Field Theory
B O T HID
U PSILON B OOKS
Electromagnetic
Field Theory
B O T HID
Swedish Institute of Space Physics
and
Department of Astronomy and Space Physics
Uppsala University, Sweden
and
School of Mathematics and Systems Engineering
Vxj University, Sweden
Also available
This book was typeset in LATEX 2 (based on TEX 3.14159 and Web2C 7.4.2) on an
HP Visualize 9000360 workstation running HPUX 11.11.
c
Copyright
1997,
1998, 1999, 2000, 2001, 2002, 2003 and 2004 by
Bo Thid
Uppsala, Sweden
All rights reserved.
Electromagnetic Field Theory
ISBN XXXXXXXXXX
Preface
The current book is an outgrowth of the lecture notes that I prepared for the fourcredit
course Electrodynamics that was introduced in the Uppsala University curriculum in
1992, to become the fivecredit course Classical Electrodynamics in 1997. To some
extent, parts of these notes were based on lecture notes prepared, in Swedish, by
B ENGT L UNDBORG who created, developed and taught the earlier, twocredit course
Electromagnetic Radiation at our faculty.
Intended primarily as a textbook for physics students at the advanced undergraduate or beginning graduate level, it is hoped that the present book may be useful for
research workers too. It provides a thorough treatment of the theory of electrodynamics, mainly from a classical field theoretical point of view, and includes such things
as formal electrostatics and magnetostatics and their unification into electrodynamics, the electromagnetic potentials, gauge transformations, covariant formulation of
classical electrodynamics, force, momentum and energy of the electromagnetic field,
radiation and scattering phenomena, electromagnetic waves and their propagation in
vacuum and in media, and covariant Lagrangian/Hamiltonian field theoretical methods for electromagnetic fields, particles and interactions. The aim has been to write
a book that can serve both as an advanced text in Classical Electrodynamics and as a
preparation for studies in Quantum Electrodynamics and related subjects.
In an attempt to encourage participation by other scientists and students in the
authoring of this book, and to ensure its quality and scope to make it useful in higher
university education anywhere in the world, it was produced within a WorldWide
Web (WWW) project. This turned out to be a rather successful move. By making an
electronic version of the book freely downloadable on the net, comments have been
only received from fellow Internet physicists around the world and from WWW hit
statistics it seems that the book serves as a frequently used Internet resource. This
way it is hoped that it will be particularly useful for students and researchers working
under financial or other circumstances that make it difficult to procure a printed copy
of the book.
Thanks are due not only to Bengt Lundborg for providing the inspiration to write
this book, but also to professor C HRISTER WAHLBERG and professor G RAN FLDT,
Uppsala University, and professor YAKOV I STOMIN, Lebedev Institute, Moscow, for
interesting discussions on electrodynamics and relativity in general and on this book in
particular. Comments from former graduate students M ATTIAS WALDENVIK, T OBIA
C AROZZI and ROGER K ARLSSON as well as A NDERS E RIKSSON, all at the Swedish
Institute of Space Physics in Uppsala and who all have participated in the teaching,
vii
P REFACE
on the material covered in the course and in this book are gratefully acknowledged.
Thanks are also due to my longterm space physics colleague H ELMUT KOPKA of
the MaxPlanckInstitut fr Aeronomie, Lindau, Germany, who not only taught me
about the practical aspects of the of highpower radio wave transmitters and transmission lines, but also about the more delicate aspects of typesetting a book in TEX
and LATEX. I am particularly indebted to Academician professor V ITALIY L AZAREVICH G INZBURG , 2003 Nobel Laureate in Physics, for his many fascinating and very
elucidating lectures, comments and historical footnotes on electromagnetic radiation
while cruising on the Volga river at our joint RussianSwedish summer schools during
the 1990s and for numerous private discussions.
Finally, I would like to thank all students and Internet users who have downloaded
and commented on the book during its life on the WorldWide Web.
Uppsala, Sweden
B O T HID
January, 2004
viii
Contents
Preface
vii
Contents
ix
List of Figures
xiii
1 Classical Electrodynamics
1.1 Electrostatics
1.1.1 Coulombs law
1.1.2 The electrostatic field
1.2 Magnetostatics
1.2.1 Ampres law
1.2.2 The magnetostatic field
1.3 Electrodynamics
1.3.1 Equation of continuity for electric charge
1.3.2 Maxwells displacement current
1.3.3 Electromotive force
1.3.4 Faradays law of induction
1.3.5 Maxwells microscopic equations
1.3.6 Maxwells macroscopic equations
1.4 Electromagnetic duality
1.5 Bibliography
1
2
2
3
6
6
7
9
9
10
10
11
14
14
15
22
2 Electromagnetic Waves
2.1 The wave equations
2.1.1 The wave equation for E
2.1.2 The wave equation for B
2.1.3 The timeindependent wave equation for E
2.2 Plane waves
2.2.1 Telegraphers equation
2.2.2 Waves in conductive media
2.3 Observables and averages
25
26
26
26
27
30
31
32
33
ix
C ONTENTS
2.4
Bibliography
34
3 Electromagnetic Potentials
3.1 The electrostatic scalar potential
3.2 The magnetostatic vector potential
3.3 The electrodynamic potentials
3.3.1 LorenzLorentz gauge
3.3.2 Coulomb gauge
3.3.3 Gauge transformations
3.4 Bibliography
35
35
36
36
38
42
42
45
4 Relativistic Electrodynamics
4.1 The special theory of relativity
4.1.1 The Lorentz transformation
4.1.2 Lorentz space
4.1.3 Minkowski space
4.2 Covariant classical mechanics
4.3 Covariant classical electrodynamics
4.3.1 The fourpotential
4.3.2 The LinardWiechert potentials
4.3.3 The electromagnetic field tensor
4.4 Bibliography
47
47
48
49
54
57
58
58
59
61
64
67
67
67
73
73
81
83
83
83
86
88
88
89
91
93
95
96
99
7.4
7.5
Radiated energy
7.4.1 Monochromatic signals
7.4.2 Finite bandwidth signals
Bibliography
101
101
102
103
105
105
106
108
112
112
115
117
118
119
120
122
133
138
145
152
F Formulae
F.1 The electromagnetic field
F.1.1 Maxwells equations
F.1.2 Fields and potentials
F.1.3 Force and energy
F.2 Electromagnetic radiation
F.2.1 Relationship between the field vectors in a plane wave
F.2.2 The far fields from an extended source distribution
F.2.3 The far fields from an electric dipole
F.2.4 The far fields from a magnetic dipole
F.2.5 The far fields from an electric quadrupole
F.2.6 The fields from a point charge in arbitrary motion
F.3 Special relativity
F.3.1 Metric tensor
F.3.2 Covariant and contravariant fourvectors
F.3.3 Lorentz transformation of a fourvector
F.3.4 Invariant line element
F.3.5 Fourvelocity
F.3.6 Fourmomentum
F.3.7 Fourcurrent density
155
155
155
155
156
156
156
156
156
157
157
157
157
157
157
158
158
158
158
158
xi
C ONTENTS
F.4
F.5
xii
F.3.8 Fourpotential
F.3.9 Field tensor
Vector relations
F.4.1 Spherical polar coordinates
F.4.2 Vector formulae
Bibliography
158
158
159
159
160
161
M Mathematical Methods
M.1 Scalars, vectors and tensors
M.1.1 Vectors
M.1.2 Fields
M.1.3 Vector algebra
M.1.4 Vector analysis
M.2 Analytical mechanics
M.2.1 Lagranges equations
M.2.2 Hamiltons equations
M.3 Bibliography
163
163
163
165
171
174
180
180
180
181
Index
183
List of Figures
1.1
1.2
1.3
1.4
3
5
7
12
4.1
4.2
4.3
48
55
56
5.1
74
7.1
100
8.1
8.2
8.3
8.4
8.5
8.6
8.7
8.8
8.9
8.10
8.11
8.12
8.13
8.14
Linear antenna
Electric dipole geometry
Loop antenna
Multipole radiation geometry
Electric dipole geometry
Radiation from a moving charge in vacuum
An accelerated charge in vacuum
Angular distribution of radiation during bremsstrahlung
Location of radiation during bremsstrahlung
Radiation from a charge in circular motion
Synchrotron radiation lobe width
The perpendicular field of a moving charge
Electronelectron scattering
VavilovCerenkov
cone
106
107
109
113
116
120
122
134
135
139
141
144
146
150
168
xiii
CHAPTER 1
Classical Electrodynamics
Classical electrodynamics deals with electric and magnetic fields and interactions
caused by macroscopic distributions of electric charges and currents. This means
that the concepts of localised electric charges and currents assume the validity of
certain mathematical limiting processes in which it is considered possible for the
charge and current distributions to be localised in infinitesimally small volumes of
space. Clearly, this is in contradiction to electromagnetism on a truly microscopic
scale, where charges and currents have to be treated as spatially extended objects and
quantum corrections must be included. However, the limiting processes used will
yield results which are correct on small as well as large macroscopic scales.
It took the genius of JAMES C LERK M AXWELL to unify electricity and magnetism into a supertheory, electromagnetism or classical electrodynamics (CED), and to
realise that optics is a subfield of this supertheory. Early in the 20th century, Nobel laureate H ENDRIK A NTOON L ORENTZ took the electrodynamics theory further to
the microscopic scale and also laid the foundation for the special theory of relativity, formulated by Nobel laureate A LBERT E INSTEIN in 1905. In the 1930s PAUL
A. M. D IRAC expanded electrodynamics to a more symmetric form, including magnetic as well as electric charges. With his relativistic quantum mechanics, he also
paved the way for the development of quantum electrodynamics (QED) for which
R ICHARD P. F EYNMAN, J ULIAN S CHWINGER, and S IN I TIRO T OMONAGA in 1965 received their Nobel prizes. Around the same time, physicists such as Nobel laureates
S HELDON G LASHOW, A BDUS S ALAM, and S TEVEN W EINBERG managed to unify
electrodynamics with the weak interaction theory to yet another supertheory, electroweak theory. The modern theory of strong interactions, quantum chromodynamics
(QCD), is influenced by QED.
In this chapter we start with the force interactions in classical electrostatics and
classical magnetostatics and introduce the static electric and magnetic fields and find
1. C LASSICAL E LECTRODYNAMICS
two uncoupled systems of equations for them. Then we see how the conservation
of electric charge and its relation to electric current leads to the dynamic connection
between electricity and magnetism and how the two can be unified into one supertheory, classical electrodynamics, described by one system of coupled dynamic field
equationsthe Maxwell equations.
At the end of the chapter we study Diracs symmetrised form of Maxwells equations by introducing (hypothetical) magnetic charges and magnetic currents into the
theory. While not identified unambiguously in experiments yet, magnetic charges and
currents make the theory much more appealing for instance by allowing for duality
transformations in a most natural way.
1.1 Electrostatics
The theory which describes physical phenomena related to the interaction between
stationary electric charges or charge distributions in space with stationary boundaries
is called electrostatics. For a long time electrostatics, under the name electricity,
was considered an independent physical theory of its own, alongside other physical
theories such as magnetism, mechanics, optics and thermodynamics. 1
F(x) =
=
=
(1.1)
40 x x0 3
40
40
x x0 
x x0 
where in the last step Formula (F.71) on page 161 was used. In SI units, which we
shall use throughout, the force F is measured in Newton (N), the electric charges q
and q0 in Coulomb (C) [= Ampreseconds (As)], and the length x x0  in metres
(m). The constant 0 = 107 /(4c2 ) 8.8542 1012 Farad per metre (F/m) is the
1 The
The whole theory of electrostatics constitutes a group of abstract ideas and general propositions, formulated in the clear and concise language of geometry and algebra, and connected
with one another by the rules of strict logic. This whole fully satisfies the reason of a French
physicist and his taste for clarity, simplicity and order. . . .
Electrostatics
x x0
q0
x0
O
F IGURE 1.1: Coulombs law describes how a static electric charge q, located at
a point x relative to the origin O, experiences an electrostatic force from a static
electric charge q0 located at x0 .
vacuum permittivity and c 2.9979 108 m/s is the speed of light in vacuum. In CGS
units 0 = 1/(4) and the force is measured in dyne, electric charge in statcoulomb,
and length in centimetres (cm).
Estat lim
q0
F
q
(1.2)
where F is the electrostatic force, as defined in Equation (1.1) on the preceding page,
from a net electric charge q0 on the test particle with a small electric net electric charge
q. Since the purpose of the limiting process is to assure that the test charge q does not
distort the field set up by q0 , the expression for Estat does not depend explicitly on q
but only on the charge q0 and the relative radius vector x x0 . This means that we can
say that any net electric charge produces an electric field in the space that surrounds
it, regardless of the existence of a second charge anywhere in this space. 2
2 In the preface to the first edition of the first volume of his book A Treatise on Electricity and Magnetism, first published in 1873, James Clerk Maxwell describes this in the following, almost poetic, manner
[9]:
For instance, Faraday, in his minds eye, saw lines of force traversing all space where the
mathematicians saw centres of force attracting at a distance: Faraday saw a medium where
they saw nothing but distance: Faraday sought the seat of the phenomena in real actions
1. C LASSICAL E LECTRODYNAMICS
Using (1.1) and Equation (1.2) on the preceding page, and Formula (F.70) on
page 160, we find that the electrostatic field Estat at the field point x (also known as
the observation point), due to a fieldproducing electric charge q 0 at the source point
x0 , is given by
q0 0
1
q0
1
q0 x x 0
stat
=
(1.3)
=
E (x) =
40 x x0 3
40
40
x x0 
x x0 
In the presence of several field producing discrete electric charges q 0i , located at
the points x0i , i = 1, 2, 3, . . . , respectively, in an otherwise empty space, the assumption
of linearity of vacuum3 allows us to superimpose their individual electrostatic fields
into a total electrostatic field
Estat (x) =
1
40
q0i
i
x x0i
3
x x0
(1.4)
If the discrete electric charges are small and numerous enough, we introduce the
electric charge density , measured in C/m3 in SI units, located at x0 within a volume
V 0 of limited extent and replace summation with integration over this volume. This
allows us to describe the total field as
Z
Z
0
1
1
1
stat
3 0
0 xx
3 0
0
E (x) =
d x (x )
=
d x (x )
40 V 0
40 V 0
x x0 
x x0 3
(1.5)
Z
0
(x )
1
d3x0
=
40
x x0 
V0
where we used Formula (F.70) on page 160 and the fact that (x0 ) does not depend on
the unprimed (field point) coordinates on which operates.
We emphasise that under the assumption of linear superposition, Equation (1.5)
above is valid for an arbitrary distribution of electric charges, including discrete
charges, in which case is expressed in terms of Dirac delta distributions:
(x0 ) = q0i (x0 x0i )
(1.6)
as illustrated in Figure 1.2 on the facing page. Inserting this expression into expression (1.5) above we recover expression (1.4).
Taking the divergence of the general Estat expression for an arbitrary electric
charge distribution, Equation (1.5) above, and using the representation of the Dirac
going on in the medium, they were satisfied that they had found it in a power of action at a
distance impressed on the electric fluids.
3 In fact, vacuum exhibits a quantum mechanical nonlinearity due to vacuum polarisation effects manifesting themselves in the momentary creation and annihilation of electronpositron pairs, but classically
this nonlinearity is negligible.
Electrostatics
q
x x0i
q0i
x0i
V0
O
F IGURE 1.2:
1
x x0
d3x0 (x0 )
40 V 0
x x0 3
Z
1
1
=
d3x0 (x0 )
40 V 0
x x0 
Z
1
1
=
d3x0 (x0 ) 2
40 V 0
x x0 
Z
1
(x)
=
d3x0 (x0 ) (x x0 ) =
0
0 V
0
Estat (x) =
(1.7)
stat
Z
0
1
3 0 (x )
(x) =
=0
dx
40
x x0 
V0
(1.8)
(1.9a)
(1.9b)
1. C LASSICAL E LECTRODYNAMICS
1.2 Magnetostatics
While electrostatics deals with static electric charges, magnetostatics deals with stationary electric currents, i.e., electric charges moving with constant speeds, and the
interaction between these currents. Here we shall discuss this theory in some detail.
0 II 0
(x x0 )
dl dl0
4 C
x x0 3
C0
I
0 I
1
0 II
0
dl dl
=
4 C
x x0 
C0
F(x) =
(1.10)
107
1
(F/m) 4 107 (H/m) = 2 (s2 /m2 )
2
4c
c
(1.11)
dl dl
4 C C 0 x x0 3
Since the integrand in the first integral is an exact differential, this integral vanishes
and we can rewrite the force expression, Equation (1.10) above, in the following symmetric way
0 II 0
F(x) =
4
I I
C
C0
x x0
dl dl0
x x0 3
(1.13)
Magnetostatics
J
C
dl
x x0
dl0
x
C0
x
J0
O
Ampres law describes how a small loop C, carrying a static
electric current I through its tangential line element dl located at x, experiences
a magnetostatic force from a small loop C 0 , carrying a static electric current I 0
through the tangential line element dl0 located at x0 . The loops can have arbitrary
shapes as long as they are simple and closed.
F IGURE 1.3:
dBstat (x)
0 I 0 0
x x0
dl
4
x x0 3
(1.14)
which expresses the small element dBstat (x) of the static magnetic field set up at the
field point x by a small line element dl0 of stationary current I 0 at the source point
x0 . The SI unit for the magnetic field, sometimes called the magnetic flux density or
magnetic induction, is Tesla (T).
If we generalise expression (1.14) to an integrated steady state electric current
density j(x), measured in A/m2 in SI units, we obtain BiotSavarts law:
Z
Z
0
0
x x0
1
3 0
0
Bstat (x) =
=
d3x0 j(x0 )
d
x
j(x
)
4 V 0
4 V 0
x x0 
x x0 3
Z
0
j(x0 )
=
d3x0
4
x x0 
V0
(1.15)
where we used Formula (F.70) on page 160, Formula (F.57) on page 160, and the
fact that j(x0 ) does not depend on the unprimed coordinates on which operates.
Comparing Equation (1.5) on page 4 with Equation (1.15), we see that there exists a
1. C LASSICAL E LECTRODYNAMICS
close analogy between the expressions for Estat and Bstat but that they differ in their
vectorial characteristics. With this definition of Bstat , Equation (1.10) on page 6 may
we written
F(x) = I
dl Bstat (x)
(1.16)
In order to assess the properties of Bstat , we determine its divergence and curl.
Taking the divergence of both sides of Equation (1.15) on the preceding page and
utilising Formula (F.63) on page 160, we obtain
Bstat (x) =
Z
0
j(x0 )
=0
d3x0
4
x x0 
V0
(1.17)
since, according to Formula (F.63) on page 160, ( a) vanishes for any vector
field a(x).
Applying the operator baccab rule, Formula (F.64) on page 160, the curl of
Equation (1.15) on the preceding page can be written
Z
0
0
3 0 j(x )
=
B (x) =
dx
4
x x0 
V0
Z
Z
0
0
1
1
3 0
0
2
3 0
0
0
0
d x j(x )
d x [j(x ) ]
=
+
4 V 0
4 V 0
x x0 
x x0 
(1.18)
stat
In the first of the two integrals on the right hand side, we use the representation of the
Dirac delta function given in Formula (F.73) on page 161, and integrate the second
one by parts, by utilising Formula (F.56) on page 160 as follows:
1
d x [j(x ) ]
x x0 
V0
Z
Z
0
0
1
1
3 0
0
0
3 0
0
= x k d x j(x )
d x j(x )
x0k x x0 
x x0 
V0
V0
Z
Z
1
1
= x k dS j(x0 ) 0
d3x0 0 j(x0 ) 0
xk x x0 
x x0 
S
V0
(1.19)
3 0
Then we note that the first integral in the result, obtained by applying Gausss theorem,
vanishes when integrated over a large sphere far away from the localised source j(x 0 ),
and that the second integral vanishes because j = 0 for stationary currents (no
charge accumulation in space). The net result is simply
Bstat (x) = 0
V0
(1.20)
Electrodynamics
1.3 Electrodynamics
As we saw in the previous sections, the laws of electrostatics and magnetostatics can
be summarised in two pairs of timeindependent, uncoupled vector partial differential
equations, namely the equations of classical electrostatics
(x)
0
Estat (x) = 0
Estat (x) =
(1.21a)
(1.21b)
stat
(1.22a)
(x) = 0 j(x)
(1.22b)
Since there is nothing a priori which connects Estat directly with Bstat , we must consider classical electrostatics and classical magnetostatics as two independent theories.
However, when we include timedependence, these theories are unified into one
theory, classical electrodynamics. This unification of the theories of electricity and
magnetism is motivated by two empirically established facts:
1. Electric charge is a conserved quantity and electric current is a transport of
electric charge. This fact manifests itself in the equation of continuity and, as a
consequence, in Maxwells displacement current.
2. A change in the magnetic flux through a loop will induce an EMF electric field
in the loop. This is the celebrated Faradays law of induction.
(1.23)
which states that the time rate of change of electric charge (t, x) is balanced by a
divergence in the electric current density j(t, x).
1. C LASSICAL E LECTRODYNAMICS
= 0 j(t, x) + 0 0 E(t, x)
t
(1.24)
where, in the last step, we have assumed that a generalisation of Equation (1.5) on
page 4 to timevarying fields allows us to make the identification4
Z
Z
1
1
1
1
3 0
0
0
3 0
0
d x (t, x )
d x (t, x )
=
40 t V 0
t
40 V 0
x x0 
x x0 
Z
0
1
(t, x )
=
= E(t, x)
d3x0
0
0
t
40
t
x x 
V
(1.25)
The result is Maxwells source equation for the B field
(1.26)
where the last term 0 E(t, x)/t is the famous displacement current. This term was
introduced, in a stroke of genius, by Maxwell[8] in order to make the right hand side
of this equation divergence free when j(t, x) is assumed to represent the density of the
total electric current, which can be split up in ordinary conduction currents, polarisation currents and magnetisation currents. The displacement current is an extra term
which behaves like a current density flowing in vacuum. As we shall see later, its
existence has farreaching physical consequences as it predicts the existence of electromagnetic radiation that can carry energy and momentum over very long distances,
even in vacuum.
10
Electrodynamics
materials, one can sometimes assume a linear relationship between the electric current
density j and E, called Ohms law:
j(t, x) = E(t, x)
(1.27)
where is the electric conductivity (S/m). In the most general cases, for instance in
an anisotropic conductor, is a tensor.
We can view Ohms law, Equation (1.27) above, as the first term in a Taylor expansion of the law j[E(t, x)]. This general law incorporates nonlinear effects such as
frequency mixing. Examples of media which are highly nonlinear are semiconductors and plasma. We draw the attention to the fact that even in cases when the linear
relation between E and j is a good approximation, we still have to use Ohms law with
care. The conductivity is, in general, timedependent (temporal dispersive media)
but then it is often the case that Equation (1.27) is valid for each individual Fourier
component of the field.
If the current is caused by an applied electric field E(t, x), this electric field will
exert work on the charges in the medium and, unless the medium is superconducting,
there will be some energy loss. The rate at which this energy is expended is j E
per unit volume. If E is irrotational (conservative), j will decay away with time.
Stationary currents therefore require that an electric field which corresponds to an
electromotive force (EMF) is present. In the presence of such a field EEMF , Ohms
law, Equation (1.27) above, takes the form
j = (Estat + EEMF )
(1.28)
dl (Estat + EEMF )
(1.29)
dl EEMF
(1.30)
It has been established experimentally that a nonconservative EMF field is produced in a closed circuit C if the magnetic flux through this circuit varies with time.
11
1. C LASSICAL E LECTRODYNAMICS
dS
B(x)
v
C
dl
B(x)
F IGURE 1.4:
dl E(t, x) =
d
dt
d
m (t, x)
dt
dS B(t, x) =
dS
B(t, x)
t
(1.31)
where m is the magnetic flux and S is the surface encircled by C which can be interpreted as a generic stationary loop and not necessarily as a conducting circuit.
Application of Stokes theorem on this integral equation, transforms it into the differential equation
E(t, x) =
B(t, x)
t
(1.32)
which is valid for arbitrary variations in the fields and constitutes the Maxwell equation which explicitly connects electricity with magnetism.
Any change of the magnetic flux m will induce an EMF. Let us therefore consider
the case, illustrated if Figure 1.4, that the loop is moved in such a way that it links
12
Electrodynamics
a magnetic field which varies during the movement. The convective derivative is
evaluated according to the wellknown operator formula
d
=
+v
dt t
(1.33)
which follows immediately from the rules of differentiation of an arbitrary differentiable function f (t, x(t)). Applying this rule to Faradays law, Equation (1.31) on the
preceding page, we obtain
E(t, x) =
d
dt
dS B =
dS
dS (v )B
(1.34)
(1.35)
(it is one of Maxwells equations) so that, according to Formula (F.59) on page 160,
(B v) = (v )B
(1.36)
dl EEMF =
dS
d
dt
S
dS B
(1.37)
dS (B v)
dl EEMF =
dS
dl (B v)
(1.38)
dl (EEMF v B) =
dS
B
t
(1.39)
where EEMF is the field which is induced in the loop, i.e., in the moving system. The
use of Stokes theorem backwards on Equation (1.39) above yields
(EEMF v B) =
B
t
(1.40)
(1.41)
13
1. C LASSICAL E LECTRODYNAMICS
(1.42)
(1.43)
Hence, we can conclude that for a stationary observer, the Maxwell equation
E=
B
t
(1.44)
E=
(1.45a)
B
t
B=0
B = 0 0
(1.45b)
(1.45c)
E
+ 0 j(t, x)
t
(1.45d)
In these equations (t, x) represents the total, possibly both time and space dependent, electric charge, i.e., free as well as induced (polarisation) charges, and j(t, x)
represents the total, possibly both time and space dependent, electric current, i.e.,
conduction currents (motion of free charges) as well as all atomistic (polarisation,
magnetisation) currents. As they stand, the equations therefore incorporate the classical interaction between all electric charges and currents in the system and are called
Maxwells microscopic equations. Another name often used for them is the MaxwellLorentz equations. Together with the appropriate constitutive relations, which relate
and j to the fields, and the initial and boundary conditions pertinent to the physical
situation at hand, they form a system of wellposed partial differential equations which
completely determine E and B.
14
Electromagnetic duality
(1.46a)
H = H[t, x; E, B]
(1.46b)
Under certain conditions, for instance for very low field strengths, we may assume
that the response of a substance to the fields is linear so that
D = E
(1.47)
(1.48)
H= B
i.e., that the derived fields are linearly proportional to the primary fields and that the
electric displacement (magnetising field) is only dependent on the electric (magnetic)
field.
The field equations expressed in terms of the derived field quantities D and H are
D = (t, x)
B
E=
t
B=0
D
H=
+ j(t, x)
t
(1.49a)
(1.49b)
(1.49c)
(1.49d)
and are called Maxwells macroscopic equations. We will study them in more detail
in Chapter 6.
15
1. C LASSICAL E LECTRODYNAMICS
E=
(1.50a)
B
(1.50b)
0 jm
t
B = 0 m
(1.50c)
E
B = 0 0
+ 0 je
(1.50d)
t
We shall call these equations Diracs symmetrised Maxwell equations or the electromagnetodynamic equations.
Taking the divergence of (1.50b), we find that
E=
( B) 0 jm 0
(1.51)
t
where we used the fact that, according to Formula (F.63) on page 160, the divergence
of a curl always vanishes. Using (1.50c) to rewrite this relation, we obtain the equation
of continuity for magnetic monopoles
( E) =
m
+ jm = 0
(1.52)
t
which has the same form as that for the electric monopoles (electric charges) and
currents, Equation (1.23) on page 9.
We notice that the new Equations (1.50) exhibit the following symmetry (recall
that 0 0 = 1/c2 ):
E cB
(1.53a)
cB E
e
(1.53b)
(1.53c)
c
e
cj j
(1.53d)
j cj
(1.53e)
(1.53f)
E = E cos + cB sin
c B = E sin + cB cos
(1.54b)
? m
(1.54d)
? e
c = c cos + sin
e
= c sin + cos
? e
c j = cj cos + j sin
? m
j = cj sin + j cos
16
(1.54a)
(1.54c)
(1.54e)
(1.54f)
Electromagnetic duality
which leaves the symmetrised Maxwell equations, and hence the physics they describe
(often referred to as electromagnetodynamics), invariant. Since E and j e are (true or
polar) vectors, B a pseudovector (axial vector), e a (true) scalar, then m and , which
behaves as a mixing angle in a twodimensional charge space, must be pseudoscalars
and jm a pseudovector.
BFARADAY S LAW AS A CONSEQUENCE OF CONSERVATION OF MAGNETIC CHARGE
E XAMPLE 1.1
Postulate 1.1 (Indestructibility of magnetic charge). Magnetic charge exists and is indestructible in the same way that electric charge exists and is indestructible. In other words we
postulate that there exists an equation of continuity for magnetic charges:
m (t, x)
+ jm (t, x) = 0
t
Use this postulate and Diracs symmetrised form of Maxwells equations to derive
Faradays law.
The assumption of the existence of magnetic charges suggests a Coulomblike law for magnetic fields:
Z
Z
0
x x0
1
0
3 0 m 0
Bstat (x) =
d3x0 m (x0 )
d
x
(x
)
=
4 V 0
4 V 0
x x0 
x x0 3
(1.55)
Z
m 0
(x )
0
= d3x0
4
x x0 
V0
[cf. Equation (1.5) on page 4 for Estat ] and, if magnetic currents exist, a BiotSavartlike law for
electric fields [cf. Equation (1.15) on page 7 for Bstat ]:
Z
Z
0
x x0
0
1
stat
3 0 m 0
3 0 m 0
E (x) =
d x j (x )
d x j (x )
=
4 V 0
4 V 0
x x0 
x x0 3
(1.56)
Z
m 0
0
j
(x
)
= d3x0
4
x x0 
V0
Taking the curl of the latter and using the operator baccab rule, Formula (F.59) on page 160,
we find that
Z
0
jm (x0 )
=
Estat (x) = d3x0
4
x x0 
V0
(1.57)
Z
Z
0
0
1
1
3 0
m 0
0
0
=
+
d3x0 jm (x0 )2
d
x
[j
(x
)
]
4 V 0
4 V 0
x x0 
x x0 
Comparing with Equation (1.18) on page 8 for Estat and the evaluation of the integrals there, we
obtain
Estat (x) = 0
V0
(1.58)
We assume that Formula (1.56) above is valid also for timevarying magnetic currents.
Then, with the use of the representation of the Dirac delta function, Equation (F.73) on page 161,
the equation of continuity for magnetic charge, Equation (1.52) on the preceding page, and
the assumption of the generalisation of Equation (1.55) to timedependent magnetic charge
17
1. C LASSICAL E LECTRODYNAMICS
V0
= 0 jm (t, x) B(t, x)
t
0
4 t
V0
d3x0 m (t, x0 )0
1
x x0 
(1.59)
[cf. Equation (1.24) on page 10] which we recognise as Equation (1.50b) on page 16. A
transformation of this electromagnetodynamic result by rotating into the electric realm of
charge space, thereby letting jm tend to zero, yields the electrodynamic Equation (1.50b) on
page 16, i.e., the Faraday law in the ordinary Maxwell equations. This process also provides
an alternative interpretation of the term B/t as a magnetic displacement current, dual to the
electric displacement current [cf. Equation (1.26) on page 10].
By postulating the indestructibility of a hypothetical magnetic charge, we have thereby been
able to replace Faradays experimental results on electromotive forces and induction in loops as
a foundation for the Maxwell equations by a more appealing one.
C E ND OF EXAMPLE 1.1
E XAMPLE 1.2
1
e cos + m sin =
=
0
c
0
?B
1
= (E cos + cB sin ) +
E sin + B cos
?E +
t
t
c
B
1 E
= 0 jm cos
cos + c0 je sin +
sin
t
c t
1 E
B
sin +
cos = 0 jm cos + c0 je sin
c t
t
= 0 (cje sin + jm cos ) = 0 ?jm
e
1
sin + 0 m cos
?B = ( E sin + B cos ) =
c
c0
(1.60)
(1.61)
(1.62)
18
Electromagnetic duality
?B
1
1
1 ? E
= ( E sin + B cos ) 2 (E cos + cB sin )
c2 t
c
c t
1 B
1 E
1
cos + 0 je cos + 2
cos
= 0 jm sin +
c
c t
c t
1 E
1 B
2
cos
sin
c t
c t
1 m
= 0
j sin + je cos = 0 ?je
c
(1.63)
QED
C E ND OF EXAMPLE 1.2
E XAMPLE 1.3
(1.64a)
j = cj tan
(1.64b)
=
cos
cos
? m
= ce sin + ce tan cos = ce sin + ce sin = 0
? e
1
1 e
(je cos2 + je sin2 ) =
j
cos
cos
? m
j = cje sin + cje tan cos = cje sin + cje sin = 0
? e
(1.65a)
(1.65b)
(1.65c)
(1.65d)
Hence, a fixed mixing angle, or, equivalently, a fixed ratio between the electric and magnetic
charges/currents, hides the magnetic monopole influence ( m and jm ) on the dynamic equations.
We notice that the inverse of the transformation given by Equation (1.54) on page 16 yields
E = ?E cos c?B sin
(1.66)
(1.67)
Furthermore, from the expressions for the transformed charges and currents above, we find that
?E =
? e
1 e
=
0
cos 0
(1.68)
and
19
1. C LASSICAL E LECTRODYNAMICS
?B = 0 ? m = 0
(1.69)
so that
E=
1 e
e
cos 0 =
cos 0
0
(1.70)
QED
C E ND OF EXAMPLE 1.3
(1.71)
(1.72)
is conserved. I.e.,
E 2 c2 B2 = Const
(1.73a)
E B = Const
(1.73b)
(1.74)
. . . there are strong theoretical reasons to believe that magnetic charge exists in nature,
and may have played an important role in the development of the universe. Searches for
magnetic charge continue at the present time, emphasizing that electromagnetism is very
far from being a closed object.
20
Electromagnetic duality
= E E c2 B B + ic(E B) + ic(B E)
(1.75)
= 0 + 0 + ic(E B) ic(E B) = 0
= E E + c2 B B ic(E B) + ic(B E)
(1.76)
E XAMPLE 1.5
Expressed in the complex field vector, introduced in Example 1.4 on the facing page, the
duality transformation Equations (1.54) on page 16 become
?
(1.77)
(1.78)
while
G ?G = e2i G G
(1.79)
Furthermore, assuming that = (t, x), we see that the spatial and temporal differentiation
of ?G leads to
?G
= i(t )ei G + ei t G
t
?G ?G = iei G + ei G
t ? G
?
(1.80a)
(1.80b)
G G = ie G + e G
(1.80c)
21
1. C LASSICAL E LECTRODYNAMICS
1.5 Bibliography
[1]
[2]
[3]
[4]
[5]
J. D. JACKSON, Classical Electrodynamics, third ed., John Wiley & Sons, Inc.,
New York, NY . . . , 1999, ISBN 047130932X.
[6]
L. D. L ANDAU AND E. M. L IFSHITZ, The Classical Theory of Fields, fourth revised English ed., vol. 2 of Course of Theoretical Physics, Pergamon Press, Ltd., Oxford . . . , 1975,
ISBN 0080250726.
[7]
F. E. L OW, Classical Field Theory, John Wiley & Sons, Inc., New York, NY . . . , 1997,
ISBN 0471595519.
[8]
J. C. M AXWELL, A dynamical theory of the electromagnetic field, Royal Society Transactions, 155 (1864).
[9]
J. C. M AXWELL, A Treatise on Electricity and Magnetism, third ed., vol. 1, Dover Publications, Inc., New York, NY, 1954, ISBN 0486606368.
[10] J. C. M AXWELL, A Treatise on Electricity and Magnetism, third ed., vol. 2, Dover Publications, Inc., New York, NY, 1954, ISBN 0486606378.
[11] D. B. M ELROSE AND R. C. M C P HEDRAN, Electromagnetic Processes in Dispersive Media, Cambridge University Press, Cambridge . . . , 1991, ISBN 0521410258.
[12] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism, second ed.,
AddisonWesley Publishing Company, Inc., Reading, MA . . . , 1962, ISBN 0201057026.
[13] F. ROHRLICH, Classical Charged Particles, Perseus Books Publishing, L.L.C., Reading,
MA . . . , 1990, ISBN 0201483009.
[14] J. S CHWINGER, A magnetic model of matter, Science, 165 (1969), pp. 757761.
[15] J. S CHWINGER , L. L. D E R AAD , J R ., K. A. M ILTON , AND W. T SAI, Classical Electrodynamics, Perseus Books, Reading, MA, 1998, ISBN 0738200565.
[16] J. A. S TRATTON, Electromagnetic Theory, McGrawHill Book Company, Inc., New
York, NY and London, 1953, ISBN 070621500.
22
Bibliography
[17] J. VANDERLINDE, Classical Electromagnetic Theory, John Wiley & Sons, Inc., New
York, Chichester, Brisbane, Toronto, and Singapore, 1993, ISBN 0471572691.
23
CHAPTER 2
Electromagnetic Waves
In this chapter we investigate the dynamical properties of the electromagnetic field by
deriving a set of equations which are alternatives to the Maxwell equations. It turns
out that these alternative equations are wave equations, indicating that electromagnetic
waves are natural and common manifestations of electrodynamics.
Maxwells microscopic equations [cf. Equations (1.45) on page 14] are
(t, x)
0
B
E=
t
B=0
E
B = 0 0
+ 0 j(t, x)
t
E=
(Coulombs/Gausss law)
(2.1a)
(Faradays law)
(2.1b)
(2.1c)
(Ampres/Maxwells law)
(2.1d)
25
2. E LECTROMAGNETIC WAVES
( E) = ( B) = 0
(2.2)
j + 0 E
t
t
t
According to the operator triple product baccab rule Equation (F.64) on page 160
( E) = ( E) 2 E
(2.3)
(2.4)
and since EEMF = 0, Ohms law, Equation (1.28) on page 11, yields
j = E
(2.5)
2 E 0
E + 0 E = 0
t
t
(2.6)
E 1 2 E
2 2 =0
t
c t
(2.7)
26
( E) = 0 E + 0 0 ( E) (2.8)
t
t
which, with the use of Equation (F.64) on page 160 and Equation (2.1c) on page 25
can be rewritten
( B) 2 B = 0
B
2
0 0 2 B
t
t
(2.9)
Using the fact that, according to (2.1c), B = 0 for any medium and rearranging, we
can rewrite this equation as
2 B 0
B 1 2 B
2 2 =0
t
c t
(2.10)
This is the wave equation for the magnetic field. We notice that it is of exactly the
same form as the wave equation for the electric field, Equation (2.7) on the facing
page.
(2.11)
and insert this into Equation (2.7) on the preceding page. This yields
1 2
E0 (x)eit 2 2 E0 (x)eit
t
c t
1
= 2 E 0 (i)E0 (x)eit 2 (i)2 E0 (x)eit
c
1
= 2 E 0 (i)E 2 (i)2 E =
c
2
E=0
= 2 E + 2 1 + i
c
0
2 E0 (x)eit 0
(2.12)
2
E=0
c2
(2.14)
27
2. E LECTROMAGNETIC WAVES
which is a timeindependent wave equation for E, representing weakly damped propagating waves. In the short limit we have instead
2 E + i0 E = 0
(2.15)
(2.16)
we can write, using the fact that c = 1/ 0 0 according to Equation (1.11) on page 6,
r
1
1
0
=
=
=
= R0
(2.17)
0 0 ck
k
0
k
where in the last step we introduced the characteristic impedance for vacuum
r
0
R0 =
376.7
(2.18)
0
E XAMPLE 2.1
(2.19b)
B = 0 0
(2.19d)
E=
E
+ 0 je
t
(2.19a)
(2.19c)
under the assumption of vanishing net electric and magnetic charge densities and in the absence
of electromotive and magnetomotive forces. Interpret this equation physically.
Taking the curl of (2.19b) and using (2.19d), and assuming, for symmetry reasons, that
there exists a linear relation between the magnetic current density j m and the magnetic field B
(the analogue of Ohms law for electric currents, je = e E)
jm = m B
(2.20)
28
( E) = 0 jm ( B) = 0 m B
t
1
E
1 E
2
= 0 m 0 e E + 2
0 e
c t
t
c
0 j e +
1 E
c2 t
2 E
t2
(2.21)
E 0
m
+ 2
c
e
E
1 2 E
2 2 20 m e E = 0
t
c t
(2.22)
e + m /c2
0
= 2 E + 2
1 2 m e + i
E=0
c
0
0
(2.23)
Realising that, according to Formula (2.18) on the preceding page, 0 /0 is the square of the
vacuum radiation resistance R0 , and rearranging a bit, we obtain the timeindependent wave
equation in Diracs symmetrised electrodynamics
2
2
e
m 2
/c
E = 0
2 E + 2 1 02 m e 1 + i
(2.24)
R2
c
1 0 m e
0
From this equation we conclude that the existence of magnetic charges (magnetic monopoles),
and nonvanishing electric and magnetic conductivities would lead to a shift in the effective
wave number of the wave. Furthermore, even if the electric conductivity e vanishes, the
imaginary term does not necessarily vanish and the wave might therefore experience damping
(or growth) according as m is positive (or negative). This would happen in a hypotethetical
medium which is a perfect insulator for electric currents but which can carry magnetic currents.
Finally, we note that in the particular case that = R0 m e , the wave equation becomes
a (timeindependent) diffusion equation
m
2 E + i0 e + 2 E = 0
(2.25)
c
and, hence, no waves exist at all!
C E ND OF EXAMPLE 2.1
29
2. E LECTROMAGNETIC WAVES
(2.26)
E
n
B
n
B
n
=0
=
(2.27a)
B
t
(2.27b)
=0
= 0 j(t, x) + 0 0
(2.27c)
E
E
= 0 E + 0 0
t
t
0 = n n
= n 0 + 0 0
E
(2.27d)
(2.28)
which simplifies to the firstorder ordinary differential equation for the normal component E n of the electric field
dE n
+ En = 0
dt
0
(2.29)
(2.30)
This, together with (2.27a), shows that the longitudinal component of E, i.e., the component which is perpendicular to the plane surface is independent of and has a time
dependence which exhibits an exponential decay, with a decrement given by the relaxation time in the medium.
Scalar multiplying (2.27b) by n,
we similarly find that
E
B
0 = n n
= n
(2.31)
t
or
B
=0
(2.32)
t
From this, and (2.27c), we conclude that the only longitudinal component of B must
be constant in both time and space. In other words, the only nonstatic solution must
consist of transverse components.
n
30
Plane waves
t
c t
(2.33)
This equation, which describes the propagation of plane waves in a conducting medium, is called the telegraphers equation. If the medium is an insulator so that = 0,
then the equation takes the form of the onedimensional wave equation
2 E 1 2 E
=0
2 c2 t2
(2.34)
As is well known, each component of this equation has a solution which can be written
E i = f ( ct) + g( + ct),
i = 1, 2, 3
(2.35)
(2.36)
k = k n = n = k
c
c
this solution can be written as
(2.37)
E = E0 ei(kxt)
(2.38)
Let us consider the lower sign in front of k in the exponent in (2.36). This corresponds to a wave which propagates in the direction of increasing . Inserting this
solution into Equation (2.27b) on the preceding page, gives
n
E
= iB = ik n E
(2.39)
k
1
1
n E = k E = k E = 0 0 n E
(2.40)
31
2. E LECTROMAGNETIC WAVES
+ K2E = 0
E
+
i
E
=
0
0
0
2
2
(2.41)
where
2
K = 0 0
1+i
0
2
= 2
c
1+i
0
=k
1+i
0
(2.42)
where, in the last step, Equation (2.16) on page 28 was used to introduce the wave
number k. Taking the square root of this expression, we obtain
r
K =k 1+i
= + i
(2.43)
0
Squaring, one finds that
= (2 2 ) + 2i
k2 1 + i
0
(2.44)
or
2 = 2 k 2
(2.45)
(2.46)
k
20
Squaring the latter and combining with the former, one obtains the second order algebraic equation (in 2 )
2 (2 k2 ) =
k 4 2
420 2
(2.47)
=k
=k
32
vr
u
2
u
u 1+
+1
t
0
2
vr
u
2
u
u 1+
1
t
0
2
(2.48a)
(2.48b)
As a consequence, the solution of the timeindependent telegraphers equation, Equation (2.41) on the preceding page, can be written
E = E0 e ei(t)
(2.49)
With the aid of Equation (2.40) on page 31 we can calculate the associated magnetic
field, and find that it is given by
B=
1
1
1
K k E = ( k E)( + i) = ( k E) A ei
(2.50)
where we have, in the last step, rewritten +i in the amplitudephase form A exp{i}.
From the above, we immediately see that E, and consequently also B, is damped, and
that E and B in the wave are out of phase.
In the limit 0 , we can approximate K as follows:
1
1
r
0 2
2
1i
=k i
k(1 + i)
K =k 1+i
0
0
20
r
r
= 0 0 (1 + i)
= (1 + i)
20
2
(2.51)
In this limit we dind that when the wave impinges perpendicularly upon the medium,
the fields are given, inside the medium, by
r
r
0
0
0
exp i
t
(2.52a)
E = E0 exp
2
2
r
0
B0 = (1 + i)
( n E0 )
(2.52b)
2
Hence, both fields fall off by a factor 1/e at a distance
s
2
=
0
(2.53)
33
2. E LECTROMAGNETIC WAVES
vectors F and G which both are timeharmonic, i.e., can be represented by Fourier
components proportional to exp{it}, then we must make the following interpretation
F(t, x) G(t, x) = Re {F} Re {G} = Re F0 (x) eit Re G0 (x) eit
(2.54)
Furthermore, letting denote complex conjugate, we can express the real part of the
complex vector F as
1
Re {F} = Re F0 (x) eit = [F0 (x) eit + F0 (x) eit ]
2
(2.55)
and similarly for G. Hence, the physically acceptable interpretation of the scalar
product of two complex vectors, representing physical observables, is
F(t, x) G(t, x) = Re F0 (x) eit Re G0 (x) eit
1
1
= [F0 (x) eit + F0 (x) eit ] [G0 (x) eit + G0 (x) eit ]
2
2
1
= F0 G0 + F0 G0 + F0 G0 e2it + F0 G0 e2it
4
1
= Re F0 G0 + F0 G0 e2it
2
1
= Re F0 eit G0 eit + F0 G0 e2it
2
1
= Re F(t, x) G (t, x) + F0 G0 e2it
2
(2.56)
Often in physics, we measure temporal averages (h i) of our physical observables.
If so, we see that the average of the product of the two physical quantities represented
by F and G can be expressed as
hF Gi hRe {F} Re {G}i =
1
1
Re {F G } = Re {F G}
2
2
(2.57)
2.4 Bibliography
[1] J. D. JACKSON, Classical Electrodynamics, third ed., John Wiley & Sons, Inc.,
New York, NY . . . , 1999, ISBN 047130932X.
[2] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism, second ed.,
AddisonWesley Publishing Company, Inc., Reading, MA . . . , 1962, ISBN 0201057026.
34
CHAPTER 3
Electromagnetic Potentials
Instead of expressing the laws of electrodynamics in terms of electric and magnetic
fields, it turns out that it is often more convenient to express the theory in terms of
potentials. This is particularly true for problems related to radiation. In this chapter
we will introduce and study the properties of such potentials and shall find that they
exhibit some remarkable properties which elucidate the fundamental apects of electromagnetism and lead naturally to the special theory of relativity.
(3.1)
Taking the divergence of this and using Equation (1.7) on page 5, we obtain Poissons
equation
2 stat (x) = Estat (x) =
(x)
0
(3.2)
A comparison with the definition of Estat , namely Equation (1.5) on page 4, shows that
this equation has the solution
stat (x) =
1
40
V0
(x0 ) 3 0
dx +
x x0 
(3.3)
35
3. E LECTROMAGNETIC P OTENTIALS
where the integration is taken over all source points x0 at which the charge density
(x0 ) is nonzero and is an arbitrary quantity which has a vanishing gradient. An
example of such a quantity is a scalar constant. The scalar function stat (x) in Equation (3.3) on the preceding page is called the electrostatic scalar potential.
(3.4)
stat
0
(x) =
4
V0
j(x0 ) 3 0
d x + a(x)
x x0 
(3.5)
where a(x) is an arbitrary vector field whose curl vanishes. From Equation (F.62) on
page 160 we know that such a vector can always be written as the gradient of a scalar
field.
36
(3.6)
Inserting this expression into the other homogeneous Maxwell equation (1.32) on page
12, we obtain
E(t, x) =
(3.7)
E(t, x) + A(t, x) = 0
t
(3.8)
As before we utilise the vanishing curl of a vector expression to write this vector expression as the gradient of a scalar function. If, in analogy with the electrostatic case, we introduce the electromagnetic scalar potential function (t, x), Equation (3.8) becomes equivalent to
E(t, x) +
A(t, x) = (t, x)
t
(3.9)
This means that in electrodynamics, E(t, x) can be calculated from the formula
E(t, x) = (t, x)
A(t, x)
t
(3.10)
and B(t, x) from Equation (3.6) on the preceding page. Hence, it is a matter of taste
whether we want to express the laws of electrodynamics in terms of the potentials
(t, x) and A(t, x), or in terms of the fields E(t, x) and B(t, x). However, there exists
an important difference between the two approaches: in classical electrodynamics the
only directly observable quantities are the fields themselves (and quantities derived
from them) and not the potentials. On the other hand, the treatment becomes significantly simpler if we use the potentials in our calculations and then, at the final stage,
use Equation (3.6) on the facing page and Equation (3.10) above to calculate the fields
or physical quantities expressed in the fields.
Inserting (3.10) and (3.6) on the facing page into Maxwells equations (1.45) on
page 14 we obtain, after some simple algebra and the use of Equation (1.11) on page 6,
the general inhomogeneous wave equations
(t, x)
( A)
0
t
2
1 A
1
2 A 2 2 ( A) = 0 j(t, x) + 2
c t
c t
2 =
A
+
=
+
c2 t2
0
t
c2 t
1
1 2 A
2
A
=
j(t,
x)
A
+
0
c2 t2
c2 t
(3.11a)
(3.11b)
(3.12a)
(3.12b)
37
3. E LECTROMAGNETIC P OTENTIALS
These two second order, coupled, partial differential equations, representing in all four
scalar equations (one for and one each for the three components A i , i = 1, 2, 3 of A)
are completely equivalent to the formulation of electrodynamics in terms of Maxwells equations, which represent eight scalar firstorder, coupled, partial differential
equations.
As they stand, Equations (3.11) on the preceding page and Equations (3.12) on the
previous page look complicated and may seem to be of limited use. However, if we
write Equation (3.6) on page 36 in the form A(t, x) = B(t, x) we can consider this
as a specification of A. But we know from Helmholtz theorem that in order to
determine the (spatial) behaviour of A completely, we must also specify A. Since
this divergence does not enter the derivation above, we are free to choose A in
whatever way we like and still obtain the same physical results!
A+
(3.13)
the coupled inhomegeneous wave Equation (3.12) on page 37 simplify into the following set of uncoupled inhomogeneous wave equations:
1 2
1 2
(t, x)
2
= 2 2 2 =
2
2
c t
c t
0
2
2
1
1 A
def
2 A = 2 2 2 A = 0 j(t, x)
2 A
c2 t2
c t
def
2
(3.14a)
(3.14b)
where 2 is the dAlembert operator discussed in Example 13.5 on page 175. Each
of these four scalar equations is an inhomogeneous wave equation of the following
generic form:
2 (t, x) = f (t, x)
(3.15)
where is a shorthand for either or one of the components Ai of the vector potential
A, and f is the pertinent generic source component, (t, x)/ 0 or 0 ji (t, x), respectively.
1 In fact, the Dutch physicist Hendrik Antoon Lorentz, who in 1903 demonstrated the covariance of
Maxwells equations, was not the original discoverer of this condition. It had been discovered by the Danish
physicist Ludvig V. Lorenz already in 1867 [5]. In the literature, this fact has sometimes been overlooked
and the condition was earlier referred to as the Lorentz gauge condition.
38
We assume that our sources are wellbehaved enough in time t so that the Fourier
transform pair for the generic source function
def
F 1 [ f (x)] f (t, x) =
def
1
2
d f (x) eit
(3.16a)
dt f (t, x) eit
(3.16b)
exists, and that the same is true for the generic potential component:
(t, x) =
(x) =
1
2
d (x) eit
(3.17a)
dt (t, x) eit
(3.17b)
Inserting the Fourier representations (3.16a) and (3.17a) into Equation (3.15) on the
facing page, and using the vacuum dispersion relation for electromagnetic waves
= ck
(3.18)
the generic 3D inhomogeneous wave equation, Equation (3.15) on the preceding page,
turns into
2 (x) + k2 (x) = f (x)
(3.19)
(3.20)
and the solution of Equation (3.19) which corresponds to the frequency is given by
the superposition
(x) =
V0
(3.21)
39
3. E LECTROMAGNETIC P OTENTIALS
polar coordinate system, and recall the expression for the Laplace operator in such a
coordinate system, Equation (3.20) on the preceding page becomes
d2
(rG) + k2 (rG) = r(r)
dr2
(3.22)
Away from r = x x0  = 0, i.e., away from the source point x0 , this equation takes the
form
d2
(rG) + k2 (rG) = 0
dr2
(3.23)
+e
ikr
+C
ikr
ikxx 
eikxx 
e
+
C
C
x x0 
x x0 
+
(3.24)
x x 0 0
1
1
+ C
,
x x0 
x x0 
(3.25)
The volume integrated Equation (3.20) on the previous page can under this assumption
be approximated by
+
C +C
Z 3 0
1
1
2
+
+
k
C
+
C
dx
dx
0
0
0
x x 
x x0 
V
V
Z
= d3x0 (x x0 )
Z
3 0
(3.26)
V0
In virtue of the fact that the volume element d3x0 in spherical polar coordinates is
proportional to x x0 2 , the second integral vanishes when x x0  0. Furthermore,
from Equation (F.73) on page 161, we find that the integrand in the first integral can
be written as 4(x x0 ) and, hence, that
C+ + C =
1
4
(3.27)
Insertion of the general solution Equation (3.24) into Equation (3.21) on the previous page gives
(x) = C
40
eikxx 
+ C
d x f (x )
x x0 
V0
3 0
eikxx 
d x f (x )
x x0 
V0
3 0
(3.28)
The inverse Fourier transform of this back to the t domain is obtained by inserting the
above expression for (x) into Equation (3.17a) on page 39:
i
h
0

Z
Z
exp i t kxx
(t, x) = C + d3x0
d f (x0 )
x x0 
V0
h
i
(3.29)
0

Z
Z
exp i t + kxx
d f (x0 )
+ C d3x0
x x0 
V0
0
0
If we introduce the retarded time tret
and the advanced time tadv
in the following way
[using the fact that in vacuum k/ = 1/c, according to Equation (3.18) on page 39]:
k x x0 
x x0 
0
0
=t
tret
= tret
(t, x x0 ) = t
c
k x x0 
x x0 
0
0
0
tadv = tadv (t, x x ) = t +
=t+
c
and use Equation (3.16a) on page 39, we obtain
(t, x) = C +
V0
d3x0
0
f (tret
, x0 )
+ C
x x0 
V0
d3x0
0
f (tadv
, x0 )
x x0 
(3.30a)
(3.30b)
(3.31)
This is a solution to the generic inhomogeneous wave equation for the potential components Equation (3.15) on page 38. We note that the solution at time t at the field
point x is dependent on the behaviour at other times t 0 of the source at x0 and that
both retarded and advanced t 0 are mathematically acceptable solutions. However, if
we assume that causality requires that the potential at (t, x) is set up by the source at
0
an earlier time, i.e., at (tret
, x0 ), we must in Equation (3.31) set C = 0 and therefore,
according to Equation (3.27) on the preceding page, C + = 1/(4).2
0
1
(tret
, x0 )
d3x0
40 V 0
x x0 
Z
0
0
j(t , x0 )
A(t, x) =
d3x0 ret 0
4 V 0
x x 
(t, x) =
(3.32a)
(3.32b)
Since these retarded potentials were obtained as solutions to the LorenzLorentz equations (3.14) on page 38 they are valid in the LorenzLorentz gauge but may be gauge
transformed according to the scheme described in subsection 3.3.3 on the next page.
As they stand, we shall use them frequently in the following.
2 In fact, inspired by a discussion by Paul A. M. Dirac, John A. Wheeler and Richard P. Feynman
derived in 1945 a fully selfconsistent electrodynamics using both the retarded and the advanced potentials
[7]; see also [3].
41
3. E LECTROMAGNETIC P OTENTIALS
2 A
(t, x)
0
(3.33)
1 2 A
1
= 0 j(t, x) + 2
c2 t2
c t
(3.34)
The first of these two is the timedependent Poissons equation which, in analogy with
Equation (3.3) on page 35, has the solution
(t, x) =
1
40
V0
d3x0
(t, x0 )
+
x x0 
(3.35)
where has vanishing gradient. We note that in the scalar potential expression the
charge density source is evaluated at time t. The retardation (and advancement) effects
occur only in the vector potential, which is the solution of the inhomogeneous wave
equation
2 A
0
1 2 A
= 0 j +
c2 t2
4 t
V0
d3x0
(t, x0 )
x x0 
(3.36)
42
(3.37a)
(3.37b)
where (t, x) is an arbitrary, differentiable scalar function called the gauge function,
and insert the transformed potentials into Equation (3.10) on page 37 for the electric
field and into Equation (3.6) on page 36 for the magnetic field, we obtain the transformed fields
() A ()
A
A0
=
+
=
t
t
t
t
t
0
0
B = A = A () = A
E0 = 0
(3.38a)
(3.38b)
where, once again Equation (F.62) on page 160 was used. We see that the fields are
unaffected by the gauge transformation (3.37). A transformation of the potentials
and A which leaves the fields, and hence Maxwells equations, invariant is called a
gauge transformation. A physical law which does not change under a gauge transformation is said to be gauge invariant. By definition, the fields themselves are, of
course, gauge invariant.
The potentials (t, x) and A(t, x) calculated from (3.11a) on page 37, with an arbitrary choice of A, can be further gauge transformed according to (3.37) on the
preceding page. If, in particular, we choose A according to the LorenzLorentz
condition, Equation (3.13) on page 38, and apply the gauge transformation (3.37) on
the resulting LorenzLorentz potential equations (3.14) on page 38, these equations
will be transformed into
1 2
(t, x)
1 2
2
2
+
=
(3.39a)
2
2
2
2
c t
t c t
0
1 2
1 2 A
2
2
= 0 j(t, x)
(3.39b)
c2 t2
c2 t2
We notice that if we require that the gauge function (t, x) itself be restricted to fulfil
the wave equation
1 2
2 = 0
c2 t2
(3.40)
these transformed LorenzLorentz equations will keep their original form. The set of
potentials which have been gauge transformed according to Equation (3.37) on the
facing page with a gauge function (t, x) restricted to fulfil Equation (3.40) above,
or, in other words, those gauge transformed potentials for which the LorenzLorentz
equations (3.14) are invariant, comprises the LorenzLorentz gauge.
The process of choosing a particular gauge condition is referred to as gauge fixing.
BE LECTROMAGNETODYNAMIC POTENTIALS
E XAMPLE 3.1
43
3. E LECTROMAGNETIC P OTENTIALS
e
0
E = 0 jm
(3.41a)
B
t
(3.41b)
B = 0 m
(3.41c)
E
B = 0 j e + 0 0
t
(3.41d)
In this theory, one derives the inhomogeneous wave equations for the usual electric scalar
and vector potentials (e , Ae ) and their magnetic counterparts (m , Am ) by assuming that the
potentials are related to the fields in the following symmetrised form:
e
A (t, x) Am
(3.42a)
t
1
1
(3.42b)
B = 2 m (t, x) 2 Am (t, x) + Ae
c
c t
In the absence of magnetic charges, or, equivalenty for m 0 and Am 0, these formulae
reduce to the usual Maxwell theory Formula (3.10) on page 37 and Formula (3.6) on page 36,
respectively, as they should.
E = e (t, x)
Inserting the symmetrised expressions (3.42) above into Equations (3.41), one obtains [cf.,
Equations (3.11a) on page 37]
e (t, x)
( Ae ) =
t
0
m (t, x)
m
2 m
+ ( A ) =
t
0
1 2 A e
1 e
2 e
e
A + A + 2
= 0 je (t, x)
c2 t2
c t
1 2 A m
1 m
2 m
m
= 0 jm (t, x)
A
+
A
+
c2 t2
c2 t
2 e +
(3.43a)
(3.43b)
(3.43c)
(3.43d)
By choosing the conditions on the vector potentials according to the LorenzLorentz prescripton
[cf., Equation (3.13) on page 38]
1 e
=0
c2 t
1
A m + 2 m = 0
c t
these coupled wave equations simplify to
Ae +
1 2 e
2 e
c2 t2
1 2 A e
2 Ae
c2 t2
1 2 m
2 m
c2 t2
1 2 A m
2 Am
c2 t2
44
(3.44)
(3.45)
e (t, x)
0
(3.46a)
= 0 je (t, x)
(3.46b)
m (t, x)
0
(3.46c)
= 0 jm (t, x)
(3.46d)
Bibliography
exhibiting once again, the striking properties of Diracs symmetrised Maxwell theory.
C E ND OF EXAMPLE 3.1
3.4 Bibliography
[1] L. D. FADEEV AND A. A. S LAVNOV, Gauge Fields: Introduction to Quantum Theory,
No. 50 in Frontiers in Physics: A Lecture Note and Reprint Series. Benjamin/Cummings
Publishing Company, Inc., Reading, MA . . . , 1980, ISBN 0805390162.
[2] M. G UIDRY, Gauge Field Theories: An Introduction with Applications, John Wiley
& Sons, Inc., New York, NY . . . , 1991, ISBN 0471631175.
[3] F. H OYLE , S IR AND J. V. NARLIKAR, Lectures on Cosmology and Action at a Distance
Electrodynamics, World Scientific Publishing Co. Pte. Ltd, Singapore, New Jersey, London
and Hong Kong, 1996, ISBN 98100225733(pbk).
[4] J. D. JACKSON, Classical Electrodynamics, third ed., John Wiley & Sons, Inc.,
New York, NY . . . , 1999, ISBN 047130932X.
[5] L. L ORENZ, Philosophical Magazine (1867), pp. 287301.
[6] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism, second ed.,
AddisonWesley Publishing Company, Inc., Reading, MA . . . , 1962, ISBN 0201057026.
[7] J. A. W HEELER AND R. P. F EYNMAN, Interaction with the absorber as a mechanism for
radiation, Reviews of Modern Physics, 17 (1945), pp. 157.
45
CHAPTER 4
Relativistic Electrodynamics
We saw in Chapter 3 how the derivation of the electrodynamic potentials led, in a
most natural way, to the introduction of a characteristic, finite speed of propagation
in vacuum that equals the speed of light c = 1/ 0 0 and which can be considered
as a constant of nature. To take this finite speed of propagation of information into
account, and to ensure that our laws of physics be independent of any specific coordinate frame, requires a treatment of electrodynamics in a relativistically covariant
(coordinate independent) form. This is the object of this chapter.
The theory of relativity is not merely a scientific development of great importance in its own
right. It is even more significant as the first stage of a radical change in our basic concepts,
which began in physics, and which is spreading into other fields of science, and indeed,
even into a great deal of thinking outside of science. For as is well known, the modern trend
is away from the notion of sure absolute truth, (i.e., one which holds independently of all
conditions, contexts, degrees, and types of approximation etc..) and toward the idea that a
given concept has significance only in relation to suitable broader forms of reference, within
which that concept can be given its full meaning.
47
4. R ELATIVISTIC E LECTRODYNAMICS
vt
y
y0
v
P(t, x, y, z)
P(t0 , x0 , y0 , z0 )
O0
x0
z0
how physical processes are interrelated when observed in different inertial systems in
uniform, rectilinear motion relative to each other and is based on two postulates:
Postulate 4.1 (Relativity principle; Poincar, 1905). All laws of physics (except the
laws of gravitation) are independent of the uniform translational motion of the system
on which they operate.
Postulate 4.2 (Einstein, 1905). The velocity of light in empty space is independent of
the motion of the source that emits the light.
A consequence of the first postulate is that all geometrical objects (vectors, tensors)
in an equation describing a physical process must transform in a covariant manner,
i.e., in the same way.
48
v
c
= p
(4.1)
1
(4.2)
1 2
where v = v. In the following, we shall make frequent use of these shorthand notations.
As shown by Einstein, the two postulates of special relativity require that the spatial coordinates and times as measured by an observer in and 0 , respectively, are
connected by the following transformation:
ct0 = (ct x)
(4.3a)
(4.3b)
(4.3c)
(4.3d)
x = (x vt)
y =y
z =z
Taking the difference between the square of (4.3a) and the square of (4.3b) we find
that
c2 t02 x02 = 2 c2 t2 2xct + x2 2 x2 + 2xvt v2 t2
v2
v2
1
2 2
2
c
t
1
x
1
=
(4.4)
c2
c2
v2
1 2
c
= c2 t 2 x2
From Equations (4.3) we see that the y and z coordinates are unaffected by the translational motion of the inertial system 0 along the x axis of system . Using this fact,
we find that we can generalise the result in Equation (4.4) above to
c2 t2 x2 y2 z2 = c2 t02 x02 y02 z02
(4.5)
which means that if a light wave is transmitted from the coinciding origins O and O 0
at time t = t0 = 0 it will arrive at an observer at (x, y, z) at time t in and an observer
at (x0 , y0 , z0 ) at time t0 in 0 in such a way that both observers conclude that the speed
(spatial distance divided by time) of light in vacuum is c. Hence, the speed of light
in and 0 is the same. A linear coordinate transformation which has this property is
called a (homogeneous) Lorentz transformation.
49
4. R ELATIVISTIC E LECTRODYNAMICS
and t is time), and the remaining components are the components of the ordinary R 3
radius vector x defined in Equation (M.1) on page 164:
x = (x0 , x1 , x2 , x3 ) = (ct, x, y, z) (ct, x)
(4.6)
We want to interpret this quadruple x as (the component form of) a radius fourvector
in a real, linear, fourdimensional vector space.2 We require that this fourdimensional
space be a Riemannian space, i.e., a metric space where a distance and a scalar
product are defined. In this space we therefore define a metric tensor, also known as
the fundamental tensor, which we denote by g .
(4.7)
where the upper index in x is summed over and is therefore a dummy index and
may be replaced by another dummy index This summation process is an example of
index contraction and is often referred to as index lowering.
(4.8)
(4.9)
2 The British mathematician and philosopher Alfred North Whitehead writes in his book The Concept
of Nature [13]:
I regret that it has been necessary for me in this lecture to administer a large dose of fourdimensional geometry. I do not apologise, because I am really not responsible for the fact
that nature in its most fundamental aspect is fourdimensional. Things are what they are. . . .
50
We notice that our space will have an indefinite norm which means that we deal with
a nonEuclidean space. We call the fourdimensional space (or spacetime) with this
property Lorentz space and denote it L4 . A corresponding real, linear 4D space with
a positive definite norm which is conserved during ordinary rotations is a Euclidean
vector space. We denote such a space R4 .
Metric tensor
By choosing the metric tensor in L4 as
if = = 0
1
g = 1 if = = i = j = 1, 2, 3
0
if ,
1 0
0 1
(g ) =
0 0
0 0
(4.10)
0
0
0
0
1 0
0 1
(4.11)
i.e., a matrix with a main diagonal that has the sign sequence, or signature, {+, , , },
the index lowering operation in our chosen flat 4D space becomes nearly trivial:
x = g x = (ct, x)
(4.12)
0
0
1
0
0 0
0
x
x
x1 x1
0
=
0 x2 x2
1
x3
x3
(4.13)
Hence, if the metric tensor is defined according to expression (4.10) above the covariant radius fourvector x is obtained from the contravariant radius fourvector x
simply by changing the sign of the last three components. These components are referred to as the space components; the zeroth component is referred to as the time
component.
As we see, for this particular choice of metric, the scalar product of x with itself
becomes
x x = (ct, x) (ct, x) = c2 t2 x2 y2 z2
(4.14)
which indeed is the desired Lorentz transformation invariance as required by Equation (4.9) on the facing page. Without changing the physics, one can alternatively
51
4. R ELATIVISTIC E LECTRODYNAMICS
choose a signature {, +, +, +}. The latter has the advantage that the transition from
3D to 4D becomes smooth, while it will introduce some annoying minus signs in the
theory. In current physics literature, the signature {+, , , } seems to be the most
commonly used one.
The L4 metric tensor Equation (4.10) on the previous page has a number of interesting properties: firstly, we see that this tensor has a trace Tr g = 2 whereas in
R4 , as in any vector space with definite norm, the trace equals the space dimensionality. Secondly, we find, after trivial algebra, that the following relations between the
contravariant, covariant and mixed forms of the metric tensor hold:
g = g
g
g g
(4.15a)
= g
=
g g =
g
g
(4.15b)
=
=
(4.15c)
(4.15d)
Here we have introduced the 4D version of the Kronecker delta , a mixed fourtensor
of rank 2 which fulfils
(
1 if =
= =
(4.16)
0 if ,
(4.17)
where the metric tensor is as in Equation (4.10) on the preceding page. As we see,
this form is indefinite as expected for a nonEuclidean space. The square root of this
expression is the invariant line element
v
"
u
2 2 2 3 2 #
u
dx
dx
1
dx 1
t
ds = c dt 1 2
+
+
c
dt
dt
dt
r
r
(4.18)
1 2
v2
2
2
= c dt 1 2 (v x ) + (vy ) + (vz ) = c dt 1 2
c
c
p
dt
2
= c dt 1 = c = c d
where we introduced
d = dt/
52
(4.19)
Since d measures the time when no spatial changes are present, it is called the proper
time.
Expressing the property of the Lorentz transformation described by Equations (4.5)
on page 49 in terms of the differential interval ds and comparing with Equation (4.17)
on the facing page, we find that
ds2 = c2 dt2 dx2 dy2 dz2
(4.20)
is invariant during a Lorentz transformation. Conversely, we may say that every coordinate transformation which preserves this differential interval is a Lorentz transformation.
If in some inertial system
dx2 + dy2 + dz2 < c2 dt2
(4.21)
(4.22)
(4.23)
is a lightlike interval; we may also say that in this case we are on the light cone. A
vector which has a lightlike interval is called a null vector. The timelike, spacelike
or lightlike aspects of an interval ds are invariant under a Lorentz transformation. I.e.,
it is not possible to change a timelike interval into a spacelike one or vice versa via
a Lorentz transformation.
Fourvector fields
Any quantity which relative to any coordinate system has a quadruple of real numbers and transforms in the same way as the radius fourvector x does, is called a
fourvector. In analogy with the notation for the radius fourvector we introduce the
notation a = (a0 , a) for a general contravariant fourvector field in L4 and find that
the lowering of index rule, Formula (4.7) on page 50, for such an arbitrary fourvector yields the dual covariant fourvector field
a (x ) = g a (x ) = (a0 (x ), a(x ))
(4.24)
The scalar product between this fourvector field and another one b (x ) is
g a (x )b (x ) = (a0 , a) (b0 , b) = a0 b0 a b
(4.25)
which is a scalar field, i.e., an invariant scalar quantity (x ) which depends on time
and space, as described by x = (ct, x, y, z).
53
4. R ELATIVISTIC E LECTRODYNAMICS
0 0
0 0
=
0
0
1 0
0
(4.26)
0 1
the linear Lorentz transformation (4.3) on page 49, i.e., the coordinate transformation
x x0 = x0 (x0 , x1 , x2 , x3 ), from one inertial system to another inertial system 0
in the standard configuration, can be written
x0 = x
(4.27)
1 + 2
1 + 1 2
(4.28)
This means that the nonempty set of Lorentz transformations constitutes a closed algebraic structure with a binary operation which is associative. Furthermore, one can
show that this set possesses at least one identity element and at least one inverse element. In other words, this set of Lorentz transformations constitutes a mathematical
group. However tempting, we shall not make any further use of group theory.
(4.29b)
(4.29c)
(4.29d)
X =x
X =x
X =x
dS = ids
54
(4.29a)
(4.29e)
X0
X
00
x01
x1
F IGURE 4.2: Minkowski space can be considered an ordinary Euclidean space
where a Lorentz transformation from (x1 , X 0 = ict) to (x01 , X 00 = ict0 ) corresponds
to an ordinary rotation through an angle . This rotation leaves the Euclidean
2
2
distance x1 + X 0 = x2 c2 t2 invariant.
where i =
(4.30)
i.e., into a 4D differential form which is positive definite just as is ordinary 3D Euclidean space R3 . We shall call the 4D Euclidean space constructed in this way the
Minkowski space M4 .3
As before, it suffices to consider the simplified case where the relative motion
between and 0 is along the x axes. Then
dS 2 = (dX 0 )2 + (dX 1 )2 = (dX 0 )2 + (dx1 )2
(4.31)
(4.32a)
x = x cos + X sin
(4.32b)
(4.33a)
x = x cosh ct sinh
(4.33b)
3 The fact that our Riemannian space can be transformed in this way into an Euclidean one means that
it is, strictly speaking, a pseudoRiemannian space.
55
4. R ELATIVISTIC E LECTRODYNAMICS
x = ct
x00
x0 = x 1
P0
ct
O=O
x01
x1 = x
(4.34a)
cosh =
(4.34b)
tanh =
(4.34c)
tanh 1 + tanh 2
1 + tanh 1 tanh 2
(4.35)
The use of ict and M4 , which leads to the interpretation of the Lorentz transformation as an ordinary rotation, may, at best, be illustrative, but is not very physical.
Besides, if we leave the flat L4 space and enter the curved space of general relativity,
the ict trick will turn out to be an impasse. Let us therefore immediately return to L 4
where all components are real valued.
56
dx
c
v
= (u0 , u)
u =
= (c, v) = q
, q
(4.36)
d
v2
v2
1
1
c2
c2
which, when multiplied with the scalar invariant m0 yields the fourmomentum
m0 v
m0 c
dx
, q
= (p0 , p)
(4.37)
= m0 (c, v) = q
p = m0
d
v2
v2
1
1
c2
c2
(4.38)
where
m = m0 = q
m0
1
(4.39)
v2
c2
We can interpret this such that the Lorentz covariance implies that the masslike term
in the ordinary 3D linear momentum is not invariant. A better way to look at this is that
p = mv = m0 v is the covariantly correct expression for the kinetic threemomentum.
Multiplying the zeroth (time) component of the fourmomentum p with the scalar
invariant c, we obtain
m0 c2
= mc2
cp0 = m0 c2 = q
v2
1 c2
(4.40)
Since this component has the dimension of energy and is the result of a covariant
description of the motion of a particle with its kinetic momentum described by the
spatial components of the fourmomentum, Equation (4.37) above, we interpret cp0
as the total energy E. Hence,
cp = (cp0 , cp) = (E, cp)
(4.41)
(4.42)
57
4. R ELATIVISTIC E LECTRODYNAMICS
Since this is an invariant, this equation holds in any inertial frame, particularly in the
frame where p = 0 and there we have
E = m 0 c2
(4.43)
dx
= 0 u = 0 (c, v) = (c, v)
d
(4.44)
where we introduced
= 0
(4.45)
1 2
2
c2 t2
(4.46)
,A
A =
c
(4.47)
where is the scalar potential and A the vector potential, defined in Section 3.3 on
page 36, we can write the uncoupled inhomogeneous wave equations, Equations (3.14)
on page 38, in the following compact (and covariant) way:
2 A = 0 j
58
(4.48)
With the help of the above, we can formulate our electrodynamic equations covariantly. For instance, the covariant form of the equation of continuity, Equation (1.23)
on page 9 is
j = 0
(4.49)
and the LorenzLorentz gauge condition, Equation (3.13) on page 38, can be written
A = 0
(4.50)
(4.51)
If only one dimension Lorentz contracts (for instance, due to relative motion along
the x direction), a 3D spatial volume element transforms according to
r
p
1
v2
3
dV = d x = dV0 = dV0 1 2 = dV0 1 2
(4.52)
where dV0 denotes the volume element as measured in the rest system, then from
Equation (4.45) on the facing page we see that
dV = 0 dV0
(4.53)
i.e., the charge in a given volume is conserved. We can therefore conclude that the
elementary charge is a universal constant.
(A )0 =
,A
=
,0
(4.54)
c
40 c x x0 0
v=0
where x x0 0 is the usual distance from the source point to the field point, evaluated
in the rest system (signified by the index 0).
Let us introduce the relative radius fourvector between the source point and the
field point:
R = x x0 = (c(t t0 ), x x0 )
(4.55)
59
4. R ELATIVISTIC E LECTRODYNAMICS
(4.56)
We know that in vacuum the signal (field) from the charge q0 at x0 propagates to
x with the speed of light c so that
x x0 = c(t t0 )
(4.57)
(4.58)
(4.59)
Now we want to find the correspondence to the rest system solution, Equation (4.54)
on the preceding page, in an arbitrary inertial system. We note from Equation (4.36)
on page 57 that in the rest system
v
c
= (c, 0)
(u )0 = q
(4.60)
, q
v2
v2
1 c2
1 c2
v=0
and
(R )0 = (x x0 , x x0 )0 = (x x0 0 , (x x0 )0 )
(4.61)
q0
u
40 cu R
(4.63)
60
= v
v
c
(4.65)
and introducing
v (x x0 )
def
x x0 (x x0 )
s x x 0
c
(4.66)
we can write
u R = cs
(4.67)
and
u
=
cu R
1 v
,
cs c2 s
(4.68)
q0
A (x ) =
=
,
,A
40 cs c2 s
c
(4.69)
where in the last step the definition of the fourpotential, Equation (4.47) on page 58,
was used. Writing the solution in the ordinary 3Dway, we conclude that for a very
localised charge volume, moving relative an observer with a velocity v, the scalar and
vector potentials are given by the expressions
q0 1
q0
1
=
40 s 40 x x0  (x x0 )
q0
q0 v
v
=
A(t, x) =
2
2
0
40 c s 40 c x x  (x x0 )
(t, x) =
(4.70a)
(4.70b)
i, j , k
(4.72)
61
4. R ELATIVISTIC E LECTRODYNAMICS
The same is true for the curl operator . For instance, the Maxwell equation
E=
B
t
(4.73)
(4.74)
We know from Chapter 3 that the fields can be derived from the electromagnetic
potentials in the following way:
B=A
E =
(4.75a)
A
t
(4.75b)
Bi j =
(4.76a)
(4.76b)
From this, we notice the clear difference between the axial vector (pseudovector) B
and the polar vector (ordinary vector) E.
Our goal is to express the electric and magnetic fields in a tensor form where the
components are functions of the covariant form of the fourpotential, Equation (4.47)
on page 58:
A =
,A
(4.77)
c
Inspection of (4.77) and Equation (4.76) makes it natural to define the fourtensor
F =
A A
= A A
x
x
(4.78)
This antisymmetric (skewsymmetric), fourtensor of rank 2 is called the electromagnetic field tensor. In matrix representation, the contravariant field tensor can be written
0
E x /c E y /c E z /c
E x /c
0
Bz
By
(F ) =
(4.79)
E y /c
Bz
0
B x
E z /c
By
Bx
The covariant field tensor is obtained from the contravariant field tensor in the usual
manner by index lowering:
F = g g F = A A
62
(4.80)
In matrix representation
0
E x /c
F =
E y /c
E z /c
E x /c
0
Bz
By
E y /c
Bz
0
Bx
E z /c
By
B x
(4.81)
It is perhaps interesting to note that the field tensor is a sort of fourdimensional curl
of the fourpotential vector A .
That the two Maxwell source equations can be written
F = 0 j
(4.82)
(4.84)
which is the Maxwell source equation for the electric field, Equation (1.45a) on page 14.
For = 1, Equation (4.83) above yields
F 01 F 11 F 21 F 31
1 E x
Bz By
+
+
+
= 2
+0
+
= 0 j1 = 0 v x
x0
x1
x2
x3
c t
y
z
(4.85)
or, using 0 0 = 1/c2 ,
By Bz
E x
0 0
= 0 j x
z
y
t
(4.86)
and similarly for = 2, 3. In summary, in threevector form, we can write the result
as
B 0 0
E
= 0 j(t, x)
t
(4.87)
which is the Maxwell source equation for the magnetic field, Equation (1.45d) on
page 14.
63
4. R ELATIVISTIC E LECTRODYNAMICS
B
t
B=0
(4.88)
(4.89)
(4.90)
Hence, Equation (4.82) on the previous page and Equation (4.90) constitute Maxwells
equations in fourdimensional formalism.
4.4 Bibliography
[1]
J. A HARONI, The Special Theory of Relativity, second, revised ed., Dover Publications, Inc., New York, 1985, ISBN 0486648702.
[2]
A. O. BARUT, Electrodynamics and Classical Theory of Fields and Particles, Dover Publications, Inc., New York, NY, 1980, ISBN 0486640388.
[3]
[4]
D. B OHM, The Special Theory of Relativity, Routledge, New York, NY, 1996, ISBN 041514809X.
[5]
[6]
L. D. L ANDAU AND E. M. L IFSHITZ, The Classical Theory of Fields, fourth revised English ed., vol. 2 of Course of Theoretical Physics, Pergamon Press, Ltd., Oxford . . . , 1975,
ISBN 0080250726.
[7]
F. E. L OW, Classical Field Theory, John Wiley & Sons, Inc., New York, NY . . . , 1997,
ISBN 0471595519.
[8]
H. M UIRHEAD, The Special Theory of Relativity, The Macmillan Press Ltd., London,
Beccles and Colchester, 1973, ISBN 333128451.
[9]
C. M LLER, The Theory of Relativity, second ed., Oxford University Press, Glasgow . . . ,
1972.
[10] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism, second ed.,
AddisonWesley Publishing Company, Inc., Reading, MA . . . , 1962, ISBN 0201057026.
64
Bibliography
[11] J. J. S AKURAI, Advanced Quantum Mechanics, AddisonWesley Publishing Company, Inc., Reading, MA . . . , 1967, ISBN 0201067102.
[12] B. S PAIN, Tensor Calculus, third ed., Oliver and Boyd, Ltd., Edinburgh and London,
1965, ISBN 050013319.
[13] A. N. W HITEHEAD, Concept of Nature, Cambridge University Press, Cambridge . . . ,
1920, ISBN 0521092450.
65
CHAPTER 5
67
Lagrange formalism
Let us now introduce a generalised action
S4 =
L4 (x , u ) d
(5.1)
where d is the proper time defined via Equation (4.18) on page 52, and L4 acts as a
kind of generalisation to the common 3D Lagrangian so that the variational principle
S 4 =
1
0
L4 (x , u ) d = 0
(5.2)
1
m0 u u
2
(5.3)
For an interaction with the electromagnetic field we can introduce the interaction with
the help of the fourpotential given by Equation (4.77) on page 62 in the following
way
L4 =
1
m0 u u + qu A (x )
2
(5.4)
We call this the fourLagrangian and shall now show how this function, together with
the variation principle, Formula (5.2), yields covariant results which are physically
correct.
The variation principle (5.2) with the 4D Lagrangian (5.4) inserted, leads to
Z
m
u u + qu A d
2
0
Z 1
m0 (u u )
u + q A u + u
x
d
=
2 u
x
0
Z 1
=
m0 u u + q A u + u A x d = 0
S 4 =
(5.5)
68
dx
d
(5.6)
which means that we can write the variation of u as a total derivative with respect to
:
dx
d
(x )
u =
=
(5.7)
d
d
Inserting this into the first two terms in the last integral in Equation (5.5) on the facing
page, we obtain
Z 1
d
d
S 4 =
(5.8)
m0 u (x ) + qA (x ) + qu A x d
d
d
0
Partial integration in the two first terms in the right hand member of (5.8) gives
Z 1
dA
du
S 4 =
x q
x + qu A x d
(5.9)
m0
d
d
0
where the integrated parts do not contribute since the variations at the endpoints vanish. A change of irrelevant summation index from to in the first two terms of
the right hand member of (5.9) yields, after moving the ensuing common factor x
outside the partenthesis, the following expression:
Z 1
dA
du
q
+ qu A x d
(5.10)
S 4 =
m0
d
d
0
Applying wellknown rules of differentiation and the expression (4.36) for the
fourvelocity, we can express dA /d as follows:
dA A dx
=
= A u
(5.11)
d
x d
By inserting this expression (5.11) into the second term in righthand member of Equation (5.10) above, and noting the common factor qu of the resulting term and the last
term, we obtain the final variational principle expression
Z 1
du
S 4 =
(5.12)
+ qu A A x d
m0
d
0
m0
du
= qu F
d
(5.14)
69
Hamiltonian formalism
The usual Hamilton equations for a 3D space are given by Equation (M.103) on
page 180 in Appendix M. These six firstorder partial differential equations are
H dqi
=
pi
dt
H
dpi
=
qi
dt
(5.15a)
(5.15b)
L4
u
(5.16)
and utilise the fourvelocity u , as given by Equation (4.36) on page 57, to define the
fourHamiltonian
H 4 = p u L4
(5.17)
With the help of these, the radius fourvector x , considered as the generalised fourcoordinate, and the invariant line element ds, defined in Equation (4.18) on page 52,
we introduce the following eight partial differential equations:
H4 dx
=
p
d
dp
H4
=
x
d
(5.18a)
(5.18b)
70
that this cp0 solves the Hamilton equations, Equations (5.15) on the preceding page.
This later consistency check is left as an exercise to the reader.
Using the definition of H4 , Equation (5.17) on the facing page, and the expression
for L4 , Equation (5.4) on page 68, we obtain
1
(5.19)
H4 = p u L4 = p u m0 u u qu A (x )
2
Furthermore, from the definition (5.16) of the canonically conjugate fourmomentum
p , we see that
L4
1
=
m0 u u + qu A (x ) = m0 u + qA
(5.20)
p =
u
u 2
Inserting this into (5.19), we obtain
1
1
H4 = m0 u u + qA u m0 u u qu A (x ) = m0 u u
2
2
(5.21)
u =
1
(p qA )
m0
(5.22)
(5.23)
1
(p qA )
p qA
H4 =
2 m0
m0
1
(p qA ) p qA
=
2m0
1
p p 2qA p + q2 A A
=
2m0
(5.24)
That this fourHamiltonian yields the correct covariant equation of motion can be
seen by inserting it into the fourdimensional Hamiltons equations (5.18) and using
the relation (5.23):
H4
q
A
= (p qA )
x
m0
x
q
A
= m0 u
m0
x
A
= qu
x
du
A
dp
= m0
q u
=
d
d
x
(5.25)
71
where in the last step Equation (5.20) on the previous page was used. Rearranging
terms, and using Equation (4.79) on page 62, we obtain
m0
du
= qu A A = qu F
d
(5.26)
which is identical to the covariant equation of motion Equation (5.14) on page 69. We
can then safely conclude that the Hamiltonian in question is correct.
Recalling expression (4.47) on page 58 and representing the canonically conjugate
fourmomentum as p = (p0 , p), we obtain the following scalar products:
p p = (p0 )2 (p)2
1
A p = p0 (p A)
c
1
A A = 2 2 (A)2
c
(5.27a)
(5.27b)
(5.27c)
Inserting these explicit expressions into Equation (5.24) on the previous page, and
using the fact that for H4 is equal to the scalar value m0 c2 /2, as derived in Equation (5.22) on the preceding page, we obtain the equation
1
q2 2
m0 c2
2
0
2
2
0 2
2
=
(5.28)
(p ) (p) qp + 2q(p A) + 2 q (A)
2
2m0
c
c
which is the second order algebraic equation in p0 :
(p0 )2
q2
2q 0 2
p (p) 2qp A + q2 (A)2 + 2 2 m20 c2 = 0
c

{z
} c
(5.29)
(pqA)2
(5.30)
Since the zeroth component (time component) p0 of a fourmomentum vector p multiplied by c represents the energy [cf. Equation (4.41) on page 57], the positive solution in Equation (5.30) must be identified with the ordinary Hamilton function H
divided by c. Consequently,
q
(5.31)
H cp0 = q + c (p qA)2 + m20 c2
is the ordinary 3D Hamilton function for a charged particle moving in scalar and
vector potentials associated with prescribed electric and magnetic fields.
The ordinary Lagrange and Hamilton functions L and H are related to each other
by the 3D transformation [cf. the 4D transformation (5.17) between L4 and H4 ]
L=pvH
72
(5.32)
Using the explicit expressions (Equation (5.31) on the preceding page) and (Equation (5.32) on the facing page), we obtain the explicit expression for the ordinary 3D
Lagrange function
q
L = p v q c (p qA)2 + m20 c2
(5.33)
and if we make the identification
m0 v
p qA = q
= mv
2
1 cv2
(5.34)
where the quantity mv is the usual kinetic momentum, we can rewrite this expression
for the ordinary Lagrangian as follows:
q
L = qA v + mv2 q c m2 v2 + m20 c2
r
(5.35)
v2
2
2
2
1 2
= mv q( A v) mc = q + qA v m0 c
c
What we have obtained is the relativstically correct (covariant) expression for the
Lagrangian describing the motion of a charged particle in scalar and vector potentials
associated with prescribed electric and magnetic fields.
73
i1
m
k
a
k
a
i+1
k
a
k
a
m
x
constants k. At equilibrium the mass points are at rest, distributed evenly with a distance a to their two nearest neighbours. After perturbation, the motion of mass point i
will be a onedimensional oscillatory motion along x . Let us denote the deviation for
mass point i from its equilibrium position by i (t) x .
The solution to this mechanical problem can be obtained if we can find a Lagrangian (Lagrange function) L which satisfies the variational equation
Z
L(i , i , t) dt = 0
(5.36)
L=
1 N 2
m i k(i+1 i )2
2
i=1
(5.37)
Let us write the Lagrangian, as given by Equation (5.37), in the following way:
N
L = aLi
(5.38)
2
1 m 2
i+1
i
Li =
ka
2 a i
a
(5.39)
i=1
Here,
is the so called linear Lagrange density. If we now let N and, at the same time,
74
let the springs become infinitesimally short according to the following scheme:
a dx
m
dm
=
a
dx
ka Y
i+1 i
a
x
we obtain
L=
(5.40b)
Youngs modulus
(5.40c)
(5.40d)
L dx
(5.41)
where
(5.40a)
, , , t
t x
"
2 #
2
1
=
Y
2
t
x
(5.42)
Notice how we made a transition from a discrete description, in which the mass points
were identified by a discrete integer variable i = 1, 2, . . . , N, to a continuous description, where the infinitesimal mass points were instead identified by a continuous real
parameter x, namely their position along x .
A consequence of this transition is that the number of degrees of freedom for the
system went from the finite number N to infinity! Another consequence is that L has
now become dependent also on the partial derivative with respect to x of the field
coordinate . But, as we shall see, the transition is well worth the price because it
allows us to treat all fields, be it classical scalar or vectorial fields, or wave functions,
spinors and other fields that appear in quantum physics, on an equal footing.
Under the assumption of time independence and fixed endpoints, the variation
principle (5.36) on the preceding page yields:
Z
L dt
=
ZZ
, ,
t x
L
dx dt
ZZ
L
L + L
=
+
dx dt
t
x
t
x
(5.43)
=0
The last integral can be integrated by parts. This results in the expression
ZZ
L
dx dt = 0
t
x
t
(5.44)
75
where the variation is arbitrary (and the endpoints fixed). This means that the integrand itself must vanish. If we introduce the functional derivative
L
L
L
=
(5.45)
x
x
L
L
=0
(5.46)
=0
(5.47)
2Y 2 =
t
x
Y t2 x2
i.e., the onedimensional wave equation for compression waves which propagate with
ZZ
L d3x dt
Z
= L , d4x
x
ZZ
L
L
=
d4x
L dt =
(5.48)
=0
where the variation is arbitrary and the endpoints are fixed. This means that the
integrand itself must vanish:
L
L
=0
(5.49)
x
x
L
L
L
=
i
x
i
(5.50)
76
L
= 0
(5.51)
L
(5.52)
H , , i ; t = L , , i
x
t
t x
(5.53)
If, as usual, we differentiate this expression and identify terms, we obtain the following
Hamilton density equations
H
t
H
(5.54a)
(5.54b)
The Hamilton density functions are in many ways similar to the ordinary Hamilton
functions and lead to similar results.
(5.55)
where the mechanical part has to do with the particle motion (kinetic energy). It is
given by L4 /V where L4 is given by Equation (5.3) on page 68 and V is the volume.
Expressed in the rest mass density %0 , the mechanical Lagrange density can be written
L mech =
1
%0 u u
2
(5.56)
77
The L inter part describes the interaction between the charged particles and the
external electromagnetic field. A convenient expression for this interaction Lagrange
density is
L inter = j A
(5.57)
For the field part L field we choose the difference between magnetic and electric
energy density (in analogy with the difference between kinetic and potential energy in
a mechanical field). Using the field tensor, we express this field Lagrange density as
L field =
1
F F
40
(5.58)
1
1
% 0 u u + j A +
F F
2
40
(5.59)
E XAMPLE 5.1
B2
0 E 2
0
(5.60)
i.e., the difference between the magnetic and electric field energy densities.
From Formula (4.79) on page 62 we recall that
0
E /c
x
(F ) =
Ey /c
Ez /c
E x /c
0
Bz
By
Ey /c
Bz
0
Bx
Ez /c
By
B x
0
(5.61)
0
E x /c
F =
Ey /c
Ez /c
E x /c
0
Bz
By
Ey /c
Bz
0
Bx
Ez /c
By
B x
0
(5.62)
where denotes the row number and the column number. Then, Einstein summation and
78
(5.63)
or
B2
1 2
E
0 c 2 0
1
2
B2
0 E 2
0
(5.64)
QED
C E ND OF EXAMPLE 5.1
tot
1
F F
40
(5.65)
inserted into the EulerLagrange equations, expression (5.49) on page 76, yields two
of Maxwells equations. To see this, we note from Equation (5.65) above and the
results in Example 5.1 that
L EM
= j
A
(5.66)
Furthermore,
L EM
1
F F
=
( A )
40
( A )
1
( A A )( A A )
40
( A )
1
=
A A A A
40
( A )
A A + A A
A A A A
=
20
( A )
(5.67)
79
But
A A = A
A + A
A
( A )
( A )
( A )
A + A
g g A
= A
( A )
( A )
A + g g A
A
= A
( A )
( A )
= A
A + A
A
( A )
( A )
(5.68)
= 2 A
Similarly,
A A = 2 A
( A )
(5.69)
so that
1
1
L EM
= ( A A ) = F
( A )
0
0
(5.70)
This means that the EulerLagrange equations, expression (5.49) on page 76, for
the Lagrangian density L EM and with A as the field quantity become
L EM
1
L EM
= j F = 0
A
( A )
0
(5.71)
F = 0 j
(5.72)
or
which, according to Equation (4.82) on page 63, is the covariant version of Maxwells
source equations.
Other fields
In general, the dynamic equations for most any fields, and not only electromagnetic
ones, can be derived from a Lagrangian density together with a variational principle
(the EulerLagrange equations). Both linear and nonlinear fields are studied with this
technique. As a simple example, consider a real, scalar field which has the following
Lagrange density:
L =
80
1
m2 2
2
(5.73)
Bibliography
Insertion into the 1D EulerLagrange equation, Equation (5.46) on page 76, yields the
dynamic equation
(2 m2 ) = 0
(5.74)
emx
x
(5.75)
which describes the Yukawa meson field for a scalar meson with mass m. With
=
1
c2 t
(5.76)
1 2 2
c + ()2 + m2 2
2
(5.77)
1
F F + m2 A A
40
(5.78)
(5.79)
This equation describes an electromagnetic field with a mass, or, in other words,
massive photons. If massive photons would exist, largescale magnetic fields, including those of the earth and galactic spiral arms, would be significantly modified to yield
measurable discrepances from their usual form. Space experiments of this kind onboard satellites have led to stringent upper bounds on the photon mass. If the photon
really has a mass, it will have an impact on electrodynamics as well as on cosmology
and astrophysics.
5.3 Bibliography
[1] A. O. BARUT, Electrodynamics and Classical Theory of Fields and Particles, Dover Publications, Inc., New York, NY, 1980, ISBN 0486640388.
[2] V. L. G INZBURG, Applications of Electrodynamics in Theoretical Physics and Astrophysics, Revised third ed., Gordon and Breach Science Publishers, New York, London, Paris,
Montreux, Tokyo and Melbourne, 1989, ISBN 2881247199.
81
[3] H. G OLDSTEIN, Classical Mechanics, second ed., AddisonWesley Publishing Company, Inc., Reading, MA . . . , 1981, ISBN 0201029189.
[4] W. T. G RANDY, Introduction to Electrodynamics and Radiation, Academic Press,
New York and London, 1970, ISBN 0122952502.
[5] L. D. L ANDAU AND E. M. L IFSHITZ, The Classical Theory of Fields, fourth revised English ed., vol. 2 of Course of Theoretical Physics, Pergamon Press, Ltd., Oxford . . . , 1975,
ISBN 0080250726.
[6] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism, second ed.,
AddisonWesley Publishing Company, Inc., Reading, MA . . . , 1962, ISBN 0201057026.
[7] J. J. S AKURAI, Advanced Quantum Mechanics, AddisonWesley Publishing Company, Inc., Reading, MA . . . , 1967, ISBN 0201067102.
[8] D. E. S OPER, Classical Field Theory, John Wiley & Sons, Inc., New York, London, Sydney
and Toronto, 1976, ISBN 0471813680.
82
CHAPTER 6
V0
d3x0 (x0 )
(6.1)
83
where the is the charge density introduced in Equation (1.7) on page 5, the electric
dipole moment vector
Z
p(x0 ) =
V0
(6.2)
V0
(6.3)
d
x
P(x
)
p (x) =
40 V 0
x x0 3
40 V 0
x x0 
Z
1
1
=
d3x0 P(x0 ) 0
0
40 V
x x0 
(6.5)
Using the expression Equation (M.85) on page 177 and applying the divergence theorem, we can rewrite this expression for the potential as follows:
Z
Z
0
0
P(x0 )
1
3 0 P(x )
d3x0 0
d
x
p (x) =
40 V 0
x x0 
x x0 
V0
I
(6.6)
Z
0
0
1
P(x0 ) n
3 0 P(x )
d
x
d2x0
=
40 S 0
x x0 
x x0 
V0
84
where the first term, which describes the effects of the induced, noncancelling dipole
moment on the surface of the volume, can be neglected, unless there is a discontinuity
in P n at the surface. Doing so, we find that the contribution from the electric dipole
moments to the potential is given by
p =
1
40
V0
d3x0
0 P(x0 )
x x0 
(6.7)
Comparing this expression with expression Equation (3.3) on page 35 for the electrostatic potential from a static charge distribution , we see that P(x) has the
characteristics of a charge density and that, to the lowest order, the effective charge
density becomes (x) P(x), in which the second term is a polarisation term.
The version of Equation (1.7) on page 5 where free, true charges and bound,
polarisation charges are separated thus becomes
E=
(6.8)
Rewriting this equation, and at the same time introducing the electric displacement
vector (C/m2 )
D = 0 E + P
(6.9)
we obtain
(0 E + P) = D = true (x)
(6.10)
where true is the true charge density in the medium. This is one of Maxwells
equations and is valid also for time varying fields. By introducing the notation pol =
P for the polarised charge density in the medium, and total = true + pol
for the total charge density, we can write down the following alternative version
of Maxwells equation (6.22a) on page 87
E=
total (x)
0
(6.11)
Often, for low enough field strengths E, the linear and isotropic relationship
between P and E
P = 0 E
(6.12)
85
Inserting the approximation (6.12) into Equation (6.9) on the preceding page, we
can write the latter
D = E
(6.13)
where, approximately,
= 0 (1 + )
(6.14)
1
2
V0
(6.15)
(6.16)
86
P
+M
t
(6.17)
B
M
0
(6.18)
and using the definition for D, Equation (6.9) on page 85, we can write this incorrect
equation in the following form
H = jtrue +
D
E
P
= jtrue +
0
t
t
t
(6.19)
As we see, in this simplistic view, we would pick up a term which makes the equation
inconsistent; the divergence of the left hand side vanishes while the divergence of the
right hand side does not. Maxwell realised this and to overcome this inconsistency he
was forced to add his famous displacement current term which precisely compensates
for the last term in the right hand side. In Chapter 1, we discussed an alternative way,
based on the postulate of conservation of electric charge, to introduce the displacement
current.
We may, in analogy with the electric case, introduce a magnetic susceptibility for
the medium. Denoting it m , we can write
H=
(6.20)
where, approximately,
= 0 (1 + m )
(6.21)
(6.22a)
B=0
(6.22b)
B
E=
t
H = j(t, x) +
(6.22c)
D
t
(6.22d)
and are called Maxwells macroscopic equations. These equations are convenient to
use in certain simple cases. Together with the boundary conditions and the constitutive
relations, they describe uniquely (but only approximately!) the properties of the electric and magnetic fields in matter.
87
(6.23)
Integration over the entire volume V and using Gausss theorem (the divergence theorem), we obtain
V0
d3x0
1
(H B + E D) =
2
V0
d3x0 j E +
S0
d2x0 (E H) n
(6.24)
(6.25)
V0
d3x0 j E =
V0
d3x0
j2
V0
d3x0 j EEMF
(6.26)
1
j2
d3x0 j EEMF =
d3x0 (E D + H B) + d2x0 (E H) n
d3x0
+
0
t V 0
2
V0
V0

{z
} S
{z
}

{z
}  {z }
Joule heat
Field energy
Radiated power
(6.27)
which is the energy theorem in Maxwells theory also known as Poyntings theorem.
It is convenient to introduce the following quantities:
Z
1
d3x0 E D
Ue =
2 V0
Z
1
Um =
d3x0 H B
2 V0
S=EH
(6.28)
(6.29)
(6.30)
where Ue is the electric field energy, U m is the magnetic field energy, both measured
in J, and S is the Poynting vector (power flux), measured in W/m2 .
88
B
(D B) + D
t
t
(6.31)
= E( D) B ( H)
(D B) D ( E) + H(
B})
 {z
t
=0
(D B)
t
One verifies easily that the ith vector components of the two terms in square brackets in the right hand member of (6.31) can be expressed as
1
D
1
E
[E( D) D ( E)]i =
E
+
E i D j E D i j
D
2
xi
xi
x j
2
(6.32)
and
[H( B) B ( H)]i =
1
2
B
H
B
xi
xi
x j
1
Hi B j B H i j
2
(6.33)
respectively.
Using these two expressions in the ith component of Equation (6.31) and reshuffling terms, we get
1
D
B
E
H
(E + j B)i
E
+ H
+ (D B)i
D
B
2
xi
xi
xi
xi
t
1
1
E i D j E D i j + Hi B j H B i j
=
x j
2
2
(6.34)
89
Introducing the electric volume force Fev via its ith component
E
H
D
B
1
D
B
(Fev )i = (E + j B)i
E
+ H
2
xi
xi
xi
xi
(6.35)
and the Maxwell stress tensor T with components
1
1
T i j = E i D j E D i j + Hi B j H B i j
2
2
(6.36)
Fev + (D B) =
= ( T)i
t
x j
i
(6.37)
If we introduce the relative electric permittivity and the relative magnetic permeability m as
D = 0 E = E
(6.38)
B = m 0 H = H
(6.39)
(6.40)
where S is the Poynting vector defined in Equation (6.28) on page 88. Integration over
the entire volume V yields
Z
d
m
d3x0 Fev +
d3x0 2 S =
d2x0 T n
dt V 0
c
V0
S0
 {z }

{z
}  {z }
Field momentum
(6.41)
Maxwell stress
which expresses the balance between the force on the matter, the rate of change of the
electromagnetic field momentum and the Maxwell stress. This equation is called the
momentum theorem in Maxwells theory.
In vacuum (6.41) becomes
Z
V0
d3x0 (E + v B) +
1 d
c2 dt
V0
d3x0 S =
S0
d2x0 T n
(6.42)
or
d mech d field
p
+ p
=
dt
dt
90
S0
d2x0 T n
(6.43)
Bibliography
6.4 Bibliography
[1] E. H ALLN, Electromagnetic Theory, Chapman & Hall, Ltd., London, 1962.
[2] J. D. JACKSON, Classical Electrodynamics, third ed., John Wiley & Sons, Inc.,
New York, NY . . . , 1999, ISBN 047130932X.
[3] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism, second ed.,
AddisonWesley Publishing Company, Inc., Reading, MA . . . , 1962, ISBN 0201057026.
[4] J. A. S TRATTON, Electromagnetic Theory, McGrawHill Book Company, Inc., New
York, NY and London, 1953, ISBN 070621500.
91
CHAPTER 7
d f (x) eit
(7.1a)
(7.1b)
1
2
dt f (t, x) eit
That such transform pairs exist is true for most physical variables which are neither
strictly monotonically increasing nor strictly monotonically decreasing with time. For
charge and current densities varying in time we can therefore, without loss of generality, work with individual Fourier components (x) and j (x), respectively. Strictly
speaking, the existence of a single Fourier component assumes a monochromatic
source (i.e., a source containing only one single frequency component), which in turn
requires that the electric and magnetic fields exist for infinitely long times. However,
by taking the proper limits, we may still use this approach even for sources and fields
of finite duration.
This is the method we shall utilise in this chapter in order to derive the electric and
magnetic fields in vacuum from arbitrary given charge densities (t, x) and current
93
(t, x) =
1
2
(x) =
d (x) eit
(7.2a)
(7.2b)
dt (t, x) eit
and
Z
j(t, x) =
1
2
j (x) =
d j (x) eit
(7.3a)
(7.3b)
dt j(t, x) eit
under the assumption that only retarded potentials produce physically acceptable solutions.
The temporal Fourier transform pair for the retarded scalar potential can then be
written
(t, x) =
(x) =
1
2
d (x) eit
dt (t, x) eit =
(7.4a)
1
40
V0
d3x0 (x0 )
eikxx 
x x0 
(7.4b)
where in the last step, we made use of the explicit expression for the temporal Fourier
transform of the generic potential component (x), Equation (3.28) on page 40.
Similarly, the following Fourier transform pair for the vector potential must exist:
A(t, x) =
A (x) =
1
2
d A (x) eit
dt A(t, x) eit =
(7.5a)
0
4
V0
d3x0 j (x0 )
eikxx 
x x0 
(7.5b)
(7.6)
in order that all physical quantities be real. Similar transform pairs and requirements
of realvaluedness exist for the fields themselves.
In the limit that the sources can be considered monochromatic containing only one
single frequency 0 , we have the much simpler expressions
(t, x) = 0 (x)ei0 t
(7.7a)
i0 t
(7.7b)
j(t, x) = j0 (x)e
(t, x) = 0 (x)e
i0 t
(7.7c)
i0 t
(7.7d)
A(t, x) = A0 (x)e
94
where again the realvaluedness of all these quantities is implied. As discussed above,
we can safely assume that all formulae derived for a general temporal Fourier representation of the source (general distribution of frequencies in the source) are valid
for these simple limiting cases. We note that in this context, we can make the formal
identification = 0 ( 0 ), j = j0 ( 0 ) etc., and that we therefore, without
any loss of stringence, let 0 mean the same as the Fourier amplitude and so on.
(7.8)
The calculations are much simplified if we work in space and, at the final stage,
inverse Fourier transform back to ordinary t space. We are working in the Lorentz
gauge and note that in space the Lorentz condition, Equation (3.13) on page 38,
takes the form
k
A i = 0
c
(7.9)
which provides a relation between (the Fourier transforms of) the vector and scalar
potentials.
Using the Fourier transformed version of Equation (7.8) and Equation (7.5b) on
the preceding page, we obtain
B (x) = A (x) =
V0
d3x0 j (x0 )
eikxx 
x x0 
(7.10)
Utilising Formula (F.57) on page 160 and recalling that j (x0 ) does not depend on x,
we can rewrite this as
ikxx0 
Z
0
e
B (x) =
d3x0 j (x0 )
4 V 0
x x0 
Z
0
x x0
0
=
d3x0 j (x0 )
eikxx 
4 V 0
x x0 3
Z
1
x x0 ikxx0 
3 0
0
e
+ d x j (x ) ik
(7.11)
x x0 
x x0 
V0
Z
0
0
j (x0 )eikxx  (x x0 )
=
d3x0
4 V 0
x x0 3
0
Z
(ik)j (x0 )eikxx  (x x0 )
+ d3x0
x x0 2
V0
95
From this expression for the magnetic field in the frequency () domain, we obtain the total magnetic field in the temporal (t) domain by taking the inverse Fourier
transform (using the identity ik = i/c):
Z
d B (x) eit
R
Z
0 i(kxx0 t)
(x x0 )
0
d j (x )e
3 0
=
dx
4
x x0 3
V0
R
Z
0 i(kxx0 t)
(x x0 )
1
d (i)j (x )e
3 0
dx
+
c V0
x x0 2
Z
Z
0
0
0
0
0 0
0
j(t , x ) (x x ) 0
3 0 j(tret , x ) (x x )
=
d3x0 ret
d
x
+
4 V 0
4c V 0
x x0 3
x x0 2

{z
} 
{z
}
B(t, x) =
Induction field
Radiation field
(7.12)
where
def
0
j(tret
, x0 )
j
t
(7.13)
0
t=tret
The first term, the induction field, dominates near the current source but falls off rapidly with distance from it, is the electrodynamic version of the BiotSavart law in
electrostatics, Formula (1.15) on page 7. The second term, the radiation field or the
far field, dominates at large distances and represents energy that is transported out to
infinity. Note how the spatial derivatives () gave rise to a time derivative ()!
1
eikxx 
eikxx  i0
+
d3x0 (x0 )
d3x0 j (x0 )
0
40
4 V 0
x x 
x x0 
V0
Z
0
1
(x0 )eikxx  (x x0 )
=
d3x0
40 V 0
x x0 3
0
Z
(x0 )(x x0 ) j (x0 ) eikxx 
3 0
ik d x
c
x x0 
x x0 
V0
=
(7.14)
96
(7.15)
(7.16)
Doing so in the last term of Equation (7.14) on the facing page, and also using the fact
that k = /c, we can rewrite this Equation as
E (x) =
Z
(x0 )eikxx  (x x0 )
x x0 3
V0
ikxx0 
Z
1
e
[0 j (x0 )](x x0 )
0
3 0
ikj (x )
dx
c V0
x x0 
x x0 

{z
}
I
1
40
d3x0
(7.17)
The last vectorvalued integral can be further rewritten in the following way:
ikxx0 
[0 j (x0 )](x x0 )
e
0
dx
ikj (x )
I =
0
0
x x 
x x0 
V
0
Z
jm xl x0l
eikxx 
0
d3x0
=
ik
j
(x
)
x
l
l
x0m x x0 
x x0 
V0
Z
3 0
(7.18)
But, since
x0m
jm
xl x0l ikxx0 
e
x x0 2
xl x0l ikxx0 
e
x x0 2
xl x0l ikxx0 
+ jm 0
e
xm x x0 2
jm
x0m
(7.19)
we can rewrite I as
0
eikxx 
xl x0l
ikxx0 
I = d x jm 0
x l e
+ ikj
xm x x0 2
x x0 
V0
Z
0
xl x l
0
+ d3x0 0 jm
x l eikxx 
2
0
0
xm
x x 
V
Z
3 0
(7.20)
0
eikxx 
2
0
0 ikxx0 
I = d x j
2 +
4 j (x x ) (x x )e
0
0
0
x x 
x x 
V
!
Z
0
0
ikxx0 
j
(x
x
)
(x
x
)
e
0
ik d3x0
eikxx  + j
x x0 
x x0 3
V0
3 0
(7.21)
97
Using the triple product baccab Formula (F.51) on page 160 backwards, and inserting the resulting expression for I into Equation (7.17) on the preceding page, we
arrive at the following final expression for the Fourier transform of the total E field:
Z
1
eikxx  i0
+
d3x0 (x0 )
E (x) =
40
4
x x0 
V0
Z
0
0 ikxx 
0
1
(x )e
(x x )
=
d3x0
40 V 0
x x0 3
Z
c V0
x x0 3
+
eikxx 
d x j (x )
x x0 
V0
3 0
(7.22)
Taking the inverse Fourier transform of Equation (7.22) above, once again using
the vacuum relation = kc, we find, at last, the expression in time domain for the
total electric field:
E(t, x) =
=
d E (x) eit
1
40

+
V0
d3x0
0
(tret
, x0 )(x x0 )
x x0 3
{z
}
1
40 c

1
40 c

V0
1
40 c2

d3x0
[j(tret , x ) (x x0 )](x x0 )
x x0 4
{z
}
(7.23)
Intermediate field
V0
0
0
3 0 [j(tret , x )
dx
(x x )] (x x )
x x0 4
{z
}
Intermediate field
V0
d3x0
0
[j(tret
, x0 ) (x x0 )] (x x0 )
x x0 3
{z
}
Radiation field
Here, the first term represents the retarded Coulomb field and the last term represents
the radiation field which carries energy over very large distances. The other two terms
represent an intermediate field which contributes only in the near zone and must be
taken into account there.
With this we have achieved our goal of finding closedform analytic expressions
for the electric and magnetic fields when the sources of the fields are completely arbitrary, prescribed distributions of charges and currents. The only assumption made
is that the advanced potentials have been discarded; recall the discussion following
Equation (3.31) on page 41 in Chapter 3.
98
1
40 c2
V0
d3x0
0
4c
V0
d3x0
0
j(tret
, x0 ) (x x0 )
x x0 2
0
[j(tret
, x0 ) (x x0 )] (x x0 )
x x0 3
(7.24a)
(7.24b)
where
def
0
j(tret
, x0 )
j
t
(7.25)
0
t=tret
Instead of studying the fields in the time domain, we can often make a spectrum
analysis into the frequency domain and study each Fourier component separately. A
superposition of all these components and a transformation back to the time domain
will then yield the complete solution.
The Fourier representation of the radiation fields Equation (7.24a) above and Equation (7.24b) were included in Equation (7.11) on page 95 and Equation (7.22) on the
facing page, respectively and are explicitly given by
Z
1
dt Brad (t, x) eit
2
Z
j (x0 ) (x x0 ) ikxx0 
k0
d3x0
e
= i
4 V 0
x x0 2
Z
0
j (x0 ) k ikxx0 
= i
e
d3x0
4 V 0
x x0 
Z
1
Erad
dt Erad (t, x) eit
(x) =
2
Z
k
[j (x0 ) (x x0 )] (x x0 ) ikxx0 
e
= i
d3x0
40 c V 0
x x0 3
Z
[j (x0 ) k] (x x0 ) ikxx0 
1
d3x0
= i
e
40 c V 0
x x0 2
Brad
(x) =
(7.26a)
(7.26b)
99
S
dS = nd
2x
k
x x0
x0
x x0
x0 x 0
x0
V0
O
Relation between the surface normal and the k vector for radiation generated at source points x0 near the point x0 in the source volume V. At
distances much larger than the extent of V, the unit vector n,
normal to the surface S which has its centre at x0 , and the unit vector k of the radiation k vector
from x0 are nearly coincident.
F IGURE 7.1:
If the source is located inside a volume V near x0 and has such a limited spatial
extent that max x0 x0  x x0 , and the integration surface S , centred on x0 , has
a large enough radius x x0  max x0 x0 , we see from Figure 7.1 that we can
approximate
k x x0 k (x x0 ) k (x x0 ) k (x0 x0 )
k x x0  k (x0 x0 )
(7.27)
Recalling from Formula (F.45) and Formula (F.46) on page 159 that
dS = x x0 2 d = x x0 2 sin d d
and noting from Figure 7.1 that k and n are nearly parallel, we see that we can approximate.
k dS
k n
=
dS d
x x0 2 x x0 2
(7.28)
100
Radiated energy
j (x0 ) k ik(x0 x0 )
0
i eikxx0  d3x0
e
4
x x0 
V0
Z
0 eikxx0 
0
d3x0 [j (x0 ) k] eik(x x0 )
i
4 x x0  V 0
Z
0
0
1
ikxx0 
3 0 [j (x ) k] (x x ) ik(x0 x0 )
Erad
(x)
i
e
d
x
e
40 c
x x0 2
V0
Z
1 eikxx0  (x x0 )
0
d3x0 [j (x0 ) k] eik(x x0 )
i
40 c x x0  x x0 
V0
Brad
(x)
(7.29a)
(7.29b)
I.e., if max x0 x0  x x0 , then the fields can be approximated as spherical waves
multiplied by dimensional and angular factors, with integrals over points in the source
volume only.
hSi = hE Hi =
(7.30)
Using the farfield approximations (7.29a) and (7.29b) and the fact that 1/c = 0 0
k)e
(7.31)
R
hSi =
0
x x0 
2
2
32
x x0 
V0
k)e
0
2
d 32
V0
(7.32)
101
dt S(t) =
=
dt (E H)
d
d0
(7.33)
dt (E H0 ) ei(+ )t
If we carry out the temporal integration first and use the fact that
Z
dt ei(+ )t = 2( + 0 )
(7.34)
d (E H )
Z
Z 0
= 2
(E H ) d +
(E H ) d
0
Z
Z
(E H ) d
= 2
(E H ) d
0
0
Z
Z
(E H ) d
= 2
(E H ) d +
dt S(t) = 2
(7.35)
2
=
(E B + E B ) d
0 0
Z
2
=
(E B + E B ) d
0 0
where the last step follows from physical requirement of realvaluedness of E and
B . We insert the Fourier transforms of the field components which dominate at large
distances, i.e., the radiation fields (7.26a) and (7.26b). The result, after integration
over the area S of a large sphere which encloses the source, is
U=
1
4
0
0
d2x n
2
Z
j k ikxx0 
e
d d3x0
k
x x0 
V0
(7.36)
Inserting the approximations (7.27) and (7.28) into Equation (7.36) and also introducing
U=
102
U d
(7.37)
Bibliography
and recalling the definition (2.18) on page 28 for the vacuum resistance R0 we obtain
2
Z
1
dU
0
d
R0 d3x0 (j k)eik(x x0 ) d
d
4
V0
(7.38)
which, at large distances, is a good approximation to the energy that is radiated per unit
solid angle d in a frequency band d. It is important to notice that Formula (7.38)
includes only source coordinates. This means that the amount of energy that is being
radiated is independent on the distance to the source (as long as it is large).
7.5 Bibliography
[1] F. H OYLE , S IR AND J. V. NARLIKAR, Lectures on Cosmology and Action at a Distance
Electrodynamics, World Scientific Publishing Co. Pte. Ltd, Singapore, New Jersey, London
and Hong Kong, 1996, ISBN 98100225733(pbk).
[2] J. D. JACKSON, Classical Electrodynamics, third ed., John Wiley & Sons, Inc.,
New York, NY . . . , 1999, ISBN 047130932X.
[3] L. D. L ANDAU AND E. M. L IFSHITZ, The Classical Theory of Fields, fourth revised English ed., vol. 2 of Course of Theoretical Physics, Pergamon Press, Ltd., Oxford . . . , 1975,
ISBN 0080250726.
[4] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism, second ed.,
AddisonWesley Publishing Company, Inc., Reading, MA . . . , 1962, ISBN 0201057026.
[5] J. A. S TRATTON, Electromagnetic Theory, McGrawHill Book Company, Inc., New
York, NY and London, 1953, ISBN 070621500.
103
CHAPTER 8
105
sin[k(L/2 x03 )]
j(t0 , x0 )
L2
L
2
F IGURE 8.1: A linear antenna used for transmission. The current in the feeder
and the antenna wire is set up by the EMF of the generator (the transmitter).
At the ends of the wire, the current is reflected back with a 180 phase shift to
produce a antenna current in the form of a standing wave.
j0 (x ) =
x03 )]
x 3
sin(kL/2)
sin[k(L/2
I0 (x01 )(x02 )
(8.1)
106
x3 = z
x
L
2
j (x0 )
x2
x1
L2
(8.2)
Inserting this expression and d = 2 sin d into Formula (7.32) on page 101 and
integrating over , we find that the total radiated power from the antenna is
P(L) = R0 I02
1
4
2
sin d
(8.3)
107
kL0
12
2
L
R0 I02
(8.4)
R (L) = 2 = 1 2 = R0
6
Ieff
I
2 0
rad
2
2
L
L
197
(8.5)
is called the radiation resistance. For the technologically important case of a halfwave antenna, i.e., for L = /2 or kL = , Formula (8.3) on the preceding page
reduces to
Z
cos2 2 cos
2 1
P(/2) = R0 I0
d
(8.6)
4 0
sin
The integral in (8.6) can always be evaluated numerically. But, it can in fact also
be evaluated analytically as follows:
Z
Z 1
cos2 2 cos
cos2 2 u
d = [cos u] =
du =
sin
1 u2
0
1
1 + cos(u)
cos2
u =
2
2
Z
1 1 1 + cos(u)
du
=
2 1 (1 + u)(1 u)
Z
Z
1 1 1 + cos(u)
1 1 1 + cos(u)
(8.7)
=
du +
du
4 1 (1 + u)
4 1 (1 u)
Z
h
1 1 1 + cos(u)
vi
=
du = 1 + u
2 1 (1 + u)
Z 2
1
1
1 cos v
dv = [ + ln 2 Ci(2)]
=
2 0
v
2
1.22
where in the last step the EulerMascheroni constant = 0.5772 . . . and the cosine
integral Ci(x) were introduced. Inserting this into the expression Equation (8.6) above
we obtain the value Rrad (/2) 73 .
108
r
x3 = z = z 0
k
x2
z 0
j (x0 )
x0
x1
0
0
F IGURE 8.3: For the loop antenna the spherical coordinate system (r, , ) describes the field point x (the radiation field) and the cylindrical coordinate system
(0 , 0 , z0 ) describes the source point x0 (the antenna current).
According to Equation (7.29a) on page 101 the Fourier component of the radiation
part of the magnetic field generated by an extended, monochromatic current source is
Brad
=
i0 eikx
4 x
V0
d3x0 eikx j k
(8.8)
In our case the generator produces a single frequency and we feed the antenna
across a small gap where the loop crosses the positive x1 axis. The circumference
of the loop is chosen to be exactly one wavelength = 2c/. This means that the
antenna current oscillates in the form of a sinusoidal standing current wave around the
circular loop with a Fourier amplitude
j = I0 cos 0 (0 a)(z0 ) 0
(8.9)
For the spherical coordinate system of the field point, we recall from subsection F.4.1 on
page 159 that the following relations between the base vectors hold:
r = sin cos x 1 + sin sin x 2 + cos x 3
= cos cos x 1 + cos sin x 2 sin x 3
= sin x 1 + cos x 2
109
and
x 1 = sin cos r + cos cos sin
x 2 = sin sin r + cos sin + cos
x 3 = cos r sin
With the use of the above transformations and trigonometric identities, we obtain for
the cylindrical coordinate system which describes the source:
0 = cos 0 x 1 + sin 0 x 2
(8.10)
0 = sin 0 x 1 + cos 0 x 2
(8.11)
(8.12)
0
(8.13)
(8.14)
and
With these expressions inserted and d3x0 = 0 d0 d0 dz0 , the source integral becomes
Z
V0
d3x0 eikx j k = a
= I0 ak
2
0
2
0
0
eika sin cos( ) cos(0 ) cos 0 d0
+ I0 ak cos
2
0
(8.15)
Utilising the periodicity of the integrands over the integration interval [0, 2], introducing the auxiliary integration variable 00 = 0 , and utilising standard trigonometric identities, the first integral in the RHS of (8.15) can be rewritten
Z
00
00
110
(8.16)
00
2
1
00
eika sin cos d00
= sin
2
0
Z 2
1
00
sin
eika sin cos cos 200 d00
2
0
(8.17)
in
ei cos cos n d =
in
2
ei cos cos n d
(8.18)
2
0
2
0
00
V0
d3x0 eikx j k = I + I
(8.20)
i0 eikr
I + I
(8.21)
4r
To obtain the desired physical magnetic field in the radiation (far) zone we must
Fourier transform back to t space and take the real part and evaluate it at the retarded
time:
0
i0 e(ikrt )
rad
B (t, x) = Re
I + I
4r
0
=
sin(kr t0 ) I + I
4r
I0 ak0
sin(kr t0 ) cos [J0 (ka sin ) J2 (ka sin )]
=
4r
+ cos sin [J0 (ka sin ) + J2 (ka sin )]
Brad
(x) =
(8.22)
From this expression for the radiated B field, we can obtain the radiated E field with
the help of Maxwells equations.
111
(8.23)
In Section 6.1.1 we introduced the electric polarisation P such that P = pol , the
polarisation charge density. If we introduce a vector field (t, x) such that
= true
(8.24a)
= jtrue
(8.24b)
t
and compare with Equation (8.23), we see that (t, x) satisfies this equation of continuity. Furthermore, if we compare with the electric polarisation [cf. Equation (6.9)
on page 85], we see that the quantity is related to the true charges in the same way
as P is related to polarised charge, namely as a dipole moment density. The quantity
is referred to as the polarisation vector since, formally, it treats also the true (free)
charges as polarisation charges so that
E =
true + pol P
=
0
0
(8.25)
(8.26a)
1
=A
(8.26b)
c2 t
where and A are the electromagnetic scalar and vector potentials, respectively. As
we see, e acts as a superpotential in the sense that it is a potential from which
we can obtain other potentials. It is called the Hertz vector or polarisation potential. Requiring that the scalar and vector potentials and A, respectively, fulfil their
inhomogeneous wave equations, one finds, using (8.24) and (8.26), that Hertz vector
must satisfy the inhomogeneous wave equation
2 e =
112
1 2 e
2 e =
c2 t2
0
(8.27)
Multipole radiation
x x0
x0
x x0
x
x0 x 0
x0
V0
O
F IGURE 8.4: Geometry of a typical multipole radiation problem where the field
point x is located some distance away from the finite source volume V 0 centred
around x0 . If k x0 x0  1 k x x0 , then the radiation at x is well approximated by a few terms in the multipole expansion.
This equation is of the same type as Equation (3.15) on page 38, and has therefore
the retarded solution
1
(t, x) =
40
e
V0
d3x0
0
(tret
, x0 )
x x0 
(8.28)
1
40
V0
d3x0
(x0 )eikxx 
x x0 
(8.29)
(8.30)
we see that we can calculate the magnetic and electric fields, respectively, as follows
1 C
c2 t
(8.31a)
E=C
(8.31b)
B=
Clearly, the last equation is valid only outside the source volume, where E = 0.
Since we are mainly interested in the fields in the far zone, a long distance from the
source region, this is no essential limitation.
Assume that the source region is a limited volume around some central point x 0
far away from the field (observation) point x illustrated in Figure 8.4. Under these
assumptions, we can expand the Hertz vector, expression (8.28), due to the presence
0
of nonvanishing (tret
, x0 ) in the vicinity of x0 , in a formal series. For this purpose we
113
eik(xx0 )(x x0 )
eikxx 
0
x x  (x x0 ) (x0 x0 )
(8.32)
= ik (2n + 1)Pn (cos ) jn (k x0 x0 )h(1)
n (k x x0 )
n=0
eikxx 
is a Green function
x x0 
is the angle between x0 x0 and x x0
h(1)
n (k x x0 ) is the spherical Hankel function of the first kind of order n
According to the addition theorem for Legendre polynomials
n
0 im( )
(1)m Pmn (cos )Pm
n (cos )e
Pn (cos ) =
(8.33)
m=n
where Pm
n is an associated Legendre polynomial and, in spherical polar coordinates,
(8.34a)
x 0 x 0 = ( x 0 x 0 , 0 , 0 )
x x0 = (x x0  , , )
(8.34b)
Inserting Equation (8.32), together with Formula (8.33) above, into Equation (8.29)
on the preceding page, we can in a formally exact way expand the Fourier component
of the Hertz vector as
e =
ik
40
V0
m
im
(2n + 1)(1)m h(1)
n (k x x0 ) Pn (cos ) e
n=0 m=n
0 im0
d x (x0 ) jn (k x0 x0 ) Pm
n (cos ) e
(8.35)
3 0
114
eikxx0 
k x x0 
(8.37)
Multipole radiation
and replace jn with the first term in its power series expansion:
jn (k x0 x0 )
n
2n n!
k x0 x 0
(2n + 1)!
(8.38)
Inserting these expansions into Equation (8.35) on the preceding page, we obtain the
multipole expansion of the Fourier component of the Hertz vector
e
e(n)
(8.39a)
n=0
where
e(n) = (i)n
1 eikxx0  2n n!
40 x x0  (2n)!
V0
d3x0 (x0 ) (k x0 x0 )n Pn (cos ) (8.39b)
eikxx0 
40 x x0 
V0
d3x0 (x0 ) =
1 eikxx0 
p
40 x x0 
(8.40)
Since represents a dipole moment density for the true charges (in the same vein as
R
P does so for the polarised charges), p = V 0 d3x0 (x0 ) is the Fourier component of
the electric dipole moment
p(t, x0 ) =
V0
d3x0 (t0 , x0 ) =
V0
(8.41)
[cf. Equation (6.2) on page 84 which describes the static dipole moment]. If a spherical coordinate system is chosen with its polar axis along p as in Figure 8.5, the
115
k
x3
Brad
Erad
x2
x1
F IGURE 8.5: If a spherical polar coordinate system (r, , ) is chosen such that
the electric dipole moment p (and thus its Fourier transform p ) is located at the
origin and directed along the polar axis, the calculations are simplified.
er e(0) r =
def
e e(0) = 0
(8.42a)
(8.42b)
(8.42c)
Evaluating Formula (8.30) on page 113 for the help vector C, with the spherically
polar components (8.42) of e(0) inserted, we obtain
ikxx0 
1
1
e
(0)
C = C,
=
ik
p sin
(8.43)
40 x x0 
x x0 
Applying this to Equation (8.31) on page 113, we obtain directly the Fourier components of the fields
ikxx0 
0
1
e
B = i
ik
p sin
(8.44a)
4
x x0 
x x0 
1
1
x x0
ik
E =
2
cos
2
40
x x0 
x x0 
x x0 
(8.44b)
ikxx0 
ik
1
2
e
sin
k
p
x x0 
x x0 2 x x0 
116
Multipole radiation
Keeping only those parts of the fields which dominate at large distances (the radiation fields) and recalling that the wave vector k = k(x x0 )/ x x0  where k = /c,
we can now write down the Fourier components of the radiation parts of the magnetic
and electric fields from the dipole:
0 eikxx0 
0 eikxx0 
p k sin =
(p k)
4 x x0 
4 x x0 
1 eikxx0 
1 eikxx0 
=
p k2 sin =
[(p k) k]
40 x x0 
40 x x0 
Brad
=
(8.45a)
Erad
(8.45b)
These fields constitute the electric dipole radiation, also known as E1 radiation.
e(1) = i
(8.46)
(8.47)
and introducing
i = xi x0,i
0i
x0i
(8.48a)
x0,i
(8.48b)
1
i , j 0i + ,i 0j
2
1
+ i , j 0i ,i 0j
2
(8.49)
i.e., as the sum of two parts, the first being symmetric and the second antisymmetric
in the indices i, j. We note that the antisymmetric part can be written as
1
1
i , j 0i ,i 0j = [, j (i 0i ) 0j (i ,i )]
2
2
1
= [ ( 0 ) 0 ( )] j
2
1
(x x0 ) [ (x0 x0 )] j
=
2
(8.50)
117
The utilisation of Equations (8.24) on page 112, and the fact that we are considering a single Fourier component,
(t, x) = eit
(8.51)
allow us to express in j as
= i
(8.52)
Hence, we can write the antisymmetric part of the integral in Formula (8.46) on the
previous page as
Z
1
(x x0 ) d3x0 (x0 ) (x0 x0 )
2
V0
Z
1
= i (x x0 ) d3x0 j (x0 ) (x0 x0 )
2
V0
1
= i (x x0 ) m
(8.53)
1
2
V0
(8.54)
The final result is that the antisymmetric, magnetic dipole, part of e(1) can be
written
e,antisym
(1)
k
eikxx0 
(x x0 ) m
40 x x0 2
(8.55)
In analogy with the electric dipole case, we insert this expression into Equation (8.30)
on page 113 to evaluate C, with which Equations (8.31) on page 113 then gives the B
and E fields. Discarding, as before, all terms belonging to the near fields and transition
fields and keeping only the terms that dominate at large distances, we obtain
0 eikxx0 
(m k) k
4 x x0 
k eikxx0 
Erad
m k
(x) =
40 c x x0 
Brad
(x) =
(8.56a)
(8.56b)
which are the fields of the magnetic dipole radiation (M1 radiation).
page 115 for the expansion of the Hertz vector can be expressed in terms of the
118
V0
(8.57)
Again we use this expression in Equation (8.30) on page 113 to calculate the fields
via Equations (8.31) on page 113. Tedious, but fairly straightforward algebra (which
we will not present here), yields the resulting fields. The radiation components of the
fields in the far field zone (wave zone) are given by
i0 eikxx0 
(k Q ) k
8 x x0 
i eikxx0 
Erad
[(k Q ) k] k
(x) =
80 x x0 
Brad
(x) =
(8.58a)
(8.58b)
0
1
(tret
, x0 )
d3x0
40 V 0
x x0 
Z
0
0
j(t , x0 )
A(t, x) =
d3x0 ret 0
4 V 0
x x 
(t, x) =
(8.59a)
(8.59b)
and consider a source region with such a limited spatial extent that the charges and
currents are well localised. Specifically, we consider a charge q0 , for instance an
electron, which, classically, can be thought of as a localised, unstructured and rigid
charge distribution with a small, finite radius. The part of this charge distribution
dq0 which we are considering is located in dV 0 = d3x0 in the sphere in Figure 8.6 on
the next page. Since we assume that the electron (or any other other similar electric
charge) moves with a velocity v whose direction is arbitrary and whose magnitude can
be almost comparable to the speed of light, we cannot say that the charge and current
R
R
0
0
to be used in (8.59) is V (tret
, x0 ) d3x0 and V v(tret
, x0 ) d3x0 , respectively, because
in the finite time interval during which the observed signal is generated, part of the
charge distribution will leak out of the volume element d3x0 .
119
x(t)
0
x0
v(t
)
dS
x
0
0 0
x (t )
dr0
dV 0
F IGURE 8.6: Signals which are observed at the field point x at time t were
generated at source points x0 (t0 ) on a sphere, centred on x and expanding, as time
increases, with the velocity c outward from the centre. The source charge element
moves with an arbitrary velocity v and gives rise to a source leakage out of the
source volume dV 0 = d3x0 .
(x x0 ) v 0 0
dS dt
x x0 
(8.60)
where the last term represents the amount of source leakage due to the fact that
the charge distribution moves with velocity v(t 0 ) = dx/dt0 . Since dt0 = dr0 /c and
dS 0 dr0 = d3x0 we can rewrite the expression for the net charge as
(x x0 ) v 3 0
0
0
dq0 = (tret
, x0 ) d3x0 (tret
, x0 )
dx
c x x0 
(x x0 ) v
0
d3x0
= (tret
, x0 ) 1
c x x0 
120
(8.61)
or
0
(tret
, x0 ) d3x0 =
dq0
1
(8.62)
(xx0 )v
cxx0 
(8.63)
This is the expression to be used in the Formulae (8.59) on page 119 for the retarded
potentials. The result is (recall that j = v)
Z
dq0
0
x (xxc )v
Z
v dq0
0
A(t, x) =
0
4 x x0  (xxc )v
(t, x) =
1
40
(8.64a)
x0 
(8.64b)
For a sufficiently small and well localised charge distribution we can, assuming that
the integrands do not change sign in the integration volume, use the mean value theR
orem and the fact that V dq0 = q0 to evaluate these expressions to become
1
q0 1
q0
=
0
40 x x0  (xxc )v
40 s
0
q0 v
q
v
v
A(t, x) =
=
(t, x)
0 )v =
(xx
2
40 c x x0  c
40 c2 s c2
(t, x) =
(8.65a)
(8.65b)
where
(x x0 (t0 )) v(t0 )
s = s(t0 , x) = x x0 (t0 )
c
x
x0 (t0 ) v(t0 )
= x x0 (t0 ) 1
c
x x0 (t0 )
0 0
0
x
x
(t
)
v(t
)
= (x x0 (t0 ))
c
x x0 (t0 )
(8.66a)
(8.66b)
(8.66c)
is the retarded relative distance. The potentials (8.65) are precisely the LinardWiechert potentials which we derived in Section 4.3.2 on page 59 by using a covariant
formalism.
It is important to realise that in the complicated derivation presented here, the observer is in a coordinate system which has an absolute meaning and the velocity v is
that of the particle, whereas in the covariant derivation two frames of equal standing
were moving relative to each other with v. Expressed in the fourpotential, Equation (4.47) on page 58, the LinardWiechert potentials become
1 v
q0
, 2
,A
(8.67)
A (x ) =
=
40 cs c s
c
121
?
x x0 
v
c
q0
x0 (t0 )
v(t0 )
x0 (t)
x x0
x x0
x(t)
(8.68a)
A(t, x)
t
(8.68b)
(8.69)
(in the interest of simplifying our notation, we drop the subscript ret on t 0 from now
on). This means that we know the trajectory of the charge q0 , i.e., x0 , for all times
up to the time t0 at which a signal was emitted in order to precisely arrive at the field
point x at time t. Because of the finite speed of propagation of the fields, the trajectory
at times later than t0 is not (yet) known.
122
dx0
dt0
a(t0 ) = v (t0 ) =
(8.70a)
dv d2 x0
= 02
dt0
dt
(8.70b)
As for the charge coordinate x0 itself, we have in general no knowledge of the velocity
and acceleration at times later than t 0 , in particular not at the time of observation t. If
we choose the field point x as fixed, application of (8.70) to the relative vector x x0
yields
d
(x x0 (t0 )) = v(t0 )
dt0
(8.71a)
d2
(x x0 (t0 )) = v(t0 )
dt0 2
(8.71b)
The retarded time t0 can, at least in principle, be calculated from the implicit relation
t0 = t0 (t, x) = t
x x0 (t0 )
c
(8.72)
and we shall see later how this relation can be taken into account in the calculations.
According to Formulae (8.68) on the preceding page the electric and magnetic
fields are determined via differentiation of the retarded potentials at the observation
time t and at the observation point x. In these formulae the unprimed , i.e., the
spatial derivative differentiation operator = x i /xi means that we differentiate with
respect to the coordinates x = (x1 , x2 , x3 ) while keeping t fixed, and the unprimed time
derivative operator /t means that we differentiate with respect to t while keeping x
fixed. But the LinardWiechert potentials and A, Equations (8.65) on page 121, are
expressed in the charge velocity v(t 0 ) given by Equation (8.70a) above and the retarded
relative distance s(t0 , x) given by Equation (8.66) on page 121. This means that the
expressions for the potentials and A contain terms which are expressed explicitly
in t0 , which in turn is expressed implicitly in t via Equation (8.72) above. Despite
this complication it is possible, as we shall see below, to determine the electric and
magnetic fields and associated quantities at the time of observation t. To this end, we
need to investigate carefully the action of differentiation on the potentials.
x x0 (t0 ) = x x
x x0 (t0 ) =
(8.73)
0
0
0
t x
t x
x x 
x x0 
123
t
x x0 
=1
(8.74)
t0 x c
t x
(x x0 ) v(t0 ) t0
=1+
c x x0 
t x
This is an algebraic equation in (t 0 /t)x which we can solve to obtain
0
t
x x0 
x x0 
=
=
0
0
0
t x x x  (x x ) v(t )/c
s
(8.75)
where s = s(t0 , x) is the retarded relative distance given by Equation (8.66) on page 121.
Making use of Equation (8.75), we obtain the following useful operator identity
0
t
x x0 
=
=
(8.76)
t x
t x t0 x
s
t0 x
Likewise, by applying ()t to Equation (8.72) on the previous page we obtain
x x0
x x0 (t0 (t, x))
=
()t (x x0 )
c
c x x0 
x x0
(x x0 ) v(t0 )
=
+
()t t0
c x x0 
c x x0 
()t t0 = ()t
(8.77)
x x0
(8.78)
cs
which gives the following operator relation when ()t is acting on an arbitrary function
of t0 and x:
x x0
0 =
()t = ()t t0
+
()
+ ()t0
(8.79)
t
t0 x
cs
t0 x
()t t0 =
With the help of the rules (8.79) and (8.76) we are now able to replace t by t 0 in the
operations which we need to perform. We find, for instance, that
1 q0
()t =
40 s
(8.80a)
0
q
x x0
v(t0 ) x x0 s
40 s2 x x0 
c
cs
t0 x
0 q0 v(t0 )
A
A
=
t
t x t 4
s
x
(8.80b)
0
q
0
0
x x sv(t ) x x0 v(t0 ) s
=
40 c2 s3
t0 x
124
Utilising these relations in the calculation of the E field from the LinardWiechert
potentials, Equations (8.65) on page 121, we obtain
A(t, x)
t
(x x0 (t0 )) x x0 (t0 ) v(t0 )/c
q0
=
40 s2 (t0 , x)
x x0 (t0 )
0 0
(x x (t )) x x0 (t0 ) v(t0 )/c s(t0 , x)
x x0 (t0 ) v (t0 )
cs(t0 , x)
t0
c2
x
(8.81)
E(t, x) = (t, x)
Starting from expression (8.66a) on page 121 for the retarded relative distance s(t 0 , x),
we see that we can evaluate (s/t 0 )x in the following way
(x x0 ) v(t0 )
s
0
xx
=
t0 x
t0 x
c
!
1 x x0 (t0 )
v(t0 )
0 0
0
0 0
v(t ) + (x x (t ))
= 0 x x (t )
t
c
t0
t0
=
+
c
c
x x0 
(8.82)
where Equation (8.73) on page 123 and Equations (8.70) on page 123, respectively,
were used. Hence, the electric field generated by an arbitrarily moving charged
particle at x0 (t0 ) is given by the expression
q0
v2 (t0 )
x x0 (t0 ) v(t0 )
0 0
E(t, x) =
(x
x
(t
))
40 s3 (t0 , x)
c
c2

{z
}
+
q
x x0 (t0 )
3
0
40 s (t , x)
c2

(x x0 (t0 ))
{z
x x0 (t0 ) v(t0 )
c
v (t0 )
(8.83)
}
The first part of the field, the velocity field, tends to the ordinary Coulomb field when
v 0 and does not contribute to the radiation. The second part of the field, the
acceleration field, is radiated into the far zone and is therefore also called the radiation
field.
From Figure 8.7 on page 122 we see that the position the charged particle would
have had if at t0 all external forces would have been switched off so that the trajectory
from then on would have been a straight line in the direction of the tangent at x 0 (t0 ) is
x0 (t), the virtual simultaneous coordinate. During the arbitrary motion, we interpret
125
xx0 as the coordinate of the field point x relative to the virtual simultaneous coordinate x0 (t). Since the time it takes for a signal to propagate (in the assumed vacuum)
from x0 (t0 ) to x is x x0  /c, this relative vector is given by
x x0 (t) = x x0 (t0 )
x x0 (t0 ) v(t0 )
c
(8.84)
This allows us to rewrite Equation (8.83) on the previous page in the following way
v2
(x x0 ) v
q0
0
(x x0 ) 1 2 + (x x )
(8.85)
E(t, x) =
40 s3
c
c2
In a similar manner we can compute the magnetic field:
x x0
cs
q0
x x0
A
x x0
=
40 c2 s2 x x0 
c x x0 
t x
t0
A
(8.86)
where we made use of Equation (8.65) on page 121 and Formula (8.76) on page 124.
But, according to (8.80a),
x x0
q0
x x0
v
()
=
t
c x x0 
40 c2 s2 x x0 
(8.87)
so that
A
x x0
()t
B(t, x) =
0
c x x 
t x
x x0
=
E(t, x)
c x x0 
(8.88)
The radiation part of the electric field is obtained from the acceleration field in
Formula (8.83) on the preceding page as
Erad (t, x) =
lim E(t, x)
xx0 
0
q
x x0  v
0
0
v
(x
x
)
(x
x
)
40 c2 s3
c
q0
=
(x x0 ) [(x x0 ) v ]
40 c2 s3
(8.89)
where in the last step we again used Formula (8.84) above. Using this formula and
Formula (8.88), the radiation part of the magnetic field can be written
Brad (t, x) =
126
x x0
Erad (t, x)
c x x0 
(8.90)
x x0 2 v2
x x0 2 v2
2 0
cos
+
sin2 0
c2
c2
x x0 2 v2
x x0 2 v2
2 0
2 0
=
(cos
+
sin
)
=
c2
c2
=
we find that
2
2
(x x0 ) v
(x x0 ) v
x x0 2 v2
=
c
c2
c
(8.92)
(8.93)
Furthermore, from Equation (8.84) on the facing page, we obtain the following identity:
(x x0 (t0 )) v = (x x0 (t)) v
(8.94)
which, when inserted into Equation (8.93) above, yields the relation
2
2
(x x0 ) v
(x x0 ) v
x x0 2 v2
=
c
c2
c
(8.95)
Inserting the above into expression (8.91) for s2 , this expression becomes
2
2
(x x0 ) v x x0 2 v2
(x x0 ) v
s 2 = x x 0 2 x x 0
+
c
c2
c
2
2
0
(x x0 ) v
x x  v
= (x x0 )
c
c
(8.96)
2
(x x0 ) v
2
= (x x0 )
c
2
(x x0 (t)) v(t0 )
x x0 (t)2
c
127
128
and v is 0 .
We note that in the case of uniform velocity v, time and space derivatives are closely related
in the following way when they operate on functions of x(t) [cf. Equation (1.33) on page 13]:
v
t
(8.99)
Hence, the E and B fields can be obtained from Formulae (8.68) on page 122, with the potentials
given by Equations (8.65) on page 121 as follows:
A
1 v
v
= 2
= 2
t
c t
c t
vv
v v
= 1 2
= +
c
vv c c
=
1
c2
v
v
v
B = A = 2 = 2 = 2
c
c
c
v
i v vv
v h v
= 2
= 2 2 1
c
c
c
c
c
v
= 2 E
c
E =
(8.100a)
(8.100b)
Here 1 = x i x i is the unit dyad and we used the fact that v v 0. What remains is just to
express in quantities evaluated at t and x.
From Equation (8.65a) on page 121 and Equation (8.98) on the preceding page we find that
q0
q0
1
=
=
s2
40
s
80 s3
(8.101)
i
q0 h
v v
=
(x
x
)
+
(x
x
)
0
0
40 s3
c
c
When this expression for is inserted into Equation (8.100a), the following result
vv
q0 vv
E(t, x) =
1 =
1 s2
2
3
2
c
80 s
c
v v
q0
(x x0 )
(x x0 ) +
=
3
40 s
c
c
vv h v v
i
v v
(x x0 ) 2
(x x0 )
c c
c
c
c
v v
v2
q0
(x
x
)
+
(x
x
)
(x
x
)
=
0
0
0
40 s3
c c
c2
v v
(x x0 )
c c
q0
v2
(x
x
)
1
=
0
40 s3
c2
(8.102)
follows. Of course, the same result also follows from Equation (8.85) on page 126 with v 0
inserted.
From Equation (8.102) we conclude that E is directed along the vector from the simultan
129
eous coordinate x0 (t) to the field (observation) coordinate x(t). In a similar way, the magnetic
field can be calculated and one finds that
0 q 0
B(t, x) =
4s3
v2
1 2
c
v (x x0 ) =
1
vE
c2
(8.103)
From these explicit formulae for the E and B fields we can discern the following cases:
1. v 0 E goes over into the Coulomb field ECoulomb
2. v 0 B goes over into the BiotSavart field
3. v c E becomes dependent on 0
4. v c, sin 0 0 E (1 v2 /c2 )ECoulomb
5. v c, sin 0 1 E (1 v2 /c2 )1/2 ECoulomb
C E ND OF EXAMPLE 8.1
E XAMPLE 8.2
Let us consider in more detail the treatment of the radiation from a uniformly moving rigid
charge distribution.
If we return to the original definition of the potentials and the inhomogeneous wave equation, Formula (3.15) on page 38, for a generic potential component (t, x) and a generic source
component f (t, x),
2 (t, x) =
1 2
2
(t, x) = f (t, x)
c2 t2
(8.104)
we find that under the assumption that v = v x 1 , this equation can be written
v2
c2
2 2 2
+ 2 + 2 = f (x)
x21
x2
x3
(8.105)
(8.106a)
2 = x 2
(8.106b)
3 = x 3
(8.106c)
def
and introducing the vectorial nabla operator in space, (/1 , /2 , /3 ), the time
130
(8.107)
in this space. This equation has the wellknown Coulomb potential solution
() =
1
4
f (0 ) 3 0
d
 0 
(8.108)
1
4
f (x0 ) 3 0
dx
s
(8.109)
(8.110)
Applying this to the explicit scalar and vector potential components, realising that for a rigid
charge distribution moving with velocity v the current is given by j = v, we obtain
Z
1
(x0 ) 3 0
dx
40 V s
Z
1
v
v(x0 ) 3 0
A(t, x) =
d x = 2 (t, x)
2
40 c V s
c
(t, x) =
(8.111a)
(8.111b)
(8.112a)
(8.112b)
which we recognise as the LinardWiechert potentials; cf. Equations (8.65) on page 121. We
notice, however, that the derivation here, based on a mathematical technique which in fact is a
Lorentz transformation, is of more general validity than the one leading to Equations (8.65) on
page 121.
Let us now consider the action of the fields produced from a moving, rigid charge distribution represented by q0 moving with velocity v, on a charged particle q, also moving with
velocity v. This force is given by the Lorentz force
F = q(E + v B)
(8.113)
With the help of Equation (8.103) on the facing page and Equations (8.111) above, and the
fact that t = v [cf. Formula (8.99) on page 129], we can rewrite expression (8.113) above
as
h
v
v
i
i
h v
v v
F=q E+v 2 E =q
(8.114)
c
c
c
c
c
Applying the baccab rule, Formula (F.51) on page 160, on the last term yields
v
v v2
v v
=
c
c
c
c c2
(8.115)
131
v2
c2
(8.116)
(8.117)
The scalar function is called the convection potential or the Heaviside potential. When the
rigid charge distribution is well localised so that we can use the potentials (8.112) the convection
potential becomes
v2
q0
= 1 2
(8.118)
c
40 s
The convection potential from a point charge is constant on flattened ellipsoids of revolution,
defined through Equation (8.110) on the previous page as
2
x1 x01
These Heaviside ellipsoids are equipotential surfaces, and since the force is proportional to the
gradient of , which means that it is perpendicular to the ellipsoid surface, the force between
two charges is in general not directed along the line which connects the charges. A consequence
of this is that a system consisting of two comoving charges connected with a rigid bar, will
experience a torque. This is the idea behind the TroutonNoble experiment, aimed at measuring
the absolute speed of the earth or the galaxy. The negative outcome of this experiment is
explained by the special theory of relativity which postulates that mechanical laws follow the
same rules as electromagnetic laws, so that a compensating torque appears due to mechanical
stresses within the chargebar system.
C E ND OF EXAMPLE 8.2
vc
(8.120)
x x0  v
x x0 ,
c
vc
(8.121)
so that the radiation field Equation (8.89) on page 126 can be approximated by
Erad (t, x) =
132
q0
(x x0 ) [(x x0 ) v ],
40 c2 x x0 3
vc
(8.122)
from which we obtain, with the use of Formula (8.88) on page 126, the magnetic field
Brad (t, x) =
40
q0
[v (x x0 )],
x x0 2
c3
vc
(8.123)
It is interesting to note the close correspondence which exists between the nonrelativistic fields (8.122) and (8.123) and the electric dipole field Equations (8.45) on
page 117 if we introduce
p = q0 x0 (t0 )
(8.124)
(8.125a)
x x = x x0
(8.125b)
The power flux in the far zone is described by the Poynting vector as a function
of Erad and Brad . We use the close correspondence with the dipole case to find that it
becomes
S=
0 q0 2 (v)2
x x0
2
2 sin x x0 
2
0
16 c x x 
(8.126)
0 q0 2 (v)2
q0 2 v2
=
6c
60 c3
(8.127)
where is the angle between v and x x0 . The total radiated power (integrated over a
closed spherical surface) becomes
P=
which is the Larmor formula for radiated power from an accelerated charge. Note that
here we are treating a charge with v c but otherwise totally unspecified motion while
we compare with formulae derived for a stationary oscillating dipole. The electric
and magnetic fields, Equation (8.122) on the preceding page and Equation (8.123),
respectively, and the expressions for the Poynting flux and power derived from them,
are here instantaneous values, dependent on the instantaneous position of the charge
at x0 (t0 ). The angular distribution is that which is frozen to the point from which the
energy is radiated.
8.3.3 Bremsstrahlung
An important special case of radiation is when the velocity v and the acceleration v are
collinear (parallel or antiparallel) so that v v = 0. This condition (for an arbitrary
magnitude of v) inserted into expression (8.89) on page 126 for the radiation field,
yields
Erad (t, x) =
q0
(x x0 ) [(x x0 ) v ],
40 c2 s3
v k v
(8.128)
133
v = 0.5c
v=0
v = 0.25c
v
F IGURE 8.8:
Polar diagram of the energy loss angular distribution factor
sin2 /(1 v cos /c)5 during bremsstrahlung for particle speeds v = 0, v = 0.25c,
and v = 0.5c.
from which we obtain, with the use of Formula (8.88) on page 126, the magnetic field
Brad (t, x) =
q0 x x0 
[v (x x0 )],
40 c3 s3
v k v
(8.129)
The difference between this case and the previous case of v c is that the approximate
expression (8.120) on page 132 for s is no longer valid; we must instead use the correct
expression (8.66) on page 121. The angular distribution of the power flux (Poynting
vector) therefore becomes
S=
0 q0 2 v2
sin2
x x0
2
6
2
0
16 c x x  1 v cos x x0 
c
(8.130)
It is interesting to note that the magnitudes of the electric and magnetic fields are the
same whether v and v are parallel or antiparallel.
We must be careful when we compute the energy (S integrated over time). The
Poynting vector is related to the time t when it is measured and to a fixed surface in
space. The radiated power into a solid angle element d, measured relative to the
particles retarded position, is given by the formula
dU rad ()
0 q0 2 v2
sin2
d = S (x x0 ) x x0 d =
d
dt
162 c 1 v cos 6
c
(8.131)
On the other hand, the radiation loss due to radiation from the charge at retarded time
t0 :
dU rad t
dU rad
d
(8.132)
d
=
dt0
dt
t0 x
134
dS
dr
x
d
0
q
0
x02 vdt x01
x x0 + c dt0
2
F IGURE 8.9: Location of radiation between two spheres as the charge moves
with velocity v from x01 to x02 during the time interval (t 0 , t0 + dt0 ). The observation
point (field point) is at the fixed location x.
(8.133)
Inserting Equation (8.130) on the facing page for S into (8.133), we obtain the
explicit expression for the energy loss due to radiation evaluated at the retarded time
0 q0 2 v2
dU rad ()
sin2
d
=
d
dt0
162 c 1 v cos 5
c
(8.134)
The angular factors of this expression, for three different particle speeds, are plotted
in Figure 8.8 on the preceding page.
Comparing expression (8.131) on the facing page with expression (8.134) above,
we see that they differ by a factor 1 v cos /c which comes from the extra factor
s/ x x0  introduced in (8.133). Let us explain this in geometrical terms.
During the interval (t0 , t0 + dt0 ) and within the solid angle element d the particle
radiates an energy [dU rad ()/dt0 ] dt0 d. As shown in 8.9 this energy is at time t located
between two spheres, one outer with its origin at x01 (t0 ) and radius c(t t0 ), and one
inner with its origin at x02 (t0 +dt0 ) = x01 (t0 )+v dt0 and radius c[t(t0 +dt0 )] = c(tt0 dt0 ).
From Figure 8.9 we see that the volume element subtending the solid angle element
dS
d =
x x0 2
(8.135)
135
is
2
d3x = dS dr = x x02 d dr
(8.136)
Here, dr denotes the differential distance between the two spheres and can be evaluated in the following way
x x02
v dt0
dr = x x02 + c dt0 x x02
x x02
 {z }
v cos
!
0
cs
x x2
v dt0 =
dt0
= c
0
x x2
x x02
(8.137)
where Formula (8.66) on page 121 was used in the last step. Hence, the volume
element under consideration is
s
dS cdt0
d3x = dS dr =
x x0
(8.138)
We see that the energy which is radiated per unit solid angle during the time interval
(t0 , t0 + dt0 ) is located in a volume element whose size is dependent. This explains
the difference between expression (8.131) on page 134 and expression (8.134) on the
preceding page.
Let the radiated energy, integrated over , be denoted U rad . After tedious, but
relatively straightforward integration of Formula (8.134) on the previous page, one
obtains
3
dU rad 0 q0 2 v2
1
v2
2 q0 2 v2
1 2
(8.139)
=
3 =
2
dt0
6c
3 40 c3
c
1 cv2
If we know v(t0 ), we can integrate this expression over t 0 and obtain the total energy
radiated during the acceleration or deceleration of the particle. This way we obtain a
classical picture of bremsstrahlung (braking radiation, freefree radiation). Often, an
atomistic treatment is required for obtaining an acceptable result.
E XAMPLE 8.3
(8.140)
136
v dt0
(8.141)
(8.142)
(8.143)
q0 sin
v (t0 t0 )
40 c2 x x0 
In this simple case B Brad is given by
(8.144)
E=
B=
E
c
(8.145)
q0 sin
v eit0
2
0
0 c x x 
(8.146)
82
We note that the magnitude of this Fourier component is independent of . This is a consequence of the infinitely short impulsive step (t 0 t0 ) in the time domain which produces an
infinite spectrum in the frequency domain.
The total radiation energy is given by the expression
Z
Z Z
dU rad 0
B
U rad =
E
dt
=
dS dt0
dt0
0
S
Z Z
Z Z
1
1
EB dt0 d2x0 =
E 2 dt0 d2x0
=
0 S
0 c S
= 0 c
Z Z
S
(8.147)
E 2 dt0 d2x0
According to Parsevals identity [cf. Equation (7.35) on page 102] the following equality
holds:
Z
E 2 dt0 = 4
E 2 d
(8.148)
(8.149)
Z
Z
sin2 2
d x d
x x0 2
Z
2
q (v)
sin2 sin d d
d
3
3
16 0 c 0
0
2
d
v
q02
=
30 c c
2
(8.150)
137
We see that the energy spectrum U rad is independent of frequency . This means that if we
would integrate it over all frequencies [0, ), a divergent integral would result.
In reality, all spectra have finite widths, with an upper cutoff limit set by the quantum
condition
1
~max = m(v)2
(8.151)
2
which expresses that the highest possible frequency max in the spectrum is that for which all
kinetic energy difference has gone into one single field quantum (photon) with energy ~ max .
If we adopt the picture that the total energy is quantised in terms of N photons radiated during
the process, we find that
U rad d
= dN
~
or, for an electron where q0 = e, where e is the elementary charge,
2
2
e2
2 v
1 2 v
d
d
dN =
40 ~c 3 c
137 3 c
(8.152)
(8.153)
where we used the value of the fine structure constant = e2 /(40 ~c) 1/137.
Even if the number of photons becomes infinite when 0, these photons have negligible
energies so that the total radiated energy is still finite.
C E ND OF EXAMPLE 8.3
(8.154a)
0
x (t ) = a[ x 1 cos (t ) + x 2 sin (t )]
0
0 0
(8.154b)
v (t ) =
v =
0 0
x (t ) = a20 [ x 1
v = a20
(8.154c)
(8.154d)
cos (t ) + x 2 sin (t )]
(8.154e)
(8.154f)
Because of the rotational symmetry we can, without loss of generality, rotate our
coordinate system around the x3 axis so the relative vector x x0 from the source
point to an arbitrary field point always lies in the x2 x3 plane, i.e.,
(8.155)
x x0 = x x0 ( x 2 sin + x 3 cos )
138
x2
(t, x)
x x0
x
v
q0 0 0
(t , x )
(t0 )
x1
x3
F IGURE 8.10: Coordinate system for the radiation from a charged particle at
x0 (t0 ) in circular motion with velocity v(t 0 ) along the tangent and constant acceleration v (t0 ) toward the origin. The x1 x2 axes are chosen so that the relative field
point vector x x0 makes an angle with the x3 axis which is normal to the plane
of the orbital motion. The radius of the orbit is a.
where is the angle between x x0 and the normal to the plane of the particle orbit
(see Figure 8.10). From the above expressions we obtain
(x x0 ) v = x x0 v sin cos
(x x0 ) v = x x0 v sin sin = x x0 v cos
(8.156a)
(8.156b)
where in the last step we simply used the definition of a scalar product and the fact
that the angle between v and x x0 is .
The power flux is given by the Poynting vector, which, with the help of Formula (8.88) on page 126, can be written
S=
x x0
1
1
(E B) =
E2
0
c0
x x0 
(8.157)
(8.158)
where the retarded distance s is given by expression (8.66) on page 121. With the
radiation part of the electric field, expression (8.89) on page 126, inserted, and using
139
(8.156a) and (8.156b) on the previous page, one finds, after some algebra, that
2
v2
v
sin2 sin2
rad
0 2 2 1 c sin cos
2
c
dU (, ) 0 q v
(8.159)
=
5
dt0
162 c
1 v sin cos
c
The angles and vary in time during the rotation, so that refers to a moving
coordinate system. But we can parametrise the solid angle d in the angle and the
(fixed) angle so that d = sin d d. Integration of Equation (8.159) over this d
gives, after some cumbersome algebra, the angular integrated expression
dU rad 0 q0 2 v2
=
dt0
6c
1
1
v2
c2
(8.160)
2
Cyclotron radiation
For a nonrelativistic speed v c, Equation (8.159) reduces to
dU rad (, ) 0 q0 2 v2
=
(1 sin2 sin2 )
dt0
162 c
(8.161)
(8.162)
where is defined in Figure 8.10 on the preceding page. This means that we can write
0 q0 2 v2
dU rad () 0 q0 2 v2
2
=
(1
cos
)
=
sin2
dt0
162 c
162 c
(8.163)
Consequently, a fixed observer near the orbit plane will observe cyclotron radiation twice per revolution in the form of two equally broad pulses of radiation with
alternating polarisation.
Synchrotron radiation
When the particle is relativistic, v . c, the denominator in Equation (8.159) above
becomes very small if sin cos 1, which defines the forward direction of the
particle motion ( /2, 0). Equation (8.159) then becomes
dU rad (/2, 0) 0 q0 2 v2
1
=
dt0
162 c 1 v 3
c
140
(8.164)
x2
(t, x)
x x0
v
q0 0 0
(t , x )
v
(t0 )
0
x1
x3
F IGURE 8.11:
which means that an observer near the orbit plane sees a very strong pulse followed,
half an orbit period later, by a much weaker pulse.
The two cases represented by Equation (8.163) on the preceding page and Equation (8.164) on the facing page are very important results since they can be used to
determine the characteristics of the particle motion both in particle accelerators and in
astrophysical objects where a direct measurement of particle velocities are impossible.
In the orbit plane ( = /2), Equation (8.159) on the preceding page gives
2
2
v
v2
dU rad (/2, ) 0 q0 2 v2 1 c cos 1 c2 sin
=
(8.165)
5
dt0
162 c
1 v cos
c
v
cr
(8.166a)
1
v2
c2
(8.166b)
Hence, the angle 0 is a measure of the synchrotron radiation lobe width ; see
Figure 8.11. For ultrarelativistic particles, defined by
r
1
v2
1,
= q
1 2 1,
(8.167)
2
c
1 v
c2
141
v2
1
=
2
c
(8.168)
(8.169)
This angular interval is swept by the charge during the time interval
t0 =
(8.170)
(8.171)
in the direction toward the observer who therefore measures a pulse width of length
l
v
vt0
v 0
v 1
t = t0
t = 1
= t0
= 1
1
c
c
c
c 0
c 0
v
v
2
1 c 1+ c
1
1
v
1 1
1 2
= 3
=
v
c
2
2
0
0
0
1+
 {z }
c
 {z }
1/2
2
(8.172)
Typically, the spectral width of a pulse of length t is . 1/t. In the ultrarelativistic synchrotron case one can therefore expect frequency components up to
max
1
= 23 0
t
(8.173)
A spectral analysis of the radiation pulse will therefore exhibit a (broadened) line
spectrum of Fourier components n0 from n = 1 up to n 23 .
When many charged particles, N say, contribute to the radiation, we can have
three different situations depending on the relative phases of the radiation fields from
the individual particles:
1. All N radiating particles are spatially much closer to each other than a typical
wavelength. Then the relative phase differences of the individual electric and
magnetic fields radiated are negligible and the total radiated fields from all individual particles will add up to become N times that from one particle. This
means that the power radiated from the N particles will be N 2 higher than for a
single charged particle. This is called coherent radiation.
142
2. The charged particles are perfectly evenly distributed in the orbit. In this case
the phases of the radiation fields cause a complete cancellation of the fields
themselves. No radiation escapes.
3. The charged particles are somewhat unevenly distributed in the orbit. This happens for an open ring current, carried initially by evenly distributed charged
particles, which is subject to thermal fluctuations. From statistical mechanics
we know that this happens for
(8.175)
Virtual photons
Let us consider a charge q0 moving with constant, high velocity v(t 0 ) along the x1
axis. According to Formula (8.102) on page 129 and Figure 8.12 on the following
page, the perpendicular component along the x3 axis of the electric field from this
moving charge is
q0
v2
E = E3 =
1
(x x0 ) x 3
(8.176)
40 s3
c2
(8.177)
143
vt
q0
v = v x 1
x x0 
B
E x 3
F IGURE 8.12:
Utilising expression (8.97a) on page 128 and simple geometrical relations, we can
rewrite this as
E =
q0
b
40 2 (vt)2 + b2 /2 3/2
(8.178)
This represents a contracted field, approaching the field of a plane wave. The passage
of this field pulse corresponds to a frequency distribution of the field energy. Fourier
transforming, we obtain
Z
q0
1
b
b
it
(t)
E , =
dt E e = 2
K1
(8.179)
2
4 0 bv
v
v
Here, K1 is the Kelvin function (Bessel function of the second kind with imaginary
argument) which behaves in such a way for small and large arguments that
q0
, b v
42 0 bv
0, b v
E ,
(8.180a)
E ,
(8.180b)
E 2 d3x = 0
bmaxZ
bmin
E 2 vdt 2b db
(8.181)
where the volume integration is over the plane perpendicular to v. With the use of
Parsevals identity for Fourier transforms, Formula (7.35) on page 102, we can rewrite
144
this as
U =
U d = 40 v
0
02
q
22 0 v
Z Z
0
bmaxZ
bmin
v/
bmin
db
d
b
E , 2 d 2b db
ln
U 2
2 0 v
bmin
(8.182)
(8.183)
bmin
where = e2 /(40 ~c) 1/137 is the fine structure constant.
Let us consider the interaction of two (classical) electrons, 1 and 2. The result of
0
this interaction is that they change their linear momenta from p1 to p01 and
p2 to p2 ,
0
respectively. Heisenbergs uncertainty principle gives bmin ~/ p1 p1 so that the
number of photons exchanged in the process is of the order
N d
d
2 c
ln
p1 p01
(8.185)
m0 c2 E 1 E 10
a formula which gives a reasonable semiclassical account of a photoninduced electronelectron interaction process. In quantum theory, including only the lowest order contributions, this process is known as Mller scattering. A diagrammatic representation
of (a semiclassical approximation of) this process is given in Figure 8.13 on the following page.
145
p02
p2
p1
F IGURE 8.13:
p01
from matter on the electromagnetic fields by introducing new, derived field quantities
D and H according to
D = (t, x)E = 0 E
(8.187)
B = (t, x)H = m 0 H
(8.188)
Expressed in terms of these derived field quantities, the Maxwell equations, often
called macroscopic Maxwell equations, take the form
D = (t, x)
(8.189a)
B
E=
(8.189b)
t
B=0
(8.189c)
D
H=
+ j(t, x)
(8.189d)
t
Assuming for simplicity that the electric permittivity and the magnetic permeability , and hence the relative permittivity and the relative permeability m all have
fixed values, independent on time and space, for each type of material we consider,
we can derive the general telegraphers equation [cf. Equation (2.33) on page 31]
2 E
E
2 E
=0
2
t
t2
(8.190)
=0
2
t2
146
(8.191)
0 m 0
m
(8.192)
i = 1, 2, 3
(8.193)
c
def
= m = c n
v
(8.194)
is called the refractive index of the medium. In general n is a function of both time and
space as are the quantities , , , and m themselves. If, in addition, the medium is
anisotropic or birefringent, all these quantities are ranktwo tensor fields. Under our
simplifying assumptions, in each medium we consider n = Const for each frequency
component of the fields.
Associated with the phase speed of a medium for a wave of a given frequency
we have a wave vector, defined as
def
k k k = kv =
v
v v
(8.195)
(8.196)
where now k is the wave vector in the medium given by Equation (8.195). With these
definitions, the vacuum formula for the associated magnetic field, Equation (2.40) on
page 31,
B=
k E =
1
1
kE= kE
v
(8.197)
is valid also in a material medium (assuming, as mentioned, that n has a fixed constant
scalar value). A consequence of a , 1 is that the electric field will, in general, have
a longitudinal component.
It is important to notice that depending on the electric and magnetic properties of
a medium, and, hence, on the value of the refractive index n, the phase speed in the
medium can be smaller or larger than the speed of light:
v =
c
=
n
k
(8.198)
147
where, in the last step, we used Equation (8.195) on the preceding page.
If the medium has a refractive index which, as is usually the case, dependent on
frequency , we say that the medium is dispersive. Because in this case also k() and
(k), so that the group velocity
vg =
(8.199)
has a unique value for each frequency component, and is different from v . Except in
regions of anomalous dispersion, v is always smaller than c. In a gas of free charges,
such as a plasma, the refractive index is given by the expression
n2 () = 1
2p
2
(8.200)
where
2p =
N q2
0 m
(8.201)
is the square of the plasma frequency p . Here m and N denote the mass and number
density, respectively, of charged particle species . In an inhomogeneous plasma,
N = N (x) so that the refractive index and also the phase and group velocities are
space dependent. As can be easily seen, for each given frequency, the phase and group
velocities in a plasma are different from each other. If the frequency is such that
it coincides with p at some point in the medium, then at that point v while
vg 0 and the wave Fourier component at is reflected there.
VavilovCerenkov
radiation
As we saw in Subsection 8.1, a charge in uniform, rectilinear motion in vacuum does
not give rise to any radiation; see in particular Equation (8.100a) on page 129. Let us
now consider a charge in uniform, rectilinear motion in a medium with electric properties which are different from those of a (classical) vacuum. Specifically, consider a
medium where
= Const > 0
(8.202a)
= 0
(8.202b)
c
1
=
<c
n
0
(8.203)
Hence, in this particular medium, the speed of propagation of (the phase planes of)
electromagnetic waves is less than the speed of light in vacuum, which we know is an
148
absolute limit for the motion of anything, including particles. A medium of this kind
has the interesting property that particles, entering into the medium at high speeds
v, which, of course, are below the phase speed in vacuum, can experience that the
particle speeds are higher than the phase speed in the medium. This is the basis for
the VavilovCerenkov
radiation that we shall now study.
If we recall the general derivation, in the vacuum case, of the retarded (and advanced) potentials in Chapter 3 and the LinardWiechert potentials, Equations (8.65)
on page 121, we realise that we obtain the latter in the medium by a simple formal
replacement c c/n in the expression (8.66) on page 121 for s. Hence, the LinardWiechert potentials in a medium characterized by a refractive index n, are
q0
1 q0
1
=
0
40 x x0  n (xxc )v 40 s
q0 v
q0 v
1
1
A(t, x) =
=
0
40 c2 x x0  n (xxc )v 40 c2 s
(t, x) =
where now
(x x0 ) v
s = x x0 n
c
(8.204a)
(8.204b)
(8.205)
The need for the absolute value of the expression for s is obvious in the case when
v/c 1/n because then the second term can be larger than the first term; if v/c 1/n
we recover the wellknown vacuum case but with modified phase speed. We also
note that the retarded and advanced times in the medium are [cf. Equation (3.30) on
page 41]
k x x0 
x x0  n
0
0
=t
tret
= tret
(t, x x0 ) = t
c
0
x0  n
k
x
x
x

0
0
tadv
= tadv
(t, x x0 ) = t +
=t+
(8.206a)
(8.206b)
so that the usual time interval t t 0 between the time measured at the point of observation and the retarded time in a medium becomes
t t0 =
x x0  n
c
(8.207)
For v/c 1/n, the retarded distance s, and therefore the denominators in Equations (8.204) above vanish when
n(x x0 )
nv
v
= x x0 cos c = x x0
c
c
(8.208)
c
nv
(8.209)
149
x(t)
q0
x0 (t0 )
F IGURE 8.14: Instantaneous picture of the expanding field spheres from a point
charge moving with constant speed v/c > 1/n in a medium where n > 1. This
generates a VavilovCerenkov
shock wave in the form of a cone.
In the direction defined by this angle c , the potentials become singular. During the
time interval t t0 given by expression (8.207) on the preceding page, the field exists
within a sphere of radius x x0  around the particle while the particle moves a distance
l = v(t t0 )
(8.210)
c
nv
(8.211)
This cone of potential singularities and field sphere circumferences propagates with
surprised that his old 1904 ideas were now becoming interesting. Tamm, Frank and Cerenkov
received
the Nobel Prize in 1958 for the discovery and the interpretation of the Cerenkov
effect [V. L. Ginzburg,
150
In order to make some quantitative estimates of this radiation, we note that we can
describe the motion of each charged particle q0 as a current density:
j = q0 v (x0 vt0 ) = q0 v (x0 vt0 )(y0 )(z0 ) x 1
(8.212)
of the VavilovCerenkov
magnetic and electric radiation fields can be calculated from
the expressions (7.11) on page 95) and (7.22) on page 98, respectively.
The total energy content is then obtained from Equation (7.35) on page 102 (integrated over a closed sphere at large distances). For a Fourier component one obtains
[cf. Equation (7.38) on page 103]
Z
2
1
ikx0 3 0
Urad d
(j
k)e
d
x
d
40 nc V
(8.215)
Z
2
0
q0 2 n2
x
0
0
2
=
kx cos
dx sin d
exp i
163 0 c3
v
where is the angle between the direction of motion, x 01 , and the direction to the
The integral in (8.215) is singular of a Dirac delta type. If we limit the
observer, k.
spatial extent of the motion of the particle to the closed interval [X, X] on the x 0 axis
we can evaluate the integral to obtain
q0 2 n2 sin2 sin2 1 nvc cos X
rad
v
(8.216)
U d =
2 d
43 0 c3
1 nv cos
c
which has a maximum in the direction c as expected. The magnitude of this maximum grows and its width narrows as X . The integration of (8.216) over
therefore picks up the main contributions from c . Consequently, we can set
sin2 sin2 c and the result of the integration is
U rad = 2
Urad () sin d
q0 2 n2 sin2 c
22 0 c3
= dcos = c = 2
X
sin2 1 + nv
c
v
2 d
1 + nv
c
v
Urad () d
(8.217)
private communication].
The first observation of this type of radiation was reported by Marie Curie in 1910, but she never pursued
the exploration of it [8].
151
The integrand in (8.217) is strongly peaked near = c/(nv), or, equivalently, near
cos c = c/(nv). This means that the integrand function is practically zero outside
the integration interval [1, 1]. Consequently, one may extend the integration
interval to (, ) without introducing too much an error. Via yet another variable
substitution we can therefore approximate
X
Z
Z 1
sin2 1 + nv
cX sin2 x
c2
2
c
v
dx
sin c
2 d 1 2 2
nv
n x2
1
1 + nv
c
v
(8.218)
c2
cX
1 2 2
=
n
n v
leading to the final approximate result for the total energy loss in the frequency interval
(, + d)
q0 2 X
c2
U rad d =
1
d
(8.219)
20 c2
n 2 v2
As mentioned earlier, the refractive index is usually frequency dependent. Realising this, we find that the radiation energy per frequency unit and per unit length
is
U rad d
q0 2
c2
=
1
d
(8.220)
2X
40 c2
n2 ()v2
This result was derived under the assumption that v/c > 1/n(), i.e., under the condition that the expression inside the parentheses in the right hand side is positive. For
all media it is true that n() 1 when , so there exist always a highest
8.4 Bibliography
[1]
H. A LFVN AND N. H ERLOFSON, Cosmic radiation and radio stars, Physical Review, 78
(1950), p. 616.
152
[2]
[3]
M. B ORN AND E. W OLF, Principles of Optics. Electromagnetic Theory of Propagation, Interference and Diffraction of Light, sixth ed., Pergamon Press, Oxford,. . . , 1980,
ISBN 0080264816.
Bibliography
[4]
V. L. G INZBURG, Applications of Electrodynamics in Theoretical Physics and Astrophysics, Revised third ed., Gordon and Breach Science Publishers, New York, London, Paris,
Montreux, Tokyo and Melbourne, 1989, ISBN 2881247199.
[5]
J. D. JACKSON, Classical Electrodynamics, third ed., John Wiley & Sons, Inc.,
New York, NY . . . , 1999, ISBN 047130932X.
[6]
J. B. M ARION AND M. A. H EALD, Classical Electromagnetic Radiation, second ed., Academic Press, Inc. (London) Ltd., Orlando, . . . , 1980, ISBN 0124722571.
[7]
[8]
J. S CHWINGER , L. L. D E R AAD , J R ., K. A. M ILTON , AND W. T SAI, Classical Electrodynamics, Perseus Books, Reading, MA, 1998, ISBN 0738200565.
[9]
[10] J. VANDERLINDE, Classical Electromagnetic Theory, John Wiley & Sons, Inc., New
York, Chichester, Brisbane, Toronto, and Singapore, 1993, ISBN 0471572691.
153
APPENDIX F
Formulae
F.1 The electromagnetic field
F.1.1 Maxwells equations
D=
(F.1)
B=0
(F.2)
E=
(F.3)
B
t
H=j+ D
t
(F.4)
Constitutive relations
D = E
B
H=
(F.5)
(F.6)
j = E
(F.7)
P = 0 E
(F.8)
(F.9)
155
F. F ORMULAE
E =
A
t
(F.10)
1
=0
c2 t
(F.11)
(F.12)
(F.13)
k E
c
(F.14)
i0 eikx
0
d3 x0 eikx j k
4 x V 0
Z
i eikx
0
d3 x0 eikx j k
Erad
x
(x) =
0
40 c x
V
Brad
(x) =
(F.15)
(F.16)
40 x
Brad
(x) =
156
(F.17)
(F.18)
Special relativity
(F.19)
(F.20)
(F.21)
(F.22)
x
)
1
+
(x
x
)
0
40 s3
c2
c2
E(t, x)
B(t, x) = (x x0 )
cx x0 
E(t, x) =
v
s = x x0 (x x0 )
c
(F.23)
(F.24)
(F.25)
v
x x0 = (x x0 ) x x0 
c
0
x x0 
t
=
t x
s
(F.26)
(F.27)
1 0
0 1
=
0 0
0 0
0
0
1
0
0
0
0
1
(F.28)
(F.29)
157
F. F ORMULAE
=
0
0
1
0
0
0
1
= p
1 2
v
=
c
(F.30)
0
0
0
1
(F.31)
(F.32)
(F.33)
dt
= c d
(F.34)
F.3.5 Fourvelocity
u =
dx
= (c, v)
d
(F.35)
F.3.6 Fourmomentum
p = m0 u =
E
,p
c
(F.36)
(F.37)
F.3.8 Fourpotential
A =
,A
c
(F.38)
158
E x /c
= A A =
E y /c
E z /c
E x /c E y /c E z /c
0
Bz
By
Bz
0
B x
By
Bx
0
(F.39)
Vector relations
x i
i=1
def
def
x i
xi
xi
(F.40)
,
,
, ,
=
(F.41)
i = i =
x1 x2 x3
x y z
(F.42a)
(F.42b)
= sin x 1 + cos x 2
(F.42c)
(F.43a)
x 3 = cos r sin
(F.43b)
(F.43c)
(F.44)
(F.45)
(F.46)
Volume element
d3x = dV = drdS = r 2 dr d
(F.47)
159
F. F ORMULAE
(F.48)
a b = b a = i jk a j bk x i
(F.49)
a (b c) = (a b) c
(F.50)
a (b c) + b (c a) + c (a b) = 0
(F.52)
a (b c) = b(a c) c(a b)
(F.51)
(a b) (c d) = (a b d)c (a b c)d
(F.53)
(F.54)
(F.55)
(a) = a a
(F.57)
(a b) = a( b) b( a) + (b )a (a )b
(F.59)
(a) = a + a
(F.56)
(a b) = b ( a) a ( b)
(F.58)
(a b) = a ( b) + b ( a) + (b )a + (a )b
(F.60)
(F.61)
= 0
(F.62)
( a) = 0
(F.63)
2
( a) = ( a) a
(F.64)
Special identities
In the following x = xi x i and x0 = x0i x i are radius vectors, k an arbitrary constant
vector, a = a(x) an arbitrary vector field, x i x i , and 0 x 0 x i .
i
x=3
(F.65)
(k x) = k
x
x =
x
x x0
x x0  =
= 0 x x0 
0
x x 
x
1
= 3
x
x
(F.67)
x=0
160
(F.66)
(F.68)
(F.69)
(F.70)
Bibliography
1
x x0
1
0
=
x x0 3
x x0 
x x0 
1
x
= 2
= 4(x)
3
x
x
1
x x0
2
=
= 4(x x0 )
x x0 3
x x0 
kx
k
1
=k
= 3
x
x
x
x
kx
k
=
if x , 0
x3
x3
1
k
= k2
= 4k(x)
2
x
x
(k a) = k( a) + k ( a) (k a)
(F.71)
(F.72)
(F.73)
(F.74)
(F.75)
(F.76)
(F.77)
Integral relations
Let V(S ) be the volume bounded by the closed surface S (V). Denote the 3dimensional
volume element by d3x( dV) and the surface element, directed along the outward
pointing surface normal unit vector n,
by dS( d2x n).
Then
Z
ZV
( a) d3x =
() d3x =
I
S
( a) d3x =
dS a
(F.78)
dS
(F.79)
(F.80)
dS a
If S (C) is an open surface bounded by the contour C(S ), whose line element is dl,
then
I
IC
C
dl =
a dl =
ZS
dS
S
(F.81)
dS ( a)
(F.82)
F.5 Bibliography
[1] G. B. A RFKEN AND H. J. W EBER, Mathematical Methods for Physicists, fourth, international ed., Academic Press, Inc., San Diego, CA . . . , 1995, ISBN 0120598167.
[2] P. M. M ORSE AND H. F ESHBACH, Methods of Theoretical Physics, Part I. McGrawHill
Book Company, Inc., New York, NY . . . , 1953, ISBN 070433168.
161
F. F ORMULAE
[3] W. K. H. PANOFSKY AND M. P HILLIPS, Classical Electricity and Magnetism, second ed.,
AddisonWesley Publishing Company, Inc., Reading, MA . . . , 1962, ISBN 0201057026.
162
APPENDIX M
Mathematical Methods
M.1 Scalars, vectors and tensors
Every physical observable can be described by a geometric object. We will describe the observables in classical electrodynamics mathematically in terms of scalars,
pseudoscalars, vectors, pseudovectors, tensors or pseudotensors and will not exploit
differential forms to any significant degree.
A scalar describes a scalar quantity which may or may not be constant in time
and/or space. A vector describes some kind of physical motion due to vection and a
tensor describes the motion or deformation due to some form of tension. However,
generalisations to more abstract notions of these quantities are commonplace. The
difference between a scalar, vector and tensor and a pseudoscalar, pseudovector and
a pseudotensor is that the latter behave differently under such coordinate transformations which cannot be reduced to pure rotations.
Throughout we adopt the convention that Latin indices i, j, k, l, . . . run over the
range 1, 2, 3 to denote vector or tensor components in the real Euclidean threedimensional (3D) configuration space R3 , and Greek indices , , , , . . . , which are
used in fourdimensional (4D) space, run over the range 0, 1, 2, 3.
M.1.1 Vectors
Radius vector
A vector can be represented mathematically in a number of different ways. One suitable representation is in terms of an ordered Ntuple, or row vector, of the coordinates
xN where N is the dimensionality of the space under consideration. The most basic
163
M. M ATHEMATICAL M ETHODS
vector is the radius vector which is the vector from the origin to the point of interest. Its Ntuple representation simply enumerates the coordinates which describe
this point. In this sense, the radius vector from the origin to a point is synonymous
with the coordinates of the point itself.
In the 3D Euclidean space R3 , we have N = 3 and the radius vector can be represented by the triplet (x1 , x2 , x3 ) of coordinates xi , i = 1, 2, 3. The coordinates xi are
scalar quantities which describe the position along the unit base vectors x i which span
R3 . Therefore a representation of the radius vector in R3 is
3
x = xi x i xi x i
def
(M.1)
i=1
where we have introduced Einsteins summation convention (E) which states that a
repeated index in a term implies summation over the range of the index in question.
Whenever possible and convenient we shall in the following always assume E and
suppress explicit summation in our formulae. Typographically, we represent a vector
in 3D Euclidean space R3 by a boldface letter or symbol in a Roman font.
Alternatively, we may describe the radius vector in component notation as follows:
def
xi (x1 , x2 , x3 ) (x, y, z)
(M.2)
x (x0 , x1 , x2 , x3 )
(M.3)
x (x0 , x1 , x2 , x3 )
(M.4)
The relation between the covariant and contravariant forms is determined by the metric tensor (also known as the fundamental tensor) whose actual form is dictated by the
properties of the vector space in question. The dual representation of vectors in contravariant and covariant forms is most convenient when we work in a nonEuclidean
vector space with an indefinite metric. An example is Lorentz space L 4 which is a 4D
Riemannian space utilised to formulate the special theory of relativity.
We note that for a change of coordinates x x0 = x0 (x0 , x1 , x2 , x3 ), due to a
transformation from a system to another system 0 , the differential radius vector dx
transforms as
dx0 =
x0
dx
x
(M.5)
164
M.1.2 Fields
A field is a physical entity which depends on one or more continuous parameters. Such
a parameter can be viewed as a continuous index which enumerates the coordinates
of the field. In particular, in a field which depends on the usual radius vector x of R 3 ,
each point in this space can be considered as one degree of freedom so that a field is a
representation of a physical entity which has an infinite number of degrees of freedom.
Scalar fields
We denote an arbitrary scalar field in R3 by
def
(M.6)
This field describes how the scalar quantity varies continuously in 3D R 3 space.
In 4D, a fourscalar field is denoted
def
(x0 , x1 , x2 , x3 ) (x )
(M.7)
which indicates that the fourscalar depends on all four coordinates spanning this
space. Since a fourscalar has the same value at a given point regardless of coordinate
system, it is also called an invariant.
Analogous to the transformation rule, Equation (M.5) on the facing page, for the
differential dx , the transformation rule for the differential operator /x under a
transformation x x0 becomes
x
= 0
0
x
x x
(M.8)
Vector fields
We can represent an arbitrary vector field a(x) in R3 as follows:
a(x) = ai (x) x i
(M.9)
(M.10)
(M.11)
(M.12)
165
M. M ATHEMATICAL M ETHODS
where x is the radius fourvector. Again, the relation between a and a is determined
by the metric of the physical 4D system under consideration.
Whether an arbitrary Ntuple fulfils the requirement of being an (Ndimensional)
contravariant vector or not, depends on its transformation properties during a change
of coordinates. For instance, in 4D an assemblage y = (y0 , y1 , y2 , y3 ) constitutes a
contravariant fourvector (or the contravariant components of a fourvector) if and
only if, during a transformation from a system with coordinates x to a system 0
with coordinates x0 , it transforms to the new system according to the rule
y0 =
x0
y
x
(M.13)
i.e., in the same way as the differential coordinate element dx transforms according
to Equation (M.5) on page 164.
The analogous requirement for a covariant fourvector is that it transforms, during
the change from to 0 , according to the rule
y0 =
x
y
x0
(M.14)
i.e., in the same way as the differential operator /x transforms according to Equation (M.8) on the preceding page.
Tensor fields
We denote an arbitrary tensor field in R3 by A(x). This tensor field can be represented
in a number of ways, for instance in the following matrix form:
Ai j (xk )
A11 (x)
A21 (x)
A31 (x)
def
A12 (x)
A22 (x)
A32 (x)
A13 (x)
A23 (x)
A33 (x)
(M.15)
Strictly speaking, the tensor field described here is a tensor of rank two.
A particularly simple ranktwo tensor in R3 is the 3D Kronecker delta symbol i j ,
with the following properties:
i j =
(
0 if i , j
(M.16)
1 if i = j
1 0 0
(i j ) = 0 1 0
0 0 1
166
(M.17)
Another common and useful tensor is the fully antisymmetric tensor of rank 3,
also known as the LeviCivita tensor
1
(M.18)
i jk = 0
if at least two of i, j, k are equal
(M.19)
In fact, tensors may have any rank n. In this picture a scalar is considered to
be a tensor of rank n = 0 and a vector a tensor of rank n = 1. Consequently, the
notation where a vector (tensor) is represented in its component form is called the
tensor notation. A tensor of rank n = 2 may be represented by a twodimensional
array or matrix whereas higher rank tensors are best represented in their component
forms (tensor notation).
BT ENSORS IN 3D SPACE
Consider a tetrahedronlike volume element V of a solid, fluid, or gaseous body, whose
atomistic structure is irrelevant for the present analysis; figure M.1 on the next page indicates
how this volume may look like. Let dS = d2x n be the directed surface element of this volume
element and let the vector T n d2x be the force that matter, lying on the side of d2x toward which
the unit normal vector n points, acts on matter which lies on the opposite side of d 2x. This force
concept is meaningful only if the forces are shortrange enough that they can be assumed to act
only in the surface proper. According to Newtons third law, this surface force fulfils
T n = T n
E XAMPLE 13.1
(M.20)
Using (M.20) and Newtons second law, we find that the matter of mass m, which at a given
instant is located in V obeys the equation of motion
T n d2x cos 1 T x 1 d2x cos 2 T x 2 d2x cos 3 T x 3 d2x + Fext = ma
(M.21)
where Fext is the external force and a is the acceleration of the volume element. In other words
Fext
m
(M.22)
T n = n1 T x 1 + n2 T x 2 + n3 T x 3 + 2 a
dx
m
Since both a and Fext /m remain finite whereas m/d2x 0 as V 0, one finds that in this limit
3
T n = ni T x i ni T x i
(M.23)
i=1
From the above derivation it is clear that Equation (M.23) is valid not only in equilibrium but
also when the matter in V is in motion.
Introducing the notation
T i j = T x i j
(M.24)
for the jth component of the vector T x i , we can write Equation (M.23) above in component
167
M. M ATHEMATICAL M ETHODS
x3
d2x
x2
V
x1
Terahedronlike volume element V containing matter.
F IGURE M.1:
form as follows
3
T n j = (T n ) j = ni T i j ni T i j
(M.25)
i=1
Using Equation (M.25), we find that the component of the vector T n in the direction of an
arbitrary unit vector m
is
T n m = T n m
= T n j m j =
j=1
j=1
ni T i j
i=1
m j ni T i j m j = n T m
(M.26)
Hence, the jth component of the vector T x i , here denoted T i j , can be interpreted as the i jth
component of a tensor T. Note that T n m is independent of the particular coordinate system used
in the derivation.
We shall now show how one can use the momentum law (force equation) to derive the
equation of motion for an arbitrary element of mass in the body. To this end we consider a
part V of the body. If the external force density (force per unit volume) is denoted by f and the
velocity for a mass element dm is denoted by v, we obtain
d
dt
168
v dm =
V
f d3x +
T n d2x
(M.27)
d
v j dm =
dt
f j d3x +
T n j d2x =
S
f j d3x +
V
ni T i j d2x
(M.28)
where, in the last step, Equation (M.25) on the preceding page was used. Setting dm = d3x
and using the divergence theorem on the last term, we can rewrite the result as
Z
d
v j d3x =
dt
f j d3x +
V
T i j 3
dx
xi
(M.29)
Since this formula is valid for any arbitrary volume, we must require that
T i j
d
vj fj
=0
dt
xi
(M.30)
or, equivalently
T i j
v j
+ v v j f j
=0
t
xi
(M.31)
Note that v j /t is the rate of change with time of the velocity component v j at a fixed point
x = (x1 , x1 , x3 ).
C E ND OF EXAMPLE 13.1
The 4D metric tensor (fundamental tensor) mentioned above is a particularly important fourtensor of rank 2. In covariant component form we shall denote it g . This
metric tensor determines the relation between an arbitrary contravariant fourvector a
and its covariant counterpart a according to the following rule:
def
a (x ) g a (x )
(M.32)
This rule is often called lowering of index. The raising of index analogue of the index
lowering rule is:
def
a (x ) g a (x )
(M.33)
More generally, the following lowering and raising rules hold for arbitrary rank n
mixed tensor fields:
1 2 ...k1
1 2 ...k1 k
gk k Ak+1
k+2 ...n (x ) = Ak k+1 ...n (x )
(M.34)
1 2 ...k1 k
2 ...k1
gk k A1k k+1
...n (x ) = Ak+1 k+2 ...n (x )
(M.35)
169
M. M ATHEMATICAL M ETHODS
Successive lowering and raising of more than one index is achieved by a repeated
application of this rule. For example, a dual application of the lowering operation on
a rank 2 tensor in contravariant form yields
A = g g A
(M.36)
i.e., the same rank 2 tensor in covariant form. This operation is also known as a tensor
contraction.
E XAMPLE 13.2
if = = 0
1
(M.37)
g = 1 if = = i = j = 1, 2, 3
0
if ,
which, in matrix notation, is represented as
1 0
0
0
0 1 0
0
(g ) =
0 0 1 0
0 0
0 1
(M.38)
1 0 0 0
0 1 0 0
(g ) =
0 0 1 0
0 0 0 1
(M.40)
i.e., a matrix with a main diagonal that has the sign sequence, or signature, {+, , , } or
1 if = = 0
g = 1
(M.39)
if = = i = j = 1, 2, 3
0
if ,
a (a0 , a1 , a2 , a3 ) = (a0 , a)
(M.41)
According to the index lowering rule, Equation (M.32) on the preceding page, we obtain the
170
a (a0 , a1 , a2 , a3 ) = g a
(M.42)
a 0 = 1 a 0 + 0 a 1 + 0 a 2 + 0 a 3 = a0
(M.43)
(M.45)
a3 = 0 a0 + 0 a1 + 0 a2 + 1 a3 = a3
(M.46)
a1 = 0 a 1 a + 0 a + 0 a = a
a2 = 0 a + 0 a 1 a + 0 a = a
(M.44)
or
a = (a0 , a1 , a2 , a3 ) = (a0 , a1 , a2 , a3 ) = (a0 , a)
(M.47)
(M.48)
(M.49)
C E ND OF EXAMPLE 13.2
(M.50)
where we used the fact that the scalar product x i x j is a representation of the Kronecker
delta i j defined in Equation (M.16) on page 166. In Russian literature, the 3D scalar
product is often denoted (ab). The scalar product of a in R3 with itself is
def
(M.51)
(M.52)
171
M. M ATHEMATICAL M ETHODS
(M.53)
where we made use of the index lowering and raising rules (M.32) and (M.33). The
result is a fourscalar, i.e., an invariant which is independent of in which 4D coordinate
system it is measured.
The quadratic differential form
ds2 = g dx dx = dx dx
(M.54)
i.e., the scalar product of the differential radius fourvector with itself, is an invariant
called the metric. It is also the square of the line element ds which is the distance
between neighbouring points with coordinates x and x + dx .
E XAMPLE 13.3
(M.55)
def
(M.56)
(M.57)
Using this in Equation (M.55), we see that we can interpret this so that the complex unit vector
is
aR
A
aI
= p 2
A =
a R + i p 2
a I
2
2
A
aR aI + 2iaR aI
aR aI + 2iaR aI
p
p
(M.58)
aR a2R a2I 2iaR aI
aI a2R a2I 2iaR aI
3
=
a
+
i
a
C
R
I
a2R + a2I
a2R + a2I
On the other hand, the definition of the scalar product in terms of the inner product of complex
vector with its own complex conjugate yields
def
(M.59)
(M.60)
C E ND OF EXAMPLE 13.3
172
E XAMPLE 13.4
In L the metric tensor attains a simple form [see Example 13.2 on page 170] and, hence,
the scalar product in Equation (M.53) on the facing page can be evaluated almost trivially. For
the {+, , , } signature it becomes
4
a b = (a0 , a) (b0 , b) = a0 b0 a b
(M.61)
The important scalar product of the L4 radius fourvector with itself becomes
x x = (x0 , x) (x0 , x) = (ct, x) (ct, x)
(M.62)
which is the indefinite, real norm of L4 . The L4 metric is the quadratic differential form
ds2 = dx dx = c2 (dt)2 (dx1 )2 (dx2 )2 (dx3 )2
(M.63)
C E ND OF EXAMPLE 13.4
Dyadic product
The dyadic product field A(x) a(x)b(x) with two juxtaposed vector fields a(x) and
b(x) is the outer product of a and b. Operating on this dyad from the right and from
the left with an inner product of an vector c one obtains
def
def
def
def
A c ab c a(b c)
(M.64a)
c A c ab (c a)b
(M.64b)
ab = x 1
x 2
x 3
a1 b1
a1 b2
a1 b3
a1 b2
a2 b2
a3 b2
a1 b3
x 1
a2 b3 x 2
a3 b3
x 3
(M.65)
which means that we can represent the tensor A(x) in matrix form as
a1 b1
Ai j (xk ) = a1 b2
a1 b3
a1 b2
a2 b2
a3 b2
a1 b3
a2 b3
a3 b3
(M.66)
which we identify with expression (M.15) on page 166, viz. a tensor in matrix notation.
173
M. M ATHEMATICAL M ETHODS
Vector product
The vector product or cross product of two arbitrary 3D vectors a and b in ordinary
R3 space is the vector
c = a b = i jk a j bk x i
(M.67)
Here i jk is the LeviCivita tensor defined in Equation (M.18) on page 167. Sometimes
the 3D vector product of a and b is denoted a b or, particularly in the Russian
literature, [ab]. Alternatively,
a b = ab sin e
(M.68)
where is the angle between a and b and e is a unit vector perpendicular to the plane
spanned by a and b.
A spatial reversal of the coordinate system (x01 , x02 , x03 ) = (x1 , x2 , x3 ) changes
sign of the components of the vectors a and b so that in the new coordinate system
a0 = a and b0 = b, which is to say that the direction of an ordinary vector is not
dependent on the choice of directions of the coordinate axes. On the other hand, as
is seen from Equation (M.67) above, the cross product vector c does not change sign.
Therefore a (or b) is an example of a true vector, or polar vector, whereas c is an
example of an axial vector, or pseudovector.
A prototype for a pseudovector is the angular momentum vector L = r p and
hence the attribute axial. Pseudovectors transform as ordinary vectors under translations and proper rotations, but reverse their sign relative to ordinary vectors for any
coordinate change involving reflection. Tensors (of any rank) which transform analogously to pseudovectors are called pseudotensors. Scalars are tensors of rank zero,
and zerorank pseudotensors are therefore also called pseudoscalars, an example being the pseudoscalar x i ( x j x k ). This triple product is a representation of the i jk
component of the LeviCivita tensor i jk which is a rank three pseudotensor.
x i
def
xi
(M.69)
where x i is the ith unit vector in a Cartesian coordinate system. Since the operator
in itself has vectorial properties, we denote it with a boldface nabla. In component
notation we can write
(M.70)
,
,
i =
x1 x2 x3
174
,
,
,
x0 x1 x2 x3
(M.71)
,
,
,
x0 x1 x2 x3
(M.72)
We can use this fourdel operator to express the transformation properties (M.13)
and (M.14) on page 166 as
and
y0 = x0 y
(M.73)
y0 = 0 x y
(M.74)
respectively.
E XAMPLE 13.5
1
,
c t
1
,
c t
(M.75)
(M.76)
1 2
2 = 2
c2 t2
(M.77)
which is the dAlembert operator, sometimes denoted , and sometimes defined with an opposite sign convention.
C E ND OF EXAMPLE 13.5
175
M. M ATHEMATICAL M ETHODS
With the help of the del operator we can define the gradient, divergence and curl
of a tensor (in the generalised sense).
The gradient
The gradient of an R3 scalar field (x), denoted (x), is an R3 vector field a(x):
(x) = (x) = x i i (x) = a(x)
(M.78)
From this we see that the boldface notation for the nabla and del operators is very
handy as it elucidates the 3D vectorial property of the gradient.
In 4D, the fourgradient is a covariant vector, formed as a derivative of a fourscalar field (x ), with the following component form:
(x ) =
E XAMPLE 13.6
(x )
x
(M.79)
= 0
x0i
(M.80)
Using this, the unprimed version, Equation (M.69) on page 174, and elementary rules of differentiation, we obtain the following two very useful results:
(x x0 ) = x i
x x0 
x x0
x x0 
=
=
i
xi
x x0 
x0i
(M.81)
= 0 (x x0 )
and
1
x x0 
x x0
= 0
x x0 3
1
x x0 
(M.82)
C E ND OF EXAMPLE 13.6
176
The divergence
We define the 3D divergence of a vector field in R3 as
a(x) = x j a j (x) = i j i a j (x) = i ai (x) =
ai (x)
= (x)
xi
(M.83)
which, as indicated by the notation (x), is a scalar field in R3 . We may think of the
divergence as a scalar product between a vectorial operator and a vector. As is the
case for any scalar product, the result of a divergence operation is a scalar. Again we
see that the boldface notation for the 3D del operator is very convenient.
The fourdivergence of a fourvector a is the following fourscalar:
a (x ) = a (x ) =
a (x )
x
(M.84)
BD IVERGENCE IN 3D
E XAMPLE 13.7
a(x0 )
x x0 
0 a(x0 )
+ a(x0 ) 0
=
x x0 
1
x x0 
(M.85)
which demonstrates how the primed divergence, defined in terms of the primed del operator in
Equation (M.80) on the preceding page, works.
C E ND OF EXAMPLE 13.7
The Laplacian
The 3D Laplace operator or Laplacian can be described as the divergence of the
gradient operator:
2 = = =
2
x i x j
= i j i j = 2i = 2 2
xi
x j
xi
i=1 xi
(M.86)
The symbol 2 is sometimes read del squared. If, for a scalar field (x), 2 < 0 at
some point in 3D space, it is a sign of concentration of at that point.
177
M. M ATHEMATICAL M ETHODS
E XAMPLE 13.8
x x0 
x x0 
(M.87)
C E ND OF EXAMPLE 13.8
The curl
In R3 the curl of a vector field a(x), denoted a(x), is another R3 vector field b(x)
which can be defined in the following way:
a(x) = i jk x i j ak (x) = i jk x i
ak (x)
= b(x)
x j
(M.88)
where use was made of the LeviCivita tensor, introduced in Equation (M.18) on
page 167.
The covariant 4D generalisation of the curl of a fourvector field a (x ) is the
antisymmetric fourtensor field
G (x ) = a (x ) a (x ) = G (x )
(M.89)
BT HE CURL OF A GRADIENT
Using the definition of the R3 curl, Equation (M.88) above, and the gradient, Equation (M.78) on page 176, we see that
[(x)] = i jk x i j k (x)
(M.90)
=
x2 x3 x3 x2
2
2
+
(x) x 2
x3 x1 x1 x3
2
2
+
(x) x 3
x1 x2 x2 x1
(M.91)
178
[(x)] 0
(M.92)
(M.93)
BT HE DIVERGENCE OF A CURL
E XAMPLE 13.10
With the use of the definitions of the divergence (M.83) and the curl, Equation (M.88) on
the preceding page, we find that
[ a(x)] = i [ a(x)]i = i jk i j ak (x)
(M.94)
Using the definition for the LeviCivita symbol, defined by Equation (M.18) on page 167, we
find that, due to the assumed wellbehavedness of a(x),
i jk
ak
xi
x j
2
2
a1 (x)
=
x2 x3 x3 x2
2
2
+
a2 (x)
x3 x1 x1 x3
2
2
+
a3 (x)
x1 x2 x2 x1
i i jk j ak (x) =
(M.95)
0
i.e., that
[ a(x)] 0
(M.96)
(M.97)
C E ND OF EXAMPLE 13.10
179
M. M ATHEMATICAL M ETHODS
Numerous vector algebra and vector analysis formulae are given in Chapter F.
Those which are not found there can often be easily derived by using the component
forms of the vectors and tensors, together with the Kronecker and LeviCivita tensors
and their generalisations to higher ranks. A short but very useful reference in this
respect is the article by A. Evett [3].
=0
(M.99)
dt q i
qi
To the generalised coordinate qi one defines a canonically conjugate momentum
pi according to
pi =
L
q i
(M.100)
(M.101)
(M.102)
After differentiating the left and right hand sides of this definition and setting them
equal we obtain
H
H
L
L
L
H
dpi +
dqi +
dt = q i dpi + pi dq i
dqi
dq i
dt (M.103)
pi
qi
t
qi
q i
t
180
Bibliography
According to the definition of pi , Equation (M.100) on the facing page, the second
and fourth terms on the right hand side cancel. Furthermore, noting that according
to Equation (M.101) on the preceding page the third term on the right hand side of
Equation (M.103) on the facing page is equal to p i dqi and identifying terms, we
obtain the Hamilton equations:
H
dqi
= q i =
pi
dt
dpi
H
= p i =
qi
dt
(M.104a)
(M.104b)
M.3 Bibliography
[1] G. B. A RFKEN AND H. J. W EBER, Mathematical Methods for Physicists, fourth, international ed., Academic Press, Inc., San Diego, CA . . . , 1995, ISBN 0120598167.
[2] R. A. D EAN, Elements of Abstract Algebra, John Wiley & Sons, Inc., New York, NY . . . ,
1967, ISBN 0471204528.
[3] A. A. E VETT, Permutation symbol approach to elementary vector analysis, American
Journal of Physics, 34 (1965), pp. 503507.
[4] P. M. M ORSE AND H. F ESHBACH, Methods of Theoretical Physics, Part I. McGrawHill
Book Company, Inc., New York, NY . . . , 1953, ISBN 070433168.
[5] B. S PAIN, Tensor Calculus, third ed., Oliver and Boyd, Ltd., Edinburgh and London, 1965,
ISBN 050013319.
[6] W. E. T HIRRING, Classical Mathematical Physics, SpringerVerlag, New York, Vienna,
1997, ISBN 0387948430.
181
Index
acceleration field, 125
advanced time, 41
Ampres law, 6
Ampreturn density, 87
anisotropic, 147
anomalous dispersion, 148
antisymmetric tensor, 61
associated Legendre polynomial, 114
associative, 54
axial gauge, 42
axial vector, 62, 174
Bessel functions, 111
BiotSavarts law, 7
birefringent, 147
braking radiation, 136
bremsstrahlung, 136, 143
canonically conjugate fourmomentum, 70
canonically conjugate momentum, 70, 180
canonically conjugate momentum density, 77
characteristic impedance, 28
classical electrodynamics, 1, 9
closed algebraic structure, 54
coherent radiation, 142
collisional interaction, 146
complex field sixvector, 20
complex notation, 33
complex vector, 172
component notation, 164
concentration, 177
conservative field, 11
conservative forces, 74
constitutive relations, 14
contravariant component form, 50, 164
contravariant field tensor, 62
contravariant fourtensor field, 169
contravariant fourvector, 166
contravariant fourvector field, 53
contravariant vector, 50
convection potential, 132
convective derivative, 13
cosine integral, 108
Coulomb gauge, 42
Coulombs law, 2
covariant, 48
covariant component form, 164
covariant field tensor, 62
covariant fourtensor field, 169
covariant fourvector, 166
covariant fourvector field, 53
covariant vector, 50
cross product, 174
curl, 178
cutoff, 138
cyclotron radiation, 140, 143
dAlembert operator, 38, 58, 175
del operator, 174
del squared, 177
differential distance, 52
differential vector operator, 174
Dirac delta, 178
Diracs symmetrised Maxwell equations, 16
dispersive, 148
displacement current, 10
divergence, 177
dot product, 171
dual vector, 50
duality transformation, 16
dummy index, 50
dyadic product, 173
dyons, 20
E1 radiation, 117
E2 radiation, 119
Einsteins summation convention, 164
electric charge conservation law, 9
183
I NDEX
184
far zone, 99
Faradays law, 12
field, 165
field Lagrange density, 78
field point, 4
field quantum, 138
fine structure constant, 138, 145
fourcurrent, 58
fourdel operator, 175
fourdimensional Hamilton equations, 70
fourdimensional vector space, 50
fourdivergence, 177
fourgradient, 176
fourHamiltonian, 70
fourLagrangian, 68
fourmomentum, 57
fourpotential, 58
fourscalar, 165
fourtensor fields, 169
fourvector, 53, 165
fourvelocity, 57
Fourier component, 27
Fourier transform, 39
freefree radiation, 136
functional derivative, 76
fundamental tensor, 50, 164, 169
Galileos law, 47
gauge fixing, 43
gauge function, 43
gauge invariant, 43
gauge transformation, 43
Gausss law of electrostatics, 5
general inhomogeneous wave equations, 37
generalised coordinate, 70, 180
generalised fourcoordinate, 70
Gibbs notation, 174
gradient, 176
Green function, 39, 114
group theory, 54
group velocity, 148
Hamilton density, 77
Hamilton density equations, 77
Hamilton equations, 70, 181
law of inertia, 47
Legendre polynomial, 114
Legendre transformation, 180
LeviCivita tensor, 167
LinardWiechert potentials, 61, 121, 131
light cone, 53
lightlike interval, 53
line element, 172
linear mass density, 75
linearly polarised wave, 31
longitudinal component, 30
Lorentz boost parameter, 55
Lorentz force, 14, 89, 131
Lorentz gauge condition, 38
Lorentz space, 51, 164
Lorentz transformation, 49, 131
LorenzLorentz gauge, 43
LorenzLorentz gauge condition, 38, 59
lowering of index, 169
M1 radiation, 118
Mller scattering, 145
Mach cone, 150
macroscopic Maxwell equations, 146
magnetic charge density, 15
magnetic current density, 15
magnetic dipole moment, 86, 118
magnetic dipole radiation, 118
magnetic displacement current, 18
magnetic field, 7
magnetic field energy, 88
magnetic field intensity, 87
magnetic flux, 12
magnetic flux density, 7
magnetic induction, 7
magnetic monopoles, 15
magnetic permeability, 146
magnetic susceptibility, 87
magnetisation, 86
magnetisation currents, 86
magnetising field, 15, 83, 87
magnetostatic vector potential, 36
magnetostatics, 6
massive photons, 81
mathematical group, 54
185
I NDEX
186
VavilovCerenkov
radiation, 149, 150
vector, 163
187
Гораздо больше, чем просто документы.
Откройте для себя все, что может предложить Scribd, включая книги и аудиокниги от крупных издательств.
Отменить можно в любой момент.