter, so that the close pairing of negative and positive charges is
disrupted. This allows electrical forces to manifest themselves on macroscopic
length-scales. Of course, very little disruption is necessary before gigantic forces
are generated. It is no coincidence that the vast majority of useful machines
which humankind has devised during the last century or so are electrical in nature.
Coulombs law and Newtons law are both examples of what are usually referred to as action at a distance theories. According to Eqs. (3.1) and (3.3), if
the first charge or mass is moved then the force acting on the second charge or
mass immediately responds. In particular, equal and opposite forces act on the
two charges or masses at all times. However, this cannot be correct according to
Einsteins theory of relativity, which implies that the maximum speed with which
information can propagate through the Universe is the speed of light in vacuum.
So, if the first charge or mass is moved then there must always be time delay (i.e.,
at least the time needed for a light signal to propagate between the two charges
or masses) before the second charge or mass responds. Consider a rather extreme
example. Suppose the first charge or mass is suddenly annihilated. The second
charge or mass only finds out about this some time later. During this time interval, the second charge or mass experiences an electrical or gravitational force
which is as if the first charge or mass were still there. So, during this period, there
is an action but no reaction, which violates Newtons third law of motion. It is
clear that action at a distance is not compatible with relativity, and, consequently,
that Newtons third law of motion is not strictly true. Of course, Newtons third
56
3.2 Coulombs law
law is intimately tied up with the conservation of linear momentum in the Universe. This is a concept which most physicists are loath to abandon. It turns out
that we can rescue momentum conservation by abandoning action at a distance
theories, and instead adopting so-called field theories in which there is a medium,
called a field, which transmits the force from one particle to another. In electromagnetism there are, in fact, two fieldsthe electric field, and the magnetic
field. Electromagnetic forces are transmitted via these fields at the speed of light,
which implies that the laws of relativity are never violated. Moreover, the fields
can soak up energy and momentum. This means that even when the actions and
reactions acting on particles are not quite equal and opposite, momentum is still
conserved. We can bypass some of the problematic aspects of action at a distance
by only considering steady-state situations. For the moment, this is how we shall
proceed.
Consider N charges, q1 though qN , which are located at position vectors r1
through rN . Electrical forces obey what is known as the principle of superposition.
The electrical force acting on a test charge q at position vector r is simply the
vector sum of all of the Coulomb law forces from each of the N charges taken in
isolation. In other words, the electrical force exerted by the ith charge (say) on
the test charge is the same as if all the other charges were not there. Thus, the
force acting on the test charge is given by
N
X
qi r r i
.
(3.8)
f(r) = q
3
4

|r
r
|
0
i
i=1
It is helpful to define a vector field E(r), called the electric field, which is the force
exerted on a unit test charge located at position vector r. So, the force on a test
charge is written
f = q E,
(3.9)
and the electric field is given by
N
X
qi r r i
E(r) =
.
(3.10)
3
4

|r
r
|
0
i
i=1
At this point, we have no reason to believe that the electric field has any real
physical existence. It is just a useful device for calculating the force which acts
on test charges placed at various locations.
57
3.2 Coulombs law
The electric field from a single charge q located at the origin is purely radial,
points outwards if the charge is positive, inwards if it is negative, and has magnitude
q
Er (r) =
,
(3.11)
4 0 r2
where r = |r|.
E
Figure 25:
We can represent an electric field by field-lines. The direction of the lines

indicates the direction of the local electric field, and the density of the lines perpendicular to this direction is proportional to the magnitude of the local electric
field. Thus, the field of a point positive charge is represented by a group of equally
spaced straight lines radiating from the charge (see Fig. 25).
The electric field from a collection of charges is simply the vector sum of the
fields from each of the charges taken in isolation. In other words, electric fields
are completely superposable. Suppose that, instead of having discrete charges,
we have a continuous distribution of charge represented by a charge density (r).
Thus, the charge at position vector r 0 is (r 0 ) d3 r 0 , where d3 r 0 is the volume element at r 0 . It follows from a simple extension of Eq. (3.10) that the electric field
generated by this charge distribution is
Z
1
r r0 3 0
0
E(r) =
(r )
dr,
(3.12)
4 0
|r r 0 |3
where the volume integral is over all space, or, at least, over all space for which
(r 0 ) is non-zero.
58
3.3 The electric scalar potential

Suppose that r = (x, y, z) and r 0 = (x 0 , y 0 , z 0 ) in Cartesian coordinates. The x
component of (r r 0 )/|r r 0 |3 is written
x x0
.
[(x x 0 )2 + (y y 0 )2 + (z z 0 )2 ] 3/2
(3.13)
However, it is easily demonstrated that

x x0
=
[(x x 0 )2 + (y y 0 )2 + (z z 0 )2 ] 3/2
(3.14)
.

x [(x x 0 )2 + (y y 0 )2 + (z z 0 )2 ] 1/2
Since there is nothing special about the x-axis, we can write

1
r r0
,
|r r 0 |3
|r r 0 |
(3.15)
E = ,
(3.16)
where (/x, /y, /z) is a differential operator which involves the components of r but not those of r 0 . It follows from Eq. (3.12) that
where
Z
(r 0 ) 3 0
1
(r) =
dr.
(3.17)
4 0 |r r 0 |
Thus, the electric field generated by a collection of fixed charges can be written as
the gradient of a scalar potential, and this potential can be expressed as a simple
volume integral involving the charge distribution.
The scalar potential generated by a charge q located at the origin is
q
(r) =
.
(3.18)
4 0 r
According to Eq. (3.10), the scalar potential generated by a set of N discrete
charges qi , located at ri , is
N
X
(r) =
i (r),
(3.19)
i=1
59
where
qi
.
(3.20)
4 0 |r ri |
Thus, the scalar potential is just the sum of the potentials generated by each of
the charges taken in isolation.
i (r) =
Suppose that a particle of charge q is taken along some path from point P to
point Q. The net work done on the particle by electrical forces is
ZQ
W=
f dl,
(3.21)
P
where f is the electrical force, and dl is a line element along the path. Making
use of Eqs. (3.9) and (3.16), we obtain
ZQ
ZQ
dl = q [ (Q) (P) ] .
(3.22)
E dl = q
W=q
P
Thus, the work done on the particle is simply minus its charge times the difference in electric potential between the end point and the beginning point. This
quantity is clearly independent of the path taken between P and Q. So, an electric field generated by stationary charges is an example of a conservative field. In
fact, this result follows immediately from vector field theory once we are told, in
Eq. (3.16), that the electric field is the gradient of a scalar potential. The work
done on the particle when it is taken around a closed loop is zero, so
I
E dl = 0
(3.23)
C
for any closed loop C. This implies from Stokes theorem that
E=0
(3.24)
for any electric field generated by stationary charges. Equation (3.24) also follows directly from Eq. (3.16), since = 0 for any scalar potential .
The SI unit of electric potential is the volt, which is equivalent to a joule per
coulomb. Thus, according to Eq. (3.22), the electrical work done on a particle
when it is taken between two points is the product of its charge and the voltage
difference between the points.
60
We are familiar with the idea that a particle moving in a gravitational field possesses potential energy as well as kinetic energy. If the particle moves from point
P to a lower point Q then the gravitational field does work on the particle causing
its kinetic energy to increase. The increase in kinetic energy of the particle is balanced by an equal decrease in its potential energy, so that the overall energy of
the particle is a conserved quantity. Therefore, the work done on the particle as it
moves from P to Q is minus the difference in its gravitational potential energy between points Q and P. Of course, it only makes sense to talk about gravitational
potential energy because the gravitational field is conservative. Thus, the work
done in taking a particle between two points is path independent, and, therefore,
well-defined. This means that the difference in potential energy of the particle
between the beginning and end points is also well-defined. We have already seen
that an electric field generated by stationary charges is a conservative field. In
follows that we can define an electrical potential energy of a particle moving in
such a field. By analogy with gravitational fields, the work done in taking a particle from point P to point Q is equal to minus the difference in potential energy of
the particle between points Q and P. It follows from Eq. (3.22), that the potential energy of the particle at a general point Q, relative to some reference point P
(where the potential energy is set to zero), is given by
E(Q) = q (Q).
(3.25)
Free particles try to move down gradients of potential energy, in order to attain a
minimum potential energy state. Thus, free particles in the Earths gravitational
field tend to fall downwards. Likewise, positive charges moving in an electric
field tend to migrate towards regions with the most negative voltage, and vice
versa for negative charges.
The scalar electric potential is undefined to an additive constant. So, the transformation
(r) (r) + c
(3.26)
leaves the electric field unchanged according to Eq. (3.16). The potential can be
fixed unambiguously by specifying its value at a single point. The usual convention is to say that the potential is zero at infinity. This convention is implicit in
Eq. (3.17),
it can be seen that 0 as |r| , provided that the total
R where
0
3 0
charge (r ) d r is finite.
61
3.4 Gauss law
3.4 Gauss law

E
S
V
q
Figure 26:
Consider a single charge located at the origin. The electric field generated by
such a charge is given by Eq. (3.11). Suppose that we surround the charge by a
concentric spherical surface S of radius r (see Fig. 26). The flux of the electric
field through this surface is given by
I
I
q
q
2
,
(3.27)
4
r
=
E dS = Er dSr = Er (r) 4 r2 =
40 r2
0
S
S
since the normal to the surface is always parallel to the local electric field. However, we also know from Gauss theorem that
Z
I
E dS = E d3 r,
(3.28)
S
where V is the volume enclosed by surface S. Let us evaluate E directly. In

Cartesian coordinates, the field is written
!
q
x y z
E=
, ,
,
(3.29)
4 0 r3 r3 r3
where r2 = x2 + y2 + z2 . So,
Ex
q
1
3xx
q r2 3 x 2
=
4
.
=
x
4 0 r3
r r
4 0
r5
Here, use has been made of
(3.30)
r
x
= .
x
r
(3.31)
62
3.4 Gauss law
Formulae analogous to Eq. (3.30) can be obtained for Ey /y and Ez /z. The
divergence of the field is thus given by
q 3 r2 3 x 2 3 y 2 3 z 2
Ex Ey Ez
+
+
=
= 0.
E=
x
y
z
4 0
r5
This is a puzzling result! We have from Eqs. (3.27) and (3.28) that
Z
q
E d3 r = ,
0
V
(3.32)
(3.33)
and yet we have just proved that E = 0. This paradox can be resolved after a
close examination of Eq. (3.32). At the origin (r = 0) we find that E = 0/0,
which means that E can take any value at this point. Thus, Eqs. (3.32) and
(3.33) can be reconciled if E is some sort of spike function: i.e., it is zero
everywhere except arbitrarily close to the origin, where it becomes very large.
This must occur in such a manner that the volume integral over the spike is
finite.
1/
/2 +/2
Figure 27:
Let us examine how we might construct a one-dimensional spike function.

Consider the box-car function

1/
for |x| < /2
g(x, ) =
(3.34)
0
otherwise
(see Fig. 27). It is clear that
Z
g(x, ) dx = 1.
63
(3.35)
3.4 Gauss law
Now consider the function

(x) = lim g(x, ).
0
(3.36)
This is zero everywhere except arbitrarily close to x = 0. According to Eq. (3.35),

it also possess a finite integral;
Z
(x) dx = 1.
(3.37)
Thus, (x) has all of the required properties of a spike function. The one-dimensional
spike function (x) is called the Dirac delta-function after the Cambridge physicist Paul Dirac who invented it in 1927 while investigating quantum mechanics.
The delta-function is an example of what mathematicians call a generalized function: it is not well-defined at x = 0, but its integral is nevertheless well-defined.
Consider the integral
Z
f(x) (x) dx,
(3.38)
where f(x) is a function which is well-behaved in the vicinity of x = 0. Since the

delta-function is zero everywhere apart from very close to x = 0, it is clear that
Z
Z
f(x) (x) dx = f(0)
(x) dx = f(0),
(3.39)
where use has been made of Eq. (3.37). The above equation, which is valid for
any well-behaved function, f(x), is effectively the definition of a delta-function. A
simple change of variables allows us to define (x x0 ), which is a spike function
centred on x = x0 . Equation (3.39) gives
Z
f(x) (x x0 ) dx = f(x0 ).
(3.40)
We actually want a three-dimensional spike function: i.e., a function which

is zero everywhere apart from arbitrarily close to the origin, and whose volume
integral is unity. If we denote this function by (r) then it is easily seen that the
three-dimensional delta-function is the product of three one-dimensional deltafunctions:
(r) = (x) (y) (z).
(3.41)
64
3.4 Gauss law
This function is clearly zero everywhere except the origin. But is its volume
integral unity? Let us integrate over a cube of dimensions 2 a which is centred on
the origin, and aligned along the Cartesian axes. This volume integral is obviously
separable, so that
Za
Za
Z
Za
(z) dz.
(3.42)
(y) dy
(x) dx
(r) d3 r =
a
The integral can be turned into an integral over all space by taking
Rthe limit a
. However, we know that for one-dimensional delta-functions (x) dx = 1,
so it follows from the above equation that
Z
(r) d3 r = 1,
(3.43)
which is the desired result. A simple generalization of previous arguments yields
Z
f(r) (r) d3 r = f(0),
(3.44)
where f(r) is any well-behaved scalar field. Finally, we can change variables and
write
(r r 0 ) = (x x 0 ) (y y 0 ) (z z 0 ),
(3.45)
which is a three-dimensional spike function centred on r = r 0 . It is easily demonstrated that
Z
f(r) (r r 0 ) d3 r = f(r 0 ).
(3.46)
Up to now, we have only considered volume integrals taken over all space. However, it should be obvious that the above result also holds for integrals over any
finite volume V which contains the point r = r 0 . Likewise, the integral is zero if
V does not contain r = r 0 .
Let us now return to the problem in hand. The electric field generated by a
charge q located at the origin has E = 0 everywhere apart from the origin,
and also satisfies
Z
q
E d3 r =
(3.47)
0
V
65
3.4 Gauss law
for a spherical volume V centered on the origin. These two facts imply that
q
E=
(r),
(3.48)
0
where use has been made of Eq. (3.43).
At this stage, vector field theory has yet to show its worth.. After all, we have
just spent an inordinately long time proving something using vector field theory
which we previously proved in one line [see Eq. (3.27) ] using conventional analysis. It is time to demonstrate the power of vector field theory. Consider, again,
a charge q at the origin surrounded by a spherical surface S which is centered on
the origin. Suppose that we now displace the surface S, so that it is no longer
centered on the origin. What is the flux of the electric field out of S? This is not
a simple problem for conventional analysis, because the normal to the surface is
no longer parallel to the local electric field. However, using vector field theory
this problem is no more difficult than the previous one. We have
I
Z
E dS = E d3 r
(3.49)
S
from Gauss theorem, plus Eq. (3.48). From these equations, it is clear that the
flux of E out of S is q/0 for a spherical surface displaced from the origin. However, the flux becomes zero when the displacement is sufficiently large that the
origin is no longer enclosed by the sphere. It is possible to prove this via conventional analysis, but it is certainly not easy. Suppose that the surface S is not
spherical but is instead highly distorted. What now is the flux of E out of S? This
is a virtually impossible problem in conventional analysis, but it is still easy using
vector field theory. Gauss theorem and Eq. (3.48) tell us that the flux is q/0
provided that the surface contains the origin, and that the flux is zero otherwise.
This result is completely independent of the shape of S.
Consider N charges qi located at ri . A simple generalization of Eq. (3.48) gives

N
X
qi
E=
(r ri ).
(3.50)

0
i=1
Thus, Gauss theorem (3.49) implies that

I
Z
Q
E dS = E d3 r = ,
0
S
V
66
(3.51)
3.4 Gauss law
where Q is the total charge enclosed by the surface S. This result is called Gauss
law, and does not depend on the shape of the surface.
Suppose, finally, that instead of having a set of discrete charges, we have a
continuous charge distribution described by a charge density (r). The charge
contained in a small rectangular volume of dimensions dx, dy, and dz located at
position r is Q = (r) dx dy dz. However, if we integrate E over this volume
element we obtain
Q
dx dy dz
E dx dy dz =
=
,
(3.52)
0
0
where use has been made of Eq. (3.51). Here, the volume element is assumed
to be sufficiently small that E does not vary significantly across it. Thus, we
obtain
E= .
(3.53)
0
This is the first of four field equations, called Maxwells equations, which together form a complete description of electromagnetism. Of course, our derivation of Eq. (3.53) is only valid for electric fields generated by stationary charge
distributions. In principle, additional terms might be required to describe fields
generated by moving charge distributions. However, it turns out that this is not
the case, and that Eq. (3.53) is universally valid.
Equation (3.53) is a differential equation describing the electric field generated by a set of charges. We already know the solution to this equation when the
charges are stationary: it is given by Eq. (3.12),
Z
1
r r0 3 0
0
E(r) =
(r )
dr.
(3.54)
4 0
|r r 0 |3
Equations (3.53) and (3.54) can be reconciled provided
1
r r0
2
= 4 (r r 0 ),
0
3
0
|r r |
|r r |
where use has been made of Eq. (3.15). It follows that
Z
0
r
r
1
d3 r 0
E(r) =
(r 0 )
0
3
4 0
|r r |
67
(3.55)
3.5 Poissons equation
(r 0 )
(r)
=
(r r 0 ) d3 r 0 =
,
(3.56)
0
0
which is the desired result. The most general form of Gauss law, Eq. (3.51), is
obtained by integrating Eq. (3.53) over a volume V surrounded by a surface S,
and making use of Gauss theorem:
I
Z
1
E dS =
(r) d3 r.
(3.57)
0 V
S
One particularly interesting application of Gauss law is Earnshaws theorem,
which states that it is impossible for a collection of charged particles to remain in
static equilibrium solely under the influence of electrostatic forces. For instance,
consider the motion of the ith particle in the electric field, E, generated by all of
the other static particles. The equilibrium position of the ith particle corresponds
to some point ri , where E(ri ) = 0. By implication, ri does not correspond to the
equilibrium position of any other particle. However, in order for r i to be a stable
equilibrium point, the particle must experience a restoring force when it is moved
a small distance away from ri in any direction. Assuming that the ith particle is
positively charged, this means that the electric field must point radially towards
ri at all neighbouring points. Hence, if we apply Gauss law to a small sphere
centred on ri , then there must be a negative flux of E through the surface of the
sphere, implying the presence of a negative charge at ri . However, there is no
such charge at ri . Hence, we conclude that E cannot point radially towards ri at
all neighbouring points. In other words, there must be some neighbouring points
at which E is directed away from ri . Hence, a positively charged particle placed
at ri can always escape by moving to such points. One corollary of Earnshaws
theorem is that classical electrostatics cannot account for the stability of atoms
and molecules.

We have seen that the electric field generated by a set of stationary charges can
be written as the gradient of a scalar potential, so that
E = .
68
(3.58)
This equation can be combined with the field equation (3.53) to give a partial
differential equation for the scalar potential:
2 = .
(3.59)
0
This is an example of a very famous type of partial differential equation known
as Poissons equation.
In its most general form, Poissons equation is written
2 u = v,
(3.60)
where u(r) is some scalar potential which is to be determined, and v(r) is a known
source function. The most common boundary condition applied to this equation
is that the potential u is zero at infinity. The solutions to Poissons equation are
completely superposable. Thus, if u1 is the potential generated by the source
function v1 , and u2 is the potential generated by the source function v2 , so that
2 u 1 = v 1 ,
2 u2 = v 2 ,
(3.61)
then the potential generated by v1 + v2 is u1 + u2 , since

2 (u1 + u2 ) = 2 u1 + 2 u2 = v1 + v2 .
(3.62)
Poissons equation has this property because it is linear in both the potential and
the source term.
The fact that the solutions to Poissons equation are superposable suggests a
general method for solving this equation. Suppose that we could construct all of
the solutions generated by point sources. Of course, these solutions must satisfy
the appropriate boundary conditions. Any general source function can be built up
out of a set of suitably weighted point sources, so the general solution of Poissons
equation must be expressible as a weighted sum over the point source solutions.
Thus, once we know all of the point source solutions we can construct any other
solution. In mathematical terminology, we require the solution to
2 G(r, r 0 ) = (r r 0 )
(3.63)
which goes to zero as |r| . The function G(r, r 0 ) is the solution generated by a
unit point source located at position r 0 . This function is known to mathematicians
69
3.6 Ampères experiments
as a Greens function. The solution generated by a general source function v(r) is

simply the appropriately weighted sum of all of the Greens function solutions:
Z
u(r) = G(r, r 0 ) v(r 0 ) d3 r 0 .
(3.64)
We can easily demonstrate that this is the correct solution:
Z
Z
h
i
2
2
0
0
3 0
u(r) = G(r, r ) v(r ) d r = (r r 0 ) v(r 0 ) d3 r 0 = v(r).
(3.65)
Let us return to Eq. (3.59):

2 =
.
0
(3.66)
The Greens function for this equation satisfies Eq. (3.63) with |G| as |r| 0.
It follows from Eq. (3.55) that
G(r, r 0 ) =
1
1
.
4 |r r 0 |
(3.67)
Note, from Eq. (3.20), that the Greens function has the same form as the potential generated by a point charge. This is hardly surprising, given the definition of
a Greens function. It follows from Eq. (3.64) and (3.67) that the general solution
to Poissons equation, (3.66), is written
Z
1
(r 0 ) 3 0
(r) =
dr.
(3.68)
4 0 |r r 0 |
In fact, we have already obtained this solution by another method [see Eq. (3.17) ].
3.6 Amp`
eres experiments
As legend has it, in 1820 the Danish physicist Hans Christian rsted was giving
a lecture demonstration of various electrical and magnetic effects. Suddenly,
much to his surprise, he noticed that the needle of a compass he was holding was deflected when he moved it close to a current carrying wire. Up until
70
current carrying
wire
I
magnetic fieldlines
Figure 28:
then, magnetism has been thought of as solely a property of some rather unusual
rocks called loadstones. Word of this discovery spread quickly along the scientific
grapevine, and the French physicist Andre Marie Ampère immediately decided to
investigate further. Ampères apparatus consisted (essentially) of a long straight
wire carrying an electric current I. Ampère quickly discovered that the needle
of a small compass maps out a series of concentric circular loops in the plane
perpendicular to a current carrying wire (see Fig. 28). The direction of circulation around these magnetic loops is conventionally taken to be the direction in
which the North pole of the compass needle points. Using this convention, the
circulation of the loops is given by a right-hand rule: if the thumb of the righthand points along the direction of the current then the fingers of the right-hand
circulate in the same sense as the magnetic loops.
Ampères next series of experiments involved bringing a short test wire, carrying a current I 0 , close to the original wire, and investigating the force exerted on
the test wire (see Fig. 29). This experiment is not quite as clear cut as Coulombs
experiment because, unlike electric charges, electric currents cannot exist as point
entitiesthey have to flow in complete circuits. We must imagine that the circuit which connects with the central wire is sufficiently far away that it has no
appreciable influence on the outcome of the experiment. The circuit which connects with the test wire is more problematic. Fortunately, if the feed wires are
twisted around each other, as indicated in Fig. 29, then they effectively cancel
one another out, and also do not influence the outcome of the experiment.
71
I
current carrying test wire
Figure 29:
Ampère discovered that the force exerted on the test wire is directly proportional to its length. He also made the following observations. If the current in the
test wire (i.e., the test current) flows parallel to the current in the central wire
then the two wires attract one another. If the current in the test wire is reversed
then the two wires repel one another. If the test current points radially towards
the central wire (and the current in the central wire flows upwards) then the test
wire is subject to a downwards force. If the test current is reversed then the force
is upwards. If the test current is rotated in a single plane, so that it starts parallel
to the central current and ends up pointing radially towards it, then the force on
the test wire is of constant magnitude, and is always at right-angles to the test
current. If the test current is parallel to a magnetic loop then there is no force
exerted on the test wire. If the test current is rotated in a single plane, so that it
starts parallel to the central current and ends up pointing along a magnetic loop,
then the magnitude of the force on the test wire attenuates like cos (where
is the angle the current is turned through = 0 corresponds to the case where
the test current is parallel to the central current), and its direction is again always
at right-angles to the test current. Finally, Ampère was able to establish that the
attractive force between two parallel current carrying wires is proportional to the
product of the two currents, and falls off like the inverse of the perpendicular
distance between the wires.
This rather complicated force law can be summed up succinctly in vector notation provided that we define a vector field B, called the magnetic field, whose
direction is always parallel to the loops mapped out by a small compass. The de72
pendence of the force per unit length, F, acting on a test wire with the different
possible orientations of the test current is described by
F = I 0 B,
(3.69)
where I 0 is a vector whose direction and magnitude are the same as those of the
test current. Incidentally, the SI unit of electric current is the ampere (A), which
is the same as a coulomb per second. The SI unit of magnetic field-strength is
the tesla (T), which is the same as a newton per ampere per meter. The variation
of the force per unit length acting on a test wire with the strength of the central
current and the perpendicular distance r to the central wire is summed up by
saying that the magnetic field-strength is proportional to I and inversely proportional to r. Thus, defining cylindrical polar coordinates aligned along the axis of
the central current, we have
0 I
B =
,
(3.70)
2 r
with Br = Bz = 0. The constant of proportionality 0 is called the permeability of
free space, and takes the value
0 = 4 107 N A2 .
(3.71)
The concept of a magnetic field allows the calculation of the force on a test
wire to be conveniently split into two parts. In the first part, we calculate the
magnetic field generated by the current flowing in the central wire. This field
circulates in the plane normal to the wire: its magnitude is proportional to the
central current, and inversely proportional to the perpendicular distance from the
wire. In the second part, we use Eq. (3.69) to calculate the force per unit length
acting on a short current carrying wire located in the magnetic field generated
by the central current. This force is perpendicular to both the magnetic field and
the direction of the test current. Note that, at this stage, we have no reason to
suppose that the magnetic field has any real physical existence. It is introduced
merely to facilitate the calculation of the force exerted on the test wire by the
central wire.
73
3.7 The Lorentz force

The flow of an electric current down a conducting wire is ultimately due to the
motion of electrically charged particles (in most cases, electrons) through the
conducting medium. It seems reasonable, therefore, that the force exerted on
the wire when it is placed in a magnetic field is really the resultant of the forces
exerted on these moving charges. Let us suppose that this is the case.
Let A be the (uniform) cross-sectional area of the wire, and let n be the number density of mobile charges in the conductor. Suppose that the mobile charges
each have charge q and velocity v. We must assume that the conductor also contains stationary charges, of charge q and number density n (say), so that the
net charge density in the wire is zero. In most conductors, the mobile charges
are electrons and the stationary charges are atomic nuclei. The magnitude of the
electric current flowing through the wire is simply the number of coulombs per
second which flow past a given point. In one second, a mobile charge moves a
distance v, so all of the charges contained in a cylinder of cross-sectional area
A and length v flow past a given point. Thus, the magnitude of the current is
q nA v. The direction of the current is the same as the direction of motion of the
charges, so the vector current is I 0 = q nA v. According to Eq. (3.69), the force
per unit length acting on the wire is
F = q nA v B.
(3.72)
However, a unit length of the wire contains nA moving charges. So, assuming
that each charge is subject to an equal force from the magnetic field (we have no
reason to suppose otherwise), the force acting on an individual charge is
f = q v B.
(3.73)
We can combine this with Eq. (3.9) to give the force acting on a charge q moving
with velocity v in an electric field E and a magnetic field B:
f = q E + q v B.
(3.74)
This is called the Lorentz force law, after the Dutch physicist Hendrik Antoon
Lorentz who first formulated it. The electric force on a charged particle is parallel
74
to the local electric field. The magnetic force, however, is perpendicular to both
the local magnetic field and the particles direction of motion. No magnetic force
is exerted on a stationary charged particle.
The equation of motion of a free particle of charge q and mass m moving in
electric and magnetic fields is
dv
= q E + q v B,
(3.75)
dt
according to the Lorentz force law. This equation of motion was first verified
in a famous experiment carried out by the Cambridge physicist J.J. Thompson
in 1897. Thompson was investigating cathode rays, a then mysterious form of
radiation emitted by a heated metal element held at a large negative voltage (i.e.,
a cathode) with respect to another metal element (i.e., an anode) in an evacuated
tube. German physicists held that cathode rays were a form of electromagnetic
radiation, whilst British and French physicists suspected that they were, in reality,
a stream of charged particles. Thompson was able to demonstrate that the latter
view was correct. In Thompsons experiment, the cathode rays passed though
a region of crossed electric and magnetic fields (still in vacuum). The fields
were perpendicular to the original trajectory of the rays, and were also mutually
perpendicular.
m
Let us analyze Thompsons experiment. Suppose that the rays are originally
traveling in the x-direction, and are subject to a uniform electric field E in the
z-direction and a uniform magnetic field B in the y-direction. Let us assume, as
Thompson did, that cathode rays are a stream of particles of mass m and charge
q. The equation of motion of the particles in the z-direction is
d2 z
m 2 = q (E vB) ,
(3.76)
dt
where v is the velocity of the particles in the x-direction. Thompson started off his
experiment by only turning on the electric field in his apparatus, and measuring
the deflection d of the ray in the z-direction after it had traveled a distance l
through the electric field. It is clear from the equation of motion that
q E t2
q E l2
d=
=
,
m 2
m 2 v2
75
(3.77)
where the time of flight t is replaced by l/v. This formula is only valid if d l,
which is assumed to be the case. Next, Thompson turned on the magnetic field
in his apparatus, and adjusted it so that the cathode ray was no longer deflected.
The lack of deflection implies that the net force on the particles in the z-direction
was zero. In other words, the electric and magnetic forces balanced exactly. It
follows from Eq. (3.76) that with a properly adjusted magnetic field strength
v=
E
.
B
(3.78)
Thus, Eqs. (3.77) and (3.78) and can be combined and rearranged to give the
charge to mass ratio of the particles in terms of measured quantities:
2dE
q
= 2 2.
m
l B
(3.79)
Using this method, Thompson inferred that cathode rays were made up of negatively charged particles (the sign of the charge is obvious from the direction of
the deflection in the electric field) with a charge to mass ratio of 1.710 11 C/kg.
A decade later, in 1908, the American Robert Millikan performed his famous oil
drop experiment, and discovered that mobile electric charges are quantized in
units of 1.6 1019 C. Assuming that mobile electric charges and the particles
which make up cathode rays are one and the same thing, Thompsons and Millikans experiments imply that the mass of these particles is 9.4 10 31 kg. Of
course, this is the mass of an electron (the modern value is 9.1 1031 kg), and
1.61019 C is the charge of an electron. Thus, cathode rays are, in fact, streams
of electrons which are emitted from a heated cathode, and then accelerated because of the large voltage difference between the cathode and anode.
Consider, now, a particle of mass m and charge q moving in a uniform magnetic field, B = B ^z. According, to Eq. (3.75), the particles equation of motion
can be written:
dv
m
= q v B.
(3.80)
dt
This reduces to
dvx
= vy ,
dt
76
(3.81)
dvy
= vx ,
(3.82)
dt
dvz
= 0.
(3.83)
dt
Here, = q B/m is called the cyclotron frequency. The above equations can be
solved to give
vx = v cos( t),
(3.84)
vy = v sin( t),
(3.85)
vz = v k ,
(3.86)
and
v
sin( t),
v
cos( t),
y =
z = vk t.
x =
(3.87)
(3.88)
(3.89)
According to these equations, the particle trajectory is a spiral whose axis is parallel to the magnetic field. The radius of the spiral is = v /, where v is the
particles constant speed in the plane perpendicular to the magnetic field. The
particle drifts parallel to the magnetic field at a constant velocity, v k . Finally, the
particle gyrates in the plane perpendicular to the magnetic field at the cyclotron
frequency.
Finally, if a particle is subject to a force f and moves a distance r in a time
interval t, then the work done on the particle by the force is
W = f r.
(3.90)
The power input to the particle from the force field is

W
= f v,
(3.91)
t0 t
where v is the particles velocity. It follows from the Lorentz force law, Eq. (3.74),
that the power input to a particle moving in electric and magnetic fields is
P = lim
P = q v E.
77
(3.92)
3.8 Ampères law
Note that a charged particle can gain (or lose) energy from an electric field, but
not from a magnetic field. This is because the magnetic force is always perpendicular to the particles direction of motion, and, therefore, does no work on the
particle [see Eq. (3.90) ]. Thus, in particle accelerators, magnetic fields are often used to guide particle motion (e.g., in a circle) but the actual acceleration is
performed by electric fields.
3.8 Amp`
eres law
Magnetic fields, like electric fields, are completely superposable. So, if a field B 1 is
generated by a current I1 flowing through some circuit, and a field B2 is generated
by a current I2 flowing through another circuit, then when the currents I1 and I2
flow through both circuits simultaneously the generated magnetic field is B1 + B2 .
I1
I2
B1
B2
Figure 30:
Consider two parallel wires separated by a perpendicular distance r and carrying electric currents I1 and I2 , respectively (see Fig. 30). The magnetic field
strength at the second wire due to the current flowing in the first wire is B =
0 I1 /2 r. This field is orientated at right-angles to the second wire, so the force
per unit length exerted on the second wire is
0 I1 I2
.
(3.93)
2 r
This follows from Eq. (3.69), which is valid for continuous wires as well as short
test wires. The force acting on the second wire is directed radially inwards toF=
78
3.9 Magnetic monopoles?
wards the first wire. The magnetic field strength at the first wire due to the
current flowing in the second wire is B = 0 I2 /2 r. This field is orientated at
right-angles to the first wire, so the force per unit length acting on the first wire
is equal and opposite to that acting on the second wire, according to Eq. (3.69).
Equation (3.93) is sometimes called Ampères law, and is clearly another example
of an action at a distance law: i.e., if the current in the first wire is suddenly
changed then the force on the second wire immediately adjusts. In reality, there
should be a short time delay, at least as long as the propagation time for a light
signal between the two wires. Clearly, Ampères law is not strictly correct. However, as long as we restrict our investigations to steady currents it is perfectly
adequate.

Suppose that we have an infinite straight wire carrying an electric current I. Let
the wire be aligned along the z-axis. The magnetic field generated by such a wire
is written
!
0 I y x
, ,0
(3.94)
B=
2 r2 r2
q
in Cartesian coordinates, where r = x2 + y2 . The divergence of this field is

0 I 2 y x 2 x y
4
= 0,
B=
2
r4
r
!
(3.95)
where use has been made of r/x = x/r, etc. We saw in Sect. 3.4 that the
divergence of the electric field appeared, at first sight, to be zero. But, in reality,
it was a delta-function, because the volume integral of E was non-zero. Does
the same sort of thing happen for the divergence
of the magnetic field? Well, if
H
we could find
R a closed surface S for which S BdS 6= 0 then, according to Gauss
theorem, V B dV 6= 0, where V is the volume enclosed by S. This would
certainly imply that B is some sort of delta-function. So, can we find such
a surface? The short answer is, no. Consider a cylindrical surface aligned with
the wire. The magnetic field is everywhere tangential to the outward surface
element, so this surface certainly has zero magnetic flux coming out of it. In fact,
79
H
it is impossible to invent any closed surface for which S B dS 6= 0 with B given
by Eq. (3.94) (if you do not believe this, try it yourselves!). This suggests that
the divergence of a magnetic field generated by steady electric currents really is
zero. Admittedly, we have only proved this for infinite straight currents, but, as
will be demonstrated presently, it is true in general.
If B = 0 then B is a solenoidal vector field. In other words, field-lines of B
never begin or end. This is certainly the case in Eq. (3.94) where the field-lines
are a set of concentric circles centred on the z-axis. What about magnetic fields
generated by permanent magnets (the modern equivalent of loadstones)? Do
they also never begin or end? Well, we know that a conventional bar magnet has
both a North and South magnetic pole (like the Earth). If we track the magnetic
field-lines with a small compass they all emanate from the South pole, spread
out, and eventually reconverge on the North pole (see Fig. 31). It appears likely
(but we cannot prove it with a compass) that the field-lines inside the magnet
connect from the North to the South pole so as to form closed loops which never
begin or end.
N
S
Figure 31:
Can we produce an isolated North or South magnetic pole: for instance, by

snapping a bar magnet in two? A compass needle would always point towards
an isolated North pole, so this would act like a negative magnetic charge. Likewise, a compass needle would always point away from an isolated South pole,
so this would act like a positive magnetic charge. It is clear from Fig. 32 that
if we take a closed surface S containing an isolated magnetic pole, which is usu80
H
ally termed a magnetic monopole, then S B dS 6= 0: the flux will be positive for
an isolated South pole, Hand negative for an isolated North pole. It follows from
Gauss theorem that if S B dS 6= 0 then B 6= 0. Thus, the statement that
magnetic fields are solenoidal, or that B = 0, is equivalent to the statement
that there are no magnetic monopoles. It is not clear, a priori, that this is a true
statement. In fact, it is quite possible to formulate electromagnetism so as to allow for magnetic monopoles. However, as far as we know, there are no magnetic
monopoles in the Universe. At least, if there are any then they are all hiding
from us! We know that if we try to make a magnetic monopole by snapping a
bar magnet in two then we just end up with two smaller bar magnets. If we snap
one of these smaller magnets in two then we end up with two even smaller bar
magnets. We can continue this process down to the atomic level without ever
producing a magnetic monopole. In fact, permanent magnetism is generated by
electric currents circulating on the atomic scale, so this type of magnetism is not
fundamentally different to the magnetism generated by macroscopic currents.
Figure 32:
In conclusion, all steady magnetic fields in the Universe are generated by circulating electric currents of some description. Such fields are solenoidal: that is,
they never begin or end, and satisfy the field equation
B = 0.
(3.96)
This, incidentally, is the second of Maxwells equations. Essentially, it says that

there is no such thing as a magnetic monopole. We have only proved that B = 0
81
3.10 Ampères circuital law
for steady magnetic fields, but, in fact, this is also the case for time-dependent
fields (see later).
3.10 Amp`
eres circuital law
Consider, again, an infinite straight wire aligned along the z-axis and carrying a
current I. The field generated by such a wire is written
0 I
(3.97)
2 r
in cylindrical polar coordinates. Consider a circular loop C in the x-y plane which
is centred on theHwire. Suppose that the radius of this loop is r. Let us evaluate
the line integral C B dl. This integral is easy to perform because the magnetic
field is always parallel to the line element. We have
I
I
B dl = B r d = 0 I.
(3.98)
B =
However, we know from Stokes theorem that

I
Z
B dl = B dS,
C
(3.99)
where S is any surface attached to the loop C.
Let us evaluate B directly. According to Eq. (3.94),
Bz By
= 0,
y
z
Bx Bz
= 0,
=
z
x
By Bx
0 I 1
2 x2
1
2 y2
=
=
4 + 2 4 = 0,
x
y
2 r2
r
r
r
( B)x =
(3.100)
( B)y
(3.101)
( B)z
(3.102)
where use has been made of r/x = x/r, etc. We now have a problem. Equations
(3.98) and (3.99) imply that
Z
B dS = 0 I.
(3.103)
S
82
But, we have just demonstrated that B = 0. This problem

is very reminiscent
R
of the difficulty we had earlier with E. Recall that V E dV = q/0 for a
volume V containing a discrete charge q, but that E = 0 at a general point. We
got around this problem by saying that E is a three-dimensional delta-function
whose spike is coincident with the location of the charge. Likewise, we can get
around our present difficulty by saying that B is a two-dimensional deltafunction. A three-dimensional delta-function is a singular (but integrable) point
in space, whereas a two-dimensional delta-function is a singular line in space. It
is clear from an examination of Eqs. (3.100)(3.102) that the only component of
B which can be singular is the z-component, and that this can only be singular
on the z-axis (i.e., r = 0). Thus, the singularity coincides with the location of the
current, and we can write
B = 0 I (x) (y) ^z.
(3.104)
The above equation certainly gives ( B)x = ( B)y = 0, and ( B)z =

0 everywhere apart from the z-axis, in accordance with Eqs. (3.100)(3.102).
Suppose that we integrate over a plane surface S connected to the loop C. The
surface element is dS = dx dy ^z, so
Z
ZZ
B dS = 0 I
(x) (y) dx dy
(3.105)
S
where the integration is performed over the region x2 + y2 r. However, since

the only part of S which actually contributes to the surface integral is the bit
which lies infinitesimally close to the z-axis, we can integrate over all x and y
without changing the result. Thus, we obtain
Z
Z
Z
B dS = 0 I
(x) dx
(y) dy = 0 I,
(3.106)
S
which is in agreement with Eq. (3.103).

But, why have we gone to so much trouble to prove something using vector
field theory which can be demonstrated in one line via conventional analysis
[see Eq. (3.98) ]? The answer, of course, is that the vector field result is easily
generalized, whereas the conventional result is just a special case. For instance,
83
it is clear that Eq. (3.106) is true for any surface attached to the loop C, not
just a plane surface. Moreover, suppose that we distort our simple circular loop
C so that it is no longer circular or even lies in one plane. What now is the line
integral of B around the loop? This is no longer a simple problem for conventional
analysis, because the magnetic field is not parallel to a line element of the loop.
However, according to Stokes theorem,
I
Z
B dl = B dS,
(3.107)
C
with B given by Eq. (3.104). Note that the only part of S which contributes
to the surface integral is an infinitesimal region centered on the z-axis. So, as
long as S actually intersects the z-axis, it does not matter what shape the rest the
surface is, and we always get the same answer for the surface integral: namely,
I
Z
B dl = B dS = 0 I.
(3.108)
C
Thus, provided the curve C circulates the z-axis, and,

H therefore, any surface
S attached to C intersects the z-axis, the line integral C B dl is equal to 0 I.
Of course, if C does not H
circulate the z-axis then an attached surface S does not
intersectHthe z-axis and C B dl is zero. There is one more proviso. The line
integral C B dl is 0 I for a loop which circulates the z-axis in a clockwise direction (looking up the z-axis). However, if the loop circulates in an anti-clockwise
direction then the integral is 0 I. This follows because in the latter case the
z-component of the surface element dS is oppositely directed to the current flow
at the point where the surface intersects the wire.
Let us now consider N wires directed along the z-axis, with coordinates (xi ,
yi ) in the x-y plane, each carrying a current Ii in the positive z-direction. It is
fairly obvious that Eq. (3.104) generalizes to
B = 0
N
X
Ii (x xi ) (y yi ) ^z.
(3.109)
i=1
If we integrate the magnetic field around some closed curve C, which can have
any shape and does not necessarily lie in one plane, then Stokes theorem and
84
the above equation imply that

I
Z
B dl = B dS = 0 I,
C
(3.110)
where I is the total current enclosed by the curve. Again, if the curve circulates
the ith wire in a clockwise direction (looking down the direction of current flow)
then the wire contributes Ii to the aggregate current I. On the other hand, if
the curve circulates in an anti-clockwise direction then the wire contributes I i .
Finally, if the curve does not circulate the wire at all then the wire contributes
nothing to I.
Equation (3.109) is a field equation describing how a set of z-directed current
carrying wires generate a magnetic field. These wires have zero-thickness, which
implies that we are trying to squeeze a finite amount of current into an infinitesimal region. This accounts for the delta-functions on the right-hand side of the
equation. Likewise, we obtained delta-functions in Sect. 3.4 because we were
dealing with point charges. Let us now generalize to the more realistic case of
diffuse currents. Suppose that the z-current flowing through a small rectangle
in the x-y plane, centred on coordinates (x, y) and of dimensions dx and dy, is
jz (x, y) dx dy. Here, jz is termed the current density in the z-direction. Let us
integrate ( B)z over this rectangle. The rectangle is assumed to be sufficiently
small that ( B)z does not vary appreciably across it. According to Eq. (3.110),
this integral is equal to 0 times the total z-current flowing through the rectangle.
Thus,
( B)z dx dy = 0 jz dx dy,
(3.111)
which implies that
( B)z = 0 jz .
(3.112)
( B)x = 0 jx ,
(3.113)
Of course, there is nothing special about the z-axis. Suppose we have a set of
diffuse currents flowing in the x-direction. The current flowing through a small
rectangle in the y-z plane, centred on coordinates (y, z) and of dimensions dy
and dz, is given by jx (y, z) dy dz, where jx is the current density in the x-direction.
It is fairly obvious that we can write
85
with a similar equation for diffuse currents flowing along the y-axis. We can
combine these equations with Eq. (3.112) to form a single vector field equation
which describes how electric currents generate magnetic fields,
B = 0 j,
(3.114)
where j = (jx , jy , jz ) is the vector current density. This is the third Maxwell equation. The electric current flowing through a small area dS located at position r is
j(r) dS. Suppose that space is filled with particles of charge q, number density
n(r), and velocity v(r). The charge density is given by (r) = q n. The current
density is given by j(r) = q n v, and is obviously a proper vector field (velocities
are proper vectors since they are ultimately derived from displacements).
If we form the line integral of B around some general closed curve C, making
use of Stokes theorem and the field equation (3.114), then we obtain
Z
I
(3.115)
B dl = 0 j dS.
S
In other words, the line integral of the magnetic field around any closed loop C
is equal to 0 times the flux of the current density through C. This result is called
Ampères circuital law. If the currents flow in zero-thickness wires then Ampères
circuital law reduces to Eq. (3.110).
The flux of the current density through C is evaluated by integrating j dS
over any surface S attached to C. Suppose that we take two different surfaces
S1 and S2 . It is clear
that if Ampères circuital law is toR make any sense then the
R
surface integral S1 j dS had better equal the integral S2 j dS. That is, when we
work out the flux of the current though C using two different attached surfaces
then we had better get the same answer, otherwise Eq. (3.115) is wrong (since
the left-hand side is clearly independent of the surface spanning C). We saw in
Sect. 2 that if the integral of a vector field A over some surface attached to a loop
depends only on the loop, and is independent of the surface which spans it, then
this implies that A = 0. The flux ofR the current density through any loop C is
calculated by evaluating the integral S j dS for any surface S which spans the
loop. According to Ampères circuital law, this integral depends only on C and is
completely independent of S (i.e., it is equal to the line integral of B around C,
86
which depends on C but not on S). This implies that j = 0. In fact, we can
obtain this relation directly from the field equation (3.114). We know that the
divergence of a curl is automatically zero, so taking the divergence of Eq. (3.114),
we obtain
j = 0.
(3.116)
We have shown that if Ampères circuital law is to make any sense then we
need j = 0. Physically, this implies that the net current flowing through any
closed surface S is zero. Up to now, we have only considered stationary charges
and steady currents. It is clear that if all charges are stationary and all currents
are steady then there can be no net current flowing through a closed surface S,
since this would imply a build up of charge in the volume V enclosed by S. In
other words, as long as we restrict our investigation to stationary charges, and
steady currents, then we expect j = 0, and Ampères circuital law makes
sense. However, suppose that we now relax this restriction. Suppose that some
of the charges in a volume V decide to move outside V. Clearly, there will be a
non-zero net flux of electric current through the bounding surface S whilst this
is happening. This implies from Gauss theorem that j 6= 0. Under these
circumstances Ampères circuital law collapses in a heap. We shall see later that
we can rescue Ampères circuital law by adding an extra term involving a time
derivative to the right-hand side of the field equation (3.114). For steady-state
situations (i.e., /t = 0), this extra term can be neglected. Thus, the field
equation B = 0 j is, in fact, only two-thirds of Maxwells third equation:
there is a term missing on the right-hand side.
We have now derived two field equations involving magnetic fields (actually,
we have only derived one and two-thirds):
B = 0,
B = 0 j.
(3.117)
(3.118)
We obtained these equations by looking at the fields generated by infinitely long,

straight, steady currents. This, of course, is a rather special class of currents. We
should now go back and repeat the process for general currents. In fact, if we
did this we would find that the above field equations still hold (provided that
87
3.11 Helmholtzs theorem
the currents are steady). Unfortunately, this demonstration is rather messy and
extremely tedious. There is a better approach. Let us assume that the above field
equations are valid for any set of steady currents. We can then, with relatively
little effort, use these equations to generate the correct formula for the magnetic
field induced by a general set of steady currents, thus proving that our assumption
is correct. More of this later.

Let us now embark on a slight mathematical digression. Up to now, we have
only studied the electric and magnetic fields generated by stationary charges and
steady currents. We have found that these fields are describable in terms of four
field equations:
,
(3.119)
E =
0
E = 0
(3.120)
for electric fields, and
B = 0,
B = 0 j
(3.121)
(3.122)
for magnetic fields. There are no other field equations. This strongly suggests that
if we know the divergence and the curl of a vector field then we know everything
there is to know about the field. In fact, this is the case. There is a mathematical
theorem which sums this up. It is called Helmholtzs theorem after the German
polymath Hermann Ludwig Ferdinand von Helmholtz.
Let us start with scalar fields. Field equations are a type of differential equation: i.e., they deal with the infinitesimal differences in quantities between neighbouring points. The question is, what differential equation completely specifies
a scalar field? This is easy. Suppose that we have a scalar field and a field
equation which tells us the gradient of this field at all points: something like
= A,
88
(3.123)
where A(r) is a vector field. Note that we need A = 0 for self consistency,
since the curl of a gradient is automatically zero. The above equation completely
specifies once we are given the value of the field at a single point, P (say).
Thus,
Z
Z
Q
(Q) = (P) +
dl = (P) +
A dl,
(3.124)
where Q is a general point. The fact that A = 0 means that A is a conservative

field, which guarantees that the above equation gives a unique value for at a
general point in space.
Suppose that we have a vector field F. How many differential equations do we

need to completely specify this field? Hopefully, we only need two: one giving
the divergence of the field, and one giving its curl. Let us test this hypothesis.
Suppose that we have two field equations:
F = D,
F = C,
(3.125)
(3.126)
where D is a scalar field and C is a vector field. For self-consistency, we need

C = 0,
(3.127)
since the divergence of a curl is automatically zero. The question is, do these
two field equations plus some suitable boundary conditions completely specify
F? Suppose that we write
F = U + W.
(3.128)
In other words, we are saying that a general field F is the sum of a conservative
field, U, and a solenoidal field, W. This sounds plausible, but it remains
to be proved. Let us start by taking the divergence of the above equation, and
making use of Eq. (3.125). We get
2 U = D.
(3.129)
Note that the vector field W does not figure in this equation, because the divergence of a curl is automatically zero. Let us now take the curl of Eq. (3.128):
F = W = ( W) 2 W = 2 W.
89
(3.130)
Here, we assume that the divergence of W is zero. This is another thing which
remains to be proved. Note that the scalar field U does not figure in this equation,
because the curl of a divergence is automatically zero. Using Eq. (3.126), we get
2 Wx = Cx ,
2 Wy = Cy ,
2 Wz = Cz ,
(3.131)
(3.132)
(3.133)
So, we have transformed our problem into four differential equations, Eq. (3.129)
and Eqs. (3.131)(3.133), which we need to solve. Let us look at these equations.
We immediately notice that they all have exactly the same form. In fact, they are
all versions of Poissons equation. We can now make use of a principle made
famous by Richard P. Feynman: the same equations have the same solutions.
Recall that earlier on we came across the following equation:
2 = ,
(3.134)
0
where is the electrostatic potential and is the charge density. We proved that
the solution to this equation, with the boundary condition that goes to zero at
infinity, is
Z
(r 0 ) 3 0
1
dr.
(3.135)
(r) =
40 |r r 0 |
Well, if the same equations have the same solutions, and Eq. (3.135) is the
solution to Eq. (3.134), then we can immediately write down the solutions to
Eq. (3.129) and Eqs. (3.131)(3.133). We get
Z
1
D(r 0 ) 3 0
U(r) =
dr,
(3.136)
4 |r r 0 |
and
1
Wx (r) =
4
1
Wy (r) =
4
1
Wz (r) =
4
Z
Z
Z
90
Cx (r 0 ) 3 0
dr,
|r r 0 |
(3.137)
Cy (r 0 ) 3 0
dr,
|r r 0 |
(3.138)
Cz (r 0 ) 3 0
dr.
|r r 0 |
(3.139)
The last three equations can be combined to form a single vector equation:
Z
C(r 0 ) 3 0
1
dr.
(3.140)
W(r) =
4 |r r 0 |
We assumed earlier that W = 0. Let us check to see if this is true. Note that
1
x x0
x0 x
1
=
,
=
=
x |r r 0 |
|r r 0 |3
|r r 0 |3
x 0 |r r 0 |
(3.141)
which implies that
1
1
0
=
|r r 0 |
|r r 0 |
(3.142)
where 0 is the operator (/x 0 , /y 0 , /z 0 ). Taking the divergence of Eq. (3.140),

and making use of the above relation, we obtain
Z
Z
1
1
1
1
d3 r 0 =
d3 r 0 . (3.143)
W =
C(r 0 )
C(r 0 ) 0
0
0
4
|r r |
4
|r r |
Now
Z
f
g
dx = [gf]
dx.
(3.144)
g
f
x
x
However, if g f 0 as x then we can neglect the first term on the righthand side of the above equation and write
Z
Z
f
g
g
f
dx =
dx.
(3.145)
x
x
A simple generalization of this result yields
Z
Z
3
g f d r = f g d3 r,
provided that gx f 0 as |r| , etc. Thus, we can deduce that

Z
1 0 C(r 0 ) 3 0
W=
dr
4
|r r 0 |
(3.146)
(3.147)
from Eq. (3.143), provided |C(r)| is bounded as |r| . However, we have

already shown that C = 0 from self-consistency arguments, so the above
equation implies that W = 0, which is the desired result.
91
We have constructed a vector field F which satisfies Eqs. (3.125) and (3.126)
and behaves sensibly at infinity: i.e., |F| 0 as |r| . But, is our solution
the only possible solution of Eqs. (3.125) and (3.126) with sensible boundary
conditions at infinity? Another way of posing this question is to ask whether
there are any solutions of
2 U = 0,
2 Wi = 0,
(3.148)
where i denotes x, y, or z, which are bounded at infinity. If there are then we are
in trouble, because we can take our solution and add to it an arbitrary amount
of a vector field with zero divergence and zero curl, and thereby obtain another
solution which also satisfies physical boundary conditions. This would imply that
our solution is not unique. In other words, it is not possible to unambiguously
reconstruct a vector field given its divergence, its curl, and physical boundary
conditions. Fortunately, the equation
2 = 0,
(3.149)
which is called Laplaces equation, has a very nice property: its solutions are
unique. That is, if we can find a solution to Laplaces equation which satisfies
the boundary conditions then we are guaranteed that this is the only solution.
We shall prove this later on in the course. Well, let us invent some solutions to
Eqs. (3.148) which are bounded at infinity. How about
U = Wi = 0?
(3.150)
These solutions certainly satisfy Laplaces equation, and are well-behaved at infinity. Because the solutions to Laplaces equations are unique, we know that
Eqs. (3.150) are the only solutions to Eqs. (3.148). This means that there is
no vector field which satisfies physical boundary equations at infinity and has
zero divergence and zero curl. In other words, our solution to Eqs. (3.125) and
(3.126) is the only solution. Thus, we have unambiguously reconstructed the
vector field F given its divergence, its curl, and sensible boundary conditions at
infinity. This is Helmholtzs theorem.
We have just proved a number of very useful, and also very important, points.
First, according to Eq. (3.128), a general vector field can be written as the sum
92
of a conservative field and a solenoidal field. Thus, we ought to be able to write

electric and magnetic fields in this form. Second, a general vector field which is
zero at infinity is completely specified once its divergence and its curl are given.
Thus, we can guess that the laws of electromagnetism can be written as four field
equations,
E = something,
(3.151)
B = something,
(3.153)
E = something,
B = something,
(3.152)
(3.154)
without knowing the first thing about electromagnetism (other than the fact that
it deals with two vector fields). Of course, Eqs. (3.119)(3.122) are of exactly
this form. We also know that there are only four field equations, since the above
equations are sufficient to completely reconstruct both E and B. Furthermore, we
know that we can solve the field equations without even knowing what the righthand sides look like. After all, we solved Eqs. (3.125)(3.126) for completely
general right-hand sides. [Actually, the right-hand sides have to go to zero at infinity, otherwise integrals like Eq. (3.136) blow up.] We also know that any solutions we find are unique. In other words, there is only one possible steady electric
and magnetic field which can be generated by a given set of stationary charges
and steady currents. The third thing which we proved was that if the right-hand
sides of the above field equations are all zero then the only physical solution is
E = B = 0. This implies that steady electric and magnetic fields cannot generate
themselves. Instead, they have to be generated by stationary charges and steady
currents. So, if we come across a steady electric field we know that if we trace
the field-lines back we shall eventually find a charge. Likewise, a steady magnetic field implies that there is a steady current flowing somewhere. All of these
results follow from vector field theory (i.e., from the general properties of fields
in three-dimensional space), prior to any investigation of electromagnetism.
93
3.12 The magnetic vector potential

Electric fields generated by stationary charges obey
E = 0.
(3.155)
This immediately allows us to write

E = ,
(3.156)
since the curl of a gradient is automatically zero. In fact, whenever we come

across an irrotational vector field in physics we can always write it as the gradient
of some scalar field. This is clearly a useful thing to do, since it enables us to
replace a vector field by a much simpler scalar field. The quantity in the above
equation is known as the electric scalar potential.
Magnetic fields generated by steady currents (and unsteady currents, for that
matter) satisfy
B = 0.
(3.157)
This immediately allows us to write
B = A,
(3.158)
since the divergence of a curl is automatically zero. In fact, whenever we come

across a solenoidal vector field in physics we can always write it as the curl of
some other vector field. This is not an obviously useful thing to do, however, since
it only allows us to replace one vector field by another. Nevertheless, Eq. (3.158)
is one of the most useful equations we shall come across in this lecture course.
The quantity A is known as the magnetic vector potential.
We know from Helmholtzs theorem that a vector field is fully specified by its
divergence and its curl. The curl of the vector potential gives us the magnetic
field via Eq. (3.158). However, the divergence of A has no physical significance.
In fact, we are completely free to choose A to be whatever we like. Note that,
according to Eq. (3.158), the magnetic field is invariant under the transformation
A A .
94
(3.159)
In other words, the vector potential is undetermined to the gradient of a scalar

field. This is just another way of saying that we are free to choose A. Recall
that the electric scalar potential is undetermined to an arbitrary additive constant,
since the transformation
+c
(3.160)
leaves the electric field invariant in Eq. (3.156). The transformations (3.159)
and (3.160) are examples of what mathematicians call gauge transformations.
The choice of a particular function or a particular constant c is referred to as
a choice of the gauge. We are free to fix the gauge to be whatever we like. The
most sensible choice is the one which makes our equations as simple as possible.
The usual gauge for the scalar potential is such that 0 at infinity. The
usual gauge for A is such that
A = 0.
(3.161)
This particular choice is known as the Coulomb gauge.
It is obvious that we can always add a constant to so as to make it zero at

infinity. But it is not at all obvious that we can always perform a gauge transformation such as to make A zero. Suppose that we have found some vector field
A whose curl gives the magnetic field but whose divergence in non-zero. Let
A = v(r).
(3.162)
The question is, can we find a scalar field such that after we perform the gauge
transformation (3.159) we are left with A = 0. Taking the divergence of
Eq. (3.159) it is clear that we need to find a function which satisfies
2 = v.
(3.163)
But this is just Poissons equation. We know that we can always find a unique
solution of this equation (see Sect. 3.11). This proves that, in practice, we can
always set the divergence of A equal to zero.
Let us again consider an infinite straight wire directed along the z-axis and
carrying a current I. The magnetic field generated by such a wire is written
0 I y x
B=
, ,0 .
2 r2 r2
!
95
(3.164)
We wish to find a vector potential A whose curl is equal to the above magnetic
field, and whose divergence is zero. It is not difficult to see that

0 I
A=
0, 0, ln[x2 + y2 ]
(3.165)
4
fits the bill. Note that the vector potential is parallel to the direction of the current. This would seem to suggest that there is a more direct relationship between
the vector potential and the current than there is between the magnetic field and
the current. The potential is not very well-behaved on the z-axis, but this is just
because we are dealing with an infinitely thin current.
Let us take the curl of Eq. (3.158). We find that
B = A = ( A) 2 A = 2 A,
(3.166)
where use has been made of the Coulomb gauge condition (3.161). We can
combine the above relation with the field equation (3.114) to give
2 A = 0 j.
(3.167)
Writing this in component form, we obtain

2 Ax = 0 jx ,
2 Ay = 0 jy ,
2 Az = 0 jz .
(3.168)
(3.169)
(3.170)
But, this is just Poissons equation three times over. We can immediately write the
unique solutions to the above equations:
Z
0 jx (r 0 ) 3 0
Ax (r) =
dr,
(3.171)
4 |r r 0 |
Z
0 jy (r 0 ) 3 0
dr,
(3.172)
Ay (r) =
4 |r r 0 |
Z
0 jz (r 0 ) 3 0
Az (r) =
dr.
(3.173)
4 |r r 0 |
These solutions can be recombined to form a single vector solution
Z
j(r 0 ) 3 0
0
dr.
A(r) =
4 |r r 0 |
96
(3.174)
Of course, we have seen a equation like this before:

Z
(r 0 ) 3 0
1
(r) =
dr.
4 0 |r r 0 |
3.13 The Biot-Savart law
(3.175)
Equations (3.174) and (3.175) are the unique solutions (given the arbitrary
choice of gauge) to the field equations (3.119)(3.122): they specify the magnetic vector and electric scalar potentials generated by a set of stationary charges,
of charge density (r), and a set of steady currents, of current density j(r). Incidentally, we can prove that Eq. (3.174) satisfies the gauge condition A = 0 by
repeating the analysis of Eqs. (3.140)(3.147) (with W A and C 0 j), and
using the fact that j = 0 for steady currents.
According to Eq. (3.156), we can obtain an expression for the electric field generated by stationary charges by taking minus the gradient of Eq. (3.175). This
yields
Z
r r0 3 0
1
(r 0 )
dr,
(3.176)
E(r) =
40
|r r 0 |3
which we recognize as Coulombs law written for a continuous charge distribution. According to Eq. (3.158), we can obtain an equivalent expression for the
magnetic field generated by steady currents by taking the curl of Eq. (3.174).
This gives
Z
0 j(r 0 ) (r r 0 ) 3 0
dr,
(3.177)
B(r) =
4
|r r 0 |3
where use has been made of the vector identity ( A) = A + A.

Equation (3.177) is known as the Biot-Savart law after the French physicists Jean
Baptiste Biot and Felix Savart: it completely specifies the magnetic field generated
by a steady (but otherwise quite general) distributed current.
Let us reduce our distributed current to an idealized zero thickness wire. We

can do this by writing
j(r) d3 r = I(r) dl,
(3.178)
97
where I is the vector current (i.e., its direction and magnitude specify the direction and magnitude of the current) and dl is an element of length along the wire.
Equations (3.177) and (3.178) can be combined to give
Z
0 I(r 0 ) (r r 0 )
dl,
(3.179)
B(r) =
4
|r r 0 |3
which is the form in which the Biot-Savart law is most usually written. This law is
to magnetostatics (i.e., the study of magnetic fields generated by steady currents)
what Coulombs law is to electrostatics (i.e., the study of electric fields generated
by stationary charges). Furthermore, it can be experimentally verified given a set
of currents, a compass, a test wire, and a great deal of skill and patience. This
justifies our earlier assumption that the field equations (3.117) and (3.118) are
valid for general current distributions (recall that we derived them by studying
the fields generated by infinite, straight wires). Note that both Coulombs law
and the Biot-Savart law are gauge independent: i.e., they do not depend on the
particular choice of gauge.
I
z
dl
l
r r
Figure 33:
Consider an infinite straight wire, directed along the z-axis, and carrying a
current I (see Fig. 33). Let us reconstruct the magnetic field generated by the
wire at point P using the Biot-Savart law. Suppose that the perpendicular distance
to the wire is . It is easily seen that
^
I (r r 0 ) = I ,
98
(3.180)
3.14 Electrostatics and magnetostatics
l = tan ,
d,
dl =
cos2
|r r 0 | =
.
cos
Thus, according to Eq. (3.179), we have
Z
I
0 /2
d
B =
4 /2 3 (cos )3 cos2
Z
0 I
0 I /2
/2
cos d =
=
[sin ]/2 ,
4 /2
4
(3.181)
(3.182)
(3.183)
(3.184)
which gives the familiar result

0 I
.
(3.185)
2
So, we have come full circle in our investigation of magnetic fields. Note that
the simple result (3.185) can only be obtained from the Biot-Savart law after
some non-trivial algebra. Examination of more complicated current distributions
using this law invariably leads to lengthy, involved, and extremely unpleasant
calculations.
B =

We have now completed our theoretical investigation of electrostatics and magnetostatics. Our next task is to incorporate time variation into our analysis. However, before we start this, let us briefly review our progress so far. We have found
that the electric fields generated by stationary charges, and the magnetic fields
generated by steady currents, are describable in terms of four field equations:
E =
,
(3.186)
0
E = 0,
(3.187)
B = 0,
B = 0 j.
99
(3.188)
(3.189)
The boundary conditions are that the fields are zero at infinity, assuming that the
generating charges and currents are localized to some region in space. According
to Helmholtzs theorem, the above field equations, plus the boundary conditions,
are sufficient to uniquely specify the electric and magnetic fields. The physical
significance of this is that divergence and curl are the only rotationally invariant
first-order differential properties of a general vector field: i.e., the only quantities
which do not change their physical characteristics when the coordinate axes are
rotated. Since physics does not depend on the orientation of the coordinate axes
(which is, after all, quite arbitrary), divergence and curl are the only quantities
which can appear in first-order differential field equations which claim to describe
physical phenomena.
The field equations can be integrated to give:
Z
I
1
dV,
E dS =

0
V
S
I
E dl = 0,
C
I
B dS = 0,
S
Z
I
B dl = 0 j dS.
(3.190)
(3.191)
(3.192)
(3.193)
S0
Here, S is a closed surface enclosing a volume V. Also, C is a closed loop, and

S 0 is some surface attached to this loop. The field equations (3.186)(3.189)
can be deduced from Eqs. (3.190)(3.193) using Gauss theorem and Stokes
theorem. Equation (3.190) is called Gauss law, and says that the flux of the
electric field out of a closed surface is proportional to the enclosed electric charge.
Equation (3.192) has no particular name, and says that there is no such things as
a magnetic monopole. Equation (3.193) is called Ampères circuital law, and says
that the line integral of the magnetic field around any closed loop is proportional
to the flux of the current through the loop. Finally. Eqs. (3.191) and (3.193) are
incomplete: each acquires an extra term on the right-hand side in time-dependent
situations.
100
The field equation (3.187) is automatically satisfied if we write

E = .
(3.194)
Likewise, the field equation (3.188) is automatically satisfied if we write

B = A.
(3.195)
Here, is the electric scalar potential, and A is the magnetic vector potential. The
electric field is clearly unchanged if we add a constant to the scalar potential:
E E as + c.
(3.196)
B B as A A + .
(3.197)
A = 0.
(3.199)
The magnetic field is similarly unchanged if we add the gradient of a scalar field
to the vector potential:
The above transformations, which leave the E and B fields invariant, are called
gauge transformations. We are free to choose c and to be whatever we like:
i.e., we are free to choose the gauge. The most sensible gauge is the one which
make our equations as simple and symmetric as possible. This corresponds to the
choice
(r) 0 as |r| ,
(3.198)
and
The latter convention is known as the Coulomb gauge.

Taking the divergence of Eq. (3.194) and the curl of Eq. (3.195), and making
use of the Coulomb gauge, we find that the four field equations (3.186)(3.189)
can be reduced to Poissons equation written four times over:
(3.200)
2 = ,
0
2 A = 0 j.
(3.201)
Poissons equation is just about the simplest rotationally invariant second-order

partial differential equation it is possible to write. Note that 2 is clearly rotationally invariant, since it is the divergence of a gradient, and both divergence
101
and gradient are rotationally invariant. We can always construct the solution
to Poissons equation, given the boundary conditions. Furthermore, we have a
uniqueness theorem which tells us that our solution is the only possible solution.
Physically, this means that there is only one electric and magnetic field which is
consistent with a given set of stationary charges and steady currents. This sounds
like an obvious, almost trivial, statement. But there are many areas of physics
(for instance, fluid mechanics and plasma physics) where we also believe, for
physical reasons, that for a given set of boundary conditions the solution should
be unique. The problem is that in most cases when we reduce the problem to a
partial differential equation we end up with something far nastier than Poissons
equation. In general, we cannot solve this equation. In fact, we usually cannot
even prove that it possess a solution for general boundary conditions, let alone
that the solution is unique. So, we are very fortunate indeed that in electrostatics
and magnetostatics the problem boils down to solving a nice partial differential
equation. When physicists make statements to the effect that electromagnetism
is the best understood theory in physics, which they often do, what they are really saying is that the partial differential equations which crop up in this theory
are soluble and have nice properties.
Poissons equation
2 u = v
(3.202)
is linear, which means that its solutions are superposable. We can exploit this
fact to construct a general solution to this equation. Suppose that we can find the
solution to
2 G(r, r 0 ) = (r r 0 )
(3.203)
which satisfies the boundary conditions. This is the solution driven by a unit
amplitude point source located at position vector r 0 . Since any general source
can be built up out of a weighted sum of point sources, it follows that a general
solution to Poissons equation can be built up out of a weighted superposition of
point source solutions. Mathematically, we can write
Z
u(r) = G(r, r 0 ) v(r 0 ) d3 r 0 .
(3.204)
The function G is called the Greens function. The Greens function for Poissons
102
equation is
G(r, r 0 ) =
1
1
.
4 |r r 0 |
(3.205)
Note that this Greens function is proportional to the scalar potential of a point
charge located at r 0 : this is hardly surprising, given the definition of a Greens
function.
According to Eqs. (3.200), (3.201), (3.202), (3.204), and (3.205), the scalar
and vector potentials generated by a set of stationary charges and steady currents
take the form
Z
1
(r 0 ) 3 0
(r) =
dr,
(3.206)
4 0 |r r 0 |
Z
j(r 0 ) 3 0
0
dr.
(3.207)
A(r) =
4 |r r 0 |
Making use of Eqs. (3.194), (3.195), (3.206), and (3.207), we obtain the fundamental force laws for electric and magnetic fields. Coulombs law,
Z
1
r r0 3 0
0
E(r) =
(r )
dr,
(3.208)
4 0
|r r 0 |3
and the Biot-Savart law,
0
B(r) =
4
j(r 0 ) (r r 0 ) 3 0
dr.
|r r 0 |3
(3.209)
Of course, both of these laws are examples of action at a distance laws, and,
therefore, violate the theory of relativity. However, this is not a problem as long as
we restrict ourselves to fields generated by time-independent charge and current
distributions.
The question, now, is by how much is this scheme which we have just worked
out going to be disrupted when we take time variation into account. The answer, somewhat surprisingly, is by very little indeed. So, in Eqs. (3.186)(3.209)
we can already discern the basic outline of classical electromagnetism. Let us
continue our investigation.
103
4 TIME-DEPENDENT MAXWELLS EQUATIONS
Time-dependent Maxwells equations
4.1 Introduction
In this section, we shall take the time-independent set Maxwells equations, derived in the previous section, and generalize it to the full set of time-dependent
Maxwells equations.
4.2 Faradays law

The history of humankinds development of physics can be thought of as the
history of the synthesis of ideas. Physicists keep finding that apparently disparate
phenomena can be understood as different aspects of some more fundamental
phenomenon. This process has continued until today all physical phenomena can
be described in terms of three fundamental forces: gravity, the electroweak force,
and the strong force. One of the main goals of modern physics is to find some way
of combining these three forces so that all of physics can be described in terms of
a single unified force. This, essentially, is the purpose of string theory.
The first great synthesis of ideas in physics took place in 1666 when Issac
Newton realised that the force which causes apples to fall downwards is the same
as the force which maintains the planets in elliptical orbits around the Sun. The
second great synthesis, which we are about to study in more detail, took place in
1830 when Michael Faraday discovered that electricity and magnetism are two
aspects of the same thing, usually referred to as electromagnetism. The third great
synthesis, which we shall discuss presently, took place in 1873 when James Clerk
Maxwell demonstrated that light and electromagnetism are intimately related.
The last (but, hopefully, not the final) great synthesis took place in 1967 when
Steve Weinberg and Abdus Salam showed that the electromagnetic force and
the weak nuclear force (i.e., the one which is responsible for decays) can be
combined to give the electroweak force. Unfortunately, Weinbergs work lies well
beyond the scope of this lecture course.
104
4.2 Faradays law
Let us now consider Faradays experiments, having put them in their proper
historical context. Prior to 1830, the only known way to make an electric current
flow through a conducting wire was to connect the ends of the wire to the positive
and negative terminals of a battery. We measure a batterys ability to push current
down a wire in terms of its voltage, by which we mean the voltage difference
between its positive and negative terminals. What does voltage correspond to
in physics? Well, volts are the units used to measure electric scalar potential, so
when we talk about a 6V battery, what we are really saying is that the difference
in electric scalar potential between its positive and negative terminals is six volts.
This insight allows us to write
Z
Z
V = () ( ) = dl =
E dl,
(4.1)
where V is the battery voltage, denotes the positive terminal, the negative
terminal, and dl is an element of length along the wire. Of course, the above
equation is a direct consequence of E = . Clearly, a voltage difference between two ends of a wire attached to a battery implies the presence of an electric
field which pushes charges through the wire. This field is directed from the positive terminal of the battery to the negative terminal, and is, therefore, such as to
force electrons to flow through the wire from the negative to the positive terminal. As expected, this means that a net positive current flows from the positive
to the negative terminal. The fact that E is a conservative field ensures that the
voltage difference V is independent of the path of the wire. In other words, two
different wires attached to the same battery develop identical voltage differences.
Let us now consider a closed loop of wire (with no battery). The voltage
around such a loop, which is sometimes called the electromotive force or e.m.f., is
I
(4.2)
V = E dl = 0.
This is a direct consequence of the field equation E = 0. So, since E is a
conservative field then the electromotive force around a closed loop of wire is
automatically zero, and no current flows around the wire. This all seems to make
sense. However, Michael Faraday is about to throw a spanner in our works! He
discovered in 1830 that a changing magnetic field can cause a current to flow
105
4.2 Faradays law
around a closed loop of wire (in the absence of a battery). Well, if current flows
through a wire then there must be an electromotive force. So,
I
V = E dl 6= 0,
(4.3)
which immediately implies that E is not a conservative field, and that E 6= 0.
Clearly, we are going to have to modify some of our ideas regarding electric fields.
Faraday continued his experiments and found that another way of generating
an electromotive force around a loop of wire is to keep the magnetic field constant and move the loop. Eventually, Faraday was able to formulate a law which
accounted for all of his experiments. The e.m.f. generated around a loop of wire
in a magnetic field is proportional to the rate of change of the flux of the magnetic field through the loop. So, if the loop is denoted C, and S is some surface
attached to the loop, then Faradays experiments can be summed up by writing
I
Z
V = E dl = A
B dS,
(4.4)
t
C
S
where A is a constant of proportionality. Thus, the changing flux of the magnetic
field through the loop creates an electric field directed around the loop. This
process is know as magnetic induction.
S.I. units have been carefully chosen so as to make |A| = 1 in the above equation. The only thing we now have to decide is whether A = +1 or A = 1. In
other words, which way around the loop does the induced e.m.f. want to drive
the current? We possess a general principle which allows us to decide questions
like this. It is called Le Chateliers principle. According to Le Chateliers principle,
every change generates a reaction which tries to minimize the change. Essentially, this means that the Universe is stable to small perturbations. When this
principle is applied to the special case of magnetic induction, it is usually called
Lenzs law. According to Lenzs law, the current induced around a closed loop is
always such that the magnetic field it produces tries to counteract the change in
magnetic flux which generates the electromotive force. From Fig. 34, it is clear
that if the magnetic field B is increasing and the current I circulates clockwise
(as seen from above) then it generates a field B 0 which opposes the increase in
106
4.2 Faradays law
Figure 34:
magnetic flux through the loop, in accordance with Lenzs law. The direction of
the current is opposite to the sense of the current loop C (assuming that the flux
of B through the loop is positive), so this implies that A = 1 in Eq. (4.4). Thus,
Faradays law takes the form
Z
I
B dS.
(4.5)
E dl =
t S
C
H
Experimentally, Faradays law is found to correctly predict the e.m.f. (i.e., E
dl) generated in any wire loop, irrespective of the position or shape of the loop.
It is reasonable to assume that the same e.m.f. would be generated in the absence
of the wire (of course, no current would flow in this case). Thus, Eq. (4.5) is valid
for any closed loop C. If Faradays law is to make any sense then it must also be
true for any surface S attached to the loop C. Clearly, if the flux of the magnetic
field through the loop depends on the surface upon which it is evaluated then
Faradays law is going to predict different e.m.f.s for different surfaces. Since
there is no preferred surface for a general non-coplanar loop, this would
not
R
make very much sense. The condition for the flux of the magnetic field, S B dS,
to depend only on the loop C to which the surface S is attached, and not on the
nature of the surface itself, is
I
B dS 0 = 0,
(4.6)
S0
107
4.3 Electric scalar potential?
for any closed surface S 0 .

Faradays law, Eq. (4.5), can be converted into a field equation using Stokes
theorem. We obtain
B
E= .
(4.7)
t
This is the final Maxwell equation. It describes how a changing magnetic field
can generate, or induce, an electric field. Gauss theorem applied to Eq. (4.6)
yields the familiar field equation
B = 0.
(4.8)
This ensures that the magnetic flux through a loop is a well-defined quantity.
The divergence of Eq. (4.7) yields
B
= 0.
(4.9)
t
Thus, the field equation (4.7) actually demands that the divergence of the magnetic field be constant in time for self-consistency (this means that the flux of the
magnetic field through a loop need not be a well-defined quantity, as long as its
time derivative is well-defined). However, a constant non-solenoidal magnetic
field can only be generated by magnetic monopoles, and magnetic monopoles
do not exist (as far as we are aware). Hence, B = 0. The absence of magnetic monopoles is an observational fact: it cannot be predicted by any theory. If
magnetic monopoles were discovered tomorrow this would not cause physicists
any problems. We know how to generalize Maxwells equations to include both
magnetic monopoles and currents of magnetic monopoles. In this generalized
formalism, Maxwells equations are completely symmetric with respect to electric
and magnetic fields, and B 6= 0. However, an extra term (involving the current of magnetic monopoles) must be added to the right-hand side of Eq. (4.7) in
order to make it self-consistent.

Now we have a problem. We can only write the electric field in terms of a scalar
potential (i.e., E = ) provided that E = 0. However, we have just found
108
that in the presence of a changing magnetic field the curl of the electric field is
non-zero. In other words, E is not, in general, a conservative field. Does this mean
that we have to abandon the concept of electric scalar potential? Fortunately, no.
It is still possible to define a scalar potential which is physically meaningful.
Let us start from the equation
B = 0,
(4.10)
which is valid for both time-varying and non time-varying magnetic fields. Since
the magnetic field is solenoidal, we can write it as the curl of a vector potential:
B = A.
(4.11)
So, there is no problem with the vector potential in the presence of time-varying
fields. Let us substitute Eq. (4.11) into the field equation (4.7). We obtain
E=
A
,
t
which can be written
(4.12)
A
= 0.
(4.13)
E+
t
We know that a curl-free vector field can always be expressed as the gradient of
a scalar potential, so let us write
!
E+
A
= ,
t
or
(4.14)
A
.
(4.15)
t
This is a very nice equation! It tells us that the scalar potential only describes
the conservative electric field generated by electric charges. The electric field
induced by time-varying magnetic fields is non-conservative, and is described by
the magnetic vector potential A.
E =
109
4.4 Gauge transformations

Electric and magnetic fields can be written in terms of scalar and vector potentials, as follows:
E =
A
,
t
B = A.
(4.16)
(4.17)
However, this prescription is not unique. There are many different potentials
which can generate the same fields. We have come across this problem before. It
is called gauge invariance. The most general transformation which leaves the E
and B fields unchanged in Eqs. (4.16) and (4.17) is
,
t
A A .
(4.18)
(4.19)
This is clearly a generalization of the gauge transformation which we found earlier for static fields:
+ c,
A A ,
(4.20)
(4.21)
where c is a constant. In fact, if (r, t) (r) + c t then Eqs. (4.18) and (4.19)
reduce to Eqs. (4.20) and (4.21).
We are free to choose the gauge so as to make our equations as simple as
possible. As before, the most sensible gauge for the scalar potential is to make it
go to zero at infinity:
(r) 0
as |r| .
(4.22)
For steady fields, we found that the optimum gauge for the vector potential was
the so-called Coulomb gauge:
A = 0.
(4.23)
We can still use this gauge for non-steady fields. The argument which we gave
earlier (see Sect. 3.12), that it is always possible to transform away the divergence of a vector potential, remains valid. One of the nice features of the
110
Coulomb gauge is that when we write the electric field,

E =
A
,
t
(4.24)
we find that the part which is generated by charges (i.e., the first term on the
right-hand side) is conservative, and the part induced by magnetic fields (i.e.,
the second term on the right-hand side) is purely solenoidal. Earlier on, we
proved mathematically that a general vector field can be written as the sum of
a conservative field and a solenoidal field (see Sect. 3.11). Now we are finding
that when we split up the electric field in this manner the two fields have different
physical origins: the conservative part of the field emanates from electric charges,
whereas the solenoidal part is induced by magnetic fields.
Equation (4.24) can be combined with the field equation
E=
0
(4.25)
(which remains valid for non-steady fields) to give

2
= .
t
0
(4.26)
With the Coulomb gauge condition, A = 0, the above expression reduces to
2 = ,
(4.27)
0
which is just Poissons equation. Thus, we can immediately write down an expression for the scalar potential generated by non-steady fields. It is exactly the same
as our previous expression for the scalar potential generated by steady fields,
namely
Z
(r 0 , t) 3 0
1
(r, t) =
dr.
(4.28)
40 |r r 0 |
However, this apparently simple result is extremely deceptive. Equation (4.28) is
a typical action at a distance law. If the charge density changes suddenly at r 0 then
the potential at r responds immediately. However, we shall see later that the full
time-dependent Maxwells equations only allow information to propagate at the
speed of light (i.e., they do not violate relativity). How can these two statements
111
be reconciled? The crucial point is that the scalar potential cannot be measured
directly, it can only be inferred from the electric field. In the time-dependent case,
there are two parts to the electric field: that part which comes from the scalar
potential, and that part which comes from the vector potential [see Eq. (4.24) ].
So, if the scalar potential responds immediately to some distance rearrangement
of charge density it does not necessarily follow that the electric field also has an
immediate response. What actually happens is that the change in the part of the
electric field which comes from the scalar potential is balanced by an equal and
opposite change in the part which comes from the vector potential, so that the
overall electric field remains unchanged. This state of affairs persists at least until
sufficient time has elapsed for a light signal to travel from the distant charges to
the region in question. Thus, relativity is not violated, since it is the electric field,
and not the scalar potential, which carries physically accessible information.
It is clear that the apparent action at a distance nature of Eq. (4.28) is highly
misleading. This suggests, very strongly, that the Coulomb gauge is not the optimum gauge in the time-dependent case. A more sensible choice is the so called
Lorentz gauge:
.
(4.29)
A = 0 0
t
It can be shown, by analogy with earlier arguments (see Sect. 3.12), that it is
always possible to make a gauge transformation, at a given instance in time, such
that the above equation is satisfied. Substituting the Lorentz gauge condition into
Eq. (4.26), we obtain
2
0 0 2 2 = .
(4.30)
t
0
It turns out that this is a three-dimensional wave equation in which information
propagates at the speed of light. But, more of this later. Note that the magnetically induced part of the electric field (i.e., A/t) is not purely solenoidal in
the Lorentz gauge. This is a slight disadvantage of the Lorentz gauge with respect
to the Coulomb gauge. However, this disadvantage is more than offset by other
advantages which will become apparent presently. Incidentally, the fact that the
part of the electric field which we ascribe to magnetic induction changes when we
change the gauge suggests that the separation of the field into magnetically induced and charge induced components is not unique in the general time-varying
112
4.5 The displacement current
case (i.e., it is a convention).

Michael Faraday revolutionized physics in 1830 by showing that electricity and
magnetism were interrelated phenomena. He achieved this breakthrough by
careful experimentation. Between 1864 and 1873, James Clerk Maxwell achieved
a similar breakthrough by pure thought. Of course, this was only possible because
he was able to take the experimental results of Faraday, Ampère, etc., as his starting point. Prior to 1864, the laws of electromagnetism were written in integral
form. Thus, Gausss law was (in S.I. units) the flux of the electric field through
a closed surface equals the total enclosed charge, divided by 0 . The no magnetic
monopole law was the flux of the magnetic field through any closed surface is zero.
Faradays law was the electromotive force generated around a closed loop equals
minus the rate of change of the magnetic flux through the loop. Finally, Ampères
circuital law was the line integral of the magnetic field around a closed loop equals
the total current flowing through the loop, multiplied by 0 . Maxwells first great
achievement was to realize that these laws could be expressed as a set of firstorder partial differential equations. Of course, he wrote his equations out in
component form, because modern vector notation did not come into vogue until
about the time of the First World War. In modern notation, Maxwell first wrote:
E =
,
(4.31)
0
B = 0,
(4.32)
B
E = ,
(4.33)
t
B = 0 j.
(4.34)
Maxwells second great achievement was to realize that these equations are wrong.
We can see that there is something slightly unusual about Eqs. (4.31)(4.34).
They are very unfair to electric fields! After all, time-varying magnetic fields can
induce electric fields, but electric fields apparently cannot affect magnetic fields
in any way. However, there is a far more serious problem associated with the
113
above equations, which we alluded to earlier on. Consider the integral form of
the last Maxwell equation (i.e., Ampères circuital law)
Z
I
(4.35)
B dl = 0 j dS.
S
This says that the line integral of the magnetic field around a closed loop C is
equal to 0 times the flux of the current density through the loop. The problem
is that the flux of the current density through a loop is not, in general, a welldefined quantity. In order for the flux to be well-defined, the integral of j dS over
some surface S attached to a loop C must depend on C, but not on the details of
S. This is only the case if
j = 0.
(4.36)
Unfortunately, the above condition is only satisfied for non time-varying fields.
Why do we say that, in general, j 6= 0? Well, consider the flux of j out of

some closed surface S enclosing a volume V. This is clearly equivalent to the rate
at which charge flows out of S. However, if charge is a conserved quantity (and
we certainly believe that it is) then the rate at which charge flows out of S must
equal the rate of decrease of the charge contained in volume V. Thus,
I
Z
j dS =
dV.
(4.37)
t V
S
Making use of Gauss theorem, this yields
j=
.
t
(4.38)
Thus, j = 0 is only true in a steady-state (i.e., when /t 0).

The problem with Ampères circuital law is well illustrated by the following
very famous example. Consider a long straight wire interrupted by a parallel
plate capacitor. Suppose that C is some loop which circles the wire. In the non
time-dependent situation, the capacitor acts like a break in the wire, so no current flows, and no magnetic field is generated. There is clearly no problem with
Ampères law in this case. However, in the time-dependent situation, a transient
current flows in the wire as the capacitor charges up, or charges down, and so
114
a transient magnetic field is generated. Thus, the line integral of the magnetic
field around C is (transiently) non-zero. According to Ampères circuital law, the
flux of the current through any surface attached to C should also be (transiently)
non-zero. Let us consider two such surfaces. The first surface, S1 , intersects the
wire. This surface causes us no problem, since the flux of j though the surface
is clearly non-zero (because it intersects a current carrying wire). The second
surface, S2 , passes between the plates of the capacitor, and, therefore, does not
intersect the wire at all. Clearly, the flux of the current through this surface is
zero. The current fluxes through surfaces S1 and S2 are obviously different. However, both surfaces are attached to the same loop C, so the fluxes should be the
same, according to Ampères law (4.35). It would appear that Ampères circuital
law is about to disintegrate! However, we notice that although the surface S 2
does not intersect any electric current, it does pass through a region of strong
changing electric field as it threads between the plates of the charging (or discharging) capacitor. Perhaps, if we add a term involving E/t to the right-hand
side of Eq. (4.34) then we can somehow fix up Ampères circuital law? This is,
essentially, how Maxwell reasoned more than one hundred years ago.
Let us try out this scheme. Suppose that we write
B = 0 j +
E
t
(4.39)
instead of Eq. (4.34). Here, is some constant. Does this resolve our problem?
We want the flux of the right-hand side of the above equation through some loop
C to be well-defined; i.e., it should only depend on C, and not the particular
surface S (which spans C) upon which it is evaluated. This is another way of
saying that we want the divergence of the right-hand side to be zero. In fact,
we can see that this is necessary for self-consistency, since the divergence of the
left-hand side is automatically zero. So, taking the divergence of Eq. (4.39), we
obtain
E
0 = 0 j +
.
(4.40)
t
But, we know that
(4.41)
E= ,
0
115
so combining the previous two equations we arrive at

0 j +

= 0.
0 t
(4.42)
Now, our charge conservation law (4.38) can be written

j+
= 0.
t
(4.43)
The previous two equations are in agreement provided = 0 0 . So, if we modify

the final Maxwell equation such that it reads
B = 0 j + 0 0
E
,
t
(4.44)
then we find that the divergence of the right-hand side is zero as a consequence
of charge conservation. The extra term is called the displacement current (this
name was invented by Maxwell). In summary, we have shown that although the
flux of the real current through a loop is not well-defined, if we form the sum of
the real current and the displacement current then the flux of this new quantity
through a loop is well-defined.
Of course, the displacement current is not a current at all. It is, in fact, associated with the generation of magnetic fields by time-varying electric fields.
Maxwell came up with this rather curious name because many of his ideas regarding electric and magnetic fields were completely wrong. For instance, Maxwell
believed in the ther, and he thought that electric and magnetic fields were some
sort of stresses in this medium. He also thought that the displacement current was
associated with displacements of the ther (hence, the name). The reason that
these misconceptions did not invalidate his equations is quite simple. Maxwell
based his equations on the results of experiments, and he added in his extra term
so as to make these equations mathematically self-consistent. Both of these steps
are valid irrespective of the existence or non-existence of the ther.
But, hang on a minute, you might say, you cant go around adding terms
to laws of physics just because you feel like it! The field equations (4.31)(4.34)
are derived directly from the results of famous nineteenth century experiments.
If there is a new term involving the time derivative of the electric field which
116
needs to be added into these equations, how come there is no corresponding

nineteenth century experiment which demonstrates this? We have Faradays law
which shows that changing magnetic fields generate electric fields. Why is there
no Joe Bloggs law that says that changing electric fields generate magnetic
fields? This is a perfectly reasonable question. The answer is that the new term
describes an effect which is far too small to have been observed in nineteenth
century experiments. Let us demonstrate this.
First, we shall show that it is comparatively easy to detect the induction of an
electric field by a changing magnetic field in a desktop laboratory experiment.
The Earths magnetic field is about 1 gauss (that is, 104 tesla). Magnetic fields
generated by electromagnets (which will fit on a laboratory desktop) are typically
about one hundred times bigger than this. Let us, therefore, consider a hypothetical experiment in which a 100 gauss magnetic field is switched on suddenly.
Suppose that the field ramps up in one tenth of a second. What electromotive
force is generated in a 10 centimeter square loop of wire located in this field?
Faradays law is written
I
BA
B dS
,
(4.45)
V =
t
t
where B = 0.01 tesla is the field-strength, A = 0.01 m2 is the area of the loop,
and t = 0.1 seconds is the ramp time. It follows that V 1 millivolt. Well, one
millivolt is easily detectable. In fact, most hand-held laboratory voltmeters are
calibrated in millivolts. It is clear that we would have no difficulty whatsoever
detecting the magnetic induction of electric fields in a nineteenth century style
laboratory experiment.
Let us now consider the electric induction of magnetic fields. Suppose that our
electric field is generated by a parallel plate capacitor of spacing one centimeter
which is charged up to 100 volts. This gives a field of 104 volts per meter. Suppose,
further, that the capacitor is discharged in one tenth of a second. The law of
electric induction is obtained by integrating Eq. (4.44), and neglecting the first
term on the right-hand side. Thus,
I
Z
B dl = 0 0
E dS.
(4.46)
t
117
Let us consider a loop 10 centimeters square. What is the magnetic field generated around this loop (we could try to measure this with a Hall probe)? Very
approximately, we find that
E l2
l B 0 0
,
(4.47)
t
where l = 0.1 meters is the dimensions of the loop, B is the magnetic fieldstrength, E = 104 volts per meter is the electric field, and t = 0.1 seconds is the
decay time of the field. We find that B 109 gauss. Modern technology is unable
to detect such a small magnetic field, so we cannot really blame Faraday for not
noticing electric induction in 1830.
So, you might say, why did you bother mentioning this displacement current thing in the first place if it is undetectable? Again, a perfectly fair question.
The answer is that the displacement current is detectable in some experiments.
Suppose that we take an FM radio signal, amplify it so that its peak voltage is one
hundred volts, and then apply it to the parallel plate capacitor in the previous
hypothetical experiment. What size of magnetic field would this generate? Well,
a typical FM signal oscillates at 109 Hz, so t in the previous example changes
from 0.1 seconds to 109 seconds. Thus, the induced magnetic field is about 101
gauss. This is certainly detectable by modern technology. So, it would seem that
if the electric field is oscillating fast then electric induction of magnetic fields is an
observable effect. In fact, there is a virtually infallible rule for deciding whether
or not the displacement current can be neglected in Eq. (4.44). If electromagnetic
radiation is important then the displacement current must be included. On the
other hand, if electromagnetic radiation is unimportant then the displacement
current can be safely neglected. Clearly, Maxwells inclusion of the displacement
current in Eq. (4.44) was a vital step in his later realization that his equations
allowed propagating wave-like solutions. These solutions are, of course, electromagnetic waves. But, more of this later.
We are now in a position to write out Maxwells equations in all their glory!
We get
E =
,
(4.48)
0
B = 0,
(4.49)
118
E =
4.6 Potential formulation
B
,
t
(4.50)
B = 0 j + 0 0
E
.
t
(4.51)
These four partial differential equations constitute a complete description of the

behaviour of electric and magnetic fields. The first equation describes how electric fields are induced by charges. The second equation says that there is no such
thing as a magnetic monopole. The third equation describes the induction of
electric fields by changing magnetic fields, and the fourth equation describes the
generation of magnetic fields by electric currents and the induction of magnetic
fields by changing electric fields. Note that with the inclusion of the displacement
current these equations treat electric and magnetic fields on an equal footing:
i.e., electric fields can induce magnetic fields, and vice versa. Equations (4.48)
(4.51) sum up the experimental results of Coulomb, Ampère, and Faraday very
succinctly: they are called Maxwells equations because James Clerk Maxwell was
the first to write them down (in component form). Maxwell also fixed them up
so that they made mathematical sense.
4.6 Potential formulation

We have seen that Eqs. (4.49) and (4.50) are automatically satisfied if we write
the electric and magnetic fields in terms of potentials:
E =
A
,
t
B = A.
(4.52)
(4.53)
This prescription is not unique, but we can make it unique by adopting the following conventions:
(r) 0 as |r| ,
A = 0 0
.
t
119
(4.54)
(4.55)
4.7 Electromagnetic waves
The above equations can be combined with Eq. (4.48) to give

2
0 0 2 2 = .
t
0
(4.56)
Let us now consider Eq. (4.51). Substitution of Eqs. (4.52) and (4.53) into
this formula yields
A ( A) 2 A = 0 j 0 0
2 A

0 0 2 ,
t
t
(4.57)
or
2 A
0 0 2 2 A = 0 j A + 0 0
.
(4.58)
t
t
We can now see quite clearly where the Lorentz gauge condition (4.29) comes
from. The above equation is, in general, very complicated, since it involves both
the vector and scalar potentials. But, if we adopt the Lorentz gauge, then the last
term on the right-hand side becomes zero, and the equation simplifies considerably, such that it only involves the vector potential. Thus, we find that Maxwells
equations reduce to the following:
!
0 0 2 2 =
,
t
0
0 0
2 A
2 A = 0 j.
2
t
(4.59)
(4.60)
This is the same (scalar) equation written four times over. In steady-state (i.e.,
/t = 0), it reduces to Poissons equation, which we know how to solve. With
the 2 /t2 terms included, it becomes a slightly more complicated equation (in
fact, a driven three-dimensional wave equation).

This is an appropriate point at which to demonstrate that Maxwells equations
possess propagating wave-like solutions. Let us start from Maxwells equations in
120
free space (i.e., with no charges and no currents):

E = 0,
(4.61)
B = 0,
(4.62)
E =
(4.63)
B
,
t
E
B = 0 0
.
t
(4.64)
Note that these equations exhibit a nice symmetry between the electric and magnetic fields.
There is an easy way to show that the above equations possess wave-like solutions, and a hard way. The easy way is to assume that the solutions are going
to be wave-like beforehand. Specifically, let us search for plane-wave solutions of
the form:
E(r, t) = E0 cos (k r t),
B(r, t) = B0 cos (k r t + ).
(4.65)
(4.66)
Here, E0 and B0 are constant vectors, k is called the wave-vector, and is the
angular frequency. The frequency in hertz, f, is related to the angular frequency
via = 2 f. The frequency is conventionally defined to be positive. The quantity
is a phase difference between the electric and magnetic fields. Actually, it is
more convenient to write
E = E0 e i (kr t) ,
(4.67)
B = B0 e i (kr t) ,
(4.68)
where, by convention, the physical solution is the real part of the above equations.
The phase difference is absorbed into the constant vector B0 by allowing it to
become complex. Thus, B0 B0 e i . In general, the vector E0 is also complex.
A wave maximum of the electric field satisfies
k r = t + n 2 + ,
121
(4.69)
where n is an integer and is some phase angle. The solution to this equation
is a set of equally spaced parallel planes (one plane for each possible value of n),
whose normals lie in the direction of the wave-vector k, and which propagate in
this direction with phase-velocity
v=
.
k
(4.70)
The spacing between adjacent planes (i.e., the wave-length) is given by

=
2
k
(4.71)
(see Fig. 35).
Figure 35:
Consider a general plane-wave vector field

A = A0 e i (kr t) .
(4.72)
What is the divergence of A? This is easy to evaluate. We have

Ax Ay Az
+
+
= (A0x i kx + A0y i ky + A0z i kz ) e i (krt)
x
y
z
= i k A.
(4.73)
A =
122
How about the curl of A? This is slightly more difficult. We have

Az Ay
= (i ky Az i kz Ay )
y
z
= i (k A)x .
( A)x =
(4.74)
This is easily generalized to

A = i k A.
(4.75)
We can see that vector field operations on a plane-wave simplify to replacing the
operator with i k.
The first Maxwell equation (4.61) reduces to
i k E0 = 0,
(4.76)
using the assumed electric and magnetic fields (4.67) and (4.68), and Eq. (4.73).
Thus, the electric field is perpendicular to the direction of propagation of the
wave. Likewise, the second Maxwell equation gives
i k B0 = 0,
(4.77)
implying that the magnetic field is also perpendicular to the direction of propagation. Clearly, the wave-like solutions of Maxwells equation are a type of transverse wave. The third Maxwell equation gives
i k E0 = i B0 ,
(4.78)
where use has been made of Eq. (4.75). Dotting this equation with E0 yields
E0 k E 0
= 0.
(4.79)
Thus, the electric and magnetic fields are mutually perpendicular. Dotting equation (4.78) with B0 yields
E0 B 0 =
B0 k E0 = B02 > 0.
(4.80)
Thus, the vectors E0 , B0 , and k are mutually perpendicular, and form a righthanded set. The final Maxwell equation gives
i k B0 = i 0 0 E0 .
123
(4.81)
Combining this with Eq. (4.78) yields

k (k E0 ) = (k E0 ) k k2 E0 = k2 E0 = 0 0 2 E0 ,
(4.82)
or
k 2 = 0 0 2 ,
(4.83)
where use has been made of Eq. (4.76). However, we know from Eq. (4.70)
that the phase-velocity c is related to the magnitude of the wave-vector and the
angular wave frequency via c = /k. Thus, we obtain
c=
1
.
0 0
(4.84)
So, we have found transverse wave solutions of the free-space Maxwell equations, propagating at some phase-velocity c, which is given by a combination of
0 and 0 . The constants 0 and 0 are easily measurable. The former is related
to the force acting between stationary electric charges, and the latter to the force
acting between steady electric currents. Both of these constants were fairly wellknown in Maxwells time. Maxwell, incidentally, was the first person to look for
wave-like solutions of his equations, and, thus, to derive Eq. (4.84). The modern
values of 0 and 0 are
0 = 8.8542 1012 C2 N1 m2 ,
0 = 4 107 N A2 .
(4.85)
(4.86)
Let us use these values to find the phase-velocity of electromagnetic waves. We

obtain
1
c=
= 2.998 108 m s1 .
(4.87)
0 0
Of course, we immediately recognize this as the velocity of light. Maxwell also
made this connection back in the 1870s. He conjectured that light, whose nature
had previously been unknown, was a form of electromagnetic radiation. This
was a remarkable prediction. After all, Maxwells equations were derived from
the results of benchtop laboratory experiments, involving charges, batteries, coils,
and currents, which apparently had nothing whatsoever to do with light.
124
Radiation Type Wave-length Range (m)

Gamma Rays
< 1011
X-Rays
1011 109
Ultraviolet
109 107
Visible
107 106
Infrared
106 104
Microwave
104 101
TV-FM
101 101
Radio
> 101
Table 1: The electromagnetic spectrum
Maxwell was able to make another remarkable prediction. The wave-length of

light was well-known in the late nineteenth century from studies of diffraction
through slits, etc. Visible light actually occupies a surprisingly narrow wavelength range. The shortest wave-length blue light which is visible has = 0.4
microns (one micron is 106 meters). The longest wave-length red light which is
visible has = 0.76 microns. However, there is nothing in our analysis which suggests that this particular range of wave-lengths is special. Electromagnetic waves
can have any wave-length. Maxwell concluded that visible light was a small part
of a vast spectrum of previously undiscovered types of electromagnetic radiation.
Since Maxwells time, virtually all of the non-visible parts of the electromagnetic
spectrum have been observed. Table 1 gives a brief guide to the electromagnetic
spectrum. Electromagnetic waves are of particular importance because they are
our only source of information regarding the universe around us. Radio waves
and microwaves (which are comparatively hard to scatter) have provided much
of our knowledge about the centre of our own galaxy. This is completely unobservable in visible light, which is strongly scattered by interstellar gas and dust
lying in the galactic plane. For the same reason, the spiral arms of our galaxy can
only be mapped out using radio waves. Infrared radiation is useful for detecting
proto-stars, which are not yet hot enough to emit visible radiation. Of course,
visible radiation is still the mainstay of astronomy. Satellite based ultraviolet
observations have yielded invaluable insights into the structure and distribution
of distant galaxies. Finally, X-ray and -ray astronomy usually concentrates on
exotic objects in the Galaxy, such as pulsars and supernova remnants.
125
Equations (4.76), (4.78), and the relation c = /k, imply that

B0 =
E0
.
c
(4.88)
Thus, the magnetic field associated with an electromagnetic wave is smaller in

magnitude than the electric field by a factor c. Consider a free charge interacting
with an electromagnetic wave. The force exerted on the charge is given by the
Lorentz formula
f = q (E + v B).
(4.89)
The ratio of the electric and magnetic forces is
fmagnetic v B0 v
.
felectric
E0
c
(4.90)
So, unless the charge is relativistic, the electric force greatly exceeds the magnetic force. Clearly, in most terrestrial situations electromagnetic waves are an
essentially electric phenomenon (as far as their interaction with matter goes). For
this reason, electromagnetic waves are usually characterized by their wave-vector
(which specifies the direction of propagation and the wave-length) and the plane
of polarization (i.e., the plane of oscillation) of the associated electric field. For
a given wave-vector k, the electric field can have any direction in the plane normal to k. However, there are only two independent directions in a plane (i.e., we
can only define two linearly independent vectors in a plane). This implies that
there are only two independent polarizations of an electromagnetic wave, once
its direction of propagation is specified.
Let us now derive the velocity of light from Maxwells equation the hard way.
Suppose that we take the curl of the fourth Maxwell equation, Eq. (4.64). We
obtain
B = ( B) 2 B = 2 B = 0 0
E
.
t
(4.91)
Here, we have used the fact that B = 0. The third Maxwell equation,
Eq. (4.63), yields
1
2
B = 0,
(4.92)
2
2
c t
126
where use has been made of Eq. (4.87). A similar equation can obtained for the
electric field by taking the curl of Eq. (4.63):
1 2
2
E = 0,
c2 t2
(4.93)
We have found that electric and magnetic fields both satisfy equations of the
form
1
A = 0
2
(4.94)
2
2
c t
in free space. As is easily verified, the most general solution to this equation (with
a positive frequency) is
Ax = Fx (k r k c t),
(4.95)
Az = Fz (k r k c t),
(4.97)
Ay = Fy (k r k c t),
(4.96)
where Fx (), Fy (), and Fz () are one-dimensional scalar functions. Looking

along the direction of the wave-vector, so that r = (k/k) r, we find that
Ax = Fx [ k (r c t) ],
(4.98)
Ay = Fy [ k (r c t) ],
(4.99)
Az = Fz [ k (r c t) ].
(4.100)
The x-component of this solution is shown schematically in Fig. 36. It clearly

propagates in r with velocity c. If we look along a direction which is perpendicular to k then k r = 0, and there is no propagation. Thus, the components of
A are arbitrarily shaped pulses which propagate, without changing shape, along
the direction of k with velocity c. These pulses can be related to the sinusoidal
plane-wave solutions which we found earlier by Fourier transformation. Thus,
any arbitrary shaped pulse propagating in the direction of k with velocity c can
be broken down into lots of sinusoidal oscillations propagating in the same direction with the same velocity.
The operator
1 2
2 2
c t
2
127
(4.101)
Fx (r, t=0)
4.8 Greens functions
Fx (r, t=t)
ct
Figure 36:
is called the dAlembertian. It is the four-dimensional equivalent of the Laplacian. Recall that the Laplacian is invariant under rotational transformation. The
dAlembertian goes one better than this, since it is both rotationally invariant
and Lorentz invariant. The dAlembertian is conventionally denoted 2 2 . Thus,
electromagnetic waves in free space satisfy the wave equations
22 E = 0,
(4.102)
22 B = 0.
(4.103)
When written in terms of the vector and scalar potentials, Maxwells equations
reduce to
22 = ,
(4.104)
0
22 A = 0 j.
(4.105)
These are clearly driven wave equations. Our next task is to find the solutions to
these equations.

Earlier on in this lecture course, we had to solve Poissons equation
2 u = v,
128
(4.106)
where v(r) is denoted the source function. The potential u(r) satisfies the boundary condition
u(r, t) 0
as |r| ,
(4.107)
provided that the source function is reasonably localized. The solutions to Poissons equation are superposable (because the equation is linear). This property
is exploited in the Greens function method of solving this equation. The Greens
function G(r, r 0 ) is the potential, which satisfies the appropriate boundary conditions, generated by a unit amplitude point source located at r 0 . Thus,
2 G(r, r 0 ) = (r r 0 ).
(4.108)
Any source function v(r) can be represented as a weighted sum of point sources
Z
v(r) = (r r 0 ) v(r 0 ) d3 r 0 .
(4.109)
It follows from superposability that the potential generated by the source v(r) can
be written as the weighted sum of point source driven potentials (i.e., Greens
functions)
Z
u(r) = G(r, r 0 ) v(r 0 ) d3 r 0 .
(4.110)
We found earlier that the Greens function for Poissons equation is
G(r, r 0 ) =
1
1
.
4 |r r 0 |
It follows that the general solution to Eq. (4.106) is written

Z
v(r 0 ) 3 0
1
dr.
u(r) =
4 |r r 0 |
(4.111)
(4.112)
Note that the point source driven potential (4.111) is perfectly sensible. It is
spherically symmetric about the source, and falls off smoothly with increasing
distance from the source.
We now need to solve the wave equation
1 2
2
u = v,
c2 t2
129
(4.113)
where v(r, t) is a time-varying source function. The potential u(r, t) satisfies the
boundary conditions
as |r| and |t| .
u(r) 0
(4.114)
The solutions to Eq. (4.113) are superposable (since the equation is linear), so a
Greens function method of solution is again appropriate. The Greens function
G(r, r 0 ; t, t 0 ) is the potential generated by a point impulse located at position r 0
and applied at time t 0 . Thus,
1 2

G(r, r 0 ; t, t 0 ) = (r r 0 ) (t t 0 ).
2
2
c t
(4.115)
Of course, the Greens function must satisfy the correct boundary conditions. A
general source v(r, t) can be built up from a weighted sum of point impulses
ZZ
v(r, t) =
(r r 0 ) (t t 0 ) v(r 0 , t 0 ) d3 r 0 dt 0 .
(4.116)
It follows that the potential generated by v(r, t) can be written as the weighted
sum of point impulse driven potentials
ZZ
u(r, t) =
G(r, r 0 ; t, t 0 ) v(r 0 , t 0 ) d3 r 0 dt 0 .
(4.117)
So, how do we find the Greens function?
Consider
F(t t 0 |r r 0 |/c)
,
G(r, r ; t, t ) =
|r r 0 |
0
(4.118)
where F() is a general scalar function. Let us try to prove the following theorem:
1 2
2
G = 4 F(t t 0 ) (r r 0 ).
2
2
c t
(4.119)
1 2
2
G = 0.
c2 t2
(4.120)
At a general point, r 6= r 0 , the above expression reduces to
130
So, we basically have to show that G is a valid solution of the free space wave
equation. We can easily show that
x x0
|r r 0 |
=
.
x
|r r 0 |
(4.121)
It follows by simple differentiation that

2 G
=
x2
3(x x 0 )2 |r r 0 |2
F
|r r 0 |5
3(x x 0 )2 |r r 0 |2 F 0 (x x 0 )2 F 00
+
+
,
|r r 0 |4
c
|r r 0 |3 c2
(4.122)
F 00
1 2 G
2 G 2 G 2 G
+
+
=
=
,
x2
y2
z2
|r r 0 | c2
c2 t2
(4.123)
where F 0 () = dF()/d. We can derive analogous equations for 2 G/y2 and

2 G/z2 . Thus,
2 G =
giving
1 2

G = 0,
c2 t2
(4.124)
which is the desired result. Consider, now, the region around r = r 0 . It is clear
from Eq. (4.122) that the dominant term on the right-hand side as |r r 0 |
0 is the first one, which is essentially F 2 (|r r 0 |1 )/x2 . It is also clear that
(1/c2 )(2 G/t2 ) is negligible compared to this term. Thus, as |r r 0 | 0 we find
that
1
1
G F(t t 0 ) 2
2
.
(4.125)
2
2
c t
|r r 0 |
However, according to Eqs. (4.108) and (4.111)
1
= 4 (r r 0 ).
2
0
|r r |
(4.126)
1 2
2
G = 4 F(t t 0 ) (r r 0 ),
2
2
c t
(4.127)
We conclude that
131
which is the desired result.

Let us now make the special choice
F() =
()
.
4
(4.128)

1 2
2
G = (r r 0 ) (t t 0 ).
2
2
c t
(4.129)
1 (t t 0 |r r 0 |/c)
G(r, r ; t, t ) =
4
|r r 0 |
(4.130)
Thus,
is the Greens function for the driven wave equation (4.113).

The time-dependent Greens function (4.130) is the same as the steady-state
Greens function (4.111), apart from the delta-function appearing in the former.
What does this delta-function do? Well, consider an observer at point r. Because
of the delta-function, our observer only measures a non-zero potential at one
particular time
|r r 0 |
.
(4.131)
t = t0 +
c
It is clear that this is the time the impulse was applied at position r 0 (i.e., t 0 ) plus
the time taken for a light signal to travel between points r 0 and r. At time t > t 0 ,
the locus of all the points at which the potential is non-zero is
|r r 0 | = c (t t 0 ).
(4.132)
In other words, it is a sphere centred on r 0 whose radius is the distance traveled

by light in the time interval since the impulse was applied at position r 0 . Thus,
the Greens function (4.130) describes a spherical wave which emanates from
position r 0 at time t 0 and propagates at the speed of light. The amplitude of the
wave is inversely proportional to the distance from the source.
132
4.9 Retarded potentials

We are now in a position to solve Maxwells equations. Recall that in steady-state,
Maxwells equations reduce to
(4.133)
2 = ,
0
2 A = 0 j.
(4.134)
The solutions to these equations are easily found using the Greens function for
Poissons equation (4.111):
Z
1
(r 0 ) 3 0
(r) =
dr
(4.135)
4 0 |r r 0 |
Z
j(r 0 ) 3 0
0
dr.
(4.136)
A(r) =
4 |r r 0 |
The time-dependent Maxwell equations reduce to
22 = ,
0
22 A = 0 j.
(4.137)
(4.138)
We can solve these equations using the time-dependent Greens function (4.130).
From Eq. (4.117) we find that
ZZ
1
(t t 0 |r r 0 |/c) (r 0 , t 0 ) 3 0 0
(r, t) =
d r dt ,
(4.139)
4 0
|r r 0 |
with a similar equation for A. Using the well-known property of delta-functions,

these equations reduce to
Z
(r 0 , t |r r 0 |/c) 3 0
1
dr
(4.140)
(r, t) =
4 0
|r r 0 |
Z
0 j(r 0 , t |r r 0 |/c) 3 0
A(r, t) =
dr.
(4.141)
4
|r r 0 |
These are the general solutions to Maxwells equations. Note that the timedependent solutions, (4.140) and (4.141), are the same as the steady-state solutions, (4.135) and (4.136), apart from the weird way in which time appears in
133
the former. According to Eqs. (4.140) and (4.141), if we want to work out the
potentials at position r and time t then we have to perform integrals of the charge
density and current density over all space (just like in the steady-state situation).
However, when we calculate the contribution of charges and currents at position
r 0 to these integrals we do not use the values at time t, instead we use the values
at some earlier time t |r r 0 |/c. What is this earlier time? It is simply the latest
time at which a light signal emitted from position r 0 would be received at position
r before time t. This is called the retarded time. Likewise, the potentials (4.140)
and (4.141) are called retarded potentials. It is often useful to adopt the following
notation
A(r 0 , t |r r 0 |/c) [A(r 0 , t)] .
(4.142)
The square brackets denote retardation (i.e., using the retarded time instead of
the real time). Using this notation Eqs. (4.140) and (4.141), become
Z
[(r 0 )] 3 0
1
dr,
(4.143)
(r) =
4 0 |r r 0 |
Z
0 [j(r 0 )] 3 0
A(r) =
dr.
(4.144)
4 |r r 0 |
The time dependence in the above equations is taken as read.
We are now in a position to understand electromagnetism at its most fundamental level. A charge distribution (r, t) can be thought of as built up out of
a collection, or series, of charges which instantaneously come into existence, at
some point r 0 and some time t 0 , and then disappear again. Mathematically, this
is written
ZZ
(r, t) =
(r r 0 )(t t 0 ) (r 0 , t 0 ) d3 r 0 dt 0 .
(4.145)
Likewise, we can think of a current distribution j(r, t) as built up out of a collection or series of currents which instantaneously appear and then disappear:
ZZ
j(r, t) =
(r r 0 )(t t 0 ) j(r 0 , t 0 ) d3 r 0 dt 0 .
(4.146)
Each of these ephemeral charges and currents excites a spherical wave in the
appropriate potential. Thus, the charge density at r 0 and t 0 sends out a wave in
134
the scalar potential:

(r, t) =
(r 0 , t 0 ) (t t 0 |r r 0 |/c)
.
4 0
|r r 0 |
(4.147)
Likewise, the current density at r 0 and t 0 sends out a wave in the vector potential:
0 j(r 0 , t 0 ) (t t 0 |r r 0 |/c)
A(r, t) =
.
4
|r r 0 |
(4.148)
These waves can be thought of as messengers which inform other charges and
currents about the charges and currents present at position r 0 and time t 0 . However, these messengers travel at a finite speed: i.e., the speed of light. So, by the
time they reach other charges and currents their message is a little out of date. Every charge and every current in the Universe emits these spherical waves. The resultant scalar and vector potential fields are given by Eqs. (4.143) and (4.144). Of
course, we can turn these fields into electric and magnetic fields using Eqs. (4.52)
and (4.53). We can then evaluate the force exerted on charges using the Lorentz
formula. We can see that we have now escaped from the apparent action at a
distance nature of Coulombs law and the Biot-Savart law. Electromagnetic information is carried by spherical waves in the vector and scalar potentials, and,
therefore, travels at the velocity of light. Thus, if we change the position of a
charge then a distant charge can only respond after a time delay sufficient for a
spherical wave to propagate from the former to the latter charge.
Let us compare the steady-state law
1
(r) =
4 0
(r 0 ) 3 0
dr
|r r 0 |
with the corresponding time-dependent law

Z
1
[(r 0 )] 3 0
(r) =
dr
4 0 |r r 0 |
(4.149)
(4.150)
These two formulae look very similar indeed, but there is an important difference. We can imagine (rather pictorially) that every charge in the Universe is
continuously performing the integral (4.150), and is also performing a similar
integral to find the vector potential. After evaluating both potentials, the charge
135
can calculate the fields, and, using the Lorentz force law, it can then work out
its equation of motion. The problem is that the information the charge receives
from the rest of the Universe is carried by our spherical waves, and is always
slightly out of date (because the waves travel at a finite speed). As the charge
considers more and more distant charges or currents, its information gets more
and more out of date. (Similarly, when astronomers look out to more and more
distant galaxies in the Universe, they are also looking backwards in time. In fact,
the light we receive from the most distant observable galaxies was emitted when
the Universe was only about one third of its present age.) So, what does our
electron do? It simply uses the most up to date information about distant charges
and currents which it possesses. So, instead of incorporating the charge density
(r, t) in its integral, the electron uses the retarded charge density [(r, t)] (i.e.,
the density evaluated at the retarded time). This is effectively what Eq. (4.150)
says.
Consider a thought experiment in which a charge q appears at position r 0 at
time t1 , persists for a while, and then disappears at time t2 . What is the electric
field generated by such a charge? Using Eq. (4.150), we find that
(r) =
1
q
4 0 |r r0 |
= 0
for t1 t |r r0 |/c t2
otherwise.
(4.151)
Now, E = (since there are no currents, and therefore no vector potential is

generated), so
E(r) =
q r r0
4 0 |r r0 |3
for t1 t |r r0 |/c t2
otherwise.
= 0
(4.152)
This solution is shown pictorially in Fig. 37. We can see that the charge effectively
emits a Coulomb electric field which propagates radially away from the charge
at the speed of light. Likewise, it is easy to show that a current carrying wire
effectively emits an Ampèrian magnetic field at the speed of light.
We can now appreciate the essential difference between time-dependent electromagnetism and the action at a distance laws of Coulomb and Biot & Savart.
136
t < t1
t1 < t < t2
t > t2
Figure 37:
In the latter theories, the field-lines act rather like rigid wires attached to charges
(or circulating around currents). If the charges (or currents) move then so do the
field-lines, leading inevitably to unphysical action at a distance type behaviour.
In the time-dependent theory, charges act rather like water sprinklers: i.e., they
spray out the Coulomb field in all directions at the speed of light. Similarly, current carrying wires throw out magnetic field loops at the speed of light. If we
move a charge (or current) then field-lines emitted beforehand are not affected,
so the field at a distant charge (or current) only responds to the change in position after a time delay sufficient for the field to propagate between the two
charges (or currents) at the speed of light.
In Coulombs law and the Biot-Savart law, it is not entirely obvious that the
electric and magnetic fields have a real existence. After all, the only measurable
quantities are the forces acting between charges and currents. We can describe
the force acting on a given charge or current, due to the other charges and currents in the Universe, in terms of the local electric and magnetic fields, but we
have no way of knowing whether these fields persist when the charge or current
is not present (i.e., we could argue that electric and magnetic fields are just a
convenient way of calculating forces, but, in reality, the forces are transmitted
directly between charges and currents by some form of magic). However, it is
patently obvious that electric and magnetic fields have a real existence in the
time-dependent theory. Consider the following thought experiment. Suppose
137
Figure 38:
that a charge q1 comes into existence for a period of time, emits a Coulomb field,
and then disappears. Suppose that a distant charge q2 interacts with this field,
but is sufficiently far from the first charge that by the time the field arrives the
first charge has already disappeared. The force exerted on the second charge is
only ascribable to the electric field: it cannot be ascribed to the first charge, because this charge no longer exists by the time the force is exerted. The electric
field clearly transmits energy and momentum between the two charges. Anything which possesses energy and momentum is real in a physical sense. Later
on in this course, we shall demonstrate that electric and magnetic fields conserve
energy and momentum.
Let us now consider a moving charge. Such a charge is continually emitting
spherical waves in the scalar potential, and the resulting wavefront pattern is
sketched in Fig. 38. Clearly, the wavefronts are more closely spaced in front of
the charge than they are behind it, suggesting that the electric field in front is
larger than the field behind. In a medium, such as water or air, where waves
travel at a finite speed, c (say), it is possible to get a very interesting effect if the
wave source travels at some velocity v which exceeds the wave speed. This is
illustrated in Fig. 39.
The locus of the outermost wave front is now a cone instead of a sphere. The
wave intensity on the cone is extremely large: this is a shock wave! The halfangle of the shock wave cone is simply sin1 (c/v). In water, shock waves are
produced by fast moving boats. We call these bow waves. In air, shock waves are
138
ct
vt
Figure 39:
produced by speeding bullets and supersonic jets. In the latter case, we call these
sonic booms. Is there any such thing as an electromagnetic shock wave? At first
sight, the answer to this question would appear to be, no. After all, electromagnetic waves travel at the speed of light, and no wave source (i.e., an electrically
charged particle) can travel faster than this velocity. This is a rather disappointing conclusion. However, when an electromagnetic wave travels through matter
a remarkable thing happens. The oscillating electric field of the wave induces a
slight separation of the positive and negative charges in the atoms which make up
the material. We call separated positive and negative charges an electric dipole.
Of course, the atomic dipoles oscillate in sympathy with the field which induces
them. However, an oscillating electric dipole radiates electromagnetic waves.
Amazingly, when we add the original wave to these induced waves, it is exactly
as if the original wave propagates through the material in question at a velocity
which is slower than the velocity of light in vacuum. Suppose, now, that we shoot
a charged particle through the material faster than the slowed down velocity of
electromagnetic waves. This is possible since the waves are traveling slower than
the velocity of light in vacuum. In practice, the particle has to be traveling pretty
close to the velocity of light in vacuum (i.e., it has to be relativistic), but modern
particle accelerators produce copious amounts of such particles. Now, we can get
an electromagnetic shock wave. We expect an intense cone of emission, just like
the bow wave produced by a fast ship. In fact, this type of radiation has been
139
4.10 Advanced potentials?
observed. It is called Cherenkov radiation, and it is very useful in high energy

physics. Cherenkov radiation is typically produced by surrounding a particle accelerator with perspex blocks. Relativistic charged particles emanating from the
accelerator pass through the perspex traveling faster than the local velocity of
light, and therefore emit Cherenkov radiation. We know the velocity of light (c ,
say) in perspex (this can be worked out from the refractive index), so if we can
measure the half angle of the radiation cone emitted by each particle then we
can evaluate the speed of the particle v via the geometric relation sin = c /v.
We have defined the retarded time
tr = t |r r 0 |/c
(4.153)
as the latest time at which a light signal emitted from position r 0 would reach position r before time t. We have also shown that a solution to Maxwells equations
can be written in terms of retarded potentials:
Z
(r 0 , tr ) 3 0
1
dr,
(4.154)
(r, t) =
4 0 |r r 0 |
etc. But, is this the most general solution? Suppose that we define the advanced
time.
ta = t + |r r 0 |/c.
(4.155)
This is the time a light signal emitted at time t from position r would reach
position r 0 . It turns out that we can also write a solution to Maxwells equations
in terms of advanced potentials:
Z
1
(r 0 , ta ) 3 0
(r, t) =
dr,
(4.156)
4 0 |r r 0 |
etc. In fact, this is just as good a solution to Maxwells equation as the one involving retarded potentials. To get some idea what is going on, let us examine the
Greens function corresponding to our retarded potential solution:
(r 0 , t 0 ) (t t 0 |r r 0 |/c)
,
(r, t) =
4 0
|r r 0 |
140
(4.157)
with a similar equation for the vector potential. This says that the charge density
present at position r 0 and time t 0 emits a spherical wave in the scalar potential
which propagates forwards in time. The Greens function corresponding to our
advanced potential solution is
(r 0 , t 0 ) (t t 0 + |r r 0 |/c)
(r, t) =
.
4 0
|r r 0 |
(4.158)
This says that the charge density present at position r 0 and time t 0 emits a spherical wave in the scalar potential which propagates backwards in time. But, hang
on a minute, you might say, everybody knows that electromagnetic waves cant
travel backwards in time. If they did then causality would be violated. Well, you
know that electromagnetic waves do not propagate backwards in time, I know
that electromagnetic waves do not propagate backwards in time, but the question is do Maxwells equations know this? Consider the wave equation for the
scalar potential:
2
1
2

=
.
(4.159)
2
2
c t
0
This equation is manifestly symmetric in time (i.e., it is invariant under the transformation t t). Thus, backward traveling waves are just as good a solution to
this equation as forward traveling waves. The equation is also symmetric in space
(i.e., it is invariant under the transformation x x). So, why do we adopt the
Greens function (4.157) which is symmetric in space (i.e., it is invariant under
x x) but asymmetric in time (i.e., it is not invariant under t t)? Would
it not be better to use the completely symmetric Greens function
(r 0 , t 0 ) 1 (t t 0 |r r 0 |/c) (t t 0 + |r r 0 |/c)
(r, t) =
+
?
4 0 2
|r r 0 |
|r r 0 |
(4.160)
In other words, a charge emits half of its waves running forwards in time (i.e.,
retarded waves), and the other half running backwards in time (i.e., advanced
waves). This sounds completely crazy! However, in the 1940s Richard P. Feynman and John A. Wheeler pointed out that under certain circumstances this prescription gives the right answer. Consider a charge interacting with the rest of
the Universe, where the rest of the Universe denotes all of the distant charges
in the Universe, and is, by implication, an awful long way away from our original
141
charge. Suppose that the rest of the Universe is a perfect reflector of advanced
waves and a perfect absorber of retarded waves. The waves emitted by the charge
can be written schematically as
1
1
F = (retarded) + (advanced).
2
2
(4.161)
The response of the rest of the universe is written

1
1
R = (retarded) (advanced).
2
2
(4.162)
This is illustrated in the space-time diagram Fig. 40. Here, A and R denote the
advanced and retarded waves emitted by the charge, respectively. The advanced
wave travels to the rest of the Universe and is reflected: i.e., the distant charges
oscillate in response to the advanced wave and emit a retarded wave a, as shown.
The retarded wave a is spherical wave which converges on the original charge,
passes through the charge, and then diverges again. The divergent wave is denoted aa. Note that a looks like a negative advanced wave emitted by the charge,
whereas aa looks like a positive retarded wave emitted by the charge. This is essentially what Eq. (4.162) says. The retarded waves R and aa are absorbed by
the rest of the Universe.
charge
time
rest of universe
A
a
aa
R
space
Figure 40:
142
4.11 Retarded fields
If we add the waves emitted by the charge to the response of the rest of the
Universe we obtain
F 0 = F + R = (retarded).
(4.163)
Thus, charges appear to emit only retarded waves, which agrees with our everyday experience. Clearly, in this model we have side-stepped the problem of a
time asymmetric Greens function by adopting time asymmetric boundary conditions to the Universe: i.e., the distant charges in the Universe absorb retarded
waves and reflect advanced waves. This is possible because the absorption takes
place at the end of the Universe (i.e., at the big crunch, or whatever) and the
reflection takes place at the beginning of the Universe (i.e., at the big bang). It
is quite plausible that the state of the Universe (and, hence, its interaction with
electromagnetic waves) is completely different at these two times. It should be
pointed out that the Feynman-Wheeler model runs into trouble when one tries to
combine electromagnetism with quantum mechanics. These difficulties have yet
to be resolved, so at present the status of this model is that it is an interesting
idea, but it is still not fully accepted into the canon of physics.

We know the solution to Maxwells equations in terms of retarded potentials. Let
us now construct the associated electric and magnetic fields using
E =
B = A.
A
,
t
(4.164)
(4.165)
It is helpful to write
R = r r 0,
(4.166)
where R = |rr 0 |. The retarded time becomes tr = tR/c, and a general retarded
quantity is written [F(r, t)] F(r, tr ). Thus, we can write the retarded potential
solutions of Maxwells equations in the especially compact form:
Z
[]
1
dV 0 ,
(4.167)
=
4 0 R
143
0
A =
4
where dV 0 d3 r 0 .
[j]
dV 0 ,
R
(4.168)
It is easily seen that
Z
[/t]
[](R1 ) +
tr dV 0
R
Z
1
[/t]
[]
=
R+
R dV 0 ,
3
2
40
R
cR
1
=
4 0
where use has been made of

R
R = ,
R
Likewise,
(R1 ) =
Z
R
R
.
,
t
=
r
R3
cR
(4.170)
tr [j/t]
(R ) [j] +
dV 0
R
Z
0 R [j] R [j/t]
dV 0 .
+
=
3
2
4
R
cR
0
A =
4
(4.169)
(4.171)
Equations (4.164), (4.165), (4.169), and (4.171) can be combined to give
Z
"
#
R
[j/t]
1
R
dV 0 ,
[]
+
E=
(4.172)
3
2
2
4 0
R
t cR
cR
which is the time-dependent generalization of Coulombs law, and
Z
0 [j] R [j/t] R
+
dV 0 ,
B=
3
2
4
R
cR
(4.173)
which is the time-dependent generalization of the Biot-Savart law.

Suppose that the typical variation time-scale of our charges and currents is t 0 .
Let us define R0 = c t0 , which is the distance a light ray travels in time t0 . We can
evaluate Eqs. (4.172) and (4.173) in two asymptotic limits: the near field region
R R0 , and the far field region R R0 . In the near field region
|t tr |
R
=
1,
t0
R0
144
(4.174)
so the difference between retarded time and standard time is relatively small.
This allows us to expand retarded quantities in a Taylor series. Thus,
[] ' +
1 2
(tr t) +
(tr t)2 + ,
2
t
2 t
(4.175)
giving
R 1 2 R2
[] '
+
+ .
t c 2 t2 c2
Expansion of the retarded quantities in the near field region yields
Z
2
R
R
1
1
j/t
E '
2 + dV 0 ,
3
2
2
4 0
R
2 t c R
cR
Z
0 j R 1 (2 j/t2 ) R
B '
+ dV 0 .
3
2
4
R
2
cR
(4.176)
(4.177)
(4.178)
In Eq. (4.177), the first term on the right-hand side corresponds to Coulombs law,
the second term is the correction due to retardation effects, and the third term
corresponds to Faraday induction. In Eq. (4.178), the first term on the right-hand
side is the Biot-Savart law, and the second term is the correction due to retardation effects. Note that the retardation corrections are only of order (R/R 0 )2 . We
might suppose, from looking at Eqs. (4.172) and (4.173), that the corrections
should be of order R/R0 . However, all of the order R/R0 terms canceled out in
the previous expansion. Suppose, then, that we have a d.c. circuit sitting on a
laboratory benchtop. Let the currents in the circuit change on a typical time-scale
of one tenth of a second. In this time, light can travel about 3 10 7 meters, so
R0 30, 000 kilometers. The length-scale of the experiment is about one meter,
so R = 1 meter. Thus, the retardation corrections are of order (3 107 )2 1015 .
It is clear that we are fairly safe just using Coulombs law, Faradays law, and the
Biot-Savart law to analyze the fields generated by this type of circuit.
In the far field region, R R0 , Eqs. (4.172) and (4.173) are dominated by the
terms which vary like R1 , so
Z
1
[j /t]
dV 0 ,
(4.179)
E '
2
4 0
c R
Z
0 [j /t] R
B '
dV 0 ,
(4.180)
2
4
cR
145
where
(j R)
R.
(4.181)
R2
Here, use has been made of [/t] = [ j] and [ j] = [j/t] R/cR +
O(1/R2 ). Suppose that our charges and currents are localized to some region in
the vicinity of r 0 = r . Let R = r r , with R = |r r |. Suppose that the extent
of the current and charge containing region is much less than R . It follows that
retarded quantities can be written
j = j
[(r, t)] ' (r, t R /c),
(4.182)
etc. Thus, the electric field reduces to

1
E'
4 0
hR
j /t dV 0
,
c 2 R
i
whereas the magnetic field is given by

hR
i
j /t dV 0 R
1
B'
.
4 0
c3 R2
Note that
(4.183)
(4.184)
E
= c,
B
(4.185)
E B = 0.
(4.186)
and
This configuration of electric and magnetic fields is characteristic of an electromagnetic wave (see Sect. 4.7). Thus, Eqs. (4.183) and (4.184) describe an electromagnetic wave propagating radially away from the charge and current containing region. Note that the wave is driven by time-varying electric currents.
Now, charges moving with a constant velocity constitute a steady current, so a
non-steady current is associated with accelerating charges. We conclude that accelerating electric charges emit electromagnetic waves. The wave fields, (4.183)
and (4.184), fall off like the inverse of the distance from the wave source. This
behaviour should be contrasted with that of Coulomb or Biot-Savart fields, which
fall off like the inverse square of the distance from the source. The fact that wave
fields attenuate fairly gently with increasing distance from the source is what
146
4.12 Summary
makes astronomy possible. If wave fields obeyed an inverse square law then no
appreciable radiation would reach us from the rest of the Universe.
In conclusion, electric and magnetic fields look simple in the near field region
(they are just Coulomb fields, etc.) and also in the far field region (they are just
electromagnetic waves). Only in the intermediate region, R R0 , do things start
getting really complicated (so we generally do not look in this region!).
4.12 Summary
This marks the end of our theoretical investigation of Maxwells equations. Let us
now summarize what we have learned so far. The field equations which govern
electric and magnetic fields are written:
E =
,
(4.187)
0
B = 0,
(4.188)
B
(4.189)
E = ,
t
1 E
(4.190)
B = 0 j + 2 .
c t
These equations can be integrated to give
I
Z
1
E dS =
dV,
(4.191)
0 V
S
I
B dS = 0,
(4.192)
S
I
Z
E dl =
B dS,
(4.193)
t S
C
Z
I
Z
1
B dl = 0 j dS + 2
E dS.
(4.194)
c t S
C
S
Equations (4.188) and (4.189) are automatically satisfied by writing
A
E =
,
(4.195)
t
147
4.12 Summary
B = A.
(4.196)
This prescription is not unique (there are many choices of and A which generate the same fields) but we can make it unique by adopting the following conventions:
(r) 0
as |r| ,
(4.197)
and
1
+ A = 0.
c2 t
Equations (4.187) and (4.190) reduce to
22 = ,
0
22 A = 0 j.
(4.198)
(4.199)
(4.200)
These are driven wave equations of the general form

1 2
22 u 2 2 2 u = v.
c t
(4.201)
1 (t t 0 |r r 0 |/c)
.
G(r, r ; t, t ) =
4
|r r 0 |
(4.202)
The Greens function for this equation which satisfies the boundary conditions
and is consistent with causality is
0
Thus, the solutions to Eqs. (4.199) and (4.200) are

Z
1
[]
dV 0 ,
(r, t) =
4 0 R
Z
0 [j]
A(r, t) =
dV 0 ,
4 R
(4.203)
(4.204)
where R = |r r 0 |, and dV 0 = d3 r 0 , with [A] A(r 0 , t R/c). These solutions can

be combined with Eqs. (4.195) and (4.196) to give
Z
"
#
R
[j/t]
1
R
dV 0 ,
[]
+
2
(4.205)
E(r, t) =
3
2
4 0
R
t c R
c R
Z
0 [j] R [j/t] R
B(r, t) =
dV 0 .
(4.206)
+
3
2
4
R
cR
148
4.12 Summary
Equations (4.187)(4.206) constitute the complete theory of classical electromagnetism. We can express the same information in terms of field equations
[Eqs. (4.187)(4.190)], integrated field equations [Eqs. (4.191)(4.194)], retarded electromagnetic potentials [Eqs. (4.203) and (4.204)], and retarded electromagnetic fields [Eqs. (4.205) and (4.206)]. Let us now consider the applications of this theory.
149
5 ELECTROSTATICS
Electrostatics
5.1 Introduction
In this section, we shall use Maxwells equations to investigate the electric fields
generated by stationary charge distributions.
5.2 Electrostatic energy

Consider a collection of N static point charges qi located at position vectors ri
(where i runs from 1 to N). What is the electrostatic energy stored in such a
collection? Another way of asking this is, how much work would we have to do
in order to assemble the charges, starting from an initial state in which they are
all at rest and very widely separated?
We know that a static electric field is conservative, and can consequently be
written in terms of a scalar potential:
E = .
(5.1)
We also know that the electric force on a charge q is written

f = q E.
(5.2)
The work we would have to do against electrical forces in order to move the
charge from point P to point Q is simply
ZQ
ZQ
ZQ
W=
f dl = q
E dl = q
dl = q [(Q) (P)] .
(5.3)
P
The negative sign in the above expression comes about because we would have
to exert a force f on the charge, in order to counteract the force exerted by the
electric field. Recall that the scalar potential generated by a point charge q 0 at
position r 0 is
q0
1
.
(5.4)
(r) =
4 0 |r r 0 |
150
5 ELECTROSTATICS
Let us build up our collection of charges one by one. It takes no work to bring
the first charge from infinity, since there is no electric field to fight against. Let
us clamp this charge in position at r1 . In order to bring the second charge into
position at r2 , we have to do work against the electric field generated by the first
charge. According to Eqs. (5.3) and Eqs. (5.4), this work is given by
W2 =
1
q2 q1
.
4 0 |r2 r1 |
(5.5)
Let us now bring the third charge into position. Since electric fields and scalar
potentials are superposable, the work done whilst moving the third charge from
infinity to r3 is simply the sum of the works done against the electric fields generated by charges 1 and 2 taken in isolation:
q3 q2
1 q3 q1
+
.
W3 =
4 0 |r3 r1 | |r3 r2 |
(5.6)
Thus, the total work done in assembling the three charges is given by
1 q2 q1
q3 q1
q3 q2
W=
+
+
.
4 0 |r2 r1 | |r3 r1 | |r3 r2 |
(5.7)
This result can easily be generalized to N charges:

N
1 X X qi qj
W=
.
4 0 i=1 j<i |ri rj |
(5.8)
The restriction that j must be less than i makes the above summation rather
messy. If we were to sum without restriction (other than j 6= i) then each pair
of charges would be counted twice. It is convenient to do just this, and then to
divide the result by two. Thus,
N
1 1 X X qi qj
.
W=
2 4 0 i=1
|ri rj |
(5.9)
j=1
j6=i
This is the potential energy (i.e., the difference between the total energy and the
kinetic energy) of a collection of charges. We can think of this as the work needed
151
5 ELECTROSTATICS
to bring static charges from infinity and assemble them in the required formation.
Alternatively, this is the kinetic energy which would be released if the collection
were dissolved, and the charges returned to infinity. But where is this potential
energy stored? Let us investigate further.
Equation (5.9) can be written
N
1X
qi i ,
W=
2 i=1
where
(5.10)
1 X qj
i =
4 0
|ri rj |
(5.11)
j=1
j6=i
is the scalar potential experienced by the i th charge due to the other charges in
the distribution.
Let us now consider the potential energy of a continuous charge distribution.
It is tempting to write
Z
1
W=
d3 r,
(5.12)
2
by analogy with Eqs. (5.10) and (5.11), where
Z
(r 0 ) 3 0
1
dr
(5.13)
(r) =
4 0 |r r 0 |
is the familiar scalar potential generated by a continuous charge distribution. Let
us try this out. We know from Maxwells equations that
= 0 E,
(5.14)
so Eq. (5.12) can be written

0
W=
2
E d3 r.
(5.15)
Vector field theory yields the standard result

(E ) = E + E.
152
(5.16)
5 ELECTROSTATICS
However, = E, so we obtain
Z
#
"Z
0
3
2 3
(E ) d r + E d r
W=
2
Application of Gauss theorem gives
I
Z
!
0
2
W=
E dS + E dV ,
2 S
V
(5.17)
(5.18)
where V is some volume which encloses all of the charges, and S is its bounding
surface. Let us assume that V is a sphere, centred on the origin, and let us take
the limit in which the radius r of this sphere goes to infinity. We know that, in
general, the electric field at large distances from a bounded charge distribution
looks like the field of a point charge, and, therefore, falls off like 1/r 2 . Likewise,
the potential falls off like 1/r. However, the surface area of the sphere increases
like r2 . Hence, it is clear that, in the limit as r , the surface integral in
Eq. (5.18) falls off like 1/r, and is consequently zero. Thus, Eq. (5.18) reduces to
Z
0
W=
E2 d3 r,
(5.19)
2
where the integral is over all space. This is a very nice result. It tells us that
the potential energy of a continuous charge distribution is stored in the electric
field. Of course, we now have to assume that an electric field possesses an energy
density
0 2
E.
(5.20)
U=
2
We can easily check that Eq. (5.19) is correct. Suppose that we have a charge
Q which is uniformly distributed within a sphere of radius a. Let us imagine
building up this charge distribution from a succession of thin spherical layers of
infinitesimal thickness. At each stage, we gather a small amount of charge from
infinity, and spread it over the surface of the sphere in a thin layer from r to
r + dr. We continue this process until the final radius of the sphere is a. If q(r)
is the charge in the sphere when it has attained radius r, then the work done in
bringing a charge dq to it is
dW =
1 q(r) dq
.
4 0
r
153
(5.21)
5 ELECTROSTATICS
This follows from Eq. (5.5), since the electric field generated by a spherical charge
distribution (outside itself) is the same as that of a point charge q(r) located at
the origin (r = 0) (see later). If the constant charge density in the sphere is
then
4
(5.22)
q(r) = r3 ,
3
and
dq = 4 r2 dr.
(5.23)
Thus, Eq. (5.21) becomes
4 2 4
r dr.
(5.24)
3 0
The total work needed to build up the sphere from nothing to radius a is plainly
Z
4 2 5
4 2 a 4
r dr =
a.
(5.25)
W=
3 0
15 0
0
dW =
This can also be written in terms of the total charge Q = (4/3) a3 as

W=
3 Q2
.
5 4 0 a
(5.26)
Now that we have evaluated the potential energy of a spherical charge distribution by the direct method, let us work it out using Eq. (5.19). We assume that
the electric field is radial and spherically symmetric, so E = Er (r) ^r. Application
of Gauss law,
Z
I
1
dV,
(5.27)
E dS =
0 V
S
where V is a sphere of radius r, yields
Er (r) =
Q r
4 0 a3
(5.28)
Er (r) =
Q
4 0 r2
(5.29)
for r < a, and
154
5 ELECTROSTATICS
for r a. Note that the electric field generated outside the charge distribution is
the same as that of a point charge Q located at the origin, r = 0. Equations (5.19),
(5.28), and (5.29) yield
Z
Z
dr
Q2 1 a 4
r
dr
+
W=
,
(5.30)
2
8 0 a6 0
a r
which reduces to
3 Q2
Q2
1
+1 =
.
W=
8 0 a 5
5 4 0 a
Thus, Eq. (5.19) gives the correct answer.
!
(5.31)
The reason we have checked Eq. (5.19) so carefully is that on close inspection
it is found to be inconsistent with Eq. (5.10), from which it was supposedly derived! For instance, the energy given by Eq. (5.19) is manifestly positive definite,
whereas the energy given by Eq. (5.10) can be negative (it is certainly negative for a collection of two point charges of opposite sign). The inconsistency
was introduced into our analysis when we replaced Eq. (5.11) by Eq. (5.13).
In Eq. (5.11), the self-interaction of the i th charge with its own electric field is
specifically excluded, whereas it is included in Eq. (5.13). Thus, the potential
energies (5.10) and (5.19) are different, because in the former we start from
ready-made point charges, whereas in the latter we build up the whole charge
distribution from scratch. Thus, if we were to work out the potential energy of a
point charge distribution using Eq. (5.19) we would obtain the energy (5.10) plus
the energy required to assemble the point charges. What is the energy required to
assemble a point charge? In fact, it is infinite. To see this, let us suppose, for the
sake of argument, that our point charges are actually made of charge uniformly
distributed over a small sphere of radius a. According to Eq. (5.26), the energy
required to assemble the i th point charge is
3 qi2
.
Wi =
5 4 0 a
(5.32)
We can think of this as the self-energy of the i th charge. Thus, we can write
Z
N
N
X
0
1X
2 3
W=
E d r=
qi i +
Wi
(5.33)
2
2 i=1
i=1
155
5 ELECTROSTATICS
5.3 Ohms law
which enables us to reconcile Eqs. (5.10) and (5.19). Unfortunately, if our point
charges really are point charges then a 0, and the self-energy of each charge
becomes infinite. Thus, the potential energies predicted by Eqs. (5.10) and (5.19)
differ by an infinite amount. What does this all mean? We have to conclude that
the idea of locating electrostatic potential energy in the electric field is inconsistent with the existence of point charges. One way out of this difficulty would
be to say that all elementary charges, such as electrons, are not points, but instead small distributions of charge. Alternatively, we could say that our classical
theory of electromagnetism breaks down on very small length-scales due to quantum effects. Unfortunately, the quantum mechanical version of electromagnetism
(quantum electrodynamics, or QED, for short) suffers from the same infinities
in the self-energies of particles as the classical version. There is a prescription,
called renormalization, for steering round these infinities, and getting finite answers which agree with experiments to extraordinary accuracy. However, nobody
really understands why this prescription works. The problem of the infinite selfenergies of elementary charged particles is still unresolved.
5.3 Ohms law

We all know the simplest version of Ohms law:
V = I R,
(5.34)
where V is the voltage drop across a resistor of resistance R when a current I

flows through it. Let us generalize this law so that it is expressed in terms of
E and j, rather than V and I. Consider a length l of a conductor of uniform
cross-sectional area A with a current I flowing down it. In general, we expect
the electrical resistance of the conductor to be proportional to its length, and
inversely proportional to its area (i.e., it is harder to push an electrical current
down a long rather than a short wire, and it is easier to push a current down a
wide rather than a narrow conducting channel.) Thus, we can write
R=
156
l
.
A
(5.35)
5 ELECTROSTATICS
5.4 Conductors
The constant is called the resistivity, and is measured in units of ohm-meters.

Ohms law becomes
l
(5.36)
V = I.
A
However, I/A = jz (supposing that the conductor is aligned along the z-axis) and
V/l = Ez , so the above equation reduces to
Ez = j z .
(5.37)
There is nothing special about the z-axis (in an isotropic conducting medium), so
the previous formula immediately generalizes to
E = j.
(5.38)
This is the vector form of Ohms law.

A charge q which moves through a voltage drop V acquires an energy q V
from the electric field. In a resistor, this energy is dissipated as heat. This type
of heating is called ohmic heating. Suppose that N charges per unit time pass
through a resistor. The current flowing is obviously I = N q. The total energy
gained by the charges, which appears as heat inside the resistor, is
P = NqV = IV
(5.39)
per unit time. Thus, the heating power is

V2
P = IV = I R =
.
R
Equations (5.39) and (5.40) generalize to
2
P = j E = j2 ,
(5.40)
(5.41)
where P is now the power dissipated per unit volume in a resistive medium.
5.4 Conductors
Most (but not all) electrical conductors obey Ohms law. Such conductors are
termed ohmic. Suppose that we apply an electric field to an ohmic conductor.
157
5 ELECTROSTATICS
5.4 Conductors
What is going to happen? According to Eq. (5.38), the electric field drives currents. These redistribute the charge inside the conductor until the original electric
field is canceled out. At this point, the currents stop flowing. It might be objected
that the currents could keep flowing
H in closed loops. According to Ohms law, this
would require a non-zero e.m.f., E dl, acting around each loop (unless the conductor is a superconductor, with = 0). However, we know that in steady-state
I
E dl = 0
(5.42)
C
around any closed loop C. This proves that a steady-state e.m.f. acting around a
closed loop inside a conductor is impossible. The only other alternative is
(5.43)
j=E=0
inside a conductor. It immediately follows from the Maxwell equation, E =

/0 , that
= 0.
(5.44)
So, there are no electric charges in the interior of a conductor. But, how can
a conductor cancel out an applied electric field if it contains no charges? The
answer is that all of the charges reside on the surface of the conductor. In reality,
the charges lie within one or two atomic layers of the surface (see any textbook
on solid-state physics). The difference in scalar potential between two points P
and Q is simply
ZQ
ZQ
E dl.
(5.45)
dl =
(Q) (P) =
P
However, if P and Q lie inside the same conductor then it is clear from Eq. (5.45)
that the potential difference between P and Q is zero. This is true no matter
where P and Q are situated inside the conductor, so we conclude that the scalar
potential must be uniform inside a conductor. A corollary of this is that the surface
of a conductor is an equipotential (i.e., = constant) surface.
Not only is the electric field inside a conductor zero. It is also possible to
demonstrate that the field within an empty cavity lying inside a conductor is also
zero, provided that there are no charges within the cavity. Let us, first of all,
158
5 ELECTROSTATICS
5.4 Conductors
conductor
S
++
+
+
cavity
Figure 41:
apply Gauss law to a surface S which surrounds the cavity, but lies wholly in the
conducting material (see Fig. 41). Since the electric field is zero in a conductor, it
follows that zero net charge is enclosed by S. This does not preclude the possibility that there are equal amounts of positive and negative charges distributed on
the inner surface of the conductor. However, we can easily rule out this possibility
using the steady-state relation
I
E dl = 0,
(5.46)
C
for any closed loop C. If there are any electric field-lines inside the cavity then
they must run from the positive to the negative surface charges. Consider a loop C
which straddles the cavity and the conductor, such as the one shown in Fig. 41. In
the presence of field-lines, it is clear that the line integral of E along that portion
of the loop which lies inside the cavity is non-zero. However, the line integral of
E along that portion of the loop which runs through the conducting material is
obviously zero (since E = 0 inside a conductor). Thus, the line integral of the
field around the closed loop C is non-zero. This, clearly contradicts Eq. (5.46).
In fact, this equation implies that the line integral of the electric field along any
path which runs through the cavity, from one point on the interior surface of the
conductor to another, is zero. This can only be the case if the electric field itself
159
5 ELECTROSTATICS
5.4 Conductors
is zero everywhere inside the cavity. There is one proviso to this argument. The
electric field inside a cavity is only zero if the cavity contains no charges. If the
cavity contains charges then our argument fails because it is possible to envisage
that the line integral of the electric field along many different paths across the
cavity could be zero without the fields along these paths necessarily being zero
(this argument is somewhat inexact: we shall improve it later on).
We have shown that if a cavity is completely enclosed by a conductor then no
stationary distribution of charges outside can ever produce any fields inside. So,
we can shield a piece of electrical equipment from stray external electric fields by
placing it inside a metal can. Using similar arguments to those given above, we
can also show that no static distribution of charges inside a closed conductor can
ever produce a field outside the conductor. Clearly, shielding works both ways!
vacuum
E
Gaussian
pillbox
conductor
Figure 42:
Let us consider some small region on the surface of a conductor. Suppose that
the local surface charge density is , and that the electric field just outside the
conductor is E. Note that this field must be directed normal to the surface of the
conductor. Any parallel component would be shorted out by surface currents.
Another way of saying this is that the surface of a conductor is an equipotential
surface. We know that is always perpendicular to equipotential surfaces, so
E = must be locally perpendicular to a conducting surface. Let us use
160
5 ELECTROSTATICS
5.4 Conductors
Gauss law,
Z
1
dV,
(5.47)
E dS =
0 V
S
where V is a so-called Gaussian pill-box (see Fig. 42). This is a pill-box shaped
volume whose two ends are aligned normal to the surface of the conductor, with
the surface running between them, and whose sides are tangential to the surface
normal. It is clear that E is parallel to the sides of the box, so the sides make
no contribution to the surface integral. The end of the box which lies inside the
conductor also makes no contribution, since E = 0 inside a conductor. Thus,
the only non-zero contribution to the surface integral comes from the end lying
in free space. This contribution is simply E A, where E denotes an outward
pointing (from the conductor) normal electric field, and A is the cross-sectional
area of the box. The charge enclosed by the box is simply A, from the definition
of a surface charge density. Thus, Gauss law yields
E =
(5.48)
0
I
as the relationship between the normal electric field immediately outside a conductor and the surface charge density.
E
Figure 43:
Let us look at the electric field generated by a sheet charge distribution a little
more carefully. Suppose that the charge per unit area is . By symmetry, we
expect the field generated below the sheet to be the mirror image of that above
161
5 ELECTROSTATICS
5.4 Conductors
the sheet (at least, locally). Thus, if we integrate Gauss law over a pill-box of
cross sectional area A, as shown in Fig. 43, then the two ends both contribute
Esheet A to the surface integral, where Esheet is the normal electric field generated
above and below the sheet. The charge enclosed by the pill-box is just A. Thus,
Gauss law yields a symmetric electric field
above,
(5.49)
Esheet = +
2 0
Esheet =
below.
(5.50)
2 0
So, how do we get the asymmetric electric field of a conducting surface, which
is zero immediately below the surface (i.e., inside the conductor) and non-zero
immediately above it? Clearly, we have to add in an external field (i.e., a field
which is not generated locally by the sheet charge). The requisite field is
Eext =
(5.51)
2 0
both above and below the charge sheet. The total field is the sum of the field
generated locally by the charge sheet and the external field. Thus, we obtain
above,
(5.52)
Etotal = +
0
Etotal = 0
below,
(5.53)
which is in agreement with Eq. (5.48).
The external field exerts a force on the charge sheet. The field generated
locally by the sheet itself obviously cannot exert a force (the sheet cannot exert
a force on itself!). The force per unit area acting on the surface of the conductor
always acts outward, and is given by
2
p = Eext =
.
(5.54)
2 0
Thus, there is an electrostatic pressure acting on any charged conductor. This
effect can be visualized by charging up soap bubbles: the additional electrostatic
pressure eventually causes them to burst. The electrostatic pressure can also be
written
0 2
p=
E,
(5.55)
2
162
5 ELECTROSTATICS
where E is the field strength immediately above the surface of the conductor.
Note that, according to the above formula, the electrostatic pressure is equivalent to the energy density of the electric field immediately outside the conductor.
This is not a coincidence. Suppose that the conductor expands by an average distance dx, due to the electrostatic pressure. The electric field is excluded from the
region into which the conductor expands. The volume of this region dV = A dx,
where A is the surface area of the conductor. Thus, the energy of the electric
field decreases by an amount dE = U dV = (0 /2) E2 dV, where U is the energy density of the field. This decrease in energy can be ascribed to the work
which the field does on the conductor in order to make it expand. This work is
dW = p A dx, where p is the force per unit area the field exerts on the conductor.
Thus, dE = dW, from energy conservation, giving
p=
0 2
E.
2
(5.56)
This technique for calculating a force given an expression for the energy of a
system as a function of some adjustable parameter is called the principle of virtual
work, and is very useful.
We have seen that an electric field is excluded from the inside of the conductor,
but not from the outside, giving rise to a net outward force. We can account for
this by saying that the field exerts a negative pressure (0 /2) E2 on the conductor.
We know that if we evacuate a metal can then the pressure difference between the
inside and the outside eventually causes it to implode. Likewise, if we place the
can in a strong electric field then the pressure difference between the inside and
the outside will eventually cause it to explode. How big a field do we need before
the electrostatic pressure difference is the same as that obtained by evacuating
the can? In other words, what field exerts a negative pressure of one atmosphere
(i.e., 105 newtons per meter squared) on conductors? The answer is a field of
strength E 108 volts per meter. Fortunately, this is a rather large field, so there
is no danger of your car exploding when you turn on the stereo!
163
5 ELECTROSTATICS
Eperp A
Epara A
P
Epara B
Gaussian pillbox
loop
Eperp B
Figure 44:

What are the most general boundary conditions satisfied by the electric field at
the interface between two media: e.g., the interface between a vacuum and a
conductor? Consider an interface P between two media A and B. Let us, first of
all, apply Gauss law,
Z
I
1
dV,
(5.57)
E dS =

0
V
S
to a Gaussian pill-box S of cross-sectional area A whose two ends are locally
parallel to the interface (see Fig. 44). The ends of the box can be made arbitrarily
close together. In this limit, the flux of the electric field out of the sides of the
box is obviously negligible. The only contribution to the flux comes from the two
ends. In fact,
I
E dS = (E A E B ) A,
(5.58)
S
where E A is the perpendicular (to the interface) electric field in medium A at

the interface, etc. The charge enclosed by the pill-box is simply A, where is
the sheet charge density on the interface. Note that any volume distribution of
charge gives rise to a negligible contribution to the right-hand side of the above
equation, in the limit where the two ends of the pill-box are very closely spaced.
Thus, Gauss law yields
E A E B =
(5.59)
0
164
5 ELECTROSTATICS
5.6 Capacitors
at the interface: i.e., the presence of a charge sheet on an interface causes a

discontinuity in the perpendicular component of the electric field. What about
the parallel electric field? Let us apply Faradays law to a rectangular loop C
whose long sides, length l, run parallel to the interface,
Z
I
B dS
(5.60)
E dl =
t S
C
(see Fig. 44). The length of the short sides is assumed to be arbitrarily small. The
dominant contribution to the loop integral comes from the long sides:
I
E dl = (Ek A Ek B ) l,
(5.61)
C
where Ek A is the parallel (to the interface) electric field in medium A at the
interface, etc. The flux of the magnetic field through the loop is approximately
B A, where B is the component of the magnetic field which is normal to the
loop, and A is the area of the loop. But, A 0 as the short sides of the loop are
shrunk to zero. So, unless the magnetic field becomes infinite (we shall assume
that it does not), the flux also tends to zero. Thus,
Ek A E k B = 0 :
(5.62)
i.e., there can be no discontinuity in the parallel component of the electric field
across an interface.
5.6 Capacitors
We can store electrical charge on the surface of a conductor. However, electric
fields will be generated immediately above this surface. The conductor can only
successfully store charge if it is electrically insulated from its surroundings. Air
is a very good insulator. Unfortunately, air ceases to be an insulator when the
electric field-strength through it exceeds some critical value which is about Ecrit
106 volts per meter. This phenomenon, which is called break-down, is associated
with the formation of sparks. The most well-known example of the break-down
of air is during a lightning strike. Clearly, a good charge storing device is one
165
5 ELECTROSTATICS
5.6 Capacitors
which holds a large amount of charge but only generates small electric fields.
Such a device is called a capacitor.
Consider two thin, parallel, conducting plates
of cross-sectional area A which
are separated by a small distance d (i.e., d A). Suppose that each plate
carries an equal and opposite charge Q. We expect this charge to spread evenly
over the plates to give an effective sheet charge density = Q/A on each
plate. Suppose that the upper plate carries a positive charge and that the lower
plate carries a negative charge. According to Eqs. (5.49) and (5.50), the field
generated by the upper plate is normal to the plate and of magnitude
20
=
20
Eupper = +
above,
(5.63)
Eupper
below.
(5.64)
Elower =
above,
(5.65)
Elower
below.
(5.66)
Likewise, the field generated by the lower plate is
20
= +
20
Note that we are neglecting any leakage of the field at the edges of the plates.
This is reasonable if the plates are closely spaced. The total field is the sum of the
two fields generated by the upper and lower plates. Thus, the net field is normal
to the plates, and of magnitude
0
= 0
E =
between,
(5.67)
otherwise.
(5.68)
Since the electric field is uniform, the potential difference between the plates is
simply
d
V = E d =
.
(5.69)
0
It is conventional to measure the capacity of a conductor, or set of conductors, to
store charge, but generate small electric fields, in terms of a parameter called the
166
5 ELECTROSTATICS
5.6 Capacitors
capacitance. This is usually denoted C. The capacitance of a charge storing device

is simply the ratio of the charge stored to the potential difference generated by
the charge. Thus,
Q
(5.70)
C= .
V
Clearly, a good charge storing device has a high capacitance. Incidentally, capacitance is measured in coulombs per volt, or farads. This is a rather unwieldy
unit, since good capacitors typically have capacitances which are only about one
millionth of a farad. For a parallel plate capacitor, it is clear that
A 0 A
=
.
(5.71)
V
d
Note that the capacitance only depends on geometric quantities, such as the area
and spacing of the plates. This is a consequence of the superposability of electric
fields. If we double the charge on conductors then we double the electric fields
generated around them, and we, therefore, double the potential difference between the conductors. Thus, the potential difference between the conductors is
always directly proportional to the charge carried: the constant of proportionality
(the inverse of the capacitance) can only depend on geometry.
C=
Suppose that the charge Q on each plate is built up gradually by transferring

small amounts of charge from one plate to another. If the instantaneous charge
on the plates is q, and an infinitesimal amount of positive charge dq is transferred from the negatively charged plate to the positively charge plate, then the
work done is dW = V dq = q dq/C, where V is the instantaneous voltage difference between the plates. Note that the voltage difference is such that it opposes
any increase in the charge on either plate. The total work done in charging the
capacitor is
Z
1
1 Q
Q2
= C V 2,
(5.72)
W=
q dq =
C 0
2C 2
where use has been made of Eq. (5.70). The energy stored in the capacitor is the
same as the work required to charge up the capacitor. Thus,
1
C V 2.
2
This is a general result which holds for all types of capacitor.
W=
167
(5.73)
5 ELECTROSTATICS
5.6 Capacitors
The energy of a charged parallel plate capacitor is actually stored in the electric
field between the plates. This field is of approximately constant magnitude E =
V/d, and occupies a region of volume A d. Thus, given the energy density of an
electric field, U = (0 /2) E2 , the energy stored in the electric field is
0 V 2
1
W=
A
d
=
C V 2,
(5.74)
2
2 d
2
where use has been made of Eq. (5.71). Note that Eqs. (5.72) and (5.74) agree.
We all know that if we connect a capacitor across the terminals of a battery then
a transient current flows as the capacitor charges up. The capacitor can then
be placed to one side, and, some time later, the stored charge can be used: for
instance, to transiently light a bulb in an electrical circuit. What is interesting
here is that the energy stored in the capacitor is stored as an electric field, so
we can think of a capacitor as a device which either stores energy in, or extracts
energy from, an electric field.
The idea, which we discussed earlier, that an electric field exerts a negative
pressure (0 /2) E2 on conductors immediately suggests that the two plates in a
parallel plate capacitor attract one another with a mutual force
0 2
1 C V2
F=
E A=
.
2
2 d
(5.75)
It is not necessary to have two oppositely charged conductors in order to make

a capacitor. Consider an isolated sphere of radius a which carries a charge Q.
The radial electric field generated outside the sphere is given by
Er (r > a) =
Q
.
4 0 r2
(5.76)
The potential difference between the sphere and infinity, or, more realistically,
some large, relatively distant reservoir of charge such as the Earth, is
Q
.
4 0 a
(5.77)
Q
= 4 0 a.
V
(5.78)
V=
Thus, the capacitance of the sphere is
C=
168
5 ELECTROSTATICS
5.6 Capacitors
The energy of a sphere when it carries a charge Q is again given by (1/2) C V 2 . It

can easily be demonstrated that this is really the energy contained in the electric
field around the sphere.
Suppose that we have two spheres of radii a and b, respectively, which are
connected by an electric wire. The wire allows charge to move back and forth
between the spheres until they reach the same potential (with respect to infinity).
Let Q be the charge on the first sphere and Q 0 the charge on the second sphere.
Of course, the total charge Q0 = Q + Q 0 carried by the two spheres is a conserved
quantity. It follows from Eq. (5.77) that
Q
a
,
=
Q0
a+b
Q0
b
=
.
Q0
a+b
(5.79)
(5.80)
Note that if one sphere is much smaller than the other one, e.g., b a, then the
large sphere grabs most of the charge:
a
Q
'
1.
Q0
b
(5.81)
The ratio of the electric fields generated just above the surfaces of the two spheres
follows from Eqs. (5.76) and (5.81):
Eb
a
' .
Ea
b
(5.82)
If b a, then the field just above the smaller sphere is far larger than that above
the larger sphere. Equation (5.82) is a simple example of a far more general rule.
The electric field above some point on the surface of a conductor is inversely
proportional to the local radius of curvature of the surface.
It is clear that if we wish to store significant amounts of charge on a conductor then the surface of the conductor must be made as smooth as possible.
Any sharp spikes on the surface will inevitably have comparatively small radii
of curvature. Intense local electric fields are generated in these regions. These
can easily exceed the critical field for the break-down of air, leading to sparking
169
5 ELECTROSTATICS
and the eventual loss of the charge on the conductor. Sparking can also be very
destructive because the associated electric currents flow through very localized
regions giving rise to intense ohmic heating.
As a final example, consider two co-axial conducting cylinders of radii a and
b, where a < b. Suppose that the charge per unit length carried by the outer
and inner cylinders is +q and q, respectively. We can safely assume that E =
Er (r) ^r, by symmetry (adopting standard cylindrical polar coordinates). Let us
apply Gauss law to a cylinder of radius r, co-axial with the conductors, and of
length l. For a < r < b, we find that
2 r l Er (r) =
ql
,
0
(5.83)
so
q
(5.84)
2 0 r
for a < r < b. It is fairly obvious that Er = 0 if r is not in the range a to b. The
potential difference between the inner and outer cylinders is
Z outer
Z inner
E dl
E dl =
V =
Er =
Zb
outer
inner
Er dr =
Zb
so
q
2 0
dr
,
r
q
b
ln .
2 0 a
Thus, the capacitance per unit length of the two cylinders is
V=
C=
q
2 0
=
.
V
ln b/a
(5.85)
(5.86)
(5.87)

We know that in steady-state we can write
E = ,
170
(5.88)
5 ELECTROSTATICS
5.8 The uniqueness theorem
with the scalar potential satisfying Poissons equation:
2 = .
0
We even know the general solution to this equation:
Z
1
(r 0 ) 3 0
(r) =
dr.
4 0 |r r 0 |
(5.89)
(5.90)
So, what else is there to say about Poissons equation? Well, consider a positive
(say) point charge in the vicinity of an uncharged, insulated, conducting sphere.
The charge attracts negative charges to the near side of the sphere, and repels
positive charges to the far side. The surface charge distribution induced on the
sphere is such that it is maintained at a constant electrical potential. We now have
a problem. We cannot use formula (5.90) to work out the potential (r) around
the sphere, since we do not know how the charges induced on the conducting
surface are distributed. The only things which we know about the surface of the
sphere are that it is an equipotential surface, and carries zero net charge. Clearly,
in the presence of conducting surfaces the solution (5.90) to Poissons equation
is completely useless. Let us now try to develop some techniques for solving Poissons equation which allow us to solve real problems (which invariably involve
conductors).

We have already seen the great value of the uniqueness theorem for Poissons
equation (or Laplaces equation) in our discussion of Helmholtzs theorem (see
Sect. 3.11). Let us now examine this theorem in detail.
Consider a volume V bounded by some surface S. Suppose that we are given
the charge density throughout V, and the value of the scalar potential S on S.
Is this sufficient information to uniquely specify the scalar potential throughout
V? Suppose, for the sake of argument, that the solution is not unique. Let there
be two potentials 1 and 2 which satisfy
2 1 = ,
(5.91)
0
171
5 ELECTROSTATICS
2 2 =
0
(5.92)
throughout V, and
1 = S ,
(5.93)
2 = S
(5.94)
on S. We can form the difference between these two potentials:

3 = 1 2 .
(5.95)
2 3 = 0
(5.96)
3 = 0
(5.97)
(3 3 ) (3 )2 + 3 2 3 .
(5.98)
The potential 3 clearly satisfies
throughout V, and
on S.
According to vector field theory,
Thus, using Gauss theorem

Z
I

(3 )2 + 3 2 3 dV = 3 3 dS.
V
(5.99)
But, 2 3 = 0 throughout V, and 3 = 0 on S, so the above equation reduces to

Z
(3 )2 dV = 0.
(5.100)
V
Note that (3 )2 is a positive definite quantity. The only way in which the volume
integral of a positive definite quantity can be zero is if that quantity itself is zero
throughout the volume. This is not necessarily the case for a non-positive definite
quantity: we could have positive and negative contributions from various regions
inside the volume which cancel one another out. Thus, since (3 )2 is positive
definite, it follows that
3 = constant
(5.101)
172
5 ELECTROSTATICS
throughout V. However, we know that 3 = 0 on S, so we get

3 = 0
(5.102)
1 = 2
(5.103)
throughout V. In other words,

throughout V and on S. Our initial assumption that 1 and 2 are two different
solutions of Poissons equation, satisfying the same boundary conditions, turns
out to be incorrect.
The fact that the solutions to Poissons equation are unique is very useful. It
means that if we find a solution to this equationno matter how contrived the
derivationthen this is the only possible solution. One immediate use of the
uniqueness theorem is to prove that the electric field inside an empty cavity in
a conductor is zero. Recall that our previous proof of this was rather involved,
and was also not particularly rigorous (see Sect. 5.4). We know that the interior
surface of the conductor is at some constant potential V, say. So, we have = V
on the boundary of the cavity, and 2 = 0 inside the cavity (since it contains
no charges). One rather obvious solution to these equations is = V throughout
the cavity. Since the solutions to Poissons equation are unique, this is the only
solution. Thus,
E = = V = 0
(5.104)
inside the cavity.
Suppose that some volume V contains a number of conductors. We know that
the surface of each conductor is an equipotential surface, but, in general, we do
not know what potential each surface is at (unless we are specifically told that it
is earthed, etc.). However, if the conductors are insulated it is plausible that we
might know the charge on each conductor. Suppose that there are N conductors,
each carrying a charge Qi (i = 1 to N), and suppose that the region V containing these conductors is filled by a known charge density , and bounded by
some surface S which is either infinity or an enclosing conductor. Is this enough
information to uniquely specify the electric field throughout V?
Well, suppose that it is not enough information, so that there are two fields E 1
173
5 ELECTROSTATICS
and E2 which satisfy
,
0
=
0
E1 =
(5.105)
E2
(5.106)
throughout V, with
I
Qi
,
0
(5.107)
Si
E1 dSi =
Qi
0
(5.108)
Si
E2 dSi =
on the surface of the ith conductor, and, finally,

I
Qtotal
E1 dS =
,
0
S
I
Qtotal
E2 dS =
0
S
(5.109)
(5.110)
over the bounding surface, where

Qtotal =
N
X
Qi +
i=1
dV
(5.111)
is the total charge contained in volume V.

Let us form the difference field
E3 = E 1 E 2 .
(5.112)
E3 = 0
(5.113)
It is clear that
throughout V, and
for all i, with
Si
E3 dSi = 0
(5.114)
E3 dS = 0.
(5.115)
174
5 ELECTROSTATICS
Now, we know that each conductor is at a constant potential, so if

(5.116)
E3 = 3 ,
then 3 is a constant on the surface of each conductor. Furthermore, if the outer

surface S is infinity then 1 = 2 = 3 = 0 on this surface. If the outer surface
is an enclosing conductor then 3 is a constant on this surface. Either way, 3 is
constant on S.
Consider the vector identity
(3 E3 ) 3 E3 + E3 3 .
(5.117)
We have E3 = 0 throughout V, and 3 = E3 , so the above identity reduces

to
(3 E3 ) = E32
(5.118)
throughout V. Integrating over V, and making use of Gauss theorem, yields

I
Z
N I
X
3 E3 dSi 3 E3 dS.
(5.119)
E32 dV =
V
Si
i=1
However, 3 is a constant on the surfaces Si and S. So, making use of Eqs. (5.114)
and (5.115), we obtain
Z
E32 dV = 0.
(5.120)
Of course,
E32
is a positive definite quantity, so the above relation implies that

E3 = 0
(5.121)
throughout V: i.e., the fields E1 and E2 are identical throughout V.

For a general electrostatic problem involving charges and conductors, it is clear
that if we are given either the potential at the surface of each conductor or the
charge carried by each conductor (plus the charge density throughout the volume, etc.) then we can uniquely determine the electric field. There are many
other uniqueness theorems which generalize this result still further: i.e., we could
be given the potential of some of the conductors and the charge carried by the
others, and the solution would still be unique.
175
5 ELECTROSTATICS
5.9 One-dimensional solution of Poissons equation
At this point, it is worth noting that there are also uniqueness theorems associated with magnetostatics. For instance, if the current density, j, is specified
throughout some volume V, and either the magnetic field, B, or the vector potential, A, is specified on the bounding surface S, then the magnetic field is uniquely
determined throughout V and on S. The proof of this proposition proceeds along
the usual lines. Suppose that the magnetic field is not uniquely determined. In
other words, suppose there are two magnetic fields, B1 and B2 , satisfying
B1 = 0 j,
(5.122)
B2 = 0 j,
(5.123)
B3 = 0
(5.124)
throughout V. Suppose, further, that either B1 = B2 = BS or A1 = A2 = AS on S.

Forming the difference field, B3 = B1 B2 , we have
throughout V, and either B3 = 0 or A3 = 0 on S. Now, according to vector field
theory,
Z
I
h
i
2
( U) U U dV U ( U) dS.
(5.125)
V
Setting U = A3 , and using B3 = A3 and Eq. (5.124), we obtain

Z
I
2
B3 dV = A3 B3 dS.
V
(5.126)
However, we know that either B3 or A3 is zero on S. Hence, we obtain

Z
B32 dV = 0.
(5.127)
Since, B32 is positive definite, the only way in which the above equation can be
satisfied is if B3 is zero throughout V. Hence, B1 = B2 throughout V, and the
solution is therefore unique.

So, how do we actually solve Poissons equation,
2 2 2
(x, y, z)
+
+
=
,
x2
y2
z2
0
176
(5.128)
5 ELECTROSTATICS
in practice? In general, the answer is that we use a computer. However, there

are a few situations, possessing a high degree of symmetry, where it is possible to
find analytic solutions. Let us discuss some of these solutions.
Suppose, first of all, that there is no variation of quantities in (say) the y- and
z-directions. In this case, Poissons equation reduces to an ordinary differential
equation in x, the solution of which is relatively straight-forward. Consider, for
instance, a vacuum diode, in which electrons are emitted from a hot cathode and
accelerated towards an anode, which is held at a large positive potential V 0 with
respect to the cathode. We can think of this as an essentially one-dimensional
problem. Suppose that the cathode is at x = 0 and the anode at x = d. Poissons
equation takes the form
(x)
d2
,
(5.129)
=
dx2
0
where (x) satisfies the boundary conditions (0) = 0 and (d) = V 0 . By energy
conservation, an electron emitted from rest at the cathode has an x-velocity v(x)
which satisfies
1
me v2 (x) e (x) = 0.
(5.130)
2
Finally, in a steady-state, the electric current I (between the anode and cathode)
is independent of x (otherwise, charge will build up at some points). In fact,
(5.131)
I = (x) v(x) A,
where A is the cross-sectional area of the diode. The previous three equations
can be combined to give
d2
I
me
=
dx2
0 A 2 e
!1/2
1/2 .
(5.132)
The solution of the above equation which satisfies the boundary conditions is
x
d
!4/3
4 0 A 2 e
I=
9 d2 m e
!1/2
V0 .
= V0
with
177
(5.133)
3/2
(5.134)
5 ELECTROSTATICS
5.10 The method of images
This relationship between the current and the voltage in a vacuum diode is called
the Child-Langmuir law.
Let us now consider the solution of Poissons equation in more than one dimension.

Suppose that we have a point charge q held a distance d from an infinite, grounded,
conducting plate. Let the plate lie in the x-y plane, and suppose that the point
charge is located at coordinates (0, 0, d). What is the scalar potential above the
plane? This is not a simple question because the point charge induces surface
charges on the plate, and we do not know how these are distributed.
What do we know in this problem? We know that the conducting plate is
an equipotential surface. In fact, the potential of the plate is zero, since it is
grounded. We also know that the potential at infinity is zero (this is our usual
boundary condition for the scalar potential). Thus, we need to solve Poissons
equation in the region z > 0, for a single point charge q at position (0, 0, d),
subject to the boundary conditions
(z = 0) = 0,
(5.135)
(5.136)
and
as x2 + y2 + z2 . Let us forget about the real problem, for a moment, and
concentrate on a slightly different one. We refer to this as the analogue problem.
In the analogue problem, we have a charge q located at (0, 0, d) and a charge
q located at (0, 0, -d), with no conductors present. We can easily find the scalar
potential for this problem, since we know where all the charges are located. We
get

q
1
q
q
q
.
analogue (x, y, z) =
4 0
x2 + y2 + (z d)2
x2 + y2 + (z + d)2
(5.137)
178
5 ELECTROSTATICS
Note, however, that

analogue (z = 0) = 0,
(5.138)
analogue 0
(5.139)
and
as x2 + y2 + z2 . In addition, analogue satisfies Poissons equation for a charge
at (0, 0, d), in the region z > 0. Thus, analogue is a solution to the problem posed
earlier, in the region z > 0. Now, the uniqueness theorem tells us that there is
only one solution to Poissons equation which satisfies a given, well-posed set of
boundary conditions. So, analogue must be the correct potential in the region
z > 0. Of course, analogue is completely wrong in the region z < 0. We know
this because the grounded plate shields the region z < 0 from the point charge,
so that = 0 in this region. Note that we are leaning pretty heavily on the
uniqueness theorem here! Without this theorem, it would be hard to convince a
skeptical person that = analogue is the correct solution in the region z > 0.
Now that we know the potential in the region z > 0, we can easily work out the
distribution of charges induced on the conducting plate. We already know that
the relation between the electric field immediately above a conducting surface
and the density of charge on the surface is
(5.140)
E = .
0
In this case,
E = Ez (z = 0+ ) =
(z = 0+ )
analogue (z = 0+ )
=
,
z
z
so
= 0
analogue (z = 0+ )
.
z

q
(z d)
(z + d)
,
=
+
z
4 0 [x2 + y2 + (z d)2 ]3/2 [x2 + y2 + (z + d)2 ]3/2
so
(x, y) =
qd
.
2 (x2 + y2 + d2 )3/2
179
(5.141)
(5.142)
(5.143)
(5.144)
5 ELECTROSTATICS
Clearly, the charge induced on the plate has the opposite sign to the point charge.
The charge density on the plate is also symmetric about the z-axis, and is largest
where the plate is closest to the point charge. The total charge induced on the
plate is
Z
dS,
(5.145)
2 r dr
,
(r2 + d2 )3/2
(5.146)
Q=
xy plane
which yields
qd
Q=
2
Z
0
where r2 = x2 + y2 . Thus,
Z
1
dk
qd
= q.
Q=
= qd
2 0 (k + d2 )3/2
(k + d2 )1/2 0
(5.147)
So, the total charge induced on the plate is equal and opposite to the point charge
which induces it.
Our point charge induces charges of the opposite sign on the conducting plate.
This, presumably, gives rise to a force of attraction between the charge and the
plate. What is this force? Well, since the potential, and, hence, the electric field,
in the vicinity of the point charge is the same as in the analogue problem, then
the force on the charge must be the same as well. In the analogue problem, there
are two charges q a net distance 2 d apart. The force on the charge at position
(0, 0, d) (i.e., the real charge) is
F=
1
q2
^z.
4 0 (2 d)2
(5.148)
What, finally, is the potential energy of the system. For the analogue problem
this is just
1 q2
.
(5.149)
Wanalogue =
4 0 2 d
Note that the fields on opposite sides of the conducting plate are mirror images of
one another in the analogue problem. So are the charges (apart from the change
in sign). This is why the technique of replacing conducting surfaces by imaginary
180
5 ELECTROSTATICS
charges is called the method of images. We know that the potential energy of a
set of charges is equivalent to the energy stored in the electric field. Thus,
Z
0
W=
E2 dV.
(5.150)
2 all space
In the analogue problem, the fields on either side of the x-y plane are mirror
images of one another, so E2 (x, y, z) = E2 (x, y, z). It follows that
Z
0
Wanalogue = 2
E2analogue dV.
(5.151)
2 z>0
In the real problem
So,
giving
0
W=
2
E(z > 0) = Eanalogue (z > 0),
(5.152)
E(z < 0) = 0.
(5.153)
0
E2 dV =
2
z>0
z>0
E2analogue dV =
1
Wanalogue ,
2
1 q2
.
W=
4 0 4 d
(5.154)
(5.155)
There is another method by which we can obtain the above result. Suppose
that the charge is gradually moved towards the plate along the z-axis from infinity
until it reaches position (0, 0, d). How much work is required to achieve this?
We know that the force of attraction acting on the charge is
1 q2
Fz =
.
4 0 4 z2
(5.156)
Thus, the work required to move this charge by dz is

1 q2
dW = Fz dz =
dz.
4 0 4 z2
The total work needed to move the charge from z = to z = d is
d
Zd 2
1 q2
1
q
1 q2
=
W=
dz
=
.
4 0 4 z2
4 0
4z
4 0 4 d
181
(5.157)
(5.158)
5 ELECTROSTATICS
Of course, this work is equivalent to the potential energy we evaluated earlier,

and is, in turn, the same as the energy contained in the electric field.
As a second example of the method of images, consider a grounded spherical
conductor of radius a placed at the origin. Suppose that a charge q is placed
outside the sphere at (b, 0, 0), where b > a. What is the force of attraction
between the sphere and the charge? In this case, we proceed by considering an
analogue problem in which the sphere is replaced by an image charge q 0 placed
somewhere on the x-axis at (c, 0, 0). The electric potential throughout space in
the analogue problem is simply
q
1
1
q0
=
.
2
2
2
1/2
2
4 0 [(x b) + y + z ]
4 0 [(x c) + y2 + z2 ]1/2
(5.159)
The image charge is chosen so as to make the surface = 0 correspond to the

surface of the sphere. Setting the above expresion to zero, and performing a little
algebra, we find that the = 0 surface satisfies
x2 +
c2 b 2
2 (c b)
x + y2 + z2 =
,
1
1
(5.160)
where = q 0 2 /q2 . Of course, the surface of the sphere satisfies

x2 + y 2 + z 2 = a 2 .
(5.161)
The above two equations can be made identical by setting = c/b and a 2 = b2 ,
or
a
(5.162)
q 0 = q,
b
and
a2
c= .
(5.163)
b
According to the uniqueness theorem, the potential in the analogue problem is
now identical with that in the real problem, outside the sphere. (Of course, in
the real problem, the potential inside the sphere is zero.) Hence, the force of
attraction between the sphere and the original charge in the real problem is the
same as the force of attraction between the two charges in the analogue problem.
It follows that
q2
ab
q q0
=
.
(5.164)
f=
4 0 (b c)2
4 0 (b2 a2 )2
182
5 ELECTROSTATICS
5.11 Complex analysis
There are many other image problems, each of which involves replacing a
conductor with an imaginary charge (or charges) which mimics the electric field
in some region (but not everywhere). Unfortunately, we do not have time to
discuss any more of these problems.

Let us now investigate another trick for solving Poissons equation (actually it
only solves Laplaces equation). Unfortunately, this method can only be applied
in two dimensions.
The complex variable is conventionally written
z = x + iy
(5.165)
(z should not be confused with a z-coordinate: this is a strictly two-dimensional

problem). We can write functions F(z) of the complex variable just like we would
write functions of a real variable. For instance,
F(z) = z2 ,
1
F(z) = .
z
(5.166)
(5.167)
For a given function, F(z), we can substitute z = x + i y and write

F(z) = U(x, y) + i V(x, y),
(5.168)
where U and V are two real two-dimensional functions. Thus, if

F(z) = z2 ,
(5.169)
F(x + i y) = (x + i y)2 = (x2 y2 ) + 2 i x y,
(5.170)
U(x, y) = x2 y2 ,
(5.171)
V(x, y) = 2 x y.
(5.172)
then
giving
183
5 ELECTROSTATICS
We can define the derivative of a complex function in just the same manner as
we would define the derivative of a real function. Thus,
dF
F(z + z) F(z)
= lim |z|
.
(5.173)
dz
z
However, we now have a slight problem. If F(z) is a well-defined function (we
shall leave it to the mathematicians to specify exactly what being well-defined
entails: suffice to say that most functions we can think of are well-defined) then it
should not matter from which direction in the complex plane we approach z when
taking the limit in Eq. (5.173). There are, of course, many different directions we
could approach z from, but if we look at a regular complex function, F(z) = z 2
(say), then
dF
= 2z
(5.174)
dz
is perfectly well-defined, and is, therefore, completely independent of the details
of how the limit is taken in Eq. (5.173).
The fact that Eq. (5.173) has to give the same result, no matter which path
we approach z from, means that there are some restrictions on the functions U
and V in Eq. (5.168). Suppose that we approach z along the real axis, so that
z = x. Then,
dF
U(x + x, y) + i V(x + x, y) U(x, y) i V(x, y)
= lim |x|0
dz
x
V
U
+i
.
(5.175)
=
x
x
Suppose that we now approach z along the imaginary axis, so that z = i y.
Then,
dF
U(x, y + y) + i V(x, y + y) U(x, y) i V(x, y)
= lim |y|0
dz
i y
U V
= i
+
.
(5.176)
y
y
If F(z) is a well-defined function then its derivative must also be well-defined,
which implies that the above two expressions are equivalent. This requires that
V
U
=
,
(5.177)
x
y
184
5 ELECTROSTATICS
V
U
=
.
x
y
(5.178)
These are called the Cauchy-Riemann relations, and are, in fact, sufficient to ensure that all possible ways of taking the limit (5.173) give the same answer.
So far, we have found that a general complex function F(z) can be written
F(z) = U(x, y) + i V(x, y),
(5.179)
where z = x + i y. If F(z) is well-defined then U and V automatically satisfy the

Cauchy-Riemann relations. But, what has all of this got to do with electrostatics?
Well, we can combine the two Cauchy-Riemann relations. We get
and
2 U
V
U
V
=
=
,
=
x2
x y
y x
y y
(5.180)
2 V
U
U
V
=
,
x2
x y
y x
y y
(5.181)
2 U 2 U
+
= 0,
x2
y2
(5.182)
2 V 2 V
+
= 0.
x2
y2
(5.183)
which reduce to
Thus, both U and V automatically satisfy Laplaces equation in two dimensions;

i.e., both U and V are possible two-dimensional scalar potentials in free space.
Consider the two-dimensional gradients of U and V:
U =
V =
Now
U V =
U U
,
,
x y
!
V V
,
.
x y
!
U V U V
+
.
x x
y y
185
(5.184)
(5.185)
(5.186)
5 ELECTROSTATICS
It follows from the Cauchy-Riemann relations that

U V =
V V V V
= 0.
y x
x y
(5.187)
Thus, the contours of U are everywhere perpendicular to the contours of V. It

follows that if U maps out the contours of some free space scalar potential then
V indicates the directions of the associated electric field-lines, and vice versa.
y
U = -1
U
V
U=1
U=1
U=0
U=0
U = -1
Figure 45:
For every well-defined complex function we can think of, we get two sets of
free space potentials, and the associated electric field-lines. For example, consider
the function F(z) = z2 , for which
U = x 2 y2 ,
(5.188)
V = 2 x y.
(5.189)
These are, in fact, the equations of two sets of orthogonal hyperboloids. So,
U(x, y) (the solid lines in Fig. 45) might represent the contours of some scalar
potential and V(x, y) (the dashed lines in Fig. 45) the associated electric field
lines, or vice versa. But, how could we actually generate a hyperboloidal potential? This is easy. Consider the contours of U at level 1. These could represent
the surfaces of four hyperboloid conductors maintained at potentials V. The
186
5 ELECTROSTATICS
scalar potential in the region between these conductors is given by V U(x, y), and
the associated electric field-lines follow the contours of V(x, y). Note that
Ex =
= V
= 2 V x
x
x
(5.190)
Thus, the x-component of the electric field is directly proportional to the distance
from the x-axis. Likewise, for y-component of the field is directly proportional to
the distance from the y-axis. This property can be exploited to make devices
(called quadrupole electrostatic lenses) which are useful for focusing particle
beams.
As a second example, consider the complex function
c2
F(z) = z ,
z
(5.191)
where c is real and positive. Writing F(z) = U(x, y) + i V(x, y), we find that
c2 x
U(x, y) = x 2
.
x + y2
(5.192)
Far from the origin, U x, which is the potential of a uniform electric field, of
unit amplitude, pointing in the x-direction. The locus of U = 0 is x = 0, and
x2 + y 2 = c 2 ,
(5.193)
which corresponds to a circle of radius c centered on the origin. Hence, we

conclude that the potential
(x, y, z) = E0 U(x, y) = E0 x + E0 c2
x2
x
+ y2
(5.194)
corresponds to that outside a grounded, infinitely long, conducting cylinder of

radius c, running parallel to the z-axis, placed in a uniform x-directed electric
field of magnitude E0 . Of course, the potential inside the cylinder (i.e., x2 + y2 <
c2 ) is zero. The induced charge density on the surface of the cylinder is simply
= 0 Er (r = c) = 0
(r = c)
= 2 0 E0 cos ,
r
187
(5.195)
5 ELECTROSTATICS
5.12 Separation of variables
where r2 = x2 + y2 , and x = r cos . Note that zero net charge is induced on the
surface.
We can think of the set of all possible well-defined complex functions as a reference library of solutions to Laplaces equation in two dimensions. We have only
considered a couple of examples, but there are, of course, very many complex
functions which generate interesting potentials. For instance, F(z) = z 1/2 generates the potential around a semi-infinite, thin, grounded, conducting plate placed
in an external field, whereas F(z) = z3/2 yields the potential outside a grounded,
rectangular, conducting corner under similar circumstances.

The method of images and complex analysis are two rather elegant techniques
for solving Poissons equation. Unfortunately, they both have an extremely limited
range of application. The final technique we shall discuss in this course, namely,
the separation of variables, is somewhat messy, but possess a far wider range of
application. Let us examine a specific example.
Consider two semi-infinite, grounded, conducting plates lying parallel to the xz plane, one at y = 0, and the other at y = (see Fig. 46). The left end, at x = 0,
is closed off by an infinite strip insulated from the two plates, and maintained
at a specified potential 0 (y). What is the potential in the region between the
plates?
conducting plate
y=
x=0
y=0
x
Figure 46:
We first of all assume that the potential is z-independent, since everything

188
5 ELECTROSTATICS
else in the problem is. This reduces the problem to two dimensions. Poissons
equation is written
2 2
+
=0
(5.196)
x2
y2
in the vacuum region between the conductors. The boundary conditions are
(x, 0) = 0,
(5.197)
(x, ) = 0
(5.198)
for x > 0, since the two plates are earthed, plus
for 0 y , and
(0, y) = 0 (y)
(5.199)
(x, y) 0
(5.200)
as x . The latter boundary condition is our usual one for the scalar potential
at infinity.
The central assumption in the method of separation of variables is that a multidimensional potential can be written as the product of one-dimensional potentials, so that
(x, y) = X(x) Y(y).
(5.201)
The above solution is obviously a very special one, and is, therefore, only likely
to satisfy a very small subset of possible boundary conditions. However, it turns
out that by adding together lots of different solutions of this form we can match
to general boundary conditions.
Substituting (5.201) into (5.196), we obtain
d2 Y
d2 X
+ X 2 = 0.
Y
dx2
dy
(5.202)
Let us now separate the variables: i.e., let us collect all of the x-dependent terms
on one side of the equation, and all of the y-dependent terms on the other side.
Thus,
1 d2 X
1 d2 Y
=
.
(5.203)
X dx2
Y dy2
189
5 ELECTROSTATICS
This equation has the form

f(x) = g(y),
(5.204)
where f and g are general functions. The only way in which the above equation can be satisfied, for general x and y, is if both sides are equal to the same
constant. Thus,
1 d2 X
1 d2 Y
2
=k =
.
(5.205)
X dx2
Y dy2
The reason why we write k2 , rather than k2 , will become apparent later on.
Equation (5.205) separates into two ordinary differential equations:
d2 X
= k2 X,
2
dx
d2 Y
= k2 Y.
2
dy
(5.206)
(5.207)
We know the general solution to these equations:

X = A exp(k x) + B exp(k x),
(5.208)
Y = C sin(k y) + D cos(k y),
(5.209)
giving
= [ A exp(k x) + B exp(k x) ] [C sin(k y) + D cos(k y)].
(5.210)
Here, A, B, C, and D are arbitrary constants. The boundary condition (5.200)

is automatically satisfied if A = 0 and k > 0. Note that the choice k 2 , instead of
k2 , in Eq. (5.205) facilitates this by making either grow or decay monotonically in the x-direction instead of oscillating. The boundary condition (5.197)
is automatically satisfied if D = 0. The boundary condition (5.198) is satisfied
provided that
sin(k ) = 0,
(5.211)
which implies that k is a positive integer, n (say). So, our solution reduces to
(x, y) = C exp(n x) sin(n y),
(5.212)
where B has been absorbed into C. Note that this solution is only able to satisfy
the final boundary condition (5.199) provided 0 (y) is proportional to sin(n y).
190
5 ELECTROSTATICS
Thus, at first sight, it would appear that the method of separation of variables
only works for a very special subset of boundary conditions. However, this is not
the case.
Now comes the clever bit! Since Poissons equation is linear, any linear combination of solutions is also a solution. We can therefore form a more general
solution than (5.212) by adding together lots of solutions involving different values of n. Thus,
X
Cn exp(n x) sin(n y),
(5.213)
(x, y) =
n=1
where the Cn are constants. This solution automatically satisfies the boundary
conditions (5.197), (5.198) and (5.200). The final boundary condition (5.199)
reduces to
X
Cn sin(n y) = 0 (y).
(5.214)
(0, y) =
n=1
The question now is what choice of the Cn fits an arbitrary function 0 (y)? To
answer this question we can make use of two very useful properties of the functions sin(n y). Namely, that they are mutually orthogonal, and form a complete
set. The orthogonality property of these functions manifests itself through the
relation
Z
(5.215)
sin(n y) sin(n 0 y) dy = nn 0 ,
2
0
where the function nn 0 = 1 if n = n 0 and 0 otherwise is called a Kroenecker delta.
The completeness property of sine functions means that any general function
0 (y) can always be adequately represented as a weighted sum of sine functions
with various different n values. Multiplying both sides of Eq. (5.214) by sin(n 0 y),
and integrating over y, we obtain
Z
Z
X
Cn sin(n y) sin(n 0 y) dy = 0 (y) sin(n 0 y) dy.
(5.216)
n=1
The orthogonality relation yields
X
Cn nn 0 = Cn 0 =
2 n=1
2
191
0 (y) sin(n 0 y) dy,
(5.217)
5 ELECTROSTATICS
so
Z
2
0 (y) sin(n y) dy.
(5.218)
Cn =
0
Thus, we now have a general solution to the problem for any driving potential
0 (y).
If the potential 0 (y) is constant then
Z
2 0
2 0
Cn =
[1 cos(n )],
sin(n y) dy =
0
n
(5.219)
giving
Cn = 0
for even n, and
Cn =
4 0
n
(5.220)
(5.221)
for odd n. Thus,

4 0 X exp(n x) sin(n y)
(x, y) =
.
n=1,3,5
n
(5.222)
In the above problem, we write the potential as the product of one-dimensional

functions. Some of these functions grow and decay monotonically (i.e., the exponential functions), and the others oscillate (i.e., the sinusoidal functions). The
success of the method depends crucially on the orthogonality and completeness
of the oscillatory functions. A set of functions fn (x) is orthogonal if the integral of
the product of two different members of the set over some range is always zero:
i.e.,
Z
b
fn (x) fm (x) dx = 0,
(5.223)
for n 6= m. A set of functions is complete if any other function can be expanded

as a weighted sum of them. It turns out that the scheme set out above can be
generalized to more complicated geometries. For instance, in spherical geometry, the monotonic functions are power law functions of the radial variable, and
the oscillatory functions are Legendre polynomials. The latter are both mutually
192
5 ELECTROSTATICS
orthogonal and form a complete set. There are also cylindrical, ellipsoidal, hyperbolic, toroidal, etc. coordinates. In all cases, the associated oscillating functions
are mutually orthogonal and form a complete set. This implies that the method
of separation of variables is of quite general applicability.
Finally, as a simple example of the solution of Poissons equation in spherical
geometry, let us consider the case of a conducting sphere of radius a, centered
on the origin, placed in a uniform z-directed electric field of magnitude E0 . The
scalar potential satisfies 2 = 0 for r a, with the boundary conditions
E0 r cos (giving E E0 ^z) as r , and = 0 at r = a. Here, r and
are spherical polar coordinates. Let us try the simplified separable solution
(r, ) = rm cos .
(5.224)
It is easily demonstrated that the above solution satisfies 2 = 0 provided m = 1

or 2. Thus, the most general solution of 2 which satisfies the boundary
condition at r is
(r, ) = E0 r cos + r2 cos .
(5.225)
The boundary condition at r = a is satisfied provided

= E 0 a3 .
(5.226)
Of course, = 0 inside the sphere (i.e., r < a). The charge sheet density induced
on the surface of the sphere is given by
= 0 Er (r = a) = 0
(r = a)
= 3 0 E0 cos .
r
193
(5.227)
6 DIELECTRIC AND MAGNETIC MEDIA
Dielectric and magnetic media
6.1 Introduction
In this section, we shall use Maxwells equations to investigate the effect of dielectric and magnetic media on electric and magnetic fields.
6.2 Polarization
The terrestrial environment is characterized by dielectric media (e.g., air, water)
which are, for the most part, electrically neutral, since they are made up of neutral atoms and molecules. However, if these atoms and molecules are placed in
an electric field then they tend to polarize. Suppose that when a given neutral
molecule is placed in an electric field E, the centre of charge of its constituent
electrons (whose total charge is q) is displaced by a distance d with respect to
the centre of charge of its constituent atomic nucleus. The dipole moment of the
molecule is defined p = q d. If there are N such molecules per unit volume then
the electric polarization P (i.e., the dipole moment per unit volume) is given by
P = N p. More generally,
X
P(r) =
Ni hpi i,
(6.1)
i
where hpi i is the average dipole moment of the ith type of molecule in the vicinity
of point r, and Ni is the average number of such molecules per unit volume at r.
Consider an infinitesimal cube of dielectric material with x-coordinates between x and x + dx, y-coordinates between y and y + dy, and z-coordinates
between z and z + dz. Suppose that the dielectric consists of electrically neutral
polar molecules, of varying number density N(r), whose electrons, charge q, displace a constant distance d from the nuclei, charge q. Thus, the dipole moment
per unit volume is P(r) = N(r) q d. Due to the polarization of the molecules,
a net charge N(x, y, z) q dx dy dz enters the bottom face of the cube, perpendicular to the x-axis, whilst a net charge N(x + dx, y, z) q dx dy dz leaves the top
face. Hence, the net charge acquired by the cube due to molecular polarization
194
6.2 Polarization
in the x-direction is dq = N(x + dx, y, z) q dx dy dz + N(x, y, z) q dx dy dz =

[N(x, y, z)/x] q dx dx dy dz = [Px (x, y, z)/x] dx dy dz. There are analogous contributions due to polarization in the y- and z-directions. Hence, the net
charge acquired by the cube due to molecular polarization is dq = [Px (x, y, z)/x+
Py (x, y, z)/y+Pz (x, y, z)/z] dx dy dz = (P) dx dy dz. Thus, it follows that
the charge density acquired by the cube due to molecular polarization is simply
P.
As explained above, it is easily demonstrated that any divergence of the polarization field P(r) of a dielectric medium gives rise to an effective charge density
b in the medium, where
b = P.
(6.2)
This charge density is attributable to bound charges (i.e., charges which arise from
the polarization of neutral atoms), and is usually distinguished from the charge
density f due to free charges, which represents a net surplus or deficit of electrons
in the medium. Thus, the total charge density in the medium is
= f + b .
(6.3)
It must be emphasized that both terms in this equation represent real physical
charge. Nevertheless, it is useful to make the distinction between bound and
free charges, especially when it comes to working out the energy associated with
electric fields in dielectric media.
Gauss law takes the differential form
f + b
=
.
E =
0
0
This expression can be rearranged to give
(6.4)
D = f ,
(6.5)
D = 0 E + P
(6.6)
where
is termed the electric displacement, and has the same dimensions as P (dipole
moment per unit volume). Gauss theorem tells us that
I
Z
DdS = f dV.
(6.7)
V
195
6.2 Polarization
In other words, the flux of D out of some closed surface S is equal to the total
free charge enclosed within that surface. Unlike the electric field E (which is
the force acting on a unit charge), or the polarization P (the dipole moment
per unit volume), the electric displacement D has no clear physical meaning.
The only reason for introducing this quantity is that it enables us to calculate
electric fields in the presence of dielectric materials without first having to know
the distribution of bound charges. However, this is only possible if we have a
constitutive relation connecting E and D. It is conventional to assume that the
induced polarization P is directly proportional to the electric field E, so that
P = 0 e E,
(6.8)
where e is termed the electric susceptibility of the medium. It follows that

D = 0 E,
(6.9)
= 1 + e
(6.10)
where
is termed the dielectric constant or relative permittivity of the medium. (Likewise,
0 is termed the permittivity of free space.) Note that is dimensionless. It follows
from Eqs. (6.5) and (6.9) that
E =
f
.
0
(6.11)
Thus, the electric fields produced by free charges in a uniform dielectric medium
are analogous to those produced by the same charges in a vacuum, except that
they are reduced by a factor . This reduction can be understood in terms of a
polarization of the atoms or molecules of the dielectric medium that produces
electric fields in opposition to those generated by the free charges. One immediate consequence of this is that the capacitance of a capacitor is increased by
a factor if the empty space between the electrodes is filled with a dielectric
medium of dielectric constant (assuming that fringing fields can be neglected).
It must be understood that Eqs. (6.8)(6.11) are just an approximation which
is generally found to hold under terrestrial conditions (provided that the fields are
196
6.3 Boundary conditions for E and D
not too large) for isotropic media. For anisotropic media (e.g., crystals), Eq. (6.9)
generalizes to
D = 0 E,
(6.12)
where is a second-rank tensor known as the dielectric tensor. For strong electric
fields, D ceases to vary linearly with E. Indeed, for sufficiently strong electric
fields, neutral molecules are disrupted, and the whole concept of a dielectric
medium becomes meaningless.
6.3 Boundary conditions for E and D

When the space surrounding a set of charges contains dielectric material of nonuniform dielectric constant then the electric field no longer has the same functional form as in vacuum. Suppose, for example, that the space is occupied by
two dielectric media whose uniform dielectric constants are 1 and 2 . What are
the boundary conditions on E and D at the interface between the two media?
Imagine a Gaussian pill-box enclosing part of the interface. The thickness of
the pill-box is allowed to tend towards zero, so that the only contribution to the
outward flux of D comes from the flat faces of the box, which are parallel to
the interface. Assuming that there is no free charge inside the pill-box (which
is reasonable in the limit in which the volume of the box tends to zero), then
Eq. (6.7) yields
D 2 D 1 = 0,
(6.13)
where D 1 is the component of the electric displacement in medium 1 which is
normal to the interface, etc. If the fields and charges are non time-varying then
the differential form of Faradays law yield E = 0, which gives the familiar
boundary condition (obtained by integrating around a small loop which straddles
the interface)
Ek 2 Ek 1 = 0.
(6.14)
Generally, there is a bound charge sheet on the interface whose density follows
from Gauss law:
b = 0 (E 2 E 1 ).
(6.15)
197
6.4 Boundary value problems with dielectrics
In conclusion, the normal component of the electric displacement, and the tangential component of the electric field, are both continuous across any interface
between two dielectric media.

Consider a point charge q embedded in a semi-infinite dielectric 1 a distance
d away from a plane interface which separates the first medium from another
semi-infinite dielectric 2 . The interface is assumed to coincide with the plane
z = 0. We need to find solutions to the equations
(6.16)
1 E =
0
for z > 0,
2 E = 0
(6.17)
for z < 0, and
E=0
(6.18)
1 Ez (z = 0+ ) = 2 Ez (z = 0 ),
(6.19)
Ex (z = 0+ ) = Ex (z = 0 ),
(6.20)
Ey (z = 0+ ) = Ey (z = 0 ).
(6.21)
everywhere, subject to the boundary conditions at z = 0 that
In order to solve this problem, we shall employ a slightly modified form of the
well-known method of images. Since E = 0 everywhere, the electric field can
be written in terms of a scalar potential. So, E = . Consider the region z > 0.
Let us assume that the scalar potential in this region is the same as that obtained
when the whole of space is filled with the dielectric 1 , and, in addition to the
real charge q at position A, there is a second charge q 0 at the image position A 0
(see Fig. 47). If this is the case, then the potential at some point P in the region
z > 0 is given by
0
q
1
q
(6.22)
(z > 0) =
+ ,
4 0 1 R1 R2
198

z=0
P
1
R1
R2
/
A
/
z >
Figure 47:
q
where R1 = r2 + (d z)2 and R2 = r2 + (d + z)2 , when written in terms of

cylindrical polar coordinates, (r, , z), aligned along the z-axis. Note that the
potential (6.22) is clearly a solution of Eq. (6.16) in the region z > 0. It gives
E = 0, with the appropriate singularity at the position of the point charge q.
Consider the region z < 0. Let us assume that the scalar potential in this region
is the same as that obtained when the whole of space is filled with the dielectric
2 , and a charge q 00 is located at the point A. If this is the case, then the potential
in this region is given by
(z < 0) =
1
q 00
.
4 0 2 R1
(6.23)
The above potential is clearly a solution of Eq. (6.17) in the region z < 0. It gives
E = 0, with no singularities.
It now remains to choose q 0 and q 00 in such a manner that the boundary conditions (6.19)(6.21) are satisfied. The boundary conditions (6.20) and (6.21) are
obviously satisfied if the scalar potential is continuous at the interface between
the two dielectric media:
(z = 0+ ) = (z = 0 ).
199
(6.24)
The boundary condition (6.19) implies a jump in the normal derivative of the
scalar potential across the interface:
(z = 0+ )
(z = 0 )
1
= 2
.
z
z
(6.25)
The first matching condition yields

q 00
q + q0
=
,
1
2
(6.26)
q q 0 = q 00 .
(6.27)
whereas the second gives

Here, use has been made of
1
z R1
z=0
1
=
z R2
=
z=0
(r2
d
.
+ d2 )3/2
(6.28)
Equations (6.26) and (6.27) imply that

2 1
q,
q =
2 + 1
!
2

2
q 00 =
q.
2 + 1
!
(6.29)
(6.30)
The bound charge density is given by b = P, However, inside either

dielectric, P = 0 e E, so P = 0 e E = 0, except at the point charge q.
Thus, there is zero bound charge density in either dielectric medium. However,
there is a bound charge sheet on the interface between the two dielectric media.
In fact, the density of this sheet is given by
b = 0 (Ez 1 Ez 2 )z=0 .
(6.31)
Hence,
b = 0
(z = 0+)
q 2 1
(z = 0)
d
0
=
.
z
z
2 1 (2 + 1 ) (r2 + d2 )3/2
(6.32)
In the limit 2 1 , the dielectric 2 behaves like a conducting medium (i.e.,

E 0 in the region z < 0), and the bound surface charge density on the interface
200
approaches that obtained in the case where the plane z = 0 coincides with a
conducting surface (see Sect. 5.10).
As a second example, consider a dielectric sphere of radius a, and uniform
dielectric constant , placed in a uniform z-directed electric field of magnitude
E0 . Suppose that the sphere is centered on the origin. Now, for an electrostatic
problem, we can always write E = . In the present problem, E = 0 both
inside and outside the sphere, since there are no free charges, and the bound
volume charge density is zero in a uniform dielectric medium (or a vacuum).
Hence, the scalar potential satisfies Laplaces equation, 2 = 0, throughout
space. Adopting spherical polar coordinates, (r, , ), aligned along the z-axis,
the boundary conditions are that E0 r cos as r , and that is wellbehaved at r = 0. At the surface of the sphere, r = a, the continuity of E k implies
that is continuous. Furthermore, the continuity of D = 0 E leads to the
matching condition

=
.
(6.33)
r r=a+
r r=a
Let us try separable solutions of the form rm cos . It is easily demonstrated
that such solutions satisfy Laplaces equation provided that m = 1 or m = 2.
Hence, the most general solution to Laplaces equation outside the sphere, which
satisfies the boundary condition at r , is
a3 cos
(r, ) = E0 r cos + E0
.
r2
(6.34)
Likewise, the most general solution inside the sphere, which satisfies the boundary condition at r = 0, is
(r, ) = E1 r cos .
(6.35)
The continuity of at r = a yields
E0 E 0 = E 1 .
(6.36)
Likewise, the matching condition (6.33) gives

E0 + 2 E 0 = E 1 .
201
(6.37)
Hence,
1
,
+2
3 E0
.
=
+2
=
E1
(6.38)
(6.39)
Note that the electric field inside the sphere is uniform, parallel to the external
electric field outside the sphere, and of magnitude E1 . Moreover, E1 < E0 , provided that > 1. Finally, the density of the bound charge sheet on the surface of
the sphere is
b = 0

1

cos .
=
3

0
r r=a+
r r=a
+2

(6.40)
As a final example, consider a spherical cavity, of radius a, in a uniform dielectric medium, of dielectric constant , in the presence of a z-directed electric field
of magnitude E0 . This problem is analogous to the previous problem, except that
the matching condition (6.33) becomes

=
.

r r=a+
r r=a

Hence,
1
,
1 + 2
3 E0
=
.
1 + 2
=
E1
(6.41)
(6.42)
(6.43)
Note that the field inside the cavity is uniform, parallel to the external electric
field outside the sphere, and of magnitude E1 . Moreover, E1 > E0 , provided that
> 1. The density of the bound charge sheet on the surface of the cavity is
b = 0
1

=
3

cos .
0
r r=a+
r r=a
1 + 2

202
(6.44)
6.5 Energy density within a dielectric medium
6.5 Energy density within a dielectric medium

Consider a system of free charges embedded in a dielectric medium. The increase
in the total energy when a small amount of free charge f is added to the system
is given by
Z
W =
f d3 r,
(6.45)
where the integral is taken over all space, and (r) is the electrostatic potential.
Here, it is assumed that the original charges and the dielectric are held fixed, so
that no mechanical work is performed. It follows from Eq. (6.5) that
Z
W = D d3 r,
(6.46)
where D is the change in the electric displacement associated with the charge
increment. Now the above equation can also be written
Z
Z
W = ( D) d3 r D d3 r,
(6.47)
giving
W =
DdS D d3 r,
(6.48)
where use has been made of Gauss theorem. If the dielectric medium is of finite
spatial extent then we can neglect the surface term to give
Z
Z
3
W = D d r = ED d3 r.
(6.49)
This energy increment cannot be integrated unless E is a known function of D.
Let us adopt the conventional approach, and assume that D = 0 E, where the
dielectric constant is independent of the electric field. The change in energy
associated with taking the displacement field from zero to D(r) at all points in
space is given by
Z
Z Z
D
W=
W=
Z ZE
0
ED d3 r,
W =
or
(6.50)
Z
1
0 (E2 ) 3
d r=
0 E2 d3 r,
2
2
203
(6.51)
6.6 Magnetization
which reduces to
Z
1
ED d3 r.
(6.52)
W=
2
Thus, the electrostatic energy density inside a dielectric medium is given by
U=
1
ED.
2
(6.53)
This is a standard result, and is often quoted in textbooks. Nevertheless, it is

important to realize that the above formula is only valid in dielectric media in
which the electric displacement D varies linearly with the electric field E. Note
that Eq. (6.53) is consistent with the expression (5.20) which we obtained earlier.
6.6 Magnetization
All matter is built up out of atoms, and each atom consists of electrons in motion.
The currents associated with this motion are termed atomic currents. Each atomic
current is a tiny closed circuit of atomic dimensions, and may therefore be appropriately described as a magnetic dipole. If the atomic currents of a given atom
all flow in the same plane then the atomic dipole moment is directed normal to
the plane (in the sense given by the right-hand rule), and its magnitude is the
product of the total circulating current and the area of the current loop. More
generally, if j(r) is the atomic current density at the point r then the magnetic
moment of the atom is
Z
1
r j d3 r,
(6.54)
m=
2
where the integral is over the volume of the atom. If there are N such atoms or
molecules per unit volume then the magnetization M (i.e., the magnetic dipole
moment per unit volume) is given by M = Nm. More generally,
X
M(r) =
Ni hmi i,
(6.55)
i
where hmi i is the average magnetic dipole moment of the ith type of molecule in
the vicinity of point r, and Ni is the average number of such molecules per unit
volume at r.
204
6.6 Magnetization
Consider a general medium which is made up of molecules which are polarizable and possess a net magnetic moment. It is easily demonstrated that any
circulation in the magnetization field M(r) gives rise to an effective current density jm in the medium. In fact,
jm = M.
(6.56)
This current density is called the magnetization current density, and is usually
distinguished from the true current density, jt , which represents the convection
of free charges in the medium. In fact, there is a third type of current called a
polarization current, which is due to the apparent convection of bound charges.
It is easily demonstrated that the polarization current density, j p , is given by
jp =
P
.
t
(6.57)
Thus, the total current density, j, in the medium takes the form
j = jt + M +
P
.
t
(6.58)
It must be emphasized that all terms on the right-hand side of the above equation
represent real physical currents, although only the first term is due to the motion
of real charges (over more than atomic dimensions).
The differential form of Ampères law is
B = 0 j + 0 0
E
,
t
(6.59)
which can also be written

B = 0 jt + 0 M + 0
D
,
t
(6.60)
where use has been made of the definition D = 0 E + P. The above expression
can be rearranged to give
D
H = jt +
,
(6.61)
t
where
B
H=
M
(6.62)
0
205
6.7 Magnetic susceptibility and permeability
is termed the magnetic intensity, and has the same dimensions as M (i.e., magnetic
dipole moment per unit volume). In a steady-state situation, Stokes theorem tell
us that
Z
I
(6.63)
Hdl = jt dS.
S
In other words, the line integral of H around some closed curve is equal to the flux
of true current through any surface attached to that curve. Unlike the magnetic
field B (which specifies the force q v B acting on a charge q moving with velocity v), or the magnetization M (the magnetic dipole moment per unit volume),
the magnetic intensity H has no clear physical meaning. The only reason for
introducing it is that it enables us to calculate fields in the presence of magnetic
materials without first having to know the distribution of magnetization currents.
However, this is only possible if we possess a constitutive relation connecting B
and H.
6.7 Magnetic susceptibility and permeability

In a large class of materials, there exists an approximately linear relationship
between M and H. If the material is isotropic then
M = m H,
(6.64)
where m is called the magnetic susceptibility. If m is positive then the material

is called paramagnetic, and the magnetic field is strengthened by the presence of
the material. On the other hand, if m is negative then the material is diamagnetic, and the magnetic field is weakened in the presence of the material. The
magnetic susceptibilities of paramagnetic and diamagnetic materials are generally extremely small. A few sample values are given in Table. 2.1
A linear relationship between M and H also implies a linear relationship between B and H. In fact, we can write
B = H,
1
(6.65)
Data obtained from the Handbook of Chemistry and Physics, Chemical Rubber Company Press, Boca Raton, FL.
206
6.8 Ferromagnetism
Material
m
Aluminium
2.3 105
Copper
0.98 105
Diamond
2.2 105
Tungsten
6.8 105
Hydrogen (1 atm) 0.21 108
Oxygen (1 atm)
209.0 108
Nitrogen (1 atm) 0.50 108
Table 2:
where
= 0 (1 + m )
(6.66)
is termed the magnetic permeability of the material in question.2 (Likewise, 0 is

termed the permeability of free space.) Note that has the same units as 0 . It is
clear from Table 2 that the permeabilities of common diamagnetic and paramagnetic materials do not differ substantially from the permeability of free space. In
fact, to all intents and purposes, the magnetic properties of such materials can be
safely neglected (i.e., = 0 ).
6.8 Ferromagnetism
There is, however, a third class of magnetic materials called ferromagnetic materials. Such materials are characterized by a possible permanent magnetization,
and generally have a profound effect on magnetic fields (i.e., 0 ). Unfortunately, ferromagnetic materials do not generally exhibit a linear dependence
between M and H, or B and H, so that we cannot employ Eqs. (6.64) and (6.65)
with constant values of m and . It is still expedient to use Eq. (6.65) as the definition of , with = (H). However, this practice can lead to difficulties under
certain circumstances. The permeability of a ferromagnetic material, as defined
by Eq. (6.65), can vary through the entire range of possible values from zero to
infinity, and may be either positive or negative. The most sensible approach is
to consider each problem involving ferromagnetic materials separately, try to determine which region of the B-H diagram is important for the particular case in
2
We could alternatively define a dimensionless relative magnetic permeability by writing B = 0 H.
207
6.8 Ferromagnetism
Figure 48:
hand, and then make approximations appropriate to this region.

First, let us consider an unmagnetized sample of ferromagnetic material. If
the magnetic intensity, which is initially zero, is increased monotonically, then
the B-H relationship traces out a curve such as that shown in Fig. 48. This is
called a magnetization curve. It is evident that the permeabilities derived from
the curve (according to the rule = B/H) are always positive, and show a wide
range of values. The maximum permeability occurs at the knee of the curve. In
some materials, this maximum permeability is as large as 105 0 . The reason for
the knee in the curve is that the magnetization M reaches a maximum value in
the material, so that
B = 0 (H + M)
(6.67)
continues to increase at large H only because of the 0 H term. The maximum
value of M is called the saturation magnetization of the material.
Next, consider a ferromagnetic sample magnetized by the above procedure. If
the magnetic intensity H is decreased, the B-H relation does not follow back down
the curve of Fig. 48, but instead moves along a new curve, sketched in Fig. 49,
to the point R. Thus, the magnetization, once established, does not disappear
with the removal of H. In fact, it takes a reversed magnetic intensity to reduce
208
6.8 Ferromagnetism
B
R
Figure 49:
the magnetization to zero. If H continues to build up in the reversed direction,

then M (and hence B) becomes increasingly negative. Finally, when H increases
again the operating point follows the lower curve of Fig. 49. Thus, the B-H curve
for increasing H is quite different to that for decreasing H. This phenomenon is
known as hysteresis.
The curve sketched in Fig. 49 called the hysteresis loop of the material in question. The value of B at the point R is called the retentivity or remanence. The
magnitude of H at the point C is called the coercivity. It is evident that is negative in the second and fourth quadrants of the loop, and positive in the first and
third quadrants. The shape of the hysteresis loop depends not only on the nature
of the ferromagnetic material, but also on the maximum value of H to which the
material has been subjected. However, once this maximum value, Hmax , becomes
sufficiently large to produce saturation in the material, the hysteresis loop does
not change shape with any further increase in Hmax .
Ferromagnetic materials are used either to channel magnetic flux (e.g., around
transformer circuits) or as sources of magnetic field (e.g., permanent magnets).
For use as a permanent magnet, the material is first magnetized by placing it in
a strong magnetic field. However, once the magnet is removed from the external field it is subject to a demagnetizing H. Thus, it is vitally important that a
permanent magnet should possess both a large remanence and a large coercivity.
209
6.9 Boundary conditions for B and H
As will become clear later on, it is generally a good idea for the ferromagnetic
materials used to channel magnetic flux around transformer circuits to possess
small remanences and small coercivities.
6.9 Boundary conditions for B and H

What are the boundary conditions for B and H at the interface between two
magnetic media? The governing equations for a steady-state situation are
B = 0,
(6.68)
H = jt .
(6.69)
B 2 B 1 = 0,
(6.70)
and
Integrating Eq. (6.68) over a Gaussian pill-box enclosing part of the interface
between the two media gives
where B denotes the component of B perpendicular to the interface. Integrating

Eq. (6.69) around a small loop which straddles the interface yields
Hk 2 Hk 1 = 0,
(6.71)
assuming that there is no true current sheet flowing in the interface. Here, H k
denotes the component of H parallel to the interface. In general, there is a magnetization current sheet flowing at the interface whose density is of amplitude
Jm =
Bk 2 B k 1
.
0
(6.72)
In conclusion, the normal component of the magnetic field and the tangential
component of the magnetic intensity are both continuous across any interface
between magnetic media.
210

Consider a ferromagnetic sphere, of uniform permeability , placed in a uniform
z-directed magnetic field of magnitude B0 . Suppose that the sphere is centred
on the origin. In the absence of any true currents, we have H = 0. Hence,
we can write H = m . Given that B = 0, and B = H, it follows that
2 m = 0 in any uniform magnetic medium (or a vacuum). Hence, 2 m = 0
throughout space. Adopting spherical polar coordinates, (r, , ), aligned along
the z-axis, the boundary conditions are that m (B0 /0 ) r cos at r ,
and that m is well-behaved at r = 0. At the surface of the sphere, r = a, the
continuity of Hk implies that m is continuous. Furthermore, the continuity of
B = H leads to the matching condition
m
m

0
=
.
r r=a+
r r=a

(6.73)
Let us try separable solutions of the form rm cos . It is easily demonstrated

that such solutions satisfy Laplaces equation provided that m = 1 or m = 2.
Hence, the most general solution to Laplaces equation outside the sphere, which
satisfies the boundary condition at r , is
a3 cos
.
m (r, ) = (B0 /0 ) r cos + (B0 /0 )
r2
(6.74)
Likewise, the most general solution inside the sphere, which satisfies the boundary condition at r = 0, is
m (r, ) = (B1 /) r cos .
(6.75)
The continuity of m at r = a yields

B0 B0 = (0 /) B1 .
(6.76)
Likewise, the matching condition (6.73) gives

B0 + 2 B 0 = B 1 .
211
(6.77)

iron core
a
wire
gap
Figure 50:
Hence,
0
,
+ 2 0
3 B0
=
.
+ 2 0
=
B1
(6.78)
(6.79)
Note that the magnetic field inside the sphere is uniform, parallel to the external
magnetic field outside the sphere, and of magnitude B1 . Moreover, B1 > B0 ,
provided that > 0 .
As a final example, consider an electromagnet of the form sketched in Fig. 50.
A wire, carrying a current I, is wrapped N times around a thin toroidal iron core
of radius a and permeability 0 . The core contains a thin gap of width
d. What is the magnetic field induced in the gap? Let us neglect any leakage
of magnetic field from the core, which is reasonable if 0 . We expect the
magnetic field, Bc , and the magnetic intensity, Hc , in the core to be both toroidal
and essentially uniform. It is also reasonable to suppose that the magnetic field,
Bg , and the magnetic intensity, Hg , in the gap are toroidal and uniform, since
d a. We have Bc = Hc and Bg = 0 Hg . Moreover, since the magnetic field
is normal to the interface between the core and the gap, the continuity of B
implies that
Bc = B g .
(6.80)
212
6.11 Magnetic energy
Thus, the magnetic field-strength in the core is the same as that in the gap.
However, the magnetic intensities in the core and the gap are quite different:
Hc = Bc / = Bg / = (0 /) Hg . Integration of Eq. (6.69) around the torus
yields
I
Z
H dl =
jt dS = N I.
(6.81)
Hence,
(2 a d) Hc + d Hg = N I.
(6.82)
NI
.
(2 a d)/ + d/0
(6.83)
It follows that
Bg =

Consider an electrical conductor. Suppose that a battery with an electromotive
field E 0 is feeding energy into this conductor. The energy is either dissipated as
heat, or is used to generate a magnetic field. Ohms law inside the conductor
gives
jt = (E + E 0 ),
(6.84)
where jt is the true current density, is the conductivity, and E is the inductive
electric field. Taking the scalar product with jt , we obtain
jt2
E jt =
Ejt .
(6.85)
The left-hand side of this equation represents the rate at which the battery does
work on the conductor. The first term on the right-hand side is the rate of ohmic
heating inside the conductor. We tentatively identify the remaining term with
the rate at which energy is fed into the magnetic field. If all fields are quasistationary (i.e., slowly varying) then the displacement current can be neglected,
and the differential form of Ampères law reduces to H = jt . Substituting this
expression into Eq. (6.85), and integrating over all space, we get
Z
Z
Z
( H)2 3
0
3
d r E( H) d3 r.
(6.86)
E ( H) d r =
213
The last term can be integrated by parts using the vector identity
(E H) H( E) E( H).
Gauss theorem plus the differential form of Faradays law yield
Z
Z
Z
B 3
3
d r (E H)dS.
E( H) d r = H
t
(6.87)
(6.88)
Since E H falls off at least as fast as 1/r5 in electrostatic and quasi-stationary

magnetic fields (1/r2 comes from electric monopole fields, and 1/r3 from magnetic dipole fields), the surface integral in the above expression can be neglected.
Of course, this is not the case for radiation fields, for which E and H fall off like
1/r. Thus, the constraint of quasi-stationarity effectively means that the fields
vary sufficiently slowly that any radiation fields can be neglected.
The total power expended by the battery can now be written
Z
Z
Z
2
(
H)
B 3
E 0 ( H) d3 r =
d3 r + H
d r.
(6.89)
The first term on the right-hand side has already been identified as the energy loss
rate due to ohmic heating. The last term is obviously the rate at which energy is
fed into the magnetic field. The variation W in the magnetic field energy can
therefore be written
Z
W = HB d3 r.
(6.90)
This result is analogous to the result (6.49) for the variation in the energy of an
electrostatic field.
In order to make Eq. (6.90) integrable, we must assume a functional relationship between H and B. For a medium which magnetizes linearly, the integration
can be carried out in much the same manner as Eq. (6.52), to give
Z
1
W=
HB d3 r.
(6.91)
2
Thus, the magnetostatic energy density inside a linear magnetic material is given
by
1
U = H B.
(6.92)
2
214
Unfortunately, most interesting magnetic materials, such as ferromagnets, exhibit

a nonlinear relationship between H and B. For such materials, Eq. (6.90) can
only be integrated between definite states, and the result, in general, depends on
the past history of the sample. For ferromagnets, the integral of Eq. (6.90) has
a finite, non-zero value when B is integrated around a complete magnetization
cycle. This cyclic energy loss is given by
ZI
W =
H dB d3 r.
(6.93)
In other words, the energy expended per unit volume when a magnetic material
is carried through a magnetization cycle is equal to the area of its hysteresis loop,
as plotted in a graph of B against H. Thus, it is particularly important to ensure
that the magnetic materials used to form transformer cores possess hysteresis
loops with comparatively small areas, otherwise the transformers are likely to be
extremely lossy.
215
7 MAGNETIC INDUCTION
Magnetic induction
7.1 Introduction
In this section, we shall use Maxwells equations to investigate magnetic induction and related phenomena.
7.2 Inductance
We have learned about e.m.f., resistance, and capacitance. Let us now investigate
inductance. Electrical engineers like to reduce all pieces of electrical apparatus to
an equivalent circuit consisting only of e.m.f. sources (e.g., batteries), inductors,
capacitors, and resistors. Clearly, once we understand inductors, we shall be
ready to apply the laws of electromagnetism to electrical circuits.
Consider two stationary loops of wire, labeled 1 and 2. Let us run a steady
current I1 around the first loop to produce a magnetic field B1 . Some of the field
lines of B1 will pass through the second loop. Let 2 be the flux of B1 through
loop 2:
Z
2 =
loop 2
B1 dS2 ,
(7.1)
where dS2 is a surface element of loop 2. This flux is generally quite difficult
to calculate exactly (unless the two loops have a particularly simple geometry).
However, we can infer from the Biot-Savart law,
I
0 I1
dl1 (r r 0 )
B1 (r) =
,
(7.2)
4 loop 1
|r r 0 |3
that the magnitude of B1 is proportional to the current I1 . This is ultimately a
consequence of the linearity of Maxwells equations. Here, dl1 is a line element
of loop 1 located at position vector r 0 . It follows that the flux 2 must also be
proportional to I1 . Thus, we can write
2 = M21 I1 ,
216
(7.3)
7.2 Inductance
where M21 is the constant of proportionality. This constant is called the mutual
inductance of the two loops.
Let us write the magnetic field B1 in terms of a vector potential A1 , so that
B1 = A 1 .
It follows from Stokes theorem that
Z
Z
2 =
B1 dS2 =
loop 2
loop 2
A1 dS2 =
(7.4)
I
loop 2
A1 dl2 ,
where dl2 is a line element of loop 2. However, we know that

I
0 I1
dl1
.
A1 (r) =
4 loop 1 |r r 0 |
The above equation is just a special case of the more general law,
Z
0
j(r 0 ) 3 0
A1 (r) =
dr,
4 all space |r r 0 |
(7.5)
(7.6)
(7.7)
for j(r 0 ) = dl1 I1 /dl1 dA and d3 r 0 = dl1 dA, where dA is the cross-sectional area
of loop 1. Thus,
I
I
dl1 dl2
0 I1
,
(7.8)
2 =
4 loop 1 loop 2 |r r 0 |
where r is now the position vector of the line element dl2 of loop 2. The above
equation implies that
I
I
0
dl1 dl2
M21 =
.
(7.9)
4 loop 1 loop 2 |r r 0 |
In fact, mutual inductances are rarely worked out from first principlesit is usually too difficult. However, the above formula tells us two important things.
Firstly, the mutual inductance of two loops is a purely geometric quantity, having to do with the sizes, shapes, and relative orientations of the loops. Secondly,
the integral is unchanged if we switch the roles of loops 1 and 2. In other words,
M21 = M12 .
217
(7.10)
7.3 Self-inductance
In fact, we can drop the subscripts, and just call these quantities M. This is a
rather surprising result. It implies that no matter what the shapes and relative
positions of the two loops, the magnetic flux through loop 2 when we run a
current I around loop 1 is exactly the same as the flux through loop 1 when we
send the same current around loop 2.
We have seen that a current I flowing around some loop, 1, generates a magnetic flux linking some other loop, 2. However, flux is also generated through the
first loop. As before, the magnetic field, and, therefore, the flux , is proportional
to the current, so we can write
= L I.
(7.11)
The constant of proportionality L is called the self-inductance. Like M it only
depends on the geometry of the loop.
Inductance is measured in S.I. units called henries (H): 1 henry is 1 volt-second
per ampere. The henry, like the farad, is a rather unwieldy unit, since most reallife inductors have a inductances of order a micro-henry.
7.3 Self-inductance
Consider a long solenoid of length l, and radius r, which has N turns per unit
length, and carries a current I. The longitudinal (i.e., directed along the axis of
the solenoid) magnetic field within the solenoid is approximately uniform, and is
given by
B = 0 N I.
(7.12)
This result is easily obtained by integrating Ampères law over a rectangular loop
whose long sides run parallel to the axis of the solenoid, one inside the solenoid,
and the other outside, and whose short sides run perpendicular to the axis. The
magnetic flux though each turn of the loop is B r2 = 0 N I r2 . The total flux
through the solenoid wire, which has N l turns, is
= N l 0 N I r2 .
218
(7.13)
7.3 Self-inductance
Thus, the self-inductance of the solenoid is
= 0 N2 r2 l.
(7.14)
I
Note that the self-inductance only depends on geometric quantities such as the
number of turns per unit length of the solenoid, and the cross-sectional area of
the turns.
L=
Suppose that the current I flowing through the solenoid changes. We have to
assume that the change is sufficiently slow that we can neglect the displacement
current, and retardation effects, in our calculations. This implies that the typical
time-scale of the change must be much longer than the time for a light-ray to
traverse the circuit. If this is the case, then the above formulae remain valid.
A change in the current implies a change in the magnetic flux linking the
solenoid wire, since = L I. According to Faradays law, this change generates
an e.m.f. in the wire. By Lenzs law, the e.m.f. is such as to oppose the change in
the currenti.e., it is a back e.m.f. We can write
V=
d
dI
= L ,
dt
dt
(7.15)
where V is the generated e.m.f.

Suppose that our solenoid has an electrical resistance R. Let us connect the
ends of the solenoid across the terminals of a battery of e.m.f. V. What is going
to happen? The equivalent circuit is shown in Fig. 51. The inductance and resistance of the solenoid are represented by a perfect inductor, L, and a perfect
resistor, R, connected in series. The voltage drop across the inductor and resistor
is equal to the e.m.f. of the battery, V. The voltage drop across the resistor is
simply I R, whereas the voltage drop across the inductor (i.e., the back e.m.f.) is
L dI/dt. Here, I is the current flowing through the solenoid. It follows that
dI
.
(7.16)
dt
This is a differential equation for the current I. We can rearrange it to give
V = IR + L
dI R
V
+ I= .
dt L
L
219
(7.17)
7.3 Self-inductance
V
R
I
Figure 51:
The general solution is

I(t) =
V
+ k exp(R t/L).
R
(7.18)
The constant k is fixed by the boundary conditions. Suppose that the battery is
connected at time t = 0, when I = 0. It follows that k = V/R, so that
I(t) =
V
[1 exp(R t/L) ] .
R
(7.19)
This curve is sketched in Fig. 52. It can be seen that, after the battery is connected, the current ramps up, and attains its steady-state value V/R (which comes
from Ohms law), on the characteristic time-scale
=
L
.
R
(7.20)
This time-scale is sometimes called the time constant of the circuit, or, somewhat
unimaginatively, the L over R time of the circuit.
We can now appreciate the significance of self-inductance. The back e.m.f.
generated in an inductor, as the current tries to change, effectively prevents the
current from rising (or falling) much faster than the L/R time. This effect is
220
7.3 Self-inductance
I
V/R
L/R
Figure 52:
sometimes advantageous, but often it is a great nuisance. All circuit elements

possess some self-inductance, as well as some resistance, and thus have a finite
L/R time. This means that when we power up a circuit, the current does not
jump up instantaneously to its steady-state value. Instead, the rise is spread out
over the L/R time of the circuit. This is a good thing. If the current were to
rise instantaneously, then extremely large electric fields would be generated by
the sudden jump in the induced magnetic field, leading, inevitably, to breakdown
and electric arcing. So, if there were no such thing as self-inductance, then every
time you switched an electric circuit on or off there would be a blue flash due
to arcing between conductors. Self-inductance can also be a bad thing. Suppose
that we possess a fancy power supply, and we wish to use it to send an electric
signal down a wire (or transmission line). Of course, the wire or transmission
line will possess both resistance and inductance, and will, therefore, have some
characteristic L/R time. Suppose that we try to send a square-wave signal down
the wire. Since the current in the wire cannot rise or fall faster than the L/R
time, the leading and trailing edges of the signal get smoothed out over an L/R
time. The typical difference between the signal fed into the wire (upper trace),
and that which comes out of the other end (lower trace), is illustrated in Fig. 53.
Clearly, there is little point having a fancy power supply unless you also possess
221
7.4 Mutual inductance
V
0
V
0
Figure 53:
a low inductance wire or transmission line, so that the signal from the power
supply can be transmitted to some load device without serious distortion.

Consider, now, two long thin solenoids, one wound on top of the other. The length
of each solenoid is l, and the common radius is r. Suppose that the bottom coil
has N1 turns per unit length, and carries a current I1 . The magnetic flux passing
through each turn of the top coil is 0 N1 I1 r2 , and the total flux linking the
top coil is therefore 2 = N2 l 0 N1 I1 r2 , where N2 is the number of turns per
unit length in the top coil. It follows that the mutual inductance of the two coils,
defined 2 = M I1 , is given by
M = 0 N1 N2 r2 l.
(7.21)
Recall that the self-inductance of the bottom coil is

L1 = 0 N12 r2 l,
(7.22)
L2 = 0 N22 r2 l.
(7.23)
and that of the top coil is

222
Hence, the mutual inductance can be written

q
M = L1 L2 .
(7.24)
Note that this result depends on the assumption that all of the flux produced by
one coil passes through the other coil. In reality, some of the flux leaks out, so that
the mutual inductance is somewhat less than that given in the above formula. We
can write
q
M = k L1 L2 ,
(7.25)
where the constant k is called the coefficient of coupling, and lies in the range
0 k 1.
Suppose that the two coils have resistances R1 and R2 . If the bottom coil has
an instantaneous current I1 flowing through it, and a total voltage drop V1 , then
the voltage drop due to its resistance is I1 R1 . The voltage drop due to the back
e.m.f. generated by the self-inductance of the coil is L1 dI1 /dt. There is also a
back e.m.f. due to inductive coupling with the top coil. We know that the flux
through the bottom coil due to the instantaneous current I2 flowing in the top
coil is
1 = M I 2 .
(7.26)
Thus, by Faradays law and Lenzs law, the e.m.f. induced in the bottom coil is
dI2
.
(7.27)
dt
The voltage drop across the bottom coil due to its mutual inductance with the top
coil is minus this expression. Thus, the circuit equation for the bottom coil is
V = M
dI2
dI1
+M
.
dt
dt
Likewise, the circuit equation for the top coil is
V1 = R 1 I 1 + L 1
dI2
dI1
+M
.
dt
dt
Here, V2 is the total voltage drop across the top coil.
V2 = R 2 I 2 + L 2
(7.28)
(7.29)
Suppose that we suddenly connect a battery of e.m.f. V1 to the bottom coil,

at time t = 0. The top coil is assumed to be open-circuited, or connected to
223
a voltmeter of very high internal resistance, so that I2 = 0. What is the e.m.f.

generated in the top coil? Since I2 = 0, the circuit equation for the bottom coil is
V1 = R 1 I 1 + L 1
dI1
,
dt
(7.30)
where V1 is constant, and I1 (t = 0) = 0. We have already seen the solution to

this equation:
V1
[1 exp(R1 t/L1 ) ] .
(7.31)
I1 =
R1
The circuit equation for the top coil is
dI1
,
dt
(7.32)
M
exp(R1 t/L1 ).
L1
(7.33)
v
u
u L2
t
exp(R1 t/L1 ).
(7.34)
N2
exp(R1 t/L1 ).
N1
(7.35)
V2 = M
giving
V2 = V 1
V2 = V 1 k
Since L1 /L2 = N12 /N22 , we obtain
V2 = V 1 k
L1
Note that V2 (t) is discontinuous at t = 0. This is not a problem, since the resistance of the top circuit is infinite, so there is no discontinuity in the current (and,
hence, in the magnetic field). But, what about the displacement current, which
is proportional to E/t? Surely, this is discontinuous at t = 0 (which is clearly
unphysical)? The crucial point, here, is that we have specifically neglected the
displacement current in all of our previous analysis, so it does not make much
sense to start worrying about it now. If we had retained the displacement current in our calculations, then we would have found that the voltage in the top
circuit jumps up, at t = 0, on a time-scale similar to the light traverse time across
the circuit (i.e., the jump is instantaneous to all intents and purposes, but the
displacement current remains finite).
224
7.5 Magnetic energy
Now,
N2
V2 (t = 0)
=k
,
(7.36)
V1
N1
so if N2 N1 then the voltage in the bottom circuit is considerably amplified in
the top circuit. This effect is the basis for old-fashioned car ignition systems. A
large voltage spike is induced in a secondary circuit (connected to a coil with very
many turns) whenever the current in a primary circuit (connected to a coil with
not so many turns) is either switched on or off. The primary circuit is connected
to the car battery (whose e.m.f. is typically 12 volts). The switching is done by
a set of points, which are mechanically opened and closed as the engine turns.
The large voltage spike induced in the secondary circuit, as the points are either
opened or closed, causes a spark to jump across a gap in this circuit. This spark
ignites a petrol/air mixture in one of the cylinders. We might think that the
optimum configuration is to have only one turn in the primary circuit, and lots
of turns in the secondary circuit, so that the ratio N2 /N1 is made as large as
possible. However, this is not the case. Most of the magnetic field lines generated
by a single turn primary coil are likely to miss the secondary coil altogether.
This means that the coefficient of coupling k is small, which reduces the voltage
induced in the secondary circuit. Thus, we need a reasonable number of turns in
the primary coil in order to localize the induced magnetic field, so that it links
effectively with the secondary coil.
7.5 Magnetic energy

Suppose that at t = 0 a coil of inductance, L, and resistance, R, is connected
across the terminals of a battery of e.m.f., V. The circuit equation is
dI
+ R I.
(7.37)
dt
The power output of the battery is V I. [Every charge q that goes around the
circuit falls through a potential difference q V. In order to raise it back to the
starting potential, so that it can perform another circuit, the battery must do
work q V. The work done per unit time (i.e., the power) is n q V, where n is
the number of charges per unit time passing a given point on the circuit. But,
V =L
225
7.5 Magnetic energy
I = n q, so the power output is V I.] The total work done by the battery in raising
the current in the circuit from zero at time t = 0 to IT at time t = T is
ZT
W = VI dt.
(7.38)
0
Using the circuit equation (7.37), we obtain

ZT
ZT
dI
dt + R I2 dt,
W=L I
0
0 dt
(7.39)
giving
ZT
1
2
(7.40)
W = L IT + R I2 dt.
2
0
The second term on the right-hand side represents the irreversible conversion of
electrical energy into heat energy in the resistor. The first term is the amount of
energy stored in the inductor at time T . This energy can be recovered after the inductor is disconnected from the battery. Suppose that the battery is disconnected
at time T . The circuit equation is now
0=L
dI
+ RI,
dt
(7.41)
giving
R
I = IT exp (t T ) ,
(7.42)
L
where we have made use of the boundary condition I(T ) = IT . Thus, the current
decays away exponentially. The energy stored in the inductor is dissipated as heat
in the resistor. The total heat energy appearing in the resistor after the battery is
disconnected is
Z
1
(7.43)
I2 R dt = L IT2 ,
2
T
where use has been made of Eq. (7.42). Thus, the heat energy appearing in the
resistor is equal to the energy stored in the inductor. This energy is actually stored
in the magnetic field generated around the inductor.
"
Consider, again, our circuit with two coils wound on top of one another. Suppose that each coil is connected to its own battery. The circuit equations are
226
7.5 Magnetic energy
thus
dI1
dI2
+M
,
dt
dt
dI1
dI2
+M
,
(7.44)
V2 = R 2 I 2 + L 2
dt
dt
where V1 is the e.m.f. of the battery in the first circuit, etc. The work done by the
two batteries in increasing the currents in the two circuits, from zero at time 0,
to I1 and I2 at time T , respectively, is
ZT
(V1 I1 + V2 I2 ) dt
W =
V1 = R 1 I 1 + L 1
ZT
(R1 I12 + R2 I22 ) dt +
+M
ZT
dI2
dI1
I1
+ I2
dt
dt
ZT
(R1 I12 + R2 I22 ) dt
Thus,
1
1
L1 I12 + L2 I22
2
2
W =
dt.
(7.45)
1
1
+ L1 I12 + L2 I22 + M I1 I2 .
2
2
Clearly, the total magnetic energy stored in the two coils is
(7.46)
1
1
(7.47)
L1 I12 + L2 I22 + M I1 I2 .
2
2
Note that the mutual inductance term increases the stored magnetic energy if
I1 and I2 are of the same signi.e., if the currents in the two coils flow in the
same direction, so that they generate magnetic fields which reinforce one another.
Conversely, the mutual inductance term decreases the stored magnetic energy if
I1 and I2 are of the opposite sign. However, the total stored energy can never be
negative, otherwise the coils would constitute a power source (a negative stored
energy is equivalent to a positive generated energy). Thus,
WB =
1
1
L1 I12 + L2 I22 + M I1 I2 0,
2
2
227
(7.48)
7.5 Magnetic energy
which can be written

2
q
q
1 q
L1 I1 + L2 I2 I1 I2 ( L1 L2 M) 0,
2
(7.49)
assuming that I1 I2 < 0. It follows that

q
M L1 L2 .
(7.50)
The equality sign corresponds to the situation where all of the flux generated by
one coil passes through the other. If some of the flux misses then the inequality
sign is appropriate. In fact, the above formula is valid for any two inductively
coupled circuits.
We intimated previously that the energy stored in an inductor is actually stored
in the surrounding magnetic field. Let us now obtain an explicit formula for the
energy stored in a magnetic field. Consider an ideal solenoid. The energy stored
in the solenoid when a current I flows through it is
W=
1 2
LI ,
2
(7.51)
where L is the self-inductance. We know that

L = 0 N2 r2 l,
(7.52)
where N is the number of turns per unit length of the solenoid, r the radius, and
l the length. The field inside the solenoid is uniform, with magnitude
B = 0 N I,
(7.53)
and is zero outside the solenoid. Equation (7.51) can be rewritten

B2
W=
V,
2 0
(7.54)
where V = r2 l is the volume of the solenoid. The above formula strongly

suggests that a magnetic field possesses an energy density
B2
U=
.
2 0
228
(7.55)
7.5 Magnetic energy
Let us now examine a more general proof of the above formula. Consider a
system of N circuits (labeled i = 1 to N), each carrying a current Ii . The magnetic
flux through the ith circuit is written [cf., Eq. (7.5) ]
Z
I
i = B dSi = A dli ,
(7.56)
where B = A, and dSi and dli denote a surface element and a line element
of this circuit, respectively. The back e.m.f. induced in the ith circuit follows from
Faradays law:
di
.
(7.57)
Vi =
dt
The rate of work of the battery which maintains the current Ii in the ith circuit
against this back e.m.f. is
di
.
(7.58)
Pi = I i
dt
Thus, the total work required to raise the currents in the N circuits from zero at
time 0, to I0 i at time T , is
N ZT
X
di
W=
Ii
dt.
(7.59)
dt
0
i=1
The above expression for the work done is, of course, equivalent to the total
energy stored in the magnetic field surrounding the various circuits. This energy
is independent of the manner in which the currents are set up. Suppose, for the
sake of simplicity, that the currents are ramped up linearly, so that
t
Ii = I 0 i .
T
(7.60)
The fluxes are proportional to the currents, so they must also ramp up linearly:
t
i = 0 i .
T
It follows that
W=
N ZT
X
i=1
I 0 i 0 i
229
(7.61)
t
dt,
T2
(7.62)
7.5 Magnetic energy
giving
N
1X
I 0 i 0 i .
W=
2 i=1
(7.63)
So, if instantaneous currents Ii flow in the the N circuits, which link instantaneous fluxes i , then the instantaneous stored energy is
N
1X
W=
I i i .
2 i=1
(7.64)
Equations (7.56) and (7.64) imply that

I
N
1X
W=
Ii A dli .
2 i=1
(7.65)
(A B) B A A B,
(7.68)
It is convenient, at this stage, to replace our N line currents by N current distributions of small, but finite, cross-sectional area. Equation (7.65) transforms
to
Z
1
W=
A j dV,
(7.66)
2 V
where V is a volume which contains all of the circuits. Note that for an element
of the ith circuit j = Ii dli /dli Ai and dV = dli Ai , where Ai is the cross-sectional
area of the circuit. Now, 0 j = B (we are neglecting the displacement current
in this calculation), so
Z
1
W=
A B dV.
(7.67)
2 0 V
According to vector field theory,
which implies that

1
W=
2 0
[ (A B) + B A] dV.
Using Gauss theorem, and B = A, we obtain

I
Z
1
1
A B dS +
B2 dV,
W=
2 0 S
2 0 V
230
(7.69)
(7.70)
where S is the bounding surface of V. Let us take this surface to infinity. It is

easily demonstrated that the magnetic field generated by a current loop falls of
like r3 at large distances. The vector potential falls off like r2 . However, the
area of surface S only increases like r2 . It follows that the surface integral is
negligible in the limit r . Thus, the above expression reduces to
Z
B2
dV.
(7.71)
W=
all space 2 0
Since this expression is valid for any magnetic field whatsoever, we can safely
conclude that the energy density of a general magnetic field is given by
B2
.
U=
2 0
(7.72)
Note, also, that the above expression is consistent with the expression (6.92)
which we obtained during our investigation of magnetic media.

Alternating current (AC) circuits are made up of e.m.f. sources and three different
types of passive element: resistors, inductors, and capacitors, Resistors satisfy
Ohms law:
V = I R,
(7.73)
where R is the resistance, I the current flowing through the resistor, and V the
voltage drop across the resistor (in the direction in which the current flows).
Inductors satisfy
dI
V=L ,
(7.74)
dt
where L is the inductance. Finally, capacitors obey
Zt
,
q
V = = I dt C,
(7.75)
C
0
where C is the capacitance, q is the charge stored on the plate with the most
positive potential, and I = 0 for t < 0. Note that any passive component of a real
231
electrical circuit can always be represented as a combination of ideal resistors,

inductors, and capacitors.
Let us consider the classic LCR circuit, which consists of an inductor, L, a
capacitor, C, and a resistor, R, all connected in series with an e.m.f. source, V.
The circuit equation is obtained by setting the input voltage V equal to the sum
of the voltage drops across the three passive elements in the circuit. Thus,
Zt
,
dI
(7.76)
V = IR + L
+ I dt C.
dt
0
This is an integro-differential equation which, in general, is quite tricky to solve.

Suppose, however, that both the voltage and the current oscillate at some fixed
angular frequency , so that
V(t) = V0 exp(i t),
(7.77)
I(t) = I0 exp(i t),
(7.78)
where the physical solution is understood to be the real part of the above expressions. The assumed behaviour of the voltage and current is clearly relevant to
electrical circuits powered by the mains voltage (which oscillates at 60 hertz).
Equations (7.76)(7.78) yield
V0 exp(i t) = I0 exp(i t)R + L i I0 exp(i t) +
I0 exp(i t)
,
iC
(7.79)
giving
1
+R .
V0 = I 0 i L +
iC
It is helpful to define the impedance of the circuit:
!
(7.80)
V
1
= iL +
+ R.
(7.81)
I
iC
Impedance is a generalization of the concept of resistance. In general, the impedance
of an AC circuit is a complex quantity.
Z=
The average power output of the e.m.f. source is

P = hV(t) I(t)i,
232
(7.82)
where the average is taken over one period of the oscillation. Let us, first of all,
calculate the power using real (rather than complex) voltages and currents. We
can write
V(t) = |V0 | cos( t),
I(t) = |I0 | cos( t ),
(7.83)
(7.84)
where is the phase-lag of the current with respect to the voltage. It follows that
Z t=2
d( t)
cos( t) cos( t )
P = |V0 | |I0 |
2
t=0
Z t=2
d( t)
,(7.85)
= |V0 | |I0 |
cos( t) [cos( t) cos + sin( t) sin ]
2
t=0
giving
1
|V0 | |I0 | cos ,
(7.86)
2
since hcos( t) sin( t)i = 0 and hcos( t) cos( t)i = 1/2. In complex representation, the voltage and the current are written
P=
V(t) = |V0 | exp(i t),
Note that
(7.87)
I(t) = |I0 | exp[i ( t )].
(7.88)
1
(V I + V I) = |V0 | |I0 | cos .
2
(7.89)
It follows that
1
1
P = (V I + V I) = Re(V I ).
4
2
Making use of Eq. (7.81), we find that
1
1 Re(Z) |V|2
2
P = Re(Z) |I| =
.
2
2
|Z|2
(7.90)
(7.91)
Note that power dissipation is associated with the real part of the impedance. For
the special case of an LCR circuit,
P=
1
R |I0 | 2 .
2
233
(7.92)
7.7 Transmission lines
It is clear that only the resistor dissipates energy in this circuit. The inductor
and the capacitor both store energy, but they eventually return it to the circuit
without dissipation.
According to Eq. (7.81), the amplitude of the current which flows in an LCR
circuit for a given amplitude of the input voltage is given by
|I0 | =
|V0 |
=
|Z|
|V0 |
.
( L 1/ C)2 + R2
(7.93)
L C, and
reachThe response
of
the
circuit
is
clearly
resonant,
peaking
at
=
1/
q
ing 1/ 2 of the peak value at = 1/ L CR/2 L (assuming that R L/C). In

fact, LCR circuits are used in radio tuners to filter out signals whose frequencies
fall outside a given band.
The phase-lag of the current with respect to the voltage is given by
= arg(Z) = tan
1 L
1/ C
.
R
(7.94)
The phase-lag varies from /2 for frequencies significantly

below the resonant
frequency, to zero at the resonant frequency ( = 1/ L C), to /2 for frequencies

significantly above the resonant frequency.
It is clear that in conventional AC circuits the circuit equation reduces to a
simple algebraic equation, and the behaviour of the circuit is summed up by
the complex impedance Z. The real part of Z tells us the power dissipated in
the circuit, the magnitude of Z gives the ratio of the peak current to the peak
voltage, and the argument of Z gives the phase-lag of the current with respect to
the voltage.

The central assumption made in the analysis of conventional AC circuits is that
the voltage (and, hence, the current) has the same phase throughout the circuit.
Unfortunately, if the circuit is sufficiently large, or the frequency of oscillation,
234
, is sufficiently high, then this assumption becomes invalid. The assumption of

a constant phase throughout the circuit is reasonable if the wave-length of the
oscillation, = 2 c/, is much larger than the dimensions of the circuit. (Here,
we assume that signals propagate around electrical circuits at about the velocity
of light. This assumption will be justified later on.) This is generally not the case
in electrical circuits which are associated with communication. The frequencies in
such circuits tend to be very high, and the dimensions are, almost by definition,
large. For instance, leased telephone lines (the type you attach computers to) run
at 56 kHz. The corresponding wave-length is about 5 km, so the constant-phase
approximation clearly breaks down for long-distance calls. Computer networks
generally run at about 100 MHz, corresponding to 3 m. Thus, the constantphase approximation also breaks down for most computer networks, since such
networks are generally significantly larger than 3 m. It turns out that you need
a special sort of wire, called a transmission line, to propagate signals around
circuits whose dimensions greatly exceed the wave-length, . Let us investigate
transmission lines.
An idealized transmission line consists of two parallel conductors of uniform
cross-sectional area. The conductors possess a capacitance per unit length, C,
and an inductance per unit length, L. Suppose that x measures the position along
the line.
Consider the voltage difference between two neighbouring points on the line,
located at positions x and x + x, respectively. The self-inductance of the portion
of the line lying between these two points is L x. This small section of the line can
be thought of as a conventional inductor, and, therefore, obeys the well-known
equation
I(x, t)
,
(7.95)
V(x, t) V(x + x, t) = L x
t
where V(x, t) is the voltage difference between the two conductors at position x
and time t, and I(x, t) is the current flowing in one of the conductors at position
x and time t [the current flowing in the other conductor is I(x, t) ]. In the limit
x 0, the above equation reduces to
I
V
= L .
x
t
235
(7.96)
Consider the difference in current between two neighbouring points on the

line, located at positions x and x+x, respectively. The capacitance of the portion
of the line lying between these two points is C x. This small section of the line
can be thought of as a conventional capacitor, and, therefore, obeys the wellknown equation
Zt
Zt
I(x, t) dt I(x + x, t) dt = C x V(x, t),
(7.97)
0
where t = 0 denotes a time at which the charge stored in either of the conductors
in the region x to x + x is zero. In the limit x 0, the above equation yields
I
V
= C
.
(7.98)
x
t
Equations (7.96) and (7.98) are generally known as the telegraphers equations,
since an old fashioned telegraph line can be thought of as a primitive transmission
line (telegraph lines consist of a single wire: the other conductor is the Earth.)
Differentiating Eq. (7.96) with respect to x, we obtain
2 V
2 I
=
L
.
x2
x t
Differentiating Eq. (7.98) with respect to t yields
2 I
2 V
= C 2 .
x t
t
The above two equations can be combined to give
2 V
2 V
LC 2 =
.
t
x2
(7.99)
(7.100)
(7.101)
This is clearly a wave equation, with wave velocity v = 1/ L C. An analogous

equation can be written for the current, I.
Consider a transmission line which is connected to a generator at one end
(x = 0), and a resistor, R, at the other (x = l). Suppose that the generator
outputs a voltage V0 cos( t). If follows that
V(0, t) = V0 cos( t).
236
(7.102)
The solution to the wave equation (7.101), subject to the above boundary condition, is
V(x, t) = V0 cos( t k x),
(7.103)
where k = /v. This clearly corresponds to a wave which propagates from the
generator towards the resistor. Equations (7.96) and (7.103) yield
I(x, t) =
V0
cos( t k x).
L/C
(7.104)
For self-consistency, the resistor at the end of the line must have a particular
value:
v
u
L
V(l, t) u
=t .
(7.105)
R=
I(l, t)
C
The so-called input impedance of the line is defined
Zin =
V(0, t)
=
I(0, t)
v
u
uL
t
(7.106)
.
q
Thus, a transmission line terminated by a resistor R = L/C acts very much like
a conventional resistor R = Zin in the circuit containing the generator. In fact, the
transmission line could be replaced by an effective resistor R = Zin in the circuit
diagram for the generator circuit. The power loss due to this effective resistor
corresponds to power which is extracted from the circuit, transmitted down the
line, and absorbed by the terminating resistor.
The most commonly occurring type of transmission line is a co-axial cable,
which consists of two co-axial cylindrical conductors of radii a and b (with b >
a). We have already shown that the capacitance per unit length of such a cable
is (see Sect. 5.6)
2 0
.
(7.107)
C=
ln(b/a)
Let us now calculate the inductance per unit length. Suppose that the inner
conductor carries a current I. According to Ampères law, the magnetic field in
the region between the conductors is given by
B =
0 I
.
2 r
237
(7.108)
The flux linking unit length of the cable is

Zb
0 I
ln(b/a).
= B dr =
2
a
(7.109)
Thus, the self-inductance per unit length is

0
=
ln(b/a).
I
2
L=
(7.110)
So, the speed of propagation of a wave down a co-axial cable is

v=
1
1
=
= c.
0 0
LC
(7.111)
Not surprisingly, the wave (which is a type of electromagnetic wave) propagates

at the speed of light. The impedance of the cable is given by
Z0 =
v
u
uL
t
0
=
C
42 0
!1/2
ln (b/a) = 60 ln (b/a) ohms.
(7.112)
If we fill the region between the two cylindrical conductors with a dielectric
of dielectric constant , then, according to the discussion in Sect. 6.2, the capacitance per unit length of the transmission line goes up by a factor . However,
the dielectric has no effect on magnetic fields, so the inductance per unit length
of the line remains unchanged. It follows that the propagation speed of signals
down a dielectric filled co-axial cable is
1
c
v=
= .

LC
(7.113)
As we shall see later, this is simply the propagation velocity of electromagnetic

waves through a dielectric medium. The impedance of the cable becomes
Z0 = 60
ln(b/a)
ohms.

(7.114)
We have seen that if a transmission line is terminated by a resistor whose

resistance R matches the impedance Z0 of the line then all of the power sent
238
down the line is absorbed by the resistor. What happens if R 6= Z 0 ? The answer
is that some of the power is reflected back down the line. Suppose that the
beginning of the line lies at x = l, and the end of the line is at x = 0. Let us
consider a solution
V(x, t) = V0 exp[i ( t k x)] + K V0 exp[i ( t + k x)].
(7.115)
This corresponds to a voltage wave of amplitude V0 which travels down the line,
and is reflected at the end of the line, with reflection coefficient K. It is easily
demonstrated from the telegraphers equations that the corresponding current
waveform is
I(x, t) =
V0
K V0
exp[i ( t k x)]
exp[i ( t + k x)].
Z0
Z0
(7.116)
Since the line is terminated by a resistance R at x = 0, we have, from Ohms law,

V(0, t)
= R.
I(0, t)
(7.117)
This yields an expression for the coefficient of reflection,

K=
R Z0
.
R + Z0
(7.118)
The input impedance of the line is given by

Zin =
R cos(k l) + i Z0 sin(k l)
V(l, t)
= Z0
.
I(l, t)
Z0 cos(k l) + i R sin(k l)
(7.119)
Clearly, if the resistor at the end of the line is properly matched, so that R = Z 0 ,
then there is no reflection (i.e., K = 0), and the input impedance of the line is Z 0 .
If the line is short-circuited, so that R = 0, then there is total reflection at the end
of the line (i.e., K = 1), and the input impedance becomes
Zin = i Z0 tan(k l).
(7.120)
This impedance is purely imaginary, implying that the transmission line absorbs
no net power from the generator circuit. In fact, the line acts rather like a pure
inductor or capacitor in the generator circuit (i.e., it can store, but cannot absorb,
239
energy). If the line is open-circuited, so that R , then there is again total

reflection at the end of the line (i.e., K = 1), and the input impedance becomes
Zin = i Z0 tan(k l /2).
(7.121)
Thus, the open-circuited line acts like a closed-circuited line which is shorter by
one quarter of a wave-length. For the special case where the length of the line is
exactly one quarter of a wave-length (i.e., k l = /2), we find
Z02
Zin =
.
R
(7.122)
Thus, a quarter-wave line looks like a pure resistor in the generator circuit. Finally, if the length of the line is much less than the wave-length (i.e., k l 1)
then we enter the constant-phase regime, and Zin ' R (i.e., we can forget about
the transmission line connecting the terminating resistor to the generator circuit).
Suppose that we wish to build a radio transmitter. We can use a standard
half-wave antenna (i.e., an antenna whose length is half the wave-length of the
emitted radiation) to emit the radiation. In electrical circuits, such an antenna
acts like a resistor of resistance 73 ohms (it is more usual to say that the antenna
has an impedance of 73 ohmssee Sect. 9.2). Suppose that we buy a 500 kW
generator to supply the power to the antenna. How do we transmit the power
from the generator to the antenna? We use a transmission line, of course. (It
is clear that if the distance between the generator and the antenna is of order
the dimensions of the antenna (i.e., /2) then the constant-phase approximation
breaks down, and so we have to use a transmission line.) Since the impedance
of the antenna is fixed at 73 ohms, we need to use a 73 ohm transmission line
(i.e., Z0 = 73 ohms) to connect the generator to the antenna, otherwise some of
the power we send down the line is reflected (i.e., not all of the power output of
the generator is converted into radio waves). If we wish to use a co-axial cable
to connect the generator to the antenna, then it is clear from Eq. (7.114) that the
radii of the inner and outer conductors need to be such that b/a = 3.38 exp( ).
Suppose, finally, that we upgrade our transmitter to use a full-wave antenna
(i.e., an antenna whose length equals the wave-length of the emitted radiation).
A full-wave antenna has a different impedance than a half-wave antenna. Does
240
this mean that we have to rip out our original co-axial cable, and replace it by
one whose impedance matches that of the new antenna? Not necessarily. Let Z 0
be the impedance of the co-axial cable, and Z1 the impedance of the antenna.
Suppose that we place a quarter-wave transmission line (i.e., one
whose length is
one quarter of a wave-length) of characteristic impedance Z1/4 = Z0 Z1 between
the
end of the cable and the antenna. According to Eq. (7.122) (with Z0
Z0 Z1 and R Z1 ), the input impedance of the quarter-wave line is Zin = Z0 ,
which matches that of the cable. The output impedance matches that of the
antenna. Consequently, there is no reflection of the power sent down the cable
to the antenna. A quarter-wave line of the appropriate impedance can easily be
fabricated from a short length of co-axial cable of the appropriate b/a.
241
8 ELECTROMAGNETIC ENERGY AND MOMENTUM
Electromagnetic energy and momentum
8.1 Introduction
In this section, we shall demonstrate that Maxwells equations conserve both energy and momentum.
8.2 Energy conservation

We have seen that the energy density of an electric field is given by [see Eq. (5.20)]
0 E 2
,
UE =
2
(8.1)
whereas the energy density of a magnetic field satisfies [see Eq. (7.55)]
UB =
B2
.
2 0
(8.2)
This suggests that the energy density of a general electromagnetic field is

U=
0 E 2
B2
+
.
2
2 0
(8.3)
We are now in a position to demonstrate that the classical theory of electromagnetism conserves energy. We have already come across one conservation law in
electromagnetism:
+ j = 0.
(8.4)
t
This is the equation of charge conservation. Integrating over some volume V,
bounded by a surface S, and making use of Gauss theorem, we obtain
Z
I
dV = j dS.
(8.5)
t V
S
In other words, the rate of decrease of the charge contained in volume V equals
the net flux of charge across surface S. This suggests that an energy conservation
242
law for electromagnetism should have the form

Z
I
U dV = u dS.
t V
S
(8.6)
Here, U is the energy density of the electromagnetic field, and u is the flux of
electromagnetic energy (i.e., energy |u| per unit time, per unit cross-sectional
area, passes a given point in the direction of u). According to the above equation,
the rate of decrease of the electromagnetic energy in volume V equals the net flux
of electromagnetic energy across surface S.
However, Eq. (8.6) is incomplete, because electromagnetic fields can gain or
lose energy by interacting with matter. We need to factor this into our analysis.
We saw earlier (see Sect. 5.3) that the rate of heat dissipation per unit volume in
a conductor (the so-called ohmic heating rate) is E j. This energy is extracted
from electromagnetic fields, so the
R rate of energy loss of the fields in volume V
due to interaction with matter is V E j dV. Thus, Eq. (8.6) generalizes to
Z
I
Z
U dV = u dS + E j dV.
(8.7)
t V
S
V
From Gauss theorem, the above equation is equivalent to
U
+ u = E j.
(8.8)
t
Let us now see if we can derive an expression of this form from Maxwells equations.
We start from the differential form of Ampères law (including the displacement current):
E
(8.9)
B = 0 j + 0 0 .
t
Dotting this equation with the electric field yields
Ej=
EB
E
+ 0 E
.
0
t
(8.10)
This can be rewritten

EB
0 E 2
.
Ej=
+
0
t
2
243
(8.11)
Now, from vector field theory

(E B) B E E B,
so
(8.12)
EB
BE
0 E 2
E j =
+
.
0
0
t
2
(8.13)
The differential form of Faradays law yields

E=
so
B
,
t
(8.14)
1
B
0 E 2
EB
+ B
+
E j =
.
0
0
t t
2
(8.15)
B2
EB
0 E 2
+
.
+
E j =
0
t
2
2 0
(8.16)
U
+ u = E j,
t
(8.17)
This can be rewritten
Thus, we obtain the desired conservation law,
where
0 E 2
B2
+
2
2 0
is the electromagnetic energy density, and
U=
u=
EB
0
(8.18)
(8.19)
is the electromagnetic energy flux. The latter quantity is usually called the Poynting flux, after its discoverer.
Let us see whether our expression for the electromagnetic energy flux makes
sense. We all know that if we stand in the sun we get hot (especially in Texas!).
This occurs because we absorb electromagnetic radiation emitted by the Sun.
So, radiation must transport energy. The electric and magnetic fields in electromagnetic radiation are mutually perpendicular, and are also perpendicular to the
244
^ (this is a unit vector). Furthermore, B = E/c. Equation

direction of propagation k
(4.78) can easily be transformed into the following relation between the electric
and magnetic fields of an electromagnetic wave:
EB=
E2 ^
k.
c
(8.20)
Thus, the Poynting flux for electromagnetic radiation is

E2 ^
^
u=
k = 0 c E2 k.
0 c
(8.21)
This expression tells us that electromagnetic waves transport energy along their
direction of propagation, which seems to make sense.
The energy density of electromagnetic radiation is
U=
0 E 2
B2
E2
0 E 2
+
+
= 0 E2 ,
=
2
2
2 0
2
2 0 c
(8.22)
using B = E/c. Note that the electric and magnetic fields have equal energy
densities. Since electromagnetic waves travel at the speed of light, we would
expect the energy flux through one square meter in one second to equal the
energy contained in a volume of length c and unit cross-sectional area: i.e., c
times the energy density. Thus,
|u| = c U = 0 c E2 ,
(8.23)
which is in accordance with Eq. (8.21).

In the presence of diamagnetic and magnetic media, starting from Eq. (6.58),
we can derive an energy conservation law of the form
U
+ u = E jt ,
t
(8.24)
via analogous steps to those used to derive Eq. (8.17). Here, the electromagnetic
energy density is written
1
1
U = E D + B H,
(8.25)
2
2
245
8.3 Electromagnetic momentum
L
E
E
x
Figure 54:
which is consistent with Eq. (8.18). The Poynting flux takes the form
u = E H,
(8.26)
which is consistent with Eq. (8.19). Of course, the above expressions are only
valid for linear dielectric and magnetic media.

We have seen that electromagnetic waves carry energy. It turns out that they also
carry momentum. Consider the following argument, due to Einstein. Suppose
that we have a railroad car of mass M and length L which is free to move in one
dimension (see Fig. 54). Suppose that electromagnetic radiation of total energy
E is emitted from one end of the car, propagates along the length of the car, and is
then absorbed at the other end. The effective mass of this radiation is m = E/c 2
(from Einsteins famous relation E = m c2 ). At first sight, the process described
above appears to cause the centre of mass of the system to spontaneously shift.
246
This violates the law of momentum conservation (assuming the railway car is
subject to no external forces). The only way in which the centre of mass of the
system can remain stationary is if the railway car moves in the opposite direction
to the direction of propagation of the radiation. In fact, if the car moves by a
distance x then the centre of mass of the system is the same before and after the
radiation pulse provided that
Mx = mL =
E
L.
c2
(8.27)
It is assumed that m M in this derivation.

But, what actually causes the car to move? If the radiation possesses momentum p then the car will recoil with the same momentum as the radiation is
emitted. When the radiation hits the other end of the car then the car acquires
momentum p in the opposite direction, which stops the motion. The time of flight
of the radiation is L/c. So, the distance traveled by a mass M with momentum p
in this time is
pL
,
(8.28)
x = vt =
Mc
giving
E
c
(8.29)
p = Mx = .
L
c
Thus, the momentum carried by electromagnetic radiation equals its energy divided by the speed of light. The same result can be obtained from the well-known
relativistic formula
E2 = p 2 c2 + m 2 c4
(8.30)
relating the energy E, momentum p, and mass m of a particle. According to
quantum theory, electromagnetic radiation is made up of massless particles called
photons. Thus,
E
p=
(8.31)
c
for individual photons, so the same must be true of electromagnetic radiation as
a whole. If follows from Eq. (8.29) that the momentum density g of electromagnetic radiation equals its energy density over c, so
U |u| 0 E2
g= = 2 =
.
c
c
c
247
(8.32)
It is reasonable to suppose that the momentum points along the direction of the
energy flow (this is obviously the case for photons), so the vector momentum
density (which gives the direction, as well as the magnitude, of the momentum
per unit volume) of electromagnetic radiation is
g=
u
.
c2
(8.33)
Thus, the momentum density equals the energy flux over c2 .

Of course, the electric field associated with an electromagnetic wave oscillates
rapidly, which implies that the previous expressions for the energy density, energy flux, and momentum density of electromagnetic radiation are also rapidly
oscillating. It is convenient to average over many periods of the oscillation (this
average is denoted h i). Thus,
0 E02
hUi =
,
2
c 0 E02 ^
^
hui =
k = c hUi k,
2
0 E02 ^ hUi ^
k=
k,
hgi =
2c
c
(8.34)
(8.35)
(8.36)
where the factor 1/2 comes from averaging cos2 ( t). Here, E0 is the peak amplitude of the electric field associated with the wave.
Since electromagnetic radiation possesses momentum then it must exert a
force on bodies which absorb (or emit) radiation. Suppose that a body is placed
in a beam of perfectly collimated radiation, which it absorbs completely. The
amount of momentum absorbed per unit time, per unit cross-sectional area, is
simply the amount of momentum contained in a volume of length c and unit
cross-sectional area: i.e., c times the momentum density g. An absorbed momentum per unit time, per unit area, is equivalent to a pressure. In other words, the
radiation exerts a pressure c g on the body. Thus, the radiation pressure is given
by
0 E 2
p=
= hUi.
(8.37)
2
248
So, the pressure exerted by collimated electromagnetic radiation is equal to its

average energy density.
Consider a cavity filled with electromagnetic radiation. What is the radiation
pressure exerted on the walls? In this situation, the radiation propagates in all
directions with equal probability. Consider radiation propagating at an angle
to the local normal to the wall. The amount of such radiation hitting the wall
per unit time, per unit area, is proportional to cos . Moreover, the component of
momentum normal to the wall which the radiation carries is also proportional to
cos . Thus, the pressure exerted on the wall is the same as in Eq. (8.37), except
that it is weighted by the average of cos2 over all solid angles, in order to take
into account the fact that obliquely propagating radiation exerts a pressure which
is cos2 times that of normal radiation. The average of cos2 over all solid angles
is 1/3, so for isotropic radiation
p=
hUi
.
3
(8.38)
Clearly, the pressure exerted by isotropic radiation is one third of its average
energy density.
The power incident on the surface of the Earth due to radiation emitted by the
Sun is about 1300 W m2 . So, what is the radiation pressure? Since,
h|u|i = c hUi = 1300 W m2 ,
(8.39)
p = hUi ' 4 106 N m2 .
(8.40)
then
Here, the radiation is assumed to be perfectly collimated. Thus, the radiation
pressure exerted on the Earth is minuscule (one atmosphere equals about 105
N m2 ). Nevertheless, this small pressure due to radiation is important in outer
space, since it is responsible for continuously sweeping dust particles out of the
Solar System. It is quite common for comets to exhibit two separate tails. One
(called the gas tail) consists of ionized gas, and is swept along by the solar wind
(a stream of charged particles and magnetic field-lines emitted by the Sun). The
other (called the dust tail) consists of uncharged dust particles, and is swept radially outward from the Sun by radiation pressure. Two separate tails are observed
249
8.4 Momentum conservation
if the local direction of the solar wind is not radially outward from the Sun (which
is quite often the case).
The radiation pressure from sunlight is very weak. However, that produced by
laser beams can be enormous (far higher than any conventional pressure which
has ever been produced in a laboratory). For instance, the lasers used in Inertial
Confinement Fusion (e.g., the NOVA experiment in Lawrence Livermore National
Laboratory) typically have energy fluxes of 1018 W m2 . This translates to a radiation pressure of about 104 atmospheres!

It follows from Eqs. (8.19) and (8.33) that the momentum density of electromagnetic fields can be written
g = 0 E B.
(8.41)
Now, a momentum conservation equation for electromagnetic fields should take

the integral form
Z
Z
Z
gi dV = Gij dSj + [ E + j B]i dV.

(8.42)
t V
S
V
Here, i and j run from 1 to 3 (1 corresponds to the x-direction, 2 to the ydirection, and 3 to the z-direction). Moreover, the Einstein summation convention is employed for repeated indices (e.g., aj aj a a). Furthermore, the tensor
Gij represents the flux of the ith component of electromagnetic momentum in the
j-direction. This tensor (a tensor is a direct generalization of a vector with two
indices instead of one) is called the momentum flux density. Hence, the above
equation states that the rate of loss of electromagnetic momentum in some volume V is equal to the flux of electromagnetic momentum across the bounding
surface S plus the rate at which momentum is transferred to matter inside V. The
latter rate is, of course, just the net electromagnetic force acting on matter inside
V: i.e., the volume integral of the electromagnetic force density, E + j B. Now,
a direct generalization of the divergence theorem states that
Z
Z
Gij
dV,
(8.43)
Gij dSj
x
j
S
V
250
where x1 x, x2 y, etc. Hence, in differential form, our momentum conservation equation for electromagnetic fields is written
Gij
+ [ E + j B]i .
[0 E B]i =
t
xj
(8.44)
Let us now attempt to derive an equation of this form from Maxwells equations.
Maxwells equations are written:
,
0
B = 0,
B
E = ,
t
E =
B = 0 j + 0 0
(8.45)
(8.46)
(8.47)
E
.
t
(8.48)
We can cross Eq. (8.48) divided by 0 with B, and rearrange, to give

0
E
B ( B)
B=
+ j B.
t
0
(8.49)
Next, let us cross E with Eq. (8.47) times 0 , rearrange, and add the result to the
above equation. We obtain
0
E
B
B ( B)
B 0 E
= 0 E ( E) +
+ j B.
t
t
0
(8.50)
Next, making use of Eqs. (8.45) and (8.46), we get
B ( B)
[0 E B] = 0 E ( E) +
t
0
1
0 ( E) E ( B) B + E + j B.
0
(8.51)
Now, from vector field theory,

(E2 /2) E ( E) + (E )E,
251
(8.52)
with a similar equation for B. Hence, Eq. (8.51) takes the form
h
i
[0 E B] = 0 (E2 /2) ( E) E (E )E

t
i
1 h
(B2 /2) ( B) B (B )B
+
0
+ E + j B.
(8.53)
Finally, when written in terms of components, the above equation becomes
h
[0 E B]i =
0 E2 ij /2 0 Ei Ej + B2 ij /2 0 Bi Bj /0
t
xj
+ [ E + j B]i ,
(8.54)
since [( E) E]i = (Ej /xj ) Ei , and [(E )E]i = Ej (Ei /xj ). Here, ij is a
Kronecker delta symbol (i.e., ij = 1 if i = j, and ij = 0 otherwise). Comparing
the above equation with Eq. (8.44), we conclude that the momentum flux density
tensor of electromagnetic fields takes the form
Gij = 0 (E2 ij /2 Ei Ej ) + (B2 ij /2 Bi Bj )/0 .
(8.55)
The momentum conservation equation (8.44) is sometimes written

[ E + j B]i =
Tij
[0 E B]i ,
xj
t
(8.56)
where
Tij = Gij = 0 (Ei Ej E2 ij /2) + (Bi Bj B2 ij /2)/0
is called the Maxwell stress tensor.
252
(8.57)
9 ELECTROMAGNETIC RADIATION
Electromagnetic radiation
9.1 Introduction
In this section, we shall use Maxwells equations to investigate electromagnetic
waves.
9.2 The Hertzian dipole

Consider two small spherical conductors connected by a wire. Suppose that electric charge flows periodically back and forth between the spheres. Let q(t) be the
instantaneous charge on one of the conductors. The system has zero net charge,
so the charge on the other conductor is q(t). Let
q(t) = q0 sin ( t).
(9.1)
We expect the oscillating current flowing in the wire connecting the two spheres
to generate electromagnetic radiation (see Sect. 4.11). Let us consider the simple
case in which the length of the wire is small compared to the wave-length of
the emitted radiation. If this is the case, then the current I flowing between the
conductors has the same phase along the whole length of the wire. It follows that
I(t) =
dq
= I0 cos( t),
dt
(9.2)
where I0 = q0 . This type of antenna is called a Hertzian dipole, after the

German physicist Heinrich Hertz.
The magnetic vector potential generated by a current distribution j(r) is given
by the well-known formula (see Sect. 4.12)
Z
[j]
0
A(r, t) =
d3 r 0 ,
(9.3)
0
4 |r r |
where
[f] = f(r 0 , t |r r 0 |/c).
253
(9.4)
Suppose that the wire is aligned along the z-axis, and extends from z = l/2 to
z = l/2. For a wire of negligible thickness, we can replace j(r 0 , t |r r 0 |/c) d3 r 0
by I(r 0 , t |r r 0 |/c) dz 0 ^z. Thus, A(r, t) = Az (r, t) ^z, and
Z
0 l/2 I(z 0 , t |r z 0 ^z|/c) 0
dz .
(9.5)
Az (r, t) =
4 l/2
|r z 0 ^z|
In the region r l,
and
|r z 0 ^z| ' r,
(9.6)
t |r z 0 ^z|/c ' t r/c.
(9.7)
The maximum error in the latter approximation is t l/c. This error (which is
a time) must be much less than a period of oscillation of the emitted radiation,
otherwise the phase of the radiation will be wrong. So
2
l

,
(9.8)
c
which implies that l , where = 2 c/ is the wave-length of the emitted radiation. However, we have already assumed that the length of the wire
l is much less than the wave-length of the radiation, so the above inequality is
automatically satisfied. Thus, in the far field region, r , we can write
Z
0 l/2 I(z 0 , t r/c) 0
Az (r, t) '
dz .
(9.9)
4 l/2
r
This integral is easy to perform, since the current is uniform along the length of
the wire. So,
0 l I(t r/c)
.
(9.10)
Az (r, t) '
4
r
The scalar potential is most conveniently evaluated using the Lorentz gauge
condition (see Sect. 4.12)
.
(9.11)
A = 0 0
t
Now,
!
!
Az
0 l I(t r/c)
z
1
A=
'
2
+O 2
(9.12)
z
4
t
r c
r
254
to leading order in r1 . Thus,

(r, t) '
l z I(t r/c)
.
4 0 c r
r
(9.13)
Given the vector and scalar potentials, Eqs. (9.10) and (9.13), respectively, we
can evaluate the associated electric and magnetic fields using (see Sect. 4.12)
A
,
t
B = A.
E =
(9.14)
(9.15)
Note that we are only interested in radiation fields, which fall off like r 1 with
increasing distance from the source. It is easily demonstrated that
E'
l I0
sin[(t r/c)] ^
sin
,
4 0 c2
r
(9.16)
and
sin[(t r/c)]
l I0
^
sin
.
(9.17)
4 0 c3
r
Here, (r, , ) are standard spherical polar coordinates aligned along the z-axis.
The above expressions for the far field (i.e., r ) electromagnetic fields generated by a localized oscillating current are also easily derived from Eqs. (4.183)
and (4.184). Note that the fields are symmetric in the azimuthal angle . There
is no radiation along the axis of the oscillating dipole (i.e., = 0), and the maximum emission is in the plane perpendicular to this axis (i.e., = /2).
B'
The average power crossing a spherical surface S (whose radius is much greater
than ) is
I
Prad =
hui dS,
(9.18)
where the average is over a single period of oscillation of the wave, and the
Poynting flux is given by (see Sect. 8.2)
2 l2 I02
sin2
EB
2
=
sin [(t r/c)] 2 ^r.
u=
0
162 0 c3
r
255
(9.19)
It follows that
2 l2 I02 sin2
^r.
hui =
(9.20)
322 0 c3 r2
Note that the energy flux is radially outwards from the source. The total power
flux across S is given by
Z
Z
sin2 2
2 l2 I02 2
d
r sin d.
(9.21)
Prad =
322 0 c3 0
r2
0
Thus,
2 l2 I02
.
(9.22)
Prad =
12 0 c3
The total flux is independent of the radius of S, as is to be expected if energy is
conserved.
Recall that for a resistor of resistance R the average ohmic heating power is
1
Pheat = hI2 Ri = I02 R,
2
(9.23)
assuming that I = I0 cos( t). It is convenient to define the radiation resistance

of a Hertzian dipole antenna:
Prad
(9.24)
Rrad = 2 ,
I0 /2
so that
!
2
l 2
Rrad =
,
(9.25)
3 0 c
where = 2 c/ is the wave-length of the radiation. In fact,
Rrad
l
= 789
!2
ohms.
(9.26)
In the theory of electrical circuits, an antenna is conventionally represented as

a resistor whose resistance is equal to the characteristic radiation resistance of
the antenna plus its real resistance. The power loss I02 Rrad /2 associated with
the radiation resistance is due to the emission of electromagnetic radiation. The
power loss I02 R/2 associated with the real resistance is due to ohmic heating of
the antenna.
256
Note that the formula (9.26) is only valid for l . This suggests that Rrad
R for most Hertzian dipole antennas: i.e., the radiated power is swamped by
the ohmic losses. Thus, antennas whose lengths are much less than that of the
emitted radiation tend to be extremely inefficient. In fact, it is necessary to have
l in order to obtain an efficient antenna. The simplest practical antenna is the
half-wave antenna, for which l = /2. This can be analyzed as a series of Hertzian
dipole antennas stacked on top of one another, each slightly out of phase with its
neighbours. The characteristic radiation resistance of a half-wave antenna is
Rrad =
2.44
= 73 ohms.
4 0 c
(9.27)
Antennas can also be used to receive electromagnetic radiation. The incoming

wave induces a voltage in the antenna, which can be detected in an electrical circuit connected to the antenna. In fact, this process is equivalent to the emission
of electromagnetic waves by the antenna viewed in reverse. It is easily demonstrated that antennas most readily detect electromagnetic radiation incident from
those directions in which they preferentially emit radiation. Thus, a Hertzian
dipole antenna is unable to detect radiation incident along its axis, and most efficiently detects radiation incident in the plane perpendicular to this axis. In the
theory of electrical circuits, a receiving antenna is represented as a voltage source
in series with a resistor. The voltage source, V0 cos( t), represents the voltage
induced in the antenna by the incoming wave. The resistor, Rrad , represents the
power re-radiated by the antenna (here, the real resistance of the antenna is
neglected). Let us represent the detector circuit as a single load resistor Rload ,
connected in series with the antenna. The question is: how can we choose R load
so that the maximum power is extracted from the wave and transmitted to the
load resistor? According to Ohms law:
V0 cos( t) = I0 cos( t) (Rrad + Rload ),
(9.28)
where I = I0 cos( t) is the current induced in the circuit.

The power input to the circuit is
V02
.
Pin = hVIi =
2 (Rrad + Rload )
257
(9.29)
The power transferred to the load is

Pload
Rload V02
.
= hI Rload i =
2 (Rrad + Rload )2
2
(9.30)
The power re-radiated by the antenna is

Prad = hI2 Rrad i =
Rrad V02
.
2 (Rrad + Rload )2
(9.31)
Note that Pin = Pload + Prad . The maximum power transfer to the load occurs
when
V02 Rrad Rload

Pload
=
= 0.
(9.32)
Rload
2 (Rrad + Rload )3
Thus, the maximum transfer rate corresponds to
Rload = Rrad .
(9.33)
In other words, the resistance of the load circuit must match the radiation resistance of the antenna. For this optimum case,
Pload = Prad =
V02
Pin
.
=
8 Rrad
2
(9.34)
So, in the optimum case half of the power absorbed by the antenna is immediately
re-radiated. Clearly, an antenna which is receiving electromagnetic radiation is
also emitting it. This is how the BBC catch people who do not pay their television
license fee in England. They have vans which can detect the radiation emitted by
a TV aerial whilst it is in use (they can even tell which channel you are watching!).
For a Hertzian dipole antenna interacting with an incoming wave whose electric field has an amplitude E0 , we expect
V0 = E0 l.
(9.35)
Here, we have used the fact that the wave-length of the radiation is much longer
than the length of the antenna. We have also assumed that the antenna is properly aligned (i.e., the radiation is incident perpendicular to the axis of the antenna). The Poynting flux of the incoming wave is [see Eq. (8.35)]
0 c E02
huin i =
,
2
258
(9.36)
9.3 Electric dipole radiation
whereas the power transferred to a properly matched detector circuit is

Pload
E02 l2
.
=
8 Rrad
(9.37)
Consider an idealized antenna in which all incoming radiation incident on some

area Aeff is absorbed, and then magically transferred to the detector circuit, with
no re-radiation. Suppose that the power absorbed from the idealized antenna
matches that absorbed from the real antenna. This implies that
Pload = huin iAeff .
(9.38)
The quantity Aeff is called the effective area of the antenna: it is the area of the
idealized antenna which absorbs as much net power from the incoming wave as
the actual antenna. Thus,
Pload =
0 c E02
E02 l2
=
Aeff ,
8 Rrad
2
(9.39)
giving
l2
3 2
Aeff =
=
.
(9.40)
4 0 c Rrad
8
It is clear that the effective area of a Hertzian dipole antenna is of order the
wave-length squared of the incoming radiation.
For a properly aligned half-wave antenna,
Aeff = 0.13 2 .
(9.41)
Thus, the antenna, which is essentially one-dimensional with length /2, acts as
if it is two-dimensional, with width 0.26 , as far as its absorption of incoming
electromagnetic radiation is concerned.
9.3 Electric dipole radiation

In the previous section, we examined the radiation emitted by a short electric
dipole of oscillating dipole moment
p(t) = p0 sin( t) ^z,
259
(9.42)
9.4 Thompson scattering
where p0 = q0 l = I0 l/. We found that, in the far field, the mean electromagnetic energy flux takes the form [see Eq. (9.20)]
4 p02 sin2
^r,
hui =
322 0 c3 r2
(9.43)
assuming that the dipole is centered on the origin of our spherical polar coordinate system. The mean power radiated into the element of solid angle d =
sin d d, centered on the angular coordinates (, ), is
dP = hu(r, , )i^r r2 d.
(9.44)
Hence, the differential power radiated into this element of solid angle is simply
4 p02
dP
=
sin2 .
2
3
d 32 0 c
(9.45)
This formula completely specifies the radiation pattern of an oscillating electric

dipole (provided that the dipole is much shorter in length than the wave-length
of the emitted radiation). Of course, the power radiated into a given element of
solid angle is independent of r, otherwise energy would not be conserved. Finally,
the total radiated power is the integral of dP/d over all solid angles.

Consider a plane electromagnetic wave of angular frequency interacting with
a free electron of mass me and charge e. Suppose that the wave is polarized
such that its associated electric field is parallel to the z-axis: i.e.,
E = E0 sin( t) ^z.
(9.46)
Recall, from Sect. 4.7, that as long as the electron remains non-relativistic, the
force exerted on it by the electromagnetic wave comes predominantly from the
associated electric field. Hence, the electrons equation of motion can be written
d2 z
me 2 = e E0 sin( t),
dt
260
(9.47)
which can be solved to give

e E0
sin( t).
(9.48)
m e 2
So, in response to the wave, the electron oscillates backward and forward in the
direction of the wave electric field. It follows that the electron can be thought of
as a sort of oscillating electric dipole, with dipole moment
z=
p = e z ^z = p0 sin( t) ^z,
(9.49)
where p0 = e2 E0 /(me 2 ). (For the moment, let us not worry about the positively
charged component of the dipole.) Now, we know that an oscillating electric
dipole emits electromagnetic radiation. Hence, it follows that a free electron
placed in the path of a plane electromagnetic wave will radiate. To be more
exact, the electron scatters electromagnetic radiation from the wave, since the
radiation emitted by the electron is not necessarily in the same direction as the
wave, and any energy radiated by the electron is ultimately extracted from the
wave. This type of scattering is called Thompson scattering.
It follows from Eq. (9.45) that the differential power scattered from a plane
electromagnetic wave by a free electron into solid angle d takes the form
dP
e4 E02
sin2 .
=
2
3
2
d 32 0 c me
(9.50)
Now, the mean energy flux of the incident electromagnetic wave is written
c 0 E02
|hui| =
.
(9.51)
2
It is helpful to introduce a quantity called the differential scattering cross-section.
This is defined
d
dP/d
=
,
(9.52)
d
|hui|
and has units of area over solid angle. Somewhat figuratively, we can think of
the electron as offering a target of area d/d to the incident wave. Any wave
energy which falls on this target is scattered into the solid angle d. Likewise,
we can also define the total scattering cross-section,
I
d
d,
(9.53)
=
d
261
which has units of area. Again, the electron effectively offers a target of area to
the incident wave. Any wave energy which falls on this target is scattered in some
direction or other. It follows from Eqs. (9.50) and (9.51) that the differential
scattering cross-section for Thompson scattering is
d
= re2 sin2 ,
d
(9.54)
e2
= 2.82 1015 m
re =
2
4 0 me c
(9.55)
where the characteristic length
is called the classical electron radius. An electron effectively acts like it has a
spatial extent re as far as its iteration with electromagnetic radiation is concerned.
As is easily demonstrated, the total Thompson scattering cross-section is
T =
8 2
re = 6.65 1029 m2 .
3
(9.56)
Note that both the differential and the total Thompson scattering cross-sections
are completely independent of the frequency (or wave-length) of the incident radiation.
A scattering cross-section of 1029 m2 does not sound like much. Nevertheless,
Thompson scattering is one of the most important types of scattering in the Universe. Consider the Sun. It turns out that the mean mass density of the Sun is
similar to that of water: i.e., about 103 kg m3 . Hence, assuming that the Sun is
made up predomintely of ionized hydrogen, the mean number density of electrons in the Sun (which, of course, is the same as the number density of protons)
is approximately ne 103 /mp 1030 m3 , where mp 1027 kg is the mass of a
proton. Let us consider how far, on average, a photon in the Sun travels before
being scattered by free electrons. If we think of an individual photon as sweeping
out a cylinder of cross-sectional area T , then the photon will travel an average
length l, such that a cylinder of area T and length l contains about one free
electron, before being scattered. Hence, T l ne 1, or
l
1
1 cm.
ne T
262
(9.57)
9.5 Rayleigh scattering
Given that the radius of the Sun is approximately 109 m, it is clear that solar
photons are very strongly scattered by free electrons. In fact, it can easily be
demonstrated that it takes a photon emitted in the solar core many thousands of
years to fight its way to the surface because of Thompson scattering.
After the Big Bang, when the Universe was very hot, it consisted predominately of ionized hydrogen (and dark matter), and was consequently opaque to
electromagnetic radiation, due to Thompson scattering. However, as the Universe expanded, it also cooled, and eventually became sufficiently cold (when
the mean temperature was about 1000 C) for any free protons and electrons to
combine to form molecular hydrogen. It turns out that molecular hydrogen does
not scatter radiation anything like as effectively as free electrons (see the next
section). Hence, as soon as the Universe became filled with molecular hydrogen,
it effectively became transparent to radiation. Indeed, the so-called cosmic microwave background is the remnant of radiation which was last scattered when
the Universe was filled with ionized hydrogen (i.e., when it was about 1000 C).
Astronomers can gain a great deal of information about the conditions in the
early Universe by studying this radiation.
Incidentally, it is clear from Eqs. (9.55) and (9.56) that the scattering crosssection of a free particle of charge q and mass m is proportional to q4 /m2 . It
follows that the scattering of electromagnetic radiation by free electrons is generally very much stronger than the scattering by free protons (assuming that the
number densities of both species are similar).

Let us now consider the scattering of electromagnetic radiation by neutral atoms.
For instance, consider a hydrogen atom. The atom consists of a light electron and
a massive proton. As we have seen, the electron scatters radiation much more
strongly than the proton, so let us concentrate on the response of the electron to
an incident electromagnetic wave. Suppose that the wave electric field is again
polarized in the z-direction, and is given by Eq (9.46). We can approximate the
263
electrons equation of motion as

d2 z
= me 02 z e E0 sin( t).
(9.58)
2
dt
Here, the second term on the right-hand side represents the perturbing force due
to the electromagnetic wave, whereas the first term represents the (linearized)
force of electrostatic attraction between the electron and the proton. Indeed, we
are very crudely modeling our hydrogen atom as a simple harmonic oscillator of
natural frequency 0 . We can think of 0 as the typical frequency of electromagnetic radiation emitted by the atom after it is transiently disturbed. In other
words, in our model, 0 should match the frequency of one of the spectral lines
of hydrogen. More generally, we can extend the above model to deal with just
about any type of atom, provided that we set 0 to the frequency of a spectral
line.
me
We can easily solve Eq. (9.58) to give

z=
e E0
sin( t).
me (2 02 )
(9.59)
Hence, the dipole moment of the electron takes the form p = p0 sin( t) ^z,
where
e2 E 0
.
(9.60)
p0 =
me (2 02 )
It follows, by analogy with the analysis in the previous section, that the differential and total scattering cross-sections of our model atom take the form
and
d
4
=
re2 sin2 ,
2
2
2
d ( 0 )
(9.61)
4
T ,
=
(2 02 )2
(9.62)
respectively.
In the limit in which the frequency of the incident radiation is much greater
than the natural frequency of the atom, Eqs. (9.61) and (9.62) reduce to the previously obtained expressions for scattering by a free electron. In other words, an
264
electron in an atom acts very much like a free electron as far as high frequency
radiation is concerned. In the opposite limit, in which the frequency of the incident radiation is much less than the natural frequency of the atom, Eqs. (9.61)
and (9.62) yield
!
4 2
d
re sin2 ,
(9.63)
=
d
0
and
!
4
=
T ,
(9.64)
0
respectively. This type of scattering is called Rayleigh scattering. There are two
features of Rayleigh scattering which are worth noting. First of all, it is much
weaker than Thompson scattering (since 0 ). Secondly, unlike Thompson
scattering, it is highly frequency dependent. Indeed, it is clear, from the above
formulae, that high frequency (short wave-length) radiation is scattered far more
effectively than low frequency (long wave-length) radiation.
The most common example of Rayleigh scattering is the scattering of visible
radiation from the Sun by neutral atoms (mostly Nitrogen and Oxygen) in the upper atmosphere. The frequency of visible radiation is much less than the typical
emission frequencies of a Nitrogen or Oxygen atom (which lie in the ultra-violet
band), so it is certainly the case that 0 . When the Sun is low in the sky, radiation from it has to traverse a comparatively long path through the atmosphere
before reaching us. Under these circumstances, the scattering of direct solar light
by neutral atoms in the atmosphere becomes noticeable (it is not noticeable when
the Sun is high is the sky, and radiation from it consequently only has to traverse
a relatively short path through the atmosphere before reaching us). According
to Eq. (9.64), blue light is scattered slightly more strongly than red light (since
blue light has a slightly higher frequency than red light). Hence, when the Sun is
low in the sky, it appears less bright, due to atmospheric scattering. However, it
also appears redder than normal, because more blue light than red light is scattered out of the solar light-rays, leaving an excess of red light. Likewise, when we
look up at the sky, it does not appear black (like the sky on the Moon) because
of light from solar radiation which grazes the atmosphere being scattered downward towards the surface of the Earth. Again, since blue light is scattered more
265
9.6 Propagation in a dielectric medium
effectively than red light, there is an excess of blue light scattered downward,
and so the sky appears blue.
Light from the Sun is unpolarized. However, when it is scattered it becomes
polarized, because light is scattered preferentially in some directions rather than
others. Consider a light-ray from the Sun which grazes the Earths atmosphere.
The light-ray contains light which is polarized such that the electric field is vertical
to the ground, and light which is polarized such that the electric field is horizontal
to the ground (and perpendicular to the path of the light-ray), in equal amounts.
However, due to the sin2 factor in the dipole emission formula (9.45) (where,
in this case, is the angle between the direction of the wave electric field and the
direction of scattering), very little light is scattered downward from the vertically
polarized light compared to the horizontally polarized light. Moreover, the light
scattered from the horizontally polarization is such that its electric field is preferentially perpendicular, rather than parallel, to the direction of propagation of the
solar light-ray (i.e., the direction to the Sun). Consequently, the blue light from
the sky is preferentially polarized in a direction perpendicular to the direction to
the Sun.
9.6 Propagation in a dielectric medium

Consider the propagation of an electromagnetic wave through a uniform dielectric medium of dielectric constant . According to Eqs. (6.8) and (6.10), the
dipole moment per unit volume induced in the medium by the wave electric field
E is
P = 0 ( 1) E.
(9.65)
There are no free charges or free currents in the medium. There is also no bound
charge density (since the medium is uniform), and no magnetization current
density (since the medium is non-magnetic). However, there is a polarization
current due to the time-variation of the induced dipole moment per unit volume.
According to Eq. (6.57), this current is given by
j=
P
.
t
266
(9.66)
9.7 Dielectric constant of a gaseous medium
Hence, Maxwells equations take the form

E = 0,
B = 0,
E =
(9.67)
(9.68)
B
,
t
B = 0 j + 0 0
(9.69)
E
.
t
(9.70)
According to Eqs. (9.65) and (9.66), the last of the above equations can be rewritten
E
E
E
+ 0 0
= 2
,
(9.71)
B = 0 0 ( 1)
t
t
c t
since c = (0 0 )1/2 . Thus, Maxwells equations for the propagation of electromagnetic waves through a dielectric medium are the same as Maxwells equations for the propagation of waves through a vacuum (see Sect. 4.7), except that
c c/n, where
(9.72)
n=
is called the refractive index of the medium in question. Hence, we conclude
that electromagnetic waves propagate through a dielectric medium slower than
through a vacuum by a factor n (assuming, of course, that n > 1). This conclusion (which was reached long before Maxwells equations were invented) is the
basis of all geometric optics involving refraction.
9.7 Dielectric constant of a gaseous medium

In Sect. 9.5, we discussed a rather crude model of an atom interacting with an
electromagnetic wave. According to this model, the dipole moment p of the atom
induced by the wave electric field E is given by
e2
p=
E,
me (02 2 )
(9.73)
where 0 is the natural frequency of the atom (i.e., the frequency of one of the
atoms spectral lines), and the frequency of the incident radiation. Suppose
267
9.8 Dielectric constant of a plasma
that there are n atoms per unit volume. It follows that the induced dipole moment per unit volume of the assemblage of atoms takes the form
n e2
E.
P=
me (02 2 )
(9.74)
Finally, a comparison with Eq. (9.65) yields the following expression for the dielectric constant of the collection of atoms,
=1+
n e2
.
0 me (02 2 )
(9.75)
The above formula works fairly well for dilute gases, although it is, of course,
necessary to sum over all species and all important spectral lines.
Note that, in general, the dielectric constant of a gaseous medium (as far
as electromagnetic radiation is concerned) is a function of the wave frequency,
. Since the effective wave speed through the medium is c/ , it follows that
waves of different frequencies traveling through a gaesous medium do so at different speeds. This phenomenon is called dispersion, since it can be shown to cause
short wave-pulses to spread out as they propagate through the medium. At low
frequencies ( 0 ), however, our expression for becomes frequency independent, so there is no dispersion of low frequency waves by a gaseous medium.

A plasma is very similar to a gaseous medium, expect that the electrons are free:
i.e., there is no restoring force due to nearby atomic nuclii. Hence, we can obtain
an expression for the dielectric constant of a plasma from Eq. (9.75) by setting
0 to zero, and n to the number density of electrons, ne . We obtain
p2
= 1 2,
where the characteristic frequency

p =
v
u
u
t
ne e2
0 m e
268
(9.76)
(9.77)
is called the plasma frequency. We can immediately see that formula (9.76) is
problematic. For frequencies above the plasma frequency, the dielectric constant
of a plasma is less than unity. Hence, the refractive index n = is also less than
unity. This would seem to imply that high frequency electromagnetic waves can
propagate through a plasma with a velocity c/n which is greater than the velocity
of light in a vacuum. Does this violate the principles of relativity? On the other
hand, for frequencies below the plasma frequency, the dielectric constant is neg
ative, which would seem to imply that the refractive index n = is imaginary.
How should we interpret this?
Consider an infinite plane-wave, of frequency, , greater than the plasma frequency, propagating through a plasma. Suppose that the wave electric field takes
the form
(9.78)
E = E0 e i (k x t) ^z,
where it is understood that the physical electric field is the real part of the above
expression. A peak or trough of the above wave travels at the so-called phase
velocity, which is given by
(9.79)
vp = .
k
Now, we have also seen that the phase velocity of electromagnetic waves in a
dielectric medium is vp = c/n = c/ , so

k2 c2
=
.

(9.80)
2 = k2 c2 + p2
(9.81)
in a plasma. The above type of expression, which effectively determines the wave
frequency, , as a function of the wave-number, k, for the medium in question,
is called a dispersion relation (since, amongst other things, it determines how fast
wave-pulses disperse in the medium). According to the above dispersion relation,
the phase velocity of high frequency waves propagating through a plasma is given
by
c
vp = q
,
(9.82)
1 p2 /2
269
which is indeed greater than c. However, the theory of relativity does not forbid
this. What the theory of relativity says is that information cannot travel at a
velocity greater than c. And the peaks and troughs of an infinite plane-wave,
such as (9.78), do not carry any information.
We now need to consider how we could transmit information through a plasma
(or any other dielectric medium) by means of electromagnetic waves. The easiest
way would be to send a series of short discrete wave-pulses through the plasma,
so that we could encode information in a sort of Morse code. We can build up
a wave-pulse from a suitable superposition of infinite plane-waves of different
frequencies and wave-lengths: e.g.,
Z
Ez (x, t) = F(k) e i (k) dk,
(9.83)
where (k) = k x (k) t, and (k) is determined from the dispersion relation
(9.81). Now, it turns out that a relatively short wave-pulse can only be built up
from a superposition of plane-waves with a relatively wide range of different k
values. Hence, for a short wave-pulse, the integrand in the above formula consists
of the product of a fairly slowly varying function, F(k), and a rapidly oscillating
function, exp[i (k)]. The latter function is rapidly oscillating because the phase
(k) varies very rapidly with k, relative to F(k). We expect the net result of integrating the product of a slowly varying function and rapidly oscillating function
to be small, since the oscillations will generally average to zero. It follows that
the integral (9.83) is dominated by those regions of k-space for which (k) varies
least rapidly with k. Hence, the peak of the wave-pulse most likely corresponds
to a maximum or minimum of (k): i.e.,
d
d
=x
t = 0.
(9.84)
dk
dk
Thus, we infer that the velocity of the wave-pulse (which corresponds to the
velocity of the peak) is given by
d
.
(9.85)
dk
This velocity is called the group velocity, and is different to the phase velocity
in dispersive media, for which is not directly proportional to k. (Of course,
vg =
270
in a vacuum, = k c, and both the phase and group velocities are equal to c.)
The upshot of the above discussion is that information (i.e., an individual wavepulse) travels through a dispersive media at the group velocity, rather than the
phase velocity. Hence, relativity demands that the group velocity, rather than the
phase velocity, must always be less than c.
What is the group velocity for high frequency waves propagating through a
plasma? Well, differentiation of the dispersion relation (9.81) yields
d
= v p vg = c 2 .
k dk
(9.86)
Hence, it follows from Eq. (9.82) that

v
u
u
t
p2
vg = c 1 2 ,
(9.87)
which is less than c. We thus conclude that the dispersion relation (9.81) is
indeed consistent with relativity.
Let us now consider the propagation of low frequency electromagnetic waves
through a plasma. We can see, from Eqs. (9.82) and (9.87), that when the wave
frequency, , falls below the plasma frequency, p , both the phase and group
velocities become imaginary. This indicates that the wave attenuates as it propagates. Consider, for instance, a plane-wave of frequency < p . According to
the dispersion relation (9.81), the associated wave-number is given by
k= i
p2
c = i |k|.
(9.88)
Hence, the wave electric field takes the form

Ez = E0 e i (i |k| x t) = E0 e|k| x ei t .
(9.89)
Indeed, it can be seen that for < p electromagnetic waves in a plasma take
the form of decaying standing waves, rather than traveling waves. We conclude
that an electromagnetic wave, of frequency less than the plasma frequency, which
is incident on a plasma will not propagate through the plasma. Instead, it will be
totally reflected.
271
We can be sure that the incident wave is reflected by the plasma, rather than
absorbed, by considering the energy flux of the wave in the plasma. It is easily
demonstrated that the energy flux of an electromagnetic wave can be written
E2
EB
=
k.
u=
0
0
(9.90)
For a wave with a real frequency and a complex k-vector, the above formula
generalizes to
|E|2
Re(k).
(9.91)
u=
0
However, according to Eq. (9.88), the k-vector for a low frequency electromagnetic wave in a plasma is purely imaginary. It follows that the associated energy
flux is zero. Hence, any low frequency wave which is incident on the plasma must
be totally reflected, since if there were any absorption of the wave energy then
there would be a net energy flux into the plasma.
The outermost layer of the Earths atmosphere consists of a partially ionized
zone known as the ionosphere. The plasma frequency in the ionosphere is about
1 MHz, which lies at the upper end of the medium-wave band of radio frequencies. It follows that low frequency radio signals (i.e., all signals in the long-wave
band, and most in the medium-wave band) are reflected off the ionosphere. For
this reason, such signals can be detected over the horizon. Indeed, long-wave
radio signals reflect multiple times off the ionosphere, with very little loss (they
also reflect multiple times off the Earth, which is enough of a conductor to act as
a mirror for radio waves), and can consequently be detected all over the world.
On the other hand, high frequency radio signals (i.e., all signals in the FM band)
pass straight through the ionosphere. For this reason, such signals cannot be detected over the horizon, which accounts for the relatively local coverage of FM
radio stations. Note, from Eq. (9.77), that the plasma frequency is proportional
to the square root of the number density of free electrons. Now, the level of ionization in the ionosphere is maintained by ultra-violet light from the Sun (which
effectively knocks electrons out of neutral atoms). Of course, there is no such
light at night, and the number density of free electrons in the ionosphere consequently drops as electrons and ions gradually recombine. It follows that the
272
9.9 Faraday rotation
plasma frequency in the ionosphere also drops at night, giving rise to a marked
deterioration in the reception of distant medium-wave radio stations.

Consider a high frequency electromagnetic wave propagating, along the z-axis,
through a plasma with a longitudinal equilibrium magnetic field, B = B 0 ^z. The
equation of motion of an individual electron making up the plasma takes the form
me
dv
= e (E + B0 v ^z),
dt
(9.92)
where the first term on the right-hand side is due to the wave electric field, and
the second to the equilibrium magnetic field. (As usual, we can neglect the wave
magnetic field, provided that the electron motion remains non-relativistic.) Of
course, v = dr/dt, where r is the electron displacement from its equilibrium
position. Suppose that all perturbed quantities vary with time like exp(i t),
where is the wave frequency. It follows that
me 2 x = e (Ex i B0 y),
(9.93)
me 2 y = e (Ey + i B0 x).
(9.94)
s = x i y,
(9.95)
It is helpful to define
E = E x i E y .
(9.96)
Using these new variables, Eqs. (9.93) and (9.94) can be rewritten
me 2 s = e (E B0 s ),
(9.97)
which can be solved to give

s =
e E
,
me ( )
273
(9.98)
where = e B0 /me is the so-called cyclotron frequency (i.e., the characteristic gyration frequency of free electrons in the equilibrium magnetic fieldsee
Sect. 3.7).
In terms of s , the electron displacement can be written
r = s+ e i (k+ z t) e+ + s e i (k z t) e ,
(9.99)
where
1
^) .
(^
x iy
2
Likewise, in terms of E , the wave electric field takes the form
e =
E = E+ e i (k+ z t) e+ + E e i (k z t) e .
(9.100)
(9.101)
Obviously, the actual displacement and electric field are the real parts of the above
expressions. It follows from Eq. (9.101) that E+ corresponds to a constant amplitude electric field which rotates anti-clockwise in the x-y plane (looking down
the z-axis) as the wave propagates in the +z-direction, whereas E corresponds
to a constant amplitude electric field which rotates clockwise. The former type of
wave is termed right-hand circularly polarized, whereas the latter is termed lefthand circularly polarized. Note also that s+ and s correspond to circular electron
motion in opposite senses. With these insights, we conclude that Eq. (9.98) indicates that individual electrons in the plasma have a slightly different response to
right- and left-hand circularly polarized waves in the presence of a longitudinal
magnetic field.
Following the analysis of Sect. 9.7, we can deduce from Eq. (9.98) that the dielectric constant of the plasma for right- and left-hand circularly polarized waves
is
p2
,
(9.102)
= 1
( )
respectively. Hence, according to Eq. (9.80), the dispersion relation for right- and
left-hand circularly polarized waves becomes
p2
2
2 2
,
k c = 1
( )
274
(9.103)
respectively. In the limit p , , we obtain

k ' k k,
(9.104)
where k = [1 (1/2) p2 /2 ]/c and k = (1/2) (p2 /2 ) /c. In other words,

in a magnetized plasma, right- and left-hand circularly polarized waves of the
same frequency have slightly different wave-numbers
Let us now consider the propagation of a linearly polarized electromagnetic
wave through the plasma. Such a wave can be constructed via a superposition of
right- and left-hand circularly polarized waves of equal amplitudes. So, the wave
electric field can be written
h
E = E0 e
i (k+ z t)
e+ + e
i (k z t)
e .
(9.105)
It can easily be seen that at z = 0 the wave electric field is aligned along the xaxis. If right- and left-hand circularly polarized waves of the same frequency have
the same wave-number (i.e., if k+ = k ) then the wave electric field will continue
to be aligned along the x-axis as the wave propagates in the +z-direction: i.e.,
we will have a standard linearly polarized wave. However, we have just demonstrated that, in the presence of a longitudinal magnetic field, the wave-numbers
k+ and k are slightly different. What effect does this have on the polarization of
the wave?
Taking the real part of Eq. (9.105), and making use of Eq. (9.104), and some
standard trigonometrical identities, we obtain
E = E0 [cos(k z t) cos(k z), cos(k z t) sin(k z), 0] .
(9.106)
The polarization angle of the wave (which is a convenient measure of its plane
of polarization) is given by
= tan1 (Ey /Ex ) = k z.
(9.107)
Thus, we conclude that in the presence of a longitudinal magnetic field the polarization angle rotates as as the wave propagates through the plasma. This effect is
known as Faraday rotation. It is clear, from the above expression, that the rate of
advance of the polarization angle with distance travelled by the wave is given by
p2
d
ne B 0
e3
= k =
=
.
dz
2 2 c 2 0 me2 c 2
275
(9.108)
9.10 Propagation in a conductor
Hence, a linearly polarized electromagnetic wave which propagates through a

plasma with a (slowly) varying electron number density, ne (z), and longitudinal
magnetic field, B0 (z), has its plane of polarization rotated through a total angle
Z
e3
1
= 0 =
ne (z) B0 (z) dz.
(9.109)
2 0 me2 c 2
Note the very strong inverse variation of with .
Pulsars are rapidly rotating neutron stars which emit regular blips of highly
polarized radio waves. Hundreds of such objects have been found in our galaxy
since the first was discovered in 1967. By measuring the variation of the angle of
polarization, , of radio emission from a pulsar with frequency, , astronomers
can effectively determine the line integral of ne B0 along the straight-line joining
the pulsar to the Earth using formula (9.109). Here, ne is the number density of
free electrons in the interstellar medium, whereas B0 is the parallel component of
the galactic magnetic field. Obviously, in order to achieve this, astronomers must
make the reasonable assumption that the radiation was emitted by the pulsar
with a common angle of polarization, 0 , over a wide range of different frequencies. By fitting Eq. (9.109) to the data, and then extrapolating to large , it is
then possible to determine 0 , and, hence, the amount, (), through which
the polarization angle of the radiation has rotated, at a given frequency, during
its passage to Earth.

Consider the propagation of an electromagnetic wave through a conducting medium
which obeys Ohms law:
j = E.
(9.110)
Here, is the conductivity of the medium in question. Maxwells equations for
the wave take the form:
E = 0,
B = 0,
276
(9.111)
(9.112)
E =
B
,
t
(9.113)
B = 0 j + 0 0
E
,
t
(9.114)
where is the dielectric constant of the medium. It follows, from the above
equations, that
E
B
=
0 E + 0 0
.
E = E =
t
t
t
"
(9.115)
Looking for a wave-like solution of the form

E = E0 e i (k z t) ,
(9.116)
we obtain the dispersion relation

k2 = 0 ( 0 + i ).
(9.117)
Consider a poor conductor for which 0 . In this limit, the dispersion

relation (9.117) yields
v
u
u 0
,
(9.118)
k'n +i t
c
2 0
where n = is the refractive index. Substitution into Eq. (9.116) gives

E = E0 ez/d e i (kr z t) ,
where
2u
u 0
d= t
,
0
(9.119)
(9.120)
and kr = n /c. Thus, we conclude that the amplitude of an electromagnetic

wave propagating through a conductor decays exponentially on some length-scale,
d, which is termed the skin-depth. Note, from Eq. (9.120), that the skin-depth
for a poor conductor is independent of the frequency of the wave. Note, also,
that kr d 1 for a poor conductor, indicating that the wave penetrates many
wave-lengths into the conductor before decaying away.
277
Consider a good conductor for which 0 . In this limit, the dispersion

relation (9.117) yields
q
k ' i 0 .
(9.121)
Substitution into Eq. (9.116) again gives Eq. (9.119), with
v
u
u
2
1
=t
.
d=
kr
0
(9.122)
It can be seen that the skin-depth for a good conductor decreases with increasing
wave frequency. The fact that kr d = 1 indicates that the wave only penetrates a
few wave-lengths into the conductor before decaying away.
Now the power per unit volume dissipated via ohmic heating in a conducting
medium takes the form
P = j E = E2 .
(9.123)
Consider an electromagnetic wave of the form (9.119). The mean power dissipated per unit area in the region z > 0 is written
v
Z
u
d
u
E02 e2 z/d dz =
(9.124)
hPi =
E02 = t
E02 ,
2 0
4
8 0
for a good conductor. Now, according to Eq. (9.91), the mean electromagnetic
power flux into the region z > 0 takes the form
E B ^z
hui =
0
*
z=0
1 E02 kr u
u
=
=t
E02 .
2 0
8 0
v
(9.125)
It is clear, from a comparison of the previous two equations, that all of the wave
energy which flows into the region z > 0 is dissipated via ohmic heating. We thus
conclude that the attenuation of an electromagnetic wave propagating through a
conductor is a direct consequence of ohmic power losses.
Consider a typical metallic conductor such as copper, whose electrical conductivity at room temperature is about 6 107 ( m)1 . Copper, therefore, acts as a
good conductor for all electromagnetic waves of frequency below about 1018 Hz.
The skin-depth in copper for such waves is thus
d=
v
u
u
t
2
'
0
278
6
cm.
(Hz)
(9.126)
9.11 Dielectric constant of a collisional plasma
It follows that the skin-depth is about 6 cm at 1 Hz, but only about 2 mm at 1 kHz.
This gives rise to the so-called skin-effect in copper wires, by which an oscillating
electromagnetic signal of increasing frequency, transmitted along such a wire, is
confined to an increasingly narrow layer (whose thickness is of order the skindepth) on the surface of the wire.
The conductivity of sea water is only about ' 5 ( m)1 . However, this is still
sufficiently high for sea water to act as a good conductor for all radio frequency
electromagnetic waves (i.e., < 109 Hz). The skin-depth at 1 MHz ( 2 km) is
about 0.2 m, whereas that at 1 kHz ( 2000 km) is still only about 7 m. This obviously poses quite severe restrictions for radio communication with submerged
submarines. Either the submarines have to come quite close to the surface to
communicate (which is dangerous), or the communication must be performed
with extremely low frequency (ELF) waves (i.e., f < 100 Hz). Unfortunately, such
waves have very large wave-lengths ( > 20, 000 km), which means that they can
only be efficiently generated by extremely large antennas.

We have now investigated electromagnetic wave propagation through two different media possessing free electrons: plasmas (see Sect. 9.8), and ohmic conductors (see Sect. 9.10). In the first case, we obtained the dispersion relation
(9.81), whereas in the second we obtained the quite different dispersion relation
(9.117). This leads us, quite naturally, to ask what the essential distinction is between the response of free electrons in a plasma to an electromagnetic wave, and
that of free electrons in an ohmic conductor. It turns out that the main distinction
is the relative strength of electron-ion collisions.
In the presence of electron-ion collisions, we can model the equation of motion
of an individual electron in a plasma or a conductor as
dv
+ me v = e E,
(9.127)
me
dt
where E is the wave electric field. The collision term (i.e., the second term on
the left-hand side) takes the form of a drag force proportional to v. In the
279
absence of the wave electric field, this force damps out any electron motion on
the typical time-scale 1 . Since, in reality, an electron loses virtually all of its
directed momentum during a collision with a much more massive ion, we can
regard as the effective electron-ion collision frequency.
Assuming the usual exp(i t) time-dependence of perturbed quantities, we
can solve Eq. (9.127) to give
v = i r =
ieE
.
me ( + i )
(9.128)
Hence, the perturbed current density can be written

j = e ne v =
i n e e2 E
,
me ( + i )
(9.129)
where ne is the number density of free electrons. It follows that the effective
conductivity of the medium takes the form
j
i n e e2
= =
.
E me ( + i )
(9.130)
Now, the mean rate of ohmic heating per unit volume in the medium is written
hPi =
1
Re() E02 ,
2
(9.131)
where E0 is the amplitude of the wave electric field. Note that only the real
part of contributes to ohmic heating, because the perturbed current must be in
phase with the wave electric field in order for there to be a net heating effect. An
imaginary gives a perturbed current which is in phase quadrature with the wave
electric field. In this case, there is zero net transfer of power between the wave
and the plasma over a wave period. We can see from Eq. (9.130) that in the limit
in which the wave frequency is much larger than the collision frequency (i.e.,
), the effective conductivity of the medium becomes purely imaginary:
i n e e2
.
'
me
280
(9.132)
9.12 Reflection at a dielectric boundary
In this limit, there is no loss of wave energy due to ohmic heating, and the
medium acts like a conventional plasma. In the opposite limit, in which the wave
frequency is much less than the collision frequency (i.e., ), the effective
conductivity becomes purely real:
ne e2
.
(9.133)
'
me
In this limit, ohmic heating losses are significant, and the medium acts like a
conventional ohmic conductor.
Following the analysis of Sect. 9.7, we can derive the following dispersion
relation from Eq. (9.128):
p2
.
(9.134)
k c =
+ i
It can be seen that, in the limit , the above dispersion relation reduces to
the dispersion relation (9.81) for a conventional (i.e., collisionless) plasma. In
the opposite limit, we obtain
2 2
p2
2
= 0 (0 + i ).
(9.135)
k = 2 +i
c
c2
where use has been made of Eq (9.133). Of course, the above dispersion relation
is identical to the dispersion relation (9.117) (with = 1) which we previously
derived for an ohmic conductor.
2
Our main conclusion from this subsection is that the dispersion relation (9.134)
can be used to describe electromagnetic wave propagation through both a collisional plasma and an ohmic conductor. We can also deduce that in the low
frequency limit, , a collisional plasma acts very much like an ohmic conductor, whereas in the high frequency limit, , an ohmic conductor acts
very much like a collisionless plasma.

An electromagnetic wave of real (positive) frequency can be written
E(r, t) = E0 e i (kr t) ,
281
(9.136)

x
reflected wave
transmitted wave
incident wave
medium 1
medium 2
z
z=0
Figure 55:
B(r, t) = B0 e i (kr t) .
(9.137)
The wave-vector, k, indicates the direction of propagation of the wave, and also
its phase-velocity, v, via
v= .
(9.138)
k
Since the wave is transverse in nature, we must have E0 k = B0 k = 0. Finally,
the familiar Maxwell equation
E=
B
t
(9.139)
leads us to the following relation between the constant vectors E0 and B0 :

B0 =
^ E0
k
.
v
(9.140)
^ = k/k is a unit vector pointing in the direction of wave propagation.

Here, k
Suppose that the plane z = 0 forms the boundary between two different dielectric media. Let medium 1, of refractive index n1 , occupy the region z < 0,
whilst medium 2, of refractive index n2 , occupies the region z > 0. Let us investigate what happens when an electromagnetic wave is incident on this boundary
from medium 1.
282
Consider, first of all, the simple case of incidence normal to the boundary (see
^ = +^z for the incident and transmitted waves, and k
^ = ^z
Fig. 55). In this case, k
for the reflected wave. Without loss of generality, we can assume that the incident
wave is polarized in the x-direction. Hence, using Eq. (9.140), the incident wave
can be written
^,
E(z, t) = Ei e i (k1 z t) x
Ei i (k1 z t)
^,
y
e
B(z, t) =
v1
(9.141)
(9.142)
where v1 = c/n1 is the phase-velocity in medium 1, and k1 = /v1 . Likewise, the

reflected wave takes the form
^,
E(z, t) = Er e i (k1 z t) x
Er
^.
B(z, t) = e i (k1 z t) y
v1
(9.143)
(9.144)
Finally, the transmitted wave can be written

^,
E(z, t) = Et e i (k2 z t) x
Et i (k2 z t)
^,
e
y
B(z, t) =
v2
(9.145)
(9.146)
where v2 = c/n2 is the phase-velocity in medium 2, and k2 = /v2 .

For the case of normal incidence, the electric and magnetic components of
all three waves are parallel to the boundary between the two dielectric media.
Hence, the appropriate boundary conditions to apply at z = 0 are
Ek 1 = E k 2 ,
(9.147)
Bk 1 = B k 2 .
(9.148)
The latter condition derives from the general boundary condition H k 1 = Hk 2 , and
the fact that B = 0 H in both media (which are assumed to be non-magnetic).
Application of the boundary condition (9.147) yields
Ei + E r = E t .
283
(9.149)
Likewise, application of the boundary condition (9.148) gives

Ei E r
Et
= ,
v1
v2
(9.150)
or
n2
v1
Et =
Et ,
(9.151)
v2
n1
since v1 /v2 = n2 /n1 . Equations (9.149) and (9.151) can be solved to give
Ei E r =
Er =
Et =
n1 n 2
Ei ,
n1 + n 2
!
2 n1
Et .
n1 + n 2
!
(9.152)
(9.153)
Thus, we have determined the amplitudes of the reflected and transmitted waves
in terms of the amplitude of the incident wave.
It can be seen, first of all, that if n1 = n2 then Er = 0 and Et = Ei . In other
words, if the two media have the same indices of refraction then there is no reflection at the boundary between them, and the transmitted wave is consequently
equal in amplitude to the incident wave. On the other hand, if n 1 6= n2 then there
is some reflection at the boundary. Indeed, the amplitude of the reflected wave
is roughly proportional to the difference between n1 and n2 . This has important
practical consequences. We can only see a clean pane of glass in a window because some of the light incident at an air/glass boundary is reflected, due to the
different refractive indicies of air and glass. As is well-known, it is a lot more
difficult to see glass when it is submerged in water. This is because the refractive
indices of glass and water are quite similar, and so there is very little reflection of
light incident on a water/glass boundary.
According to Eq. (9.152), Er /Ei < 0 when n2 > n1 . The negative sign indicates
a 180 phase-shift of the reflected wave, with respect to the incident wave. We
conclude that there is a 180 phase-shift of the reflected wave, relative to the
incident wave, on reflection from a boundary with a medium of greater refractive
index. Conversely, there is no phase-shift on reflection from a boundary with a
medium of lesser refractive index.
284
The mean electromagnetic energy flux, or intensity, in the z-direction is simply

hE B ^zi E0 B0
E02
.
I=
=
=
0
2 0
2 0 v
(9.154)
The coefficient of reflection, R, is defined as the ratio of the intensities of the

reflected and incident waves:
Ir
Er
R= =
Ii
Ei
!2
(9.155)
Likewise, the coefficient of transmission, T , is the ratio of the intensities of the

transmitted and incident waves:
v1 Et
It
T= =
Ii
v2 Ei
!2
n2 E t
=
n1 E i
!2
(9.156)
Equations (9.152), (9.153), (9.155), and (9.156) yield

R =
n1 n 2
n1 + n 2
!2
(9.157)
2 n1
n2
T =
n1 n1 + n 2
!2
(9.158)
Note that R + T = 1. In other words, any wave energy which is not reflected at
the boundary is transmitted, and vice versa.
Let us now consider the case of incidence oblique to the boundary (see Fig. 56).
Suppose that the incident wave subtends an angle i with the normal to the
boundary, whereas the reflected and transmitted waves subtend angles r and t ,
respectively.
The incident wave can be written
E(r, t) = Ei e i (ki r t) ,
(9.159)
B(r, t) = Bi e i (ki r t) ,
(9.160)
with analogous expressions for the reflected and transmitted waves. Since, in
the case of oblique incidence, the electric and magnetic components of the wave
285
reflected wave
x
transmitted wave
i
medium 1
medium 2
incident wave
z=0
Figure 56:
fields are no longer necessarily parallel to the boundary, the boundary conditions
(9.147) and (9.148) at z = 0 must be supplemented by the additional boundary
conditions
1 E 1 = 2 E 2 ,
(9.161)
B 1 = B 2 .
(9.162)
Equation (9.161) derives from the general boundary condition D 1 = D 2 .

It follows from Eqs. (9.148) and (9.162) that both components of the magnetic
field are continuous at the boundary. Hence, we can write
Bi e i (ki r t) + Br e i (kr r t) = Bt e i (kt r t)
(9.163)
at z = 0. Given that Bi , Br , and Bt are constant vectors, the only way in which
the above equation can be satisfied for all values of x and y is if
ki r = k r r = k t r
(9.164)
throughout the z = 0 plane. This, in turn, implies that

ki x = k r x = k t x
(9.165)
ki y = k r y = k t y .
(9.166)
and
286
It immediately follows that if ki y = 0 then kr y = kt y = 0. In other words, if the

incident wave lies in the x-z plane then the reflected and transmitted waves also
lie in the x-z plane. Another way of putting this is that the incident, reflected,
and transmitted waves all lie in the same plane, know as the plane of incidence.
This, of course, is one of the laws of geometric optics. From now on, we shall
assume that the plane of incidence is the x-z plane.
Now, ki = kr = /v1 and kt = /v2 . Moreover,
kx i
,
(9.167)
ki
with similar expressions for r and t . Hence, according to Eq. (9.165),
sin i =
sin r = sin i ,
(9.168)
which implies that r = i . Moreover,

v2
n1
sin t
=
= .
(9.169)
sin i
v1
n2
Of course, the above expressions correspond to the law of reflection and Snells
law of refraction, respectively.
For the case of oblique incidence, we need to consider two independent wave
polarizations separately. The first polarization has all the wave electric fields perpendicular to the plane of incidence, whilst the second has all the wave magnetic
fields perpendicular to the plane of incidence.
Let us consider the first wave polarization. We can write unit vectors in the
directions of propagation of the incident, reflected, and transmitted waves likso:
î = (sin i , 0, cos i ) ,
k
(9.170)
^r = (sin i , 0, cos i ) ,
k
^t = (sin t , 0, cos t ) .
k
(9.171)
(9.172)
The constant vectors associated with the incident wave are written
^,
Ei = E i y
Ei
Bi =
( cos i , 0, sin i ) ,
v1
287
(9.173)
(9.174)
where use has been made of Eq. (9.140). Likewise, the constant vectors associated with the reflected and transmitted waves are
^,
Er = E r y
Er
Br =
(cos i , 0, sin i ) ,
v1
(9.175)
(9.176)
and
^,
Et = E t y
Et
Bt =
( cos t , 0, sin t ) ,
v2
(9.177)
(9.178)
respectively.
Now, the boundary condition (9.147) yields Ey 1 = Ey 2 , or
Ei + E r = E t .
(9.179)
Likewise, the boundary condition (9.162) gives Bz 1 = Bz 2 , or

sin i
sin t
(Ei + Er )
= Et
.
(9.180)
v1
v2
However, using Snells law, (9.169), the above expression reduces to Eq. (9.179).
Finally, the boundary condition (9.148) yields Bx 1 = Bx 2 , or
cos i
cos t
(Ei Er )
= Et
.
(9.181)
v1
v2
It is convenient to define the parameters
cos t
=
,
(9.182)
cos i
and
v1
n2
=
= .
(9.183)
v2
n1
Equations (9.179) and (9.181) can be solved in terms of these parameters to give
!
1
Er =
Ei ,
(9.184)
1 +
!
2
Et =
Ei .
(9.185)
1 +
288
These relations are known as Fresnel equations.

The wave intensity in the z-direction is given by
hE B ^zi E0 B0 cos E02 cos
=
=
.
Iz =
0
2 0
2 0 v
(9.186)
Hence, the coefficient of reflection is written

Er
R=
Ei
!2
1
=
1 +
!2
(9.187)
whereas the coefficient of transmission takes the form

cos t v1 Et
T=
cos i v2 Ei
!2
2
=
1 +
!2
(9.188)
Note that it is again the case that R + T = 1.

Let us now consider the second wave polarization. In this case, the constant
vectors associated with the incident, reflected, and transmitted waves are written
Ei = Ei (cos i , 0, sin i ),
Ei
^,
y
Bi =
v1
(9.189)
Er = Er (cos i , 0, sin i ),
Er
^,
Br = y
v1
(9.191)
Et = Et (cos t , 0, sin t ),
Et
^,
Bt =
y
v2
(9.193)
(9.190)
and
(9.192)
and
(9.194)
respectively. The boundary condition (9.148) yields By 1 = By 2 , or

Ei E r
Et
= .
v1
v2
289
(9.195)
Likewise, the boundary condition (9.147) gives Ex 1 = Ex 2 , or

(Ei + Er ) cos i = Et cos t .
(9.196)
Finally, the boundary condition (9.161) yields 1 Ez 1 = 2 Ez 2 , or

1 (Ei Er ) sin i = 2 Ei sin t .
(9.197)
Making use of Snells law, and the fact that = n2 , the above expression reduces
to Eq. (9.195).
Solving Eqs. (9.165) and (9.196), we obtain
Er =
Et =
Ei ,
+
!
2
Ei .
+
!
(9.198)
(9.199)
The associated coefficients of reflection and transmission take the form

R =
!2
(9.200)
2
T =
+
!2
(9.201)
respectively. As usual, R + T = 1.
Note that at oblique incidence the Fresnel equations, (9.184) and (9.185), for
the wave polarization in which the electric field is parallel to the boundary are
different to the Fresnel equations, (9.198) and (9.199), for the wave polarization
in which the magnetic field is parallel to the boundary. This implies that the
coefficients of reflection and transmission for these two wave polarizations are,
in general, different.
Figure 57 shows the coefficients of reflection (solid curves) and transmission
(dashed curves) for oblique incidence from air (n1 = 1.0) to glass (n2 = 1.5). The
left-hand panel shows the wave polarization for which the electric field is parallel
to the boundary, whereas the right-hand panel shows the wave polarization for
which the magnetic field is parallel to the boundary. In general, it can be seen
290
Figure 57:
that the coefficient of reflection rises, and the coefficient of transmission falls,
as the angle of incidence increases. Note, however, that for the second wave
polarization there is a particular angle of incidence, know as the Brewster angle,
at which the reflected intensity is zero. There is no similar behaviour for the first
wave polarization.
It follows from Eq. (9.198) that the Brewster angle corresponds to the condition
= ,
(9.202)
or
cos2 t
1 sin2 t
1 sin2 i /2
=
=
=
,
(9.203)
cos2 i
1 sin2 i
1 sin2 i
where use has been made of Snells law. The above expression reduces to
2
sin i =
,
1 + 2
(9.204)
or tan i = = n2 /n1 . Hence, the Brewster angle satisfies

B = tan
291
n2
.
n1
!
(9.205)
9.13 Wave-guides
If unpolarized light is incident on an air/glass (say) boundary at the Brewster

angle then the reflected beam is 100% plane polarized.
9.13 Wave-guides
A wave-guide is a hollow conducting pipe, of uniform cross-section, used to transport high frequency electromagnetic waves (generally, in the microwave band)
from one point to another. The main advantage of wave-guides is their relatively
low level of radiation losses (since the electric and magnetic fields are completely
enclosed by a conducting wall) compared to transmission lines.
Consider a vacuum-filled wave-guide which runs parallel to the z-axis. An electromagnetic wave trapped inside the wave-guide satisfies Maxwells equations for
free space:
E = 0,
B = 0,
(9.206)
(9.207)
B
,
(9.208)
t
1 E
(9.209)
B = 2 .
c t
Let /t i , and /z i k, where is the wave frequency, and k the
wave-number parallel to the axis of the wave-guide. It follows that
E =
Ex Ey
+
+ i k Ez = 0,
x
y
Bx By
+
+ i k Bz = 0,
x
y
Ez
i Bx =
i k Ey ,
y
Ez
i By =
+ i k Ex ,
x
Ey Ex
i Bz =
,
x
y
292
(9.210)
(9.211)
(9.212)
(9.213)
(9.214)
9.13 Wave-guides
Bz
E
=
+ i k By ,
x
c2
y
Bz
i 2 Ey =
i k Bx ,
c
x
By Bx
i 2 Ez =
+
.
c
x
y
(9.215)
(9.216)
(9.217)
Equations (9.213) and (9.215) yield

Ez
Bz
+k
Ex = i
y
x
! 2
c2
k2
and
Bz
Ez
+k
By = i 2
c x
y
Likewise, Eqs. (9.212) and (9.216) yield
! 2
c2
Ez
Bz
+k
Ey = i
x
y
These equations can be combined to give
(9.219)
(9.220)
(9.221)
(9.222)
k2
2
c
and
Ez
Bz
Bx = i 2
+k
c y
x
k2
(9.218)
k2
2
c
Et
2
= i ( Bz ^z + k Ez ) 2 k2
c
Bt
= i 2 Ez ^z + k Bz
c
! 2
c2
k2
(9.223)
Here, Et and Bt are the transverse electric and magnetic fields: i.e., the electric
and magnetic fields in the x-y plane. It is clear, from Eqs. (9.222) and (9.223),
that the transverse fields are fully determined once the longitudinal fields, E z and
Bz , are known.
293
9.13 Wave-guides
Substitution of Eqs. (9.222) and (9.223) into Eqs. (9.214) and (9.217) yields
the equations satisfied by the longitudinal fields:
2
2
2
+ 2 Ez +
k2 Ez = 0,
2
2
x
y
c
2
2
2
+
k2 Bz = 0.
Bz +
x2 y2
c2
(9.224)
(9.225)
The remaining equations, (9.210) and (9.211), are automatically satisfied provided Eqs. (9.222)(9.225) are satisfied.
We expect E = B = 0 inside the walls of the wave-guide, assuming that
they are perfectly conducting. Hence, the appropriate boundary conditions at
the walls are
Ek = 0,
(9.226)
B = 0.
(9.227)
It follows, by inspection of Eqs. (9.222) and (9.223), that these boundary conditions are satisfied provided
Ez = 0,
(9.228)
^ Bz = 0,
n
(9.229)
^ is the normal vector to the walls. Hence, the electromagat the walls. Here, n
netic fields inside the wave-guide are fully specified by solving Eqs. (9.224) and
(9.225), subject to the boundary conditions (9.228) and (9.229), respectively.
Equations (9.224) and (9.225) support two independent types of solution. The
first type has Ez = 0, and is consequently called a transverse electric, or TE, mode.
Conversely, the second type has Bz = 0, and is called a transverse magnetic, or
TM, mode.
Consider the specific example of a rectangular wave-guide, with conducting
walls at x = 0, a, and y = 0, b. For a TE mode, the longitudinal magnetic field
can be written
Bz (x, y) = B0 cos(kx x) cos(ky y),
(9.230)
294
9.13 Wave-guides
The boundary condition (9.229) requires that Bz /x = 0 at x = 0, a, and

Bz /y = 0 at y = 0, b. It follows that
n
,
(9.231)
kx =
a
m
,
(9.232)
ky =
b
where n = 0, 1, 2, , and m = 0, 1, 2, . Clearly, there are many different kinds
of TE mode, corresponding to the many different choices of m and n. Let us
refer to a mode corresponding to a particular choice of m, n as a TEmn mode.
Note, however, that there is no TE00 mode, since Bz (x, y) is uniform in this case.
According to Eq. (9.225), the dispersion relation for the TEmn mode is given by
2
,
k2 c2 = 2 mn
(9.233)
v
u 2
un
ct 2
a
(9.234)
where
mn =
m2
+ 2.
b
According to the dispersion relation (9.233), k is imaginary for < mn . In

other words, for wave frequencies below mn , the TEmn mode fails to propagate
down the wave-guide, and is instead attenuated. Hence, mn is termed the cutoff frequency for the TEmn mode. Assuming that a > b, the TE mode with the
lowest cut-off frequency is the TE10 mode, where
c
.
(9.235)
10 =
a
For frequencies above the cut-off frequency, the phase-velocity of the TE mn
mode is given by
c
vp =
=q
,
(9.236)
2 /2
k
1 mn
which is greater than c. However, the group-velocity takes the form

q
d
2 /2 ,
= c 1 mn
vg =
(9.237)
dk
which is always less than c. Of course, energy is transmitted down the wave-guide
at the group-velocity, rather than the phase-velocity. Note that the group-velocity
goes to zero as the wave frequency approaches the cut-off frequency.
295
9.13 Wave-guides
For a TM mode, the longitudinal electric field can be written

Ez (x, y) = E0 sin(kx x) sin(ky y),
(9.238)
The boundary condition (9.228) requires that Ez = 0 at x = 0, a, and y = 0, b. It

follows that
n
,
(9.239)
kx =
a
m
ky =
,
(9.240)
b
where n = 1, 2, , and m = 1, 2, . The dispersion relation for the TM mn mode
is also given by Eq. (9.233). Hence, Eqs. (9.236) and (9.237) also apply to TM
modes. However, the TM mode with the lowest cut-off frequency is the TM11
mode, where
v
u
u 1
1
(9.241)
11 = c t 2 + 2 > 10 .
a
b
It follows that the mode with the lowest cut-off frequency is always a TE mode.
There is, in principle, a third type of mode which can propagate down a waveguide. This third mode type is characterized by Ez = Bz = 0, and is consequently
called a transverse electromagnetic, or TEM, mode. It is easily seen, from an inspection of Eqs. (9.212)(9.217), that a TEM mode satisfies
2 = k 2 c 2 ,
(9.242)
and
Et = ,
(9.243)
Bt = c1 ^z,
(9.244)
2 = 0.
(9.245)
= constant
(9.246)
where (x, y) satisfies

The boundary conditions (9.228) and (9.229) imply that
296
9.13 Wave-guides
at the walls. However, there is no non-trivial solution of Eqs. (9.245) and (9.246)
for a conventional wave-guide. In other words, conventional wave-guides do not
support TEM modes. In fact, it turns out that only wave-guides with central conductors support TEM modes. Consider, for instance, a co-axial wave-guide in
which the electric and magnetic fields are trapped between two parallel concentric cylindrical conductors of radius a and b (with b > a). In this case, = (r),
and Eq. (9.245) reduces to
!
1
r
= 0,
(9.247)
r r
r
where r is a standard cylindrical polar coordinate. The boundary condition
(9.246) is automatically satisfied at r = a and r = b. The above equation has the
following non-trivial solution:
(r) = b ln(r/b).
(9.248)
Note, however, that the inner conductor must be present, otherwise

as r 0, which is unphysical. According to the dispersion relation (9.242),
TEM modes have no cut-off frequency, and have the phase-velocity (and groupvelocity) c. Indeed, this type of mode is the same as that supported by a transmission line (see Sect. 7.7).
297
10 RELATIVITY AND ELECTROMAGNETISM
10
Relativity and electromagnetism
10.1 Introdunction
In this section, we shall discuss Maxwells equations in the light of Einsteins
special theory of relativity.
10.2 The relativity principle

Physical phenomena are conventionally described relative to some frame of reference which allows us to define fundamental quantities such as position and time.
Of course, there are very many different ways of choosing a reference frame, but
it is generally convenient to restrict our choice to the set of rigid inertial frames.
A classical rigid reference frame is the imagined extension of a rigid body. For
instance, the Earth determines a rigid frame throughout all space, consisting of
all those points which remain rigidly at rest relative to the Earth, and to each
other. We can associate an orthogonal Cartesian coordinate system S with such
a frame, by choosing three mutually orthogonal planes within it, and measuring
x, y, and z as perpendicular distances from these planes. A time coordinate must
also be defined, in order that the system can be used to specify events. A rigid
frame, endowed with such properties, is called a Cartesian frame. The description given above presupposes that the underlying geometry of space is Euclidian,
which is reasonable provided that gravitational effects are negligible (we shall
assume that this is the case). An inertial frame is a Cartesian frame in which
free particles move without acceleration, in accordance with Newtons first law
of motion. There are an infinite number of different inertial frames, moving with
some constant velocity with respect to one another.
The key to understanding special relativity is Einsteins relativity principle,
which states that:
All inertial frames are totally equivalent for the performance of all physical
experiments.
298
10.2 The relativity principle
In other words, it is impossible to perform a physical experiment which differentiates in any fundamental sense between different inertial frames. By definition,
Newtons laws of motion take the same form in all inertial frames. Einstein generalized this result in his special theory of relativity by asserting that all laws of
physics take the same form in all inertial frames.
Consider a wave-like disturbance. In general, such a disturbance propagates
at a fixed velocity with respect to the medium in which the disturbance takes
place. For instance, sound waves (at S.T.P.) propagate at 343 meters per second
with respect to air. So, in the inertial frame in which air is stationary, sound
waves appear to propagate at 343 meters per second. Sound waves appear to
propagate at a different velocity any inertial frame which is moving with respect
to the air. However, this does not violate the relativity principle, since if the air
were stationary in the second frame then sound waves would appear to propagate
at 343 meters per second in this frame as well. In other words, exactly the same
experiment (e.g., the determination of the speed of sound relative to stationary
air) performed in two different inertial frames of reference yields exactly the
same result, in accordance with the relativity principle.
Consider, now, a wave-like disturbance which is self-regenerating, and does
not require a medium through which to propagate. The most well-known example of such a disturbance is a light wave. Another example is a gravity wave.
According to electromagnetic theory, the speed of propagation of a light wave
through a vacuum is
1
= 2.99729 108 meters per second,
c=
0 0
(10.1)
where 0 and 0 are physical constants which can be evaluated by performing

two simple experiments which involve measuring the force of attraction between
two fixed changes and two fixed parallel current carrying wires. According to the
relativity principle, these experiments must yield the same values for 0 and 0 in
all inertial frames. Thus, the speed of light must be the same in all inertial frames.
In fact, any disturbance which does not require a medium to propagate through
must appear to travel at the same velocity in all inertial frames, otherwise we
could differentiate inertial frames using the apparent propagation speed of the
299
10.3 The Lorentz transformation
disturbance, which would violate the relativity principle.

Consider two Cartesian frames S(x, y, z, t) and S 0 (x 0 , y 0 , z 0 , t 0 ) in the standard configuration, in which S 0 moves in the x-direction of S with uniform velocity v, and
the corresponding axes of S and S 0 remain parallel throughout the motion, having
coincided at t = t 0 = 0. It is assumed that the same units of distance and time are
adopted in both frames. Suppose that an event (e.g., the flashing of a light-bulb,
or the collision of two point particles) has coordinates (x, y, z, t) relative to S,
and (x 0 , y 0 , z 0 , t 0 ) relative to S 0 . The common sense relationship between these
two sets of coordinates is given by the Galilean transformation:
x 0 = x v t,
(10.2)
y 0 = y,
(10.3)
z 0 = z,
(10.4)
t 0 = t.
(10.5)
This transformation is tried and tested, and provides a very accurate description
of our everyday experience. Nevertheless, it must be wrong! Consider a light
wave which propagates along the x-axis in S with velocity c. According to the
Galilean transformation, the apparent speed of propagation in S 0 is c v, which
violates the relativity principle. Can we construct a new transformation which
makes the velocity of light invariant between different inertial frames, in accordance with the relativity principle, but reduces to the Galilean transformation at
low velocities, in accordance with our everyday experience?
Consider an event P, and a neighbouring event Q, whose coordinates differ by
dx, dy, dz, dt in S, and by dx 0 , dy 0 , dz 0 , dt 0 in S 0 . Suppose that at the event P
a flash of light is emitted, and that Q is an event in which some particle in space
is illuminated by the flash. In accordance with the laws of light-propagation,
and the invariance of the velocity of light between different inertial frames, an
300
observer in S will find that

dx2 + dy2 + dz2 c2 dt2 = 0
(10.6)
for dt > 0, and an observer in S 0 will find that

dx 02 + dy 02 + dz 02 c2 dt 02 = 0
(10.7)
for dt 0 > 0. Any event near P whose coordinates satisfy either (10.6) or (10.7) is
illuminated by the flash from P, and, therefore, its coordinates must satisfy both
(10.6) and (10.7). Now, no matter what form the transformation between coordinates in the two inertial frames takes, the transformation between differentials
at any fixed event P is linear and homogeneous. In other words, if
x 0 = F(x, y, z, t),
(10.8)
where F is a general function, then

dx 0 =
F
F
F
F
dx +
dy +
dz +
dt.
x
y
z
t
(10.9)
It follows that
dx 02 + dy 02 + dz 02 c2 dt 02 = a dx2 + b dy2 + c dz2 + d dt2 + g dx dt + h dy dt
+k dz dt + l dy dz + m dx dz + n dx dy, (10.10)
where a, b, c, etc. are functions of x, y, z, and t. We know that the right-hand
side of the above expression vanishes for all real values of the differentials which
satisfy Eq. (10.6). It follows that the right-hand side is a multiple of the quadratic
in Eq. (10.6): i.e.,
dx 02 + dy 02 + dz 02 c2 dt 02 = K (dx2 + dy2 + dz2 c2 dt2 ),
(10.11)
where K is a function of x, y, z, and t. [We can prove this by substituting into

in
Eq. (10.10) the following obvious
zeros
of the quadratic
Eq. (10.6):
(1,
0, 0, 1),
(0, 1, 0, 1), (0, 0, 1, 1), (0, 1/ 2, 1/ 2, 1), (1/ 2, 0, 1/ 2, 1), (1/ 2, 1/ 2, 0, 1):
and solving the resulting conditions on the coefficients.] Note that K at P is also
independent of the choice of standard coordinates in S and S 0 . Since the frames
are Euclidian, the values of dx2 + dy2 + dz2 and dx 02 + dy 02 + dz 02 relevant to P
301
and Q are independent of the choice of axes. Furthermore, the values of dt2 and
dt 02 are independent of the choice of the origins of time. Thus, without affecting
the value of K at P, we can choose coordinates such that P = (0, 0, 0, 0) in both
S and S 0 . Since the orientations of the axes in S and S 0 are, at present, arbitrary,
and since inertial frames are isotropic, the relation of S and S 0 relative to each
other, to the event P, and to the locus of possible events Q, is now completely
symmetric. Thus, we can write
dx2 + dy2 + dz2 c2 dt2 = K (dx 02 + dy 02 + dz 02 c2 dt 02 ),
(10.12)
in addition to Eq. (10.11). It follows that K = 1. K = 1 can be dismissed

immediately, since the intervals dx2 + dy2 + dz2 c2 dt2 and dx 02 + dy 02 + dz 02
c2 dt 02 must coincide exactly when there is no motion of S 0 relative to S. Thus,
dx 02 + dy 02 + dz 02 c2 dt 02 = dx2 + dy2 + dz2 c2 dt2 .
(10.13)
Equation (10.13) implies that the transformation equations between primed and
unprimed coordinates must be linear. The proof of this statement is postponed
until Sect. 10.7.
The linearity of the transformation allows the coordinate axes in the two
frames to be orientated so as to give the standard configuration mentioned earlier.
Consider a fixed plane in S with the equation l x+m y+n z+p = 0. In S 0 , this becomes (say) l (a1 x 0 +b1 y 0 +c1 z 0 +d1 t 0 +e1 )+m (a2 x 0 + )+n (a3 x 0 + )+p = 0,
which represents a moving plane unless l d1 + m d2 + n d3 = 0. That is, unless the normal vector to the plane in S, (l, m, n), is perpendicular to the vector
(d1 , d2 , d3 ). All such planes intersect in lines which are fixed in both S and S 0 ,
and which are parallel to the vector (d1 , d2 , d3 ) in S. These lines must correspond
to the direction of relative motion of the frames. By symmetry, two such frames
which are orthogonal in S must also be orthogonal in S 0 . This allows the choice
of two common coordinate planes.
Under a linear transformation, the finite coordinate differences satisfy the
same transformation equations as the differentials. It follows from Eq. (10.13),
assuming that the events (0, 0, 0, 0) coincide in both frames, that for any event
with coordinates (x, y, z, t) in S and (x 0 , y 0 , z 0 , t 0 ) in S 0 , the following relation
holds:
x2 + y2 + z2 c2 t2 = x 02 + y 02 + z 02 c2 t 02 .
(10.14)
302
By hypothesis, the coordinate planes y = 0 and y 0 = 0 coincide permanently.

Thus, y = 0 must imply y 0 = 0, which suggests that
y 0 = A y,
(10.15)
where A is a constant. We can reverse the directions of the x- and z-axes in S and
S 0 , which has the effect of interchanging the roles of these frames. This procedure
does not affect Eq. (10.15), but by symmetry we also have
y = A y 0.
(10.16)
It is clear that A = 1. The negative sign can again be dismissed, since y = y 0

when there is no motion between S and S 0 . The argument for z is similar. Thus,
we have
y 0 = y,
(10.17)
z 0 = z,
(10.18)
as in the Galilean transformation.

Equations (10.14), (10.17) and (10.18) yield
x2 c2 t2 = x 02 c2 t 02 .
(10.19)
Since, x 0 = 0 must imply x = v t, we can write

x 0 = B (x v t),
(10.20)
where B is a constant (possibly depending on v). It follows from the previous two
equations that
t 0 = C x + D t,
(10.21)
where C and D are constants (possibly depending on v). Substituting Eqs. (10.20)
and (10.21) into Eq. (10.19), and comparing the coefficients of x2 , x t, and t2 ,
we obtain
1
B=D =
,
(10.22)
(1 v2 /c2 )1/2
v/c2
.
C =
(1 v2 /c2 )1/2
303
(10.23)
10.4 Transformation of velocities
We must choose the positive sign in order to ensure that x 0 x as v/c 0.

Thus, collecting our results, the transformation between coordinates in S and S 0
is given by
x0 =
x vt
,
(1 v2 /c2 )1/2
(10.24)
y 0 = y,
(10.25)
z 0 = z,
(10.26)
t v x/c2
t =
.
(1 v2 /c2 )1/2
(10.27)
This is the famous Lorentz transformation. It ensures that the velocity of light is
invariant between different inertial frames, and also reduces to the more familiar
Galilean transform in the limit v c. We can solve Eqs. (10.24)(10.27) for x,
y, z, and t, to obtain the inverse Lorentz transformation:
x0 + v t0
x =
,
(1 v2 /c2 )1/2
(10.28)
y = y 0,
(10.29)
z = z 0,
t 0 + v x 0 /c2
.
t =
(1 v2 /c2 )1/2
(10.30)
(10.31)
Not surprizingly, the inverse transformation is equivalent to a Lorentz transformation in which the velocity of the moving frame is v along the x-axis, instead
of +v.

Consider two frames, S and S 0 , in the standard configuration. Let u be the velocity
of a particle in S. What is the particles velocity in S 0 ? The components of u are
u1 =
304
dx
,
dt
(10.32)
dy
,
dt
dz
=
.
dt
u2 =
(10.33)
u3
(10.34)
Similarly, the components of u 0 are

dx 0
,
(10.35)
=
dt 0
dy 0
0
,
(10.36)
u2 =
dt 0
dz 0
u30 =
.
(10.37)
dt 0
Now we can write Eqs. (10.24)(10.27) in the form dx 0 = (dxv dt), dy 0 = dy,
dz 0 = dz, and dt 0 = (dt v dx/c2 ), where
u10
= (1 v2 /c2 )1/2
(10.38)
is the well-known Lorentz factor. If we substitute these differentials into Eqs. (10.32)
(10.34), and make use of Eqs. (10.35)(10.37), we obtain the transformation
rules
u1 v
u10 =
,
(10.39)
1 u1 v/c2
u2
,
(10.40)
u20 =
(1 u1 v/c2 )
u3
u30 =
.
(10.41)
(1 u1 v/c2 )
As in the transformation of coordinates, we can obtain the inverse transform by
interchanging primed and unprimed symbols, and replacing +v with v. Thus,
u1
u10 + v
,
=
1 + u10 v/c2
u20
,
(1 + u10 v/c2 )
(10.43)
u30
.
=
(1 + u10 v/c2 )
(10.44)
u2 =
u3
(10.42)
305
Equations (10.42)(10.44) can be regarded as giving the resultant, u = (u1 , u2 , u3 ),

of two velocities, v = (v, 0, 0) and u 0 = (u10 , u20 , u30 ), and are therefore usually
referred to as the relativistic velocity addition formulae. The following relation
between the magnitudes u = (u12 + u22 + u32 )1/2 and u 0 = (u10 2 + u20 2 + u30 2 )1/2 of
the velocities is easily demonstrated:
c2 (c2 u 02 ) (c2 v2 )
c u =
.
(c2 + u10 v)2
2
(10.45)
If u 0 < c and v < c then the right-hand side is positive, implying that u < c.
In other words, the resultant of two subluminal velocities is another subluminal
velocity. It is evident that a particle can never attain the velocity of light relative
to a given inertial frame, no matter how many subluminal velocity increments it
is given. It follows that no inertial frame can ever appear to propagate with a
superluminal velocity with respect to any other inertial frame (since we can track
a given inertial frame using a particle which remains at rest at the origin of that
frame).
According to Eq. (10.45), if u 0 = c then u = c, no matter what value v takes:
i.e., the velocity of light is invariant between different inertial frames. Note that
the Lorentz transform only allows one such invariant velocity [i.e., the velocity
c which appears in Eqs. (10.24)(10.27)]. Einsteins relativity principle tells us
that any disturbance which propagates through a vacuum must appear to propagate at the same velocity in all inertial frames. It is now evident that all such
disturbances must propagate at the velocity c. It follows immediately that all
electromagnetic waves must propagate through the vacuum with this velocity,
irrespective of their wavelength. In other words, it is impossible for there to be
any dispersion of electromagnetic waves propagating through a vacuum. Furthermore, gravity waves must also propagate with the velocity c.
The Lorentz transformation implies that the velocities of propagation of all
physical effects are limited by c in deterministic physics. Consider a general
process by which an event P causes an event Q at a velocity U > c in some frame
S. In other words, information about the event P appears to propagate to the
event Q with a superluminal velocity. Let us choose coordinates such that these
two events occur on the x-axis with (finite) time and distance separations t > 0
306
10.5 Tensors
and x > 0, respectively. The time separation in some other inertial frame S 0 is
given by [see Eq. (10.27)]
t 0 = (t v x/c2 ) = t (1 v U/c2 ).
(10.46)
Thus, for sufficiently large v < c we obtain t 0 < 0: i.e., there exist inertial frames
in which cause and effect appear to be reversed. Of course, this is impossible in
deterministic physics. It follows, therefore, that information can never appear to
propagate with a superluminal velocity in any inertial frame, otherwise causality
would be violated.
10.5 Tensors
It is now convenient to briefly review the mathematics of tensors. Tensors are
of primary importance in connection with coordinate transforms. They serve to
isolate intrinsic geometric and physical properties from those that merely depend
on coordinates.
A tensor of rank r in an n-dimensional space possesses nr components which
are, in general, functions of position in that space. A tensor of rank zero has one
component, A, and is called a scalar. A tensor of rank one has n components,
(A1 , A2 , , An ), and is called a vector. A tensor of rank two has n2 components,
which can be exhibited in matrix format. Unfortunately, there is no convenient
way of exhibiting a higher rank tensor. Consequently, tensors are usually represented by a typical component: e.g., the tensor Aijk (rank 3), or the tensor Aijkl
(rank 4), etc. The suffixes i, j, k, are always understood to range from 1 to n.
For reasons which will become apparent later on, we shall represent tensor
components using both superscripts and subscripts. Thus, a typical tensor might
look like Aij (rank 2), or Bij (rank 2), etc. It is convenient to adopt the Einstein
summation convention. Namely, if any suffix appears twice in a given term, once
as a subscript and once as a superscript, a summation over that suffix (from 1 to
n) is implied.
To distinguish between various different coordinate systems, we shall use primed
and multiply primed suffixes. A first system of coordinates (x1 , x2 , , xn ) can
307
10.5 Tensors
0
then be denoted by xi , a second system (x1 , x2 , , xn ) by xi , etc. Similarly, the

general components of a tensor in various coordinate systems are distinguished
by their suffixes. Thus, the components of some third rank tensor are denoted
0
Aijk in the xi system, by Ai 0 j 0 k 0 in the xi system, etc.
When making a coordinate transformation from one set of coordinates, x i ,
0
to another, xi , it is assumed that the transformation in non-singular. In other
0
words, the equations which express the xi in terms of the xi can be inverted to
0
express the xi in terms of the xi . It is also assumed that the functions specifying
a transformation are differentiable. It is convenient to write
0
xi
i0
=
p
i ,
xi
xi
= pii 0 .
0
i
x
(10.47)
(10.48)
Note that
0
pii 0 pij = ij ,
(10.49)
by the chain rule, where ij (the Kronecker delta ) equals 1 or 0 when i = j or

i 6= j, respectively.
The formal definition of a tensor is as follows:
0
1. An entity having components Aijk in the xi system and Ai 0 j 0 k 0 in the xi

system is said to behave as a covariant tensor under the transformation x i
0
xi if
(10.50)
Ai 0 j 0 k 0 = Aijk pii 0 pjj 0 pkk 0 .
0
2. Similarly, Aijk is said to behave as a contravariant tensor under xi xi if

0 0
Ai j k = Aijk pii pjj pkk .
(10.51)
ij
is said to behave as a mixed tensor (contravariant in i j and
3. Finally, Akl
0
covariant in k l) under xi xi if
0
ij i
j
k
l
Aki 0j
l 0 = Akl pi pj pk 0 pl 0 .
308
(10.52)
10.5 Tensors
When an entity is described as a tensor it is generally understood that it behaves as a tensor under all non-singular differentiable transformations of the
relevant coordinates. An entity which only behaves as a tensor under a certain subgroup of non-singular differentiable coordinate transformations is called
a qualified tensor, because its name is conventionally qualified by an adjective
recalling the subgroup in question. For instance, an entity which only exhibits
tensor behaviour under Lorentz transformations is called a Lorentz tensor, or,
more commonly, a 4-tensor.
When applied to a tensor of rank zero (a scalar), the above definitions imply
that A 0 = A. Thus, a scalar is a function of position only, and is independent of
the coordinate system. A scalar is often termed an invariant.
The main theorem of tensor calculus is as follows:
If two tensors of the same type are equal in one coordinate system, then they
are equal in all coordinate systems.
The simplest example of a contravariant vector (tensor of rank one) is provided
by the differentials of the coordinates, dxi , since
0
xi
0
i
dxi = dxi pii .
dx =
i
x
0
(10.53)
The coordinates themselves do not behave as tensors under all coordinate transformations. However, since they transform like their differentials under linear
homogeneous coordinate transformations, they do behave as tensors under such
transformations.
The simplest example of a covariant vector is provided by the gradient of a
function of position = (x1 , , xn ), since if we write
i =
then we have
,
xi
xi
= i 0 = i i 0 = i pii 0 .
x
x x
i0
309
(10.54)
(10.55)
10.5 Tensors
An important example of a mixed second-rank tensor is provided by the Kronecker delta introduced previously, since
0
ij pii pjj 0 = pij pjj 0 = ij 0 .
(10.56)
Tensors of the same type can be added or subtracted to form new tensors. Thus,
if Aij and Bij are tensors, then Cij = Aij Bij is a tensor of the same type. Note
that the sum of tensors at different points in space is not a tensor if the ps are
position dependent. However, under linear coordinate transformations the ps
are constant, so the sum of tensors at different points behaves as a tensor under
this particular type of coordinate transformation.
If Aij and Bijk are tensors, then Cijklm = Aij Bklm is a tensor of the type indicated by the suffixes. The process illustrated by this example is called outer
multiplication of tensors.
Tensors can also be combined by inner multiplication, which implies at least
one dummy suffix link. Thus, Cjkl = Aij Bikl and Ck = Aij Bijk are tensors of the
type indicated by the suffixes.
Finally, tensors can be formed by contraction from tensors of higher rank. Thus,
if Aijklm is a tensor then Cjkl = Aijikl and Ck = Aijkij are tensors of the type indicated
by the suffixes. The most important type of contraction occurs when no free
suffixes remain: the result is a scalar. Thus, Aii is a scalar provided that Aji is a
tensor.
Although we cannot usefully divide tensors, one by another, an entity like
C in the equation Aj = Cij Bi , where Ai and Bi are tensors, can be formally
regarded as the quotient of Ai and Bi . This gives the name to a particularly
useful rule for recognizing tensors, the quotient rule. This rule states that if a set
of components, when combined by a given type of multiplication with all tensors of
a given type yields a tensor, then the set is itself a tensor. In other words, if the
product Ai = Cij Bj transforms like a tensor for all tensors Bi then it follows that
Cij is a tensor.
ij
310
Let
ij
Then if Akl
(10.52) yields
0
10.5 Tensors
ij
Akl
ij
= Akl,m
.
(10.57)
m
x
is a tensor, differentiation of the general tensor transformation
0
ij
j k
i
l
m
Aik 0j
l 0 ,m 0 = Akl,m pi pj pk 0 pl 0 pm 0 + P1 + P2 + ,
(10.58)
where P1 , P2 , etc., are terms involving derivatives of the ps. Clearly, Aij
kl is not
a tensor under a general coordinate transformation. However, under a linear
0
0
coordinate transformation (ps constant) Aik 0j
l 0 ,m 0 behaves as a tensor of the type
indicated by the suffixes, since the P1 , P2 , etc., all vanish. Similarly, all higher
partial derivatives,
ij
Akl
ij
(10.59)
Akl,mn = m n
x x
etc., also behave as tensors under linear transformations. Each partial differentiation has the effect of adding a new covariant suffix.
So far, the space to which the coordinates xi refer has been without structure.
We can impose a structure on it by defining the distance between all pairs of
neighbouring points by means of a metric,
ds2 = gij dxi dxj ,
(10.60)
where the gij are functions of position. We can assume that gij = gji without loss
of generality. The above metric is analogous to, but more general than, the metric
of Euclidian n-space, ds2 = (dx1 )2 +(dx2 )2 + +(dxn )2 . A space whose structure
is determined by a metric of the type (10.60) is called Riemannian. Since ds2 is
invariant, it follows from a simple extension of the quotient rule that g ij must be
a tensor. It is called the metric tensor.
The elements of the inverse of the matrix gij are denoted by gij . These elements are uniquely defined by the equations
gij gjk = ik .
(10.61)
It is easily seen that the gij constitute the elements of a contravariant tensor. This
tensor is said to be conjugate to gij . The conjugate metric tensor is symmetric
(i.e., gij = gji ) just like the metric tensor itself.
311
10.5 Tensors
The tensors gij and gij allow us to introduce the important operations of raising
and lowering suffixes. These operations consist of forming inner products of a
given tensor with gij or gij . For example, given a contravariant vector Ai , we
define its covariant components Ai by the equation
Ai = gij Aj .
(10.62)
Conversely, given a covariant vector Bi , we can define its contravariant components Bi by the equation
Bi = gij Bj .
(10.63)
More generally, we can raise or lower any or all of the free suffixes of any given
tensor. Thus, if Aij is a tensor we define Ai j by the equation
Ai j = gip Apj .
(10.64)
Note that once the operations of raising and lowering suffixes has been defined,
the order of raised suffixes relative to lowered suffixes becomes significant.
By analogy with Euclidian space, we define the squared magnitude (A)2 of a
vector Ai with respect to the metric gij dxi dxj by the equation
(A)2 = gij Ai Aj = Ai Ai .
(10.65)
A vector Ai termed a null vector if (A)2 = 0. Two vectors Ai and Bi are said to be
orthogonal if their inner product vanishes: i.e., if
gij Ai Bj = Ai Bi = Ai Bi = 0.
(10.66)
Finally, let us consider differentiation with respect to an invariant distance, s.

The vector dxi /ds is a contravariant tensor, since
0
dxi
xi dxi
dxi i 0
=
=
p .
ds
xi ds
ds i
(10.67)
The derivative d(Aij kl )/ds of some tensor with respect to s is not, in general,
a tensor, since
d(Aij kl )
dxm
ij
= A kl,m
,
(10.68)
ds
ds
312
10.6 The physical significance of tensors
and, as we have seen, the first factor on the right-hand side is not generally a tensor. However, under linear transformations it behaves as a tensor, so under linear
transformations the derivative of a tensor with respect to an invariant distance
behaves as a tensor of the same type.
10.6 The physical significance of tensors

In this course, we shall only concern ourselves with coordinate transformations
which transform an inertial frame into another inertial frame. This limits us to
four classes of transformations: displacements of the coordinate axes, rotations
of the coordinate axes, parity reversals (i.e., x, y, z x, y, z), and Lorentz
transformations.
One of the central tenets of physics is that experiments should be reproducible.

In other words, if somebody performs a physical experiment today, and obtains
a certain result, then somebody else performing the same experiment next week
ought to obtain the same result, within the experimental errors. Presumably, in
performing these hypothetical experiments, both experimentalists find it necessary to set up a coordinate frame. Usually, these two frames do not coincide. After
all, the experiments are, in general, performed in different places and at different times. Also, the two experimentalists are likely to orientate their coordinate
axes differently. Nevertheless, we still expect both experiments to yield the same
result. What exactly do we mean by this statement? We do not mean that both experimentalists will obtain the same numbers when they measure something. For
instance, the numbers used to denote the position of a point (i.e., the coordinates
of the point) are, in general, different in different coordinate frames. What we
do expect is that any physically significant interrelation between physical quantities (i.e., position, velocity, etc.) which appears to hold in the coordinate system
of the first experimentalist will also appear to hold in the coordinate system of
the second experimentalist. We usually refer to such interrelationships as laws of
physics. So, what we are really saying is that the laws of physics do not depend
on our choice of coordinate system. In particular, if a law of physics is true in one
coordinate system then it is automatically true in every other coordinate system,
313
10.7 Space-time
subject to the proviso that both coordinate systems are inertial.

Recall that tensors are geometric objects which possess the property that if
a certain interrelationship holds between various tensors in one particular coordinate system, then the same interrelationship holds in any other coordinate
system which is related to the first system by a certain class of transformations.
It follows that the laws of physics are expressible as interrelationships between tensors. In special relativity, the laws of physics are only required to exhibit tensor
behaviour under transformations between different inertial frames: i.e., translations, rotations, and Lorentz transformations. Parity inversion is a special type of
transformation, and will be dealt with later on. In general relativity, the laws of
physics are required to exhibit tensor behaviour under all non-singular coordinate
transformations.
10.7 Space-time
In special relativity, we are only allowed to use inertial frames to assign coordinates to events. There are many different types of inertial frames. However, it
is convenient to adhere to those with standard coordinates. That is, spatial coordinates which are right-handed rectilinear Cartesians based on a standard unit
of length, and time-scales based on a standard unit of time. We shall continue
to assume that we are employing standard coordinates. However, from now on,
we shall make no assumptions about the relative configuration of the two sets
of spatial axes, and the origins of time, when dealing with two inertial frames.
Thus, the most general transformation between two inertial frames consists of a
Lorentz transformation in the standard configuration plus a translation (this includes a translation in time) and a rotation of the coordinate axes. The resulting
transformation is called a general Lorentz transformation, as opposed to a Lorentz
transformation in the standard configuration, which will henceforth be termed a
standard Lorentz transformation.
In Sect. 10.3, we proved quite generally that corresponding differentials in two
inertial frames S and S 0 satisfy the relation
2
dx2 + dy2 + dz2 c2 dt2 = dx 0 + dy 0 + dz 0 c2 dt 0 .

314
(10.69)
10.7 Space-time
Thus, we expect this relation to remain invariant under a general Lorentz transformation. Since such a transformation is linear, it follows that
(x2 x1 )2 + (y2 y1 )2 + (z2 z1 )2 c2 (t2 t1 )2 =
(x20 x10 )2 + (y20 y10 )2 + (z20 z10 )2 c2 (t20 t10 )2 ,
(10.70)
where (x1 , y1 , z1 , t1 ) and (x2 , y2 , z2 , t2 ) are the coordinates of any two events in
S, and the primed symbols denote the corresponding coordinates in S 0 . It is
convenient to write
dx2 dy2 dz2 + c2 dt2 = ds2 ,
(10.71)
and
(x2 x1 )2 (y2 y1 )2 (z2 z1 )2 + c2 (t2 t1 )2 = s2 .
(10.72)
The differential ds, or the finite number s, defined by these equations is called
the interval between the corresponding events. Equations (10.71) and (10.72)
express the fact that the interval between two events is invariant, in the sense that
it has the same value in all inertial frames. In other words, the interval between
two events is invariant under a general Lorentz transformation.
Let us consider entities defined in terms of four variables,
x1 = x, x2 = y, x3 = z, x4 = c t,
(10.73)
and which transform as tensors under a general Lorentz transformation. From now
on, such entities will be referred to as 4-tensors.
Tensor analysis cannot proceed very far without the introduction of a nonsingular tensor gij , the so-called fundamental tensor, which is used to define the
operations of raising and lowering suffixes. The fundamental tensor is usually
introduced using a metric ds2 = gij dxi dxj , where ds2 is a differential invariant.
We have already come across such an invariant, namely
ds2 = dx2 dy2 dz2 + c2 dt2
= (dx1 )2 (dx2 )2 (dx3 )2 + (dx4 )2
= g dx dx ,
315
(10.74)
10.7 Space-time
where , run from 1 to 4. Note that the use of Greek suffixes is conventional
in 4-tensor theory. Roman suffixes are reserved for tensors in three-dimensional
Euclidian space, so-called 3-tensors. The 4-tensor g has the components g11 =
g22 = g33 = 1, g44 = 1, and g = 0 when 6= , in all permissible coordinate
frames. From now on, g , as defined above, is adopted as the fundamental
tensor for 4-tensors. g can be thought of as the metric tensor of the space
whose points are the events (x1 , x2 , x3 , x4 ). This space is usually referred to as
space-time, for obvious reasons. Note that space-time cannot be regarded as a
straightforward generalization of Euclidian 3-space to four dimensions, with time
as the fourth dimension. The distribution of signs in the metric ensures that the
time coordinate x4 is not on the same footing as the three space coordinates.
Thus, space-time has a non-isotropic nature which is quite unlike Euclidian space,
with its positive definite metric. According to the relativity principle, all physical
laws are expressible as interrelationships between 4-tensors in space-time.
A tensor of rank one is called a 4-vector. We shall also have occasion to use
ordinary vectors in three-dimensional Euclidian space. Such vectors are called 3vectors, and are conventionally represented by boldface symbols. We shall use the
Latin suffixes i, j, k, etc., to denote the components of a 3-vector: these suffixes
are understood to range from 1 to 3. Thus, u = ui = dxi /dt denotes a velocity
vector. For 3-vectors, we shall use the notation ui = ui interchangeably: i.e., the
level of the suffix has no physical significance.
When tensor transformations from one frame to another actually have to be
computed, we shall usually find it possible to choose coordinates in the standard configuration, so that the standard Lorentz transform applies. Under such a
transformation, any contravariant 4-vector, T , transforms according to the same
scheme as the difference in coordinates x2 x1 between two points in space-time.
It follows that
0
T 1 = (T 1 T 4 ),
0
T 2 = T 2,
0
T 3 = T 3,
0
T 4 = (T 4 T 1 ),
316
(10.75)
(10.76)
(10.77)
(10.78)
10.7 Space-time
where = v/c. Higher rank 4-tensors transform according to the rules (10.50)
(10.52). The transformation coefficients take the form
p =
p 0 =
+
0
0

0
1
0
0
0
0
0
1
0
0 +
+
0
0
+
0
1
0
0
0 +
0
0
1
0
0 +
(10.79)
(10.80)
Often the first three components of a 4-vector coincide with the components of
a 3-vector. For example, the x1 , x2 , x3 in R = (x1 , x2 , x3 , x4 ) are the components
of r, the position 3-vector of the point at which the event occurs. In such cases,
we adopt the notation exemplified by R = (r, c t). The covariant form of such
a vector is simply R = (r, c t). The squared magnitude of the vector is (R)2 =
R R = r2 + c2 t2 . The inner product g R Q = R Q of R with a similar
vector Q = (q, k) is given by R Q = rq + c t k. The vectors R and Q are
said to be orthogonal if R Q = 0.
Since a general Lorentz transformation is a linear transformation, the partial
derivative of a 4-tensor is also a 4-tensor:
A
= A , .
(10.81)
Clearly, a general 4-tensor acquires an extra covariant index after partial differentiation with respect to the contravariant coordinate x . It is helpful to define a
covariant derivative operator
1
= ,
,
x
c t
(10.82)
A A , .
(10.83)
where
317
10.7 Space-time
There is a corresponding contravariant derivative operator

1
= ,

,
x
c t
(10.84)
A g A , .
(10.85)
1 A0
A = A = A +
.
c t
(10.86)
where
The 4-divergence of a 4-vector, A = (A, A0 ), is the invariant
The four-dimensional Laplacian operator, or dAlembertian, is equivalent to the

invariant contraction
1 2
(10.87)
2 = 2 + 2 2 .
c t
Recall that we still need to prove (from Sect. 10.3) that the invariance of the
differential metric,
2
ds2 = dx 0 + dy 0 + dz 0 c2 dt 0 = dx2 + dy2 + dz2 c2 dt2 ,
(10.88)
between two general inertial frames implies that the coordinate transformation
between such frames is necessarily linear. To put it another way, we need to
demonstrate that a transformation which transforms a metric g dx dx with
0
0
constant coefficients into a metric g 0 0 dx dx with constant coefficients must
be linear. Now
0
0
(10.89)
g = g 0 0 p p .
Differentiating with respect to x , we get
0
g 0 0 p p + g 0 0 p p = 0,
where
p
2 x
0
p =
= = p ,
x
x x
etc. Interchanging the indices and yields
0
(10.90)
g 0 0 p p + g 0 0 p p = 0.
318
(10.91)
(10.92)
10.8 Proper time
Interchanging the indices and gives

0
g 0 0 p p + g 0 0 p p = 0,
(10.93)
where the indices 0 and 0 have been interchanged in the first term. It follows
from Eqs. (10.90), (10.92), and (10.93) that
0
g 0 0 p p = 0.
(10.94)
Multiplication by p 0 yields
0
g 0 0 p p p 0 = g 0 0 p = 0.
0
(10.95)
Finally, multiplication by g gives

0
g 0 0 g p = p = 0.
(10.96)
This proves that the coefficients p are constants, and, hence, that the transformation is linear.
10.8 Proper time

It is often helpful to write the invariant differential interval ds2 in the form
ds2 = c2 d2 .
(10.97)
The quantity d is called the proper time. It follows that

dx2 + dy2 + dz2
d =
+ dt2 .
2
c
2
(10.98)
Consider a series of events on the world-line of some material particle. If the

particle has speed u then
2
2
2
u2
2
2 dx + dy + dz
2
d = dt
+ 1 = dt 1 2 ,
c2 dt2
c
(10.99)
implying that
dt
= (u).
d
319
(10.100)
10.8 Proper time
It is clear that dt = d in the particles rest frame. Thus, d corresponds to the

time difference between two neighbouring events on the particles world-line,
as measured by a clock attached to the particle (hence, the name proper time).
According to Eq. (10.100), the particles clock appears to run slow, by a factor
(u), in an inertial frame in which the particle is moving with velocity u. This is
the celebrated time dilation effect.
Let us consider how a small 4-dimensional volume element in space-time
transforms under a general Lorentz transformation. We have
d4 x 0 = J d4 x,
where
(10.101)
0
(x1 , x2 , x3 , x4 )
J =
(x1 , x2 , x3 , x4 )
(10.102)
is the Jacobian of the transformation: i.e., the determinant of the transformation

0
matrix p . A general Lorentz transformation is made up of a standard Lorentz
transformation plus a displacement and a rotation. Thus, the transformation matrix is the product of that for a standard Lorentz transformation, a translation,
and a rotation. It follows that the Jacobian of a general Lorentz transformation
is the product of that for a standard Lorentz transformation, a translation, and a
rotation. It is well-known that the Jacobian of the latter two transformations is
unity, since they are both volume preserving transformations which do not affect
time. Likewise, it is easily seen [e.g., by taking the determinant of the transformation matrix (10.79)] that the Jacobian of a standard Lorentz transformation is
also unity. It follows that
d4 x 0 = d 4 x
(10.103)
for a general Lorentz transformation. In other words, a general Lorentz transformation preserves the volume of space-time. Since time is dilated by a factor in
a moving frame, the volume of space-time can only be preserved if the volume of
ordinary 3-space is reduced by the same factor. As is well-known, this is achieved
by length contraction along the direction of motion by a factor .
320
10.9 4-velocity and 4-acceleration
10.9 4-velocity and 4-acceleration

We have seen that the quantity dx /ds transforms as a 4-vector under a general
Lorentz transformation [see Eq. (10.67)]. Since ds d it follows that
dx
U =
d
(10.104)
also transforms as a 4-vector. This quantity is known as the 4-velocity. Likewise,

the quantity
d2 x
dU
A =
=
(10.105)
d2
d
is a 4-vector, and is called the 4-acceleration.
For events along the world-line of a particle traveling with 3-velocity u, we
have
dx dt
dx
=
= (u) (u, c),
(10.106)
U =
d
dt d
where use has been made of Eq. (10.100). This gives the relationship between a
particles 3-velocity and its 4-velocity. The relationship between the 3-acceleration
and the 4-acceleration is less straightforward. We have
d
dU
d
d
dU
=
= ( u, c) =
u + a, c
A =
,
d
dt
dt
dt
dt
!
(10.107)
where a = du/dt is the 3-acceleration. In the rest frame of the particle, U =

(0, c) and A = (a, 0). It follows that
U A = 0
(10.108)
(note that U A is an invariant quantity). In other words, the 4-acceleration of a

particle is always orthogonal to its 4-velocity.
10.10 The current density 4-vector

Let us now consider the laws of electromagnetism. We wish to demonstrate that
these laws are compatible with the relativity principle. In order to achieve this, it
321
10.10 The current density 4-vector
is necessary for us to make an assumption about the transformation properties of

electric charge. The assumption we shall make, which is well substantiated experimentally, is that charge, unlike mass, is invariant. That is, the charge carried
by a given particle has the same measure in all inertial frames. In particular, the
charge carried by a particle does not vary with the particles velocity.
Let us suppose, following Lorentz, that all charge is made up of elementary
particles, each carrying the invariant amount e. Suppose that n is the number
density of such charges at some given point and time, moving with velocity u,
as observed in a frame S. Let n0 be the number density of charges in the frame
S0 in which the charges are momentarily at rest. As is well-known, a volume of
measure V in S has measure (u) V in S0 (because of length contraction). Since
observers in both frames must agree on how many particles are contained in the
volume, and, hence, on how much charge it contains, it follows that n = (u) n 0 .
If = e n and 0 = e n0 are the charge densities in S and S0 , respectively, then
= (u) 0 .
(10.109)
The quantity 0 is called the proper density, and is obviously Lorentz invariant.
Suppose that x are the coordinates of the moving charge in S. The current
density 4-vector is constructed as follows:
dx
J = 0
= 0 U .
(10.110)
d
Thus,
J = 0 (u) (u, c) = (j, c),
(10.111)
where j = u is the current density 3-vector. Clearly, charge density and current density transform as the time-like and space-like components of the same
4-vector.
Consider the invariant 4-divergence of J :
.
(10.112)
t
We know that one of the caveats of Maxwells equations is the charge conservation law
+ j = 0.
(10.113)
t
J = j +
322
10.11 The potential 4-vector
It is clear that this expression can be rewritten in the manifestly Lorentz invariant
form
J = 0.
(10.114)
This equation tells us that there are no net sources or sinks of electric charge in
nature: i.e., electric charge is neither created nor destroyed.
10.11 The potential 4-vector

There are many ways of writing the laws of electromagnetism. However, the most
obviously Lorentz invariant way is to write them in terms of the vector and scalar
potentials (see Sect. 4.6). When written in this fashion, Maxwells equations
reduce to
1 2
2 +
=
,
c2 t2
0
1 2
2 +
A = 0 j,
c2 t2
(10.115)
(10.116)
where is the scalar potential, and A the vector potential. Note that the differential operator appearing in these equations is the Lorentz invariant dAlembertian,
defined in Eq. (10.87). Thus, the above pair of equations can be rewritten in the
form
c
2 =
,
(10.117)
c 0
j
2cA =
.
(10.118)
c 0
Maxwells equations can be written in Lorentz invariant form provided that the
entity
= (c A, )
(10.119)
transforms as a contravariant 4-vector. This entity is known as the potential 4vector. It follows from Eqs. (10.111), (10.115), and (10.116) that
J
2 =
.
c 0
323
(10.120)
10.12 Gauge invariance
Thus, the field equations which govern classical electromagnetism can all be
summed up in a single 4-vector equation.
10.12 Gauge invariance

The electric and magnetic fields are obtained from the vector and scalar potentials
according to the prescription (see Sect. 4.3)
E =
A
,
t
B = A.
(10.121)
(10.122)
These fields are important, because they determine the electromagnetic forces
exerted on charged particles. Note that the above prescription does not uniquely
determine the two potentials. It is possible to make the following transformation, known as a gauge transformation, which leaves the fields unaltered (see
Sect. 4.4):
,
t
A A ,
(10.123)
(10.124)
where (r, t) is a general scalar field. It is necessary to adopt some form of convention, generally known as a gauge condition, to fully specify the two potentials.
In fact, there is only one gauge condition which is consistent with Eqs. (10.114).
This is the Lorentz gauge condition,
1
+ A = 0.
c2 t
Note that this condition can be written in the Lorentz invariant form
= 0.
(10.125)
(10.126)
This implies that if the Lorentz gauge holds in one particular inertial frame then
it automatically holds in all other inertial frames. A general gauge transformation
can be written
+ c .
(10.127)
324
Note that even after the Lorentz gauge has been adopted, the potentials are undetermined to a gauge transformation using a scalar field, , which satisfies the
sourceless wave equation
2 = 0.
(10.128)
However, if we adopt sensible boundary conditions in both space and time then
the only solution to the above equation is = 0.

We already know the solutions to Eqs. (10.117) and (10.118). They take the
form (see Sect. 4.9)
Z
[(r 0 )]
1
dV 0 ,
(10.129)
(r, t) =
0
4 0 |r r |
Z
0 [j(r 0 )]
A(r, t) =
dV 0 .
(10.130)
0
4 |r r |
The above equations can be combined to form the solution of the 4-vector wave
equation (10.120),
Z
[J ]
1
=
dV.
(10.131)
4 0 c
r
Here, the components of the 4-potential are evaluated at some event P in spacetime, r is the distance of the volume element dV from P, and the square brackets
indicate that the 4-current is to be evaluated at the retarded time: i.e., at a time
r/c before P.
But, does the right-hand side of Eq. (10.131) really transform as a contravariant 4-vector? This is not a trivial question, since volume integrals in 3-space are
not, in general, Lorentz invariant due to the length contraction effect. However,
the integral in Eq. (10.131) is not a straightforward volume integral, because the
integrand is evaluated at the retarded time. In fact, the integral is best regarded
as an integral over events in space-time. The events which enter the integral
are those which intersect a spherical light wave launched from the event P and
evolved backwards in time. In other words, the events occur before the event P,
325
and have zero interval with respect to P. It is clear that observers in all inertial
frames will, at least, agree on which events are to be included in the integral,
since both the interval between events, and the absolute order in which events
occur, are invariant under a general Lorentz transformation.
We shall now demonstrate that all observers obtain the same value of dV/r
for each elementary contribution to the integral. Suppose that S and S 0 are two
inertial frames in the standard configuration. Let unprimed and primed symbols
denote corresponding quantities in S and S 0 , respectively. Let us assign coordinates (0, 0, 0, 0) to P, and (x, y, z, c t) to the retarded event Q for which r and dV
are evaluated. Using the standard Lorentz transformation, (10.24)(10.27), the
fact that the interval between events P and Q is zero, and the fact that both t and
t 0 are negative, we obtain
vx
r = c t = c t 2 ,
c
0
(10.132)
where v q
is the relative velocity between frames S 0 and S, is the Lorentz factor,
and r = x2 + y2 + z2 , etc. It follows that
v
ct vx
= r 1 + cos ,
r = r +
r
cr
c
!
(10.133)
where is the angle (in 3-space) subtended between the line PQ and the x-axis.
We now know the transformation for r. What about the transformation for
dV? We might be tempted to set dV 0 = dV, according to the usual length
contraction rule. However, this is incorrect. The contraction by a factor only
applies if the whole of the volume is measured at the same time, which is not
the case in the present problem. Now, the dimensions of dV along the y- and
z-axes are the same in both S and S 0 , according to Eqs. (10.24)(10.27). For the
x-dimension these equations give dx 0 = (dx v dt). The extremities of dx are
measured at times differing by dt, where
dr
dx
=
cos .
c
c
(10.134)
v
dx = 1 + cos dx,
c
(10.135)
dt =
Thus,
326
10.14 Tensors and pseudo-tensors
giving
v
dV = 1 + cos dV.
(10.136)
c
It follows from Eqs. (10.133) and (10.136) that dV 0 /r 0 = dV/r. This result will
clearly remain valid even when S and S 0 are not in the standard configuration.
!
Thus, dV/r is an invariant and, therefore, [J ] dV/r is a contravariant 4-vector.

For linear transformations, such as a general Lorentz transformation, the result
of adding 4-tensors evaluated at different 4-points is itself a 4-tensor. It follows
that the right-hand side of Eq. (10.131) is indeed a contravariant 4-vector. Thus,
this 4-vector equation can be properly regarded as the solution to the 4-vector
wave equation (10.120).

The totally antisymmetric fourth rank tensor is defined
for , , , any even permutation of 1, 2, 3, 4
+1
=
1
for , , , any odd permutation of 1, 2, 3, 4
0
otherwise
(10.137)
The components of this tensor are invariant under a general Lorentz transformation, since
0
0 0
0 0
p p p p = |p | = ,
(10.138)
where |p | denotes the determinant of the transformation matrix, or the Jacobian

of the transformation, which we have already established is unity for a general
Lorentz transformation. We can also define a totally antisymmetric third rank
tensor ijk which stands in the same relation to 3-space as does to spacetime. It is easily demonstrated that the elements of ijk are invariant under a
general translation or rotation of the coordinate axes. The totally antisymmetric
third rank tensor is used to define the cross product of two 3-vectors,
(a b)i = ijk aj bk ,
327
(10.139)
and the curl of a 3-vector field,

( A)i = ijk
Ak
.
xj
(10.140)
The following two rules are often useful in deriving vector identities
ijk iab = ja kb jb ka ,
(10.141)
ijk ijb = 2 kb .
(10.142)
Up to now, we have restricted ourselves to three basic types of coordinate

transformation: namely, translations, rotations, and standard Lorentz transformations. An arbitrary combination of these three transformations constitutes
a general Lorentz transformation. Let us now extend our investigations to include a fourth type of transformation known as a parity inversion: i.e., x, y, z,
x, y, z. A reflection is a combination of a parity inversion and a rotation. As
is easily demonstrated, the Jacobian of a parity inversion is 1, unlike a translation, rotation, or standard Lorentz transformation, which all possess Jacobians of
+1.
The prototype of all 3-vectors is the difference in coordinates between two
points in space, r. Likewise, the prototype of all 4-vectors is the difference in
coordinates between two events in space-time, R = (r, c t). It is not difficult to
appreciate that both of these objects are invariant under a parity transformation
(in the sense that they correspond to the same geometric object before and after
the transformation). It follows that any 3- or 4-tensor which is directly related to
r and R , respectively, is also invariant under a parity inversion. Such tensors include the distance between two points in 3-space, the interval between two points
in space-time, 3-velocity, 3-acceleration, 4-velocity, 4-acceleration, and the metric tensor. Tensors which exhibit tensor behaviour under translations, rotations,
special Lorentz transformations, and are invariant under parity inversions, are
termed proper tensors, or sometimes polar tensors. Since electric charge is clearly
invariant under such transformations (i.e., it is a proper scalar), it follows that
3-current and 4-current are proper vectors. It is also clear from Eq. (10.120) that
the scalar potential, the vector potential, and the potential 4-vector, are proper
tensors.
328
It follows from Eq. (10.137) that under a parity inversion.

Tensors such as this, which exhibit tensor behaviour under translations, rotations,
and special Lorentz transformations, but are not invariant under parity inversions
(in the sense that they correspond to different geometric objects before and after
the transformation), are called pseudo-tensors, or sometimes axial tensors. Equations (10.139) and (10.140) imply that the cross product of two proper vectors
is a pseudo-vector, and the curl of a proper vector field is a pseudo-vector field.
One particularly simple way of performing a parity transformation is to exchange positive and negative numbers on the three Cartesian axes. A proper
vector is unaffected by such a procedure (i.e., its magnitude and direction are the
same before and after). On the other hand, a pseudo-vector ends up pointing in
the opposite direction after the axes are renumbered.
What is the fundamental difference between proper tensors and pseudo-tensors?
The answer is that all pseudo-tensors are defined according to a handedness convention. For instance, the cross product between two vectors is conventionally
defined according to a right-hand rule. The only reason for this is that the majority of human beings are right-handed. Presumably, if the opposite were true then
cross products etc. would be defined according to a left-hand rule, and would,
therefore, take minus their conventional values. The totally antisymmetric tensor
is the prototype pseudo-tensor, and is, of course, conventionally defined with respect to a right-handed spatial coordinate system. A parity inversion converts left
into right, and vice versa, and, thereby, effectively swaps left- and right-handed
conventions.
The use of conventions in physics is perfectly acceptable provided that we recognize that they are conventions, and are consistent in our use of them. It follows
that laws of physics cannot contain mixtures of tensors and pseudo-tensors, otherwise they would depend our choice of handedness convention.3
3
Here, we are assuming that the laws of physics do not possess an intrinsic handedness. This is certainly the case
for mechanics and electromagnetism. However, the weak interaction does possess an intrinsic handedness: i.e., it is
fundamentally different in a parity inverted universe. So, the equations governing the weak interaction do actually
contain mixtures of tensors and pseudo-tensors.
329
Let us now consider electric and magnetic fields. We know that

E =
B = A.
A
,
t
(10.143)
(10.144)
We have already seen that the scalar and the vector potential are proper scalars
and vectors, respectively. It follows that E is a proper vector, but that B is a
pseudo-vector (since it is the curl of a proper vector). In order to fully appreciate the difference between electric and magnetic fields, let us consider a thought
experiment first proposed by Richard Feynman. Suppose that we are in radio contact with a race of aliens, and are trying to explain to them our system of physics.
Suppose, further, that the aliens live sufficiently far away from us that there are
no common objects which we both can see. The question is this: could we unambiguously explain to these aliens our concepts of electric and magnetic fields? We
could certainly explain electric and magnetic lines of force. The former are the
paths of charged particles (assuming that the particles are subject only to electric
fields), and the latter can be mapped out using small test magnets. We could also
explain how we put arrows on electric lines of force to convert them into electric
field-lines: the arrows run from positive charges (i.e., charges with the same sign
as atomic nuclei) to negative charges. This explanation is unambiguous provided
that our aliens live in a matter (rather than an anti-matter) dominated part of the
Universe. But, could we explain how we put arrows on magnetic lines of force
in order to convert them into magnetic field-lines? The answer is, no. By definition, magnetic field-lines emerge from the North poles of permanent magnets
and converge on the corresponding South poles. The definition of the North pole
of a magnet is simply that it possesses the same magnetic polarity as the South
(geographic) pole of the Earth. This is obviously a convention. In fact, we could
redefine magnetic field-lines to run from the South poles to the North poles of
magnets without significantly altering our laws of physics (we would just have to
replace B by B in all our equations). In a parity inverted Universe, a North pole
becomes a South pole, and vice versa, so it is hardly surprising that B B.
330
10.15 The electromagnetic field tensor

Let us now investigate whether we can write the components of the electric and
magnetic fields as the components of some proper 4-tensor. There is an obvious
problem here. How can we identify the components of the magnetic field, which
is a pseudo-vector, with any of the components of a proper-4-tensor? The former
components transform differently under parity inversion than the latter components. Consider a proper-3-tensor whose covariant components are written B ik ,
and which is antisymmetric:
Bij = Bji .
(10.145)
This immediately implies that all of the diagonal components of the tensor are
zero. In fact, there are only three independent non-zero components of such a
tensor. Could we, perhaps, use these components to represent the components of
a pseudo-3-vector? Let us write
Bi =
1 ijk
Bjk .
2
(10.146)
It is clear that Bi transforms as a contravariant pseudo-3-vector. It is easily seen

that
0
B
B
z
y
ij
Bx
B = Bij = Bz 0
(10.147)
,
By Bx 0
where B1 = B1 Bx , etc. In this manner, we can actually write the components

of a pseudo-3-vector as the components of an antisymmetric proper-3-tensor. In
particular, we can write the components of the magnetic field B in terms of an
antisymmetric proper magnetic field 3-tensor which we shall denote Bij .
Let us now examine Eqs. (10.143) and (10.144) more carefully. Recall that
= (c A, ) and = (, c1 /t). It follows that we can write Eq. (10.143)
in the form
Ei = i 4 + 4 i .
(10.148)
Likewise, Eq. (10.144) can be written
c Bi =
1 ijk
c Bjk = ijk j k .
2
331
(10.149)
Let us multiply this expression by iab , making use of the identity
We obtain
iab ijk = ja kb jb ka .
(10.150)
c
(Bab Bba ) = a b + b a ,
2
(10.151)
c Bij = i j + j i ,
(10.152)
or
since Bij = Bji .
Let us define a proper-4-tensor whose covariant components are given by
F = .
(10.153)
It is clear that this tensor is antisymmetric:

F = F .
(10.154)
This implies that the tensor only possesses six independent non-zero components.
Maybe it can be used to specify the components of E and B?
Equations (10.148) and (10.153) yield
F4i = 4 i i 4 = Ei .
(10.155)
Likewise, Eqs. (10.152) and (10.153) imply that

Fij = i j j i = c Bij .
(10.156)
Thus,
Fi4 = F4i = Ei ,
(10.157)
Fij = Fji = c Bij .
(10.158)
In other words, the completely space-like components of the tensor specify the
components of the magnetic field, whereas the hybrid space and time-like components specify the components of the electric field. The covariant components
332
of the tensor can be written
F =
0
c Bz +c By Ex
+c Bz
0
c Bx Ey
c By +c Bx
0
Ez
+Ex +Ey +Ez
0
(10.159)
Not surprisingly, F is usually called the electromagnetic field tensor. The above
expression, which appears in all standard textbooks, is very misleading. Taken
at face value, it is simply wrong! We cannot form a proper-4-tensor from the
components of a proper-3-vector and a pseudo-3-vector. The expression only
makes sense if we interpret Bx (say) as representing the component B23 of the
proper magnetic field 3-tensor Bij
The contravariant components of the electromagnetic field tensor are given by
or
F =
Fi4 = F4i = +Ei ,
(10.160)
Fij = Fji = c Bij ,
(10.161)
0
c Bz +c By +Ex
+c Bz
0
c Bx +Ey
c By +c Bx
0
+Ez
Ex Ey Ez
0
Let us now consider two of Maxwells equations:

E =
,
0
B = 0
(10.162)
(10.163)
E
j + 0
.
t
!
(10.164)
Recall that the 4-current is defined J = (j, c). The first of these equations can
be written
J4
i
i4
44
i E = i F + 4 F =
.
(10.165)
c 0
333
10.16 The dual electromagnetic field tensor
since F44 = 0. The second of these equations takes the form

ijk
j (c Bk ) 4 E =
ijk
Ji
.
j (1/2 kab c B ) + 4 F =
c 0
ab
4i
(10.166)
Making use of Eq. (10.150), the above expression reduces to

Ji
1
.
j (c Bij c Bji ) + 4 F4i = j Fji + 4 F4i =
2
c 0
(10.167)

F
J
.
=
c 0
(10.168)
This equation is consistent with the equation of charge continuity, J = 0,

because of the antisymmetry of the electromagnetic field tensor.

We have seen that it is possible to write the components of the electric and magnetic fields as the components of a proper-4-tensor. Is it also possible to write the
components of these fields as the components of some pseudo-4-tensor? It is obvious that we cannot identify the components of the proper-3-vector E with any of
the components of a pseudo-tensor. However, we can represent the components
of E in terms of those of an antisymmetric pseudo-3-tensor Eij by writing
Ei =
1 ijk
Ejk .
2
(10.169)
It is easily demonstrated that

ij
E = Eij =
0
Ez Ey
Ez 0
Ex
Ey Ex 0
in a right-handed coordinate system.

334
(10.170)
Consider the dual electromagnetic field tensor, G , which is defined

G =
1

F .
2
(10.171)
This tensor is clearly an antisymmetric pseudo-4-tensor. We have

G4i =
1 4ijk
1
1
Fjk = ijk4 Fjk = ijk c Bjk = c Bi ,
2
2
2
(10.172)
plus
1 ijk4
( Fk4 + ij4k F4k ) = ijk Fk4 ,
(10.173)
2
where use has been made of F = F . The above expression yields
Gij =
1
Gij = ijk Ek = ijk kab Eab = Eij .
2
(10.174)
Gi4 = G4i = c Bi ,
(10.175)
Gij = Gji = Eij ,
(10.176)
It follows that
or
G =
0
Ez +Ey c Bx
+Ez
0
Ex c By
Ey +Ex
0
c Bz
+c Bx +c By +c Bz
0
(10.177)
The above expression is, again, slightly misleading, since Ex stands for the component E23 of the pseudo-3-tensor Eij , and not for an element of the proper-3vector E. Of course, in this case, Bx really does represent the first element of the
pseudo-3-vector B. Note that the elements of G are obtained from those of F
by making the transformation c Bij Eij and Ei c Bi .
The covariant elements of the dual electromagnetic field tensor are given by
Gi4 = G4i = +cBi ,
(10.178)
Gij = Gji = Eij ,
(10.179)
335
or
G =
0
Ez +Ey +c Bx
+Ez
0
Ex +c By
Ey +Ex
0
+c Bz
c Bx c By c Bz
0
(10.180)
The elements of G are obtained from those of F by making the transformation

c Bij Eij and Ei c Bi .
Let us now consider the two Maxwell equations
B = 0,
E =
(10.181)
B
.
t
(10.182)
The first of these equations can be written

i (c Bi ) = i Gi4 + 4 G44 = 0,
(10.183)
since G44 = 0. The second equation takes the form

ijk j Ek = ijk j (1/2 kab Eab ) = j Eij = 4 (c Bi ),
(10.184)
or
j Gji + 4 G4i = 0.
(10.185)

G = 0.
(10.186)
Thus, we conclude that Maxwells equations for the electromagnetic fields are
equivalent to the following pair of 4-tensor equations:
F
J
,
=
c 0
= 0.
(10.187)
(10.188)
It is obvious from the form of these equations that the laws of electromagnetism
are invariant under translations, rotations, special Lorentz transformations, parity inversions, or any combination of these transformations.
336
10.17 Transformation of fields
10.17 Transformation of fields

The electromagnetic field tensor transforms according to the standard rule
0
F = F p p .
(10.189)
This easily yields the celebrated rules for transforming electromagnetic fields:
Ek0 = Ek ,
(10.190)
Bk0 = Bk ,
(10.191)
E0 = (E + v B),
B0 = (B v E/c2 ),
(10.192)
(10.193)
where v is the relative velocity between the primed and unprimed frames, and the
perpendicular and parallel directions are, respectively, perpendicular and parallel
to v.
At this stage, we may conveniently note two important invariants of the electromagnetic field. They are
1
F F = c2 B2 E2 ,
2
(10.194)
and
1
G F = c EB.
(10.195)
4
The first of these quantities is a proper-scalar, and the second a pseudo-scalar.
10.18 Potential due to a moving charge

Suppose that a particle carrying a charge e moves with uniform velocity u through
a frame S. Let us evaluate the vector potential, A, and the scalar potential, , due
to this charge at a given event P in S.
Let us choose coordinates in S so that P = (0, 0, 0, 0) and u = (u, 0, 0). Let S 0
be that frame in the standard configuration with respect to S in which the charge
337
10.18 Potential due to a moving charge
is (permanently) at rest at (say) the point (x 0 , y 0 , z 0 ). In S 0 , the potential at P is

the usual potential due to a stationary charge,
A 0 = 0,
0 =
e
,
4 0 r 0
(10.196)
(10.197)
where r 0 = x 0 2 + y 0 2 + z 0 2 . Let us now transform these equations directly into

the frame S. Since A = (c A, ) is a contravariant 4-vector, its components
transform according to the standard rules (10.75)(10.78). Thus,
!
u 0
ue
0
c A1 = c A1 + =
,
(10.198)
c
4 0 c r 0
c A2 = c A20 = 0,
(10.199)
(10.200)
c A3 = c A30 = 0,
!
e
u
= 0 + c A10 =
,
(10.201)
c
4 0 r 0
since = u/c in this case. It remains to express the quantity r 0 in terms of
quantities measured in S. The most physically meaningful way of doing this is
to express r 0 in terms of retarded values in S. Consider the retarded event at the
charge for which, by definition, r 0 = c t 0 and r = c t. Using the standard
Lorentz transformation, (10.24)(10.27), we find that
r 0 = c t 0 = c (t u x/c2 ) = r (1 + ur /c),
(10.202)
where ur = u x/r = ru/r denotes the radial velocity of the change in S. We can
now rewrite Eqs. (10.198)(10.201) in the form
0 e
[u]
,
4 [r + ru/c]
1
e
,
=
4 0 [r + ru/c]
A =
(10.203)
(10.204)
where the square brackets, as usual, indicate that the enclosed quantities must
be retarded. For a uniformly moving charge, the retardation of u is, of course,
superfluous. However, since
Z
1
[J ]
=
dV,
(10.205)
4 0 c
r
338
10.19 Fields due to a moving charge
it is clear that the potentials depend only on the (retarded) velocity of the charge,
and not on its acceleration. Consequently, the expressions (10.203) and (10.204)
give the correct potentials for an arbitrarily moving charge. They are known as
the Lienard-Wiechert potentials.

Although the fields generated by a uniformly moving charge can be calculated
from the expressions (10.203) and (10.204) for the potentials, it is simpler to
calculate them from first principles.
Let a charge e, whose position vector at time t = 0 is r, move with uniform
velocity u in a frame S whose x-axis has been chosen in the direction of u. We
require to find the field strengths E and B at the event P = (0, 0, 0, 0). Let S 0 be
that frame in standard configuration with S in which the charge is permanently
at rest. In S 0 , the field is given by
B 0 = 0,
E0 =
e r0
.
4 0 r 0 3
(10.206)
(10.207)
This field must now be transformed into the frame S. The direct method, using
Eqs. (10.190)(10.193), is somewhat simpler here, but we shall use a somewhat
indirect method because of its intrinsic interest.
In order to express Eq. (10.206) and (10.207) in tensor form, we need the
electromagnetic field tensor F on the left-hand side, and the position 4-vector
R = (r, c t) and the scalar e/(4 0 r 0 3 ) on the right-hand side. (We regard r 0
as an invariant for all observers.) To get a vanishing magnetic field in S 0 , we
multiply on the right by the 4-velocity U = (u) (u, c), thus tentatively arriving
at the equation
e

F =
U
R .
(10.208)
3
4 0 c r 0
Recall that F4i = Ei and Fij = c Bij . However, this equation cannot be correct,
because the antisymmetric tensor F can only be equated to another antisym339
metric tensor. Consequently, let us try

F =
e

3 (U R U R ).
0
4 0 c r
(10.209)
This is found to give the correct field at P in S 0 , as long as R refers to any

event whatsoever at the charge. It only remains to interpret (10.209) in S. It
is convenient to choose for R that event at the charge at which t = 0 (not the
retarded event). Thus,
Fjk = c Bjk =
e
j k
k j
3 (u) (u r u r ),
0
4 0 c r
(10.210)
giving
Bi =
or
1
0 e
j k
ijk Bjk =
3 (u) ijk u r ,
0
2
4 r
B=
0 e
u r.
4 r 0 3
Likewise,
F4i = Ei =
or
E=
(10.211)
(10.212)
e
i
3 r ,
0
4 0 r
(10.213)
e
r.
4 0 r 0 3
(10.214)
Lastly, we must find an expression for r 0 3 in terms of quantities measured in S at

time t = 0. If t 0 is the corresponding time in S 0 at the charge, we have
2 ur2
2 u2 x2
02
2
2 02
2
2
r =r +c t =r +
=r 1+
.
c2
c2
(10.215)
Thus,
e
r,
4 0 r3 (1 + ur2 2 /c2 )3/2
0 e
1
B =
u
r
=
u E.
4 r3 (1 + ur2 2 /c2 )3/2
c2
E =
(10.216)
(10.217)
Note that E acts in line with the point which the charge occupies at the instant of
measurement, despite the fact that, owing to the finite speed of propagation of all
340
10.20 Relativistic particle dynamics
physical effects, the behaviour of the charge during a finite period before that instant can no longer affect the measurement. Note also that, unlike Eqs. (10.203)
and (10.204), the above expressions for the fields are not valid for an arbitrarily
moving charge, not can they be made valid by merely using retarded values. For
whereas acceleration does not affect the potentials, it does affect the fields, which
involve the derivatives of the potential.
For low velocities, u/c 0, Eqs. (10.216) and (10.217) reduce to the wellknown Coulomb and Biot-Savart fields. However, at high velocities, (u) 1,
the fields exhibit some interesting behaviour. The peak electric field, which occurs
at the point of closest approach of the charge to the observation point, becomes
equal to times its non-relativistic value. However, the duration of appreciable
field strength at the point P is decreased. A measure of the time interval over
which the field is appreciable is
t
b
,
c
(10.218)
where b is the distance of closest approach (assuming 1). As increases, the

peak field increases in proportion, but its duration goes in the inverse proportion.
The time integral of the field is independent of . As , the observer at P
sees electric and magnetic fields which are indistinguishable from the fields of a
pulse of plane polarized radiation propagating in the x-direction. The direction
of polarization is along the radius vector pointing towards the particles actual
position at the time of observation.

Consider a particle which, in its instantaneous rest frame S0 , has mass m0 and
constant acceleration in the x-direction a0 . Let us transform to a frame S, in the
standard configuration with respect to S0 , in which the particles instantaneous
velocity is u. What is the value of a, the particles instantaneous x-acceleration,
in S?
The easiest way in which to answer this question is to consider the acceleration
341
4-vector [see Eq. (10.107)]

d
d
u + a, c
.
A =
dt
dt
!
(10.219)
Using the standard transformation, (10.75)(10.78), for 4-vectors, we obtain

a0 = 3 a,
d
u a0
=
.
dt
c2
(10.220)
(10.221)

f = m 0 3
du
,
dt
(10.222)
where f = m0 a0 is the constant force (in the x-direction) acting on the particle
in S0 .
Equation (10.222) is equivalent to
d(m u)
,
dt
(10.223)
m = m0 .
(10.224)
f=
where
Thus, we can account for the ever decreasing acceleration of a particle subject
to a constant force [see Eq. (10.220)] by supposing that the inertial mass of the
particle increases with its velocity according to the rule (10.224). Henceforth,
m0 is termed the rest mass, and m the inertial mass.
The rate of increase of the particles energy E satisfies
du
dE
= f u = m 0 3 u
.
dt
dt
(10.225)
dE d(m c2 )
=
,
dt
dt
(10.226)
This equation can be written
342
10.21 The force on a moving charge
which can be integrated to yield Einsteins famous formula

E = m c2 .
(10.227)
The 3-momentum of a particle is defined

(10.228)
p = m u,
where u is its 3-velocity. Thus, by analogy with Eq. (10.223), Newtons law of
motion can be written
dp
f=
,
(10.229)
dt
where f is the 3-force acting on the particle.
The 4-momentum of a particle is defined
P = m0 U = m0 (u, c) = (p, E/c),
(10.230)
where U is its 4-velocity. The 4-force acting on the particle obeys

dP
= m 0 A ,
F =
d
(10.231)
where A is its 4-acceleration. It is easily demonstrated that
(10.232)
dE
= fu.
dt
(10.233)
since
dm
fu
F = f, c
,
= f,
dt
c

The electromagnetic 3-force acting on a charge e moving with 3-velocity u is
given by the well-known formula
f = e (E + u B).
343
(10.234)
When written in component form this expression becomes

fi = e (Ei + ijk uj Bk ),
(10.235)
fi = e (Ei + Bij uj ),
(10.236)
or
where use has been made of Eq. (10.147).
Recall that the components of the E and B fields can be written in terms of an
antisymmetric electromagnetic field tensor F via
Fi4 = F4i = Ei ,
(10.237)
Fij = Fji = c Bij .
(10.238)

fi =
e
(Fi4 U4 + Fij Uj ),
c
(10.239)
where U = (u, c) is the particles 4-velocity. It is easily demonstrated that

fu e
e
e
= Eu = Ei ui =
(F4i Ui + F44 U4 ).
c
c
c
c
(10.240)
Thus, the 4-force acting on the particle,
can be written in the form
fu
F = f,
,
c
F =
e
F U .
c
(10.241)
(10.242)
The skew symmetry of the electromagnetic field tensor ensures that

e
F U = F U U = 0.
(10.243)
c
This is an important result, since it ensures that electromagnetic fields do not
change the rest mass of charged particles. In order to appreciate this, let us
assume that the rest mass m0 is not a constant. Since
F =
d(m0 U )
dm0
= m 0 A +
U ,
d
d
344
(10.244)
we can use the standard results U U = c2 and A U = 0 to give

F U = c 2
dm0
.
d
(10.245)
Thus, if rest mass is to remain an invariant, it is imperative that all laws of

physics predict 4-forces acting on particles which are orthogonal to the particles
4-velocities. The laws of electromagnetism pass this test.

Consider a continuous volume distribution of charged matter in the presence of
an electromagnetic field. Let there be n0 particles per unit proper volume (unit
volume determined in the local rest frame), each carrying a charge e. Consider
an inertial frame in which the 3-velocity field of the particles is u. The number
density of the particles in this frame is n = (u) n0 . The charge density and
the 3-current due to the particles are = e n and j = e n u, respectively. Multiplying Eq. (10.242) by the proper number density of particles, n0 , we obtain an
expression
f = c1 F J
(10.246)
for the 4-force f acting on unit proper volume of the distribution due to the ambient electromagnetic fields. Here, we have made use of the definition J = e n0 U .
It is easily demonstrated, using some of the results obtained in the previous section, that
!
Ej
f = E + j B,
.
(10.247)
c
The above expression remains valid when there are many charge species (e.g.,
electrons and ions) possessing different number density and 3-velocity fields. The
4-vector f is usually called the Lorentz force density.
We know that Maxwells equations reduce to
F
J
=
,
c 0
= 0,
345
(10.248)
(10.249)
where F is the electromagnetic field tensor, and G is its dual. As is easily

verified, Eq. (10.249) can also be written in the form
(10.250)
F + F + F = 0.
f = 0 F F .
(10.251)
This expression can also be written

f = 0 [ (F F ) F F ] .
(10.252)
Now,
1
F ( F + F ),
(10.253)
2
where use has been made of the antisymmetry of the electromagnetic field tensor.
1
1
F F = F F = (F F ).
(10.254)
2
4
Thus,
"
#
1
(10.255)
f = 0 (F F ) (F F ) .
4
F F =
The above expression can also be written

f = T ,
where
(10.256)
1
(10.257)
T = 0 F F + (F F )
4
is called the electromagnetic energy tensor. Note that T is a proper-4-tensor. It
follows from Eqs. (10.159), (10.162), and (10.194) that
"
T ij
T
T44
Bi Bj
Bk Bk
i
i 1
k
= 0 E Ej +
j 0 E E k +
,
0
2
0
= T
ijk Ej Bk
=
,
0 c
346
(10.258)
(10.259)
Bk Bk
1
0 E k E k +
.
=
2
0
(10.260)
Equation (10.256) can also be written

f = T ,
(10.261)
where T is a symmetric tensor whose elements are

T
T
ij
i4
T 44
Bi Bj
B2
ij 1
2
+
= 0 E E
,
0 E +
0
2
0
i j
= T
4i
(E B)i
,
=
0 c
B2
1
0 E 2 + .
=
2
0
(10.262)
(10.263)
(10.264)
Consider the time-like component of Eq. (10.261). It follows from Eq. (10.247)
that
Ej
= i T i4 4 T 44 .
(10.265)
c
This equation can be rearranged to give
U
+ u = Ej,
t
(10.266)
where U = T 44 and ui = cT i4 , so that

0 E 2
B2
U=
+
,
2
2 0
and
(10.267)
EB
.
(10.268)
0
The right-hand side of Eq. (10.266) represents the rate per unit volume at which
energy is transferred from the electromagnetic field to charged particles. It is
clear, therefore, that Eq. (10.266) is an energy conservation equation for the electromagnetic field (see Sect. 8.2). The proper-3-scalar U can be identified as the
energy density of the electromagnetic field, whereas the proper-3-vector u is the
energy flux due to the electromagnetic field: i.e., the Poynting flux.
u=
347
10.23 Accelerated charges
Consider the space-like components of Eq. (10.261). It is easily demonstrated

that these reduce to
g
+ G = E j B,
(10.269)
t
where Gij = T ij and gi = T 4i /c, or
B2
Bi Bj
ij 1
ij
i j
2
+
,
G = 0 E E
0 E +
0
2
0
(10.270)
and
u
= 0 E B.
(10.271)
c2
Equation (10.269) is basically a momentum conservation equation for the electromagnetic field (see Sect. 8.4). The right-hand side represents the rate per
unit volume at which momentum is transferred from the electromagnetic field to
charged particles. The symmetric proper-3-tensor Gij specifies the flux of electromagnetic momentum parallel to the ith axis crossing a surface normal to the
jth axis. The proper-3-vector g represents the momentum density of the electromagnetic field. It is clear that the energy conservation law (10.266) and the
momentum conservation law (10.269) can be combined together to give the relativistically invariant energy-momentum conservation law (10.261).
g=

Let us calculate the electric and magnetic fields observed at position xi and time
0
t due to a charge e whose retarded position and time are xi and t 0 , respectively.
0
From now on (xi , t) is termed the field point and (xi , t 0 ) is termed the source
point. It is assumed that we are given the retarded position of the charge as a
0
function of its retarded time: i.e., xi (t 0 ). The retarded velocity and acceleration
of the charge are
0
dxi
i
u =
,
(10.272)
dt 0
and
dui
u
i =
,
(10.273)
dt 0
348
respectively. The radius vector r is defined to extend from the retarded position
0
of the charge to the field point, so that ri = xi xi . (Note that this is the opposite
convention to that adopted in Sects. 10.18 and 10.19). It follows that
dr
= u.
(10.274)
dt 0
The field and the source point variables are connected by the retardation condition

1/2
i i0
i
i0
0
r(x , x ) = (x x ) (xi xi )
= c (t t 0 ).
(10.275)
The potentials generated by the charge are given by the Lienard-Wiechert formulae
0 e u
A(xi , t) =
,
(10.276)
4 s
e 1
,
(10.277)
(xi , t) =
40 s
where s = r r u/c is a function both of the field point and the source point
variables. Recall that the Lienard-Wiechert potentials are valid for accelerating,
as well as uniformly moving, charges.
The fields E and B are derived from the potentials in the usual manner:
E =
B = A.
A
,
t
(10.278)
(10.279)
However, the components of the gradient operator are partial derivatives at

constant time, t, and not at constant time, t 0 . Partial differentiation with respect to the xi compares the potentials at neighbouring points at the same time,
but these potential signals originate from the charge at different retarded times.
Similarly, the partial derivative with respect to t implies constant x i , and, hence,
refers to the comparison of the potentials at a given field point over an interval
of time during which the retarded coordinates of the source have changed. Since
we only know the time variation of the particles retarded position with respect
to t 0 we must transform /t|xi and /xi |t to expressions involving /t 0 |xi and
/xi |t 0 .
349
Now, since xi is assumed to be given as a function of t 0 , we have

0
r(xi , xi (t 0 ) ) r(xi , t 0 ) = c (t t 0 ),
(10.280)
which is a functional relationship between xi , t, and t 0 . Note that

r
t 0
It follows that
=
xi
ru
.
r
(10.281)
t 0
r t 0
r
ru t 0
= 0
= c 1
=
,
t
t
t t
r t
(10.282)
t 0
1
r
=
= ,
t
1 ru/r c s
(10.283)
r
=
.
t s t 0
(10.284)
r ru
r
0
t
=
t 0 ,
0
t
r
r
(10.285)
where all differentiation is at constant xi . Thus,
giving
Similarly,
r = c t 0 = 0 r +
where 0 denotes differentiation with respect to xi at constant t 0 . It follows that

r
,
sc
(10.286)
r
.
s c t 0
(10.287)
t 0 =
so that
= 0
Equation (10.278) yields
or
4 0
s
u
E= 2
,
e
s
t s c2
(10.288)
4 0
0s
r s
r
r u s
E= 2 3
.
u
+
e
s
s c t 0 s2 c2
s3 c2 t 0
(10.289)
350
However,
r u
,
r c
0s =
and
(10.290)
u u
u2
r
r u
r u r u
s
=
+
=
+ .
t 0
t 0
c
c
r
c
c
(10.291)
Thus,
4 0
1
ru
1
ru
E= 2 r
+ 3 r
e
s r
c
s c
c
!
u2 r u
r
, (10.292)
+
2 2u
r
c
c
s c
!
ru
which reduces to
4 0
1
ru
E= 3 r
e
s
c
!
1
u2
1
2 + 3 2 r
c
s c
"
ru
u
r
c
!
#!
(10.293)
Similarly,
u
4
u
0s u
r
u s
B= =
2 0 ,
2
0 e
s
s
sc
s s t
(10.294)
u2
4
ru
r
u ru r u
u
B= 2
+
+
,
0 e
s r
sc
s s2 r
c
c
(10.295)
or
which reduces to
ur
4
u2
1 r
B=
r
1
+
0 e
s3
c2
s3 c r
"
ru
r
u
c
!
#!
(10.296)
A comparison of Eqs. (10.293) and (10.296) yields

B=
rE
.
rc
(10.297)
Thus, the magnetic field is always perpendicular to E and the retarded radius
vector r. Note that all terms appearing in the above formulae are retarded.
The electric field is composed of two separate parts. The first term in Eq. (10.293)
varies as 1/r2 for large distances from the charge. We can think of ru = r r u/c
351
as the virtual present radius vector: i.e., the radius vector directed from the position the charge would occupy at time t if it had continued with uniform velocity
from its retarded position to the field point. In terms of ru , the 1/r2 field is simply
Einduction
e 1 u2 /c2
=
ru .
4 0
s3
(10.298)
We can rewrite the expression (10.216) for the electric field generated by a uniformly moving charge in the form
1 u2 /c2
e
r0 ,
E=
4 0 r03 (1 u2 /c2 + ur2 /c2 )3/2
(10.299)
where r0 is the radius vector directed from the present position of the charge at
time t to the field point, and ur = ur0 /r0 . For the case of uniform motion, the
relationship between the retarded radius vector r and the actual radius vector r 0
is simply
r
r0 = r u.
(10.300)
c
It is straightforward to demonstrate that
q
s = r0 1 u2 /c2 + ur2 /c2
(10.301)
in this case. Thus, the electric field generated by a uniformly moving charge can
be written
e 1 u2 /c2
r0 .
(10.302)
E=
4 0
s3
Since ru = r0 for the case of a uniformly moving charge, it is clear that Eq. (10.298)
is equivalent to the electric field generated by a uniformly moving charge located
at the position the charge would occupy if it had continued with uniform velocity
from its retarded position.
The second term in Eq. (10.293),
Eradiation =
)
r (ru u
e
,
4 0 c2
s3
(10.303)
is of order 1/r, and, therefore, represents a radiation field. Similar considerations

hold for the two terms of Eq. (10.296).
352
10.24 The Larmor formula

Let us transform to the inertial frame in which the charge is instantaneously at
rest at the origin at time t = 0. In this frame, u c, so that ru ' r and s ' r for
events which are sufficiently close to the origin at t = 0 that the retarded charge
still appears to travel with a velocity which is small compared to that of light. It
follows from the previous section that
)
r (r u
e
,
4 0 c2
r3
r
e
u
'
.
4 0 c3 r2
Erad '
(10.304)
Brad
(10.305)
Let us define spherical polar coordinates whose axis points along the direction of
instantaneous acceleration of the charge. It is easily demonstrated that
e
sin
u
,
4 0 c2 r
e
sin
'
u
.
4 0 c3 r
E '
(10.306)
(10.307)
These fields are identical to those of a radiating dipole whose axis is aligned along
the direction of instantaneous acceleration (see Sect. 9.2). The Poynting flux is
given by
e2
sin2 2
E B
=
u
.
(10.308)
0
162 0 c3 r2
We can integrate this expression to obtain the instantaneous power radiated by
the charge
e2
P=
u
2.
(10.309)
3
6 0 c
This is known as Larmors formula. Note that zero net momentum is carried off
by the fields (10.306) and (10.307).
In order to proceed further, it is necessary to prove two useful theorems. The
first theorem states that if a 4-vector field T satisfies
T = 0,
353
(10.310)
and if the components of T are non-zero only in a finite spatial region, then the
integral over 3-space,
Z
I=
T 4 d3 x,
(10.311)
is an invariant. In order to prove this theorem, we need to use the 4-dimensional

analog of Gausss theorem, which states that
Z
I
4
T d x = T dS ,
(10.312)
V
where dS is an element of the 3-dimensional surface S bounding the 4-dimensional

volume V. The particular volume over which the integration is performed is indicated in Fig. 58. The surfaces A and C are chosen so that the spatial components
of T vanish on A and C. This is always possible because it is assumed that the
region over which the components of T are non-zero is of finite extent. The surface B is chosen normal to the x4 -axis, whereas the surface D is chosen normal
0
0
to the x4 -axis. Here, the x and the x are coordinates in two arbitrarily chosen
inertial frames. It follows from Eq. (10.312) that
Z
Z
0
T 4 dS4 + T 4 dS4 0 = 0.
(10.313)
Here, we have made use of the fact that T dS is a scalar and, therefore, has the
3
3 0
same value
R 4in 3all inertial frames. Since dS4 = d x and dS4 0 = d x it follows
that I = T d x is an invariant under a Lorentz transformation. Incidentally, the
above argument also demonstrates that I is constant in time (just take the limit
in which the two inertial frames are identical).
The second theorem is an extension of the first. Suppose that a 4-tensor field
Q satisfies
Q = 0,
(10.314)
and has components which are only non-zero in a finite spatial region. Let A be
a 4-vector whose coefficients do not vary with position in space-time. It follows
that T = A Q satisfies Eq. (10.310). Therefore,
Z
I = A Q4 d3 x
(10.315)
354

x4

dS4
B
C
D
dS4
1/
x1
Figure 58:
is an invariant. However, we can write

I = A B ,
where
B =
Q4 d3 x.
(10.316)
(10.317)
It follows from the quotient rule that if A B is an invariant for arbitrary A then
B must transform as a constant (in time) 4-vector.
These two theorems enable us to convert differential conservation laws into
integral conservation laws. For instance, in differential form, the conservation of
electrical charge is written
J = 0.
(10.318)
However, from Eq. (10.313) this immediately implies that
Z
Z
1 4 3
J d x = d3 x
Q=
c
(10.319)
is an invariant. In other words, the total electrical charge contained in space is

both constant in time, and the same in all inertial frames.
355
Suppose that S is the instantaneous rest frame of the charge. Let us consider
the electromagnetic energy tensor T associated with all of the radiation emitted
by the charge between times t = 0 and t = dt. According to Eq. (10.261), this
tensor field satisfies
T = 0,
(10.320)
apart from a region of space of measure zero in the vicinity of the charge. Furthermore, the region of space over which T is non-zero is clearly finite, since we
are only considering the fields emitted by the charge in a small time interval, and
these fields propagate at a finite velocity. Thus, according to the second theorem
Z
1
P =
T 4 d3 x
(10.321)
c
is a 4-vector. It follows from Sect. 10.22 that we can write P = (dp, dE/c), where
dp and dE are the total momentum and energy carried off by the radiation emitted between times t = 0 and t = dt, respectively. As we have already mentioned,
dp = 0 in the instantaneous rest frame S. Transforming to an arbitrary inertial
frame S 0 , in which the instantaneous velocity of the charge is u, we obtain
0
dE = (u) dE + u dp1 = dE.

(10.322)
However, the time interval over which the radiation is emitted in S 0 is dt 0 = dt.
Thus, the instantaneous power radiated by the charge,
P0 =
dE
dE 0
=
= P,
dt 0
dt
(10.323)
is the same in all inertial frames.

We can make use of the fact that the power radiated by an accelerating charge
is Lorentz invariant to find a relativistic generalization of the Larmor formula,
(10.309), which is valid in all inertial frames. We expect the power emitted by
the charge to depend only on its 4-velocity and 4-acceleration. It follows that the
Larmor formula can be written in Lorentz invariant form as
e2
A A ,
P=
3
6 0 c
356
(10.324)
10.25 Radiation losses
since the 4-acceleration takes the form A = (

u, 0) in the instantaneous rest
frame. In a general inertial frame,
2
2
d
2 2 d
c
A A =
,
(10.325)
u + u
dt
dt
where use has been made of Eq. (10.107). Furthermore, it is easily demonstrated
that
u u
d
= 3 2 .
(10.326)
dt
c
It follows, after a little algebra, that the relativistic generalization of Larmors
formula takes the form
e2
(u
u
)
.
2
6 u
(10.327)
P=
3
2
6 0 c
c

Radiation losses often determine the maximum achievable energy in a charged
particle accelerator. Let us investigate radiation losses in various different types
of accelerator device using the relativistic Larmor formula.
For a linear accelerator, the motion is one-dimensional. In this case, it is easily
demonstrated that
dp
= m 0 3 u
,
(10.328)
dt
where use has been made of Eq. (10.326), and p = m0 u is the particle momentum in the direction of acceleration (the x-direction, say). Here, m 0 is the
particle rest mass. Thus, Eq. (10.327) yields
e2
dp
P=
2
6 0 m0 c3 dt
!2
(10.329)
The rate of change of momentum is equal to the force exerted on the particle in
the x-direction, which, in turn, equals the change in the energy, E, of the particle
per unit distance. Consequently,
dE
e2
P=
6 0 m02 c3 dx
357
!2
(10.330)
Thus, in a linear accelerator, the radiated power depends on the external force
acting on the particle, and not on the actual energy or momentum of the particle.
It is obvious, from the above formula, that light particles, such as electrons, are
going to radiate a lot more than heavier particles, such as protons. The ratio of
the power radiated to the power supplied by the external sources is
P
e2
e2
1 dE
1 dE
=
'
,
dE/dt 6 0 m02 c3 u dx
6 0 m0 c2 m0 c2 dx
(10.331)
since u ' c for a highly relativistic particle. It is clear, from the above expression, that the radiation losses in an electron linear accelerator are negligible unless the gain in energy is of order me c2 = 0.511 MeV in a distance of
e2 /(6 0 me c2 ) = 1.28 1015 meters. That is 3 1014 MeV/meter. Typical
energy gains are less that 10 MeV/meter. It follows, therefore, that radiation
losses are completely negligible in linear accelerators, whether for electrons, or
for other heavier particles.
The situation is quite different in circular accelerator devices, such as the synchrotron and the betatron. In such machines, the momentum p changes rapidly
in direction as the particle rotates, but the change in energy per revolution is
small. Furthermore, the direction of acceleration is always perpendicular to the
direction of motion. It follows from Eq. (10.327) that
e2 4 u4
e2
4 2
u
=
,
P=
6 0 c3
6 0 c3 2
(10.332)
where is the orbit radius. Here, use has been made of the standard result
u
= u2 / for circular motion. The radiative energy loss per revolution is given by
2
e2 4 u3
E =
P=
.
u
3 0 c 3
(10.333)
For highly relativistic (u ' c) electrons, this expression yields

E(MeV) = 8.85 102
[E(GeV)]4
.
(meters)
(10.334)
In the first electron synchrotrons, 1 meter, Emax 0.3 GeV. Hence, Emax 1
keV per revolution. This was less than, but not negligible compared to, the energy gain of a few keV per turn. For modern electron synchrotrons, the limitation
358
10.26 Angular distribution of radiation
on the available radio-frequency power needed to overcome radiation losses becomes a major consideration, as is clear from the E4 dependence of the radiated
power per turn.
10.26 Angular distribution of radiation

In order to calculate the angular distribution of the energy radiated by an accelerated charge, we must think carefully about what is meant by the rate of radiation
of the charge. This quantity is actually the amount of energy lost by the charge
in a retarded time interval dt 0 during the emission of the signal. Thus,
dE
,
dt 0
where E is the energy of the charge. The Poynting vector
P(t 0 ) =
Erad Brad
2 r
= 0 c Erad
,
0
r
(10.335)
(10.336)
where use has been made of Brad = (r Erad )/r c [see Eq. (10.297)], represents
the energy flux per unit actual time, t. Thus, the energy loss rate of the charge
into a given element of solid angle d is
dP(t 0 )
dE(, )
dE(, ) dt 2
2 s 2
d =
d
=
r
d
=

c
E
0
rad r d, (10.337)
d
dt 0
dt
dt 0
r
where use has been made of Eq. (10.283). Here, and are angular coordinates
used to locate the element of solid angle. It follows from Eq. (10.303) that
)]2
dP(t 0 )
e2 r [r (ru u
=
.
d
162 0 c3
s5
(10.338)
Consider the special case in which the direction of acceleration coincides with
the direction of motion. Let us define spherical polar coordinates whose axis
points along this common direction. It is easily demonstrated that, in this case,
the above expression reduces to
e2 u
2
sin2
dP(t 0 )
=
.
d
162 0 c3 [1 (u/c) cos ]5
359
(10.339)
10.27 Synchrotron radiation
In the non-relativistic limit, u/c 0, the radiation pattern has the familiar sin 2
dependence of dipole radiation. In particular, the pattern is symmetric in the
forward ( < /2) and backward ( > /2) directions. However, as u/c 1, the
radiation pattern becomes more and more concentrated in the forward direction.
The angle max for which the intensity is a maximum is
max
1 q
= cos1
( 1 + 15 u2 /c2 1) .
3 u/c
(10.340)
2 e2 u
2 8 ( )2
dP(t 0 )
' 2
.
d
0 c3 [1 + ( )2 ]5
(10.341)
This expression yields max /2 as u/c 0, and max 1/(2 ) as u/c 1.

Thus, for a highly relativistic charge, the radiation is emitted in a narrow cone
whose axis is aligned along the direction of motion. In this case, the angular
distribution (10.339) reduces to
The total power radiated by the charge is obtained by integrating Eq. (10.339)
over all solid angles. We obtain
Z
Z
e2 u
2
e2 u
2 +1 (1 2 ) d
sin3 d
0
P(t ) =
=
. (10.342)
8 0 c3 0 [1 (u/c) cos ]5
8 0 c3 1 [1 (u/c) ]5
It is easily verified that
Z +1
1
Hence,
(1 2 ) d
4
= 6 .
5
[1 (u/c) ]
3
e2
6 u
2,
P(t ) =
3
6 0 c
= 0.
which agrees with Eq. (10.327), provided that u u
0
(10.343)
(10.344)

Synchrotron radiation (i.e., radiation emitted by a charged particle constrained
to follow a circular orbit by a magnetic field) is of particular importance in astrophysics, since much of the observed radio frequency emission from supernova
remnants and active galactic nuclei is thought to be of this type.
360
Consider a charged particle moving in a circle of radius a with constant angular velocity 0 . Suppose that the orbit lies in the x-y plane. The radius vector
pointing from the centre of the orbit to the retarded position of the charge is
defined
= a (cos , sin , 0),
(10.345)
where = 0 t 0 is the angle subtended between this vector and the x-axis. The
retarded velocity and acceleration of the charge take the form
d
= u ( sin , cos , 0),
dt 0
du
=
=
u (cos , sin , 0),
u
dt 0
(10.346)
u =
(10.347)
where u = a 0 and u
= a 02 . The observation point is chosen such that the radius vector r, pointing from the retarded position of the charge to the observation
point, is parallel to the y-z plane. Thus, we can write
r = r (0, sin , cos ),
(10.348)
where is the angle subtended between this vector and the z-axis. As usual, we
define as the angle subtended between the retarded radius vector r and the
retarded direction of motion of the charge u. It follows that
cos =
ur
= sin cos .
ur
(10.349)
It is easily seen that

r =
u
u r sin sin .
(10.350)
A little vector algebra shows that

)]2 = (r u
)2 r2 (1 u2 /c2 ) + u
[r (ru u
2 r4 (1 ru/r c)2 ,
(10.351)
giving
)]2 = u
[r (ru u
2 r4
u
1 cos
c
!2
u2
1 2 tan2 cos2 .
c
361
(10.352)
Making use of Eq. (10.337), we obtain

dP(t 0 )
e2 u
2 [1 (u/c) cos )]2 (1 u2 /c2 ) tan2 cos2
=
.
d
162 0 c3
[1 (u/c) cos ]5
(10.353)
It is convenient to write this result in terms of the angles and , instead of

and . After a little algebra we obtain
dP(t 0 )
e2 u
2 [1 (u2 /c2 )] cos2 + [(u/c) sin cos ]2
=
.
d
162 0 c3
[1 (u/c) sin cos ]5
(10.354)
Let us consider the radiation pattern emitted in the plane of the orbit: i.e.,
= /2, with cos = cos . It is easily seen that
e2 u
2
[(u/c) cos ]2
dP(t 0 )
=
.
d
162 0 c3 [1 (u/c) cos ]5
(10.355)
In the non-relativistic limit, the radiation pattern has a cos 2 dependence. Thus,
the pattern is like that of dipole radiation where the axis is aligned along the
instantaneous direction of acceleration. As the charge becomes more relativistic,
the radiation lobe in the forward direction (i.e., 0 < < /2) becomes more
more focused and more intense. Likewise, the radiation lobe in the backward
direction (i.e., /2 < < ) becomes more diffuse. The radiation pattern has
zero intensity at the angles
0 = cos1 (u/c).
(10.356)
These angles demark the boundaries between the two radiation lobes. In the
non-relativistic limit, 0 = /2, so the two lobes are of equal angular extents.
In the highly relativistic limit, 0 1/, so the forward lobe becomes highly
concentrated about the forward direction ( = 0). In the latter limit, Eq. (10.355)
reduces to
2 2
dP(t 0 )
e2 u
2
6 [1 ( ) ]
' 2
.
(10.357)
d
2 0 c3
[1 + ( )2 ]5
Thus, the radiation emitted by a highly relativistic charge is focused into an intense beam, of angular extent 1/, pointing in the instantaneous direction of
motion. The maximum intensity of the beam scales like 6 .
362
Integration of Eq. (10.354) over all solid angle (using d = sin d d)

yields
e2
P(t 0 ) =
4 u
2,
(10.358)
3
6 0 c
= 0. This expression can also
which agrees with Eq. (10.327), provided that u u
be written
P
2 02 r0 2 4
=
,
(10.359)
m0 c2
3 c
where r0 = e2 /(4 0 m0 c2 ) = 2.82 1015 meters is the classical electron radius,
m0 is the rest mass of the charge, and = u/c. If the circular motion takes place
in an orbit of radius a, perpendicular to a magnetic field B, then 0 satisfies
0 = e B/m0 . Thus, the radiated power is
2 eB
P
=
m0 c2
3 m0
!2
r0
( )2 ,
c
(10.360)
and the radiated energy E per revolution is

4 r0 3 4
E
=
.
m0 c2
3 a
(10.361)
Let us consider the frequency distribution of the emitted radiation in the highly
relativistic limit. Suppose, for the sake of simplicity, that the observation point lies
in the plane of the orbit (i.e., = /2). Since the radiation emitted by the charge
is beamed very strongly in the charges instantaneous direction of motion, a fixed
observer is only going to see radiation (at some later time) when this direction
points almost directly towards the point of observation. This occurs once every
rotation period, when ' 0, assuming that 0 > 0. Note that the point of
observation is located many orbit radii away from the centre of the orbit along
the positive y-axis. Thus, our observer sees short periodic pulses of radiation
from the charge. The repetition frequency of the pulses (in radians per second) is
0 . Let us calculate the duration of each pulse. Since the radiation emitted by the
charge is focused into a narrow beam of angular extent 1/, our observer
only sees radiation from the charge when < . Thus, the observed pulse is
emitted during a time interval t 0 = /0 . However, the pulse is received in a
363
somewhat shorter time interval
u
t =
,
1
0
c
!
(10.362)
because the charge is slightly closer to the point of observation at the end of the
pulse than at the beginning. The above equation reduces to
t '
,
2
2 0
0 3
(10.363)
since 1 and 1/. The width of the pulse in frequency space obeys
t 1. Hence,
= 3 0 .
(10.364)
In other words, the emitted frequency spectrum contains harmonics up to 3
times that of the cyclotron frequency, 0 .
364

Classical Electromagnetism by R-Fitzpatrick

Uploaded by

Copyright:

Available Formats

Classical Electromagnetism by R-Fitzpatrick

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Classical Electromagnetism by R-Fitzpatrick

Uploaded by

Copyright:

Available Formats

Classical Electromagnetism:

An intermediate level course

3 Time-independent Maxwell equations

The electric scalar potential . . .

4 Time-dependent Maxwells equations

The uniqueness theorem . . . . . . . . . . . . .

6 Dielectric and magnetic media

8 Electromagnetic energy and momentum

10 Relativity and electromagnetism

10.20 Relativistic particle dynamics . . .

1.1 Intended audience

1.2 Major sources

1.4 Outline of course

1.4 Outline of course

2.2 Vector algebra

In applied mathematics, physical quantities are (predominately) represented

2.2 Vector algebra

It is clear that vector algebra is distributive with respect to scalar multiplication:

2.2 Vector algebra

Suppose that we transform to a new orthogonal basis, the x 0 -, y 0 -, and z 0 -axes,

2.3 Vector areas

2.3 Vector areas

2.3 Vector areas

2.4 The scalar product

2.4 The scalar product

the basis through

about the z-axis. In the new basis, a = (1/ 2, 1/ 2, 0)

2.4 The scalar product

+(ax sin + ay cos ) (bx sin + by cos ) + az bz

if a is the position vector of P relative to the origin O. So, the invariance of a a

2.5 The vector product

(AB)2 = (OA)2 + (OB)2 2 (OA) (OB) cos ,

However, according to the cosine rule of trigonometry,

where (AB) denotes the length of side AB. It follows that

Clearly, the invariance of ab under transformation is equivalent to the invariance

The work W performed by a constant force F moving an object through a

2.5 The vector product

2.5 The vector product

Is a x b a proper vector? Suppose that a = (1, 0, 0), b = (0, 1, 0). Clearly,

thebasis through 45 about the z-axis then

Thus, the x-component of a b transforms correctly. It can easily be shown that

2.5 The vector product

Therefore, a b is perpendicular to a. Likewise, it can be demonstrated that a b

= (ax2 + ay2 + az2 ) (bx2 + by2 + bz2 ) (ax bx + ay by + az bz )2

= |a|2 |b|2 (a b)2

= |a|2 |b|2 |a|2 |b|2 cos2 = |a|2 |b|2 sin2 .

normal to the plane of the parallelogram, which means that it is perpendicular to

2.7 The scalar triple product

must be a vector as well. Also, if a 0 is interpreted as a(t+t) in the above equation

2.8 The vector triple product

but any anti-cyclic permutation causes it to change sign,

2.8 The vector triple product

2.9 Vector calculus

which proves the theorem.

2.9 Vector calculus

which implies that

2.10 Line integrals

It is easily demonstrated that

2.10 Line integrals