Basic Laws of Electromagnetism Irodov PDF

Besides the theoretical descrip-
tion of the basic laws of electro-

magnetism, this book contains
a detailed analysis of a large
number of examples and prob-
lems. These problems are
closely connected with the
main text and often serve to
develop or complement the
ideas presented in the text.
Whenever possible, the material
of the book has been kept free
from excessive mathematical
considerations, and the physical
aspects of the phenomena have
been given prime importance.
Intended as a textbook for
university students.
Igor lrodov is a professor of
physics. He graduated from the
Moscow Physical Engineering
Institute, defended his thesis in
1953 and since then has been
teaching physics at the same in-
stitute. He has published a
number of scientific papers and
books of which MIR
PUBLISHERS has translated in-
to English Fundamental Laws of
Mechanics, Problems in
General Physics, and Problem
Book on Atomic and Nuclear
Physics. The latter had six Rus-
sian editions and was published
in Poland, Romania, Great Bri-
tain (Pergamon Press), and the
USA. A Problem Book on
Genera! Physics (with
I. Savelyev and 0. Zamsha as
co-authors) had three Russian
editions and was published in
Poland and translated by folk
PUBLISHERS into French and
Arabic.
MIR PUBLISHERS of Moscow publishes Soviet scien-
tific and technical literature in twenty four languages
including all those most widely used. MIR translates
texts into Russian, and from Russian originals pro-
duces books in English, German, French, Italian,
Spanish, Portuguese, Czech, Slovak, Finnish,
Hungarian, Arabic, Persian, Hindi, Tamil, Kannada,
Bengali, Marathi, Telugu, Khmer, Laotian, Viet-
namese, and Dart Titles include textbooks fo;- higher
technical and vocational schools, literature on the
natural sciences and medicine (including textbooks
for medical schools), popular science and science fic-
tion. MIR's authors are leading Soviet scientists and
engineers from all fields of science and technology,
among them more than forty Members and Cor-
responding Members of the USSR Academy of
Sciences. Skilled translators provide a high standard
of translation from the original Russian. Many of the
titles already issued by MIR PUBLISHERS have been
adopted as textbooks and manuals at educational
establishments in France, Switzerland, Cuba, Syria,
India, Brasil and many other countries.
MIR PUBLISHERS' books can be purchased or
ordered through booksellers in your country dealing
with v/0 "Mezhdunarodnaya Kniga", the authorised
exporters.
Basic laws of electromagnetism
H. E. HP0g0B
OCHOBHbIE 3AHOHbI
3JIEKTPOMATHET143MA
HmaTeabemo «Bbicmaff unto.na*

Mocitsa
I.E.Irodov
Basic laws
of electromagnetism
Translated from Russian
by
Natasha Deineko and Ram Wadhwa
Mir Publishers Moscow

First published 1986
Revised from the 1983 Russian edition
Ha atia.altaCKOM Raune
143AaTeJlECTBO d3LICIllaFf inaoza*, 1983

(4) English translation, Mir Publishers, 1986
Preface
The main idea behind this book is to amalgamate the de-

scription of the basic concepts of the theory and the practical
methods of solving problems in one book. Therefore, each
chapter contains first a description of the theory of the
subject being considered (illustrated by concrete examples)
and then a set of selected problems with solutions. The prob-
lems are closely related to the text and often complement it.
Hence they should be analy:ed together with the text. In
author's opinion, the selected problems should enable the
reader to attain a deeper understanding of many important
topics and to visualize (even without solving the problems
but just by going through them) the wide range of applica-
tions of the ideas presented in this book.
In order to emphasize the most important laws of electro-
magnetism, and especially to clarify the most difficult top-
ics, the author has endeavoured to exclude the less impor-
tant topics. In an attempt to describe the main ideas concise-
ly, clearly and at the same time correctly, the text has been
kept free from superfluous mathematical formulas, and the
main stress has been laid on the physical aspects of the phe-
nomena. With the same end in view, various model repre-
sentations, simplifying factors, special cases, symmetry con-
siderations, etc. have been employed wherever possible.
SI units of measurements are used throughout the book.
However, considering that the Gaussian system of units is
still widely used, we have included in Appendices 3 and 4
the tables of conversion of the most important quantities
and formulas from SI to Gaussian units.
The most important statements and terms are given in
italics. More complicated material and problems involving
cumbersome mathematical calculations are set in brevier
type. This material can be omitted on first reading without
any loss of continuity. The brevier type is also used for
problems and examples.
6 Preface
The book is intended as a textbook for undergraduate stu-

dents specializing in physics (in the framework of the course
on general physics). It can also be used by university
teachers.
The author is grateful to Prof. A.A. Detlaf and Reader
L.N. Kaptsov who reviewed the manuscript and made a
number of valuable comments and suggestions.
I. Irodov
Contents
Preface 5
List of Notations 10
1. Electrostatic Field in a Vacuum . 11

1.1. Electric Field 11
1.2. The Gauss Theorem 16
1.3. Applications of the Gauss Theorem 19
1.4. Differential Form of the Gauss Theorem 23
1.5. Circulation of Vector E. Potential 25
1.6. Relation Between Potential and Vector E 30
1.7. Electric Dipole 34
Problems 39
2. A Conductor in an Electrostatic Field • 45

2.1. Field in a Substance 45
2.2. Fields Inside and Outside a Conductor 47
2.3. Forces Acting on the Surface of a Conductor . 49
2.4. Properties of a Closed Conducting Shell . . . . . 51
2.5. General Problem of Electrostatics. Image Method 54
2.6. Capacitance. Capacitors 57
Problems 60
3. Electric Field in Dielectrics 67

3.1. Polarization of Dielectrics 67
3.2. Polarization 70
3.3. Properties of the Field of P 72
3.4. Vector D 76
3.5. Boundary Conditions 80
3.6. Field in a Homogeneous Dielectric 84
Problems 86
4. Energy of Electric Field 94

4.1. Electric Energy of a System of Charges 94
4.2. Energies of a Charged Conductor and a Charged
Capacitor 99
4.3. Energy of Electric Field 100
4.4. A System of Two Charged Bodies 105
4.5. Forces Acting in a Dielectric 106
Problems 110
8 Contents
5. Direct Current 116

5.1. Current Density. Continuity Equation 116
5.2. Ohm's Law for a Homogeneous Conductor 119
5.3. Generalized Ohm's Law 122
5.4. Branched Circuits. Kirchhoff's Laws 126
5.5. Joule's Law 129
5.6. Transient Processes in a Capacitor Circuit . 133
Problems 136
6. Magnetic Field in a Vacuum 141

6.1. Lorentz Force. Field B 141
6.2. The Biot-Savart Law 145
6.3. Basic Laws of Magnetic Field 148
6.4. Applications of the Theorem on Circulation of Vec-
tor B 150
6.5. Differential Forms of Basic Laws of Magnetic Field 154
6.6. Ampere's Force 155
6.7. Torque Acting on a Current Loop 159
6.8. Work Done upon Displacement of Current Loop 161
Problems 163
7. Magnetic Field in a Substance 172

7.1. Magnetization of a Substance. Magnetization Vec-
tor J 172
7.2. Circulation of Vector J 176
7.3. Vector H 178
7.4. Boundary Conditions for B and H 182
7.5. Field in a Homogeneous Magnetic 185
7.6. Ferromagnetism 188
Problems 192
8. Relative Nature of Electric and Magnetic Fields 198

8.1. Electromagnetic Field. Charge Invariance . . . 198
8.2. Laws of Transformation for Fields E and B . . . 200
8.3. Corollaries of the Laws of Field Transformation 206
8.4. Electromagnetic Field Invariants 208
Problems 209
9. Electromagnetic Induction 217

9.1. Faraday's Law of Electromagnetic Induction. Lenz's
Law 217
9.2. Origin of Electromagnetic Induction 220
9.3. Self-induction 226
9.4. Mutual Induction 231
9.5. Magnetic Field Energy 234
9.6. Magnetic Energy of Two Current Loops 238
Contents 9
9.7. Energy and Forces in Magnetic Field 240

Problems 244
10. Maxwell's Equations. Energy of Electromagnetic Field 253

10.1. Displacement Current 253
10.2. Maxwell's Equations 257
10.3. Properties of Maxwell's Equations 261
10.4. Energy and Energy Flux. Poynting's Vector . 264
10.5. Electromagnetic Field Momentum 268
Problems 271
11. Electric Oscillations 277

11.1. Equation of an Oscillatory Circuit 277
11.2. Free Electric Oscillations 280
11.3. Forced Electric Oscillations 285
11.4. Alternating Current , 290
Problems 293
Appendices 299
1. Notations for Units of Measurement 299
2. Decimal Prefixes for Units of Measurement 299
3. Units of Measurement of Electric and Magnetic Quan-
tities in SI and Gaussian Systems 299
4. Basic Formulas of Electricity and Magnetism in SI
and Gaussian Systems 300
5. Some Physical Constants 304
Subject Index 305
List of Notations
Vectors are denoted by the bold-face upright letters

(e.g. r, E). The same letter in the standard-type print (r, E)
denotes the magnitude of the vector.
Average quantities are denoted by angle brackets ( ),
e.g. (v ), (P ).
The symbols in front of the quantities denote:
A is the finite increment of the quantity, viz. the difference
between its final and initial values, e.g. AE = E2 — El,
= (P2 - (Pl;
d is the differential (infinitesimal increment), e.g. dE, dcp;
6 is the incremental value of a quantity, e.g. 6A is the ele-
mentary work.
Unit vectors:
ex, ey , ez (or i, j, k) are the unit vectors of Cartesian coordi-
nates x, y, z;
▪ ecp e Z are the unit vectors of cylindrical coordinates
ID, q, z;
n is the unit vector of the normal to a surface element;
t is the unit vector of the tangent to the contour or to an
interface.
Time derivative of an arbitrary function f is denoted by
afiat or by the dot above the letter denoting the function, f.
Integrals of any multiplicity are denoted by a single sym-
bol .r and differ only in the notation of the element of inte-
gration: dV is the volume element, dS is the surface ele-
ment and dl is the element of length. Symbol 1) denotes the
integration over a closed contour or a closed surface.
Vector operator y (nabla). The operations involving this
operator are denoted as follows: ycp is the gradient of
c (grad c), V .E is the divergence of E (div E), and y x E
is the curl of E (curl E).
Vector product is denoted [a x b], where a and b are
vectors.
1. Electrostatic Field in a Vacuum
1.1. Electric Field

Electric Charge. At present it is known that diverse phe-
nomena in nature are based on four fundamental interactions
among elementary particles, viz. strong, electromagnetic,
weak, and gravitational interactions. Each type of inter-
action is associated with a certain characteristic of a particle.
For example, the gravitational interaction depends on the
mass of particles, while the electromagnetic interaction is
determined by electric charges.
The electric charge of a particle is one of its basic charac-
teristics. It has the following fundamental properties:
(1) electric charge exists in two forms, i.e. it can be posi-
tive or negative;
(2) the algebraic sum of charges in any electrically insu-
lated system does not change (this statement expresses the
law of conservation of electric charge);
(3) electric charge is a relativistic invariant: its magni-
tude does not depend on the reference system, in other words,
it does not depend on whether the charge moves or is fixed.
It will be shown later that these fundamental properties
have far-reaching consequences.
Electric Field. In accordance with modern theory, inter-
action among charges is accomplished through a field. Any
electric charge q alters in a certain way the properties of
the space surrounding it, i.e. creates an electric field. This
means that another, "test" charge placed at some point of
the field experiences the action of a force.
Experiments show that the force F acting on a fixed test
point charge q' can always be represented in the form
F = q'E, (1.1)
where vector E is called the intensity of the electric field at
a given point. Equation (1.1) shows that vector E can be
defined as the force acting on a positive fixed unit charge.
Here we assume that the test charge q' is sufficiently small so
12 1. Electrostatic Field in a Vacuum
that its introduction does not noticeably distort the field

under investigation (as a result of possible redistribution of
charges creating the field).
The Field of a Point Charge. It follows directly from exper-
iment (Coulomb's law) that the intensity of the field of a
fixed point charge q at a distance r from it can be represent-
ed in the form
E= l e (1.2)
eo 2 r'
where co is the electric constant and er is the unit vector of

the radius vector r drawn from the centre of the field, where
the charge q is located, to the point we are interested in.
Formula (1.2) is written in SI. Here the coefficient
1/4aco = 9 X 109 m/F,
the charge q is measured in coulombs (C) and the field inten-
sity E in volts per metre (V/m). Vector E is directed along r
or oppositely to it depending on the sign of the charge q.
Formula (1.2) expresses Coulomb's law in the "field" form.
It is important that the intensity E of the field created by a
point charge is inversely proportional to the square of the
distance r. All experimental results indicate that this law
is valid for distances from 10-13cm to several kilometres,
and there are no grounds to expect that this law will be
violated for larger distances.
It should also be noted that the force acting on a test
charge in the field created by a fixed point charge does not
depend on whether the test charge is at rest or moves. This
refers to a system of fixed charges as well.
Principle of Superposition. Besides the law expressed by
(1.2), it also follows from experiments that the intensity of
the field of a system of fixed point charges is equal to the
vector sum of the intensities of the fields that would be
created by each charge separately:
E, (1.3)
42-E80 Ti,
where ri is the distance between the charge q i and the point

under consideration.
This statement is called the principle of _superposition of
electric fields. It expresses one of the most remarkable prop-
erties of fields and allows us to calculate field intensity of
any system of charges by representing this system as an ag-
gregate of point charges whose individual contributions are
given by formula (1.2).
Charge Distribution. In order to simplify mathematical
calculations, it is often convenient to ignore the fact that
charges have a discrete structure (electrons, nuclei) and
assume that they are "smeared" in a certain way in space. In
other words, it is convenient to replace the actual distribu-
tion of discrete point charges by a fictitious continuous
distribution. This makes it possible to simplify calculations
considerably without introducing any significant error.
While going over to a continuous distribution, the concept
of charge density is introduced, viz. the volume density p,
surface density a, and linear density X,. By definition,
dq dq dq
(1.4)
P =dV' a= dS ' = dl
where dq is the charge contained in the volume dV, on the

surface dS, and in the length dl respectively.
Taking these distributions into consideration we can repre-
sent formula (1.3) in a different form. For example, if the
charge is distributed over the volume, we must replace qi
by dq = p dV and summation by integration. This gives
1 pe dV 1 pr dV
E= r3 (1.5)
42u, j r2 4aso
where the integration is performed over the entire space with
nonzero values of p.
Thus, knowing the distribution of charges, we can complete-
ly solve the problem of finding the electric field intensity
by formula (1.3) if the distribution is discrete or by formu-
la (1.5) or by a similar formula if it is continuous. In the
general case, the calculation involves certain difficulties
(although they are not of principle nature). Indeed, in order
to find vector E, we must first calculate its projections E x,
E y, and E z, which means that we must take three integrals
of the type (1.5). As a rule, the problem becomes much

simpler in cases when a system of charges has a certain sym-
metry. Let us consider two examples.
Example 1. The field on the axis of a thin uniformly charged ring.
A charge q > 0 is uniformly distributed over a thin ring of radius a.
Find the electric field intensity E on the axis of the ring as a function
of the distance z from its centre.
It can be easily shown that vector E in this case must be directed
along the axis of the ring (Fig. 1.1). Let us isolate element dl on the
Fig. 1.1 Fig. 1.2
ring in the vicinity of point A . We write the expression for the comp-
onent dE, of the field created by this element at a point C:
1 Xl
d
dE = cos a,
z 4ats, r2
where ? = q/2aa. The values of r and a will be the same for all the
elements of the ring, and hence the integration of this equation is sim-
ply reduced to the replacement of 2■ , dl by q. As a result, we obtain
z
E—
4nso (a2 -1- Z2)3/ 2
It can be seen that for z>> a the field E z q/43-isoz2, i.e. at large dis-
tances the system behaves as a point charge.
Example 2. The field of a uniformly charged straight filament. A
thin straight filament of length 2/ is uniformly charged by a charge q.
Find the field intensity E at a point separated by a distance x from
the midpoint of the filament and located symmetrically with respect
to its ends.
It is clear from symmetry considerations that vector E must be
directed as shown in Fig. 1.2. This shows the way of solving this
problem: we must find the component dE, of the field created by the
element dl of the filament, having the charge dq, and then integrate
the result over all the elements of the filament. In this case
1 k dl
dEx =dE cos a= 43. , —cos a,
r2
where k = qI21 is the linear charge density. Let us reduce this equation
to the form convenient for integration. Figure 1.2 shows that
dl cos a = rda and r = x/cos a; consequently,
1
dE x = cos a da.
4ne 0 r2 41[80x
This expression can be easily integrated:
ao
k
E= 2 cos a da= 2 sin a°,
4neox 4:18,x
D
where a, is the maximum value of the angle a, sin a0 = 1/ 1n2 + x2.

Thus,
qI21 9
1
E—,
4380X 02_1_ x2 4oisox 1/12 +x2
In this case also E q/4:-ce,x2 for x » 1 as the field of a point charge.
Geometrical Description of Electric Field. If we know

vector E at each point, the electric field can be visually re-
presented with the help of field lines, or lines of E. Such a
line is drawn so that a tangent to it at each point coincides
with the direction of vector E. The density of the lines,
i.e. the number of lines per unit area normal to the lines is
proportional to the magnitude of vector E. Besides, the lines
are directed like vector E. This pattern gives the idea about
the configuration of a given electric field, i.e. about the di-
rection and magnitude of vector E at each point of the field.
On the General Properties of Field E. The field E defined
above has two very important properties. The knowledge of
these properties helped to deeper understand the very con-
cept of the field and formulate its laws, and also made it pos-
sible to solve a number of problems in a simple and elegant
way. These properties, viz. the Gauss theorem and the theo-
rem on circulation of vector E, are associated with two most
important mathematical characteristics of all vector fields:
the flux and the circulation. It will be shown below that in
16 1. Electr.ostatic Field in a Vacuum
terms of these two concepts not only all the laws of electri-
city but also all the laws of magnetism can be described. Let
us go over to a systematic description of these properties.
1.2. The Gauss Theorem

Flux of E. For the sake of clarity, we shall use the geomet-
rical description of electric field (with the 'help of lines
of E). Moreover, to simplify the
analysis, we shall assume that the
density of lines of field E is equal
to the magnitude of vector E. Then
the number of lines piercing the
dS area element dS, the normal n to
which forms angle a with vector E,
Fig. 1.3 is determined as E•dS cos a (see
Fig. 1.3). This quantity is just the
flux dO of E through the area element dS. In a more compact
form, this can be written as
dal) dS = E•dS,
where En is the projection of vector E onto the normal n to
the area element dS, and dS is the vector whose magnitude
is equal to dS and the direction coincides with the direction
of the normal. It should be noted that the choice of the di-
rection of n (and hence of dS) is arbitrary. This vector could
be directed oppositely.
If we have an arbitrary surface S, the flux of E through
it can be expressed as
= S E dS. (1.6)
This is an algebraic quantity, since it depends not only on

the configuration of the field E but also on the choice of the
normal. If a surface is closed it is customary to direct the
normal n outside the region enveloped by this surface, i.e.
to choose the outward normal. Henceforth we shall always
assume that this is the case.
Although we considered here the flux of E, the concept of
flux is applicable to any vector field as well.
1.2. The Gauss Theorem 17
The Gauss Theorem. The flux of E through an arbitrary

closed surface S has a remarkable property: it depends only
on the algebraic sum of the charges embraced by this sur-
face, i.e.
E dS= gin, (1.7)

co
where the circle on the integral symbol indicates that the

integration is performed over a closed surface.
This expression is essentially the Gauss theorem: the flux of
Fig. 1.4 Fig. 1.5
E through a closed surface is equal to the algebraic sum of the

charges enclosed by this surface, divided by Bo.
Proof. Let us first consider the field of a single point charge
q. We enclose this charge by an arbitrary closed surface S
(Fig. 1.4) and find the flux of E through the area element dS:
q
dcI)=E dS cos cc —
42160 —
r2
dS cos cc =-- 4nqeoc/52, (1.8)
where d52 is the solid angle resting on the area element dS
and having the vertex at the point where the charge q is lo-
cated. The integration of this expression over the entire
surface S is equivalent to the integration over the entire
solid angle, i.e. to the replacement of dS2 by 4n. Thus we
obtain CD = ego, as is defined by formula (1.7).
It should be noted that for a more complicated shape of a
closed surface, the angles a may be greater than n/2, and
hence cos a and °IQ in (1.8) generally assume either positive
or negative values. Thus, ci52 is an algebraic quantity: if
dS2 rests on the inner side of the surface S, d52 > 0, while if
it rests on the outer side, d52 < 0.
2—i81
48 1. Electrostatic Field in a Vacuant
In particular, this leads to the following conclusion: if

the charge q is located outside a closed surface S, the flux
of E through this surface is equal to zero. In order to prove
this, it is sufficient to draw through the charge q a conical
surface tangent to the closed surface S. Then the integration
of Eq. (1.8) over the surface S is equivalent to the integra-
tion over Q (Fig. 1.5): the outer side of the surface S will be
seen from the point q at an angle Q > 0, while the inner side,
at an angle —Q (the two angles being equal in magnitude).
The sum is equal to zero, and (I) = 0, which also agrees
with (1.7). In terms of field lines or lines of E, this means
that the number of lines entering the volume enclosed by
the surface S is equal to the number of lines emerging from
this surface.
Let us now consider the case when the electric field is
created by a system of point charges q1 q2, . . . . In this
,
case, in accordance with the principle of superposition E =

= El E 2 . , where El is the field created by the
charge q1,, etc. Then the flux of E can be written in the form
E dS (1 (Ei E2 dS
= dS+ E2 dS ... =01+02+ ...
In accordance with what was said above, each integral on the

right-hand side is equal to qileo if the charge qi is inside
the closed surface S and is equal to zero if it is outside the
surface S. Thus, the right-hand side will contain the alge-
braic sum of only those charges that lie inside the surface S.
To complete the proof of the theorem, it remains for us
to consider the case when the charges are distributed contin-
uously with the volume density depending on coordinates.
In this case, we may assume that each volume element dV
contains a "point" charge p dV. Then on the right-hand side
of (1.7) we have
pdV, (1.9)
where the integration is performed only over the volume

contained within the closed surface S.
We must pay attention to the following important circum-
stance: while the field E itself depends on the mutual con-

figuration of all the charges, the flux of E through an arbi-
trary closed surface S is determined by the. algebraic sum
of the charges inside the surface S. This means that if we
displace the charges, the field E will be changed everywhere,
and in particular on the surf ace S. Generally, the flux of E
through the surface S will also change. However, if the
displacement of charges did not involve their crossing of
the surface S, the flux of E through this surface would re-
main unchanged, although, we stress again, the field E itself
may change considerably. What a remarkable property of
electric field!
1.3. Applications of the Gauss Theorem

Since the field E depends on the configuration of all charges,,
the Gauss theorem generally does not allow us to deter-
mine this field. However, in certain cases the Gauss theorem
proves to be a very effective analytical instrument since it
gives answers to certain principle questions without solving
the problem and allows us to determine the field E in a very
simple way. Let us consider some examples and then formu-
late several general conclusions about the cases when appli-
cation of the Gauss theorem is the most expedient.
Example 1. On the impossibility of stable equilibrium of a charge
in an electric field. Suppose that we have in vacuum a system of fixed
point charges in equilibrium. Let us consider one of these charges,
e.g. a charge q. Can its equilibrium be stable?
In order to answer this question, let us envelop the charge q by a
small closed surface S (Fig. 1.6). For the sake of definiteness, we as-
sume that q > 0. For the equilibrium of this charge to be stable, it is
necessary that the field E created by all the remaining charges of the
system at all the points of the surface S be directed towards the charge
q. Only in this case any small displacement of the charge q from the
equilibrium position will give rise to a restoring force, and the equilib-
rium state will actually be stable. But such a configuration of the
field E around the charge q is in contradiction to the Gauss theorem:
the flux of E through the surface S is negative, while in accordance
with the Gauss theorem it must be equal to zero since it is created by
charges lying outside the surface S. On the other hand, the fact that
E is equal to zero indicates that at some points of the surface S vector
E is directed inside it and at some other points it is directed outside.
2*
Hence it follows that in any electrostatic field a charge cannot be in

stable equilibrium.
Example 2. The field of a uniformly charged plane. Suppose that
the surface charge density is a. It is clear from the symmetry of the
problem that vector E can only be normal to the charged plane. More-
over, at points symmetric with respect to this plane, vectors E obvi-
ously have the same magnitude but opposite directions. Such a configu-
ration of the field indicates that a right cylinder should be chosen for
the closed surface as shown in Fig. 1.7, where we assume that a > 0.
Fig. 1.6 Fig. 1.7
The flux through the lateral surface of this cylinder is equal to

zero, and hence the total flux through the entire cylindrical surface
is 2E AS, where AS is the area of each endface. A charge a AS is
enclosed within the cylinder. According to the Gauss theorem, 2E AS
= a AS/co, whence E = cd2go. In a more exact form, this expression
must be written as
En= a/2so, (1.10)
where En is the projection of vector E onto the normal n to the charged
plane, the normal n being directed away from this plane. If a > 0
then En >0, and hence vector E is directed away from the charged
plane, as shown in Fig. 1.7. On the other hand, if o- < 0 then En < 0,
and vector E is directed towards the charged plane. The fact that E is
the same at any distance from the plane indicates that) the corre-
sponding electric field is uniform (both on the right and on the left of
the plane).
The obtained result is valid only for an infinite plane surface,
since only in this case we can use the symmetry considerations dis-
cussed above. However, this result is approximately valid for the region
near the middle of a finite uniformly charged plane surface far from
its ends.
Example 3. The field of two parallel planes charged uniformly with
densities o- and —cr by unlike charges.
This field can be easily found as superposition of the fields created
by each plane separately (Fig. 1.8). Here the upper arrows correspond
to the field from the positively charged plane, while the lower arrows,
to that from the negatively charged plane. In the space between the
planes the intensities of the fields being added have the same direc-
tion, hence the result (1.10) will be doubled, and the resultant field
intensity will be
E = also, (1.11)
where a stands for the magnitude of the surface charge density. It can
be easily seen that outside this space the field is equal to zero. Thus,
E= 0 E = -2—
co E =0
Fig. 1.8 Fig. 1.9

in the given case the field is located between the planes and is uni-
form.
This result is approximately valid for the plates of finite dimen-
sions as well, if only the separation between the plates is considerably
smaller than their linear dimensions (parallel-plate capacitor). In this
case, noticeable deviations of the field from uniformity are observed
only near the edges of the plates (these distortions are often ignored in
calculations).
Example 4. The field of an infinite circular cylinder uniformly
charged over the surface so that the charge X, corresponds to its unit
length.
In this case, as follows from symmetry considerations, the field
is of a radial nature, i.e. vector E at each point is perpendicular to
the cylinder axis, and its magnitude depends only on the distance r
from the cylinder axis to the point. This indicates that a closed sur-
face here should be taken in the form of a coaxial right cylinder
(Fig. 1.9). Then the flux of E through the endfaces of the cylinder is
equal to zero, while the flux through the lateral surface is Er 2 arh,
where Er is the projection of vector E onto the radius vector r coincid-
ing with the normal n to the lateral surface of the cylinder of radius
r and height h. According to the Gauss theorem, Er 2arh = Mleo for
> a, whence
Er-= (r > a). (1.12)
2n8Ur
For X > 0, Er >0 as well, i.e. vector E is directed away from

the charged cylinder, and vice versa.
If r < a, the closed surface does not contain any charge since in
this region E = 0 irrespective of r. Thus, inside a circular infinite
cylinder uniformly charged over the surface the field is absent.
Example 5. The field of a spherical surface uniformly charged by
the charge q.
This field is obviously centrally symmetric: vector E from any
point passes through the centre of the sphere, while its magnitude
must depend only on the distance r from the point to the centre of the
sphere. It is clear that for such a configuration of the field we should
take a concentric sphere as a closed surface. Let the radius of this
sphere be r > a. Then, in accordance with the Gauss theorem, Er41-cr2
= q/so, whence
Er= q (r> a), (L13)

4a cso r2
-
where Er is the projection of vector E onto the radius vector r coinciding

with the normal n to the surface at each of its points. The sign of the
charge q determines the sign of the projection ET in this case as well.
Hence it determines the direction of vector E itself: either away from
the sphere (for q > 0) or towards it (for q < 0).
If r < a, the closed surface does not contain any charge and hence
within this region E = 0 everywhere. In other words, inside a uni-
formly charged spherical surface the electric field is absent. Outside
this surface the field decreases with the distance r in accordance with
the same law as for a point charge.
Example 6. The field of a uniformly charged sphere. Suppose that
a charge q is uniformly distributed over a sphere of radius a. Obvious-
ly, the field of such a system is centrally symmetric, and hence for de-
termining the field we must take a concentric sphere as a closed sur-
face. It can be easily seen that for the field outside the sphere we obtain
the same result as in the previous example [see (1.13)]. However,
inside the sphere the expression for the field will be different. The
sphere of radius r < a encloses the charge q' = q (r1a)3since in our
case the ratio of charges is equal to the ratio of volumes and is pro-
portional to the radii to the third power. Hence, in accordance with
the Gauss theorem we have
1 3
Er • 4nr2 = q( a ) ,
—
eo
whence
E = r (r < a), (1.14)
r 43-cso
i.e. inside a uniformly charged sphere the field intensity grows linear-
ly with the distance r from its centre. The curve representing the de-
pendence of E on r is shown in Fig. 1.10.
General Conclusions. The results obtained in the above

1.4. Differential Form of the Gauss Theorem 23
examples could be found by direct integration (1.5) as well.

However, it is clear that these problems can be solved in a
much simpler way by using the Gauss theorem.
The simple solution of the problems considered above may
create an illusive impression about the strength of the method
based on the application of
the Gauss theorem and
about the possibility of
solving many other prob-
lems by using this the-
orem. Unfortunately, it is
not the case. The number of
problems that can be easily
solved with the help of the
Gauss theorem is limited. Fig. 1.10
We cannot use it even to
solve the problem of finding the field of a symmetric charge
distribution on a uniformly 'charged disc. In this case, the
field configuration is rather complicated, and a closed
surface for a simple calculation of the flux of E cannot be
found.
The Gauss theorem can be effectively applied to calcula-
tion of fields only when a field has a special symmetry (in
most cases plane, cylindrical, or circular). The symmetry,
and hence the field configuration, must be such that, firstly,
a sufficiently simple closed surface S can be found and,
secondly, the calculation of the' flux of E can be reduced
to a simple multiplication of E (or Er) by the area of the
surface S or its part. If these conditions are not satisfied,
the problem of finding the field should be solved either di-
rectly by formula (1.5) or by using other methods which
will be discussed below.
1.4. Differential Form of the Gauss Theorem

A remarkable property of electric field expressed by the Gauss
theorem suggests that this theorem be represented in a different form
which would broaden its possibilities as an instrument for analysis
and calculation.
In contrast to (1.7) which is called the integral form we shall seek
the differential form of the Gauss theorem, which establishes the rela-
tion between the volume charge density p and the changes in the field
intensity E in the vicinity of a given point in space.
For this purpose, we first represent the charge q in the volume V

enveloped by a closed surface S in the form qint = (p) V, where (p )
is the volume charge density, averaged over the volume V. Then we
substitute this expression into Eq. (1.7) and divide both its sides by
V, which gives
E dS = (p)/80. (1.15)
1
V
We now make the volume V tend to zero by contracting it to the
point we are interested in. In this case, (p) will obviously tend to the
value of p at the given point of the field, and hence the ratio on the
left-hand side of Eq. (1.15) will tend to pis,.
The quantity which is the limit of the ratio of E dS to V as V –›- 0
is called the divergence of the field E and is denoted by div E. Thus,
by definition,
1
div E = lim — E dS. (1.16)
V
The divergence of any other vector field is determined in a similar
way. It follows from definition (1.16) that divergence is a scalar
function of coordinates.
In order to obtain the expression for the divergence of the field E,
we must, in accordance with (1.16), take an infinitely small volume V,
determine the flux of E through the closed surface enveloping this
volume, and find the ratio of this flux to the volume. The expression
obtained for the divergence will depend on the choice of the coordi-
nate system (in different systems of coordinates it turns out to be
different). For example, in Cartesian coordinates it is given by
aE x °Ey OEz (1.17)
div
Ox ay -L
1 az
Thus, we have found that as V -± 0 in (1.15), its right-hand side

tends to p/so, while the left-hand side tends to div E. Consequently,
the divergence of the field E is related to the charge density at the
same point through the equation
I div E = p/so. I (1.18)
This equation expresses the Gauss theorem in the differential form.

The form of many expressions and their applications can be consid-
erably simplified if we introduce the vector differential operator v.
In Cartesian coordinates, the operator v has the form
0 a (1.19)
v=i + j -Ty +k 79; '
where i, j, and k are the unit vectors of the X-, Y-, and Z-axes. The
eperator v itself does not have any meaning. It becomes meaningfu I
1.5. Circulation of Vector K Potential 25
only in combination with a scalar or vector function by which it is

symbolically multiplied. For example, if we form the scalar product
of vector v and vector E, we obtain
0 0 0
E = V,,E x f Vy Ey+VzEz=w Ex+w. Ey+ Ez.
It follows from (1.17) that this is just the divergence of E.
Thus, the divergence of the field E can be written as div E or v •E
(in both cases it is read as "the divergence of E"). We shall be using
the latter, more convenient notation. Then, for example, the Gauss
theorem (1.18) will have the form
I V•E= pie°. I (1.20)
The Gauss theorem in the differential form is a local theorem:

the divergence of the field E at a given point depends only on the
electric charge density p at this point. This is one of the remarkable
properties of electric field. For example, the field E of a point charge
is different at different points. Generally, this refers to the spatial
derivatives 0Exiax, 0Ey/8y, and 0Ezi8z as well. However, the Gauss
theorem states that the sum of these derivatives, which determines the
divergence of E, turns out to be equal to zero at all points of the field
(outside the charge itself).
At the points of the field where the divergence of E is positive, we
have the sources of the field (positive charges), while at the points
where it is negative, we have sinks (negative charges). The field lines
emerge from the field sources and terminate at the sinks.
1.5. Circulation of Vector E. Potential

Theorem on Circulation of Vector E. It is known from
mechanics that any stationary field of central forces is con-
servative, i.e. the work done by the forces of this field is
independent of the path and depends only on the position
of the initial and final points. This property is inherent in
the electrostatic field, viz. the field created by a system of
fixed charges. If we take a unit positive charge for the test
charge and carry it from point 1 of a given field E to point 2,
the elementary work of t he forces of the field done over the
distance dl is equal to E•dl, and the total work of the field
forces over the distance between points 1 and 2 is defined as
2
E dl. (1.21)
This integral is taken along a certain line (path) and is

therefore called the line integral.
We shall now show that from the independence of line
integral (1.21) of the path between two points it follows
that when taken along an arbitrary closed path, this inte-
gral is equal to zero. Integral (1.21) over a closed contour is
called the circulation of vector E and is denoted by
Thus we state that circulation of vector E in any electro-
static field is equal to zero, i.e.
cciEd1=0. (1.22)
1248
This statement is called the theorem on circulation of
vector E.
In order to prove this theorem, we break an arbitrary closed
path into two parts 1a2 and 2b1
2 (Fig. 1.11). Since line integral (1.21)
cr c i) (we denote it by ) does not
depend on the path hetween points1
C 12
(a) (b)
and 2, we have . On the

12 12
other hand, it is clear that
(b) (b) (b)
Fig. 1.11
where f is the
,
12 v 21 . 21
integral over the same segment b but taken in the opposite
direction. Therefore
(a)
:c
(a) (b) (b)
12 4- 21 .1. 11 — ,J 12 = o '
Q.E.D.
A field having property (1.22) is called the potential field.
Hence, any electrostatic field is a potential field.
The theorem on circulation of vector E makes it possible
to draw a number of important conclusions without re-
sorting to calculations. Let us consider two examples.
Example 1. The field lines of an electrostatic field E cannot be
closed.
Indeed, if the opposite were true and some lines of field E were
closed, then taking the circulation of vector E along this line we would
immediately come to contradiction with theorem (1.22). This means
that actually there are no closed lines of E in an electrostatic field:
the lines emerge from positive charges
and terminate on negative ones (or go
to infinity).
Example 2. Is the configuration of an I ). I
electrostatic field shown in Fig. 1.12
possible?
No, it is not. This immediately becomes
clear if we apply the theorem on circula-
tion of vector E to the closed contour
shown in the figure by the dashed line.
The arrows on the contour indicate the
direction of circumvention. With such a
special choice of the contour, the contri- Fig. 1.12
bution to the circulation from its ver-
tical parts is equal to zero, since in
this case E 1 dl and E 0. It remains for us to consider the
two horizontal segments of equal lengths. The figure shows that the
contributions to the circulation from these regions are opposite in
sign, and unequal in magnitude (the contribution from the upper seg-
ment is larger since the field lines are denser, and hence the value of
E is larger). Therefore, the circulation of E differs from zero, which
contradicts to (1.22).
Potential. Till now we considered the description of elec-

tric field with the help of vector E. However, there exists
another adequate way of describing it by using potential q)
(it should be noted at the very outset that there is a one-to-
one correspondence between the two methods). It will be
shown that the second method has a number of significant
advantages.
The fact that line integral (1.21) representing the work of
the field forces done in the displacement of a unit positive
charge from point 1 to point 2 does not depend on the path
allows us to state that for electric field there exists a certain
scalar function cp (r) of coordinates such that its decrease is
given by
2
(pi— cp2 = E dl, (1.23)
where (p 1and cp2 are the values of the function cp at the

points 1 and 2. The quantity cp (r) defined in this way is
called the field potential. A comparison of (1.23) with the

expression for the work done by the forces of the potential
field (the work being equal to the decrease in the potential
energy of a particle in the field) leads to the conclusion
that the potential is the quantity numerically equal to the po-
tential energy of a unit positive charge at a given point of the
field.
We can conditionally ascribe to an arbitrary point 0 of the
field any value cpo of the potential. Then the potentials of
all other points of the field will be unambiguously deter-
mined by formula (1.23). If we change wo by a certain value
AT, the potentials of all other points of the field will change
by the same value.
Thus, potential cp is determined to within an arbitrary
additive constant. The value of this constant does not play
any role, since all electric phenomena depend only on the
electric field strength. It is determined, as will be shown
later, not by the potential at a given point but by the poten-
tial difference between neighbouring points of the field.
The unit of potential is the volt (V).
Potential of the Field of a Point Charge. Formula (1.23)

contains, in addition to the definition of potential cp, the
method of finding this function. For this purpose, it is suf-
ficient to evaluate the integral f E dl over any path be-
tween two points and then represent the obtained result as
a decrease in a certain function which is just cp (r). We can
make it even simpler. Let us use the fact that formula
(1.23) is valid not only for finite displacements but for ele-
mentary displacements dl as well. Then, in accordance with
this formula, the elementary decrease in the potential over
this displacement is
—dcp = E•dl. (1.24)
In other words, if we know the field E (r), then to find cp
we must represent E•d1 (with the help of appropriate trans-
formations) as a decrease in a certain function. This function
will be the potential cp.
Let us apply this method for finding the potential of the
field of a fixed point charge:

q dr
E • d1= er • dl — q
r2.
43-Leo r 2 4aeo
1 q
=—d
4a80 r + const) ,
where we took into account that er dl = 1. (dl),. = dr, since
the projection of dl onto er , and hence on r, is equal to the
increment of the magnitude of vector r, i.e. dr. The quantity
appearing in the parentheses under the differential is exact-
ly y (r). Since the additive constant contained in the for-
mula does not play any physical role, it is usually omitted in
order to simplify the expression for cp. Thus, the potential
of the field of a point charge is given by
1 q
P (1.25)
12 4a80 r •
The absence of an additive constant in this expression
indicates that we conventionally assume that the potential
is equal to zero at infinity (for r co).
Potential of the Field of a System of Charges. Let a system
consist of fixed point charges q1, q2, . . . . In accordance
with the principle of superposition, the field intensity at any
point of the field is given by E = E1 E 2 . . ,where
E1is the field intensity from the charge q1, etc. By using
formula (1.24), we can then write
E•d1 = (E1 E2 .) dl = + E2 .dl + . . .
dy2 — . . . =
where cp = ETi , i.e. the principle of superposition turns out
to be valid for potential as well. Thus, the potential of a
system of fixed point charges is given by
(1.26)
4nso
where r i is the distance from the point charge q i to the point

under consideration. Here we also omitted an arbitrary con-
stant. This is in complete agreement with the fact that any
real system of charges is bounded in space, and hence its
potential can be taken equal to zero at infinity.
If the charges forming the system are distributed contin-

uously, then, as before, we assume that each volume ele-
ment dV contains a "point" charge p dV, where p is the charge
density in the volume dV . Taking this into consideration,
we can write formula (1.26) in a different form:
1
4a80 p dV
r 9 (1.27)
where the integration is performed either over the entire

space or over its part containing the charges. If the charges
are located only on the surface S. we can write
c dS
(1.28)
= 43180 J r
where a is the surface charge density and dS is the element
of the surface S. A similar expression corresponds to the
case when the charges have a linear distribution.
Thus, if we know the charge distribution (discrete or con-
tinuous), we can, in principle, find the potential of any
system.
1.6. Relation Between Potential and Vector E
It is known that electric field is completely described by
vector function E (r). Knowing this function, we can find
the force acting on a charge under investigation at any point
of the field, calculate the work of field forces for any displace-
ment of the charge, and so on. And what do we get by
introducing potential? First of all, it turns out that if we
know the potential cp (r) of a given electric field, we can re-
construct the field E (r) quite easily. Let us consider this
question in greater detail.
The relation between cp and E can be established with the
help of Eq. (1.24). Let the displacement dl be parallel to
the X-axis; then dl = i dx, where i is the unit vector along
the X-axis and dx is the increment of the coordinate x. In
this case
E • dl = E • i dx = E„ dx,
where E x is the projection of vector E onto the unit vector i
(and not on the displacement dl!), A comparison of this ex-
1.6. Relation between Potential and Vector t 81
pression with formula (1.24) gives

-ex = —8(p/ax, (1.29)
where the symbol of partial derivative emphasizes that the
function (p (x, y, z) must be differentiated only with respect
to x, assuming that y and z are constant in this case.
In a similar way, we can obtain the corresponding expres-
sions for the projections E y and E z. Having determined
E x, Ey, and E z, we can easily find vector E itself:
E= — act' j 2_k).
ay (1.30)
The quantity in the parentheses is the gradient of the po-
tential (p (grad cp or p(p). We shall be using the latter, more
convenient notation and will formally consider vcp as the
product of a symbolic vector v and the scalar (p. Then
Eq. (1.30) can be represented in the form
E (1.31)
= VT'
i.e. the field intensity E is equal to the potential gradient
with the minus sign. This is exactly the formula that can
be used for reconstructing the field E if we know the
function (p (r).
Example. Find the field intensity E if the field potential has the
form: (1) cp (x, y) — axy, where a is a constant; (2) cp (r) = —a •r,
where a is a constant vector and r is the radius vector of a point under
consideration.
(1) By using formula (1.30), we obtain E = a (yi xi).
(2) Let us first represent the function cp as cp = axx
— — ay y
azz, where ax, ay and azare constants. Then with the help of for-
mula (1.30) we find E = ay j azk = a. It can be seen that in
this case the field E is uniform.
Let us derive one more useful formula. We write the right-

hand side of (1.24) in the form E•d1 = E 1 dl, where dl
Idl l is an elementary displacement and E t is the projec-
tion of vector E onto the displacement dl. Hence
El = —OTIOl, (1.32)
i.e. the projection of vector E onto the direction of the dis-

placement dl is equal to the directional derivative of the

potential (this is emphasized by the symbol of partial de-
rivative).
Equipotential Surfaces. Let us introduce the concept of
equipotential surface, viz. the surface at all points of which
potential cp has the same value. We shall show that vector E
at each point of the surface is directed along the normal to
the equipotential surface and towards the decrease in the
potential. Indeed, it follows
vi
4°2 k I4I
from formula (1.32) that the
projection of vector E onto
any direction tangent to the
equipotential surface at a given
point is equal to zero. This
means that vector E is norm-
al to the given surface. Fur-
ther, let us take a displace-
ment dl along the normal to the
surface, towards decreasing
Fig. 1.13 (p. Then Op < 0, and accord-
ing to (1.32), E 1 > 0, i.e.
vector E is directed towards decreasing IT, or in the direc-
tion opposite to that of the vector vp.
It is expedient to draw equipotential surfaces in such a
way that the potential difference between two neighbouring
surfaces be the same. Then the density of equipotential
surfaces will visually indicate the magnitudes of field inten-
sities at different points. Field intensity will be higher in
the regions where equipotential surfaces are denser ("the
potential relief is steeper").
Since vector E is normal to an equipotential surface every-
where, the field lines are orthogonal to these surfaces.
Figure 1.13 shows a two-dimensional pattern of an electric field.
The dashed lines correspond to equipotential surfaces, while the solid
lines to the lines of E. Such a representation can be easily visualized.
It immediately shows the direction of vector E, the regions where field
intensity is higher and where it is lower, as well as the regions with
greater steepness of the potential relief. Such a pattern can be used to
obtain qualitative answers to a number of questions, such as "In what
direction will a charge placed at a certain point move? Where is the
magnitude of the potential gradient higher? At which point will
the force acting on the charge be greater?" etc.
1.6. Relation Between Potential and Vector E 33
On the Advantages of Potential. It was noted earlier that

electrostatic field is completely characterized by vector
function E (r). Then what is the use of introducing potential?
There are several sound reasons for doing that. The concept
of potential is indeed very useful, and it is not by chance
that this concept is widely used not only in physics but in
engineering as well.
1. If we know the potential cp (r), we can easily calculate
the work of field forces done in the displacement of a point
charge q' from point 1 to point 2:
A l2 = (W1 - W2)1 (1.33)
where cp1and cp2 are the potentials at points 1 and 2. This
means that the required work is equal to the decrease in the
potential energy of the charge q' upon its displacement from
point 1 to 2. Calculation of the work of the field forces with
the help of formula (1.33) is not just very simple, but is
in some cases the only possible resort.
Example. A charge q is distributed over a thin ring of radius a.
Find the work of the field forces done in the displacement of a point
charge q' from the centre of the ring to infinity.
Since the distribution of the charge q over the ring is unknown, we
cannot say anything definite about the intensity E of the field created
by this charge. This means that we cannot calculate the work by eval-
uating the integral q E dl in this case. This problem can be easily
solved with the help of potential. Indeed, since all elements of the
ring are at the same distance a from the centre of the ring, the poten-
tial of this point is y, = ql4neo a. And we know that p = 0 at infinity.
Consequently, the work A = q'yo =q'q/4asoa.
2. It turns out in many cases that in order to find electric
field intensity E, it is easier first to calculate the potential
cp and then take its gradient than to calculate the value of E
directly. This is a considerable advantage of potential. In-
deed, for calculating cp, we must evaluate only one integral,
while for calculating E we must take three integrals (since
it is a vector). Moreover, the integrals for calculating cp are
usually simpler than those for Ex, Ey, and E.
Let us note here that this does not apply to a comparative-
ly small number of problems with high symmetry, in
which the calculation of the field E directly or with the help
of the Gauss theorem turns out to be much simpler.
3-0181
There are some other advantages in using potential which

will be discussed later.
1.7. Electric Dipole

The Field of a Dipole. The electric dipole is a system of two
equal in magnitude unlike charges + q and — q, separated
by a certain distance 1. When the dipole field is considered,
it is assumed that the dipole itself is pointlike, i.e. the dis-

tance r from the dipole to the points under consideration is
assumed to be much greater than 1.
The dipole field is axisymmetric. Therefore, in any plane
passing through the dipole axis the pattern of the field is
the same, vector E lying in this plane.
Let us first find the potential of the dipole field and then
its intensity. According to (1.25), the potential of the dipole
field at the point P (Fig. 1.14a) is defined as
1 q (r_-- r+)
= 4nso k q
r_ 42160
Since r » 1, it can be seen from Fig. 1.14a that r_. ^ r+
= 1 cos 15 and r+r_ = r2, where r is the distance from the
point P to the dipole (it is pointlikel). Taking this into
account, we get
p cos15
(1.34)
(1)— 43-teo r2 2
where p = ql is the electric moment of the dipole. This quan-
tity corresponds to a vector directed along the dipole axis
from the negative to the positive charge:

p = ql, (1.35)
where q > 0 and 1 is the vector directed as p.
It can be seen from formula (1.34) that the dipole field
depends on its electric moment p. It will be shown below
that the behaviour of the dipole in an external field also
depends on p. Consequently, p is an important characteristic
of the dipole.
It should also be noted that the potential of the dipole field
decreases with the distance r faster than the potential of the
field of a point charge (in proportion to 1/r2 instead of 1/r).
In order to find the dipole field, we shall use formula (1.32)
and calculate the projections of vector E onto two mutually
perpendicular directions along the unit vectors e,. and eo
(Fig. 1.14b):
ay 1 2p cos*
Er = — ar
4aso r3
ay
(1.36)
1 p sin A5
-
ra* 43-ceo r3
Hence, the modulus of vector E will be
E ET--+E — 431180 713,3 + 3 cosz (1.37)
In particular, for 0 = 0 and 0 It/2 we obtain the ex-
pressions for the field intensity on the dipole axis (E ) and
on the normal to it (E l):
l 2p 1
ll = 4neo r3
E = 4a8o (1.38)
r39 E
i.e. for the same r the intensity E 11 is twice as high as El.

The Force Acting on a Dipole. Let us place a dipole into
a nonuniform electric field. Suppose that E + and E_ are
the intensities of the external field at the points where the
positive and negative dipole charges are located. Then the
resultant force F acting on the dipole is (Fig. 1.15a):
F qE+ qE_ =- q (E+— E_).
The difference E+— E_ is the increment AE of vector E
on the segment equal to the dipole length 1 in the direction
3*
of vector 1. Since the length of this segment is small, we can

write
E+ = of 1 = 00E1 1.
AE =
Substituting this expression into the formula for F, we

find that the force acting on the dipole is equal to
aE
F=p (1.39)
where p = ql is the dipole electric moment. The derivative

appearing in this expression is called the directional deriva-
tive of the vector. The symbol of
(a) partial derivative indicates that it
qE÷
is taken with respect to a certain
direction, viz. the direction coincid-
ing with vector 1 or p.
Unfortunately, the simplicity of
formula (1.39) is delusive: taking
the derivative OE/0/ is a rather com-
plicated mathematical operation.
(b) We shall not discuss this question
in detail here but pay attention to
E_ the essence of the obtained result.
Fig. 1.15
First of all, note that in a uniform
field oval = 0 and F = 0. This
means that generally the force acts
on a dipole only in a nonuniform field. Next, in the gener-
al case the direction of F coincides neither with vector E
nor with vector p. Vector F coincides in direction only
with the elementary increment of vector E, taken along the
direction of 1 or p (Fig. 1.15b).
Figure 1.16 shows the directions of the force F acting on a dipole
in the field of a point charge q for three different dipole orientations.
We suggest that the reader prove independently that it is really so.
If we are interested in the projection of force F onto a
certain direction X, it is sufficient to write equation (1.39)
in terms of the projections onto this direction, and we get
OE„
F x =-- p al , (1.40)
where 0Ex/0/ is the derivative of the corresponding projec-

tion of vector E again onto the direction of vector 1 or p.
Let a dipole with moment p be oriented along the symme-
try axis of a certain nonuniform field E. We take the positive
direction of the X-axis, for example,
as shown in Fig. 1A7. Since the incre- q• .)..---F
ment of the projection Ea, in the direc-
tion of vector p will be negative, Fx<0, ft
and hence vector F is directed to the t
left, i.e. towards increasing field in-
tensity. If we rotate vector p shown in
the figure through 90° so that the dipole F
q
centre coincides with the symmetry
axis of the field, it can be easily seen
that in this position F, = 0. Fig. 1.16
The Moment of Forces Acting on a
Dipole. Let us consider behaviour of a dipole in an exter-
nal electric field in its centre-of-mass system and find out
whether the dipole will rotate or not. For this purpose, we
must find the moment of external forces with respect to the
dipole centre of mass*.
By definition, the moment of forces F+ = qE+ and F_ =
= —qE_ with respect to the centre of mass C (Fig. 1.18) is
equal to
M = [r+ x F+] -1- [r_ X F_] = [r+ X qE+] — [r_ X qE_],
where r + and r_ are the radius vectors of the charges + q
and — q relative to the point C. For a sufficiently small di-
pole length, E+ E_ and M = [(r+ — r_) X qE]. It re-
mains for us to take into account that r+ — r_ = 1 and
ql = p, which gives
I M=
[p x El.
This moment of force tends to rotate the dipole so that

(1.41)
its electric moment p is oriented along the external field E.

Such a position of the dipole is stable.
Thus, in a nonuniform electric field a dipole behaves as
* We take the moment with respect to the centre of mass in order
to eliminate the moment of inertial forces.
follows: under the action of the moment of force (1.41),

the dipole tends to get oriented along the field (p fifi E),
while under the action of the resultant force (1.39) it is
Fig. 1.17 Fig. 1.18

displaced towards the region where the field E has larger
magnitude. Both these motions are simultaneous.
The Energy of a Dipole in an External Field. We know
that the energy of a point charge q in an external field is
W = qcp, where cp is the field potential at the point of lo-
cation of the charge q. A dipole is the system of two charges,
and hence its energy in an external field is
W = q+cp + =-. q ((p+
where cp+and cp_ are the potentials of the external field at
the points of location of the charges +q and —q. To within
a quantity of the second order of smallness, we can write
acp
(14-9)-- —a--1,
where Ow/a/ is the derivative of the potential in the direc-
tion of the vector I. According to (1.32), 4101 = —E 1,
and hence IT + — = —E 11 = —E-1, from which we get
W = — p•E. (1.42)
It follows from this formula that the dipole has the mini-
mum energy (Wniin =—pE) in the position p if E (the
position of stable equilibrium). If it is displaced from this
Problems 39
position, the moment of external forces will return the di-

pole to the equilibrium position.
Problems
• 1.1. A very thin disc is uniformly charged with surface charge
density a > 0. Find the electric field intensity E on the axis of this
disc at the point from which the disc is seen at an angle Q.
Fig. 1.19
Solution. It is clear from symmetry considerations that on the disc

axis vector E must coincide with the direction of this axis (Fig. 1.19).
Hence, it is sufficient to find the component dE, from the charge of
the area element dS at the point A and then integrate the obtained
expression over the entire surface of the disc. It can be easily seen
that
1 cr
—dS cos O. (1)
4n80 r 2
In our case (dS cos 0)/r2 = d52 is the solid angle at which the area
element dS is seen from the point A, and expression (1) can be writ-
ten as
1
dE, = ad52.
4neo
Hence, the required quantity is
1
E— aa52.
4n eo
It should be noted that at large distances from the disc, 52 = SIr2,
where S is the area of the disc and E = q/4nco r2 just as the field of the
point charge q = aS. In the immediate vicinity of the point 0, the
solid angle 52 = 2n and E = a/280.
• 1.2. A thin nonconducting ring of radius R is charged with a

linear density X, = 2,„ cos cp, where 7,„ is a positive constant and q,
is the azimuth angle. Find the electric field intensity E at the centre
of the ring.
Solution. The given charge distribution is shown in Fig. 1.20.
The symmetry of this distribution implies that vector E at the point 0
is directed to the right, and its magnitude is equal to the sum of the
projections onto the direction of E of vectors dE from elementary
charges dq. The projection of vector dE onto vector E is
1 dq
dE cos cp= —
43180 cos cp, (1)
where dq = 2%.R dcp = X„R cos cp dcp. Integrating (1) over cp between 0
and 2n, we find the magnitude of the vector E:
2n
E= X° S.cost cp dcp — .X°
asof? 480R
0
It should be noted that this integral is evaluated in the most simple
way if we take into account that (cost cp) = 1/2. Then
2n
cosy Tdcp= (cos2cp) 2n = n.
0
• 1.3. A semi-infinite straight uniformly charged filament has a
charge k per unit length. Find the magnitude and the direction of the
field intensity at the point separated from the filament by a distance y
and lying on the normal to the filament, passing through its end.
Solution. The problem is reduced to finding Ex and Ey, viz. the
projections of vector E (Fig. 1.21, where it is assumed that > 0).
Let us start with Ex. The contribution to Ex from the charge element
of the segment dx is
dx
dE x -= sin a. (1)
4n1 s, r2
Let us reduce this expression to the form convenient for integration.
In our case, dx = r da/cos a, r = y/cos a. Then
dEx= da.
4 nsoy sin a
Integrating this expression over a between 0 and n/2, we find
Ex =X/4nse.
In order to find the projection Ey, it is sufficient to recall that
dEydiffers from dE, in that sin a in (1) is simply replaced by cos a.
Problems 41
This gives
dEy = (X, cos a dot)143180 and Ey=- 2143-ceo y.
We have obtained an interesting result: Ex = Eyindependently of y,
i.e. vector E is oriented at the angle of 45° to the filament. The modu-
lus of vector E is
Ef/ = 17274n e 0y
1.4. The Gauss theorem. The intensity of an electric field de-
pends only on the coordinates x and y as follows:
E = a (xi Yi)I(x2 + Y2),
where a is a constant, and i and j are the unit vectors of the X and
-
Y-axes. Find the charge within a sphere of radius' R with the centre
at the origin.
Solution. In accordance with the Gauss theorem, the required
charge is equal to the flux of E through this sphere, divided by so.
In our case, we can determine the flux as follows. Since the field E is
axisymmetric (as the field of a uniformly charged filament), we arrive
at the conclusion that the flux through the sphere of radius R is equal
to the flux through the lateral surface of a cylinder having the same
radius and the height 2R, and arranged as shown in Fig. 1.22. Then
q= E dS = so ErS ,
where Er = alit and S = 2nR .2R = 4xR2. Finally, we get

q = 4nsoaR.
• 1.5. A system consists of a uniformly charged sphere of radius
R and a surrounding medium filled by a charge with the volume den-
sity p = oar, where a is a positive constant and r is the distance from
the centre of the sphere. Find the charge of the sphere for which the
electric field intensity E outside the sphere is independent of r. Find
the value of E.
Solution. Let the sought charge of the sphere be q. Then, using the
Fig. 1.22 Fig. 1.23

Gauss theorem, we can write the following expression for a spherical
surface of radius r (outside the sphere with the charge q):
E•43cr2 =–-1- 1 f 4nr 2 dr.

6o 6o
After integration, we transform this equation to

E.4nr 2 = (q 2inzR2)/e,
— 4itar2/280.
The intensity E does not depend on r when the expression in the paren-
theses is equal to zero. Hence
q = 2aaR2 and E =--- cd2so .
1.6. Find the electric field intensity E in the region of inter-
section of two spheres uniformly charged by unlike charges with the
volume densities p and —p, if the distance between the centres of the
spheres is determined by vector I (Fig. 1.23).
Solution. Using the Gauss theorem, we can easily show that the
electric field intensity within a uniformly charged sphere is
E = (p/380) r,
where r is the radius vector relative to the centre of the sphere. We
can consider the field in the region of intersection of the spheres as the
superposition of the fields of two iiniforlmy charged spheres. Then at
an arbitrary point A (Fig. 1.24) of this region we have
E = E+ E_ = p (r+ — r_)/380 = p1/3so.
Thus, in the region of intersection of these spheres the field is uni-
form. This conclusion is valid regardless of the ratio between the radii
Problems 43
of the spheres and of the distance between their centres. In particular,

it is valid when one sphere is completely within the other or, in other
words, when there is a spherical cavity in a sphere (Fig. 1.25).
• 1.7. Using the solution of the previous problem, find the field
intensity E inside the sphere over which a charge is distributed with
Fig. 1.24 Fig. 1.25

the surface density a = a, cos 0, where a, is a constant and 0 is
the polar angle.
Solution. Let us consider two spheres of the same radius, having
uniformly distributed volume charges with the densities p and —p.
Suppose that the centres of the spheres are sep-
arated by the distance 1 (Fig. 1.26). Then, in
accordance with the solution of the previous
problem, the field in the region of intersection
of these spheres will be uniform:
E (p/38,) 1. (1)
In our case, the volume charge differs from zero
only in the surface layer. For a very small 1,
we shall arrive at the concept of the surface
charge density on the sphere. The thickness of
the charged layer at the points determined by
angle 15 (Fig. 1.26) is equal to 1 cos 0. Hence,
in this region the charge per unit area is
a = pl cos 0 = a, cos 0, where a, = pl, and
expression (1) can be represented in the form Fig. 1.26
E = —(a0 /380) k,
where k is the unit vector of the Z-axis from which the angle 19 is
measured.
• 1.8. Potential. The potential of a certain electric field has the
form cp = a (xy — z2). Find the projection of vector E onto the di-
rection of the vector a = i 3k at the point M (2, 1, —3).
Solution. Let us first find vector E:
E= = —a (yi 4 - 2zk).
The sought projection is

— a (yi—xj— 2zk) (1+ 3k) —a(y — 6z)
Ea— E • a
+ 32 1/16 •
At the point M we have
—a (1+18) 19
Ea — a.
VIM 110
• 1.9. Find the potential cp at the edge of a thin disc of radius R
with a charge uniformly distributed over one of its sides with the sur•
face density a.
Solution. By definition, the potential in the case of a surface charge
distribution is defined by integral (1.28). In order to simplify in-
tegration, we shall choose the area element dS in the form of a part
of the ring of radius r and width dr (Fig. 1.27). Then dS = 215r dr,
r = 2R cos 0, and dr = —2R sin * dO. After substituting these ex-
pressions into integral (1.28), we obtain the expression for p at the
point 0:
0
uR
=- — sin * dO.
ago
n/2
We integrate by parts, denoting = u
and sin = dv:
.f 0 sin dft = — cos
Fig. 1.27 + f cos 19 dO = —0 cos 15+ sin 0,

which gives —1 after substituting the limits of integration. As a re-
sult, we obtain
aRhteo.
• 1.10. The potential of the field inside a charged sphere depends
only on the distance r from its centre to the point under consideration
in the following way: fp = ar2 b, where a and b are constants.
Find the distribution of the volume charge p (r) within the sphere.
Solution. Let us first find the field intensity. According to (1.32),
we have
Er = —8(plar = —2ar. (1)
Then we use the Gauss theorem: 4o-cr2E7. = q/so. The differential of
this expression is
1 1
4n d(r2Er)= — dq = — p•4ar2 dr,
so
2.1. Field in a Substance 45
where dq is the charge contained between the spheres of radii r and

r ± dr. Hence
OE r 2E p
r2 dEr +2rE, dr =- 1 pre dr,
813 Or -1- r e0
Substituting (1) into the last equation, we obtain
p = —6c0a,
i.e. the charge is distributed uniformly within the sphere.
• 1.11. Dipole. Find the force of interaction between two point
dipoles with moments pi and p2, if the vectors pi and P2 are directed
along the straight line connecting the dipoles and the distance be-
tween the dipoles is 1.
Solution. According to (1.39), we have
F = Pr 18E101 I,
where E is the field intensity from the dipole p2, determined by the
first of formulas (1.38):
1 2/32
E —
4;te0 /3 •
Taking the derivative of this expression with respect to 1 and sub-
stituting it into the formula for F, we obtain
1 6pi p2
F=
43180 /4 '
It should be noted that the dipoles will be attracted when pi t t p2 and
repulsed when pi ti p2.
2. A Conductor in an Electrostatic Field

2.1. Field in a Substance
Micro- and Macroscopic Fields. The real electric field in
any substance (which is called the microscopic field) varies
abruptly both in space and in time. It is different at differ-
ent points of atoms and in the interstices. In order to find
the intensity E of a real field at a certain point at a given
instant, we should sum up the intensities of the fields of all
individual charged particles of the substance, viz, electrons
and nuclei. The solution of this problem is obviously not
feasible. In any case, the result would be so complicated
46 2. A Conductor in an Electrostatic Field
that it would be impossible to use it. Moreover, the knowledge

of this field is not required for the solution of macroscop-
ic problems. In many cases it is sufficient to have a simpler
and rougher description which we shall be using henceforth.
Under the electric field E in a substance (which is called
the macroscopic field) we shall understand the microscopic
field averaged over space (in this case time averaging is
superfluous). This averaging is performed over what is called
a physically infinitesimal volume, viz, the volume contain-
ing a large number of atoms and having the dimensions
that are many times smaller than the distances over which
the macroscopic field noticeably changes. The averaging over
such volumes smoothens all irregular and rapidly varying
fluctuations of the microscopic field over the distances of
the order of atomic ones, but retains smooth variations of
the macroscopic field over macroscopic distances.
Thus, the field in the substance is
E = Ema cro =(EmIcro)• (2.1)
The Influence of a Substance on a Field. If any substance is
introduced into an electric field, the positive and negative
charges (nuclei and electrons) are displaced, which in turn
leads to a partial separation of these charges. In certain
regions of the substance, uncompensated charges of different
signs appear. This phenomenon is called the electrostatic
induction, while the charges appearing as a result of sepa-
ration are called induced charges.
Induced charges create an additional electric field which
in combination with the initial (external) field forms the
resultant field. Knowing the external field and the distribu-
tion of induced charges, we can forget about the presence of
the substance itself while calculating the resultant field,
since the role of the substance has already been taken into
account with the help of induced charges.
Thus, the resultant field in the presence of a substance is
determined simply as the superposition of the external field
and the field of induced charges. However, in many cases
the situation is complicated by the fact that we do not know
beforehand how all these charges are distributed in space,
and the problem turns out to be not as simple as it could
seem at first sight. It will be shown later that the distri-
2.2. Fields Inside and Outside a Conductor 47
bution of induced charges is mainly determined by the prop-

erties of the substance, i.e. its physical nature and the
shape of the bodies. We shall have to consider these ques-
tions in greater detail.
2.2. Fields Inside and Outside a Conductor

Inside a Conductor E = 0. Let us place a metallic conduc-
tor into an external electrostatic field or impart a certain
charge to it. In both cases, the electric field will act on all
the charges of the conductor, and as a result all the negative
charges (electrons) will be displaced in the direction against
the field. This displacement (current) will continue until
(this practically takes a small fraction of a second) a cer-
tain charge distribution sets in, at which the electric field
at all the points inside the conductor vanishes. Thus, in the
static case the electric field inside a conductor is absent
(E = 0).
Further, since E = 0 everywhere in the conductor, the
density of excess (uncompensated) charges inside the conduc-
tor is also equal to zero at all points (p = 0). This can be
easily explained with the help of the Gauss theorem. Indeed,
since inside the conductor E = 0, the flux of E through any
closed surface inside the conductor is also equal to zero.
And this means that there are no excess charges inside the
conductor.
Excess charges appear only on the conductor surface with
a certain density o- which is generally different for different
points of the surface. It should be noted that the excess
surface charge is located in a very thin surface layer (whose
thickness amounts to one or two interatomic distances).
The absence of a field inside a conductor indicates,, in
accordance with (1.31), that potential cp in the conductor
has the same value for all its points, i.e. any conductor in
an electrostatic field is an equipotential region, its surface
being an equipotential surface.
The fact that the surface of a conductor is equipotential
implies that in the immediate vicinity of this surface the
field E at each point is directed along the normal to the sur-
face. If the opposite were true, the tangential component of
E would make the charges move over the surface of the
conductor, i.e. charge equilibrium would be impossible.

Example. Find the potential of an uncharged conducting sphere
provided that a point charge q is located at a distance r from its centre
(Fig. 2.1).
Potential cp is the same for all points
of the sphere. Thus we can calculate
its value at the centre 0 of the sphere,
because only for this point it can
be calculated in the most simple way:
1 q
(1)= (1)
4 3-ce o r
where the first term is the potential of
Fig. 2.1 the charge q, while the second is the
potential of the charges induced on
the surface of the sphere. But since all induced charges are at the same
distance a from the point 0 and the total induced charge is equal to
zero, cp' =0 as well. Thus, in this case the potential of the sphere will
be determined only by the first term in (1).
Figure 2.2 shows the field and the charge distributions for
a system consisting of two conducting spheres one of which
(left) is charged. As a result of electric induction, the charges
of the opposite sign appear on the surface of the right
uncharged sphere. The field of these charges will in turn
cause a redistribution of charges on the surface of the left
sphere, and their surface distribution will become nonuni-
form. The solid lines in the figure are the lines of E, while
the dashed lines show the intersection of equipotential sur-
faces with the plane of the figure. As we move away from
this system, the equipotential surfaces become closer and
closer to spherical, and the field lines become closer to ra-
dial. The field itself in this case resembles more and more the
field of a point charge q, viz. the total charge of the given
system.
The Field Near a Conductor Surface. We shall show that
the electric field intensity in the immediate vicinity of the
surface of a conductor is connected with the local charge
density at the conductor surface through a simple relation.
This relation can be established with the help of the Gauss
theorem.
Suppose that the region of the conductor surface we are
interested in borders on a vacuum. The field lines are nor-
mal to the conductor surface. Hence for a closed surface we
Forces Acting on the Surface of a Conductor 49
shall take a small cylinder and arrange it as is shown in

Fig. 2.3. Then the flux of E through this surface will be
equal only to the flux through the "outer" endface of the
cylinder (the fluxes through the lateral surface and the
inner endface are equal to zero). Thus we obtain En AS =
Fig. 2.2 Fig. 2.3

= aAS/co, where En is the projection of vector E onto
the outward normal n (with respect to the conductor), AS
is the cross-sectional area of the cylinder and a is the local
surface charge density of the conductor. Cancelling both
sides of this expression by AS , we get
En =a/80. I (2.2)
If a > 0, then En >0, i.e. vector E is directed from the
conductor surface (coincides in direction with the normal n).
If a < 0, then En < 0, and vector E is directed towards the
conductor surface.
Relation (2.2) may lead to the erroneous conclusion that
the field E in the vicinity of a conductor depends only on
the local charge density a. This is not so. The intensity E
is determined by all the charges of the system under con-
sideration as well as the value of a itself.
2.3. Forces Acting on the Surface of a Conductor
Let us consider the case when a charged region of the sur-
face of a conductor borders on a vacuum. The force acting
on a small area AS of the conductor surface is
AF = aAS.E0, (2.3)
4 — 0 18 1
2. A Conductor in an Electrostatic Field
where a AS is the charge of this element and E0 is the,.-.field

created by all the other charges of the system din the region
where the charge a AS is located. It should be noted at the
very outset that E0 is not equal to the field intensity E in
the vicinity of the given surface element of the conductor,
although there exists a certain relation between them. Let
us find this relation, i.e. express E0 through E.
Let Ea be the intensity of the field created by the charge
on the area element AS at the points that are very close to
this element. In this region, it behaves as an infinite uni-
formly charged plane. Then, in accordance with (1.10), E„ =
= a/280.
The resultant field both inside and outside the conductor
(near the area element AS) is the superposition of the fields
E0 and E0. On both sides of the area element AS the field E0
is practically the same, while the field E, has opposite di-
rections (see Fig. 2.4 where for the sake of definiteness it is
assumed that a > 0). From the condition E = 0 inside the
conductor, it follows that E, = E0, and then outside the
conductor, near its surface, E = E0 E, =2E0. Thus,
E0 = E/2, (2.4)
E =2E0
and Eq. (2.3) becomes
1
4F ----- — u4S•E.
2
(2.5)
Dividing both sides of this

equation by AS, we obtain the
expression for the force acting
on unit surface of a conductor:
1
Fu =nE. (2.6)
Fig. 2.4 We can write this expression
in a different form since the
quantities a andtE appearing in it are interconnected. Indeed,
in accordance with (2.2), Er, = also, or E = (a/e0) n, where
n is the outward normal to the surface element at a given
point of the conductor. Hence
80E 2
n= 2 n, (2.7)
2A. Properties of a Closed Conducting Shell
where we took into account that a = e 0En1 and El, = E2.

The quantity Fu is called the surface density of force. Equa-
tion (2.7) shows that regardless of the sign of a, and hence
of the direction of E, the force Fuis always directed outside
the conductor, tending to stretch it.
Example. Find the expression for the electric force acting in a
vacuum on a conductor as a whole, assuming that the field intensities
E are known at all points in the vicinity of the conductor surface.
Multiplying (2.7) by dS, we obtain the expression for the force dF
acting on the surface element dS:
1
dF=— E2 dS
2 ° ,
where dS = n dS. The resultant force acting on the entire conductor
can be found by integrating this equation over the entire conductor
surface:
0
F= 1) E2 dS.
2
2.4. Properties of a Closed Conducting Shell

It was shown that in equilibrium there are no excess charges
inside a conductor, viz. the material inside the conduc-
tor is electrically neutral. Consequently, if the substance is
removed from a certain volume inside a conductor (a closed
cavity is created), this does not change the field anywhere,
i.e. does not affect the equilibrium distribution of charges.
This means that the excess charge is distributed on a
conductor with a cavity in the same way as on a uniform
conductor, viz. on its outer surface.
Thus, in the absence of electric charges within the cavity
the electric field is equal to zero in it. External charges,
including the charges on the outer surface of the conductor, do
not create any electric field in the cavity inside the conductor.
This forms the basis of electrostatic shielding, i.e. the screening
of bodies, e.g. measuring instruments, from the influence
of external electrostatic fields. In practice, a solid conduct-
ing shell can be replaced by a sufficiently dense metallic
grating.
That there is no electric field inside an empty cavity can
be proved in a different way. Let us take a closed surface S
enveloping the cavity and lying completely in the material
4*
of the conductor. Since the field E is equal to zero inside the

conductor, the flux of E through the turface S is also equal
to zero. Hence, in accordance with the Gauss theorem, the
total charge inside S is equal to zero as well. This does not
exclude the situation depicted in Fig 2.5, when the surface
Fig. 2.5 Fig. 2.6
of the cavity itself contains equal quantities of positive

and negative charges. However, this assumption is prohibit-
ed by another theorem, viz. the theorem on circulation of
vector E. Indeed, let the contour F cross the cavity along
one of the lines of E and be closed in the conductor material.
It is clear that the line integral of vector E along this contour
differs from zero, which is in contradiction with the theorem
on circulation.
Let us now consider the case when the cavity is not empty
but contains a certain electric charge q (or several charges).
Suppose also that the entire external space is filled by a
conducting medium. In equilibrium, the field in this medium
is equal to zero, which means that the medium is electrical-
ly neutral and contains no excess charges.
Since E = 0 inside the conductor, the field flux through a
closed surface surrounding the cavity is also equal to zero.
According to the Gauss theorem, this means that the alge-
braic sum of the charges within this closed surface is equal to
zero as well. Thus, the algebraic sum of the charges induced
on the cavity surface is equal in magnitude and opposite in
sign to the algebraic sum of the charges inside the cavity.
In equilibrium the charges induced on the surface of the
cavity are arranged so as to compensate completely, in the
2.4. Properties of a Closed Conducting Shell 53
space outside the cavity, the field created by the charges

located inside the cavity.
Since the conducting medium is electrically neutral every-
where, it does not influence the electric field in any way.
Therefore, if we remove the medium, leaving only a conduc-
ting shell around the cavity, the field will not be changed
anywhere, and will remain equal to zero beyond this shell.
Thus, the field of the charges surrounded by a conducting
shell and of the charges induced on the surface of the cavity
(on the inner surface of the shell) is equal to zero in the en-
tire outer space.
We arrive at the following important conclusion: a closed
conducting shell divides the entire space into the inner and
outer parts which are, completely independent of one another
in respect of electric fields. This must be interpreted as follows:
any arbitrary displacement of charges inside the shell does
not introduce any change in the field of the Outer space,
and hence the charge distribution on the outer surface of the
shell remains unchanged. The same refers to the field inside
the cavity (if it contains charges) and to the distribution of
charges induced on the cavity walls. They will also remain
unchanged upon the displacement of charges outside the
shell. Naturally, the above arguments are applicable only
in the framework of electrostatics.
Example. A point charge q is within an electrically neutral shell
whose outer surface has spherical shape (Fig. 2.6). Find the potential
cp at the point P lying outside the shell at a distance r from the
centre 0 of the outer surface.
The field at the point P is determined only by charges induced on
the outer spherical surface since, as was shown above, the field of the
point charge q and of the charges induced on the inner surface of the
sphere is equal to zero everywhere outside the cavity. Next, in view of
symmetry, the charge on the outer surface of the shell is distributed
uniformly, and hence
(19 4nso r•
An infinite conducting plane is a special case of a closed
conducting shell. The space on one side of this plane is
electrically independent of the space on its other side.
We shall repeatedly use this property of a closed con-
ducting shell.
2.5. General Problem of Electrostatics.

Image Method
Frequently, we must solve problems in which the charge
distribution is unknown but the potentials of conductors,
their shape and relative arrangement are given. We must
find the potential cp (r) at any point of the field between the
conductors. It should be recalled that if we know the poten-
tial cp (r), the field E (r) itself can be easily reconstructed
and then its value in the immediate vicinity of the conductor
surfaces can be used for determining the surface charge
distribution for the conductors.
The Poisson and Laplace equations. Let us derive the differential
equation for the function cp (potential). For this purpose, we substitute
into the left-hand side of (1.20) the expression for E in terms of IT,
i.e. E = —vcp. As a result, we obtain the general differential equation
for potential, which is called the Poisson equation:
VT= — P/go, (2.8)
where v2 is the Laplace operator (Laplacian). In Cartesian coordinates

it has the form
02 02
v2 = 02
ax2 + ay2 + az2 '
i.e. is the scalar product v -v [see (1.19)].
If there are no charges between the conductors (p = 0), Eq. (2.8)
is transformed into a simpler equation viz. the Laplace equation:
V2 cp =O. (2.9)
To determine potential, we must find a function cp which satisfies

Eqs. (2.8) or (2.9) in the entire" space between the conductors and ac-
quires the given values (pi, (1)2, . . . on the surfaces of the conductors.
It can be proved theoretically that this problem has a uni-

que solution. This statement is called the uniqueness theo-
rem. From the physical point of view, this conclusion is
quite obvious: if there are more than one solution, there
will be several potential "reliefs", and hence the field E at
each point generally has not a single value. Thus we arrive
at a physically absurd conclusion.
Using the uniqueness theorem, we can state that in a
2.5. General Problem of Electrostatics 55
static case the charge is distributed over the surface of a

conductor in a unique way as well. Indeed, there is a one-to-
one correspondence (2.2) between the charges on the conduc-
tor and the electric field in the vicinity of its surface: r =
q
(a) (b) (c)
Fig. 2.7
= coEn. Hence it immediately follows that the uniqueness

of the field E determines the uniqueness of the charge
distribution over the conductor surface.
The solution of Eqs. (2.8) and (2.9) in the general case is a compli-
cated and laborious problem. The analytic solutions of these equations
were obtained only for a few particular cases. As for the uniqueness
theorem, it simplifies the solution of a number of electrostatic prob-
lems. If a solution of the problem satisfies the Laplace (or Poisson)
equation and the boundary conditions, we can state that it is correct
and unique regardless of the methods by which it was obtained (if
only by guess).
Example. Prove that in an empty cavity of a conductor the field
is absent.
The potential cp must satisfy the Laplace equation (2.9) inside the
cavity and acquire a certain value Toat the cavity's walls. The solution
of the Laplace equation satisfying this condition can immediately be
found. It is cp = To. In accordance with the uniqueness theorem, there
can be no other solutions. Hence, E = —VT = 0.
Image Method. This is an artificial method that makes it
possible to calculate in a simple way the electric field in
some (unfortunately few) cases. Let us consider this method
by using a simple example of a point charge q near an infinite
conducting plane (Fig. 2.7a).
The idea of this method lies in that we must find another
problem which can be easily solved and whose solution or a
part of it can be used in our problem. In our case such a sim-
ple problem is the problem about two charges: q and —q.
The field of this system is well known (its equipotential

surfaces and field lines are shown in Fig. 2.7b).
Let us make the conducting plane coincide with the mid-
dle equipotential surface (its potential cp = 0) and remove
the charge — q. According to the uniqueness theorem, the
field in the upper half-space will remain unchanged. Indeed,
0 on the conducting plane and everywhere at infinity.
The point charge q can be considered to be the limiting case
of a small spherical conductor whose radius tends to zero
and potential to infinity. Thus, the boundary conditions for
the potential in the upper half-space remain the same, and
hence the field in this region is also the same (Fig. 2.7c).
It should be noted that we can arrive at this conclusion
proceeding from the properties of a closed conducting shell
[see Sec. 2.41, since both half-spaces separated by the con-
ducting plane are electrically independent of one another,
and the removal of the charge q will not affect the field
—
in the upper half-space.

Thus, in the case under consideration the field differs from
zero only in the upper half-space. In order to calculate this
field, it is sufficient to introduce a fictitious image charge
= q, opposite in sign to the charge q, by placing it on
—
the other side of the conducting plane at the same distance

as the distance from q to the plane. The fictitious charge q'
creates in the upper half-space the same field as that of the
charges induced on the plane. This is precisely what' is
meant when we say that the fictitious charge produces the same
"effect" as all the induced ,charges. We must only bear in
mind that the "effect" of the fictitious charge extends only to
the half-space where the real charge q is located. In another
half-space the field is absent.
Summing up, we can say that the image method is essen-
tially based on driving the potential to the boundary con-
ditions, i.e. we strive to find' another problem (configura-
tion of charges) in which the field configuration in the region
of space we are interested in is the same. The image method
proves to be very effective if this can be done with the help
of sufficiently simple configurations. Let us consider one
more example.
Example. A point charge g is placed between two mutually perpen-
dicular half-planes (Fig. 2.8a). Find the location of fictitious point
charges whose action on the charge q is equivalent to the action of all

charges induced on these half planes.
We have to find a system of point charges for which the equipoten-
tial surfaces with IT = 0 would
coincide with the conducting (a) (1)
half-planes. One or two fictitious
charges are insufficient in this
case; there should be three of
them (Fig. 2.8b). Only with such a
configuration of the system of four
charges we can realize the re-
quired "trimming", i.e. ensure
that the potential on the con-
ducting half-planes be equal to
zero. These three fictitious char-
ges create just the same field
within the "right angle" as the Fig. 2.8
field of the charges induced on
the conducting planes.
Having found this configuration of point charges (another problem),
we can easily answer a number of other questions, for example, find
the potential and field intensity on any point within the "right angle"
or determine the force acting on the charge q.
2.6. Capacitance. Capacitors

Capacitance of an Isolated Conductor. Let us consider a
solitary conductor, i.e. the conductor removed from other
conductors, bodies, and charges. Experiments show that the
charge q of this conductor is directly proportional to its
potential cp (we assumed that at infinity potential is equal to
zero): cp oc q. Consequently, the ratio Op does not depend
on the charge q and has a certain value for each solitary
conductor. The quantity
C (2.10)
is called the electrostatic capacitance of an isolated conductor
(or simply capacitance). It is numerically equal to the charge
that must be supplied to the conductor in order to increase
its potential by unity. The capacitance depends on the size
and shape of the conductor.
Example. Find the capacitance of an isolated conductor which has
the shape of a sphere of radius R.
It can be seen from formula (2.10) that for this purpose we mut t
mentally charge the conductor by a charge q and calculate its potert-
tial cp. In accordance with (1.23), the potential of a sphere is

00 00
1 1
=- Er dr — q dr .-------
—
4nso r2 43-ceo R •
Substituting this result into (2.10), we find

C = 4nsoR. (2.11)
The unit of capacitance is the capacitance of a conductor
whose potential changes by 1 V when a charge of 1 C is
supplied to it. This unit of capacitance is called the fa-
rad (F).
The farad is a very large quantity. It corresponds to the
capacitance of an isolated sphere 9 x 106 km in radius,
which is 1500 times the radius of the Earth (the capacitance
of the Earth is 0.7 mF). In actual practice, we encounter
capacitances between 1 tF and 1 pF.
Capacitors. If a conductor is not isolated, its capacitance
will considerably increase as other bodies approach it. This
is due to the fact that the field of the given conductor causes
a redistribution of charges on the surrounding bodies, i.e.
induces charges on them. Let the charge of the conductor
be q > 0. Then negative induced charges will be nearer to
the conductor than the positive charges. For this reason, the
potential of the conductor, which is the algebraic sum of the
potentials of its own charge and of the charges induced on
other bodies will decrease when other uncharged bodies
approach it. This means that its capacitance increases.
This circumstance made it possible to create the system
of conductors, which has a considerably higher capacitance
than that of an isolated conductor. Moreover, the capacitance
of this system does not depend on surrounding bodies.
Such a system is called a capacitor. The simplest capacitor
consists of two conductors (plates) separated by a small
distance.
In order to exclude the effect of external bodies on the
capacitance of a capacitor, its plates are arranged with re-
spect to one another in such a way that the field created by
the charges accumulated on them is concentrated almost
completely inside the capacitor. This means that the lines
of E emerging on one plate must terminate on the other,
i.e. the charges on the plates must be equal in magnitude

and opposite in sign (q and —q).
The basic characteristic of a capacitor is its capacitance.
Unlike the capacitance of an isolated conductor, the capaci-
tance of a capacitor is defined as the ratio of its charge to
the potential difference between the plates (this difference is
called the voltage):
C=q1U. (2.12)
The charge q of a capacitor is the charge of its positively

charged plate.
Naturally, the capacitance of a capacitor is also measured
in farads.
The capacitance of a capacitor depends on its geometry
(size and shape of its plates), the gap between the plates,
and the material that fills the capacitor. Let us derive
the expressions for the capacitances of some capacitors
assuming that there is a vacuum between their plates.
Capacitance of a Parallel-plate Capacitor. This capacitor
consists of two parallel plates separated by a gap of width h.
If the charge of the capacitor is q, then, according to (1.11),
the intensity of the field between its plates is E = u/so,
where u = qlS and S is the area of each plate. Consequent-
ly, the voltage between the plates is
U = Eh = qhleoS.
Substituting this expression into (2.12), we obtain
C = NAM. (2.13)
This calculation was made without taking into account
field distortions near the edges of the plates (edge effects).
The capacitance of a real plane capacitor is determined by
this formula the more accurately the smaller the gap h in
comparison with the linear dimensions of the plates.
Capacitance of a Spherical Capacitor. Let the radii of the
inner and outer capacitor plates be a and b respectively. If
the charge of the capacitor is q, field intensity between the
plates is determined by the Gauss theorem:
Er 4nso r-
The voltage of the capacitor is

q 1 1 1
U= E dr = 4ne, k a b) .
a
It can be easily seen that the capacitance of a spherical
capacitor is given by
ab
C = 4n8c, (2.14)
u —a •
It is interesting that when the gap between the plates is
small, i.e. when b — a < a (or b), this expression is re-
duced to (2.13), viz. the expression for the capacitance of a
parallel-plate capacitor.
Capacitance of a Cylindrical Capacitor. By using the same
line of reasoning as in the case of a spherical capacitor, we
obtain
2318
C — In (bla)
° (2.15)
where 1 is the capacitor's length, a and b are the radii of the

inner and outer cylindrical plates. Like in the previous
case, the obtained expression is reduced to (2.13) when the
gap between the plates is small.
The influence of the medium on the capacitance of a ca-
pacitor will be discussed in Sec. 3.7.
Problems
• 2.1. On the determination of potential. A point charge q is at
a distance r from the centre 0 of an uncharged spherical conducting
layer, whose inner and outer radii are equal to a and b respectively.
Find the potential at the point 0 if r < a.
Solution. As a result of electrostatic induction, say, negative charges
will be induced on the inner surface of the layer and positive char-
ges on its outer surface (Fig. 2.9). According to the principle of super-
position, the sought potential at the point 0 can be represented in
the form
1 ( q , a_ dS , C o-+ dS
q)= 4nso r+ 1-C
yb ) '
where the first integral is taken over all the charges induced on the
inner surface of the layer, while the second integral, over all the
Problems 61
charges on the outer surface. It follows from this expression that

I1 1 1\
(P— 4nso r a 7- b ) •
It should be noted that the potential can be found in such a simple
form only at the point 0 since all the like induced charges are at the
same distance from this point and their
distribution (which is unknown to us)
does not play any role.
• 2.2. A system consists of two
concentric spheres, the inner sphere of
radius R1having a charge q1. What cha-
rge q2must be placed onto the outer
sphere of radius R2 to make the poten-
tial of the inner sphere equal to zero?
What will be the dependence of poten-
tial cp on the distance r from the cen-
tre of the system? Plot schematically
the graph of this dependence, assuming
that ql < 0. Fig. 2.9
Solution. We write the expressions
for potentials outside the system ((pH) and in the region between the
spheres (cpi):
1 qi+42 1 q1
TII = -
41L80 r ' (Pi = 42.teo TO,
where To is a certain constant. Its value can be easily found from the
boundary condition: for r = R2, (PH = Ti. Hence
To = q2/4ate0R2.
From the condition TI (R1) = 0 we find that q2 = —q11/2/R1.
The cp (r), dependence (Fig. 2.10) will have the form:
q1 1—R21R1 41 f 1 1
Wii — 43180 r , (Pi= 41teo k r )
RI 1 •
• 2.3. The force acting on a surface charge.. An uncharged me-
tallic sphere of radius R is placed into an external uniform field, as a
result of which an induced charge appears on the sphere with surface
density a = a, cos *, where ao is a positive constant and is a polar
angle. Find the magnitude of the resultant electric force acting on
like charges.
Solution. According to (2.5), the force acting on the area element
dS is
1
dF — 6E dS. (1)
2
It follows from symmetry considerations that the resultant force F
is directed along the Z-axis (Fig. 2.11), and hence it can be represented
as the sum (integral) of the projections of elementary forces (1) onto

the Z-axis:
dF, = dF cos 0. (2)
It is expedient to take for the area element dS a spherical zone dS =
Z
Fig. 2.11
2xR sin 0.R c/0. Considering that E a/80, we transform (2) as

follows:
dF,= (2-t 0-2R2 ,80%) sin cos d* — (acFS1121s 0) cos3 0 d (cos 0).
Integrating this expression over the half-sphere (i.e. with respect to
cos 0 between 1 and 0), we obtain
F = maaR2/480.
• 2.4. Image method. A point charge q is at a distance / from an
infinite conducting plane. Find the density of surface charges induced
on the plane as a function of the distance r from the base of the per-
pendicular dropped from the charge q onto the plane.
Solution. According to (2.2), the surface charge density on a conduc-
tor is connected with the electric field near its surface (in vacuum)
through the relation a = co En. Consequently, the problem is reduced
to determining the field E in the vicinity of the conducting plane.
Using the image method, we find that the field at the point P
(Fig. 2.12) which is at a distance r from the point 0 is
1
E=2E cos cc — 2
4:-ceox2 x'
Hence
ql
a=
2:1 (12 + r2)3/2
Problems
where the minus sign indicates that the induced charge is opposite
to sign to the point charge q.
• 2.5. A point charge q is at a distance 1 from an infinite conduc-
ting plane. Find the work of the electric force acting on the charge q
done upon its slow removal to a very large distance from the plane.
Fig. 2.12 Fig. 2.13
Solution. By definition, the work of this force done upon an ele-

mentary displacement dx (Fig. 2.13) is given by
q2
621.=-Fxdx — dx,
4ns 0 (2x)2
where the expression for the force is obtained with the help of the
image method. Integrating this equation over x between / and oo, we
find
A
q2
16aso
r
00
dx
x2
q2
16nsol
Remark. An attempt to solve this problem in a different way (through

potential) leads to an erroneous result which differs from what
was obtained by us by a factor of two. This is because the relation
A = q (q)i. — cp2) is valid only for potential fields. However, in the
reference system fixed to the conducting plane, the electric field of
induced charges is not a potential field: a displacement of the charge
qleads to a redistribution of the induced charges, and their field turns
out to be time-dependent.
• 2.6. A thin conducting ring of radius R, having a charge q, is
arranged so that it is parallel to an infinite conducting plane at a dis-
tance 1 from it. Find (1) the surface charge density at a point of the
plane, which is symmetric with respect to the ring and (2) the electric
field potential at the centre of the ring.
Solution. It can be easily seen that in accordance with the image
method, a fictitious charge —q must be located on a similar ring but
2.A Conductor in an ElectrostaticField
on the other side of the conducting plane (Fig. 2.14). Indeed, only in
this case the potential of the midplane between these rings is equal to
zero, i.e. it coincides with the potential of the conducting plane. Let
us now use the formulas we already know.
(1) In order to find a at the point 0, we must, according to (2.2),
find the field E at this point (Fig. 2.14). The expression for E on the
R
F1
a
I a 12
(A I
:////////4/,.° Ia A -I
I I
al 2
q - q 0
1————
B IB
(a) (b)
Fig. 2.14 Fig. 2.15
axis of a ring was obtained in Example 1 (see p. 14). In our case, this
expression must be doubled. As a result, we obtain
ql
23 (R2+12)312 •
-c
(2) The potential at the centre of the ring is equal to the algebraic
sum of the potentials at this point created by the charges q and q: —
1 (q q
4neo 1R ••V VR2+412 1 •
2.7. Three unlike point charges are arranged as shown in

Fig. 2.15a, where AOB is the right angle formed by two conducting
half-planes. The magnitude of each of the charges is I q I and the dis-
tances between them are shown in the figure. Find (1) the total charge
induced on the conducting half-planes and (2) the force acting on the
charge — q.
Solution. The half-planes forming the angle AOB go to infinity,
and hence their potential p = 0. It can be easily seen that a system
having equipotential surfaces with cp = 0 coinciding with the conduc-
ting half-planes has the form shown in Fig. 2.15b. Hence the action
of the charges induced on the conducting half-planes is equivalent to
the action of the fictitious charge q placed in the lower left corner of
—
the dashed square.

Thus we have already answered the first question: q. —
By reducing the system to four point charges, we can easily find

Problems 65
the required force (see Fig. 2.15b)

2 1 q2
F= F2 — F1= 4neo 2a 2 9
and answer the second question.

• 2.8. Capacitance of parallel wires. Two long straight wires
with the same cross section are arranged in air parallel to one another.
Fig. 2.16 Fig. 2.17
The distance between the wires is times larger than the radius of the
wires' cross section. Find the capacitance of the wires per unit length
provided that >> 1.
Solution. Let us mentally charge the two wires by charges of the
same magnitude and opposite signs so that the charge per unit length
is equal to X. Then, by definition, the required capacitance is
Cu = X/U, (1)
and it remains for us to find the potential difference between the wires.
It follows from Fig. 2.16 showing the dependences of the potentials
cp +and cp_ on the distance between the plates that the sought potential
difference is
IATA-1+ 164-1= 2 IAT+1. (2)
The intensity of the electric field created by one of the wires at a dis-
tance x from its axis can be easily found with the help of the Gauss
theorem: E = X/2as0x. Then
b— a
b a
I AT+ I = J E dx —
2aso
In
—
a (3)
a
where a is the radius of the wires' cross section and b is the separation
between the axes of the wires.
5-0181
It follows from (1), (2) and (3) that

Cu = 80/111
where we took into account that b >> a.

• 2.9. Four identical metallic plates are arranged in air at the
same distance h from each other. The outer plates are connected by a
conductor. The area of each plate is equal to S. Find the capacitance
of this system (between points 1 and 2, Fig. 2.17).
Solution. Let us charge the plates 1 and 2 by charges go and — go.
Under the action of the dissipation field appearing between these
plates (edge effect), a charge will move in the connecting wire, after
which the plate A will be charged negatively while the plate B will
acquire a positive charge. An electric field appears in the gaps be-
tween the plates, accompanied by the corresponding distribution of
potential cp (Fig. 2.18). It should be noted that as follows from the
symmetry of the system, the potentials at the Lmiddle of the system as
well as on its outer plates are equal to zero.
By definition, the capacitance of the system in this case is
C = golU (1)
where U is the required potential difference between the points 1
and 2. Figure 2.18 shows that the potential difference U between the
inner plates is twice as large as the potential difference between the
outside pair of plates (both on the right and on the left). This also
refers to the field intensity:
E = 2E'. (2)
And since E cr, we can state that according to (2) the charge go on
plate 1 is divided into two parts: q0 /3 on the left side of the plate
and 40 /3 on its right side. Hence
U = Eh =- alas, = 2q,121380S,
and the capacitance of the system (between points 1 and 2) is
3s S
C— °
2h '
• 2.10. Distribution of an induced charge. A point charge q is
placed between two large parallel conducting plates 1 and 2 sepa-
rated by a distance 1. Find the total charges q1and q2induced on each
plate, if the plates are connected by a wire and the charge q is located
at a distance llfrom the left plate 1 (Fig. 2.19a).
Solution. Let us use the superposition principle. We mentally
place somewhere on a plane P the same charge q. Clearly, this will
double the surface charge on each plate. If we now distribute uniformly
on the surface P a certain charge with surface density o, the electric
field can be easily calculated (Fig. 2.19b).
The plates are connected by the wire, and hence the potential
difference between them is equal to zero. Consequently,

Eix/i E2x(I — 11) = 0,
where E1 and E2, are the projections of vector E onto the X-axis to
the left and to the right of the plane P (Fig. 2.19b).
On the other hand, it is clear that
if = a2),
where, in accordance with (2.2), 61 = 8oE1n1 = 80 E1, and a2 =
= soE2n2 = ---E,E2x (the minus sign indicates that the normal n2
is directed oppositely to the unit vector of the X-axis).
Eliminating Eix and E2, from these equations, we obtain
= (1— /i)//, 62 = — a1111.
The formulas for charges q1 and q2 in terms of q have a similar form.
It would be difficult, however, to solve this problem with the help
P
:11
X
(a)
Fig. 2.18 Fig. 2.19
of the image method, since it would require an infinite series of ficti-

tious charges arranged on both sides of the charge q, and to find the
field of such a system is a complicated problem.
3. Electric Field in Dielectrics

3.1. Polarization of Dielectrics
Dielectrics. Dielectrics (or insulators) are substances that
practically do not conduct electric current. This means that
in contrast, for example, to conductors dielectrics do not
contain charges that can move over considerable distances
and create electric current.
5*
68 3. Electric Field in Dielectrics
When even a neutral dielectric is introduced into an exter-

nal electric field, appreciable changes are observed in the
field and in the dielectric itself. This is because the dielec-
tric is acted upon by a force, the capacitance of a capacitor
increases when it is filled by a dielectric, and so on.
In order to understand the nature of these phenomena, we
must take into consideration that dielectrics consist either
of neutral molecules or of charged ions located at the sites
of a crystal lattice (ionic crystals, for example, of the NaC1
type). The molecules can be either polar or nonpolar. In a
polar molecule, the centre of "mass" of the negative charge
is displaced relative to the centre of "mass" of the positive
charge. As a result, the molecule acquires an intrinsic dipole
moment p. Nonpolar molecules do not have intrinsic dipole
moments, since the "centres of mass" of the positive and ne-
gative charges in them coincide.
Polarization. Under the action of an external electric
field, dielectric is polarized. This phenomenon consists in
the following. If a dielectric is made up by nonpolar molecu-
les, the positive charge in each molecule is shifted along the
field and the negative, in the opposite direction. If a dielec-
tric consists of polar molecules, then in the absence of the
field their dipole moments are oriented at random (due to
thermal motion). Under the action of an external field, the
dipole moments acquire predominant orientation in the
direction of the external field. Finally, in dielectric crystals
of the NaC1 type, an external field displaces all the positive
ions along the field and the negative ions, against the field. *
Thus, the mechanism of polarization depends on the struc-
ture of a dielectric. For further discussion it is only impor-
tant that regardless of the polarization mechanism, all the
positive charges during this process are displaced along the
field, while the negative charges, against the field. It should
be noted that under normal conditions the displacements of
charges are very small even in comparison with the dimen-
sions of the molecules. This is due to the fact that the inten-
sity of the external field acting on the dielectric is consider-
* There exist ionic crystals polarized even in the absence of an

external field. This property is inherent in dielectrics which are
called electrets (they resemble permanent magnets).
ably lower than the intensities of internal electric fields in

the molecules.
Bulk and Surface Bound Charges. As a result of polariza-
tion, uncompensated charges appear on the dielectric surface
as well as in its bulk. To understand better the mechanism
E.0
(a) (b) (c)

Fig. 3.1
of emergence of these charges (and especially bulk charges),

let us consider the following model. Suppose that we have a
plate made of a neutral inhomogeneous dielectric (Fig. 3.1a)
whose density increases with the coordinate x according
to a certain law. We denote by and V the magnitudes of
the volume densities of the positive and negative charges in
the material (these charges are associated with nuclei and
electrons).
In the absence of an external field, = p' at each point
of the dielectric, since the dielectric is electrically neutral.
However, as well as p' increase with x due to inhomoge-
neity of the dielectric (Fig. 3.1b). This figure shows that in
the absence of external field, these two distributions ex-
actly coincide (the distribution of p+' (x) is shown by the
solid line, while that of p' (x), by the dashed line).
Switching on of the external field leads to a displacement
of the positive charges along the field and of the negative
charges against the field, and the two distributions will be
shifted relative to one another (Fig. 3.1c). As a result, un-
compensated charges will appear on the dielectric surface as
well as in the bulk (in Fig. 3.1 an uncompensated negative
charge appears in the bulk). It should be noted that the re-
version of the field direction changes the sign of all these

charges. It can be easily seen that in the case of a plate made
of a homogeneous dielectric, the distributions p,' (x) and
p' (x) would be H-shaped, and only uncompensated surface
charges would appear upon their relative displacement in
the field E.
Uncompensated charges appearing as a result of polariza-
tion of a dielectric are called polarization, or bound, charges.
The latter term emphasizes that the displacements of these
charges are limited. They can move only within electrically
neutral molecules. We shall denote bound charges by a prime
cr ,).
Thus, in the general case the polarization of a dielectric
leads to the appearance of surface and bulk bound charges
in it.
We shall call the charges that do not constitute dielectric
molecules the extraneous charges.* These charges may be
located both inside and outside the dielectric.
The Field in a Dielectric. The field E in a dielectric is the
term applied to the superposition of the field E0 of extra-
neous charges and the field E' of bound charges:
E E', (3.1)
where E, and E' are macroscopic fields, i.e. the microscopic
fields of extraneous and bound charges, averaged over a
physically infinitesimal volume. Clearly, the field E in the
dielectric defined in this way is also a macroscopic field.
3.2. Polarization
Definition. It is natural to describe polarization of a dielec-
tric with the help of the dipole moment of a unit volume.
If an external field or a dielectric (or both) are nonuniform,
polarization turns out to be different at different points of
the dielectric. In order to characterize the polarization at a
given point, we must mentally isolate an infinitesimal volume
A V containing this point and then find the vector sum of the
* Extraneous charges are frequently called free charges, but this

term is not convenient in some cases since extraneous charges may be
not free.
3.2. Polarization 71
dipole moments of the molecules in this volume and write

the ratio
NI
P= AV (3.2)
Vector P defined in this way is called the polarization of a

dielectric. This vector is numerically equal to the dipole mo-
ment of a unit volume of the substance.
There are two more useful representations of vector P.
Let a volume AV contain AN dipoles. We multiply and di-
vide the right-hand side of (3.2) by AN. Then we can write
P = n (p), (3.3)
where n = AN/AV is the concentration of molecules (their
number in a unit volume) and (p ) = (Epi)/AN is the mean
dipole moment of a molecule.
Another expression for P corresponds to the model of
a dielectric as a mixture of positive and negative "fluids".
Let us isolate a very small volume AV inside the dielectric.
Upon polarization, the positive charge pip AV contained
in this volume will be displaced relative to the negative
charge by a distance 1, and these charges will acquire the
dipole moment Ap = p+ A V.1. Dividing both sides of this
formula by AV, we obtain the expression for the dipole mo-
ment of a unit volume, i.e. vector P:
P = 141. (3.4)
The unit of polarization P is the coulomb per square meter
(C/m2).
Relation Between P and E. Experiments show that for
a large number of dielectrics and a broad class of phenomena,
polarization P linearly depends on the field E in a dielectric.
For an isotropic dielectric and for not very large E, there
exists a relation
P (3.5)
where x is a dimensionless quantity called the dielectric
susceptibility of a substance. This quantity is independent
of E and characterizes the properties of the dielectric itself.
x is always greater than zero.
Henceforth, if the opposite is not stipulated, we shall
consider only isotropic dielectrics for which relation (3.5)

is valid.
However, there exist dielectrics for which (3.5) is not
applicable. These are some ionic crystals (see footnote on
page 68) and ferroelectrics. The relation between P and E
for ferroelectrics is nonlinear and depends on the history
of the dielectric, i.e. on the previous values of E (this phe-
nomenon is called hysteresis).
3.3. Properties of the Field of P

The Gauss Theorem for the Field of P. We shall show
that the field of P has the following remarkable and impor-
tant property. It turns out that the flux of P through an
P
(a) (b)
Fig. 3.2
arbitrary closed surface S is equal to the excess bound charge

(with the reverse sign) of the dielectric in the volume en-
closed by the surface S, i.e.
P dS = — girt (3.6)
This equation expresses the Gauss theorem for vector P.

Proof of the theorem. Let an arbitrary closed surface S
envelope a part of a dielectric (Fig. 3.2a, the dielectric is
hatched). When an external electric field is switched on,
the dielectric is polarized—its positive charges are displaced
relative to the negative charges. Let us find the charge which
passes through an element dS of the closed surface S in the
outward direction (Fig. 3.2b).
Let 1+and 1_ be vectors characterizing the displacement
of the positive and negative bound charges as a result of

polarization. Then it is clear that the positive charge
13,'1+dS cos a inclosed in the "inner" part of the oblique
cylinder will pass through the area element dS from the
surface S outwards (Fig. 3.2b). Besides, the negative charge
to'l_dS cos a enclosed in the "outer" part of the oblique cyl-
inder will enter the surface S through the area element dS.
But be know that the transport of a negative charge in a
certain direction is equivalent to the transport of the posi-
tive charge in the opposite direction. Taking this into account,
we can write the expression for the total bound charge pass-
ing through the area element dS of the surface S in the out-
ward direction:
dq' = dS co5 a + I p' IL dS cos a.
Since I p' = p: we have
dq' (1+ l_) dS cos a = p'+ 1 dS cos a,
(3.7)
where 1 = 1+ L is the relative displacement of positive
and negative bound charges in the dielectric during polari-
zation.
Next, according to (3.4), p+/ = P and dq' = P dS cos a,
or
dq' = P, dS = P•dS. (3.8)
Integrating this expression over the entire closed surface
S, we find the total charge that left the volume enclosed
by the surface S upon polarization. This charge is equal to
P• dS. As a result, a certain excess bound charge q' will
be left inside the surface S. Clearly, the charge leaving the
volume must be equal to the excess bound charge remaining
within the surface S, taken with the opposite sign. Thus, we
arrive at (3.6).
Differential form of Eq. (3.6). Equation (3.6), viz. the Gauss
theorem for the field of vector P, can be written in the differential form
as follows:
V •P= (3.9)
i.e. the divergence of the field of vector P is equal to the volume density
of the excess bound charge at the same point, but taken with the oppo-
site sign. This equation can be obtained from (3.6) in the same manner
as the similar expression for vector E was obtained (see p. 24). For
this purpose, it is sufficient to replace E by P and p by p'.
When Is p' Equal to Zero in a Dielectric? We shall show
that the volume density of excess bound charges in a dielec-
tric is equal to zero if two conditions are simultaneously
satisfied: (1) the dielectric is homogeneous and (2) there
are no extraneous charges within it (p = 0).
Indeed, it follows from the main property (3.6) of the
field of vector P that in the case of a homogeneous dielectric,
we can substitute xsoE for P in accordance with (3.5), take
x out of the integral, and write
x e0E • dS — q' .
The remaining integral is just the algebraic sum of all the

charges—extraneous and bound—inside the closed surface
S under consideration, i.e. it is equal to q q'. Hence,
x (q q') =- — q', from which we obtain
qf
(3.10)
4+H q.
This relation between the excess bound charge q' and

the extraneous charge q is valid for any volume inside the
dielectric, in particular, for a physically infinitesimal vol-
ume, when q' -->- dq' = p' dV and q dq = pdV . Then, after
cancelling out dV , Eq. (3.10) becomes
(3.11)
Hence it follows that in a homogeneous dielectric p' = 0

when p = 0.
Thus, if we place a homogeneous isotropic dielectric of
any shape into an arbitrary electric field, we can be sure
that its polarization will give rise only to the surface bound
charge, while the bulk excess bound charge will be zero at
all points of such a dielectric.
Boundary Conditions for Vector P. Let us consider the
behaviour of vector P at the interface between two homo-
geneous isotropic dielectrics. We have just shown that in
such a dielectric there is no excess bound bulk charge and

only a surface bound charge appears as a result of polari-
zation.
Let us find the relation between polarization P and the
surface density a' of the bound charge at the interface be-
tween the dielectrics. For this purpose, let us use property
(3.6) of the field of vector P. We
choose the closed surface in the
form of a flat cylinder whose end-
faces are on different sides of the
interface (Fig. 3.3). We shall as-
sume that the height of the cyl-
inder is negligibly small and the fp,
area AS of each endface is so small Fig. 3.3
that vector P is the same at all
points of each endface (this also refers to the surface density
a' of the bound charge). Let n be the common normal to the
interface at a given point. We shall always draw vector n
from dielectric 1 to dielectric 2.
Disregarding the flux of P through the lateral surface
of the cylinder, we can write, in accordance with (3.6):
Pen AS + Pi.n,AS = — AS,
where P„ and P„• are the projections of vector P in dielec-
tric 2 onto the normal n and in dielectric 1 onto the normal
n' (Fig. 3.3). Considering that the projection of vector P onto
the normal n' is equal to the projection of this vector onto
the opposite (common) normal n, taken with the opposite
sign, i.e. P„, = — P„, we can write the previous equation
in the following form (after cancelling AS):
Pen — P ln a'. (3.12)
This means that at the interface between dielectrics the
normal component of vector P has a discontinuity, whose
magnitude depends on a'. In particular, if medium 2 is a
vacuum, then P en = 0, and condition (3.12) acquires a
simpler form:
a' = Pn, (3.13)
where Pn is the projection of vector P onto the outward nor-
mal to the surface of a given dielectric. The sign of the pro-
jection 137, determines the sign of the surface bound charge

a' at a given point. Formula (3.13) can be written in a differ-
ent form. In accordance with (3.5), we can write
a' = xeoEn, (3.14)
where Enis the projection of vector E (inside the dielectric
and in the vicinity of its surface) onto the outward normal.
Here, too, the sign of En determines the sign of a'.
A Remark about the Field of Vector P. Relations (3.6)
and (3.13) may lead to the erroneous conclusion that the
field of vector P depends only on the bound charge. Actually,
this is not true. The field of vector P, as well as the field of E,
depends on all the charges, both bound and extraneous.
This can be proved if only by the fact that vectors P and
E are connected through the relation P = xe0E. The bound
charge determines the flux of vector P through a closed sur-
face S rather than the field of P. Moreover, this flux is de-
termined not by the whole bound charge but by its part
enclosed by the surface S.
3.4. Vector D
The Gauss Theorem for Field D. Since the sources of an
electric field E are all the electric charges—extraneous and
bound, we can write the Gauss theorem for the field E in
the following form:
80E dS = (q+ A nt, (3.15)
where q and q' are the extraneous and bound charges en-
closed by the surface S. The appearance of the bound charge
q' complicates the analysis, and formula (3.15) turns out
to be of little use for finding the field E in a dielectric even
in the case of a "sufficiently good" symmetry. Indeed, this
formula expresses the properties of unknown field E in terms
of the bound charge q' which in turn is determined by un-
known field E.
This difficulty, however, can be overcome by expressing
the charge q' in terms of the flux of P by formula (3.6). Then
3.4. Vector 15 77
expression (3.15) can be transformed as follows:

(80E P) dS = gins. (3.16)
The quantity in the parentheses in the integrand is denoted

by D. Thus, we have defined an auxiliary vector D
D = soE P, (3.17)
whose flux through an arbitrary closed surface is equal to the

algebraic sum of extraneous charges enclosed by this sur-
face:
D dS = a-.int • (3.18)
This statement is called the Gauss theorem for field D.

It should be noted that vector D is the sum of two complete-
ly different quantities: soE and P. For this reason, it is
indeed an auxiliary vector which does not have any deep
physical meaning. However, the property of the field of
vector D, expressed by equation (3.18), justifies the intro-
duction of this vector: in many cases it considerably simpli-
fies the analysis of the field in dielectrics.*
Relations (3.17) and (3.18) are valid for any dielectric,
both isotropic and anisotropic.
Expression (3.17) shows that the dimensions of vector
D are the same as those of vector P. The quantity D is
measured in coulombs per square metre (C/m2).
Differential form of Eq. (3.18) is
V•D =P, (3.19)
i.e. the divergence of the field D is equal to the volume density of an

extraneous charge at the same point.
This equation can be obtained from (3.18) in the same way as it
was done for the field E (see p. 24). It suffices to replace E by D and
take into account only extraneous charges.
* The quantity D is often called dielectric displacement, or electro-

static induction. We shall not be using this term, however, in order to
emphasize the auxiliary nature of vector D.
3. Electric Field in Dielectrics
At the points where the divergence of vector D is positive we have

the sources of the field D (p > 0), while at the points where the diver-
gence is negative, the sinks of the field D (p < 0).
Relation Between Vectors D and E. In the case of isotrop-
ic dielectrics, polarization P = xe0E. Substituting this
expression into (3.17), we obtain D = Bo (1 x) E, or
D = 808E, (3.20)
where a is the dielectric constant of a substance:

8 = 1 ± x. (3.21)
The dielectric constant a (as well as x) is the basic electric
characteristic of a dielectric. For materials a > 1, while
for vacuum a = 1. The value of a depends on the nature of
the dielectric and varies between the values slightly differ-
ing from unity (for gases) and several thousands (for some
ceramics). The value of a for water is rather high (a = 81).
Formula (3.20) shows that in isotropic dielectrics vector
D is collinear to vector E. For anisotropic dielectrics, these
vectors are generally noncollinear.
The field D can be graphically represented by the lines
of vector D, whose direction and density are determined in
the same way as for vector E. The lines of E may emerge and
terminate on extraneous as well as bound charges. We say
that any charges may be the sources and sinks of vector E.
The sources and sinks of field D, however, are only extraneous
charges, since only on these charges the lines of D emerge and
terminate. The lines of D pass without discontinuities
through the regions of the field containing bound charges.
A Remark about the Field of Vector D. The field of vector
D generally depends on extraneous as well as bound charges
(just as the field of vector E). This follows if only from the
relation D = 808E. However, in certain cases the field of
vector D is determined only by extraneous charges. It is
just the cases for which vector D is especially useful. At
the same time, this may lead to the erroneous conclusion
that vector D always depends only on extraneous charges
and to an incorrect interpretation of the laws (3.18) and
(3.19). These laws express only a certain property of field
D but do not determine this field proper.
.9.4. Vector D 70
Let us illustrate what was said above by several examples.

Example 1. An extraneous point charge q is located at the centre
of a sphere of radius a, made of an isotropic dielectric with a dielectric
1 +q •
rtiA
•
Fig. 3.4 Fig. 3.5
constant e. Find the projection Er of field intensity E as a function of

the distance r from the centre of this sphere.
The symmetry of the system allows us to use the Gauss theorem
for vector D for solving the problem (we cannot use here the similar
theorem for vector E, since the bound charge is unknown to us). For a
sphere of radius r with the centre at the point of location of the charge
q, we can write the following relation: 4nr2Dr = q. Hence we can find
Dr and then, using formula (3.20), the required quantity Er:
1 g2 1 q
Er (r < a) = 4as 0 er 9 Er (r > a) = 4ae, 2
Figure 3.4 shows the curves D (r) and E (r).
Example 2. Suppose that a system consists of a point charge q > 0
and an arbitrary sample of a homogeneous isotropic dielectric
(Fig. 3.5), where S is a certain closed surface. Find out what will hap-
pen to the fields of vectors E and D (and to their fluxes through the
surface S) if the dielectric is removed.
The field E at any point of space is determined by the charge g
and by bound charges of the polarized dielectric. Since in our case
D = co 8E, this refers to the field D as well: it is also determined by
the extraneous charge q and by the bound charges of the dielectric.
The removal of the dielectric will change the field E, and hence the
field D. The flux of vector E through the surface S will also change,
since negative bound charges will vanish from inside this surface.
However, the flux of vector D through the surface S will remain the
same in spite of the change in the field D.
Example 3. Let us consider a system containing no extraneous
charges and having only bound charges. Such a system can, for
example, be a sphere made of an electret (see p. 68). Figure 3.6a shows
the field E of this system. What can we say about the field D?
80 g. Pilectric Field in Dielectrics
First of all, the absence of extraneous charges means that the field
D has no sources: the lines of D do not emerge or terminate anywhere.
However, the field D exists and is shown in Fig. 3.6b. The lines of E
Lines of E Lines of D
(a) (6)
Fig. 3.6
and D coincide outside the sphere, but inside the sphere they have
opposite directions, since here the relation D = 80 8E is no longer va-
lid and D = 80E + P.
3.5. Boundary Conditions

Let us first consider the behaviour of vectors E and D
at the interface between two homogeneous isotropic dielec-
trics. Suppose that, for greater generality, an extraneous
surface charge exists at the interface between these dielec-
trics. The required conditions can be easily obtained with
the help of two theorems: the theorem on circulation of
vector E and the Gauss theorem for vector D:
dl — 0 and D dS = qint.
Boundary Condition for Vector E. Let the field near the

interface be E1in dielectric 1 and E2 in dielectric 2. We
choose a small elongated rectangular contour and orient it as
shown in Fig. 3.7. The sides of the contour parallel to the
interface must have such a length that the field E over this
length in each dielectric can be assumed constant. The
"height" of the contour must be negligibly small. Then, in
accordance with the theorem on circulation of vector E,
we have
E 2,1 Erel = 0,
where the projections of vector Ware taken on the direction

of the circumvention of the contour, shown by arrows in
the figure. If in the lower region of the contour we take the
n
r ,............„)....
...
2 i-7-1-7i
''''' T' T
1
Fig. 3.7 Fig. 3.8
projection of vector E not onto the unit vector T'but onto
the common unit vector T, then E„. =- — Er , and it fol-
lows from the above equation that
= E2r, (3.22)
i.e. the tangential component of vector E turns out to be

the same on both sides of the interface (it does not have a
discontinuity).
Boundary Condition for Vector D. Let us take a cylinder
of a very small height and arrange it at the interface be-
tween two dielectrics (Fig. 3.8). The cross section of the
cylinder must be such that vector D is the same within each
of its endfaces. Then, in accordance with the Gauss theorem
for vector D, we have
D211 • AS + Dm. .AS = szTAS,
where o- is the surface density of the extraneous charge at

the interface. Taking both projections of vector D onto the
common normal n (which is directed from dielectric 1 to
dielectric 2), we obtain D1n • — D17„ and the previous
equation can be reduced to the form
Den Din = a. (3.23)
It follows from this relation that the normal component

of vector D generally has a discontinuity when passing
6-0181
through the interface. If, however, there are no extraneous

charges at the interface (a = 0), we obtain
Di„ = D2„ (3.24)
In this case the normal components of vector D do not have

a discontinuity and turn out to be the same on different
sides of the interface.
Thus, in the absence of extraneous charges at the inter-
face between two homogeneous isotropic dielectrics, the
components Er and D, vary continuously during a transi-
tion through this interface, while the components En, and
D, have discontinuities.
Refraction of E and D Lines. The boundary conditions
which we obtained for the components of vectors E and D
at the interface between two dielectrics indicate (as will
be shown later) that these vectors have a break at this in-
terface, i.e. are refracted (Fig. 3.9). Let us find the relation
between the angles al and a2.
In the absence of extraneous charges at the interface,
we have, in accordance with (3.22) and (3.24), E„ =
and s2E2n = eiEin• Figure 3.9 shows that
tan a2E2T/E2n
tan al. EiT /Eir, •
Taking into account the above conditions, we obtain the
law of refraction of lines E, and hence of lines D:
tan a2 82
tan a l '
(3.25)
This means that lines of D and E will form a larger angle

with the normal to the interface in the dielectric with a
larger value of e (in Fig. 3.9, 82 > ei).
Example. Let us represent graphically the fields E and D at the
interface between two homogeneous dielectrics 1 and 2, assuming that
62 > siand there is no extraneous charge on this surface.
Since 82 > 81, in accordance with (3.25) a2 > al (Fig. 3.10).
Considering that the tangential component of vector E remains un-
changed and using Fig. 3.9, we can easily show that E2 < F1 in
magnitude, i.e. the lines of E in dielectric 1 must be denser than in
dielectric 2, as is shown in Fig. 3.10. The fact that the normal corn-
ponents of vectors D are equal leads to the conclusion that D 2 > D1

in magnitude, i.e. the lines of D must be denser in dielectric 2.
Fig 3.9 Fig. 3.10
We see that in the case under consideration, the lines of E are

refracted and undergo discontinuities (due to the presence of bound
charges), while the lines of D are only refracted, without discontinu-
ities (since there are no extraneous charges at the interface).
Boundary Condition on the Conductor-Dielectric Inter-
face. If medium 1 is a conductor and medium 2 is a dielectric
(see Fig. 3.8), it follows from formula (3.23) that
Dn = a, (3.26)
where n is the conductor's outward normal (we omitted the

subscript 2 since it is inessential in the given case). Let us
verify formula (3.26). In equilibrium, the electric field
inside a conductor is E = 0, and hence the polarization
P = 0. This means, according to (3.17), that vector D = 0
inside the conductor, i.e. in the notations of formula (3.23)
D1= 0 and D1 = 0. Hence D,„ = a.
Bound Charge at the Conductor Surface. If a homogeneous
dielectric adjoins a charged region of the surface of a con-
ductor, bound charges of a certain density a' appear at the
conductor-dielectric interface (recall that the volume den-
sity of bound charges p' = 0 for a homogeneous dielectric).
Let us now apply the Gauss theorem to vector E in the same
way as it was done while deriving formula (2.2). Consider-
ing that there are both bound and extraneous charges (a
6*
and a') at the conductor-dielectric interface we arrive at

the following expression: En =(a + a')/c0. On the other
hand, according to (3.26) E„ = D„1880 = a/ca0. Combin-
ing these two equations, we obtain a/c = a + a', whence
(y r 8—
8
a . (3.27)
It can be seen that the surface density a' of the bound

charge in the dielectric is unambiguously connected with
the surface density a of the extraneous charge on the con-
ductor, the signs of these charges being opposite.
3.6. Field in a Homogeneous Dielectric

It was noted in Sec. 2.1 that the determination of the
resultant field E in a substance is associated with consider-
able difficulties, since the distribution of induced charges in
the substance is not known beforehand. It is only clear
that the distribution of these charges depends on the nature
and shape of the substance as well as on the configuration
of the external field E0.
Consequently, in the general case, while solving the prob-
lem about the resultant field E in a dielectric, we encounter
serious difficulties: determination of the macroscopic field
E' of bound charges in each specific case is generally a com-
plicated independent problem, since unfortunately there
is no universal formula for finding E'.
An exception is the case when the entire space where there
is a field E0is filled by a homogeneous isotropic dielectric.
Let us consider this case in greater detail. Suppose that we
have a charged conductor (or several conductors) in a va-
cuum. Normally, extraneous charges are located on conduc-
tors. As we already know, in equilibrium the field E inside
the conductor is zero, which corresponds to a certain
unique distribution of the surface charge a. Let the field
created in the space surrounding the conductor be E0.
Let us now fill the entire space of the field with a homo-
geneous dielectric. As a result of polarization, only surf ace
bound charges a will appear in this dielectric at the inter-
face with the conductor. According to (3.27) the charges
3.6. Field in a Homogeneous Dielectric 85
a' are unambiguously connected with the extraneous charges

a on the surface of the conductor.
As before, there will be no field inside the conductor (E =
= 0). This means that the distribution of surface charges
(extraneous charges a and bound charges a') at the conductor-
dielectric interface will be similar to the previous distri-
bution of extraneous charges (a), and the configuration of
the resultant field E in the dielectric will remain the same
as in the absence of the dielectric. Only the magnitude of
the field at each point will be different.
In accordance with the Gauss theorem, a + a' = 80E„,
where En = Dn I880 =a/e80, and hence
a + a' = a/c. (3.28)
But if the charges creating the field have decreased by a factor
of a everywhere at the interface, the field E itself has become
less than the field E0 by the same factor:
E = Eo/8. (3.29)
Multiplying both sides of this equation by ccc, we obtain
D = Do, (3.30)
i.e. the field of vector D does not change in this case.
It turns out that formulas (3.29) and (3.30) are also valid
in a more general case when a homogeneous dielectric fills
the volume enclosed between the equipotential surfaces
of the field E, of extraneous charges (or of an external field).
In this case also E = E0/8 and D = Doinside the dielectric.
In the cases indicated above, the intensity E of the field
of bound charges is connected by a simple relation with
the polarization P of the dielectric, namely,
E' —P/a0. (3.31)
This relation can be easily obtained from the formula E
= E0 E' if we take into account that E0 =8E and
P = xa0E.
As was mentioned above, in other cases the situation
is much more complicated, and formulas (3.29)-(3.31) are
inapplicable.
Corollaries. Thus, if a homogeneous dielectric fills the
entire space occupied by a field, the intensity E of the field
will be lower than the intensity E, of the field of the same

extraneous charges, but in the absence of dielectric, by a
factor of a. Hence it follows that potential cp at all points
will also decrease by a factor of a:
(1) = co/a, (3.32)
where cp0is the field potential in the absence of the dielectric.
The same applies to the potential difference:
U = UoIg, (3.33)
where U0 is the potential difference in a vacuum, in the
absence of dielectric.
In the simplest case, when a homogeneous dielectric
fills the entire space between the plates of a capacitor, the
potential difference U between its plates will be by a factor
of a less than that in the absence of dielectric (naturally,
at the same magnitude of the charge q on the plates). And
since it is so, the capacitance C = girl of the capacitor
filled by dielectric will increase a times
C' = eC, (3.34)
where C is the capacitance of the capacitor in the absence
of dielectric. It should be noted that this formula is valid
when the entire space between the plates is filled and edge
effects are ignored.
Problems
• 3.1. Polarization of a dielectric and the bound charge. An extra-
neous point charge q is at the centre of a spherical layer of a heteroge-
neous isotropic dielectric whose dielectric constant varies only in the
radial direction as s = alr, where a is a constant and r is the distance
from the centre of the system. Find the volume density p' of a bound
charge as a function of r within the layer.
Solution. We shall use Eq. (3.6), taking a sphere of radius r as
the closed surface, the centre of the sphere coinciding with the centre
of the system. Then
42-0. Pr = q' (r),
—
where q' (r) is the bound charge inside the sphere. Let us take the
differential of this expression:
431 d(r2 •Pr) = — dq'. (1)
Problems 87
Here dq' is the bound charge in a thin layer between the spheres with
radii r and r + dr. Considering that dq' = p'4itr2 dr, we transform (1)
as follows:
r2 dPr 2rPr dr = —p'r2 dr,
whence
p, ddPrr+ r2 pr
(2)
In the case under consideration we have

s-1 a - 1 q
Pr =x80 Er — Dr =
a 42tr2
and after certain transformation expression (2) will have the form
, 1 q
P 4ncx, r2
which is just the required result.
3.2. The Gauss theorem for vector D. An infinitely large plate
made of a homogeneous dielectric with the dielectric constant a is
uniformly charged by an extraneous
charge with volume density p > 0.
The thickness of the plate is 2a. (1)
Find the magnitude of vector E and
the potential cp as functions of the di-
stance / from the middle of the plate
(assume that the potential is zero in
the middle of the plate), choosing the
X-axis perpendicular to the plate.
Plot schematic curves for the pro-
jection Ex (x) of vector E and the po-
tential y (x). (2) Find the surface and
volume densities of the bound charge.
Solution. (1) From symmetry con-
siderations it is clear that E = 0 in Fig. 3.11
the middle of the plate, while at all
other points vectors E are perpend-
icular to the surface of the plate. In order to determine E, we shall
use the Gauss theorem for vector D (since we know the distribu-
tion of only extraneous charges). We take for the closed surface a
right cylinder of height 1, one of whose endfaces coincides with the
midplane (x = 0). Let the cross-sectional area of this cylinder be S.
Then
DS = pS1, D = pl, E = p//aao (1 < a),
DS = pSa, D = pa, E = pa/so (/ a).
The graphs of the funtions E (x) and cp (x) are shown in Fig. 3.11.
It is useful to verify that the graph of Ex (x) corresponds to the de-
rivative —acp/Ox.
(2) In accordance with (3.13), the surface density of the bound
charge is
8 1 pa >0.
= Pn=xFoEn= (s— 1) pa/E= 6-
This result is valid for both sides of the plate. Thus, if the extraneous
charge p > 0, the bound charges appearing on both surfaces of the
plate are also positive.
In order to find the volume density of the bound charge, we use
Eq. (3.9) which in our case will have a simpler form:
013 x
p __= a ( 8-I 8-1
ax ax P
It can be seen that the bound charge is uniformly distributed over

the bulk and has the sign opposite to that of the extraneous charge
• 3.3. A homogeneous dielectric has the shape of a spheric al
layer whose inner and outer radii are 'a and b. Plot schematically the
curves of intensity E and potential IT of the electric field as functions
of the distance r from the centre of the system, if the dielectric is
charged by a positive extraneous charge distributed uniformly (1)
over the inner surface of the layer, (2) over the layer's bulk.
Solution. (1) We use the Gauss theorem for vector D, taking
for the closed surface a sphere of radius r:
43tr2D = q,
where q is the extraneous charge within this sphere. Hence it follows
that
D (r < a) = 0, D (r > a) = ql4nr2.
The required intensity is
E (r < a) = 0, E (r > a) = Dies°.
The curve for E (r) is shown in Fig. 3.12a. The curve for q) (r) is also
shown in this figure. The curve q) (r) must have such a shape that the
derivative acplar taken with the opposite sign corresponds to the curve
of the function E (r). Besides, we must take into account the normali-
zation condition: cp 0 as r –›- co.
It should be noted that the curve corresponding to the function
cp (r) is continuous. At the points where the function E (r) has finite
discontinuities, the function q) (r) is only broken.
(2) In this case, in accordance with the Gauss theorem, we have
4
4nr2D=- — x (r3 —a3) p,
3
Problems 89
where p is the volume density of the extraneous charge. Hence

p rs — as
es° 3808 r2
The corresponding curves for E (r) and up (r) are shown in Fig. 3.12b.
• 3.4. Extraneous charge is uniformly distributed with the volume

density p > 0 over a sphere of radius a, made of a homogeneous dielec-
tric with the permittivity e. Find (1) the magnitude of vector E as a
function of the distance r from the centre of the sphere and plot the
curves of the functions E (r) and cp (r); (2) the surface and volume den-
sities of bound charge.
Solution. (1) In order to determine E, we shall use the Gauss theo-
rem for vector D since we know the distribution of only extraneous
charge:
4 p
r < a, 4rtr2D= nr3p, D= r, E=—D—=— r,
ss o 3sso
4
r > a, 4nr2D= --3— na3p, D= pa3 D pas 1
80 38o r2 •
The curves for the functions E (r) and cp (r) are shown in Fig. 3.13.
(2) The surface density of the bound charge is
8-1 pa
=P n— 3•
In order to find the volume density of bound charge, it is sufficient to
repeat the reasoning that led us to formula (3.11), and we get
8-
= 8 1 P. (1)
This result can be obtained in a different way, viz. by using

Eq. (3.9). Since P = xeoE and x does not depend on coordinates (in-.
side the sphere), we obtain

p' = —7 .P = —xsoV •E,
where soV •E = p p'. Hence p' =- —x (p p'), which gives for-
mula (1).
• 3.5. Capacitance of a conductor. Find the capacitance of a
spherical conductor of radius a, surrounded by a layer of a homoge-
;E•
0 a b r
Fig. 3.13 Fig. 3.14
neous dielectric and having the outer radius b and the dielectric con-
stant E. Plot approximate curves for E (r) and cp (r), where r is the dis-
tance from the centre of the sphere, if the sphere is charged positi-
vely.
Solution. By definition, the capacitance C = q/cp. Let us find the
potential cp of the conductor, supplying mentally a charge q to it:
00 b 00
1 q 1
cp= ET dr — dr + d r.
4neo sr2 4n8o :2
a a b
Integrating this expression, we obtain

q 1 , 8-1. 1 4neoea
C=
(P = 4neos ( a 1 b /' 1+ (8-1) alb •
The curves for E (r) and cp (r) are shown in Fig. 3.14.
• 3.6. Capacitance of a capacitor. A spherical capacitor with
the radii of the plates a and b, where a < b, is filled with an isotropic
heterogeneous dielectric whose permittivity depends on the distance r
from the centre of the system as a = oar, where a is a constant. Find
the capacitance of the capacitor.
Solution. In accordance with the definition of the capacitance of a
capacitor (C = qiU), the problem is reduced to determination of the
potential difference U for a given charge q:
b
U= E dr, (1)
a
Problems 91
where we assume that the charge of the inner plate is q > 0. Let us
find E with the help of the Gauss theorem for vector D:
1 q1
4nr2 D= q, E = D=
880 4ae, E rg 418, ar
Substituting the latter expression into (1) and integrating, we find
4na a
U= In , C —°
40180a a In (b/a)
3.7. The Gauss theorem and the principle of superposition.
Suppose that we have a dielectric sphere which retains polarization
Fig. 3.15 Fig. 3.16
after an external electric field is switched off. If the sphere is polar-

ized uniformly, the field intensity inside it is E' = —P/38,, where P
is the polarization. (1) Derive this formula assuming that the sphere is
polarized as a result of a small displacement of all positive charges of
the dielectric with respect to all its negative charges. (2) Using this
formula, find the intensity E, of the field in the spherical cavity inside
an infinite homogeneous dielectric with the permittivity a if the field
intensity in the dielectric away from the cavity is E.
Solution. (1) Let us represent this sphere as a combination of two
spheres of the same radii, bearing uniformly distributed charges with
volume densities p and —p. Suppose that as a result of a small shift,
the centres of the spheres are displaced relative to one another by a
distance 1 (Fig. 3.15). Then at an arbitrary point A inside the sphere
we have
p
E'=WE -F EL= P .+ —r_)=
(
38o 38lo
where we used the fact that field intensity inside a uniformly charged
sphere is E = pr/3a0, which directly follows from the Gauss theorem.
It remains for us to take into account that in accordance with (3.4),
p1 = P.
(2) The creation of a spherical cavity in a dielectric is equivalent
to the removal from a sphere of a ball made of a polarized material.
Consequently, in accordance with the principle of superposition, the

field E inside the dielectric can be represented as the sum E = E' +
+ E0. Hence
E o = E — E' = E + P/380.
Considering that P = (a — 1) soE, we obtain
E0 = (2 + a) E13.
• 3.8. Boundary conditions. In the vicinity of point A (Fig. 3.16)
belonging to a dielectric-vacuum interface, the electric field intensity
in vacuum is equal to E0, and the vector E, forms the angle ao with
the normal to the interface at the given point. The dielectric permittiv-
ity is a. Find the ratio E/E0, where E is the intensity of the field in-
side the dielectric in the vicinity of point A.
Solution. The field intensity inside the dielectric is
E=17- q+q. (1)
Using conditions (3.22) and (3.24) at the interface between dielectrics,
we find
Eti = E,, sin a0, En =- Dn1880 = Eon/8 = E,, cos a0/e,
where E0„ is the normal component of the vector E0 in a vacuum.
Substituting these expressions into (1), we obtain
2
sin a0+ cos2 a,< 1,
Eo 8
i.e. E <
• 3.9. A point charge q is in a vacuum at a distance 1 from the
plane surface of a homogeneous dielectric filling the half-space below
the plane. The dielectric's permittivity is a. Find (1) the surface den-
sity of the bound charge as a function of the distance r from the point
charge q and analyse the obtained result; (2) the total bound charge on
the surface of the dielectric.
Solution. Let us use the continuity of the normal component of
vector D at the dielectric-vacuum interface (Fig. 3.17):
D en = Dlnt E2n = 8E1n
or
1 q cos 15+ a' =a ( 1 a' \
— 4rcs0 — —q cos 15
r2 2a 0 41(80 r2 280 j'
where the term a'/280 is the component of the electric field created
near the region of the plane under consideration, where the surface
charge density is a'. From the last equality it follows that
a-1 ql
= (1)
a+1 2xr3 •
Problems 93
Here we took into account that cos = 11r. As 1 —> 0 the quantity
a' ---).- 0, i.e. if the charge q is just at the interface, there is no surface
charge on the plane.
(2) Let us imagine at the interface a thin ring with the centre at
the point 0 (Fig. 3.17). Suppose that the inner and outer radii of this
ring are r' and r' + dr'. The surface
q bound charge within this ring is
1. dq'=--a' • 2a-cr' dr'. It can be seen
from the figure that r2 = /2 + r'2,
whence r dr = r' dr', and the expres-
sion for dq' combined with (1) gives
/# .7 dq' =
r2
Integrating this equation over r be-
tween 1 and oo, we obtain
Fig. 3.17 -
e + 1 q.
• 3.10. A point charge q is on the plane separating a vacuum
from an infinite homogeneous dielectric with the dielectric permittivi-
ty 8. Find the magnitudes of vectors D and E in the entire space.
Solution. In this case, it follows from the continuity of the normal
component of vector D that E2,,, = eEin. Only the surface charge a'
will contribute to the normal component of vector E in the vicinity
of the point under consideration. Hence the above equality can be
written in the form
a'/2e0 = e (—aV2E0).
We immediately find that a' = 0.
Thus, in the given case there is no bound surface charge (with the
exception of the points in direct contact with the extraneous point
charge q). This means that the electric field in the surrounding space
is the field of the point charge q q', and E depends only on the dis-
tance r from this charge. But the charge q' is unknown, and hence
we must use the Gauss theorem for vector D. Taking for the closed
surface a sphere of radius r with the centre at the point of location of
the charge q, we can write
2nr2D, 2nr2D = q,
where D o and D are the magnitudes of vector D at the distance r
from the charge q in the vacuum and in the dielectric respectively.
Besides, from the continuity of the tangential component of vector
E it follows that
D = eDo.
Combining these two conditions, we find
eq
Do= 2a D=
(1+ e) r 2 231 (1+0r2
94 4. Energy of Electric Field
and the electric field intensity in the entire space is
E = 80° = 2n (1 + a) so r2 •
It can be seen that for e = 1 these formulas are reduced to the al-
ready familiar expressions for D and E of the point charge in a vac-
uum.
The obtained results are represented graphically in Fig. 3.18. It
Field E Field D
Fig. 3.18
should be noted that the field D in this case is not determined by the
extraneous charge alone (otherwise it would have the form of the field
of a point charge).
4. Energy of Electric Field

4.1. Electric Energy of a System of Charges
Energy Approach to Interaction. The energy approach
to interaction between electric charges is, as will be shown,
rather fruitful in respect of its applications. Besides, this
approach makes it possible to consider the electric field
from a different point of view.
First of all, let us find out how we can arrive at the con-
cept of the energy of interaction in a system of charges.
1. Let us first consider a system of two point charges 1
and 2. We shall find the algebraic sum of the elementary
works of the forces F, and F2 of interaction between the
charges. Suppose that in a certain system of reference K
the charges were displaced by dl, and dl, during the time
dt. Then the corresponding work of these forces is

SA1, 2 = •dli + • d12.
Considering that F2 = — F1(according to the Newton
third law), we can write this expression in the form
6A1, 2 =F1 (dI1— d12) .
The quantity in the parentheses is the displacement of
charge 1 relative to charge 2. In other words, it is the dis-
placement of charge 1 in the system of reference K', which
is rigidly fixed to charge 2 and accomplishes with it a trans-
lational motion relative to the initial reference system K.
Indeed, the displacement d11of charge 1 in system K can
be represented as the displacement d12 of system K' plus
the displacement d11 of charge 1 relative to system
K' : d11= d12 Hence, d1 1— d12 = d11 and,
"1, 2 — F1•d1;.
Thus, it turns out that the sum of the elementary works
done by two charges in an arbitrary system of reference K
is always equal to the elementary work done by the force
acting on one charge in another reference system (K') in
which the other charge is at rest. In other words, the work
6A1, 2 does not depend on the choice of the initial system of
reference.
The force F1acting on charge 1 from charge 2 is conserva-
tive (as a central force). Consequently, the work of the given
force in the displacement d1i can be represented as the de-
crease in the potential energy of interaction between the
pair of charges under consideration:
6 A1, 2 = dW127
where W12 is the quantity depending only on the distance
between these charges.
2. Let us now go over to a system of three point charges
(the result obtained for this case can be easily generalized
for a system of an arbitrary number of charges). The work
performed by all forces of interaction during elementary
displacements of all the charges can be represented as the
sum of the works of three pairs of interactions, i.e. SA =
= SA 1,2 ± M1,3 ± 6A 2,3. But as it has just been shown,
for each pair of interactions Mi, h = — dWih, and hence

SA = — d (W12 +W13 + W23) = — dW,
where W is the energy of interaction for the given system of
charges:
W = W12 W13 + W23.
Each term of this sum depends on the distance between
corresponding charges, and hence the energy W of the given
system of charges is a function of its configuration.
Similar arguments are obviously valid for a system of
any number of charges. Consequently, we can state that to
each configuration of an arbitrary system of charges, there
corresponds a certain value of energy W, and the work of
all the forces of interaction upon a change in this configu-
ration is equal to the decrease in the energy W
6A. = — dW. (4.1)
Energy of Interaction. Let us find the expression for the
energy W. We again first consider a system of three point
charges, for which we have found that W = W12 + W13 +
± W23. This sum can be transformed as follows. We
represent each term Wih in a symmetric form: Wih
1
= - (W Wh i), since Wel, = Whi. Then
2 ih
W =—2- (W12 + W21 f W13 + W31 + W23 + W32).
Let us group the terms with similar first indices:

W =4. i(W12-1- W13) -F (W21+ W23) + (W3i±W32)].
Each sum in the parentheses is the energy Wi of interaction
between the ith charge with all the remaining charges.
Hence the latter expression can be written in the form
3
W =-7 + W + W ) -= -T
2 3
wi.
i=i
This expression can be generalized to a system of an arbi-

trary number of charges, since the above line of reasoning
obviously does not depend on the number of charges con-

stituting the system. Thus, the energy of interaction for
a system of point charges is
W= —
2 2 wi. (4.2)
Considering that Wi = qicp i, where qi is the ith charge of

the system and (pi is the potential created at the point of
location of the ith charge by all the remaining charges, we
obtain the final expression for the energy of interaction of
the system of point charges:
‘
-1
W= 2 Qat• (4.3)
Example. Four similar point charges q are located at the vertices

of a tetrahedron with an edge a (Fig. 4.1). Find the energy of inter-
action of charges in this system.
The energy of interaction for each pair of charges of this system is
the same and equal to W1 = q2/43-ccoa. It can be seen from the figure
that the total number of interacting pairs is six, and hence the energy
of interaction of all point charges of the system is
W = 6W1= 6q2Matsoa.
Another approach to the solution of this
problem is based on formula (4.3). The pote-
ntial IT at the point of location of one of the
charges, created by the field of all the other
charges, is IT = 3q/4neoa. Hence
4
1 1 1 6q2
q
W= 4 q(P = 4318 0 a •
i=i Fig. 4.1
Total Energy of Interaction. If the charges are arranged
continuously, then, representing the system of charges as
a combination of elementary charges dq = p dV and going
over in (4.3) from summation to integration, we obtain
= pcp dV , (4.4)
7-0181
where cp is the potential created by all the charges of the

system in the volume element dV. A similar expression can
be written, for example, for a surface distribution of charges.
For this purpose, we must replace in (4.4) p by a and dV by
dS.
Expression (4.4) can be erroneously interpreted (and
this often leads to misunderstanding) as a modification of
expression (4.3), corresponding to the replacement of the
concept of point charges by that of a continuously distrib-
uted charge. Actually, this is not so since the two expres-
sions differ essentially. The origin of this difference is in
different meanings of the potential cp appearing in these ex-
pressions. Let us explain this difference with the help of
the following example.
Suppose a system consists of two small balls having
charges q1and q2. The distance between the balls is consi-
derably larger than their dimensions, hence q1and q2can be
assumed to be point charges. Let us find the energy W of
the given system by using both formulas.
According to (4.3), we have
W (qi(pi =q1~1 q2q3,2,
where (pi(W2) is the potential created by charge q2 (q1) at

the point where charge q1(q2) is located.
On the other hand, according to formula (4.4) we must
split the charge of each ball into infinitely small elements
P dV and multiply each of them by the potential cp created
by not only the charge elements of another ball but by the
charge elements of this ball as well. Clearly, the result will
be completely different:
W= + W2 + W12, (4.5)
where W1is the energy of interaction of the charge elements
of the first ball with each other, W2 the same but for the
second ball, and W12 the energy of interaction between the
charge elements of the first ball and the charge elements of
the second ball. The energies W1 and W2 are called the
intrinsic energies of charges q1and q2, while W12 is the energy
of interaction between charge q, and charge q2.
Thus, we see that the energy W calculated by formula (4.3)
4.2. Charged Conductor and Capacitor Energies 99
corresponds only to the energy W„, while calculation by

formula (4.4) gives the total energy of interaction: in addi-
tion to W12, it gives intrinsic energies W1and W2. Disre-
gard of this circumstance is frequently a 'cause of gross
errors.
We shall return to this question in Sec. 4.4. Now, we
shall use formula (4.4) for obtaining several important re-
sults.
4.2. Energies of a Charged Conductor and

a Charged Capacitor
Energy of an Isolated Conductor. Let a conductor have
a charge q and a potential' cp. Since the value of cp is the
same at all points where charge is located, we can take cp
in formula (4.4) out of the integral. The remaining integral
is just the charge q on the conductor, and we obtain
Cy2 q2
W =-2
2 2 2C '
(4.6)
These three expressions are written assuming that C = q/cp.

Energy of a Capacitor. Let q and (p i_ be the charge and
potential of the positively charged plate of a capacitor.
According to (4.4), the integral can be split into two parts
(for the two plates). Then
W = 2( q±(P+ q-W-)*
Since q_ = — q+, we have
,
W +((pi_ q)_) =- — qu ,
2q 2
where q = q+is the charge of the capacitor and U is the
potential difference across its plates. Considering that
C = q/ U, we obtain the following expression for the energy
of the capacitor:
U CU2 q2
W
2 2
(4.7)
2C •
7*
It should be noted here that these formulas determine the

total energy of interaction, viz. not only the energy of in-
teraction between the charges of one plate and those of the
other plate, but also the energy of interaction of charges
within each plate.
And What if We Have a Dielectric? We shall show that
formulas (4.6) and (4.7) are valid in the presence of a dielectric.
For this purpose, we consider the process of charging a
capacitor as a transport of small portions of charge (dq')
from one plate to the other.
The elementary work performed against the forces of the
field in this case is
SA = U' dq' = (q'IC) dq',
where U' is the potential difference between the plates at
the moment when the next portion of charge dq' is being
transferred.
Integrating this expression over q' between 0 and q, we
obtain
A = q2/2C,
which coincides with the expression for the total energy of
a capacitor. Consequently, the work done against the forces
of the electric field is completely spent for accumulating
the energy W of the charged capacitor. Moreover, the expres-
sion obtained for the work A is also valid in the case when
there is a dielectric between the plates of a capacitor. Thus,
we have proved the validity of (4.7) in the presence of a
dielectric.
Obviously, all this applies to (4.6) as well.
4.3. Energy of Electric Field

On Localization of Energy. Formula (4.4) defines electric
energy W of any system in terms of charge and potential.
It turns out, however, that energy W can also be expressed
through another quantity characterizing the field itself,
viz, through field intensity E. Let us at first demonstrate
this by using the simple example of a parallel-plate capaci-
tor, ignoring field distortions near the edges of the plates
(edge effect). Substituting into the formula W = CU2I2
the expression C = ee0SM, we obtain

W = CU 2 se0SU2 880 U 12 eh
2— 2 —2 k '4- f'"
And since Ulh = E and Sh = V (the volume between the

capacitor plates), we get
1
W — e sE2V
2 * (4.8)
This formula is valid for the uniform field which fills the
volume V.
In the general theory, it is proved that the energy W
can be expressed in terms of E (in the case of an isotropic
dielectric) through the formula
E D
W – —2 dV
-68E dV.• (4.9)
2
The integrand in this equation has the meaning of the energy

contained in the volume dV. This leads us to a very impor-
tant and fruitful physical idea about localization of energy
in the field. This assumption was confirmed in experiments
with fields varying in time. It is the domain where we en-
counter phenomena that can be explained with the help of
the notion of energy localization in the field. These varying
fields may exist independently of electric charges which
have generated them and may propagate in space in the
form of electromagnetic waves. Experiments show that
electromagnetic waves carry energy. This circumstance con-
firms the idea that the field itself is a carrier of energy.
The last two formulas show that the electric energy is
distributed in space with the volume density
2 E•D
808E
w= (4.10)
2 2 •
It should be noted that this formula is valid only in the

case of isotropic dielectrics for which the relation P = xeoE
holds.
For anisotropic dielectrics the situation is more complicat-
ed.
Another substantiation for formula (4.9). It is known that the

energy of an isolated charged conductor is W = qcp/2. Let us show that
this formula is correct, proceeding from the idea of localization of
energy in the field.
Let us consider an arbitrary positively charged conductor. We
mentally isolate a tube of infinitesimal cross section, which is bound-
ed by lines of E (Fig. 4.2), and take in it an elementary volume dV =
=dS dl.This volume contains the energy
dS
ED as dl D E dl.
2
Let us now find the energy local-
ized in the entire isolated tube. For
this purpose, we integrate the last ex-
pression, considering that the product
D dS is the same in all cross sections
of the tube and hence it can be taken
out of the integral:
00
dw _
Fig. 4.2 D dS (E dl= D2 (1),
2 2
A
wh ere A is a point at the beginning of the imaginary tube.
It remains for us to make the last step, i.e. integrate the obtained
expression over all the tubes, and find the energy localized in the entire
field. Considering that potential q is the same at the endf aces of all
the tubes (since they originate on the surface of the conductor), we
write
(I) .D dS,
HT= 2—
where the integration is performed over a closed surface coinciding

with one of the equipotential surfaces. In accordance with the Gauss
theorem, this integral is equal to the charge q of the conductor, and
we finally get
W = qtp/2,
Q .E.D.
Let us consider two examples illustrating the advantages
we get by using the idea of energy localization in the field.
Example 1. A point charge q is at the centre of a spherical layer
made of a homogeneous dielectric with the dielectric constant e. The
inner and outer radii of the layer are equal to a and b respectively.
Find the electric energy contained in this dielectric layer.
We mentally isolate in the dielectric a very thin concentric spher-
ical layer of radius from r to r + dr. The energy localized in this
layer is given by
808E 2
dW = 2 43-cr2dr,
where E = q2/43-E80 6r2. Integrating this expression over r between a

and b, we obtain
q2 1 1
W =_-.
8as0s ( a b
Example 2. Find the work that must be done against the electric
forces in order to remove a dielectric plate from a parallel-plate char-
ged capacitor. It is assumed that the charge q of the capacitor remains
unchanged and that the dielectric fills the entire space between the
capacitor plates. The capacitance of the capacitor in the absence of
the dielectric is C.
The work against the electric forces in this system is equal to the
increment of the electric energy, of the system:
A= 4W= W2 — W1,
where W1is the energy of the field between the capacitor plates in
the presence of the dielectric and W2is the same quantity in the absence
of the dielectric. Bearing in mind that the magnitude of vector D
will not change as a result of the removal of the plate, i.e. that D2 =
= D1 = a, we can write
D2 D2 ) q2 8 )
1 7
A— W2 •••• W1 — 280 2808
2C (1
where V = Sh and C= 8O Slh, S being the area of each plate and h
the distance between them.
The Work of the Field During Polarization of a Dielectric.

An analysis of formula (4.10) for the volume energy density
reveals that for the same value of E, the quantity w is a
times greater in the presence of a dielectric than when it is
absent. At first glance this may seem strange: field intensity
in both cases is maintained the same. As a matter of fact,
when a field is induced in a dielectric, it does an additional
work associated with polarization. Therefore, under the
energy of the field in the dielectric we must understand the
sum of the electric energy proper and an additional work
which is accomplished during polarization of the dielectric.
In order to prove this, let us substitute into (4.10) the
quantity a 0E P for D, which gives
E,E E •P
w— 2 1- 2 •
(4.11)
The first term on the right-hand side coincides with the

energy density of the field E in a vacuum. It remains for
us to verify that the "additional" energy E•P/2 is associated
with polarization.
Let us calculate the work done by the electric field for
polarization of a unit volume of the dielectric, i.e. for the
dl- 1+ d1+
1=1+— 1-
Fig. 4.3
displacements of charges p+ and p'_ respectively along and

against the field upon an increase in the field intensity from
E to E -F- dE. Neglecting the second-order terms, we write
SA = p'+E • d1+ p'E • dl_
where dl +and dl_ are additional displacements due to an
increase in the field by dE (Fig. 4.3). Considering that
p'_ = — p'+ , we obtain
SA = (d1+— -E = (4_ dl •E,
where dl = dl + — dl_ is the additional displacement of
the positive charges relative to the negative charges. Accord-
ing to (3.4), plFdl = dP and we get
SA = E• dP. (4.12)
Since P = x80E, we have
A = E• xec, dE = d X8 2E2 =d E 2) .
Hence, the total work spent for the polarization of a
unit volume of a dielectric is
A = E•P/2, (4.13)
which coincides with the second term in formula (4.11).
Thus, the volume energy density w = E•D/2 includes
the energy E0E2/2 of the field proper and the energy E•P/2
associated with the polarization of the substance.
4.4. A System of Two Charged Bodies 105
4.4. A System of Two Charged Bodies

Suppose that we have a system of two charged bodies in a
vacuum. Let one body create in the surrounding space the
field E1, while the other body, the field E2. The resultant
field is1 E E 2,
=andEthe square of this quantity is
Ez = Kt' E22 +2Ei .E2.
Therefore, according to (4.9), the total energy W of the

given system is equal to the sum of three integrals:
W= E"2
2
dV 8°E1dV
2
ecEi• E2dV , (4.14)
which coincides with formula (4.5) and reveals the field

meaning of the terms appearing in this sum. The first two
integrals in (4.14) are the intrinsic energies of the first and
second charged bodies (W1and W2), while the last integral
is the energy of their interaction (W12).
The following important circumstances should be men-
tioned in connection with formula (4.14).
1. The intrinsic energy of each charged body is an essen-
tially positive quantity. The total energy (4.9) is also always
positive. This can be readily seen from the fact that the in-
tegrand contains essentially positive quantities. However,
the energy of interaction can be either positive or negative.
2. The intrinsic energy of bodies remains constant upon
all possible displacements that do not change the configu-
ration of charges on each body, and consequently this energy
can be assumed to be an additive constant in the expression
for the total energy W. In such cases, the changes in W are
completely determined only by the changes in the interac-
tion energy W12. In particular, this is just the mode of be-
haviour of the energy of a system consisting of two point
charges upon a change in the distance between them.
3. Unlike vector E, the energy of the electric field is
not an additive quantity, i.e. the energy of a field E which
is the sum of fields E1and E2 is generally not equal to the
sum of the energies of these fields in view of the presence of
the interaction energy W12. In particular, if E increases n
times everywhere, the energy of the field increases n2 times.
4.5. Forces Acting in a Dielectric

Electrostriction. Experiments show that a dielectric in
an electric field experiences the action of forces (sometimes
these forces are called ponderomotive). These forces appear
when the dielectric is neutral as a whole. Ponderomotive
forces appear in the long run due to the action of a nonuni-
form electric field on dipole molecules of the polarized dielec-
tric (it is known that a dipole in a nonuniform electric
field is acted upon by a force directed towards the increasing
field). In this case, the forces are caused by the nonuniformi-
ty of not only the macroscopic field but the microscopic
field as well, which is created mainly by the nearest mole-
cules of the polarized dielectric.
Under the action of these electric forces, the polarized
dielectric is deformed. This phenomenon is called electro-
striction. As a result of electrostriction, mechanical stresses
appear in the dielectric.
Owing to electrostriction, not only the electric force
(which depends on the charges) acts on a conductor in a
polarized dielectric, but also an additional mechanical
force caused by the dielectric. In the general case, the effect
of a dielectric on the resultant force acting on a conductor
cannot be taken into account by any simple relations, and
the problem of calculating the forces with simultaneous
analysis of the mechanism of their appearance is, as a rule,
rather complicated. However, in many cases these forces
can be calculated in a sufficiently simple way without a
detailed analysis of their origin by using the law of conser-
vation of energy.
Energy Method for Calculating Forces. This method is
the most general. It allows us to take into account automat-
ically all force interactions (both electric and mechanical)
ignoring their origin, and hence leads to a correct result.
Let us consider the essence of the energy method for cal-
culating forces. The simplest case corresponds to a situa-
tion when charged conductors are disconnected from the
power supply. In this case, the charges on the conductors
remain unchanged, and we may state that the work A of
all internal forces of the system upon slow displacements
of the conductors and dielectrics is done completely at the
expense of a decrease in the electric energy W of the system

(or its field). Here we assume that these displacements do
not cause the transformation of electric energy into other
kinds of energy. To be more precise, it is assumed that such
transformations are negligibly small. Thus, for infinitesimal
displacements we can write
SA= —dWi g, (4.15)
where the symbol q emphasizes that the decrease in the
energy of the system must be calculated when charges on
the conductors are constant.
Equation (4.15) is the initial equation for determining
the forces acting on conductors and dielectrics in the electric
field. This can be done as follows. Suppose that we are in-
terested in the force acting on a given body (a conductor or
a dielectric). Let us displace this body by an infinitesimal
distance dx in the direction X we are interested in. Then
the work of the required force F over the distance dx is
SA = Fxdx, where Fx is the projection of the force F onto
the positive direction of the X-axis. Substituting this ex-
pression for SA into (4.15) and dividing both parts of (4.15)
by dx, we obtain
aw
Fx = Ox
(4.16)
,a •
We must pay attention to the following circumstance.
It is well known that the force depends only on the position
of bodies and on the distribution of charges at a given in-
stant. It cannot depend on how the energy process will deve-
lope if the system starts to move under the action of forces.
And this means that in order to calculate Fx by formula
(4.16), we do not have to select conditions under which all
the charges of the conductor are necessarily constant
(q = const). We must simply find the increment dW under
the condition that q = const, which is a purely mathemati-
cal operation.
It should be noted that if a displacement is performed at
constant potential on the conductors, the corresponding
calculation leads to another expression for the force: Fx =-
= OWIOx 1,. However (and it is important!) the result
of the calculation of F x with the help of this formula or

(4A6) is the same, as should be expected. Therefore, hence-
forth we shall confine ourselves to the application of only
formula (4.16) and will use it for any conditions, including
those where q const upon small displacements. We must
not be confused: the derivative OW/Ox will be calculated
at q = const in such cases as well.
Example. Find the force acting on one of the plates of a parallel-
plate capacitor in a liquid dielectric, if the distance between the
plates is h, the capacitance of the capacitor under given conditions is
C and the voltage U is maintained across its plates.
In this case, if we mentally move the plates apart, the voltage U
remains constant, while the charge q of the capacitor changes (this
follows from the relation C = q/U). In spite of this, we shall calculate
the force under the assumption that q =
X = const, i.e. with the help of formula
x4dx (4.16). Here the most convenient expres-
-a sion for the energy of the capacitor is
x 2
e F
+q W = x
2C 2880S '
0
where e is the permittivity of the dielec-
Fig. 4.4 tric, S is the area of each plate, and x is
the distance between them (x = h).
Next, let us choose the positive direction of the X-axis as is shown
in Fig. 4.4. According to (4.16), the force acting on the upper plate
of the capacitor is
Fx -
ow q2
(1)
Ox 2880S •
The minus sign in this formula indicates that vector F is directed
towards the negative values on the X-axis, i.e. the force is attractive
by nature. Considering that q = GS = DS =- seoES and E = Ulh,
we transform (1) to
Fx = —CU2/2h.
Forces in a Liquid Dielectric. Formula (1) of the last
example shows that the force of interaction between the
plates of a parallel-plate capacitor in a liquid dielectric
is smaller than the corresponding force in a vacuum by a
factor of a (in vacuum a = I). Experiment shows that this
result can be generalized: if the entire space occupied by a
field is filled by a liquid or gaseous dielectric, the forces of
interaction between charged conductors (at constant charges
on them) decrease by a factor of a:

F = Fo18. (4.17)
Hence it follows that two point charges q, and q2 separat-
ed by a distance r and placed into an infinite liquid or gas-
eous dielectric interact with the force
F= Im21 (4.18)
43-c&O ere
which is smaller than the force in a vacuum by a factor of

a. This formula expresses the Coulomb law for point charges
in an infinite dielectric.
We must pay a special attention to the fact that point
charges in this law are extraneous charges concentrated on
macroscopic bodies whose dimensions are small in comparison
with the distance between them. Thus, the law (4.18) has,
unlike the Coulomb law in vacuum, a very narrow field
of application: the dielectric must be homogeneous, infinite,
liquid or gaseous, while the interacting bodies must be
pointlike in the macroscopic sense.
It is interesting to note that the electric field intensity
E, as well as the force F acting on a point charge q in a
homogeneous liquid or gaseous dielectric filling the entire
space of the field, are by a factor of a smaller than the values
of E0 and Fo in the absence of dielectric. This means that
the force F acting on the point charge q in this case is de-
termined by the same formula as in vacuum:
F = qE, (4.19)
where E is the field intensity in the dielectric at the point
where the extraneous charge q is placed. Only in this case
formula (4.19) makes it possible to determine the field E
in the dielectric from the known force F. It should be noted
that another field differing from the field in the dielectric
will be acting on the extraneous charge itself (which is
located on some small body). Nevertheless, formula (4.19)
gives the correct result, strange as it may seem.
Surface Density of a Force. We shall be speaking of the
force acting on a unit surface area of a charged conductor
in a liquid or gaseous dielectric. For this purpose, let us
consider a parallel-plate capacitor in a liquid dielectric.
Suppose that the capacitor is charged and then disconnected

from the power supply to maintain the charge and the field
E of the capacitor constant when the plates are moved apart.
Let us consider once again Fig. 4.4. The energy of the
capacitor is the energy of the field within it. In accordance
with (4.9), this energy is W = (1/2) EDSx, where S is the
surface area of each plate and x is the distance between them
(Sx is the volume occupied by the field). By formula (4.16),
the force acting on the upper plate is
F ,= —OW 10x1 q=
whence the surface density of the force is
E•D
Fu = 2 '
(4.20)
We have obtained an interesting and important result

of a general nature (for a liquid or gaseous dielectric). It
turns out that the surface density of the force acting on a
conductor is equal to the volume density of the electric
energy near the surface. This force is directed always out-
ward along the normal to the surface of the conductor (tend-
ing to stretch it) regardless of the sign of the surface charge.
Problems
• 4.1. Energy of interaction. A point charge q is at a distance /
from an infinite conducting plane. Find the energy W of interaction
between this charge and the charges induced on the plane.
Solution. Let us mentally "freeze" the charge distributed over the
plane and displace under these conditions the point charge q to in-
finity. In this case the charge q will move in the potential field which
is equivalent to the field of a fixed fictitious point charge q, located
—
at a fixed distance 1 on the other side of the plane. We can write

straightaway
1 q2
W=
4a80 21 •
• 4.2. Intrinsic, mutual, and total energies. A system consists
of two concentric metallic shells of radii R1and R2 with charges q1
and q2respectively. Find intrinsic energies W1and W2 of each shell,
the energy W12 of interaction between the shells, and the total elec-
tric energy of the system W, if R2 > R1.
Problems 111
Solution. In accordance with (4.6), the intrinsic energy of each

shell is equal to qw/2, where cp is the potential of a shell created only
by the charge q on it, i.e. cp = q/4nsoR, where R is the shell radius.
Thus, the intrinsic energy of each shell is
2 2
1 41.
W12 -
4ns, 2R1, 2 •
The energy of interaction between the charged shells is equal to the

charge q of one shell multiplied by the potential cp created by the charge
of the other shell in the point of location of the charge q: W12 =
In our case (R2 > R), we have
1 q2 1 q1q2
W12 = q1 ==
41180 R2 43180 R2
The total electric energy of the system is
1 / q2 4 q1(12
W W 2+ 4ns0 2R1 ± 2R2 Rz
• 4.3. Two small metallic balls of radii R1and R2 are in vacuum
at a distance considerably exceeding their dimensions and have a
certain total charge. Find the ratio q1/q2between the charges of the
balls at which the energy of the system is minimal. What is the po-
tential difference between the balls in this case?
Solution. The electric energy of this system is
1
w= wi+ w2+wi2 - 4ns, ,( 2Ri ql -F 2R2
q / ,) '
where W1and W2are the intrinsic electric energies of the balls (qp/2),
W12 is the energy of their interaction (q1cp2 or q2cp1), and / is the dis-
tance between the balls. Since q2 = q —q1, where q is the total charge
of the system, we have
w_ r. qi (q-q1)2 ,qi (q-q1)

L 2R1 2R2
The energy W is minimal when Wag'=- 0. Hence
R R2
q 1 and q2 q
R1+ R 2 Ri+ R2
where we took into account that Ri and R2 are considerably smaller
than / and
q11q2 = R1/R2•
The potential of each ball (they can be considered isolated) is cp qIR.
Hence it follows from the above relation that cp, = cp2, i.e. the poten-
tial difference is equal to zero for such a distribution.
• 4.4. Energy localization in the Held. A charge q is uniformly

distributed inside a sphere of radius R. Assuming that the dielectric
constant is equal to unity everywhere, find the intrinsic electric energy
of the sphere and the ratio of the energy W1localized inside the sphere
to the energy W2 in the surrounding space.
Solution. Let us first find the fields inside and outside the sphere
with the help of the Gauss theorem:
1
El = q r (r R), E2 = ,q (r _›- R).
43-csoR3 ,ITE80 r2
We can now calculate the intrinsic electric energy of the sphere:

R 00
W W i +W 14nr2 dr + 6()2E 4nr2dr= 2 + 1)

EinsR
o 5
0
Hence, it follows that
1 3q2 IV1 1
W=
4aso 5R ' W2 5'
It is interesting to note that the ratio 1471/W2 does not depend on

the radius of the sphere.
• 4.5. A spherical shell is uniformly charged by a charge q.
A point charge gois at its centre. Find the work of electric forces upon
the expansion of the shell from radius R1 to R2.
Solution. The work of electric forces is equal to the decrease in
the electric energy of the system:
A= - — W2*
In order to find the difference W1 W2, we note that upon the expan-
—
sion of the shell (Fig. 4.5), the electric field, and hence the energy
localized in it, changed only in the hatched spherical layer. Conse-
quently,
R2
go
W1—W2= (EI— 4str2 dr,
where El and E2 are the field intensities (in the hatched region at a
distance r from the centre of the system) before and after the expan-
sion of the shell. By using the Gauss theorem, we find
q q. E =
„
qo
47(80 r2 41(8, r2 '
Problems 113
As a result of integration, we obtain

1\
A— (40+q/2)
4a80 \ R1 R2 •
Remark. If we try to calculate the work in terms of the potential

as A = q (cp, — cp,), where cp is the potential created by the charge qo
at the point of location of the charge q, the answer would be different
Fig. 4.5 Fig. 4.6
and incorrect. This is due to the fact that this approach does not take
into account the additional work performed by electric forces upon
the change in the configuration of the charge q located on the expand-
ing shell.
e 4.6. A point charge q is at the centre of a spherical uncharged
conducting layer whose inner and outer radii are a and b respectively.
Find the work done by electric forces in this system upon the remov-
al of the charge q from its original position through a small hole
(Fig. 4.6) to a very large distance from the spherical layer.
Solution. We shall proceed from the fact that the work of electric
forces is equal to the decrease in the electric energy of the system. As
is well known, the latter is localized in the field itself. Thus, the prob-
lem is reduced to determination of the change in the field as a result
of this process.
It can be easily seen that the field around the charge q will change
only within the spherical layer with the inner radius a and the outer
radius b. Indeed, in the initial position of the charge there was no
field in this region, while in the final position there is a certain field
(since the conducting spherical layer is far from the charge q). Conse-
quently, the required work is
b
so E 2
A=0— W1=- — r 2 dV.
a
Considering that E = q/43-ceor2 and dV = 4nr2 dr, and integrating,
we obtain
q2 a—b
A— <O.
87mo ab
8-0181
• 4.7. The work done upon moving capacitor plates apart.

A parallel-plate air capacitor has the plates of area S each. Find the
work A' against the electric forces, done to increase the distance be-
tween the plates from x1 to x 2, if (1) the charge q of the capacitor and
(2) its voltage U are maintained constant. Find the increments of the
electric energy of the capacitor in the two cases.
Solution. (1) The required work is
q2
A' = gEl (x2 — x1)----- 280s(x2—x1),
where E1is the intensity of the field created by one plate (E = 0/2e0).
It is in this field that the charge located on the other plate moves.
This work is completely spent for increasing the electric energy:
AW = A'.
(2) In this case, the force acting on each capacitor plate will de-
pend on the distance between the plates. Let us write the elementary
work of the force acting on a plate during its displacement over a dis-
tance dx relative to the other plate:
60SU 2 dx
— c/Ei
2 x2
where we took into account that q = CU, El = U/2x, and C = eoSix.

After integration, we obtain
A, = 80SU 2 I 1 1 )
>O.
2 k x, x2
The increment of the electric energy of the capacitor is
(C2 — CO U2 80SU 2 1 1 \
OW
2 2 k x2 x,
< 0.
It should be noted that AW = — A'.

Thus, by moving the plates apart, we perform a positive work
(against the electric forces). The energy of the capacitor decreases in
this case. In order to understand this, we must consider a source main-
taining the potential difference of the capacitor at a constant value.
This source also accomplishes the work A s. According to the law of
conservation of energy, A s + A' = AW, whence A R = AW — A' =
= — 2A' <O.
• 4.8. Forces acting between conductors in a dielectric. A paral-
lel-plate capacitor is immersed, in the horizontal position, into a liq-
uid dielectric with the dielectric constant a, filling the gap of width h
between the plates. Then the capacitor is connected to a source of
Problems 115
permanent voltage U. Find the force f' acting on a unit surface of the
plate from the dielectric.
Solution. The resultant force f acting per unit area of each plate
can be represented as
= to — f', (1)
where fo is the electric force acting per unit area of a plate from the
other plate (it is just the force per unit area when the dielectric is
absent). In our case, we have
f = 1018, f o = GE = a2/280, (2)
where E is the field intensity in the region occupied by one plate, creat-
ed by the charges of the other plate. Considering that a = D =
= seoU/12, and substituting (2) into (1), we obtain
fo (1 — 1/e) = a (e — 1) 80 U2/2h2.
For example, for U = 500 V, h = 1.0 mm and a = 81 (water), we get
= 7 kPa (0.07 atm).
• 4.9. The force acting on a dielectric. A cylindrical layer of a
homogeneous dielectric with the dielectric constant a is introduced
into a cylindrical capacitor so that the layer fills the gap of width d
between the plates. The mean radius of the plates is 12 such that R >> d.
The capacitor is connected to a source of a permanent voltage U. Find
the force pulling the dielectric inside the capacitor.
Solution. Using the formula W = q2/2C for the energy of a capac-
itor, we find that, in accordance with (4.16), the required force is
OW q2 OC/Ox U2 OC
Fx — (1)
q= 2 C2 — 2 Ox '
Since d << R the capacitance of the given capacitor can be calculated
by the formula for a parallel-plate capacitor. Therefore, if the diele-
ctric is introduced to a depth x and the capacitor length is 1, we have
eeox • 2a12 80 (/— x)• 2nR so• 2nR
(ex+1 — x). (2)
Substituting (2) into (1), we obtain

Fs = so (a — 1) nRU 2Id.
• 4.10. A capacitor consists of two fixed plates in the form of a
semicircle of radius R and a movable plate of thickness h made of a
dielectric with the dielectric constant a, placed between them. The
latter plate can freely rotate about the axis 0 (Fig. 4.7) and practical-
ly fills the entire gap between the fixed plates. A constant voltage U is
maintained between the plates. Find the moment M about the axis 0
of forces acting on the movable plate when it is placed as shown in the
figure.
Solution. The work performed by the moment of forces M upon
8•
116 5. Direct Current
the rotation of the plate through an angle element da is equal to the

decrease in the electric energy of the system at q = const [see (4.16)]:
M z da= — dWl q ,
where W = q2/2C. Hence
aW q2 OC/aa
Mz (1)
as q 2 C2 •
In the case under consideration, C = C1 + CE, where C1 and Cc
are the capacitances of the parts of the capacitor with and without the
dielectric. The area of a sector with an angle a is determined as S =
=
- aR2/2, and hence
C = coaR2/2h eeo (a — a) R2/2h.
Differentiating with respect to a, we find
OCIaa = ( e oR2/2h) (1 — a) . Substituting
this expression into formula (1) and con-
sidering that C = qIU, we obtain
U2 s oR2
M1= 2 21z (I 8)
Fig. 4.7 = — (8-1) ,oRzuz
4h < 0.
The negative sign of Mzindicates that the moment of the force is

acting clockwise (oppositely to the positive direction of the angle a;
see Fig. 4.7). This moment tends to pull the dielectric inside the ca-
pacitor.
It should be noted that M1is independent of the angle a. However,
in equilibrium, when a = 0, the moment Mz = 0. This discrepancy
is due to the fact that for small values of a we cannot ignore edge effects
as was done in the solution of this problem.
5. Direct Current
5.1. Current Density. Continuity Equation
Electric Current. In this chapter, we shall confine our-
selves to an analysis of conduction current in a conducting
medium, especially in metals. It is well known that electric
current is the transfer of charge through a certain surface
S (say, the cross section of a conductor).
In a conducting medium, current can be carried by elec-
trons (in metals), ions (in electrolytes), or some other parti-
cles. In the absence of electric field, current carriers perform
chaotic motion and on average the same number of carriers
of either sign passes through each side of any imaginary
surface S. Thus, the current passing through S in this case
5.1. Current Density. Continuity Equation 117
is zero. However, when an electric field is applied, an orde-

red motion with a certain average velocity u is imposed on
the chaotic motion of the carriers, and a current flows
through the surface S. Thus, an electric current is essentially
an ordered transfer of electric charges.
The quantitative measure of electric current is intensity
I, defined as the charge transferred across the surface S in
a unit time:
I = dQ/dt.
The unit of current is the ampere (A).
Current Density. Electric current may be distributed non-
uniformly over the surface through which it passes. Hence,
in order to characterize the current in greater detail, current
density vector j is introduced. The magnitude of this vector
is equal to the ratio of the current dI through a surface
element perpendicular to the direction of motion of charge
carriers to the area dS_L of this element: j = dildS±. For
the direction of vector j we take the direction of velocity
vector u of the ordered motion of positive carriers (or the
direction opposite to that of the velocity vector of the or-
dered motion of negative carriers). If the current consists of
positive and negative charges, its density is defined by the
formula
j = P+u+ P-u-, (5.1)
where P+and p_ are the volume densities of positive and
negative charge carriers respectively, and u +and u_ are the
velocities of their ordered motion. In conductors, where the
charge is carried only by electrons (p_ < 0 and u + = 0),
the current density is given by
j = p_u_. (5.2)
The field of vector j can be graphically represented with
the help of lines of current (lines of vector j), which are
drawn in the same way as for vector E.
Knowing the current density vector at each point of the
surface S under consideration, we can also find the current
passing through this surface as the flux of vector j:
I= j•dS. (5 . 3)
Current I is a scalar algebraic quantity. It can be seen

from formula (5.3) that besides other factors, the sign of
the current is determined by the choice of the direction of
the normal at each point on the surface S, i.e. by the choice
of the direction of vectors dS. If the directions of all the
vectors dS are reversed, the current I changes sign.
Continuity Equation. Let us imagine a closed surface S
in a conducting medium through which a current is passing.
It is customary to choose the outward direction for vectors
normal to closed surfaces, and hence, for vectors dS. Con-
sequently, the integral j dS gives the charge leaving
the volume V (enveloped by the surface S) per unit time.
According to the law of conservation of charge, this integral
is equal to the decrease in the charge inside volume V per
unit time:
j dS = at (5.4)
This relation is called the continuity equation. It is an ex-

pression for the law of conservation of charge.
In the case of a steady-state (direct) current, the charge
distribution in space must remain unchanged. In other
words, on the right-hand side of Eq. (5.4) dq/dt = 0. Con-
sequently, for direct current we have.
j dS -= O. (5.5)
This means that in this case, the lines of vector j do not

start or terminate anywhere. The field of vector j is said
to have no sources in the case of direct current.
Differential form of the continuity equation. Let us write Eqs.
(5.4) and (5.5) in the differential form. For this purpose, we express the
—
charge q as ,c p d V and the right-hand side of Eq. (5.4) as
c0
Op
—
t
--
dt
p dV
dV. Here, we have taken the partial time derivative of p,

since p can depend on time as well as on coordinates. Thus,
j dS= — ap dV.
at
Proceeding in the same way as for the flux of vector E in Sec. 1.4, we
find that the divergence of vector j at a certain point is equal to the
decrease in the charge density per unit time at the same point:
--aplat. (5.6)
This leads to the steady-state condition (when ap/at = 0):
p.i=o. I (5.7 )
This means that for direct current the field of vector j does not have
sources.
5.2. Ohm's Law for a Homogeneous Conductor

Ohm's law, which was discovered experimentally, states:
the current passing through a homogeneous conductor is pro-
portional to the potential difference across its terminals (or
to the voltage U):
I (5.8)
where R is the electric resistance of the conductor.
The unit of resistance is the ohm (Q).
Resistance R depends on the shape and size of the conduct-
or, on its material and temperature, and above all on the
configuration (distribution) of the current through the con-
ductor. The meaning of resistance is quite clear when we
are dealing with a wire. In a more general case of the volume
distribution of current, it is meaningless to speak of re-
sistance until we know either the position of the leads at-
tached to the conductor or the current configuration.
In the simplest case of a homogeneous cylindrical con-
ductor, resistance is given by
R =-- p-s,- , (5.9)
where 1 is the length of the conductor, S is its cross-sectional

area, and p is the resistivity, which depends on the material
of the conductor and its temperature. Resistivity is measured
in ohm •metres (Q • m).
Resistivity of the best conductors like copper and alumi-
nium is of the order of 10-8Q•m at room temperature.
Differential Form of Ohm's Law. Let us establish a re-

lation between the current density j and the field E at a
point of a conducting medium. We shall confine ourselves
to an isotropic conductor in which the directions of vectors
j and E coincide.
Let us mentally isolate around a certain point in a con-
ducting medium a cylindrical volume element with genera-
trices parallel to vector j, and hence to vector E. If the cylind-
er has a cross-sectional area dS and length dl, we can write
on the basis of (5.8) and (5.9) the following expression for
such cylindrical element:
E dl
j dS —
p dl/dS
After cancelling out common terms we obtain the following

equation in vector form:
j= — E = GE, (5.10)
where a = 1/p is the conductivity of the medium. The unit

reciprocal to an ohm is called a mho (Siemens). Consequently,
the unit for the measurement of a is mho per metre.
Equation (5.10) is the differential form of Ohm's law.
It does not contain any differentials (derivatives), but is
called the differential form since it establishes a relation
between quantities corresponding to the same point in a
conductor. In other words, Eq. (5.10) is an expression for
Ohm's law at a point.
Methods for Calculating Resistance (R). There are several
methods for measuring resistance and all of them are ulti-
mately based on the application of relations (5.8)-(5.10).
The expedience of using a particular method is determined
by the formulation of the problem and by its symmetry.
The practical applications of these methods are described in
Problems 5.1-5.3 and 5.6.
Charge in a Current-carrying Conductor. If the current is
direct (constant) the excess charge inside a homogeneous
conductor is equal to zero everywhere. Indeed, Eq. (5.5)
is valid for direct current. Taking into account Ohm's law
in the form (5.10) we can rewrite (5.5) as follows:

oE dS = 0,
where the integration is performed over any closed surface

S inside the conductor. For a homogeneous conductor, the
quantity a can be taken out of the integral:
a E dS = O.
According to the Gauss theorem, the remaining integral is

proportional to the algebraic sum of the charges inside the
closed surface S, i.e. proportional to
the excess charge in this surface. How-
ever, it can be directly seen from the
last equation that this integral is
equal to zero (since a / 0). This means
that the excess charge is also equal to
zero. Since the surface S is chosen ar-
bitrarily, the excess charge under these
conditions is always equal to zero in-
side a conductor.
The excess charge can appear only Fig. 5.1
at the surface of a homogeneous con-
ductor at places where it comes in contact with other
conductors, as well as in the regions where the conductor
has inhomogeneities.
Electric Field of a Current-carrying Conductor. Thus,
when current flows, an excess charge appears on the surface
of a conductor (inhomogeneity region). In accordance with
(2.2), this means that the normal component of vector E
appears at the outer surface of the conductor. Further, from
the continuity of the tangential component of vector E,
we conclude that a tangential component of this vector also
exists near the surface of the conductor. Thus, near the sur-
face of the conductor, vector E forms (in the presence of the
current) a nonzero angle a with the vertical (Fig. 5.1). When
the current is absent a = 0.
If the current flows in a steady state, the distribution
of electric charges in the conducting medium (generally
speaking, inhomogeneous) does not change in time, although
the charges are in motion: at each point, new charges con-

tinuously replace the departing ones. These moving charges
create a Coulomb field which is identical to the one created
by stationary charges of the same configuration. This means
that the electric field of stationary currents is a potential
field.
At the same time, electric field in the case of stationary
currents differs significantly from electrostatic, or Coulomb,
field of fixed charges. If charges are in equilibrium, the
electrostatic field inside a conductor is always equal to zero.
The electric field of stationary currents is also an electro-
static (Coulomb) field, although the charges inducing this
field are in motion. Hence, the field E of steady-state currents
exists also inside a current-carrying conductor.
5.3. Generalized Ohm's Law

Extraneous Forces. If all the forces acting on charge
carriers were reduced to electrostatic forces, the positive
carriers would move under the action of these forces from
regions with a higher potential to those with a lower poten-
tial, while the negative carriers would move in the opposite
direction. This would lead to the equalization of potentials,
and the potentials of all mutually connected conductors
would become the same, thus terminating the current. In
other words, under the action of Coulomb forces alone, a
stationary field must be a static field.
In order to prevent this, a direct current circuit must
contain, besides the subcircuits in which positive carriers
move in the direction of decreasing potential cp, subcircuits
in which positive carriers move in the direction of increas-
ing cp or against the electric field forces. In these subcircuits,
the transfer of carriers is possible only through forces that
are not electrostatic. We shall call such forces extraneous.
Thus, a direct current can be sustained only with the
help of extraneous forces which are applied either to cer-
tain parts of the circuit, or to the entire circuit. The
physical nature of extraneous forces may be quite diverse.
Such forces may be created by physical or chemical inho-
mogeneities in a conductor, for example, forces resulting
from a contact of conductors of different types (galvanic
cells, accumulators) or conductors at different temperatures

(thermocouples).
Generalized Ohm's Law. In order to describe extraneous
forces quantitatively, the concepts of extraneous force field
and its strength E* are introduced. The latter is numerically
equal to the extraneous force acting on a unit positive charge.
Let us now turn to current density. If an electric field E
generates in a conductor a current of density j = GE, it
is obvious that under the combined action of the field E
and the extraneous force field E*, the current density will
be given by
j = a (E E*). (5.11)
This equation generalizes the law (5.10) to the case of

inhomogeneous regions of a conducting medium, and is
called the generalized Ohm's law in the differential form.
Ohm's Law for a Nonuniform Subcircuit. Nonuniform
subcircuits are characterized by the presence of extraneous
forces in them.
Let us consider a special, but practically quite important
case when a current flows along a thin wire. In this case, the
direction of the current will coincide with the axis of the
wire and the current density j can be assumed to be the same
at all points on the cross section of the wire. Suppose that
the cross-sectional area of the wire is equal to S (S need not
be the same over the entire length of the wire).
Dividing both sides of Eq. (5.11) by G, forming a scalar
product of the resulting expression with the element dl
of the wire along its axis from cross section 1 to cross section
2 (this direction is taken as the positive direction) and in-
tegrating over the length of the wire between the two cross
sections, we obtain
2 2 2
X dl
E dl E* dl. (5.12)
j a
Let us transform the integrand in the first integral: we

replace 6 by 1/p and j dl by /1d1, where ji is the projection
of vector j onto the direction of vector dl. Further, we note
that ji is an algebraic quantity and depends on the orienta-
tion of vector j relative to dl: if j fit dl then j i > 0; if j dl,

then / 1 < 0. Finally, we replace :7 / by ES, where I, the
current, is also an algebraic quantity (like j /). Since I is
the same in all sections of the circuit when we are dealing
with direct current, this quantity can be taken out of the
integral. Cosequently, we get
2 2
dl dl
I P (5.13)
a
The expression p dl/S defines the resistance of the sub-

circuit of length dl, while the integral of this expression is
the total resistance R of the circuit between cross sections
and 2.
Let us now consider the right-hand side of (5.12). Here,
R the first integral is the poten-
tial difference (pi— W2, while
1 2 the second integral is the elec-
tromotive force (e.m.f.) act- e„
ing in the given subcircuit:
2
612— E* dl. (5.14)
Fig. 5.2 Like the current I, the electro-

motive force is an algebraic
quantity: if e.m.f. facilitates the motion of positive carriers
in a certain direction, then en > 0; if, the e.m.f. hinders
this motion X12 <0.
As a result of all the transformations described above,
(5.12) assumes the following form:
RI =cp1 — cPz (5.15)
This equation is the integral form of Ohm's law for a non-

uniform subcircuit [cf. Eq. (5.11) which describes the same
law in the differential form].
Example. Consider the subcircuit shown in Fig. 5.2. The resistance
is nonzero only in the segment R. The lower part of the figure shows
the variation of potential IT over this region. Let us analyse the course
of events in this circuit.
Since the potential decreases over the segment R from left to right,
I > 0. This means that the current flows in the positive direction
(from 1 to 2). In the present case, pi< 11)2, but the current flows from
point 1 to point 2, i.e. towards increasing potential. This is possible
only because an e.m.f. is acting in this circuit frdm 1 to 2, i.e. in
the positive direction.
Let us consider Eq. (5.15). It follows from this equation

that points 1 and 2 are identical for a closed circuit, i.e.
yi = y2. In this case, the equation acquires a simpler form:
RI = g, (5.16)
where R is the total resistance of the closed circuit, and g
is the algebraic sum of all the e.m.f.'s in the circuit.
Next, we consider the subcircuit containing the source
of the e.m.f. between terminals 1 and 2. Then, R in (5.15)
will be the internal resistance of the e.m.f. source in the
subcircuit under consideration, while yl — y, the potential
difference across its terminals. If the source is disconnected,
I = 0 and e= cp 2 In other words, the e.m.f. of
the source can be defined as the potential difference across
its terminals in the open circuit.
The potential difference across the terminals of an e.m.f.
source connected to an external resistance is always less
than its e.m.f. and depends on the load.
Example. The external resistance of a circuit is ii times higher than

the internal resistance of the source. Find the ratio of the potential
difference across the terminals of the source to its e.m.f.
Let Ribe the internal resistance of the source and Ra the external
resistance of the circuit. According to (5.15), cp2 — = e — RiI,
while according to (5.16), (111 + Ra) I e. From these two equa-
tions, we get
Ri R
RH-Ra — Ri -1-11„ — 1+11'
It follows from here that the higher the value of 1, the closer the po-
tential difference across the current terminals to its e.m.f., and vice
versa.
Concluding this section, let us consider a picture illus-

trating the flow of direct current in a closed circuit. Fig-
ure 5.3 shows the distribution of potential along a closed
circuit containing the e.m.f. source on segment AB. For

the sake of clarity of representation, potential cp is plotted
along the generatrices of a cylindri-
9A
cal surface resting on the current con-
tour. Points A and B correspond to
the positive and negative terminals
of the source. It can be seen from the
figure that the flow of current can be
described as follows: the positive
A B charge carriers "slide" along the in-
Fig. 5.3 clined "chute" from point WA to point
cp B along the external subcircuit,
while inside the source, carriers "rise" from point cp Bto WA
with the help of extraneous force shown by an arrow.
5.4. Branched Circuits. Kirchhoff's Laws

Calculations of branched circuits, for example, determina-
tion of current in individual branches, can be considerably
simplified by using the following two Kirchhoff's laws.
Kirchhoff's First Law pertains to the junctions, i.e.
branch points in a circuit, and states that the algebraic
sum of the currents meeting at a junction is equal to zero:
`11-7 ,= 0. (5.17)
Here, currents converging at a junction and diverging from
it are supposed to have opposite signs; for example, we can
assume the former to be positive and the latter to be negative
(or vice versa, this is immaterial). When applied to Fig. 5.4,
Eq. (5.17) assumes the form Il — I2 + /3 = O.
Equation (5.17) is a consequence of the steady-state con-
dition (5.7); otherwise, the charge at a junction would change
and the currents would not be in a steady state.
Kirchhoff's Second Law. This law is applicable to any
closed contour in a branched circuit and states that the
algebraic sum of the products of current and resistance in each
part of a network is equal to the algebraic sum of e.m.f.'s in
the circuit:
E rkR, = E eh. (5.18)
5.4. Branched Circuits. Kirchhoff's Laws 127
In order to prove this law, it is sufficient to consider the

case when the circuit consists of three subcircuits (Fig. 5.5).
Let us assume the direction of circumvention to be clockwise,
Fig. 5.4 Fig. 5.5

as shown in the figure. Applying Ohm's law (5.15) to each
of the three subcircuits, we obtain
= (P2 - (P3 el,
12112= (1)3 - (PI ± t2?
/ 3R3 (P2 e3.
Adding these equations and cancelling all potentials, we
arrive at formula (5.18), i.e. Kirchhoff's second law.
Thus, Eq. (5.18) is a consequence of Ohm's law for non-
uniform subcircuits.
Setting up of a System of Equations. In each specific case,
Kirchhoff's laws lead to a complete system of algebraic
equations which can be used, say, for finding all the un-
known currents in the circuit.
The number of equations of the form (5.17) and (5.18)
must be equal to the number of unknown quantities. Care
should be taken to ensure that none of the equations is a
corollary of any other equation in the system:
(1) if a branched circuit has N junctions, independent
equations of type (5.17) can be set up only for N — 1 junc-
tions, and the equation for the last junction will be a corol-
lary of the preceding equations;
(2) if a branched circuit contains several closed loops,
the independent equations of type (5.18) can be set up only
for loops which cannot be obtained by superimposing
the loops considered before. For example, such equa-
tions will be independent for loops 124 and 234 of the
circuit shown in Fig. 5.6. Equation for the loop 1234

will follow from the two preceding ones. It is possible to
set up autonomous equations for two other loops, say,
124 and 1234. Then the equation for contour 234 will be a
Fig. 5.6 Fig. 5.7

consequence of these two equations. The number of auto-
nomous equations of type (5.18) will be equal to the smallest
number of discontinuities which must be created in a cir-
cuit in order to break all the loops. This number is
equal to the number of subcircuits bounded by conductors
if the circuit can be drawn in a plane without intersections.
For example, it is necessary to set up three equations of
type (5.17) and three of type (5.18) for a circuit (Fig. 5.7)
containing four junctions. This is so because the number of
discontinuities (marked by crosses in the circuit) breaking
all the loops is three (the number of subcircuits is also
equal to three in this case). If we assume the currents to
be unknown the number of discontinuities will be six in
accordance with the number of subcircuits between junctions,
which corresponds to the number of independent equations.
The following procedure should be adopted while set-
ting up equations of type (5.17) and (5.18).
1. The directions of currents are marked hypothetically
by arrows, without caring for the direction of these arrows.
If a certain current turns out to be positive as a result of
calculations, this means that the direction chosen for it is
correct. If, however, the current is found to be negative, its
actual direction will be opposite to the one pointed by the
arrow.
2. Having selected an arbitrary closed loop, we cir-
cumvent its sections in one direction, say, clockwise. If
the assumed direction for any current coincides with the
direction of circumvention, the corresponding term IR in

(5.18) should be taken with the positive sign; in the opposite
case, the minus sign should be used. The same procedure is
applicable to 6: if an e.m.f. increases potential in the di-
rection of circumvention, it should be assigned the positive
sign; the negative sign should be used in the opposite case.
Example. Find the magnitude and direction of the current passing
through resistor R in the circuit shown in Fig. 5.8. All resistances and
e.m.f.'s are assumed to be known.
R
There are three subcircuits, and hence,
three unknown currents /, I and /2 in
this circuit. We mark (arbitrarily) the sup-
posed directions of these currents by ar-
rows (at the right junction).
The circuit contains N = 2 junctions.
This means that there is only one inde-
pendent equation of type (5.17):
I // + /2 = 0.
Let us now set up equations of type
(5.18). According to the number of sub- G2 R2
circuits, there should be two of them. Let
us consider the loop containing R and Fig. 5.8
R1and the loop with R and R2. Taking
the clockwise direction for circumventing each of these loops, we
can write
— IR + I1111 = —IR + I 2R2 =
We can verify that the corresponding equation for the loop with
R1and R2 can be obtained from these two equations. Solving the
system of three equations, we obtain
—R1e2 -1-R2g1
/—
R 1R2-FRRI +RR2 -
If we find that as a result of substitution of numerical values into
this equation I > 0, the current actually flows as shown in Fig. 5.8.
If I < 0, the current flows in the opposite direction.
5.5. Joule's Law

The passage of current through a conductor having a
resistance is invariably accompanied by liberation of heat
(heating of conductors). Our task is to find the quantity
of heat liberated per unit time in a certain part of the cir-
9-0181
cuit. We shall consider two cases which are possible, viz.

uniform and nonuniform subcircuits. The problem is consid-
ered on the basis of the law of conservation of energy and
Ohm's law.
Uniform Subcircuit. Suppose that we are interested in
the region between cross sections 1 and 2 of a conductor
(Fig. 5.9). Let us find the work
done by the field during a time
interval dt in the region 12.
2 If the current through the con-
ductor is equal to I, a charge
dq = I dt will pass through each
cross section of the conductor
Fig. 5.9 during the time dt. In particular,
this charge dq will enter cross
section 1 and the same charge will leave cross section 2.
Since charge distribution in the conductor remains unchanged
in this case (the current is direct), the whole process
is equivalent to a transfer of charge dq from section 1 to
section 2 with potentials (piand cp, respectively.
Hence the work done by the field in such a charge trans-
fer is
6.4 = dq (cpi — cp2) = I (cpi — cp2) dt.
According to the law of conservation of energy, an equiv-
alent amount of energy must be liberated in another form.
If the conductor is stationary and no chemical transforma-
tions take place in it, this energy must be liberated in the
form of internal (thermal) energy. Consequently, the con-
ductor gets heated. The mechanism of this transformation
is quite simple: as a result of the work done by the field,
charge carriers (for example, electrons in metals) acquire
an additional kinetic energy which is then spent on exciting
lattice vibrations due to collisions of the carriers with the
lattice sites' atoms.
Thus, in accordance with the law of conservation of ener-
gy, the elementary work SA = Q dt, where Q is the heat
liberated per unit time (thermal power). A comparison of
this equation with the preceding one gives
Q=I((Pi—TO).
Since in accordance with Ohm's law cp,cp2 = RI,

we get
RI 2. (5.19)
This is the expression for the well-known Joule's law.

We shall now derive an expression for the differential
form of this law, characterizing the liberation of heat at
different parts of a conducting medium. For this purpose,
we isolate a volume element of this medium in the form of
a cylinder with its generatrices parallel to vector j, viz.
the current density at the given point. Let dS and dl be the
cross-sectional area and length of the small cylinder re-
spectively. The amount of heat liberated in this volume during
the time dt will be given, in accordance with Joule's law, by
8Q= RI 2 dt = pdsdi (j dS)2dt = pj2dV dt,

where dV = dS dl is the volume of the small cylinder. Di-
viding the last equation by dV dt, we obtain a formula for
the amount of heat liberated per unit time in a unit volume
of the conducting medium, or the thermal power density of
the current:
Qd= P12 . (5.20)
This formula expresses Joule's law in the differential form:

the thermal power density of current at any point is propor-
tional to the square of the current density and to the resistivity
of the medium at that point.
Equation (5.20) is the most general form of Joule's law,
applicable to all conductors irrespective of their shape or
homogeneity, as well as the nature of forces which induce
the electric current. If the charge carriers are subjected to
electric forces only, we can write, on the basis of Ohm's law
(5.10),
= • E = o-E2. (5.21)
This equation is of a less general nature than (5.20).
Nonuniform Subcircuit. If a subcircuit contains a source
0*
of e.m.f., the charge carriers will be subjected not only to

electric forces, but to extraneous forces as well. In accordance
with the law of conservation of energy, the amount of heat
liberated in this case will be equal to the algebraic sum of
the works done by the electric and extraneous forces. The
same applies to the corresponding powers: the thermal power
must be equal to the algebraic sum of the powers due to
electric and extraneous forces. This can be easily verified
by multiplying (5.15) by I:
RI2 = ((pi— cp,) I 6I. (5.22)
The left-hand side of this expression is the thermal power
Q liberated in the subcircuit. In the presence of extraneous
forces, the value of Q is determined by the same formula
(5.19) which is applied to a uniform subcircuit. The last
term on the right-hand side is the power generated by ex-
traneous forces in the subcircuit under consideration. It
should also be noted that the last term (I) is an algebraic
quantity and, unlike RI2, it reverses its sign when the di-
rection of the current I is reversed.
Thus, (5.22) indicates that the heat liberated in the re-
gion between points 1 and 2 is equal to the algebraic sum of
the electric and extraneous powers. The sum of these powers,
i.e. the right-hand side of this equation, is called the electric
power developed in the subcircuit. It can then be stated
that for a stationary subcircuit, the thermal power evolved
is equal to the electric power developed in the subcircuit
by the current.
Applying (5.22) to the entire unbranched circuit ((pi
(1)2), we get
(2 ti.
- (5.23)
In other words, the total Joule heat liberated per unit time
in the entire circuit is equal to the power developed by
extraneous forces only. This means that heat is generated
only by extraneous forces. The role of the electric field
is reduced to redistribution of this heat over different parts
of the circuit.
We shall now derive Eq. (5.22) in the differential form.
For this purpose, we multiply both sides of Eq. (5.11) by
5.6. Transient Processes in a Capacitor Circuit 133
j, and consider that = 1/p and pj2 =Qd (see (5.20)).

The thermal power density of current in an inhomogeneous
medium can then be written in the form
Qd = p12 =j•(E -I- E*). (5.24)
5.6. Transient Processes in a Capacitor Circuit

Transient Processes. These are the processes involving
a transition from one stationary regime of the circuit to
another. An example of such a process-
is the charging and discharging of a
capacitor, which will be considered in
detail in this section. Cis+
So far, we have considered only di-
rect currents. It turns out, however, that
most of the laws obtained in this case • •
are also applicable to alternating cur-
rents. This applies to all cases in which Fig. 5.10
the variation of current is not too rapid.
The instantaneous value of current will then practically
be the same at all cross sections of the circuit. Such currents
and fields associated with them are called quasistationary
(a more exact criterion for quasistationary currents and
fields is given in Sec. 11.1).
Quasistationary currents can be described by the laws
obtained for constant currents by applying these laws to
instantaneous values of quantities.
Let us now consider the discharging and charging of a
capacitor, assuming that the currents in these processes
are quasistationary.
Discharging of a Capacitor. If the plates of a charged
capacitor with capacitance C are connected through a resis-
tor R, a current will flow through the resistor. Let I, q
and U be the instantaneous values of the current, charge
of the positive plate, and the potential difference between
the plates (voltage).
Assuming that the current I is positive when it flows
from the positive plate to the negative plate (Fig. 5.10),
we can write I = — dqldt. According to Ohm's law, we

obtain the following expression for the external subcircuit
having resistance R:
RI = U.
Considering that I = — dqldt and U = qIC , we can
transform the above equation as follows:
dq
dt
(5.25)
RC U.
After separating the variables in this differential equation

and integrating, we get
goo — t (5.26)
where q0 is the initial charge on the capacitor and t is a
constant having dimensions of time:
T= RC. (5.27)
This constant is called relaxation time. It can be seen from
(5.26) that T is the time during which the capacitor charge
decreases to lie of its initial value.
Differentiating (5.26) with respect to time, we obtain
the law of variation of current:
dq
— — I e- (5.28)
dt °
where /0 = qo/T is the current at the instant t = 0.

Figure 5.11 shows the dependence of the capacitor charge
q on time t. The I vs. t dependence also has the same form.
Charging of a Capacitor. Consider a circuit in which
a capacitor C, a resistor R and a source of e.m.f. are
connected in series (Fig. 5.12). To begin with, the capacitor
is not charged (key K is open). At the instant t = 0 the
key is closed, and a current starts to flow through the circuit,
thus charging the capacitor. The increasing charge on the
capacitor plates will obstruct the passage of the current
and gradually decrease it.
The current itt the circuit will now be assumed positive
if it flows toward the positive plate of the capacitor: I
dq/dt. Applying Ohm's law for a nonuniform subcircuit
5.6. Transient Processes in a Capacitor Circuit 135
to the loop 1e112, we get

RI = pi — cP2 e,
where R is the total resistance of the circuit, including
the internal resistance of the e.m.f. source. Considering
0
Fig. 5.11 Fig. 5.12
that I == dqldt and cp, — ch, = U = q/C, we can rewrite

the last equation in the form
dq g — qIC
dt R •
Separating the variables, we get

R dy
— dt.
g —q1C
Fig. 5.13
Integrating this equation under
the initial condition q = 0 at t = 0, we obtain
RC In ( 1 — e
q — t,
C
whence
qm (1—e-tIT). (5.29)
Here q„, = eC is the limiting value of the capacitor charge
(for t —›- oo), and r = RC.
The current varies with time according to the following
law:
dq
I = ---= I°e -tIt (5.30)
dt
where /0 =
The dependences of q and I on t are shown in Fig. 5.13.
Problems
• 5.1. Resistance of a conducting medium. A metallic sphere of
radius a is surrounded by a concentric thin metallic shell of radius b.
The space between these two electrodes is filled with a homogeneous
poorly conducting medium of resistivity p. Find the resistance of the
gap between the two electrodes.
Solution. Let us mentally isolate a thin spherical layer between
radii r and r + dr. Lines of current at all points of this layer are per-
pendicular to it, and therefore such a layer can be treated as a cylin-
drical conductor of length dr and a cross-sectional area 4Tcr2. Using
formula (5.9), we can write
dr
dR—p
4nr2 '
Integrating this expression with respect to r between a and b, we get
R p 1 1
4a a b)•
• 5.2. Two identical metallic balls of radius a are placed in a
homogeneous poorly conducting medium with resistivity p. Find the
resistance of the medium between the balls under the condition that
the distance between them is much larger than their size.
Solution. Let us mentally impart charges +q and — q to the balls.
Since the balls are at a large distance from one another, electric field
near the surface of each ball is practically determined only by the
charge of the nearest sphere, and its charge can be considered to be
uniformly distributed over the surface. Surrounding the positively
charged ball by a concentric sphere adjoining directly the ball's sur-
face, we write the expression for the current through this sphere;
I = 4na2 j,
where j is the current density. Using Ohm's law (j = Elp) and the
formula E = q/43180a2, we obtain
I = q/eop.
Let us now find the potential difference between the balls:
U = cp+ — 2q/4nsoa.
The sought resistance is given by
11 = U// = p/2na.
This result is valid regardless of the magnitude of the dielectric con-
stant of the medium.
O 5.3. Two conductors of arbitrary shape are placed into an in-
finite homogeneous slightly conducting medium with the resistivity p
and the dielectric constant 8. Find the value of the product RC for
this system, where .R is the resistance of the medium between the
Problems 137
conductors and C is the mutual capacitance of the conductors in the

presence of the medium.
Solution. Let us mentally supply the charges +q and —q to the
conductors. Since the medium between them is Poorly conducting,
the surfaces of the conductors are equipotential, and the field configu-
ration is the same as in the absence of the medium.
Let us surround, for example, the positively charged conductor by
a closed surface S directly adjoining the conductor and calculate R
and C separately:
X= --= ,
I jn dS o (j) E dS
DT, dS sso dS
C =q = —
U
where the integrals are taken over the given surface S. While calcu-
lating R, we used Ohm's law in the form j = oE, and C was calcula-
ted with the help of the Gauss. theorem.
Multiplying these expressions, we obtain
RC = noglo = coep•
• 5.4. Conditions on the boundary of a conductor. A conductor
with resistivity p borders on a dielectric whose dielectric constant is
8. The electric induction at a certain point A at the surface of the con-
ductor is equal to D, vector D being directed away from the conductor
and forming angle a with the normal to the surface. Find the surface
charge density on the conductor and the current density near the
point A.
Solution. The surface charge density on the conductor is given by
a = DT, = D cos a.
The current density can be found with the help of Ohm's law: j =
= E/p. It follows from the continuity equation (5.5) that the normal
components of vector j are equal, and since in the dielectric in =0
(there is no current), in the conductor we also have im = 0. Hence,
vector j in the conductor is tangent to its surface. The same applies
to vector E inside the conductor.
On the other hand, it follows from the theorem on circulation
of vector E that its tangential components on different sides of the
interface are equal, and hence E=ET =D sin a/BE0, where Et is
the tangential component of vector E in the dielectric. Taking these
arguments into account, we obtain
E D sin a
j= =
8o6P
• 5.5. The gap between the plates of a parallel-plate capacitor is

filled consecutively by two dielectric layers .1 and 2 with thicknesses
and /2, dielectric constants si and 82, and resistivities pl and p2.
The capacitor is under a permanent voltage U, the electric field being
directed from layer 1 to 2. Find the surface density of extraneous

charges at the interface between the dielectric layers.
Solution. The required surface charge density is given by
CY =D2n—Din=eo82E2-8osiEl. (1)
In order to determine E1and E2, we shall make use of two conditions:
since ji = j2, it follows that E/ /pi = E2/p2, and besides E111 E212=
= U. Solving the two equations, we find E1and E2. Substituting these
values into (1), we obtain
o 82132 -8191
s0 U.
Piii+P2/2
Hence it follows that a = 0 for sipi =- 2. 2'
• 5.6. Nonhomogeneous conductor. A long conductor of a cir-
cular cross section of area S is made of a material whose resistivity
depends only on the distance r from the conductor axis as p = a/r2,
where a is a constant. The conductor carries current I. Find (1) field
intensity E in the conductor and (2) the resistance of the unit length
of the conductor.
Solution. (1) In accordance with Ohm's law, field intensity E is
related to current density j, while j is related to current I. Hence we
can write
.c j2nr dr= (Elp)2ar dr.
The field intensity E is the same at all poi is of the cross section of
the given conductor, i.e. is independent of r. We can easily verify
this by selecting a rectangular contour in the conductor so that one
side of the contour coincides with, for example, the conductor's axis,
and then applying to this contour the theorem on circulation of
vector E.
Thus, E can be taken out of the integral, and as a result of inte-
gration we obtain
E = 2nal/S2.
(2) The resistance of a unit length of the conductor can be found
with the help of the formula R = Dividing both sides of this
equation by length 1 of the section of the conductor, having resistance
I? and voltage U, we find
Ru = 2sta/S2.
0 5.7. Ohm's law for nonuniform subcircuit. In the circuit
shown in Fig. 5.14 the e.m.f.'s e and eo of sources, the resistances
R and R 0, and the capacitance C of a capacitor are known. The inter-
nal resistances of sources are negligibly small. Find the charge on
plate 1 of the capacitor.
Solution. In accordance with Ohm's law for a closed circuit con-
Problems 139
taming resistors R and Ro, we can write

(R Ro ) I = 8 — 80,
where the positive direction is chosen clockwise. On the other hand, for
the nonuniform subcircuit aRb of the circuit we have
RI — (pa — (Ph ± 87 R
while for the subcircuil (ICI) we have
+ (p2 — (pi (pb — wa• 6I 1 2
a b
The joint solution of these equations Ro
gives go
R
(P2 R Ro Fig. 5.14
The charge on plate 1 is determined by the formula q1 = C ((pi— cps)•

Therefore, the final result is
RC
R Ro
e0).
It can be seen that q1 >0 for 8 > 8o and vice versa.
5.8. The work of an ean.f. source. A glass plate completely
fills the gap between the plates of a parallel-plate capacitor whose
capacitance is equal to Co when the plate is absent. The capacitor is
connected to a source of permanent voltage U. Find the mechanical
work which must be done against electric forces for extracting the
plate out of the capacitor.
Solution. According to the law of conservation of energy, we can
write
A m + A s = AW, (1)
where A mis the mechanical work accomplished by extraneous forces
against electric forces, A sis the work of the voltage source in this
process, and AW is the corresponding increment in the energy of the
capacitor (we assume that contributions of other forms of energy to
the change in the energy of the system is negligibly small).
Let us find. A W and A s. It follows from the formula W = C U2/2 =
qUI2 for the energy of a capacitor that for U = const
AW = AC U2/2 = Aq U/2. (2)
Since the capacitance of the capacitor decreases upon the removal of
the plate (AC < 0), the charge of the capacitor also decreases (Aq <
< 0). This means that the charge has passed through the source against
the direction of the action of extraneous forces, and the source has
done negative work
A s = Aq•U. (3)
Comparing formulas (3) and (2), we obtain

A s =2A W.
Substitution of this expression into (1) gives
A m = —A W or Am = (8 — 1) COU2.
Thus, extracting the plate out of the capacitor, we (extraneous

forces) do a positive work (against electric forces). The e.m.f. source in
this case accomplishes a negative work, and the energy of the capaci-
tor decreases:
A m > 0, As < 0, AW < O.
• 5.9. Transient processes. A circuit consists of a permanent
source of e.m.f. e, and a resistor R and capacitor C connected in se-
ries. The internal resistance of the source is negligibly small. At the
moment t = 0, the capacitance of the capacitor was abruptly (jump-
wise) decreased by a factor of rl. Find the current in the circuit as a
function of time.
I Solution. We write Ohm's law for the
inhomogeneous part leR2 of the circuit
lC (Fig. 5.15):
RI — (pi— (p2 -e _ u — e.
Considering that U = qIC', where C' =
R = Ch), we obtain
Fig. 5.15 RI = riqIC — e. (1)
We differentiate this equation with respect to time, considering that
in our case (q decreases) dqhlt = —I:
dI 11 AT dI TI
R
n. = -- RC dt.
at C ' I =—
Integration of this equation gives
,
in —= —
Is RC '
I—re-lime
where /0 is determined by condition (1). Indeed, we can write
RI 0 = rig 01C — g,
where q o = eC is the charge of the capacitor before its capacitance
has changed. Therefore,
/0 = (ri — 1) eiR.
• 5.10. A charge qo was supplied to a capacitor with capacitance
C, which at the moment t = 0 was shunted by resistor R. Find
the amount of heat liberated in the resistor as a function of time t.
Solution. The required amount of heat is given by
RI 2 dt, (1)
from which it follows that first of all we must find the time dependence
I (t). For this purpose, we shall use Ohm's law for the part 1R2 of
the circuit (Fig. 5.16):
RI = cpt— cp2 = U,
Or
RI = gIC. (2)
Let us differentiate (2) with respect to
time:
dI 1 r dI dt
R
C I — RC •
R
dt
Integrating the last equation, we obtain
/0e - tIRC, Fig. 5.16
In = (3)
I o RC
where /0 is determined by condition (2) for q = go, i.e. /0 = go/RC.
Substituting (3) into (1) and integrating over time, we get
qg (1_ 2t/RC).
2C e-
6. Magnetic Field in a Vacuum

6.1. Lorentz Force. Field B
Lorentz Force. Experiments show that the force F acting
on a point charge q generally depends not only on the posi-
tion of this charge but also on its velocity v. Accordingly,
the force F is decomposed into two components, viz. the
electric component Fe (which does not depend on the motion
of the charge) and the magnetic component Fm (which
depends on the charge velocity). The direction and magni-
tude of the magnetic force at any point of space depend
on the velocity v of the charge, this force being always
perpendicular to vector v. Besides, at any point, the magne-
tic force is perpendicular to the direction specified at this
point. Finally, the magnitude of this force is proportional
to the velocity component which is perpendicular to this
direction.
142 6. Magnetic Field in a Vacuum .
All these properties of magnetic force can be described

by introducing the concept of magnetic field. Characterizing
this field by vector B which determines the specific direction
at each point of space, we can write the expression for
magnetic force in the form
Fin = q [v x Bl. (6.1)
Then the total electromagnetic force acting on charge q
will be given by
[v x13]. (6. 2)
It is called the Lorentz force. Expression (6.2) is universal:

it is valid for constant as well as for varying electric and
magnetic fields for any velocity v of the charge.
The action of the Lorentz force on a charge can in prin-
ciple be used for determining the magnitudes and directions
of vectors E and B. Hence, the expression for the Lorentz
force can be considered as the definition of electric and
magnetic fields (as was done in the case of electric field).*
It should be emphasized that magnetic field does not act
on an electric charge at rest. In this respect, magnetic field
essentially differs from electric field. Magnetic field acts
only on moving charges.
Vector B characterizes the force acting due to magnetic
field on a moving charge, and hence in this respect it is
an analog of vector E characterizing the force acting due
to electric field.
A distinctive feature of magnetic force is that it is always
perpendicular to the velocity vector of the charge. There-
fore, no work is done over the charge. This means that the
energy of a charged particle moving in a permanent magne-
tic field always remains unchanged irrespective of the
motion of the particle.
In the nonrelativistic approximation, the Lorentz force
(6.2), like any other force, does not depend on the choice
* Several methods of measuring field B have been worked out. In

the final analysis, they all are based on the phenomena that can be
described by Eq. (6.2).
of the (inertial) reference system. On the other hand, the

magnetic component of the Lorentz force varies upon a tran-
sition from one frame of reference to another (because of v).
Therefore, the electric component qE must also change.
Hence it follows that the decomposition of the total force F
(the Lorentz force) into the electric and magnetic compo-
nents depends on the choice of the reference system. Without
specifying the reference system, such a decomposition is
meaningless.
Magnetic Field of a Uniformly Moving Charge. Experi-
ments show that magnetic field is generated by moving
charges (currents). As a result of the generalization of expe-
rimental results, the elementary law defining the field B
of a point charge q moving at a constant nonrelativistic
velocity v was obtained. This law is written in the form*
B = [to q [v X 11 (6.3)
4x r3
where [to is the magnetic constant, the coefficient

µo/4n = 10-7 H/m,
and r is the radius vector from the point charge q to the
point of observation. The tail of the radius vector r is
fixed in a given frame of refe-
rence, while its tip moves with
velocity v (Fig. 6.1). Conse-
quently, vector B in the given
reference frame depends nott
only on the position of the
point of observation but on
time as well.
In accordance with formu- Fig. 6.1
la (6.3), vector B is directed
normally to the plane containing vectors v andir, the rota-
tion around vector v in the direction of vector B forming
* Formula (6.3) is also valid in the case when the charge moves
with an acceleration, but only at sufficiently small distances r from
the charge (so small that the velocity v of the charge does not noti-
ceably change during the time rIc).
144 6. Magnetic Field in a Vacuum
the right-handed system with vector v (Fig. 6.1). It should

be noted that vector B is axial (pseudovector).
The quantity B is called magnetic induction.
Magnetic induction is measured in teslas (T).
The electric field of a point charge q moving at a non-
relativistic velocity is described by the same law (1.2).
Hence expression (6.3) can be written in the form
B = 8o tto [v x El = [v x El/c2, (6.4)
where c is the electrodynamic constant = 1111801o), equal

.v V
1 2q
Fig. 6.2
to the velocity of light in vacuum (this coincidence is not

accidental).
Example. Comparison of forces of magnetic and electric interaction
between moving charges. Let two point charges q of a sufficiently large
mass move in parallel to one another with the same nonrelativistic
velocity v as is shown in Fig. 6.2. Find the ratio between the magnetic
Fm and electric Fe forces acting, for example, from charge 1 on
charge 2.
According to (6.2), Fm = qvB and Fe = qE where v is the velocity
of charge 2 and B and E are the induction of the magnetic and the
intensity of the electric fields created by charge 1 at the point of loca-
tion of charge 2.
The ratio Fm/Fe = vBIE. According to (6.4), in our case B =
vElc2, and hence
Fm/Fe = (v/c)2. (6.5)
Even at sufficiently high velocities, e.g. v = 300 km/s, this ratio

is equal to 10-6, i.e. the magnetic component of the force is a milli-
onth fraction of the electric component and constitutes a negligible
correction to the electric force.
This example may give rise to the following question:
Are such forces worth investigating? It turns out that they
are, and there are two sound reasons behind this.
6.2. The Blot-Savart Law 145
First, we have to deal with beams of particles moving

at velocities close to the velocity of light, for which this
"correction" and the electric force become comparable (it
should be noted that relation (6.5) is also valid for relati-
vistic velocities).
Second, during the motion, say, of electrons, along wires,
their directional velocity amounts to several tens of a mil-
limeter per second at normal densities, while the ratio
(v/c)2 10-24. It is indeed a negligible correction to the
electric force! But as a matter of fact, in this case the magne-
tic force is practically the only force since electric forces
disappeared as a result of an almost ideal balance (much
more perfect than 10-24) of negative and positive charges
in the wires. The participation of a vast number of charges
in creating current compensates for the smallness of this
term.
In other words, the excess charges on the wires are negli-
gibly small in comparison with the total charge of carriers.
For this reason, magnetic forces in this case considerably
exceed the electric forces acting on excess charges of the
wires.
6.2. The Biot-Savart Law

The Principle of Superposition. Experiments show that
magnetic fields, as well as electric fields, obey the principle
of superposition: the magnetic field created by several
moving charges or currents is equal to the vector sum of the
magnetic fields created by each charge or current separately:
B= Bi . (6.6)
The Biot-Savart Law. Let us consider the problem

of determining the magnetic field created by a direct electric
current. We shall solve this problem on the basis of law (6.3)
determining the induction B of the field of a uniformly
moving point charge. We substitute into (6.3) the charge
p dV for q (where dV is the volume element and p is the
volume charge density) and take into account that in accord-
10-0181
ante with (5.2) pv = j. Then formula (6.3) becomes
dB= 1-to [ixr] dV (6.7)

4n r3 •
If the current I flows along a thin wire with the cross-

sectional area AS, we have
j dV j AS dl = I dl,
where dl is an element of the length of the wire. Intro-
ducing vector dl in the direction of the current I, we can
write this expression as follows:
j dV = I dl. (6.8)
Vectors j dV and I dl are called volume and linear current
elements respectively. Replacing in formula (6.7) the vol-
ume current element by the linear one, we obtain
/ [d1 x
dB —
— P.°
r3
(6.9)
Formulas (6.7) and (6.9) express the Biot-Savart law .
In accordance with the principle of superposition the

total field B is found as a result of integration of Eq. (6.7)
or (6.9) over all current elements:
B= [j x dV B = µo I[d1 x r]
4n j r3
(6.10)
4n `.-Y r3
Generally, calculation of the magnetic induction of
a current of an arbitrary configuration by these formulas
is complicated. However, the calculations can be consider-
ably simplified if current distribution has a certain sym-
metry. Let us consider several simple examples of deter-
mining the magnetic induction of current.
Example 1. Magnetic field of the line current, i.e. the current
flowing along a thin straight wire of infinite length (Fig. 6.3).
In accordance with (6.9), vectors dB from all current elements have
at a point A the same direction, viz. are directed behind the plane of
the figure. Therefore, the summation of vectors dB can be replaced by
the summation of their magnitudes dB, where
uo I dl cos a
dB- •
4n rz
6.2. The Btot-Savart Law 147
It is clear from the figure that dl cos a = r da and r = Nees a. Hence

da
dB— N I cos
4n
Integrating this expression over all current elements, which is equiva-
lent to the integration over a between —102 and n/2, we find
110 /
B— (6.11)
4n b •
Example 2. Magnetic field at the axis of circular current. Figu-
re 6.4 shows vector dB from the current element I dl located to the
dB B
I
Fig. 6.3 Fig. 6.4
right. All current elements will form the cone of vectors dB, and it can
be easily seen that the resultant vector B at point A will be directed
upwards along the Z-axis. This means that in order to find the magni-
tude of vector B, it is sufficient to sum up the components of vectors
dB along the Z-axis. Each such projection has the form
[o Idl
dBz = dB cos — cos 13,
4 7.2
where we took into account that the angle between the element dl
and the radius vector r is equal to n12, and hence the sine is equal to
unity. Integrating this expression over dl (which gives 2nR) and taking
into account that cos f3 = 111r and r = (z 2 R2)1127 we get
p., 2nR2/
B= (6.12)
4n (z2 + R2)3/2 •
Hence it follows that the magnitude of vector B at the centre of the
current ring (z = 0) and at a distance z >> R is given by
u,, 2n I luo 2Rflat
Bz=0 =--43.-c - R__ ,Bz.>.:>R — (6.13)
4n z3 •
10*
6.3. Basic Laws of Magnetic Field

Like electric field, magnetic field has two very important
properties. These properties, which are also related to the
flux and circulation of a vector field, express the basic laws
of magnetic field.
Before analysing these laws, we should consider the
graphic representation of field B. Just as any vector field,
field B can be visually represented with the help of the
lines of vector B. They are drawn in a conventional way,
viz. sothat the tangent to these lines at any point coincides
with the direction of vector B, and the density of the
lines is proportional to the magnitude of vector B at a given
point.
The geometrical pattern obtained in this way makes it
possible to easily judge about the configuration of a given
magnetic field and considerably simplifies the analysis of
some situations.
Let us now consider the basic laws of magnetic field,
i.e. the Gauss theorem and the theorem on circulation.
The Gauss Theorem for Field B. The flux of B through
any closed surface is equal to zero:
BdS= O. (6.14)
This theorem is essentially a generalization of experience.

It expresses in the form of a postulate the experimental
result that the lines of vector B have neither beginning
nor end. Therefore, the number of lines of vector B, emerg-
ing from any volume bounded by a closed surface S, is
always equal to the number of lines entering this volume.
Hence follows an important corollary which will be repeat-
edly used below: the flux of B through a closed surface S
bounded by a certain contour does not depend on the shape
of the surface S. This can be easily grasped with the help
of the concept of the lines of vector B. Since these lines
are not discontinued anywhere, their number through the
surface S bounded by a given contour (i.e. the flux of B)
indeed must be independent of the shape of the surf ace S.
Law (6.14) also expresses the fact that there are no magne-
tic charges in nature, on which the lines of vector B begin
6.3. Basic Laws of Magnetic Field 149
or terminate. In other words, magnetic field has no sources,

in contrast to electric field.
Theorem on Circulation of Vector B (for the magnetic
field of a direct current in a vacuum). Circuldtion of vector B
around an arbitrary contour F is equal to the product of t 0
by the algebraic sum of the currents enveloped by the contour I':
B dl = 110/, (6.15)
where I = E/h , and Ih are algebraic quantities. The cur-

rent is assumed positive if its direction is connected with
the direction of the circumven-
tion of the contour through the
right-hand screw rule. The cur-
rent having the opposite direc-
tion is considered to be negative.
This rule is illustrated in Fig.
6.5: here currents /, and I3 are
positive since their directions are
connected with the direction of /1> o
contour circumvention through <0
the right-hand screw rule, while Fig. 6.5
current / 2 is negative.
The theorem on circulation (6.15) can be proved on the
basis of the Biot-Savart law. In the general case of arbitra-
ry currents, the proof is rather cumbersome and will not
be considered here. We shall treat statement (6.15) as
a postulate verified experimentally.
Let us make one more remark. If current I in (6.15) is
distributed over the volume where contour F is located,
it can be represented in the form
j dS. (6.16)
In this expression, the integral is taken over an arbitrary

surface S stretched on contour F. Current density j in
the integrand corresponds to the point where the area
element dS is located, vector dS forming the right-handed
system with the direction of the contour circumvention.
Thus, in the general case Eq. (6.15) can be written a,5
follows:
Bdl=µo dS = 1.1,0 hi dS. (6.17)
The fact that circulation of vector B generally differs
from zero indicates that in contrast to electrostatic field,
field B is not a potential field. Such fields are called
vortex, or solenoidal fields.
Since circulation of vector B is proportional to the cur-
rent I enveloped by the contour, in general we cannot
ascribe to magnetic field a scalar potential which would be
related to vector B through an expression similar to E
—vp. This potential would not be single-valued: upon
each circumvention of the current and return to the initial
point, this potential would acquire an increment equal to
[to/. However, magnetic potential win can be introduced and
effectively used in the region of space where the currents
are absent.
The Role of the Theorem on Circulation of Vector B.
This theorem plays almost the same role as the Gauss
theorem for vectors E and D. It is well known that field B
is determined by all currents, while circulation of vector B,
only by the currents enveloped by the given contour. In
spite of this, in certain cases (in the presence of a special
symmetry) the theorem on circulation proves to be quite
effective since it allows us to determine B in a very simple
way.
This can be done in the cases when the calculation of
circulation of vector B can be reduced, by an appropriate
choice of the contour, to the product of B (or B 1) by the
length of the contour or its part. Otherwise, field B must
be calculated by some other methods, for example, with
the help of the Biot-Savart law or by solving the cor-
responding differential equations, and the calculation be-
comes much more difficult.
6.4. Applications of the Theorem

on Circulation of Vector B
Let us consider several practically important examples
illustrating the effectiveness of the application of the theo-
6.4. Theorem on Circulation of Vector B 151
rem on circulation of vector B and then see whether this

method is universal.
Example 1. Magnetic field of a straight current-. Direct current I
flows along an infinitely long straight wire with a circular cross section
of radius a. Find the magnetic field induction B outside and inside the
wire.
It follows from the symmetry of the problem that the lines of
vector B in this case have the form of circles with the centre at the
Fig. 6.6 Fig. 6.7

wire axis. The magnitude of vector B must be the same at all points
at a distance r from the axis of the wire. Therefore, in accordance with
the theorem on circulation of vector B for a circular contour Pi
(Fig. 6.6), we have B •23-Er = µo/, whence it follows that outside the
wire
B = (A0/2n) Ilr (r > a). (6.18)
It should be noted that a direct solution of this problem (with the help
of the Biot-Savart law) turns out to be more complicated.
From the same symmetry considerations it follows that inside the
wire the lines of vector B are also circles. According to the theorem on
circulation of vector B for a circular contour P2 (see Fig. 6.6),
B•23tr = µo/r, where I,. = I (r1a)2is the current enveloped by the gi-
ven contour. Whence we find that inside the wire
B = (u0/2n) Ir a2 (r < a) (6.19)
The dependence B (r) is shown in Fig. 6.7.
If the wire has the shape of a tube, the induction B outside it is
determined by formula (6.18), while inside the tube magnetic field is
absent. This can also be easily shown with the help of the theorem on
circulation of vector B.
Example 2. Magnetic field of a solenoid. Let current I flow along a
conductor helically wound on the surface of a cylinder. Such cylinder
streamlined by the current is called a solenoid. Suppose that a unit
length of the solenoid contains n turns of the conductor. If the pitch
of the helix is sufficiently small, each turn of the solenoid can be approx-
imately replaced by a closed loop. We shall also assume that the con-
duetor cross-sectional area is so small that the current in the solenoid

can be considered flowing over its surface.
Experiments and calculations show that the longer the solenoid,
the less the magnetic induction outside it. For an infinitely long sole-
noid, magnetic field outside it is absent at all.
It is clear from symmetry considerations that the lines of vector B
inside the solenoid are directed along its axis, vector B forming the
right-handed system with the direction of the current in the solenoid.
4
a
Fig. 6.8
Even the above considerations about the magnetic field configura-

tion of the solenoid indicate that we must choose a rectangular contour
as shown in Fig. 6.8. Circulation of vector B around this contour is
equal to Bl, and the contour envelops the current nil. According to
the theorem on circulation, Bl = Ro n//, whence it follows that inside
a long solenoid
B = Ron/, (6.20)
i.e. the field inside a long solenoid is uniform (with the exception of
the regions adjoining the solenoid's endfaces, which is often ignored
in calculations). The product nI is called the number of ampere-turns.
For n = 2000 turns/m and I = 2 A, the magnetic field inside the so-
lenoid is equal to 5 mT.
Example 3. Magnetic field of a toroid. A toroid is a wire wound
around a torus (Fig. 6.9).
It can be easily seen from symmetry considerations that the lines
of vector B must be circles whose centres lie on the axis 00' of the
toroid. Hence it is clear that for the contour we must take one of such
circles.
If a contour lies inside the toroid, it envelops the current Ni, where
N is the number of turns in the toroidal coil and I is the current
in the wire. Let the contour radius be r; then, according to the theorem
on circulation B.2xtr = RoNI, whence it follows that inside the toroid
B = (R0/231) NI/r. (6.21)
A comparison of (6.21) and (6.18) shows that the magnetic field inside
the toroid coincides with the magnetic field NI of the straight cur-
rent flowing along the 00' axis. Tending N and radius R of the toroid
6.4. Theorem on Circulation of Vector B 153
to infinity (at a constant toroid cross section), we obtain in the limit

the expression (6.20) for the magnetic field of an infinitely long so-
lenoid.
If the chosen circular contour passes outside the toroid, it does
not envelop any currents, and hence B •2:tr = 0 for such a contour.
This means that outside the toroid magnetic field is absent.
In the above considerations it was assumed that the lines of cur-
rent lie in meridional planes, i.e. the planes passing through the axis
00' of the toroid. In real toroids, the lines of cur-
rent (turns) do not lie strictly in these planes, and
hence there is a current component around
the 00' axis. This component creates an addi-
tional field similar to the field of a circular cur-
rent.
Example 4. Magnetic field of a current-car-
rying plane. Let us consider an infinite conduct-
ing plane with undirectional current uniformly
distributed over its surface. Figure 6.10 shows
the cross section of such a plane when the cur-
rent flows behind the plane of the figure (this
is marked by crosses). Let us introduce the con-
cept of linear current density i directed along the
lines of current. The magnitude of this vector Fig. 6.10
is the current per unit length which plays the
role of the "cross-sectional area".
Mentally dividing the current-carrying plane into thin current
filaments, we can easily see that the resultant field B will be directed
parallel to the plane, downwards to the right of the plane and upwards
to the left of it (Fig. 6.10). These directions can be easily established
with the help of the right-hand screw rule.
In order to determine induction B of the field, we shall use the
theorem on circulation of vector B. Knowing the arrangement of the
lines of vector B, we shall choose the contour in the form of rectangle
1234 (Fig. 6.10). Then, in accordance with the theorem on circulation,
2B1 = Roil, where / is the length of the contour side parallel to the
conducting plane. This gives
1
B= soi. (6.22)
2
This formula shows that the magnetic field is uniform on both sides
of the plane. This result is also valid for a bounded current-carrying
plane, but only for the points lying near the plane and far from its
ends.
General Considerations. The results obtained in the above
examples could be found directly with the help of the
Biot-Savart law. However, the theorem on circulation
makes it possible to obtain these results much more simply
and quickly.
However, the simplicity with which the field was cal-
culated in these examples must not produce an erroneous
impression about the potentialities of the method based

on the application of the theorem on circulation. Just as
in the case of the Gauss theorem for electric field, the num-
ber of problems that can be easily solved by using the theo-
rem on circulation of vector B -is quite limited. It is suffi-
cient to say that even for such a symmetric current configu-
ration as the current ring, the theorem on circulation be-
comes helpless. Despite an apparently high symmetry, the
magnetic field configuration does not allow us, however,
to find a simple contour required for calculations, and the
problem has to be solved by other, much more cumber-
some, methods.
6.5. Differential Forms of Basic Laws

of Magnetic Field
Divergence of field B. The differential form of the Gauss theorem
(6.14) for field B can be written as
V • B —0, (6.23)
i.e. the divergence of field B is equal to zero everywhere. As was men-
tioned above, this means that magnetic field has no sources (magnetic
charges). Magnetic field is generated by electric currents and not by
magnetic charges which do not exist in nature.
Law (6.23) is of fundamental nature: it is valid not only for con-
stant but for varying fields as well.
Curl of field B. The important property of magnetic field expressed
by the theorem on circulation of vector B motivates the representation
of this theorem in the differential form which broadens its potential-
ities as a tool for investigations and calculations.
For this purpose, we consider the ratio of circulation of vector B
to the area S bounded by a contour. It turns out that this ratio tends
to a certain limit as S 0. This limit depends on the contour orienta-
tion at a given point of space. The contour orientation is specified by
vector n (the normal to the plane of the contour around which the
circulation is calculated) the direction of n being connected with the
direction of contour circumvention through the right-hand screw rule.
The limit obtained as a result of this operation is a scalar quantity
which behaves as the projection of a certain vector onto the direction
of the normal n to the plane of the contour around which the circula-
tion is taken. This vector is called the curl of field B and is denoted as
curl B. Thus, we can write
B dl
lim (curl B),, (6.24)
S-0 S
where on the right-hand side we have the projection of curl B onto

normal n.
Hence, each point of vector field B can be characterized by curl B
whose direction and magnitude are determined by the properties of
the field itself at a given point. The direction of curl B is determined
by the direction of normal n to the surface element S, for which the
quantity (6.24), which defines the magnitude of curl B, attains its
maximum value.
In mathematics, curl B is obtained in coordinate representa-
tion. However, for our purposes another fact is more important: it
turns out that formally curl B can be considered as the cross product
of the operator V by vector B, i.e. V X B. We shall be using the
latter, more convenient notation since it allows us to represent the
cross product V X B with the help of the determinant:
ex ey ez
V X B= d/ax a/ay alaz (6.25)
Bx B y B z
where ex, ey, e, are the unit vectors of Cartesian coordinates. This
expression is valid for the curl of not only field B, but of any other
vector field also, in particular, of field E.
Let us now consider the theorem on circulation of vector B. Accord-
ing to (6.24), Eq. (6.17) can be represented in the form
B dl
I im = 1101n
S
or (V X B)„ = Rojn. Hence
V X B= tto j. (6.26)
This is precisely the differential form of the theorem on circulation
of vector B. Obviously, the direction of curl B coincides with that of
vector j, viz. the current density at this point, while the magnitude
of V X B is equal to [to/.
In an electrostatic field, circulation of vector E is equal to zero,
and hence
V xE=0. (6.27)
A vector field whose curl is equal to zero everywhere is a potential
field. Otherwise it is called the solenoidal field. Consequently, electro-
static field is a potential field, while magnetic field is a solenoidal field.
6.6. Ampere's Force

Ampere's Law. Each charge carrier experiences the
action of a magnetic force. The action of this force is trans-
mitted to a conductor through which charges are moving.
As a result, magnetic field acts with a certain force on a cur-

rent-carrying conductor. Let us find this force.
Suppose that the volume density of charges which are
the carriers of current (e.g. electrons in a metal) is equal
to p. We mentally isolate a volume element dV in the
conductor. It contains the charge (current carrier) equal
to p dV. Then the force acting on the volume element dV
of the conductor can be written by formula (6.1) as follows:
dF = p Eu x 131 dV.
Since j = pu, we have
dF x 13] dV. (6.28)
If the current flows along a thin conductor, then, ac-

cording to (6.8), j dV = I dl, and
dF I [d1 X B], (6.29)
where dl is the vector coinciding in direction with the

current and characterizing an element of length of the thin
conductor.
Formulas (6.28) and (6.29) express Ampere's law. Inte-
grating these expressions over the current elements (volume
or linear), we can find the magnetic force acting on a cer-
tain volume of a conductor or on its linear part.
The forces acting on currents in a magnetic field are
called Ampere's forces.
Example. The force of interaction between parallel currents.
Find the Ampere's force with which two infinitely long wires with
currents Il and /2 interact in a vacuum, if the distance between them
is b. Calculate the force per unit length of the system.
Each current element of 12 is in the magnetic field of the current /1,
i.e. according to (6.19), the field B1 =(p4/41i) 211/b. The angle be-
tween the element of current 72 and vector B1is 7112. Hence, it follows
from formula (6.29) that the force acting per unit length of the con-
ductor with current /2 is Fu = 12B1, or
11,0 2/112
F — (6.30)
u 431 b
Naturally, the same expression can be obtained for the force acting
per unit length of the conductor with current I.
Finally, it can be easily seen that currents of the same

direction are attracted, while the opposite currents repell
each other. Here we speak only about a magnetic force.
It should be recalled, however, that besides magnetic forces
there are electric forces, i.e. the forces due to excess charges
on the surface of conductors. Consequently, if we speak about
the total force of interaction between the wires, it can either
be attractive or repulsive, depending on the ratio between
the magnetic and electric components of the total force
(see Problem 6.7).
The Force Acting on a Current Loop. The resultant Ampe-
re's force acting on a current loop in a magnetic field is
defined, in accordance with (6.29), by
F = / [di x B], (6.31)
where the integration is performed along the given loop
with current I.
If the magnetic field is uniform, vector B can be taken
out of the integral, and the problem is reduced to the calcu-
lation of the integral dl. This integral is a closed chain
of elementary vectors dl and hence is equal to zero. Conse-
quently, F = 0, i.e. the resultant Ampere's force in a uni-
form magnetic field is equal to zero.
If, however, the magnetic field is nonuniform, the resul-
tant force (6.31) generally differs from zero and in each
concrete case is determined with the help of (6.31). The
most interesting case for further analysis is that of a plane
current loop of sufficiently small size. Such a current loop
is called elementary.
The behaviour of an elementary current loop can be
conveniently described with the help of magnetic moment pm.
By definition, magnetic moment pm is given by
pm = ISn, (6.32)
where I is the current, S is the area bounded by the loop
and n is the normal to the loop whose direction is connected
with the direction of the current in the loop through the
right-hand screw rule (Fig. 6.11). In terms of magnetic field,
the elementary loop is completely characterized by its
magnetic moment pm.
An involved calculation with the help of formula (6.31),

taking into account the small size of the loop, yields the
following expression for the force acting on an elementary
current loop in a nonuniform magnetic field:
OB
F = Pm a n 9 (6.33)
where p,„ is the magnitude of the magnetic moment of the

loop, and OB/an is the derivative of vector B along the
F
Jo ® Pm
Jo Pm
F
Fig. 6.11 Fig. 6.12
direction of normal n or along the direction of vector pm.

This formula is similar to expression (1.39) for the force
acting on an electric dipole in an electric field.
It follows from formula (6.33) that, like in the case of
electric dipole,
(1) in a uniform magnetic field F = 0 since OB/On = 0;
(2) the direction of vector F generally does not coincide
with vector B or with vector pm; vector F coincides in
direction only with the elementary increment of vector B
taken in the direction of vector pm at the locus of the loop.
This is illustrated by Fig. 6.12, where three arrangements
of the loop in the magnetic field of straight current /0
are shown. The vector of the resultant force F acting on the
loop in each case is also shown in the figure (it is useful
to verify independently that it is really so).
If we are interested in the projection of force F onto
a certain direction X, it is sufficient to write expression
6.7. Torque Acting on a Current Loop 159
(6.33) in terms of the projections onto this direction. This

gives
oBx
F = Pin on (6.34)
where 013„lan is the derivative of the corresponding pro

jection of vector B again with re-
spect to normal n (or with respect
to pm) to the loop.
Example. An elementary current loop
having a magnetic moment pm is ar-
ranged perpendicularly to the symmetry
axis of a nonuniform magnetic field, vec-
tor pmbeing directed along vector B. Let
us choose the positive direction of the
1C-axis as is shown in Fig. 6.13. Since
the increment of projection Bx will be Fig. 6.13
negative in the direction of vector pm,
Fs <0. Hence vector F is directed to
the left, i.e. to the side where B is greater. If we rotate the cur-
rent loop (and vector pm) through 90° so that the loop centre coin-
cides with the symmetry axis of field B, in this position Fx --- 0, and
vector F is directed perpendicularly to the X-axis and to the same
side as vector pm.
6.7. Torque Acting on a Current Loop

Let us consider a plane loop with a current I in a uniform
magnetic field B. It was shown above (see p. 157) that
the resultant force (6.31) acting on the current loop in
a uniform magnetic field is equal to zero. It is known from
mechanics that if the resultant of the forces acting on any
system is equal to zero, the resultant moment of these forces
does not depend on the position of point 0 relative to which
the torques are determined. Therefore, we can speak about
the resultant moment of Ampere's forces in our case.
By definition, the resultant moment of Ampere's forces
is given by
M = [r x dF], (6.35)
where dF is defined by formula (6.29). If we perform cal-

culation by formula (6.35) (it is cumbersome and hardly
interesting, hence we omit it here), it turns out that the
16,1 6. Magnetic Field in a Vacuum
moment of forces for an arbitrary shape of the loop can

be represented in the form
= [pm x B], (6.36)
where pm is the magnetic moment of the current loop (pm =

= ISn for a plane loop).*
B It is clear from (6.36) that the mom-
ent M of Ampere's forces acting on
a current loop in a uniform magne-
pm field is perpendicular both to vec-
tor pm and to vector B. The mag-
nitude of vector M is M= p m B sin a,
where a is the angle between vec-
tors pm and B. When pm B, M = 0
and it is not difficult to see that the
Fig. 6.14 position of the loop is stable. When
pm B, M also equals zero but such
a position of the loop is unstable: the slightest deviation
from it leads to the appearance of the moment of force that
tends to deviate the loop from the initial position still
further.
Example. Let us verify the validity of formula (6.36) by using

a simple example of a rectangular current loop (Fig. 6.14).
It can be seen from the figure that the forces acting on sides a are
perpendicular to them and to vector B. Hence these forces are directed
along the horizontal (they are not shown in the figure) and only strive
to stretch (or compress) the loop. Sides b are perpendicular to B,
and hence each of them is acted upon by the force
F = IbB.
These forces tend to rotate the loop so that vector pmbecomes directed
like vector B. Hence, the loop is acted upon by a couple of forces
whose torque is equal to the product of the arm a sin a of the couple
by the force F, i.e.
M = IbBa sin a.
* If the loop is not plane, its magnetic moment is pm = 1"f dS,

where the integral is taken over the surface S stretched over the cur-
rent loop. This integral does not depend on the choice of the surface S
and depends only on the loop over which it is stretched.
6.8. Work Done upon Displacement of Current Loop 161
Considering that ab is the area bounded by the loop and Iba = pm,
we obtain
M = pmB sin a,
which in the vector form is written as (6.36).
Concluding this section, we note that expression (6.36)
is valid for nonuniform magnetic fields as well. The only
requirement is that the size of the current loop must be
sufficiently small. Then the effect of nonuniformity on the
rotational moment (torque) M can be ignored. This pre-
cisely applies to the elementary current loop.
An elementary current loop behaves in a nonuniform
magnetic field in the same way as an electric dipole in
an external nonuniform electric field: it will be rotated
towards the position of stable equilibrium (for which
pm tt B) and, besides, under the action of the resultant
force F it will be pulled into the region where induction B
is larger.
6.8. Work Done upon Displacement of Current Loop

When a current loop is in an external magnetic field (we
shall assume that this field is constant), certain elements
of the loop are acted upon by Ampere's forces which hence
will do work during the displacement of the loop. We shall
show here that the work done by Ampere's forces upon an
elementary displacement of the
I
loop with current I is given by
SA = /c/(13, (6.37)
where dcll) is the increment of the
magnetic flux through the loop
upon the given displacement.
We shall prove this formula Fig. 6.15
in three stages.
1. Let us first consider a particular case: a circuit (Fig. 6.15)
with a movable jumper of length 1 is in a uniform magnetic
field perpendicular to the circuit plane and directed behind
the plane of the figure. In accordance with (6.29), the jumper
is acted upon by Ampere's force F = IlB . Displacing the
jumper to the right by dx, this force accomplishes positive
11-0181
work
SA = F dx = IB1 dx = IB dS, (6.38)
where dS is the increment of the area bounded by the con-
tour.
In order to determine the sign of magnetic flux 0, we
shall agree to take the normal n to the surface bounded by
the contour always in such a way that it forms with the
direction of the current in the circuit a right-handed system
(Fig. 6.15). In this case, the current I will always be a posi-
tive quantity. On the other hand, the flux 0 may be either
positive or negative. In our case, however, both 0 and
d0 = B dS are positive quantities (if the field B were
directed towards us or the jumper were displaced to the
left, d'D < 0). In any of these cases expression (6.38) can
be represented in the form (6.37).
2. The obtained result is valid for an arbitrary direction
of the field B as well. To prove this, let us decompose vec-
tor B into three components: B = Bn + B1 Bx. The
component 131 directed along the jumper is parallel to the
current in it and does not produce any force acting on the
jumper. The component 13, (along the displacement) is
responsible for the force perpendicular to the displacement,
which does not accomplish any work. The only remaining
component is Bn, which is normal to the plane in which
the jumper moves. Hence, in formula (6.38) instead of B
we must take only B,.. But B, dS = c/0, and we again
arrive at (6.37).
3. Let us now consider any current loop which is arbi-
trarily displaced in a constant nonuniform magnetic field
(the contour of this loop may be arbitrarily deformed in
this process). We mentally divide the given loop into infinite-
ly small current elements and consider their infinitesimal
displacements. Under these conditions, the magnetic field
in which each element of current is displaced can be as-
sumed uniform. For such a displacement, we can apply to
each element of current the expression dA =I d'0 for the
elementary work, where d'D expresses the contribution of
a given element of current to the increment of the flux
through the contour. Summing up elementary works for
all the elements of the loop, we again obtain expression
Problems 163
(6.37), where c/(1) is the increment of the magnetic flux

through the entire circuit.
In order to find the work done by Ampere's forces upon
the total displacement of the current loop from the initial
position 1 to the final position 2, it is sufficient to integrate
expression (6.37):
A= I d(1). (6.39)
If the current I is maintained constant during this dis-

placement we obtain
A = /(0 2 — (6.40)
where (I), and 02 are magnetic fluxes through the contour
in the initial and final positions. Thus, the work of Ampere's
forces in this case is equal to the product of the current
by the increment of the magnetic flux through the circuit.
Expression (6.40) gives not only the magnitude but also
the sign of the accomplished work.
Example. A plane loop with current I is rotated in magnetic
field B from the position in which n 1f B to the position in which
n f t B, n being the normal to the loop (it should be recalled that the
direction of the normal is connected with the direction of the current
through the right-hand screw rule). The area bounded by the loop is S.
Find the work of Ampere's forces upon such a displacement, assuming
that the current I is maintained constant.
In accordance with (6.40), we have
A = I [BS — ( BS)] = 2IBS.
—
In the given case, the work A > 0, while upon the reverse rotation
A < 0.
It should be noted that work (6.40) is accomplished not at the
expense of the energy of the external magnetic field (which does not
change) but at the expense of the e.m.f. source maintaining the current
in the loop. (This question will be considered in greater detail in
Chap. 9.)
Problems
• 6.1. Direct calculation of induction B. Current I flows along
a thin conductor bent as is shown in Fig. 6.16. Find the magnetic
induction B at point 0. The required data are presented in the figure.
Solution. The required quantity B = B_ B_, where B_ is the
11*
magnetic field created by the rectilinear part of the loop and

by its curvilinear part. According to the Biot-Savart law (see Example 1
on p. 146), we have
ao
110 / cos a da
B_= 2 tg
4xa cos a0 — 2xa a°,
0
I (2n — 2a0) a
B. = (I —ao).
4n a2 2xa
As a result, we obtain
B = (x — a, tg ao) Ro//2na.
It is interesting to show that for a, 0, we arrive at the familiar
expression (6.13).
Fig. 6.16 Fig. 6.17
• 6.2. A thin isolated wire forms a plane spiral consisting of

a large number N of closely packed turns through which a direct cur-
rent I is flowing. The radii of the internal and external loops are a
and b (Fig. 6.17). Find (1) the magnetic induction B at the centre
of the spiral (point 0), and (2) the magnetic moment of the spiral for
the given current.
Solution. (1) According to (6.13), the contribution of one turn of
radius r to the magnetic induction is equal to
B1 = tt 0 //2r, (1)
and of all the turns,
B= dN, (2)
where dN is the number of turns in the interval (r, r+ dr),

N
dN= b— a dr . (3)
Problems 165
Substituting (1) and (3) into (2) and integrating the result over r
between a and b, we obtain
b
B =2 In
(b— a) a•
(2) The magnetic moment of a turn of radius r is Pmi
and of all the turns, pm = Pmi dN, here dN is defined by formula
(3). The integration gives
nIN b3— a3 TIN
Pm= b— a (a2 + ab b2).
3 3
• 6.3. Current I flows in a long straight conductor having thn
shape of a groove with a cross section in the form of a thin half-ring
Fig. 6.18 Fig. 6.19
of radius R (Fig. 6.18). The current is directed from the reader behind
the plane of the drawing. Find the magnetic induction B on the axis 0.
Solution. First of all, let us determine the direction of vector B
at point 0. For this purpose, we mentally divide the entire conductor
into elementary filaments with currents dI. Then it is clear that the
sum of any two symmetric filaments gives vector dB directed to the
right (Fig. 6.19). Consequently, vector B will also be directed to the
right.
Therefore, for calculating the field B at point 0 it is sufficient to
find the sum of the projections of elementary vectors dB from each
current filament onto the direction of vector B:
B= c dB sin cp (I)
In accordance with (6.11), we have
dB = (2)
where dl = (IITc) dcp (see Fig. 6.19). Substituting (2) into (1), we get
n
B
2n 2R , sin (I) cl(P= '°I •
0
6.4. Theorem on circulation of B and principle of superposi-

tion. Inside a homogeneous long straight wire of circular cross section,
there is a circular cylindrical cavity whose axis is parallel to the con-
ductor axis and displaced relative to it by a distance 1. A direct cur-
rent of density j flows along the wire. Find magnetic induction B
inside the cavity.
Solution. In accordance with the principle of superposition, the
required quantity can be represented as follows:
B = Bo — B', (1)
where Bo is the magnetic induction of the conductor without cavity,
Fig. 6.20 Fig. 6.21
while B' is the magnetic induction of the field at the same point due
to the current flowing through the part of the conductor which has
been removed in order to create the cavity.
Thus, the problem implies first of all the calculation of magnetic
induction B inside the solid conductor at a distance r from its axis.
Using the theorem on circulation, we can write 2TfrB = [toar2j, whence
B = (1/2) Rorj. This expression can be represented with the help of
Fig. 6.20 in the vector form:
1
[j X r].
Using now this formula for Bo and B', we find their difference (1):
1-to [j xrj—±—
21L )-[jxr1=_7_[ix(r---111.
Figure 6.21 shows that r = I r', whence r — r' = 1, and

1
B----7141i X 11.
Thus, in our case the magnetic field B in the cavity is uniform, and
if the current is flowing towards us (Fig. 6.21), the field B lies in the
plane of the figure and is directed upwards.
Problems 167
• 6.5. Principle of superposition. Current I flows through a long

solenoid, whose cross-sectional area is S and the number of turns n.
Find the magnetic flux through the endface of this solenoid.
Solution. Let the flux of B through the solenoid endface be O. If
we place another similar solenoid in contact with the endface of our
(a) (b )
Fig. 6.22 Fig. 6.23

solenoid, the flux through the contacting endfaces will be cri cro —
= cDo, where (Do is the flux through the solenoid's cross section far
from its endface. Then we have
01:1 = 00/2 = pion/S/2.
By the way, we must pay attention to the following propertie,
of the field B at the endface of a long solenoid.
1. The lines of B are arranged as shown in Fig. 6.22. This can be
easily shown with the help of superposition principle: if we place on
the right one more solenoid, field B outside the thus formed solenoid
must vanish, which is possible only with the field configuration shown
in the figure.
2. From the principle of superposition, it follows that the normal
component B, will be the same over the area of the endface, since when
we form a composite solenoid, B, Bn = Bo, where Bo is the field
inside the solenoid away from its endfaces. At the centre of the endface
B = B0, and we obtain B = B0/2.
* 6.6. The field of a solenoid. The winding of a long solenoid of
radius a is a thin conducting band of width h, wound in one layer prac-
tically without gaps. Direct current I flows along the band. Find
magnetic field B inside and outside the solenoid as a function of dis-
tance r from its axis.
Solution. The vector of linear current density i can be represented
as the sum of two components:
The meaning of vectors i1and ill is clear from Fig. 6.23b. In our
case, the magnitudes of these vectors can be found with the help of
Fig. 6.23a by the formulas

= i cos a = i 1/1— sin2 a = (I Ih) (h/2na)2.
i sin a =
The magnetic induction B inside the solenoid is determined, in

accordance with (6.20), by the quantity /1, while outside the solenoid,
by in:
Bi =µoi l = (kith) 171—(11/22-ca) 2 (r < a),
Ba =kili alr =110//2ar (r > a),
where we used the theorem on circulation while calculating Ba out-
side the solenoid: 2arBa =
Thus, having represented the current in the solenoid as the super-
position of "transverse" and "longitudinal" components, we arrive at
the conclusion that only the longitudinal component of the field B
exists inside such a solenoid and only the transverse component out-
side it (as in the case of a straight current).
Besides, if we decrease the band width maintaining the current
density unchanged, I —. 0 as h –›- 0, but //h = const. In this case,
only the field inside the solenoid remains, i.e. the solenoid becomes
"ideal".
• 6.7. Interaction of parallel currents. Two long wires with
negligible resistance are shunted at their ends by resistor B and at
the other ends are connected to a
+ + + + + source of constant voltage. The
radius of the cross section of each
wire is smaller than the distance
between their axes by a factor of
rl = 20. Find the value of the re-
sistance B at which the resultant
force of interaction between the
wires vanishes.
Fig.
Solution. There are excess sur-
face charges on each wire (irrespec-
tive of whether or not the current is flowing through them) (Fig. 6.24).
Hence, in addition to the magnetic force Fm, we must take into account
the electric force Fe. Suppose that an excess charge X corresponds to
a unit length of the wire. Then the electric force exerted per unit
length of the wire by the other wire can be found with the help of the
Gauss theorem:
1 2X 2X2
Fe =2■,E
4xe0 / = 4nso / '
where 1 is the distance between the axes of the wires. The magnetic
force acting per unit length of the wire can be found with the help
Problems 169
of the theorem on cirulation of vector B:

Fm = (p,0/4a) 2/211,
where I is the current in the wire.
It should be noted that the two forces, electric and magnetic, are
directed oppositely. The electric force is responsible for the attraction
between the wires, while the magnetic force causes their repulsion.
Let us find the ratio of these forces:
Fm/Fe = op,o/2/X2. (1)
There is a certain relation between the quantities I and X (see
Problem 2.8):
X= C1U— 1n£6 (2)
where U = RI. Hence it follows from relation (2) that

//X = In rihnoR. (3'
Substituting (3) into (1), we obtain
Fm [10 1n2 11
(4)
Fe eo n2R2 •
The resultant force of interaction vanishes when this ratio is equal
to unity. This is possible when R = R e, where
R 0= V go 11111 — 360 Q.
go a
If R < 131) then Fm > Fe, and the wires repel each other. If, on
the contrary, R > R 0, Fm < Fe, and the wires attract each other.
This can be observed experimentally.
Thus, the statement that current-carrying wires attract each other
is true only in the case when the electric component of the interac-
tion can be neglected, i.e. for a sufficiently small resistance R in the
circuit shown in Fig. 6.24.
Besides, by measuring the force of interaction between the wires
(which is always resultant), we generally cannot determine current I.
This should be borne in mind to avoid confusion.
• 6.8. The moment of Ampere's forces. A loop with current I
is in the field of a long straight wire with current / 0 (Fig. 6.25). The
plane of the loop is perpendicular to the straight wire. Find the moment
of Ampere's forces acting on this loop. The required dimensions of the
system are given in the figure.
Solution. Ampere's forces acting on curvilinear parts of the loop
are equal to zero. On the other hand, the forces acting on the rectilinear
parts form a couple of forces. We must calculate the torque of this
couple.
Let us isolate two small elements of the loop (Fig. 6.26). It can be
seen from the figure that the torque of the couple of forces correspond
ing to these elements is

dM = 2x tan cp dF, (1)
where the elementary Ampere's force is given by
dF = I dlB. (2)
The dependence of the magnetic induction B on the distance r
from the straight wire can be found with the help of the theorem on
circulation:
B = ji0//2otr. (3)
Let us now substitute (3) into (2), then (2) into (1) and, conside-
ing that dl = dr and x = r cos cp, integrate the obtained expression
over r between a and b. This gives
M = (10/n) II o (b — a) sin cp,
vector M being directed to the left (Fig. 6.26).
6.9. A small coil with current, having the magnetic moment
pm, is placed on the axis of a circular loop of radius R, along which
current I is flowing. Find the force F acting on the coil if its distance
from the centre of the loop is / and vector pm is oriented as is shown
in Fig. 6.27.
Solution. According to (6.33), the required force is defined as
F = pm013/On, (1)
where B is the magnetic induction of the field created by the loop
at the locus of the coil. Let us select Z-axis in the direction of the vec-
tor pm. Then the projection of (1) onto this axis will be
F = pm aBz Iaz = pm0.1310z,
where we took into account that Bz = B for the given direction of
current in the loop. The magnetic induction B is defined by formula
(6.12), whence
OB 3 RoR2//
Oz = 2 (0+//95/2'
Problems 171
Since 8B/oz < 0, the projection of the force F2 <0, i.e. vector F
is directed towards the loop with current I. The obtained result can
be represented in the vector form as follows:
3 p,OR2//
F=
2 (12 + R2)5/2 pm' Pm
It should be noted that if pm(and hence

Z-axis as well) were directed oppositely, then
B5 = — B and 0./35/az > 0, and hence F5 >0
and vector F would be directed to the right, Fig. 6.27
i.e. again oppositely to the direction of pm.
Hence, the obtained direction for F is valid for both orientations of pm
• 6.10. Current I flows in a long thin-walled circular cylinder
of radius R. Find the pressure exerted on the cylinder walls.
Solution. Let us consider a surface current element i dS, where
i is the linear current density and dS is the surface element. We shall
dB
g
dS 2
..e0.-•=1**
Bi
Fig. 6.28 Fig. 6.29
find the relation between the surface and volume elements of the cur-
rent:
j dV = j6li•61) dl = i dS.
The meaning of the quantities appearing in this relation is clarified
in Fig. 6.28. In the vector form, we can write
j dV = i dS. (1)
The Ampere's force acting on the surface current element in this
case is determined by the formula obtained from (6.28) with the help
of substitution (1):
dF = X B'] dS, (2)
where B' is the magnetic induction of the field at the point of location
of the given current element, but created by all other current elements
excluding the given one. In order to find B', we proceed as for cal-
culating electric force (Sec. 2.3). Let Bi be the magnetic induction
172 7. Magnetic Field in a Substance
of the field created by the surface element itself at a point very close
to its surface (see Fig. 6.29, where it is assumed that the current is
flowing from us). According to (6.22), Bi = (1/2) tio i.
Further, using the theorem on circulation of vector B and symmetry
considerations, we can easily see that the magnetic induction of the
field outside the cylinder near its surface is equal to
B = pio//2aR, (3)
while inside the cylinder the field is absent.
The latter fact indicates that the field B' from the remaining cur-
rent elements at points 1 and 2 very close to the cylinder surface (see
Fig. 6.29) must be the same and satisfy the following conditions inside
and outside the cylinder surface:
B' = Bi and B = B' Bi = 2B'.
Hence it follows that
B' = B/2. (4)
Substituting this result into (2), we obtain the following expression
for the required pressure:
dF 2B' B2
B'
19 = dS =119, = 2R0
Taking into account (3), we finally get
P = ito/2/8n2R2.
It is clear from formula (2) that the cylinder experiences lateral com-
pression.
7. Magnetic Field in a Substance

7.1. Magnetization of a Substance.
Magnetization Vector J
Field in Magnetics. If a certain substance is introduced
into a magnetic field formed by currents in conductors,
the field will change. This is explained by the fact that each
substance is a magnetic, i.e. it is magnetized (acquires
a magnetic moment) under the action of magnetic field.
A magnetized substance creates its own magnetic field B',
which forms, together with the primary field B, created
by conduction currents, the resultant field
B = B, + B'. (7.1)
Here B' and B stand for the fields averaged over a
physical infinitely small volume.
7.1. Magnetization of a Substance. Vector J 173
The field B', like the field B, of conduction currents,

has no sources (magnetic charges). Hence, for the resultant
field B in the presence of a magnetic, the Gauss theorem
is applicable:
(, , BdS=0. (7.2)
This means that the field lines of B remain continuous

everywhere in the presence of a substance.
Mechanism of Magnetization. At present, it is estab-
lished that molecules of many substances have intrinsic
magnetic moments due to the motion of intrinsic charges
in them. Each magnetic moment corresponds to a circular
current creating a magnetic field in the surrounding space.
In the absence of external magnetic field, the magnetic
moments of molecules are oriented at random, and hence
the resultant magnetic field due to these moments is equal
to zero, as well as the total magnetic moment of the sub-
stance. This also applies to the substances that have no
magnetic moments in the absence of an external field.
If a substance is placed into an external magnetic field,
under the action of this field the magnetic moments of the
molecules acquire a predominant orientation, and the sub-
stance is magnetized, viz. its resultant magnetic moment
becomes other than zero. In this case, the magnetic mo-
ments of individual molecules no longer compensate each
other, and as a result the field B' appears.
Magnetization of a substance whose molecules have no
magnetic moments in the absence of external field proceeds
differently. When such materials are introduced into an
external field elementary circular currents are induced
in the molecules, and the entire substance acquires a mag-
netic moment, which also results in the generation of the
field B'.
Most of materials are weakly magnetized when introduced
into a magnetic field. Only ferromagnetic materials such
as iron, nickel, cobalt and their many alloys have clearly
pronounced magnetic properties.
Magnetization. The degree of magnetization of a magnet-
ic is characterized by the magnetic moment of a unit
volume. This quantity is called magnetization and denoted

by J. By definition,
= I NI AV Pm,
(7.3)
where AV is a physical infinitely small volume surround-

ing a given point and pm is the magnetic moment of an
individual molecule. The summation is performed over
all molecules in volume AV.
By analogy with what was done for polarization P
[see (3.3)1, magnetization can be represented in the form
J = n (pm ), (7.4)
where n is molecular concentration and (pm )) is the average
magnetic moment of a molecule. This formula shows that
vector J is collinear precisely to the average vector (pm ).
Hence for further analysis it is sufficient to know the behav-
iour of vector (pm ) and assume that all the molecules
within the volume AV have the same magnetic moment
(pm ). This will considerably simplify the understanding of
questions associated with the phenomenon of magnetiza-
tion. For example, an increase in magnetization J of a mate-
rial implies the corresponding increase in vector (pm ): if
J = 0, (pm )) = 0 as well.
If vector J is the same at each point of a substance, it is
said that the substance is uniformly magnetized.
Magnetization Currents I'. As was mentioned above, mag-
netization of a substance is caused by the preferential orienta-
tion of magnetic moments of individual molecules in one
direction. The same can be said about the elementary circular
currents associated with each molecule and called molecular
currents. It will be shown that such a behaviour of molecular
currents leads to the appearance of macroscopic currents I'
called magnetization currents. Ordinary currents flowing in
the conductors are associated with the motion of charge
carriers in a substance and are called conduction currents (I).
In order to understand how magnetization currents appear,
let us first imagine a cylinder made of a homogeneous mag-
netic whose magnetization J is uniform and directed along
the axis. Molecular currents in the magnetized magnetic
are oriented as shown in Fig. 7.1. Molecular currents of
7.1. Magnetization of a Substance. Vector J. 175
adjacent molecules at the points of their contact have op-

posite directions and compensate each other. The only uncom-
pensated molecular currents are those emerging on the lateral
surface of the cylinder. These currents form macroscopic
surface magnetization current I' circulating over the lateral
ri
f f
(mos
•Ake
111+0+0+0
• 0+0+0
9006
+
Fig. 7.1 Fig. 7.2
surface of the cylinder. The current I' induces the same mac-
roscopic magnetic field as that of all molecular currents
taken together.
Let us now consider another case: a magnetized magnetic
is nonhomogeneous. Let, for instance, the molecular currents
be arranged as shown in Fig. 7.2, where the line thickness
corresponds to the intensity of molecular currents. It follows
from the figure that vector J is directed behind the plane
of the figure and increases in magnitude with the coordi-
nate x. It can be seen that the molecular currents are not
compensated inside a nonhomogeneous magnetic, and as
a result, the macroscopic volume magnetization current I'
appears, which flows in the positive direction of the Y-axis.
Accordingly, we can speak about the linear i' and surface j'
current densities, i' being measured in A/m and j' in A/m2.
Calculation of Field B in a Magnetic. It can be stated that
the contribution to field B from a magnetized magnetic
is equal to the contribution that would be created by the
same distribution of currents I' in a vacuum. In other
words, having established the distribution of magnetization
currents I', we can find, with the help of the Biot-
Savart law, the field B' corresponding to them, and then
calculate the resultant field B by formula (7.1).
Unfortunately, the distribution of currents I' depends

not only on the configuration and the properties of a mag-
netic but on the field B as well. For this reason, in the general
case the problem of finding B in a magnetic cannot be solved
directly. It remains for us to try and find another approach
to the solution of this problem. The first step in this direction
is the establishment of an important relation between mag-
netization current I' and a certain property of vector J,
viz. its circulation.
7.2. Circulation of Vector J

It turns out that for a stationary case the circulation of
magnetization J around an arbitrary contour F is equal
Fig. 7.3 Fig. 7.4

to the algebraic sum of magnetization currents I' enveloped
by the contour F:
J dl = /', (7.5)
,
where I' j' dS and integration is performed over an
arbitrary surface stretched on the contour F.
In order to prove this theorem, we shall calculate the
algebraic sum of molecular currents enveloped by contour F.
Let us stretch an arbitrary surface S on the contour F
(Fig. 7.3). It can be seen from the figure that some molecular
currents intersect the surface S twice (once in one direction
and for the second time in the opposite direction). Hence
such currents make no contribution to the resultant mag-
netization current through the surface S.
However, the molecular currents that are wound around
7.2. Circulation of Vector J 177
the contour P intersect the surface S only once. Such molec-

ular currents create a macroscopic magnetization current I'
piercing the surface S.
Let each molecular current be equal to Im and the area
embraced by it be Sm. Then, as is shown in Fig. 7.4, the
element dl of the contour P is wound by those molecular
currents whose centres get into the oblique cylinder with
the volume dV = Sm cos a dl, where a is the angle between
the contour element dl and the direction of vector J at the
point under consideration. All these molecular currents
intersect the surface S once, and their contribution to the
magnetization current is dI' = Imn dV, where n is the molec-
ular concentration. Substituting into this formula the
expression for dV, we obtain
= ImSmn cos a dl = J cos a dl = J dl,
where we took into account that imS„, = pmis the magnetic
moment of an individual molecular current, and ImSmn
is the magnetic moment of a unit volume of the material.
Integrating this expression over the entire contour F,
we obtain (7.5). The theorem is proved.
It remains for us to note that if a magnetic is nonhomo-
geneous, magnetization current I' generally pierces the
entire surface (see Fig. 7.3) and not only the region near its
boundary, adjoining the contour P. This is why we can use
the expression I' = j' dS, where integration is performed
over the entire surface S, bounded by the contour P. In
the above proof, we managed as if to "drive" the entire
current I' to the boundary of the surface S. The only goal
of this method is to simplify the calculation of the current I'.
Differential form of Eq.(7.5):
I VXJ=j', (7.6)
i.e. the curl of magnetization J is equal to the magnetization current
density at the same point of space.
A Remark about the Field of J. The properties of the
field of vector J, expressed by Eqs. (7.5) and (7.6), do not
imply at all that the field J itself is determined only by the
currents I'. The field of J (which is bounded only by the
12-0181
spatial region filled by magnetic) depends on all currents,

viz. both on the magnetization currents I' and the conduc-
tion currents I. However, in some cases
with a definite symmetry the situation
is such as if the field of vector J were
determined only by currents I'.
Example. Find the surface magnetization

current per unit length of a cylinder made
of a homogeneous magnetic, if its magnet-
ization is J, vector J being directed every-
where along the cylinder axis.
Let us apply Eq. (7.5) to the contour
chosen as is shown in Fig. 7.5. It can be
easily seen that the circulation of vector J
around this contour is equal to the product
Fig. 7.5 Jl. In this case, the magnetization current
is the surface current. If we denote its linear
density by i', the contour under consideration embraces the magne-
tization current i'l. It follows from the equality Jl = i'l that
= J. (7.7)
It should be noted, by the way, that vectors i' and J are mutually
perpendicular: i' I J.
7.3. Vector H
Theorem on Circulation of Vector H (for magnetic fields
of direct currents). In magnetics placed into an external
magnetic field magnetization currents are induced. Hence
circulation of vector B will now be determined not only
by conduction currents but by magnetization currents as
well:
B dl = [1,0 (/ r), (7.8)
where I and I' are the conduction and magnetization cur-

rents embraced by a given contour F.
Since the determination of currents I' in the general case
is a complicated problem, formula (7.8) becomes inap-
plicable in practical respect. It turns out, however, that
we can find a certain auxiliary vector whose circulation
will be determined only by the conduction currents embraced
by the contour P. Indeed, as we know, the current I' is
7.3. Vector 11 179
related to the circulation of the magnetization:

J dl = /'. (7.9)
Assuming that the circulation of vectors B and J is taken

over the same contour P we express I' in Eq. (7.8) through
,
formula (7.9). Then

t1,30 J dl — /. (7.10)
Let us denote the integrand of this expression by H.

Thus, we have found an auxiliary vector H,
B
H — —J (7.11)
Feu
whose circulation around an arbitrary contour I' is equal

to the algebraic sum of the conduction currents I embraced
by this contour:
Hd1=/. (7.12)
This formula expresses the theorem on circulation of vec-

tor H: the circulation of vector H around an arbitrary closed
contour is equal to the algebraic sum of the conduction currents
embraced by this contour.
The sign rule for currents is the same as in the case of
circulation of vector B (see p. 149).
It should be noted that vector H is a combination of two
quite different quantities, B/u0 and J. Hence, vector H
is indeed an auxiliary vector which does not have any deep
physical meaning.* However, the important property of
vector H expressed in the theorem on its circulation justifies
the introduction of this vector: in many cases, it consider-
ably simplifies the investigation of the field in magnetics.
Relations (7.11) and (7.12) are valid for any magnetics,
including anisotropic ones.
* The quantity H is often called magnetic field intensity, but we

shall not be using this term to emphasize again its auxiliary nature.
12*
It follows from formula (7.12) that the magnitude of

vector H has the dimensions of the current per unit length.
Hence, H is measured in amperes per metre (Alm).
Differential form of the theorem on circulation of vector II:
x H=j, (7.13)
i.e. the curl of vector H is equal to the density of the conduction

current at the same point of the substance.
Relation Between Vectors J and H. It was shown above
that magnetization J depends on the magnetic induction B
at a given point of the substance. However, customarily
vector J is related not with B but with H. We shall confine
ourselves to the analysis of only those magnetics for which
the dependence between J and H is linear, viz.
J = xH, (7.14)
where x is the magnetic susceptibility. It is a dimensionless
quantity typical of each magnetic (that x is dimensionless
follows from the fact that, according to (7.11), the dimen-
sions of H and J are the same).
Unlike electric susceptibility x which is always positive,
magnetic susceptibility x may be either positive or negative.
Accordingly, the magnetics satisfying formula (7.14) are
divided into paramagnetics (x > 0) and diamagnetics (x < 0).
In paramagnetics, J fit H, while in diamagnetic J 44 H.
It should be noted that in addition to these magnetics,
there exist ferromagnetics for which the dependence J (H)
has a rather complicated form: it is not linear, and besides,
it has a hysteresis, i.e. J depends on the prehistory of a mag-
netic. (Ferromagnetics will be described in greater detail
in Sec. 7.6.)
Relation Between B and H. For magnetics obeying (7.14),
expression (7.11) acquires the form (1 + x) H = 13/110.
Hence
B = p,p,oH, (7.15)
where p, is the permeability of the medium,
= 1 ± x. (7.16)
7.3. Vector H 181
For paramagnetics µ > 1, while for diamagnetics µ < 1.

It should be noted that in both cases p, differs from unity
only slightly, i.e. the magnetic properties of these magnetics
are manifested very weakly.
A Remark about the Field of H. Let us consider a question
which often involves misunderstanding: which currents
determine the field of vector H? Generally, the field of H
(just as the field of B) depends on all currents, both con-
duction and magnetization. This can be seen even from
(7.15). However, in some cases the field H is determined
only by conduction currents. Vector H is very helpful for
these cases in particular. At the same time, this makes
grounds for the erroneous conclusion that the field of H
always depends on conduction currents and for incorrect
interpretations of the theorem on circulation of vector H
and Eq. (7.13). This theorem expresses only a certain prop-
erty of the field of H without defining the field itself.
Example. A system consists of a long straight wire with current I
and an arbitrary piece of a paramagnetic (Fig. 7.6). Let us find the
changes in the fields of vectors B and H
and in the circulation of vector H around
a certain fixed contour P upon the
removal of the magnetic.
The field B at each point of space is
determined by the conduction current I
as well as by the magnetization currents
in the paramagnetic. Since in our case,
in accordance with (7.15), H =
this also applies to the field of vector
H, i.e. it also depends on the conduction Fig. 7.6
current I and the magnetization cur-
rents.
The removal of the piece of paramagnetic will lead to a change
in the field B, and hence in the field H. The circulation of vector B
around the contour r will also change, since the surface stretched on
the contour P will no longer be pierced by the magnetization currents.
Only the conduction current will remain. However, the circulation
of vector H around the contour P will remain unchanged in spite of
the variation of the field H itself.
When j' = 0 Inside a Magnetic? We shall show now that
magnetization currents inside a magnetic are absent if (1)
the magnetic is homogeneous and (2) there are no conduction
currents in it (j = 0). In this case, for any shape of a mag-
netic and any configuration of the magnetic field, we can be
sure that volume magnetization currents are equal to zero,

and only surface magnetization currents remain.
In order to prove this, we shall use the theorem on cir-
culation of vector J around an arbitrary contour I' com-
pletely lying inside the magnetic. In the case of a homo-
geneous magnetic, we can, substituting xH for J, take x
in Eq. (7.5) out of the integral and write
/' = x (Cy' H dl.
In accordance with (7.12), the remaining integral is equal

to the algebraic sum of the conduction currents I embraced
by the contour F. Hence for a homogeneous magnetic, we
have
= x/. (7.17)
This relationship between the currents I' and I is valid for
any contour inside the magnetic, in particular, for a very
small contour, when I' dI' = jn dS and I dl = jn dS.
Then :in dS = xjr, dS, and after cancelling dS, we obtain
In = xjn . This equation holds for any orientation of the
small contour, i.e. for any direction of the normal n to it.
Hence the vectors j' and j themselves are related through
the same equation:
I •
J = XJ. (7.18)
Thus it follows that in a homogeneous magnetic j' = 0
if j = 0, Q.E.D.
7.4. Boundary Conditions for B and H

We shall analyse the conditions for vectors B and H at
the interface between two homogeneous magnetics. These
conditions will be obtained, just as in the case of dielectrics,
with the help of the Gauss theorem and the theorem on cir-
culation. We recall that for vectors B and H these theorems
have the form
B dS — 0, () H dl = /. (7.19)
Condition for Vector B. Let us imagine a very low cylinder

located at the interface between two magnetics as is shown
7.4. Boundary Conditions for a and H 183
in Fig. 7.7. Then the flux of B in the outward direction (we

ignore the flux through the lateral surface) can be written
in the form
1327, AS -1-Bln,AS = 0.
Taking the two projections of vector B onto the normal
n to the interface, we obtain B1n,= —.B1„,, and after can-
celling AS, the previous equation becomes
B20 — B10 , (7.20)

i.e. the normal component of vector B turns out to be the
n
2
n : AS
2
1
I
yy
Fig. 7.7 Fig. 7.8

same on both sides of the interface. This quantity does not
have a discontinuity.
Condition for Vector H. We shall assume, for higher
generality that a surface conduction current with linear
densiy i flows over the interfacial surface of the magnetics.
Let us apply the theorem on circulation of vector H to
a very small rectangular contour whose height is negligibly
small in comparison with its length 1 (the arrangement of
the contour is shown in Fig. 7.8). Neglecting the contribu-
tions from smaller sides of the contour to the circulation,
we can write for the entire contour
H2,/ H „•1= 1,1,
where iN is the projection of vector i onto the normal N
to the contour (vector N forms a right-handed system with
the direction of contour circumvention). Taking the two
projections of vector H onto the common unit vector of the
tangent T (in medium 2), we obtain H„‘ = —H„. Can-
celling 1 out of the previous equation, we get
1 H2 ti — H it = 1,, (7.21)
i.e. the tangential component of vector H generally has

a discontinuity upon a transition through the interface,
which is due to the presence of conduction currents.
If, however, there are no conduction currents at the inter-
face between magnetics (i = 0), the tangential component
of vector H turns out to be the same on both sides of the
interface:
H2T -H1T• (7.22)
Thus, if there is no conduction current at the interface

between two homogeneous magnetics, the components B„,
B25
Ben I a 2 B2 1
2
, Bin
Field B Field H
Fig. 7.9 Fig. 7.10
and HT vary continuously (without a jump) upon a transition

through this interface. On the other hand, the components
B, and HT, in this case have discontinuities.
It should be noted that vector B behaves at the interface
like vector D, while vector H behaves like E.
Refraction of Lines of B. The lines of B undergo refraction
at the interface between two magnetics (Fig. 7.9). As in
the case of dielectrics, we shall find the ratio of the tangents
of angles a1 and a,:
tan a2 B2 5/B2 n
tan a,
We shall confine ourselves to the case when there is no con-
duction current at the interface. According to (7.22) and
(7.20), in this case we have
B2,/p,2 = Bit /R1 and B2, =
Taking these relations into account, we obtain the law of

refraction of the lines of vector B (and hence of vector H
as well) similar to (2.25):
tanoc, _= 1.4
(7.23)
tan a, th
Figure 7.10 depicts the fields of vectors B and H near the
interface between two magnetics (in the absence of con-
duction currents). Here > n,,. A comparison of the den-
sities of the lines shows that B 2 >B1, while H 2 < H1.
The lines of B do not have a discontinuity upon a transition
through the interface, while the H lines do (due to the sur-
face magnetization currents).
The refraction of magnetic field lines is used in magnetic protection.
If, for example, a closed iron shell (layer) is introduced into an exter-
nal magnetic field, the field lines will be concentrated (condensed)
mainly on the shell itself. Inside the shell (in the cavity) the magnetic
field turns out to be considerably weakened in comparison with the
external field. In other words, the iron shell acts like a screen. This
is used in order to protect sensitive devices from external magnetic
fields.
7.5. Field in a Homogeneous Magnetic

It was mentioned in Sec. 7.1 that determination of the
resultant magnetic field in the presence of arbitrary mag-
netics is generally a complicated problem. Indeed, for this
purpose, in accordance with (7.1), the field Bo of conduction
currents must be supplemented by the macroscopic field B'
created by magnetization currents. The problem is that we
do not know beforehand the configuration of magnetization
currents. We can only state that the distribution of these
currents depends on the nature and configuration of the
magnetic as well as on the configuration of the external
field Bo, viz. the field of conduction currents. And since we
do not know the distribution of magnetization currents, we
cannot calculate the field B'.
The only exception is the case when the entire space
occupied by the field. B is filled by a homogeneous isotropic
dielectric. Let us consider this case in greater detail. But
first of all, we shall analyse the phenomena observed when
conduction current flows along a homogeneous conductor
in a vacuum. Since each conductor is a magnetic, magneti-

zation currents also flow through it, viz. volume currents
given by (7.18) and surface currents. Let us take a contour
embracing our current-carrying conductor. In accordance
with the theorem on circulation of vector J (7.5), the alge-
braic sum of magnetization (volume and surface) currents is
equal to zero everywhere since J = 0 at all points of the
contour, i.e. I' = Iv + 1-; = 0. Hence IC = —1.;, i.e. the
volume and surface magnetization currents are equal in
magnitude and opposite in direction.
Thus, it can be stated that in an ordinary case, when cur-
rents flow along sufficiently thin wires, the magnetic field
in the surrounding space (in a vacuum) is determined only
by conduction currents, since magnetization currents com-
pensate each other (except for, perhaps, the points lying
very close to the wire).
Let us now fill the space surrounding the conductor by
a homogeneous nonconducting magnetic (for the sake of
definiteness, we assume that it is a paramagnetic, x > 0).
At the interface between this magnetic and the wire, surface
magnetization current I' will appear. It can be easily seen
that this current has the same direction as the conduction
current I (when x > 0).
As a result, we shall have the conduction current 1, the
surface and volume magnetization currents in the conductor
(the magnetic fields of these currents compensate one another
and hence can be disregarded), and the surface magnetiza-
tion current I' on the nonconducting magnetic. For sufficient-
ly thin wires, the magnetic field B in the magnetic will be
determined as the field of the current I + I'.
Thus, the problem is reduced to finding the current I'.
For this purpose, we surround the conductor by a contour
arranged in the surface layer of the nonconducting magnetic.
Let the plane of the contour be perpendicular to the wire
axis, i.e. to the magnetization currents. Then, taking (7.7)
and (7.14) into consideration, we can write
/' = dl dl= x - dl.
According to (7.12), it follows that I' = x/.

The configurations of the magnetization current I' and
of the conduction current I practically coincide (the wires

are thin), and hence at all points the induction B' of the
field of the magnetization currents differs from the induc-
tion Bo of the field of the conduction currents only in mag-
nitude, these vectors being related in the same way as the
corresponding currents, viz.
B' = xBo. (7.24)
Then the induction of the resultant field B = Bo B' =
—(1 + x) B, or
B = µB°. (7.25)
This means that when the entire space is filled with a homo-
geneous magnetic, B increases times. In other words, the
quantityµ shows the increase in the magnetic field B upon
filling the entire space occupied by the field with a magnetic.
If we divide both parts of (7.25) by titio, we obtain
H = Ho (7.26)
(in the case under consideration, field H turns out to be
the same as in a vacuum).
Formulas (7.24)-(7.26) are also valid when a homogeneous
magnetic fills the entire volume bounded by the surfaces formed
by the lines of Bo (the field of the conduction current). In
this case too the magnetic induction B inside the magnetic
is 11 times larger than Bo.
In the cases considered above, the magnetic induction B'
of the field of magnetization currents is connected with
magnetization J of the magnetic through the simple relation
B' = (7.27)
This expression can be easily obtained from the formula
B = Bo B'if we take into account that B' = xBo and
B = iutioH, where H = J/x.
As has already been mentioned above, in other cases the
situation is much more complicated, and formulas (7.24)-
(7.27) become inapplicable.
Concluding the section, let us consider two simple exam-
ples.
Example 1. Field B in a solenoid. A solenoid having nI ampere-
turns per unit length is filled with a homogeneous magnetic of the
permeability R > 1. Find the magnetic induction B of the field in

the magnetic.
According to (6.20), in the absence of magnetic the magnetic
induction Bo in the solenoid is equal to Ron/. Since the magnetic fills
the entire space where the field differs from zero (we ignore the edge
effects), the magnetic induction B must be R times larger:
B = RRon/. (7.28)
In this case, the field of vector H remains the same as in the absence
of the magnetic, i.e. H = Ho.
The change in field B is caused by the appearance of magnetization
currents flowing over the surface of the
magnetic in the same direction as the
conduction currents in the solenoid
winding when R > 1. If, however, R< 1,
the directions of these currents will be
opposite.
The obtained results are also valid
when the magnetic has the form of a
very long rod arranged inside the sole-
noid so that it is parallel to the sole-
noid's axis.
Example 2. The field of straight
current in the presence of a magnetic.
Suppose that a magnetic fills a long
cylinder of radius a along whose axis Fig. 7.11
a given current I flows. The permea-
bility of the magnetic R > 1. Find the
magnetic induction B as a function of the distance r from the cylin-
der axis.
We cannot use directly the theorem on circulation of vector B
since magnetization currents are unknown. In this situation, vector H
is very helpful: its circulation is determined only by conduction cur-
rents. For a circle of radius r, we have 2xrH = I, whence
B = p,110H = Rp.,//2nr.
Upon a transition through the magnetic-vacuum interface, the
magnetic induction B, unlike H, undergoes a discontinuity
(Fig. 7.11).
An increase in B inside the magnetic is caused by surface magnetiza-
tion currents. These currents coincide in direction with the current I
in the wire on the axis of the system and hence "amplify" this current.
On the other hand, outside the cylinder the surface magnetization
current is directed oppositely, but it does not produce any effect
on field B in the magnetic. Outside the magnetic, the magnetic fields
of both currents compensate one another.
7.6. Ferromagnetism
Ferromagnetics. In magnetic respect, all substances can
be divided into weakly magnetic (paramagnetics and dia-
magnetics) and strongly magnetic (ferromagnetics). It is

well known that in the absence of magnetic field, para-
and diamagnetics are not magnetized and are characterized
by a one-to-one correspondence between magnetization J
and vector H.
Ferromagnetics are the substances (solids) that may possess
spontaneous magnetization, i.e. which are magnetized even
J
Js
Fig. 7.12 Fig. 7.13
in the absence of external magnetic field. Typical ferro-

magnetics are iron, cobalt, and many of their alloys.
Basic Magnetization Curve. A typical feature of ferro-
magnetics is the complex nonlinear dependence J (H) or
B (H). Figure 7.12 shows the magnetization curve for a
ferromagnetic whose magnetization for H = 0 is also zero.
This curve is called the basic magnetization curve. Even at
comparatively small values of H, magnetization J attains
saturation (J 5). Magnetic induction B = µo (H J) also
increases with H. After attaining saturation, B continues
to grow with H according to the linear law B = [10H +
+const, where const = NJ,. Figure 7.13 represents the basic
magnetization curve on the B-H diagram.
In view of the nonlinear dependence B (H), the permeabil-
ity µ for ferromagnetics cannot be defined as a constant
characterizing the magnetic properties of a specific ferro-
magnetic. However, as before, it is assumed that Ix = B/ [1,0H,
but here tA, is a function of H (Fig. 7.14). The value max
of permeability for ferromagnetics may be very large. For
example, for pure iron p,max =---- 5,000, while for supermalloy
[imax = 800,000.
It should be noted that the concept of permeability is
applicable only to the basic magnetization curve since, as

will be shown, the B (H) dependence is nonunique.
Magnetic Hysteresis. Besides the nonlinear dependence
B (H) or J (H), ferromagnetics also exhibit magnetic hystere-
sis: the relation between. B and H or J and II turns out to
Fig. 7.14 Fig. 7.15
be ambiguous and is determined by the history of the ferro-

magnetic's magnetization. If we magnetize an initially
nonmagnetized ferromagnetic by increasing H from zero
to the value at which saturation sets in (point 1 in Fig. 7.15),
and then reduce H from +H1to —H1, the magnetization
curve will not follow path 10, but will take path 1 2 3 4.
If we then change H from —H1 to +H, the magnetization
curve will lie below, and take path 4 5 6 1.
The obtained closed curve is called the hysteresis loop.
The maximum hysteresis loop is obtained when saturation
is attained at points 1 and 4. If, however, there is no satu-
ration at extreme points, the hysteresis loops have similar
form but smaller size, as if inscribed into the maximum
hysteresis loop.
Figure 7.15 shows that for H = 0 magnetization does not
vanish (point 2) and is characterized by the quantity Br
called the residual induction. It corresponds to the residual
magnetization J r. The existence of permanent magnets is
associated with this residual magnetization. The quantity B
vanishes (point 3) only under the action of the field H,
which has the direction opposite to that of the field causing
the magnetization. The quantity H, is called the coercive
force.
The values of Br and Hefor different ferromagne tics vary

over broad ranges. For transformer steel, the hysteresis
loop is narrow (He is small), while for ferromagnetics used
for manufacturing permanent magnets it is wide (H, is
large). For example, for alnico alloy, He =50,000 A/m
and Br = 0.9 T.
These peculiarities of magnetization curves are used in a convenient
practical method for demagnetizing ferromagnetics. A magnetized
sample is placed into a coil through which an alternating current is
passed, its amplitude being gradually reduced to zero. In this case,
the ferromagnetic is subjected to multiple cyclic reverse magnetiza-
tions in which hysteresis loops gradually decrease, contracting to the
point where magnetization is zero.
Experiments show that a ferromagnetic is heated upon
reverse magnetization. It can be shown that in this case the
amount of heat Quliberated per unit volume of the ferromag-
netic is numerically equal to the "area" Su of the hysteresis
loop:
Qu = H dB= Sn . (7.29)
Curie Point. As the temperature increases, the ability of

ferromagnetics to get magnetized becomes weaker, and in
particular, saturation magnetization becomes lower. At a
certain temperature called the Curie point, ferromagnetic
properties of the material disappear.
At temperatures above the Curie point, a ferromagnetic
becomes a paramagnetic.
On the Theory of Ferromagnetism. The physical nature
of ferromagnetism could be explained only with the help of
quantum mechanics. The so-called exchange forces may
appear in crystals under certain conditions. These forces
orient parallel to each other the magnetic moments of elec-
trons. As a result, regions (1-10 tm in size) of spontaneous
magnetization, which are called domains, appear in crystals.
Within the limits of each domain, a ferromagnetic is mag-
netized to saturation and has a certain magnetic moment.
The directions of these moments are different for different
domains. Hence, in the absence of external field, the resultant
magnetic moment of the sample is equal to zero, and the
sample as a whole is macroscopically nonmagnetized.
When an external magnetic field is switched on, the
domains oriented along the field grow at the expense of

domains oriented against the field. In weak fields, this
growth is reversible. In stronger fields, the simultaneous
reorientation of magnetic moments within the entire domain
takes place. This process is irreversible, which explains
hysteresis and residual magnetization.
Problems
• 7.1. Conditions at the interface. In the vicinity of point A
(Fig. 7.16) on a magnetic-vacuum interface, the magnetic induction
Fig. 7.16 Fig. 7.17
in the vacuum is B,, vector Bo forming an angle a, with the normal

to the interface. The permeability of the magnetic is Find the
magnetic induction B in the magnetic in the vicinity of point A.
Solution. The required quantity
B= . (1)
Considering conditions (7.20) and (7.22) at the interface, we find
Lin = B0 cos ao,
.13.r -=t4t0 H.,=ftp,ollot =i1B0.5 = pB, sin a0,
where Hot is the tangential component of vector Ho in the vacuum.
Substituting these expressions into (1), we obtain
B=B0 ycos2 a, p.2 sine ao.
• 7.2. Surface magnetization current. A long thin current-carry-
ing conductor with current I is arranged perpendicularly to a plane
vacuum-magnetic interface (Fig. 7.17). The permeability of the mag-
netic is tt. Find the linear current density i' of the surface magnetiza-
tion current on this interface as a function of the distance r from the
conductor.
Solution. Let us first consider the configuration of the surface mag-
netization current. It is clear from Fig. 7.17 that this current has the
Problems 193
radial direction. We shall use the theorem on circulation of magnetiz-

ation vector J, taking as a contour a small rectangle whose plane is
perpendicular to the magnetization current in this region. The arrange-
ment of this contour is shown in Fig. 7.18, where the crosses indicate
the direction of the surface magnetization current. From the equality
= i'/ we get i' = .1.
Then we can write T = xH, where H can be determined from the
circulation of vector II around the circle of radius r with the centre
Fig. 7.18
on the conductor axis: 2rtrH = I (it is clear from symmetry consider-

ations that the lines of H must be circles lying in the planes per-
pendicular to the conductor with the current I). This gives
= (p. 1) I 12nr.
—
• 7.3. Circulation of vector H. A long thin conductor with cur-

rent I lies in the plane separating the space filled with a nonconduct-
ing magnetic of permeability p. from a vacuum. Find the magnetic
induction B in the entire space as a function of the distance r from
the conductor. Bear in mind that the lines of B are circles with centres
on the conductor axis.
Solution. The lines of H are clearly also circles. Unlike vector B,
vector H undergoes a discontinuity at the vacuum-magnetic interface.
We denote by H and Ho magnetic fields in the magnetic and vacuum
respectively. Then, in accordance with the theorem on circultaion
of vector H around the contour in the form of a circle of radius r with
the centre on the conductor axis, we have
a-crH rtrHo = I. (1)
Besides, B = Bo at the interface, or
= Ho. (2)
Solving Eqs. (1) and (2) together, we obtain
ti = , B=p,p,,H
(1 +11) r (1- - F tt) ar •
13-0181
The configurations of the fields B and H in this case are shown in

Fig. 7.19. We advise the reader to convince himself that for ti = 1,
we arrive at the well-known formulas for B and II in a vacuum.
• 7.4. Circulation of vectors H and J. k direct current I flows
along a long homogeneous cylindrical wire with a circular cross section
of radius R. The wire is made of a paramagnetic with the susceptibil-
ity x. Find (1) field B as a function of the distance r from the axis
of the wire and (2) magnetization current density j' inside the wire.
Solution. (1) From circulation of vector H around the circle of
radius r with the centre at the wire axis, it follows that
r < R, 2xrH = I (r1R)2 (H QC r),
r > R, 2xrH = I (H CX 11r).
Figure 7.20 shows the plots of the dependences H (r) and B (r).
(2) Let us use the theorem on circulation of magnetization J
around the circle of radius r (see Fig. 7.20): 2nrJ = I', where I' is
the magnetization current embraced by this contour. We find the
differential of this expression (going over from r to r dr):
tat d (rJ) = dI'.
Since dI' = j'22-Er dr, the above equation can be transformed as
follows:
J , dJ
I r I- dr •
Let us now consider that J = xH = (xI12xtR 2) r. This gives
j' =
It can be easily seen that this current has the same direction as the
conduction current (unlike the surface magnetization current which
has the opposite direction).
• 7.5. A long solenoid is filled with a nonhomogeneous isotropic
paramagnetic whose susceptibility depends only on the distance r
from the solenoid axis as x = ar2, where a is a constant. The magnetic
induction on the solenoid axis is Bo. Find (1) magnetization I and (2)
magnetization current density j' as functions of r.
Solution. (1) J = xH. In our case, H is independent of r (this di-
rectly follows from circulation of vector H around the contour shown
in the left part of Fig. 7.21). Hence H = Ho on the solenoid's axis,
and we obtain
I ,,--- ar2H0 = ar2Boitto.
(2) From the theorem on circulation of magnetization J around
an infinitely narrow contour shown in Fig. 7.21 on the right, it follows
that
Jl — (.1 + dJ) 1 = .4,'1 dr,
Problems 195
where l is the contour height and dr its width. Hence

d.1" 2aBo
Jn= — dr r.
The minus sign indicates that vector j' is directed against the normal
vector n which forms the right-hand system with the direction of cir-
Fig. 7.20 Fig. 7.21
cumvention of the contour. In other words, j' is directed towards us

in the region of the right contour in the figure, i.e. volume magnetiza-
tion currents form with vector B, the left-hand system.
e 7.6. A permanent magnet has the shape of a ring with a narrow

gap between the poles. The middle diameter of the ring is d. The gap
width is b, the magnetic induction of the field in the gap is B. Find
the magnitudes of vectors H and J inside the magnet, ignoring the
dissipation of the field at the edges of the gap.
Solution. Using the theorem on circulation of vector II around the
dashed circle of diameter d (Fig. 7.22) and considering that there is
no conduction current, we can write
(ad — b) Hi + = 0,
where IL, is the projection of vector H onto the direction of contour
circumvention (which is taken so that it coincides with the direction
of vector B in the gap). Hence
bB bB
HT = (1)
(ad — b) Road •
The minus sign indicates that the direction of vector H inside the
magnetic material is opposite to the direction of vector B at the same
point. It should be noted that as b 0, H -- 0 as well.
The magnitude of magnetization J can be found by formula (7.11),
13*
taking into account (1):

Blito B
J
1—blad 11,0
The relationship between vectors B/p,o, H, and J at any point
inside the magnet is shown in Fig. 7.23.
13/ P o
J
Fig. 7.22 Fig. 7.23
• 7.7. A winding containing N turns is wound on an iron core
in the form of a torus with the middle diameter d. A narrow transverse
slot of width b (Fig. 7.22) is cut in the core. When current I is passed
through the winding, the magnetic induction in the slot is B. Find
the permeability of iron under these conditions, neglecting the dis-
sipation of the field at the edges of the slot.
Solution. In accordance with the theorem on circulation of vector H
around the circle of diameter d (see Fig. 7.22), we have
(ad b) H bH0 = NI,
—
where II and Ho are the magnitudes of H in iron and in the slot re-
spectively. Besides, the absence of field dissipation at the edges means
that
B = Bo.
Using these two equations and taking into account that B
= ttli0H and b < d, we obtain
a dB
11= ix oNI—bB •
• 7.8. The force acting on a magnetic. The device shown in

Fig. 7.24 is used for measuring (with the help of a balance) a force with
which a small paramagnetic sphere of volume V is attracted to the
pole of magnet M. The magnetic induction on the axis of the pole
shoe depends on the height as B = Bo e-ax2, where Bo and a are con-
stants. Find (1) the height am at which the sphere must be placed
so that the attractive force be maximum and (2) the magnetic sus-
ceptibility of the paramagnetic if the maximum force of attraction
is Fm.
Problems 197
Solution. (1) Let, for the sake of definiteness, vector B on the axis
be directed upwards and coincide with the X-axis. Then, in accordance
with (6.34), Fx = pm0.1310x, where we took into account that the
magnetic moment pm is directed as vector B (for paramagnetics) and
hence replaced On by Ox.
Next, since pm = JV = yHV and.
OB/Ox = —2aBoxe-Qxz, we have
F x = — Axe-2ax 2, (1)
where A = 2aB2xVINuo.
Having calculated the derivative
dF,Idx and equating it to zero, we ob-
tain the following equation for deter-
mining xm: 1 — 4ax2 = 0, whence
xm -1/2 17a- . (2)
(2) Substituting (2) into (1), we
find 11111111111 M
Ro Frn e Fig. 7.24
=
B8v V a
where we took into account that p. ^ 1 for paramagnetics.
• 7.9. A long thin cylindrical rod made of a paramagnetic with
the magnetic susceptibility x and cross-sectional area S is arranged
Fig. 7.25
along the axis of a current-carrying coil. One end of the rod is at the
centre of the coil, where the magnetic field is B, while the other end
lies in the region where the magnetic field is practically absent. Find
the force with which the coil acts on the rod.
Solution. Let us mentally isolate an element of the rod of length dx
(Fig. 7.25). The force acting on this element is given by
aB
dFx =dpm
Onx •
Suppose that vector B on the axis of the coil is directed to the right
(see the figure), towards positive values of x. Then B, = B, On = Ox,
198 8. Relative Nature of Electric and Magnetic Fields
and since dpm = JS dx = xHS dx, we get
dF,,==xHS dx OR = XS B dB.
Ox u[to
Having integrated this equation, we obtain
= XS c XSB 2
B dB=
2iutp4 •
B
The minus sign indicates that vector F is directed to the left, i.e. the
rod is attracted to the current-carrying coil.
•7.10. A small sphere of volume V, made of a paramagnetic with
magnetic susceptibility x, was displaced along the axis of a current-
carrying coil from the point where the magnetic induction is B to the
region where the magnetic field is practically absent. Find the work
against the magnetic forces accomplished in this case.
Solution. Let us direct the X-axis along the axis of the coil. Then
the elementary work done against the magnetic forces upon the dis-
placement of the sphere by dx is equal to
aB
6A —F x dx= dx, (1)
P m any
where Fx is the projection of the magnetic force (6.34) onto the X-axis
The minus sign indicates that the work is done against this force.
Suppose that vector B on the axis is directed towards positive
values of x. Then Bx = B and an = Ox (otherwise, Bx = — B and
On = — Ox, i.e. the derivative OBx/an does not depend on the direction
of vector B). Considering that pm = JV = xHV, we rewrite Eq. (1)
in the form
OB , xV
ti A= —X12 V - ar = B dB.
Ox Iltto
Integrating this equation between B and 0, we obtain
0
xV xB2V
A=— B dB =-_--
14to J 211,[10 •
B
8. Relative Nature of Electric

and Magnetic Fields
8.1. Electromagnetic Field. Charge Invariance
So far, we have considered electric and magnetic fields
separately, without establishing any clear relation between
8.1. Electromagnetic Field. Charge Invariance 199
them. This could be done only because the two fields were
static. In other cases, however, it is impossible.
It will be shown that electric and magnetic fields must
always be considered together as a single total electromagnetic
field. In other words, it turns out that electric and magnetic
fields are in a certain sense the components of a single physi-
cal object which we call the electromagnetic field.
The division of the electromagnetic field into electric
and magnetic fields is of relative nature since it depends to
a very large extent on the reference system in which the
phenomena are considered. The field which is constant in
one reference frame in the general case is found to vary in
another reference frame.
Let us consider some examples.
A charge moves in an inertial system K at a constant velocity v.
In this system, we shall observe both the electric and magnetic fields
of this charge, the two fields varying with time. If, however, we go
over to an inertial system K' moving together with the charge, the
charge will be at rest in this system, and we shall observe only the
electric field.
Suppose two identical charges move in the system K towards each
other at the same velocity v. In this system, we shall observe both
electric and magnetic fields, both these fields varying. In this case it
is impossible to find such a system K' where only one of the fields
would be observed.
In the system K, there exists a permanent nonuniform magnetic
field (e.g., the field of a fixed permanent magnet). Then in the system
K' moving relative to the system K we shall observe a varying mag-
netic field and, as will be shown below, an electric field as well.
Thus, it is clear that in different reference frames different
relations are observed between electric and magnetic fields.
Before considering the main points of this chapter, viz.
the laws of transformation of fields upon a transition from
one reference system to another, we must answer the
question which is important for further discussion: how do
an electric charge q itself and the Gauss theorem for vector E
behave upon such transitions?
Charge Invariance. At present, there is exhaustive evi-
dence that the total charge of an isolated system does not
change with the change in the motion of charge carriers.
This may be proved by the neutrality of a gas consisting
of hydrogen molecules. Electrons in these molecules move
at much higher velocities than protons. Therefore, if the
charge depended on the velocity, the charges of the elec-

trons and protons would not compensate each other and the
gas would be charged. However, no charge (to 20 decimal
places!) has been observed in hydrogen gas.
Another proof can be obtained while observing the heat-
ing of a piece of a substance. Since the electron mass is
considerably smaller than the mass of nuclei, the velocity
of electrons upon heating must increase more than that of
nuclei. And if the charge depended on the velocity, the sub-
stance would become charged upon heating. Nothing of
this kind was ever observed.
Further, if the electron charge depended on its velocity,
then the total charge of a substance would change as a result
of chemical reactions, since the average velocities of electrons
in a substance depend on its chemical composition. Cal-
culations show that even a weak dependence of the charge
on the velocity would result in extremely strong electric
fields even in the simplest chemical reactions. But nothing
of this kind was observed in this case either.
Finally, the design and operation of all modern particle
accelerators are based on the assumption that a particle's
charge is invariable when its velocity is changed.
Thus, we arrive at the conclusion that the charge of any
particle is a relativistic invariant quantity and does not
depend on the particle's velocity or on the choice of the
reference system.
Invariance of the Gauss Theorem for Field E. It turns
out, as follows from the generalization of experimental
results, that the Gauss theorem E dS = q/go is valid
not only for fixed charges but for moving charges also. In
the latter case, the surface integral must be taken for a def-
inite instant of time in a given reference system.
Besides, since different inertial reference systems are
physically equivalent (in accordance with the relativity
principle), we can state that the Gauss theorem is valid
for all inertial systems of reference.
8.2. Laws of Transformation for Fields E and B

While going over from one reference system to another,
fields E and B are transformed in a certain way. The laws
8.2. Laws of Transformation for Fields E and B 201
of this transformation are established in the special theory

of relativity in a rather complicated way. For this reason,
we shall not present here the corresponding derivations and
will pay main attention to the content of these laws, the
corollaries following from them, and the application of
these laws for solving some specific problems.
Formulation of the Problem. Suppose we have two inertial
systems of reference: the system K and the system K' moving
relative to the first system at a velocity vo. We know the
magnitudes of the fields E and B at a certain point in space
and time in the system K. What are the magnitudes of the
fields E' and B' at the same point in space and time in the
system K'? (Recall that the same point in space and time
is a point whose coordinates and time in the two reference
systems are related through the Lorentz transformations*)
As was mentioned above, the answer to this question is
given in the theory of relativity from which it follows that
the laws of transformation of fields are expressed by the
formulas
E'm =
[vo >03] xEl/c2 (8.1)
= Bl 0_32
--132 '
Here the symbols and mark the longitudinal and trans-

verse (relative to the vector vo) components of the electric
and magnetic fields, i3 = vo/c, and c is the velocity of light
in a vacuum (c2 == 1/84•
Written in terms of projections, these formulas have the form
B c = Bx ,
Ey — voBz By 4- V o E z IC 2
9y —
17.1-02 (8.2)
= Ez+ voBy .trz .8,—vo Eyle2
pa 0.-132
where it is assumed that the X- and X'-axes of coordinates are directed
x – vo t t—xvoica
y'=--y,
Y-1 – (vgle)2 iri —(voic)2
along the vector vo, the Y'-axis is parallel to the Y-axis and the Z'-axis
to the Z-axis.
It follows from Eqs. (8.1) and (8.2) that each of the vectors
E' and B' is expressed both in terms of E and B. This is
an evidence of a common nature of electric and magnetic
fields. Taken separately, each of these fields has no absolute
meaning: speaking of electric and magnetic fields, we must
indicate the system of reference in which these fields are
considered.
It should be emphasized that the properties of electro-
magnetic field expressed by the laws of its transformation
are local; the values of E' and B' at a certain point in space
and time in the system K' are uniquely determined only by
the values of E and B at the same point in space and time
in the system K.
The following features of the laws of field transformation
are also very important.
1. Unlike the transverse components of E and B, which
change upon a transition from one reference system to an-
other, their longitudinal components do not change and remain
the same in all systems.
2. Vectors E and B in different systems are connected
with each other through highly symmetric relations. This
can be seen most clearly in the form of the laws of trans-
formation for the projections of the fields [see (8.2)].
3. In order to obtain the formulas of the inverse trans-
formation (from K' to K), it is sufficient to replace in for-
mulas (8.1) and (8.2) all primed quantities by unprimed
ones (and vice versa) as well as the sign of vo.
Special Case of Field Transformation (vo <c). If the
system K' moves relative to the system K at a velocity
vo ec, the radicals in the denominators of formulas (8.1)
can be replaced by unity, and we obtain
Eii = Eii, B6 =
Ei = E1 + [vo x B1— [vo x E]/c2. (8.3)
E' = E + [vo x B], B' =13— [vo E]/c2. (8.4)

It should be noted that the first formula in (8.4) can be

directly obtained in a very simple way. Let a charge q
in the system K have a velocity vo at a certain instant t.
The Lorentz force acting on this charge is F = qE
-F q [vo X B]. We go over to the inertial system K' moving
relative to the system K at the same velocity as the charge q
at the instant t, viz. the velocity vo. At this instant, the
charge q is at rest in the system K', and the force acting
on the immobile charge is purely electric: F' = qE'. If
v < c (as in the case under consideration), this force is
invariant (F' = F), whence we obtain the first formula in
(8.4).
However, the transformation formulas for magnetic field
can be obtained only with the help of the theory of relativity
as a result of rather cumbersome calculations.
Let us consider a simple example of the application of
formulas (8.4).
Example. A large metallic plate moves at a constant nonrelativistic
velocity v in a uniform magnetic field B (Fig. 8.1). Find the surface
density of charges induced on the surfaces of the plate as a result Of
its motion.
Let us go over to the reference sys-
tem fixed to the plate. In accordance
with the first of formulas (8.4), in this
system of reference a constant uni-
form electric field
E' = [v X B]
will be observed. It is directed towards
us. Under the action of this external Fig. 8.1
field, the charges will be displaced so
that positive charges will appear on
the front surface of the plate and negative charges, on the hind sur-
face.
The surface density a of these charges will be such that the field
created by these charges inside the plate completely compensates the
external field E', since in equilibrium the resultant electric field inside
the plate must be equal to zero. Taking into account relation (1.11),
we obtain
a = 60E' = eovB.
I t should be noted that while solving this problem, we could follow
another line of reasoning, viz. from the point of view of the reference
frame in which the plate moves at the velocity v. In this system there
will be an electric field inside the plate. It appears as a result of the
action of the magnetic component of the Lorentz force causing the
displacement of all the electrons in the plate behind the plane of

Fig. 8.1. As a result, the front surface of the plate will be charged
positively and the hind surface, negatively, and the electric field
appearing in the plate will be such that the electric force qE compen-
sates the magnetic component of the Lorentz force, q Iv X B], whence
E —[v X B]. This field is connected with the surface charge density
through the same formula a = co vB.
Both approaches to the solution of the given problem are equally
justified.
Relativistic Nature of Magnetism. Formulas (8.1) and

(8.2) for the transformation of fields lead to a remarkable
conclusion: the appearance of the magnetic field is a purely
relativistic effect following from the existence in nature
of the limiting velocity c equal to the velocity of light in
a vacuum.
If this velocity (and hence the velocity of propagation of
interactions) were infinite, no magnetism would exist at all.
Indeed, let us consider a free electric charge. Only electric
field exists in the system K where this charge is'at rest. Con-
sequently, according to (8.1), no magnetic field would
appear in any other system K' if c tended to infinity. This
field appears only due to the fact that c is finite, i.e. in the
long run due to the relativistic effect.
The relativistic nature of magnetism is a universal physical
fact, and its origin is associated with the absence of mag-
netic charges.
Unlike most of relativistic effects, magnetism can be
easily observed in many cases, for example, the magnetic
field of a current-carrying conductor. The reason behind
such favourable circumstances is that the magnetic field
may be created by a very large number of moving electric
charges under the conditions of the almost complete vanish-
ing of the electric field due to practically ideal balance of
the numbers of electrons and protons in conductors. In
this case the magnetic interaction is predominant.
The almost complete compensation of electric charges
made it possible for physicists to study relativistic effects
(i.e. magnetism) and discover correct laws. For this reason,
the laws of electromagnetism, unlike Newton's laws, have
not' been refined after the creation of the theory of relativity.
Field Varies Rather than Moves. Since the relations be-
tween electric and magnetic fields vary upon a transition
to another reference system, the fields E and B must be

treated with care. For instance, even the concept of the
force exerted on a moving charge by a moving magnetic
field does not have any precise meaning. The force is deter-
mined by the values of the quantities E and B at the point
where the charge is located. If as a result of motion of the
sources of the fields E and B their values at this point change,
the force will change as well. Otherwise the motion of the
sources does not affect the magnitude of the force.
Thus, while solving the problem about the force acting
on a charge, we must know E and B at the point of location
of the charge as well as its velocity v. All these quantities
must be taken relative to the inertial reference frame we are
interested in.
The expression "mov- K B /1(11
ing field" (if it is used)
should be interpreted as wo
a convenient way of the
verbal description of a
Are
varying field under cer-
tain conditions, and noth-
ing more. Fig. 8.2
The following simple
example will illustrate the fact that upon a transition
from one reference frame to another the field must be
treated with care.
Example. A charged particle is at rest between the poles of a mag-
net fixed in the system K. Let us go over to the system K' that moves
to the right (Fig. 8.2) at a nonrelativistic velocity vo relative to the
system K. (1) Can we state that the charged particle moves in the mag-
netic field in the system K'? (2) Find the force acting on this particle
in the system K'.
(1) Yes, the particle moves in the magnetic field. It should be noted,
however, that it moves in the magnetic field and not relative to the
magnetic field. It is meaningful to say that a particle moves relative
to the reference system, a magnet, or other objects. However, we cannot
state that a particle moves relative to the magnetic field. The latter
statement has no physical meaning. All this refers not only to magnetic
but to electric fields as well.
(2) In order to find the force, we must take into account that the
electric field E' = [vo X B] directed towards us (Fig. 8.2), will
appear in the system K'. In this system, the charge will move to the
left at the velocity —vo, and this motion will occur in crossed electric
and magnetic fields. For the sake of definiteness, we suppose that the
charge of the particle is q > 0. Then the Lorentz force in the system K'
is
F ' a' + q[—v 0 B] =q [v 0X B] — q[vo x B] = 0
which, however, directly follows from the invariance of the force upon
nonrelativistic transformations from one reference frame to the other.
8.3. Corollaries of the Laws of Field Transformation

Some Simple Corollaries. Several simple and useful rela-
tions follow in some cases from transformation formulas
(8.1).
1. If the electric field E alone is present in the system K
(the magnetic field B = 0), the following relation exists
between the fields E' and B' in the system K':
B' = —Ivo x E'l/c2. (8.5)
Indeed, if B = 0, then El = E,_/V1 — (32 and I311 = 0,
131 = —[vo x El/c2 V1— = --tvo x E']/c2, where we
took into account that in the cross product we can write
either E or El.(the same applies to the primed quantities).
Considering that B' = B,_' = Bi, we arrive at for-
mula (8.5).
2. If the magnetic field B alone is present in the system K
(while the electric field E = 0), then for the system K'
we have
E' = [vc, X (8.6)
Indeed, if E = 0, then Bd. ' = 131/V-1 — (32 and Eli = 0
Ei = [v, x B]/V1 — (32. Substituting B1for B and then
B' for BI in the last cross product, we arrive at formula
(8.6).
Formulas (8.5) and (8.6) lead to the following important
conclusion: If only one of the fields (E or B) is present in the
system K, the electric and magnetic fields in the system K'
are mutually perpendicular (E' 1 B'). It should be noted
that the opposite statement is true only under certain ad-
ditional limitations imposed on the magnitudes of the vec-
tors E and B.
Finally, since Eqs. (8.5) and (8.6) contain only the quan-
tities referring to the same system of reference, these equa-
8.3. Field Transformation Laws. Corollaries 207
tions can be easily applied to the fields varying in space and

time. An illustrative example is the field of a uniformly
moving point charge.
Field of Freely Moving Relativistic Charge,. The formulas
for field transformation are interesting above all by that
they express remarkable properties of electromagnetic field.
Besides, they are important from a
purely practical point of view since
they allow us to solve some problems
in a simpler way.
For example, the problem of deter-
mining the field of a uniformly mov-
ing point charge can be solved as a
result of transformation of a purely
Coulomb field which is observed in the
reference system fixed to the charge.
Calculations show (see Problem 8.10)
that the lines of the field E created by
a freely moving charge q have the form
shown in Fig. 8.3, where v is the ve- Fig. 8.3
locity of the charge. This pattern cor-
responds to the instant "picture" of the electric field config-
uration. At an arbitrary point P of the reference system, vec-
tor E is directed along the radius vector r drawn from the
point of location of the charge q to point P.
The magnitude of vector E is determined by the formula
q 1—(32
E = 4aeo
1 r2(1- 02 sine 0)3/ 2 (8.7)
where p = v/c, ft is the angle between the radius vector r

and the vector v corresponding to the velocity of the charge.
The electric field is "flattened" in the direction of motion
of the charge (see Fig. 8.3), the degree of flattening being
the larger the closer the velocity v of the charge to the veloc-
ity c of light. It should be also borne in mind that the field
shown in the figure "moves" together with the charge. That
is why the field E in the reference frame relative to which
the charge moves varies with time.
Knowing the field E, we can find the field B in the same
reference system:
1 [to q[v 1-132
B=— cz ry x EI = 0 (1-132sing '0)3/2
(8.8)
This formula is a corollary of relation (8.5) in which we have

replaced the primed quantities by unprimed ones and v
by —v.
When v < c ((3 < 1), expressions (8.7) and (8.8) are trans-
formed to (1.2) and (6.3) respectively.
8.4. Electromagnetic Field Invariants

Since vectors E and B characterizing electromagnetic field
depend on the system of reference (at the same point in
space and time), a natural question arises concerning the
invariants of electromagnetic field, i.e. the quantitative
characteristics independent of the reference system.
It can be shown that there exist two such invariants which
are the combinations of vectors E and B, viz.
E •B = inv. and E2— c2B2 =inv. (8.9)
The invariance of these quantities (relative to the Lorentz

transformations) is a consequence of field transformation
formulas (8.1) or (8.2). This question is considered in greater
detail in Problem 8.9.
The application of these invariants allows us in some
cases to find the solution quickly and simply and make the
corresponding conclusions and predictions. We shall list
the most important conclusions.
1. The invariance of E • B immediately implies that if E B
in some reference system, i.e. E • B = 0, in all other inertial systems
E' L B' as well.
2. It follows from the invariance of E2 — c2B2 that if E = cB
(i.e. E2 c2B2 = 0), in any other inertial system E' = cB' as well.
3. If in a certain reference system the angle between vectors E
and B is acute (obtuse) (this means that E • B is greater (less) than zero),
in any other reference system the angle between vectors E' and B'
will be acute (obtuse) as well.
4. If in a certain reference system E > cB (or E < cB), (which
means that E2 — c2B2 is greater (or less) than zero), in any other
reference system E' > cB' (or E' < cB') as well.
5. If both invariants are equal to zero, then E 1.B and E = cB
in all reference systems. It will be shown later that this is precisely
observed in electromagnetic waves.
6. If the invariant E • B alone is equal to zero, we can find a refer-
ence system in which either E' = 0 or B' = 0. The sign of the other
invariant determines which of these vectors is equal to zero. The
Problems 209
opposite is also true: if E = 0 or B = 0 in some reference system,

then E' 1 B' in any other system. (This conclusion has already been
.
drawn in Sec. 8.3.)

And finally one more remark: it should be borne in mind
that generally fields E and B depend both on coordinates
- vo
q
Fig. 8.4 Fig 8.5
and on time. Consequently, each invariant of (8.9) refers

to the same point in space and time of the field, the coordi-
nates and time of the point in different systems of reference
being connected through the Lorentz transformations.
Problems
• 8.1. Special case of field transformation. A nonrelativistic
point charge q moves at a constant velocity v. Using the transforma-
tion formulas, find the magnetic field B of this charge at a point whose
position relative to the charge is defined by the radius vector r.
Solution. Let us go over to the system of reference K' fixed to the
charge. In this system, only the Coulomb field with the intensity
E' = r,
4aso r
3
is present, where we took into account that the radius vector r' = r
in the system K' (nonrelativistic case). Let us now return from the
system K' to the system K that moves relative to the system K' at
the velocity —v. For this purpose, we shall use the formula for the
field B from (8.4), in which the role of primed quantities will be played
by unprimed quantities (and vice versa), and replace the velocity vo
by —vo (Fig. 8.4). In the case under consideration, vo = v, and hence
B = B' [v X E']/c2. Considering that in the system K' B' = 0
and that c2 = lisop.o, we find
110 q [v x r]
B= •
4a-c r3
We have obtained formula (6.3) which was earlier postulated as
a result of the generalization of experimental data.
14-0181
• 8.2. A large plate made of homogeneous dielectric with the

permittivity a moves at a constant nonrelativistic velocity v in a uni-
form magnetic field B as shown in Fig. 8.5. Find the polarization P
of the dielectric and the surface density a' of bound charges.
Solution. In the reference system fixed to the plate, in addition to
the magnetic field, there will be an electric field which we shall denote
by E. In accordance with formulas (8.4) of field transformation, we
have
Eo = [v X /3].
The polarization of the dielectric is given by
8-1
P xe0E' =8, 8 [V XII],
where we took into account that according to (3.29), E' = EL/8. The
surface density of bound charges is
8-1
10'I -P —go vB.
8
At the front surface of the plate (Fig. 8.5), a' > 0, while on the opposite
face a' < 0.
• 8.3. Suppose we have an uncharged long straight wire with
a current I. Find the charge per unit length of this wire in the reference
system moving translationally with a nonrelativistic velocity vo
along the conductor in the direction of the current I.
Solution. In accordance with the transformation formulas (8.4),
the electric field E' = No X B] will appear in the moving reference
system, or
= —v,p,0/72nr (1)
where r is the distance from the wire axis.
The expression for B in this formula was obtained with the help
of the theorem on circulation.
On the other hand, according to the Gauss theorem (in the moving
system) we have
E. = XV2a80r, (2)
where X' is the charge per unit length of the wire.
A comparison of (1) and (2) gives
= —v0//c2,
where c2 = 1/80µo. The origin of this charge is associated with the
Lorentz contraction experienced by the "chains" of positive and nega-
tive charges (they have different velocities!).
• 8.4. There is a narrow beam of protons moving with a relativis-
tic velocity v in the system of reference K. At a certain distance from
the beam, the intensity of the electric field is equal to E. Find the
Problems 211
induction B' of the magnetic field at the same distance from the beam
in the system K' that moves relative to the system K at the velocity Co
in the direction of the proton beam.
Solution. This problem can be solved in the most'simple way with
the help of formulas (8.1). But first we must find the induction B in
the system K at the same distance from the beam where the intensity E
is given.
Using the theorem on circulation of vector B and the Gauss theo-
rem for vector E, we find
B = i.tb//2nr, E = X123-Egor,
where r is the distance from the beam, I = Xv is the current, and X
is the charge per unit length of the beam. It follows from these formulas
that
B/E = sotto//X = v/c2,
where c2 = 1/sotto. Substituting the expression for B from this for-
mula into the last of transformation formulas (8.1), we obtain
Ely—Vol
B'
c2 1 (vo/c)2
—
If in this case Co < v, the lines of B' form the right-handed system
with the vector vo. If vo > v, they form the left-handed system (since
the current I' in the system K' in this case will flow in the opposite
direction).
• 8.5. A relativistic charged particle moves in the space occupied
by uniform mutually perpendicular electric and magnetic fields E
and B. The particle moves rectilinearly in the direction perpendicular
to the vectors E and B. Find E' and B' in the reference system moving
translationally with the particle.
Solution. It follows from the description of motion of the particle
that its velocity must satisfy the condition
vB = E. (1)
In accordance with the transformation formulas (8.1), we have
E+ [v X13]
E' =0,
1/1-132
since in the case under consideration the Lorentz force (and hence
the quantity E [v X B]) is equal to zero.
According to the same transformation formulas, the magnetic
field is given by
B— [v X E]/c2
B'
The arrangement of vectors is shown in Fig. 8.6, from which it follows

that [v X E] tt B. Consequently, considering that according to (1)
14*
v = EIB, we can write

B—E21Bc2 B (1-02)
1/1- p2
- - (2
or in the vector form
B' =-. B /1— (E/cB)2.
We advise the reader to verify that the obtained expressions satisfy
both field invariants.
• 8.6. The motion of a charge in crossed fields E and B. A non-
relativistic particle with a specific charge aim moves in a region where
Y
E
Fig. 8.6 Fig. 8.7
uniform, mutually perpendicular fields E and B have been created

(Fig. 8.7). At the moment t = 0 the particle was at point 0 and its
velocity was equal to zero. Find the law of motion of the particle,
x (t) and y (t).
Solution. The particle moves under the action of the Lorentz force.
It can be easily seen that the particle always remains in the plane XY.
Its motion can be described most easily in a certain system K' where
only the magnetic field is present. Let us find this reference system.
It follows from transformations (8.4) that E' = 0 in a reference
system that moves with a velocity vo satisfying the relation E
—[vo X B]. It is more convenient to choose the system K' whose
velocity vo is directed towards the positive values on the X-axis
(Fig. 8.7), since in such a system the particle will move perpendicularly
to vector B' and its motion will be the simplest.
Thus, in the system K' that moves to the right at the velocity
vo = EIB, the field E' = 0 and only the field B' is observed. In accord-
ance with (8.4) and Fig. 8.7, we have
B' = B — [vo X E]/c2 = B (1 — vS/c2).
Since for a nonrelativistic particle vo < c, we can assume that B' = B.
In the system K', the particle will move only in the magnetic field,
its velocity being perpendicular to this field. The equation of motion
for this particle in the system K' will have the form
mvani = qvoB. (1)
Problems 213
This equation is written for the instant t = 0, when the particle

moved in the system K' as is shown in Fig. 8.8. Since the Lorentz
force F is always perpendicular to the velocity of the particle, vo =
= const, and it follows from (1) that in the system. K' the particle
will move in a circle of radius
R = mvolqB.
Thus, the particle moves uniformly with the velocity vo in a circle
in the system K' which, in turn, moves uniformly to the right with
the same velocity vo = EIB. This motion can be compared with the
motion of the point q at the rim of a wheel (Fig. 8.9) rolling with the
angular velocity o = vo/R = qB1m.
Figure 8.9 readily shows that the coordinates of the particle q
at the instant t are given by
x = vot—R sin cot = a (cot—sin cot),
g= R — R cos on = a (1— cos cot),
where a = mElqB2 and co = qB1m.
• 8.7. There is only a uniform electric field E in an inertial sys
tem K. Find the magnitudes and directions of vectors E' and B' in
the system K' moving relative to the system K at a constant relativistic
velocity vo at an angle a to vector E.
Solution. According to transformation formulas (8.1) and taking
into account that B = 0 in the system K, we obtain
El l = E cos a, El --- E sin a/ -1/-1 — (32, p=voic.
Hence the magnitude of vector E' is given by
E' =j1E;; E? = E 17(1— p2 cost a) /(1 _pa)
and the angle a' between vectors E' and vo can be determined from
the formula
tan a' = tan a/3/1— 32.
The magnitude and direction of vector B' can be found in a similar
way:
[vo XE]/(c2
— — pa),
This means that B' 1vo, and
B' = vo E sin a/(c2V1 — fi2).
• 8.8. Uniform electric and magnetic fields E and B of the same
direction exist in a system of reference K. Find the magnitudes of
vectors E' and B' and the angle between these vectors in the system K'
moving at a constant relativistic velocity vo in the direction perpen-
dicular to vectors E and B.
Solution. In accordance with formulas (8.1), in the system K'
the two vectors E' and B' will be also perpendicular to the vector v 0
(Fig. 8.10). The magnitudes of the vectors E' and B' can be found by
the formulas
E , =- E 2±(v0B)2 B2 +(V0E/C2)2
V 1—(volc)2 ' V I.- (710/0 2 •
The angle between the vectors E' and B' can be found from its tan-
gent through the formula
tan a' = tan (a/+a'B)= (tan a'E + tan a'B)/(1 — tan aE tan a'B).
Since tan ce = voB/E and tan a'B = voE/c2B (Fig. 8.10), we obtain
vo (B2+ E2/0)
tan a' (1—R2) EB •
This formula shows that as vo c (r3 1), the angle a' ri/2. The
IN
2/ I
/ I (DB
R
Vo 0 X'
Fig. 8.8 Fig. 8.9
opposite is also true: if we know E and B in one reference system, and

the angle between these vectors is less than 90°, there exist reference
systems where the two vectors E' and B' are parallel to one another.
• 8.9. Invariant E•B. Using transformation formulas (8.1)
show that the quantity E •B is an invariant.
Solution. In the system K', this product will be
E'•B'---(Eji r E1)•(13±B1)=Ell•BIH El.B1.
-- - (1)
Let us write the last term with the help of formulas (8.1):
(E± + [voX BID• — fvo X E j/c2)
-131 — (2)
1-132
Considering that the vectors E l and IIIare perpendicular to the
vector vo, we transform the numerator of (2) to the form
Ej_•131—(vo /c)2 E1.131 =E±-131 (1—n, (3)
Problems 215
where we used the fact that [vo X Bi]gy o X E1] = 1E1 X

X cos a = vBBi•El(Fig. 8.11). The remaining two scalar products
in (2) are equal to zero since the vectors are mutually perpendicular.
Thus, the right-hand side of (1) becomes
E;,-13(i +El•BI=E il •B11 H-E j_•Bi ---E•B
Q.E.D.
• 8.10. Field E of a uniformly moving charge. A point charge q
moves uniformly and rectilinearly with a relativistic velocity v.
Fig. 8.10 Fig. 8.11
Find the intensity E of the field created by this charge at a point whose
radius vector relative to the charge is equal to r and forms an angle 15
with the vector v.
Solution. Suppose that the charge moves in the positive direction
of the X-axis in the system of reference K. Let us go over to the sys-
tem K' at whose origin this charge is at rest (the axes X' and X of
these systems coincide, while the Y'- and Y-axes are parallel). In the
system K' the field E' of the charge has a simpler form:
1 q r,
E' =
43180 r'3
and in the X'Y' plane we have
1 q 1 q
E' x', E' = (1)
x 4nso r'3 43180 r' 3
Let us now perform the reverse transition to the initial system K,
which moves relative to the system K' with the velocity —v. At the
moment when the charge passes through the origin of the system K,
the projections x and y of vector r are connected with the projections
x' and y' of vector r' through the following relations:
x= r cos 0=x'171-132 y=r sin 0=y' (2)
where R = v/c. Here we took into account that the longitudinal dimen-
sions undergo the Lorentz contraction, while the transverse dimensions
do not change. Besides, in accordance with the transformations inverse
to (8.2), we obtain
E x = E. , Ey -= EYI 171 — (32.
Substituting into these expressions (1) where x' and y' are replaced
by the corresponding expressions from formulas (2), we get
Ex —
1 q x 1 q u
E, -
4318 0 r' 3 ,/1—S2 ' .1 4a-cso r '3 -Vi_pz
It should be noted that ExlEy := - xly, i.e. vector E is directed

radially along the vector r. The situation is such as if the effect of lag
were absent altogether. But this is true only when v = const. If,
however, the charge moves with an acceleration, the field E is no longer
radial.
K K1 It remains for us to find the magni-
tude of vector E:
q.—...- v
1 q Vx2+y2
E r=11El+ 2 = 3 1-132
11 4nso r'
q4,--o- v Since x2 + y2 =r2 and
(x, 2 y,93/2
Fig. 8.12 ( 1— [32 sine 0 )3/2 ,
=Ts
1-132
in accordance with (2), the electric field intensity will be given by
1 q 1-132
E—
lingo r2 (1— (32 sine 0)3 / 2
• 8.11. Interaction between two moving charges. The relativistic
particles having the same charge q move parallel to each other at the
same velocity v as is shown in Fig. 8.12. The separation between the
particles is 1. Using expression (8.7), find the force of interaction be-
tween the particles.
Solution. In this case, the angle between the vector v of one of the
particles and the direction towards the other particle is 15 = 90°.
Hence, in accordance with formula (8.7), the electric component of
the Lorentz force is given by
1 q2
Fe — qE — , (1)
iingo 12 1,2
while the magnetic component of the Lorentz force is

q 2v2
FM = qvB 11° (2)
43 /2 -1/1 pz '
where we took into account that in the case under consideration B
is related to E through formula (8.5) from which it follows that B
9.1. Faraday's Law. Lenz's Law 217
= vE/c2, c2 = 1/80µo• It should be noted that the ratio

FnalFe = eoluov2 = (v/c)2,
just as in the nonrelativistic case (6.5). It can be seen that as v —>- c,
the magnetic component of the force Fm tends to Fe.
The resultant force of interaction (repulsive force) is given by
q2
F= Fm = 42180 12 1/1— 132.
9. Electromagnetic Induction
9.1. Faraday's Law of Electromagnetic Induction.
Lenz's Law
In the previous chapter it was established that the relation
between the "components" of an electromagnetic field, viz.
electric and magnetic fields, is
mainly determined by the C
choice of the reference system. In
other words, the two components
of electromagnetic field are in-
terrelated. We shall show here
that there exists a still closer
connection between the E- and
B-fields, which is exhibited in
phenomena of electromagnetic
induction.
Faraday's Discovery. In 1831 Fig. 9.1
Faraday made one of the most
fundamental discoveries in electrodynamics—the pheno-
menon of electromagnetic induction. It consists in that
an electric current (called the induced current) appears in
a closed conducting loop upon a change in the magnetic flux
(i.e. the flux of B) enclosed by this loop.
The appearance of the induced current indicates that the
ei
induced electromotive force appears in the loop as a result
of a change of the magnetic flux. It is important here that
el does not at all depend on how the change in the magnetic
flux (T) was realized and is determined only by the rate of
its variation, i.e. by dD/dt. Besides, a change in the sign
218 9. Electromagnetic Induction
of the derivative daoldt leads to a change in the sign or

"direction" of
Faraday found out that the induced current may be gener-
ated in two different ways. This is illustrated in Fig. 9.1,
which shows the coil C with current I (the coil creates the
magnetic field) and the loop L connected to the galvano-
meter G which indicates the induced current.
The first way consists in the displacement of the loop L
(or its separate parts) in the field of the fixed coil C.
In the second case, the loop is fixed but the magnetic
field varies either due to the motion of the coil C or as a result
of a change in the current I in it, or due to both reasons.
In all these cases the galvanometer G indicates the pres-
ence of induced current in the loop L.
Lenz's Law. The direction of the induced current (and
hence the sign of the induced e.m.f.) is determined by Lenz's
law: the induced current is directed so that it counteracts the
cause generating it. In other words, the induced current
creates a magnetic flux which prevents the variation of the
magnetic flux generating the induced e.m.f.
If, for example, the loop L (Fig. 9.1) is drawn to the
coil C, the magnetic flux through the loop increases. The
current induced in the loop in this case is in the clockwise
direction (if we look at the loop from the right). This current
creates the magnetic flux "directed" to the left, which pre-
vents an increase in the magnetic flux, which causes the
induced current.
The same situation takes place if we increase the current
in the coil C, keeping fixed the coil and the loop L. On the
other hand, if we decrease the current in the coil C, the cur-
rent induced in the loop L will reverse its direction (it will
now be directed counterclockwise if we look from the right).
Lenz's law expresses an important physical fact, viz. the
tendency of a system to counteract the change in its state
(electromagnetic inertia).
Faraday's Law of Electromagnetic Induction. According
to this law, the e.m.f. induced in a loop is defined by the
formula
dolo
dt
(9.1)
91. Faraday's Law. Lenz's Law 219
regardless of the cause of the variation of the magnetic flux

embraced by the closed conducting loop. The minus sign
in this equation is connected with a certain sign rule. The
sign of the magnetic flux 0 is determined by the choice of
the normal to the surface S bounded by the loop under
consideration, while the sign of ti depends on the choice
of the positive direction of circumvention of the loop.
As before, we assume that the direction of the normal n
to the surface S and the positive direction of the loop cir-
cumvention are connected through the right-hand screw rule*
(Fig. 9.2). Consequently, when we choose
(arbitrarily) the direction of the normal,
we define the sign of the flux 0 as well as
the sign (and hence the "direction") of the
induced e.m.f.
With such a choice of positive directions
(in accordance with the right-hand screw
ei
rule), the quantities and dO/dt have
opposite signs.
The unit of the magnetic flux is the weber (Wb). If the
rate of variation of the magnetic flux is 1 Wb/s, the e. m.f.
induced in the loop is equal to 1 V [see (9.1)].
Total Magnetic Flux (Magnetic-flux Linkage). If a closed
loop in which an e.m.f. is induced contains not one but N
turns (like a coil), t i will be equal to the sum of the e.m.f.s
induced in each turn. And if the magnetic flux embraced by
each turn is the same and equal to 01, the total flux 0 through
the surface stretched over such a complex loop can be
represented as
N01. (9.2)
This quantity is called the total magnetic flux, or the
magnetic-flux linkage. In this case, the e.m.f. induced in
the loop is defined, in accordance with (9.1), by the formula
— N dtD1 . (9.3)
* If these two directions were connected through the left-hand

screw rule, Eq. (9.1) would not contain the minus sign.
9.2. Origin of Electromagnetic Induction

Let us now consider the physical reasons behind the
appearance of induced e.m.f. and try to derive the law of
induction (9.1) from what we already know. Let us consider
two cases.
A Loop Moves in a Permanent Magnetic Field. Let us
first take a loop with a movable jumper of length / (Fig. 9.3).
Suppose that it is in a uniform mag
netic field perpendicular to the plane
of the loop and directed behind
the plane of the figure. We start
CO-) moving the jumper to the right with
velocity v. The charge carriers in
the jumper, viz, the electrons, will
also start to move with the same
Fig. 9.3 velocity. As a result, each electron
will be acted upon by a magnetic
force F = —e [v x B] directed along the jumper. The
electrons will start moving downwards along the jumper,
which is equivalent to electric current directed upwards.
This is just the induced current. The charges redistributed
over the surfaces of the conductors will create an electric
field which will induce a current in the remaining parts of
the loop.
The magnetic force F plays the role of an extraneous force.
The field corresponding to it is E* = F/(—e) = Iv x B].
It should be noted that this expression can also be obtained
with the help of field transformation formulas (8.4).
Circulation of vector E* around a closed contour gives,
by definition, the magnitude of the induced e.m.f. In the
case under consideration, we have
= —vBl, (9.4)
where the minus sign is taken in accordance with the adopted
sign rule: the normal n to the surface stretched over the
loop was chosen so that it is directed behind the plane of
Fig. 9.3 (towards the field B), and hence, according to
the right-hand screw rule, the positive direction of circum-
vention is the clockwise direction, as is shown in the figure.
In this case, the extraneous field E* is directed against the
9.2. Origin of Electromagnetic induction 221
positive direction of loop circumvention, and hence ei

is a negative quantity.
The product vl in (9.4) is the increment of the area bounded
by the loop per unit time (dS 1 dt). Hence vB1 = B dS 1dt =
= al:01 dt , where at,is the increment of the magnetic flux
through the area of the loop (in our case, c/(1) > 0). Thus,
= —aDldt. (9.5)
It can be proved in the general form that law (9.1) is
valid for any loop moving arbitrarily in a permanent non-
uniform magnetic field (see Problem 9.2).
Thus, the excitation of induced e.m.f. during the motion
of a loop in a permanent magnetic field is explained by the
action of the magnetic force proportional to [v x B],
which appears during the motion of the conductor.
It should be noted by the way that the idea of the circuit
shown in Fig. 9.3 forms the basis of all induction generators
in which a rotor with a winding rotates in an external mag-
netic field.
A Loop Is at Rest in a Varying Magnetic Field. In this
case also, the induced current is an evidence of the fact that
the magnetic field varying with time creates extraneous
forces in the loop. But what are these forces? What is their
origin? Clearly, they cannot be magnetic forces proportional
to [v X since these forces cannot set in motion the
charges that have been at rest (v = 0). But there are no
other forces besides qE and q [v x B]! This leaves the only
conclusion that the induced current is due to the appearance
of an electric field E in the wire. It is this field which is
responsible for the induced e.m.f. in a fixed loop placed
into a magnetic field varying with time.
Maxwell assumed that a magnetic field varying with time
leads to the appearance in space of an electric field regardless
of the presence of a conducting loop. The latter just allows
us to reveal the existence of this electric field due to the
current induced in the loop.
Thus, according to Maxwell, a magnetic field varying
with time generates an electric field. Circulation of vector E
of this field around any fixed loop is defined as
act)
E at . (9.6)
Here the symbol of partial derivative with respect to time

(8A) emphasizes the fact that the loop and the surface
stretched on it are fixed. Since the flux cli) = B dS (the
integration is performed around an arbitrary surface stretch-
ed over the loop we are interested in), we have
a aB
at — [GO —
at dS.
—
In this equality, we exchanged the order of differentiation

with respect to time and integration over the surface, which
is possible since the loop and the surface are fixed. Then
Eq. (9.6) can be represented in the form
()E dl --
aB
at dS. (9.7)
This equation has the same structure as Eq. (6.17) the role of the
vector j being played by the vector —a Blat. Consequently, it can
be written in the differential form the same as Eq. (6.26), i.e.
X E = ---0 B/Ot. (9.8)
This equation expresses a local relation between electric and magnetic
fields: the time variation of magnetic field B at a given point determines
the curl of vector E at the same point. The fact that V X E differs from
zero is an evidence of the presence of the electric field itself.
The fact that the circulation of the electric field induced
by a varying magnetic field differs from zero indicates that
this electric field is not a potential field. Like magnetic field,
it is a vortex field. Thus, electric field can be either a potential
field (in electrostatics) or a vortex field.
In the general case, electric field E may be the sum of the
electrostatic field and the field induced by a magnetic
field varying in time. Since the circulation of the electrostatic
field is equal to zero, Eqs. (9.6)-(9.8) turn out to be valid
for the general case as well, when the field E is the vector sum
of these two fields.
Betatron. The vortex electric field has found a brilliant application
in the induction electron accelerator, viz. betatron. This accelerator
consists of a toroidal evacuated chamber arranged between the poles
of an electromagnet (Fig. 9.4). The variation of the current in the
electromagnet winding creates a varying magnetic field which induces
a vortex electric field. The latter accelerates the electrons and simul-
taneously retains them on an equilibrium circular orbit of a certain
9.2. Origin of Electromagnetic Induction
radius (see Problem 9.5). Since the electric field is a vortex field-
the direction of the force acting on the electrons always coincides with
the direction of motion, and the electrons
continuously increase their energy. During the
time when the magnetic field increases 1
ms), the electrons manage to make about a
million turns and acquire an energy up to 400
MeV (at such energies, the electrons' veloc-
ity is almost equal to the velocity of light
c in vacuum).
The induction accelerator (betatron) re-
sembles a transformer in which the role of
the secondary winding consisting of a single Fig. 9.4
turn is played by an electron beam.
Conclusion. Thus, the law of electromagnetic induction

(9.1) is valid when the magnetic flux through a loop varies
either due to the motion of the loop or the time variation
of the magnetic field (or due to both reasons). However,
we had to resort to two quite different phenomena for ex-
plaining the law in these two cases: for the moving loop we
used the action of the magnetic force proportional to [v X B]
while for the magnetic field varying with time, Divot,
the concept of vortex electric field E was employed.
Since there is no unique and profound principle which
would combine these phenomena, we must interpret the law
of electromagnetic induction as the combined effect of two
entirely different phenomena. These phenomena are gen-
erally independent and nevertheless (which is astonishing)
the e.m.f. induced in a loop is always equal to the rate of va-
riation of the magnetic flux through this loop.
In other words, when the field B varies with time as well
as the configuration or arrangement of the loop in the field,
the induced e.m.f. should be calculated by formula (9.1)
containing on its right-hand side the total time derivative
dOldt that automatically takes into consideration the two
factors. In this connection, law (9.1) can be represented
in the form
E dl = — 04: [v dl. (9.9)
The expression on the right-hand side of this equation is

the total derivative &ill:11dt. Here the first term is due to
the time variation of the magnetic field, while the second
is due to the motion of the loop. The origin of the second

term is explained in greater detail in Problem 9.2.
Possible Difficulties. Sometimes we come across a situ-
ation when the law of electromagnetic induction in the form
(9A) turns out to be inapplicable (mainly because of the
difficulties associated with the choice of the contour itself).
Fig. 9.5 Fig. 9.6
In such cases it is necessary to resort to the basic laws, viz.

the Lorentz force qE q [v x B] and the law 17 x E =
—aB/at. These laws express the physical meaning of the
law of electromagnetic induction in all cases.
Let us consider two instructive examples.
Example 1. A conducting strip is displaced at a velocity v through
a region in which a magnetic field B is applied (Fig. 9.5). In the
figure, this region is shown by the dotted circle where the field B is
directed towards us. A galvanometer G is connected to fixed contacts
(arrows) touching the moving strip. Will the galvanometer indicate
the presence of a current?
At first sight, the question seems not so simple, since in this case it
is difficult to choose the contour itself: it is not clear where it should
be "closed" in the strip and how this part of the contour will behave
during the motion of the strip. If, however, we resort to the Lorentz
force, it becomes immediately clear that the electrons in the moving
strip will be displaced upwards, which will create an electric current
in the galvanometer circuit, directed clockwise.
It should be noted that the idea of this experiment formed the basis
of MHD generator in which the internal (thermal) energy is directly
converted into electric energy. Instead of a conducting strip, a plasma
(consisting of electrons and positive ions) is blown at a high speed
through the generator. In all other respects, the situation is the same
as in the case of the conducting strip.
Example 2. The dotted circle in Fig. 9.6. shows the region in which
9.2. Origin of Electromagnetic Induction 225
a permanent magnetic field B is localized (i t is directed perpendicularly

to the plane of the figure). This region is encircled by a fixed metallic
ring R. By moving the sliding contacts to the other side of the ring
(see Fig. 9.6), we introduce the magnetic flux 0 into, the closed con-
tour containing a galvanometer G (1—initial position, 2—final posi-
tion). Will the galvanometer show a current pulse?
Applying formally the law (9.1), we should conclude that there
will be an induced current. But it is not so. There is no current since
in this case both aslat and the Lorentz force are equal to zero: the
magnetic field B is constant and the closed loop moves in the region
where there is no magnetic field. Thus, we have none of the physical
reasons underlying the law of electromagnetic induction.
On an Apparent Paradox. We know that the force acting
on an electric charge in a magnetic field is perpendicular
to its velocity and hence accomplishes no work. Meanwhile,
during the motion of a current-carrying conductor (moving
charges!) Ampere's forces undoubtedly accomplish some
work (electric motor!). What is the matter?
This seeming contradiction disappears if we take into
account that the motion of the current-carrying conductor
in the magnetic field is inevitably accompanied by electro-
magnetic induction. And since the e.m.f. induced in the
conductor performs work on the charges, the total work of
the forces of the magnetic field (the work done by Ampere's
forces and by the induced e.m.f.) is equal to zero. Indeed,
upon an elementary displacement of a current loop in the
magnetic field, Ampere's forces perform the work (see
Sec. 6.8)
SA A = I dIJ (9.10)
while the induced e.m.f. accomplishes during the same time
the work
SA1 = dt —I dcli) (9.11)
where we took into account that ei = — dOldt. It follows
from these two formulas that the total work is
SA A + SA1 = 0. (9.12)
Thus, the work of the forces of the magnetic field includes
not only the mechanical work (due to Ampere's forces)
but also the work done by the e.m.f. induced during the
motion of the loop. These works are equal in magnitude and
opposite in sign, and hence their sum is equal to zero.
15-0181
The work of Ampere's forces is done not at the expense

of the energy of the external magnetic field but at the expense
of the power supply which maintains the current in the loop.
In this case, the source performs additional work against
the induced e.m.f., equal to 6A ad = —6iI dt = I dcl),
which turns out to be the same as the work SA A of Ampere's
forces.
The work SA which is done during the displacement of
the loop against the impeding Ampere's forces (which
appear due to the current induced in accordance with Lenz's
law) is transformed into the work of the induced e.m.f.:
SA = —SA A = SA i. (9.13)
From the point of view of energy, this is the essence of the
operation of all induction generators.
9.3. Self-induction
Electromagnetic induction appears in all cases when the
magnetic flux through a loop changes. It is not important
at all what causes the variation of this flux. If in a certain
loop there is a time-varying current the magnetic field of
this current will also change. And this leads to the variation
of the magnetic flux through the contour, and hence to the
appearance of induced e.m.f.
Thus, the variation of current in a circuit leads to the
appearance of an induced e.m.f. in this circuit. This phe-
nomenon is called self-induction.
Inductance. If there are no ferromagnetics in the space
where a loop with current I is located, field B and hence
total magnetic flux T through the contour are proportional
to the current I, and we can write
€I30= LI (9.14)
where L is the coefficient called the inductance of the circuit.
In accordance with the adopted sign rule for the quantities
and I, it turns out that al and I always have the same
signs. This means that the inductance L is essentially a
positive quantity.
The inductance L depends on the shape and size of the
loop as well as on the magnetic properties of the surrounding
medium. If the loop is rigid and there are no ferromagnetics

in its neighbourhood, the inductance is a constant quantity
independent of the current I.
The unit of inductance is henry (H). In accordance with
(9.14) the inductance of 1 H corresponds to the loop the
magnetic flux through which is equal to 1 Wb at the current
of 1 A, which means that 1 H = 1 Wb/A.
Example. Find the inductance of a solenoid, neglecting edge effects.
Let the volume of the solenoid be V, the number of turns per unit
length n, and the magnetic permeability of a substance inside the
solenoid R.
According to (9.14), L = V/. Consequently, the problem is re-
duced to the determination of the total magnetic flux (1), assuming
that the current I is given. For the current I, the magnetic field in the
solenoid is B = pp,on/. The magnetic flux through a turn of the sole-
noid is (I), = BS = lip onIS, while the total magnetic flux piercing N
turns is
(1) = N(1)1 = nl..BS = p.p,on2 V7,
where V = Si. Hence the inductance of the solenoid is
L = Kton2V. (9.15)
On Certain Difficulties. It should be noted that the deter-
mination of inductance with the help of formula L
is fraught with certain difficulties. No matter how thin the
wire may be, its cross-sectional area is finite, and we simply
do not know how to draw in the body of the conductor a geo-
metrical contour required for calculating D. The result
becomes ambiguous. For a sufficiently thin wire this ambigu-
ity is immaterial. However, for thick wires the situation is
entirely different: in this case, the result of calculation of L
may contain a gross error due to the indeterminacy in the
choice of the geometrical contour. This should be always
kept in mind. We shall later show (see Sec. 9.5) that there
exists another method of determining L, which is not subject
to the indicated drawbacks.
E.M.F. of Self-induction. According to (9.1), a variation
of the current in the circuit leads to the appearance of the
e.m.f. of self-induction es
c/(I) „
j i )• (9.16)
68 = —
d t — —
dt (
i
If the inductance L remains constant upon a variation of

the current (the configuration of the circuit does not change
15*
and there are no ferromagnetics), then

=—L dt (L =const). (9.17)
The minus sign here indicates that es is always directed

so that it prevents a change in the current (in accordance
with the Lenz's law). This e.m.f. tends to keep the current
constant: it counteracts the current when it increases and
sustains it when it decreases. In self-induction phenomena,
the current has "inertia". Therefore, the induction effects
strive to retain the magnetic flux constant just in the same
way as mechanical inertia strives to preserve the velocity
of a body.
Examples of Self-induction Effects. Typical manifesta-
tions of self-induction are observed at the moments of con-
nection or disconnection of an electric circuit. The increase
in the current before it reaches the steady-state value after
closing the circuit and the decrease in the current when
the circuit is disconnected occur gradually and not instant-
aneously. These effects of lag are the more pronounced the
larger the inductance of the circuit.
Any large electromagnet has a large inductance. If its
winding is disconnected from the source, the current rapidly
diminishes to zero and creates during this process a huge
e.m.f. This often leads to the appearance of a Voltaic arc
between the contacts of the switch which is not only very
dangerous for the electromagnet winding but may even prove
fatal. For this reason, a bulb with the resistance of the
same order of magnitude as that of the wire is normally
connected in parallel to the electromagnet winding. In this
case, the current in the winding decreases slowly and is not
a hazard.
Let us now consider in greater detail the mode of vanishing
and establishment of the current in a circuit.
Example 1. Current collapse upon the disconnection of a circuit.
Suppose that a circuit consists of a coil of constant inductance L,
a resistor R, an ammeter A, an e.m.f. source '6 and a special key K
(Fig. 9.7a). Initially, the key K is in the lower position (Fig. 9.7b)
and a current /0 = bIR flows in the circuit (we assume that the re-
sistance of the source of e.m.f. is negligibly small).
At the instant t = 0, we rapidly turn key K clockwise from the
lower to the upper position (Fig. 9.7a). This leads to the following:
the key short-circuits the source for a very short time and then dis-
connects it from the circuit without breaking the latter.
The current through the inductance coil L starts to decrease, which
leads to the appearance of self-induction e.m.f. = —L dlldt
Fig. 9.8
which, in accordance with Lenz's law, counteracts the decrease in

the current. At each instant of time, the current in the circuit will be
determined by Ohm's law I = W S IR, or
RI = - h (9.18)
dt
Separating the variables, we obtain
dl R
--a,t.
I L
Integration of this equation over I (between / 0 and I) and over t
(between 0 and t) gives In (///b) = —RLIt, or
1= _re-"' (9.19)
where v is a constant having the dimension of time:
= Vil. (9.20)
It is called the time constant (or the relaxaton time). This quantity
characterizes the rate of decrease in the current: it follows from (9.19)
that v is the time during which the current decreases to (1/e) times its
initial value. The larger the value of v, the slower the decrease in the
current. Figure 9.8 shows the curve of the dependence I(t) describing
the decrease of current with time (curve 1).
Example 2. Stabilization of current upon closure of a circuit.
At the instant t = 0, we rapidly turn the switch S counterclock-
wise from the upper to the lower position (Fig. 9.7b). We thus con-
nected the source to the inductance coil L. The current in the cir-
cuit starts to grow, and a self-induction e.m.f., counteracting this
increase, will again appear. In accordance with Ohm's law, RI =
=-- Ws, or
dI
RI=W—L (9.21)
Tir •
We transpose W to the left-hand side of this equation and introduce
a new variable u = RI — W, du = R dI. After that, we transform
the equation thus obtained to
dulu = —dart
where T = Lai is the time constant.
Integration over u(between — W and RI — W) and over t (be-
tween 0 and t) gives In [(RI — g)I( — W)] — tit, or
= I, (1—e-0), (9.22)
where /0 = W/.// is the value of the steady-state current (for t 00).
Equation (9.22) shows that the rate of stabilization of the currentis
determined by the same constant T. The curve I(t) characterizing the
increase in the current with time is shown in Fig. 9.8 (curve 2).
On the Conservation of Magnetic Flux. Let a current
loop move and be deformed in an arbitrary external mag-
netic field (permanent or varying). The current induced in
the loop in this case is given by
•+ di:130
= —
R R dt •
If the resistance of the circuit R = 0, c101dt must also be
equal to zero, since the current I cannot be infinitely large.
Hence it follows that cI) = const.
Thus, when a superconducting loop moves in a magnetic
field, the magnetic flux through its contour remains constant.
This conservation of the flux is ensured by induced currents
which, according to Lenz's law, prevent any change in the
magnetic flux through the contour.
The tendency to conserve the magnetic flux through
a contour always exists but is exhibited in the clearest form
in the circuits of superconductors.
Example. A superconducting ring of radius a and inductance L
is in a uniform magnetic field B. In the initial position, the plane of
the ring is parallel to vector B, and the current in the ring is equal
to zero. The ring is turned to the position perpendicular to vector B.
Find the current in the ring in the final position and the magnetic
induction at its centre.
The magnetic flux through the ring does not change upon its rota-
tion and remains equal to zero. This means that the magnetic fluxes
of the field of the induced current and of the external current through
the ring are equal in magnitude and opposite in sign. Hence LI
rta2B, whence
I = na2BIL.
This current, in accordance with (6.13), creates a field B1 =- agoa/3/2L
at the centre of the ring. The resultant magnetic induction at this
point is given by
Bres= B— B1= B (1 — riot/24
9.4. Mutual Induction

Mutual Inductance. Let us consider two fixed loops 1
and 2 (Fig. 9.9) arranged sufficiently close to each other.
If current I, flows in loop 1, it creates through loop 2
the total magnetic flux 0 2 proportional (in the absence of
ferromagnetics) to the current /1:
02 = L21/1. (9.23)
Similarly, if current /2 flows in
loop 2, it creates through the
contour 1 the total magnetic
flux
= L1212. (9 . 24)
The proportionality factors L12 Fig. 9.9
and L21 are called mutual induc-
tances of the loops. Clearly, mutual inductance is nu-
merically equal to the magnetic flux through one of the
loops created by a unit current in the other loop. The coef-
ficients L 21 and L12 depend on the shape, size, and mutual
arrangement of the loops, as well as on the magnetic perme-
ability of the medium surrounding the loops. These coef-
ficients are measured in the same units as the inductance L.
Reciprocity Theorem. Calculations show (and experiments
confirm) that in the absence of ferromagnetics, the coef-
ficients L12 and L21are equal:
IL 12 = L21. I (9.25)
This remarkable property of mutual inductance is usually
called the reciprocity theorem. Owing to this theorem, we
do not have to distinguish between L12 and L 21and can

simply speak of the mutual inductance of two circuits.
The meaning of equality (9.25) is that in any case the
magnetic flux 01through loop 1, created by current I
in loop 2, is equal to the magnetic flux 0 2 through loop 2,
created by the same current I in loop 1. This circumstance
often allows us to considerably simplify the calculation,
for example, of magnetic fluxes. Here are two examples.
Example 1. Two circular loops 1
and 2 whose centres coincide lie in a
plane (Fig. 9.10). The radii of the loops
are al and a2. Current I flows in
loop 1. Find the magnetic flux 02 em-
braced by loop 2, if al<< a2.
The direct calculation of the flux (132
is clearly a rather complicated problem
since the configuration of the field
itself is complicated. However, the
application of the reciprocity theorem
IA c greatly simplifies the solution of the
problem. Indeed, let us pass the same
current I through loop 2. Then the
magnetic flux (D i created by this cur-
/ rent through loop 1 can be easily
eiZZA found, provided that al<< a2: it is suf-
Fig. 9.11 ficient to multiply the magnetic induc-
tion B at the centre of the loop (B =
=R0//2a2) by the area ytai of the circle
and take into account that cl)2 = (131in accordance with the reci-
procity theorem.
Example 2. A loop with current I has the shape of a rectangle.
Find the magnetic flux 1 through the hatched half-plane (Fig. 9.11)
whose boundary is at a given distance from the contour. Assume that
this half-plane and the loop are in the same plane.
In this case too, the magnetic field of current I has a complex con-
figuration, and hence it is very difficult to directly calculate the flux el)
in which we are interested. However, the solution can be considerably
simplified by using the reciprocity theorem.
Suppose that current I flows not around the rectangular contour
but along the boundary of the half-plane, enveloping it at infinity
The magnetic field created by this current in the region of the rectan-
gular loop has a simple configuration—this is the field of a straight
current. Hence we can easily find the magnetic flux c131' through the
rectangular contour (with the help of simple integration). In accordance
with the reciprocity theorem, the required flux a) = c1' and the
problem is thus solved.
However, the presence of ferromagnetics changes the
situation, and the reciprocity theorem becomes inapplicable.

Let us verify this with the help of the following concrete
example.
Example. A long ferromagnetic cylinder of volume V has two wind-
ings (one over the other). One winding contains n1turns per unit
length, while the other contains n2turns. Find their mutual inductance
ignoring the edge effects.
According to (9.23), L21 = 02//i. This means that we must create
current I1in winding 1 and calculate the total magnetic flux through
all the turns of winding 2. If winding 2 contains N2 turns, we obtain
= N2B1S,
where S is the cross-sectional area of the cylinder. Considering that
N2 = n21, where 1 is the cylinder length, and B1 = p,ittoni/,, where
p1is the magnetic permeability for current Il, we write '2 =
= ttitto nin2 V/1, where V = IS. Hence
L21 = mon1n2 V.
Similarly, we can find L12:
L12 = 1121-Lon1n2V.
Since the values of th and 112 in the last two expressions are generally
different (they depend on currents 71and 72 in ferromagnetics), the
values of L21 and L12 do not coincide.
Mutual Induction. The presence of a magnetic coupling
between circuits is manifested in that any variation of
current in one of the circuits leads to the appearance of
an induced e.m.f. in the other circuit. This phenomenon
is called mutual induction.
According to Faraday's law, the e.m.f.'s appearing in
contours 1 and 2 are given by
dI
1 —L 2
dt 12 dt
aq) 2 dI1
dt
L zi dt
(9.26)
respectively. Here we assume that the contours are fixed

and there are no ferromagnetics in the surroundings.
Taking into account self-induction, the current appearing,
say, in contour 1 upon a change in currents of both circuits
is determined according to Ohm's law through the formula
di1 r dI2
R1I1= - LI
dt - dt
where ei is the extraneous e.m.f. in contour 1 (besides

induced e.m.f.$) and L1is the inductance of contour 1.
A similar equation can be written for the current I2 in
contour 2.
It should be noted that transformers, devices intended
for transforming currents and voltages, are based on the
principle of induction.
On the Sign of L22. Unlike inductance L which, as was
mentioned earlier, is essentially a positive quantity, mutual
inductance L12 is an algebraic quantity (which may, in
particular, be equal to zero). This is
due to the fact that, for instance, in
(9.23) the quantities '2and refer
to different contours. It immediately
follows from Fig. 9.9 that the sign
L12 >0 L12 < 0 of the magnetic flux cl), for a given
direction of current Il will depend on
(a) (b) the choice of the normal to the sur-
face bounded by contour 2 (or on the
Fig. 9.12 choice of the positive direction of cir-
cumvention of this contour).
The positive directions for currents (and e.m.f.$) in both
contours can always be chosen arbitrarily (and the positive
direction of contour circumvention is uniquely (through
the right-hand screw rule) related to the direction of the
normal n to the surface bounded by the contour, i.e. in the
long run with the sign of the magnetic flux). As soon as
these directions are specified, the quantity L12 should be
assumed to be positive if the fluxes of mutual induction
through the contours are positive for positive currents,
i.e. when they coincide in sign with self-induction fluxes.
In other words, L12 > 0 if for positive currents the con-
tours "magnetize" each other. Otherwise, L12 <0. In special
cases, the positive directions of circumvention of the con-
tours can be established beforehand so that the required
sign of the quantity L12 can be obtained (Fig. 9.12).
9.5. Magnetic Field Energy

Magnetic Energy of Current. Let us connect a fixed circuit
containing induction coil L and resistor R to a source of
e.m.f. eo. As we already know, the current in the circuit

will start increasing. This leads to the appearance of a self-
induced e.m.f. es. In accordance with Ohm's law, RI
=o+ es , whence
to = RI — es.
Let us find the elementary work done by extraneous forces
(i.e. the source of eo) during time dt. For this purpose, we
multiply the above equality by I dt:
dt = RI2dt — es/ dt.
Taking into account the meaning of each term and the rela-
tion es = _doidt, we can write
OA extr = 6Q / c/(I).
It can be seen that in the process of current stabilization,
when the flux cl) varies so that da) > 0 (if I > 0), the work
done by the source of eoturns out to be greater than the
Joule heat liberated in the circuit. A part of this work
(additional work) is performed against the self-induced
e.m.f. It should be noted that after the current has been
stabilized, dD = 0, and the entire work of the source of to
will be spent for the liberation of the Joule heat.
Thus, the additional work accomplished by extraneous
forces against the self-induced e.m.f. in the process of current
stabilization is given by
6A add =-l d(1). (9.27)
This relation is of a general nature. It is valid in the

presence of ferromagnetics as well, since while deriving it
no assumptions have been made concerning the magnetic
properties of the surroundings.
Here (and below) we shall assume that there are no ferro-
magnetics. Then c/(13 = L dI, and
6A add = LI dl. (9.28)
Having integrated this equation, we obtain A add =
= LI212. According to the law of conservation of energy,
any work is equal to the increment of any kind of energy.
It is clear that a part of the work done by extraneous forces
(to) is spent on increasing the internal energy of conductors

(it is associated with the liberation of the Joule heat), while
another part (during current stabilization) is spent on some-
thing else. This "something" is just the magnetic field, since
its appearance is associated with the appearance of the
current.
Thus, we arrive at the conclusion that in the absence of
ferromagnetics, the contour with the induction coil L
and current I has the energy
T.2
W= 2
LI2
2 2L •
(9.29)
This energy is called the magnetic, or intrinsic, energy of

current. It can be completely converted into the internal
energy of conductors if we disconnect the source of eo
from the circuit as shown in Fig. 9.7, i.e. if we rapidly turn
the key K from position b to position a.
Magnetic Field Energy. Formula (9.29) expresses the
magnetic energy of current in terms of inductance and cur-
rent (in the absence of ferromagnetics). In this case, however,
the energy like the electric energy of charged bodies, can be
directly expressed in terms of magnetic induction B. Let
us show this first for a simple case of a long solenoid, ignoring
field distortions at its ends (edge effects). Substituting the
expression L = uuon2 V into (9.29) we obtain
W = (1/2)L/2 = (1/2)u.Ron2/2 V.
And since nI = H = B/uuo, we have
B2 , B • H
= (9.30)
21410 v 2
This formula is valid for a uniform field filling the volume V
(as in the case of the solenoid under consideration).
The general theory shows that the energy W can be ex-
pressed in terms of B and H in any case (but in the absence
of ferromagnetics) through the formula
(' B • H
— dV " (9.31)
2
The integrand in this equation has the meaning of the

energy contained in the volume element dV.
Thus, as in the case of the electric field energy, we arrive

at the conclusion that the magnetic energy is also localized
in the space occupied by a magnetic field.
It follows from formulas (9.30) and (9.31) that the mag-
netic energy is distributed in space with the volume density
B•H B2
w .= (9.32)
2 2p110 •
It should be noted that this expression can be applied

only to the media for which the B vs. H dependence is
linear, i.e. p. in the relation B = 1,14t 0H is independent of H.
In other words, expressions (9.31) and (9.32) are applicable
only to paramagnetics and diamagne tics. In the case of
ferromagnetics, these expressions are inapplicable*.
it should also be noted that
the magnetic energy is an essen- dl
tially positive quantity. This
can be easily seen from the last
two formulas.
Another Approach to the Sub-
stantiation of Formula (9.32). Let
us prove the validity of this formula
by proceeding "the other way round",
i.e. let us show that if formula (9.32) Fig. 9.13
is valid, the magnetic energy of the
current loop is W = L1212.
For this purpose, we consider the magnetic field of an arbitrary
loop with current I (Fig. 9.13). Let us imagine that the entire field is
divided into elementary tubes whose generatrices are the field lines
of B. We isolate in one such tube a volume element dV = dl dS.
1
According to formula (9.32), the energy 2 BH dl dS is localized in
—
this volume.
Let us now find the energy dW in the volume of the entire elementa-
ry tube. For this purpose, we integrate the latter expression along
the tube axis. The flux c/(13 = B dS through the tube cross-section is
* This is due to the fact that in the long run expressions (9.31)
and (9.32) are the consequences of the formula 6,4 add = I dt and of
the fact that, in the absence of hysteresis, the work 6./1 add is spent only
on increasing the magnetic energy dW. For a ferromagnetic medium,
the situation is different: the work 611. addis also spent on increasing
the internal energy of the medium, i.e. on its heating.
constant along the tube, and hence c/0 can be taken out of the integral:
d0 da
dW H
2 , 2•
Here we used the theorem on circulation of vector H (in our case, pro-
jection H1=--- H) .
Finally, we sum up the energy of all elementary tubes:
W= cc/0 = /0/2= L/2/2,

2
where 0 = LI is the total magnetic flux enveloped by the current
loop, Q.E.D.
Determination of Inductance from the Expression for

Energy. We have introduced the inductance L as the pro-
portionality factor between the total magnetic flux (I) and
current I. There is, however, another possibility of calculat-
ing L from the expression for energy. Indeed, comparing
formula's (9.31) and (9.29), we see that in the absence of
ferromagnetics,
L= PJ C11.110
B2 dV. (9.33)
The value of L found in this way is free of indeterminacy

associated with the calculation of the magnetic flux (I)
in formula (9.14) (see p. 227). The discrepancy appearing
sometimes in determining L through formula (9.33) and
from the expression (9.14) for the flux is illustrated in
Problem 9.9 about a coaxial cable.
9.6. Magnetic Energy of Two Current Loops

Intrinsic and Mutual Energies. Let us consider two fixed
loops 1 and 2, arranged near one another at a sufficiently
small distance (so that there is a magnetic coupling between
them). We assume that each loop includes its own source
of constant e.m.f. Let us close each loop at the instant
t = 0. In each loop, its own current will start being estab-
lished, and hence self-induced e.m.f. es and the e.m.f. ei
of mutual induction will appear. The additional work done
in this case by the sources of constant e.m.f. against es
ei
and is spent, as was shown above, for creating the mag-
netic energy.
9.6. Magnetic Energy of Two Current Loops 239
Let us find this work over the time dt:

Madd = + ea) 11 dt --I- w, 1 122 dt = dT/V
s-s2 . -12, dW .
Considering that es, = —L1d/1/dt, eii = —L12 d12/dt,
etc., we transform this formula as follows:
dW = L111 d11 +L1211 d12 L212 d12 + L2112 d11.
Taking into account the fact that L12 = L21, we represent
this expression in the form
dW = d (L1I;.12) d (L 21:12)
d (L121112),
whence
L L
147 - 22 2 +L121112: (9.34)
The first two terms in this expression are called the intrinsic
energies of currents I1and 12, while the last term, the mutual
energy of the currents. Unlike the intrinsic energies of cur-
rents, the mutual energy is an algebraic quantity. A change
in the direction of one of the currents leads to the reversal
of the sign of the last term in (9.34), viz. the mutual energy.
Example. Suppose that we have two concentric loops with currents
Il and /2whose directions are shown in Fig. 9.14. The mutual energy
of these currents ( W12 = L121112) depends on three algebraic quantities
whose signs are determined by the choice of positive directions of cir-
cumvention of the two loops. It is useful to verify, however, that the
sign of W12 (in this case W12 > 0) depends only on the mutual orien-
tation of the currents themselves and is independent of the choice of
the positive directions of circumvention of the loops. We recall that
the sign of the quantity L12 was considered in Sec. 9.4.
Field Treatment of Energy (9.34). There are some more
important problems that can be solved by calculating the
magnetic energy of two loops in a different way, viz. from
the point of view of localization of energy in the field.
Let B1 be the magnetic field of current /1, and B2 the
field of current /2. Then, in accordance with the principle
of superposition, the field at each point is B = B1 +B2 ,
and according to (9.31), the field energy of this system of
currents is W = (B2/2p,Ro) dV. Substituting into this
formula B2 = /3,2 ± -22 + 213032, we obtain
W=1 dV 2Nto
dV 131.132 dV . (9.35)
The correspondence between individual terms in formulas

(9.35) and (9.34) is beyond doubt.
Formulas (9.34) and (9.35) lead to the following important
sequences.
1. The magnetic energy of a system of two (or more)
currents is an essentially positive quantity (W > 0). This
follows from the fact that W OC B2 dV, where the in tegrand
contains positive quantities.
2. The energy of currents is a nonadditive quantity (due
to the existence of mutual energy).
3. The last integral in (9.35) is proportional to the pro-
duct I „I: 2of currents since B1(:)( I, and B OC
2 12. The pro-
portionality factor (i.e. the remaining integral) turns out
to be symmetric with respect to indices 1 and 2 and hence
can be denoted by L12 or L21in accordance with formula
(9.34). Thus, L12 =L21 indeed.
4. Expression (9.35) leads to another definition of mutual
inductance L12. Indeed, a comparison of (9.35) and (9.34)
gives
1 B 1 .B2 dv
L i2 = 1112 (9.36)
J 1-1110 •
9.7. Energy and Forces in Magnetic Field

The most general method for determining the forces acting
in a magnetic field is the energy method, in which the expres-
sion for the magnetic field energy is used.
We shall confine ourselves to the case when the system
consists of two loops with currents /1 and /2. The magnetic
energy of such a system can be represented in the form
W = (/).€121, + 102), (9.37)
where 1J1 and 4:1)1 are the total magnetic fluxes piercing
loops 1 and 2 respectively. This expression can be easily
obtained from formula (9.34) by representing the last term
in it as the sum (1/2) L121112 +(1/2) L „I 211and then taking
into account that

(1) = L11, + L1212, (1)2 = L2-12 L2,/1. (9.38)
According to the law of conservation of energy, the work
611* done by the sources of e.m.f. included into circuits 1
and 2 is converted into the heat SQ and is spent to increase
the magnetic energy of the system by dW (due to the motion
of the loops or the variation of currents in them as well as
to accomplish the mechanical work SA„, (as a result of a
displacement or deformation of the loops):
SA* = SQ dW SAm. (9.39)
We assumed that the capacitance of the loops is negligibly
small and hence the electric energy can be ignored.
Henceforth, we shall be interested not in the entire work
SA* of the source of e.m.f. but only in its part which is done
against the induced and self-induced e.m.f.s (in each loop).
This work (which was called the additional work) is equal
to Ma" = esi) dt — (82 ei2
+ 6 ) 12 dt.
Considering that ei+ es = —daildt for each loop, we
can write the expression for the additional work in the form
SA add =Il da1 + /2 d02. (9.40)
It is just this part of the work of the e.m.f. sources (the
work against the induced and self-induced e.m.f.$) that
is associated with the variation of the fluxes 01 and 02
and spent for increasing the magnetic energy of the system
as well as for accomplishing the mechanical work:
Il del + dW SA,,,. (9.41)
This formula forms the basis for calculating the mechanical
work SA,„ and then the forces acting in a magnetic field.
Formula (9.41) can be used for obtaining simpler expres-
sions for SA,, by assuming that either all the magnetic fluxes
through the loops, or currents flowing in them remain un-
changed during a displacement. Let us consider this in
greater detail.
1. If the fluxes are constant (0h = const), it immediately
16-0181
follows from (9.41) that

I SA RI = — dW 10, I (9.42)
where the subscript €1) indicates that the increment of the
magnetic energy of the system must be calculated at constant
fluxes through the loops. The obtained formula is similar
to the corresponding formula (4.15) for the work in an elec-
tric field.
2. If the currents are constant (b, = const), we have
dW 11.1 (9.43)
Indeed, for Ik = const, it follows from formula (9.37) that
dW I I = d(D, + I2 d(I) 2),
i.e. in this case, in accordance with (9.40), the increment

of the magnetic energy of the system is equal to half of the
additional work done by the sources of e.m.f. Another half
of this work is spent for doing mechanical work. In other
words, when the currents are constant, 01711= &A m,
Q.E.D.
It should be emphasized that the two expressions (9.42)
and (9.43) obtained by us define the mechanical work of
the same force, i.e. we can write
F•d1 = —dW10 =dWI I. (9.44)
In order to calculate the force with the help of these
formulas, there is no need, of course, to choose a regime in
which either magnetic fluxes or currents would necessarily
remain constant. We must simply find the increment dW
of magnetic energy of the system provided that either Oh =
= const or Ih = const, which is a purely mathematical
operation.
The value of the obtained expressions (9.42) and (9.43)
lies in their generality: they are applicable to systems con-
sisting of any number of contours—one, two, or more.
Let us consider several examples illustrating the applica-
tion of these formulas.
Example 1. Force in the case of one current loop. Suppose that
we have a current loop, where AB is a movable jumper (Fig. 9.15).
The inductance of this loop depends in a certain way on the coor-
dinate x, i.e. L (x) is known. Find the Ampere's force acting on the
jumper, in two ways: for I = const and for (13 = const.
A
0 x X
B
Fig. 9.15 Fig. 9.16
In the case under consideration, the magnetic energy of the system

can be represented, in accordance with (9.29), as follows:
W = L/2/2 = cP2/2L,
where cl) = LI. Let us displace the jumper, for example, to the right
by dx. Since 6Am = Fx dx, we have
OW _ 12 al,
Fx=
ax 1 2 ax
Or
aw _ aL 12 at,
Fx ax icp 2L2 ax -= 2 ax
i.e. the calculations carried out with the help of the two formulas
give, in accordance with (9.44), the same result.
Example 2. Interaction between two current-carrying coils. Two
coils 1 and 2 with currents Il and /2 are mounted on a magnetic core
(Fig. 9.16). Suppose that the mutual inductance of the coils depends
on their separation x according to a known law L12 (x). Find the force
of interaction between the coils.
The magnetic energy of the system of two coils is given by formula
(9.34). For determining the force of interaction, we shall use expres-
sion (9.43). Let us displace coil 2 through a distance dx at constant
currents /1 and /2. The corresponding increment of the magnetic
energy of the system is
dW i i =1112 dLi2 (x)•
Since the elementary mechanical work 6Am = F2x dx, according to
(9.43) we obtain
r aLi2(x)
F2X It- 2
ax •
Let currents /1and 72 magnetize each other. Then L12 > 0, and for
dx > 0 the increment dL12 < 0, i.e. F2x <0. Consequently, the force
exerted on coil 2 by coil 1 is the attractive force: vector F2 is directed
to the left in the figure.
16*
244 9. Electromagnetic induction
Example 3. Magnetic pressure acting on solenoid winding.

Let us mentally increase the radius of the solenoid cross section
by dr, retaining the current I through the winding constant. Then
Ampere's forces will accomplish the work 13,4m = dW In the case
under consideration, we have
6A m = pS dr,
where p is the required pressure and S is the lateral surface of the
solenoid, and
B2 B2
dW l i =d(- 21 0 V) = 2110 S dr.
Here we took into account the fact that when I = coast, B = coast
as well. Equating these two expressions, we obtain
p = B2/2110.
Magnetic Pressure. The expression for pressure obtained
in Example 3 can be generalized for the case when the mag-
netic field is different (B1and B2) on different sides of a sur-
face with current (conduction current or magnetization cur-
rent). In this case, the magnetic pressure is given by
p = Bi .Hi B2.H2
(9.45)
2 2
The situation is such as if the region with a higher magnetic
energy density were the region of higher pressure.
Relation (9.45) is one of the basic relations in magneto-
hydrodynamics which studies the behaviour of electroconduct-
ing liquids (in electrical engineering and astrophysics).
Problems
• 9.1. Induced e.m.f. A wire in the shape of a parabola y = kx2
is in a uniform magnetic field B perpendicular to the plane XY.
A jumper translates without initial velocity and at a constant acceler-
ation a from the apex of the parabola (Fig. 9.17). Find the e.m.f.
induced in the formed contour as a function of the coordinate y.
Solution. By definition, el = daVdt. Having chosen the normal n
to the plane of the contour in the direction of vector B, we write c/(1) =
B dS, where dS = 2x dy. Considering now that x = Vylic, we
obtain
gi= —B-2 j/ylk dyldt.
During the motion with a constant acceleration, the velocity dyldt =
= ]/2ay, and hence
—By "Ii8allc•
Problems 245
This formula shows that O( y. The minus sign indicates that 4'i
in the figure acts counterclockwise
• 9.2. A loop moves arbitrarily. A closed conducting loop is moved
arbitrarily (even with a deformation) in a constant nonuniform mag-
netic field. Show that Faraday's law (9.1) will be fulfilled in this case.
0 x
Fig. 9.17
Solution. Let us consider an element dl of loop length, which at

a given instant moves at a velocity v in the magnetic field B. In
accordance with formulas (8.4) of field transformation, in the reference
system fixed to the given element the electric field E = [v X B] will
be observed. It should be noted that this expression can be also ob-
tained with the help of the Lorentz force, as it was done at the beginn-
ing of Sec. 9.2.
Circulation of vector E around the entire contour is, by definition,
the induced e.m.f.:
— [v x B] dl (1)
Let us now find the corresponding increment of the magnetic flux
through the loop. For this purpose, we turn to Fig. 9.18. Suppose that
the loop was displaced from position r1 to ra during time dt. If in
the first position the magnetic flux through the surface Si stretched
over the loop was cl)i, the corresponding magnetic flux in the second
position can be represented as c1301 dcp, i.e. as the flux through the
surface S dS. Here d(I) is the required increment of the magnetic
flux through the narrow strip dS between contours and r2.
Using Fig. 9.18, we can write
B • ds = B • [dr X dl]:= — [dr x 13] dl. (2)
Here (1) the direction of normal n is matched with the direction of

circumvention of the contour, viz. with vector dl (right-handed sys-
tem); (2) the direction of vector ds, viz. the area element of the strip,
is matched with the choice of normals n, and (3) the following cyclic
transposition is used in the scalar triple product:
a• [b x c] = b•[c X a] = c•[a X b] = —[b X a] •c.
246 9. Electromaknecic Induction
Dividing expression (2) by de, we find
dOldt= — [v x B] dl, (3)

where v = dr/dt. It remains for us to compare (3) with (1), which
gives `6i = —dt/dt.
• 9.3. A flat coil with a large number N of tightly wound turns
is in a uniform magnetic field perpendicular to the plane of the coil
(Fig. 9.19). The external radius of the coil is equal to a. The magnetic
field varies in time according to the law B = Bo sin tot. Find the
maximum value of the e.m.f. induced in the coil.
Solution. Since each turn of the coil
practically does not differ from a circle, the
e.m.f. induced in it is given by
Ei = — dt/dt = --ytr2Bo co cos cot,
where r is the radius of the turn under con-
sideration. The number of turns corresponding
to the interval dr of the values of the radius
is dN = (N la) dr. The turns are connected in
series, hence the total e.m.f. induced in the
coil is
Fig. 9.19 Et (r) dN .
Having integrated this expression, we obtain the following for the

maximum value of the induced e.m.f.:
Wuxi = (1/3)2-ca2 NBoco.
• 9.4. A coil of N turns with the cross-sectional area S is placed
inside a long solenoid. The coil is rotated at a constant angular velocity
co around the axis coinciding with its diameter and perpendicular
to the axis of the solenoid. The magnetic field in the solenoid varies
according to the law B = Bo sin cot. Find the e.m.f. induced in the
coil, if at the instant t = 0 the coil axis coincided with the axis of
the solenoid.
Solution. At the instant t, the total magnetic flux through the
coil is
4:11 = NB S cos cot = NBoS sin cot •cos cot = (1/2)NBoS sin aot.
In accordance with Faraday's law, we have
_davat = —(1/2)NB-os • 2o) cos 2tot = —NB0S to cos 2cot.
• 9.5. The betatron condition. Show that electrons in a betatron
will move in an orbit of a constant radius 7-0provided that the magnetic
field Bo on the orbit is equal to half of the value (B) of the magnetic
(B)
field averaged over the area inside the orbit, i.e. Bo =
2 •
Problems 247
Solution. Let us write the relativistic equation for motion of an

electron,
dpIdt = eE e [v X Bo], (1)
where E is the vortex electric field, in terms of the projections onto
the tangent t and the normal n to the trajectory. For this purpose,
I
2
Fig. 9.21
we write the electron momentum as p = pr and find its time derivative

dp (Pr dp ,2
= dp (2)
dt P11-7n--
dt dt ro n'
where we took into account that p = my, m being the relativistic
mass, and &chit = (v/ro)n (this can be easily seen from Fig. 9.20).
Indeed, dv = cicp•n = (v daro)n, and the rest is obvious.
Besides, according to Faraday's law, 2nroE = jc/Vdt1 , where
= nrg (B). Hence
r d
—( (3)
2 2t (B).
Let us now write Eq. (1), taking into account (2) and (3), in terms
of the projections onto the tangent and the normal to the trajectory:
dP -=eE e (4)
dt dt (B)'
777, —=evBo .
ro
The last equation can be written, after cancelling v, in the form
p = eroBo.
We differentiate this equation with respect to time, considering
that ro = const:
dB
dP er ° (5)
dt ° dt •
A comparison of Eqs. (5) and (4) gives

d 1 d ,,
B
– 2. 17 •
In particular, the last condition will be satisfied when
Bo= 2 (11).
Practically, it is attained by manufacturing pole pieces in a special

form (in the form of a blunt-nosed cones).
• 9.6. Induced current. A square wire frame with side a and
a long straight conductor with a direct current / 0 lie in the same plane
(Fig. 9.21). The inductance of the frame is L and its resistance is R.
The frame has been rotated through 180° around the axis 00' and
then stopped. Find the amount of electricity that has passed through
the frame. The distance b between the axis 00' and the straight wire
is assumed to be known.
Solution. According to Ohm's law, the current I appearing in the
frame during its rotation is determined by the formula
dfl20 dI
RI —
dt dt •
Hence the required amount of electricity (charge) is
1 1
q= I dt= (d0 + L dI)= — (AGI)+ L AI).
Since the frame has been stopped after rotation, the current in it
vanishes, and hence Al = 0. It remains for us to find the increment
of the flux AO through the frame (AO = —
Let us choose the normal n to the plane of the frame, for instance,
so that in the final position n is directed behind the plane of the figure
(along B). Then it can be easily seen that in the final position 02 > 0,
while in the initial position cl)i < 0 (the normal is opposite to B),
and AO turns out to be simply equal to the flux through the surface
bounded by the final and initial positions of the frame:
b+a
Ael=02+ I Pl1= Ba dr,
b -a
where B is a function of r, whose form can be easily found with the

help of the theorem on circulation.
Finally, omitting the minus sign, we obtain
AO IA04 b +a
q= 13 = 2zuR In 1)— a •
Problems 249
It can be seen that the obtained quantity is independent of the induct-

ance of the circuit (the situation would be different if the circuit were
superconducting).
• 9.7. A jumper 12 of mass m slides without friction along two long
conducting rails separated by a distance 1 (Fig. 9.22). The system is
in a uniform magnetic field perpendicular to the plane of the circuit.
2 8 K
Fig. 9.22 Fig. 9.23
The left ends of the rails are shunted through resistor R. At the instant
t = 0, the jumper received the initial velocity vo directed to the right.
Find the velocity of the jumper as a function of time, ignoring the
resistances of the jumper and of the rails as well as self-inductance of
the circuit.
Solution. Let us choose the positive direction of the normal n
to the plane of the circuit away from us. This means that the positive
direction of circumvention of the circuit (for induced e.m.f. and cur-
rent) is chosen clockwise, in accordance with the right-hand screw
rule. It follows from Ohm's law that
c/( dS
RI=— — — –= (1)
dt dt
where we took into account that when the jumper moves to the right,
d(I3 > 0. According to Lenz's law, the induced current I causes the
Ampere force counteracting the motion, directed to the left.
Having chosen the X-axis to the right, we write the equation of
motion of the jumper
m dvIdt = I1B, (2)
where the right-hand side is the projection of the Ampere force onto
the X-axis (this quantity is negative, but we omit the minus sign
since, as can be seen from (1), current I < 0).
Eliminating I from Eqs. (1) and (2), we obtain
day = — a dt, a = B2121mR.
The integration of this expression, taking into account the initial
condition, gives
In (vivo) =-- — at, U voe—at.
• 9.8. The role of transient processes. In the circuit shown in

Fig. 9.23, the e.m.f. d of the source, its internal resistance R and the
inductances L1and L2 of superconducting coils are known. Find the
currents established in the coils after key K has been closed.
Solution. Let us use Kirchhoff's laws for electric circuits 6L1
and WL2:
d11 dI,
=-W —Li , RI=W — L2
dt dt •
A comparison of these expressions shows that L1 d11 = L2 d12, while

for stabilized currents we have
L1110 = L2/ 20. (1)
Besides,
/10 + /20 = ./() = (2)
From these equations, we find
`6 L2 'Co' Li
R Li + L2 ,
-120
—R L 1 +L2 •
• 9.9. Calculation of inductance. A coaxial cable consists of
internal solid conductor of radius a and external thin-wall conducting
tube of radius b. Find the inductance of a unit length of the cable,
considering that the current distribution over the cross section of the
internal conductor is uniform. The permeability is equal to unity
everywhere.
Solution. In the case under consideration, the inner conductor is
not thin, and hence the inductance should be determined in terms of
the energy rather than through the magnetic flux. In accordance with
(9.33), we can write
b
L❑ =-- p —
B 22nr dr, (1)
Fro
where r is the distance from the cable axis. In order to evaluate this
integral, we must find the dependence B(r). Using the theorem on
circulation, we have
* NI 1
B — r Ba<r<b Br>b =0• (2)
r<a 22-ca2 ' 231 r '
The form of these dependences is shown in Fig. 9.24. On account of

(2), integral (1) is split into two parts. As a result of integration, we
obtain
Lu 2p,no ( 1 _on b \
aI
It should be noted that an attempt to determine this quantity in
terms of magnetic flux by the formula LI, = (Icy/ leads to a different
Problems 251
(erroneous) result, namely, instead of 1/4, 1/2 is obtained in the paren-

theses. The thinner the central wire, i.e. the larger the ratio b/a,
the smaller the relative difference in the results of calculation by these
two methods, viz. through the energy and through the flux.
• 9.10. Mutual induction. A long straight wire is arranged along
the symmetry axis of a toroidal coil of rectangular cross section, whose
Fig. 9.24 Fig. 9.25
dimensions are given in Fig. 9.25. The number of turns on the coil
is N, and the permeability of the surrounding medium is unity. Find
the amplitude of the e.m.f. induced in this coil if the current I =
= Im cos cot flows along the straight wire.
Solution. The required e.m.f. ei= doidt, where c = ND1
—
and D1is the magnetic flux through the cross section of the coil:
1.10h/ in b
(Di = Bn dS — i 11° lh dr
2n r 22E a
a
where Ba is determined with the help of the theorem on circulation
of vector B. Taking the time derivative of I and multiplying the
obtained result by N, we find the following expression for the amplitude
value of induced e.m.f.:
dim=
nolKoIrn N , b
in a —
• 9.11. Calculation of mutual inductance. Two solenoids of the
same length and of practically the same cross section are completely
inserted one into the other. The inductances of solenoids are L1and L2.
Neglecting edge effects, find their mutual inductance (modulo).
Solution. By definition, the mutual inductance is
L12 (13142, (1)
where Tiis the total magnetic flux through all the turns of solenoid 1
provided that current 12 flows in solenoid 2. The flux D1 =
where N1is the number of turns in solenoid 1, S is the cross-sectional
area of the solenoid, and B2 = pLuon2/2. Hence formula (1) can be

written (after cancelling 72) as follows:
I L12 I = p.p.on2N1S= Rtonin,V, (2)
where we took into account that N1 = n11, where 1 is the solenoid
length and /S = V is its volume. Expression (2) can be represented
in terms of L1and L2 as follows:
I L12 I = 171.41,0nTy Vuuon417=
It should be noted that this expression defines the limiting (maximum)
value of I L12 I. In general, I L12 I < 1/ L1L2.
9.12. Reciprocity theorem. A small cylindrical magnet M is
placed at the centre of a thin coil of radius a, containing N turns
(Fig. 9.26). The coil is connected to a ballistic galvanometer. The
resistance of the circuit is R. After the magnet had been rapidly re-
moved from the coil, a charge q passed through the galvanometer.
Find the magnetic moment of the magnet.
Solution. In the process of removal of the magnet, the total mag-
netic flux through the coil was changing, which resulted in the emerg-
ence of induced current defined by the following equation:
c/(I) dl
RI L
dt dt •
We multiply both sides of this
equation by dt and take into
account that I dt = dq. This
Fig. 9.26 gives
R dq = c/113— L dI.
—
Having integrated this expression, we obtain Rq — L

Considering now that Al = 0 (the current was equal to zero at the
beginning as well as at the end of the process), we obtain
q= AT/R = 0/R, (1)
where cI) is the magnetic flux through the coil at the beginning of the
process (we omitted the minus sign since it is immaterial).
Thus, the problem is reduced to the calculation of the flux D through
the coil. This quantity cannot be determined directly. This difficulty,
however, can be overcome by using the reciprocity theorem. We men-
tally replace the magnet by a small current loop creating in the sur-
rounding space the same magnetic field as the magnet. If the area of
the loop is S and the current is I, their product will give the magnetic
moment pm of the magnet: pm = IS. According to the reciprocity
theorem, L12I = L231 and the problem is reduced to determining the
magnetic flux through the area S of the loop, which creates the same
current I, but flowing in the coil. Assuming the field to be uniform
within the loop, we obtain
(I) = BS = RoN/S/2a. (2)
10.1 Displacement Current 253
It remains for us to substitute (2) into (1) and recall that IS --- pm.
Then q = ttoNp„,I2aR, and
pm = 2aRqltioN.
10. Maxwell's Equations.

Energy of Electromagnetic Field
10.1. Displacement Current
Maxwell's Discovery. The theory of electromagnetic field
founded by Faraday was mathematically completed by
Maxwell. One of the
most important new S' Si
ideas put forward by
Maxwell was the idea
off symmetry in the
mutual dependence
( 1 (
n
of electric and mag-

netic fields. Name-
ly, since a time-var- Fig. 10.1
ying magnetic field
(OBI at) creates an electric field, it should be expected that
a time-varying electric field (aE/at) creates a magnetic field.
This idea necessitating the existence of a new in principle
phenomenon of induction can be approached, for example,
by the following line of reasoning. We know that according
to the theorem on circulation of vector H,
H dl= j dS. (10.1)
Let us apply this theorem to the case when a preliminarily

charged parallel plate capacitor is being discharged through
a certain external resistance (Fig. 10.1a). For the contour I'
we take a curve embracing the wire. We can stretch various
surfaces on this contour, for example, S and S'. These two
surfaces have "equal rights". However, current I flows through
surface S, while through surface S' there is no current!
It turns out that the circulation of vector H depends on
the surface stretched over a given contour (?!), which is
254 10. Maxwell's Equations. Electromagnetic Field Energy
obviously ruled out (and never happened in the case of

direct currents).
Is it possible to change somehow the right-hand side of
(10.1) to avoid this inconvenience? Fortunately, this can
be done in the following way.
First, we note that surface S' "cuts" only the electric
field. In accordance with the Gauss theorem, the flux of
vector D through a closed surface is t D dS = q, whence
(c) aD dS — aq (10.2)
at at '
On the other hand, according to the continuity equation
(5.4), we have
dS= — at ' (10.3)
Summing up the left- and right-hand sides of Eqs. (10.2)

and (10.3) separately, we obtain
(j dS= O. (10.4)
This equation is similar to the continuity equation for

direct current. It can be seen that in addition to the density j
of the conduction current, there is one more term OD/at
whose dimensions are the same as for current density. Max-
well termed this addend the density of displacement current:
jd = ODIdt. (10.5)
The sum of the conduction and displacement currents is
called the total current. Its density is given by
it = +ap (10.6)
at
According to (10.4), the lines of the total current are con-
tinuous in contrast to the lines of conduction current. If
conduction currents are not closed, displacement currents
close them.
We shall show now that the introduction of the concept of
total current eliminates the difficulty associated with the
dependence of circulation of vector H on the choice of the
surface stretched over contour F. It turns out that for this
10.1. Displacement Current 255
it is sufficient to introduce on the right-hand side of Eq.

(10.1) the total current instead of conduction current, viz.
the quantity
I1 = (j (31)
--
at
dS. (10.7)
Indeed, the right-hand side of (10.7) is the sum of conduc-

tion current I and displacement current Id :It = I + Id.
Let us show that the total current It will be the same for
surfaces S and S' stretched over the same contour P. For
this purpose, we apply formula (10.4) to the closed surface
composed of surfaces S and S' (Fig. 10.1b). Taking into ac-
count the fact that for a closed surface the normal n is di-
rected outwards, we write
It (S') +./t (S) = 0.
Now, if we direct the normal n' for surface S' to the same
side as for surface S, the first term in this expression will
change the sign, and we obtain
It (S') = It (S),
Q.E.D.
Thus, the theorem on circulation of vector H, which was
established for direct currents, can be generalized for an ar-
bitrary case in the following form:
I . OD
Hdl = (3+ 7-t--) dS. (10.8)
In this form, the theorem on circulation of vector H is al-

ways valid, which is confirmed by the agreement of this
equation with the results of experiments in all cases without
any exception.
Differential form of Eg. (10.8):
OD
VXH=j+ 0T, (10.9)
where the curl of vector H is determined by the density j of conduction

current and aD/at of displacement current at the same point.
Several Remarks about Displacement Current. It should

be kept in mind that displacement current is equivalent to

conduction current only from the point of view of its ability
of creating a magnetic field.
Displacement currents exist only when an electric field
varies with time. In dielectrics, displacement current con-
sists of two essentially different
components. Since vector D
80E+ P, it follows that the den-
sity OD/Ot of displacement cur-
rent is the sum of the densities
of the "true" displacement current,
co0E/Ot, and of the polarization
current °P/at. The latter quantity
appears due to the motion of
bound charges. There is nothing un-
Fig. 10.2 expected in the fact that polari-
zation currents excite a magnetic
field, since these currents do not differ in nature from con-
duction currents. A new in principle statement consists in
that the other component of the displacement current,
a00E/Ot, which is not connected with any motion of charges
and is only due to the variation of the electric field, also
excites a magnetic field. Even in a vacuum, any temporal
change of an electric field excites in the surrounding space
a magnetic field.
The discovery of this phenomenon was the most essential
and decisive step made by Maxwell while constructing the
electromagnetic field theory. This discovery was as valuable
as the discovery of electromagnetic induction, according to
which a varying magnetic field excites a vortex electric field.
It should also be noted that the discovery of displacement
current by Maxwell was a purely theoretical discovery of
primary importance.
Let us consider an example in which displacement cur-
rents are manifested.
Example. A metallic positively charged sphere is in an infinite
homogeneous conducting medium (Fig. 10.2). Electric currents flowing
in radial directions should excite a magnetic field. Let us find the
direction of vector B at an arbitrary point P.
First of all, it is clear that vector B cannot have a radial compo-
nent. Otherwise, the flux of B through the surface S of the sphere
(Fig. 10.2) would differ from zero, which is in contradiction with

Eq. (7.2). Consequently, vector B must be perpendicular to the radial
direction at the point P. But this is also impossible, since all directions
perpendicular to the radial direction are absolutely equivalent. It
remains for us to conclude that magnetic field is equal to zero every-
where.
The absence of magnetic field in the presence of electric current
of density j indicates that in addition to conduction current j, displace-
ment current jd is also present in the system. This current is such that
the total current is equal to zero everywhere, i.e. jd = —j at each
point. This means that
I q ap
d =/ = 4nr2 = 4nr2 =-
Ot
where we took into account that D = q/4xr2 in accordance with the
Gauss theorem.
10.2. Maxwell's Equations

Maxwell's Equations in the Integral Form. The introduction
of displacement current brilliantly completed the macro-
scopic theory of electromagnetic field. The discovery of dis-
placement current (0D/Ot) allowed Maxwell to create a
unified theory of electric and magnetic phenomena. Maxwell's
theory not only explained all individual phenomena of elec-
tricity and magnetism from a single point of view, but also
predicted a number of new phenomena whose existence was
subsequently confirmed.
Hitherto, we considered separate parts of this theory.
Now we can represent the entire theory in the form of a sys-
tem of fundamental equations in electrodynamics, called
Maxwell's equations for stationary media. In all, there are
four Maxwell's equations (we are already acquainted with
each of these equations separately from previous sections;
here we simply gather them together). In the integral form,
the system of Maxwell's equations is written as follows:
E dl = — a13 dS
— D dS = p dV, (10.10)
at
H dl = j dS, B dS = 0, (10.11)
where p is the volume density of extraneous charges and j

is the density of conduction current.
17-0181
These equations express in a concise form all our know-

ledge about electromagnetic field. The content of these equa-
tions consists in the following.
1. Circulation of vector E around any closed contour is
equal, with the minus sign, to the time derivative of the magnet-
ic flux through an arbitrary surface bounded by the given
contour. Here E is not only vortex electric field but also elec-
trostatic field (whose circulation is known to be equal to
zero).
2. The flux of D through any closed surface is equal to
the algebraic sum of extraneous charges embraced by this
surface.
3. Circulation of vector H around an arbitrary closed con-
tour is equal to the total current (conduction current plus
displacement current) through an arbitrary surface bounded
by this contour.
4. The flux of B through an arbitrary closed surface is
always equal to zero.
It follows from Maxwell's equations for circulations of vec-
tors E and H that electric and magnetic fields cannot be
treated as independent: a change in time in one of these fields
leads to the emergence of the other. Hence it is only a combi-
nation of these fields, describing a unique electromagnetic
field, • that has a sense.
If the fields are stationary (E = const and B = const),
Maxwell's equations split into two groups of independent
equations
E dl = 0, D dS = q,
(10.12)
() dl =I, B clS = O.
In this case, the electric and magnetic fields are indepen-
dent of one another, which allowed us to study first a con-
stant electric field and then, independently, constant mag-
netic field.
It should be emphasized that the line of reasoning which
has led us to Maxwell's equations can by no means be con-
sidered their proof. These equations cannot be"derived" since
they are the fundamental axioms, or postulates of electro-
dynamics, obtained with the help of generalization of exper-
imental results. These postulates play in electrodynamics

the same role as Newton's laws in classical mechanics or the
laws in thermodynamics.
Differential Form of Maxwell's Equations. Equations
(10.10) and (11.11) can be represented in differential form,
i.e. in the form of a system of differential equations, viz.
aB
V X E= at , V •D = p, (10.13)
an
VxH=j+ V O. (10.14)
Equations (10.13) show that there may be two causes for

an electric field. First, its sources are electric charges, both
extraneous and bound (which follows from the equation
V • D = p, if we take into account that D =- goE P and
V •P = — p' ,thenV •EOC p p'). Second, the field E always
emerges when a magnetic field varies in time (which ex-
presses Faraday's law of electromagnetic induction).
On the other hand, Eqs. (10.14) indicate that the magnet-
ic field B can be excited by moving electric charges (elec-
tric currents), by varying electric fields, or by both factors
simultaneously (this follows from the equation V x H =
j ODA, if we take into account that H = B/po—J and
V X J = j'; then V x BOCj±j'± aP/at 800E1 at,
where j' is the magnetization current density and OP/at is
the polarization current density. The first three currents are
associated with motion of charges, while the last current
is due to a time-varying field E). As follows from the equa-
tion V• B = 0, there are no sources of magnetic field in na-
ture (called, by analogy, magnetic charges) that would be
similar to electric charges.
The importance of Maxwell's equations in differential
form consists not only in that they express the basic laws
of electromagnetic field, but also in that the fields E and B
themselves can be found by solving (integrating) these equa-
tions.
Together with the equation of motion of charged particles
under the action of the Lorentz force,
dp/dt = qE q [v x B], (10.15)
17*
Maxwell's equations form a fundamental system of equa-

tions. In principle, this system is sufficient for describing all
electromagnetic phenomena in which quantum effects are
not exhibited.
Boundary Conditions. Maxwell's equations in integral
form are of a more general nature than their differential
counterpart, since they remain valid in the presence of a
surface of discontinuity, viz. the surface on which the proper-
ties of the medium or fields change abruptly. On the other
hand, Maxwell's equations in the differential form presume
that all quantities vary continuously in space and time.
However, the differential form of Maxwell's equations
can be imparted the same generality by supplementing them
with the boundary conditions which must be observed for an
electromagnetic field at the interface between two media.
These conditions are contained in the integral form of
Maxwell's equations and are already familiar to us:
Dln = D2„, El, = E„, B,„ = B2n) HIT = 112T
(10.16)
(here the first and last conditions refer to the cases when there
are neither extraneous charges nor conduction currents on the
interface). It should also be noted that these boundary con-
ditions are valid both for constant and for varying fields.
Material Equations. Maxwell's fundamental equations
still do not form a complete system of equations for the elec-
tromagnetic field. These equations are insufficient for de-
termining the fields from given distributions of charges and
currents.
Maxwell's equations should be supplemented with rela-
tions which would contain the quantities characterizing in-
dividual properties of a medium. These relations are called
material equations. Generally, these equations are rather com-
plicated and are not as general and fundamental as Max-
well's equations.
Material equations have the simplest form for sufficiently
weak electric fields which vary in time and space at a com-
paratively low rate. In this case, for isotropic media
containing no ferroelectrics and ferromagnetics, the material
equations have the following (already familiar to us) form:
D = 880E, B = j = a (E E*), (10.17)
where a, t, and a are the well-known constants characteriz-

ing the electric and magnetic properties of the medium
(permittivity, permeability, and electroconductivity), and
E* is the strength of the extraneous force field due to chemi-
cal or thermal processes
10.3. Properties of Maxwell's Equations

Maxwell's Equations Are Linear. They contain only the first
derivatives of fields E and B with respect to time and spa-
tial coordinates, and the first powers of densities p and j
Fig. 10.3 Fig. 10.4
of electric charges and currents. The linearity of Maxwell's

equations is directly connected with the principle of super-
position: if any two fields satisfy Maxwell's equations,
the sum of these fields will also satisfy these equations.
Maxwell's Equations Include the Continuity Equation. This
expresses the law of conservation of electric charge. In
order to make sure of this, let us take an infinitely small
contour P, stretch over it an arbitrary finite surface S
(Fig. 10.3), and then contract this contour to a point, so
that surface S remains finite. In the limit, 01 dl vanishes,
surface S becomes closed, and the first of Eqs. (10.11) be-
comes
+ dS = O.
j dS = — td (1 D dS = — aq
at
which is just the continuity equation (5.4). This equation
states that the current flowing from a volume V through a
closed surface S is equal to the decrease in charge in this

volume per unit time.
The same law (viz. the continuity equation) can also be
obtained from Maxwell's differential equations. For this
purpose, it is sufficient to take the divergence of both sides
of the first of Eqs. (10.14) and make use of the second of Eqs.
(10.13). This gives V•j = — °p/a.
Maxwel's Equations Are Satisfied in All Inertial Systems of
Reference. They are relativistic invariants. This is a conse-
quence of the relativity principle, according to which all
inertial systems of reference are equivalent from the physi-
cal point of view. The invariance of Maxwell's equations
(with respect to Lorentz' transformations) is confirmed by
numerous experimental data. The form of Maxwell's equa-
tions remains unchanged upon a transition from one inertial
reference system to another, but the quantities contained in
them are transformed in accordance with certain rules. The
transformation of vectors E and B was considered in Sec. 8.
Thus, unlike, for example, Newton's equations in mechan-
ics, Maxwell's equations are correct relativistic equations.
On the Symmetry of Maxwell's Equations. Maxwell's
equations are not symmetric relative to electric and magnet-
ic fields. This is again due to the fact that there are electric
charges in nature, but as far as it is known at present, mag-
netic charges do not exist. However, for a neutral homogen-
eous nonconducting medium, where p = 0 and j = 0,
Maxwell's equations acquire a symmetric form, since E
is related to aB/at in the same way as B to 0E/0t:
V x E = — aBlot, V • D=0,
V XH=- arot, v .13 = 0. (10.18)
The symmetry of equations relative to electric and magnet-
ic fields does not involve only the sign of the derivatives
ail/at and OD/Ot. The difference in signs of these derivatives
indicates that the lines of the vortex electric field induced
by a variation of field B form a left-hand screw system with
vector OBIdt, while the lines of the magnetic field induced
by a variation of D form a right-hand screw system with
vector OD/at (Fig. 10.4).
On Electromagnetic Waves. Maxwell's equations lead to
the following important conclusion about the existence of a
new in principle physical phenomenon: electromagnetic field

may exist independently, without electric charges and cur-
rents. A change in its state in this case necessarily has a
wave nature. Such fields are called electromagnetic waves.
In a vacuum, these waves always propagate at a velocity
equal to the velocity of light c.
It also turned out that displacement current (aD/at)
plays in this phenomenon a primary role. It is its presence,
Fig. 10.5 Fig. 10.6
along with the quantity OB/Ot, that makes the appearance

of electromagnetic waves possible. Any time variation of a
magnetic field induces an electric field, while a variation of
an electric field, in its turn, induces a magnetic field. This
continuous interconversion or interaction of the fields pre-
serves them, and an electromagnetic pertu-bation propa-
gates in space.
Maxwell's theory not only predicted the existence of elec-
tromagnetic waves, but also made it possible to establish
all their basic properties. Any electromagnetic wave, regard-
less of its specific form (it can be a harmonic wave or an
electromagnetic perturbation of any form) is characterized
by the following properties:
(1) the velocity of its propagation in a nonconducting
neutral nonferromagnetic medium is equal to
v = vjlell, where c = 1/1/8o[to; (10.19)
(2) vectors E, B, and v (wave velocity) are mutually per
pendicular and form a right-hand screw system (Fig. 10.5)
This right-handed relation is an intrinsic property of an
electromagnetic wave, independent of the choice of coordi-
nate system;
(3) vectors E and B in an electromagnetic wave always os-

cillate in phase (see Fig. 10.6 showing a "photograph" of
a wave), the instantaneous values E and B at any point be-
ing connected through the relation E = vB, or
V-88o E = jlTN H. (10.20)
This means that E and H (or B) simultaneously attain their
maximum values, vanish, etc.
Maxwell's brilliant success in the development of the elec-
tromagnetic theory of light was due to his understanding of
the possibility of existence of electromagnetic waves, fol-
lowing from differential equations (10.18).
10.4. Energy and Energy Flux. Poynting's Vector

Poynting's Theorem. Proceeding from the idea of energy
localization in the field itself and on the basis of the prin-
ciple of energy conservation, we should conclude that if
energy decreases in a certain region, this may only occur
due to its "flowing out" through the boundaries of the region
under consideration (the medium is assumed to be sta-
tionary).
In this respect, theme is a formal analogy with the law
of conservation of charge expressed by Eq. (5.4). The mean-
ing of this law consists in that a decrease in charge per unit
time in a given volume is equal to the flux of vector j through
the surface enclosing this volume.
In a similar manner, for the law of conservation of energy
we should assume that in addition to the energy density w
in a given region, there exists a certain vector S characteri-
zing the energy flux density.
If we speak only of the energy of an electromagnetic field,
its total energy in a given volume will change due to its flow-
ing out of this volume as well as due to the fact that the
field transfers its energy to a substance (charged particles),
i.e. accomplishes work over the substance. In macroscopic
form, this statement can be written as follows:
dW
SdA+P, (10.21)
dt
where dA is a surface element.

10.4. Energy and Energy Flux, Poynting's Vector 265
This equation expresses Poynting's theorem: a decrease

in energy per unit time in a given volume is equal to the energy
flux through the surface bounding this volume plus the work
per unit time (i.e. power P) which is accomplished by the field
over the charges of the substance inside the given volume.
In Eq. (10.21), W = f w dV, where w is the field
energy density, and P = j • E dV, where j is the current

density and E is the electric field strength. The above ex-
pression for P can be obtained as follows. During the time
dt, field E will do the work SA = qE• u dt over a point charge
q, where u is the charge velocity. Hence the power of force
qE is P = qu•E. Going over to the volume distribution of
charges, we replace q by p dV, where p is the volume charge
density. Then dP = pu•E dV = j • E dV. It remains for us
to integrate dP over the volume under consideration.
It should be noted that power P in (10.21) can be either
positive or negative. The latter takes place when positive
charges of the substance move against field E, or negative
charges move in the opposite direction. For example, such
a situation is observed at the points of a medium where in
addition to an electric field E, the field E* of extraneous
forces is also acting. At such points, j = a (E E*),
and if E* E, and E* > E in magnitude, the product
j • E in the expression for P turns out to be negative.
Poynting obtained the expressions for the energy density
w and vector S by using Maxwell's equations (we shall not
present this derivation here). If a medium contains no ferro-
electrics and ferromagnetics (i.e. in the absence of hysteresis),
the energy density of an electromagnetic field is given by
E•D B•H
tv
2 11- 2 •
(10.22)
It should be noted that individual terms of this expres-

sion have been obtained by us earlier [see (4.10) and (9.32)].
The density of the electromagnetic energy flux, which is
called Poynting vector is defined as
S [E X II]. I (10.23)
Strictly speaking, Maxwell's equations cannot give un-

ambiguous expressions for the two quantities w and S.
The expressions presented above are the simplest from an in-
finite number of possible equations. Therefore, we should
treat these expressions as postulates whose validity should
be confirmed by the agreement with experiments of their
corollaries.
Several following examples will demonstrate that although
the results obtained with the help of formulas (10.22)
and (10.23) may sometimes seem strange, we will not be
able to find in them anything incredible, any disagreement
with experiment. And this is just the evidence that these
expressions are correct.
Example 1. Energy flux in an electromagnetic wave (in vacuum).
Let us calculate energy dW passing during time dt through a unit area
perpendicular to the direction of propagation of the wave.
If we know the values of E and B in the region of location of the
unit area, then
dW = we dt,
where w is the energy density, w = coE2/2 p01/2/2. In accordance
with (10.20), for an electromagnetic wave we have
613E2 = p,oH2.
This means that the electric energy density in the electromagnetic
wave at any instant is equal to the magnetic energy density at the
same point, so we can write for the energy density
w = coE2.
Then
dW = e0 E2c dt= 17
r 801110 E 2 dt.
Let us now see what we shall obtain with the help of Poynting
vector. The same quantity dW can be represented in terms of the mag-
nitude of vector S as follows:
dW =S dt= EH dt= E2 dt.
Thus, both expressions (for w and for S) lead to the same result
(the last two formulas).
Example 2. Evolution of heat in a conductor. Let a current I flow
through a straight circular wire of radius a (Fig. 10.7). Since the wire
has a resistance, a certain electric field E is acting along it. The same
value of E will be at the wire surface in a vacuum. Besides, the pres-
ence of current generates a magnetic field. According to the theorem
on circulation of vector H, near the surface of the wire we have
23-taH = I, H = II2na. Vectors E and H are arranged so that the
10.4. Energy and Energy Flux. Poynting's Vector 267
Poynting vector is directed inside the wire normally to its lateral sur-
face (Fig. 10.7). Consequently, the electromagnetic energy flows into
the wire from the surrounding space! But does it agree with the amount
of heat liberated in the conductor? Let us
calculate the flux of electromagnetic energy
through the lateral surface of a section of
the wire of length 1:
ER •27tal = 22-taH•El = I = RI2,
where U is the potential difference across
the ends of this section and R is its resist-
ance. Thus, we arrive at the conclusion that
the electromagnetic energy flows into the
wire from outside and is completely con-
verted into Joule's heat. We must agree
that this is a rather unexpected conclu-
sion. Fig 10.7
It should be noted that in a power
source, vector E is directed againgt current I
I, and hence in the vicinity of the source 92
the Poynting vector is directed outside:
in this region, electromagnetic energy flows W1
into the surrounding space. In other words, I
it turns out that the energy from the source Fig. 10.8
is transmitted not along the wires but
through the space surrounding a conductor
in the form of the flux of electromagnetic energy, viz. the flux of vec-
tor S.
Example 3. Figure 10.8 shows a section of a balanced (twin) line.
The directions of currents in the wires are known as well as the fact
that the wires' potentials are irp, < w2. Can
we find where (to the right or to the left) the
power source (generator) is?
The answer to this question can be ob-
tained with the help of the Poynting vector.
In the case under consideration, between the
wires, vector E is directed downwards while
vector H is directed behind the plane of the
figure. Hence, vector S =[E X H] is directed
to the right, i.e. the power source is on the Fig. 10.9
left and a consumer is on the right.
Example 4. Charging of a capacitor. Let us take a parallel-plate
capacitor with circular plates of radius a. Ignoring edge effects (field
dissipation), find the electromagnetic energy flux through the lateral
"surface" of the capacitor, since only in this region the Poynting vector
is directed inside the capacitor (Fig. 10.9).
On this surface, there is a varying electric field E and a magnetic
field H generated by its variation. According to the theorem on cir-
culation of vector I3, it follows that 25taH = Ita201:110t, where the
right-hand side contains the displacement current through the contour

shown in Fig. 10.9 by the dashed line. Hence H = (1/2) a aDlat. If
the distance between the plates is h, the flux of vector S through the
lateral surface is
aD OD
EHITah= E Li — 2nah = E — V (1)
2 Ot Ot '
where V = rta2h is the volume of the capacitor. We shall assume that
this flux is completely spent for increasing the capacitor's energy.
Then, multiplying Eq. (1) by dt, we obtain the increment of the capa-
citor's energy during the time dt:
dW=E dD•V =d
es2
o E2
7 =d
)
2DD v) .
1
Having integrated this equation, we find the formula for the energy W
of a charged capacitor. Thus, everything is all right in this case too.
10.5. Electromagnetic Field Momentum

Pressure of Electromagnetic Wave. Maxwell showed theo-
retically that electromagnetic waves, being reflected or ab-
sorbed by bodies on which they are incident, exert pressure
on these bodies. This pressure appears as a result of the action
of the magnetic field of a wave on the electric currents
induced by the electric field of the same wave.
Suppose an electromagnetic wave propagates in a homo-
geneous medium capable of absorption. The presence of ab-
sorption means that Joule's heat with the volume density
6E2 will be liberated in the medium. Hence a 0, i.e.
the absorbing medium is conductive.
The electric field of the wave excites in this medium an elec-
tric current with a density j = aE. As a result, Ampere's
force Fu = Ij x B] = a [E x B] will act on a unit volume
of the medium. This force is directed towards the propaga-
tion of the wave (Fig. 10.10) and causes the pressure of the
electromagnetic wave.
In the absence of absorption, the conductivity a = 0
and Fu= 0, i.e. in this case the electromagnetic wave does
not exert any pressure on the medium.
Electromagnetic Field Momentum. Since an electromag-
netic wave exerts a pressure on a substance, the latter ac-
quires a certain momentum. However, if in a closed system
consisting of a substance and an electromagnetic wave only
the substance possessed a momentum, the law of conserva-
tion of momentum would be violated,
10.5. Electromagnetic Field Momentum 269
The momentum of such a system can be conserved only

provided that the electromagnetic field (wave) also possesses
a momentum: the substance acquires the momentum due
to that transferred to it by the electromagnetic field.
Po
Fig. 10.11
Let us introduce the concept of momentum density G of

electromagnetic field as the quantity numerically equal to
the momentum of the field in a unit volume. Calculations
which will be omitted here show that the momentum density
is given by
G = S/c2, (10.24)
where S = FE x 111 is Poynting's vector. Just as vector

S, momentum density G is generally a function of time and
coordinates.
In accordance with (10.20), for an electromagnetic wave
in a vacuum we have -V80E IlTioH. Hence the energy den-
sity w and the magnitude S of Poynting's vector are re-
spectively equal to
w = goE2/2 RoH2/2 = e0E2, S = EH = co/No E 2.
It follows that S = 80µo• And since Yeo µo = 1/c,
where c is the velocity of light in vacuum, S = wc. Taking
into account (10.24), we obtain that for an electromagnetic
wave in a vacuum
G—wIc. I (10.25)
The same relation between energy and momentum is inhe-

rent (as is shown in the theory of relativity) in particles
with zero rest mass. This is quite natural since, according to
quantum-mechanical notions, an electromagnetic wave is

equivalent to the flow of photons, viz. the particles with zero
rest mass.
Some More Remarks on the Pressure of Electromagnetic
Waves. Let us calculate, using formula (10.25), the pres-
sure exerted by an electromagnetic wave on a body when the
wave is incident normal to its surface and is partially reflect-
ed in the opposite direction. In accordance with the law of
conservation of momentum, Po = p' o p,where Po, p;„ are
the momenta of the incident and reflected waves, while p
is the momentum transferred to the body (Fig. 10.11). Hav-
ing projected this equality onto the direction of the inci-
dent wave and referring all the quantities to unit time and
unit cross-sectional area, we obtain
P = Po = (G) c (G') c,
where (G) and (G' ) are the average values of the momentum
density for the incident and reflected waves. It remains for
us to take into account relation (10.25) between (G) and
(w ) as well as the fact that (w' ) = p (w ), where p is the
reflection coefficient. As a result, the previous expression be-
comes
P = (1+ P) (ID) (10.26)

Here the quantity p is just the pressure of the electromagnet-
ic wave on the body. In the case of total reflection, p =
= 1 and pressure p = 2 (w ) while for total absorption,
p = 0 and p = (w ).
It should be added that the pressure of electromagnetic
radiation is usually very low (the exception is the pressure
of high-power laser radiation beams, especially after beam
focusing, and the pressure of radiation inside hot stars).
For example, the pressure of solar radiation on the Earth
amounts to about 10-6 Pa, which is smaller than the atmo-
spheric pressure by a factor of 1010. In spite of insignificant
values of these quantities, experimental proof of the
existence of electromagnetic waves, viz. the pressure of
light, was obtained by Lebedev. The results of these experi-
ments were in agreement with the electromagnetic theory of
light.
Problems 271
Problems
• 10.1. Displacement current. A point charge q moves uniformly
along a straight line with a nonrelativistic velocity v. Find the vector
of the displacement current density at a point P lying at a distance r
from the charge on a straight line (1) coinciding with its trajectory;
(2) perpendicular to its trajectory and passing through the charge.
-ud t
Fig. 10.12 Fig. 10.13
Solution. The displacement current density jd = aD/Ot. Hence,

the solution of the problem is reduced to determining vector D at
the indicated points and to finding its time derivative. In both cases,
D = qe„/43-ir2, where e, is the unit vector of r. Let us find the derivative
OD/at
(1) At point P1(Fig. 10.12, where it is assumed that q > 0), we
have
OD 2q dr 2qv
at 43-tr3 at er - 41tr3
where we took into account that for point P1the derivative Or/at = —v.
If point P1were not in front of the charge (in the direction of its motion)
but behind it, vector jd would be directed in the same way and would
have the same magnitude.
Thus, for q > 0, vector jd tt v, and vice versa.
(2) At point P2 (Fig. 10.12), I dD = v der, and hence we can
write
apiat = —qvA1tr3.
If q > 0, jd T1v, and vice versa.
• 10.2. A current flowing in a long straight solenoid with the
radius R of cross section is varied so that the magnetic field inside the
solenoid increases with time according to the law B = Pt2, where p
is a constant. Find the displacement current density as a function of
the distance r from the solenoid axis.
Solution. In order to find the displacement current density, we

must, in accordance with (10.5), first find the electric field strength
(here it will be a vortex field). Using Maxwell's equation for circula-
tion of vector E, we write
2o-crE = xr2 01310t, E = rPt (r < R);
2xrE xR2 aBlat, E = R9307- (r > R).
Now, using the formula jd = Co aElat, we can find the displacement
current density:
jd = 800r (r < R); jd = gof3R 2/r (r > R).
The plot of the dependence id (r) is shown in Fig. 10.13.
• 10.3. A parallel-plate capacitor is formed by two discs the space
between which is filled with a homogeneous, poorly conducting medium.
The capacitor was charged and then disconnected from a power source.
Ignoring edge effects, show that the magnetic field inside the capacitor
is absent.
Solution. Magnetic field will be absent since the total current (con-
duction current plus displacement current) is equal to zero. Let us
prove this. We consider the current density. Suppose that at a certain
instant the density of conduction current is j. Obviously, j CC D and
D = an, where a is the surface charge density on the positively charged
plate and n is the normal (Fig. 10.14).
The presence of conduction current leads to a decrease in the surface
charge density a, and hence in D as well. This means that conduction
current will be accompanied by the displacement current whose den-
sity is
jd = aplat = (a6/at) n = --fn = —j.
Hence it follows that, indeed
it = jd = 0.
• 10.4. The space between the plates of a parallel-plate capacitor
in the form of circular discs is filled with a homogeneous poorly con-
ducting medium with a conductivity a and permittivity a. Ignoring
edge effects, find the magnitude of vector H between the plates at
a distance r from their axes, if the electric field strength between the
plates varies with time in accordance with the law E =
Em cos cot.
Solution. From Maxwell's equation for circulation of vector H,
it follows that
E n ) ar2
2nrH= (j n + 880 O
at
Taking into account Ohm's law jn = aEn (t), we obtain
r EEO aEn rEm ,
H an, cos cot ( —se0co sin cot).
2 G at 2
Problems 273
Let us transform the expression in the parentheses to cosine. For this

purpose, we multiply and divide this expression by f 62 + (88,02
and then introduce angle 6 through the formulas al f = cos 6, eso coif =
= sin 6. This gives
1
H rEm 0 2+(mot )r cos (wn+ (5)•
• 10.5. A point charge q moves in a vacuum uniformly and

rectilinearly with a nonrelativistic velocity v. Using Maxwell's
Fig. 10.14 Fig. 10.15
equation for the circulation of vector H, obtain the expression for H

at a point P whose position relative to the charge is characterized by
radius vector r (Fig. 10.15).
Solution. It is clear from symmetry considerations that for a con-
tour around which the circulation of vector H should be taken, we
must choose a circle with centre 0 (its trace is shown in Fig. 10.16
by the dashed line). Then
2aRH= at 13, dS,

—
(1)
where R is the radius of the circle.
Let us find the flux of vector D through a surface bounded by this
circle. For the sake of simplicity, we shall take a spherical surface
with the radius of curvature r (Fig. 10.16). Then the flux of D through
an elementary ring of this spherical surface is
D dS = 2ar sin cc' • r dce = I sin a' da' ,
while the total flux through the selected surface is
D dS =÷ (1 — cos a). (2)
Now, in accordance with (1), we differentiate (2) with respect to

time:
a
07 D dS=-- I sin da
2 a dt • (3)
18-0181
For the displacement of the charge from point 1 to point 2 (Fig. 10.17)
over the distance v dt, we have v dt•sin a = r da, whence
da v sin a
dt r (4)
Substituting (4) into (3) and then (3) into (1), we obtain
H =-- qvr sin agnr3, (5)
where we took into account that R = r sin a. Relation (5) in vector
form can be written as follows:
H= q [v X r]
4J-c r3 •
Thus we see that expression (6.3) which we have postulated earlier
is a corollary of Maxwell's equations.
• 10.6. Curl of E. A certain region of an inertial system of reference
contains a magnetic field of magnitude B = const, rotating at an
Fig. 10.16
angular velocity co. Find V X E in this region as a function of vectors
w and B.
Solution. It follows from the equation V X E -OBI& that
vector V x E is directed oppositely to vector dB. The magnitude of
this vector can be calculated with the help of Fig. 10.18:
dB I = Bco dt, I dB/dt = Bw.
Hence
V X E = --[J) X B].
• 10.7. Poynting's vector. Protons having the same velocity v
form a beam of a circular cross section with current I. Find the direc-
tion and magnitude of Poynting's vector S outside the beam at a
distance r from its axis.
Problems 275
Solution. Figure 10.19 shows that S f v. Let us find the mag-

nitude of S: S = EH, where E and H depend on r. According to the
Gauss theorem, we have
2nrE = Also,
where is the charge per unit length of the beam. Besides, it follows
from the theorem on circulation of vector H that
22trH
Having determined E and H from the last two equations and taking
into account that I = Xv, we obtain
S = EH = 12/431280yr2.
• 10.8. A current flowing through the winding of a long straight
dB
Fig. 10.18
solenoid is being increased. Show that the rate of increase in the energy
of the magnetic field in the solenoid is equal to the flux of Poynting's
vector through its lateral surface.
Solution. As the current increases, the magnetic field in the solenoid
also increases, and hence a vortex electric field appears. Suppose that
the radius of the solenoid cross section is equal to a. Then the strength
of the vortex electric field near the lateral surface of the solenoid
can be determined with the help of Maxwell's equation that expresses
the law of electromagnetic induction:
9 OB a aB
2naE =Ica- — E= —- ---
at 2 a
The energy flux through the lateral surface of the solenoid can be re-
presented as follows:
8 ( B2
(120= EH •22-cal=2-ca21-
0t 2110
where 1 is the solenoid length and 2-ca21 is its volume.
Thus, we see that the energy flux through the lateral surface of the
solenoid (the flux of vector S) is equal to the rate of variation of the
magnetic energy inside the solenoid:
cl) = S•2Ical = OW/0t.
i8*
• 10.9. The energy from a source of constant voltage U is trans-

mitted to a consumer via a long coaxial cable with a negligibly small
resistance. The current in the cable is I. Find the energy flux through
the cable cross section, assuming that the outer conducting shell of
the cable has thin walls.
E Solution. The required ener-
gy flux is defined by the formula
b
r
a: =
- S•2zir dr, (1)
a
Fig. 10.20 where S=EH is the flux density,
2xr dr is the elementary ring of
width dr within which the value of S is constant, and a and b are the
radii of the internal wire and of the outer shell of the cable (Fig. 10.20).
In order to evaluate this integral, we must know the dependence
S (r), or E (r) and H (r).
Using the Gauss theorem, we obtain
2arE = ?Jco, (2)
where is the charge per unit length of the wire. Further, by the theo-
rem on the circulation we have
2xrH = I. (3)
After substituting E and H from formulas (2) and (3) into expression
(1) and integrating, we get
A'I b (4)
(1) = 2 n *
The values of )■
,, a, and b are not given in the problem. Instead,
we know U. Let us find the relation between these quantities:
b
U= E dr= In — (5)
2aso a
a
A comparison of (4) and (5) gives
(I) = U/.
This coincides with the value of power liberated in the load.
• 10.10. A parallel-plate air capacitor whose plates are made
in the form of disks of radius a are connected to a source of varying
harmonic voltage of frequency (a. Find the ratio of the maximum values
of magnetic and electric energy inside the capacitor.
Solution. Let the voltage across the capacitor vary in accordance
with the law U = Umcos cot and the distance between the capacitor
plates be h. Then the electric energy of the capacitor is equal to

80E2 2h sona2 (1)
We 2 na = 2h
U2 COO Wt.
The magnetic energy can be determined through the formula

B2
Wm = v- 10 dV. (2)
The quantity B required for evaluating this integral can be found

from the theorem on the circulation of vector H: 2scrH = scr201,10t.
Hence, considering that H = Bitto and oD/Ot = —so (Um/h) co sin cot,
we obtain
1 r coUrn
B2 7sop, h I sin cot I. (3)
It remains for us to substitute (3) into (2), where for dV we must
take an elementary volume in the form of a ring for which dV =
= 2nr dr•h. As a result of integration, we obtain
1,08vosou
wm — sin2cot. (4)
16
The ratio of the maximum values of magnetic energy (4) and elec-
tric energy (1) is given by
Wm max I
i1080a26)2.
We max 8
For example, for a = 6 cm and ai = 1000 s-1, this ratio is equal to
5 X 10-15.
11. Electric Oscillations

11.1. Equation of an Oscillatory Circuit
Quasi-steady Conditions. When electric oscillations occur,
the current in a circuit varies with time and, generally speak-
ing, turns out to be different at each instant of time in diff-
erent sections of the circuit (due to the fact that electromag-
netic perturbations propagate although at a very high but
still finite velocity). There are, however, many cases when
instantaneous values of current prove to be practically the
same in all sections of the circuit (such a current is called
quasistationary). In this case, all time variations should occur
so slowly that the propagation of electromagnetic perturba-
278 11. Electric Oscillations
tions could be considered instantaneous. If 1 is the length of

a circuit, the time required for an electromagnetic pertur-
bation to cover the distance 1 is of the order of T = //C.
For a periodically varying current, quasi-steady condition
will be observed if
T = //C « T,
where T is the period of variations.
For example, for a circuit of length 1 =- 3 m, the time
T = 10-8 s, and the current can be assumed to be quasista-
(a) (b)
Fig. 11.1 Fig. 11.2
tionary down to frequencies of 106Hz (which corresponds to

T = 10-6 s).
In this chapter, we shall assume everywhere that in the
cases under consideration quasi-steady conditions are ob-
served and currents are quasistationary. This will allow us to
use formulas obtained for static fields. In particular, we shall
be using the fact that instantaneous values of quasista-
tionary currents obey Ohm's law.
Oscillatory Circuit. In a circuit including a coil of indu-
ctance L and a capacitor of capacitance C, electric oscilla-
tions may appear. For this reason, such a circuit is called an
oscillatory circuit. Let us find out how electric oscillations
emerge and are sustained in an oscillatory circuit.
Suppose that initially, the upper plate of the capacitor
is charged positively and the lower plate, negatively (Fig.
11.1a). In this case, the entire energy of the oscillatory cir-
cuit is concentrated in the capacitor. Let us close key K.
The capacitor starts to discharge, and a current flows through
coil L. The electric energy of the capacitor is converted in-
to the magnetic energy of the coil. This process terminates
when the capacitor is discharged completely, while current
in the circuit attains its maximum value (Fig. 11.1b).

Starting from this moment, the current begins to decrease,
retaining its direction. It will not, however, cease immediate-
ly since it will be sustained by self-induced e.m.f. The cur-
rent recharges the capacitor, and the appearing electric field
will tend to reduce the current. Finally, the current ceases,
while the charge on the capacitor attains its maximum value.
From this moment, the capacitor starts to discharge again,
the current flows in the opposite direction, and the process
is repeated.
If the conductors constituting the oscillatory circuit have
no resistance, strictly periodic oscillations will be observed
in the circuit. In the course of the process, the charge on the
capacitor plates, the voltage across the capacitor and the
current in the induction coil vary periodically. The oscil-
lations are accompanied by mutual conversion of the energy
of electric and magnetic fields.
If, however, the resistance of conductors R / 0, then, in
addition to the process described above, electromagnetic
energy will be transformed into Joule's heat.
Equation of an Oscillatory Circuit. Let us derive the equa-
tion describing oscillations in a circuit containing series-
connected capacitor C, induction coil L, resistor R, and
varying external e.m.f. 6 (Fig. 1.1.2).
First, we choose the positive direction of circumvention,
e.g. clockwise. We denote by q the charge on the capacitor
plate the direction from which to the other plate coincides
with the chosen direction of circumvention. Then current in
the circuit is defined as
I = dq/dt. (11.1)
Consequently, if I > 0, then dq > 0 as well, and vice versa
(the sign of I coincides with that of dq).
In accordance with Ohm's law for section 1RL2 of the
circuit, we have
RI = (Pi — (P2 + es + ', (11.2)
where es is the self-induced e.m.f. In the case under con-
sideration,
6, = — L dI/dt and cp,— (pi = q/C
(the sign of q must coincide with the sign of the potential

difference y, — C jsince C > 0). Hence Eq. (11.2) can be
written in the form
L dI -- 1"r -r = (11.3)
or, taking into account (11.1),

d 2q , „ dq 1
L —
dt2
—dt
—C
(11.4)
This is the equation of an oscillatory circuit, which is a li-

near second-order nonhomogeneous differential equation with
constant coefficients. Using this equation for calculating
q (t), we can easily obtain the voltage across the capacitor
as Ue = y, — cpl = qIC and current I by formula (11.1).
The equation of an oscillatory circuit can be given a differ-
ent form:
q+2Pq (14q tIL, (11.5)
where the following notation is introduced:

2P = RIL, (14=11LC. (11.6)
The quantity to is called the natural frequency of the circuit
and p is the damping factor. The meaning of these quantities
will be explained below.
If = 0, the oscillations are usually called free oscilla-
tions. They will be undamped for R = 0 and damped for
R 0. Let us consider consecutively all these cases.
11.2. Free Electric Oscillations

Free Undamped Oscillations. If a circuit contains no exter-
nal e.m.f. g and if its resistance R = 0, the oscillations
in such a circuit will be free and undamped. The equation
describing these oscillations is a particular case of Eq. (11.5)
when 6 = 0 and R = 0:
(11.7)
The solution of this equation is the function

q = qn., cos (coot + a), (11.8)
where qn, is the maximum value of the charge on capacitor
plates, coo is the natural frequency of the oscillatory cir-
cuit, and a is the initial phase. The value of coo is determined
only by the properties of the circuit itself, while the values
of qmand a depend on the initial conditions. For these con-
ditions we can take, for example, the value of charge q
and current I = q at the moment t = 0.
According to (11.6), (0 0 =ITV-LC; hence the period of
free undamped oscillations is given by
To = 2n V" 'LC (11.9)
(Thomson's formula).
Having found current I [by differentiating (11.8) with
respect to time] and bearing in mind that the voltage across
the capacitor plates is in phase with charge q, we can easily
see that in free undamped oscillations current I leads in
phase the voltage across the capacitor plates by 31/2.
While solving certain problems, energy approach can al-
so be used.
Example. In an oscillatory circuit, free undamped oscillations
occur with energy W. The capacitor plates are slowly moved apart
so that the frequency of oscillations decreases by a factor of II. Which
work against electric forces is done in this case?
The required work can be represented as an increment of the energy
of the circuit:
A=w, qn \ w C 4
2 k C' C / = C' /•
On the other hand, coo O( 1/1re, and hence 11 = coo/coo

and consequently
A = W N 2— 1).
Free Damped Oscillations. Every real oscillatory circuit
has a resistance, and the energy stored in the circuit is grad-
ually spent on heating. Free oscillations will be damped.
We can obtain the equation for a given oscillatory circuit
by putting g = 0 in (11.5). This gives
q +213q.c4q= O. (11.10)
It can be shown (we shall not do it here since we are inter-

ested in another aspect of the problem) that for 132 < co(2),
the solution of this homogeneous differential equation has
the form
q= qme-Pt cos (cot + cc),
(11.11)
where
= yo)20 [32
Fig. 11.3
17A -- (2)2, (11.12)
and qm and a are arbitrary constants determined from the

initial conditions. The plot of the function (11.11) is shown
in Fig. 11.3. It can be seen that this is not a periodic
function since it determines damped oscillations.
Nevertheless, the quantity T =- 2nho is called the pe-
riod of damped oscillations:
23-c T,
T— (11.13)
1/€04-132 1(1— ( 3 /00)2 '
where To is the period of free undamped oscillations.
The factor qme -13/ in (11.11) is called the amplitude of
damped oscillations. Its dependence on time is shown in Fig.
11.3 by the dashed line.
Voltage across a Capacitor and Current in an Oscillatory
Circuit. Knowing q (t), we can find the voltage across a ca-
pacitor and the current in a circuit. The voltage across a
capacitor is given by
Ue = = e-Ot cos (cot + a). (11.14)
The current in a circuit is

= dq qme — Pt [— (3 cos (cot + a) — co sin (cot ± a)].
We transform the expression in the brackets to cosine. For

this purpose, we multiply and divide this expression by
+ (32 = coo and then introduce angle 6 by the formu-
las
— 6/co, = cos 6, co/coo = sin 6. (11.15)
After this, the expression for I takes the following form:
coqme-3t cos (cot + a + 6). (11.16)
It follows from (11.15) that angle 6 lies in the second quad-
rant (n/2 < 6 < it). This means that in the case of a non-
zero resistance, the current in the circuit leads voltage in phase
(11.14) across the capacitor by more than n/2. It should
be noted that for R = 0 this advance is 6 = 102.
The plots of the dependences Uc (t) and I (t) have the form
similar to that shown in Fig. 11.3 for q (t).
Example. An oscillatory circuit contains a capacitor of capacitance
C and a coil with resistance R and inductance L. Find the ratio be-
tween the energies of the magnetic and electric fields in the circuit at
the moment when the current is at a maximum.
According to equation (11.3) for an oscillatory circuit,
dI q
L
dt
When the current is at a maximum, dI/dt = 0, and RI =
Hence the required ratio is
Wm/We = L/CR2.
Quantities Characterizing Damping.
1. Damping factor 13 and relaxation time t, viz. the time
during which the amplitude of oscillations decreases by a
factor of e. It can be easily seen from formula (11.11) that
= 1/(3. (11.17)
2. Logarithmic decrement X of damping. It is defined as the
Naperian logarithm of two successive values of amplitudes
measured in a period T of oscillations:
a (t)
X — ln a(t+T) =13T ' (11.18)
where a is the amplitude of the corresponding quantity

(q, U, or I). In a different form, we can write
= 1/Ne, (11.19)
where Ne is the number of oscillations during the time r,

i.e. the time required for the amplitude of oscillations to
decrease to 1/e of its initial value. This expression can be
easily obtained from formulas (11.17) and (11.18).
If damping is small (132<< coD, then w coo = 1/V LC
and, in accordance with (11.18), we have
(3 • 2a/coo = (11.20)
3. Q-factor of an oscillatory circuit. By definition,
Q = nNe, (11.21)
where 2 is the logarithmic decrement. The smaller the damp-
ing, the higher the Q-factor. For a weak damping (132 <
< 64), in accordance with (11.20), we have
1.7 c
L
(11.22)
Here is one more useful formula for the Q-factor in the case of
a weak damping:
Q 2:t
SW
(11.23)
where W is the energy stored in the circuit and SW is the de-

crease in this energy during the period 1' of oscillations. In-
deed, the energy W is proportional to the square of the am-
plitude value of the capacitor charge, i.e. W OC e-20t.
Hence the relative decrease in the energy over the period is
SW/W = 213T = 27k. It remains for us to take into account
the fact that according to (11.21), k nIQ.
Concluding the section, it should be noted that when
132 > (14, an aperiodic discharge of the capacitor will occur
instead of oscillations. The resistance of the circuit for which
the aperiodic process sets in is called the critical resistance:
R, =--- 2 ji LIC. (11.24)
Let us consider two examples.
Example 1. An oscillatory circuit has a capacitance C, inductance L,
and resistance R. Find the number of oscillations after which the
amplitude of current decreases to 1/e of its initial value.
The current amplitude (in, e —13t) decreases to 1/e of its initial
value during the time T = 1/(3. During this time, Ne oscillations will
occur. If T is the period of damped oscillations, then
N e -= =
1113 1 V (00
T 23-0/ o_13 2- La /
Considering that coo = 1/LC and p= R I2L, we obtain
1 4L
Ne
2a V CR2
Example 2. Find the time during which the amplitude of current
oscillations in a circuit with a given Q-factor will decrease to 1/11
of its initial value, if the frequency of damped oscillations is equal
to co.
Since the current amplitude 4, (X e – Ott the time to during which
the amplitude decreases by a factor of ri is determined by the equation
rl = 034. Hence
to = (ln Ti)/13.
On the other hand, the Q-factor is also related to [3:
Q= = at/(3T = co/2[3.
Eliminating 13 from the last two equations, we obtain
2Q ,
to = — ni 1/.
11.3. Forced Electric Oscillations

Steady-state Oscillations. Let us return to equations (11.3)
and (11.4) for an oscillatory circuit and consider the case
when the circuit includes a varying external e.m.f. 6 whose
dependence on time is decreased by the harmonic law:
emcos wt. (11.25)
This law occupies a special place owing to the properties of
the oscillatory circuit itself to retain a harmonic form of os-
cillations under the action of external harmonic e.m.f.
In this case, the equation for an oscillatory circuit is
written in the form
dI
L— dt
+ RI + = 6mcos cot, (11.26)
or
q 2§q wag = (6 nil L) cos cot. (11.27)
286 //. Electric Oscillations
As is known from mathematics, the solution of this equa-

tion is the sum of the general solution of the homogeneous
equation (with the zero right-hand side) and a particular
solution of the nonhomogeneous equation.
We shall be interested only in steady-state oscillations,
i.e. in the particular solution of this equation (the general
solution of the homogeneous equation exponentially attenu-
ates and after the elapse of a certain time it virtually van-
ishes). It can be easily seen that this solution has the form
q = qn, cos (cot —), (11.28)
where qn, is the amplitude of charge on the capacitor and
is the phase shift between oscillations of the charge and of
the external e.m.f. 6 (11.25). It will be shown that qm
and 'tp depend only on the properties of the circuit itself and
the driving e.m.f. 6. It turns out that > 0, and hence q
always lags behind 6 in phase.
In order to determine the constants qn, and V, we must sub-
stitute (11.28) into the initial equation (11.27) and trans-
form the result. However, for the sake of simplicity, we shall
proceed in a different way: first find current I and then sub-
stitute its expression into (11.26). By the way, we shall sol-
ve the problem of determining the constants qm and V.
Differentiation of (11.28) with respect to time t gives
I = — coqmsin (cot —1p) = cagmcos (cot — p + a/2).
Let us write this expression as follows:
I = I n, cos (cot — cp), (11.29)
where Im is the current amplitude and cp is the phase shift
between the current and external e.m.f. 6,
= coqm , cP = — 5t/ 2. (11.30)
We aim at finding I n, and cp. For this purpose, we proceed as
follows. Let us represent the initial equation (11.26) in the
form
UL + UR + (IC = en, cos cot, (11.31)
where the left-hand side is the sum of voltages across induc-
tion coil L, capacitor C and resistor R. Thus, we see that
at each instant of time, the sum of these voltages is equal
to the external e.m.f. 6. Taking into account relation (11.30),

we write
U R =RI = RI „ cos (cot — (I)), (11.32)
Ue C = Cn—
/ cos (cot —11)) =-- dCcos( cot — , (11.33)
U L = L -di—t= — (61,1„sin (cot — cp)
coLI,,cos (cot — (I) + 2) . (11.34)

Vector Diagram. The last three formulas show that UR
is in phase with current I, Uclags behind I by n/2, and UL
(dLlm
Im
coC
Fig. 11.4 Fig. 11.5
leads I by rc/2. This can be visually represented with the

help of a vector diagram, depicting the amplitudes of vol-
tages
URm =111„, Ucm = I „I coC , U L„ = coL/77,
and their vector sum which, according to (11.31), is equal
to vector ern (Fig. 11.4).
Considering the right triangle of this diagram, we can
easily obtain the following expressions for I, and cp:
/in— (11.35)
17R2 -1- (c0L-11coC)2
-
coL —Ri/coc .
tan cp = (11.36)
Thus, the problem is solved.

It should be noted that the vector diagram obtained above
proves to be very convenient for solving many specific

problems. It permits a visual, easy, and rapid analysis of
various situations.
Resonance Curves. This is the name given to the plots of
dependences of the following quantities on the frequency co
of external e.m.f. 6: current I, charge q on a capacitor, and
wo
Fig. 11.6 Fig. 11.7
voltages UR, Um and UL defined by formulas (11.32)-
(11.34).
Resonance curves for current I Tr, ((o) are shown in Fig.
11.5. Expression (11.35) shows that the current amplitude
has the maximum value for coL — 1/(oC = 0. Consequently,
the resonance frequency for current coincides with the na-
tural frequency of the oscillatory circuit:
or res = wo =1/1LC. (11.37)
The maximum at resonance is the higher and the sharper
the smaller the damping factor [3 = R/2L.
Resonance curves for the charge qm, (w) on a capacitor are
shown in Fig. 11.6 (resonance curves for voltage Ucm
across the capacitor have the same shape). The maximum of
charge amplitude is attained at the resonance frequency
(09 res =- 1764 — 2132, (11.38)
which comes closer and closer to w() with decreasing
In order to obtain (11.38), we must represent qm, in accord-
ance with (11.30), in the form qm. =- / m/co where I m is
defined by (11.35). Then
qm =-. (11.39)
✓o,a _or+4132,0
The maximum of this function or, which is the same, the

minimum of the radicand can be found by equating to zero
the derivative of the radicand with respect to co. This gives
the resonance frequency (11.38).
Let us now see how the amplitudes of voltages UR, Uc,
and UL are redistributed depending on the frequency co
of the external e.m.f. This pattern is depicted in Fig. 11.7.
The resonance frequencies for UR, Uc and UL are deter-
mined by the following formulas:
(0R res = (00,
(0c res = (00V-1— 2((3/(00)2, (11.40)
(1).r. res = (00 V1— 2([3/(00)2.

The smaller the value of (3, the closer are the resonance fre-
quencies for all quantities to the value coo.
Resonance Curves and Q-factor. The shape of resonance
curves is connected in a certain way with the Q-factor of an
oscillatory circuit. This connection has the simplest form
for the case of weak damping, i.e. for (32 «wo. In this case,
Ucres/em = Q (11.41)
(Fig. 11.7). Indeed, for (32 << co:, the quantity ores 00
and, according to (11.33) and (11.35), Ucm res = im/WoC =
= emhooCR , or Ucm res, m = VLC/CR = (1/R) liLiC
which is, in accordance with formula (11.22), just the Q-
factor.
Thus, the Q-factor of a circuit (for (32<< (1)02 ) shows how
many times the maximum value of the voltage amplitude
across the capacitor (and induction coil) exceeds the ampli-
tude of the external e.m.f.
The Q-factor of a circuit is also connected with another
important characteristic of the resonance curve, viz. its
width. It turns out that for 132 <<
Q = (00180), (11.42)
where coois the resonance frequency and 6o) is the width of
the resonance curve at a "height" equal to 0.7 of the peak
height, i.e. at resonance.
Resonance. Resonance in the case under consideration is
1/2 19-0181
the excitation of strong oscillations at the frequency of ex-

ternal e.m.f. or voltage. This frequency is equal to the na-
tural frequency of an oscillatory circuit. Resonance is used
for singling out a required component from a composite volt-
age. The entire radio reception technique is based on reso-
nance. In order to receive with a given radio receiver the
station we are interested in, the receiver must be tuned. In
other words, by varying C and L of the oscillatory circuit,
we must attain the coincidence between its natural frequency
and the frequency of radio waves emitted by the radio
station.
The phenomenon of resonance is also associated with a cer-
tain danger: the external e.m.f. or voltage may be small, but
the voltages across individual elements of the circuit (the
capacitor or induction coil) may attain the values dangerous
for people. This should always be remembered.
11.4. Alternating Current

Total Resistance (Impedance). Steady-state forced electric
oscillations can be treated as an alternating current flow-
ing in a circuit having a capacitance, inductance, and resist-
ance. Under the action of external voltage (which plays
the role of external e.m.f.)
U = U„ cos cot, (11.43)
the current in the circuit varies according to the law
I = I, cos (cot — cp), (11.44)
where
Um olL-1/0.)C . (11.45)
tan (1) — R
im R2+ (coL —1/wC)2
The problem is reduced to determining the current ampli-
tude and the phase shift of the current relative to U.
The obtained expression for the current amplitude i n, (o))
can be formally interpreted as Ohm's law for the amplitude
values of current and voltage. The quantity in the denomi-
nator of this expression, which has the dimension of resist-
ance, is denoted by Z and is called the total resistance, or
11.4. Alternating Current 291
impedance:
Z =1/ R2+ (coL — 1RoC)2. (11.46)
It can be seen that for co = coo = 1/V-LC, the impedance
has the minimum value and is equal to the resistance R.
The quantity appearing in the parentheses in formula
(11.46) is denoted by X and is called the reactance:
X = coL — 1/6)C. (11.47)
Here the quantity coL is called the inductive reactance,
while the quantity 1/coC is called the capacitive reactance.
They are denoted X L and Xcrespectively. Thus,
XL = Xc= 1/coC, X = X L— Xc,
(11.48)
z if R2+ x2.
It should be noted that the inductive reactance grows with
the frequency co, while the capacitive reactance decreases
with increasing co. When it is said that a circuit has no ca-
pacitance, this must be understood so that there is no capac-
itive reactance which is equal to 1/coC and hence vanishes
if C 00, (when a capacitor is replaced by a short-cir-
cuited section).
Finally, although the reactance is measured in the same
units as the resistance, they differ in principle. This difference
consists in that only the resistance determines irreversible
processes in a circuit such as, for example, the conversion
of electromagnetic energy into Joule's heat.
Power Liberated in an A.C. Circuit. The instantaneous
value of power is equal to the product of instantaneous
values of voltage and current:
P (t) = UI = UmI racos cot cos (cot — cp). (11.49)
Using the formula cos (cot — cp) = cos cot cos +
sin cot sin cp, we transform (11.49) as follows:
P (t)U mlm (cos2cot cos cp + sin wt cos cot sin cp).
Of practical importance is the value of power averaged
over a period of oscillation. Considering that (cos2cot
= 1/2 and (sin cot cos cot) = 0, we obtain
p) Clam
(
cos T. (11.50)
2
19*
This expression can be written in a different form if we

take into account that, as follows from the vector diagram
(see Fig. 11.4), Um cos cp = RI „2. Hence,
(P)-= —12-Rnn. (11.51)
This power is equal to that of the direct current I =

The quantities
/ = /m/V2, (11.52)
are called effective (root-mean-square) values of current and
voltage. All ammeters and voltmeters are graduated for
r.m.s. values of current and voltage.
The expression for the average power (11.50) in terms of
r.m.s. values of current and voltage has the form
(P) = UI cos cp, (11.53)
where the factor cos cp is usually called the power factor.
Thus, the power developed in an a.c. circuit depends not
only on the voltage and current but also on the phase angle
between them.
When cp = a/2, the value (P) = 0 regardless of the val-
ues of U and I. In this case, the energy transmitted over a
quarter of a period from a generator to an external circuit
is exactly equal to the energy transmitted from the external
circuit to the generator during the next quarter of the period
so that the entire energy "oscillates" uselessly between the
generator and the external circuit.
The dependence of the power on cos cp should be taken into
account while designing transmission lines for alternating
current. If loads being fed have a high reactance X, cos w
may be considerably less than one. In such cases, in order
to transmit the required power to a consumer (for a given
voltage of the generator), it is necessary to increase current
I, which leads to an increase of useless energy losses in feed-
ing wires. Hence, loads, inductances and capacitances should
be always distributed so as to make cos (1) as close to one
as possible. For this purpose, it is sufficient to make reactance
X as small as possible, i.e. to ensure the equality of in-
ductive and capacitive reactances (X L = X0).
When the conductors forming an a.c. circuit are in motion,
Problems 293
the concept of electric resistance becomes wider since in ad-

dition to the conversion of electric energy into Joule's
heat, other types of energy transformations are also possi-
ble. For example, a part of electric energy can be converted
into mechanical work (electric motors).
Problems
11.1. Free undamped oscillations. Free undamped oscillations
occur in an oscillatory circuit consisting of a capacitor with capacitance
C and an induction coil with inductance L. The voltage amplitude
on the capacitor is Um. Find the e.m.f. of self-induction in the coil at
the moments when its magnetic energy is equal to the electric energy
of the capacitor.
Solution. According to Ohm's law,
RI = U `60
where U is the voltage across the capacitor (U = (pi — y2). In our
case, R = 0 and hence is = —U.
It remains for us to find voltage U at the moments when the electric
energy of the capacitor is equal to the magnetic energy of the coil.
Under this condition, we can write
CU 2 1 L/2 CU 2
2 2 2 2 /
whence I U I = Um/ 1727
es
As a result, we have I I = Um/ 1127
11.2. An oscillatory circuit consists of an induction coil with
inductance L and an uncharged capacitor of capacitance C. The resist-
ance of the circuit R = 0. The coil is in a permanent magnetic field.
The total magnetic flux piercing all the turns of the coil is cp. At the
moment t = 0, the magnetic field was abruptly switched off. Find
the current in the circuit as a function of time t.
Solution. Upon an abrupt switching off of the external magnetic
field at the moment t = 0, an induced current appears, but the capac-
itor still remains uncharged. In accordance with Ohm's law, we have
d€13 r dl
RI — —
dt dt •
•
In the given case R = 0, and hence (P + LI = 0. This gives (P = L/0,
where /0 is the initial current (immediately after switching off the
field).
After the external field has been switched off, the process is de-
scribed by the following equation:
dI
0=— —L (1)
dt
Differentiation of this equation with respect to time gives
This is the equation of harmonic oscillations. We seek its solution

in the form
/ = /m cos (coot + a).
The constants /, and a can be found from the initial conditions
(o) —1,, I(0)=0
(the second condition follows from Eq. (1), since at the initial moment
t = 0 the capacitor was uncharged). From these conditions we find
a = 0 and /m = /0. As a result, we obtain
I = /0 cos coo t = ((D /L) cos coot,
where coo = 1/ VLC.
• 11.3. Q-factor of a circuit. An oscillatory circuit with a low
damping has a capacitance C and inductance L. In order to sustain
in it undamped harmonic oscillations with the voltage amplitude
across the capacitor, it is necessary to supply the average power (P).
Find the Q-factor of the circuit.
Solution. Since damping is low, we can make use of formula (11.23):
Q = 2nW/6W, (1)
where W = CULI2 and 6147 = (P) T, T being the period of damped
oscillations. In our case, T T o = 2a-c }" LC. Having substituted
these expressions into (1), we obtain
Q — 2 (P) V L•
• 11.4. Damped oscillations. An oscillatory circuit includes a
capacitor of capacitance C, an induction coil of inductance L, a resistor
of resistance R, and a key. The capacitor was charged with the key
open. When the key was closed, a current began to flow. Find the ratio
of the voltage across the capacitor at time t to the voltage at the initial
moment (immediately after closing the key).
Solution. The voltage across the capacitor depends on time in the
same way as the charge does. Hence we can write
U=Ume -(3t cos (cot+ a). (1)
At the initial moment t = 0, the voltage U (0) = Um, cos a, where
Um is the amplitude at this moment. We must find U (0)IU,, i.e.
cos a.
For this purpose, we shall use another initial condition: at the
moment t = 0, current / = q• = 0. Since q = CU, it is sufficient to
Problems 295
differentiate (1) with respect to time and equate the obtained expres-
sion to zero at t = 0. We obtain —13 cos a — co sin a = 0, whence
tan a -= —f3/(o. The required ratio is
U (0) 1 1
— cos a — (2)
Urn 471±tan2 1/1±(13/6))2
The quantities Um and U (0) are shown in Fig. 11.8.
Considering that (o2= cog — 132, we transform (2) as follows:
U (0)/Um = 1/1— 0/0)02 = — R2C/4L,
where we took into account that p = RI2L and coa = 1/LC.
• 11.5. In an oscillatory circuit with a capacitance C and induc-
U
Um
U(0)
Fig. 11.8 Fig. 11.9
tance L damped oscillations occur, in which the current varies with

time in accordance with the law I (t) = /7,,e-f3t sin wt. Find the volt-
age across the capacitor as a function of time.
Solution. Let us choose the clockwise direction as the positive
direction of circumvention (Fig. 11.9). According to Ohm's law for
section 1RL2 of the circuit, we have RI = Ti — W2 ± In our
case, = —LI and q)2 — cPi = q/C = tic, where q is the charge on
plate 2. Hence Ohm's law can be written as
(lc = —RI—Li.
Having substituted into this formula the expression for I (t) and
its derivative, we obtain
RIme -13i
Uc 2/3 ( 3 sin on—co cos cot) .
Let us transform the expression in the parentheses to sine. For
this purpose, we multiply and divide it by 17(o2 + 132 = (Lk, and in-
troduce angle 6 through the formulas
—13hoo = cos 6, co/60 = sin 6. (1)
Then
RITyy(Do e'Rt
c— 2/3 sin (cot — 6)= In, LIde-13tsin (cot — 6),
where angle 6 is, in accordance with (1), in the second quadrant, i.e.
assumes the values 31/2 < 6 < at. Thus, the voltage across the capacitor
lags behind the current in phase.
11.6. Steady-state oscillations. An induction coil of inductance
L and resistance R was connected at the moment t = 0 to a source of
external voltage U = Urn cos wt. Find the current in the circuit as
a function of time.
Solution. In our case, RI = U — LI, or
I (RIL) I =Urncos wt.
The solution of this equation is the general solution of the homogeneous
equation plus the particular solution of the nonhomogeneous equation:
I (t)= Ae-(R/L)t—
VI?UM
2 + 02L2 cos (6) t —(1)),
where A is an arbitrary constant and angle IT is defined by condition

(11.36): tan cp = coLIR.
The constant A is found from the initial condition I (0) = 0.
Hence A --- —(Ur„l }/R2+ (D2L2) cos cp. This gives
I (t)-- U rn [cos (cot — cp)—e-(R/ L)t cos (0.
1/R2 + CO2L2
For a sufficiently large t, the second term in the brackets becomes
negligibly small, and we obtain the steady-state solution I (t) 0(
a cos (tat — cp).
011.7. Forced oscillations. A section of a circuit consisting of
a series-connected capacitor and resistor R, is connected to a source
of varying voltage with the amplitude Um. The amplitude of the
steady-state current turned out to be /m. Find the phase angle between
the current and external voltage.
Solution. In the case under consideration,
U = Urn cos cot, I = /in cos (wt — (p),
where cp is defined by formula (11.36): tan cp = —1/o)CR.
The unknown value of capacitance C can be found from the expres-
sion for the current amplitude: /,n = Und VR2 + (1/0)02, whence
C 1/e) V(Um //m )2— R2.
Having substituted this expression into formula for tan q), we obtain
tan cp.-- — ii(Um/R/m)2-1.
Problems 297
In our case cp < 0, which means that the current leads the external
voltage (Fig. 11.10).
• 11.8. An a.c. circuit, containing a series-connected capacitor
and induction coil with a certain resistance, is connected to the source
1,,
co C
Fig. 11.10 Fig. 11.11
of external alternating voltage whose frequency can be altered without

changing its amplitude. At frequencies (piand co, the current ampli-
tudes in the circuit proved to be the same. Find the resonance frequency
of the current.
Solution. According to (11.35), the amplitudes are equal under the
condition
1
col L o.)2L — I. (1)
co1C
The maximum of the resonance curve for current corresponds to the
frequency equal to the natural frequency col) = 1l VIC. Further,
let col < coo < 0)2 (the opposite inequality is also possible, it will
not affect the final result). Equality (1) can then be written in a more
general form: w8/0)1— wl = w2 — (4/(02, or
9 1 1
— = (06 — )•
wi
Cancelling out (02 - col on both sides of this equality, we obtain
1= (4/(1)0)2, whence
oo=
• 11.9. Vector diagram. A circuit consisting of a series-connected
capacitor of capacitance C and induction coil with resistance R and
inductance L is connected to an external voltage with the amplitude
Urn and frequency co. Assuming that current in the circuit leads in
phase the external voltage, construct the vector diagram and use it
for determining the amplitude of voltage across the coil.
Solution. The vector diagram for the case under consideration is
shown in Fig. 11.11. It is readily seen from this diagram that the
20-0181
amplitude of voltage across the coil is

Ur..rtm=im R 2-{' OP,
where /m = Umil/R2 + (c)L — 1/(oC)2. In the presence of resistance
the voltage across the coil leads the current by less than Tr/2.
•11.10. Power in an a.c. circuit. A circuit consisting of a series-
connected resistor R with no inductance and an induction coil with
a certain resistance is connected to the main system with the r.m.s.
voltage U. Find the thermal power developed in the coil if the r.m.s.
values of voltage across the resistor R and the coil are equal to U1
and U2 respectively.
Solution. Let us use the vector diagram shown in Fig. 11.12. In
accordance with the law of cosines we obtain from this diagram
U2 = Uf 2U1U2cos (pL. (1)
The power developed in the coil is
P2 = 1U2 cos (pL, (2)
where I
Combining Eqs. (1) and (2), we obtain
P 2 = (U2— UT U2)/2R.
Appendices
1. Notations for Units of Measurement
A ampere min minute

C coulomb Mx maxwell
eV electronvolt N newton
F farad Oe oersted
G gauss ohm
g gram rad radian
H henry s second
hr hour S siemens
Hz hertz T tesla
joule V volt
K kelvin W watt
m meter Wb weber
2. Decimal Prefixes for Units of Measurement
T, thera-, 1012 d, deci-, 10-1 n, nano, 10-9

G, giga-, 109 c, centi-, 10-2 p, pico-, 10-22
M, mega-, 106 m, milli-, 10-3
k, kilo-, 103 IA, micro-, 10-6
h, hecto-, 103
da, deca-, 103
3. Units of Measurement of Electric and Magnetic

Quantities in SI and Gaussian Systems
Unit of mea-
Nota- surement SI unit
Quantity tion Ratio
unit
SI I CGS
Force F N dyne 102

Work, energy A, W erg 107
Charge q C CGSE unit 3X 109
20*
300 Appendices
Table 3. Cont.
Unit of mea-
Nota- surement SI unit
Quantity tion Ratio SI
unit
SI CGS
Electric field strength E V/m CGSE unit 1/(3 x 104)

Potential, voltage cp, U V CGSE unit 1/300
Electric moment P C •ra CGSE unit 3 X 1011
Polarization P C/M 2 CGSE unit 3 X 105
Vector D D C/m2 CGSE unit 12a x 105
Capacitance C F cm 9X 1011
Current I A CGSE unit 3 x 109
Current density j A/m2 CGSE unit 3 x 105
Resistance R Q CGSE unit 1/(9 x 1011)
Resistivity p Q • m CGSE unit 1/(9 x 109)
Conductance 2 S CGSE unit 9 x 1011
Conductivity a S/m CGSE unit 9 x 109
Magnetic induction B T G 104
Magnetic flux, magnetic-
flux linkage (l), 111 Wb Mx 108
Magnetic moment Pm A•m2 CGSE unit 103
Magnetization J A/m CGSE unit 10-3
Vector H H A/m Oe 4a-c x10-3
Inductance L H cm 109
4. Basic Formulas of Electricity and Magnetism

in SI and Gaussian Systems
Relation SI Gaussian system
Field E of a point 1 q q
charge --
4neo r2 r2
Field E in a parallel-
plate capacitor and 4na
E- E-
at the surface of a 8 o8
conductor
Potential of the field 1
of a point charge (1)2= 43teo r = r
2
Relation between E E= E dl
and cp •(1), (P1 —(1)2=
Appendices 301
Table 4. Cont.
Circulation of vector
E in an electros- E d1=0
tatic field
Electric moment of a P=q1
dipole
Electric dipole p in
field E F=p OE
0/
, M=[p X E], W= —p-E
Relation between po-
larization and field P = xe0E P =x E
strength
Relation between o-', (5' -= Pn=x80En o =P,=xE,
P and E
Definition of vector D D=e0E-FP D =E -F4nP
Relation between a 8=1±X e =1+47m
and x
Relation between D D = eaoE D = eE
and E
Gauss theorem for
vector D DdS= q D dS= 4nq
Capacitance of a ca- C =--
pacitor
Capacitance of a par- 880S 8S
allel-plate capaci- C= C
tor h 41th
Energy of system of
charges Tv= T2 qi(pt
Total energy of in-
teraction pq) dV
Energy of capacitor W = qU 12 = CU 2/2 = q2/2C
Electric field energy E•D E •D
density w=
w= 2 8n
Continuity equation j dS= — .j= — :
Ohm's law RI = (Pit-cP2+412, j= (E± E*)
Joule-Lenz law Q = RI 2, Qs p= p12
Lorentz force F =qE.-1-9 [v XBJ F = qE +.÷[v X II]
302 Appendices
Table 4. Cont.
Field B of a moving o q[vX r]

charge B= g B= 1 q[v xr]
4n r3 c r3
Biot-Savart law dB= Ro
[i X r] dV I [i X r ] dV
dB =
4n r3 c r3
µo I [d1 x r]
dB = 4rt dB 1 /[d] X r]
r3 r3
Field B
(a) of straight 21 1 21
current 43c b c b
(b) at the centre of B ._ Pi 2311 1 2311
a loop 431 R c R
(c) in solenoid B=p,onI B— nI
Ampere's law dF=/ [d1 X B] dF=—[d1 X B]

1
dF -= [j x B] dV dF= [j x B] dV
c
Force of interaction
between parallel Ro 2/1/2 2 /112
currents 4n b Fu=
Magnetic moment of
a current loop pm = IS
Magnetic dipole pm OB
in field B F — Pm —aTi M = [Pm X B]
Work done in displace-
ment of a current A= I (02 -191)
loop A= 1 I (02-01)
Circulation of mag-
netization J dI = /' J d1=—
1 I'
c
Definition of vector H H--= B/Ro — H = B —4nJ
Circulation of vector
H in a stationary H dl= 4n
() H dl = — I
field c
Relations between
J and H J xH
p, and x g=1-Fx p, = 1 +43-cx
B and H B = p,g0 H
- B=p,H
Appendices 303
Table 4. Cont.
Laws of transforma- 1
tion of fields E and B E' =E+[voX 13] E' =E+ [vo x BI
for vo<< c 1 1
B' B— — [vo xE]
[v o XE]
c
Electromagnetic field E.B= inv
invariants E 2— c2B2 =inv E 2— B2 =inv
Induced e.m.f. &XI 1 dc1:1
— =—
dt c dt
Inductance L=-
- L=c0II
Inductance of a so-
lenoid Rixon2V L= 4nun2V
E.m.f.. of self-induc- dI dl
tion dt c2 L dt
Energy of the mag- LI 2 ,7 1 LI 2
netic field of a cur- W= vy=---
2 c 2
rent
Magnetic field energy B•H B•H
w- w=
density 2 8n
Displacement current OD it OD
density jd -
at jd= 4n at
Maxwell's equations r
in integral form E dl = — 13 dS (1Edl= dS
D dS= p dV D OS= 4n p dV
H dl = c (j+D)dS H dl
•
(j+-T dS ht
--
BdS=0 B dS= 0
Maxwell's equations
in differential form VXE=—B V X E — 113
•D= p V •D =4.np
4n
V X 11 =i+i) cXH= c 0+2. )
V•B=0 V•B= 0
304 Appendices
Tabl. 4. Cont.
Relation between E
and H in electro- E V soe-= HVI.41.0 E "V" =-- H
magnetic wave
Poynting's vector S=[EXH] [E X HI
43t
Density of electro- 1
magnetic field mo- G=— [E X Ill G— [E X11]
mentum c2 4c
n
5. Some Physical Constants
Velocity of light in vacuum c= 2.998 x 108 m/s

G=. I 6.67 x 10-n m3 /(kg•s2)
Gravitational constant!
6.67 x 10-8 cm3/(g•s2)
Acceleration of free fall g= 9.807 m/s2
Avogadro constant NA= 6.022 x 1023 mole-1
1.602 x 10-19 C
Charge of electron or proton e=
{ 4.80 x 10-1° CGSE units

Rest mass of electron me = 0.911 x 10-30 kg
e { 1.76 x 1011C/kg
Specific charge of electron
me 5.27 x 1017 CGSE units/g
Rest mass of proton mp=1.672 x 10-27 kg
Electric constant 80 =- 0.885 X 10-11F/m
1/42-cs0 =- 9 X 109 m/F
Magnetic constant tto =1.257 x 10-6 H/m
R0 /41t =10-7 H/m
Relation between velocity c= 1/ V sop°
of light and so and flo
Subject Index
Ampere, 117 electrodynamic, 144

magnetic, 143
Coulomb, 12
Betatron, 222 current,
boundary conditions, 80, 92, 192 alternating, 290ff
for B and H, 182ff conduction, 174
direct, 178
displacement, 253f, 271
Capacitance, 57ff, 86, 90
induced, 248
of cylindrical capacitor, 60
magnetization, 174
of isolated conductor, 57
surface, 175, 178, 192
of parallel-plate capacitor, 59
volume, 175
of parallel wires, 65
molecular, 174, 177
of spherical capacitor, 59
polarization, 256
unit of, 58
density of, 259
capacitor, 57ff, 86, 90
quasistationary, 278
charging, 134
straight, 188
discharging, 133
total, 254
energy of, 110
current element(s),
parallel-plate, 103, 108f, 114f,
137 linear, 146
charge invariance, 198f volume, 146
circulation, curve,
of vector B, 166 magnetization, 189
of vector E, 245 resonance, 288f
of vector H, 179, 193
of vector J, 176ff, 193
condition, betatron, 246, 260 Damping factor, 280, 283
quasi-steady, 277 diamagnetism, 180
conductor in electrostatic field, 45 dielectric(s),
homogeneous, 120
nonhomogeneous, 138 anisotropic, 101
constant, isotropic, 101
dielectric (electric), 12, 78f, 86, liquid, 108f
102 polarization of, 67
306 Subject Index
dielectric susceptibility, 71 electrostatic shielding, 51

dipole, e.m.f.,
electric, 34, 44 induced, 244
energy of, 38 of self-induction, 227
moment of, 34ff energy,
dipole moment, 71 of charged capacitor, 99f, 139
intrinsic, 68 of charged conductor, 99f
domain(s), 191 of current,
intrinsic, 238
magnetic, 236
Edge effects, 59, 66, 86, 100 mutual, 238
electret(s), 68, 79 of electric field, 94, 100, 103,
electric charge(s), 11, 83, 88f, 92 105, 107, 237, 276
bound (polarization), 69f, 731 of electromagnetic field, 253
bulk, 69f localization, 10011, 112
at conductor surface, 83 magnetic, 243
surface, 69f of two current loops, 238
in current-carrying conductor, of interaction, 96ff
120 for system of point charges,
extraneous, 78, 88f, 109, 138 97, 105, 110f
fictitious, 63 total, 97, 99f, 110
induced, 46, 66 intrinsic, 98, 105, 110f
surface, density of, 63 of magnetic field, 234, 236f,
electric circuit, 275f
a.c., 291, 297 of current, 234
branched, 126ff energy flux, 266
oscillatory, 293 density of, 264
electric current, 116, 119 equation(s),
density of, 117 continuity, 116, 118, 261
direct, 116, 120 in differential form, 118
linear, density of, 153 Laplace, 54f
quasistationary, 133 material, 260
steady-state, 118 Maxwell, 253ff, 257f, 272, 274
electric oscillations, 277ff in differential form, 259
electric resistance, 119 in integral form, 257
methods of calculation, 120 properties of, 261f
electromagnetic induction, 217ff, symmetry of, 262
225 of oscillatory circuit, 277, 279f
electromagnetic wave(s), 101, 262f Poisson, 54f
pressure of, 268, 270 equipotential surfaces, 32, 471
in vacuum, 269 57
Subject Index 307
Farad, 58f potential, 26, 28, 155

ferroelectric(s), 72 sinks of, 25
ferromagnetic(s), 180, 188f, 232 solenoidal, 155
ferromagnetism, 188, 191 sources of, 25
theory of, 191 vector,
field, circulation of, 15, 25
Coulomb, 122 divergence of, 24
electric, 11, 198, 206 flux of, 15
in charge-carrying conductor, force(s),
121 Ampere, 155, 157, 159, 169, 171,
in dielectrics, 61 225
intensity (strength) of, 11ff, moment of, 160, 169
28
work of, 161
macroscopic, 45ff
coercive, 190
microscopic, 45
Coulomb, 122
of point charge, 12
in dielectric, 106ff
electromagnetic, 198f
electric, 157
momentum of, 268
extraneous, 122f, 235
density of, 269
induced electromotive, 217
electrostatic, 155
force(s),
in vacuum, 11
of interaction,
of freely moving relativistic
electric, 144f
charge, 207
magnetic, 144f
in homogeneous magnetic, 185
Lorentz, 141ff, 203, 212f
magnetic, 142, 145, 155, 198,
206 magnetic, 157
on the axis of circular cur- in magnetic, 196
rent, 147 surface density of, 109f
of current-carrying plane, 153 flux,
energy of, 240 magnetic,
forces in, 240 conservation of, 230
in magnetic, 175 total, 212
permanent, 220 of vector D, 273
of solenoid, 151, 167, 181 formula, Thomson, 281
of straight current, 151 frequency,
in substance, 172ff natural, 280
of toroid, 152 resonance, 289
of uniformly moving charge, Generator, MHD, 224
143
in vacuum, 141ff Henry, 227
varying, 221 hysteresis, 72, 180
vortex, 222, 275 magnetic, 190
308 Subject Index
Image charge, 56 Newton's third, 95

impedance, 290f Ohm, 119, 127, 131, 134, 136ff,
inductance, 226, 238 140f, 233, 249, 293, 295
calculation of, 250f in differential form, 120
mutual, 231, 240, 251 generalized, 123
induction, in integral form, 124
electrostatic, 46 for nonuniform subcircuit,
123, 138
magnetic, 1431f
of transformation for fields E
direct calculation of, 163
and B, 200ff, 206, 213
mutual, 233, 251
logarithmic decrement of
residual, 190 damping, 283
interaction(s), 11
electromagnetic, 11
gravitational, 11 Magnet, permanent, 68, 195
of parallel currents, 168 magnetic(s), 172, 174, 181
strong, 11 homogeneous, 174
weak, 11 nonhomogeneous, 175, 177
invariants of electromagnetic magnetic constant, 143
field, 208, 214 magnetic field intensity, 179
magnetic flux, 167
linkage of, 219
Joule heat, 235
magnetic induction, 143f, 146,
163, 166, 168, 187
magnetic protection, 185
Law(s),
magnetism, relative nature of, 204
Ampere, 155ff
magnetization, 172ff, 195
Biot-Savart, 145ff, 149f, 153,
164 mechanism of, 177
of conservation, residual, 190
of electric charge, 11, 118 magnetohydrodynamics, 244
of energy, 106, 114, 130, 139, method,
235, 241 energy, 106, 240
Coulomb, 109 image, 54ff, 62f, 67
in field form, 12 molecule(s),
Joule, 129ff nonpolar, 68
in differential form, 131 polar, 68
Kirchhoff, 126, 256 moment, magnetic, 157ff
first, 126
second, 126f
Lenz, 217ff, 243 Ohm, 119
of magnetic field, in differen- oscillation(s),
tial form, 194 damped, 281f, 294
Subject Index 309
oscillation(s), Self-induction, 226

forced, 294 solenoid, 151
free, 280 superconductor(s), 230
steady-state, 285, 296 susceptibility, magnetic, 180
undamped, 280
Tesla, 144
theorem,
Paramagnetic(s), 180, 194, 197 on circulation of field vectors,
permeability, 180, 189 15, 25, 52, 80f, 137f, 148f,
point, Curie, 191 151ff, 155, 166, 178f, 193f,
polarization of dielectric, 68, 73, 195f, 211, 248, 255
86, 103 Gauss, 15ff, 19, 21ff, 25, 33,
polarization, 70ff 41, 44, 47f, 59, 65, 72,
76f, 79ff, 83, 85, 87ff, 91,
unit of, 71 93, 102, 112, 121, 137,
potential, 25, 27, 30, 33, 43f, 148, 150, 157, 168, 173,
47f, 53f, 57, 61, 64, 66, 200, 211
86f, 88, 97f, 102, 124 application of, 19
of point charge, 28 in differential form, 23ff, 73
of sphere, 58 Poynting's 264
of system of charges, 29 of reciprocity, 231, 252
potential difference, 86, 90, 99 uniqueness, 541f
power liberated in a.c. circuit, theory of relativity, 269
291f, 298
thermal power, 132
power factor, 292
pressure, magnetic, 244 transformation(s), Lorentz, 201
principle of superposition, 12f,
18, 29, 60, 66, 146, 166f
process, transient, 133, 140 Vector,
D, 76ff
H, 178
Poynting's, 265, 269, 274f
Q-factor, 284, 289, 294 vector diagram, 287, 297
volt, 28
voltage, 59
Reflection coefficient, 270
refraction of B-lines, 184
relaxation time, 134, 283 Work,
resistance of conducting medium of displacement of current
136 loop, 161ff
resistivity, 119 of electric field, 103f
resonance, 289f of e.m.f. source, 139
TO THE READER .
Mir Publishers would be grateful for

your comments on the content, translation
and design of this book. We would also be
pleased to receive any other suggestions
you may wish to make.
Our address is:
Mir Publishers
2 Pervy Rizhsky Pereulok
I-110, GSP, Moscow, 129820
USSR
Printed in the Union of Soviet Socialist Republics

Also from Mir Publishers
Fundamental Laws of Mechanics
I. lrodov
The book considers the basic laws of mechanics—laws of mo-

tion and laws of conservation of energy, impulse and moment
of impulse. The possibilities of their applications are indicated.
A large number of problems and exercises are included.
The material is arranged in accordance with the syllabus of
a course in general physics for university students. The book
is meant for university students who intend to specialize in
physics.
Handbook of Physics
B. Yavorsky and A. Deflaf
A companion volume to Vygodsky's Handbook of Higher Math-

ematics, designed for use by engineers, technicians, research
workers, students, and teachers of physics. Includes definitions
of basic physical concepts, brief formulations of physical laws,
concise descriptions of phenomena, tables of physical quanti-
ties in various systems of units, universal physical constants,
etc.

Basic Laws of Electromagnetism Irodov PDF

Uploaded by

Copyright:

Available Formats

Basic Laws of Electromagnetism Irodov PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Basic Laws of Electromagnetism Irodov PDF

Uploaded by

Copyright:

Available Formats

What is the purpose of the book?

What is the purpose of the book?

Who is the author of the book?

Who is the author of the book?

Besides the theoretical descrip-

tion of the basic laws of electro-

HmaTeabemo «Bbicmaff unto.na*

Mir Publishers Moscow

143AaTeJlECTBO d3LICIllaFf inaoza*, 1983

The main idea behind this book is to amalgamate the de-

The book is intended as a textbook for undergraduate stu-

1. Electrostatic Field in a Vacuum . 11

2. A Conductor in an Electrostatic Field • 45

3. Electric Field in Dielectrics 67

4. Energy of Electric Field 94

5. Direct Current 116

6. Magnetic Field in a Vacuum 141

7. Magnetic Field in a Substance 172

8. Relative Nature of Electric and Magnetic Fields 198

9. Electromagnetic Induction 217

9.7. Energy and Forces in Magnetic Field 240

10. Maxwell's Equations. Energy of Electromagnetic Field 253

11. Electric Oscillations 277

Vectors are denoted by the bold-face upright letters

1.1. Electric Field

that its introduction does not noticeably distort the field

where co is the electric constant and er is the unit vector of

where ri is the distance between the charge q i and the point

where dq is the charge contained in the volume dV, on the

of the type (1.5). As a rule, the problem becomes much

Fig. 1.1 Fig. 1.2

where a, is the maximum value of the angle a, sin a0 = 1/ 1n2 + x2.

In this case also E q/4:-ce,x2 for x » 1 as the field of a point charge.

Geometrical Description of Electric Field. If we know

1.2. The Gauss Theorem

This is an algebraic quantity, since it depends not only on

The Gauss Theorem. The flux of E through an arbitrary

E dS= gin, (1.7)

where the circle on the integral symbol indicates that the

Fig. 1.4 Fig. 1.5

E through a closed surface is equal to the algebraic sum of the

In particular, this leads to the following conclusion: if

case, in accordance with the principle of superposition E =

= dS+ E2 dS ... =01+02+ ...

In accordance with what was said above, each integral on the

where the integration is performed only over the volume

stance: while the field E itself depends on the mutual con-

1.3. Applications of the Gauss Theorem

Hence it follows that in any electrostatic field a charge cannot be in

Fig. 1.6 Fig. 1.7

The flux through the lateral surface of this cylinder is equal to

Fig. 1.8 Fig. 1.9

For X > 0, Er >0 as well, i.e. vector E is directed away from

Er= q (r> a), (L13)

where Er is the projection of vector E onto the radius vector r coinciding

General Conclusions. The results obtained in the above

examples could be found by direct integration (1.5) as well.

1.4. Differential Form of the Gauss Theorem

For this purpose, we first represent the charge q in the volume V