1

ELECTROMAGNETIC FIELDS FOR

ENGINEERS AND SCIENTISTS, Vol.1

1) A supplementary textbook, applications oriented, for Engineers,

Physicists, and Mathematicians, ideal for independent study.

2) Covers in depth, Electrostatics, Magnetostatics, Steady Currents,

Time-Varying Fields and Maxwell’s equations.

3) 245 fully worked illustrative examples and 867 problems for

solution.

4) Odd numbered problems are provided with answers.

 2

ABOUT THE AUTHOR

Demetrios P. Kanoussis received his bachelor’s degree in Electrical
Engineering from the National Technical University of Athens (Greece), in 1980.
After that, he received his master’s degree in Mathematics, in 1983, and his
Ph.D in Electrical Engineering, in 1986, both from Tennessee Technological
University (TTU), U.S.A.

Demetrios P. Kanoussis has a long teaching experience and has taught

various courses in the areas of Applied Mathematics and Electrical Engineering.
During his stay at Tennessee Technological University, he taught several
courses, in the EE Department, mainly in the areas of Electromagnetic Fields,
Physical Electronics and Plasma Engineering, Junior and Senior level.

Regarding his work experience, Demetrios P. Kanoussis has been actively

involved, as a professional Electrical Engineer, in the design and the
implementation of various projects, primarily in the field of Integrated Control
Systems.

Dr. Kanoussis is the author or coauthor of numerous research articles. His

original scientific research and contribution to Mathematics and Electrical
Engineering has been published, among other places, in various, high impact
international journals.

In addition to his professional activities, teaching and research, Demetrios

Kanoussis is the author of numerous textbooks in Electrical Engineering and
Applied Mathematics.
 3

ELECTROMAGNETIC FIELDS FOR ENGINEERS AND

SCIENTISTS, Vol. 1
All rights reserved. No part of this publication may be reproduced,
distributed or transmitted in any form or by any means, electronic or
mechanical, without the written permission of the author, except in the case of
brief quotations and certain other noncommercial uses permitted by copyright
law.

Thank you for respecting the work of this author.

Demetrios P. Kanoussis, Ph.D

 4

PREFACE
Electromagnetism constitutes a core subject in an engineering curriculum
for many reasons; it provides an in depth understanding of electric and
magnetic phenomena, predicts the existence and the propagation of EM
waves, investigates the propagation of EM waves in unbounded and in
bounded media, unifies optics with electromagnetic phenomena, etc.

A great number of applications follow directly from Maxwell’s equations,

which are the fundamental equations of the EM fields. Capacitors, Inductors,
derivation of circuit theory from field theory, transmission lines, radiating
systems (antennas), microwaves, lasers, motion of particles in electric and
magnetic fields, transmission of EM energy and wireless communication, just
to mention a few, are investigated and analyzed by means of Maxwell’s
equations. All these applications require a solid understanding of
Electromagnetics.

The central theme in this book is the development of Maxwell’s equations.

Newton’s equations in Mechanics and Maxwell’s equations in
Electromagnetism, are among the most influential equations in science and
technology. Every macroscopically observed electromagnetic phenomenon can
be analyzed and explained, by means of Maxwell’s equations.

In this book we use the historical approach of developing Maxwell’s

equations; that is we start with the relevant experimental laws, (Coulomb’s
law, Gauss’s law, Biot-Savart law, Faraday’s law of induction, etc), and
gradually build, step by step, Maxwell’s equations. The last, but very decisive
step towards the final formulation of the equations (today known as Maxwell’s
Equations), was made by Maxwell himself, by his ingenious introduction of the
“displacement current density”, in order to make the, then, existing equations
compatible with the “continuity equation”, which expresses the conservation
of the electric charge. This method of developing Maxwell’s equations has the
advantage of introducing the student, gradually, step by step, to ideas,
principles, methods and analytical techniques, relevant to the subject matter
under study.

For the interested reader, in the Appendix, we present an axiomatic

derivation of Maxwell’s equations. Taking this approach, the totality of our
 5

knowledge and experience, related to electromagnetic phenomena, is

summarized into four postulates, which, in conjunction with the principle of
conservation of energy, lead to Maxwell’s equations.

We assume that the readers of this book have already been introduced to
vector calculus and differential equations. However, since it is likely that many
readers are not familiar, or, at least, have little experience with cylindrical and
spherical coordinates, vector algebra, unit vectors, multiple integrals, line
integrals, surface integrals, gradient, divergence, curl, etc., a concise and
systematic review of vector calculus is given in Chapter 1.

A summary of numerous important formulas and vector identities,

(derived in Chapter 1), is presented at the beginning of the book, right after the
Table of contents. The reader may thus find there, concentrated in one place,
all the formulas and vector identities, used repeatedly throughout this
textbook. This textbook is mathematically self-contained.

Solved examples and problems for solution are provided at the end of each
section. A number of supplementary problems, (usually more difficult), are
provided at the end of each Chapter. Hints are often given for the more
involved problems.

The book contains 245 illustrative, fully solved examples and 867 problems
for solution.

The International System of Units (SI), a standard system in Engineering and

Physics, is used throughout this textbook. This system is based on six basic
units: the meter ( ) for length, the kilogram ( ) for mass, the second ( )
for time, the amber ( ) for electric current, the degree Kelvin ( ) for
temperature and the candela ( ) for luminous intensity. All the other units,
called secondary units, are derived from these six, fundamental units.

Chapter 1, as already mentioned, is devoted to a systematic summary of

vector analysis, (concepts, theorems, formulas, etc), necessary for the
understanding of the material under study.

The rest of the book is divided in three parts.

Part A: Electrostatics (Chapters 2 12)

 6

Part B: Magnetostatics (Chapters 13 23)

Part C: Time – Varying Fields and Maxwell’s Equations (Chapters 24, 25)

A brief description of the topics covered in each part is given below.

CHAPTER 1: MATHEMATICAL INTRODUCTION

(Brief description: For analytic description see Table of Contents)

Review of vector algebra, coordinate systems (Cartesian, Cylindrical and

Spherical coordinates), transformations between coordinate systems, unit
vectors in the three coordinate systems, scalar and vector fields, conversion of
a scalar or a vector field from one coordinate system to another,
differentiation and integration of vector fields, multiple integrals (double,
triple, line and surface integrals), the gradient of a scalar field, divergence and
curl of vector fields, the Del operator ( ), Laplacian of a scalar field, physical
meaning of gradient, divergence and curl, Gauss-Ostrogradsky theorem,
Stokes’ theorem, expressions of the gradient, the divergence, the curl and the
Laplacian in Cartesian, Cylindrical and Spherical coordinates, solenoidal and
irrotational fields, source-point and field-point notation.

PART A: ELECTROSTATICS (CHAPTERS 2 12)

(Brief description: For analytic description see Table of Contents)

Discrete and continuous charge distributions, Coulomb’s law, the electric

field , calculation of the electric field using the superposition principle, the
two fundamental laws of static electric fields, electric flux, Gausssian surfaces
and applications, the electric potential , Poisson’s and Laplace’s equation,
energy stored in electric fields, multipole expansion, conductors in static
electric fields, dielectrics in electric fields, polarization of electric charge and
bound charges, capacitors, special methods and techniques in Electrostatics,
(Orthogonal Functions, Fourier expansion, separation of variables method to
solve Laplace’s equation in Cartesian, Spherical and Cylindrical coordinates,
method of images).
 7

PART B: MAGNETOSTATICS (CHAPTERS 13 23)

(Brief description: For analytic description see Table of Contents)

Steady currents, current density, resistance, electromotive force, steady

currents and Laplace’s equation, the field of steady currents, Biot-Savart law,
calculation of the magnetic field using the superposition principle, the
magnetic potential , the two fundamental laws of static magnetic fields,
magnetic flux, Amperian loops and applications, solenoids and toroids,
magnetic forces and torques, motion of charged particles in electric and
magnetic fields, magnetic fields in matter, permanent magnets, self-inductance
and mutual inductance, energy stored in magnetic fields, Maxwell’s equations
for static electric and magnetic fields.

PART C: TIME-VARYING FIELDS AND MAXWELL’S EQUATIONS

(CHAPTERS 24, 25)

(Brief description: For analytic description see Table of Contents)

Faraday’s law of induction, motional emf, displacement current, Maxwell’s

equations for time - varying fields, retarded potentials, the Poynting vector and
transmission of electromagnetic energy.

APPENDIX: AXIOMATIC DERIVATION OF MAXWELL’S EQUATIONS

A second volume, covering EM waves in unbounded and in bounded media,

(waveguides), propagation of EM waves in dielectric and in conducting media,
electromagnetic radiation, simple radiating systems, derivation of Kirchhoff’s
laws (in circuit theory) from Maxwell’s equations, transmission of EM energy,
and various other applications of Maxwell’s equations, will appear in the near
future.
 8

TABLE OF CONTENTS
A SUMMARY OF IMPORTANT FORMULAS AND VECTOR IDENTITIES

A) Transformation between coordinate systems 18

(Rectangular, Cylindrical, Spherical)

B) Two important theorems 20

(Gauss-Ostrogradsky (or Divergence) Theorem, Stokes’ Theorem)

C) Useful vector identities 20

D) Expressions for the Gradient, the Divergence, the Curl and the
Laplacian in Rectangular, Cylindrical and Spherical Coordinates 21

CHAPTER 1: MATHEMATICAL INTRODUCTION

1-1) Review of basic definitions and properties of vectors 23

1-2) Coordinate systems, (Cartesian, Cylindrical, Spherical) 43

1-3) Transformations between coordinate systems 52

1-4) Transformation of a field, (scalar or vector), from one coordinate

system to another 62

1-5) Differential arc length and differential volume in Rectangular,

Cylindrical and Spherical coordinates 71

1-6) Differentiation and integration of vector fields 74

1-7) Multiple integrals, (double, triple, line and surface integrals) 80

1-8) Gradient of scalar fields, divergence and curl of vector fields, the Del
operator , the Laplacian of scalar fields 90

1-9) The Gauss-Ostrogradsky theorem, (divergence theorem) and Stokes

theorem 95

1-10) The physical meaning of the divergence, the curl and the gradient 101
 9

1-11) Expressions for the gradient, the divergence, the curl and the
Laplacian in Cylindrical and in Spherical coordinates 108

1-12) Solenoidal and irrotational fields, potential fields and vector

potentials 113

1-13) Source – point and Field – point notation 120

PART A: ELECTROSTATICS
CHAPTER 2: COULOMB’S LAW

2-1) Electric charge: Units and Properties 131

2-2) Distributions of electric charges: Discrete – Continuous 133

2-3) Coulomb’s law 139

2-4) Dielectric materials 141

2-5) Superposition principle and the general problem in Electrostatics 147

CHAPTER 3: THE ELECTRIC FIELD

3-1) Electric fields – Test charge 164

3-2) The electric field of a given charge distribution 164

3-3) Calculation of the electric field for some important charge

distributions, (point charge, uniform line charge distribution, uniform sheet of
charge, circular ring of charge, uniformly charged sphere, etc) 166

3-4) Field lines of the electric field 181

CHAPTER 4: THE TWO FUNDAMENTAL LAWS OF STATIC ELECTRIC

FIELDS

4-1) Introduction 188

4-2) The static electric field is irrotational, ( ) 189

 10

4-3) The electric displacement vector 196

4-4) Gauss’s law, ( ), Electric flux 198

4-5) Applications of Gauss’s law: Gaussian surfaces, (uniform line charge

distribution, a spherical ball of charge, cylindrical charge distribution, etc) 209

CHAPTER 5: THE ELECTROSTATIC POTENTIAL

5-1) Introductory concepts and definitions 222

5-2) Poisson’s equation 223

5-3) Laplace’s equation 225

5-4) Potential of a point charge , (Coulomb’s potential) 225

5-5) Potential of a charge distribution, (discrete and/or continuous) 235

5-6) Electric dipole and electric dipole moment 242

5-7) The physical meaning of the potential 246

5-8) Equipotential surfaces, (definition and properties) 255

5-9) Earnshaw’s theorem 257

CHAPTER 6: ELECTRIC FIELD ENERGY AND ENERGY DENSITY

6-1) Introductory concepts and definitions 265

6-2) Energy of a discrete charge distribution 266

6-3) Energy of a continuous charge distribution, energy density 270

CHAPTER 7: A SYSTEMATIC METHOD FOR CALCULATING

POTENTIALS OF AN ARBITRARY CHARGE DISTRIBUTION (MULTIPOLE
EXPANSIONS)

7-1) Fundamental concepts and definitions 282

 11

7-2) A systematic approach for the calculation of the potential of an

arbitrary charge distribution 283

CHAPTER 8: INTERACTION OF AN ARBITRARY CHARGE

DISTRIBUTION AND AN EXTERNALLY APPLIED ELECTRIC FIELD

8-1) Introduction 293

8-2) Calculation of the potential energy of a charge distribution formed

within an externally applied field 293

CHAPTER 9: CONDUCTORS IN STATIC ELECTRIC FIELDS

9-1) Conductors and dielectrics 302

9-2) Basic properties of conductors 303

9-3) Boundary conditions between conductors and dielectrics 305

9-4) Electrostatic pressure 309

9-5) Cavities within conductors 311

9-6) Green’s reciprocity theorem 314

9-7) Induced charges 318

9-8) Energy of a system of conductors 324

CHAPTER 10: DIELECTRICS IN STATIC ELECTRIC FIELDS

(POLARIZATION AND BOUND CHARGES)

10-1) Polar and nonpolar atoms and molecules 329

10-2) Dielectrics in the presence of an electric field 330

10-3) The field of a polarized dielectric 335

10-4) The field inside a dielectric, the displacement vector 338

10-5) Properties of the fields and in the case of non homogeneous

dielectrics 348
 12

10-6) Boundary conditions between perfect dielectrics, refraction of the

electric lines 350

10-7) Ferroelectric materials 357

CHAPTER 11: CAPACITORS

11-1) General concepts and definitions 361

11-2) The parallel plate capacitor 363

11-3) Spherical capacitors 371

11-4) Cylindrical capacitors 375

11-5) Capacitors connected in parallel or in series 381

11-6) Forces on dielectric materials moving in the field of a capacitor 385

CHAPTER 12: SPECIAL METHODS AND TECHNIQUES IN

ELECTROSTATICS

12-1) A general approach in solving problems in Electrostatics 397

12-2) Uniqueness theorem 400

12-3) Laplace’s equation in one dimension 402

12-4) Laplace’s equation in two dimensions, Rectangular coordinates:

Orthogonal functions, Fourier expansion, Separation of variables method,
(Potential in various rectangular regions, bounded or unbounded, etc) 407

12-5) Laplace’s equation in two dimensions, Spherical coordinates:

Legendre’s functions, (Dielectric and conducting spheres in a uniform electric
field, etc) 426

12-6) Laplace’s equation in two dimensions, Cylindrical coordinates:

Bessel’s functions, (Potential in various cylindrical regions, etc) 441

12-7) Method of images, charges and conducting planes, images with

cylinders and spheres 450
 13

PART B: MAGNETOSTATICS
CHAPTER 13: STEADY CURRENTS

13-1) General concepts and definitions 472

13-2) Electric current 473

13-3) A convention about the current direction 475

13-4) Electric current , current density and the point form of

Ohm’s law 477

13-5) The continuity equation 490

13-6) Decay of free charges 495

13-7) Joule heat loss in conducting materials 497

13-8) The concept of the resistance, resistors connected in series or in

parallel 499

13-9) A more general definition of the resistance 504

13-10) Breakdown of Ohm’s law, Vacuum diode, p – n diode 510

13-11) Voltage sources, electromotive force (emf) 514

13-12) Electric circuits and Kirchhoff’s laws 520

13-13) Steady currents and Laplace’s equation 525

CHAPTER 14: THE STATIC MAGNETIC FIELD IN FREE SPACE

14-1) Introduction 541

14-2) Biot – Savart law 543

14-3) Magnetic field produced by various current distributions,

(filamentary conductors, surface currents, volume currents) 545
 14

14-4) Calculation of the Magnetic field of some important current

distributions, (infinite filamentary conductors, infinite sheet of planar current,
current carrying circular loop, etc) 549

14-5) Magnetic field lines and magnetic flux 562

CHAPTER 15: THE MAGNETIC VECTOR POTENTIAL

15-1) Introduction 573

15-2) Magnetic vector potential of a current distribution, (Vector potential

and Magnetic field of a spinning, charged spherical surface, of a finite current
element, etc) 573

15-3) Magnetic flux through a surface, in terms of the magnetic vector

potential 584

15-4) Magnetic dipole and magnetic dipole moment 586

CHAPTER 16: THE TWO FUNDAMENTAL LAWS OF STATIC MAGNETIC

FIELDS

16-1) Introduction 595

16-2) The magnetic field is solenoidal, ( ) 596

16-3) Ampere’s circuital law, ( ) 601

16-4) The magnetic field intensity , Amperian loops, (long filamentary

current carrying conductors, cylindrical conductors, etc) 605

16-5) Rigorous derivation of Ampere’s circuital law from the Biot – Savart
law 615

CHAPTER 17: SOLENOIDS AND TOROIDS

17-1) Magnetic field of a solenoid 623

17-2) Toroids 628

 15

CHAPTER 18: MAGNETIC FORCES AND TORQUES

18-1) Magnetic force 630

18-2) Magnetic torque 643

CHAPTER 19: MOTION OF CHARGED PARTICLES IN ELECTRIC AND

MAGNETIC FIELDS

19-1) Introduction 655

19-2) Motion of a charged particle in a uniform electric field 656

19-3) Motion of a charged particle in a uniform magnetic field , Larmor’s

frequency 657

19-4) Motion of a charged particle in perpendicular, static, uniform electric

and magnetic fields 669

19-5) An important case of a non uniform magnetic field – The magnetic

bottle configuration 673

CHAPTER 20: STATIC MAGNETIC FIELDS IN MATTER

20-1) Diamagnetic, Paramagnetic and Ferromagnetic materials 682

20-2) The physical origin of paramagnetism, diamagnetism and

ferromagnetism 684

20-3) Magnetic fields produced by a magnetized object 690

20-4) The magnetic field intensity 697

20-5) Boundary conditions between two magnetic media, refraction of the

magnetic lines 705

20-6) Hysteresis loop in ferromagnetic materials 713

CHAPTER 21: PERMANENT MAGNETS – MAGNETIC CIRCUITS

21-1) Permanent magnets 717

 16

21-2) Scalar magnetic potential, (Magnetic field of a uniformly magnetized

sphere, field of a slab of permanent magnet, etc) 719

21-3) Magnetic circuits, Magnetomotive force (mmf) 726

CHAPTER 22: SELF – INDUCTANCE, MUTUAL INDUCTANCE, ENERGY

STORED IN MAGNETIC FIELDS

22-1) Self – Inductance 740

22-2) Mutual Inductance, Neumann’s formula 743

22-3) Energy stored in magnetic fields, magnetic energy density, Inductance

in terms of the magnetic energy 752

22-4) Lifting force of a magnet 760

CHAPTER 23: MAXWELL’S EQUATIONS FOR STATIC ELECTRIC AND

MAGNETIC FIELDS

23-1) Maxwell’s equations for static fields 768

23-2) Potential formulation of Electrostatics and Magnetostatics 769

23-3) Boundary conditions 770

23-4) The continuity equation 771

PART C: TIME – VARYING FIELDS AND

MAXWELL’S EQUATIONS
CHAPTER 24: FARADAY’S LAW OF INDUCTION – DISPLACEMENT
CURRENT

24-1) Introduction 774

24-2) Faraday’s law – Lenz’s law 775

24-3) Conductors moving in a uniform magnetic field, (Motional emf) 786

 17

24-4) Displacement current 793

CHAPTER 25: MAXWELL’S EQUATIONS (TIME – VARYING FIELDS)

25-1) Point and Integral form of Maxwell’s equations 805

25-2) Maxwell’s equations in free space (vacuum) 806

25-3) The wave equation for and 808

25-4) Boundary conditions for time – varying fields 813

25-5) Scalar potential and vector potential for time – varying fields, the
Lorentz condition 818

25-6) The retarded potentials 823

25-7) The Poynting vector and transmission of electromagnetic energy 826

25-8) Energy associated with a set of currents 833

APPENDIX: AXIOMATIC DERIVATION OF MAXWELL’S EQUATIONS

a) The four fundamental postulates 843

b) Derivation of Maxwell’s equations 845

c) Electrostatics, Magnetostatics and the Field of Steady Currents, as special

cases of Maxwell’s equations 850

d) Quasi-static approximation 851

e) Derivation of Coulomb’s law from Maxwell’s equation 852

 18

A SUMMARY OF IMPORTANT
FORMULAS AND VECTOR IDENTITIES
A) TRANSFORMATION BETWEEN COORDINATE SYSTEMS

1) RECTANGULAR (CARTESIAN) COORDINATES

a) Independent coordinates: , unit vectors .

b) Differential vector: .

c) Differential volume: .

2) CYLINDRICAL COORDINATES

a) Independent coordinates: , unit vectors .

b) Differential vector: .

c) Differential volume: .

d) Relation between Cylindrical and Rectangular coordinates:

e) Unit vectors in Cylindrical coordinates in terms of the unit vectors

in Rectangular coordinates:

f) Dot product of unit vectors in Cylindrical and Rectangular

coordinate systems:
 19

3) SPHERICAL COORDINATES

a) Independent coordinates: , unit vectors .

b) Differential vector: .

c) Differential volume: .

d) Relation between Spherical and Rectangular coordinates:

e) Unit vectors in Spherical coordinates in terms of the unit vectors in

Rectangular coordinates:

f) Dot product of unit vectors in Spherical and Rectangular coordinate

systems:
 20

B) TWO IMPORTANT THEOREMS

THE GAUSS-OSTROGRADSKY THEOREM (Divergence Theorem):

STOKES’ THEOREM:

C) USEFUL VECTOR IDENTITIES

 21

D) EXPRESSIONS OF THE GRADIENT, THE DIVERGENCE, THE CURL

AND THE LAPLACIAN IN RECTACULAR, CYLINDRICAL AND SPHERICAL
COORDINATES

1) In rectangular coordinates :

2) In cylindrical coordinates :
 22

3) In spherical coordinates :
 23

CHAPTER 1: MATHEMATICAL INTRODUCTION

1-1) Review of Basic Definitions and Properties of Vectors

In Engineering and Physics we consider various types of physical quantities.

One class of quantities is completely determined by just one single number
(the magnitude) and the corresponding unit of measurement. Such quantities
are called scalar quantities or simply scalars. For example, length, area,
volume, mass, density, temperature, electric charge, electric potential, energy
and work are all scalar quantities. If the area of a plane figure is , the
number (magnitude) together with the corresponding unit ( ) completely
specifies the area of the figure. If the mass of a body is , this completely
specifies the mass of the body, etc.

Another class of physical quantities consists of quantities the complete

determination of which requires both magnitude and direction. Such
quantities are called vector quantities or simply vectors. For example, velocity,
acceleration, force, torque of a force, momentum, electric and magnetic fields,
etc, are vector quantities.

A vector is represented by a directed line segment as shown in Figure 1-1.

Fig. 1-1: The vector concept.

Symbolically the three vectors shown in Fig. 1-1 are denoted as , and
, respectively. In the first vector the point is the initial point or the
origin of the vector and the point is the terminal point of the vector. The
magnitude or the length of the vector is denoted by , and similarly
and are the magnitudes (lengths) of the vectors and
respectively. A second notation for vectors is often used, which consists of
 24

single small letters beneath an arrow, such as , ,…., and in this notation,
,…., will represent the magnitudes of the corresponding vectors.

A vector of particular interest is the null or zero vector, which is defined as

a vector whose magnitude (length) is zero. The null vector is denoted by . The
null vector has no direction.

Equal and opposite vectors

Two vectors and are said to be equal, if and only if, is parallel to
with the same orientation and .

Note that two vectors which are not parallel cannot possibly be equal,
even if they have the same length (magnitude).

Two vectors and are said to be opposite if they are parallel, have the
same magnitude but opposite orientations.

Fig. 1-2: Equal and opposite vectors.

The sum of two opposite vectors is the null vector, for example in Fig. 1-2,
since and are opposite vectors,

Note: In equations involving vector quantities, it is customary to denote the

null vector by , (instead of the full notation ).
 25

A vector which is restricted to pass through a given, fixed point, is called a

bound or localized vector. In this case the line of action of the vector is fixed.
Force acting on a body is a bound vector since its effect depends on the point
of its application.

On the other hand, a vector which is not restricted to pass through a fixed
point in space is called a free vector. Free vectors have the same magnitude
and direction and act at different points in the same line or parallel lines. They
are all equivalent to each other. The moment of a couple of forces, for
example, is a free vector.

In order to be able to apply the theory of vectors in various problems arising

in Engineering and Physics, we have to develop an Algebra of Vectors, i.e. how
to add and subtract vectors, how to multiply a vector by a scalar (number),
what is meant by “ a linear combination of vectors”, etc.

Vector addition and subtraction

Let and be two vectors as shown in Fig. 1-3.

Fig. 1-3: Addition of two vectors, “The parallelogram law of addition”.

The vector which is the diagonal of the parallelogram is

defined to be the sum of the two vectors and and is written as

Alternatively, the sum of the two vectors can be obtained from the
equivalent “triangle law of addition”. In Fig. 1-3, the vectors and are
equal (since they are parallel, have the same orientation and equal
 26

magnitudes). If we therefore consider the triangle having sides

and , then the third side of the triangle is the sum of the two
vectors and .

The “triangle law of addition” can be applied to obtain easily the sum of
more than two vectors, as shown in Fig. 1-4.

Fig. 1-4: The vector addition of five vectors.

In Fig. 1-4, the initial point of coincides with the terminal point of , the
initial point of coincides with the terminal point of , the initial point of
coincides with the terminal point of , and the initial point of coincides with
the terminal point of . In this case, the vector with initial point the initial
point of the first vector and terminal point the terminal point of the last
vector is defined to be the sum of the five vectors and is written as

If the vectors are not arranged so that the terminal point of one coincides
with the initial point of the next, then we make a parallel translation of the
vectors, so that the terminal point of one coincides with the initial point of
the next, and then perform the addition. Let us, for definiteness, consider four
vectors and as shown in Fig. 1-5. We consider a point (the origin)
and form the “successive” vectors and .
Then the vector is the sum of the four vectors and , i.e.
 27

Fig. 1-5: The vector addition of four, arbitrary vectors.

Let and be any two vectors. The subtraction of from is defined to

be the addition of the vector to , i.e. . The
subtraction of two vectors is shown in Fig. 1-6.

Fig. 1-6: The subtraction of two vectors .

Multiplication of a vector by a scalar

Let be any real number (scalar) and be any vector ( ). Then the
product is a vector such that:

a) The magnitude of is ,

b) If then the vector has the same direction with , while if

then the vector has direction opposite to , (i.e. has the direction of
),
 28

c) If then the vector .

Note that we attribute no direction to the null vector.

In Fig. 1-7 we see the multiplication of a given vector by the scalars

(numbers) and , respectively.

Fig. 1-7: Multiplication of a vector by a scalar.

The multiplication of a vector by a scalar has the following properties:

The unit vector

If we divide a vector by its magnitude (positive number) then we obtain a

new vector which has the same direction and orientation with the original
vector and magnitude equal to . For this reason the thus obtained vector is
called “a unit vector”. To emphasize that a vector is a unit vector we use the
symbol . The magnitude of any unit vector is (by definition) equal to , i.e.
.
 29

Fig. 1-8: Unit vectors.

In Fig. 1-8, are the unit vectors corresponding to the vectors

and respectively.

The position vector

The position vector of a point with respect to an arbitrarily chosen origin

, is the vector . The vector uniquely specifies the position of the point
relative to the origin .

Fig. 1-9: The position vectors.

 30

In Fig. 1-9, let and be the position vectors of the points and
respectively, relative to the same origin . Then,

In words: Every vector is equal to the difference of the position vector

of its terminal point minus the position vector of its initial point .

Vector products

Two kinds of vector multiplication are defined, the dot or scalar or inner
product and the cross or outer product.

The dot product: Let and be two vectors and be the

smallest angle between the two vectors, ( ).

Fig. 1-10: The dot (inner or scalar) product of two vectors.

The inner product or dot product of the two vectors and is defined to be
the scalar quantity (real number),

where and are the magnitudes of the vectors and

respectively.

1) If and are parallel and have the same orientation ( ), then

. If and are parallel and of opposite orientation ( ),
then .
 31

2) If and are perpendicular to each other ( ), then , and

conversely, if then , meaning that .

3) If then and formula (1-1-6) implies,

4) The dot product of two vectors is commutative and distributive, i.e.

5) If and are any two vectors and and are any two real numbers,
then it is easily shown that

6) Let and be the unit vectors along the and axes

respectively, in an orthonormal Cartesian system . The magnitude of each
one of these vectors is one, (unit vectors), and also these vectors are pair-wise
perpendicular, which means that

7) The Cartesian expression of the dot product of two vectors.

Let and be the Cartesian

expression of two vectors and . Since the dot product is distributive, we
have,

and taking into account formulas (1-1-10) equation (*) finally simplifies to
the following, (let the reader check it),
 32

Formula (1-1-11) is called the Cartesian expression of the dot product of

the two vectors and . Formula (1-1-11) implies that two vectors
and are perpendicular, if and only if

8) The length of a vector in a Cartesian system of axes.

If formula (1-1-11) implies

9) Angle between two vectors expressed by their Cartesian coordinates,

( ).

Since in general, , (where ),

Equation (1-1-14) gives the angle between two vectors expressed by their
Cartesian coordinates,
 33

10) The direction cosines of the position vector .

Fig. 1-11: Direction angles and direction cosines.

Let and be the unit vectors along the axes and

respectively, as shown in Fig. 1-11.The angles between the vector and the
unit vectors are called direction angles, while the cosines of the direction
angles are called direction cosines. Let for definiteness be the angle
between and , the angle between and and the angle between
and . From the fundamental definition of the dot product,

and similarly for the other two direction cosines. In summary, the direction
cosines of the vector , are given by the formulas:

It is evident, (from (1-1-15) that the sum of the squares of the direction
cosines is equal to , i.e.
 34

The cross product: Let and be two vectors and be the

smallest angle between the two vectors, ( ).

Fig. 1-12: The cross (outer) product of two vectors.

The outer or cross product of two vectors and , designated and

read cross , is a third vector such that:

a) Is perpendicular to the plane , determined by the two vectors and

, i.e. is perpendicular to both and .

b) Its direction is such that the system (in this order) is right-
handed, i.e. the vector points in the direction a right-handed screw advances
when its head is rotated from to through the angle .

c) Its magnitude is

where and are the magnitudes of the vectors and

respectively.
 35

1) The cross product of two collinear (parallel) vectors ( or ) is

zero, since in both cases . In particular, the cross product of any
vector by itself is zero, i.e. . And conversely, if the cross product of
two non zero vectors and is zero, then the two vectors are parallel, (since
in this case ).

2) The magnitude is equal to the geometric area of the

parallelogram formed by the two vectors and , (see
Fig. 1-12). Indeed, if is the perpendicular distance between the two parallel
lines and , then , and the area of the parallelogram is

3) The cross product is anti commutative, i.e.

Indeed, the vectors and have the same magnitude but opposite
directions. This implies that when working with cross products of vectors, the
order of the vectors is important.

4) If is any real number (not zero, ), then

5) The cross product is distributive with respect to the addition, i.e.

Let us for example consider the cross product . We

may expand this cross product with the aid of (1-1-19) and (1-1-20), i.e.

6) Let and be the unit vectors along the and axes

respectively, of an orthonormal right-handed Cartesian system . The
magnitude of each one of these vectors is one, (unit vectors), and also these
vectors are pair wise perpendicular.
 36

Fig. 1-13: Cross products of the unit vectors.

The following vector identities hold true, (let the reader check it):

Since the cross product is anti commutative, , etc.

7) The Cartesian expression of the cross product.

Let us consider the two vectors and

. The cross product of these two vectors is

which by virtue of (1-1-19) and (1-1-20) yields,

(let the reader verify the calculations). Formula (**) can be expressed,
equivalently, in a convenient form, using determinants notation, i.e.
 37

(The reader is supposed to be familiar with the elementary properties of

determinants).

Scalar and vector fields

Let us assume that at each point (determined by its position vector

relative to a given coordinate system ) of a domain
in space, a certain value of a scalar quantity is specified. In this case we say
that there is a scalar point function defined in the domain and we write
or simply .

The domain over which the scalar function is defined may be the whole
space or be a part of it.

The scalar point function is called a scalar field defined over

the domain .

As an example, we may speak about the temperature field at the points of a

heated body or about the potential field around an electrified object, etc.

Similarly, if at each point of a domain a certain value of a vector

quantity is specified, then this vector point function is
called a vector field defined over the domain .

Examples of vector fields defined in space are the electric and the magnetic
field intensities ( and respectively), the field of gravitational forces around
a mass , etc.

In very general terms, a field (scalar or vector) is the distribution of a scalar

or a vector quantity respectively, in space and time. For brevity we may write,
 38

The notation means the scalar quantity at the point

at the time instant , and the same applies for the vector field
.

In case a field (scalar or vector) does not depend on the time , the field is
called static or stationary field. In this case the field depends only on the
spatial coordinates and we may just write or , (i.e.
the spatial distribution of the fields remains time invariant). In the opposite
case, the field is called “time varying”. In the special case where the time
dependence of a field is a sinusoidal function of time, the field is called “time
harmonic field”. For example the fields

are time harmonic fields.

A level surface of a scalar field is the set of points in space at which the
field assumes a given constant value , that is

A level surface of a scalar field passing through a given point

has equation .

The surfaces of constant electric potential, produced by a static charge

distribution are referred to as equipotential surfaces.

A special type of scalar fields are the so called “plane fields” that is fields
which depend on two variables (coordinates) only, for instance the and the
coordinates. In such cases we speak of “level lines” of the scalar field, which
are lines over which the scalar assumes constant values .

Regarding vector fields, a vector line (or field line) of a vector field is a
curve in space at every point of which, the direction of the tangent of the
curve coincides with the direction of the vector of the field. Vector lines are
used to help us visualize vector fields. As an example, the vector lines of the
velocity field of a fluid flow (known as “stream lines”) may be interpreted as
the trajectories of motion of the fluid particles, while the “field lines” of a
 39

static electric field are the lines traversed by a positive, point charge, placed
inside the electric field region and left to move under the action of electric
forces only.

Fig. 1-14: Vector or field lines of a vector field.

In order to find the equation of the field lines, we note that the field line of
a vector field defined in space, and passing through a given point , is
collinear with the differential vector directed along the tangent to the field
line through , and therefore,

Equation (1-1-25) is the sought for differential equation for the vector lines
of a vector field . Let us further assume that the expression of the vector field
in a rectangular system of coordinates is

Then, since , equation (1-1-25) implies

Strictly speaking, (1-1-26) is a system of differential equations, the solution

of which (integration) yields the sought for vector lines of the field .
(For a proof of equation (1-1-26) see Example 1-1-5).

Example 1-1-1: Find the dot and the cross product of the vectors
and . What is the angle between the vectors
and ?
 40

Solution

Example 1-1-2: For what value of the vectors and

are perpendicular?

Solution

The vectors and will be perpendicular if and only if their dot product
vanishes, i.e.

Example 1-1-3: If find the magnitude and the unit

vector in the direction of .

Solution
 41

Example 1-1-4: If and , find the unit vector

perpendicular to both and .

Solution

The vector is perpendicular to the plane determined by and , i.e. is

perpendicular to both vectors, and the sought for unit vector is

Example 1-1-5: Starting with equation (1-1-25) show equation (1-1-26).

Solution
 42

PROBLEMS

1-1-1) Find the direction cosines of .

(Ans: ( )).

1-1-2) If , evaluate: ,
.

1-1-3) Find the angles between the vectors and

, (Ans: ).

1-1-4) If show that .

1-1-5) Find the area of the triangle formed by the vectors

and , (Ans: ).

1-1-6) If is the centroid of a triangle , show that .

1-1-7) The position vectors of three points in space are

respectively. Show that the triangle is a
right triangle.

1-1-8) If and , show that , (provided that

).

1-1-9) Three forces , with magnitudes and act on

a particle in the directions of and
respectively. If the particle is displaced from the point to the point
, find the work done by the forces, (distances are expressed in ).

(Ans: ).

Hint: Work .

1-1-10) Find the field lines of the field , (Ans: ).

Hint: The differential equation of the field lines is .

1-1-11) Find the level surfaces of the scalar field .

 43

(Ans: , concentric spheres, centered at the origin).

1-1-12) For what value of the vectors and

are parallel? (It suffices ).

1-2) Coordinate Systems, (Cartesian, Cylindrical, Spherical)

Even though vectors and vector operations are defined independently of

any coordinate system, it is very useful and convenient to use coordinate
systems while working in solving various practical problems.

A coordinate system serves mainly, two purposes:

a) To locate points in space and

b) To provide analytic expressions for scalar and vector fields defined in

space.

In particular, when we consider vector functions defined in space (vector

fields), the representation of the vector function is accomplished in terms of
three unit vectors which are mutually orthogonal and form a right-handed
system.

The most familiar coordinate system is the Rectangular Coordinate System.

The characteristic property of a Rectangular system is that the base unit
vectors are constant, i.e. they do not vary as we move from one point to
another, in space. This property is not shared by other coordinate systems.

In this chapter we will study in details the three most commonly used
coordinate systems, the rectangular, the cylindrical and the spherical
coordinate system. The choice of a particular system depends upon the
symmetry of the problem at hand. In problems with cylindrical symmetry we
prefer the cylindrical coordinate system, while in problems with spherical
symmetry we choose the spherical coordinate system. By doing so, we greatly
simplify the mathematics involved.

Rectangular Coordinate Systems

The independent variables in a rectangular system are and , as shown

in Fig. 1-15. Each variable varies from to , and each point in space is
 44

uniquely determined by an ordered triad of numbers. If is a given (fixed)

real number, the totality of points in space, such that

define a plane perpendicular to the axis, at . The plane is

a coordinate surface. Similarly we define the coordinate surface (plane)
as the totality of points in space, such that

and the coordinate surface (plane) as the set of points in space, such
that

We note that is a plane perpendicular to the axis at and

is a plane perpendicular to the axis at .

Fig. 1-15: Rectangular coordinate system.

We may consider the point in space, as the intersection of the

three coordinate surfaces (planes), and .

The unit vector at the arbitrary point is defined as a unit

vector ( ), perpendicular to the coordinate surface and pointing
in the direction of increasing . Similar definitions are made for and . We
note that the unit vectors, thus defined, are constant for all points in space. As
 45

we have already mentioned, this is a characteristic property of a Rectangular

system. In the Cylindrical and the Spherical systems, this is no longer valid. In
these systems the unit vector change direction while moving from one point
to another, (but their magnitude is always equal to ).

The system of unit vectors , in this order, forms a right-handed

orthogonal system, meaning that .

Cylindrical Coordinate Systems

The cylindrical coordinate system is used in problems with cylindrical

symmetry. When analyzing problems with cylindrical structures, for example
the flow of a fluid in a round pipe or the propagation of an Electromagnetic
wave in cylindrical waveguides, the mathematics is simpler if we use cylindrical
coordinates.

The independent variables in a cylindrical coordinate system are and

as shown in Fig. 1-16.

Fig. 1-16: Cylindrical coordinate system.

If are fixed
numbers, within their range of variation, then the coordinate surface
is a cylindrical surface of radius concentric with the axis, the coordinate
surface is a plane passing through the axis and forming angle
with the plane and the coordinate surface is a plane perpendicular
to the axis at . The intersection of these coordinate surfaces defines
 46

the point . As an example, the point is the

intersection of the three coordinate surfaces and .

The unit vectors in cylindrical coordinates, and , at the arbitrary

point are defined in a manner analogous to the definition of the
unit vectors in rectangular coordinates. The unit vector at is
perpendicular to the coordinate surface , pointing in the direction of
increasing coordinate. The unit vector at is perpendicular to
the coordinate surface and points in the direction of increasing
coordinate, while the unit vector at is perpendicular to the
coordinate surface and points in the direction of increasing
coordinate. Evidently the unit vectors and change their direction as we
move from one point to another, in space, while their magnitude remains
always constant and equal to . The vector remains constant and coincides
with the unit vector in rectangular coordinates.

The unit vectors , in this order, form an orthogonal right-handed

system, meaning that .

Fig.1-17: Unit vectors in cylindrical coordinates.

Spherical Coordinate Systems

The spherical coordinate system is used in problems with spherical

symmetry. When analyzing problems with spherical structures, for example
 47

when we need to determine the electrostatic potential produced by a spherical

charge distribution or to find the gravitational attraction between two
spherical masses, the mathematics is simpler in spherical coordinates.

The independent variables in spherical coordinates are and as shown

in Fig.1-18.

Fig. 1-18: Spherical coordinate system.

If are given, fixed

numbers, within their range of variation, then the coordinate surface is
the surface of a sphere of radius centered at the origin, the coordinate
surface is a cone with vertex at the origin , axis , and generating
angle and the coordinate surface is a plane passing through the
axis and forming angle with the plane.

In the spherical system of coordinates the positive axis corresponds to

while the negative axis corresponds to . The plane
corresponds to , while for we obtain cones with generating angles
greater than .

The intersection of the three coordinate surfaces defines the point

, for instance, the point is the intersection of the
coordinate surfaces and .
 48

The unit vectors in spherical coordinates, and , at the arbitrary point

are defined in a manner similar to the definition of the unit
vectors in rectangular and cylindrical coordinates. The unit vector at
is perpendicular to the coordinate surface , pointing in the
direction of increasing coordinate. The unit vector at is
perpendicular to the coordinate surface and points in the direction of
increasing coordinate, while the unit vector at is
perpendicular to the coordinate surface and points in the direction of
increasing coordinate. Evidently the unit vectors and , change their
direction as we move from one point to another, in space.

The unit vectors , in this order, form an orthogonal right-handed

system, meaning that .

Fig. 1-19: Unit vectors in spherical coordinates.

In summary:

In order to locate points in space and to provide analytic expressions for

scalar and vector functions defined in space, we have defined the three, most
commonly used coordinate systems; the Rectangular, the Cylindrical and the
Spherical coordinate systems. For each one of these systems we define the
independent variables and the unit vectors.

In the Rectangular system the unit vectors are constant vectors,

meaning that they do not change as we move from one point to another.
 49

These unit vectors are translated parallel to themselves as we move from point
to point in space.

In the other two systems, the Cylindrical and the Spherical, the unit vectors
and , respectively, change from point to point, i.e. these
vectors are functions of the position. Of course, as being unit vectors, their
magnitude is constant and always equal to , i.e. and
, but their direction changes from point to point.

In general, cylindrical coordinates are used when analyzing problems with

cylindrical structures while spherical coordinates are used in problems
involving spherical structures.

Example 1-2-1: In cylindrical coordinates plot the coordinate surfaces

.

Solution
 50

Fig. 1-20: Coordinate surfaces in cylindrical coordinates.

 51

Example 1-2-2: In spherical coordinates plot the coordinate surfaces

.

Solution

Fig. 1-21: Coordinate surfaces in spherical coordinates.

 52

PROBLEMS

1-2-1) In rectangular coordinates plot the coordinate surfaces

.

1-2-2) In cylindrical coordinates plot the coordinate surfaces

.

1-2-3) In spherical coordinates plot the coordinate surfaces

.

1-3) Transformations Between Coordinate Systems

The location of a given point in space can be expressed either by its

rectangular coordinates or by its cylindrical coordinates or even more by its
spherical coordinates. However, all these expressions refer to the same point
. Similar considerations are obviously true for a scalar or a vector function of
the spatial coordinates. It will be therefore convenient to obtain the
representation of the coordinates of a point or a scalar or a vector function in
one coordinate system from its representation in another coordinate system.

Relation between Rectangular and Cylindrical Coordinates

The relations between rectangular and cylindrical coordinates are easily

obtained with the aid of Fig. 1-16 and simple trigonometry.

By means of formula (1-3-1), given the rectangular coordinates of a

point , we may find the cylindrical coordinates of the same point
and vice versa.

Relation between Rectangular and Spherical Coordinates

Similarly the relations between rectangular and spherical coordinates are

obtained with the aid of Fig. 1-18.
 53

By means of formula (1-3-2) we may find the spherical coordinates

of a point from its rectangular coordinates and vice versa.

It will be useful in the subsequent development of the theory to obtain

expressions of the unit vectors and in terms of the constant
unit vectors .

The Unit Vectors in terms of

The unit vectors in cylindrical coordinates are expressed in terms

of the constant unit vectors in rectangular coordinates as follows:

Equation (1-3-3) can be written in the equivalent matrix form:

Let the point and be the unit vectors at the point . If

we call , the vectors and are collinear, since they both lie on the
plane and are perpendicular to the axis, i.e. , and therefore

However,

From (*) and (**) it follows easily that

 54

Also, since is a right-handed orthogonal system, it follows that

and this completes the proof.

Fig. 1-22: The unit vectors in terms of .

The Unit Vectors in terms of

The unit vectors in spherical coordinates are expressed in terms of

the constant unit vectors in rectangular coordinates as follows:

or in an equivalent matrix form,

 55

In Fig. 1-23, note that the unit vector is perpendicular to the plane
formed by the vectors and , while the unit vector is perpendicular to
the vector . Using simple trigonometry we have:

and this proves the first equation in (1-3-5).

Fig. 1-23: The unit vectors in terms of .

The unit vector lies in the plane formed by the vectors and and is
perpendicular to . From Fig. 1-23, we have:
 56

and since , (recall that is a unit vector) and

(equation (1-3-3), we finally obtain

and this proves the second equation in (1-3-5).

Finally, since it follows readily that and

this completes the proof, (for a detailed proof see Pr. 1-3-7).

Example 1-3-1: Give the rectangular coordinates of the point

.

Solution

The point is determined by its cylindrical coordinate and its

rectangular coordinates are found from equations (1-3-1), i.e.

Example 1-3-2: Give the rectangular coordinates of the point

.

Solution

The point is determined by its spherical coordinates and its

rectangular coordinates are found from equations (1-3-2), i.e.

Example 1-3-3: Give the cylindrical coordinates of the point

.

Solution
 57

The point is expressed in rectangular coordinates. To find its cylindrical

coordinates we apply formula (1-3-1), i.e.

Note: Caution should be given when inverting the .

If we just blindly apply the formula to obtain , the

answer supplied by the calculator will be , since the calculator
gives the “principal value” of the Inverse Tangent function, which takes values
in the open interval .

Fig. 1-24: Proper evaluation of the .

However, in Fig. 1-24 it is shown that the sought for angle

.

Similar considerations apply for the “principal values” of the Inverse Sine
and the Inverse Cosine functions. The principal value of the ,
(supplied by the calculator) lies in the closed interval , while the
 58

principal value of the , (supplied by the calculator) lies in the closed

interval .

Example 1-3-4: Give the spherical coordinates of the point

.

Solution

The point is given in Cartesian coordinates. The spherical coordinates of

the same point are obtained from equation (1-3-2), i.e.

Example 1-3-5: Determine the unit vectors at the point

in terms of the unit vectors .

Solution

Application of formula (1-3-3) yields,

 59

Example 1-3-6: Determine the unit vectors at the point

in terms of the unit vectors .

Solution

Application of formula (1-3-5) yields,

Example 1-3-7: Let the points and be determined by their cylindrical

coordinates, and respectively. Express the vector in
terms of the rectangular unit vectors .

Solution

The Cartesian coordinates of the point are (formula (1-3-1):

while the Cartesian coordinates of are:

 60

The vector in rectangular coordinates is expressed as

Example 1-3-8: Give the vector in rectangular coordinates that extends

from to and then
determine the length of the vector .

Solution

The rectangular coordinates of the point are (formula (1-3-2):

Similarly the rectangular coordinates of are:

The vector in rectangular coordinates is

The length of the vector is

PROBLEMS

1-3-1) Given that , find and .

(Ans: ).
 61

1-3-2) Given that , find and .

1-3-3) Find and , (see Fig. 1-24).

(Ans: ).

1-3-4) The rectangular coordinates of a point are .

Find the cylindrical and the spherical coordinates of .

Hint: Use formulas (1-3-1) and (1-3-2).

1-3-5) Given that the spherical coordinates of a point are

find the rectangular and the cylindrical coordinates of .

(Ans: ).

1-3-6) In Problem 1-3-5 find the distance of the point from the origin.

1-3-7) Using the expressions found for and (in terms of ), find the
expression for in terms of the same unit vectors , starting with
.

1-3-8) Using formulas (1-3-3) and (1-3-5) verify by direct calculations that:

a) and .

b) and .

This verifies that the unit vectors in the cylindrical and the spherical
coordinates are mutually perpendicular.

1-3-9) The points and are determined by their spherical coordinates,

and . Find the distance .

(Ans: ).

1-3-10) Points and

are located on the surface of a radius sphere. a) What is their
 62

separation, using a path on the spherical surface? b) What is their separation,

using a straight-line path?

1-3-11) Give the cylindrical coordinates of the point .

(Ans: ).

1-3-12) In the previous problem, find the spherical coordinates of point .

(Ans: ).

1-4) Transformation of a Field, (Scalar or Vector), from one Coordinate

System to Another

Depending upon the particular coordinate system being used, a field (scalar
or vector) can be expressed either in rectangular or in cylindrical or even in
spherical coordinates. For example, a vector field in rectangular
coordinates can be can be expressed as

The same field, in cylindrical coordinates, can be written as

and in spherical coordinates as

Suppose now that we wish to transform a scalar static field in

cylindrical coordinates, i.e. to obtain an expression . All we have to
do is to use equations (1-3-1) which relate rectangular and cylindrical
coordinates. Similarly if we wish to express the scalar field in spherical
coordinates we have to apply equations (1-3-2) which relate rectangular and
spherical coordinates.

The transformation of a vector field from, say, rectangular to cylindrical

coordinates, is more complicated. Let us, for definiteness, consider the vector
static field
 63

and suppose that we wish to represent this field in a cylindrical coordinates

system as

There are two steps in the process of transforming:

a) must be obtained in terms of and

and

b) The and must be expressed in terms of and , so that

eventually the field components and will be functions of and
rather than and .

As we shall see in the examples to follow, (see Example 1-4-2), in order to

complete the transformation, we need to know the dot product of the unit
vectors associated with the various coordinate systems, for example we need
to know or or , etc. The following two tables shall be used
extensively when transforming a vector field from one coordinate system to
another.

Dot products of unit vectors in cylindrical and rectangular coordinates

Dot products of unit vectors in spherical and rectangular coordinates

Since two matrices are equal when all the corresponding entries are equal,
equation (1-4-3) implies that , , etc. while equation
(1-4-4) implies that , , etc.
 64

The proof of (1-4-3) and (1-4-4) is not difficult. For example, in order to
prove (1-4-3) we consider equation (1-3-3) which expresses the unit vectors
and in terms of the unit vectors and , i.e.

Dot multiplying both sides of the first equation in (*) by , we obtain

since ( is a unit vector) while ( and are orthogonal to

each other). Similarly, dot multiplying both sides of the second equation in (*)
by , we obtain

In a similar fashion the proof of (1-4-3) is completed.

To prove (1-4-4) we consider equation (1-3-5) which expresses the unit

vectors and in terms of the unit vectors and and dot multiply each
one of the three equations by the unit vectors and , (for detailed
calculations see Problem 1-4-2).

Example 1-4-1: Express the scalar field in

cylindrical and in spherical coordinates.

Solution

a) Cylindrical coordinates:

The rectangular coordinates relate to the cylindrical coordinates by means

of equations (1-3-1), i.e.

and therefore
 65

b) Spherical coordinates:

The rectangular coordinates relate to the spherical coordinates by means of

equations (1-3-2), i.e.

and therefore,

We note that the given field assumes its simplest form in spherical
coordinates.

Example 1-4-2: Transform from rectangular

coordinates to spherical coordinates.

Solution

We want to express the given vector field in the form

Dot multiplying both sides of (*) by , and noting that ,

(since the unit vectors are mutually orthogonal) and that , we
obtain,

The dot products are obtained from (1-4-4) while the are expressed
in terms of by means of equations (1-3-2), and equation (**) implies,

, or even more,

Similarly, the component is found if we dot multiply both sides of (*) by

, i.e.
 66

and using equation (1-4-4) for the dot products, we find,

Finally, the component is found in a similar manner, i.e.

From (*), (***), (****) and (*****) we find the expression of the given
vector field in spherical coordinates,

Again, we note that the given vector field assumes its simplest
form when expressed in spherical coordinates.

Example 1-4-3: A vector field is given as

Specify the locus of all points at which, a) b) , c) .

Solution

a) The component of the field is .

 67

b) The component of the field is .

c) The component of the field is .

Note: Equations (*) represent planes, equation (**) is the equation of a

parabolic cylinder while equation (3) is the equation of a circular cylinder.

Example 1-4-4: Let the points and be determined by their cylindrical

coordinates and . Express the
vector in cylindrical coordinates.

Solution

In Example 1-3-7 we found that the rectangular expression of the vector

is . In cylindrical coordinates,

Following the procedure developed in Example 1-4-2, we have,

Similarly, we find,
 68

From (*), (**), (***) and (****) we get,

and at the point ,

Example 1-4-5: Find the field lines of the plane vector field
.

Solution

The vector or field lines of the vector field are given by (1-1-26), i.e.

Fig. 1-25: Field lines of the vector field .

Example 1-4-6: Find the level surfaces of the scalar field

.

Solution
 69

The level surfaces of the given scalar field are

We note that the level surfaces are concentric spheres of radius

centered at the origin.

PROBLEMS

1-4-1) Express the vector field in cylindrical coordinates.

(Ans: ).

Hint: See Example 1-4-2.

1-4-2) Show equations (1-4-4).

1-4-3) Find the field lines of the field .

(Ans: ).

1-4-4) Two points and are determined by their cylindrical coordinates

and .

Express the vector : a) In rectangular coordinates and b) In cylindrical

coordinates.

Hint: See Examples 1-3-7 and 1-4-4.

1-4-5) Find the level lines of the plane field .

(Ans: . When , we get a pair of straight lines, ,

while for we obtain a family of hyperbolas).

1-4-6) Find the level surfaces of the field .

1-4-7) Find the level surfaces of the field .

(Ans: , a family of ellipsoids).

 70

1-4-8) Find the level surfaces of the field , where is constant

vector and is the radius vector of a point in space.

1-4-9) Find the vector line of the field which passes

through the point .

(Ans: The parametric equation of the sought for vector line is

where is a parameter).

1-4-10) Find the vector lines of the field

.

1-4-11) Express the field in cylindrical

coordinates (variables and components). Also, find at the point
.

(Ans:
).

1-4-12) Express the vector field : a) In cylindrical

coordinates and b) In spherical coordinates.

Hint: See Example 1-4-2.

1-4-13) Find in spherical coordinates at the point

. Also find in rectangular coordinates at the same point .

(Ans: ).

1-4-14) Two points and are determined by their spherical coordinates,

and . Express the vector : a) In rectangular
coordinates, b) In cylindrical coordinates and c) In spherical coordinates.

Hint: See Examples 1-4-2 and 1-4-4.

1-4-15) In Problem 1-4-14 find the distance .

(Ans: ).
 71

1-5) Differential Arc Length and Differential Volume in Rectangular,

Cylindrical and Spherical Coordinates

Rectangular Coordinates

Fig. 1-26: Differential box in rectangular coordinates.

Let us consider the “differential box” in Fig. 1-26, having sides and
. The differential length (the diagonal of the differential box) is

while the differential volume is

Cylindrical Coordinates

Fig. 1-27: Differential box in cylindrical coordinates.

 72

The sides of the cylindrical differential box are and . The

differential arc length and the differential volume are given respectively
by the formulas:

Spherical Coordinates

Fig. 1-28: Differential box in spherical coordinates.

The sides of the spherical differential box are and . The

differential arc length and the differential volume are given respectively
by the formulas:

In the plane , let us consider the polar coordinates which are

related to the Cartesian coordinates by means of the formulas,

Any point in the plane can be determined either by its rectangular

coordinates or by its polar coordinates , as shown in Fig. 1-29.
 73

Fig. 1-29: Polar coordinates.

In polar coordinates the differential area is

Example 1-5-1: Show that the differential area in polar coordinates is

.

Solution

The sides of the “differential square” in polar coordinates are and ,

and therefore .

Fig. 1-30: Differential area in polar coordinates.

Example 1-5-2: Find the area of a circle of radius .

Solution
 74

PROBLEMS

1-5-1) Use polar coordinates to show that the circumference of a circle of

radius is .

1-5-2) Show that the surface element on the surface of a sphere of radius
is and then integrate w.r.t. the variables and to show
that the surface area of the sphere is .

1-6) Differentiation and Integration of Vector Fields

a) Derivatives of vector fields: Let be a vector function of a scalar

argument . We assume that takes values within an interval , which
could be finite or infinite. The derivative of with respect to the scalar
argument is analogous to the usual definition of the derivative of a real
function , i.e.

The derivative exists provided that the limit of the “differential

quotient”

exists and is independent of the manner in which the increment

approaches zero. In this case we say that the vector function is
differentiable.

If it is obvious from (1-6-1) that

This is in fact due to the fact that the unit vectors and , in a
rectangular coordinate system are constant vectors.
 75

Given a vector function its derivative is another vector function, so

we can differentiate it once more with respect to , to get the second
derivative of with respect to , i.e.

Again, assuming that , the second

derivative will be,

In a similar manner higher order derivatives of may be obtained.

Rules for vector differentiation

Let be a scalar function of , and and be two

vector functions of the same scalar argument . Then:

The rules of differentiation are obtained easily by straightforward

application of the definition.

b) Derivatives of the unit vectors in cylindrical and spherical systems with

respect to their corresponding coordinates: The rectangular system is the only
coordinate system where the unit vectors are constant, i.e. independent from
the position. This is actually the reason why the rectangular system is the most
 76

frequently used in practice. The derivatives of the unit vectors and with
respect to the rectangular variables and , are therefore equal to zero.

In the other two coordinate systems, this is no longer valid. The unit
vectors do vary with the position, (the unit vectors are functions of the
position) and as such, their partial derivatives with respect to the spatial
coordinates are not zero. The partial derivatives of the unit vectors in the
cylindrical and the spherical systems are given below:

Cylindrical coordinates ( and unit vectors

Equations (1-6-7) are obtained easily from (1-3-3) which gives an expression
of the unit cylindrical vectors in terms of the constant unit vectors
.

Spherical coordinates and unit vectors

Equations (1-6-8) are derived from (1-3-5) where the unit vectors in
spherical coordinates , are expressed in terms of the constant unit
vectors .

We point out, once more, that the unit vectors in the cylindrical and the
spherical coordinates are not constant, on the contrary they depend on the
position. When differentiating or integrating vector functions expressed in
 77

cylindrical or in spherical coordinates, we must be very cautious as how to

treat the differentiation or the integration of the unit vectors.

c) Integration of vector fields: A vector function is called a primitive of

the vector function , defined in the interval , if is
differentiable and

The collection of all primitives of is called “the indefinite integral or

the antiderivative” of the vector function , and we write

where is an arbitrary constant vector.

The following properties for the indefinite integrals are proved easily:

In the third formula in (1-6-11) and are arbitrary numerical constants.

Also, assuming that is a constant vector, then:

Finally, if , then

Equation (1-6-13) shows that the integration of a vector function of a scalar

argument reduces to three ordinary integrations, one for each component.
 78

The definite integral of over the integral is defined as

where is an antiderivative of .

If , then,

Example 1-6-1: Given that find: a) , b) ,

c) , d) .

Solution

Example 1-6-2: Given that ,

find and .

Solution

Example 1-6-3: If , find .

 79

Solution

Example 1-6-4: Show that a vector field has a constant magnitude if

and only if .

Solution

Assume that (constant). Then , i.e.

PROBLEMS

1-6-1) Evaluate the integral .

(Ans: ).

1-6-2) In spherical coordinates find: a) and b)

.

Hint: Use equations (1-6-8).

1-6-3) Solve the differential equation .

(Ans: , are arbitrary constant vectors.

1-6-4) Solve the differential equation .

1-6-5) If ,
where are functions of , show that
 80

1-6-6) If and find

and . Work the problem with two different methods, a)
Find the dot and cross products of and and then evaluate the
derivatives, and b) use formulas (4) and (5) in equation (1-6-6).

1-7) Multiple Integrals (Double, Triple, Line and Surface Integrals)

1) Double Integrals: Let be a function of two variables and

defined over a plane domain in the plane. Then the double integral of
over , denoted by

is the limiting value of the sum as the “elementary

surface areas” approach zero. In equation (1-7-1), is the
differential surface area.

In case the function is expressed in polar coordinates, i.e. if

then the differential area in polar coordinates is , (see equation
(1-5-8)), and the double integral becomes,

Example 1-7-1: Evaluate the double integral over the

domain , bounded by the curves and .

Solution

The domain of integration is shown in Fig. 1-31. Let us call

.

We note that the variable varies between while the variable varies
from to .
 81

Fig. 1-31: Domain of integration.

and performing the integrations w.r.t. ,

Example 1-7-2: Evaluate the double integral .

Solution

The integral is evaluated easier if we pass to polar coordinates. The domain

of integration is the interior of the circle . In polar coordinates the
 82

domain of integration is described by ,

and .

2) Triple Integrals: Let be a function of three variables

and defined over some volume in space. Then the triple integral of
over , denoted by

is the limiting value of the sum as the “elementary

volumes” approach zero. In equation (1-7-3) is the
differential volume.

In case the function is expressed in cylindrical coordinates , the

differential volume is and (1-7-3) becomes

If the function is expressed in spherical coordinates , the differential

volume is and (1-7-3) becomes

Example 1-7-3: Evaluate the integral where

is the domain bounded by the planes and the plane
.
 83

Solution

The integrand is , while the domain is a pyramid

with vertices and , as shown in Fig. 1-32.

We have:

Fig. 1-32: Domain of integration.

In Fig. 1-32, the surface , (i.e. is the triangle ), the surface

while and , (note that the
equation of is ).

We first evaluate the inner integral in (*):

Having computed the inner integral in (*) the integration w.r.t. yields:
 84

and finally, by virtue of (*), (**) and (***) we have,

and this is the sought for value of the triple integral.

Example 1-7-4: Find the volume of the sphere .

Solution

In spherical coordinates the equation of the spherical surface takes the

simple form . The volume of the sphere is found by integrating the
differential volume , i.e.

The limits of integration, in spherical coordinates, are,

, and formula (*) becomes,

3) Line Integrals: There are two types of line integrals, the line integral of
the first type and the line integral of the second type.

a) Line integral of the first type: Let be a scalar function (scalar

field) defined for all points lying on a space curve . If is the
differential length of the curve, then the line integral of the first type of
over is
 85

In practice, to evaluate the line integral in (1-7-6) we must have a

parametric representation of the curve, i.e. all the points of the curve should
be expressed in terms of a parameter , ; in this
case the differential length is

and the integral (1-7-6) reduces to an ordinary integration w.r.t. the

variable .

b) Line integral of the second type: Let

be a vector field defined in space and be an
oriented curve, as shown in Fig. 1-33. Then the line integral of the second type
of the vector field over is a scalar quantity, defined as follows:

Fig. 1-33: Line integral of the second type.

 86

Notice that if we traverse the curve in the opposite direction, the line
integral changes sign, (since each picks a negative sign), i.e.

The work done by a force while moving its point of application from to
over the curve is, (as we know from Physics), .

c) The circulation of a vector field over a closed curve : Assuming that

the curve is closed, then the line integral of over is symbolized as
and is called the circulation of around , i.e.

Example 1-7-5: Find the work done by the force field

when a material point is moving from to along the helical
curve where .

Solution

4) Surface Integrals: There are two types of surface integrals, the surface
integrals of the first type and the surface integrals of the second type.

a) Surface integrals of the first type: Let be a scalar function

(scalar field) defined for all points lying on a surface . If is the
differential area of the surface, then the surface integral of the first type of
over is
 87

Methods to evaluate surface integrals of the first type are developed in

courses in multiple integrals and vector analysis. These types of integrals
appear quite often in various problems in Physics and Engineering. Suppose,
for example, that the surface charge density ( ) is a known function of
the points on a surface . Then the total amount of charge residing
on the surface will be .

b) Surface integrals of the second type: Let be a vector field

defined in space and be an oriented surface. If is the unit normal to the
surface at the arbitrary point of the surface, then the surface integral
of the second type of over is a scalar quantity defined as

Fig. 1-34: Surface integral of the second type.

In case the surface is a closed surface, the notation we use is

 88

c) The flux of a vector field over a surface : The surface integral of the
second type of a vector field over a surface is called the flux of through
the surface , i.e.

d) An important integral: If is the position vector of the points of a closed

surface relative to an origin , then:

Example 1-7-6: Evaluate the surface integral

over the surface of the unit sphere in the first octant.

Solution

On the surface of the unit sphere ( ) the normal unit vector

coincides with the unit vector , i.e. . Then,

The evaluation of the integral is simplified if we pass to spherical

coordinates . On the unit sphere ( ) in the first quadrant, we
have:

Also from equation (1-4-4) we get,

 89

By virtue of (**) and (***) the integrand in (*) becomes,

and the integral in (*) is

Note: To evaluate the first integral make the substitution .

PROBLEMS

1-7-1) Find the double integral , where is the domain

, (Ans: ).

1-7-2) Find the double integral over the

rectangle , (Ans: ).

1-7-3) Evaluate the triple integral over the

parallelepiped defined by .

(Ans: ).

1-7-4) Show that .

1-7-5) The linear charge density along the axis is .

Find the electric charge between the points a) and , and b)
and , (Ans: a) , b) ).
 90

1-7-6) Find the integral along the curve

from to if the vector field is
.

1-7-7) A material point traverses the unit circle in the positive direction
(counterclockwise) within a force field . Find the
work done by the force during one complete revolution, (Ans: ).

1-7-8) Evaluate the integral , a) over the

straight line segment joining the points and , and b) along the
parabola between the same points and .

(Ans: a) , b) ).

1-7-9) Find the flux of the vector field through the

surface of the cylinder , (Ans: ).

1-7-10) Find the flux of the vector field through a sphere of radius
centered at the origin, (Ans: ).

1-8) Gradient of Scalar Fields, Divergence and Curl of Vector Fields, the Del
Operator , the Laplacian of Scalar Fields

a) Let be a scalar field, defined in some domain in space. If

we may write (for brevity), . We assume that the
function is differentiable.

The gradient of the scalar field at a given point is a vector denoted as

and defined by the equation

For example, if , then the gradient of is:

 91

Let now be a vector

field defined in a domain in space. We assume that the functions are
continuous and have continuous partial derivatives at all points within the
domain .

b) The divergence of the vector field at a given point is a scalar

denoted as and defined by the equation

For example, if , then the divergence of is,

c) The curl of the vector field at a given point is a vector denoted as

and defined by the equation

Equation (1-8-3) can be expressed in an easy to remember symbolic form,

using the notation of determinants, i.e.

This determinant is to be expanded along the terms of the first row while
the operations of multiplication of the elements of the second row by the
elements of the third row are to be regarded as operations of differentiation,
for example, , etc.

For example, if then the curl of is,

 92

Note: Some authors use the symbol (the rotation of ), for the
, as defined in (1-8-4).

d) In vector analysis it is convenient to define the so called operator ,

by means of the formula,

This operator combines both differential and vectorial properties. In

vector algebra we perform formal operations involving as if it were a vector.
In terms of the operator, we may express the and as
follows:

To find the we simply operate the operator on , to find the

we dot multiply the by , while for the we cross multiply by
, (let the reader verify it).

In the formalism developed, using the operator , we must bear in mind

that the operator is not really a vector in the strict sense of the word; it
rather operates on a quantity appearing to its right. This means that while
is the divergence of , the term has no meaning.
 93

In summary:

1) With a given scalar field we may associate another vector field, the
gradient of , symbolized as .

2) With a given vector field we may associate two new fields, a scalar
field called the divergence of , symbolized as and a vector field called
the curl of , symbolized as .

e) Suppose that we have a scalar field . When the operator

applies on it generates a vector field, the gradient of , i.e.

The divergence of the vector field is, (according to the fundamental

definition 1-8-2),

We note that the divergence of the gradient of a scalar field is a scalar

field. The expression is called the Laplacian of the scalar
field . At this point we may define the Laplace operator as the dot
product of the del operator by itself. This operator is also symbolized as , i.e.

The Laplacian of a scalar field is obtained when the Laplace operator

operates on , i.e.

This operator appears quite often and plays a fundamental role in

Mathematical Physics.
 94

We point out that formula (1-8-9) expresses Laplace’s operator in

rectangular coordinates. Expressions for the Laplacian in cylindrical and
spherical coordinates will be developed in Section 1-11.

Example 1-8-1: Find the gradient of the scalar field .

Solution

Example 1-8-2: Find the gradient of at the point

.

Solution

and at the point , .

Example 1-8-3: Find the divergence and the curl of the vector field
.

Solution
 95

This shows that the curl of is identically equal to zero, at all points. Such
fields, characterized by the property that their curl is zero in some domain (in
our example the whole space), are termed irrotational fields.

PROBLEMS

1-8-1) Consider the time varying field and show that

.

1-8-2) If and are scalar fields, show that a)

and b) .

1-8-3) If , find and .

(Ans:
).

1-8-4) If , find and .

1-8-5) If show that

.

1-9) The Gauss-Ostrogradsky Theorem (Divergence Theorem) and Stokes

Theorem

The Gauss-Ostrogradsky theorem (known also as the Divergence Theorem)

and the Stokes’ Theorem are two of the most important and powerful
Theorems in Vector Analysis, with a wide range of applications in various
branches of Natural Sciences, (Fluid dynamics, Electromagnetic fields theory,
Heat transfer, etc.). For instance, these theorems are often used in order to
derive the general laws of the electromagnetic fields, the so called Maxwell’s
equations.
 96

a) The Gauss-Ostrogradsky Theorem (the Divergence Theorem).

Let us consider a vector field ,

defined in some region in space. We assume that the vector field is
continuous and has continuous partial derivatives in . Then the Divergence
theorem may be stated as follows:

The Divergence Theorem: The flux of the vector field through any
closed, piecewise smooth surface located in the region , is equal to the
triple integral of the divergence of over the volume bounded by the
closed surface . In symbols:

The normal unit vector to the surface is taken to be the outer normal.

The divergence theorem transforms a surface integral of the second type

(over a closed surface) to a volume integral over the volume bounded by this
closed surface, (see Fig. 1-35). It is a powerful theorem with many applications.

Fig. 1-35: The divergence theorem of Gauss.

 97

b) Stokes’ Theorem

a) Let be an open, two sided surface, bounded by a simple closed (non-

intersecting) curve . We further assume that the surface is oriented w.r.t.
the direction of traversing its boundary and let be a unit vector normal to
the surface , pointing from the inner towards the outer side of the surface.

Fig. 1-36: Oriented surface and outward unit normal.

Assuming that a vector field

is continuous and has continuous partial derivatives, Stokes’
Theorem may be stated as follows:

Strokes’ Theorem: The circulation of the vector field around a closed

contour is equal to the flux of the curl of the vector field through any
surface which is bounded by the contour . In symbols:

Stokes’ Theorem transforms a line integral of the second type (over a

closed path) to a surface integral of the second type.

Example 1-9-1: Evaluate the flux of the vector field

through the surface of the sphere of radius .
 98

Solution

Application of the Divergence theorem implies,

Example 1-9-2: Verify Stokes’ theorem if the vector field is

and the surface is the upper half of the sphere .

Solution

a) Direct calculation of the circulation: The closed contour of integration is

the circle , which lies on the plane.

Fig. 1-37: Verification of Stokes’ theorem.

The points on the closed path admit the following

parameterization, . The dot product
on the closed path is (since ),
 99

b) Computation using Stokes’ theorem: The curl of is

On the surface of the sphere, (the unit vector in spherical

coordinates), and

The dot products are obtained from equation (1-4-4), and

substituting into (**) we get,

and since , we have:

 100

Direct evaluation of the integrals and shows that and

, and formula (***) implies that , which is
identical to the result found in (*) and this verifies Stokes’ theorem.

PROBLEMS

1-9-1) Compute the flux of the vector field through

the closed surface bounded by the top half of the spherical surface and
the plane.

(Ans: ).

1-9-2) Verify the divergence theorem assuming that

and that the closed surface is the rectangular
parallelepiped .

1-9-3) Verify the divergence theorem for taken over

the surface of the sphere .

(Ans: ).

1-9-4) If , , show that

.

1-9-5) Compute the circulation of the vector field along

the closed path which lies on the plane, (a) by direct
calculation and (b) using Stokes’ theorem.
 101

(Ans: ).

1-9-6) Compute the circulation of the vector field along

the closed path formed by the intersection of the plane with the
coordinate planes and , (a) by direct calculation and (b)
using Stokes’ theorem.

1-9-7) Use Stokes’ theorem to compute the circulation of along the

contour .

(Ans: ).

1-10) The Physical Meaning of the Divergence, the Curl and the Gradient

a) The physical meaning of the divergence of a vector field.

The Divergence theorem can be used to assign a physical meaning to the

divergence of a vector field at an arbitrary point in the field. Let us for
definiteness consider a point in space, and surround this point by a surface
of arbitrary shape, for instance by a sphere of a small radius , as shown in
Fig. 1-38. Let be the small volume enclosed by the closed surface .

Fig. 1-38: The physical meaning of the divergence.

Application of the divergence theorem yields,

 102

According to the properties of triple integrals, the triple integral in (*) is

equal to the value of the integrand at a point , inside the small
volume , multiplied by , i.e.

Assuming now that , the point tends to coincide

with the point , and in this case formula (**) implies that

This formula gives an invariant definition of the divergence of a vector field

at an arbitrary point and reveals its physical meaning; the divergence of a
vector field at an arbitrary point represents in fact the volume density of
the flux of the vector field at this particular point. This property, which can
also be considered as the definition of the divergence, is independent of the
coordinate system being used. We may say that the expression of the
divergence in a rectangular coordinate system is expressed by .
In other coordinate systems, for instance the cylindrical and the spherical
systems, the divergence assumes different expressions, (see Sec. 1-11).

The points at which the divergence of a vector field is positive are termed
“sources” while the points where the divergence of a vector field is negative
are termed “sinks” of the vector field.

If at all points of a domain in space, the divergence of a vector field is

zero, the vector field is called “solenoidal field”.
 103

Fig. 1-39: Sources and sinks of vector fields.

b) The physical meaning of the curl of a vector field.

Stokes’ theorem can be used to assign a physical meaning to the curl

(rotation) of a vector field, independent from the coordinate system being
used. This invariant definition of is useful in obtaining an expression for
the in other coordinate systems, for example the cylindrical and the
spherical systems, (see Sec. 1-11). Let us for definiteness consider a small
surface bounded by a simple closed curve and let us choose a point
on , (not on the boundary ). If is a unit vector normal to at the point
, then application of Stokes’ theorem implies that

Fig. 1-40: The physical meaning of the curl.

 104

According to the properties of the surface integrals, the surface integral

is equal to the value of the integrand at some
point on the surface multiplied by , i.e.

Assuming now that the surface shrinks to zero, but always enclosing the
point , the point tends to coincide with the point , and in the limit as
, formula (**) yield the exact value of the quantity at the
point , i.e.

Note that in formula (1-10-3) the small surface is perpendicular to at

the point . In words, the projection of the vector on any direction is
independent from the coordinate system being used and is equal to the
surface density of the circulation of the vector around the simple closed
curve of the area perpendicular to that direction.

Taking for instance , then formula (1-10-3) yields the

component of the at the arbitrary point , and similarly we may find the
and components of . Knowing the three components of we
may easily construct the vector .

c) The physical meaning of the gradient of a scalar field.

Let us now consider a scalar field and a unit vector , which actually
defines a direction in space. If we start from a point of the level surface
of the scalar field and move to another the point which lies on the level
surface , the motion being parallel to the unit vector (i.e. so that
), then the limit
 105

is called the directional derivative of the scalar field in the direction of .

This limit actually determines the rate of change of the scalar field in the
direction of , at the point .

Fig. 1-41: Directional derivatives.

Evaluation of the directional derivative: If are the

direction cosines of the unit vector , (see Sec. 1-1, formula (1-1-15)), and call
, then

and application of formula (1-10-4) yields,

However, since represents an arbitrary point in space, we may drop the

subscript , and state that in general, the directional derivative of a scalar field
at an arbitrary point in space is given by the formula
 106

Since and ,
(remember , since is a unit vector), formula (1-10-6) can be written in
a different but equivalent form as follows,

i.e. the directional derivative of a scalar field in the direction determined by

a unit vector , is equal to the dot product of the gradient of the field by
the unit vector .

Formula (1-10-7) is important since it yields the physical meaning of the

gradient of a scalar field, independent from any chosen system of
coordinates. As a matter of fact, we shall use this invariant characteristic of
the gradient to find its expression in other coordinate systems, for example
the cylindrical and the spherical systems.

To reveal the physical meaning of the gradient of a scalar field , let

us consider Fig. 1-42. If is the angle between and at an arbitrary point
, then

and this shows that .

Fig. 1-42: Physical meaning of the .

 107

From the inequality the following properties of the gradient

are implied:

1) The direction of the most rapid increase of at a point is the

direction of the gradient at this point. The derivative in this direction is equal
to the modulus of the gradient at the given point, i.e.

2) If is a differential vector displacement given by

and is a scalar field, then the differential of the
vector field can be expressed as

Indeed,

3) The gradient (as a vector) is in the direction of the normal to the

level surface passing through the particular point. In case of a plane field, the
gradient is normal to the level lines.

Indeed, let be the equation of a level surface. Let

be the position vector of a point on the surface and be an arbitrary
differential vector lying on the plane tangent to the level surface
at the point and whose initial point is the terminal point of .
Since , the differential , and equation (1-10-10)
implies that , which in turn means that is perpendicular to
the level surface at the point .

These properties reveal an invariant characteristic of the gradient, which is

independent from the chosen coordinate system.
 108

1-11) Expressions for the Gradient, the Divergence, the Curl and the
Laplacian in Cylindrical and in Spherical coordinates

In section 1-8 we gave the expressions for the gradient, the divergence, the
curl and the Laplacian in rectangular coordinates, where the independent
variables are and the unit vectors are . In this section we shall give
expressions for these quantities in cylindrical coordinates (independent
variables and unit vectors ), and in spherical coordinates
(independent variables and unit vectors ).

In problems with cylindrical symmetry (axial symmetry), calculations are

usually simplified when working in cylindrical coordinates. In problems with
central symmetry (spherical symmetry), calculations are simplified when
working in spherical coordinates.

The great advantage of the rectangular system, as compared with the

cylindrical and the spherical systems, is that while and are constant
vectors, (not depending on the position) in cylindrical and in spherical
coordinates the corresponding unit vectors do depend on the position (are
functions of the position).

Let us for definiteness consider a scalar field . Depending on the

coordinate system being used we have:

Similarly, for a vector field ,

Note that denotes the component of and not the partial derivative
w.r.t. , and similarly for all the other components of the field.
 109

a) Expressions for the in cylindrical

coordinates.

b) Expressions for the in spherical

coordinates.

Example 1-11-1: A vector field is expressed in cylindrical coordinates as

. Find the and the .

Solution
 110

The components of the vector field are and

. Application of (1-11-1) yields,

Example 1-11-2: For the vector field in example 1-11-1, verify the
vector identity .

Solution

If we set , the components of are and

. The divergence of is obtained from formula (1-11-1),
i.e.
 111

Example 1-11-3: For the field show that .

Solution

The given field, in spherical coordinates, depends only on , and this means
that and . The Laplacian in spherical coordinates is obtained
from (1-11-2),

Example 1-11-4: Find the gradient of the scalar field .

Solution

From formula (1-11-2) we have,

Example 1-11-5: Find the curl of the vector field

.

Solution

The components of the field are and

. The curl of is obtained from the corresponding formula in
equation (1-11-2), i.e.
 112

(let the reader verify the calculations).

Example 1-11-6: For the vector field in Example 1-11-5, verify the vector
identity .

Solution

The components of the vector field are ,

and . The divergence of in spherical
coordinates is

and substituting the expressions for and simplifying we get

, (let the reader verify the calculations).

PROBLEMS

1-11-1) If find and then verify the vector identity

.

(Ans: ).

1-11-2) Find the curl of the vector field

.

1-11-3) In Problem 1-11-2 show the vector identity .

1-11-4) The expression of a vector field in cylindrical coordinates is

. Find and verify the vector identity
.

1-11-5) If , find and then verify

Stokes’ theorem on the surface .

(Ans: ).
 113

1-11-6) In problem 1-11-5 verify the identity .

1-11-7) Find the directional derivative of the scalar field at

the point in the direction of .

(Ans: ).

1-11-8) Find the directional derivative of at

the point in the direction of .

1-11-9) What is the greatest rate of increase of at the point

. What is the direction of the greatest rate of increase?

(Ans: ).

1-11-10) Find the angle between the gradients of the functions

and at the point .

1-12) Solenoidal and Irrotational Fields, Potential Fields and Vector

Potentials

Definition 1: If at all points of a domain the divergence of a vector field

(defined in ) is zero, then we say that the field is solenoidal in this domain.

Definition 2: If at all points of a domain the curl of a vector field

(defined in ) is zero, then we say that the field is irrotational in this domain.

Another name for irrotational fields is conservative fields.

For example the vector field is solenoidal, since

. The vector field
is irrotational, since , (let the reader verify it).
 114

a) PROPERTIES OF SOLENOIDAL FIELDS:

1) A solenoidal field is free of sources and sinks (see section 1-10). This is
so since .

2) In a solenoidal field the total flux of the field through any closed
surface lying in is zero, i.e. .

This is proved easily with the aid of the divergence theorem,

3) Since , where is an arbitrary vector field, it

follows that if is solenoidal, then there exists a vector field such
that

In other words, every solenoidal field can be represented as the curl of

another vector field . This vector field is called the vector potential of the
field .

4) The vector field of a solenoidal field is not unique. On the contrary, if

is any scalar function, then the vector function is also
a vector potential of , since

b) PROPERTIES OF IRROTATIONAL (CONSERVATIVE) FIELDS:

1) If is an irrotational field, then there exists a scalar function

such that , i.e.
 115

This scalar function is called the scalar potential of the

irrotational field . The scalar potential of is determined up to an arbitrary
constant , since .

2) If is irrotational in a domain and is any closed, simple curve in ,

then

Fig. 1-43: Irrotational fields, (Conservative fields).

Equation (1-12-6) shows that the circulation of an irrotational field around

any closed contour lying in is zero.

In Fig. 1-43, let be any surface bounded by the closed curve .

Application of Stokes’ Theorem implies,

and this proves equation (1-12-6).

3) If is an irrotational field, then the line integral is independent

from the path connecting the points and . Furthermore, if is the
scalar potential of , then
 116

Proof: By virtue of equation (1-12-6) and considering Fig. 1-43, we have,

and this shows that the integral is independent from the path connecting
the points and , it depends only on the initial and the terminal points. Also,
if is the scalar potential of , then

and the line integral from to is,

and this completes the proof.

 117

c) HARMONIC FIELDS: A vector field which is simultaneously irrotational

and solenoidal, is called harmonic. Since is irrotational, then , where
is its scalar potential. Also since is solenoidal, then , i.e.

and this shows that the potential of an harmonic field is an harmonic

function (i.e. it satisfies Laplace’s equation).

Theorem 11-1 (Helmholtz’s Theorem): Every vector field can be

expressed as the sum of a solenoidal field and an irrotational field , i.e.

Note: A distinction should be made between Harmonic fields and Time

Harmonic Fields. The term Time Harmonic Fields, refers to fields that do
depend on time and the time variation is a sinusoidal function of time, (see
section 1-1). The term harmonic refer to functions that satisfy Laplace’s
equation.

Example 1-12-1: Show that the field

has neither sources nor sinks, i.e. it is a solenoidal field.

Solution

Example 1-12-2: Show that the plane field

is irrotational and determine its scalar potential .
 118

Solution

and this shows that is irrotational. If is its scalar potential,

then we must have,

From the first equation in (*) we get (with partial integration w.r.t. )

and taking the partial derivative of w.r.t. , (notice that in this case
), we get,

Note that is determined up to an arbitrary constant .

Example 1-12-3: For the vector field in Example 1-12-2, evaluate the
integral .
 119

Solution

Since is irrotational, we may apply formula (1-12-7) to compute the

integral (which is independent of the path of integration), i.e.

Note that in the difference, the arbitrary constant is eliminated.

PROBLEMS

1-12-1) Show that the vector field is

irrotational and find its scalar potential .

(Ans: ).

1-12-2) For the vector field in Problem 1-12-1, find the integral
.

Hint: See Example 1-12-2.

1-12-3) For the vector field in Problem 1-12-1, evaluate the integral
along the straight line connecting and , and verify that the
answer obtained is identical to the answer found in Problem 1-12-2.

(Ans: ).

1-12-4) Show that the line integral of the vector field

does depend on the line path connecting the points and
.

Hint: It suffices to show that is not irrotational.

1-12-5) Show that the vector field

is solenoidal but not irrotational, ( are constants, not zero).

1-12-6) For the solenoidal field in Pr. 1-12-5 find its vector potential .
 120

(Ans: ).

1-12-7) Show that the vector is perpendicular to the level surfaces

. In electric fields, this means that the electric field is
perpendicular to the surfaces of constant potential.

1-12-8) Check whether the following scalar fields are harmonic or not: a)
, b) , c) (cylindrical
coordinates), d) (spherical coordinates), e)
(spherical coordinates).

Hint: A function is harmonic if it satisfies Laplace’s equation .

1-12-9) Show that the field is solenoidal and find its

vector potential .

(Ans: ).

1-13) Source-Point and Field-Point Notation

In the theory of electromagnetic fields the main problem is to determine

the fields (electric or/and magnetic) generated by a given charge distribution
(or a given current distribution). Charges and currents are the sources of
electromagnetic fields. The coordinate points needed to designate the location
of the sources of the fields, i.e. the points where charges and/or currents are
located, are called source points. The points where the field is to be evaluated
are called field points. Usually the fields are found by a proper integration
over the regions where the sources are located. In order to avoid confusion,
we shall use the notation for the source points and for the
field points (where the field is to be found). Source points and field points are
two sets of points referring to the same coordinate system. In shorthand
notation stands for and stands for .

Let be the distance vector between the source points and the field
points, as shown in Fig. 1-44. Both sets of points are referring to the same
coordinate system .
 121

Fig. 1-44: Source points and field points.

In rectangular coordinates, while and

the distance vector is

The magnitude of is

The unit vector in the direction of is

We note that actually depends on six variables, the three field

coordinates and the three source coordinates . We may now
define the operator (Del operator) which operates on the source
coordinates, as follows:
 122

The Del operator operates on the field coordinates. It is easy to show the
following two useful relations,

For a proof, see Example 1-13-1.

Example 1-13-1: Show the two formulas in equation (1-13-5).

Solution

Similarly, we find that

and substituting in (*) we get,

and this completes the proof of the first equation in (1-13-5).

To show the second formula in equation (1-13-5) we apply the operator

on the scalar :
 123

Similarly we find,

and substituting in (**) we get,

and this completes the proof.

PROBLEMS

1-13-1) Show that and .

1-13-2) Show that .

1-13-3) Show that and .

1-13-4) Show that .

1-13-5) Show that if is a scalar field of , then .

 124

SUPPLEMENTARY PROBLEMS

1-1) Find the dot product and the cross product of the vectors,
and .

(Ans: ).

1-2) Find the unit vector in the direction of .

Hint: .

1-3) If and find the unit vector

perpendicular to the plane defined by the vectors and .

(Ans: ).

1-4) If and , find: (a) the magnitude of

, (b) the unit vector in the direction of , and (c) the angle
between the vectors and .

1-5) Given the points and find: (a) the

unit vector directed from to , (b) the angle between the vectors and
, and (c) the length of the perimeter of the triangle .

(Ans: ).

1-6) Find the Cartesian and the spherical coordinates of the point
.

1-7) Find the Cartesian and the cylindrical coordinates of the point
.

(Ans: ).

1-8) Express the vector field in cylindrical

coordinates, (i.e. express as ) and then find at the point
.

Hint: See Example 1-4-2.

 125

1-9) Express the scalar field in cylindrical and in spherical

coordinates.

(Ans: ).

1-10) Provided that and

, find and .

1-11) Find the indefinite integral of the vector function

.

(Ans: ).

1-12) Compute the definite integral , given that

.

1-13) Compute the flux of the vector through the lateral

surface of the cylindrical surface , bounded by the planes and
, (Ans: ).

Hint: The calculations are facilitated if we transform in cylindrical

coordinates.

1-14) Show that the vector field is solenoidal throughout any

region that does not contain the origin (i.e. ). Notice that at , the
field is not even defined.

1-15) Show that the vector field is solenoidal and find its
vector potential , (Ans: ).

1-16) Show that the vector field is solenoidal and find

its vector potential .

1-17) Show that the vector field (in spherical coordinates),

, is a potential field, i.e. .
 126

1-18) Provided the is a differentiable function, show that the vector

field is a potential field and find its scalar potential .

1-19) Show that the field is irrotational and find its scalar
potential , (Ans: ( arbitrary constant)).

1-20) Compute the circulation of the field over

the closed curve , (a) directly and (b) via
the Stokes Theorem, (Ans: ).

1-21) Compute the circulation of the vector field

around the circle , (a) directly and (b) via the Stokes
theorem, (the positive direction of (c) is in the direction of increasing ).

(Ans: ).

1-22) Find the flux of the vector field through the surface of
the sphere , (a) by direct computation and (b) using Gauss-Ostrogradsky
theorem, ( ).

1-23) Find the flux of the vector field through the

surface of the sphere , (a) by direct computation and (b) using Gauss-
Ostrogradsky theorem, ( ), (Ans: ).

1-24) Show that the field is irrotational and find its scalar
potential , (Ans: , is an arbitrary constant).

1-25) Find the Laplacian of the scalar field .

(Ans: ).

1-26) Show that the field is harmonic, i.e. satisfies Laplace’s

equation .

1-27) Find all the solutions of Laplace’s equation , (in spherical

coordinates), that depend only on , (Ans: ).
 127

ELECTROMAGNETICS: A BRIEF HISTORY

Newton’s second law of motion, ( ) is, perhaps, the most

fundamental equation in Mechanics. It shows how a body of mass is moving
when it is acted upon by a force . In nature there are four kinds of forces:
Gravitational, Electromagnetic, Strong nuclear and Weak nuclear forces.
Nuclear forces have very short range and are important within the nucleus of
the atoms. These forces keep the nucleus together. Gravitational forces are
dominant on a planetary to cosmic scale, and are responsible for the motion of
planets and all other heavenly bodies. Gravitational forces have long range, as
opposed to the nuclear forces. Electromagnetic forces are very important and
are somewhere in between. Virtually, all forces we experience in our everyday
life are either electric or magnetic, (even though this is not so obvious). For
example, friction, chemical forces that bind the atoms to form molecules, etc,
are all electromagnetic forces.

It all started about six hundred years before Christ, when a Greek
mathematician and philosopher, Thales of Miletus (640-546 BC), noticed that,
when amber is rubbed with silk, it acquires the very interesting (and curious)
property to attract light weight particles of straw and fluff. The Greek word for
amber is “ηλεκτρον” (electron), and from this we get the contemporary terms,
electron, electricity, electronics, etc. Most probably, Thales was also aware of
the action of the “lodestone” (a natural magnetic substance), found at a place
called “Magnesia” (Μαγνησια); the lodestone attracted small pieces of iron,
and since this substance was found in Magnesia, Thales named it “μαγνητη”
(magnet).

For many centuries, after Thales, there was actually no progress regarding
the knowledge of electric and magnetic phenomena.

In 1551 AD, the Italian Gardano Girolamo (1501-1576), found that the
electrical forces, due to the amber, are of different nature from the magnetic
forces due to magnets.

In 1600 AD, 2200 years after Thales, the British scientist William Gilbert
(1544-1603), made a systematic study of various electric and magnetic
phenomena. In his book “De Magnete”, he quotes that “…other bodies
acquire, by rubbing, the attractive power of the amber…” .
 128

In 1750, the American scientist Benjamin Franklin, (mostly known for his
invention of the lightning rod), determined that there are two kinds of electric
charges, positive charges and negative charges and furthermore, he
established the law of conservation of charge.

In 1785, the French scientist Charles-August de Coulomb (1736-1806),

established the first quantitative formula for the electric forces between two
charged bodies, (the famous Coulomb’s law).

In 1800, Alessandro Volta, an Italian Physicist, invented the “voltaic cell”

and then, connecting voltaic cells in series, we was able to make the first
“electric battery”. Connecting a wire to the poles of the battery, an “electric
current” flows through the wire.

Up to this point, scientists assumed that electric and magnetic phenomena

were not related at all. Electric and magnetic phenomena were considered to
be independent from each other.

However, in 1819, the Danish Physicist Hans Christian Oersted (1787-1851),

accidentally, in the course of performing an experiment, noticed that a current-
carrying wire deflected a nearby magnetic compass. But, what deflects a
magnetic compass is a magnetic field. The conclusion was inevitable: Electric
currents, (i.e. motion of electric charges), generate magnetic fields. This was
the first indication that electricity and magnetism are, somehow, interrelated.

It was only a few years later, in 1831, when the British Physicist Michael
Faraday (1791-1867), established exactly the reverse result; time- varying
magnetic fields can generate electric currents. That was a second indication
that electricity and magnetism are interrelated.

James Clerk Maxwell (1831-1879), a prominent Physicist at Cambridge

University, compiled all experimental and theoretical data on electricity and
magnetism, available at his time, put it all together and built a compact theory,
as it is known today. In his great work “Treatise on Electricity and Magnetism”
in 1873, Maxwell introduced a new electromagnetic theory. Maxwell’s
equations are four partial differential equations, relating the electric and the
magnetic fields with their sources, (electric charges and electric currents). In
 129

fact, Maxwell’s great contribution was the ingenious introduction of the

“electric displacement current” into his equations.

By a proper transformation, Maxwell’s equations lead to a wave type of

equation for the electric and the magnetic fields. These fields, the
Electromagnetic fields, travel in vacuum, (free space), with the speed of light.
In other words, Maxwell’s equations predict the existence of electromagnetic
waves.

In 1888, 15 years after Maxwell published his equations, the German

Physicist H. Hertz, succeeded in generating and detecting radio waves of
wavelength of about 5 meters, with the aid of a spark transmitter and receiver.

In 1901, Guglielmo Marconi (1874-1937), improved Hertz’s transmitter, and

was able to send radio waves over the Atlantic Ocean. The era of radio
communication had just begun.

One last comment: In the regime of Maxwell’s equations, the speed of light
, which is an electromagnetic wave of extremely short wavelength, is given by
the formula:

where is the permittivity of free space and

is the permeability of free space. Since and are
constants of nature, the speed of light must also be a constant of nature,
independent of the motion of the source emitting the light. This was in a direct
conflict with the physical laws, before 1900, since, by that time, the notion of
an absolute speed was impossible.

Einstein was motivated by Maxwell’s theory of EM fields, and in fact, the

constancy of the speed of light was the second postulate in his Special Theory
of Relativity, published in 1905.
 130

PART A: ELECTROSTATICS
 131

CHAPTER 2: COULOMB’S LAW

2-1) Electric Charge: Units and Properties

a) According to the current atomic theory, matter is made up of discrete

units called “atoms”. The word “atom” originates from the Greek word
“ατομο” meaning “uncuttable”. The ancient Greek philosopher Democritus
(460 BC-370 BC) taught that matter is not indefinitely divisible; instead, we
reach a point where further division is not possible.

Large experimental evidence, primarily towards the end of the 19th-

beginning of the 20th century, revealed that atoms consist of a heavy,
positively charged nucleus and some negatively charged lighter particles,
called electrons revolving about the nucleus.

The nucleus of an atom consists of two kinds of particles, the protons which
are positively charged and the neutrons which are electrically neutral (they
bear no electric charge).

Experimentally it has been found that, in nature, there are two kinds of
electric charge. One kind is called “positive charge” and the other kind is “the
negative charge”. The electron is considered to be the carrier of the
elementary negative charge in nature, while the proton the carrier of the
elementary positive charge.

In the S.I. system of units, the unit of the electric charge is (Coulomb).
Experimentally, it has been determined that the mass and the charge of the
electrons and the protons are:

The mass of the proton is about the same with the mass of the neutron;
also, the mass of the proton is about 1835 times larger than the mass of the
electron.
 132

Under normal conditions, an atom contains the same number of protons

and electrons, i.e. the same number of positive and negative charges and the
atom appears to be (macroscopically) electrically neutral. In Fig. 2-1 we see a
simple model of the Hydrogen and the Carbon atoms.

Fig. 2-1: Hydrogen and Carbon atoms.

The Hydrogen atom, (the simplest atom in nature), has one proton (positive
charge) and one orbiting electron (negative charge). The atom is electrically
neutral. The Carbon atom has six protons in its nucleus and six orbiting
electrons. The atom is electrically neutral. The carbon atom has also six
neutrons (in the nucleus) but the neutrons do not carry electric charge.

If a neutral atom looses one or more of its electrons, then the number of
protons will exceed the number of electrons, and macroscopically the atom
will have a surplus of positive charge. If a neutral atom gains on or more
electrons, then the number of electrons will exceed the number of protons and
the atom will have a surplus of negative charge. Positively charged atoms are
called positive ions, while negatively charged atoms are called negative ions.

b) What is “electric charge” is not really known. All we can say is that
charge is a fundamental property of matter, and appears in nature in two
species, which we (arbitrarily) call negative and positive. Electrons are the
carriers of negative charge while protons are the carriers of positive charge.
 133

Extensive research has revealed the following fundamental properties of

electric charge:

1) Like charges repel each other while unlike charges attract.

2) Charge is quantized: This simply means that there is a basic unit of

charge in nature, (either the electron’s charge or the proton’s charge), and all
other charge quantities are integral multiples of this basic unit. So if we call
the charge of the electron, we may have a charge of or a charge of
, etc, but never a charge of or , etc.

3) Charge is conserved: Electric charge can neither be created nor

destroyed. This property implies that the total amount of charge in the
universe remains constant (does not vary with time).

4) A body (bulk matter) which consists of neutral atoms shall be electrically

neutral. When this body receives electrons then the body becomes negatively
charged, while when it loses electrons becomes positively charged. In either
case the body becomes electrically charged. It is worth mentioning that due to
the extremely small mass of the electron ( ), the mass of
the charged body is practically the same with the mass of the neutral body.

2-2) Distributions of Electric Charges: Discrete - Continuous

In our efforts to find the force between two charged bodies, (electric
repulsion or attraction), it facilitates the calculations involved if we consider
the following two basic distributions of charges, the discrete distribution and
the continuous distribution.

a) Discrete distribution of charges.

Let us consider point charges , whose positions relative to a

given coordinate system are given by their position vectors ,
respectively, as shown in Fig. 2-2.

We say that these point charges form a discrete charge distribution.

The term “point charge”, which we shall refer to, frequently, from now on,
means that the charges are distributed over material bodies
 134

respectively, whose sizes, however, are negligible as compared

to their separation distances.

Fig. 2-2: Discrete charge distribution.

b) Continuous distributions of charges.

There are cases where we consider charges that are distributed

continuously over some region. Depending on the region (line, surface or
volume), we consider the following cases:

1) Line charge distribution when charge is distributed along a line , (not

necessarily a straight line), as shown in Fig. 2-3. We define the line charge
density as

where is the differential charge over the differential length . From

equation (2-2-1) we obtain,

(line integral of the first type).

 135

2) Surface charge density when charge is distributed over a surface , as

shown in Fig. 2-3. The surface charge density is defined as

where is the differential charge over the differential area . From

equation (2-2-3) we get,

(surface integral of the first type).

3) Volume charge density when charge is distributed throughout a volume

, as shown in Fig. 2-3. The volume charge density is defined as

where is the differential charge in the differential volume . From

equation (2-2-5) we get,

(triple integral).

Note: As we shall show in the sequel, (Chapter 9), volume charges when
placed within a material body of volume , remain at the position they are
introduced, provided that the body is a perfect insulator. However, in nature,
there are no perfect insulators (nor perfect conductors). Even the slightest
conductivity results in the movement of volume charges towards the surface of
the body, where they are now distributed as surface charges.

It is possible, of course, in the general case, to have both discrete and

continuous charge distributions.
 136

Fig. 2-3: Continuous charge distributions.

Example 2-2-1: In a cylindrical coordinate system the volume

charge density is . a) Find the total charge inside
the cylinder , and b) Find if the total
 137

charge found in the volume is half the

value found in part (a) above.

Solution

a) The differential volume in cylindrical coordinates is , and

application of formula (2-2-6) yields:

To evaluate the first integral we make the substitution , then

, and

and the charge is

b) According to the statement of problem we must have:

 138

Example 2-2-2: In a spherical coordinate system ( ) the volume charge

density is . a) Find the total charge inside a sphere of
radius , b) Find the total charge inside the cone
, and c) Find the total charge inside the region
.

Solution

a) The total charge inside the sphere is

b) Inside the cone the total charge is

found similarly, i.e.

c) Inside the region the total charge is

 139

PROBLEMS

2-2-1) The line charge density along the axis is

where . Find the total charge on the line, (Ans: ).

Hint: .

2-2-2) Find the total charge inside the volume

if .

2-2-3) For positive, let . Find the total

charge inside the cube , (Ans: ).

2-2-4) The surface charge density over the plane is

. Find the total charge inside the circle .

Hint: To evaluate the integral, pass to polar coordinates, (see eq. (1-7-2)).

2-2-5) Assuming that

, find the total charge over the whole plane, (Ans: ).

Hint: In the answer found in Pr. 2-2-4, take the limit as .

2-2-6) Assuming that

find the total charge throughout the whole space.

2-3) Coulomb’s Law

a) Let us consider two point charges and in free space (vacuum), at

rest w.r.t. a given coordinate system, whose positions are determined by their
position vectors and respectively, (see Fig. 2-4). The force on the point
charge due to the charge is given by Coulomb’s law,
 140

Fig. 2-4: Coulomb’s Law.

In formula (2-3-1), is the vector extending from to (the

charge on which we want to determine the force , due to ), and is a
universal constant, called the permittivity of free space (vacuum). The value
of , measured in ( , as obtained from (2-3-1), is

In Chapter 11, we shall define the Farad ( ), the unit of the capacitance and
will show that it has dimensions of ( ). In terms of the Farad, the
permittivity of free space is written, equivalently as

The second expression for facilitates the calculations, in practice, since

.

By virtue of Newton’s Third law (Action=Reaction), the charge exerts a

force on , equal to
 141

From Coulomb’s formula we may easily draw the following conclusions:

1) The force between two charged particles lies in the line determined by
the two charges.

2) This force is proportional to the product of the charges and inversely

proportional to the square of their separation distance .

3) Like charges (both positive or both negative) are repelled whereas

unlike charges (one positive and one negative) are attracted. This is so since
in the former case and is in the direction of , while in the latter
case and is in the direction of ( ).

b) Some remarks on Coulomb’s formula:

1) The magnitude of the force between two charged particles bears a close
resemblance to the magnitude of the gravitational attraction between two
particles with masses and , separated by a distance ; the force of
attraction is given by the formula

where is the universal constant of gravitation; the numerical value of

is . However, while the gravitational force
is always attractive, electric forces can be either attractive (unlike charges) or
repulsive (like charges).

2) Strictly speaking, Coulomb’s formula holds true for “point charges”, i.e.
for charged particles whose dimensions are negligible as compared to their
separation distance. In all other cases, Coulomb’s law gives an “approximate
value” of the electric force.

2-4) Dielectric Materials

Coulomb’s formula (2-3-1) gives the electric force between two charged
particles in vacuum (or air, approximately). Vacuum is space devoid of matter.
 142

However, electric forces between charged particles do exist even in cases

where charges are placed inside matter. For the time being, we shall assume
that the material within which the charged particles are placed is a “perfect
insulator”. Practically, this means that the charged particles remain at the
positions they are initially placed, even though each one of the charges
experiences electric forces from all the other charges.

The property of material bodies to allow the transmission of electric forces

through their mass is called “dielectricity”. All insulators are dielectrics and
quite often the terms dielectrics and insulators are used indiscriminately. This
may lead to the (erroneous) conclusion that conducting bodies are not
dielectrics, i.e. they do not allow the transmission of electric forces through
them; however, this is not true, since even conducting bodies do allow electric
forces through their bodies, and hence they are described by a dielectric
constant.

Assume now that two point charges and are placed within an
homogeneous dielectric material (not vacuum), and let be the vector from
to (see Fig. 2-4). Then the force on due to , is given by the
formula

where is the “relative dielectric constant or the relative permittivity” of

the dielectric material. Note that is a pure number (no units). Usually, we
define the “permittivity ” of the dielectric material as

and formula (2-4-1) can now be expressed as

In fact, permittivity takes into account the polarization of electric charges,

when a dielectric material is placed within an electrostatic field already existing
in vacuum. Polarization of dielectrics is studied in Chapter 10, and a full
justification of formulas (2-4-1) and (2-4-2) is given in the same Chapter.
 143

The permittivity of various materials is found either experimentally or is

evaluated theoretically. The relative permittivity for some representative
materials is given in the following Table.

Material Relative Permittivity

Air 1.000590
Hydrogen 1.000264
Methane 1.000944
Carbon dioxide 1.000985

Amber 2.9
Bakelite 4.8
Germanium 16
Glass 5…16.5
Ice 3.1
Mica 4.5…6
Nylon 3.5
Paper 3
Plexiglas 3.5
Porcelain 6
Quartz 4.3…4.7
Rubber 2.5…2.8
Silicon 11.8
Snow 3.3
Soil (dry) 2.8
Titanium dioxide 100
Water (distilled) 80
Wood (dry) 2.5…6.8

We notice that the relative permittivity of gases (under normal conditions)

is practically equal to 1, (the first four entries in the Table).

Example 2-4-1: In free space (vacuum) consider at

and at . Find the vector force on due to and the
force on due to .
 144

Solution

Application of Coulomb’s formula (2-3-1) yields,

The vector , (directed from to ) is

The magnitude of is, , and substituting in

(*) we get,

The force on due to is , by

virtue of Newton’s Third Law (Action=Reaction). The same is found from
equation (*) if is replaced by .

Example 2-4-2: Repeat Example 2-4-1, assuming now that the two charges
are placed inside distilled water, whose relative permittivity is .

Solution

Application of formula (2-4-1) implies that

Example 2-4-3: The repulsive force between two identical positive ions
separated by a distance of is . Find the charge of each
ion and determine how many electrons are missing from each ion.
 145

Solution

The magnitude of the force between the two positive ions (assuming free
space conditions) is

and solving for we obtain,

Since the charge of each electron is (in magnitude) ,

the number of electrons missing from each ion is electrons.

Example 2-4-4: A certain charge is divided into two parts and

which are then placed a distance apart, inside a homogeneous dielectric of
relative permittivity . Find and so that the repulsive force between
them is the maximum possible. What is this maximum repulsive force?

Solution

If we call then , and the repulsive force between them is

The problem thus reduces to the following: For what the function
attains its maximum value? Setting the derivative equal
to zero we find:
 146

Thus the charge has to be divided into two equal parts in order to have
the maximum repulsive force possible. The maximum force is

PROBLEMS

2-4-1) The force between two charged particles in free space reduces by a
factor of when the charges are placed inside a homogeneous dielectric. Find
the relative permittivity of the dielectric. (Ans: ).

2-4-2) In free space (vacuum) consider at and

at . Find the vector force on due to and the force
on due to .

2-4-3) Compare the electric force versus the gravitational force between
two electrons placed a distance apart.

(Ans: ).

2-4-4) Two small identical balls of mass are hung from silk threads of
length and carry like charge , as shown in Fig. 2-5. Show that
. Assuming now that is very small, make the approximation

and show that .

Fig. 2-5: Force between two electric charges.

 147

2-4-5) In Pr. 2-4-4 what will be assuming that the system of the two
charges in Fig. 2-5 is placed inside a dielectric material of .

(Ans: ).

2-4-6) What equal amounts of like charges would have to be placed on two
planets with masses and to neutralize their gravitational attraction?
Assume that their sizes are very small as compared to their separation distance
so that we can safely apply Coulomb’s law.

Hint: See equation (2-3-5).

2-4-7) Two equal like charges are at a distance

apart, (in free space). The repulsive force on is
. If is positioned at the point find the
position of the second charge. (Ans: ).

2-5) Superposition Principle and the General Problem in Electrostatics

a) Let us assume that we have a discrete distribution of point charges

positioned at the points respectively, as
shown in Fig. 2-6. We further assume that a point charge , (usually called a
Test Charge), is placed at the point determined by its position vector .

Fig. 2-6: Discrete distribution of point charges.

According to Coulomb’s law, the force on , due to charge is

 148

Here we have assumed that all electric charges exist in free space (vacuum).
In case charges are placed inside a homogeneous dielectric with relative
permittivity , then as we have quoted in section 2-4, (equation (2-4-1)), the
force is found from equation (2-5-1), if is replaced by .

Note also that, in equation (2-5-1), is the vector directed from

towards , the charge on which the electric force is to be determined.

In a similar fashion we may find the forces on , due to the rest of the
charges, . Then, the total force on the test charge , due to all
charges is found with the aid of the principle of superposition,
i.e.

Notice that the total electric force on , due to all other charges is
proportional to ; this implies that the ratio is independent of , it
depends solely on the other charges and their relative disposition in space.

Equation (2-5-2) is a vector equation and as such can be split in three scalar
equations, one for each axis. For example, the component of is:

where of course (position vector of test charge ) and

(position vector of ). Similar expressions can be
found for and . Once the three components of have been found, then
the magnitude of and the angles this force makes with the coordinate axes
are given by the formulas:
 149

b) Continuous charge distributions

In case we have a continuous static charge distribution (line, surface or

volume charge distribution) then to find the force on the test charge the
summation in (2-5-2) is replaced by a proper integration as shown below:

1) Line charge distribution .

Fig. 2-7: Line charge distribution.

Let be the charge on the differential arc . The differential

force on the test charge , due to is, (Coulomb’s law),
 150

and the total force is found using superposition; however, now, since the
charge distribution is continuous, the summation is replaced by a proper line
integration along ( ), i.e.

2) Surface charge distribution .

Fig. 2-8: Surface charge distribution.

Reasoning as in part (1), the charge on the differential surface is

and the total force on the test charge due to the surface
charge distribution is
 151

3) Volume charge distribution .

Fig. 2-9: Volume charge distribution.

Reasoning as in part (1), the total force on the test charge due to the
volume charge distribution is

Some general remarks:

1) In the general case, the integrals in (2-5-5), (2-5-6) and (2-5-7) are very
difficult (if not impossible) to be evaluated. This is due to the fact that the
charge densities (line, surface and volume densities) are usually functions of
the position and also the vector (the vector from towards the position of
the test charge ) is not constant; on the contrary, the initial point of spans
the charge distribution, while the terminal point of is fixed (at the point
where the test charge is located). To clarify the situation, let us consider the
volume charge distribution in Fig. 2-9. Let be the coordinates of the
center of the differential volume , and be the location of
the test charge . Then , and formula (2-
5-7) implies that
 152

In equation (*) the integration is performed w.r.t. the source variables

, and the result, therefore, will be a function of , i.e.
, as expected. Similar considerations hold true for line and surface
charge distributions.

2) The evaluation of the complicated integrals is somehow facilitated if the

charge distributions are uniform, i.e. independent of the spatial coordinates or
in cases where there is some kind of symmetry involved.

3) Notice that in all cases, the force on the test charge is proportional
to , i.e. the ratio is independent of . We shall use this fact, in Chapter
3, to define the electric field of a given charge distribution.

c) The general problem in Electrostatics

The general problem in Electrostatics is the following: Given a static charge

distribution, either discrete or continuous or both, find the electric force on a
test charge placed at a specific point in space, due to all the other charges.
Coulomb’s law and the superposition principle are used in order to solve (at
least theoretically) the problem. The important thing to notice is that the
force on is directly proportional to , which means that the ratio
at any point in space, does not depend on , depends only on the
charge distribution. We shall use this fact in Chapter 3 to define the so called
“electric field ” generated by a given charge distribution.

Example 2-5-1: Three point charges and

are placed at the points and
respectively. Find the force on a test charge at the origin .
 153

Solution

Fig. 2-10: Discrete charge distribution.

Let be the vectors from to the origin , where the test

charge is positioned. We have:

The total force on is, (see formula (2-5-2):

 154

Example 2-5-2: Let us assume a uniform straight line charge distribution

extending along the axis in a cylindrical coordinate system from
up to , as shown in Fig. 2-11. Find the force on a test
charge , placed at a point .

Solution

Fig. 2-11: Infinite, uniform line charge distribution.

Uniform charge density means that the charge per length is constant, i.e.
. Let us define the plane as the plane which passes
through and is perpendicular to the axis. The cylindrical coordinates of
are . The differential charge on the differential length is
and the force on the test charge is (see formula (2-5-5),

since is constant and can be pulled out of the integral. From Fig. 2-11, we
have,

and substituting in (*) we get,

 155

Integral A: Since is a constant unit vector, it can be pulled out of the

integral, i.e.

Integral B: Similarly, in this integration with respect to along the axis,

and remain constant, and hence,

(For the evaluation of integral B, see Problem 2-5-1).

Having found integrals A and B, formula (***) yields:

Example 2-5-3: A uniform surface charge density exists inside

the circle , situated on the plane. Find the force on a
test charge , placed at the point .

Solution

The sought for force is given by equation (2-5-6), i.e.

since is constant (uniform charge distribution).

 156

Fig. 2-12: Uniform surface charge distribution.

In Fig. 2-12, we see that . We notice

that when we integrate over the surface of the circle, the unit vector
changes direction and therefore, is not a constant vector, and hence we
cannot take it out of the integral. However, the unit vector and the variable
remain constant in the integration and we can of course pull them out of the
integral. Substituting the expressions for and in (*) we obtain,

Integral is zero, while integral ,

and substituting in (**) we finally get,

 157

We notice that is in the direction, i.e. is perpendicular to the plane

.

Some remarks regarding the evaluation of the integrals and :

a) To evaluate integral , we express the differential area in polar

coordinates, (see equation (1-5-8)), and the integral becomes,

and this integral is evaluated easily if we make the substitution

, etc.

b) One possible method to evaluate integral is to express in terms of

the constant unit vectors and , i.e. set , (see
equation (1-3-3)) and then carry out the integration. Since and are constant
vectors, they can be taken out of the integral, etc. For a detailed proof, see Pr.
2-5-2.

Another method of evaluation is the following:

Fig. 2-13: Evaluation of the integral in equation (**).

 158

Consider the two differential areas which are symmetric w.r.t. the center of
the circle, as shown in Fig. 2-13. The quantities on these two
differential areas are opposite, since they both have the same and , the
same , but opposite , so these two quantities add up to zero.
Now recalling that the surface integral is actually the sum of an infinite
number of infinitesimal quantities of the form , and since we
have, as explained, pair wise cancellation, we conclude that integral .

Example 2-5-4: A uniform surface charge density exists on an

infinite, planar charged sheet, located in the plane. Find the force on
a test charge , placed at the point .

Solution

Fig. 2-14: Infinite, planar uniform charged sheet.

In previous Example 2-5-3, if we consider that the radius of the circle

increases indefinitely, then the circle covers the whole plane. So, the
sought for force may be obtained from equation (***) in Example 2-5-3, if
we take the limit of as , i.e.
 159

Note: The force on a test charge, as shown in (*) does not depend on the
distance of the charge from the plane. If the test charge is placed or
from the plane , the force is the same. This is contrary to our
intuition that the force should be larger close to the plane and weakens as we
move away. Evidently, this is not the case, and this is due to the fact that the
plane is infinite. We also note that the force is always perpendicular to the
plane and is directed away from the plane, (for a positive charge ).

PROBLEMS

2-5-1) Show that the value of integral in example 2-5-2 is .

2-5-2) In integral in equation (**), example 2-5-3, set

, carry out the integration and show that .

2-5-3) Find the force of attraction between the nucleus of a Hydrogen atom
and an electron. The radius of the Hydrogen atom is , and the
charge of the nucleus is equal in magnitude and opposite in sign to that of the
electron. (Ans: ).

2-5-4) How many times is the force of gravitational attraction between two
protons smaller than the force of Coulomb repulsion?

2-5-5) An electron rotates in a circular orbit around a nucleus of charge

, where is a positive integer. Determine the speed of the electron in its
orbit. (Ans: ).

Hint: Coulomb’s force acts as the centripetal force on the electron.

2-5-6) A negative charge is placed in the center of a square, each vertex of

which contains a positive charge . Find the magnitude of the negative charge
if the resulting force acting on each charge is equal to zero.
 160

2-5-7) Two small identical balls of mass are suspended on two threads so
that their surfaces are in contact. What charge should be supplied to the balls
for the tension of the threads to become equal to ? Assume that the length of
each thread is .

(Ans: ).

2-5-8) Four point charges

are placed at the points
respectively. Find the force on a test charge placed at the point
.

Hint: See Example 2-5-1.

2-5-9) Electric charge is uniformly distributed along the axis, from

up to , with a line charge density , and is zero everywhere
else. Find the force on a test charge placed at the point .

(Ans: ).

2-5-10) In Pr. 2-5-9, take the limit of as , derive the expression for
due an infinite, uniform line charge distribution and thus verify formula
(****) in Example 2-5-2.

2-5-11) The distance between two point charges and is . How should
a positive charge be arranged so that it will be in equilibrium?

(Ans: Between the two charges, at a distance from ).

2-5-12) The distance between two small, identical balls, having unlike
charges and is , while the attractive force between the balls is . After
the balls have been connected by a wire and the latter has been removed, the
balls repel each other with a force . Determine the original charges on the
balls. (Ans: ).

2-5-13) In Fig. 2-15, line is a plane with a uniform, positive surface

charge density and is a small sphere with a like charge,
 161

having mass and charge . What angle will be

formed between plane and the thread which the sphere is suspended
from?

(Ans: ).

Fig. 2-15: Small electric sphere suspended from a thread.

Hint: See formula (*) in Example 2-5-4.

2-5-14) Two point charges in air at a distance from each other

interact with a force . At what distance from each other should these charges
be placed in oil ( ) to obtain the same force of interaction?

(Ans: ).

SUPPLEMENTARY PROBLEMS

2-1) A point charge is located at in free space, and

is another point charge located at , (the coordinates
are expressed in ). Find the vector force exerted on by .

(Ans: ).

2-2) In Problem 2-1 find the coordinates of a point at which a third

charge experiences no force.

2-3) The volume charge density in a region in space is given by the

expression . (a) Find the charge contained within an
 162

infinitesimal volume centered at the point

. (b) What is the total charge contained within the volume
bounded by the surfaces ?

(Ans: a) ), b) ).

2-4) The portion of the axis carries a nonuniform

line charge density and is zero everywhere else. Find the total
charge contained in the segment .

2-5) A surface charge density is distributed in the

interior of the circle and is zero everywhere else. Find
the total charge contained in the circle, (Ans: ).

2-6) In Problem 2-5, a point charge is placed at the point

. Find the vector force exerted on by the charge distribution.

Hint: Since the point is located far away from the charge distribution, we
may safely treat the charge on the circle as a point charge , placed at the
origin, and then apply Coulomb’s Law to find the force.

2-7) The infinite plane carries a uniform charge density

. Find the vector force on a point charge placed at
, (Ans: ).

2-8) The magnitude of the force between two charged particles, in free
space, is . What will the magnitude be, if the system of the two
charges is embedded in a homogeneous dielectric with relative permittivity
?

2-9) The infinite axis carries a uniform line charge density

. Find the vector force on a point charge placed (a) at
, and (b) at .

(Ans: (a) , (b) ).

2-10) In Problem 2-9, a second point charge is placed at

. Find the vector force exerted on at
the points and , (as defined in the previous problem).
 163

2-11) The infinite plane carries a uniform charge density

while the infinite axis carries a uniform line charge density
. Find the vector force on a point charge placed at
the point .

(Ans: ).

2-12) In Problem 2-11, find the force on provided that a second point
charge is placed at the point
.

2-13) The portion of the axis carries a line

charge density and is zero everywhere else. Find the
vector force on a point charge placed at the point .

(Ans: ).

2-14) In Problem 2-13 find the vector force on , provided that a second
point charge is placed at the point .
 164

CHAPTER 3: THE ELECTRIC FIELD

3-1) Electric Fields - Test Charge

Electric field, in general, is the space within which forces are exerted on
electric charges, even if these charges are at rest. To investigate if some space
is an electric field, we place a small charge (called a test charge) within the
space and observe if forces are exerted on this charge.

Since forces on electric charges are due to the presence of other electric
charges, (Coulomb’s interaction), we are thus led to the inevitable conclusion
that the space around electric charges is an electric field. In other words,
electric charges equip their surrounding space with a characteristic property,
which is to exert forces on any other charge introduced in the space.

3-2) The Electric Field of a Given Charge Distribution

Let us consider a charge distribution, existing within an infinite,

homogeneous dielectric space. In the most general case, this charge
distribution may consists of discrete point charges and
continuous charge distributions, (line, surface and volume distributions), as
shown in Fig. 3-1.

Fig. 3-1: Discrete and continuous charge distributions.

Let a small test charge be now placed within the electric field (generated
by the given charge distribution) at a point .
 165

The total electric force on , is (by virtue of Coulomb’s law and the
superposition principle),

where and the vector designates the vector from the position
of the charge towards the point , where the test charge is situated. As we
have already pointed out in Chapter 2, the force is proportional to , and
this means that does not depend on , i.e. the ratio is constant and
depends solely on the charge distribution which generates the electric field.

We may thus define the electric field at the point as follows:

The physical meaning of the limit as , is that the test charge should
be very small, so that the presence of at the point will not alter the original
charge distribution (which produces ). Recall that due to the action-reaction
law, also exerts a force on the charges of the original distribution, and if
were not assumed to be very small, in magnitude, then, the forces it would
exert on the charge distribution might change and redistribute the original
charge distribution, and thus change the field of the original distribution.

From equations (3-2-2) and (3-2-1) we obtain the following expression for
the electric field :

Equation (3-2-3) shows that at every point in space there corresponds a

unique value of , i.e. is a vector function of the position. If
we may write, . The magnitude of
the vector is usually called the electric field intensity.
 166

The electric field is measured in , as obtained from (3-2-2), (Force

( ) per Charge ( )). As we shall show in Chapter 5, is equivalent to
, where is the unit for the electric potential, so is actually
measured in , (Volts per meter).

Assuming that at a given point is known, then if we place a test charge

at the point , the force exerted on shall be, (see equation (3-2-2)),

The direction of obviously depends on .

Fig. 3-2: Force on a test charge .

Uniform electric fields: As we have already mentioned, the electric field

varies from point to point in space, i.e. is a function of the spatial coordinates
. In cases where the vector is the same at all points of the field, the
field is called uniform, ( is a constant vector).

3-3) Calculation of the Electric Field for Some Important Charge

Distributions

All the information in calculating the electric field is contained in equation

(3-2-3). Of course, this formula covers the most general case. In practice, for
arbitrary charge distributions, the integrals involved are very difficult (or even
impossible) to be evaluated analytically. Hereunder we calculate the electric
fields due to some rather simple, but important, charge distributions.
 167

a) The electric field due to a point charge , (Coulomb’s field).

Fig. 3-3: Electric field of a point charge.

Let us assume that we have just one point charge , positioned at the origin
of a spherical coordinate system, as shown in Fig. 3-3. In this case (the
position vector in spherical coordinates) and

If , the electric field points radially outwards; if the electric

field points radially inwards.

The magnitude of the electric field at some distance from the point
charge is (as obtained from (3-3-1)),

The magnitude remains constant over the surface of a

sphere with radius . However, as a vector, is not constant, since its
direction changes from point to point on the spherical surface, always being
perpendicular to the spherical surface.
 168

b) The electric field of an infinite, uniform line charge distribution .

Fig. 3-4: The electric field of an infinite, uniform, line charge distribution.

We assume that a uniform line charge density extends along the axis,
in a cylindrical coordinate system, from to , as shown in
Figure 3-4. If at a point we place a test charge , then the force exerted on
from the given line distribution is given by formula (****) in Example 2-5-2,
i.e.

The electric field, as shown in (3-3-3) at a point is inversely proportional

to the distance of from the axis. If , is in the direction of ,
while if , is in the direction of .
 169

c) The electric field of an infinite, uniform planar surface charge

distribution .

The force on a test charge due to an infinite sheet of charge having a

uniform density has been calculated in Example 2-5-4:

Fig. 3-5: The electric field of an infinite, uniform, sheet of charge.

In Ex. 2-5-4, the infinite sheet of charge was assumed to be the plane,
and in that sense, the quantity appearing in equation (*) is a unit
vector, perpendicular to the sheet of charge, i.e.

The important thing to notice is that is always perpendicular to the planar

sheet of charge. So if we define a unit vector , perpendicular to the planar
sheet of charge and pointing away from it, then in general, we may write,
 170

We notice that the electric field is constant in magnitude and direction,

i.e. the field produced by an infinite planar charge with constant charge
density is a uniform field. This is a startling result, since it shows that the
field close to the plane is identical to the field far away from the plane.

d) The electric field of a circular ring of charge with a uniform line charge
density .

Fig. 3-6: The electric field of a uniform, circular ring of charge.

Let an electric charge be distributed, uniformly, over the circular ring of

radius , lying on the plane, as shown in Fig. 3-6. Then the electric field
at a point is given by the formula

For a proof, see Example 3-3-4.

e) The electric field of a spherical surface of radius which carries a total

charge , uniformly distributed over its surface.

Let a total charge be uniformly distributed over the surface of a sphere of

radius , as shown in Fig. 3-7. Then, the electric field at the point
 171

determined by its radius vector , in spherical coordinates, is given by the

formula,

Fig. 3-7: The electric field of a uniformly charged spherical surface.

For a proof see Example 3-3-5.

Remark: From equation (3-3-6) we see that the field of a spherical surface
of radius , which carries a total charge , uniformly distributed over its
surface, is identical to the field which would result, should the charge were
positioned at the center of the sphere, (for ) and is everywhere zero,
inside the sphere.

In problems with cylindrical symmetry (like an infinite straight line with

uniform charge density) or spherical symmetry (like the spherical surface with
uniform charge density), it is possible to avoid the evaluations of the
complicated integrals to determine the electric field ; instead, in such cases,
the evaluation of is much simpler if we make use of Gauss’s Law, which will
be studied in details in Chapter 4.
 172

Example 3-3-1: Three point charges ,

in free space, are placed at the points respectively.
Find the electric field at the point .

Solution

Let be the vectors from the position of the charges

respectively, to the point , where the electric field is to be found. Then,

The magnitudes of these vectors are:

The electric field at the point is, (see the first term in eq. (3-2-3)),

Example 3-3-2: Two parallel, infinite planes, and

are charged with uniform surface charge densities and
, respectively. Find the electric field : a) in the space
between the two planes, , b) in the space and c) in the space
, (assume free space).
 173

Solution

Fig. 3-8: The electric field of two uniformly charged, infinite planes.

a) In space the electric field is found by superposition of the

fields due to and , i.e. (see equation (3-3-4)),

b) In space , similarly, the superposition of the two fields yields,

c) In space ,

In summary:
 174

Example 3-3-3: A sheet of charge , is present at the plane

in free space and a line charge is located on the line
. Find the electric field at the origin.

Solution

Fig. 3-9: Calculation of the electric field due to a surface and a line
distribution.

Let and be the electric fields due to the surface and the line charges,
respectively. Then, at the origin,

The total field results from the superposition of and , i.e.

 175

Example 3-3-4: Show formula (3-3-5).

Solution

Fig. 3-10: The electric field of a circular ring uniformly charged.

Since the charge is uniformly distributed over the circular ring, the line
charge density is .

The electric field at is

and the total electric field at is found by integration, i.e.

Term A:

since we have pair wise cancelation, (see Fig. 3-10).

 176

Term B:

and from equation (*) we obtain,

Example 3-3-5: Show formula (3-3-6).

Solution

Due to the spherical symmetry involved, , where is the unit

vector in spherical coordinates. In Fig. 3-11, at the point , (the point where
the field is to be found), the unit vector coincides with .

Fig. 3-11: The electric field of a spherical, uniform surface distribution.

Since the charge is uniformly distributed over the surface of the sphere of
radius , the surface charge density is .
 177

Let us consider the spherical ring, as shown in Fig. 3-11. The charge
existing on the ring is,

We may think of this spherical ring, as a circular ring, of radius ,

and according to formula (3-3-5) the electric field at , due to the
charge on the circular ring, is

From Fig. 3-11, we obtain,

The second equation in (***) is obtained if we apply the law of cosines in

triangle . Substituting equations (***) into (**) and integrating w.r.t. ,
from to , we get,

Integral is found to be, (see Problem 3-3-17).

and finally, from equation (****) we obtain,

or, since ,
 178

Due to the spherical symmetry, in general, the electric field at any

point is

where is the unit vector in spherical coordinates, at the point .

PROBLEMS

3-3-1) A point charge is located at in free space. a)

Find at the point , and b) Find the locus of all points at
which .

(Ans:
).

3-3-2) Three point charges are

located at the points respectively. Find the
electric field at the point .

Hint: See Example 3-3-1.

3-3-3) Eight identical point charges are located at the corners of a cube in
free space. Show that the electric field at the center of the cube is zero.

3-3-4) A particle with mass is charged to a net charge

and placed in an electric field. If the acceleration of the particle is
, find the electric field .

Hint: .
 179

3-3-5) Express in Cartesian coordinates the electric field of an infinite

straight line coinciding with the axis, charged with a uniform charge density
.

(Ans: ).

3-3-6) A point particle of mass and charge is suspended in equilibrium

above an infinite plane charged with a uniform charge density .
Determine the charge of the particle, assuming free space conditions.

3-3-7) An infinite sheet of uniform charge density is situated coincident

with the plane. The sheet has a hole of radius centered at the origin.
Determine the electric field , (assume free space, ).

(Ans: ).

3-3-8) The square surface is charged with a

surface charge density . Determine the electric field at the
point .

3-3-9) How many electrons should a spherical conductor, accept, so that the
magnitude of the electric field at its surface be ? Assume
that the radius of the conductor is . Which is the total mass of the
electrons?

(Ans: electrons, total mass ).

3-3-10) Two infinite straight lines in free space charged with a uniform
each, are parallel to the axis at the points .
Find .

3-3-11) The plane is uniformly charged to

. Find the electric field at the origin, assuming free space conditions.

(Ans: ).
 180

3-3-12) The square surface is charged

with a surface charge density . Find the electric
field .

3-3-13) Eight identical particles each, are placed at the corners

of a cube of side . Find the magnitude of the Coulomb force at each
particle, assuming free space conditions.

(Ans: ).

3-3-14) Four identical particles each, are placed at the vertices

of a square of side situated in the plane and centered at the
origin. Find .

3-3-15) An electron is placed at a height above an infinite plane with a

uniform, positive, charge distribution . The electron is attracted
towards the plane. Find the time it takes the electron to reach the plane,
(assume that the electron has zero initial velocity).

(Ans: .

3-3-16) An electron, located at , is at rest, (at ), near to a heavy

positive point charge , placed at . Find the speed of the electron as a
function of , for .

Fig. 3-12: An electron close to a proton.

Hint: Since the positive ion is much heavier than the electron, assume that
it stays at rest, and the electron is attracted towards . To obtain the
differential equation for , note that , etc.
 181

3-3-17) In Example 3-3-5 evaluate integral in formula (****).

Hint: Make the substitution ,

etc.

3-3-18) Two point charges and are placed at the

points and respectively, while the axis is charged
with a uniform line charge density . Find and
.

3-3-19) Two infinite planar sheets of charge at and have

uniform charge densities and ,
respectively. Find the electric field everywhere.

(Ans: ).

Hint: See Example 3-3-2.

3-3-20) An infinite plane at is charged with a uniform surface

charge density . The axis is charged with a uniform line
charge density . Find and
.

3-3-21) Volume charge density is given as for

and for ; a) Find the total electric charge inside
the sphere of radius , and b) Find the electric field
. Since , we may safely consider the charge within
the sphere as a point charge situated at the origin, and then apply formula (3-
3-1) to find .

(Ans: ).

3-4) Field Lines of the Electric Field

A vector line (or field line) of an electric field , is a curve in space at every
point of which the direction of the tangent of the curve coincides with the
direction of the electric field .
 182

Fig. 3-13: Field lines.

A positive test charge placed at the point and free to move would
accelerate in the direction of the field line passing through . The positive
direction of a field line is defined to be the direction which coincides with the
direction of the field. For example, the positive direction of the field line
passing through the point , is from to . We designate the positive
direction by a small arrowhead on the field line. A negative test charge placed
at and free to move, would accelerate along the field line passing from ,
but in the negative direction.

To find the equation of the field line, we note that if

is the differential tangent vector of a field line at an arbitrary
point of the line, then the vectors and (at the given point) are parallel
and consequently their cross product must be zero, i.e.

This is the vector differential equation of the field lines.

Recall the following expressions of in rectangular, cylindrical and

spherical coordinates:
 183

Let us assume that in rectangular coordinates . Then

equation (3-4-1) implies that

Equation (3-4-3) is the analytic form of the differential equation of the field
lines, equivalent to equation (3-4-1).

Similar expressions are obtained in the other two coordinate systems.

Example 3-4-1: Find the equation of the field line that passes through the
point , for the field .

Solution

The given field is a plane field, i.e. lies on the plane. In this case,
equation (3-4-3) becomes:

where is the arbitrary constant of integration. Since this field line must
pass through , we have,
 184

Example 3-4-2: Consider the electric field ,

expressed in cylindrical coordinates. Show that equation (3-4-1) assumes the
form , and then find the equation of the field line passing
through for the field .

Solution

In deriving (*) we have used the fact that .

For the given electric field, and , and

equation (*) implies,

This is, in general, the equation of the field lines in polar coordinates.

The particular field line that passes through is obtained

from (*) as follows:
 185

PROBLEMS

3-4-1) Let be a plane electric field. Find the field

line passing through , (Ans: ).

3-4-2) Find the field line that passes through the point for the
electric field .

3-4-3) Find the field line that passes through the point for the
electric field , (Ans: ).

3-4-4) Find the equation of the field lines of the electric field generated by a
point charge , (see equation (3-3-1)).

(Ans: The intersection of the coordinate surfaces, and

. The intersections of these two coordinate surfaces are
rays issued from the origin, see Fig. 1-21).

SUPPLEMENTARY PROBLEMS

3-1) Four positively charged particles with charge and mass each, are
fixed at the four corners of a square of side , centered at the origin (in the
plane). A negatively charged particle with charge and mass is placed
at the center of the square and is free to move. (a) Show that the negative
charge is at equilibrium at this point, and then show that this equilibrium is
unstable for small displacements in the plane , (b) Show that the equilibrium
is stable for small displacements perpendicular to the plane , and (c) Show

that the period of small vertical oscillations is .

Remark: Charged particles, in general, cannot be held in stable equilibrium,

by electrostatic forces alone. This important remark follows from the
Earnshaw’s Theorem, studied in Chapter 5, Section 5-9.

3-2) Find the electric field at the origin, caused by the three point charges
in free space, placed at the points
 186

and respectively, (distances are in

).

3-3) The plane carries a uniform charge density and

the plane carries a uniform charge density . Find the
electric field at the points and , (distances
are expressed in ).

(Ans: ).

3-4) In Problem 3-3, an additional point charge is placed at the

point . Find the electric field at the same points and .

3-5) The plane carries a uniform charge density ,

while an infinite line perpendicular to the plane at
carries a uniform line charge density . Find the electric field at
the points and .

(Ans: ).

3-6) Let be a plane electric field. Find the

equation of the field line passing through the point .

(Ans: ).

3-7) Electric charge is distributed on the plane with a uniform surface

charge density . A point is at a distance above the plane. As
shown (equation (3-3-4)), the electric field at is . Show that one
half of the field arises from those points of the plane that are less than
from the point .

3-8) A point charge is placed at the origin and a negative charge

is placed at the point . Find the electric field at a
point which is at a distance of from the positive charge and
from the negative charge.
 187

3-9) Two identical small spheres are suspended on two threads of length
, so that their surfaces are in conduct. A charge is
supplied to the spheres, making then to repel each other to an angle of .
Find the mass of each sphere, (Ans: ).

3-10) In Problem 3-9, what is the tension of the threads ?

3-11) In Example 3-3-4, (the field of a circular ring, uniformly charged) at

what the magnitude of the electric field becomes maximum ? What is the
maximum magnitude of the electric field ?

(Ans: ).

3-12) A ring made of a wire of radius caries a uniform charge

density . Find the electric field on the axis of the ring at
points lying a distance and from the center of
the ring. At what the magnitude becomes maximum and what is this
maximum magnitude ?

3-13) The infinite plane carries a uniform charge density

, while an infinite line perpendicular to the plabe carries a
uniform charge density . Find the force per unit meter acting
on the line charge, (Ans: ).

3-14) Two infinite lines each carrying are spaced

apart. Find the magnitude of the electric field at a point from each line.

3-15) A negative charge is placed at the center of a square each vertex

of which contains a positive charge . Find the magnitude of the negative
charge if the net force on each positive charge is zero.

(Ans: ).
 188

CHAPTER 4: THE TWO FUNDAMENTAL LAWS OF STATIC ELECTRIC

FIELDS

4-1) Introduction

In the preceding chapters we have developed a systematic method to

calculate the electric field generated by a static charge distribution, either
discrete or/and continuous. In this chapter we shall study the two
fundamental laws which hold true for any static electric field . These two,
fundamental laws, are the following:

a) The static electric field is irrotational:

b) Gauss’s Law:

In the rest of this chapter we shall derive, rigorously, equations (4-1-1) and
(4-1-2) and study some of their most important applications.

Some general remarks:

1) In each equation, the so called “integral form” is equivalent to the

corresponding “point form”. One can go from the point form to the integral
form, and vice versa, by a judicious application of proper theorems of vector
analysis.

2) The vector appearing in equation (4-1-2) is the “electric displacement”

or just the “ vector” which is defined as , being the permittivity of
the medium, (see Section 2-4).
 189

3) While equation (4-1-1) holds true strictly for static fields, Gauss’s law in
equation (4-1-2) is of general validity; it holds true for static and time varying
fields as well, and is actually one of the four Maxwell’s equations. It is a
general law applying to all Electromagnetic fields, not just to the static ones.

We now proceed to the proof of equation (4-1-1).

4-2) The Static Electric Field is Irrotational: (

Fig. 4-1: Proof that static electric fields are irrotational.

Let us consider a point charge situated at the origin, within a

homogeneous dielectric of permittivity , as shown in Fig. 4-1. The electric field
at a point , determined by its position vector , is given by equation (3-3-1),
i.e.

Let us consider the line integral (of the second type) of the electric field
along a curve in space, from an arbitrary point to another arbitrary point
, as shown in Fig. 4-1, i.e. let us consider the integral .

The electric field is given by equation (*) while the differential length in
spherical coordinates is, , (see
equation (3-4-2)), so we have:
 190

since .

Equation (4-2-1) shows that the line integral depends only on

and , i.e. the end points and of the curve and is independent from
the particular path taken to connect the two points and . If we now
consider a closed, smooth curve , then the two end points coincide, ( )
and hence , and equation (4-2-1) implies,

Equation (4-2-2) was proved for a point charge . However, it still holds
true for arbitrary, static charge distributions. To justify this assertion, recall
that by virtue of the superposition principle, the electric field in such a case
(many charges) is the vector sum of their individual fields, and since for each
individual field (4-2-2) holds true, it remains true for their vector sum, i.e. for
the total electric field.

According to the terminology introduced in section 1-7, (see eq. (1-7-10) we

may say that the circulation of any static electric field around any closed
curve is equal to zero. We have thus shown the first equation in (4-1-1).

To prove the second equation ( ) in (4-1-1) we may transform the

line integral to a surface integral with the aid of Stokes’ theorem, (see section
1-9, equation (1-9-2)).

Let us for definiteness consider a surface bounded by a closed, smooth

curve as shown in Fig. 4-2.
 191

Fig. 4-2: Proof of the second formula in Equation (4-1-1).

Application of Stokes’ theorem yields:

and since for any closed curve , equation (**) implies that
. This equation must hold true, for an arbitrary surface
, bounded by an arbitrary closed curve , and in order for this to be true,
the integrand must be identically equal to zero, i.e. , and this
completes the proof.

In the three most commonly used coordinate systems, rectangular,

cylindrical and spherical, assumes the following form, (sect. 1-11):
 192

According to the terminology introduced in section 1-12, the static electric

field is an irrotational field (or a conservative field).

Also, since the static electric field is irrotational, there exists a scalar
function such that , (see section 1-12). This scalar
function is of extreme importance in static fields, it is call the potential of the
electric field, and it will be studied in details, in Chapter 5.

Notice: Since the line integral between two points and does
not depend on the path connecting the points and , we may just write
, without specifying the particular form of the path. In practice, we
choose this path from to which makes our calculations as simple as
possible.

Example 4-2-1: Which one of the following two fields is an impossible static
electric field? a) , b) .

Solution

A static electric field is impossible to exist if its curl is not identically

equal to zero, i.e. if then the static electric field cannot exist.
 193

Case a:

Case b:

and this shows that this electric field is impossible exist.

Example 4-2-2: Consider the static field . a) Verify that

, b) Evaluate the line integral along: 1) the path ACB
(see Fig. 4-3), 2) the straight line joining A and B and 3) along the parabola
and thus verify that the integral is path independent.

Solution
 194

Fig. 4-3: Evaluation of the integral .

1) Along the path ACB: Along , while along

.

and by virtue of (**) and (***), equation (*) yields, .

2) Along the straight line AB:

.
 195

3) Along the parabola :

.

The answer obtained is the same in all three cases. The value of the integral
does not depend on the path taken from A to B.

PROBLEMS

4-2-1) Consider the static electric field and verify that

. Then find the integral along the straight line joining
the points A and B, (distances are given in ), (Ans: ).

4-2-2) In Problem 4-2-1 evaluate the integral along the path

and verify that you obtain the same answer.

4-2-3) Consider the Coulomb field of a point charge situated at the

origin (formula (3-3-1)) and verify that except at the origin ,
where the field is not even defined, (work in spherical coordinates, see
equation (4-2-5)).

4-2-4) The axis is charged with a uniform line charge density

. Find the integral from to
, (assume free space conditions).
 196

Hint: The electric field is given by equation (3-3-3) and the differential
length in cylindrical coordinates is .

4-2-5) For the electric field of an infinite, uniform line charge distribution,
show that except on the axis , where the field is not even
defined.

4-3) The Electric Displacement Vector

Let be the electric field generated by a given charge distribution, inside a

homogeneous dielectric with permittivity . We may define the so called
“electric displacement ” or the “vector field ” or just “the electric flux
density ”, as

The term “electric displacement” is due to J. C. Maxwell, who first

introduced the vector in Electromagnetism.

The vector of a point charge is easily obtained from equation (3-3-1),

i.e.

Similarly, the field due to a general charge distribution (discrete and/or

continuous) is obtained easily from the definition of , (equation (4-3-1)), and
equation (3-2-3), i.e.

In free space (vacuum), and hence, .

 197

As obtained from (4-3-2) the unit of is , i.e. “charge per unit

area”.

Remark: As follows from the definition (4-3-1), in general, the vectors and
are collinear (parallel) and point in the same direction. This is true, provided
that the medium is isotropic. A material is isotropic if it has the same
properties in every direction. A material which is not isotropic is called
anisotropic.

Anisotropic materials cannot be described by a simple permittivity . The

vectors and are no longer parallel. Instead, it is found that in anisotropic
materials, each component of is a function of every component of . In this
case, assuming that and , we
may write,

The matrix

is called the permittivity matrix or the permittivity tensor. In terms of this

matrix, we may write

From (4-3-4) it follows that

and this clearly shows that the component , for example, depends on the
components and and .
 198

It can be shown that the permittivity matrix is symmetric, i.e.

.

We further note that the elements of do depend on the selection of the

coordinate axes in the anisotropic material. It can be shown that there exists a
system of coordinate axes, called “the principal axes” in which all the off the
main diagonal elements in the permittivity matrix are zero, i.e.
. In this system of axes, equations (4-3-7) are simplified to the
following,

4-4) Gauss’s Law, ( ), Electric Flux

a) Solid angles: Recall that the radian is the measure of an angle that,
when drawn as a central angle, subtends an arc whose length is equal to the
length of the radius of the circle.

Fig. 4-4: Measure of solid angle-Steradians.

In the first Figure, the central angle is equal to , if the length of the
arc is equal to the radius of the circle. Given that the length of the
circumference of the circle is , it follows that the circle has radians. It
also follows that the length of an arc corresponding to a central angle ,
expressed in radians, is .

Angles are also measured in degrees ( ). A circle has , and this shows
that correspond to . This equivalency is important when we
have to convert to degrees and vice versa.
 199

In the three dimensional space, we deal with solid angles. A solid angle is
made up of all the lines from a closed curve in space, meeting at a fixed point
, termed the vertex of the solid angle, as depicted in Fig. 4-5. Solid angles are
measured in Steradians (symbolized ).

The definition of the Steradian is analogous to the definition of the Radians

for plane angles. Let us consider the sphere of radius in Fig. 4-4. The
Steradian is defined as the solid angle at the center of the sphere,
subtending a portion on the surface equal in area to the square of the radius
of the sphere, i.e. equal to . Given that the area of the surface of a sphere
of radius is , it follows that a sphere has Steradians.

Fig. 4-5: Differential solid angle .

From the very definition of the solid angle, it follows that the surface area
of a portion on the surface of the sphere with radius , corresponding to a
central angle , expressed in Steradians, is .

Let us now consider a surface element (a differential surface) whose

position is determined by the radius vector , relative to the origin . The
totality of rays connecting all the points on the boundary of with the origin
, form a cone, as shown in Fig. 4-5. Let be the area of that portion of a
sphere with center and radius which is cut out by the cone. Then the
solid angle subtended by at , is . However,
 200

Formula (4-4-1) is fundamental and appears quite often in various problems

in Engineering and Physics. By integration we can find the total solid angle
subtended by a surface at . For instance, in the first figure in Fig. 4-5,

We thus see that solid angles are defined in terms of surface integrals of
the second type.

As an immediate consequence of the definitions and formulas in the

foregoing discussion, it follows that

assuming that the origin lies inside the closed surface .

However, if the origin lies outside the closed surface , then

, (for a proof see Problem 4-4-13).
In summary: If is the position vector of the points of a closed surface
relative to a point , then:

b) Electric flux: The flux of the vector field over some surface is
defined by the surface integral, (see equation (1-7-14)).
 201

c) Let us now consider one point charge positioned at the origin and let
us calculate the flux of the field over a closed surface enclosing the
charge.

Fig. 4-6: Flux of the electric displacement through a closed surface.

Since the field of a point charge is given by equation (4-3-2) the flux of
the vector field over the closed surface is

If we now assume that the closed surface encloses two charges and
, then the total field , by virtue of the principle of
superposition, and the flux of over shall be
 202

Finally, let us consider another example, where we have three charges

, two charges, say and enclosed by the closed surface and
the third charge being outside of the closed surface . Then
, and the total flux over will be

The generalization is obvious and we are thus led to the following general
conclusion: The flux of the vector over every closed surface is equal to the
total electric charge enclosed by the surface. This is the famous Gauss’s law,
which in mathematical terms is expressed as follows:

Let us now assume that we have a continuous volume charge distribution,

with a volume charge density ).

Fig. 4-7: Gauss’s law for a continuous charge distribution.

 203

We imagine a closed surface within the continuous charge distribution,

which encloses a volume . The surface and hence the enclosed volume
are arbitrary. The total charge enclosed by , is .

Application of Gauss’s law, over yields,

and transforming the surface integral into a triple integral, by virtue of the
divergence theorem, (see equation (1-9-1)), we obtain,

and since this equation holds true for an arbitrary closed surface
enclosing an arbitrary volume , the integrand must be identically
equal to zero, i.e.

As we have already mentioned, formula (4-4-7) holds true even in cases

where the fields are time varying, and not only for static fields. Eq. (4-4-7) is
actually one of the four, famous, Maxwell’s equations.

In the three, most commonly used coordinate systems, i.e. rectangular,

cylindrical and spherical, equation (4-4-7) assumes the following form (section
1-11):
 204

Example 4-4-1: Let in free space. Determine

the total electric charge within the cube
and the electric field .

Solution

The volume charge density is obtained from Gauss’s law in point form
(equation (4-4-7), i.e.

The total charge inside the volume is

The electric field is

 205

Example 4-4-2: Let . a) Find the charge density

, and b) find the total charge within the unit sphere ( ), centered at
the origin.

Solution

a) Since , and ,

b) The total charge within the unit sphere is

Alternatively, we may find the charge within the unit sphere using Gauss’s
law in integral form. On the surface of the unit sphere ,
, and .

Example 4-4-3: Within the cylindrical region the vector

( ). a) Find the volume charge density at , b) How much
electric flux is leaves the cylinder ?, c) How much charge
is contained in the cylinder ?
 206

Solution

At the point

The charge contained in the cylinder is

Alternatively, we may compute the charge contained in the cylinder using

Gauss’s law in integral form.

Fig. 4-8: Total flux through a cylinder.

Surface : , and since

, (recall ), the first integral in (*) is zero.

Surface : For similar reasons ( ), the second integral in (*) is zero.

 207

Surface : On the cylindrical surface , and

the third integral in (*) becomes,

and this shows that the total flux leaving the cylinder is .

PROBLEMS

4-4-1) If find the volume charge density .

(Ans: ).

4-4-2) If for and for , find the

volume charge density in both regions.

4-4-3) In cylindrical coordinates the field is given by the following

expression

Find the volume charge density in both regions.

(Ans: and ).

4-4-4) In Problem 4-4-3 find the flux leaving the cylinder .

4-4-5) A circular disk of radius and surface charge density

is contained inside a closed surface . Find the total flux
leaving the surface, (Ans: ).

4-4-6) The volume charge density in space is given by the following

expression
 208

Find the total electric flux leaving the surface of a sphere of radius
centered at the origin.

4-4-7) If , find a) the volume charge

density at the point , b) the flux through a small surface by
on the cylindrical surface at the point .

(Ans: ).

Hint: Convert in cylindrical coordinates; see Section 1-4).

4-4-8) In spherical coordinates for and is

zero everywhere else. a) Find the charge density at the point
, b) Find the flux through the spherical surface , c) Find
the flux through the spherical surface , (assume free space conditions).

Hint: .

4-4-9) Given that , a) Show that

, b) Find the charge density and c) Find the flux through the
spherical surface .

(Ans: ).

4-4-10) An electric field exists in free

space. a) Show that and b) Find the charge density everywhere.

4-4-11) If find the flux through a unit sphere centered at

the origin, (Ans: ).

4-4-12) If find the total charge

within the volume .
 209

4-4-13) Show formula (4-4-4) for the case where the origin lies outside
the closed surface .

4-5) Applications of Gauss’s Law: Gaussian Surfaces

One of the most important applications of Gauss’s law, in integral form, is

the computation of the electric fields in terms of the source charges.
Symmetry in the charge distribution (which generates the electric fields) is
the crucial factor in this application of Gauss’s law. In such cases, Gauss’s law
provides an easy method to compute the electric field . Recall that Gauss’s
law in integral form states that

Strictly speaking, for a known equation (*) is an integral

equation for the unknown . The solution for is simplified if we can choose
a closed surface , termed “Gaussian surface”, which satisfies the following
two conditions:

a) is either tangential to the Gaussian surface (in which case ) or

is perpendicular to the Gaussian surface (in which case ),

b) On that portion of the Gaussian surface where is perpendicular to the

Gaussian surface, remains constant, and therefore can be pulled out
of the integral in (*), (this is the key idea).

The proper choice of a Gaussian surface depends on the symmetry of the

charge distribution of the problem at hand. In very general terms, if the charge
distribution exhibits spherical symmetry we choose as Gaussian surfaces
concentric spherical surfaces; for cylindrical symmetries we choose as
Gaussian surfaces coaxial cylindrical surfaces; for plane symmetry we choose
as Gaussian surfaces appropriate “pillboxes”, (see Problem 4-5-2).

Remark: As we have already seen, equation (4-3-3) may be used to

compute the electric field , (and then compute ), by summing or
 210

integrating over the source charges. However, these integrations are quite
complicated. In cases where symmetry allows, Gauss’s law provides an
alternative method of computation, by far, easiest and quickest.

The method is best illustrated by means of a few examples.

Example 4-5-1: Find the electric field of an infinite, uniform line charge
distribution , extending from to , in free space.

Solution

Fig. 4-9: Electric field of an infinite, uniform line charge distribution.

Due to the cylindrical symmetry involved, the vector at a point at a

distance from the axis, points in the direction of the local unit vector (at
the point ) and its magnitude depends only on , the distance of from the
axis. In other words, at the arbitrary point the vector .

Let us choose as a Gaussian surface a cylindrical surface of height , coaxial

with the axis, as shown in Fig. 4-9.

Application of Gauss’s law implies:

 211

since the total charge enclosed by the closed cylindrical surface is .

On the top and bottom surfaces and respectively, the dot product at
each point is zero, since is perpendicular to the unit normal .

On the cylindrical surface , of radius , since , the dot product

We note that on the cylindrical surface of radius the magnitude of

the vector remains constant, so we can take it out of the integral, and finally
equation (*) yields,

As a vector,

Having found the field follows readily,

Note that the expression for is identical to the expression found by

integration over the line charge distribution (see equation (3-3-3)).

Example 4-5-2: Find the electric field produced by the following volume
charge distribution in free space,
 212

Solution

Due to the spherical symmetry involved, the vector points in the radial
direction and its magnitude depends only on , i.e. .

Let, in this problem, consider as Gaussian surfaces spherical surfaces

concentric with the charge distribution, as shown in Fig. 4-10.

Fig. 4-10: Gaussian surfaces for a symmetric, spherical charge distribution.

Region I: , Gaussian Surface .

Total charge enclosed by is and application of Gauss’s

law yields,

and since on the surface , remains constant

over , it can be pulled out of the integral, and equation (*) implies,
 213

Region II: , Gaussian Surface .

Reasoning similarly, and noting that the total charge enclosed by is

, we have,

In summary:

The field , i.e.

Remark 1: In principle, the electric field of the given charge distribution

can be computed by proper integration over the volume charge sources, (see
equation (3-2-3)). However, this approach involves a quite complicated
integration. On the other hand, application of Gauss’s law yields the field
rather easily, and from then the field follows readily.
 214

The plot of the magnitude of versus , (the distance of the point

where the field is to be evaluated from the center of the charge distribution),
is shown in Fig. 4-11.

Fig. 4-11: Plot of versus .

We see that inside the region of the charge the electric field increases
linearly with , while outside of the charge distribution the electric field falls
off as . As a matter of fact, the electric field in region II, (outside of the
spherical charge distribution), as obtained from the second equation in (*****)
is

Notice that the field outside the sphere of radius is the same as it would
have been if all the charge were concentrated at the origin,
(compare with equation (3-3-1)).

Remark 2: Usually, we think of an electron as a material point carrying a

charge . In this model, the radius of the electron is , and as an inevitable
conclusion, the electric field close to the electron (as ), as obtained from
equation (3-3-1), blows up (goes to infinity). This, physically, makes no sense.
We cannot have an infinite electric field close to the electron.
 215

We may remedy this problem, if instead, we consider the electron to be a

small sphere, with an extremely small, but still finite radius , and assume
that the charge is uniformly distributed throughout the volume of the sphere.
Then equation (******) shows that the field at the center of the electron is
zero, increases linearly with the distance, up to the boundary , and then
falls off as . This is a more realistic, classical model of the electron. Of
course, this model solves one problem, but at the same time creates a new
one; assuming the electron to be a small sphere of negative charge, uniformly
distributed inside the sphere of radius , the mutual repulsion between like
(negative) charges should make the charges to fly apart, something which
evidently is not the case. So what exactly happens here? We have to admit that
this is a rather embarrassing situation, with no clear explanation, whatsoever.

Example 4-5-3: An infinite cylinder coaxial to the axis, of radius , carries a

charge density , in free space, where is a given constant. The
charge outside the cylinder is zero. Find the electric field everywhere.

Solution

Fig. 4-12: Electric field of an infinite cylindrical charge distribution.

Due to the cylindrical symmetry involved, the field will point radially
outward and its magnitude will depend only on , the distance from the axis,
i.e. . We choose as Gaussian surfaces cylindrical surfaces, coaxial to
 216

the axis. Let us consider the Gaussian surface shown in Fig. 4-12. The unit
normal on the top surface is , while on the bottom surface is . In
both cases, the dot product is zero, since . On the lateral surface
of the Gaussian surface, , so which is constant
over the cylindrical surface.

a) Region I: :

b) Region II: :

In this case the Gaussian surface will be similar to the Gaussian surface in
region I, with radius . The total charge enclosed by the Gaussian surface
is

(since for the charge density is zero), and working as in part (a) we
find

Again, in principle, the vector , could be computed by a proper integration

over the charge sources, but the integrations involved would be quite
complicated. On the contrary, Gauss’s law provides an easy and
straightforward method for the computation of .
 217

In summary:

PROBLEMS

4-5-1) Find the electric field produced by a charge , uniformly distributed

over the surface of a sphere of radius , and compare your answer with
formula (3-3-6).

Hint: Take Gaussian surfaces concentric to the sphere.

4-5-2) Find the electric field of an infinite sheet of charge with uniform
surface charge density , and compare your answer with formula (3-
3-4).

Hint: Take a Gaussian surfaces in the form of a “pillbox” as shown in Figure

4-13.

Fig. 4-13: Electric field of an infinite, uniform sheet of charge.

4-5-3) Consider two concentric spherical surfaces with radii and

. A positive charge is uniformly distributed over the inner
 218

spherical surface , while a negative charge – is distributed uniformly over

the outer spherical surface . Find the electric field everywhere, (assume
free space conditions).

Ans: .

4-5-4) The space between two infinite cylindrical surfaces and

, coaxial to the axis, carries a uniform volume charge density
. Everywhere else, . Find the electric field as a function
of the distance from the axis. Also, plot the magnitude versus .

Hint: There are three regions, and .

Consider as Gaussian surfaces cylindrical surfaces coaxial to the axis.

SUPPLEMENTARY PROBLEMS

4-1) In cylindrical coordinates the electric flux density .

Find the volume charge density at the point .

(Ans: ).

4-2) In Problem 4-1 find the total charge within the cylinder
, a) By integration of over the volume of the cylinder and b)
by integration of over the surface of the cylinder, (Ans: ).

4-3) In spherical coordinates the electric flux density is

. Find the total charge inside the unit sphere.

(Ans: ).

4-4) Charge is placed at the center of the unit sphere. Find the
electric flux through the surface .

4-5) For the plane field , find: a) the

general equation of the field lines and b) The equation of the field line passing
through the origin.
 219

(Ans: ).

4-6) In spherical coordinates . Find the volume charge

density and the electric charge within the volume:
.

4-7) Which one of the following field is an impossible static electric field and
why?

a)

b)

c)

(Ans: Field (b) is impossible. Recall that for any static electric field,
must be zero, identically).

4-8) If , in free space, show that

and find the charge density .

4-9) The volume charge density in free space is , Find

the electric field everywhere.

(Ans: ).

4-10) Consider the electric field , in free

space. Show that and find the general form of the field lines.

4-11) In Example 4-5-3, show by direct computation that .

4-12) The electric displacement of a static electric field, in free space, is

. Show that this is a valid static field
(show that ) and find the volume charge density . Next,
consider an infinitesimal volume at the point
and find the charge contained in .
 220

4-13) Four point charges and are enclosed by a smooth,

closed surface . Find the flux of the electric field through the closed surface
, assuming that: a) , and
b) . Can we, in this
problem, use Gauss’s law to compute the electric field ?

(Ans: a) , b) . We cannot use Gauss’s law to compute since there is

no symmetry).

4-14) Consider the volume charge density of Example 4-5-2, and let us
further assume that . If an electron is placed within the sphere of radius
, show that the motion of the electron is a simple harmonic motion about the

center of the sphere with period ,( =mass of electron and

= charge of electron).

4-15) In spherical coordinates, , (in free space). (a) Show that

, (b) Find the total charge contained in the region
using two methods, 1) integrating over the surface of
the volume and 2) application of the divergence theorem, and verify that by
both methods the results obtained are identical.

(Ans: ).

4-16) Find the equation of the field lines corresponding to a charge

placed at and to a charge placed at .

Hint: To solve the differential equation , make the substitution

and .

4-17) The spherical region contains a volume charge density

. Find the electric field everywhere.

(Ans: ,
).
 221

Hint: Consider the three regions, 1) , 2) and 3)

, and apply Gauss’s law.

4-18) For the electric field in Problem 4-17, show that , and
then evaluate the integral , where and
.

4-19) The field (in cylindrical coordinates) of a given charge distribution is

given by the formula

(a) Show that the corresponding electric field is irrotational, i.e.

and (b) Find the charge density everywhere, (Ans: ).

4-20) In Problem 4-19, assume , and find the total

charge contained in the region .

4-21) Two spheres each of radius and carrying uniform charge densities
and – respectively, are partially overlapped, as shown in Fig. 4-14. Show that
the electric field in the region of overlap is uniform, (Ans: ).

Fig. 4-14: The field in the region of overlap.

 222

CHAPTER 5: THE ELECTROSTATIC POTENTIAL

5-1) Introductory Concepts and Definitions

The electric potential is one of the most fundamental, important and useful
concepts in the theory of electromagnetic fields. There are several ways to
introduce the concept of the potential; however, in this book, we shall
introduce the potential as an immediate consequence of the fundamental law
, valid for all static electric fields. Having followed this way of
introducing the potential, we shall then investigate its properties and its
physical significance.

As proved in Chapter 3, the static electric field is an irrotational field,

meaning that . As shown in Section 1-12, any irrotational field can
be expressed as the gradient of a scalar field (function) , (see
equation 1-12-5). We have further shown that the scalar field is not
unique; on the contrary it is determined up to an arbitrary constant.

So, in our case (static electric fields), the fundamental law

implies that there exists a scalar function , called electric
potential , or just for brevity potential, such that the field is obtained as
the gradient of , i.e.

Since is measured in , it follows from (5-1-1) that the potential is

measured in Volts.

The physical meaning of the negative sign in (5-1-1) will become apparent
soon, (actually it shows that the field points from higher towards lower
potentials). The fact that is determined up to an arbitrary constant is
insignificant in the determination of the field , since the gradient of any
constant is zero. Again, the physical significance of the arbitrary constant
hidden in , will be explained shortly, (Section 5-4).

If we can somehow, determine the potential of a given charge

distribution, in terms of the spatial coordinates, then equation (5-1-1)
furnishes a convenient method to find .
 223

In the three most common coordinate systems, rectangular, cylindrical and

spherical, formula assumes the following form, (see Section 1-11):

5-2) Poisson’s Equation

Let us consider an electric field in free space, generated by a given

volume charge distribution , existing in space. The electric displacement
and from Gauss’s law in point form, , we have,

since the divergence of the gradient of a scalar field is the Laplacian of the
scalar field, (see equation (1-8-9)).

Equation (5-2-1) is known as Poisson’s equation. Strictly speaking, Poisson’s

equation is a Partial Differential Equation (PDE) which relates the potential
with the sources of the field, (electric charge distribution).
 224

In each coordinate system, rectangular, cylindrical and spherical, the

Laplacian has its own expression in terms of the corresponding
independent variables, (see Section 1-11), i.e.

Poisson’s equation, being a PDE, admits an infinite number of solutions. In

order to hope for a unique solution, Poisson’s equation should be coupled
with some appropriate Boundary Conditions. Then the solution of Poisson’s
equation which satisfies the given (in each problem) boundary conditions,
leads to a unique solution for the potential , and then from we determine
the electric field , (see uniqueness theorem in Chapter 12).

In case where the electric field is generated within a homogeneous

dielectric with relative permittivity , then the corresponding Poisson’s
equation is
 225

5-3) Laplace’s Equation

In cases where the electric potential is to be determined in a region in

space where (region free of charges), then Poisson’s equation (5-2-1)
reduces to the so called Laplace’s equation

Remark: Poisson’s and Laplace’s equations are usually difficult to be solved

in the general case, where all spatial coordinates are present. However, if there
is some kind of symmetry (cylindrical or spherical, etc) in the problem at hand,
then, this usually leads to a considerable simplification of the aforementioned
equations, (they are reduced to ordinary differential equations).

5-4) Potential of a Point Charge , (Coulomb’s Potential)

a) Let us consider a point charge positioned at the origin , in free space,

as shown in Figure 5-1. We want to determine the potential of this charge
distribution (a single point charge), which is known as “Coulomb’s potential”.

Fig. 5-1: Potential of a point charge (Coulomb’s potential).

Assuming that there is no any other charge in space ( ), the sought for
potential must satisfy Laplace’s equation
 226

Due to the spherical symmetry involved, should depend on the variable

only, (and not on and ), i.e. and in this case Laplace’s equation
reduces to (see formula (5-2-4)):

Notice that since there is no dependence on and , the partial derivatives

and vanish, while the partial derivative with respect to
becomes an ordinary derivative.

The solution of equation (**) is determined easily:

where, up to this point, and are the arbitrary constants of integration.

The field, (according to formulas (5-1-1) and (5-1-4), is

and comparing this expression with the expression for the field of a point
charge (equation (3-3-1)), we find that , and finally, from
equation (***), the following expression for the potential is obtained:

Notice that the arbitrary constant is still undetermined and in fact cannot
be determined unless we are given one more condition for . At the same
time, notice that the electric field does not depend at all from this (as yet
undetermined) constant , since the derivative (gradient) of any constant is
zero.

b) The reference point for the potential: Equation (5-4-1) reveals a simple
physical fact: in nature, there is no such a thing as “absolute value of the
 227

potential”; on the contrary, the potential is always determined with respect to

“a reference point” at which we arbitrarily assign a value to the potential.

For example, for the potential of a point charge , in equation (5-4-1), we

take the infinity ( ) as the reference point and assign the value zero to
the potential at infinity, i.e. we assume that .
Making this assumption, we have,

and finally,

Equation (5-4-2) expresses the potential of a point charge assuming of

course, potential zero at infinity.

This is a typical convention; we assume zero potential at infinity. We may

always set potential zero at infinity, provided that the charge distribution is
finite, not extending to infinity. However, for charge distributions extending to
infinity, the convention of potential zero at infinity fails. In such cases, we have
to make another choice for the reference point of the potential, (see Example
5-4-3).

In some practical problems, it facilitates the calculations if we choose as a

reference point the Earth, (for example in a power transmission system),
assigning zero potential to Earth and then measuring the potential of all the
other points with respect to the Earth.

Example 5-4-1: Find the potential of a point charge at the

point whose distance from is . Assume . Then find the
potential at the same point assuming that . Show that in
both cases, the electric field is the same, (even though the expressions for
the potential are different, since they refer to different reference points).

Solution

In the first case, , the potential is given by equation (5-4-2), i.e.

 228

In the second case . From equation (5-4-1) we obtain,

We notice that with respect to the reference point

, while with respect to the reference point
.

From equation (5-4-1) we have:

in both cases, since . This verifies that, even though the

potential does depend on the choice of the reference point for the potential,
the electric field does not.

Example 5-4-2: Find the potential of the following volume charge

distribution in free space, and then determine the electric field everywhere.
Compare the resulting with that found in Example 4-5-2, (set ).
 229

Solution

Fig. 5-2: Potential of a uniformly charged sphere.

If we call the potential in region 1 ( ) and the potential in

region 2 ( ), then will satisfy Poisson’s equation while will
satisfy Laplace’s equation, i.e.

Due to the spherical symmetry involved, the potential in both regions

depends solely on , that is and and equations in (*)
become,

The solution of the first equation in (**) is

 230

while the solution of the second equation in (**) is

where are arbitrary constants of integration, (for the detailed

calculation of the solutions, see Problem 5-4-3).

In region 1, since the potential should remain finite at the center of the
distribution, i.e. as , the constant must be zero ( ); in region 2,
since the potential should approach zero as the constant must be
zero ( ). Having thus determined and , equations (***) and (****)
become:

The two constants and are determined from the following two
conditions:

a) , (continuity of the potential at the boundary

between the two regions) and

b) , (continuity of the electric field at the boundary

between the two regions).

From these two conditions, we find easily that and

, (for detailed calculations see Pr. 5-4-12), and finally,

The electric field , (equation (5-1-1)).

 231

The expressions for the electric field are identical to the ones found in
Example 4-5-2, equation (*****).

Example 5-4-3: Find the potential and the electric field of an infinite,
uniform, cylindrical charge distribution, with its axis coincident with the axis,
assuming that

Solution

Fig. 5-3: Potential of an infinite, cylindrical, uniform charge distribution.

Due to the cylindrical symmetry involved . If we call the

potential in region 1 and the potential in region 2, then satisfies Poisson’s
equation and satisfies Laplace’s equation (since in region 2, ), i.e.
 232

(For the derivation of the solutions in (*) see Pr. 5-4-13). The constants
are arbitrary constants which need to be determined from the
appropriate boundary conditions.

Since the potential in region 1 must remain finite everywhere

within this region, the constant must be zero ( ), otherwise on the
axis ( ) the potential would be infinite, ( ). Now, since
this charge distribution extends to infinite, we cannot assign a potential zero
to infinity, (simply because charge exists at infinity). Let us, arbitrarily, assume
that the potential is zero on the cylindrical surface , i.e. let us assume
that . Then the first equation in (*) yields (recall that ),

The constants and in the second equation in (*) will be determined

from the requirement that the potential and the electric field are continuous
on the cylindrical surface , i.e.

Tedious but straightforward calculations show that

 233

and finally, in summary:

Notice that these expressions for the potential have been obtained on the
assumption that the potential is zero on the cylindrical surface , (in
formula (***), notice that ).

The electric field is obtained from formula (***) as the negative gradient of
the potential, i.e.

PROBLEMS

5-4-1) Find the electric charge density associated with the potential
, (Ans: ).

Hint: From Poisson’s equation, .

5-4-2) Find the electric charge density associated with the potential a)
and b) .

5-4-3) In Example 5-4-2, solve the differential equations in formula (**). The
solution of the second equation in (**) is obtained easily, since
implies that (constant), etc. The first equation in (**) is a non
homogeneous equation and its general solution is
, where is a solution of the corresponding homogeneous
equation and is a particular solution of the equation. We assume
 234

that and try to determine the constants so

that satisfies identically the differential equation.

5-4-4) Work Example 5-4-3, assuming that the potential is zero at the
cylindrical surface , where is an arbitrary positive number. Find
expressions for and then determine the electric field and in
regions 1 and 2, respectively. Notice that the expressions for the electric fields
thus obtained are identical to the expressions found in Example 5-4-3,
(equation (****)).

5-4-5) In a charge free region of free space the potential is given by the
expression . Find if and are zero at the
origin, (Ans: ).

5-4-6) In Problem 5-4-5, find the volume charge density , and the electric
field associated with the given potential at the point .

5-4-7) Find all solutions of the Poisson equation , in spherical

coordinates, provided that .

(Ans:

5-4-8) Find all possible solutions of the Laplace equation , in

spherical coordinates, that depend solely on .

(Ans: ).

5-4-9) Find all possible solutions of the Laplace equation , in

spherical coordinates, that depend solely on , (Ans: ).

5-4-10) The potential field in free space is given by the expression

. Find the electric field and the volume
charge density .
 235

5-4-11) The potential field in free space is given by the expression

. Find the electric field and the volume charge density .

(Ans: ).

5-4-12) In Example 5-4-2 determine the constants and appearing in the

solution of the differential equations in (*****).

5-4-13) In Example 5-4-3 solve the differential equations involved to obtain

the solutions in (*).

5-5) Potential of a Charge Distribution, (Discrete or/and Continuous)

In section 5-4 we found the potential of certain symmetric charge

distributions, by solving either Poisson’s or Laplace’s equations, subject to the
appropriate boundary conditions.

However, if there is no symmetry involved in the charge distribution, then

we may find the potential using the principle of superposition.

a) Discrete charge distribution.

Let us, for definiteness, consider a discrete charge distribution

positioned at respectively, as shown in Fig. 5-4.

Fig. 5-4: Potential of point charges.

 236

Suppose now that we want to find the potential , of this charge

distribution at a point in space, whose position is determined by the position
vector .

Let us define the vectors from the position of the charges towards the
point , where the potential is to be determined:

Invoking the principle of superposition and formula (5-4-2), the potential at

will be,

or in a more analytic form,

Having found the electric field

, according to formula (5-5-1).

b) Continuous charge distribution.

Fig. 5-5: Potential of continuous charge distributions.

 237

For a continuous charge distribution, the charge element may be

treated as a point charge, the summation in formula (5-5-2) is replaced by
integration, and the formulas yielding the potential at the arbitrary point in
space, assume the following form:

Remarks:

1) Formulas (5-5-2) and (5-5-4) hold true on the assumption that the
reference point for the potential is at infinity; this is so since we have derived
these formulas from the potential of a point charge (formula (5-4-2)), and this
formula was obtained assuming that the reference point for the potential is at
infinity. In short, formulas (5-5-2) and (5-5-4) hold true for charge distributions
not extending to infinity and assuming that .

2) is the magnitude of the vector from the differential element (line,

surface or volume) towards the point , where the potential is to be
determined.

3) So far we have been using the same symbol for the potential and the
volume. In order to avoid any confusion in our calculations, we keep the
symbol for the potential and use the symbols for the differential volume
and for the volume occupied by the charge distribution.

4) Formulas (5-5-2) and (5-5-4) are valid for charge distributions in free
space (vacuum). If charges are within an homogeneous dielectric with
 238

permittivity , the same formulas are used to calculate the potential,

provided that is replaced by .

Example 5-5-1: Three point charges and

are positioned at the points and
respectively. Find the potential at an arbitrary point in space.

Solution

Let denote the vectors from respectively, to the point

. We have:

The potential at is given by formula (5-5-2), that is

Notice that this expression for is a well defined field of the spatial
coordinates, i.e. . If we want to determine the electric field
, all we have to do is to take the negative gradient of , i.e.
 239

The calculations are tedious, but straightforward.

Example 5-5-2: Electric charge Q in free space, is uniformly distributed over

the circumference , lying on the plane. Determine the
potential .

Solution

Fig. 5-6: Potential of a uniform ring of charge.

Since the charge is uniformly distributed over the circumference, the line
charge density is constant and equal to . The potential at
is given by the integral (see formula (5-5-4)),

and since and remain constant in the integration

over the circumference,

Example 5-5-3: Charge in free space is uniformly distributed over the

surface of the circle , lying on the plane. Determine the
potential at .
 240

Solution

Fig. 5-7: Potential of a uniform disk of charge.

Since the charge is uniformly distributed over the surface of the circle, the
surface charge density is constant and equal to . The potential at
is given by the integral, (see formula (5-5-4)),

To evaluate the first integral, we may use the substitution .

Notice that if , formula (*) reduces to , which

makes sense and is justified easily, since in this case (far away from the disk)
the charge may be considered as a point charge, (to show it use the
approximation provided that ).
 241

PROBLEMS

5-5-1) A circular disk of radius carries a surface charge density

. Find the potential at the point , assuming that
.

(Ans: ).

5-5-2) Point charges of and are located at and

. Find the potential at , (distances in ).

5-5-3) The charge density exists along the axis for

and is zero for . Find the potential at .

(Ans: ).

5-5-4) Four equal charges in free space are located at the

points . Find the potential at .

5-5-5) The ring , caries a uniform charge density in

free space. Find the potential at the origin.

(Ans: ).

5-5-6) Four point charges

in free space, are positioned at the corners of a square with side
. Find the potential at the center of the square.

5-5-7) Charge is uniformly distributed over the spherical surface .

Find the potential everywhere.

(Ans: Outside the sphere , within the sphere

).

5-5-8) Consider a square lying on the plane and centered at

the origin. Each side of the square has a length and carries a charge
, uniformly distributed. Find the potential at the point .
 242

5-5-9) A uniform charge density exists along the axis from to

. Find the potential at .

(Ans: ).

5-6) Electric Dipole and Electric Dipole Moment

An electric dipole is a system of two charges of equal magnitude and

opposite signs, separated by a distance , as shown in Fig. 5-8.

Fig. 5-8: Electric dipole.

We want to find the potential and the electric field at a point

determined by its position vector .

The potential at , due to the charges and – is

From Fig. 5-8, and assuming that , we have,

where is the angle between and , and therefore,

 243

Assuming that the point is far away from the dipole, ( ), then
and formula (**) implies that . Using this
approximation and substituting in formula (*) we get,

We define the electric dipole moment as a vector directed from

to and having magnitude equal to , i.e.

In terms of , the numerator in (5-6-1) can be expressed as

and finally (5-6-1) assumes the equivalent form, (note that ):

The electric field at is found by taking the negative gradient of , i.e.

Notice that formulas (5-6-3) and (5-6-4) in fact represent the far field of the
dipole (valid for ). It is interesting to note that the potential of the dipole
falls off as while the electric field falls off as ; recall that the potential
and the electric field of a point charge fall off as and respectively.

Equation (5-6-4) can be expressed in a “coordinate free” form,

 244

For a proof see Problem 5-6-1.

Example 5-6-1: A dipole moment is located at

. Find the potential at the point .

Solution

Fig. 5-9: Potential of an electric dipole.

The vector ,
and application of formula (5-6-3) yields,

Example 5-6-2: Point charges and are located at

and respectively. Find the dipole moment , the
potential and the field at the point .

Solution

The position vector of the point is

The potential of the dipole at is given by formula (5-6-3), i.e.

 245

The electric field is given by formula (5-6-4), i.e.

In formula (**) is the angle between and , and can be determined by

taking the dot product of and , i.e.

and formula (**) implies,

PROBLEMS

5-6-1) Starting with equation (5-6-4) derive equation (5-6-5).

5-6-2) Point charges and – are located at

and respectively. Find the dipole moment , the
potential and the field at the point .

5-6-3) A dipole moment is located at

. Find the potential at the point , (distances in ).

(Ans: ).

5-6-4) In Problem 5-6-3 find the electric field at the point .

 246

5-6-5) Show that the field lines of the electric dipole are ,
, where are arbitrary constants.

Hint: , where in spherical coordinates

and is given by equation (5-6-4).

5-7) The Physical Meaning of the Potential

Let us assume that we have a volume charge distribution, in free space,

which generates an electric field in the surrounding space, as shown in Figure
5-10.

Fig. 5-10: The physical meaning of the potential.

Suppose that we want to move a point charge from point to another

point along the curve (path) . The volume charge distribution exerts an
electric force on , , where is the electric field due the charge
distribution at the position of the point charge . An external force (not due to
the electric field) which is equal in magnitude and opposite in direction to the
field force must be applied to the charge , in order to overcome the electric
force . Let us call this external force , (force applied externally).
Obviously, .

The work done by the external force, in moving the point charge from
point to point , along the path , is (as we know from Mechanics),
 247

If is the potential of the given charge distribution, then as we

know , and substituting in (5-7-1) yields,

We note that

and substituting in (*) we find,

where and are the potentials of the points and respectively.

Since, as we have shown, the static electric field is irrotational, the path
integral is independent of the path taken from to , depends only
on the end points and . We may therefore write the integral in (5-7-1) as
, and equation (5-7-2) assumes the following form:
 248

Formula (5-7-3) expresses the electric voltage or potential difference

between two points and , in terms of the electric field . The electric
voltage is independent of the path connecting the end points and . Formula
(5-7-2) shows that the voltage is the work per unit charge, required
to move the charge from to . The same formula also shows that the unit
.

Finally, let us assume that the point is at an infinite distance from the
(finite) charge distribution, and set our reference point .
Then the potential of an arbitrary point in space is (according to formulas (5-
7-2) and (5-7-3)),

It also follows from (5-7-4) that the work required to move a point charge
from to a point with potential , is

This formula is important, and will be the starting point in Chapter 6, in

order to determine the energy stored in an electric field of a given charge
distribution (discrete or continuous).

A general remark:

The work needed to move a point charge around any closed path is,
according to formula (5-7-2),

since , for any static electric field. Fields having this property
are said to be “conservative”. Thus, the static electric field is a conservative
field. The gravitational field is also a conservative field. Time varying fields, on
the other hand, are not conservative.
 249

Example 5-7-1: Find the work required to move a point charge

from the origin to the point in the electric field
.

Solution

Application of formula (5-7-2) yields,

Since the electric field is irrotational, the integral is path independent. We

may, therefore choose the path , as shown in Fig. 5-11.

Fig. 5-11: Work needed to move a charge from to .

On path :

On path :

In view of (**) and (***) equation (*) yields,

 250

Alternative solution: Since is irrotational, there exists a scalar field (the

potential), such that

From the second equation in (****) we obtain,

and comparing with the first equation in (****) we find

where is an arbitrary constant, and finally, ,

(recall that the potential is determined up to an arbitrary constant).

Application of (5-7-2) yields,

The results obtained by both methods, are of course identical. The given
electric field is conservative.

Example 5-7-2: Four identical point charges each, are located

at the corners of a square on side, in free space. How much work is
required to move a charge at the center of the square?
 251

Solution

Fig. 5-12: Potential of four point charges.

Application of formula (5-7-5) yields,

where is the potential at , due to the four charges at the corners of

the square.

Example 5-7-3: For the charge distribution in Example 4-5-2, we found,

(with the aid of Gauss’s Law), that

Using this expression for the electric field , and formula (5-7-4), find the
potential , at an arbitrary point in space, (assuming ), and
then compare the answer found with the expression for obtained for the
 252

same charge distribution by solving Poisson’s equation, (equation (******) in

Example 5-4-2).

Solution

Fig. 5-13: Potential of a uniform, spherical charge distribution.

Region II, :

Region I, :
 253

In summary:

This expression for the potential is identical with that found in Ex. 5-4-2,
formula (******).

Example 5-7-4: In Example 4-5-1 we found that the electric field of an

infinite, uniform line charge distribution in free space, extending
from to , is given by the formula . Find the electric
voltage between the points and .

Solution

Application of formula (5-7-3) yields,

PROBLEMS

5-7-1) In Example 5-7-4 find the electric voltage between the points
and , assuming , (Ans: ).
 254

5-7-2) Four identical point charges each are located at the

corners of a square on side, in free space. How much work must is
required to move a charge at the center of the square?

5-7-3) For the charge distribution in Example 5-7-3, assume ,

. Find the potential at the points and
. How much work is needed to move a point charge
to the point ? To the point ?

(Ans: .

5-7-4) Three point charges are

located at the corners respectively, of a square , in free space.
The side of the square is . How much work is needed to move a
point charge to the fourth corner ? To the center of the square ?

5-7-5) Find the work needed to carry an electron

from the point to the point
within the electric field , (Ans: ).

5-7-6) The electric field in the vicinity of points and is

. Find the differential work needed to move a point charge
from to .

Hint: , where .

5-7-7) Show that is a valid expression for an

electrostatic field (show that ) and find the work required to move a
point charge from to .

(Ans: ).

5-7-8) Show that is a valid expression for a static electric

field, and find the work needed to move a point charge from
and .

5-7-9) Any field such that around any closed path , is

called “conservative field”. The name follows from the fact that no work is
 255

done around a closed path, which means that the energy is conserved. a)
Show that the gravitational field is conservative, b) Show that any static
electric field is conservative and c) Show that the field is not
conservative.

5-8) Equipotential Surfaces, (Definition and Properties)

a) Let be the potential of a given charge distribution. In general, is a

function of the spatial coordinates, i.e. in rectangular
coordinates, in cylindrical coordinates and in
spherical coordinates.

An equipotential surface is a surface composed of all those points having

the same value of the potential. For example, we have found that the
potential of a point charge a distance from the charge is ,
(assuming ), see formula (5-4-2). Assuming that (constant),
we find , and this shows that equipotential surfaces are
concentric spherical surfaces with center the charge . For we obtain
the equipotential surface , for we obtain another
equipotential surface , etc.

In general we may say that the general equation of the equipotential

surfaces is

(Similar expressions define equipotential surfaces in cylindrical and

spherical coordinates).

b) Some general properties of equipotential surfaces and field lines.

1) The potential difference (electric voltage) between two points

and belonging to the same equipotential surface is
zero. Indeed, .

2) The equipotential surfaces do not intersect; in other words, from each

point in space, passes one and only one equipotential surface. Indeed,
assuming that two equipotential surfaces and
intersect along a line , then we would have for all the points on ,
 256

and ; this would mean that the potential , at one and the same point
, has two different values and , i.e. that the potential
would not be uniquely determined. On physical grounds this is not acceptable.
We are therefore forced to accept that two equipotential surfaces do not
intersect.

3) The field lines are perpendicular to the equipotential surfaces.

Since , the field is perpendicular to the surfaces,

(see Sect. 1-10, formula (1-10-10)), and since, by definition, at each point in the
field, the field line at this point is tangent to the field at the same point, we
conclude that the field lines are perpendicular to the equipotential surfaces.

4) The field (and therefore the field lines) is directed from higher
potentials towards lower potentials. This follows from the definition
.

Let us, for reasons of simplicity, assume that the potential depends only
on , i.e. , and let us further assume that decreases with .
Then , and points in the positive
direction, i.e. the fields points in the direction of decreasing potentials.

5) From each point in space there passes one and only one field line (proof
similar to the property 2).

6) In static electric fields, the field lines cannot be closed. On the contrary,
the field lines have a beginning and an end. The field lines emerge from
positive charges (sources) and dive in negative charges (sinks). It is of course
possible, some of the charges to be at infinity.

Actually, property 6 follows directly from the two fundamental laws of

static electric field.

For example, if we assume that a static field line is closed, then we would
have that the integral would be either positive (in case and
point in the same direction) or negative (if and point in opposite
direction). In either case, would not be zero, and this is impossible
 257

for static electric fields. We are therefore forced to conclude that the field
lines, of static electric fields, cannot be closed. Note that in time varying fields
this is not true; in such cases we may have closed field lines.

Also, the fact that positive charges are sources of field lines and negative
charges are sinks of field lines is explained easily from Gauss’s law
, (see Figure 1-39).

c) Using field lines and equipotential surfaces we may represent graphically

electrostatic fields (approximately), generated by charge distributions. If we
make the assumption that through a small surface of a fixed area (for example
) placed perpendicularly to the field , at an arbitrary point in space,
there pass as many field lines as is the integral value of the magnitude of at
, then we can obtain an approximate graphical representation of the field. At
regions where the field lines are dense the field is strong, while in regions
where the field lines are sparse the field is weak.

The equipotential surfaces are then plotted so that they are always
perpendicular to the field lines.

5-9) Earnshaw’s Theorem

Earnshaw’s theorem states that a charged particle subject to electrostatic

forces only, cannot be held in stable equilibrium.

Earnshaw’s theorem has some far reaching consequences, regarding the

confinement of charged particles by electric forces only, which will be
discussed in the sequel.

If we place four identical positive charges at the corners of a square and a

negative charge at the center, then the total force on the negative charge is
zero and the particle will be in equilibrium (at the center). But is this
equilibrium position stable? The answer (as implied by Earnshaw’s theorem) is
negative.

We may justify Earnshaw’s theorem by the following arguments.

Let us consider a spherical surface of radius and center , inside an

electrostatic field in free space, as shown in Fig. 5-14.
 258

Fig. 5-14: Proof of Earnshaw’s Theorem.

If is the potential on the surface of the sphere, then the

average potential of the surface is,

Taking the derivatives of both sides with respect to yields,

since, by Gauss’s law in integral form, the integral appearing in (5-9-2) is

equal to the total charge enclosed by the spherical surface of radius .

Integration of (5-9-2) implies that

 259

If , (no charge enclosed by the surface ), equation (5-9-3) shows

that the average value of the potential (constant), no matter how
small the radius can be. In other words, the average value of the potential on
the surface of the spherical surface is equal to the potential at the center ,
provided that there are no charges included. This, in turn, implies that the
potential at the center cannot attain neither a maximum nor a minimum
value at , and since is an arbitrary point in space we conclude that, in the
absence of electric charges, the potential function cannot attain maximum
or minimum values.

The conclusion is that a freely movable point charge which is brought

within an electric field, (generated by another charge distribution) and
subject to electrostatic forces only, cannot be in stable equilibrium. To justify
our assertion, let us consider Fig. 5-15.

Fig. 5-15: Conditions for stable equilibrium of a point charge.

A negative charge will be in stable equilibrium, provided that is placed at

a point where the potential has a local maximum, (recall that the field is
directed from higher towards lower potentials), so if , then the restoring
force will have the direction shown in the first figure, and would attract back
to its equilibrium position. However, as we have just proved, the potential
cannot attain maximum, so a negative charge cannot be held in equilibrium
position.
 260

Similarly a positive charge will be in equilibrium position provided that is

placed at a point where the potential has a local minimum, and since such
points do not exist, a positive charge cannot be held in equilibrium position.

We are thus led to the conclusion that electric charges cannot be held in
stable equilibrium by static electric forces alone. Electric charges either will
move to infinity or will be held at their positions with the aid of other
mechanical forces. This means that electric charges, which generate
electrostatic fields, are always attached to material bodies.

In the microscopic world, Earnshaw’s theorem implies that the electrons in

an atom or a molecule can never be at rest, on the contrary, these elementary
particles are subject not only to static electric forces, but to other kind of
forces (magnetic forces) as well.

Finally, Earnshaw’s theorem rules out the possibility of constructing

“electrostatic shields”, i.e. structures by means of which electric charges could
be confined to some region in space under their own influence, i.e. by
electrostatic forces only. However, as we shall see in Section 19-5, electric
charges can be confined in space, by properly designed magnetic fields,
(magnetic bottle configuration).

SUPPLEMENTARY PROBLEMS

5-1) In Problem 5-5-1, find the total charge of the disk, and show that if
the potential is . Give a physical explanation of this
answer.

5-2) Graph (approximately) the field lines of the field of two charges and
, a distance apart, in the following cases:

a) , b) , c)
.

5-3) In Example 5-4-3 find the voltage between the points and
, a) using the expression for the potential obtained in Example 5-4-3
and b) using the formula (5-7-3), i.e. ( ). The expression for
the electric field may be obtained from formula (****) in Example 5-4-3.
 261

(Ans: ).

5-4) In Example 5-4-3 find the equipotential surfaces, (assume ).

5-5) An electric dipole with moment is

located at the point in free space. Find the potential at the
point , (Ans: ).

5-6) A dipole moment is placed at the origin

and another dipole moment is placed at
. Find the potential and the electric field at the
point . Assume free space conditions.

Hint: Find and for each dipole separately, and then use superposition.

5-7) The sides of a square, of side each, carry a uniform line charge
density . Find the potential at the center of the square.

(Ans: ).

5-8) A square with sides carries a uniform surface charge density

. Find the potential at the center of the square.

(Ans: ).

5-9) Four identical point charges each, are placed at the

corners of a square lying on the plane and centered at the origin. The side
of the square is . Find the potential at the point . What is
the potential when ?

(Ans: , at ).

5-10) In Problem 5-9, what is the energy required to move a point charge
from infinity to the point ?

5-11) A point charge is placed at the origin, in free space. Find

the work required to move another point charge from
to the point .
 262

(Ans: ).

5-12) The axis carries a uniform line charge density ,

while two point charges and are placed at the points
and respectively. Find the potential difference
between the points and .

5-13) A dipole moment is located at the origin,

while a point charge is placed at . (a) What is the work
required to move a point charge from infinity to the point
? and (b) What is the potential difference between the
points and ?

(Ans: ).

Hint: Use the superposition principle.

5-14) The potential of a charge distribution in free space is given by

. Find the volume charge density , the fields and
and the total charge within the sphere .

Hint: .

5-15) A small sphere with mass having a positive charge

, moves at a speed of . Up to what distance can the
ball approach a positive point charge ?, (Ans: ).

5-16) Up to what distance can two electrons approach each other if they
are moving towards each other with a relative speed of ?

5-17) Two small spheres with charges and

respectively, are at a distance . What work is required to reduce
their distance to ?

(Ans: ).

5-18) Six equal charges are placed at the vertices of a regular

hexagon of side . Find the potential and the electric field at the
center of the hexagon.
 263

5-19) Determine the potential at a point at a distance from

the center of a small charged sphere of radius , provided that the
surface charge density of the sphere is , (Ans: ).

5-20) (a) Justify why an electron, when placed in an electric field, is

accelerated from regions of lower potential towards regions of higher
potential, (b) Find the speed of an electron which passes through a potential
difference equal to and .

Hint: (a) Recall that the electric field is directed from higher potentials
towards lower potentials and that the force on the electron is , where
the charge of the electron is negative, (b) In general, the speed of the
electron is given by the formula, , where is the potential
difference.

5-21) A point charge is at a distance from an

infinitely long filamentary conductor, carrying a line charge density
( . Under the action of the field, the point charge moves to a distance
and a work is performed. Determine the electric charge
density , (Ans: ).

5-22) A point charge is near an infinite charge sheet with

. Under the action of the field the charge moves over a
distance of and a work is performed. Find .

5-23) A small sphere with mass and charge moves

from a point where the potential is towards a point where the
potential is . What is the speed at when the speed at the point is
?, (Ans: ).

5-24) In Problem 3-1 (supplementary problems on Chapter 3), the negative

charge at the center of the square is in stable equilibrium for small vertical
displacements and in unstable equilibrium for small displacements in the plane
. Does this contradict Earnshaw’s theorem? Justify your answer.

5-25) A circle of radius , lying on the axis and centered at the origin,
has a positive point charge placed at the point . (a) Find the
 264

potential at the point , (b) Find the flux of through the surface of
the circle, (take as unit out normal ).

(Ans: , ).

5-26) In problem 5-25, a second positive charge is placed at the point

. Given that , find (in terms of ) if the net flux of
(due to and ), through the surface of the circle, is zero.

5-27) In quantum Physics, it is shown that in the ground state of the

hydrogen atom, the electronic cloud has a charge density ,
where (the charge of the electron) and is the Bohr radius.
(a) Show that the total charge of the atom is , (b) Show that the potential
due to this charge distribution is given by the formula

Hint: Solve Poisson’s equation in spherical coordinates.

5-28) Show that an electric dipole , when placed in an homogeneous

electric field , experiences a torque . In addition, when is placed
in an inhomogeneous field , it experiences a net force .

Remark: For a more systematic treatment on torques on electric dipoles,

see Chapter 10, Section 10-2. For the force on an electric dipole, see Chapter 8,
equation (8-2-9).

5-29) A system of charges at the origin, at and at

exists in free space. Show that the potential far away from this
charge configuration , the potential is (approximately) given by the
expression .

Hint: The given charge configuration is known as “a qudrupole”. For a

systematic method to obtain the potential of a quadrupole, see Chapter 7.
 265

CHAPTER 6: ELECTRIC FIELD ENERGY AND ENERGY DENSITY

6-1) introductory Concepts and Definitions

Every electric field generated by a charge distribution (discrete and/or

continuous) contains energy, called “electric energy”, which is actually the
energy expended in building up the charge distribution, which in turn,
produces the electric field in space.

The key idea to understand why “energy is stored in electric fields” is to

recall that charges interact by Coulomb forces. Let us, for definiteness, assume
that we have a positive point charge in space and that we want to bring
another positive charge in the vicinity of the first one, and at a distance
apart. Due to the Coulomb interaction, the initial charge repels , while
approaching, and an external agent should actually push against the
repulsive force exerted on by , to bring a distance from , and then
hold it there. The work done by the external agent, cannot disappear, on the
contrary, it is found redistributed as “electric energy” stored in the electric
field of the final charge configuration (the system of two charges and , a
distance apart). In simple terms, the electric energy stored in any electric
field is a consequence of the “conservation of energy”.

If the external agency releases its hold on the charge , then this charge
would accelerate away from (assumed to be fixed in space), and its kinetic
energy far from , (an infinite distance from ) would be equal to the energy
expended in bringing at a distance from .

The same argument applies in cases where we have the formation of a

discrete charge distribution of more than two charges or even the formation
of a continuous charge distribution (which may be considered as a
conglomeration of infinitesimal charges brought to their respective
position, one by one, from infinity, where the Coulomb interaction is practically
zero).

In summary, in order to find the electric energy stored in an electric field,

we must find the work (energy) done by an external agent in positioning the
charges, that is, in forming the charge distribution.
 266

We shall first find the energy of a discrete distribution and then we shall
generalize to the case of a continuous charge distribution.

6-2) Energy of a Discrete Charge Distribution

Let us, for definiteness, consider four charges positioned

initially, somewhere at infinity, as shown in Fig. 6-1.

Fig.6-1: Energy of a discrete charge distribution.

We start forming our charge distribution by bringing first, the charge

from infinity to its position . In doing so, there is no work required, since
there is no field present. Next we move the point charge from infinity to its
position . The work required is (according to formula (5-7-5)),

where is the potential at the point due the charge , (as usually, we
assume ). So far the charge configuration consists of two point
charges positioned at and .

We now move the third charge from infinity to its position , and the
work required is
 267

since the potential now at is the potential due to and

(superposition of potentials).

Finally, in moving from infinity to its position , the work needed is

where is the potential at due to the charges and

.

The total work required to form the given charge distribution is therefore
equal to

Formula (****) may assume a more symmetric form, using the following
argument.

We assume that we form the same exactly distribution, but with a slightly
different order; we move first charge at , then at , next at and
finally at . The work required is obviously the same as the one in (****),
because in both cases we have formed the same charge distribution, but now
the work is given by the expression

Adding together equations (****) and (*****) we find,

We note that is the total potential at , the position of ,

due to all the other charges( except positioned at ), is the
total potential at , due to all the other charges (except positioned at ),
etc. We may thus write:
 268

and in general, for a discrete charge distribution consisting of point

charges ,

where is the total potential at the position of the charge , due to all
the other charges (all except ).

If the relative distances between the point charges are known, then we may
apply the formula for the potential of a point charge (Coulomb’s potential) to
determine the potentials and then, with the aid of formula (6-2-1), to find the
energy stored in the charge distribution. For example, in Fig. 6-1,

and similarly we find and .

Example 6-2-1: Find the energy required to place four charges

each, at the corners of a square of side .

Solution

Fig. 6-2: Energy of a discrete charge distribution.

The potential at due to all the other charges (at ) is,

 269

and similarly, .

The sought for work is (according to formula (6-2-1)) is,

Example 6-2-2: Find the energy required to gather three equal point
charges Coulomb each, at the vertices of an equilateral triangle of side .

Solution

Fig. 6-3: Energy of a discrete charge distribution.

The potential at due to the other two charges is

and due to the symmetry,

The work required to gather the three charges is

 270

PROBLEMS

6-2-1) Find the energy required to gather three equal charges Coulomb
each, to the axis, at the points and .

(Ans: ).

6-2-2) Find the energy required to gather four equal charges Coulomb
each, at the three vertices and at the centroid of an equilateral triangle of side
.

6-2-3) An electron at rest is accelerated from point of potential

to the point of potential . Explain why the electron is accelerated
from lower towards higher potentials and find the speed of the electron at
point , (neglect gravitation).

(Ans: ).

Hint: Recall that the electric field is directed from higher towards lower
potentials and that the charge of the electron is negative. Conservation of
energy requires, , etc.

6-3) Energy of a Continuous Charge Distribution, Energy Density

Equation (6-2-1) may be generalized for a continuous charge distribution

inside a volume , if we make the transformations

and replace the summation by integration. Equation (6-2-1) assumes now

the form:

where and are considered to be functions of the spatial coordinates.

 271

Fig. 6-4: Energy of a continuous charge distribution.

It will be convenient in our subsequent analysis, if we convert formula (6-3-

1) into an expression involving the electric field alone.

In Fig. 6-4, let be the volume enclosing the charge, in free space, and let
also and be the volume charge density and the potential, respectively, at
the center of the differential volume . Since , where in free space
, equation (6-3-1) is expressed equivalently as

By using the vector identity

equation (*) becomes

 272

and transforming the first integral in (**) using the Gauss-Ostrogradsky

theorem (the Divergence theorem), we find,

where is the closed surface enclosing the volume .

Note that, if we integrate over a bigger surface, for example the surface
in Fig. 6-4, which encloses the surface , (the surface boundary of the volume
), the total energy will not change, since between and , (see
equation (*)). This implies that while the second term in (***) increases as the
volume increases from to , since is a positive
quantity, the first term (the surface integral) has to decrease in such a way so
that the sum of the two integrals in (***) remains constant, and equal to .

We may now assume that the surface keeps increasing, for example, we
may assume that becomes a spherical surface of some radius , where
. Then from the points of this surface the charge distribution inside the
volume will look like a point charge, and therefore on this spherical surface
of radius

while the spherical surface increases proportionally to , and this means

that the first (surface integral) in (***) is proportional to and therefore,
approaches zero as . At the same time, as the surface area keeps
increasing, the volume enclosed covers actually the whole space, and formula
(***) yields,

From this formula we may define “the electric field energy density ” as
 273

The electric field energy density is “the electric energy per unit volume”, is
measured in , and when integrated over all space gives the total
energy stored in the field generated by a continuous charge distribution.

Formula (6-3-2) is very important since it associates the electric energy with
the electric field. The work required to build the continuous charge
distribution is found redistributed everywhere in space, in regions where
.

Important comments:

1) For line charge distribution and surface charge distribution

, the fundamental formula (6-3-1) becomes,

2) Equation (6-3-1) shows that the energy density exists where charges
exist ( ) and is zero at points where . However, equation (6-3-2)
which was derived from (6-3-1), shows that electric energy exists wherever an
electric field exists. The two interpretations seem to be conflicting, since in
the first case ( ), , while in the second case even in places
where . However, at the same time, notice that the domains of
integration are different in the two cases. In equation (6-3-1) the volume of
integration is the volume occupied by the charge, while in the second case the
integration is over all space. Both integrations give the same answer for the
total energy stored in the electric field, which is equal to the work done to
form the charge distribution, (see Example 6-3-1).

For our subsequent analysis it will be convenient to think that the energy is
stored in the electric field and thus use formula (6-3-2) in order to find the
energy stored in the field of a continuous charge distribution.
 274

3) The energy density, as shown, is . In free space,

and ; in a homogeneous dielectric characterized by a permittivity
, and . For anisotropic materials (see Section 4-3), where
the vectors and are no longer parallel, the energy density is still given by
the formula .

4) To compute the energy of a given (continuous) charge distribution, it

suffices to know the potential of the distribution, since from we can find
, then we can find the energy density , and finally,
integrating over all space we may compute the total energy stored in the
field.

Example 6-3-1: Find the energy associated with the charge distribution

Work the problem using: a) formula (6-3-1) and b) formula (6-3-2) and
verify that both methods yield the same result for the energy stored.

Solution

Fig. 6-5: Energy of a spherical ball of charge.

The potential of this charge distribution was found in Example 5-4-2 to be

 275

and the field of the same charge distribution was found in Example 4-5-2
to be

a) Find using equation (6-3-1):

and evaluating the integrals we find

b) Find using formula (6-3-2):

 276

Equations (***) and (****) verify that both methods yield the same answer
for the energy stored in the field.

Example 6-3-2: Find the energy of a uniformly charged spherical surface of

total charge and radius .

Solution

In Pr. 5-5-7 we found that the potential of this spherical charge distribution
is given by the formula

while the electric field was computed in Ex. 3-3-5,

 277

a) Find using the modified formula (6-3-4) for surface charge density:

Notice that formula (***) applies on the spherical surface since the charge
exists on the surface and is zero everywhere else. Also, since the charge
is uniformly distributed, , and formula (***) yields,

b) Find using formula (6-3-2):

The results in (***) and (****) are identical.

Example 6-3-3: The potential of a charge distribution, in free space, is given

by the expression . Find the electric field , the energy
density and the total energy stored inside a sphere of radius ,
centered at the origin.

Solution
 278

PROBLEMS

6-3-1) A potential field is expressed as in free space.

Find the electric field , the energy density and the total energy stored in
the region , (Ans: ).

6-3-2) A potential field is expressed as in free space.

Find: a) the electric field , b) the energy density , c) the
charge density , d) the energy stored inside a small volume element
centered at and e) the field line passing through the
point .

Hint: .

6-3-3) In Example 5-4-3 find the energy stored inside the cylinder
, (Ans: ).

6-3-4) A volume charge distribution in free space is given as

where and are given constants. Find the electric field everywhere
( ) and the electric energy stored in space.
 279

6-3-5) Assume that the electron is a spherical ball of charge , uniformly

distributed within a sphere of radius , (see note in Example 4-5-2). If we call
the energy of the electron inside the sphere of radius and the
total energy of the electron (the self energy of the electron) show that
. Note that, about of the energy, is stored outside
the electron.

6-3-6) If , find the charge density , the potential

and the total energy of the field.

SUPLEMENTARY PROBLEMS

6-1) Two separate charge distributions generate electric fields and ,

respectively. These two charge distributions are brought in the same region of
space. Show that the total energy of the system is equal to the sum of their
individual energies plus a cross energy term equal to . This
problem shows that the principle of the superposition does not apply for the
energy.

6-2) The potential in a region in space is .

Find: (a) The potential at the point , (b) The
electric field at the same point , (c) The total energy stored within the cube
, (d) The energy stored within the infinitesimal volume
centered at the point .

6-3) Four charges each, are placed at the corners of a square with
side , in free space. (a) Find the energy stored in this charge
configuration, (b) Assuming that the charge configuration is placed in a
dielectric medium with , recalculate the energy.

(Ans: ).

6-4) Given that , find (a) The total charge within the
volume , (b) The energy stored within the same
volume.

6-5) Find the energy associated with six equal charges each,
placed at the vertices of a regular hexagon with side .
 280

(Ans: ).

6-6) In Problem 6-5, find the work required to bring a charge ,

from infinity to the center of the hexagon, (assume free space conditions).

6-7) The potential in a region in free space is . Find

the charge density , the electric field , the electric flux density and the
energy stored in the volume .

(Ans: ).

6-8) For the charge distribution in Example 6-3-1, find the energy stored in
the volume .

6-9) Let be a closed surface which does not enclose any electric charge.
Show that the electric energy stored in that part of the field, enclosed by is
given by the formula, ,( is the directional derivative of
the potential in the direction of the out normal unit vector ).

6-10) The potential in a region in free space is

KVolts. (a) Find the electric field , (b) the energy within an infinitesimal
volume centered at the point ,
(c) the total energy stored in the volume .

6-11) For the charge distribution in Example 5-4-3, find the energy stored in
the volume , (Ans: ).

6-12) For the charge distribution in Example 5-4-3, find the energy stored in
the volume .

6-13) Find the energy density of the electric field (a) at a distance
from the surface of a charged spherical shell with radius and
, (b) at a distance from an infinitely long charged
plane with , and (c) at a distance from an infinitely
long charged filament with . For all the three cases, .

(Ans: ).
 281

6-14) Find the energy required to place three equal charges at

the vertices of an equilateral triangle of side and a fourth charge
at the centroid of the triangle.

6-15) Two concentric conducting spherical surfaces of radii and

have equal charges on the inner and on the
outer. Assuming that the region between the spheres is filled with a dielectric
with , find the total energy stored between the two spheres.

(Ans: ).

6-16) An electric dipole is located at the

corner of a square , with side . Find the work required to
place three equal charges at the other three corners and .

6-17) The potential in a region in space is . (a)

Find the total energy stored within a cube of side , (b) Find the radius
of a sphere within which the total energy stored is .

(Ans: ).

6-18) (a) Find the energy associated with three equal charges
each, located at the points and , (b) Repeat if
the charge at is replaced by .

6-19) The potential in a region in space is .

(a) Find the charge density , (b) The electric field and (c) the total energy
stored within the cube .

(Ans: , ).

6-20) (a) Show that the expression , is a valid expression

for a static electric field, (b) Find the total energy stored in the volume
.
 282

CHAPTER 7: A SYSTEMATIC METHOD FOR CALCULATING

POTENTIALS OF AN ARBITRARY CHARGE DISTRIBUTION (MULTIPOLE
EXPANSIONS)

7-1) Fundamental Concepts and Definitions

In this chapter we shall develop a general, systematic method to compute

the potential of a given, arbitrary charge distribution. This method is known as
“the multipole expansions method”.

In order to keep our (rather complicated) calculations as neat and concise

as possible, we shall make a small change in our so far notations, regarding the
symbols of the independent variables. This will eventually prove to facilitate
our subsequent analysis.

We shall be using the notation instead of to represent

the independent variables of the field points, i.e. the points where the field is
to be evaluated, and the notation instead of to designate
the source points, i.e. the points in space where the charges are located, (see
Section 1-13). The potential and the field at a field point, is found by a
proper integration over the source points. The differential volume element
within the source space will now be .

Fig. 7-1: A systematic method to compute .

 283

7-2) A Systematic Approach for the Calculation of the Potential Generated

by an Arbitrary Charge Distribution

Let us consider a volume charge distribution , in free space,

inside a volume bounded by the surface , as shown in Fig. 7-1. The potential
at the field point , (according to formula (5-
5-4)), is given by the integral, (superposition principle):

or equivalently,

Let us consider the function

The function , obviously, is a function of the six variables

. Let us now consider the point in the
six dimensional space, and expand in a Taylor’s series about . We find that,

In formula (**) we can make the following remarks:

The first term in the right-hand side in equation (*) is

 284

Regarding the partial derivatives appearing in (**) we note that

This formula results easily from the definition of the function

, (let the reader check it).

In view of formulas (***) and (****), the second term in the right-hand side
in equation (**) assumes the following form:

(since ).

Formula (**) now is written equivalently as,

and substituting this expression in (*), results in the following expression for
:
 285

where each one of the summation indices and takes the values ,
and finally,

We note that the integral is the total charge inside

the volume . At this point we make the following definitions:

The first moments of the charge distribution (dipole moment) is defined as

while the second moments of the charge distribution (quadrupole

moment) is defined as

In view of formulas (7-2-3) and (7-2-4), equation (7-2-2) assumes the form
 286

This formula can be simplified further, if we notice that (for a proof see
Problem 7-2-1)

where the symbol is “the Kronecker’s Delta” defined by the formula

and finally, formula (7-2-5) takes the form:

This formula is called “the multipole expansion for the potential at the
point , due to an arbitrary charge distribution”, and admits a simple
physical interpretation. Far away from the charge distribution ( )
the dominant term in equation (7-2-8) is the term . Indeed, far away

from the charge distribution, the charge looks like a point charge , , and,
to a first approximation, its potential at is , (the monopole term).

It is possible, however, the total charge inside the volume to be zero,

(consider for example an electric dipole consisting of two charges and – ).
Then the first term in (7-2-8) is zero, and the dominant term, in this formula, is
the second term , (the dipole term), unless of course it
 287

also vanishes, and in this case, the dominant term in (7-2-8) is the third term
, (the quadrupole term), etc.

Equation (7-2-8) provides a systematic procedure to determine the

potential at a point , due to an arbitrary charge distribution, as a series of
successive approximations, with each term of the series being smaller of the
preceding one.

Discrete charge distributions:

Let us consider point charges positioned at the points

, , … ,
, respectively. In this notation, each component of the vector
, has two indices. The first index
refers to the corresponding coordinate, while the second index refers to the
corresponding point charge. For example, the component means the “
component of the charge ”, the component means the “ component
of the charge ”, etc.

For this discrete charge distribution , the moment formulas

(7-2-3) and (7-2-4) assume the following form:

Dipole and quadrupole configurations are important in studying the forces

between atoms and molecules.

Remarks:

1) Consider the dipole term in equation (7-2-8). We

may define the dipole moment vector by means of the formula
 288

Then, the dipole term for the potential may be written, equivalently, as

and this expression is identical to the potential of an electric dipole in

equation (5-6-3).

2) The quadrupole moment in (7-2-8) is conveniently represented by a

three by three matrix of the form

Note that the terms depend on the charges and their relative
disposition, while the terms depend on the field coordinates, i.e. the
coordinates of the point where the potential is to be determined.

Example 7-2-1: Consider the following system of point charges:

Find the total charge , the first moments and the second moments
of this charge distribution.

Solution
 289

Application of formula (7-2-9) yields:

Similarly, for the second moments:

The calculations are straightforward, and the final result is (let the reader
check it),

Example 7-2-2: A molecule placed at the origin ( ) has a quadrupole

moment
 290

Find the potential at ,( ).

Solution

From formula (7-2-8) the potential of the molecule is

In our problem, , and since all the terms of are zero, except
and , formula (*) implies that

The terms are evaluated from formula (7-2-6), i.e.

and from equation (**),

 291

PROBLEMS

7-2-1) If show that

where is the Kronecker’s delta.

Hint: Notice that , since are independent variables.

7-2-2) An electric dipole is placed at the origin

. Find the potential at the point ., (Ans: ).

7-2-3) Consider the following discrete charge distribution:

where and . Find the total charge , the first

moments and the second moments of the charge distribution.

(Ans: ).

7-2-4) For the charge distribution in Problem 7-2-3 find the potential
and .

Hint: See Example 7-2-2.

 292

SUPPLEMENTARY PROBLEMS

7-1) For the charge configuration in Problem 5-29: (a) Find the dipole
moment vector , (b) the quadrupole moment , (c) all the terms .

(Ans: , all other terms are zero,

, all other terms are zero).

7-2) Using the results of Problem 7-1, re derive the formula for the potential
, found in Problem 5-29, i.e. , where is the
angle between the axis (where the charges are placed) and the position
vector .

7-3) Three charges and are located at the

points and , respectively. Use the
formula for the potential obtained in Problem 7-2, find the potential at the
point , (Ans: ).

7-4) For the charge configuration in Problem 7-3, find the potential at the
point using the exact formula for the potential of a point
charge (Coulomb’s potential) and the superposition principle. Compare your
answer with the result obtained in Problem 7-3.

(Ans: , practically the same with the answer in Problem 7-3).

 293

CHAPTER 8: INTERACTION OF AN ARBITRARY CHARGE

DISTRIBUTION AND AN EXTERNALLY APPLIED ELECTRIC FIELD

8-1) Introduction

In Chapter 6 we calculated the energy stored in an electric field, which is

actually the energy expended to form the charge distribution. The work done
to “build” the charge distribution is found redistributed in the electric field
generated by this charge distribution.

In this chapter we shall study a slightly different problem; we assume that

an electric field already exists in free space, (let us call it, an externally
applied field), and the problem is to find the energy needed to “build” a new
charge distribution, within the already existing field , which obviously has
been formed by another charge distribution, positioned somewhere else in
space.

We shall keep the notation introduced in Chapter 7, i.e. for the

field points and for the source points.

8-2) Calculation of the Potential Energy of a Charge Distribution Formed

within an Externally Applied Field

Let us assume that within an already existing static electric field (formed
by a charge distribution located somewhere else in space), we have placed a
volume charge distribution , inside a volume , bounded by a
surface , as shown in Fig. 8-1. We may assume that the charge distribution
is being built gradually, by moving elementary charges , one
after the other, from infinity, to their terminal position, inside the volume .

If we call the differential charge inside the differential volume

, at the point , then the work required to move
from infinity to the point , is (according to formula (5-7-5)),

assuming that . Notice that is the potential at the point

due to the charge distribution, which generates the applied field .
 294

Fig. 8-1: Calculation of the interaction energy.

The work done in moving from infinity to the position , is by definition

the potential energy of the charge , i.e.

and the total potential energy of the charge distribution inside the
volume is obtained by integration over , i.e.

We further proceed by expanding in a Taylor’s series about the

origin :

and substituting in (8-2-2) yields,

 295

Noting that (charge inside the volume ),

is the first moments of the distribution (see
formula (7-2-3)) and is the second moments
of the distribution, (see formula (7-2-4)), equation (8-2-3) becomes:

In formula (8-2-4), each one of the summation indices and takes on the
values . Also, in the same formula, for simplicity, we use the notation
instead of since , and

therefore, their partial derivatives with respect to their corresponding

independent variables, when evaluated at the origin , yield the
same number.

If we further define:
 296

where (see formula (7-2-10), formula (8-2-4) assumes

the form

The force in general, is related to the potential energy, by means of the

formula

a) Let us now consider an electric monopole (a charge ) at the origin

. Then the force on due to the external field is

(We recover the well known formula for the force exerted on a point charge
).

b) Similarly, the force exerted on an electric dipole placed at the

origin, is given by the formula . The component of this force
in the direction of ( ) is,
 297

The total force (as a vector) on the electric dipole, is therefore equal to

The meaning of the term is to be understood in the following

sense: This term is actually an operator, which operates on the electric field
which appears to its right side. Notice that, formally,

and this term clearly, is an operator, which makes sense when it operates
on the quantity to its right, (the electric field in our case).

For an alternative proof of formula (8-2-9), see Problem 8-2-5.

Example 8-2-1: Show that the force exerted on a dipole placed in a uniform
electric field is zero.

Solution

Application of formula (8-2-9) implies that

 298

This makes sense, since in a uniform field , the forces on and on

have equal magnitudes ( ) but opposite directions, and the
net force is zero. However, in a non uniform field the two forces do not
balance, and the net force is given by formula (8-2-9).

Example 8-2-2: A beam of molecules with dipole moment

, mass and initial velocity
passes through a small hole on screen and is directed to the
screen , (see Fig. 8-2). Between and there is a potential difference
where . Determine the deflection of the
molecules when they reach .

Solution

Fig. 8-2: The deflection of a beam of polarized molecules.

The force on the dipole is given by equation (8-2-9), i.e.

 299

According to Newton’s Law of motion,

where are arbitrary constants, to be determined from the initial

conditions

Straightforward calculations show that and

equations (**) take the form

and eliminating the time between the two equations yields the following
expression: . The deflection of the beam at is

PROBLEMS

8-2-1) Show that the interaction energy between two dipole and
separated by a distance , in free space, is given by
 300

Fig. 8-3: The interaction energy between two electric dipoles.

Hint: The interaction energy is the energy required to place at a distance

apart from . According to formula (8-2-6) this energy is given by

where is the potential at the position of due to , i.e. ,

(see formula (5-6-3)).

8-2-2) In Problem (8-2-1) assume that

is located at while
is located at . Find the interaction energy .

Hint: Notice that .

8-2-3) A dipole is located at the origin .

What is the work required to move a point charge from infinity to
the point ?, (Ans: ).

8-2-4) What is the force exerted on a dipole ,

placed in the potential field , at the point
.

Hint: , where .
 301

8-2-5) Consider an electric dipole , in a non uniform field , as

shown in Fig. 8-4.

Fig. 8-4: Electric dipole in a non uniform electric field.

The net force, on the dipole, is . If we call,

, (assuming to be very small), then we have,
,
i.e. , and the net force is, , or since
, .

8-2-6) (a) Show that the torque a dipole experiences when placed in a
uniform field is , and (b) find the torque on the dipole
( ), when placed the uniform electric field .

8-2-7) (a) Show that the potential energy of a dipole in an electric field
is , (b) Find the energy of the dipole in
the electric field , (Ans: ).

Hint: Use the second equation in formula (8-2-5).

8-2-8) Consider the dipole placed at the origin and the

dipole placed at the point . Find the force and the
torque on the dipole due to the dipole .
 302

CHAPTER 9: CONDUCTORS IN STATIC ELECTRIC FIELDS

9-1) Conductors and Dielectrics

As we know from elementary Physics courses, materials, as far as their

electrical properties, are divided in two main categories, conductors and
dielectrics or insulators.

A conductor is a material body which allows electric charges to move

freely through its mass. The carriers of electric charge are electrons. To explain
the free motion of electric charges in conductors, we assume that the
conductor contains a huge number of “free electrons” or “conduction
electrons”. Recall that the Coulomb’s force which attracts an electron to the
nucleus is inversely proportional to the square of the radius of the orbit of the
electron. Electrons orbiting at large distances from the nucleus (the most
remote ones) are therefore loosely bound to the nucleus and they can easily
escape from their parent atom. They become “free electrons”. Each atom in a
conductor may contribute one or more free electrons which move around
through the crystal lattice of the conducting medium. Free electrons are no
longer associated with any particular atom, and due to their huge number, we
usually talk about “the cloud of free or conduction electrons”. A typical
number of the density of free electrons in conductors is of the order of
. All metals are good conductors.

A dielectric or an insulator is a substance which does not allow the motion

of electric charges through its mass. If an electric charge is placed or somehow
developed in a region inside an insulator, it will remain there forever (at least
theoretically, assuming a perfect insulator). In dielectrics, there are no free
charges; on the contrary, dielectric materials contain “bound charges” which
are bound more or less tightly to their parent atoms. Glass, mica, Bakelite, and
porcelain are good insulators.

In real life there are neither “perfect conductors” (an unlimited supply of
free electrons) nor “perfect insulators” (zero number of free electrons), but
there are substances which come amazingly close to these two ideal models.
 303

The study of the behavior and the properties of conductors and insulators is
an important topic for Electrical Engineers, and a number of interesting
applications follow from their thorough study and understanding.

9-2) Basic Properties of Conductors

For a perfect conductor, as defined in Section 9-1, the following properties

hold true:

1) The field inside a conductor is zero ( ).

Fig. 9-1: The electric field inside a conductor.

Indeed, let us assume that a conductor is placed in a static electric field .

Then, since the conductor contains a huge number of free electrons, which can
move freely within its body, the free electrons under the influence of the
electric field are subject to an electric force , and since the charge
of the electron is negative, the electrons will move towards the left side of the
conductor. This results in an accumulation of negative charge on the side (a),
so the side (a) is charged negatively. On the other hand, on side (b) will be a
deficiency of negative charge (since all the electrons have moved to the left)
and this side will be charged positively. Notice that even though the
conductor, as a whole, is still electrically neutral, the separation of charges
(negative charge on side (a) and positive charge on side (b)) produces an
electric field directed from the positive side towards the negative side. The
 304

total field inside the conductor is the superposition of the initial field plus
the induced field , i.e.

The induced field keeps increasing, until the total field becomes zero,
and from this point on the movement of free electrons stops, (since there is no
electric force now to move the electrons, ).

Outside of the conductor, the field still exists and is uninfluenced by .

The process just described is almost instantaneous.

2) The charge density inside a conductor is zero ( ).

As we proved, inside the conductor , which implies that ,

and since it follows that .

3) Electric charges on conductors are distributed over their outer surface.

Indeed, this is the only place where charges can be. If a conductor is
charged in a static electric field, then the charge must reside on its surface and
since there is no other place to go (free space around the conductor is a
perfect insulator) it remains there forever. In general, the surface charge
distribution on the outer surface of a conductor is not uniform; it depends on
the geometrical shape of the conductor.

4) A conductor in Electrostatics is an equipotential body, i.e. any two

points inside or on the surface of a conductor have the same potential.

Indeed, if and are any two points inside a conductor, then from formula
(5-7-3) we have,

and this justifies our assertion, (notice that the integral is zero, since
everywhere inside the conductor).
 305

By virtue of this property, we may talk about “the potential of a

conductor”.

9-3) Boundary Conditions Between Conductors and Dielectrics

In a single medium the electric fields and are continuous functions of

the spatial coordinates. However, at the boundary between two different
media both fields may experience abrupt changes (discontinuities) in
magnitude and direction. Knowing how the fields change at the boundary
between two different media is important information, when solving various
problems in Electrostatics.

In this Section we shall study the boundary conditions between a

conductor and free space (which is considered to be a perfect insulator). In
Section 10-6 we shall study the boundary conditions between two different
dielectrics.

Fig. 9-2: Boundary conditions: conductor-free space.

In Fig. 9-2, let be the boundary between a conductor and free space. As
we have shown in the previous section, the electric field and the
field are zero, inside the conductor. The field on the boundary may be
decomposed into two components, one tangential ( ) and one normal ( ) to
the conductor surface, as shown in Fig. 9-2. We consider a small rectangle with
sides and , with being parallel to and being parallel to .
 306

Application of the fundamental law , where is the perimeter

of the small rectangle, results in

and assuming that , equation (*) implies that . The tangential

component of the field at the surface of a conductor is zero. In other words,
the field is always perpendicular to the surface of the conductor. We
expected that, since the field lines are perpendicular to the equipotential
surfaces, and since the surface of a conductor is an equipotential surface, the
field must necessarily by perpendicular to this equipotential surface.

Let us now consider a small cylindrical surface as the Gaussian surface, as

shown in Fig. 9-2. Application of the second fundamental law
, on the cylindrical surface and recalling that inside the
conductor and on the surface of the conductor, we find

In summary:

a) The electric field is always perpendicular to the surface of the

conductor, and

b) The normal component of the electric flux density at any

point on the surface of a conductor is equal to the surface charge density
at this point.

Example 9-3-1: Electric charge is uniformly distributed over the surface of

a conducting sphere of radius . Plot the potential and the magnitude of the
electric field as a function of .

Solution

The electric field of the given charge distribution is given by formula (3-3-
6), while the potential for the same distribution has been evaluated in
Problem 5-5-7.
 307

Fig. 9-3: Variation of the field and the potential versus .

The surface charge density is , as

expected.

Example 9-3-2: Given the electric field in free

space and the information that point lies on a conducting surface, find
and at .
 308

Solution

Since the point lies on a conducting surface,

and .

Example 9-3-3: A potential field in free space is given as

. The point is located at the conductor-
free space boundary. Find and at the point .

Solution

and at the point , .

Since the point lies on the conductor-free space boundary, ,

, and
.

PROBLEMS

9-3-1) Given the electric field in free space and

the information that point lies on a conducting surface find and at
the point , (Ans: ).

9-3-2) A potential field in free space is given as

. The point is located at the
conductor-free space boundary. Find and at the point .

9-3-3) Consider two conductors and with potentials and

respectively. Find the work needed to move a point charge
from to , (Ans: ).

Hint: Apply formula (5-7-2).

 309

9-4) Electrostatic Pressure

It is known that a conductor in a static electric field is subject to attractive

forces, under the influence of which the conductor tends to expand (increase
in volume). With the theory developed thus far, we are in a position to offer a
satisfactory explanation of this phenomenon.

Fig. 9-4: Electrostatic pressure.

Let be the unit out normal vector on the surface of the conductor. This
unit vector is perpendicular to the surface of the conductor and points from
the conductor towards the free space.
 310

As we know the field just immediately outside of the conductor is

If we assume that at a point the surface charge density is positive,

then the electric field (at the point ) will be and since the charge
, the force on is , i.e. points
outwards (is an attractive force).

If we assume that at a point the surface charge density is negative,

then the electric field (at the point ) will be , and since the charge
, the force on is ,

i.e. points outwards again (is an attractive force).

We have thus shown that in all cases, the electric forces tend to attract the
surface of the conductor, regardless of the sign of on the surface.

Let us now assume that under the influence of the attractive force a
small surface area experiences an infinitesimal displacement , as shown
in the second figure in Fig. 9-4. Then the volume of the (original) electrostatic
field will decrease by an infinitesimal volume and this means
that the original energy of the field will decrease by ,
where is the electric field at the point . This decrease in the electrostatic
energy should be equal to the work done by , i.e.

The electrostatic pressure at is defined as the force per unit area, i.e.

Equation (9-4-2) shows that the electrostatic pressure at the surface of a

conductor at a point lying on its surface is equal to the energy density of the
field at this point.
 311

Formula (9-4-2) may assume the following equivalent forms:

As a vector, the force per unit area is .

The total attractive force on the conductor, in principle, is found by

integration over the surface of the conductor, i.e.

(For an alternative derivation of (9-4-2), see Problem 9-2).

9-5) Cavities Within Conductors

Let us assume that within a conductor we have carved out a cavity, of an

arbitrary shape, and that there are no charges placed within the cavity.

We shall show that, inside the cavity, the electric field is zero, and
furthermore, the surface charge density on the surface of the cavity is zero.

Fig. 9-5: Cavities within conductors.

Let be the volume bounded by the Gaussian surface , which lies entirely
within the cavity.
 312

If we call and the potential and the electric field, respectively, inside
the cavity, we form the vector function and apply the divergence
theorem in the volume bounded by the Gaussian surface :

Making use of the vector identity

(since inside the cavity, by assumption), equation (*) implies,

Notice that formula (***) holds true for an arbitrary Gaussian surface, lying
entirely inside the cavity. If we now imagine that the Gaussian surface
increases continuously, until it coincides with the surface of the cavity, then
on this Gaussian surface the potential becomes equal to the constant
potential of the conductor, and as a constant, can be taken out of the
integral in (***), which now takes the form,

However, the surface integral in (****) is equal to the charge

enclosed by , which is zero (by assumption) and this leads to .
Since is a non negative quantity, the volume integral is zero if is
identically zero, everywhere inside the cavity.

Also, since , and as well, (on the surface of

the cavity).

If a charge exists inside the cavity, the situation is different, (Ex. 9-5-1).
 313

A Faraday cage (or shield) is a conducting enclosure used to block

Electromagnetic fields. The enclosure may be formed either by a solid
conducting material or by a mesh of conducting materials.

Example 9-5-1: A positive point charge is placed at the center of an

uncharged thick metal spherical shell, with inner radius and outer radius .
Find the electric field everywhere.

Solution

Fig. 9-6: Evaluation of the electric field versus .

Due to the spherical symmetry involved, the electric field is of the form
. Inside the cavity, , we do have an electric field, since we
have a point charge at the center, and the electric field is the field of a point
charge , which as we know is

Inside the conducting shell, .

Outside the conducting shell, , application of Gauss’s law over

the Gaussian surface yields,
 314

In summary:

PROBLEMS

9-5-1) In Example 9-5-1 find the surface charge densities on the surfaces
and . What is the total charge on the surface ? On the surface
? (Ans: ).

9-5-2) The cavity in Fig. 9-6 contains a uniform charge density .

Find the electric field everywhere.

9-5-3) In Problem 9-5-2 what are the surface charge densities on the
surfaces and ? (Ans: ).

9-6) Green’s Reciprocity Theorem

This theorem, which refers to a system of conductors in free space or in a

homogeneous dielectric (insulator), gives a simple equation relating two
electrostatic states of the system.

Let us for definiteness assume that we have conductors in free space. In

the first state we consider that the conductors are being charged with charges
. The electric field thus produced assigns potentials
on the respective conductors.

In the second state, we assume that the conductors are charged with
whereas the assigned potentials are respectively.

Green’s theorem states that

 315

or, in a compact form, using the notation

In order to keep our calculations as simple as possible, we shall show

Green’s theorem for , i.e. for four conductors; the generalization for any
number of conductors follows readily.

a) Let us consider four points in space, as shown in Fig. 9-7.

Fig. 9-7: Green’s reciprocity theorem.

As state 1, we consider four point charges placed at

respectively. In this charge configuration, we call the potential at the point
, (the position of ) due to all other charges (all except ); in other words,
is the potential at the position of due to the point charges and .
 316

Similarly, is the potential at the position of due to the point charges

and , etc.

If we replace by respectively, we obtain the

state 2 of the system. In this case we designate by the
corresponding potentials (defined exactly as in state 1).

If we call the distance between and , the distance between

and , etc, then we have:

Multiplying the first equation in (*) by , the second by , the third by

the fourth by , adding term wise the resulting equations and grouping
properly, we get:

and this finally simplifies to the following equation,

 317

b) Thus far, we have shown Green’s theorem for point charges. This
theorem can easily be extended from point charges to arbitrary conductors,
charged or neutral, in free space or in a homogeneous dielectric. As we know,
the charge on a conductor resides on its surface in the form of surface charge
density; we may thus assume that the charge of a conductor consists of a huge
number of infinitesimal charges , each one of which may be
treated as a point charge, and therefore equation (9-6-1) may be applied.
However, in this case, for the terms corresponding
to charges on the same conductor, the potentials ,
since the conductor is an equipotential body and this leads to the following
equation,

The corresponding term (the same conductor in the other state) would
likewise be , and this actually completes the proof of Green’s
theorem for conductors.

c) An interesting corollary of Green’s theorem: Assuming that we have two

conductors and , then equation (9-6-1) implies that

Assuming further that and that , formula (**)

leads to the unexpected result . This shows that the potential an
uncharged conductor ( ) acquires when another conductor ( ) is charged by a
charge , is equal to the potential of the conductor ( ), assumed to be
uncharged, when the first conductor ( ) is charged by the same charge .

d) In our so far analysis, we have shown Green’s reciprocity theorem for

conductors in free space or in homogeneous dielectrics. However, it can be
shown, that the theorem is still valid, even in case where the charges or the
conductors are within non homogeneous dielectrics.

e) One of the most important applications of Green’s theorem is the

computation of the “induced charges”, and this is the subject of the next
Section.
 318

9-7) Induced Charges

It is known that if a point charge is brought in the vicinity of an

uncharged conductor, then the two bodies attract one another. How do we
explain this, given that the conductor is assumed to be uncharged?

As the point charge approaches the uncharged conductor, the free

electrons inside the conductor, are attracted to the side of the conductor
which is closer to , as shown in Fig. 9-8.

Fig. 9-8: Separation of charges due to the field of a point charge.

As a result, we have a surplus of negative charge on the side of the

conductor closer to the charge and a deficiency of negative charge (i.e. a
surplus of positive charge) on the other side of the conductor. The distribution
of the charges on the surface of the conductor is such as to produce
inside the conductor. The surface charges on the surface of the conductor are
called “induced charges”. Since the negative induced charge is closer to
than the positive induced charge, the attractive force will be greater in
magnitude than the repulsive force; the net effect is an attractive force, which
explains the attraction between the two bodies. Same reasoning applies if a
negative charge – is brought close to the conductor. In this case, the positive
charges would be closer to the negative point charge than the negative
charges, and attractive forces would still prevail.

If, in Fig. 9-8, the point charge is removed, then the electrons will move
to the left as to neutralize the positive charges, (otherwise there would be an
 319

electric field inside the conductor, directed from the positive towards
the negative charges), something which is not allowed in electrostatics.

Let us now examine a slightly different situation. Let us assume that while
is held at the vicinity of the uncharged conductor, we connect the
conductor to the “Earth” by means of a metallic wire, or as we say we
“ground” the conductor, as shown in Fig. 9-9.

Fig. 9-9: Charging an uncharged conductor.

Then what happens is that since Earth has an unlimited supply of electrons,
electrons will rush into the conductor and will neutralize the positive charges.
If we discontinue the connection to the Earth, and remove the point charge
, the conductor now has a real surplus of negative charge, it becomes a
“charged conductor”.

For practical problems, we assume that the potential of the ground is zero,
so by saying “we ground a conductor” we mean that we assign potential zero
to the conductor.

Let us now consider the following problem, (computation of induced

charges).

Suppose that a point charge is placed at a point in space, close to ,

initially uncharged conductors, which are then grounded. As explained earlier,
“induced charges” of sign opposite to that of will now appear on the
conductors, (charges of the same sign with will escape through the
grounding wire to the Earth), and the problem is to compute the “induced
 320

charge ” on the conductor. Green’s reciprocity theorem provides a help

towards this direction.

As state 1, we consider the following situation: the charge is at position

, the induced charges are , while the potentials are at the
point and , since the conductors are grounded.

As state 2, we consider the following situation: the charge is removed

(i.e. ), the grounding of the conductor is removed, while all the other
conductors remain grounded. In this state, let be the potential of the
conductor, (the ungrounded one); the potential of all the other conductors will
still be zero (since they remain grounded) and let also be the potential at
the point . Finally, let be the induced charges on the
conductors, in this state.

Application of equation (9-6-1) yields:

Formula (9-7-1) shows that in order to compute the induced charge on

the conductor, it suffices to find the potential at the point (where the
point charge is initially placed), when the conductor is charged to
potential , while all the other conductors are grounded. Of course, the
point charge is assumed to be known.

This is a simple, general and efficient method for computing induced

charges on conductors, in the presence of a point charge. Of course, we still
have to compute the potentials and , but in many cases, this is easy to be
done.

For an application of this method, see Example 9-7-2.

 321

A general remark:

Fig. 9-10: Surface charge density on a conductor.

If the potential in space is a known function of the spatial coordinates,

then, in principle, the surface charge density is given by the formula

where is the normal derivative of at the surface (for the notion of

the directional derivative, see section 1-10). Of course, to find the potential of
a given charge distribution is not an easy task, especially if we are working in
two or three dimensions. Such types of problems will be considered in Chapter
12. But in any case, if we can, somehow, compute , then the charge on the
conductor will be .

In particular, when working in two or three dimensions, the potential

function is obtained by solving Laplace’s equation, or Poisson’s equation,
subject to the appropriate boundary conditions. This leads to a unique solution
for the potential .

The solution of Laplace’s or Poisson’s equation in two or three dimensions

is treated in Chapter 12.

Example 9-7-1: Consider two infinite conducting planes at and .

The plane is held at potential while the plane is grounded. Find
the potential and the electric field in the region between the planes,
(assume zero charge density everywhere between the planes).
 322

Solution

Fig. 9-11: Potential distribution between two conducting, parallel planes.

Since there is no variation in the and direction, the potential depends

only on the variable , i.e. . The potential must satisfy Laplace’s
equation , which in our case ( ), reduces to the ordinary
differential equation , the solution of which is

If we apply the Boundary conditions and we can

determine the constants of integration and , , , and

The electric field is now evaluated easily,

The electric field between the two plates is a uniform field (constant
magnitude and direction).

(Again, notice that the electric field points from higher towards lower
potentials).
 323

Example 9-7-2: Consider two infinite, grounded conducting planes at

and . A point charge is placed at a point between the two planes, a
distance from the plane . Find the induced charges and on the
planes and respectively.

Solution

Fig. 9-12: Computation of induced charges.

The induced charge on the grounded plane , is found by a suitable

application of formula (9-7-1), i.e.

If we assume potential on the plane, then according to formula

(**) in Example 9-7-1,

and formula (*) yields . Similarly we may find the charge

induced on the plane, , (let the reader verify the
calculations). Notice that (as expected).

PROBLEMS

9-7-1) Work Example 9-7-1, assuming ,

. Find and in the region between the two planes.

(Ans: ).
 324

9-7-2) In Example 9-7-2, assume and . Find

the induced charges and .

9-8) Energy of a System of Conductors

Let us assume that we have conductors in free space, of arbitrary shape,

carrying charges respectively. If is the potential of the first
conductor, the potential of the second, etc, then the electric energy stored
in the field of the conductors is given by the formula

We notice that this formula is similar to the formula giving the energy of a
discrete charge distribution, (equation (6-2-1)).

To justify equation (9-8-1) we recall that the electric field produced by the
charged conductors, is actually the field generated by the surface charges,
residing on the surface of the conductors. Also, each conductor is an
equipotential body. According to formula (6-3-4) the energy stored in the field
produced by the surface charges of the conductors, is

where and are the surface charge density and the potential of the
conductor, and since is constant on the surface of the conductor,
equation (*) implies that

since the charge of the conductor.

 325

SUPPLEMENTARY PROBLEMS

9-1) Explain why an uncharged piece of paper when released closed to a

static charge is attracted to it, but after touching it, is repelled, (think in terms
of Fig. 9-8).

9-2) In deriving equation (9-4-2) we have assumed that the charge changes
abruptly, from zero (inside the conductor) to the value on the surface.
However, this is not an accurate picture. What happens, is that the charge on
the surface of the conductor occupies a very thin layer of width , (a few
atomic diameters), as shown in Fig. 9-13.

Fig. 9-13: The charge layer on the surface of a conductor.

The field is normal to the surface of the conductor. Assuming to be the

charge density within the small cylinder in Fig. 9-13, Gauss’s law leads to the
equation . Now the charge within the cylinder, between the two layers
at and , is , and the force on this infinitesimal amount of
charge is , and the total force on the charge within the
cylinder is

This force is exerted on the charges within the infinitesimal cylinder, and
since these charges cannot escape from the surface of the conductor, it
appears, eventually, to be exerted on the surface of the conductor.
 326

The electrostatic pressure is .

9-3) Given the electric field in free space and the

information that a point lies on a conducting surface find and at
the point , (Ans: ).

9-4) A point charge is placed between two concentric conducting spheres,

with radii and , as shown in Fig. 9-14.

Fig. 9-14: Calculation of the induced charge.

If the outer surface is grounded, show that the induced charge

on the inner surface is .

Hint: Use formula (9-7-1).

9-5) A conducting sphere of radius receives a charge . (a) Find the

expression for the potential everywhere, (b) Find the electric field
everywhere and (c) Find the energy stored in the electric field of the
conductor.

(Ans: ,
 327

Hint: Integrate the energy density over the whole space.

9-6) The conductors and are held at potentials

and carry charges and
. Find the energy stored in the field of the conductors.

Hint: Apply formula (9-8-1).

9-7) A spherical conductor of radius , centered at the origin, has a cavity of

an arbitrary shape, carved out of it, as shown in Fig. 9-15. Assume the
conductor to be initially uncharged.

Fig. 9-15: Cavity with a point charge inside.

(a) A point charge is placed inside the cavity. Justify why the induced
charge density on the surface of the spherical conductor is uniform and find
its expression, (b) Will the surface charge density on the surface of the cavity
will be uniform ?, (c) If the point charge moves to another point within the
cavity, will that have an effect on the surface charge density of the spherical
conductor?

(Ans: (a) On the spherical conductor: , (b) No, (c) No, still
the surface charge density will be uniform).
 328

Hint: The field, just outside of the conductor will be normal to the
spherical surface. Then, since , show that this eventually implies that
must be constant on the spherical surface.

9-8) In Problem 9-7, assume that and . Find: (a) The

surface charge density on the surface of the spherical conductor, (b) The
potential of the spherical conductor and (c) The electric field at the point
.

Hint: To find the potential and the electric field, use the results obtained in
Problem 9-5.

9-9) In Problem 9-8, what is the work required to move a point charge
, from infinity to the point ?, (Ans: ).

9-10) Repeat Problems 9-8 and 9-9, provided that the spherical conductor is
given a charge .

Hint: Apply the superposition principle.

9-11) Find the total electric force on a spherical conductor of radius and
carrying a total charge , due to the electric field generated by .

(Ans: ).
 329

CHAPTER 10: DIELECTRICS IN STATIC ELECTRIC FIELDS

(POLARIZATION AND BOUND CHARGES)

In Chapter 9 we studied the behavior of conductors in electrostatic fields. In

this Chapter we shall study the behavior and the fundamental properties of
dielectrics (insulators) when placed in static electric fields.

While conductors contain a huge number of free or conduction electrons,

(typical value ), that can move freely through the
conductor, dielectrics do not contain free electrons. In dielectric materials,
the electrons are tightly bound, by means of electrical forces, to their parent
atoms or molecules. This basic difference between conductors and dielectrics
explains their different, macroscopic, electrical behavior. For example, while
the electric field inside a conductor is zero, this is not necessarily true for
dielectrics.

10-1) Polar and Nonpolar Atoms and Molecules

Dielectric materials are divided into two broad categories. In the first
category (nonpolar) belong these materials, every molecule of which, in the
absence of an external electric field, is electrically neutral. In the second
category (polar) belong these materials, the molecules of which, even in the
absence of an external electric field, form electric dipoles, (Section 5-6).

Fig. 10-1: Polar and nonpolar molecules.

In the second case, we say that the atoms or molecules are “electrically
polarized”.
 330

If we consider a very small volume , within the dielectric, which however

is large enough so as to contain a huge number of polar or nonpolar molecules,
then, due to the random, thermal motion of the molecules, the net charge of
the volume is practically zero, assuming of course that there is no external
electric field applied.

Notice that in many dielectrics, polar and nonpolar molecules may exist
simultaneously.

10-2) Dielectrics in the Presence of an Electric Field

Let us now examine what happens when a dielectric material is placed

within a static electric field.

a) A nonpolar molecule, under the action of an externally applied field

undergoes a separation of its charges, since the positive charge is pushed in
the direction of the field , while the negative charge is pushed in the opposite
direction. The atom or molecule is still electrically neutral, but now, under the
influence of the electric field there is a separation of the positive and
negative charges. The atom or molecule becomes an electric dipole. This
phenomenon is called “polarization”. As a result of the polarization, the
nonpolar atom becomes “polarized”, i.e. becomes an electric dipole and as
such, it acquires a dipole moment , where is the vector displacement
from – towards , (see equation (5-6-2)). Even though the displacement
is extremely small, the cumulative effect of a huge number of these tiny
dipole moments is responsible for the behavior of the dielectric materials. An
estimation of the displacement is given in Example 10-2-1.

Notice that, if the applied field is strong enough, then, from a certain
point on, this might lead to a violent extraction of the electrons (negative
charge) from the parent atom or molecule. In this case we have a “dielectric
breakdown”, the bound electrons become “free electrons” and the dielectric
becomes a conductor.

In summary, for relatively weak electric fields, dielectric materials get

polarized when placed inside an electric field .
 331

Fig. 10-2: A non polar molecule in externally electric field.

b) Let us now examine the behavior of an electric dipole in an electric field.

Fig. 10-3: Stable and unstable equilibrium of electric dipoles.

In Fig. 10-3 we see that in general, the dipole is in equilibrium when the
electric dipole is parallel to the applied field . However, due the
everlasting thermal, random motion of the atoms and molecules inside the
 332

dielectric material, the vector will slightly deviate from its equilibrium
position, (imagine that oscillates up and down from its equilibrium position).
Then, as it is shown in the first figure in Fig. 10-3, if is anti parallel to , then
there is a torque tending to rotate clockwise, so that is aligned to , i.e. so
that the vectors and become parallel and of the same direction. This
equilibrium position in the first figure is an “unstable equilibrium”, since the
slightest deviation from the equilibrium generates a torque which does not
allow the atom or molecule to return to its initial equilibrium position.

In the second figure, when and are parallel and of the same direction,
“the equilibrium is stable”, since the slightest deviation from the equilibrium
generates a torque tending to restore the system back to its initial
equilibrium position.

In conclusion, we may say that when a dielectric material is placed within an

electric field then all the tiny dipole moments (which already exist if the
material consists of polar atoms or the ones which are induced if the material
consists of nonpolar atoms as explained in part (a)), are aligned to the
externally applied field . Of course, the random thermal motion, makes these
dipole moments to swing about their equilibrium position, but still, now, there
is a dominant direction of orientation, (all are aligned to the field ).

c) Let us now consider a simple example, in order to understand, physically,

what happens when a dielectric material is placed within an electric field .

Let be the electric field in free space, generated by some charge

distribution. Suppose now that we place a dielectric material within this
electric field, and that we want to examine the electric field inside the
dielectric.

As explained in part (b) above, all the dipole moments will be aligned to
the field , so the picture, inside the dielectric, will look as shown in Fig. 10-4.
Assuming for simplicity that the dielectric is homogeneous, we have
accumulation of negative charges in the left side and accumulation of
positive charges on the right side. This separation of charges (which did not
exist before we place the dielectric in the electric field ), gives rise to a new
electric field, an induced electric field , inside the dielectric, due to the
 333

polarization. So the total electric field inside the dielectric will be the
superposition of the initial field in free space and the induced electric field
(due to the polarization of the bound charges inside the dielectric), i.e.

Fig. 10-4: Electric field inside the dielectric.

As we have shown the electric dipole is aligned to the field and thus,
the charge configuration is as shown in Fig. 10-4. At the position (a) there is
cancelation of the positive charges from the adjacent negative ones, and the
same happens at position (b), (assuming that the dielectric is homogeneous).
We are thus left with negative charges on the left side (since there no
positive adjacent ones) and with positive charges on the right side. This, in
turn, generates an induced field inside the dielectric, which is then
superimposed to the existing field in free space , to yield the total electric
field inside the dielectric as given in equation (9-2-1).

Notice that if the dielectric is not homogeneous, then, we may not have
cancelation of charges at the positions (a), (b), etc, and there will be additional
fields generated, due to the accumulation of bound charges within the
dielectric, which must also be included in the superposition. A systematic
method to include all the effects of the polarization will be presented in the
next section.
 334

Example 10-2-1: A crude model of an atom consists of a heavy positive

nucleus ( ) surrounded by a uniform negative cloud – ), of radius , (see
Notice in Example 4-5-2). Calculate the displacement when an electric field
applies.

Solution

Fig. 10-5: Separation of positive and negative charges in an atom.

With zero field applied the “centers” of the positive and negative charges
coincide. However, when a field applies, there will be a charge separation,
since the positive nucleus is pushed in the direction of , while the negative
charge is pushed in the opposite direction. Since the separation, in general, is
very small, we may assume that the negative cloud retains its spherical shape
as it moves to the left. Equilibrium is established when , where is
the force on the positive nucleus due to the applied field , while is the
force on the positive nucleus due to the negative electronic cloud.

In Example 4-5-2 we found that the electric field produced by a charge ,

uniformly distributed within a sphere of radius , is given by
 335

Assuming (the radius of the atom), , we find

from (**) that the ratio

Even with a relatively high value of , which

shows that the displacement is about times smaller than the
atomic radius.

10-3) The Field of a Polarized Dielectric

Having acquired a physical knowledge as to what happens inside a dielectric

material we proceed further to a more systematic method in order to
compute the field of a polarized dielectric. Recall that when a dielectric
material is placed inside an electric field it becomes polarized, that is a huge
number of tiny dipole moments are aligned to the field . As we shall show
below, this effect has some far reaching consequences, regarding the electrical
behavior of the dielectric.

Let us consider a small volume inside the dielectric and let us assume
that within this small volume there are electric dipoles . We
may define “the average dipole moment per unit volume ” as follows:

The physical meaning of the limit in (10-3-1), is that the volume should
be very small, but still, at the same time, be large enough so as to include a
huge number of tiny dipole moments . The vector is called “the
polarization vector” and accounts for the average orientation of the atomic
dipole moments inside . Notice that the unit of is the same as
that of the “electric flux density ”.

Let us now assume that a dielectric material exists within a volume

bounded by a surface , as shown in Fig. 10-6 and that is a known function
 336

of the spatial coordinates , (we use barred coordinates for the source
points and unbarred coordinates for the field points, see Section 1-13).

Fig. 10-6: The field of a polarized dielectric.

Considering ( ) to be an electric dipole placed at , the potential at

the point , caused by this dipole, is given by equation (5-6-3),
i.e.

and the total potential at is found by integration over the source

points, i.e.

Notice that in (10-3-2) we use the permittivity of free space , since the
charges due to the polarization are taken into consideration in the vector .

In order to transform the volume integral in (10-3-2) in a more convenient

form, we notice that
 337

and formula (10-3-2) becomes:

In (***), recall that is a function of the barred coordinates, .

Now we consider the vector identity

and substituting , as obtained from the identity, in (***) we have:

Transforming the first integral to a surface integral, by using the Divergence

theorem, we get,

Equation (10-3-3) admits a simple physical interpretation:

a) The first term corresponds to the potential due to a surface charge

density

b) The second term corresponds to the potential due to a volume charge

density

where the subscript stands for bound or polarization charges (notice

that we have switched back to the unbarred coordinates, for convenience).
 338

Equations (10-3-4) and (10-3-5) show that the potential of a polarized

dielectric is the same as that produced by a volume charge density and a
surface charge density . Once the potential is determined, then
the electric field .

10-4) The Field Inside a Dielectric, the Displacement Vector

Let us assume, once more, that we have an electric field in free space
and then we place a dielectric material within the electric field. The main
problem is to determine the field inside the dielectric. Recall that when the
dielectric is placed within an electric field, the dielectric gets polarized; the
accumulation of bound charges produces a volume charge density
within the dielectric and a surface charge density , on the surface
of the dielectric, ( points from the dielectric towards the free space).

The vector field was originally defined, in free space, by the equation
, where is the permittivity of free space. Gauss’s law, (eq. (4-4-7)),
states that , where is the volume charge density. This
charge density was deliberately placed in space, and it is actually the source of
the electric field. This charge does not include bound charges, since in free
space, there are no bound charges to be polarized.

In the presence of a dielectric material however, the total charge density

inside the dielectric is the sum of two terms, one term being the initial charge
density attributable to all charges except bound charges, plus the charge
density caused by bound charges alone, i.e.

and therefore, Gauss’s law has to be modified accordingly, so as to include

the bound charges as well, i.e.

If we now define the displacement vector in the presence of a dielectric

as
 339

then formula (10-4-2) assumes again the form

where is the volume charge density due to all charges except the bound
charges. This is very important, since we do have control over , (we may
deliberately place charges in space), but we do not have any control on the
bound charges (which are induced when a dielectric is placed in an electric
field). For homogeneous and isotropic dielectrics (and not very strong electric
fields), the following linear relationship exists between and ,

where is the “electric susceptibility of the dielectric”. Using this

relationship in equation (10-4-3), we have

The expression within the parentheses is defined to be the “relative

permittivity of the dielectric ” , i.e.

and equation (10-4-6) assumes the form

where is the permittivity of the dielectric. This expression for the

permittivity was introduced in Section 2-4, (equation (2-4-2). Typical values of
the relative permittivity for some important for applications dielectrics are
given in the Table in Section (2-4).

For anisotropic materials, the relation between and is not so simple, in

such cases the permittivity turns out to be a matrix, which means that
the , for example, component of the field , may depend on and
and , (see remark in Section 4-3).

The integral form of equation (10-4-4) is

 340

where the total charge enclosed contains all charges except bound
charges. Formula (10-4-9) results easily from formula (10-4-4), by using the
divergence theorem.

In summary: The effects of the polarization or bound charges induced when

a dielectric material is placed inside an electric field are described in terms of
the “permittivity ” of the dielectric, which for many dielectrics can be
determined experimentally. All the effects of the polarization are included in
the permittivity .

All formulas obtained for fields in free space, are still valid if the
permittivity of free space is replaced by the permittivity of the dielectric.
For example, the energy density in free space is while the energy
density in a dielectric of permittivity is ; Poisson’s equation in
free space is while in a dielectric of permittivity Poisson’s
equation is , etc.

A general remark: When we want to determine the field inside a dielectric,

usually we determine first the field by using Gauss’s law in integral form
(equation (10-4-9)), and then find . Recall that the
includes all charges except the bound charges.

The following examples illustrate the main points of our analysis.

Example 10-4-1: A point charge is placed at the center of two

concentric spheres of radii and respectively, ( ). The space between
the two spheres is filled with a homogeneous dielectric with permittivity .
Find the electric field everywhere.

Solution

The only real charge (not bound) in our problem is the point charge ,
placed at the center. Due to the spherical symmetry involved, the field
depends only on , the distance from the center, and points radially outwards,
 341

i.e. . As Gaussian surfaces, we therefore choose spherical surfaces,

concentric with the two spheres, as shown in Fig. 10-7.

Application of Gauss’s law (equation (10-4-9) yields

Gaussian
surfaces

Fig. 10-7: The electric field inside a dielectric.

We notice that in equation (*) depends only on and .

Case 1: , (free space).

Case 2: , dielectric material with dielectric constant .

 342

Case 3: (free space).

In summary:

Example 10-4-2: In Example 10-4-1, find the polarization inside the

dielectric.

Solution

From equation (10-4-3) we have

and using the expression for found in Example 10-4-1, equation (*****),
we find the following expression for :

Of course, the polarization is zero exterior to the dielectric (free space)

since in free space there are no charges to be polarized. Inside the dielectric
the polarization is given by the second equation in (*).

The volume charge density of the bound charges, inside the dielectric is
given by equation (10-3-5), i.e.
 343

This means that, within any infinitesimal volume , inside the dielectric,
the elementary electric dipole moments are oriented so that the of one
dipole cancels the – of an adjacent dipole.

However, at the boundary of the dielectric, ( and ), we should

expect surface bound charges, since outside of the dielectric there exists free
space, and the bound charges, for example, very close to the surface ,
and inside the dielectric, do not have any adjacent charges in free space to
cancel.

By using formula (10-3-4) we have, .

On the surface , (see Fig. 10-7), and

Similarly, on the surface we have,

Example 10-4-3: The polarization vector within a homogeneous dielectric,

at a point is given by , while the
component of at the same point is . Find the electric
energy density at the point .

Solution

Using equation (10-4-5) we have:

Also, using equation (10-4-3) we obtain

 344

This is the magnitude of the electric field at the point .

The energy density at the point is

Example 10-4-4: A long coaxial cable with its axis coincident with the axis
is partially filled with a dielectric of relative permittivity . The cross
section of the coax is shown in Fig. 10-8. The surface charge density on the
inner conductor ( ), is .

Fig. 10-8: Cross section of the coaxial cable partially filled with dielectric.

Assuming find: a) the electric field inside

the cable, b) the surface charge density on the outer conductor, c) the
polarization inside the dielectric, d) the charge density of the bound charges,
e) the surface charge density of the bound charges on the boundary between
 345

dielectric and free space ( ), and f) the voltage between the two
conductors.

Solution

a) Due to the cylindrical symmetry involved, (assuming that the length of

the wire is much greater than ), and . Taking as
Gaussian surfaces cylindrical surfaces, coaxial to the axis, and applying
Gauss’s theorem in integral form, yields the following expression for , (let the
reader verify the result),

and the electric field is given by the formula:

Within the inner conductor ( ), .

b) The surface charge density on the outer conductor is

c) Inside the dielectric ( ), the polarization is

d) The charge density of bound charges, inside the dielectric is

e) The surface charge density on the boundary “dielectric-free space”

( ) is
 346

f) The voltage between the two conductors is

PROBLEMS

10-4-1) The polarization within a homogeneous dielectric with is

. Find and .

(Ans: ).

10-4-2) The relative permittivity of a homogeneous dielectric is . If

in the dielectric, find and .

10-4-3) A dielectric material in the form of a small cube of volume

and relative permittivity , is placed perpendicularly to a uniform field
. Assuming that the dielectric is polarized uniformly, find the
polarization vector and the moment of the electric dipole.

(Ans: ).

10-4-4) A perfect dielectric contains . Assuming that

the electric susceptibility is with find: a) ,
b) the polarization vector , c) the atomic dipole moment and d) the
effective dipole length.

10-4-5) A conducting sphere of radius carries a charge uniformly

distributed over its surface. The sphere is surrounded, out to a radius , by a
homogeneous dielectric of relative permittivity . Find: a) the field
 347

everywhere, b) The field everywhere, c) the surface charge density on the

surface of the conductor, d) the surface charge density on the boundary
“dielectric-free space”, e) the volume charge density of bound charges inside
the dielectric, f) the polarization vector inside the dielectric and g) The
potential of the conducting sphere, (assuming ).

(Ans: a) for , (inside conducting sphere, ),

b) in the dielectric , in free space , c) surface charge
density on the conductor , d) , e) , f)

, g) .

10-4-6) The region between two concentric, conducting spherical surfaces

of radii and is filled with two dielectrics with relative permittivity and
, as shown in Fig. 10-9. A charge is uniformly distributed over the inner
spherical surface. In the space between the two conducting spheres find: a) the
field , b) the polarization , c) the potential between the two conducting
spherical surfaces, d) the surface charge density at the boundary between the
two dielectrics ( ), e) the volume charge density of the bound charges, f)
the energy density and g) the total energy stored in the region .

Fig. 10-9: Two dielectrics between the spherical surfaces.

 348

10-4-7) The electric susceptibility of a homogeneous dielectric is

with . Find and .

(Ans: ).

10-4-8) In a homogeneous dielectric with the field

. Find and .

10-5) Properties of the Fields and in the case of Non Homogeneous

Dielectrics

a) Recall that, in general, the static electric field is irrotational, that is

(the fundamental property of the static electric fields), while
, where is the volume charge density of all charges, except bound
charges, which are induced when the dielectric is placed within an electric
field. Gauss’s law in point form ( ) holds true for static and time
varying fields.

In a homogeneous dielectric ( ), and since is

constant,

which shows that the field is also irrotational.

Also, in this case,

b) Non homogeneous dielectrics:

In this case, the permittivity is a function of the spatial coordinates, and

application of Gauss’s law implies,

Using the vector identity , (for a proof see

Problem 10-5-1), with and , equation (***) yields,
 349

This equation shows that the sources (or sinks) of the field are found not
only at the places of volume charge densities inside the dielectric, but also at
the places where the dielectric is non homogeneous, i.e. at places where
. For homogeneous dielectrics, equation (10-5-1) reduces to
equation (**).

Also, regarding the curl (rotation) of the field we note that in a non
homogeneous dielectric, . To justify our assertion, we consider the
curl of , which is

and using the vector identity , (for a

proof see Problem 10-5-2), equation (****) yields

Equation (10-5-2) shows clearly that, in general, the field is not

irrotational, due to the term . In a homogeneous dielectric, equation (10-
5-2) reduces to equation (*).

Example 10-5-1: A region in space is filled with a non homogeneous

dielectric with . Assuming and that at the point
the electric field , find and at the
point .

Solution

From equation (10-5-1), with , we have:

 350

The is obtained from equation (10-5-2),

which at the point is found to be

(let the reader verify the calculations).

PROBLEMS

10-5-1) Show the vector identity , ( is a

scalar field, is a vector field).

10-5-2) Show the vector identity .

10-5-3) Show that equation (10-5-2) may be expressed equivalently as

.

10-5-4) A region in space is filled with a non homogeneous dielectric with

. Assuming and that at the point the
electric field , find and at the point .

Hint: See Example 10-5-1.

10-6) Boundary Conditions between Perfect Dielectrics

a) In Section 9-3 we have established the boundary conditions at a

conductor-perfect dielectric boundary, (or conductor-free space
boundary). In this Section we shall derive the boundary conditions at the
boundary between two perfect dielectrics with permittivity and
respectively.
 351

Fig. 10-10: Boundary conditions between two dielectrics.

Working exactly as we did in Section 9-3 we find that

The first equation in (10-6-1) shows that the tangential component of the
field is continuous across the boundary between the two dielectrics. This
condition results from the fundamental equation , when applied
around the small rectangle in Fig. 10-10.

The second equation in (10-6-1) shows that the normal component of the
field is, in general, discontinuous. The “discontinuity jump”, is equal to the
surface charge density at this point. This condition results from Gauss’s law
applied in the infinitesimal cylinder of Fig. 10-10.

The surface charge density cannot be bound charge, since the effects of
bound charges are taken into account in the vector ; includes all charges
except bound charges. In other words, has been placed there deliberately,
by an external agent. This rarely occurs in practice, so we may safely assume
that, in most cases, , and under this assumption, the second equation in
(10-6-1) takes the form
 352

Equation (10-6-2) shows that the normal component of the field is

discontinuous across the boundary between two different dielectrics, as
opposed to the tangential component of the field , which is continuous.

b) Refraction of the electric field lines at the boundary between two

dielectrics.

Equations (10-6-1) and (10-6-2) may be combined to show how the field
lines are refracted at the boundary between two dielectric materials. Recall
that, by definition, the field lines are always tangent to the field.

Let us, for definiteness, consider Fig. 10-11, where at the point on the
boundary, the electric field is in the dielectric and is in the dielectric
.

Fig. 10-11: Refraction of the field lines between two dielectrics.

We assume that on the boundary. Let and be the angles

between the normal to the surface and the vectors and respectively.
Equations (10-6-1) and (10-6-2) may be used in order to find the relation
between and .
 353

Dividing (*) and (**) term wise, results in

Also, squaring (*) and (**) and adding term wise, results in

These equations allow us to find the fields and on one side of the
boundary, if we know the corresponding fields on the other side.

Also equation (10-6-3) shows that if , then (why?), which

shows that the field diverges more from the normal to the surface as it
enters into regions of increasing permittivity.

c) Contact voltage between two dielectrics.

An important consequence of the continuity of the tangential component of

the electric field at the boundary between two dielectrics is that the
potential difference across the boundary is constant. This constant potential
difference is called “the contact voltage” and in general, it depends on the
temperature and the properties of the dielectrics.

Let us, for simplicity, assume that the boundary between the two dielectrics
in Fig. 10-11, is the plane. Then equation implies that

If we call and the potential in regions ( ) and ( )

respectively, across the boundary, then equations (***) imply that
 354

Equation (****) shows that the difference of potentials retains

the same value at all points on the boundary, (the contact voltage).

In most cases of practical interest, the contact voltage is very small, we may
approximately take it to be zero, and in such cases . This means that
the potential function is uniquely defined and continuous in all
regions occupied by dielectrics.

Example 10-6-1: In Fig. 10-11, . If the angle of incidence is

find the refraction angle .

Solution

Application of formula (10-5-3) yields

Example 10-6-2: In Example 10-6-1, assume that and

. Find and .

Solution

Since it follows that .

Application of formula (10-6-4) yields,

In the dielectric : .
 355

In the dielectric : .

Example 10-6-3: Region 1 ( ) contains a dielectric with , while

region 2 ( ) contains a different dielectric with . If
, find and , (assume on the boundary).

Solution

Let and
.

On the boundary we must have:

Also, on we must have,

PROBLEMS

10-6-1) In Example 10-6-3 determine the angle of incidence and the

angle of refraction , and verify that they satisfy equation (10-6-3).

(Ans: ).

10-6-2) Three infinite, homogeneous and isotropic dielectric slabs with

relative dielectric constants and , are placed as shown
in Fig. 10-12. If , find , (assume
on the boundaries). What is the angle between and the axis?
 356

Fig. 10-12: Three dielectric slabs.

10-6-3) The electric field at a certain point within a homogeneous

dielectric material with is . a) Find the
electric susceptibility of the dielectric and b) Calculate and at the point
of question, (Ans: ).

10-6-4) In Example 10-4-1 show that the and fields obtained, satisfy the
proper boundary conditions.

10-6-5) Region 1 ( ) contains a homogeneous dielectric with ,

while region 2 ( ) is free space. At the boundary “dielectric-free space”
. If find
at the boundary.

(Ans: ,

).

10-6-6) The potential in a homogeneous dielectric with is given

by . Find and in the material.

Hint: .
 357

10-7) Ferroelectric Materials

There are some materials in which the polarization vector is permanent.

These materials are called ferroelectrics. In ferroelectric materials the
relationship between the applied field and the polarization is non-linear,
and furthermore, the polarization caused by a given electric field depends on
the past history of the material. As we say, the ferroelectric material shows
“hysteresis effects”. In practice, there are a few special materials, like the
barium titanate and the Rochelle salt, which exhibit permanent electric
polarization . If a ferroelectric bar is placed within a field , the bar will
experience a torque, which tends to align the polarization vector to the
externally applied field .

SUPPLEMENTARY PROBLEMS

10-1) Region is filled with a homogeneous dielectric with ,

while the region is free space. A uniform electric field of magnitude
and in a direction and exists in the region
. Find and , everywhere in rectangular coordinates, (assume
on the boundary).

(Ans: a) In free space (region 1, ):

b) In the dielectric region (region 2, ):

Hint: .

10-2) Find the susceptibility of a dielectric material for which the flux
density is seven times its polarization.

10-3) A dipole is located symmetrically on the axis. An

electric field is applied. Find the energy required to rotate the
dipole to its stable position, (Ans: ).
 358

10-4) Show that the energy density expended in polarizing a dielectric is

.

Hint: .

10-5) A linear, homogeneous and isotropic dielectric has a relative

permittivity of . If inside the dielectric, find
.

(Ans: ,
).

10-6) An infinite slab is surrounded by free space, as shown in Fig. 10-13.

Assuming that the field inside the dielectric is ,( )
find the field in regions 1 and 3.

Fig. 10-13: Infinite slab surrounded by free space.

10-7) The polarization vector within an homogeneous dielectric having

is . Find and . If
and are two points within the dielectric
find the voltage .

(Ans: ,
).
 359

10-8) A region in space is filled with a non homogeneous dielectric with

. Assuming and that at the point the
electric field , find and at the point .

Hint: See Example 10-5-1.

10-9) The plane is the separating surface between free

space (Region 1, the region containing the origin ) and a
homogeneous dielectric with (Region 2). Given that the electric field in
region 1 is find: and the energy
density in both regions. Assume on the separating surface.

(Ans: ).

Hint:

Fig. 10-14: Plane (ABC): Separating surface between two dielectrics.

The unit out normal (pointing from free space towards the dielectric) is
given by the formula, .
, etc.

10-10) Three infinite slabs of dielectrics are surrounded by free space, as

shown in Fig. 10-15. In region 1, . (a) Find the
fields and in all regions, (b) What is the angle formed between the
field and the axis? (c) What is the angle formed by the two fields and
 360

? (d) What is the angle formed between the two fields and ? (Assume
that surface charge density ).

Fig. 10-15: Three infinite slabs surrounded by free space.

10-11) Region is free space with ,

while is filled with a homogeneous dielectric with relative permittivity
. Determine in the region .

(Ans: ).

10-12) If is the charge density introduced inside a homogeneous dielectric

with relative permittivity (not including bound charges) and is the bound
charge density, show that .
 361

CHAPTER 11: CAPACITORS

11-1) General Concepts and Definitions

a) A capacitor (also known as a condenser) is an electric device for storing

charge (and therefore energy), and consists of two conductors of arbitrary
shape, known as “the plates” of the capacitor, surrounded by a dielectric
material with dielectric constant , as shown in Fig. 11-1. The plates carry
charges of equal magnitude but of opposite sign. The positive plate carries a
charge , while the negative plate carries a negative charge . The total
charge of the system is zero.

Fig. 11-1: Capacitor.

As we know, the conductors and are equipotential bodies, and the

surfaces and are equipotential surfaces. The charge on the positive
plate appears as a positive, surface charge density and similarly, the charge
on the negative plate appears as a negative surface charge density. The field
lines emanate from the positive plate, perpendicularly to the equipotential
surface and sink in the negative plate, perpendicularly to the equipotential
surface .

The capacitance of the capacitor is defined as the ratio of the charge

of the positive plate to the potential difference (electric voltage) between
the conductors and , i.e.
 362

From this definition the capacitance is a positive quantity.

The unit of the capacitance, as it results from the definition (11-1-1) is

. This unit is called “Farad” so the capacitance is measured in Farads.
Since common values of the capacitance are very small fractions of a Farad,
more practical units for the capacitance are submultiples of the Farad, i.e.
, or , or even .

b) Let us assume that the potential function of the electric field

generated by the charges and is a known function of the spatial
coordinates. Then, the electric field .

The surface charge density on the positive conductor , is (see eq. (9-3-1))

where is the unit out normal to the surface . Also, the voltage between
the two conductors is (see eq. (5-7-3))

Notice that integral (**) does not depend on the path; we may consider
any, arbitrary path, connecting a point of the conductor to a point of the
conductor .

By virtue of (*) and (**), equation (11-1-1) assumes the form

Notice that the capacitance is independent of the total charge and the
potential . If the surface charge density increases by a factor of , then
Gauss’s law implies that the flux density increases by , and this in turn implies
that the field increases by the same factor as well, and formula (11-1-2)
shows that the capacitance remains unchanged. The capacitance depends only
 363

on the geometry of the conductors and the dielectric constant of the

surrounding dielectric.

c) The capacitance of a single conductor.

In Fig. 11-1, if we imagine that the conductor is removed to infinity, then

we are left only with the conductor . Then becomes just , i.e. the
potential of the conductor with respect to infinity.

In this case we may define “the capacitance of a conductor” as

where is the charge of the conductor and is the potential of the

conductor with respect to infinity.

In the following sections we shall study the three most common types of
capacitors, the parallel plate, the spherical and the cylindrical capacitor.

In very general terms, to calculate the capacitance of a given capacitor,

we have to solve the equation for the potential , and this method is
illustrated in the subsequent sections. Once the potential function is
determined as a function of the spatial coordinates, then, in principle, we may
determine the electric field , and then, application of formula (11-1-2) yields
the sought for capacitance of the capacitor.

11-2) The Parallel Plate Capacitor

a) Let us consider two equal, parallel plates of area , a distance apart,

filled with a homogeneous dielectric with a dielectric constant , as shown in
Fig. 11-2. Assuming that the linear dimensions of the plates are much greater
than their separation distance , the field between the plates will be
uniform.

To show this, we assume that the plate is held at a positive potential ,

while the plate is held at potential zero (is grounded).
 364

Fig. 11-2: A parallel plate capacitor.

Since the potential between the two plates depends only on the variable,
i.e. , Laplace’s equation becomes, (see also Example 9-7-1),

The solution of this equation is

where and are the arbitrary constants of integration. To determine

these constants we apply the boundary conditions and
and we find easily that and . The potential
between the plates is

The electric field is determined as the negative gradient of , i.e.

Assuming dielectric constant between the plates, the surface charge

density on the positive plate is
 365

and formula (11-1-1) implies,

This formula shows explicitly that the capacitance of a parallel plate

capacitor depends solely on the geometrical characteristics of the capacitor
and the dielectric constant. If there is no dielectric between the plates, then
.

b) The energy density inside the capacitor is

and the total energy stored in the capacitor is

Formula (11-2-2) is important, since it expresses the energy stored in a

capacitor in terms of the capacitance.

Comments:

1) It may seems that we have restricted our problem by assuming potential

at the plate and potential zero at the plate . What if plate were held at
potential and plate at potential ? In this case the expression for
the potential would be different than the one in equation (**), (same
differential equation but different boundary conditions), but the electric field
would be identical to that in equation (***), where now , (the
voltage between the two plates). Let the reader verify it.

2) In a parallel plate capacitor, we have the so called “fringe effect”. In Fig.

11-2 we have assumed that the electric field is uniform inside the capacitor
and is zero outside. This is, of course, a reasonable assumption, provided that
 366

the linear dimensions of the plates are much greater that their separation
distance. However, strictly speaking, this is not entirely correct.

Fig. 11-3: Fringing effects at the ends of a capacitor.

Let us consider the capacitor in Fig. 11-3. The electric field at the open ends
of the capacitor, deviates from uniformity, and looks like as shown in the
figure. This is an inevitable consequence of the fact that the field is
irrotational. For if the field were uniform inside and terminated abruptly
outside, then the line integral around the rectangle would
be strictly positive, (the contribution would be from the integration along
only, the contribution from the path would be zero, if the field were zero
outside). But this violates the fundamental law that must be zero,
around any closed path . So the field outside the capacitor, should be as
shown in Fig. 11-3, so that the positive contribution is cancelled by
the negative contribution , (notice that on , and are
opposite vectors, and their dot product is a negative number).

3) The fundamental formula (11-2-2) may be considered as an alternative

definition for the capacitance, i.e.

Several times, this formula provides an easy and straightforward method to

calculate the capacitance, especially in cases where we have multiple
dielectrics. The idea is to determine the electric field , then find the energy
 367

density , integrate over the whole volume to find the total energy stored in
the field of the capacitor, and finally obtain , using equation (11-2-3). The
following examples illustrate the method.

Example 11-2-1: a) Find the relative permittivity of the dielectric material

used in a parallel plate capacitor if , b)
For an applied voltage , what is the charge of the capacitor? What
is the energy stored in the capacitor?

Solution

The uniform field inside the capacitor is ,

directed from the positive plate towards the negative. The surface charge
density is

The charge of the capacitor is

The energy of the capacitor is

Example 11-2-2: The capacitor in Fig. 11-4 contains two different dielectrics
with permittivity and respectively. Find its capacitance .

Solution

If we call the potential in region 1 (dielectric ) and

the potential in region 2 (dielectric ), then both and must satisfy
Laplace’s equation (since we have not placed any free charges inside the
dielectrics), i.e.
 368

Fig. 11-4: A plate capacitor with two dielectrics.

The solutions of the equations in (*) are:

where the arbitrary constants of integration , must be determined

from the appropriate boundary conditions, which are:

The second equation in (***) results from the continuity of the potential at
the interface between the two dielectrics, (we assume zero contact voltage,
see Section 10-6), while the fourth one expresses the continuity of the normal
 369

component of the field , since there are no free charges on the boundary. In
principle we are done. We have four equations to determine the four arbitrary
constants in equation (**). Straightforward calculations show that (for detailed
calculations see Problem 11-2-3),

and the potential functions (and the electric fields) in regions 1 and 2 are:

The energy density in regions 1 and 2 is,

The total energy stored in the capacitor is

Remark: We may express formula (******) in the following form:

 370

or,

This formula is known as the formula for the total capacitance of two
capacitors connected in series. It can obviously be generalized to any number
of capacitors in series.

PROBLEMS

11-2-1) A parallel plate capacitor has , and is filled

with a dielectric . Find its capacitance , (Ans: ).

11-2-2) A voltage applies on the capacitor of problem 11-2-1. Find

the charge of the capacitor , the surface charge density on the positive
plate and the total energy stored in the capacitor.

11-2-3) In Example 11-2-2, perform analytic calculations to derive the

constants of integration and .

11-2-4) Find the capacitance of the capacitor shown in Fig. 11-5.

Fig. 11-5: A plate capacitor with two dielectrics.

(Ans: , where and ).

 371

Hint: Consider Laplace’s equation in each region and apply the appropriate
boundary conditions, (see Example 11-2-2).

11-2-5) Two conductors and with capacitances and , carry charges

and respectively. When the two conductors are connected with a
metallic wire, they acquire the same potential , (why?). a) Find , b) If
is the energy of the system before the connection and is the energy after
the connection, show that , What happens to the energy ? c)
Under what conditions is ? Justify your answer.

(Ans: ).

11-2-6) For the capacitor in Figure 11-4 assume,

. Find and the energy
stored in the capacitor.

11-3) Spherical Capacitors

Let us consider a spherical capacitor formed of two concentric spherical

conducting shells of radii and , ( ), as shown in Fig. 11-6.

Fig. 11-6: Spherical capacitor.

The space between the two conducting shells is filled with a homogeneous
dielectric with permittivity . The outer shell is held at potential , while the
inner shell is grounded.
 372

Due to the spherical symmetry the potential function will be a function of

the radial distance only, i.e. . In this case, Laplace’s equation
becomes

The solution of differential equation (*) is found easily to be,

where and are the arbitrary constants of integration. To determine

these two constants we apply the boundary conditions, , ,
and we easily find

The potential function in the region is

Having found , we may compute the field ,

Since , the field is directed towards the center, as we

expected, since the field points from higher towards lower potentials.

The energy density is

and the total energy stored in the capacitor is

 373

Having found the energy stored in the capacitor, the capacitance is given
by (11-2-3), i.e.

Notice that the capacitance depends on the dielectric and the geometry
only.

If we assume that , then formula (11-3-4) shows that the capacitance

of a spherical conductor of radius is given by the formula

Example 11-3-1: Assuming that the Earth is a sphere of radius

, find its capacitance .

Solution

Application of formula (11-3-5) with yields

Example 11-3-2: Find the capacitance of a spherical capacitor with

, , filled with a dielectric material having .
 374

Solution

Example 11-3-3: What is the energy required to charge an initially

uncharged conducting sphere in air ( ), of radius , to potential
?

Solution

The required energy is

The capacitance , according to formula (11-3-5) is

and substituting in (*) we find .

PROBLEMS

11-3-1) A spherical capacitor with is filled with a

dielectric . Find the capacitance of the capacitor.

(Ans: ).

11-3-2) For the spherical capacitor in Fig. 11-6, find the surface charge
density on the surfaces and and an expression for the
polarization inside the dielectric.

11-3-3) The energy stored in the spherical capacitor of Fig. 11-6 is

. Assuming that and find the applied
voltage , (Ans: ).

11-3-4) Find the capacitance of a spherical capacitor, filled with two

dielectric materials having permittivity and , as shown in Fig. 11-7. Assume
that the outer shell ( ) is held at potential while the inner shell ( ) is
grounded.
 375

Hint: Solve Laplace’s equation in each region separately and then apply the
appropriate boundary conditions, (see Example 11- 2-2).

Fig. 11-7: Spherical capacitor with two dielectrics.

11-3-5) For the spherical capacitor of Fig. 11-6, assume

, and . Find the electric fields and inside the
dielectric.

(Ans: ).

11-4) Cylindrical Capacitors

A cylindrical capacitor consists of two coaxial conducting cylindrical

surfaces with radii and , ( ). The space between the two conducting
surfaces is filled with a homogeneous dielectric with permittivity . We further
assume that the length of the capacitor is much greater than , and this
allows us to neglect the fringing elects at the open ends of the capacitor. In
such case, due to the cylindrical symmetry, the potential function depends
only on the variable , i.e. .

Notice that a coaxial cable may be consider as a cylindrical capacitor.

 376

Let us consider the cylindrical capacitor of fig. 11-8. The outer surface is
held at potential while the inner surface is grounded.

Fig. 11-8: Cylindrical capacitor.

The potential function inside the dielectric is obtained from Laplace’s

equation, coupled with the appropriate boundary conditions. Laplace’s
equation in cylindrical coordinates is (recalling that ),

whose solution is found to be

The boundary conditions are and , from which

we find and , and substituting in
equation (*) we obtain the following expression for :
 377

For detailed calculations in deriving formula (11-4-1) see Problem 11-4-1.

The electric field inside the dielectric is

The energy density is

The total energy stored in the capacitor of Fig. 11-8 is

Equation (11-4-3) gives the capacitance of a cylindrical capacitor of length

. The “capacitance per unit length” is
 378

Example 11-4-1: Find the capacitance per unit length of a coaxial cable
having and .

Solution

Application of formula (11-4-4) yields

Example 11-4-2: Find the capacitance of a coaxial cable having an

inner conductor in diameter, an outer conductor having an inside
diameter , filled with a dielectric with .

Solution

In our problem we have and the capacitance per

unit length is given by equation (11-4-4), i.e.

The capacitance of a long cable is

Example 11-4-3: Consider a cylindrical capacitor having

and , (see Fig. 11-8). If find and on
the outer surface. What is the charge per unit length on the surface ?

Solution

The electric field inside the dielectric is given by (11-4-2), i.e.

 379

On the outer conducting surface , and

The charge per unit length on the cylindrical surface is

PROBLEMS

11-4-1) Show that the expression for in formula (11-4-1) is a solution

of the differential equation (*) that satisfies the boundary conditions
and .

11-4-2) Find the capacitance of a coaxial cable having an inner

conductor in diameter, an outer conductor having an inside diameter
, filled with a dielectric with .

11-4-3) Consider the cylindrical capacitor in Fig. 11-8. The region between
the conducting cylinders contains two different dielectrics: and
. Find the capacitance per unit length.

(Ans: ).

11-4-4) For the capacitor in Problem 11-4-3, the inner cylinder is held at
potential while the outer one is grounded. Assuming
, find: a)
everywhere inside the capacitor, and b) the energy stored in the capacitor.
 380

11-4-5) Find the capacitance per unit length of the capacitor in Pr. 11-4-3,
assuming .

(Ans: ).

11-4-6) Find the capacitance of the capacitor shown in Fig. 11-9. Assume
that and are conducting surfaces and that is held at potential , while
is grounded. The space between and is filled with a dielectric with
dielectric constant .

Fig. 11-9: Computation of the capacitance of a truncated wedge.

(Ans: , expressed in radians).

Hint: Solve Laplace’s equation in the region , subject to the

boundary conditions . Next find and then
find the energy stored in the capacitor, etc.

11-4-7) In Problem 11-4-6, find and on, assuming

. Also, find
the density of bound charges inside the dielectric and the total amount of
charges on the surfaces and .
 381

11-5) Capacitors Connected in Parallel or in Series

Arrays of capacitors with their plates connected as in Fig. 11-10 (a) are said
to be connected in parallel, while capacitors connected as in Fig. 11-10 (b) are
said to be connected in series.

In the “parallel connection” all capacitors are under the same voltage ; in
the “series connection” all capacitors have the same charge .

Fig. 11-10: Capacitors connected in parallel (a) and in series (b).

In the first case, (parallel connection), all the capacitors are under the same
voltage . If we call their respective charges, then we have:

and term wise addition results in

and this equation shows that the array of parallel capacitors may be
replaced by one capacitor of total capacity equal to
 382

For the connection in series, we assume that the capacitors are initially
uncharged. If we place a charge on the left plate of , this will induce a
charge – on the right plate of and this, in turn, will induce a on the left
plate of , etc. All the capacitors carry the same charge , so

or, equivalently,

Adding term wise and noting that results in

and this equation shows that the array of capacitors in series may be
replaced by one capacitor of total capacitance equal to

Example 11-5-1: Find the total capacitance of three capacitors

if the capacitors are connected, (a) in
parallel, and (b) in series.

Solution

Parallel connection:

Connection in series:

Example 11-5-2: Five capacitors

and are connected as in Fig. 11-11. Find the total
capacitance.
 383

Solution

Capacitors and are connected in parallel, and the same holds true for
and .

Fig. 11-11: Capacitors connected in parallel and in series.

The capacitors and are connected in series, so the total

capacitance is given by

Example 11-5-3: Consider a parallel plate capacitor having area of plates

and distance between the plates . The capacitor is
connected to a voltage source and then is disconnected, (so that
the electric charge on the plates cannot escape). Then the region between the
plates is partially filled with a dielectric slab of thickness and
permittivity , as shown in Fig. 11-12. Find: a) the capacity in free space
(prior to the insertion of the dielectric), b) the charge with no and with the
dielectric, c) the fields and everywhere inside the capacitor, d) the
surface charge density on the positive plate, and e) the voltage of the
capacitor after the insertion of the dielectric.

Solution

The capacitance of the capacitor in free space is (see formula (11-2-1)),

 384

The charge prior and after to the insertion of the dielectric is

Fig. 11-12: Capacitor with a dielectric slab inserted.

The surface charge density, on the positive plate, is

The field everywhere inside the capacitor, (which depends only on ), is

The field in free space between the plates of the capacitor is

The field inside the dielectric is

The polarization is zero in free space, and inside the dielectric is

 385

The voltage between the plates (with the dielectric inserted) is

For an alternative computation of the voltage with the dielectric inserted,

see Problem 11-7, in the Supplementary problems.

PROBLEMS

11-5-1) Find the total capacitance of three capacitors and

connected a) in parallel and b) in series, (Ans: ).

11-5-2) In Fig. 11-11 find the total capacitance assuming

.

11-5-3) A capacitor of is charged to and a capacitor of is

charged to . The two charged capacitors are connected in parallel. What
is the potential difference between the plates after the connection?

(Ans: ).

11-5-4) Three concentric, conducting spherical surfaces in free space, have

radii . The surfaces and are connected together
by a metallic wire and form one plate of a capacitor. The surface is the
other plate. Show that the capacitance of the thus formed capacitor is
.

Hint: Regard this system as two separate spherical capacitors connected in

parallel. The capacitance of a spherical capacitor is given by (11-3-4).

11-6) Forces on Dielectric Materials Moving in the Field of a Capacitor

The force exerted on a dielectric moving in the field of a capacitor, is usually

found by energy methods. We shall consider two cases: a) The charge of the
capacitor remains constant and b) The voltage of the capacitor remains
constant.

a) The charge of the capacitor remains constant, (isolated capacitor).

Let and be the plates of a capacitor whose charge remains constant. If

the dielectric body, under the action of the electrical force (due to the field
 386

of the capacitor) is displaced by along some direction , then the work

done by (positive) should be at the expense of the electric energy stored in
the field of the capacitor. i.e.

where is the infinitesimal change in the energy of the capacitor.

Fig. 11-13: Forces on dielectric materials.

For a capacitor with a constant charge , the energy is

and from equation (11-6-1) we find:

Notice that the same formula yields the force exerted on any surface
whose displacement causes changes in the capacity, (see Example 11-6-1).

b) The capacitor voltage remains constant, (non-isolated capacitor).

In this case the capacitor is not isolated, since it is connected to the voltage
source . Energy balance leads to the following equation,
 387

and since equation (*) yields

We notice that half of the energy supplied by the voltage source ( ) is

expended in mechanical work ( ) and the other half to increase the
energy stored in the capacitor.

Example 11-6-1: Find the force between the plates of a parallel plate
capacitor, assuming to be constant.

Solution

Fig. 11-14: Force between the plates of a capacitor.

The force on the right plate, is (according to formula (11-6-2))

(The negative sign means that the force points to the left).

A force of equal magnitude but in the opposite direction applies on the left
plate, (the positive plate).
 388

Alternative solution: The electrostatic pressure on the right plate is (see

formula (9-4-3)),

This force is directed to the left since and points in the positive
direction.

Example 11-6-2: A long dielectric slab of relative permittivity can slide

freely (no friction) between the plates of a parallel plate capacitor, as in Fig. 11-
15. If the dielectric is placed just below the capacitor, and between the plates,
which are then connected to a voltage source , show that the slab will elevate
if , where is the mass of the dielectric slab.

Fig. 11-15: Motion of a dielectric slab between the plates of a capacitor.

Solution

Let us assume that the dielectric has been elevated to a height , inside the
capacitor. According to Pr. 11-2-4, the capacitance of the capacitor is
 389

The electric force on the dielectric will be (see formula (11-6-3)):

This equation shows that the dielectric will elevate provided that

and this completes the proof.

PROBLEMS

11-6-1) In Example 11-6-2 show that the maximum elevation is

11-6-2) Show that the “force per unit volume” on a polarized dielectric is
and thus deduce that the total force on a piece of dielectric is
, (where is the volume of the dielectric).

Hint: See equation (8-2-9), the force on an electric dipole.

11-6-3) Using the vector identity ,

together with the fact that , show that the total force on a piece of a
homogeneous dielectric, ( ), is given by the formula

This formula shows that the force on the dielectric, as being proportional
to , pulls the dielectric in the direction of increasing field strength.

11-6-4) Apply the formula for obtained in Pr. 11-6-3, to derive formula
(**) in Example 11-6-2. Note that if we assume that the field exist only
between the plates and is zero everywhere else, then should be zero (since
the field inside the capacitor is uniform). Where is the pitfall? Well, in this
case, is the fringing field (see Fig. 11-3) that is responsible for pulling the
 390

dielectric into the capacitor. If we call the field inside the capacitor
and the fringing field, then (since ) show that

which is identical to formula (**) obtained in Example 11-6-2, (recall

).

SUPPLEMENTARY PROBLEMS

11-1) A parallel plate capacitor in free space is charged to a potential .

Then the space between the plates is filled with a dielectric. The charge of the
capacitor must then be increased by 5 times to restore the former potential
difference. Find the relative dielectric constant of the dielectric, (Ans: ).

11-2) The parallel combination of two capacitors and is

connected in series with a capacitor . Find if the total capacitance of the
array is .

11-3) Breakdown voltage of a dielectric material: When the electric field

strength is strong enough, then what happens is that electrons may be
violently removed from their parent atoms; the bound electrons become free
electrons and usually pick up enough kinetic energy to knock off other bound
electrons from their atoms. The dielectric becomes a conductor. In an
extremely short period of time we get a cascade of electrons, giving rise to a
spark across the plates of the capacitor. This is known as the dielectric
breakdown. Since strong fields are produced by high voltages, usually there is
a limit on the voltage imposed on capacitors, and this is a crucial factor when
designing electrostatic systems. Let us consider the following problem:

Design a coaxial cable with capacity per unit length to

operate at a voltage . The relative permittivity of the dielectric is
while the maximum electric strength allowed is .

(Ans: , see Fig. 11-8).

Hint: Use formulas (11-4-2) and (11-4-4). Notice that the maximum field
occurs at .
 391

11-4) Calculate the capacitance of an orbiting satellite of diameter .

11-5) Consider a parallel plate capacitor with one plate at and the
other plate at . The area of each plate is . The dielectric constant
between the plates is . Show that the capacitance of
the capacitor is .

Hint: The capacitance of a capacitor with spacing between its plates, is

and since all these infinitesimal capacitors are
connected in series, the total capacitance will be , etc.

11-6) In the capacitor of Problem 11-5, the plate is held at potential

while the plate is grounded. Show that the potential inside
the capacitor is given by the expression , and then
find the energy density and the total energy stored in the capacitor. Using
the expression obtained for find the capacitance and verify
that your answer is identical to the one found in Problem 11-5.

Hint: and in the one dimensional case,

, i.e. , etc. Notice that
since there are no free charges inside the capacitor.

11-7) In Example 11-5-3, consider the given configuration as three

capacitors connected in series, and show that the total capacitance is
. Then, the voltage between the two plates is . Verify
that application of this formula yields , same result as that
obtained in Example 11-5-3.

11-8) Two conductors and have capacitances and respectively.

The two conductors are then connected by a fine wire. We thus obtain one
conductor. Show that the capacitance of this conductor is .

11-9) Consider two conducting spheres with radii and

. The two spheres are connected by a fine wire, and then a total charge
is supplied to the system. What is the potential of the system?

(Ans: ).
 392

Hint: Use the results obtained in Problem 11-8 and equation (11-3-5)).

11-10) A conducting sphere with radius is charged to a potential , while

another conducting sphere of radius is charged to a potential . If we
connect the two spheres by a fine wire, show that their common potential is
given by the expression .

11-11) The number next to each capacitor expresses its capacitance in .

Find the capacitance of the system and the charge of each capacitor.

Fig. 11-16: A system of capacitors.

(Ans: ,
).

11-12) Conductor with capacitance is charged and contains energy

, while another conductor is uncharged. The conductor comes
into contact with the conductor , (for example, by means of a fine wire), and
then is disconnected. The energy of the conductor after the disconnection is
. What is the ratio . Given that the initial charge of the
conductor was find the energy expended for the transfer of
charges from conductor to the conductor .

(Ans: ).

Hint: See Problem 11-8.

 393

11-13) A spherical capacitor consists of two concentric conducting spherical

surfaces with radii and , ( ). The inner sphere carries a charge while
the outer sphere is grounded. Due to the electrostatic forces, the outer
conductor contracts to a new radius . Show that the work done by the
electrostatic forces is .

11-14) Design a spherical capacitor of capacitance and operating

voltage . To avoid dielectric breakdown, the electric field must not
exceed the value . Assume that the relative dielectric
constant is .

(Ans: Inner radius , outer radius ).

Hint: The maximum field intensity takes place at , inner surface, see
Section 11-3.

11-15) Stresses on Dielectrics:

Let us consider a homogeneous dielectric with permittivity and charge

density , as shown in Fig. 11-17.

Fig. 11-17: Force on a dielectric material.

Assuming that the charge density within is , and that the electric field
at the center of the infinitesimal cube is , the differential force on the
charge is, , and since ,

The total force on the dielectric is found by in integration over the volume
, i.e.
 394

By appropriate mathematical manipulation, (see Problem 11-17), the

volume integral can be transformed to the following surface integral, over the
surface of the dielectric,

In formula (***), the quantity may be interpreted

as the “electric surface force density”. When this force density is integrated
over the whole surface, yields the total force exerted on the dielectric enclosed
by the surface . Assuming that the electric field at the surface of the
dielectric (but inside the dielectric) forms an angle with the outward normal
unit vector , then it is easily shown that (see Problem 11-18).

where is the tangential unit vector, as shown in Fig. 11-18.

Fig. 11-18: The surface force density at the surface of the dielectric.

Note that is at an angle from , which is twice the angle the electric
field at this point , forms with . Since depends on , the electric
 395

force density pulls on the surface, no matter if the electric field points
towards or away from the surface.

11-16) Starting with equation (****) of Problem 11-15, show that the
electric pressure on the surface of a dielectric, surrounded by free space, is
given by the formula

where and are the tangential and the normal components of the
electric field in the dielectric, (recall that if is the electric field in free space,
then from the boundary conditions on the surface of the dielectric,
and ). Since , there is a pressure, and therefore a
mechanical force, directed from the dielectric into the vacuum.

11-17) (a) Show the vector identity: , where is

a scalar field. Note that both sides are vector quantities. To show this identity,
apply the divergence theorem to the vector field , where is an
arbitrary, constant vector, and use the identity
, (b) Use the vector identity (see Pr. 11-6-3), (c)
The integrand in the volume integral in equation (***) in Problem 11-15 can be
written as . Next we use the
vector identity, from which we obtain,
, and similar expressions are obtained for
the and components, etc. From this point on, the manipulations are
tedious but straightforward.

11-18) Starting with the expression , show that

can be written, equivalently as .

Hint: Resolve into components normal and tangential to the surface, i.e.
, (see Figure 11-18).

11-19) Show that if at the surface of the dielectric is at an angle

from , then the electric pressure at this point is zero.
 396

11-20) Starting with the expression for the electric pressure , obtained in
Problem 11-16, work Example 11-6-2, (note that in this Problem and
).

11-21) A spherical conducting surface of radius , in free space, is centered

at the origin, and receives a total charge . The spherical surface is divided
into two parts by the plane, and then, the two hemispherical surfaces are
slightly separated. Show that the upper hemispherical surface is repelled from
the lower one by a force . What is the force exerted on the lower
part by the upper one?

11-22) The region between the two conducting spherical surfaces and
,( ), is filled with a dielectric with relative permittivity ,
where is a positive constant. The inner surface is kept at potential
while the outer surface is grounded (potential zero). (a) Find the
potential everywhere inside the spherical capacitor, (b) Find the electric
field inside the capacitor and (c) Find the capacitance of the capacitor.

(Ans:
).

Hint: Starting with equation (10-5-1), show that the Poisson’s equation is
, where .

11-23) In Problem 11-22, assume that when , i.e., when , then

the capacitance assumes the familiar form , (see formula
(11-3-4)).

Hint: In the expression for in Problem 11-22, take the limit as .

Recall that
 397

CHAPTER 12: SPECIAL METHODS AND TECHNIQUES IN

ELECTROSTATICS

In the preceding chapters we have developed some general methods to

determine the electric field , using either Coulomb’s or Gauss’s law when
the charge distribution is known, or using the relation when the
potential function is a known function of the spatial coordinates. However, in
most cases, neither the charge distribution nor the potential are known in
advance.

Many of the charge distributions we have considered so far are “artificial

distributions” serving the sole purpose to familiarize the student with the basic
laws and techniques in Electrostatics. For instance, we cannot have point
charges or line charges, since the electric field close to the point or to the line
charge would be so intense that would cause a dielectric breakdown. Recall
that for a point charge and as , while for an
infinite, uniform line charge distribution and as
.

Regarding volume charges in dielectrics, given that there are no perfect

insulators, a volume charge placed inside a dielectric, sooner or later would
“leak out” to the surface, where it will be found redistributed as a surface
charge.

Free space (vacuum), of course, is considered to be a perfect dielectric;

however, even in this case, we cannot have volume charges in stable
equilibrium, according to Earnshaw’s theorem.

We are thus led to the conclusion that electrostatic fields, in practice, are
generated by surface charges distributed either on surfaces of conductors or
surfaces of dielectrics.

12-1) A General Approach in Solving Problems in Electrostatics

Let us, for definiteness, consider the following problem: We have

conductors , charged to potentials respectively, as in
Fig. 12-1. The surrounding space is a homogeneous and linear dielectric of
permittivity .
 398

The electrostatic problem will be fully solved, if we can determine the

potential as a function of the spatial coordinates, ( in rectangular
coordinates, or in cylindrical coordinates, or in spherical
coordinates). Then, all the other field quantities will be uniquely determined.
For instance, , etc.

Fig. 12-1: Charged conductors within a homogeneous dielectric.

The surface charge density, on the surface of the conductors, is given by the
equation

and knowing we may (in principle) find the total charge on a conductor
by integrating over the surface of the conductor, etc.

As we have shown in Chapter 5, the potential function satisfies Poisson’s

equation
 399

where is the volume charge density of the free charges (charges that we
have, somehow, placed at this position). Notice that does not include bound
charges, which are induced within the dielectric, and on which we do not have
any control at all.

In a charge free region, , and Poisson’s equation reduces to Laplace’s

equation

We point out that equations (12-1-2) and (12-1-3) hold true in

homogeneous dielectrics. For non homogeneous dielectrics, i.e. for dielectrics
for which is a function of the spatial coordinates, Poisson’s equation
becomes

while the corresponding Laplace’s equation becomes

Equation (12-1-4) is obtained from equation (10-5-1) if is substituted by

.

Since as we have already mentioned, there are no perfect dielectrics in

nature, even if we place a free charge inside a dielectric, sooner or later this
charge will leak out to the surface, and therefore, in most electrostatic
problems we assume that , and this means that in order to determine
the potential we have to solve Laplace’s equation.

Any function which satisfies Laplace’s equation is called “harmonic

function”, (see Section 1-12 (c)).

Laplace’s equation, being a Partial Differential Equation (P.D.E), admits an

infinite number of solutions. Among all the solutions we have to choose the
one which satisfies the Boundary Conditions as well. Usually, the boundary
conditions are the potential of the conducting bodies.
 400

The problem of determining the solution of a P.D.E. subject to certain

boundary conditions is an important problem in Mathematics, known as “a
boundary value problem”.

In summary, to solve an Electrostatic problem, we seek a function which is

harmonic in the region of interest and satisfies the boundary conditions of
the problem at hand.

From practical experience, it is known that the solution of a problem in

Electrostatics is unique. We shall show this fact mathematically, i.e. we shall
show that if a function satisfies Laplace’s equation and also satisfies the
appropriate boundary conditions, then this function is the only possible
solution. This means that any solution of Laplace’s which satisfies also the
boundary conditions is unique. This is known as the uniqueness theorem and
is proved in the next section.

12-2) Uniqueness Theorem

Let us, for simplicity, consider the simple system in Fig. 12-2.

Fig. 12-2: Uniqueness theorem.

The system consists of a conductor charged to a potential with respect

to the conducting surface , assumed to be grounded. The space between the
conductor and the surface is filled with a homogeneous dielectric of
dielectric constant .
 401

Let us assume that there are two solutions and that satisfy both,
Laplace’s equation and the boundary conditions, i.e.

Let us define a function . We notice that also satisfies

Laplace’s equation, since

Also, on the boundaries and , . Indeed,

and similarly .

If we now consider the vector identity

which is true for any scalar and any vector , and apply this identity for
and we obtain:

Integrating this identity over the volume and applying the Divergence
theorem to transform a volume to a surface integral, we get:
 402

Since , the integral , and similarly, since

, the integral as well, and this implies that

The volume integral in (****) must therefore be equal to zero, i.e.

Equation (*****) implies that , (see Remark), and this identity holds
true if (constant), everywhere in the region. The constant can be
evaluated by considering a point on the boundary, say on the boundary . At
this point,

and this completes the proof.

Remark : The integrand in (*****) is a non negative quantity, i.e.

can be either positive or zero. If we assume that at some point , then
would be different from zero, in a small neighborhood about this point,
and hence for all the points in this neighborhood. Then the
integral in (*****) would be positive and not zero. We are therefore forced to
conclude that (*****) implies , (think about this).

The proof presented here about the uniqueness of the solution, is easily
extended to more complicated configurations.

The same reasoning and approach also apply to prove uniqueness of the
solution of the Poisson’s equation, (see Pr. 12-1 in Supplementary Problems).

12-3) Laplace’s Equation in One Dimension

In a one dimensional problem, Laplace’s equation becomes an ordinary

differential equation. In this case the potential depends on one variable.
 403

For instance, in rectangular coordinates, could depend only on , or only

on , or only on , and similarly for the cylindrical and the spherical
coordinates.

In chapter 11 we solved Laplace’s equation in one dimension, in order to

find the capacitance of a parallel plate capacitor ( , of a spherical
capacitor ( ) and of a cylindrical capacitor ( ).

Also, in Chapter 5, we have solved Laplace’s and Poisson’s equations to find

the potential of certain spherical and cylindrical charge distributions (see
Examples 5-4-2 and 5-4-3).

We now consider two more examples; in the first one the potential
depends on the cylindrical coordinate and in the second depends on the
spherical coordinate .

Example 12-3-1: In a cylindrical system the conducting plane is

charged to potential while the conducting plane is grounded. Along
the axis the two planes are separated by an infinitesimal insulating gap. Find
the potential and the electric field everywhere between the planes.

Solution

Fig. 12-3: Computation of the field between two planes.

 404

Assuming that depends only on , i.e. , Laplace’s equation in

cylindrical coordinates becomes

From the boundary conditions and we find

, and equation (*) yields

The electric field is

Example 12-3-2: Find the potential in the region between a conducting

cone charged to a potential and the grounded
plane , insulated from the vertex of the cone by means of a small
insulating gap.

Solution

Fig. 12-4: Computation of the field between a cone and a plane.

Since the potential depends only on , i.e. , Laplace’s equation in

spherical coordinates assumes the form
 405

(for detailed calculations see Problem 12-3-1). From the boundary

conditions and , we obtain and
, and equation (*) yields:

The electric field in the region is

Notice: In Example 12-3-2 we assume that the vertex of the cone is

separated from the plane by a small insulating gap, otherwise the plane
and the cone would be one conducting body and as such, all its points would
have the same potential. Similar remark holds for the insulating infinitesimal
gap in Example 12-3-1.

PROBLEMS

12-3-1) Show that the solution of the differential equation in Ex. 12-3-2 is
given by equation (*).

Hint: , set ,
etc.

12-3-2) In Ex. 12-3-2, starting with equation (**) for the potential , show
that the electric field is given by equation (***).

12-3-3) Consider the two coaxial, conducting cones in Fig. 12-5. The cone
is held at a potential while the cone is grounded ( ). Find
 406

the potential in the region . The cones are separated by

an infinitesimal gap at the origin.

Fig. 12-5: The field between two coaxial, conducting cones.

(Ans: ).

12-3-4) In Problem 12-3-3 find the electric field in the region .

12-3-5) In Problem 12-3-3 assume . Find the

potential and the electric field at the point .

(Ans: , ).

12-3-6) In Problem 12-3-5 find the surface charge density on the cone
, at the point .

Hint: at the point .

12-3-7) The region between two concentric conducting spherical surfaces

and has a constant volume charge density . The
surface is held at a potential while the surface is grounded
( ). Solve Poisson’s equation to determine the potential in the
region , (Ans: , where:

and ).
 407

12-3-8) In Problem 12-3-7, find the electric field , the energy density
and the total energy stored between the two conducting spherical surfaces.

12-3-9) The region between two long, coaxial conducting cylindrical

surfaces and has a constant volume charge density
. The surface is held at potential while the surface is grounded
( ). Solve Poisson’s equation to determine the potential in the
region , (Ans: , where:

and ).

12-3-10) In Problem 12-3-9 find the electric field and the total energy
stored per unit length.

12-4) Laplace’s Equation in Two Dimensions (Rectangular Coordinates)

In the preceding section we solved Laplace’s equation in one dimension. In

such a case the Partial Differential Equation reduces to an Ordinary Equation.
In this section we shall develop a general (and very powerful method) of
solving PDE’s in two dimensions. The method can be extended in cases where
all dimensions ( ) are present. Usually, the method of approach is based
on a mathematical technique, known as “Fourier Analysis”. Even though the
reader is supposed to be familiar with Fourier analysis methods, for the
completeness of the book, we present an elementary introduction to Fourier
analysis, which will prove to be very useful for our subsequent analysis.

a) Orthogonal functions.

A set of functions is said to be orthogonal

over some interval if

Notice that orthogonality is defined relative to “an interval ”. A set of

functions may be orthogonal over some interval , but not orthogonal
over another interval .
 408

The norm of a function is defined by the integral

Using Kronecker’s delta (see equation (7-2-7)), equations (12-4-1) and (12-
4-2) may be expressed in the compact form

As an example, let us show that the functions

are orthogonal over the interval , where is a real number.

Assuming , equation (12-4-4) implies:

The norm of is
 409

We may therefore write:

b) Expansion of a function defined over , in terms of a series of

orthogonal functions defined over the same interval.

Let us assume that a function is defined over an interval and

that is a set of functions, orthogonal over the same interval. We seek an
expansion of the form

where are coefficients to be properly evaluated. An expansion

of the form (12-4-6) is known as “a Fourier expansion of ” over the
interval , named after the French Mathematician Joseph Fourier, who
first developed and applied this method while working on a problem in heat
transfer. If are trigonometric functions (sines and cosines), which is the
most common case, (12-4-6) represents a “trigonometric Fourier expansion”
of .

As we shall show, the “key idea” in evaluating the expansion coefficients

is the orthogonality property of the functions . Note
that, in a sense, expansion (12-4-6) resembles the expansion of a vector into its
three components, along the orthogonal unit vectors and .

Very soon, we shall see why expansion (12-4-6) is important, when we try
to solve Laplace’s equation in two dimensions.

Suppose now that we want to determine the coefficient , .

Multiplying both sides of equation (12-4-6) by and integrating both sides
from to we get:
 410

and using equation (12-4-3) we find that

As an example, let us consider the function (constant) over

and a family of functions, orthogonal over the
same interval , as we proved earlier.

We want to determine coefficients such that

Application of formula (12-4-7) yields

By virtue of equation (12-4-8) we may thus write:

 411

c) The method of separation of the variables.

The method of separation of the variables is a very general method of

solving P.D.E. coupled with boundary conditions. It is a very popular method
in Engineering and Physics.

Let us, for definiteness, assume that we have to solve Laplace’s equation in
two dimensions (say and ). In this case Laplace’s equation is

The first step is to assume that can be written as a product

where is a function of only and is a function of only.

Substituting (***) into (**) results in

Dividing by and rearranging results in the following equation

We come now to the crucial point; equation (****) must be true for all
values of and , and since the left member is a function of only and the
right member is a function of only, each member of the equation must be
equal to a constant number, independent of and , i.e.

It seems that the original P.D.E has been split into two independent
ordinary differential equations, but this is not true, since the two equations in
(12-4-10) are coupled through the constant .

Well, what is ? Up to this point, it is still not known. In general, the

separation constant is determined from the boundary conditions. As we
 412

shall see soon, usually, may take an infinite number of allowed values,
determined from the boundary conditions of the problem. To each allowed
we have a product solution , (as obtained from
equation (12-4-10), and finally the solution of the original P.D.E in (**) is
obtained by a suitable combination of these particular solutions.

Let us illustrate the method by means of a few examples.

Example 12-4-1: Consider the rectangular trough in Fig. 12-6. The side
is held at a potential while the other three sides
are grounded. Solve for the potential everywhere inside the trough.

Solution

Fig. 12-6: The potential distribution inside the trough.

The potential satisfies Laplace’s equation coupled with the appropriate

boundary conditions:
 413

Assuming that , Laplace’s equation reduces to the

following system of equations, (see formula (12-4-10),

Regarding the boundary conditions, we first consider the homogeneous

conditions, and leave the non homogeneous one for the very last step.

We have thus obtained some boundary conditions for the functions

and . Summarizing, we have:

For the function

and for the function

1) Let us first consider equation (***), (since for the function we have
more information-two boundary conditions).

To begin with, at this point, we do not know anything about the separation
constant .

a) Let us assume that is a positive number. We may set

. Then the differential equation becomes

since the hyperbolic functions are defined in terms of the exponential

functions by means of the formulas
 414

(Let the reader verify the equivalent form of the solutions for ).

The first boundary condition in (***), , implies that

The second boundary condition, , implies and

since when , the conclusion is that , and . This
of course, is a solution, but obviously it is a trivial solution, and we are not
interested in such solutions. The conclusion is that the separation constant
cannot be a positive number.

b) Let . Equation (***) becomes . The general solution of this

equation is and imposing the boundary conditions
and results in and , i.e. which is again a trivial
solution. We conclude that the separation constant cannot be zero.

c) Finally, let us consider the case where is a negative number. We may

thus set . Then the differential equation becomes

The first boundary condition implies

The second boundary condition implies

The subscript indicates that (and therefore ) depends on the integer

, i.e. .
 415

We see that in this case (separation constant negative), the problem

admits non trivial solutions. As a matter of fact, admits an infinite number of
non trivial solutions. For each integer value of we have a valid solution. For
example,

2) Recall that the two ordinary equations are coupled through the
separation constant. Since , equation (****) becomes

and since , and

(For each we have one valid solution ), i.e.

Recalling that the potential was expressed as a product

we see that we have an infinite number of valid expressions for , each
one of which satisfies Laplace’s equation and the three homogeneous
boundary conditions in formula (*), i.e.
 416

Here comes the second crucial point in our analysis. Since Laplace’s
equation is a linear equation, the sum of all the particular solutions in (*****),
(an infinite number of solutions), will also satisfy Laplace’s equation, i.e. the
most general form of the solution of Laplace’s equation will be

We point out that the solution of Laplace’s equation in (******) satisfies

Laplace’s equation and the three homogeneous boundary conditions in (*).
This solution will be a solution to our problem if we can determine the
constants (which up to this point are arbitrary, unknown constants) in such
a way as to satisfy the fourth, non homogeneous condition
, i.e. if
 417

from which and for . The only surviving

term is the one corresponding to , and finally equation (******) assumes
the form

This is the solution of Laplace’s equation which satisfies all the boundary
conditions in (*), i.e. is a solution to our problem, (let the reader verify by
direct calculations that the found expression for satisfies Laplace’s
equation and all the boundary conditions).

Comments:

1) Using “the separation of variables method” we have found one solution

to our problem. Any other method of solution would yield the same answer for
the potential , by virtue of the uniqueness theorem.

2) The fourth boundary condition in (*), ( ) is a rather

artificial boundary condition, which however, has the advantage of illustrating
the main points of the separation of variables method.

In the next Example we shall consider a more realistic problem, which is to

consider again the rectangular trough in Fig. 12-6 and assume that the top side
is kept at a constant potential .

3) As we see in this Example, the separation constant is actually determined

from the boundaries of the problem at hand. The allowed values of the
separation constant are called “the eigenvalues of the problem”. In our
Example, we have “a discrete set of eigenvalues”, (
). There are problems, where the separation constant may take on
any real value. For such a problem, see Example 12-4-4.

Example 12-4-2: Consider the rectangular trough in Fig. 12-6. The side
is held at a constant potential while the other three sides
are grounded. Solve for the potential everywhere inside the trough.
 418

Solution

The potential satisfies Laplace’s equation, coupled with the

appropriate boundary conditions,

The solution of Laplace’s equation, which satisfies the three homogeneous

conditions, as found in Example 12-4-1, (see equation (******) in previous
example), is

In order this expression to be the solution of our problem, as stated in (*),

we have to (somehow) determine the expansion coefficients so that the
fourth, non homogeneous condition, will be satisfied as well, i.e.

The expansion coefficients in (***) are given by equation (12-4-8), i.e.

 419

and substituting in (**) we get:

Example 12-4-3: Consider the infinite rectangular trough in Fig. 12-7. The
sides and are grounded while the side is kept at a constant
potential . Solve for the potential everywhere inside the
trough.

Solution

Fig. 12-7: The potential inside an infinite trough.

The sought for potential is determined from the solution of

Laplace’s equation subject to the appropriate boundary conditions
 420

In this problem, the domain of solution is infinite, ( ). However,

the potential should remain finite at all points within the region, and
this is actually the fourth condition in equation (*).

Assuming that , equation (*) reduces to the following

two ordinary differential equations, (see Example 12-4-1):

where (to be determined) is the separation constant.

Equation (**) admits non trivial solutions, provided that

(discrete eigenvalues), while the corresponding functions are
given by the expression, (see Example 12-4-1),

In equation (****) are arbitrary constants.

With the allowed values of the separation constant , equation (***)

becomes:

and since should remain finite as , we choose , (i.e.

the term should not appear in our solution, otherwise the potential
would approach infinity as , something which, on physical
grounds, is not allowed). On the other hand, as , the term
approaches zero. With this choice of ,( ),
 421

The general solution of Laplace’s equation which remains bounded in the

region of interest and satisfies the two homogeneous boundary conditions in
equation (*) is:

where we have renamed .

It remains to determine the coefficients so that the expression for

in equation (******) satisfies the non homogeneous condition in (*),
i.e.

and substituting in (******) we find

It can be shown that this expression for the potential can be written in
closed form as follows, (for a proof see Problem 12-4-4),

Example 12-4-4: Find the solution of Laplace’s equation in the half

plane provided that the potential takes on the value
(known function) on the axis.
 422

Solution

Fig. 12-8: The potential in the half plane .

The potential is obtained as the solution of the following boundary

value problem:

The third equation in (*) means that the potential function remains
bounded everywhere in the region .

Setting , Laplace’s equation reduces to

or equivalently,

Since , in the region , the separation constant

must be negative, say , so that the equation becomes
. (Let the reader think what would happen if were positive or
zero). The general solution of this equation is

which remains bounded for all . With this choice of , the second
equation becomes , from which we obtain
 423

. Again from the condition as , the constant must

be zero, and hence

The potential function will be

where we have renamed and .

Notice that in this case, (the eigenvalues of the problem) can be any
positive number, i.e. may vary continuously from to .

The expression for in (****) satisfies Laplace’s equation and the

condition that remains bounded everywhere in the region
. This expression will be the solution to our problem, provided
that we can determine the coefficients and so that the boundary
conditions will be satisfied.

In the previous examples, (case of discrete eigenvalues), we used the

superposition principle to construct the general solution from the particular
solutions; (each particular solution corresponded to one eigenvalue).

In our problem, since varies continuously from to , we can,

still, apply the superposition principle, using the following trick.

Since there is no restriction on , we may replace by , by and

integrate over from to , to obtain:

The condition yields

 424

The functions and can now be determined with the aid of

“Fourier’s integral theorem”, according to which

and substituting these expressions into (*****) and simplifying we find:

(For detailed calculations see Problem 12-4-5).

The expression in (******) may be written in the form (see Pr. 12-4-6),

This expression is known as the “Poisson’s solution of Laplace’s equation

for the half plane”.

PROBLEMS

12-4-1) Find the potential inside the trough in Fig. 12-6, assuming
that .

(Ans: ).

12-4-2) In Problem 12-4-1 find the electric field . Then assuming that
, find the potential and the electric field
at the points and .

Hint: .

12-4-3) Consider the infinite rectangular trough in Fig. 12-9. The sides
and are grounded while the side is held at a constant potential
. Solve for the potential everywhere inside the trough.
 425

Fig. 12-9: The potential inside an infinite trough.

(Ans: ).

Hint: See Example 12-4-3.

12-4-4) In Example 12-4-3 derive expression (*******) for the potential

, starting with the infinite series solution.

12-4-5) In Example 12-4-4 show equation (******).

12-4-6) In Example 12-4-4 derive Poisson’s solution for the half plane,
starting with equation (******).

12-4-7) In Example 12-4-4, use Poisson’s formula to find the potential

, assuming that

(Ans: ).

12-4-8) Repeat Problem 12-4-7, if

(Ans: ).
 426

12-5) Laplace’s Equation in Two Dimensions (Spherical Coordinates)

The method of separation of variables, which we have introduced in the

preceding section for rectangular coordinates, can be applied to spherical
coordinates as well, as we shall show in the sequel.

We shall consider a number of interesting problems possessing “azimuthal

symmetry”, that is the potential does not depend on , it depends only on
and , i.e. . Since is independent, , and Laplace’s
equation in spherical coordinates, for the potential becomes:

It turns out that the solution of this P.D.E is expressed in terms of the so
called “Legendre functions”. These functions have been studied extensively
and their properties have been investigated thoroughly. A very brief account
on Legendre’s functions and related properties is given below.

a) Legendre’s functions.

Legendre’s functions are obtained as the solutions of the “Legendre’s

differential equation” which is of the form

where is a non negative integer number (can be either zero or a positive

integer).

As we shall see shortly, solving equation (12-5-1) using separation of

variables, leads to the Legendre’s equation (12-5-2).

A popular (and powerful) method of solving linear ordinary differential

equations, when the coefficients are functions of , is the so called “ method
of Frobenious”. That is, we assume a power series solution of the form
 427

and try to determine the expansion coefficients and the constant , so

that the assumed solution satisfies identically the given differential equation.
Doing so, we find that the general solution of Legendre’s equation can be
expressed as follows:

In this solution, are polynomials in , of degree , bounded

everywhere in the closed interval , while are functions
bounded in the open interval , but at the end points ,
blow up (go to infinity), ( ). This is an important remark, since, if
we want to find the bounded solutions of Legendre’s differential equation, in
the closed interval , we must take , so that the term
does not appear in the solution.

The polynomials are normalized, so that for all integer

values of .

It can be shown that

This formula is known as the Rodrigues’ formula. By means of this formula,

we may, for example, find the first few Legendres’ polynomials:

Regarding the functions , it can be show, for example, that

Notice that as we have mentioned, at , and approach infinity.

 428

Legendre’s polynomials satisfy the following, useful, recurrence formulas:

Same recurrence formulas apply for the functions .

The following orthogonality property of the Legendre’s polynomials, over

the interval , is useful, since by virtue of this property we may
expand a given function defined over the interval in terms
of Legendre’s polynomials:

As an example, let us assume that is defined over and

that we want to expand in a series of Legendre polynomials, i.e. to
express in the form

To determine the coefficient we multiply both sides of (*) by and

integrate from to . We have:
 429

Remark: In deriving formula (12-5-7) we assume that the function

satisfies “Dirichlets’ conditions” i.e. over the interval , has at
most a finite number of jump discontinuities and is square integrable,
( ). Most of the functions we are working with, meet these
two conditions and we can therefore safely apply (12-5-7) to determine the
expansion coefficients .

b) Solving Laplace’s equation in spherical coordinates.

Assuming and substituting in (12-5-1) we obtain, (for

brevity we write and instead of and , respectively),

or dividing through by and rearranging,

for reasons explained in the preceding section 12-4. Why we have called the
separation constant (instead of ), will become apparent soon. At
this point, we do not know anything about . The allowed values of will be
determined (as we perhaps anticipate) from the appropriate boundary
conditions. The original P.D.E splits into the following two ordinary differential
equations, (coupled through the separation constant),

The first equation in (12-5-8) takes the form

 430

This is a Cauchy-Euler type of differential equation for which the method of

solution is known. We set and try to determine so that the
assumed expression satisfies the differential equation identically. Simple
calculations show that either or and hence

Regarding the second equation in (12-5-8) if we make the substitution

, the equation becomes (let the reader check it),

which is a Legendre’s type of equation and hence its solution is

or, in terms of the original variable ,

The potential function will be:

Note that if the origin ( ) belongs in the domain of the solution, then
we must take , (otherwise the potential would be unbounded at the
origin). Reasoning similarly, if the potential is to remain bounded on the axis
( or , i.e., either or ) then we must take , since
as we have explained .

As it was the case with the rectangular coordinates, we have an infinite

number of solutions, one solution for each .

Since Laplace’s equation is a linear equation, the general solution of

Laplace’s equation is the linear combination of the separable solutions, i.e.
 431

The following examples illustrate the method.

Example 12-5-1: (Dielectric sphere in a uniform electric field).

Consider a uniform field existing in a homogeneous dielectric material

(of infinite extend) with dielectric constant . A dielectric sphere of radius
and dielectric constant is placed inside the electric field . Find the electric
field inside and outside the sphere.

Solution

Fig. 12-10: Dielectric sphere in a uniform electric field.

Let us set up a spherical coordinate system, with the axis parallel to the
field . Close to the sphere, the field ceases from being uniform (due to the
presence of the dielectric sphere). However, far away from the sphere, we
expect the field to be unaffected. This means that far away from the sphere,
the field is still uniform and equal to .
 432

Let us call region 1 the region inside the sphere ( ) and region 2 the
region outside ( ). We shall use the subscript 1 for all quantities referring
to region 1, and the subscript 2 for the quantities referring to region 2.

Region 1: Since the problem has azimuthal symmetry, (no dependence on

), the simplest expression for the potential is given by equation (12-
5-9) with , i.e.

Notice that since must remain bounded at the center , . Also,

since the potential must remain bounded on the axis, . The term
in (*) is actually the term , (recall ).

Region 2: The potential in region 2, is given by (12-5-9) with ,

i.e.

Since does not belong in region 2, our solution includes the term
.

Now we have to check whether the assumed expressions for the potential
in regions 1 and 2 satisfy the following boundary conditions on the surface of
the sphere:

1) Continuity of the potential function

2) Continuity of the tangential component of the electric field

3) Continuity of the normal component of the electric field

Expressed in Mathematical terms, these three conditions imply the

following equations,
 433

Using the expressions of the potentials and , equations in (***) yield

Since we have two equations and three unknowns, we need one more
equation. This may be obtained from the following considerations.

The electric field in region 2, is

Far away from the sphere ( ), the field should be unaffected by the
presence of the sphere, and this means that

Having determined , the constants and are found from equation

(****) , (for detailed calculations see Pr. 12-5-2), and substituting in (*) and
(**) we obtain the sought for expressions for the potential in regions 1 and 2:

Curiously enough, the electric field inside the sphere (region 1) is uniform
and is given by the formula (let the reader prove it)

The components of the electric field in region 2 are

 434

(For a proof, see Problem 12-5-3).

Taking the case we have the case of a spherical hole in a dielectric

, while taking the case we have the case of a dielectric sphere in
vacuum.

Example 12-5-2: A hollow conducting sphere is biased as shown in Figure

12-11. Determine the potential outside the sphere.

Solution

Fig. 12-11: The potential of a biased conducting sphere.

Since the problem possesses azimuthal symmetry (no dependence on ),

the general solution of Laplace’s equation is given by equation (12-5-10), i.e.

The potential should approach zero as and this implies that .

Also, since the potential has to be bounded on the axis ( ),
and the expression for the potential assumes the form
 435

At the surface of the sphere ( ), we have:

By virtue of (**) equation (*) implies

or if we define ,

Multiplying both sides by and integrating between

and , we get:

and taking into account the orthogonality of Legendre’s polynomials as

expressed in equation (12-5-6),
 436

This formula can be simplified even further if we consider the following

properties of Legendre’s polynomials:

by means of which we obtain, and

By virtue of formula (***) the potential in (*) assumes the form:

Evaluating the first few coefficients by direct calculations (see Pr. 12-5-4) we
find

For the potential inside the sphere, see Problem 12-5-10.

Example 12-5-3: (Spherical conductor in a uniform electric field).

Consider a conducting sphere of radius , grounded (potential zero), placed

in a uniform electric field parallel to the axis, ( ). Find the potential
everywhere outside the sphere.
 437

Solution

Close to the conducting spherical surface, the field lines must be

perpendicular to the surface. However, far away from the conducting sphere,
the field must be uniform and equal to . This picture is shown in Fig 12-12.

Fig. 12-12: Spherical conductor in a uniform electric field.

We notice that far away from the conducting surface, , since in

this case, , as expected. The rectangular
coordinate , and this implies that

The potential outside the sphere satisfies Laplace’s equation

whose solution is given by equation (12-5-10), i.e.

Since the potential has to be bounded for and , (i.e. along the
axis), and the solution becomes
 438

where we have renamed and . From equation (**)

we have

and this shows that the only surviving term is the one corresponding to
and is zero for all other values of , (recall that ), i.e.
and for . Then equation (**) becomes

To determine we use the condition from which we obtain

and finally the potential outside the conducting sphere is given by

Example 12-5-4: In Example 12-5-3, find the induced charge density on the
surface of the sphere.

Solution

and using the expression for the potential in equation (***) in the
preceding Example, we find

This equation shows that positive charge exists on the upper half of the
sphere ( ) while negative charge exists on the lower half of the
sphere ( ). The total charge on the surface is zero, as it should be
(since the sphere was initially uncharged and isolated). In Problem 12-5-11 we
shall work the same problem but with the addition of an initial charge on the
spherical surface.
 439

Remark: Equation (***) in Example 12-5-3, shows that the potential may be
considered as the superposition of two potentials, the first term
being the potential of the uniform field and the second term
being the potential of a dipole moment
placed at the origin, (see Equation (5-6-3)). The dipole moment is due to the
separation of charges as explained in Example 12-5-4.

PROBLEMS

12-5-1) Use Rodrigues’ formula to find and .

12-5-2) In Example 12-5-1, perform analytic calculations to find the

constants and .

12-5-3) In Example 12-5-1, perform analytic calculations to find the

components and in region 2.

12-5-4) In Example 12-5-2, starting with equation (****) derive the formula
(*****) for the potential .

12-5-5) Show that .

12-5-6) Find the solution of the differential equation

, which is bounded in the closed interval .

Hint: This is a Legendre’s equation with and its solution is .

12-5-7) Expand the function in a series of

Legendre polynomials.

(Ans: ).

Hint: Use formula (12-5-7). Notice that is an odd function of , while

is odd for odd and is even for even, and therefore is odd
for even and is even for odd. As a result, for even,
while for
.
 440

12-5-8) Show that .

Hint: Notice that if (orthogonality of the

Legendre’s polynomials), and this completes the proof since .

12-5-9) A hollow conducting, grounded sphere of radius is placed

at the mid plane between two parallel plates, as in Fig. 12-13. Determine the
potential at .

Solution

Fig. 12-13: A conducting sphere between two parallel charged planes.

(Ans: ).

Hint: Since the electric field far away from the

sphere will practically remain uniform. The problem reduces to that of Example
12-5-3 and hence we may use formula (***) found in Example 12-5-3, for the
potential . Notice that .

12-5-10) In Example 12-5-2 determine the potential inside the

sphere.

(Ans:

Hint: The potential must remain bounded at the origin ( ) and on the
axis.
 441

12-5-11) A conducting sphere of radius , carrying a charge is placed in a

uniform electric field parallel to the axis, ( ). Find the potential
outside the sphere and the induced surface charge density.

(Ans: ).

Hint: Consider the expression (***) for the potential in Example 12-5-3 and
apply the superposition principle.

12-6) Laplace’s Equation in Two Dimensions (Cylindrical Coordinates)

When a problem possesses cylindrical symmetry, it is convenient to work in

cylindrical coordinates. The method of separation of variables in cylindrical
coordinates leads to the so called “Bessel’s functions”, which are the solutions
of the “Bessel’s differential equation”, named after the German
Mathematician F.W. Bessel. A very brief account of these functions and their
properties (which will be useful in the sequel), is given below.

a) Bessel’s functions.

Bessel’s functions result as the solutions of “the Bessel’s differential

equation” which is of the form

where is a real number.

The solution of equation (12-6-1) is obtained using the method of

Frobenious, which was also used to solve Legendre’s differential equation in
section 12-5. The computations, tedious but straightforward, show that the
general solution of (12-6-1) in the interval is given by

where is the Bessel’s function of the first kind of order and

is the Bessel’s function of the second kind of order . It can be shown (by
using the method of Frobenious) that is given by the infinite series
 442

If , then since the Gamma function

, formula (12-6-3) assumes the form

For we obtain the Bessel function of the first kind of order zero

Notice that while for

The function is the bounded solution of the differential equation

, in the interval , (obtained from (12-6-1) with
).

The expression for the Bessel’s function of the second kind is, by far,
more complicated as compared to . It can be shown that goes to
infinity as , i.e. . This implies that the bounded solution of
Bessel’s differential equation on the interval is obtained from
equation (12-6-2) by setting , and therefore, in such a case

However, the solution of Bessel’s equation on the interval , (the

point is not included), must contain both and .

Bessel’s functions satisfy certain orthogonality conditions, which, as we

know, can be used in order to expand a given function in a series of Bessel’s
functions. We shall not cover the general case here, but we confine ourselves
to the zero order Bessel function of the first kind and just state the
following result:

A function defined on the interval can be expanded in a

series of Bessel’s functions of the first kind and of zero order as follows:
 443

where are the zeros of , and the expansion

coefficients are given by the formula

Notice that has an infinite number of real roots , in

the interval ; it can be shown that, the large positive roots of
are approximated by , where is a large positive integer.

b) Solution of Laplace’s equation in Cylindrical coordinates.

Laplace’s equation in cylindrical coordinates is (see equation (1-11-1))

In problems where there exists azimuthal symmetry, i.e. no dependence on

, , and Laplace’s equation assumes the form

Assuming that , substituting in (*) and separating the

variables leads to the following equation, ( is the separation constant),

which leads to the following system of two equations ( )

 444

The solution of the first equation in (**) is ,

while the solution of the second equation is
and a solution of Laplace’s equation is

Notice that if the axis ( ) belongs in the domain of solution, then,

since the potential must remain bounded, we have to take , so that
the term which blows up at , is not included in the solution.
Therefore, in this case, the bounded solution of Laplace’s equation is

where we have renamed and .

Of course, up to this point, is not known. The separation constant , as

usually, is determined from the appropriate boundary conditions. For each ,
we have one valid, bounded solution. The general solution of Laplace’s
equation is obtained by superposition, over all the allowed values of .

The following examples illustrate the method.

Example 12-6-1: Consider the hollow conducting cylinder in Fig. 12-14. The
top surface is kept at potential , while the rest of its surface is
grounded (potential zero). Find the potential everywhere inside the cylinder.
(Assume that the top surface is isolated from the rest of the cylinder by an
infinitesimal isolating gap).

Solution

The potential obviously, does not depend on , and since the potential
has to be bounded everywhere inside the cylinder, the potential inside
the cylinder is given by equation (12-6-10):

The boundary condition implies that , i.e. and

therefore
 445

Also, on the lateral surface , and this condition implies

that , i.e. must be one of the (infinite in number) positive
roots of . If we call the root of , we must have

Fig. 12-14: The potential inside a hollow conducting cylinder.

For each we have one valid solution for , satisfying Laplace’s

equation and the homogeneous boundary conditions (
). The general solution is obtained by superposition, i.e.

The coefficients must now be determined, so that , i.e.

Making use of formula (12-6-8) and the identities

 446

we may find first and then , and substituting in (****) we finally find

Example 12-6-2: Solve Laplace’s equation in cylindrical coordinates,

assuming that there is no dependence on , i.e. the potential depends only on
and . (As we shall see, the solution in this case is much simpler, as it does
not contain Bessel functions).

Solution

Assuming , the term and Laplace’s equation

becomes,

Setting and separating the variables, results in the

following two equations (let the reader check it):

where is the separation constant. The first equation in (*) is a Cauchy-

Euler type of equation, (see Section 12-5 (b)), and its solution is
, if and is if , while the
solution of the second one is if and is
if .

The constant is determined, as always, from the associated boundary

conditions. For each we have a solution of Laplace’s equation, i.e.
 447

The general solution of Laplace’s equation is obtained by superposition, i.e.

As an application of this general solution, let us consider the following

problem.

Example 12-6-3: Consider an infinitely long, conducting, hollow cylinder,

coaxial to the axis. The half cylinder is kept at potential while
the other half is kept at potential . Find the potential
everywhere inside the cylinder.

Solution

Fig. 12-15: The potential inside a hollow conducting cylinder.

The two parts of the cylinder are separated by infinitesimal gaps, one gap
being the straight line parallel to the axis and perpendicular to the plane
at and the other gap being perpendicular to the plane at – , as
shown in Fig. 12-15.

Since the potential in this case does not depend on , the solution of
Laplace’s equation inside the cylinder ( ) is given by
equation (**) in Example 12-6-2, i.e.
 448

Since the potential must be finite on the axis, (at ), and

, assuming , and equation (*) reduces to the following:

The periodicity condition implies that and

integer, i.e. the allowed value of are , and equation (**)
becomes

where . At the boundary ,

and equation (***) implies

This expression is the Fourier series expansion of the function

, and the expansion coefficients are given by the formulas
 449

The first formula yields that , while the second

formula yields

(let the reader check it), and substituting in formula (***) we find:

or, since only the odd terms ( ) contribute to the sum,

For an expression of the potential outside the cylinder, see Problem 12-6-4.

PROBLEMS

12-6-1) An infinitely long dielectric rod in the form of a cylinder of radius ,

with its axis coincident with the axis, is placed in a uniform electric field
in free space. Find the potential inside and outside the cylinder.
(Assume that the dielectric constant of the dielectric is ).

(Ans: , ).

Hint: The potential does not depend on . Use the expression for found
in Example 12-6-3 and also, see Example 12-5-1.

12-6-2) In the preceding problem find the electric field everywhere.

Hint: .

12-6-3) In Problem 12-6-1, show that the field inside the cylinder is
uniform, find its expression and give a physical interpretation.

(Ans: ).
 450

12-6-4) In Example 12-6-3, find the potential outside the cylinder ( ).

12-7) Method of Images

Problems involving point charges and conducting boundaries can

(sometimes) be solved by a method (technique) called “the method of
images”. This method of solution, introduced by W. Thomson, is best
illustrated by means of a few examples. By virtue of the “uniqueness theorem”
discussed in section 12-2, any expression for the potential , which satisfies
Laplace’s equation and the boundary conditions is a unique solution. Based on
this theorem, any solution obtained by the method of images, will be the
unique solution to the electrostatic problem at hand.

Example 12-7-1: Find the field produced by a positive point charge placed
a distance above the infinite, grounded conducting plane at .

Solution

Fig. 12-16: A positive charge above an infinite, grounded plane.

In Fig. 12-16 (a), the field lines begin from the positive charge and end
(perpendicularly) on the grounded plane , whose potential is zero. The
field in the region , is identical to the field produced by the two charges
 451

and – , and no conducting plane, as shown in Fig. 12-16 (b). Notice that
the charge is not a real charge; it is a fictitious charge, located
symmetrically below the plane , which is called the electric image of .

Fig. 12-17: The field of the two charges and – , in the region .

The potential at the point , due to the two charges and is

This expression for the potential, in the region , satisfies Laplace’s

equation (since the term due to each charge satisfies Laplace’s equation) and
also for points lying on the plane, and hence .
Therefore, expression (12-7-1) satisfies Laplace’s equation and the boundary
condition on the plane , and according to the uniqueness theorem,
is the only solution to our problem (in the region ).

If the conducting plane, (in Fig. 12-16(a)), were held at a potential

, then the solution of the problem would be
 452

since this expression satisfies Laplace’s equation and the boundary

condition .

Having determined the potential , we may calculate the electric field

and the surface charge density on the conducting plane
.

Example 12-7-2: A positive point charge is located at above the

infinite grounded plane . Find the surface charge density , induced on
the conducting plane.

Solution

Fig. 12-18: Induced charge on the conducting plane .

Application of formula (12-7-1) yields,

and the component of the electric field at the plane is

 453

Notice that on the plane, as expected (why?). Let the

reader verify by direct calculations.

The surface charge density is

The integral , as expected. For detailed calculations

see Problem 12-7-1.

The method of images may be extended to more complicated problems. In

Example 12-7-3 we find the potential of a point charge located between two
perpendicular conducting planes.

Example 12-7-3: A point charge is located between the two conducting,

perpendicular planes and , each one of which is grounded. Find the
potential everywhere in the region .

Solution

Fig. 12-19: A point charge , placed between two perpendicular

conducting planes held at potential zero.

The original charge is imaged first in the plane to and then

in the plane to , as shown in Fig. 12-19. However, this charge
configuration, original charge and the two image charges does not
 454

make the potential to be zero on the two planes, (as it should be). In order to
make the potential of the two planes to be zero, the two image charges must
also be imaged to the image charge , as shown in Fig. 12-19. Then the
potential at the point is

This expression of the potential satisfies Laplace’s equation and the

boundary conditions , and according to the
uniqueness theorem, is the unique solution to our problem.

Remark: If the angle between the two conducting planes is , where

is a positive integer, then the number of image charges is . For
instance, in our example and we have image charges, (see
Pr. 10-26, in Supplementary Problems). However, if the angle between the two
planes is not a sub multiple of , then an infinite number of image charges is
required. In this case, the method of images breaks down.

The method of images can be applied to find the potential due to a line
charge parallel to an infinitely long conducting cylindrical surface or to a
point charge and a conducting sphere. Let us consider the following examples.

Example 12-7-4: Images with cylinders and spheres.

a) Line charge parallel to an infinitely long, grounded, conducting

cylindrical surface.

Fig. 12-20: A line charge parallel to a grounded cylindrical surface.

 455

In Fig. 12-20, the image line charge and its position are given by the
formulas:

In the equivalent system, ( and ), the conducting surface does not

exist. However, the potential on the cylindrical surface due to and , as
defined in (12-7-4), is zero, as it were in the presence of the conducting
cylindrical surface.

For a proof, see Problem 12-7-2.

b) Point charge close to a grounded, conducting sphere.

Fig. 12-21: A point charge close to a grounded conducting sphere.

In Fig. 12-21, the image point charge and its position are given by the
formulas:

In the equivalent system, ( and ), the conducting surface does not

exist. However, the potential on the spherical surface due to and , as
defined in (12-7-5), is zero, as it were in the presence of the conducting
spherical surface.

For a proof see Problem 12-7-3.

 456

Remarks: (a) In both cases, (line charge parallel to a grounded cylindrical

surface and point charge close to a grounded spherical surface), the image
technique provides the potential to the exterior of the cylindrical or the
spherical surface, i.e. the region containing the line charge or the point
charge , respectively.

b) In general, we must never place an image charge within the region

where we want to compute the potential, since then we simply alter the
original charge distribution. Recall that in the region where we want to
compute the potential, we must have , where is the original
charge distribution.

Example 12-7-5: In Fig. 12-21, and . Find

the potential at the point . Assume that is positioned at
the origin and that the center of the sphere lies on the axis.

Solution

By virtue of equation (12-7-5) we have:

Thus, the original problem reduces to the one shown in Fig. 12-22.

Fig. 12-22: The field of a point charge close to a conducting sphere.

 457

where and
, and substituting in (*) we find, .

Example 12-7-6: Point charge outside a semi-infinite dielectric slab.

A point charge , is placed at the point in free space, in front of

the semi-infinite dielectric slab ( ) of
permittivity , as shown in Fig. 12-23. Find the potential everywhere.

Solution

Fig. 12-23: A point charge close to a dielectric slab.

We shall solve this problem using a method, similar to the method of

images.

We assume that the image of is at , while the image of

is at , as shown in Fig. 12-23.

Notice that, as yet, we do not know the charges and , which are still
to be determined.

We assume that the potential in free space, (region 2, ), due to

and , is given by the expression

or equivalently,
 458

We further assume that the potential in the dielectric, (region 1, ),

is produced by the charge , placed at the position of the original charge, and
is therefore given by the formula

The continuity of the potential function at (the boundary between

free space and dielectric), yields

where is the relative permittivity of the dielectric.

The continuity of the tangential component of the field on the boundary

, leads again to equation (***), (let the reader check it).

From the condition we obtain,

Solving the system of equations (***) and (****) for and we find

Having determined and we obtain the potential in free space from

equation (*) and the potential in the dielectric region from equation (**),
(see Problem 12-7-6).

PROBLEMS

12-7-1) Starting with equation (12-7-3) derived in Example 12-7-2, show

that .
 459

Hint: Work in polar coordinates ,

etc.

12-7-2) Prove equation (12-7-4), in Example 12-7-4.

Hint: It suffices to show that the potential due to and satisfies

Laplace’s equation and the boundary condition .

12-7-3) Prove equation (12-7-5), in Example 12-7-4.

Fig. 12-24: Proof of equation (12-7-5).

Hint: The potential at the point , on the spherical surface is

. Using the law of cosines, we have,

and , and if we
use the expressions for and as given in equation (12-7-5), we find
, (same potential as in the spherical conducting surface).

12-7-4) A point charge of is located on the axis from a

grounded, conducting plane at . Find the surface charge density at
and the electric field at .

Hint: See Example 12-7-2.

12-7-5) In Example 12-7-1 show that the positive charge is attracted to

the conducting plane by a force .
 460

12-7-6) In Example 12-7-6 an electron is placed at outside of a

uniform dielectric of relative permittivity , (region 1 in Fig. 12-23). Find
the potential at the points and .

(Ans: ).

Hint: Find and from equation (*****) in Example 12-7-6, and then
find the potential using equations (*) and (**).

12-7-7) In Problem 12-7-6, find the force on the electron.

(Ans: ).

Hint: The force on the electron is , where ).

SUPPLEMENTARY PROBLEMS

12-1) In Fig. 12-2, show that the solution of Poisson’s equation ,

which satisfies the boundary conditions is unique. This shows
that the potential in the region is uniquely determined if the charge density in
the region and the potential on the conducting surfaces are specified.

12-2) A second form of the uniqueness theorem: If the total charge on

each conductor is given, show that the electric field of the system is uniquely
determined. Consider the system in Fig. 12-2, total charge on conducting
surface is and total charge on is . Assume that the region between
the two conductors is charge fee ( ).

Hint: Assume that there are two electric field and , and show that
.

12-3) A potential field , exists between two

parallel conducting plates, each having an area of . The separation
distance between the plates is . Find the voltage between the
plates, the energy density and the capacitance, assuming that a dielectric of
exists between the plates.

(Ans: ).
 461

12-4) For what positive value of is a solution of

Laplace’s equation in cylindrical coordinates? For this value of , find
.

12-5) Find all solutions of Laplace’s equation (in spherical coordinates), that
depend solely on .

(Ans: ).

12-6) Show that the potential is a solution of

Laplace’s equation, provided that .

12-7) Consider the rectangular trough in Fig. 12-25. The side is kept
at a potential , the side is kept at potential , while the other two
sides are grounded. Find the potential at all points inside the trough.

(Ans:

Fig. 12-25: The potential inside a rectangular trough.

 462

Hint: Find the potential assuming that the side is kept at

potential while the other three sides grounded, and then find the potential
assuming that the side is kept at potential while the other
three sides grounded. If both and apply simultaneously, then by
superposition, .

12-8) In the rectangular trough of Fig. 12-25 the side is kept at

potential , while the other three sides are grounded. Assuming that
, find the potential at all points inside the trough
and plot the equipotential lines .

Ans: , and
measured in .

12-9) In the rectangular trough of Fig. 12-25, the side is held at a

potential , while the other three sides are grounded. Find the
potential inside the trough.

(Ans: ).

12-10) In Example 12-5-2 assume and find the potential at the

points and .

Hint: Use formula (*****) in Example 12-5-2. In our case, ,

, and

Similarly, , since and

.

12-11) In Problem 12-5-10, assume and find the potential at the

points and , (keep the first two terms of the series).
 463

12-12) Find the acceleration of an electron, placed a distance above an

infinite, grounded conducting plane at .

12-13) A homogeneous infinite dielectric with is placed in a

uniform field . A spherical cavity of radius is cut out of the dielectric.
Find the electric field at the center of the cavity.

(Ans: ).

Hint: Apply formula (*****) in Example 12-5-1.

12-14) A dielectric sphere of radius and relative permittivity is placed

in a uniform field in free space. Show that the sphere gets polarized
and behaves as an electric dipole .

Hint: See formula (5-6-3) and the expression for the potential
found in Example 12-5-1.

12-15) Consider two infinitely long, conducting cylindrical surfaces, coaxial

to the axis, with radii and respectively. The inner
surface is charged to potential while the outer one is grounded. A
dielectric sphere of radius and permittivity is placed at
. Find the force exerted on the sphere.

(Ans: ).

Hint: Solve Laplace’s equation in cylindrical coordinates, find the potential

between the two cylindrical surfaces, and then show that
. At the position of the sphere ( ), the electric field is

. Due to the very small size of the sphere, we my safely

assume that the field lines, in the neighborhood of the sphere, are almost
parallel, in other words we may assume that the dielectric sphere is placed in a
uniform field , (in the vicinity of the dielectric sphere). The
sphere gets polarized and behaves as an electric dipole , as shown in Problem
12-14. The force on an electric dipole is given by equation (8-2-9), i.e.
 464

12-16) A line charge density is placed at a distance above an infinite,

grounded conducting plane, at . Show that the induced surface charge
density at a point on the plane is , where is the
shortest distance from to the line charge.

12-17) A conducting sphere of radius is given a charge and then is

placed in a uniform field . Find the surface charge density of the
induced charge and show that the least amount of such that no part of the
sphere is negatively charged is .

Hint: For an expression of the potential outside the sphere see Pr. 12-5-11.
The charge density .

12-18) In Problem 12-17, assume that and . Find

at the points and .

12-19) The radius of a nucleus is given by , where

is the nuclear mass number. Determine the electrostatic potential at the
center of a nucleus, (92 protons), (Ans: ).

Hint: Assume that the charge within the nucleus is uniformly distributed,
with density , and then use the expression for the
potential found in Example 5-4-2, formula (******).

12-20) Find the potential and the electric field produced by a uniformly
polarized sphere of radius , ( inside the sphere and is zero outside).
 465

(Ans: .
The electric field . Note that the electric field inside the sphere is
uniform ).

Hint: The bound surface charge density is ,

while the bound volume charge density is . The potential is
thus given by the integral , etc.

Fig. 12-26: The potential and the electric field of a uniformly polarized
sphere.

Note that , . The

angle varies from to , while varies from to .

12-21) An electric field is applied to a uniform dielectric with

relative dielectric constant . Determine the electric field in the
neighborhood of a molecule, (Ans: ).

Hint: Assume that we cut out an imaginary sphere in the neighborhood of

the molecule, and since the radius of the sphere is extremely small, we may
safely assume that the polarization vector is the same in the small
neighborhood around the molecule. The polarization surface charge density is
, since . The electric field at the point , due to the
 466

bound surface charge density is

where . Note that the vector is not constant, in the

integration, so we have to express in terms of the constant unit vectors
and , (see formula (1-3-5), etc. The result of the integration is .

Fig. 12-27: The field in the neighborhood of a molecule.

The total field at the point is found by superposition, i.e. is the

superposition of the applied field plus the field due to the superposition, i.e.

since .

12-22) A uniform dielectric slab is placed between the plates of a capacitor

which are separated by a distance . The electric field in the
neighborhood of a given molecule is . Assuming that the voltage
between the plates is , determine the relative dielectric constant
of the dielectric slab, (Ans: ).

Hint: Use the result obtained in Problem 12-21.

 467

12-23) An electron is located at within a uniform dielectric

that has . Determine and at .

(Ans: ).

12-24) Two equal charges are placed a distance apart. Each charge is
at a distance above an infinite conducting plane, kept at potential zero
(grounded). Find the magnitude of the force between the charges. What is the
expression for the vector force ?

(Ans: ).

12-25) A point charge is placed at a distance form the center of a

grounded conducting sphere of radius , as shown in Fig. 12-28. Find the force
on the charge .

Fig. 12-28: Force between a point charge and a conducting, grounded

sphere.

(Ans: Attractive force ).

Hint: Use formula (12-7-5).

12-26) Outline the method to find the potential due to a point charge
placed between two conducting planes at , kept at zero potential.

Hint: , , so we will have images, (see Remark

in Section 12-7).

12-27) A point charge is placed between two perpendicular, infinite,

grounded, conducting planes, as in Fig. 12-29. Find the force on the charge.
 468

Fig. 12-29: A point charge between to perpendicular, grounded planes.

(Ans: ).

12-28) Consider an infinite, grounded conducting cylinder of radius

, with its axis coincident to the axis and an infinite line charge
, parallel to the axis at . Find the electric
field at the point .

Hint: Use formula (12-7-4).

12-29) An electric dipole of dipole moment is at a distance from the

center of a grounded conducting sphere of radius , as shown in Fig. 12-30.
Show that the image is a dipole with moment and a charge
at a distance from the center of the sphere.

Fig. 12-30: The image of a dipole close to a grounded, conducting sphere.

 469

12-30) In Problem 10-29, , and .

Find the potential and the electric field at the point
.

12-31) A dielectric sphere of radius and relative permittivity

is placed at the point , while a positive charge
is placed at the origin, as shown in Fig. 12-31. Show that the
dielectric sphere is attracted towards the point charge by a force
.

Fig. 12-31: Force of attraction between a dielectric sphere and a point

charge.

Hint: Due to the extremely small size of the dielectric sphere, the electric
field in the neighborhood of the sphere is approximately uniform, and equal to
. The dielectric sphere behaves as an electric dipole with

, (see Problem 12-14). The force on the sphere is

, (see Pr. 12-15).

12-32) An infinite conducting cylinder of radius , with its axis coincident to

the axis, is placed inside a uniform field . Assuming that the
potential of the cylinder is zero, find the potential outside the cylinder.

(Ans: ).
 470

Hint: See Examples 12-6-2 and 12-6-3. Note that, far away from the
cylinder, i.e. as , , etc.

12-33) In Problem 12-32, find the electric field on the surface of the
cylinder and the surface charge density .

(Ans: ).

12-34) In Problem 12-32, consider a small dielectric sphere with radius

and dielectric constant , placed at the point .
Assume and . Using the method employed in
Problems 12-14 and 12-31, find the force on the dielectric sphere.

12-35) In Problem 12-32, find the electric field everywhere, outside the
cylinder.

Hint: .
 471

PART B: MAGNETOSTATICS
 472

CHAPTER 13: STEADY CURRENTS

13-1) General Concepts and Definitions

a) In the preceding chapters we have developed the general laws of the

static electric fields. Recall that an electric field is called static, if:

1) All the characteristic quantities related to the fields, (potential , electric

field , charge density , etc), do not depend on time, i.e. are time
independent, though they may be functions of the spatial coordinates, and

2) For the maintenance of this static state there is no need for a continuous
energy supply.

A static electric field, once installed in a region in space, containing

conductors and dielectrics, stays there forever, (at least theoretically), all the
involved quantities do not vary with time, and for the maintenance of this
state there is no need for any energy supply. Of course, this scheme describes
an ideal system, with perfect conductors and perfect insulators.

In reality, perfect insulators do not exist in nature. For this reason, a fully
charged car battery, when disconnected, gradually loses its strength, and
after some time, its voltage may drop to or even less, (the battery is
discharged through air). In order the battery to maintain its original voltage of
, the battery must be continuously charged, i.e. energy must be supplied,
continuously, to the battery. In this case, we note that the battery voltage
remains constant ( ), but for the maintenance of this state, there is a need
for a continuous energy supply. This state is called “steady state”.

So, in general, a state of a system is called steady, if:

1) All the characteristic quantities do not depend on time (are time

invariant), though they may be functions of the spatial coordinates, and

2) For the maintenance of this state there is a need for a continuous

energy supply.

As a characteristic example, let us consider a Direct Current circuit (DC

circuit). In such a circuit, the currents flowing in the circuit and the voltages
between any two points in the circuit do not vary with time, but for the
 473

maintenance of this state, the energy sources of the circuit (batteries), supply
energy, continuously, to the circuit.

13-2) Electric Current

In very general terms, the oriented motion of electric charges, is called

“electric current” or just, for brevity, “current”. In conductors, the current is
due to the motion of free or conduction electrons, in semiconductors the
current is due to the motion of electrons and holes, (as described in Section
13-3), while in electrolytes the current results from the motion of positive and
negative ions.

Let us, for example, consider a filamentary conductor. By definition, the

conductor contains a huge number of “free or conduction electrons” (of the
order of ), (see Section 9-1). These remote electrons are
loosely bound to their parent atoms and are therefore, almost free and not
associated with any particular atom. They form a “cloud of free electrons”.
These electrons, due to their thermal motion, are wandering within the
conductor, executing “random walks”. We may assume that the free electrons,
while moving randomly within the conductor, collide elastically with each
other and with the heavy nuclei of the atoms, exchanging thus their energies.

In this picture, there is no any preferred direction of motion. If we imagine a

small surface within the conductor, we expect that, during a time interval
, as many electrons cross in one direction, the same number of electrons
to cross the same surface in the opposite direction. Since the electron is the
carrier of the electric charge, we conclude that the amount of charge crossing
in one direction is the same as the amount of charge crossing the same area
in the opposite direction, within the same time interval , i.e. the net
amount of charge crossing is practically zero.

If we want to give a prominent orientation to the motion of the electrons,

we must help by exerting on them an electric force , with the aid of an
externally applied electric field . Then the electric force will be ,
where is the charge of the electron. Given that is negative, the force
points in a direction opposite to that of the electric field .

Question: How can we produce the electric field ?

 474

Answer: By imposing a potential difference at the ends of the conductor.

In this case, the actual motion of the electrons will be the superposition of
two motions, the random thermal motion and the motion due to the force ,
caused by the externally applied field . This actual motion, has now a
prominent orientation (the direction of the applied field ), and as a result we
have now a non zero electric charge through the surface , during the time
interval .

Since the protons are about times heavier than the electrons, to a
first approximation, we may safely assume that under the influence of an
externally applied electric field , only the electrons are moving while the
heavy nuclei remain at rest (even though they also experience an electric
force).

If we focus our attention to a particular electron, this electron, under the

influence of the electric force is accelerated, and then it collides with
either another electron or a heavy nucleus and momentarily comes to a stop;
right after the collision, the electron, under the influence of the electric force,
is accelerated again, until a new collision takes place, etc. As a result of this
repeated process, (acceleration-collision-abrupt stop, acceleration-collision-
abrupt stop, etc-etc), the electrons, macroscopically, seem to move with an
average constant velocity , called “drift velocity”, which is found to be
proportional to the applied field , i.e.

The positive coefficient is called “the mobility of the electrons” and

depends on the material of the conductor and the temperature. The mobility is
measured in , as obtained readily from (13-2-1). The negative sign
in formula (13-2-1) shows that the drift velocity, (as a vector), is in the opposite
direction of the applied field .

For example, the mobility of Copper is , the

mobility of Silver is , while the mobility of
Aluminum is .
 475

It is worth mentioning that within the conductors, the electrons are moving
at very low speeds, of the order of , or even lower, (see Ex. 13-4-1).

13-3) A Convention about the Current Direction

Good conductors are “unipolar materials”. This simply means that their
conductivity, i.e. their ability to allow electric currents to flow through them, is
due to one type of charge carriers, free electrons in our case.

However, there are other materials which are “bipolar”. This means that
their conductivity is due to two types of charge carriers. As an example,
semiconductors are bipolar elements, in the sense that their conductivity is
due to the motion of “free electrons” (carriers of negative charge) and to the
motion of “holes” which are the carriers of positive charge. Strictly speaking,
the consideration of a “hole” as a particle of mass and carrying a
positive charge , equal in magnitude with the charge of the electron, is fully
justified by Quantum Mechanics, and the analytical calculations are in
complete agreement with the experimental results.

Electrolytes and Plasma (an ionized gas consisting of negative electrons and
positive ions), are also examples of “bipolar materials”.

For such bipolar materials, where the conductivity is due to two types of
carriers, one type being the carrier of negative charge and the other type being
the carrier of positive charge, we must take into account the following
convention:

1) The motion of positive charge in one direction results in a current flow in

the same direction, (see Fig. 13-1 (b)), and

2) The motion of negative charge in one direction results in a current flow

in the opposite direction, (see Fig. 13-1 (a)).

Let us consider Fig. 13-1. For the given direction of the electric field , the
electrons drift to the left with velocity , and this is equivalent to a
current flow in the opposite direction, i.e. in the direction of the applied field
. At the same time, the positively charged particles, (holes in semiconductors
or positive ions in electrolytes), drift to the right with velocity ,
 476

where is the mobility of the positively charged particles, and this is

equivalent to a current flow in the direction of the applied field .

Fig. 13-1: Flow of current in relation to the charge of the particles.

The important thing to notice is that, in bipolar elements, even though the
charged particles (electrons and holes or positive ions) drift in opposite
directions, still, their corresponding currents flow is in the same direction,
that dictated by the applied electric field .

The two most common types of semiconductors are Germanium (Ge) and
Silicon (Si). The mobility of the electrons and the holes in these two
semiconductors has been determined experimentally to be:

These values for the corresponding mobility refer to pure semiconductors,

at room temperature. In practice, in order to increase the conductivity, the
 477

semiconductors are doped with certain types of impurities, so that we may

have either “n - type of semiconductors”, where the majority of the carriers
are electrons, or “p – type of semiconductors”, where the majority of the
carriers are holes.

13-4) Electric Current ( ), Current Density ( ), and the Point Form

of Ohm’s Law

(a) Let us consider a small, differential cylinder of base and height , as

shown in Fig. 13-2, which is part of a filamentary conductor.

Fig. 13-2: The definition of current.

Under the action of an electric field , the free electrons in the conductor
acquire a drift velocity , where is the mobility of free electrons,
(see equation (13-2-1)). As explained in Section 13-3, Fig. 13-1 (a), this motion
of the electrons constitutes a current, flowing in the direction of the applied
field .

The electric current or just current , flowing in the filamentary conductor,

is defined as the time rate at which electric charge cross the area . If
we assume that charge crosses the area within a time interval , then
 478

The unit of the current in the SI system is the Amber, named after the
French physicist Ampere.

In formula (13-4-1) we assume that the current does not depend on time.
However, in the most general case, the current may depend on time, and in
this case the instantaneous value of the current is given by the formula

(b) In Fig. 13-2, let us call the unit vector in the direction of , (which
coincides with in the direction of the field inside the conductor), and
consider the product . We have:

or, since ,

If we define , i.e. the electrons per

unit volume, then the charge within the infinitesimal volume , in Figure
13-2, is

where is the (negative) charge of the electron. By virtue of (*), equation

(13-4-3) yields,

In formula (**), the quantity represents the volume charge density

, i.e. represents the charge per unit volume, and is measured in . In
our case, is negative, since the charge of the electron is negative.
 479

(c) The current density :

We define the vector current density by means of the formula

The units of the current density in the SI system of units is , i.e. is

current per unit area, as it is easily obtained from the definition in (13-4-4).
Indeed, from the definition of , the unit of the current density is

Equation (13-4-4) clearly shows that electric charge in motion constitutes

electric current. If the current is due to the motion of electrons within a
metallic conductor, it is called “conduction current”. If the current is due to the
motion of electrons or protons in free space (a beam of electrons, for example,
in free space), it is called “convection current”.

In view of the definition of , in equation (13-4-4), formula (**) takes the

form,

This equivalency will prove to be very useful in the next chapter,

(computation of the magnetic fields produced by given current distributions).

(d) Ohm’s Law in point form:

By virtue of equation (13-2-1), formula (13-4-4) assumes the form,

or, if we define the “conductivity of free electrons in the metallic

conductor” as ,
 480

Formula (13-4-6) is known as the point form of Ohm’s Law. A more familiar
form of Ohm’s law will be derived shortly.

Notice that the conductivity is a positive quantity, since the charge of

the electron is negative.

The unit of the conductivity is

given that . The is symbolized as , (sometimes

written and as ), and finally, conductivity is measured in , or
equivalently as . Typically, the conductivity of metallic conductors is of the
order of . For example, for copper ,
for silver , and for aluminum
.

The reciprocal of the conductivity is called resistivity , i.e. , and

is measured in .

Formula (13-4-4) is easily generalized as follows:

If a charge density , (either positive or negative), is moving with a velocity

, then

where is the density of the charge carriers and is the charge of each
particle. If , then and point in the same direction, while if , the
and point in opposite directions.

(e) The current density in a semiconductor:

In a metallic conductor, as we have already explained, the current is due

solely to the motion of free electrons. However, in a semiconductor, (a bipolar
element), the current is due to the motion of free electrons and holes
(considered to be positively charged particles with mass , positive charge
and density ), and therefore the current density will be
 481

It is worth mentioning that, even though under the influence of an

externally applied field , the electrons and the holes drift in opposite
directions, still their currents add, (see Fig. 13-1).

(f) Isotropic and anisotropic materials:

Equation (13-4-6) shows that in a metallic conductor, the current density

depends linearly on the applied field , the constant of proportionality being
the conductivity . This is so, since a metallic conductor is an isotropic
medium, i.e. it has the same properties in all directions. However, there are
materials which are not isotropic, and these materials are called anisotropic. A
typical example is “the magneto plasma”, i.e. the plasma in a magnetic field. A
plasma is an ionized gas consisting of a quasineutral mixture of charged
particles (negative electrons and positive ions), whose dynamical behavior is
dominated by electromagnetic forces. In such cases (anisotropic materials),
Ohm’s law is still valid, however, the conductivity now is not just a number, on
the contrary it is a matrix, or as we say it is a tensor , having three rows
and three columns, i.e.

Ohm’s law in point form now takes the form,

Formula (13-4-10) shows that in an anisotropic material, it is possible, the

current density to flow in a direction other than the direction of the applied
field . For example, from (13-4-10), the component of the current density is
, etc.
 482

(g) The current through a surface in terms of the current density :

Let us consider Fig. 13-3 (a), where the current density is perpendicular to
the surface . Since is expressed in , the product expresses
the current through .

Fig.13-3: Current in terms of the current density .

Indeed, in Fig. 13-3 (a) we have:

since the differential volume .

If is not perpendicular to , then only the normal component of

contributes to the current through the surface , i.e. in this case,

In general, if the current density is given as a function of the spatial

coordinates, i.e. if , then the current through a surface is given
by the integral
 483

where is the unit out normal to the surface .

Fig. 13-4: Total current through a surface .

The total current flowing out from a closed surface , is .

(h) Surface current density :

It will facilitate our subsequent calculations if we consider surface current

densities, i.e. currents flowing over a sheet of vanishingly small thickness,
(ideally, currents flowing over a surface , as shown in Fig. 13-5).

Fig. 13-5: Surface current density .

Let us consider a differential rectangle with sides and , as in Fig.

13-5, and let be the current crossing, perpendicularly, the side . The
unit vector points in the direction of the current flow.
 484

The surface current density is defined as

This shows that , the magnitude of , is the current per unit length,
perpendicular to the current flow. The surface current density, as follows from
formula (13-4-12), is measured in .

From this definition it also follows that

Taking into account equations (13-4-5) and (13-4-13), we see that in general
we may write:

Question: Why do we need these equivalencies?

Answer: As we shall show in the next chapter, electric charges in motion,

i.e. electric currents, produce magnetic fields. While the existence of electric
charges in some region in space produces electric fields, the motion of electric
charges, i.e. electric currents, produce, in addition, magnetic fields in the
surrounding space.

The currents responsible for the magnetic fields could be linear elements
, (filamentary conductors), or surface elements , (for example a
rotating charged disk), or even volume elements , (for example a beam of
electrons in free space).

Recall that in static electric fields, we have considered three types of charge
distributions, the line charge density , the surface charge density
and the volume charge density , and in each case we
computed the electric field produced, (using the superposition principle).

Using a analogous approach, in Chapter 14, we shall determine the

magnetic fields produced by either linear elements , or by surface
elements , or even by volume elements .
 485

Example 13-4-1: Electric current flows though a Copper filamentary

conductor of cross section . Find the drift velocity of the electrons
inside the conductor.

Given:

1) The Avogadro number

2) The Copper density

3) The Copper atomic mass

Solution

Fig. 13-6: Drift velocity of electrons inside a conductor.

The current density: .

As known from Chemistry, the atomic mass of Copper, expressed in ,

contains atoms (the Avogadro number), i.e. of Copper contain
atoms and therefore conduction electrons, assuming that each atom
contributes one conduction electron. The density of conduction (free)
electrons is therefore

From equation (13-4-4) we obtain:

 486

We see that the conduction electrons move very slowly, it takes about
for an electron to travel . This drift speed of the electrons, should
not be confused with the speed of installation of the electric field inside the
conductor (this field actually sets the electrons in motion), which practically is
equal to the speed of light ( ).

Example 13-4-2: Find the voltage across a Copper filamentary conductor

of diameter and length , carrying a current . The
conductivity of Copper is .

Solution

The current density is

From Ohm’s law in point form, (equation (13-4-6)), we obtain:

Example 13-4-3: The current density in space is given by

( ). Find the total current flowing through:

1) The surface

2) The spherical surface of radius .

Solution

1) Application of formula (13-4-11) yields:

 487

2) The total current through the spherical surface is

since the term .

Example 13-4-4: Total current flows within a strip of width

, situated on the plane, in the direction defined by the vector
. Find the surface current density .

Solution

Fig. 13-7: Determination of the surface current density .

Assuming that the current flows uniformly through , the magnitude of

is . As a vector, the surface current density is
 488

, where is the unit vector in the direction defined by , i.e.

and .

PROBLEMS

13-4-1) Current flows in the negative axis towards the origin and
is channeled, uniformly, over the conducting plane, as shown in Fig. 13-8.
Determine the surface current density and the total current flowing in the
sector , (Ans: ).

Fig. 13-8: Determination of the surface current density .

13-4-2) The conductivity of aluminum is and the

density of the conduction electrons is . Find
the mobility of the electrons.

Hint: Use formula (13-4-6) for the conductivity .

13-4-3) A conducting spherical surface of radius is centered at the

origin. Current flowing in the positive axis towards the origin, enters
the spherical surface at the point and exits at the point
 489

. Find the surface current density on the spherical

surface.

Fig. 13-9: Surface current density on the spherical surface.

(Ans: ).

13-4-4) A certain current density is given as . How

much current flows through the square lying on the plane, whose
vertices are .

13-4-5) Consider an infinite hollow conducting cylinder whose axis coincides

with the axis. The inner radius of the cylinder is and the outer
radius is . Given that the current density is ,
, find the total current flowing in the cylinder.

(Ans: ).

13-4-6) Both the electron and hole concentrations in a pure sample of

Silicon (Si) are , (since electrons and holes are
generated by pairs). Given that the mobility of the electrons and the holes are
, find the conductivity of the
Silicon.

Hint: Apply formula (13-4-8). Notice that .

 490

13-5) The Continuity Equation

The continuity equation is one of the most fundamental equations in the

theory of Electromagnetic Fields and expresses, in mathematical form, the
principle of the conservation of charge, that is, electric charge can neither be
created nor destroyed, (see Section 2-1, properties of electric charge).

Let us consider an arbitrary, closed surface bounding a volume , as

shown in Fig. 13-10. Electric current is allowed to flow through the closed
surface , which, in principle, can be computed if the current density vector
in space is known.

Fig. 13-10: The continuity equation.

The total net current leaving the volume , through the closed surface , is
, while the total charge within the volume is ,
where is the volume charge density (in ) within . The total charge
leaving the volume per unit time, must be accounted for by a decrease in the
total charge stored within , i.e.

The negative sign in (13-5-1) is needed in order to make both sides of the
equation positive, since, assuming that net current flows through , the charge
 491

stored within decreases, and is a negative number, so that is

positive. Application of the divergence theorem converts the surface integral
to a volume integral, i.e.

and since this equation must hold true for an arbitrary volume bounded
by an arbitrary closed surface , we conclude that the integrand must be zero
identically, i.e.

This fundamental equation is known as “the continuity equation”.

For the case of a steady state, where there are no time variations of the
quantities involved, the partial derivative , and the continuity equation
becomes,

The integral form of equation (13-5-3) is

Equation (13-5-4) expresses the well known “Kirchhoff’s Current Law”, (the
algebraic sum of the currents leaving a node is zero), (see Example 13-5-3). The
other, also well known from circuits’ theory, “Kirchhoff’s Voltage Law”, is
expressed by the equation , (the algebraic sum of the voltages
around any closed loop is zero).

Example 13-5-1: The current density in a region in space is given as

. Assuming that the charge density as
, determine .
 492

Solution

From the continuity equation (13-5-2) we obtain:

This expression for results from the “partial integration” with

respect to . The function is an arbitrary function of , (notice that
). Imposing the condition as , we find ,
and finally,

Example 13-5-2: In Example 13-5-1, determine the drift velocity of the

charge and the total current crossing the surface , at the time instant
.

Solution

a) Since in general , (see equation (13-4-7)), we find

b) The total current through the surface , at time , is:

 493

Alternatively, we could have computed the total charge stored within the
sphere , and then find the current , at . Indeed:

which is identical to the answer obtained in (*).

Example 13-5-3: Show that formula (13-5-4) expresses Kirchhoff’s current

law.

Solution

Fig. 13-11: Kirchhoff’s Current Law.

Let us, for simplicity, assume that we have five filamentary tubes, emerging
from the node K.
 494

Let be the cross sections of the tubes and be

the corresponding currents, assumed to flow outwards (positive currents).
Application of formula 13-5-4 yields:

and since , formula (*) implies

i.e. the algebraic sum of the currents leaving a node is zero (Kirchhoff’s
current law).

An important remark: Kirchhoff’s current and voltage laws, strictly

speaking, apply for steady states only, i.e. for cases where there is no time
variation of the quantities involved. However, these laws are extremely good
approximations even in time varying cases, provided that the time variations
are reasonably slow. In particular, these two laws are applied safely to
household currents which alternate 50 times a second (in Europe), or 60 times
per second in the United States.

PROBLEMS

13-5-1) A current density in some region in space is given as

, (in cylindrical coordinates). Find the charge density
assuming that .

(Ans: ).

13-5-2) In Problem 13-5-1, find the drift velocity of the charge and the
total current crossing the cylindrical surface at the
time instant .
 495

Hint: See Example 13-5-2.

13-5-3) A current density exists in free

space. Find the total current through the surface of a unit sphere centered at
the origin.

(Ans: ).

13-5-4) A current density exists in free

space. Determine the total current flowing out of the unit cube
.

13-5-5) A current density exists in

free space. Calculate the total current flowing out of the cylindrical surface
, (Ans: ).

13-5-6) A current density exists in

free space. Calculate the total current in the direction, through the surface
.

13-6) Decay of Free Charges

In our so far analysis we have tacitly assumed that, in nature, there are two
types of materials, perfect dielectrics (insulators) and perfect conductors. A
charge placed inside an insulator, stays there forever, while a charge placed
inside a perfect conductor leaks out to the surface.

However, in reality, there are neither perfect insulators nor perfect

conductors. Every insulator possesses a small conductivity , which in turn
means that if an amount of charge is placed inside the insulator, sooner or
later (depending on ), the charge will leak out to the surface. This process is
known as the decay of free charges. We can explain this decay using the
continuity equation.

Assuming that is the dielectric constant and is the conductivity of a

dielectric material, we have:
 496

By virtue of the first equation, the continuity equation becomes:

and assuming that and are constants, equation (*) implies that

The number is known as the relaxation time of the material.

Notice that the relaxation time is measured in seconds. Indeed, the unit of ,
as obtained from the definition, is

The solution of (**) is easily obtained to be,

where , i.e. is the charge density at the point at the time

instant . Equation (16-3-1) shows that any charge distribution placed
inside a dielectric material, decays exponentially to zero. The rate of decay
depends on the relaxation constant , i.e. on and .

For good conductors (metals) is too small, for example for Copper,
, but for other dielectrics is considerably larger, for
example, for glass .

The time decay of charges as a function of the time , is shown in Fig. 13-12.

Notice the two extreme cases:

1) Perfect conductors, ( , and .

2) Perfect insulators, ( ), and .

 497

Fig. 13-12: Decay of free charges.

In summary: A free charge placed inside a perfect insulator, stays there

forever; a free charge placed inside a perfect conductor, almost
instantaneously, leaks out to the surface, i.e. inside the conductor. For a
more realistic model of the materials ( and finite), the charge decays
according to equation (13-6-1).

Example 13-6-1: At the charge density inside a piece of Copper is

. Given that the conductivity of copper is,
approximately, and , find the relaxation
constant and the charge density at .

Solution

The relaxation constant is

At the charge density is

13-7) Joule Heat Loss in Conducting Materials

It is a well known fact that electric currents flowing inside conductors

generate heat. This heat is known as “Joule’s heat”. Heat, as we know, is a
form of energy. This energy is supplied by the voltage or current sources which
supply current to the circuit.
 498

Let us assume that we have a current density within a conducting material

with conductivity .

Fig. 13-13: Heat loss in a conductor.

Let be the electric charge within the infinitesimal volume . If is the

density of the free electrons in the conductor, then , where is
the charge of the electron. As we have already explained, (Section 13-2), the
charge under the action of the electric force , finally attains
an average drift velocity , where is the mobility of the electrons.

The differential power delivered in the volume (and which

eventually is converted to heat) is

and since the conductivity , (see formula (13-4-6)),

The total amount of power delivered to the volume (and converted to

heat), is found by integration over , i.e.
 499

13-8) The Concept of the Resistance

a) An elementary definition of the resistance.

Let us consider a cylindrical conductor of length and cross section , as in

Fig. 13-14.

Fig. 13-14: Definition of the resistance.

An applied voltage across the conductor, results in a current flowing

through the conductor. In 1825, the German Physicist Georg Ohm, discovered
(experimentally) that the current is directly proportional to the applied
voltage , i.e. the ratio is constant. The graph of versus is shown in
Fig. 13-15.

Fig. 13-15: Ohm’s Law, .

The constant ratio , which depends only on the geometry and the
material of the conductor, is called the resistance of the conductor, i.e.
 500

The resistance is measured in Ohm ( ), where .

Equation (13-8-1) is a fundamental formula, applies well for filamentary

conductors and is the starting point in circuit analysis. However, we must point
out that there are cases where Ohm’s law breaks down, i.e. the relation
between the applied voltage and the current flowing through is not linear. We
shall consider such cases in Section 13-10.

b) A theoretical justification of Ohm’s Law.

Based on the theory we have developed so far, we are in a position to

justify Ohm’s law. Indeed, if we call the current density in the cylindrical
conductor of Fig. 13-14, then, assuming that the total current flowing in the
conductor, is uniformly distributed over its cross section , .

The voltage across the conductor is

and if we define the resistance of the conductor as , then

equation (*) takes the form (Ohm’s law).

As mentioned in section 13-4, the reciprocal of the conductivity is called

resistivity (measured in ), and in terms of the resistivity we have:

The resistivity depends on the material of the conductor, and eq. (13-8-2)
shows that the resistance in general, depends on the material and the
geometry of the conductor.

c) Heat loss in a conductor.

To compute the heat losses in the conductor of Fig. 13-14, we start with
equation (13-7-2), i.e.
 501

and since (Ohm’s Law), we finally find that the power is

Formula (13-8-3) is known as the Joule’s heating law.

d) The dependence of the resistance on the temperature.

In very general terms, the resistance of a conductor varies with the

temperature. It is found that if is the resistance of a conductor at
temperature , (usually room temperature ), and is the resistance
at temperature , then

The coefficient in (13-8-4) is known as the temperature coefficient of the

resistance and is measured in .

1) All metallic conductors have positive , meaning that their resistance

increases with temperature. At temperatures close to the absolute zero
, the resistance of some metals vanishes. This
phenomenon is known as “superconductivity” and the corresponding
materials are called “superconductors”. Superconductivity was observed for
the first time in 1911, by the Dutch Physicist Onnes, when he discovered that
the resistance of the mercury drops to zero at absolute temperature
.

2) For carbon and solutions of acids, bases and salts in water, the
temperature coefficient of the resistance is negative, and this means that
their resistance decreases as the temperature increases. This is explained
easily, since as the temperature increases the number of positive and negative
 502

pairs of ions increases, i.e. in this case, we have an increase in the number of
carriers of the electric charge.

3) Finally, there are some alloys, for example the constantan, having .
Constantan is the name for a copper-nickel alloy, ( copper, nickel).
Since , the resistance of conductors made from constantan remains
practically constant over a wide range of temperature variations.

e) The conductance .

The reciprocal of a resistance is called conductance , i.e.

The conductance is measured in . The unit is also

called Siemen ( ).

In terms of the conductance , Ohm’s law (13-8-1) may be written,

equivalently, as

f) Connection of resistors.

Conductors characterized by their resistance (measured in ) are called

resistors. Resistors may be connected in series (when the same current flows
through each one of them), or in parallel (when the same voltage applies
across them).

If resistors are connected in series, then the total resistance

is given by the formula

If resistors are connected in parallel, then the total

resistance is given by the formula
 503

In terms of the conductance, formula (13-8-8) takes the form:

Formulas (13-8-7) and (13-8-8) are known from circuit analysis.

For a proof, see Problem 13-8-1.

Example 13-8-1: An aluminum conductor is long and has a circular

cross section with radius . If there is a DC voltage of between
the ends, find: (a) the resistance , (b) the current , (c) the current density
and (d) the power dissipated, (the resistivity of the Aluminum is
).

Solution

(a) By virtue of formula (13-8-2) we find:

(b) The current is (see formula (13-8-1)),

(c) The current density is

(d) The power dissipated is (formula (13-8-3)),

 504

Example 13-8-2: What resistance must be connected in parallel to a

resistance of , given that the total resistance of the parallel connection is
?

Solution

PROBLEMS

13-8-1) Show formulas (13-8-7) and (13-8-8).

13-8-2) Three resistors and are

connected in parallel. (a) Find the total resistance , (b) If a voltage
applies across them, find the current in each resistor, (c) find the
power dissipated in each resistor.

13-8-3) The resistance of a conductor at is . When the temperature

increases from to , the resistance increases by . Determine the
temperature coefficient of the resistance , (Ans: ).

13-8-4) Two resistors when connected in series have a total resistance

, while, when the same resistors are connected in parallel have a total
resistance . Find the resistance of each resistor.

13-8-5) A filamentary conductor of length has a resistance . The

conductor is cut into two equal pieces, each one of length , which are then
connected in parallel. Find the total resistance of the parallel connection.

(Ans: ).

13-9) A More General Definition of the Resistance

(a) Let us consider two perfect conductors and , connected to a voltage

source (battery), as shown in Fig. 13-16. The conductivity of the perfect
conductors is . The space between the two conductors is filled with a
homogeneous conducting material with a finite conductivity .

A current flows through the filamentary conductor, from the positive

terminal of the battery, to the conductor , from where it diffuses throughout
 505

the region between the two conductors, according to the relation , it is

collected to the conductor , and from there it flows back to the negative
terminal of the battery, through the filamentary conductor connecting the
conductor with the negative terminal of the battery. This cycle is repeated
over and over.

Fig. 13-16: A more general definition of the resistance.

If we consider a closed surface , containing the conductor , then the

total current which leaves the conductor, is

The voltage between the two conductors is

 506

Notice that the integral in (**) does not depend on the particular path
chosen, since the electric field is irrotational.

The resistance in this case, is, (according to Ohm’s law),

assuming that is constant. We see that, if the electric field is a known

function of the spatial coordinates, then, in principle, the resistance can be
evaluated.

Let us note that, as a closed surface , we could have taken the outer
surface of the conductor .

(b) The relation between the resistance and the capacitance.

The two conductors in Fig. 13-16, form a capacitor (Chapter 11). If is the
dielectric constant of the material surrounding the two conductors, then, the
capacitance of the thus formed capacitor is given by equation (11-1-2), i.e.

From equations (13-9-1) and (13-9-2) we find that

By virtue of this formula we may determine the resistance between two

conductors, if the capacitance of the corresponding capacitor is known.

For instance, the capacitance of a cylindrical capacitor of length , inner

radius and outer radius , was found to be, (see equation (11-4-3), Fig. 11-8),
 507

According to (13-9-3), the resistance between the outer and the inner
cylindrical surfaces is

where is the finite conductivity of the material between the two

cylindrical surfaces.

Example 13-9-1: The region between the two spherical surfaces and
,( ), is filled with a conducting material with dielectric constant
and finite conductivity . Find the resistance of the conductor.

Solution

The capacitance of the corresponding capacitor is (see formula (11-3-4)),

By virtue of formula (13-9-3) we find,

Example 13-9-2: A perfectly conducting sphere of radius , is

embedded into an infinite, homogeneous material, with dielectric constant
and finite conductivity . Find the electrical resistance of the sphere.

Solution

The sought for resistance may be obtained from formula (**), in Example
13-9-1, if we let , i.e.

Example 13-9-3: For the conductor in Fig. 13-17, find the resistance ,
assuming that the conducting material has a finite conductivity .
 508

Solution

Fig. 13-17: Resistance of a conductor with variable cross section.

The resistance of the infinitesimal cylinder with base and

height , is (see formula (13-8-2)),

The total resistance , between and , is

since all the infinitesimal cylinders are connected in series.

PROBLEMS

13-9-1) (a) Using formula (**) in Example 13-9-3, find the resistance of the
truncated conducting cone of Fig. 13-18. Assume to be the conductivity of
the cone, (b) assuming that , show that the answer obtained in part
(a) reduces to the known formula for a cylindrical conductor (formula (13-8-2)).

(Ans: ).
 509

Fig. 13-18: Resistance of a truncated cone.

13-9-2) Show that the resistance of a conductor can be expressed as

where is the volume of the conductor.

Hint: Use equation (13-7-20).

13-9-3) Find the resistance of the conductor in Fig. 13-17, assuming that
, (Ans: ).

13-9-4) A conductor in the form of a truncated wedge, ,

, , has a conductivity . Within
the conductor the electric field is . Find the current density ,
the current flowing through the conductor, the voltage between the two
surfaces and and the resistance of the conductor.

13-9-5) In Problem 13-9-4, find the resistance using the expression

obtained in Problem 13-9-2, which must be identical to the value found in
Problem 13-9-4.

(Ans: ).

13-9-6) Repeat problem 13-9-4, if the electric field within the conductor is
.
 510

13-10) Breakdown of Ohm’s Law, (Vacuum Diode and p –n Diode)

Ohm’s law , derived in Section 13-8, is a fundamental law and

applies well for a great variety of conductors. However, we must point out that
there are situations where Ohm’s law does not apply. We shall consider two
such, characteristic cases, the vacuum diode and the junction (the basic
semiconductor device).

a) The vacuum diode.

Fig. 13-19: The vacuum diode.

In a vacuum diode, electrons are “boiled” off a hot cathode (thermionic

emission of electrons), which is grounded ( ), and are accelerated
towards the anode, held at a potential . During the operation of the diode,
the cloud of the thermionic electrons builds up a negative space charge in
the gap, which limits the flow of the current, (repels the approaching electrons
back). In this case, the potential and the charge density are functions of ,
i.e. and . However, when equilibrium is established, a
steady current (constant), flows from the anode towards the cathode.

If we call and the potential and the charge density, respectively,

at distance from the anode, we have (from Poisson’s equation),
 511

Also, the current density is given by the formula

where is the area of the plates, and from energy considerations,

From equations (*), (**) and (***) we find the following differential
equation for the potential , (Problem 13-10-1),

Assuming that the electric field at the cathode is zero, ( ), the

solution of the differential equation for the potential is found to be (for the
derivation see Problem 13-10-2),

This formula gives the potential as a function of the distance from the
cathode and the flowing steady current .

At , and substituting in (13-10-2) and solving for the

current in terms of the applied voltage we find:

Equation (13-10-3) is known as “the Child-Langmuir law”. Since the current

is proportional to the , the vacuum diode is a non-linear device, it does
not obey the simple Ohm’s law.

b) The diode.

Another characteristic example of a non-linear device is the diode.

 512

Fig. 13-20: Volt-amber characteristic of the diode.

It can be shown that the diode current , in terms of the applied voltage ,
is given by the formula:

The symbol , known as “the volt equivalent of temperature” is about

, at room temperature. Notice that the current varies, in
general, exponentially with the diode voltage . For a large reverse bias
( ), , and . In this case, the reverse current is practically
constant, independent of the applied reverse voltage. The current is usually
called “the reverse saturation current”.

We must point out that the great utility of the p-n diode, in practical
applications, is due to its non-linear volt-amber characteristic.

c) A general remark about non-linear devices.

Since, for a non-linear device, the voltage does not vary linearly with the
current , we cannot, in general, speak about the resistance of the device.
Instead, in such a case, it is more convenient to define “the dynamic resistance
of the device” at its operating point ( ), defined as
 513

Example 13-10-1: Assuming that the reverse saturation current of a p-n

diode is , find the diode current for an applied voltage .
What is the dynamic resistance of the diode at its operating point?

Solution

a) Application of formula (13-10-4) yields,

b) The dynamic resistance of the diode is

PROBLEMS

13-10-1) Starting with equations (*), (**) and (***), derive the differential
equation (13-10-1).

13-10-2) Show that the solution of the differential equation in (13-10-1) is

given by the expression for in equation (13-10-2), (assume that the
electric field at the cathode is zero, i.e. ).

Hint: Multiply both sides of the differential equation by and integrate.

The constants of integration are found from the boundary conditions,
and .

13-10-3) Starting with equation (13-10-3), show that the dynamic resistance
of the vacuum diode can be expressed as
 514

13-10-4) The vacuum diode of Fig. 13-19 has . Find

the diode current, given that .

Hint: Use equation (13-10-3).

13-10-5) Starting with the expression for found in Example 13-10-1,

show that for a large reverse bias ( ), , and thus conclude that the
diode, in this case, behaves as an open circuit.

13-11) Voltage Sources, Electromotive Force (emf)

a) Every electric circuit consists of “passive elements” which either

consume energy (resistors) or store energy (capacitors and inductors), and
“active elements” which produce electric energy. The energy is transferred
from the active elements to the passive elements via the “transmission lines”.

The active elements are termed “electric energy sources”. An energy

source, obviously, does not generate energy from nothing. What it does, is that
it convert some form of energy, (chemical, kinetic, nuclear, etc) to electric
energy, which is then transmitted to the passive elements, via the transmission
lines.

In every electric energy source we have two terminals, called “the

electrodes or the poles” of the source, and these are the points where the rest
of the circuit is connected to the source.

In very general terms, electric energy sources are divided into two broad
categories, the DC sources (where the voltage provided by the source is time
invariant) and the AC sources (where the source voltage is a sinusoidal
function of time). DC circuits contain only resistors, (in DC circuits capacitors
behave as open circuits while inductors behave as short circuits). AC circuits
may contain resistors and capacitors and inductors.

All over the world, large amounts of electric power are generated in AC
form, (power plants-transmission lines-consumers). However, for everyday,
practical applications, requiring small amounts of power, DC sources (batteries)
are of common use. Thus a car operates on a DC 12 V or a DC 24 V battery,
while a pocket calculator uses a DC 3 V battery.
 515

Every electric energy source presents a resistance in the flow of current

from one electrode to the other, through the interior of the source. This
resistance is called “the internal resistance of the source”. The mathematical
model of a DC voltage source consists of an ideal voltage source in series
with its internal resistance , as shown in Fig. 13-21.

Fig. 13-21: Mathematical model of a voltage source.

In Fig. 13-21 (a), no external load (resistance) is connected to the electrodes

(terminals) A and B of the source. The terminal voltage is . Since this
is an open circuit, the current is . In Fig. 13-21 (b), there is an external
load connected at the terminals A and B. In this case there is a current
flowing in the circuit. The terminal voltage , in this case, is less than , and
is given by the formula

The terminal voltage of the voltage source, with no external load

connected, is termed the “electromotive force of the source, (emf)”. The term
is misleading, since emf is not actually a force, it is a voltage under open
circuited conditions.

In very general terms, we may say that any real DC voltage source, such as
the voltage source in Fig. 13-21 (b), must contain an ideal, constant voltage
 516

source . This constant voltage source is necessary for a steady current to

flow through the circuit. The negative terminal of the ideal voltage source
( ) has a surplus of free electrons, which flow through the resistance
towards the positive terminal of the source. This constitutes the current
flowing in the circuit. When the electrons arrive at the positive pole, they are
propelled to the negative pole, and the cycle starts over again.

We may therefore, think of the source as being a pump of electric charges,

where charges are being pumped from the positive pole to the negative pole,
through the source. It is important to realize at this point, that inside the
source, the transfer of charges from the positive pole to the negative pole is
caused by forces of non electric nature. This requires an expenditure of energy
of some form (chemical, mechanical, etc) which is subsequently converted to
electric energy, supplied by the source.

b) In order to gain a deeper understanding of the operation of a voltage

source, let us consider the following simple circuit.

Fig. 13-22: Operation of a voltage source.

The circuit contains an ideal voltage source (assume internal resistance

), A being the positive terminal and B being the negative terminal of the
source. The two terminals are connected with a flexible, filamentary
 517

conductor, with a finite conductivity . If we call the electric field produced

by the charge distribution in the two poles, then, as we know, the field lines
emerge from positive charges and dive into the negative charges, as shown in
Fig. 13-22. Inside the conductor, the current density . Notice that the
field lines exist both inside and outside the source. This is an inevitable
consequence of the irrotationality of the electric field .

If we call the closed path AKLBA, then, since the electric field is
irrotational, we must have,

as expected.

Let us now imagine an electron which has started from the terminal B and is
moving in the direction , inside the conductor. As the electron
approaches the positive terminal , it is assisted, in a way, from the attraction
of the positively charged terminal , (the terminal has a deficiency of
electrons and appears to be positively charged), but what exactly happens
when the electron reaches the terminal and is just outside the positive
terminal, inside the source?

As shown in Fig. 13-22, since the charge of the electron is negative ( ),

the electric force on the electron points towards the positive terminal ,
and this means that the electron, under the action of the electric force, cannot
possibly continue its trip, inside the source, to reach the negative terminal .

In summary, in order the electron to be able to move from the positive

terminal to the negative terminal, inside the source, some other type of force,
another agent, not related to the electric field , must carry the electron from
the positive to the negative terminal. The force exerted on the electron, by the
external agent, must be equal in magnitude and of opposite direction to the
 518

electric force . In a car battery, for example, it is some kind of

chemical process that pushes the electrons from the positive towards the
negative terminal, inside the battery. In a photoelectric cell it is the incident
light that does the job.

For our subsequent analysis, let us call the force, of not electric nature,
per unit charge, which is responsible for the pushing of the electrons (against
the electric field ), from the positive terminal to the negative one. We see
that in fact, there are two forces involved in driving the current around a
circuit. The force due the external agent, (of not electric nature), confined in
the neighborhood of the battery (voltage source) and the electrostatic force
due to the field , produced by the charge distribution on the two terminals of
the source. The total force, per unit charge, exerted on the electron shall be

Ohm’s law in point form can now be generalized as,

The electromotive force of the circuit (emf), measured in Volts, is defined

as the line integral of around the circuit, i.e.

Using the expression of from (13-11-2), we find:

since , for electrostatic fields.

Using formula (13-11-3), we find another expression for the electromotive

force of the circuit, i.e.
 519

or, since the resistance of the loop is , we finally find that

Equation (13-11-6) shows that the current in a circuit is determined by the

electromotive force .

In general, the action of is confined in the region between the terminals

of the source, (see Fig. 13-22), and formula (13-11-5) reduces to the following:

Under open circuited conditions, , and from equation (13-11-3) it

follows that , and then formula (13-11-7) implies that

This expression justifies the alternative definition of the emf , given in

section (a), as the terminal voltage of the voltage source when there is
no external load connected.

Finally, another physical interpretation of the emf , which sometimes is

used as the fundamental definition of the emf, is obtained from (13-11-5),
which shows that is in fact the work done per unit charge, by the electric
source, ( ). This is so, since, by definition, is the force, of not
electric nature, per charge.

The power supplied by a voltage source in a circuit is given by the

fundamental formula

This is easily justified, since the power is

 520

13-12) Electric Circuits and Kirchhoff’s Laws

As it is known from circuit theory courses, a circuit is composed of closed

conducting paths containing resistors and sources. The sources supply energy
which is consumed by the resistors.

The model of a voltage source is the one shown in Fig. 13-21 (a). The
analysis of a circuit is based on the two Kirchhoff’s laws, obtained
experimentally by the German Physicist Gustav Kirchhoff in 1847.

The two Kirchhoff’s laws are the following:

1) Kirchhoff’s current law: The algebraic sum of currents flowing towards a

node is zero, (results from the conservation of charge), and

2) Kirchhoff’s voltage law: The algebraic sum of all the voltages around any
closed loop (including source voltages and voltage drops across resistors) is
zero, (results from the conservation of the energy in the circuit).

These two laws, originally obtained for circuits, may be obtained from
Maxwell’s equations, which are the most general equations for the
Electromagnetic fields. However, the same two laws still apply for circuits,
provided that the frequency is not too high. In particular, Kirchhoff’s laws can
be safely applied for household currents which alternate times a second (in
Europe) or times a second (in United States).

Kirchhoff’s laws together with Ohm’s law are used to analyze any circuit, i.e.
to determine all the currents and all the voltages in the circuit. For
complicated circuits, the algebra may become very tedious. However, some
very powerful and systematic methods and techniques are developed in
circuit analysis courses, (Mesh method, Nodal method, Thevenin’s and
Norton’s Theorem, etc), by means of which one may write by inspection the
circuit equations and then proceed to the solution. Most of these systematic
methods make extensive use of Matrix algebra.

Example 13-12-1: A light bulb is connected to a battery by means of two

long conducting cables, as in Fig. 13-23. From experience, we know that the
 521

current is the same all the way around the circuit. However, given that the
action of the battery is confined inside the battery, (the non electric force per
unit charge , introduced in Sect. 13-11), one should expect the current to be
large close to the battery and diminish as we move away from the battery. But
this does not happen. The current remains the same all around the circuit. Why
does this happen? Give a physical explanation.

Solution

Fig. 13-23: Why the current remains the same along the conductor.

Let us assume that, right after the connection of the lamp, the current is not
the same around the closed loop (battery-conductors-lamp). Then, the charge
must be accumulated somewhere. Let us, for definiteness, assume that the
current entering at the point is greater than the current leaving at the point
. Since , there will be an accumulation of positive charge , in the
region between and , as shown in Fig. 13-23. In this case, (net positive
charge in the region), an electric field emerges, in the direction shown in the
Figure. This electric field slows down the incoming current and promotes the
outgoing current, until the charge unbalance vanishes, equilibrium is
established, and the incoming current equals the outgoing current ( ).
So, any charge unbalance is automatically self adjusted, by the generated
electric field caused by this unbalance. This induced electric field (caused by
any possible charge unbalance) is always of such a direction as to smooth out
 522

the current flow. In practice, this process is so fast, almost instantaneous, so

that we do not notice any difference in the current flowing in the circuit.

Example 13-12-2: In the circuit of Fig. 13-24, find the currents and
and the power supplied by the voltage source .

Given: .

Solution

Fig. 13-24: A two loops circuit.

We use (as we usually do in circuits) the symbol for the emf of the first
voltage source and for the emf of the second voltage source.

Application of KCL at the node A, yields:

Application of KVL in the two loops of the circuit yields:

Equations (*), (**) and (***) constitute a system, three equations in the
three unknowns, and . Solving the system we find:
 523

The negative sign in , simply means that this current actually flows in the
opposite direction from the one shown in the figure.

Example 13-12-3: (The maximum power transfer Theorem).

In Fig. 13-25, a variable load (resistance) is connected at the terminals of

a voltage source of emf and internal resistance . For what value of the
variable resistance , the power dissipated on is the maximum possible?
What is this maximum power?

Fig. 13-25: The maximum power transfer theorem.

Solution

From KVL in the loop of the circuit we find that

The power dissipated on is:

Obviously, the power dissipated on is a function of . To find for what

value of the power becomes maximum, we set the derivative .
Simple calculations show that the power becomes maximum when
(internal resistance of the source), and this maximum power is
 524

PROBLEMS

13-12-1) In Example 13-12-2 find the powers supplied and absorbed by each
circuit element, and verify that the total power supplied equals the total power
absorbed.

13-12-2) In Example 13-12-3, assume that and .

For what we have maximum power dissipation on ? What is ?

13-12-3) Total current flows towards the parallel combination of two

resistors and . Find the currents and in and respectively.

(Ans: ).

13-12-4) A long, uniform, two conductors line, of length , has a distributed

resistance per unit length and a distributed leak conductance per unit
length. Assuming that at a distance from the beginning of the line the voltage
is and the current is , show that and .
Then show that the voltage satisfies the differential equation
and the current satisfies .

Hint:

Fig. 13-26: A long transmission line.

Apply Kirchhoff’s laws in the infinitesimal segment .

 525

13-12-5) For the uniform line of Fig. 13-26, find the voltage and the current
at , assuming that and .

(Ans:
, where and ).

Hint: The solution of the differential equation is

, . The current relates to , since
. The arbitrary constants are determined from the given
conditions and .

13-12-6) In Problem 13-12-5, assume that:

. Find and .

13-13) Steady Currents and Laplace’s Equation

a) In part A, Electrostatics was defined as the study of the electric fields

generated by charges which are at rest with respect to a stationary observer.
However, we must mention that Electrostatics is more than that. Actually,
Electrostatics applies whenever the charge distributions, producing the
electric fields, are independent of time. The charge distributions could be, as
we know, either discrete or continuous. Similarly, as we shall show in the
following chapters, Magnetostatics, (the study of magnetic fields produced by
current distributions), applies whenever the current distributions are
independent of time. For example, a rotating conducting sphere of radius
with a uniform surface charge density , produces the same electric field as
if it were at rest, ( ), where .

However, there is one exception to this most general definition of

Electrostatics, that is, “The electric field inside a conductor is zero when the
charges are indeed stationary”, (see Section 9-2, basic properties of
conductors). On the contrary, when we have steady currents flowing inside
conductors, the electric field inside the conductor need not be zero, and as a
matter of fact, cannot be zero, since, if it were zero, then we would not have
any current in the conductor. Recall that the current density , and if
 526

were zero, the current density would be zero and then, there would be no
current flowing in the conductor.

The electric field inside a conductor, in steady state, is an electrostatic field,

and as such satisfies the fundamental law , (irrationality of the
electrostatic field). This shows that the electric field may be derived as the
gradient of a scalar function , called the potential of the field, i.e. we
may set . The further development of the theory is similar to that
presented in Chapter 5.

b) We may thus summarize and write down the differential equations for
the electric field for the flow of steady currents:

Combining these equations we find:

and assuming that the conductivity is constant, we get,

We thus see that in the steady state (currents not depending on time), the
potential satisfies Laplace’s equation. Once the potential is determined,
then the electric field and the current density are obtained easily from
equations (13-13-1) and (13-13-2).

In addition, we have the following boundary conditions, which are

necessary in order to have a unique solution of Laplace’s equation:

1) The potential function is a continuous function of the spatial

coordinates,
 527

2) At the boundary between a conductor and free space (or perfect

insulator), there is no normal flow of current, i.e. the normal component of
the current density must be zero, i.e.

3) The potential must remain finite at all points in space.

All methods and techniques developed in Chapter 12, for solving Laplace’s
equation in Cartesian, Cylindrical and Spherical coordinates apply equally well
for steady flow problems. Let us consider a few illustrative examples.

Example 13-13-1: Let us consider a very small electrode made of Copper

(very good conductor) placed at the origin, within an infinite, homogeneous
conducting medium, with conductivity . A current flows to the electrode,
through a thin, insulated filamentary conductor, which is then diffused to the
conducting medium. (This diffused current is collected by the surface of a
sphere, “at infinite distance” from the electrode, and then, from there, it
returns back to its source). Determine the potential , the electric field and
the current density .

Solution

Fig. 13-27: The field of a point current source.

 528

Since the dimensions of the electrode are very small, we may treat it as “a
point current source”, supplying current to the surrounding conducting
medium. Due the spherical symmetry involved, the potential depends only on
, the distance from the origin, where the electrode is located. In this case,
Laplace’s equation becomes:

Assuming that , it follows that , and then, .

To determine the constant , we think as follows:

The total current diffuses with a current density , or, as a

vector, , and since , it follows that

However, , from which, .

We have thus found:

This expression for the potential is with respect to the infinity, ( ).

Remark: Assume that “a differential current element” is diffused, at

some point , within an infinite, homogeneous conducting medium, with
conductivity . Then, the expressions for the potential, the electric field and
the current density at a point whose vector distance from the source is ,
( ), are (in accordance to the expressions in (***)):

If the current distribution on the surface of an electrode is known, then, in

principle, we may compute the potential, the electric field and the current
 529

density, everywhere inside the conducting medium, by integrating the

expressions in (****) over the electrode’s surface (superposition principle).

Recall that we have used the superposition principle to calculate the

electrostatic potential, due to discrete or continuous charge distributions,
(equation (5-5-4)).

Example 13-13-2: A perfectly conducting hemisphere of radius (electrode)

is located within an infinite, homogeneous conducting medium, with constant
conductivity , as shown in Fig. 13-28. A current driven to the hemisphere,
diffuses uniformly in the conducting medium. Find the potential , the electric
field and the current density .

Solution

Fig. 13-28: The field of a conducting hemisphere.

The current density within the conducting medium is , (and is

zero in air). Solving Laplace’s equation in the region of the conducting
medium, (the procedure is identical to the procedure in Example 13-13-1), we
find:

Example 13-13-3: a) In the preceding example, find the potential of the

electrode with respect to the infinity, and then compute its resistance ,
 530

(known as “the grounding resistance of the electrode”), b) Assuming that the

conducting medium is ground with conductivity , compute
the grounding resistance of the electrode if .

Solution

a) If is the potential of the hemispherical electrode with respect to the

infinity, we know that

The grounding resistance of the electrode is

b) For , we find .

Example 13-13-4: The plane , i.e. the plane, separates two infinite,
conducting, homogeneous media with conductivities and . A current
driven to the origin, through a thin, insulated filamentary conductor, diffuses
uniformly in the two conducting media. Find the potential and the electric field
everywhere.

Solution

Fig. 13-29: The field in two adjacent conducting media.

 531

Let us call and the potential in regions 1 and 2, respectively.

Due to the spherical symmetry involved, in each region, and depend only
on the radial distance , and as it was derived in Example 13-13-1, the
potentials in regions 1 and 2 will be given by the expressions:

where are constants to be determined from the appropriate

boundary conditions:

1) The potential must approach zero as . This condition leads to the

determination of the constants and . Using the expressions for the
potential in (*), we find and , and therefore,

2) The tangential component of the electric field must be continuous on

the boundary between the two media, i.e. , and since on the
boundary we have, and , we find,
(using equation (**)),

and the expressions for the potential in equation (**) become,

where the constant is still to be determined. This shows that the potential
has the same expression in both regions, . As a result of this, the
electric field in both regions has the same expression,

3) The current density in region 1 is and in region 2 is

, or taking into account equation (****),
 532

On the spherical surface of radius (see Fig. 13-29), we have:

and from equation (***) we obtain:

The electric field is the negative gradient of the potential, i.e.

PROBLEMS

13-13-1) A cylindrical beam of positively charged particles flowing in the

positive direction, with speed , carries a total current . If the beam radius is
, find the electric field everywhere, (inside and outside the beam).

(Ans: ).

Hint: Since the charge density is time independent, the configuration is

electrostatic, and due to the cylindrical symmetry involved, . The
current density is , and the charge density is . To find
apply Gauss’s law in integral form, over cylindrical surfaces, coaxial to the
axis, etc.

13-13-2) In Example 13-13-4, assume that and

. Find the voltage between the points
and .

Hint: , where is given in formula in

Example 13-13-4.
 533

13-13-3) In Fig. 13-28 find the potential difference between two points
and , on the boundary between air and conducting medium, if the distance of
from the center of the hemispherical conductor is and the distance of
from is , ( ).

(Ans: , where is the potential of the

hemispherical electrode with respect to the infinity, (equation (*) in Example
13-13-2).

13-13-4) A high voltage line of potential , comes into contact with

a hemispherical electrode of radius , immersed into ground with
conductivity . A person is walking at a distance
from the electrode. Assuming that the distance between its feet is ,
find the voltage applied to the walker, (step voltage). What would the step
voltage be if ?

Fig. 13-30: Computation of the step voltage.

(Ans: ).

SUPPLEMENTARY PROBLEMS

13-1) What is the average current of a lightning stroke, provided that it

delivers at a potential of . Assume that the discharge duration is
about , (Ans: ).

13-2) Show that at the boundary between two conductors with

conductivities and we have:
 534

( stands for the tangential component and for the normal).

Hint: On the boundary: , .

13-3) Show that at a conductor-conductor boundary, the normal

component of the current density is continuous, i.e. .

13-4) Show that the surface current density between two conducting
media with permittivities and and conductivities and respectively, is
given by the formula

where is the normal component of the current density .

Hint: , since
.

13-5) For a certain Silicon sample it is given:

, , and
. Find the conductivity and the resistivity of the sample.

(Ans: ).

13-6) In Example 13-9-4, having already computed the resistance , find the
capacitance of the truncated wedge, a) Using the formula (formula
(13-9-3)) and b) Verify your answer using the formula for the capacitance
obtained in Prob. 11-4-6.

13-7) A current density in free space is given as

a) Find the everywhere, and b) Find the total current in the direction.
 535

(Ans: ).

13-8) A truncated cone is concentric with the axis and has a conductivity
. The two faces have radii and and
are a distance apart. Find the resistance between the two end faces.

Hint: See Problem 13-9-1.

13-9) For the planar vacuum diode in Fig. 13-19 it is given:

and . Find the diode current , (Ans: ).

Hint: Use formula (13-10-3).

13-10) For the planar vacuum diode of Problem 13-9, find the potential and
the electric field at .

13-11) For the long transmission line of Problem 13-12-4, find and
if it is given that and .

(Ans: , where ).

Hint: See Problem 13-12-5.

13-12) The reverse saturation current of a p-n is . Find the current

for applied voltages: . For the indicated values
of the voltage, find the dynamic resistance .

Hint: Use formulas (13-10-4) and (13-10-5).

13-13) In pure Silicon , the electron and hole mobilities at room

temperature are and .
Calculate what fraction of the total conductivity is due to electrons. Repeat for
Germanium with conductivities and
.

(Ans: For Silicon , for Germanium ).

13-14) The potential inside a unit cube with one corner at the origin and
another corner at the point is given by .
a) Find the electric field and the charge density inside the cube, b)
 536

Provided that the charge density has a velocity of ,

find the total current crossing the side of the cube at and at .

13-15) A surface current density flows on the plane.

Find the total current crossing the segment .

Fig. 13-31: Current through the segment .

(Ans: ).

Hint: , ).

13-16) The current density in space is given by .

a) Find the charge density , b) The charge within the region at
, c) The current crossing the area at and d) The
velocity of the charge density.

Hint: See Examples 13-5-1 and 13-5-2.

13-17) The field of a linear current source: A filamentary conductor (with

negligible cross section) of length is embedded into an infinite,
homogeneous conducting medium of conductivity . A total current diffuses
(through the linear conductor) to the conducting medium. Assuming that the
conductor is placed along the axis, from to , find the potential
everywhere inside the conducting medium, (see Fig. 13-32).

(Ans: ).
 537

Hint: A differential element of the conductor, of length , can be treated

as a point current source, diffusing current into the
conducting medium. The potential of this point current source , was found in
Example 13-13-1 to be:

and the potential at the point is found by integration over the

current source, i.e. , (see Remark in Example 13-13-1).

13-18) Far away from the linear current source, show that the potential
reduces to the potential of a point current source, diffusing total current in
the medium, (see Example 13-13-1).

Fig. 13-32: The potential of a linear current source.

13-19) Let us consider an electrode at a position , at depth below the

surface of the ground of conductivity , as shown in Fig. 13-33. Through a thin,
well insulated filamentary conductor, total current is driven to the electrode,
from where it diffuses into the ground. Find the potential everywhere within
the ground, (the space over ground (air) is considered to be a perfect insulator,
i.e. on air is zero).

(Ans: ).
 538

Hint: The field inside the original configuration (conducting ground-

insulating air) may be determined using the method of images. We assume
that the conducting ground extends to infinity (covers the whole space). The
original electrode is still placed at while the image electrode, supplying
current to the conducting medium, is placed at , the image of with
respect to the boundary surface between ground-air.

Fig. 13-33: The potential of a grounded electrode.

Notice that the expression for the potential is valid

inside the conducting ground. This expression satisfies Laplace’s equation,
approaches zero far away from the electrode and on the boundary ( )
yields zero normal component of the electric field and current density.

13-20) In Problem 13-19, assume that the electrode at is a spherical

conductor with radius . Using the expression for the potential found in the
preceding problem, with and , show that the potential of the
electrode is given, (approximately), by the expression

Next, show that in terms of (the potential of the electrode), the potential
at the point can be written as
 539

13-21) In Problem 13-20, show that the “grounding resistance” of the

electrode is .

13-22) a) In Problem 13-20, if is the distance of a point from the point

, (on the boundary surface between the conducting medium and air, see Fig.
13-33), show that the electric field at is given by the expression

b) Show that the maximum magnitude of the electric field occurs at a

distance , and this maximum value of the electric field is

c) Assuming , and , find the maximum

step voltage a walker will experience, provided that its two feet are at a
distance apart.

Hint: At the point in Fig. 13-33, . The potential at is

(from Problems 13-19 and 13-13-20), and the electric field is

, etc.

13-23) Consider a uniform electric field existing within a

homogeneous conducting medium of infinite extend, with conductivity . A
conducting sphere of radius and conductivity is placed in the electric field.
Find the potential inside and outside the conducting sphere.

(Ans: , ).
 540

Hint: We seek a solution of Laplace’s equation satisfying the following

boundary conditions, (see also Example 12-5-1): The potential function must
remain bounded everywhere, far away from the conducting sphere the electric
field should be unaffected by the presence of the sphere, and should therefore
be , at the potential must be continuous and at the
normal component of the current density must be continuous.

Fig. 13-34: Conducting sphere in a uniform electric field.

13-24) In Problem 13-23 find the current density everywhere and show
that inside the sphere the field is uniform and equal to .

13-25) In Problem 13-24, show that if , then and

.
 541

CHAPTER 14: THE STATIC MAGNETIC FIELD IN FREE SPACE

14-1) Introduction

The ancient Greek mathematician and philosopher Thales of Miletus (640-

546 BC) was probably aware of the action of the “lodestone”, which is a
natural magnetic substance, found in a place called “Magnesia”. He discovered
that the lodestone has the curious property of attracting small pieces of iron,
and since this substance was found in Magnesia, Thales named it “magnet”.

For many centuries, the exact nature of these attractive forces (now known
as magnetic forces) was not known, and at any rate, magnetic phenomena
were considered to be completely independent from electric phenomena.

However, in 1819, an event occurred, which showed, (for the first time),
that electric and magnetic phenomena, must, somehow, be related. During
the course of an experiment, the Danish Physicist Hans Christian Oersted
discovered that electric currents exert magnetic forces on a magnetic compass
needle (deflect the magnetic compass); but, electric current is nothing else but
electric charge in motion. Oersted’s experiment demonstrates, very clearly,
that electric charges in motion produce magnetic fields (since they deflect a
magnetic compass).

A few years later, the American Physicist Henry Rowland performed

another experiment; he set a charged disk in fast rotating motion and, again,
detected a magnetic field which deflected a nearby magnetic compass.
Rowland’s experiment showed that a moving charge in free space (convection
current) produces magnetic fields in the same way as a conduction current
flowing in a filamentary wire.

Today it is known that magnetic fields are produced by various

distributions of electric currents, i.e. by moving electric charges. “Magnetic
charges” similar to electric charges, (which produce electric fields), do not
exist in nature. This fact, when expressed in mathematical form, constitutes
one of the four Maxwell’s equations.

The only “real substance” which produces electric and magnetic fields, is
the electric charge. Stationary charges produce only electric fields in the
 542

space around them, while charges in motion (i.e. electric currents) produce,
in addition, magnetic fields in the surrounding space.

As a manifestation of the simplicity and the symmetry of nature, we shall

show a little later that, a magnet in motion can generate electric currents,
(Faraday’s law of induction).

We shall also show that the macroscopic magnetic properties of matter

can be deduced if we assume that the fundamental constituents of matter, i.e.
the atoms (negatively charged electrons revolving around the heavy, positively
charged nuclei) behave as tiny magnets. Depending on the orientation of these
tiny magnets, it is possible, a body to exhibit macroscopic magnetic properties,
i.e. to behave as a permanent magnet.

In this chapter we shall develop methods and techniques to calculate the

magnetic fields produced by steady electric currents, i.e. by currents that do
not depend on time, though they may depend on the spatial coordinates, i.e.
. Steady currents produce static magnetic fields, i.e. magnetic fields
not depending on time. These fields are known as “static magnetic fields”. The
study of static magnetic fields is called “Magnetostatics”.

The method of approach is similar to the method followed in the

development of Electrostatics. We will find, first, the magnetic field produced
by a differential current element, and then we shall integrate over the volume
of the source. The starting point in Electrostatics was Coulomb’s Law. The
starting point in Magnetostatics is the so called Biot-Savart Law, to be
developed in the next Section.

Initially we shall determine the magnetic field produced by given, steady

current distributions, in free space (vacuum), practically in air. Then we shall
investigate what happens when a piece of a material is placed in the, already
existing, magnetic field in free space. Recall the analogous development of
Electrostatics. We first determined the electric field in free space produced by
a given charge distribution, and then we investigated what happens when a
material body is embedded in the electric field (polarization of the dielectric).
We shall follow a similar approach for the magnetic fields.
 543

14-2) Biot-Savart Law

Magnetic field is, in general, a region in space, at every point of which,

there are forces exerted on permanent magnets or on moving electric
charges or even on current carrying conductors.

a) Let us consider a point charge , in free space, moving with a constant

velocity (not accelerated motion), with respect to a given reference system.

Fig. 14-1: Electric and magnetic field of a moving charge, ( constant).

The moving charge , at the point whose vector distance from is ,

produces an electric field , (just because of its presence), and, at the same
time, produces a magnetic field , because of its motion.

It can be shown that the electric and the magnetic fields are given by the
expressions:

where is the speed of light in free

space.
 544

In formula (14-2-1), is a universal constant, called “the permeability of

free space” and its numerical value in the International System of Units is

In a next chapter we shall show that the unit is equivalent to ,

where is the unit for the inductance. Notice that is actually
.

The magnetic field , which is the fundamental vector in Magnetostatics, is

measured in , as obtained easily from the defining equation in
formula (14-2-1). This unit is given the official name “Tesla”, i.e.

In summary, the magnetic field is measured in Tesla. Notice that, in

particular for the magnetic field , its unit in the system CGS is , (in
honor of the great German Mathematician C. F. Gauss). The Gauss is,
sometimes used, instead of the Tesla. It suffices to know that
. The earth’s magnetic field is approximately
. In a laboratory, we may produce a strong magnetic field of the
order of a few thousands of Gauss.

b) Assuming that the speed of the charge is much smaller than the speed
of light, i.e. assuming that , (non relativistic speeds), the two
equations in (14-2-1) take the form:

The first formula in (14-2-4) gives the electric field at a distance from
the point charge , and this formula, recall, was the starting point in the
development of Electrostatics, in Chapter 3.
 545

The second formula in (14-2-4) will, likewise, be the starting point for the
development of Magnetostatics. This formula is known as the Biot-Savart law
for a moving point charge.

c) Recall that, in Electrostatics, we have considered point charges, line

charge distributions, surface charge distributions and volume charge
distributions. In Magnetostatics, we shall, likewise, consider point charges in
motion, line currents, (for example currents in filamentary conductors),
surface currents and volume currents.

The “connecting point” between these current distributions is eq. (13-4-14),

which is repeated here for convenience:

These equivalencies shall be used repeatedly in the sequel, in order to

determine the magnetic field , produced by various current distributions.

14-3) Magnetic Field Produced by Various Current Distributions

a) The magnetic field of moving point charge.

The magnetic field produced is given by equation (14-2-4).

b) The magnetic field of a filamentary conductor.

Let us consider a filamentary conductor, i.e. a conductor whose diameter of

its cross section is much smaller as compared to its length.

We further assume that a current flows in the filamentary conductor, as

shown in Fig. 14-2. The current is actually due to the motion of electrons,
inside the conductor. But, as we already know, charges in motion produce a
magnetic field in the space around the conductor. In summary, in the space
around the conductor, we expect to have a magnetic field, (Oersted’s
experiment).

The problem we have to solve, is then to determine the magnetic field , at

any point in space, produced by the current carrying filamentary conductor.
 546

Fig. 14-2: Magnetic field of a filamentary conductor.

Let us assume that the charge within the infinitesimal cylinder, in Fig. 14-2,
is . This charge is moving with velocity , and according to formula (14-2-4),
the magnetic field at the point , whose vector distance from is , is

and since , equation (*) can be written, equivalently, as

This formula is the classic form of Biot-Savart’s law, as applied to a

filamentary conductor, carrying a current . It gives an expression for the
magnetic field produced at a point in space, in terms of the “differential
source” and the vector distance between the source and the field
point. Recall that, as it was the case in Electrostatics, the vector always
points from the “source element” towards the “field point”, i.e. the point
where the magnetic field is to be calculated.

Due to the cross product appearing in (14-3-1), the computation

of magnetic fields is, in general, by far, more difficult as compared to the
computation of electric fields generated by charge distributions.
 547

The magnetic field produced by a filamentary conductor, of a finite length

, is found by superposition, i.e. by integration of (14-3-1) over the length of
the conductor, i.e.

c) The magnetic field of a surface current density.

Let us assume that we have a current flowing over a surface , with a

known surface current density .

Fig. 14-3: Magnetic field of a surface current density.

Since , the differential magnetic field at the point ,

produced by the current element is obtained from formula (14-3-1), if
is replaced by , i.e.

The total magnetic field at the point , produced by a finite surface , is

found, similarly, by integration of (14-3-3) over the whole surface, i.e.
 548

d) The magnetic field of a volume current density.

Let us assume that within a volume there is a current distribution, with a

known current density .

Fig. 14-4: Magnetic field of a volume current density.

Since , the differential magnetic field at the point ,

produced by the current element is obtained from formula (14-3-3), if
is replaced by , i.e.

The total magnetic field at the point , produced by a finite volume is

found by integration of (14-3-5) over the whole volume, i.e.

e) The general problem in Magnetostatics.

The most general problem in Magnetostatics, is to determine the magnetic

field everywhere in space, produced by given current distributions. Using the
Biot-Savart law and the principle of superposition, the problem is, at least in
 549

principle, solved. However, in most cases, in practice, the integrals we have

developed in this chapter cannot be evaluated for arbitrary shapes of
filamentary conductors, surfaces or volumes. In such cases we have to resort to
suitable numerical techniques for the computation of the corresponding
integrals.

14-4) Calculation of the Magnetic Field of Some Important Current

Distributions

In this section we shall use the formulas developed in the preceding section
in order to compute the magnetic field produced by some important current
distributions.

a) The magnetic field of an infinitely long filamentary conductor.

The magnetic field produced by an infinitely long, straight, filamentary

conductor, carrying a steady current , is given by the formula

Fig. 14-5: The magnetic field of an infinitely long, straight filament.

 550

Let us, for definiteness, consider an infinitely long, straight, filamentary

conductor, coincident with the axis, carrying a current , as in Fig. 14-5.

We want to determine the magnetic field at a point , whose distance from

the filament is . The magnetic field at the point , is given by eq. (14-3-2), i.e.

Since and , the cross product

(recall that and , in cylindrical coordinates). At the

point , where we want to determine the magnetic field, and remain
constant with respect to the integration and therefore equation (*) yields:

The value of the integral in (***) is , (see Problem 14-4-1), and finally
we obtain the expression for the magnetic field as given in formula (14-4-1).

It follows that the field depends only on the radial distance of the point
from the filamentary conductor. If, for example, consider the plane ,
( ), then we see that:

1) The magnitude remains constant over the circumference of a

circle of constant radius , and

2) The direction of , at any particular point, is in the direction of the local

unit vector at this point.

So, on the circumference , (Fig. 14-6) all the vectors have the same
magnitude , however, their direction changes and is in the
direction of the local vector, at the particular point, (see, for example, the
fields ).
 551

On the circumference , all the vectors have the same magnitude

, smaller than the magnitude of on the circumference
.

In general, as we move away from the conductor, the magnetic field falls off
as .

Fig. 14-6: Variation of the field as a function of the radial distance .

Even though formula (14-4-1) is relatively simple, in practice, one may

encounter certain difficulties in its application, when the geometry deviates
from the geometry of Fig. 14-5. For example, it is not so simple to find the
magnetic field at the point produced by a filamentary contactor,
lying in the plane, passing through the origin and forming an angle with
the axis. Let the reader try to work this problem, using formula (14-4-1), in
order to realize the difficulty involved.

In such cases, we may use another formula, (obtained from (14-4-1)), which
is more convenient for the determination of the magnetic field.
 552

Assuming that the filamentary conductor passes through the origin, let
be the unit vector in the direction of the current , and be
the vector radius of a point where the magnetic field is to be determined,
(see Fig. 14-7).

Fig. 14-7: An alternative expression for the field of an infinite filament.

The magnetic field at is given by the formula:

For a proof see Problem 14-4-2.

Some important remarks:

1) The direction of magnetic field at a point in space, depends on the

direction of the current . For instance, if in the conductor of Fig. 14-5, the
current flows in the opposite direction, i.e. from towards , then the
magnitude of remains unaffected, but its direction is reversed, i.e. in this
case, .

2) Several times, it is desirable to have a physical intuition about the

direction of the magnetic field produced by a filamentary conductor, without
actually evaluating the field. The direction is given by the “right hand rule”.

If we grab the wire with our right hand, the thumb in the direction of the
current, then the figures curl around in the direction of the magnetic field.
 553

Fig. 14-8: The right hand rule.

3) For the filamentary conductor, we notice that the magnetic field lines
are closed. This fact has a deep physical meaning. Recall that in static electric
fields, the field lines are open, emerging from positive charges and diving in
negative charges, ( , the electric charge density). Even though we shall
examine this issue in depth, in the following chapter, it suffices at this point to
mention that the magnetic field lines are closed, since in nature there are no
“magnetic charges”. This important fact, no existence of magnetic charges,
when expressed in mathematical terms, leads to one of the four Maxwell’s
equations.

While electric charges do exist in nature, and are the sources of the
electric fields, magnetic charges do not exist. The magnetic fields are
produced by electric currents, i.e. by charges in motion.

This general remark holds true for static fields, i.e. fields where all the
characteristic quantities involved ( , etc) do not vary with time, but
they may vary as functions of the spatial coordinates.

However, as we shall show in the sequel, in the time varying

electromagnetic fields, a time varying electric field produces a magnetic field
with closed lines and a time varying magnetic field produces an electric field
with closed lines.
 554

b) The magnetic field of a uniform, surface current density.

A surface current density exists on the plane. Determine the

magnetic field at a point in space.

The solution of the problem is achieved by a suitable application of formula

(14-3-4), i.e.

Fig. 14-9: Magnetic field of a uniform surface current density.

Since , the cross product

and substituting in (*) we obtain:

 555

The integrals in (**) seem, and indeed are, difficult to be evaluated. The
computations are facilitated if we pass to polar coordinates, i.e. if we set:

The differential area in polar coordinates is . By means of this

transformation, we find that the first integral is zero, while the second integral
is . For the evaluation of the integrals see Problem 14-3. We thus find
that the magnetic field at , produced by the surface current
density , over the whole plane, is given by the formula

If we call a unit vector, perpendicular to the plane where the surface

current density exists, and pointing in the region where the magnetic field is to
be evaluated, (in Fig. 14-9, if the magnetic field is to be evaluated in the
upper half region, or if the field is to be evaluated in the lower half
region), then formula (14-4-3) may be expressed in the following “coordinate
free form”

As it follows from (14-4-3) the magnetic field in the upper half region
( ) has constant magnitude and a constant direction (the
direction of ), while in the lower half region ( ) the field has constant
magnitude and a constant direction (the direction of ).

Magnetic field for which the vector remains constant in space, are
called “uniform magnetic fields”. In Fig. 14-9, we have a uniform field in the
upper half region and another uniform field in the lower half region.

Recall that the electric field produced by an infinite, planar, uniform charge
density is uniform, (see formula (3-3-4)).
 556

c) The magnetic field of a current carrying circular loop.

Let us consider a circular loop of radius , which carries a steady current .

The loop is situated on the plane, with its center at the origin. Find the
magnetic field at the point .

Fig. 14-10: Magnetic field of a circular current.

Application of formula (14-3-2) yields:

The differential length of the loop is , while

, and . The cross product in the
numerator of formula (*) is

since and . In view of formula (**), equation (*)

yields:
 557

The term C in (***) is zero. This is due to the fact that is not constant, on
the contrary, varies as a function of the position of on the loop. If we
consider the integral as the sum of infinitesimal quantities, then, we see that
the term corresponding to some , is cancelled by the similar term
at , since at these two points the unit vectors are opposite. In
summary, we have pair wise cancellation, resulting in a zero value of the
integral. Another method of approach, would be to express in terms of and
, (formula (1-3-3), which are constant unit vectors, and then integrate. Let
the reader try it.

In the term A, in (***), the unit vector remains constant in the integration,
and it can be taken out of the integral, the result being,

At the center of the loop, the magnetic field is obtained from (14-4-5) for
, i.e.

Remark 1: If we consider a coil consisting of tightly wound turns, each of

radius , then the magnetic field at is found from (14-4-5) if is
replaced by , i.e.
 558

Remark 2: If more than one current distribution exists in a region in space,

for instance an infinite, current carrying filamentary conductor and a current
carrying circular loop, then we may apply the superposition principle to
determine the magnetic field, at any point in the region.

Example 14-4-1: An infinite, straight, filamentary conductor, lies on the

plane and forms an angle of with the axis. Assuming that , find
the magnetic field at the point .

Solution

Fig. 14-11: Magnetic field of an infinite, straight filament.

We shall find the magnetic field at , using formula (14-4-2), i.e.

The vector , while the unit vector ,

i.e. . The cross product is

and . The magnetic field at , as it is found

from formula (*), is
 559

Note that, should we use formula (14-4-1) instead of (14-4-2), the

calculations would be, by far, more complicated. Let the reader try it!

Example 14-4-2: Two planes ( ) and ( ) are perpendicular to the axis, at

and , respectively. A surface current exists on ( )
and a surface current exists on ( ), (assume ). Find the
magnetic field everywhere.

Solution

Fig. 14-12: Magnetic field produced by two current sheets.

The two planes divide the whole space in three regions: Region A, ,
region B, , (the region between the planes), and region C: .

The magnetic field in space, produced by only, is (according to (14-4-4)):

 560

Similarly, the magnetic field in space, produced by only, is

The total magnetic field , produced when both surface currents and
exist simultaneously in space, is found using the principle of the superposition,
i.e. , i.e.

In Region A, :

In Region B,

In Region C,

We note that in the region between the two planes the magnetic field is
uniform, while everywhere else is zero.

PROBLEMS

14-4-1) Using the substitution , evaluate the integral

, (Ans: ).

14-4-2) Starting with equation (14-4-1), show formula (14-4-2).

Hint: In Fig. 14-7, .

14-4-3) Using polar coordinates show:

 561

14-4-4) A straight filamentary conductor, of infinite extend, coinciding with

the axis, carries a current , in the positive direction. Find the
magnetic field at the point .

Hint: See Example 14-4-1.

14-4-5) A square coil of side , lying in the plane, carries a current in

the counterclockwise direction. Find the magnetic field at the center of the
square, (Ans: ).

14-4-6) Work Example 14-4-2, assuming that and .

14-4-7) A surface current density flows in the plane,

in the region bounded by . Find the magnetic
field at the origin, (Ans: ).

14-4-8) A filamentary current flows from infinity towards the

origin on the negative semi axis and then back to infinity along the positive
semi axis. Find the magnetic field at the point .

14-4-9) Find the differential field at produced by a source

element , placed at .

(Ans: ).

14-4-10) In Fig. 14-13, find the magnetic field at the point K, if .

Fig. 14-13: Magnetic field at K, for the conductor shown.

14-4-11) A infinite filamentary conductor, coinciding with the axis, carries

a current . The conductor is placed within a uniform magnetic field
 562

. Find the total magnetic field at the point

.

(Ans: ).

14-5) Magnetic Field Lines and Magnetic Flux

a) Magnetic field lines.

As already explained, any steady current distribution (currents do not vary

with time), produces a static magnetic field in space, (determined, in
principle, from the Biot-Savart law and the principle of superposition). As
usually, the symbolization stands for the magnetic field at
the point in space. Just as electric field lines are convenient
for picturing electrostatic fields, magnetic field lines are convenient for
picturing magnetostatic fields. A vector line (or field line) of a magnetic field
, is a curve in space, at every point of which the direction of the tangent of
the curve coincides with the direction of the magnetic field . The positive
direction of a field line is defined to be the direction which coincides with the
direction of the field.

If we call a differential vector, tangent to a field line passing through a

point , (Fig. 14-14), the differential vector equation of the field line is

For expressions of in cylindrical and spherical coordinates, see equation

(3-4-2). For example, in cylindrical coordinates,

and if , then implies, (let the reader

verify it),
 563

This is the equation of magnetic field lines in cylindrical coordinates, and

similarly, one may obtain the equation of the field lines in spherical
coordinates.

Fig. 14-14: Definition of the magnetic field lines.

In general, the components are functions of the spatial

coordinates. Equation (14-5-1) has two solutions (integrals), let us say,

Each equation represents a surface in space, and the intersection of any

two such surfaces specifies a magnetic field line.

There is a fundamental difference between field lines of Electrostatic fields

and field lines of Magnetostatic fields. The field lines of a static electric field
emerge from positive charges and dive in negative charges, meaning that the
field lines are open. However, the field lines of a static magnetic field are
always closed, (since in nature there are no magnetic charges, something
similar to electric charges). As we shall see in time varying fields, a time varying
magnetic field produces an electric field with closed field lines. But, for static
fields (charge densities and currents not depending on time), the field lines of
the electric field are open while the field lines of the magnetic field are closed.

Using field lines we may represent graphically magnetostatic fields

(approximately), generated by steady current distributions. If we make the
assumption that through a small surface of a fixed area (for example )
 564

placed perpendicularly to the field , at an arbitrary point in space, there

pass as many field lines as is the integral value of the magnitude of at ,
then we can obtain an approximate graphical representation of the field. At
regions where the field lines are dense the field is strong, while in regions
where the field lines are sparse the field is weak.

For example, let us consider the magnetic field configuration in Fig. 14-15.

Fig. 14-15: Magnetic bottle configuration.

The magnetic field is more intense (stronger) at the ends AA and BB and is
weaker in the middle. This important configuration, is called “magnetic
bottle”, (because of its shape), and is used frequently in order to trap charged
particles in space (in the region between AA and BB). Recall that charged
particles cannot be confined in space by electrostatic forces alone (Earnshaw’s
Theorem). In section 19-5, we shall explain how this field configuration can
 565

indeed trap charge particles in the region between AA and BB, by means of
magnetic forces.

b) Magnetic flux through a surface.

Let us consider a surface within a magnetic field , as shown in

Fig. 14-16.

Fig. 14-16: Magnetic flux through a surface.

The magnetic flux through the surface is defined as

where is the unit out normal to the surface at the position of . Since
, the normal component of the field , we see that (14-5-4) is
written equivalently as . Recall that the electric flux through
a surface is defined similarly, .

The unit of the magnetic flux is the , which according to the

definition in (14-5-4) is

Alternatively, we may say that the magnetic field is measured in

. Also, from the defining equation (14-5-4), the magnetic field is
 566

sometimes called “the magnetic flux density”, i.e. the magnetic flux per unit
area.

The magnetic flux through a closed surface is given by the integral

An important remark: Formula (14-5-5) is similar to the formula expressing

the total electric flux through a closed surface, , which
according to Gauss’s law is equal to the total electric charge enclosed by the
surface. So, we expect, in analogy to the electric flux, the total magnetic flux
through a closed surface to be equal to the total magnetic charge enclosed by
the surface. However, since magnetic charges do not exist in nature, we
expect that, for an arbitrary closed surface , . This is not a
proof, of course, it is just an indication. However, as we shall show rigorously in
Chapter 16, this is indeed true, and this fact, when formulated in mathematical
terms, leads to one of the four Maxwell’s equations.

Example 14-5-1: Find the magnetic flux through the area of the rectangle
, in Fig. 14-17.

Solution

Fig. 14-17: Magnetic flux through ABCD.

 567

We know that, in general, the magnetic field of an infinitely long,

filamentary conductor, carrying a current , is

Over the surface ABCD, which lies on the plane, ,

, and the magnetic flux is:

PROBLEMS

14-5-1) A magnetic field in free space is given, in cylindrical coordinates, as

, a) Show that and b) Find the magnetic flux

through the open cylindrical surface .

(Ans: ).

14-5-2) In Problem 14-5-1, show that the magnetic flux through the closed
surface of the cylinder is zero.

14-5-3) In Example 14-5-1, find the magnetic flux through ABCD, assuming
that .

(Ans: ).

SUPPLEMENTARY PROBLEMS

14-1) A point charge is moving along the axis, at a constant

speed . When the particle passes from the origin,
determine the electric field and the magnetic field at the point .
(Since , we may safely use the non relativistic eq. (14-2-4)).

(Ans: ,
).
 568

14-2) In Problem 14-1, assume that , and determine the and

fields, at the same point . Now we have to use the relativistic formulas
in (14-2-1).

Hint: . The angle , (see Fig. 14-1), is found from the

equation , etc.

14-3) Two infinite, straight, filamentary conductors are parallel to the axis.
One passes through and carries a current of in the positive
direction, while the other passes through and carries a current
of in the negative direction. Find the magnetic field at the origin.

(Ans: ).

14-4) a) Find the magnetic field at the point due to a

differential element placed at , and b)
Find the magnetic field at the origin due to a differential element
, placed at the point .

14-5) A current flows in the positive direction, along the infinite

axis. Find the magnetic flux through the surface of the rectangle ABCD, shown
in Fig. 14-18. Given: , (Ans: ).

Fig. 14-18: Magnetic flux through ABCD.

14-6) A uniform magnetic field exists in free space.

Find the magnetic flux through the surface of the triangle , provided that
.
 569

14-7) a) A straight, filamentary conductor of length , carries a current , as

shown in Fig. 14-19. Show that the magnetic field at the point is given by the
formula , b) Show that, in the limiting case, as
and , the magnetic field at reduces to formula (14-4-1), (the
field of an infinite filament).

Fig. 14-19: Magnetic field of a straight, finite filament.

Hint: As , , and as , .

14-8) A regular polygon with sides, inscribed in a circle of radius , lies in

the plane and carries a current , in the counterclockwise direction. a) Show
that the magnetic field at the center of the polygon is given by the formula

b) Show that as , (the regular polygon becomes a circle of radius ),

the magnetic field at the center reduces to formula (14-4-6), (the magnetic
field at the center of a circular current), (Recall that, ).

c) Apply the formula obtained in (a) to find the magnetic field at the center
of a regular hexagon, provided that and .
 570

14-9) A magnetic field , in cylindrical coordinates,

exists in free space. a) Show that , and b) Find the magnetic flux
through the surface .

(Ans: ).

14-10) A circular loop of radius , lying in the plane and

centered at the origin, carries a current in the clockwise direction.
Find the magnetic field at .

14-11) A surface current density exists in the region .

Find the magnetic field at .

(Ans: ).

Fig. 14-20: Magnetic field at the point A, due to the surface current .

14-12) In the problem 14-11, find the magnetic field at , assuming

that and . (Note that as and as .

14-13) A circular loop of radius , produces at its center a

magnetic field when a voltage is applied at its ends. Find the
resistance of the loop, (Ans: ).

14-14) Current flows in the conductor ABCD, shown in Fig. 14-21.

Find the magnetic field at the point D, due to three small segments, each of
 571

length , in the middle of the conductors AB, BC and CD. The side of the
cube is .

Fig. 14-21: Magnetic field at the point D.

14-15) a) The electron in the hydrogen atom moves on a circular orbit of

radius , with frequency . Find the magnetic field at
the center of the circular orbit, b) Assuming that the magnetic field remains
practically constant over the circular surface of the orbit, find the magnetic flux
through the surface.

(Ans: ).

Hint: If is the period of the electron, then the orbiting electron produces a
current , where is the charge of the electron, .
The magnetic flux is .

14-16) a) A uniform magnetic field exists in free space.

Find the magnetic flux through the surfaces ABOE, BCDO, and ACDE, b) Repeat
if the magnetic field is , c) Show that, in both
cases, the total magnetic flux through the closed surface of the prism is zero,
i.e. show that , (as usually, is the unit out normal to the
surface).
 572

Fig. 14-22: Magnetic flux through a surface.

 573

CHAPTER 15: THE MAGNETIC VECTOR POTENTIAL

15-1) Introduction

In Chapter 14, we have developed some methods and techniques to

determine the magnetic field , produced by various steady current
distributions. The starting point in our analysis was Biot-Savart’s law, which
actually gives the differential magnetic field, due to a current element, and
then, by integration over the source volume (superposition), we determined
the magnetic field everywhere in space.

Recall that, in Electrostatics, we have followed a similar approach. We first

found the electric field due to a charge element and then, by integration over
the source volume, we determined the electric field at any point in space.
However, as we have shown, the calculations are facilitated, if we find first the
electric potential and then compute the electric field as the negative
gradient of the potential, i.e. .

An analogous situation exists in Magnetostatics. To find the magnetic field

produced by a given current distribution, the calculations are facilitated if we
find first the so called “magnetic vector potential ” of the given current
distribution, and then obtain the magnetic field as the curl of , i.e.
. This is shown in the next section.

15-2) Magnetic Vector Potential of a Current Distribution

Let us assume that a current distribution (source currents), exists within a

volume , bounded by a surface as shown in Fig. 15-1. We shall use the
notation for the source points, (the points in space where the
sources are distributed) and for the field points, (the points
where the field is to be determined), (see Section 1-13).

For definiteness, we assume that at each point within the volume

, determined by the vector radius , there exists a current
density , assumed to be a known function of the spatial
coordinates . The problem is to find the magnetic field
 574

at the point in space, determined by its radius vector

.

Fig. 15-1: Magnetic field produced by a current distribution.

As we know, the magnetic field at is given by formula (14-3-6), i.e.

where, as always, , is the vector directed from the source point

towards the field point. Note that in the integration in (*), the vector is not
constant; its origin spans the volume , and in the integration process
covers the whole volume , while its terminal point remains constant.

The vector is a function of six

variables, ( ), while is a function of the three
variables . The integration with respect to in
equation (*) yields a result depending only on , (the are the
dummy variables in the integration), i.e. finally, the magnetic field shall be a
function of only, i.e. ,as expected. Invoking the vector
identities (see formula (1-13-5)),
 575

equation (*) takes the equivalent form:

At this point, if we apply the vector identity

(which is true for any scalar function and any vector function ), for
and , we obtain:

The term in (****) is identically equal to zero, i.e. ,

since is a function of only, while the operator
operates on the variables . Also, since the
integration in (****) is with respect to the variables, the operator
can be removed from the integral, and thus equation (****) assumes the form:

We should point out that, the integral inside the braces, is a function of
only, and that the curl of this function yields the magnetic field at
the point in space.

If we define “the magnetic vector potential ” as

then, the magnetic field is

 576

Since, as we know, , the magnetic vector potential

due to filamentary currents or to current sheets, is given by the following
formulas (obtained from (15-2-1):

The magnetic vector potential is measured in as it follows from

equation (15-2-2).

Question: Why do we have to introduce the magnetic vector potential in

our calculations?

Answer: The potential facilitates the mathematical calculations, when we

try to compute the magnetic field produced by given current distributions.
Recall that the direct computation of by means of formula (14-3-6), involves
integrals which are quite difficult to be evaluated. However, the corresponding
integral for the evaluation of in equation (15-2-1) is much simpler. Having
found , the magnetic field , i.e. is obtained by means of
differentiations, which as a rule, are much easier as compared to integrations.

We have actually followed a similar approach for the determination of static

electric fields. Instead of computing the electric field directly from the charge
distribution, we found first the potential and then obtained as the
negative gradient of , i.e. .

As it was the case with the electric potential , the magnetic vector
potential is not uniquely determined. If is the magnetic vector potential of
a current distribution, then , where is an arbitrary
scalar field, may also serve as a vector potential, since
 577

(the curl of the gradient is identically equal to zero). We may say that the
vector potential is determined up to the gradient of a scalar function.

In Magnetostatics, we choose so that . In time varying fields, as

we shall show, a different choice of is appropriate.

Remark 1: In deriving formula (15-2-1) we have tacitly assumed that all the
current distributions, producing the magnetic field , are contained within a
finite volume. So, in general, we should avoid cases where currents exist to
infinity, which, anyways, are “hypothetical situations”.

Remark 2: Since the divergence of the curl of any vector field is identically
zero, it follows from (15-2-2) that

Even though formula (15-2-5) was derived for static fields, it is true for time
varying fields as well, and constitutes one of the four Maxwell’s equations. In
all cases, (static and time varying), the magnetic field is “solenoidal”, (see
Section 1-12).

Example 15-2-1: The magnetic field of a straight, infinite, filamentary

conductor, coincident with the axis and carrying a current , is
, (see equation (14-4-1)). Show, by direct computation,
that .

Solution

In this case, and . The divergence of , in

cylindrical coordinates is
 578

Example 15-2-2: An axially symmetric magnetic field exists in the region

. Given that the component of the field is ,
find the radial component of the field.

Solution

Due to the axial symmetry, the magnetic field does not depend on , i.e.
. The divergence of must be zero, (equation (15-2-5)), and
expressing the divergence in cylindrical coordinates, we have:

Equation (*) is a differential equation for . The solution of this equation,

which remains bounded for , i.e. on the axis, is found to be

(Let the reader check it).

Example 15-2-3: A short, thin filament of length , at the origin, carries a

current in the positive direction. Find the potential and the magnetic field
at the point , assuming , i.e. far away from the current
element.

Solution

Fig. 15-2: Magnetic potential of a finite current element.

Application of formula (15-2-4) yields:

 579

Assuming that the point is far away from the current element, we may
set, , and formula (*) implies:

and since (in spherical coordinates), we find the

following expression of in spherical coordinates:

Let the reader verify that the magnetic field at the point is:

We also note that, by direct computation, we find , as expected,

(see Problem 15-2-1).

Example 15-2-4: Find the magnetic vector potential of an infinite,

straight, filamentary conductor, carrying a current in the positive direction,
and extending from up to .

Solution

In this example, the current is not contained within a finite region in space,
on the contrary, extends to infinity, and this creates certain problems when we
try to apply formula (15-2-4) to compute the potential , (the corresponding
integral diverges). Recall that we encountered a similar problem, in
Electrostatics, when we tried to compute the potential of charge distributions
extending to infinity. In this case, in Electrostatics, we were not allowed to
assign potential zero at infinity, simply because we do have charges at infinity.
 580

A similar situation arises in this example, and we must solve the problem
using a slightly different approach.

As we know, (equation (14-4-1), the magnetic field of the infinitely long

filamentary conductor, is , and since, in general, the
magnetic field , we have:

This is a differential equation for . Due to the cylindrical symmetry

involved, must depend only on , and equating the component of
with the component of the magnetic field, we find:

where is the arbitrary constant of integration. To determine , we need a

“reference point” for the vector potential. We may, arbitrarily, assume that
at , i.e. on the cylindrical surface , where, again, is an
arbitrary positive number. With this choice of “reference for the magnetic
potential”, we find from (**) that , and finally,

Note that, since , the magnetic field , does not depend on the
choice of the reference surface .

Example 15-2-5: A charge uniformly distributed over a spherical surface

of radius , is set spinning at angular speed . Find the vector potential and
the magnetic field everywhere in space.

Solution

The surface charge density is .

 581

The spinning charge produces a surface current density, at a point on the

spherical surface, determined by its radius vector , equal to

Fig. 15-3: Magnetic field of a spinning charged spherical surface.

The magnetic potential at the point , according to eq. (15-2-3), is

and since and are constants,

The evaluation of the integral in (**) is shown in Fig. 15-3(b).

Let be the intersection of a plane perpendicular to , at the point

, with the spherical surface .
 582

The vector , where is the component of in the direction of

, which is a constant direction in space, and is the projection of in the
direction , which is perpendicular to . All the points of the circumference
are equidistant from the point , where the field is to be determined.
We thus have:

The term , in equation (***) is zero. This is so, since, to each there
corresponds a term, so we have pair wise cancelation, and as a
result, the value of the integral is zero. To determine the value of the term ,
in equation (***), we note that

and since is constant with respect to the integration, we finally

obtain:

and if we make the substitution , the integral

becomes, (see Problem 15-2-2),

Note that, inside the spherical surface, ,

while outside the spherical surface, , and
finally we obtain the following expression for the vector potential at the
point :
 583

Assuming further that the spherical shell spins about the axis, ( ),
formula (****) yields:

It turns out that, the magnetic field inside the sphere is uniform, since

For the magnetic field outside the spherical shell, see Problem 15-2-4.

PROBLEMS

15-2-1) For the magnetic field , (formula (***) in Example 15-2-3), show
that .

15-2-2) Evaluate the integral , using the substitution

,( ).

15-2-3) A uniform surface current density exists on the plane

, while another uniform surface current density exists on the
plane . Find the magnetic vector potential in the region between the
two planes, (Ans: ).

Hint: See Examples 14-4-2 and 15-2-4.

15-2-4) In Example 15-2-5, find the magnetic field outside the spherical
shell.
 584

15-3) Magnetic Flux Through a Surface, in Terms of the Magnetic Vector

Potential

Fig. 15-4: Magnetic flux in terms of the magnetic vector potential.

Let ( ) be a surface in space, bounded by a closed curve ( ), as in Fig. 15-4.

The surface exists inside a magnetic field . The magnetic flux through the
surface ( ), is defined by the surface integral , where is the
unit out normal to the surface. If is the magnetic vector potential, associated
with the magnetic field , then,

(using Stokes Theorem). Formula (15-3-1) expresses the magnetic flux

through a surface, in terms of the magnetic vector potential .

Example 15-3-1: In Example 15-2-4, we found that the magnetic field

produced by an infinite filamentary conductor, carrying current in the
positive direction, is given by the formula , where is
the “reference cylindrical surface” at which vanishes. Using this expression
for , work Example 14-5-1, (Fig. 14-17).
 585

Solution

With reference to Fig. 14-17, we have:

Over the sides and the dot product , since , and

formula (*) yields:

which is simplified to the following,

The result is identical to the one obtained in Example 14-5-1. Note that is
independent from the “reference surface ”, as expected.

PROBLEMS

15-3-1) Find the vector magnetic potential associated with the magnetic
field , (in cylindrical coordinates).

(Ans: ).

Hint: Assume ).

15-3-2) For the magnetic field of Problem 15-3-1, find the magnetic flux
through the circle , lying on the plane, a) Using the formula
, and b) Using the formula (15-3-1), and thus verify that in
both cases, the results are identical, (Ans: ).
 586

15-4) Magnetic Dipole and Magnetic Dipole Moment

A small, current carrying closed loop is called “magnetic dipole”. The main
problem, in this section, is to find the magnetic field produced by the
magnetic dipole.

Fig. 15-5: Magnetic dipole.

Let us assume that the small current loop lies on the plane, and carries a
current , as shown in Fig. 15-5. We want to find the magnetic field at a point
in space, determined by its vector radius . It will be easier to find at , first
the vector magnetic potential , and then find the magnetic field .

The source points are the points on the loop , and we shall use the
notation for the infinitesimal current source. According to eq. (15-2-4), the
vector potential will be

To evaluate the integral, we shall use the vector identity, (for a proof, see
Pr. 14-5-1),
 587

where the closed contour is the boundary of an open surface , (see

for example Fig. 15-4) and is a scalar function of the spatial coordinates. As
always, is the unit out normal to the surface. Application of this identity with
and , yields:

since , (see formula (**) in Section 15-2).

For the “far field” of the magnetic dipole, i.e. for the field far away from
the magnetic dipole, we may safely assume that , and formula
(**) reduces to the following:

and since is constant with respect to the integration, we finally

obtain,

where is the unit vector along and is the area of the loop.

At this point we define the “magnetic moment” of the magnetic dipole as

or, more generally,

where is the surface of the loop, is the loop current and is the unit
vector, normal to the loop surface.
 588

Fig. 15-6: Magnetic dipole moment.

Note the direction of the unit vector , in reference to the direction of the
current flow. The direction of is the direction a right handed screw advances,
when it is rotated in the direction of the current flow.

In terms of the magnetic moment, the vector potential in formula (***)

assumes the form:

Having determined , we may find the “far magnetic field” easily, as the
curl of , (see Problem 15-4-2), i.e.

Formula (15-4-4) may be written in a “coordinate free form” as

where is the vector extending from the magnetic dipole to the point
, and is the unit vector, (for a proof, see Problem 15-4-3).

Notice the close similarity between formulas (5-6-4) and (5-6-5), (expressing
the far electric field of an electric dipole ), and formulas (15-4-4) and (15-4-5)
respectively, (expressing the far magnetic field of a magnetic dipole ).

Remark: In Chapter 10, we studied static electric fields inside matter, with
the aid of electric dipoles. We may, therefore, anticipate that magnetic dipoles
 589

shall be used, in a similar way, to study static magnetic fields inside matter.
This is indeed true, and this topic will be developed in details in Chapter 20.

Example 15-4-1: A magnetic dipole

is placed at . Find the magnetic field at the point
.

Solution

Fig. 15-7: Magnetic field due to a magnetic dipole.

The vector radius from to is

Application of formula (15-4-5) yields:

Example 15-4-2: Show by direct calculations that, for the magnetic field
of a magnetic dipole, .

Solution

From equation (15-4-4) we see that the components of the magnetic field
are:
 590

The divergence of in spherical coordinates is:

PROBLEMS

15-4-1) Show vector identity (15-4-1) for the particular case where the
closed contour lies on the plane.

15-4-2) Starting with equation (15-4-3) show equation (15-4-4).

15-4-3) Starting with equation (15-4-3) show equation (15-4-5).

15-4-4) A magnetic dipole is placed

at . Find the magnetic field at the point .

Hint: See Example 15-4-1.

15-4-5) A simple model of the Hydrogen atom, consists of a heavy, positive

nucleus of charge and a negatively charged electron of charge ,
revolving about the nucleus. Assuming that the nucleus is located at the origin
and that the orbit of the electron lies on the plane, (see Fig. 15-8), find the
magnetic dipole moment of the atom.

Given: Radius of orbiting electron , period of circular motion

, charge .

(Ans: ).
 591

Fig. 15-8: Magnetic dipole moment of an atom.

Hint: The orbiting electron, in the direction shown (counterclockwise), is

equivalent to a current , in the opposite direction, (since the charge
of the electron is negative).

SUPPLEMENTARY PROBLEMS

15-1) Find the vector magnetic potential associated with:

a) A uniform magnetic field , (Cartesian coordinates),

b) A uniform magnetic field , (Cylindrical coordinates),

c) The magnetic field , (Spherical coordinates).

(Ans: a) , where is an arbitrary

constant, b) , c) ).

15-2) Two magnetic dipoles and

, are placed at the points
and respectively. Find the magnetic field at the
point .

Hint: Use formula (15-4-5) and the superposition principle.

 592

15-3) The magnetic field of a spinning charged sphere: A uniformly charged

sphere of radius , centered at the origin and carrying a total charge , is
rotating about the axis, at a constant angular speed . Find the magnetic
field everywhere.

Fig. 15-9: Magnetic field of a spinning charged sphere.

(Ans: Outside ( ):

Inside ( ):

Hint: Use the results obtained in Example 15-2-5 and Problem 15-2-4.

15-4) a) Use the results obtained in Problem 15-3, to show that the
magnetic field outside the sphere is given by the formula:

b) Compare this expression with the expression for the magnetic field of a
magnetic dipole (formula (15-4-4)), and thus deduce that the spinning charged
sphere behaves as a magnetic dipole with dipole moment .

15-5) For the spinning sphere of Fig. 15-9, it is given:

 593

. Apply the formula found in

Problem 15-4, to find the magnetic field at the point whose Cartesian
coordinates are

Hint: Find the spherical coordinates of the point .

15-6) Find the vector magnetic potential of a loop of radius , lying on the
plane, centered at the origin and carrying a current .

Fig. 15-10: Magnetic vector potential of a circular loop.

(Ans: , while

This integral, in general, cannot be evaluated in closed form. However, if

the point is near to the axis, meaning that , then we can find a
useful, approximate expression of , as shown in Problem 15-7.

15-7) In Fig. 15-10, show that the vector magnetic potential near the axis,
is given (approximately) by the expression
 594

Hint:
. Then use the approximation:

with , since, by assumption, . Use this

approximation in the expression for , in Problem 15-6, to obtain the
desired approximation for the vector magnetic potential, near to the axis.

15-8) Use the expression for obtained in Problem 15-7, to find the
(approximate) expression for the magnetic field , near to the axis.

Hint: .

15-9) The magnetic field exists in free space. a)

Show that , and b) Find the vector magnetic potential of the field .

(Ans: ).

Hint: Due to the arbitrariness of , (recall that if is a vector potential of ,

then , where is an arbitrary scalar field, is also a vector potential,
since ), we may assume , etc.

15-10) Find the vector magnetic potential of the magnetic field ,

produced by a charge that is moving with at constant velocity , (equation
(14-2-4)).

(Ans: ).
 595

CHAPTER 16: THE TWO FUNDAMENTAL LAWS OF STATIC MAGNETIC

FIELDS

16-1) Introduction

In the preceding chapters we have developed a systematic method to

calculate the static magnetic field , generated by various distributions of
steady currents. In this chapter, we shall study the two fundamental laws
which hold true for any static magnetic field . These two, fundamental laws,
are the following:

a) The static magnetic field is solenoidal:

b) Ampere’s circuital law:

In the rest of this chapter, we shall derive rigorously equations (16-1-1) and
(16-1-2), and study some of their most important applications.

Some general remarks:

1) In each equation, the so called “integral form” is equivalent to the

corresponding “point form”. One can go from the point form to the integral
form, and vice versa, by a suitable application of proper theorems of vector
analysis.

2) While equation (16-1-2) holds true strictly for static fields, eq. (16-1-1) is
of general validity; it holds true for static and time varying fields as well, and
 596

is actually one of the four Maxwell’s equations. It is a general law applying to

all Electromagnetic fields, not just to the static ones.

We now proceed with the proof of equation (16-1-1).

16-2) The Magnetic Field is Solenoidal ( )

As we have shown in Section 15-2, equation (15-2-2), the magnetic field

can be expressed as the curl of the vector magnetic potential , i.e. .
The divergence of is

since the divergence of the curl, of any vector field, is identically equal to
zero. Even though equation results from equation , which
was proved for static magnetic fields, vast experimental data shows that
for all magnetic fields, static and time varying. It is a fundamental
law, applying to all Electromagnetic fields.

According to the terminology introduced in Section 1-12, the magnetic field

is solenoidal. No expression for a magnetic field is valid, unless it satisfies
. Equation constitutes one of the four Maxwell’s
equations. It expresses the deep and fundamental fact that in nature, there
are no magnetic charges (something analogues to electric charges).

To show the integral form, in equation (16-1-1), we consider a closed

surface within a magnetic field . The total magnetic flux through the
closed surface , is

In equation (*) we have applied the divergence theorem, to transform the

surface integral to a volume integral.
 597

Equation (*) shows that the total magnetic flux through any closed surface
is zero.

Fig. 16-1: Magnetic flux through a closed surface .

In Fig. 16-1, the flux at the point is negative, since the angle between
and is obtuse, while at the point the flux is positive, since the angle
between and is acute. At the point flux enters into the closed surface,
while at the point flux exits from the closed surface. The physical meaning of
the equation , is that the amount of flux entering ( ),
(negative flux), is equal to the amount of flux exiting the surface ( ), (positive
flux). Physically, this is easily explained. Recall that in electric fields, the electric
flux through a closed surface ( ), , could be positive if
encloses a net amount of positive charge, or it could be negative if
encloses a net amount of negative charge, or even it could be zero, if the net
positive charge enclosed is equal to the net negative charge enclosed.
However, in magnetic fields, we do not have “magnetic charges” at which
magnetic field lines can either emerge from or dive to, so the number of
magnetic field lines entering must necessarily be equal to the number of
magnetic field lines exiting from the closed surface, in other words, the total
magnetic flux through any closed surface must be zero.
 598

Example 16-2-1: Which one of the following two fields is an impossible

magnetic field? a) , b) .

Solution

Any magnetic field must satisfy the equation . An expression for

which does not satisfy , is an impossible magnetic field. For the
magnetic field in (a), we have:

This shows that is a valid expression for a magnetic

field.

Regarding the expression of in (b), we have:

and since is not identically zero, the expression of in (b) is an

impossible magnetic field.

Example 16-2-2: Consider the magnetic field . a)

Show that , and b) Show by direct calculation that the magnetic flux
through the surface of a unit cube, whose one corner is at the origin and the
opposite corner is at the point , (Fig. 16-2), is zero.

Solution

The total flux through the closed surface of the unit cube is:
 599

Fig. 16-2: Magnetic flux through the surface of a unit cube.

1) On the surface : , , , , and

2) On the surface : , , , , and as in

part (1), we find

3) On the surface : , , , , and

 600

4) On the surface : , , (obtained from the

general expression of for ), , , (since
), and

5) Working similarly we find:

The total magnetic flux through the closed surface of the unit cube, is thus

as expected.

Alternative solution: Application of the divergence theorem, yields,

PROBLEMS

16-2-1) Show that it is not possible to have a magnetic field of the form
, (in spherical coordinates).
 601

16-2-2) The magnetic field

exists in free space. a) Show that , and b) Show that the total
magnetic flux through the closed surface of the unit cube in Fig. 16-2, is zero.

16-2-3) For the magnetic field of a moving point charge , (eq. (14-2-4)),
show that .

16-2-4) If is the magnetic field of a spinning charged sphere, (see

Supplementary Problem 15-3), show that .

16-3) Ampere’s Circuital Law ( )

In this section we shall give a formal derivation of Ampere’s circuital law,

starting with the magnetic field of an infinite filamentary conductor, carrying a
current , (formula (14-4-1)). Then, in the next section, we will work several
problems and applications to show that, for currents with special symmetry,
Ampere’s circuital law in integral form, provides an extremely powerful
method for finding the produced magnetic fields, without having to resort to
the evaluation of the complicated integrals appearing in formulas (14-3-2), (14-
3-4) and (14-3-6).

A rigorous derivation of Ampere’s circuital law, starting with the Biot-Savart

law, is given in section 16-5.

a) The magnetic field of an infinite filamentary conductor, coinciding with

the axis, and carrying a current , is . Let us evaluate the
integral around: 1) a circle of radius , centered at the
conductor, and 2) around another circle not enclosing the conductor.

In the first case, (circle centered at the conductor), we have:

In the second case, (circle not enclosing the conductor), we have:

 602

Fig. 16-3: Evaluation of the circulation , (one conductor).

since , and , (see Fig. 16-3).

b) The results obtained in (a) are still valid, for arbitrary closed space
curves, and not just for plane circles, lying on planes perpendicular to the
filamentary conductor, i.e.

Indeed, if is an arbitrary closed contour, enclosing the filamentary

conductor, then, since in cylindrical coordinates the differential length is
, we have:
 603

(since and ). The integral if the

closed path encloses the conductor, and is zero if does not enclose the
conductor, and this completes the proof of formula (16-3-1).

c) We may, now, easily extend formula (16-3-1), in cases where we have

many conductors. Let us, for definiteness, consider a case with three
conductors, as in Fig. 16-4.

Fig. 16-4: Evaluation of the circulation , (many conductors).

Two of the conductors are enclosed by the arbitrary, closed, space curve
, while the third conductor is not enclosed by .

The total magnetic field of the given current configuration, is, according to
the superposition principle, , where are the
magnetic fields produced by the currents and respectively. The line
integral of over the closed contour is:

which, by virtue of (16-3-1) yields,

 604

and in general,

where is the total current enclosed by the arbitrary closed

path . Note that the current , not enclosed by , contributes nothing to
the integral. Formula (16-3-2) is Ampere’s circuital law, in integral form.

d) Let us further assume that there is a current density , defined in space,

and is an arbitrary closed path, bounding a surface . The current density
, in general, is a function of the spatial coordinates.

Fig. 16- 5: Derivation of Ampere’s circuital law in point form.

Application of (16-3-2) yields:

and transforming the line integral to a surface integral, using Stokes

theorem, we have:
 605

and since this must be true for an arbitrary surface , enclosed by an

arbitrary closed path , the integrand must be identically equal to zero, i.e.

Formula (16-3-3) is Ampere’s circuital law, in point form.

16-4) The Magnetic Field Intensity , Amperian Loops

In free space (vacuum), we introduce a new vector, called “the magnetic

field intensity ”, or just the “field ”, defined as

In terms of the field , Ampere’s circuital law, in vacuum, assumes the

form:

As it follows from (16-4-2), is measured in .

Some general remarks:

1) In the general case, both and , may be functions of the spatial

coordinates, but not of time, since we are still in the regime of Magnetostatics.

2) In vacuum, , as it follows immediately from equation (16-4-1)

and the fact that . However, as we shall see in Chapter 21, for
magnetic fields inside matter, we may have . However, even in this
 606

case, (magnetic fields inside matter), the divergence of is still zero,

identically, ( ).

3) The current density in (16-4-2) is due to the motion of electrons within

conductors, or more generally, to the motion of charged particles within
conducting media. As we shall see in Chapter 20, (magnetic fields inside
matter), the current density must include another term, which accounts for
the effects of “the bound currents”, induced when a piece of material is placed
inside an existing magnetic field.

4) When symmetry permits, Ampere’s circuital law in integral form, is used

to provide an easy and straightforward method for the determination of the
field . In a sense, Ampere’s circuital law in integral form is the magnetic
analogue of Gauss’s law in Electrostatics, ( ). Just
as Gauss’s law allows the easy determination of the electric field without
resorting to the evaluation of complicated integrals, so Ampere’s circuital law
provides an easy and powerful method to determine the magnetic field .

5) Strictly speaking, for a known , the second eq. in (16-4-2) is

an integral equation for the unknown . If symmetry allows, the solution for
is simplified if we can choose a closed path , termed “Amperian loop”,
which satisfies the following two conditions:

a) is tangential to the Amperian loop (in which case ,

being the tangential component of ), and

b) The tangential component remains constant along the Amperian

loop, and as such, it can be pulled out of the integral in (16-4-2), (this is the key
idea).

The proper choice of an Amperian loop depends on the symmetry of the

current distribution of the problem at hand.
 607

6) A convention about the direction of the currents and the fields.

Several times, in order to avoid three dimensional figures, which are usually
difficult to be drawn, we prefer to work with cross sections of a system with
planes, which are two dimensional figures. In this case, we have to adopt a
convenient method, to indicate the direction of the currents or the fields.

A convention which is widely used is the following:

a) We use the symbol to indicate that the field or the current, is

perpendicular to the page, and is directed from the reader towards the page,
and

b) The symbol indicates that the field or the current, is perpendicular to

the page, and is directed from the page towards the reader.

Example 16-4-1: Four parallel, infinite, filamentary conductors, parallel to

the axis, carry currents and . The cross
section of the conductors with the plane, is shown in Fig. 16-6. Find the
path integrals: a) , b) , c) .

Solution

Fig. 16-6: Evaluation of the integrals .

 608

The indicated direction of integration, along the closed paths, defines the
direction of in each path. Having defined the direction of traversing the
closed paths, we have to decide which currents are considered as positive and
which as negative. The answer is given by the “right hand rule”. If the figures
of the right hand show the direction of the integration around the closed path,
then the direction of the thumb defines the positive currents, (see Fig. 14-8). A
current flowing in the opposite direction is regarded as a negative current.

For example, in the integration around , and are positive, while

is negative.

Question: Can we use, for example, , to find the magnetic

field , at the points of the closed path ?

Answer: No, since ( being the tangential component of

), does not remain constant over . Equation is a correct
equation, but it does not help to determine the field .

Example 16-4-2: Find the magnetic field of an infinitely long, filamentary

conductor, coincident to the axis, and carrying a current in the positive
direction.
 609

Solution

We have actually worked this problem (see equation (14-4-1)), using the
Biot- Savart law and the principle of superposition, (integration over the source
currents).

In this example, we shall derive the magnetic field using Ampere’s circuital
law.

Fig. 16-7: Magnetic field of an infinitely long filamentary conductor.

Due to the cylindrical symmetry involved, the magnetic field depends

only on , the distance of the field point from the filamentary conductor, and
points in the direction of the local unit vector, i.e. .

A suitable Amperian loop, for this problem, is a circle of radius , lying in a

plane perpendicular to the filamentary conductor, with its center on the axis.
For such an Amperian loop, as in Fig. 16-7, we have:

(since ). Note that, since is constant over the Amperian

loop ( ), application of Amber’s circuital law around ( ) yields:
 610

As a vector, , and , (identical

with the expression of in equation (14-4-1)).

Note: An ideal filamentary conductor, cannot really exist, since in this case,
as , i.e. very close to the conductor, the magnetic field would blow up
( as ). We faced a similar problem with point charges. To remedy
the problem, we may assume that the filament has a very small, but still, finite
radius , and that the current is uniformly distributed over the cross section
of the filament. This problem is examined in the next example.

Example 16-4-3: A straight filamentary conductor, of infinite length and

finite radius , carries a current in the positive direction. Assuming that the
current is uniformly distributed over the cross section of the conductor, find
the magnetic field everywhere.

Solution

Fig. 16-8: Magnetic field of a cylindrical conductor.

The second figure in Fig. 16-8, shows the cross section of the conductor,
with a plane perpendicular to the axis. Due to the cylindrical symmetry, the
 611

magnetic field, at every point in space, will depend only on , and its direction
will be in the direction of the local vector, i.e. .

Since the current is uniformly distributed over the cross section of the
conductor, the current density .

1) The magnetic field inside the conductor, .

Application of Ampere’s circuital law over the closed path , yields:

2) The magnetic field outside the conductor, .

Application of Ampere’s circuital law over the closed path , yields:

In summary:
 612

The graph of as a function of , is shown in Fig. 16-9.

Fig. 16-9: Graph of the magnetic field versus .

In this model of the filamentary conductor, the magnetic field is zero at the
center, increases linearly up to the boundary , and then falls off as
.

Example 16-4-4: A magnetic field ( ) exists in free

space, a) Show that , b) Find the current density at the
point in space, and c) Find the current through the surface
.

Solution

For the given magnetic field: .

a) Using the expression of the divergence in spherical coordinates, we find:

b) From the point form of Ampere’s circuital law:

 613

c) The total current through is:

Example 16-4-5: In Example 16-4-4, find the current through ( ), using the
integral form of Ampere’s circuital law.

Solution

Fig. 16-10: Total current through the surface .

 614

From Ampere’s circuital law in integral form, we have:

Over the paths and , the differential vector is perpendicular to ,

meaning that the dot product . Over the path , ,
the dot product , and therefore, equation (*) yields:

same result, of course, with the result obtained in Example 16-4-4 (c).

PROBLEMS

16-4-1) For the magnetic field , in Example 16-4-3, (formula (***)), show
by direct calculation that and , (zero outside and
inside).

16-4-2) The magnetic field inside a cylindrical conductor is

, ( ), where are given constants. a) Show that
and b) Find the current density everywhere inside the conductor.

Hint: .

16-4-3) For the magnetic field of a magnetic dipole, (formula (15-4-4)),

show that and .

16-4-4) A magnetic field ( ),

exists in free space. Find the total current passing through the plane, in the
 615

direction, inside the square . Use two different

methods and compare the results.

Hint: One method is to find and integrate over the surface of the
square, the other method is to evaluate around the perimeter of the
square.

16-4-5) A magnetic field exists in free

space: a) Show that , b) Find the current density everywhere in
space, and c) Find the total current through the conical surface
, (Ans: ).

16-4-6) A magnetic field exists in free space, in

the region : a) Show that , b) Find the current density ,
c) Find the total current, in the direction, through a circle of radius
, lying on the plane and centered at the origin.

16-5) Rigorous Derivation of Ampere’s Circuital Law from the Biot-Savart

Law

In this section, we give a rigorous proof of Ampere’s circuital law, starting

with the fundamental Biot-Savart law.

Fig. 16-11: Derivation of Ampere’s circuital law from the Biot-Savart law.
 616

By virtue of equations (15-2-1) and (15-2-2), the vector magnetic potential

and the magnetic field at the point are:

The curl of the magnetic field is therefore,

or, making use of the vector identity,

(which is true for any vector field , in general), we find:

a) Let us consider, first, the term .

since the integration is with respect to the source coordinates ,

while the operator operates on the field coordinates .

If we now apply the vector identity , which

is true for any scalar field and any vector field , for and , we
obtain:
 617

since the term is zero identically, (note that is a function of

, while operates on ). As we have also proved, (see
equation (**), in section 15-2), , and equation (***) becomes:

and substituting in (**) we get:

Using, again, the same identity in (***), with replaced with , we find

or, since for steady state currents , (see equation (13-5-3)), we

find that , and thus equation (****) becomes,

(by application of Gauss-Ostrogradsky theorem).

At this point we make the reasonable assumption that all currents are
contained within the volume . This means that over the closed surface , the
boundary surface of , the current density is either zero, or it is tangent to
the surface, i.e. . In either case, , and therefore, equation
(*****) implies that . Equation (16-5-1), now, simplifies to the
following:

or, if we assume that , then

 618

b) We now consider the terms , where .

From the first equation in (*) we find that

At this point, we recall the analogous problem in Electrostatics.

A volume charge density , within a volume , generates a potential

in space, given by the formula (see equation (5-5-4)),

At the same time, as we know, the potential satisfies Poisson’s

equation

In summary, the last two equations imply that if a scalar field satisfies
equation (16-5-3), then its Laplacian satisfies eq. (16-5-4), and vice
versa. As a result, as expressed in (******) must also satisfy

and substituting in equation (16-5-2) we find:

and this completes the proof.

 619

SUPPLEMENTARY PROBLEMS

16-1) A magnetic field exists in

free space: a) Show that , b) Find the current density and c) Find the
total current passing through the plane, in the positive direction, inside
the square .

(Ans: ).

16-2) The vector magnetic potential associated with a current distribution

in free space, is ( ): a) Show that
, b) Find the magnetic field , and c) Find the current density .

16-3) Within an infinitely long, cylindrical conductor of radius , the current

density is and it is zero outside: a) Find the magnetic
field everywhere in space, and b) By direct calculation verify that .

(Ans: ).

16-4) A magnetic field exists in free

space: a) Show that , b) Find the current density , and c) Find the
total current through the surface of a circle of radius , lying on the
plane and having its center on the axis.

16-5) The planes and are perfect conductors. A voltage

is maintained between them, (the plane is held at
while the is grounded). The space between the two
perfectly conducting planes, is filled with a conducting material with
conductivity . a) Find the electric field and the current
density in the region between the two planes, b) Find the total current in the
positive direction, through a cylinder of radius , and c) Exploit the
cylindrical symmetry to determine the magnetic field , using Ampere’s
circuital law in integral form.
 620

(Ans:

, for and it is zero everywhere

else).

16-6) An infinite cylindrical shell of radius carries a uniform surface

current density . Find the magnetic field , using two different
methods, a) Using formula (14-3-4) and b) Ampere’s circuital law. Which
method is easier?

(Ans: ).

16-7) The magnetic field inside an infinitely long cylindrical conductor is

: a) Find the current density , and b) Find the current
flowing through a circle of radius , lying on the plane and
centered at the origin.

(Ans: ).

16-8) A current density exists in the region

, and is zero everywhere else. On physical grounds, justify
why the magnetic field cannot depend on and , i.e. depends only on , and
furthermore, why . Then, apply Ampere’s circuital law, or any
other method, to find the magnetic field everywhere in space.

(Ans: ).

16-9) The cross section of an infinitely long coaxial cable, whose axis
coincides with the axis, is shown in Fig. 16-12. A current flows in the inner
conductor ( ) in the positive direction, and returns back through the
outer conductor ( ). Assuming that the current is uniformly
distributed over the cross sections of the inner and the outer conductors, find
the magnetic field everywhere in space, and plot versus .
 621

Fig. 16-12: Cross section of a current carrying coaxial cable.

Hint: The current density in the inner conductor is and in the

outer conductor is . Due to the cylindrical symmetry,
the magnetic field . As Amperian loops consider circles concentric
to the axis, and apply Ampere’s circuital law in integral form.

(Ans: ).

Fig. 16-13: Plot of versus .

 622

16-10) In Problem 16-9, by direct calculation show that .

16-11) A current flows in a cylindrical conductor of radius , coaxial with the

axis. The current density is for (inside the
conductor), and is zero for (outside the conductor). Determine the
magnetic field everywhere.

( is the Bessel’s function of the first kind, of order zero, and is

the approximate value of the first positive root of , see section 12-6
for an elementary introduction to Bessel’s functions).

(Ans: , where is the Bessel’s

function of the first kind, of order one).

Hint: . To show this, it suffices to show that

. For the expressions of and , see section 12-6,
formulas (12-6-4) and (12-6-5).

16-12) In Problem 16-11, by direct calculation show that .

 623

CHAPTER 17: SOLENOIDS AND TOROIDS

A solenoid consists of a circular cylinder of radius , around which a wire is

tightly wrapped. Similarly, a toroid consists of a toroid (a circular ring) around
which a wire is tightly wrapped. If the wire carries a steady current , then the
magnetic field inside a solenoid is (approximately) uniform and it is zero
outside. In a toroid, the magnetic field is circumferential inside the toroid, of
approximately, constant magnitude, and it is zero outside.

17-1) Magnetic Field of a Solenoid

Let us first consider an “ideal solenoid”, consisting of an infinite cylindrical

shell with a circular current sheet , as shown in Fig. 17-1 (a). The
practical solenoid consists of closely wound turns on a cylindrical surface of
finite length , as in Fig. 17-1 (b).

Fig. 17-1: a) Ideal Solenoid, b) Practical Solenoid.

As we shall show shortly, the magnetic field of an ideal solenoid with a

surface current density is uniform (constant magnitude and
direction) inside the solenoid and it is zero outside. The magnetic field is given
by the formula
 624

Ideal Solenoid:

For the practical solenoid in Fig. 17-1 (b), assuming that the solenoid has
closely wound turns, then, , and if we define , (turns per
unit length), it follows that , and from formula (17-1-1) we obtain the
magnetic field for the practical solenoid.

Practical Solenoid:

This approximation for the magnetic field is valid, on the assumption that
the radius is much less than the length , i.e. when .

To justify formula (17-1-1), we shall show, step by step, the following:

1) The magnetic field produced by the current density , is in the

direction, i.e. in the direction of the axis of the solenoid. The proof of this
statement is left as a problem, (see Problem 17-1-3).

2) From the “right hand rule”, inside the solenoid the magnetic field points
in the positive direction, while outside the solenoid, the magnetic field points
in the negative direction.

3) Due to the cylindrical symmetry involved, will depend only on

the radial distance , i.e.

4) The magnetic field outside the solenoid is zero. Indeed, let us consider a
section of the solenoid with a plane passing through its axis, as in Figure 17-2.
 625

Fig. 17-2: Computation of the magnetic field of a solenoid.

Application of Ampere’s circuital law in loop (a) yields:

Notice that on the horizontal sides of the loop, is perpendicular to ,

and therefore, their dot product is zero.

Formula (**) shows that outside the solenoid, the magnetic field has a
constant magnitude, independent from . Far away, however, from the
solenoid, the magnetic field must be zero, and since its magnitude must
remain constant, it must be zero everywhere outside. We are thus led to the
conclusion that everywhere outside.

To find the magnetic field inside the solenoid, we apply Ampere’s circuital
law in loop (b), and since the field is zero outside, we have:
 626

or, as a vector, inside the solenoid, and this completes the proof
of equation (17-1-1).

Let us now consider a practical solenoid with turns and a finite length .
Assuming that , we may set , where is
the number of turns per unit length, and substituting this expression of in
(17-1-1) we obtain equation (17-1-2).

For an alternative derivation of formula (17-1-2), see Example 17-1-2.

Example 17-1-1: a) A current flows in a solenoid and produces a

magnetic field in its interior. Find the number of
turns per unit length, b) If the radius of the solenoid is , find the
magnetic flux through a cross section of the solenoid.

Solution

Application of formula (17-1-2) yields:

The magnetic flux through a cross section of the solenoid is

Example 17-1-2: For the solenoid of Fig. 17-3, show that the magnetic field
on the axis is given by the formula , where
is the number of turns per unit length. Then, assuming that , show
that , (formula (17-1-2)).
 627

Solution

Fig. 17-3: Alternative derivation for the magnetic field of a solenoid.

The magnetic field at points on the axis of a circular loop of radius and
carrying a current is given by equation (14-4-5), i.e.

We may consider that the solenoid consists of turns of wire, closely

wound on a cylinder of radius , so that the number of turns in an elementary
length is . It follows that the total field at the point is
(using superposition),

To evaluate this integral we use the substitution , the

result being, (let the reader check it),
 628

From Fig. 17-3, we see that and

, and finally, equation (**) yields:

For a long solenoid, , , and equation (****)

yields . This approximation, of course, is valid for points not close to
the bottom and the top surface), i.e. for point well inside the solenoid.

PROBLEMS

17-1-1) A solenoid with turns and length , carries a current

. Find the magnetic field inside the solenoid.

(Ans: ).

17-1-2) The magnetic field inside a long solenoid is . Given

that and , find the number of turns.

17-1-3) Show that the magnetic field of an ideal solenoid with

has only a component.

17-2) Toroids

As we did with cylindrical solenoids, we shall first consider an ideal toroid,

with current density at , as shown in Fig. 17-4.

Fig. 17-4: An ideal toroid.

The surface current density at the points is .

 629

For the ideal toroid in Fig. 17-4, it can be shown that the magnetic field is
given by the formula:

A practical toroid, consists of a toroid, around which a wire carrying a

current , is closely wrapped. In this case, assuming that there are turns, then
, and formula (17-2-1) leads to the following:

As it is the case for cylindrical solenoids, the number is the

number of turns per unit length. Notice that the field inside the toroid is
circumferential and of constant magnitude .

Example 17-2-1: A toroid of radius has turns and carries a

current . Find the magnetic field inside the toroid.

Solution

PROBLEMS

17-2-1) A toroid has: and turns. Find the

magnetic field inside and the magnetic flux through a cross section of the
toroid, assuming that the radius of the cross section is .

(Ans: ).
 630

CHAPTER 18: MAGNETIC FORCES AND TORQUES

18-1) Magnetic Force

It is a well known fact that electric charges, (positive or negative), in

motion, within a magnetic field, experience a force called “magnetic force”.
Also, current carrying wires, inside magnetic fields, experience forces, known
as “Laplace’s forces”. This is an extremely important fact, since the principle of
operation of electric motors is based exactly on these Laplace’s forces.

a) Magnetic force on a point charge moving inside a magnetic field.

Fig. 18-1: Magnetic force on a moving charge (Lorentz force).

Let us assume that a point charge moves inside a magnetic field with
velocity . It is found that there exists a magnetic force exerted on ,
known as the Lorentz force, given by the fundamental formula

The magnetic force is perpendicular to the plane defined by and (due

to the cross product), and its magnitude is . Notice that the
magnetic force depends on the sign of the charge .

If a charge is moving in a combined electric field and a magnetic field ,

then the total force on the particle is found by superposition, i.e.
 631

An important remark: Let us assume that a point charge , under the

action of a magnetic force , is displaced from a point to another point ,
along the path ( ), inside a magnetic field .

Fig. 18-2: Magnetic forces do no work.

If the kinetic energy of the particle at the point is and at the point
is , then the variation of the kinetic energy, when the point charge goes
from to , is:

However, since the cross product is perpendicular to the velocity

, at each point on the trajectory , the dot product is zero, and
from equation (*) it follows that

This shows that the kinetic energy of a charged particle cannot change as
the particle moves inside a static magnetic field, and this, in turn, means that
the speed of a particle, (i.e. the magnitude of the velocity ), cannot
change, it remains constant while the charged particle is moving within a
magnetic field. Of course, magnetic forces can change the direction of a
particle, but they cannot speed it up or slow it down, (change their speed).
 632

On the contrary, electric fields can change the kinetic energy of charged
particles, either by increasing or decreasing their speeds, and therefore
increasing or decreasing their kinetic energy.

In summary, electric forces do work, but magnetic forces do no work.

b) Magnetic forces on current carrying conductors (Laplace’s forces).

Exploiting the equivalency , (see equation

(13-4-14)), we may start with the Lorentz force on a point charge to find the
magnetic force exerted on current carrying conductors.

1) Filamentary conductors.

Let us assume that we have a filamentary conductor, carrying a steady

current within a magnetic field , as in Fig. 18-3.

Fig. 18-3: Magnetic force on a filamentary conductor.

Since each one of the moving electrons, inside the filamentary conductor,
experiences a magnetic force (a Lorentz force), and since within a current
carrying conductor there exists a huge number of conduction electrons, we
expect the filamentary conductor, as a whole, to experience a magnetic force.
Of course, within the conductor, there are also positive ions, but since they are
mach heavier than the electrons, we may assume that they are not moving,
and therefore they do not experience magnetic forces. So, the magnetic forces
 633

on current carrying filamentary conductors are due, primarily, to the motion of

the free, conduction electrons.

In Fig. 18-3, let be the charge passing through the cross section of the
filamentary conductor, at the point . Within a time interval , an amount of
charge , will pass through the cross section at , to the interior of
the infinitesimal cylinder. Assuming that in the neighborhood of the
infinitesimal cylinder, the drift velocity of the electrons is and the magnetic
field is , then the magnetic force on is

Formula (18-1-3) is a fundamental formula, expressing the magnetic force

on a differential length of a filamentary conductor, carrying a current ,
placed in a magnetic field . The magnetic force in (18-1-3) is known as “the
Laplace’s force”.

The total magnetic force on a finite segment of a filamentary conductor of

length , is found by superposition, i.e.

The total magnetic force on a closed loop, carrying a current , is

In practice, it is quite difficult to evaluate the integrals in formulas (18-1-4)

and (18-1-5), since in the general case, both and may be functions of the
position.
 634

2) Surface currents .

Fig. 18-4: Magnetic force on a current sheet.

Assuming now that we have a surface current density , specified on a

surface ( ), the magnetic force on is obtained from formula (18-1-3), if
is replaced by , i.e.

The total magnetic force exerted on a finite surface , is found by

superposition, i.e.

3) Volume currents .

Fig. 18-5: Magnetic force on a volume current density.

 635

If a current density is specified at all points within a volume , then the

magnetic force on a differential volume is found from (18-1-6), if is
replaced by , i.e.

The total magnetic force exerted on a finite volume is found by

superposition, i.e.

If a differential volume exists within a combined electric field and a

magnetic field , then, assuming that is the electric charge within , the
electromagnetic force exerted on , will be

or, since , ( being the volume charge density at the position of

the infinitesimal volume ),

It is convenient to define the “volume force density” in accordance with

the equation

Equation (18-1-11) gives the total volume force density at a point in space,
in terms of and at this point.

The right hand rule for the Laplace force: The Laplace force on a
differential element is, as know, , i.e. is the cross product of
and . This allows us to picture the direction of the magnetic force in space,
using the following rule; if we grab the wire with our right hand, so that the
 636

pointer finger is in the direction of the current and the middle finger is in
the direction of the magnetic field , then the thumb points in the direction
of the magnetic force , (middle finger , pointer finger , thumb ). In
mathematical terms, the system forms a right handed system.

Equation (18-1-10) is the starting point for the study of electrically

conducting fluids, such as plasma or liquid metals. This subject is known as
“Magneto hydrodynamics”.

An important application of formula (18-1-10), namely the “Hall Effect” will

be studied in Example 18-1-3.

c) Magnetic force between two current carrying closed loops, (Ampere’s

law for the magnetic force).

Fig. 18-6: Magnetic force between two current carrying closed loops.

Let us consider two closed loops, carrying a current and carrying

a current , as in Fig. 18-6. The current produces in the surrounding space a
magnetic field . The second loop, carrying a current , experiences Laplace
forces, due to the magnetic field , generated by . We want to find the total
magnetic force exerted on loop , caused by the current flowing in
loop .

The magnetic force on a differential element , on loop is:

 637

The magnetic field , at the position of , due to the current in loop

is, according to the Biot-Savart law and the superposition principle,

From equations (*) and (**) we find,

The total magnetic force is found by integration (superposition) of

around the loop , i.e.

Formula (18-1-12) is known as “the Ampere’s law for the magnetic force
between two, current carrying closed loops”. It follows that the magnetic
force is proportional to the product of the two currents and also, it
depends on the relative disposition of the two loops, i.e. on the geometry of
the configuration.

It can be shown (see Problem 18-1-1) that formula (18-1-12) can take the
equivalent form:

If we call the total magnetic force exerted on loop due to the

current in loop , then, as it follows from (18-1-13), , as it
should, of course, be, according to Newton’s third law, (Action=Reaction), (see
Problem 18-1-2).
 638

Example 18-1-1: Using the expression of the Laplace force, explain the
classic experiment in Electromagnetism: Two parallel, filamentary conductors
carrying currents in the same direction attract, while, when carrying currents in
the opposite direction repel.

Solution

Fig. 18-7: Magnetic force between two parallel, filamentary conductors.

In Fig. 18-7 (a) the currents flow in the same direction, (both flow up). The
magnetic field produced by the current , in the vicinity of the conductor
, will be in the direction , (from the reader towards the page), according to
the right hand rule. The element , (the force on due
to the magnetic field ), is directed from the right towards the left, and
therefore, the resultant force will have the direction shown in Fig. 18-7 (a).
Reasoning similarly, the force (the force on the conductor due to the
magnetic field , produced by the current ), is directed from the left
towards the right. The two conductors attract.

Similarly, in Fig. 18-7 (b), we can show that the two conductors repel.

Example 18-1-2: Show that the total magnetic force on a current carrying
closed loop (filamentary conductor), placed in a uniform magnetic field , is
zero.
 639

Solution

The total magnetic force on the closed loop is: , and since
is uniform, by assumption,

since . Notice that , but , the length

of the loop, (think about this).

Example 18-1-3: (The Hall Effect).

It has been experimentally observed, that, when a material, conductor or

semiconductor, with a current flowing through, is placed within a magnetic
field , perpendicular to , then an electric field is produced, perpendicularly
to both and . As a result, a potential difference (voltage) is produced.
This phenomenon was discovered by the American Physicist Edwin Hall, in
1879, and since then it bears its name. The induced voltage is called, “the
Hall Voltage”. Making use of Lorentz force, explain the Hall Effect.

Fig. 18-8: The Hall Effect.

In order to understand the Physics behind the Hall Effect, let us consider a
conductor in the form of a parallelepiped, as in Fig. 18-8.
 640

A current flows in the positive direction, in the presence of an applied

magnetic field , pointing in the positive direction.

a) Assuming that the parallelepiped is a metallic conductor, we know that

the current is due to the motion of “free or conduction electrons”, with a
drift velocity in the negative direction, as shown in Fig. 18-8 (b). The Lorentz
force on the electrons, , points in the positive direction, since
the charge of the electron is negative. As a result, the electrons are
deflected to the right and soon, there will be a concentration of negative
charge on the side . The side becomes negatively charged while side will
be positively charged. It follows that an electric field is produced, pointing
from the positive side towards the negative side , as shown in Fig. 18-8 (b).

Each electron now, experiences two forces, the electric force and
the magnetic force . The deflection of the electrons will stop,
when the electric force balances the magnetic force, i.e. when

or, since ,

If we call the current density, then , ( being the charge density),

and the current , in terms of is (see Fig. 18-8),

Formula (**) is important, since can be measured, and then

from formula (**), the charge density can be determined.

b) A similar analysis holds true if the current is due to the motion of

positively charged particles, for instance “holes” in a “p-type” semiconductor.
The situation is depicted in Fig. 18-8 (a). In this case ( positive), positive ions
are deflected to the side , which becomes positively charged, while side
becomes negatively charged. The electric field now, points from the side
towards the side .
 641

c) The Hall Effect can be used to determine whether a semiconductor is “p-

type” (majority carriers are holes), or “n-type” (majority carriers are
electrons). If is positive, then the semiconductor is “p-type”,
while if is negative, then the semiconductor is “n-type”, (see
Fig. 18-8 (a) and (b)).

Example 18-1-4: A current density exists in a uniform

magnetic field . Find the total magnetic force on the unit cube
.

Solution

Application of formula (18-1-9) yields:

Example 18-1-5: A surface current density ( ) flows over the

plane , in the presence of a uniform magnetic field . Find the
force per unit area exerted on the current density.

Solution

Application of formula (18-1-6) yields:

PROBLEMS

18-1-1) Starting with formula (18-1-12) show formula (18-1-13).

Hint: Use the vector identity: .

 642

18-1-2) Starting with formula (18-1-13) verify that .

18-1-3) A point charge is moving within a combined electric field

and a magnetic field . At some time instant , the velocity of the
particle is . Find the force acting on the particle, at the
time instant .

(Ans: ).

18-1-4) A cube of side is placed within a uniform magnetic field

. A current flows in the conductor . Find the
magnetic forces exerted on the sides and .

Fig. 18-9: Magnetic forces on filamentary conductors.

18-1-5) a) A current density exists in a uniform

magnetic field . Find the magnetic force exerted on the cube
, where is the side of the cube, b) For
which the magnetic force on the cube is zero?

(Ans: a) , b) ).

18-1-6) A magnetic field exists in free space. Find

the net magnetic force on the rectangular loop , ,
lying on the plane and carrying a current in the counter
clockwise direction.
 643

18-1-7) Find the force of attraction per unit length between two filamentary
conductors, with currents at and at .

(Ans: ).

18-1-8) A current carrying conductor, joins the points and

. A current flows from towards . If a uniform
magnetic field is applied, find the force exerted on
the conductor.

Hint: , and the

magnetic force is .

18-2) Magnetic Torque

a) Let us consider a force applied at a point , whose position with

respect to a coordinate system is determined by its radius vector .

Fig. 18-10: Torque of a force.

From Mechanics, we know that the torque of the force with respect to
the origin, is a vector expressed by the cross product

The vector is sometimes called “the lever arm” of the force .

If we consider a system of forces with lever arms

respectively, then the total torque of the system of forces with
respect to the origin is:
 644

b) Magnetic torque on a current carrying closed loop, inside a magnetic

field.

Let us consider a wire in the form of a small, planar square loop, of side ,
lying on the plane, centered at the origin and carrying a current , as in Fig.
18-11.

Fig. 18-11: Torque on a current carrying square loop.

We further assume that a magnetic field exists in the

neighbourhood of the square loop. The four sides of the loop, carrying the
current , experience Laplace forces, according to formula (18-1-3), i.e.:

We notice that the total magnetic force on the loop is zero, meaning that
the loop does not undergo any translation; however, the total magnetic
torque on the loop is not zero. The two forces and form a “couple of
 645

forces”, under the action of which the current carrying square loop tends to
rotate about the axis, as shown in Fig. 18-11.

The total magnetic torque of the couple of forces and according

to (18-2-1) and (18-2-2) is:

Taking into account formula (15-4-2), the magnetic dipole moment of the
square loop is , and equation (**) becomes:

Formula (18-2-3) expresses the magnetic torque exerted on a dipole

moment , in the presence of a magnetic field .

Some general remarks: A magnetic dipole when placed in a magnetic

field experiences a torque (due to Laplace’s forces), given by .
Under the action of this magnetic torque, the magnetic dipole will rotate, until
becomes parallel to , in which case becomes zero.

So far, we have assumed that the magnetic field is uniform. However, in the
general case, may be a function of the spatial coordinates, i.e. .
However, if we assume that the loop is very small, i.e. it is an infinitesimal loop
with magnetic dipole moment , then, in the vicinity of the loop, we may
consider the magnetic field to be uniform, and in this case the differential
magnetic torque exerted on is, (according to formula (18-2-3)),

A position of the dipole, at which is parallel to , is an equilibrium

position, since then . As it was the case with the equilibrium of electric
 646

dipoles, (see section 10-2), when and are parallel and of the same
direction, the equilibrium is stable; If and are parallel and of opposite
direction ( and are antiparallel vectors), the equilibrium is unstable. This is
shown in Fig. 18-12.

Fig. 18-12: Equilibrium of a magnetic dipole in a magnetic field.

The situation is analogues to the situation of electric dipoles, placed within

electric fields (see section 10-2).

Electric dipoles are used to develop the theory of electric fields inside
matter. Likewise, magnetic dipoles are used to develop the theory of static
magnetic fields inside matter, (Chapter 20).

Example 18-2-1: Find the magnetic torque on a dipole

when placed in a uniform magnetic field .

Solution

Application of formula (18-2-3) yields:

 647

Example 18-2-2: A wire in the form of a rectangle carries a current

. Find the torque on the wire when a uniform magnetic field
is applied.

Solution

Fig. 18-13: Torque on a current carrying rectangle.

The magnetic moment of the loop is

(for the indicated direction of the current flow).

The torque acting on the dipole is:

Example 18-2-3: For the orbiting electron of Fig. 15-8, in Example 15-4-5,
find the torque on the atom, if a magnetic field is
applied.

Solution

In Example 15-4-5, we found that the dipole moment of the orbiting

electron is . The torque on the atom when the
magnetic field is applied, is
 648

Example 18-2-4: (The ideal electrical motor)

The exploitation of magnetic forces on current carrying conductors within

magnetic fields, finds many applications, of great practical importance. One of
the most important applications is the electrical motor, (E.M.). In an E.M. the
magnetic forces on current carrying conductors, produce a torque which
rotates a cylindrical drum (the rotor) about an axis, delivering thus power to
the external world. In other words, an E.M. converts electric energy to
mechanical energy.

In this example we shall consider the principles of operation of an ideal

motor.

Fig. 18-14: The ideal motor.

Let us consider a conducting shell of radius and length , as in Fig. 18-14.

On the cylindrical surface , there exists a surface current density
. Also, by means of an external current distribution, we have
managed to produce a magnetic field , at the cylindrical surface, which,
approximately, is given by the expression .
 649

The current , flowing in the strip of height and width , in the

direction, is

The magnetic force exerted on the current carrying strip is

where is the local unit vector, in the direction, at the midpoint of the
strip. The torque of with respect to the axis is

and the total torque, rotating the cylindrical shell about the axis is

If, under the action of the torque , the cylinder revolves about the axis
with angular frequency , then the mechanical power delivered by
the motor is

Example 18-2-5: A solenoid consisting of circular turns of radius , carries

a current , as in Fig. 18-15. Find the torque on the solenoid, when a magnetic
field is applied.

Solution

We may consider that the solenoid consists of circular turns, stacked on

each other, and since each turn behaves as a magnetic dipole of moment
, the moment of the solenoid will be
 650

Fig. 18-15: Torque on a solenoid.

The torque exerted on the solenoid is

PROBLEMS

18-2-1) For the ideal motor of Example 18-2-4, it is given: ,

, and . a) Find the total torque on the
cylinder, b) If the cylinder rotates at , (rotations per minute), find the
power delivered.

(Ans: , ).

Hint: .

18-2-2) A conductor of length is oriented along the axis and carries a

current in the positive direction. What is the energy needed to rotate the
conductor by , at a constant speed, assuming that the magnetic field is
?

Fig. 18-16: Energy required rotating a conductor inside a magnetic field.

(Ans: ).
 651

18-2-3) A small, planar circular loop of radius , lies a distance above an

infinite current sheet , flowing on the plane. A current flows in
the circular loop. a) Find the total force and the total torque on the loop,
assuming that the loop is parallel to the plane, b) Repeat if loop is
parallel to the plane).

(Ans: a) , b) (depending on the

direction of the current ).

Hint: The magnetic field of the current sheet is uniform (formula (14-4-3)).

18-2-4) The current carrying conductor in Fig. 18-17, carries a current

in the positive direction. In the presence of a magnetic field
, find the power required for the conductor to rotate about the
axis with , (the lower end of the conductor traces the circle
).

Fig. 18-17: Rotation of a conductor in a magnetic field.

(Ans: ).

18-2-5) Two small loops are positioned as shown in Figure 18-18. Loop 1
carries a current and has an area , while loop 2 carries a current and has
an area . Find the torque on loop 2 due to the loop 1. Assume that the
separation distance is large as compared with the dimensions of the loops.

(Ans: ).
 652

Fig. 18-18: Torque on a magnetic dipole due to another dipole.

Hint: To find the magnetic field at the vicinity of the loop 2, due to
, use formula (15-4-4) with and . The sought for
torque on loop 2, is .

SUPPLEMENTARY PROBLEMS

18-1) For a conductor in the form of a parallelepiped as in Fig. 18-8, it is

given: . Find the Hall
Voltage , (Ans: ).

18-2) The solenoid in Fig. 18-15 has turns, each of radius ,

and carries a current . If a magnetic field is
applied, find the torque on the solenoid.

Hint: See Example 18-2-5.

18-3) A loop, placed at the origin, carries a current . A second

loop , carrying a current is situated at the point
. Both loops are oriented with their axes in the direction. Find
the torque exerted on the second loop. Assume that both currents flow in the
counter clockwise direction.

(Ans: ).

18-4) A conductor of differential length is placed at the point

and carries a current in the direction . Find the
force on the conductor , due to another infinitely long filamentary
 653

conductor, lying on the plane, passing through the origin, forming an angle
with the axis and carrying a current in the direction of increasing .

Hint: To find the magnetic field at , due to the infinite

filamentary conductor, use formula (14-4-2).

18-5) In the preceding Problem, find the torque on a dipole

placed at the point .

18-6) A circular loop of radius , lying on the plane and centered at the
origin, carries a current in the counter clockwise direction. If a magnetic field
is applied, find the force and the torque exerted on the
loop.

18-7) A surface current density flows on the plane and a

current density flows on the plane . Find the force per unit
area, between the two planes. Is the force attractive or repulsive?

(Ans: , Attractive force).

18-8) An infinite filamentary conductor, carrying a current , is parallel to a

surface current density , as in Fig. 18-19. Find the force per unit
length, exerted on the filamentary conductor.

Fig. 18-19: Force per unit length on the filamentary conductor.

 654

(Ans: ).

Hint: The magnetic field produced by the surface current density is given by
equation (14-4-3).

18-9) A filamentary conductor coincident with the axis, carries a current

in the positive direction. If a magnetic field is
applied, the force per unit length exerted on the filamentary conductor is
. Find the magnetic field, (Ans: ).

18-10) For the ideal motor of Example 18-2-4, Fig. 18-14, it is given:
, , and . a) Find the total torque
on the cylinder, b) If the cylinder rotates at , (rotations per minute),
find the power delivered.

18-11) A charge uniformly distributed over a spherical surface of radius

is set spinning at an angular velocity . a) Use the expression for the
vector magnetic potential in formula (****) in Example 15-2-5, find
the magnetic field outside the spinning sphere, (the field inside is uniform,
as found in Example 15-2-5), b) Find the force of attraction between the upper
and the lower hemispheres of the spinning spherical surface.

(Ans: ).

Hint: If is the magnetic field outside the spherical surface, the force on a
differential surface of the upper hemisphere is , where
, , (see Example 15-2-5). Does the uniform
magnetic field inside the spherical shell contribute to the attractive force?
 655

CHAPTER 19: MOTION OF CHARGED PARTICLES IN ELECTRIC AND

MAGNETIC FIELDS

19-1) Introduction

The motion of charged particles, (electrons or protons), in electric and

magnetic fields, is a subject of great theoretical and practical importance, with
many applications.

In studying the motion of charged particles, we shall assume that the speed
of the particles is much less than the speed of light , (i.e. , non-
relativistic particles), and as a consequence, we shall consider that the mass
of the particles remains constant during their motion.

If we, therefore, consider a charged particle of mass and charge , in

motion within an electric field and a magnetic field , then the force exerted
on the particle is given by the “Lorentz formula”

According to the Newton’s second law of motion, the net force on a particle
is given by the formula

where is the acceleration of the particle.

From these two equations it follows that

Equation (19-1-3) is a vector differential equation for the velocity of

the particle. To solve this equation for we need one initial condition,
. Once has been found, then, since , we
may solve for the trajectory of the particle.
 656

In the most general case, the electric field and the magnetic field may
be functions of the spatial coordinates. However, in this case, the solution of
the differential equation in (19-1-3) could be quite complicated.

To avoid complicated mathematical calculations, at this introductory level,

we shall assume that both fields, and are uniform. We start our analysis,
assuming first that the particle is moving inside an electric field alone, (no
magnetic field present); then we shall study the motion of the particle in a
magnetic field alone (no electric field present), and finally we shall study the
motion of a particle in a combined electric and magnetic field.

19-2) Motion of a Charged Particle in a Uniform Electric Field

Fig. 19-1: Motion of a particle in a uniform electric field.

Let us assume that at the time instant , when the velocity of a charged
particle of mass and charge is , a uniform electric field is
applied. The force acting on the particle is the electric force .

The equation of the motion of the particle is:

Since is a constant vector (since was assumed to be uniform), the

solution of the differential equation in (19-2-1) is found easily to be:
 657

If is the vector position of the particle at the time instant , then, as we

know, , and equation (19-2-2) becomes,

The solution of equation (19-2-3) is:

where is the initial position of the particle.

If, for simplicity, we assume that at the time instant , the particle is at
the origin ( ) with zero initial velocity, ( ), then formulas (19-2-2)
and (19-2-4) are simplified to the following:

19-3) Motion of a Charged Particle in a Uniform Magnetic Field

a) Let us assume that a charged particle is moving within a uniform

magnetic field with velocity . Then a magnetic force (the Lorentz force)
acts on the particle. The equation of motion, in this case, is:

As we did in section 19-2, we assume that , the initial velocity

of the particle is known.

We may resolve the velocity of the particle in two components, being

parallel to the magnetic field and being perpendicular to , as shown in
Fig. 19-2.
 658

Fig. 19-2: Resolution of into two components, and .

Since , equation (19-3-1) becomes:

since , (parallel vectors), and equation (19-3-2) yields:

From the first equation in (19-3-3) it follows that the parallel component
of the velocity does not vary with time, i.e. remains constant. The magnetic
field has no effect on the component of the velocity which is parallel to the
magnetic field. Only the vertical component of the velocity does vary with
time, in accordance to the second equation in (19-3-3).

It suffices, therefore, to study the motion of a charged particle, within a

uniform magnetic field , with its initial velocity being perpendicular to the
magnetic field, ( ).

b) Let us, for definiteness, consider a charged particle entering a magnetic

field with initial velocity , as shown in Fig. 19-3.

If is the velocity of the particle, at any subsequent time instant , then

the equation of motion of the particle is:
 659

or, since ,

Fig. 19-3: Motion of a particle with velocity perpendicular to .

In equation (19-3-4) we have a system of two (coupled) differential

equations. Differentiating the first equation in (19-3-4), with respect to time ,
and taking into account the second equation, yields:

We define the “cyclotron frequency ”, sometimes also referred to as

“the Larmor frequency” as
 660

In terms of the cyclotron frequency, equation (*) is written as

The general solution of the differential equation in (**) is:

where and are the arbitrary constants of integration.

From the first equation in (19-3-4) we find that , which by

virtue of (***) yields:

Applying the initial conditions, and , we find

and , and finally,

(For detailed calculations see Problem 19-3-1).

Remark: Note that at any instant of time , the speed of the particle is

which, in turn, means that the kinetic energy of the particle remains
constant during its motion. We expected this, since as we know, magnetic
forces do no work.

Having determined and , we may now find the trajectory of the

particle. Let us assume that at the particle is at the origin, i.e.
and . The trajectory of the particle is found from the following set of
equations:
 661

The solution of the system in (19-3-7) is, (see Problem 19-3-2).

Eliminating between the two equations in (19-3-8) results in the following:

Equation (19-3-9), in the plane, represents a circle, with center at

and radius .

Fig. 19-4: Circular trajectory of a charged particle in a magnetic field.

c) So far we have studied the motion of a charged particle in a uniform

magnetic field , under the assumption that the velocity of the particle is
perpendicular to the magnetic field.
 662

We have shown that the trajectory of the particle is a circle.

Let us now assume that the velocity of the particle is not perpendicular to
the magnetic field . The velocity may be decomposed into one component
parallel to and another component perpendicular to .

The component remains unaffected by the magnetic field, as explained

in part (a), while the vertical component makes the particle to move on a
circular trajectory. In other words, the particle now participates in two
motions, one translational motion with velocity and, at the same time, in a
circular motion. The trajectory of the particle is a “helical curve” as shown in
Fig. 19-5.

Fig. 19-5: Helical trajectory of a charged particle.

d) The radius of the circular motion is (from equation (19-3-9)), ,

or more generally,

where is the component of the velocity, perpendicular to the magnetic

field . The period of the circular motion is

It follows that the period of the circular orbit is independent from the
initial speed of the particle. If a number of identical particles, say electrons,
 663

enter a magnetic field, with different initial speeds, all of them will execute
circular motion with the same period .

e) Assume that a charged particle, say an electron, of mass and charge

, enters in a uniform magnetic field , with velocity . As we know, the
circular motion of the electron, due to , the component of the velocity
perpendicular to the magnetic field, will have a period of ,
(equation (19-3-11)). The circular motion of the orbiting electron, can be
considered as magnetic dipole , whose magnitude is , where is
the radius of the circular orbit and , i.e.

which, by virtue of equations (19-3-10) and (19-3-11), is written

equivalently as,

where is the kinetic energy of the electron, associated with

its motion perpendicularly to the magnetic field .

f) Assuming that an electron enters a magnetic field with its velocity

perpendicular to , the magnetic flux through the circular orbit, (of radius
), of the charged particle, is , and taking into account (19-3-11)
and (19-3-12) we find:

g) Electron Volt (eV) - The unit of energy in the atomic scale.

As we know, the unit of energy in the S.I system of units is the Joule.
However, when working with electrons, protons, atoms, etc, Joule is a very
large number. For this reason, in Atomic Physics, we prefer, and use another,
more convenient unit for the energy of the involved particles, called “electron
volt”, (eV), and defined as follows:
 664

The energy acquired by an electron ( ) when falling

through a potential difference , is by definition equal to .

Fig. 19-6: Definition of the electron volt (eV).

In Fig. 19-6, the electron, when moving from point A to point B, within the
electric field , gains energy, from the field, equal to .

The relation between Joule and electron volt is obtained from the definition
of the electron volt:

In Fig. 19-6, assuming that the electron starts from rest at the point A, the
energy gained from the field, when it reaches the point B has been converted
to kinetic energy, i.e.

It is thus obvious that a voltage of imparts a large velocity to the

electron. Still, this speed is a non relativistic speed, since ,
( is the speed of light in free space).

Example 19-3-1: An electron has a velocity normal to a

magnetic field . Calculate the radius of the circular orbit, the period,
and the cyclotron frequency of the electron.

Solution

Application of formula (19-3-5) yields:

 665

The radius of the circular orbit is obtained from formula (19-3-10),

The period is given from equation (19-3-1), i.e.

Example 19-3-2: An electron is at rest in a field-free region. If a uniform

electric field of magnitude is applied for , find the speed
of the particle. Then, the electron enters a magnetic field , with its
velocity perpendicular to the magnetic field. Determine the radius of the
circular orbit.

Solution

The acceleration of the electron is , and its speed

acquired in is

From equation (19-3-10), , where , (eq. 19-3-5),

we find:

Remark: Another way to obtain the expression for the radius is the
following: The magnetic force acting on the electron ) is actually
the centripetal force, causing the circular motion, i.e.
 666

Example 19-3-3: An electron enters the region between the two plates of a
capacitor with velocity , as shown in Fig. 19.7. Assuming that is the length of
the plates, is their separation distance and is the voltage applied between
the plates, find the value of so that the electron will fall on the lower plate.

Solution

Let and be the charge and the mass of the electron.

Fig. 19-7: Motion of an electron in the electric field of a capacitor.

In general, depending on the voltage , there are three cases: the electron
hits the upper plate (trajectory ( )), or escapes the capacitor (trajectory ( )), or
it falls back on the lower plate (trajectory ( )).

When the electron enters the capacitor, experiences an electric force

, and since the charge is negative, the electric force points in the
negative direction, (we neglect the weight of the electron, since
).

a) Motion along the axis:

Since the net force in the direction is zero, the particle executes
rectilinear motion, with constant speed . The equation of motion
of the electron, in the direction is:
 667

Motion along the axis:

The net force in the direction is directed in the negative direction, and is
equal to , where is the absolute value of the charge
of the electron. Since the force is constant, the acceleration is also
constant. The motion of the electron, in the direction, is rectilinear with
constant acceleration and initial speed , and therefore the equations of
motion are:

The maximum height attained is obtained as follows: At the time

instant at which is attained, the vertical component of the velocity
must be zero, and from the first equation in (**) we find:

and substituting this value in the second equation in (**) yields,

or, since ,

To find the maximum range , in the direction, we need to find the time
it takes for the electron to travel from the origin up to the distance , (see
Fig. 19-7). When , the corresponding is zero, and from the second
equation in (**) we find:

and substituting in (*) we obtain,

 668

Conclusion: The electron will fall back on the lower plate, provided that the
following two inequalities hold true simultaneously:

The voltage must satisfy simultaneously both equations in (*****), for

the electron to fall back on the lower plate.

PROBLEMS

19-3-1) Starting with equations (***) and (****), apply the initial conditions
and to derive equation (19-3-6).

19-3-2) Show that the solution of the differential system in (19-3-7) is given
by the equations in (19-3-8).

19-3-3) Give a physical interpretation of the minus sign, in the vector

definition of the cyclotron frequency , in equation (19-3-5).

19-3-4) In Example 19-3-3, it is given: , ,

, and . What is the least value of the applied voltage , for
the electron to fall on the lower plate.
 669

19-4-5) In the Hydrogen atom, the radius of the electronic orbit is

,( ). The orbiting electron produces a magnetic field at
the center . What is the electron’s angular frequency of revolution?

(Ans: ).

19-4-6) An electron is accelerated through a potential difference of

and then enters at right angle a uniform magnetic field . What is the
radius of the circular orbit? What is the magnetic flux through the circular
orbit? The orbiting electron is equivalent to a magnetic dipole. Find the
magnetic dipole moment.

19-4-7) A electron is injected into a magnetic field , at an

angle of with respect to . As we have explained, the electron’s trajectory
is a helical path, (Fig. 19-5). Calculate the cyclotron frequency , the radius ,
the period and the distance between two adjacent turns of the helix.

(Ans: ).

19-4) Motion of a Charged Particle in Perpendicular, Static, Uniform

Electric and Magnetic Fields

So far, we have studied the motion of charged particles in static electric

fields alone or in static magnetic fields alone. In this section we shall study the
motion of a charged particle in a region in space, where there are two static
fields; an electric field and a magnetic field , perpendicular to . We
assume that both fields are static and uniform.

Let us, for definiteness, assume that a particle of mass and charge is
released from rest at the origin, in a region in space, where there are two
fields, an electric field and a magnetic field , as shown in Figure
19-8.

Before we start solving the equations of motion, let us try, qualitatively, to

understand the behaviour of the particle. Since the particle is released from
rest, the initial motion of the particle will be in the direction of the electric field
. As the particle gains speed, in the direction of , the particle will start
“feeling” a magnetic force , and therefore will start moving in the
 670

direction. Eventually, the particle will be forced by the magnetic field to return
to the axis, and at this instant of time, the speed of the particle will be zero.
The whole process starts over again. The path of the motion is a curve called
“cycloidal”.

Fig. 19-8: A charged particle in crossed electric and magnetic fields.

Let us now study the motion of the particle, in quantitative terms.

If is the velocity of the particle at time , then the equation of motion

of the particle is:

We shall assume that the solution (for ), in equation (*), can be

expressed as

where is to be determined shortly.

It is easy to verify that the unit of the term is , (see

Problem 19-4-1). This shows that the term represents a constant
velocity, which we shall call “drift velocity ”. For the field configuration in
Fig. 19-8,
 671

Substituting the assumed expression for the solution in (**), in equation (*),
results in the following:

and since , (see Problem 19-4-2), it follows that

Equation (19-4-2) is identical to equation (19-3-1), which was solved in

section 19-3. As we found, equation (19-3-1) represents a circular or a helical
motion. Taking into account, also, equation (**) we conclude that

Equation (19-4-3) shows that the motion of the particle in the crossed
electric and magnetic fields is the superposition of two motions; one circular
or helical motion and a drift of constant velocity in the direction
perpendicular to both fields. Note that the drift velocity does not depend on
the sign of the charge and the mass of the particle. Electrons and protons,
placed in the crossed field of Fig. 19-8, attain the same drift velocity. However,
the circular motion, does depend on and .

The trajectory of a positively charged particle, released from rest at the

origin, is shown in Fig. 19-9. This trajectory is called “a cycloidal”.

(A cycloidal is the curve described by a fixed point on the circumference of a

wheel, as the wheel rolls along a straight line, without slipping).
 672

Fig. 19-9: Cycloidal trajectory of a charged particle.

Example 19-4-1: What is the drift velocity of an electron in the crossed

fields and ? Show that the drift
velocity is perpendicular to both and .

Solution

The drift velocity, in general, is given by the formula .

To show that is perpendicular to , it suffices to show that .

Indeed,

Similarly, we show that , (let the reader check it).

PROBLEMS

19-4-1) Verify that the unit of is .

19-4-2) Show that , provided that and are

perpendicular vectors.
 673

Hint: You may use the vector identity .

19-4-3) A charged particle of charge and mass , is released from rest at

the origin, in the crossed fields and . Find the equation of the
trajectory described by the particle. What is the drift velocity ?

(Ans: , where the cyclotron frequency

, , ).

19-5) An Important Case of a Non Uniform Magnetic Field - The Magnetic

Bottle Configuration

So far, we have studied the motion of charged particles in static field

configurations, assuming that all fields involved are uniform. Of course, this
assumption simplifies the mathematics, but, on the other hand, the most
important applications involve magnetic fields that do vary with the spatial
coordinates, i.e. involve non uniform magnetic fields.

One of the most important applications is the trapping of charged particles

in a finite region in space, which may then lead to the containment of “hot
plasma” in the region, long enough for thermonuclear fusion to occur. Our
hopes for cheap and clean production of electric energy, rest on this simple
idea.

As shown in section 19-4, crossed uniform fields and is not a proper

configuration for plasma containment, since all charged particles, electrons
and protons, will drift with the same velocity , and eventually they will
escape. Only static, non uniform magnetic fields (of special configuration) are
suitable for particle containment.

However, we may resort to the simple cases studied in the preceding

sections, in order to develop a qualitative understanding of the behaviour of
charged particles in non uniform fields.

Let us consider the magnetic field in Fig. 19-10. Obviously, this is not a
uniform field. At the position AA the field is weak, (sparse field lines), while at
the positions KK and MM the field is strong (dense field lines). To study
analytically the behaviour of a charged particle in this magnetic field would be
 674

extremely difficult, or even impossible. However, using our so far experience,

we may draw some general, very useful conclusions.

Fig. 19-10: Magnetic bottle configuration.

To start our analysis, we assume that there exists a strong, uniform

magnetic field in the direction, plus small perturbation fields
superimposed. The total magnetic field will be,
 675

. The perturbation fields are so chosen, as to

produce the field lines in Fig. 19-10 (a). This important, for practical
applications configuration is known as “the magnetic bottle” configuration, for
reasons to be explained shortly. Briefly, the magnetic bottle configuration is a
suitable configuration for trapping charged particles in the region between KK
and MM. It is a useful configuration for plasma containment.

Assuming cylindrical symmetry, and . The

fundamental law , implies:

and since must remain finite at , i.e. on the axis, we must take
, and therefore,

The force acting on a charged particle will be:

The component of the force is

Let us now consider the motion of one particle (assume positive). Since
the perturbation fields are small as compared to the strong field, we
expect the trajectory of the particle (circle) to lie (approximately) on planes
perpendicular to the axis, as shown in Fig. 19-11.
 676

What is of practical interest, is the time average of the force over

one period of gyration, (we use the symbol to denote the average of a
quantity).

Fig. 19-11: Trajectory of a particle on a plane perpendicular to axis.

From Fig. 19-11, we have:

Using these two formulas, equation (19-5-1) becomes:

 677

Also, again from Fig. 19-11, we have:

and since the average of over one gyration is zero, .

Equation (***) now yields , and finally (19-5-2) becomes:

If is the cyclotron frequency of the orbiting particle,

and equation (19-5-3) is written as

where is the dipole moment of the

orbiting particle. (To avoid confusion, we shall use for the mass of
electrons and for the mass of protons).

Equation (19-5-4) shows that the magnetic force, always, points in the
direction of the weaker magnetic field, as shown in Fig. 19-10 (c). The field
configuration in Fig. 19-10 “reflects back” charged particles approaching either
end of the configuration, and hence the name “magnetic bottle”.

One final comment: Let us assume that the particle, under the action of
moves a distance , in the direction. If we call the energy associated with
the motion parallel to the axis and the energy associated with the motion
perpendicular to the axis, then

The total energy of the particle remains constant, since magnetic forces
do no work. This implies that
 678

From equation (19-3-13), , and the derivative of with

respect to is:

From equations (*****) and (19-5-5) it follows that in the magnetic bottle
configuration, we must necessarily have

which shows that the magnetic moment is independent of . The

magnetic moment is a constant of the motion.

The magnetic flux through the orbit of the particle is given by (19-3-14), i.e.

where is the mass of the particle. Since remains constant, the flux
through the orbit must remain constant. As the particle drifts towards regions
of increasing field strength, its orbit must get smaller so that the flux will
remain constant.

Example 19-5-1: A particle with initial velocity enters in the

magnetic field , where ,
and . What is the reflection point
? (Assume that the particle enters in the magnetic field at ).

Solution

We call , ( ), the component of the velocity (energy) parallel to the

magnetic field and , ( ), the component of the velocity (energy)
perpendicular to the magnetic field. At the reflection point, .

Since magnetic forces do no work,

 679

Also, as we know, the magnetic moment of the particle, does not change
with , i.e.

PROBLEMS

19-5-1) In Example 19-5-1, what is the radius of orbit, a) at the reflection

point, and b) at , (Assume that the particle is an electron).

(Ans: , ).

19-5-2) An electron with initial velocity ,

enters the magnetic field at the origin. Provided that the
average force on the particle, at the point of reflection , is
, find and at the origin. Given: and
.

(Ans: , ).

Hint: Use the fact that the magnetic dipole moment is constant, i.e.
and that at the reflection point , (see Ex. 19-5-1).
The expression for the average force is given in equation (19-5-4).
 680

SUPPLEMENTARY PROBLEMS

19-1) An electron in a magnetic field makes a complete revolution in

. What is the magnitude of the magnetic field?

(Ans: ).

19-2) An electron is injected in a crossed fields configuration, with

and . What is the drift velocity of the electron?
What is the displacement of the electron in the direction during one
complete revolution?

19-3) When an electron exists at the origin with velocity

, an electric field is applied. a) Find
the location where , b) The time it takes the electron to go from the
origin to the point where .

(Ans: ).

19-4) A particle with initial velocity enters, at , the

magnetic field , where ,
and . What is the reflection point ?

(Ans: ).

Hint: See Example 19-5-1.

19-5) At time , when a proton exists at the origin with initial velocity
, a magnetic field is applied. At
the proton is located at . Find .

(Ans: ).

19-6) An electron has a velocity normal to a magnetic

field . Calculate the radius of the circular orbit, the period, and the
cyclotron frequency of the electron.

19-7) When an electron exists at the origin with velocity a

magnetic field is applied. Find the number of gyrations the
 681

electron makes in the time it takes to move a distance of . (Assume

), (Ans: gyrations).

19-8) When an electron is at the origin with velocity , an

electric field and a magnetic field are
applied. What is the total displacement in 100 gyrations?

(Ans: ).

19-9) A proton is directed towards a second stationary proton with speed .

What should the minimum speed be for the protons to collide? Assume that
the proton is a particle of mass and charge , uniformly distributed within
a sphere of finite radius .

(Ans: ).

19-10) A electron enters a uniform magnetic field

with its velocity making an angle of with the vector . Find the period
and the radius of the helical trajectory.

19-11) In a magnetic field of , an electron circulates at the

speed of light. What are the radius of orbit and the energy of the electron?

(Ans: ).

19-12) The total force acting on an electron by an electric field of

and a magnetic field of is zero. Find the minimum electron
speed and draw the vectors and .

Hint: .
 682

CHAPTER 20: STATIC MAGNETIC FIELDS IN MATTER

20-1) Diamagnetic, Paramagnetic and Ferromagnetic Materials

In chapter 10 we studied how a static electric field, produced by a charge

distribution in free space, is affected by the presence of matter. The basic idea
was that, due to the polarization of the bound charges, inside the dielectric,
the total electric field inside matter is the superposition of the electric field
existing in free space plus the induced electric field due to the polarization
effects.

For linear, homogeneous and isotropic materials, the macroscopic electrical

properties of matter are taken into account by the relative dielectric constant,
(or relative permittivity), .

We shall adopt a similar approach in investigating the magnetic fields in the

presence of matter. Recall that in chapter 14 we have studied magnetic fields
in free space (vacuum), produced by various steady current distributions. In
free space (vacuum), by definition, there is no matter present. The magnetic
fields in vacuum are expressed in terms of the permeability of free space
(just like electric fields in vacuum are expressed in terms of the permittivity
of free space ).

As we know, magnetic charges or magnetic monopoles do not exist in

nature. This is exactly the physical meaning of the equation . Static
magnetic fields are generated by electric charges in motion, i.e. by steady
electric currents. This is the physical meaning of the equation .

Of course, if we take a permanent magnet, even though we notice that the

magnet does produce a magnetic field in the surrounding space, (deflects a
compass, for example), we do not notice, macroscopically, any electric
currents. This seems to contradict our previous statement that magnetic fields
are generated by electric currents.

However, if we could focus our attention on a very small piece of a magnet,

or any other piece of material, in the atomic scale, we would notice that inside
the material there are tiny currents, generated by:
 683

1) The motion of negatively charged electrons, orbiting around their

respective positive nuclei, and

2) The intrinsic angular momentum of the electrons (the spin of the

electrons). We should mention, at this point, that the spin of the electron is a
“Quantum Characteristic” which cannot be explained by Classical Physics. Even
though this is not entirely correct, we may consider the electron as a tiny
charged sphere, revolving about the axis, producing thus a current and hence
a magnetic field.

Our notion that the magnetic properties of matter are due to minute
currents within the matter, associated with the circular motion and the spin of
the electrons was firstly formulated by Ampere, and since then it has been
fully confirmed.

An electron revolving around the nucleus can be treated as a tiny magnetic

dipole with dipole moment , (see Example 15-4-5 and Fig. 15-8).
Each tiny magnetic dipole produces a minute magnetic field, (formula (15-4-4).
If we picture an elementary volume within a material, then, inside
we expect to have a huge number of tiny magnetic dipoles, randomly
oriented, due to the random thermal motion of the atoms. As a result, the
average magnetic moment within is approximately zero, and the
resulting macroscopic magnetic field is, likewise, zero. This explains why we do
not notice appreciable magnetic fields in most of the materials, (even though
minute magnetic fields do exist inside).

However, if an external magnetic field is applied, then, all the tiny dipole
moments (originally at random orientations), tend to align with the magnetic
field , (see section 18-2, Fig. 18-12). Even though the random, thermal
motion still exists, now there is certainly a prominent direction of orientation
of the tiny dipole moments. The material now becomes “magnetically
polarized” or, as we say, becomes “magnetized”. The total magnetic dipole
moment, within , now, is not zero. The material exhibits macroscopic
magnetic properties, which cease to exist once the applied field is removed.

In some materials the induced magnetic dipole moments are in the

direction of the applied field , and these materials are called “paramagnetic
materials”. In some other materials, the induced magnetic dipole moments are
 684

in a direction opposite to that of , and these materials are called

“diamagnetic materials”. Paramagnetism and diamagnetism cease to exist
when the applied field is removed, (the tiny magnetic moments return back to
their random thermal motion).

Finally, there are a few substances, called “ferromagnetic materials”, like

iron, nickel, and cobalt, together with some alloys, which when brought in the
vicinity of a permanent magnet are attracted strongly. Even if the magnet is
removed, i.e. if the external field is removed, the ferromagnetic material
retains its magnetic properties. Its behaviour is completely different from
paramagnets and diamagnets. The magnetization of a ferromagnet depends
not only from the present magnetic field but also, from the “magnetic past
history” of the material, during the magnetization process.

20-2) The Physical Origin of Paramagnetism, Diamagnetism and

Ferromagnetism

In the preceding section we mentioned that the magnetization of the

materials is originated in the tiny loop currents, inside the materials, or the
same, the minute magnetic dipole moments due:

1) To the orbital motion of the electrons around their positive nuclei and

2) To the spin of the electrons.

The magnetic dipole moment associated with the spin of the electrons has
been found to be . The plus and minus signs correspond
to the two spin states, “spin up” and “spin down”.

When a dipole moment (orbital moment or spin moment) is placed inside a

magnetic field , then, the dipole experiences a magnetic torque
(see equation (18-2-3)), which tends line the dipole moment parallel to the
field , (in this equilibrium position ).

a) Paramagnetic materials - The origin of paramagnetism.

In Quantum Mechanics, there is the famous “Pauli’s Exclusion Principle”.

According to this principle, electrons in atoms combine together in pairs with
opposite spin. This, in turn, implies that the magnetic moment per pair of
 685

electrons, due to spin, is practically zero. But, if the material has an odd
number of electrons in its atoms, then the pairing of electrons is incomplete,
and when the material is introduced in a magnetic field the “unpaired”
electrons tend, under the action of the torque , to align their
magnetic dipole moments due to the spin, with the external field . In this
case, the induced magnetism enhances the external field , and this is the
characteristic behaviour of paramagnetic materials.

Obviously, paramagnetism does depend on the temperature. The lower the

temperature, the more pronounced the paramagnetic properties of the
material will be, since lower temperature implies less random thermal motion
of the electrons.

Paramagnetic materials, in general, tend to move from weaker towards the

stronger part of the field. For this reason, if a piece of paramagnetic material
approaches either pole (North or South) of a strong magnet, it will be attracted
towards the magnet.

Aluminium, Platinum, Sodium, Lithium, Manganese, Tungsten, Oxygen, and

Calcium are some of the most known Paramagnetic materials.

b) Diamagnetic materials - The origin of diamagnetism.

Diamagnetism is due to the interaction of the orbital motion of the

electrons around their positive nuclei with an externally applied magnetic field.
When an orbiting electron is placed in a magnetic field, its speed changes, in a
way that the magnetic dipole moment, due to the orbital motion, increases in
a direction antiparallel to the applied field . This means that the magnetic
dipole moment of each atom increases a little, and these small increments are
all antiparallel to the applied field .

Indeed, let us consider an orbiting electron, of mass , in a circular motion

of radius and speed , around the positive nucleus, with no magnetic field
applied, ( ), as shown in Fig. 20-1 (a). The Coulomb’s force acts as a
centripetal force, and therefore
 686

If a magnetic field is now applied, (Fig. 20-1 (b)), there will be an

additional magnetic force applied on the electron, directed
towards the center of the orbit, since . The total force on the electron,
acting as a centripetal force, now is:

Fig. 20-1: The origin the diamagnetism.

We assume that in both cases, the radius of the orbit is the same and equal
to , since according to Bohr, the orbits are quantized to certain, well defined
values, i.e. the electron cannot move in an orbit of a little bigger or of a little
less radius, after the field is applied.

Subtracting equation (20-2-1) from (20-2-2) results in the following:

Equation (*) shows that , i.e. the speed of the particle increases
when a magnetic field is applied.
 687

The magnetic dipole moment with no magnetic field present ( )

(Fig. 20-1 (a)), is:

Similarly, in Fig. 20-1 (b), with magnetic field applied, ( ), we find:

and since , it follows that . The conclusion is that, when

a magnetic field is applied, the magnetic dipole moment of each atom,
increases a little, but in a direction opposite to the applied field .

The increment in the magnetic moment of the atom is

Assuming further that (justify this assumption using (*)), then,

, and finally equation (**) becomes:

Equation (20-2-5) shows clearly that the increments of the magnetic dipole
moments are all antiparallel to the applied field .

Equation (15-4-4) shows that the magnetic field , due to , is in a

direction opposite to the direction of the applied field .

We reach the same conclusion if the applied field is in a direction

opposite to the direction shown in Fig. 20-1 (b). Let the reader verify it.
 688

Let us now consider a small volume , containing a huge number of atoms.

With no magnetic field applied ( ), due to the random thermal motion of
the atoms, we expect the total magnetic dipole moment within to be zero.
However, when a field is applied, each one of the atoms picks a little
magnetic dipole moment, and all these dipole moments are antiparallel to the
applied field , or at least have a component antiparallel to .
Macroscopically, this results in a decrease of the magnetic field inside the
material. This mechanism explains the origin of diamagnetism. As it is the case
with paramagnetic materials, diamagnetism ceases to exist once the applied
magnetic field is removed.

Diamagnetism is a phenomenon affecting, of course, all atoms. But, since

usually, it is weaker than paramagnetism, it is mainly observed in atoms with
even number of electrons, where paramagnetism is not present (see part (a)).

If a diamagnetic material approaches either pole of a permanent magnet

it is repelled, (the opposite of what happens with paramagnetic materials).
This phenomenon was observed for the first time by Michael Faraday, when he
approached a piece of Bismuth on a permanent magnet.

Some of the most common diamagnetic materials are: Zinc, Copper, Silver,
Bismuth, Gold, Water, Glass, Wood, etc.

c) Ferromagnetic materials - The origin of Ferromagnetism.

There is a third kind of magnetism, known as Ferromagnetism. Strictly

speaking, Ferromagnetism can be analysed and understood only with the aid
of Quantum Mechanics. The main Ferromagnetic materials, Iron, Nickel and
Cobalt, when brought in the vicinity of a permanent magnet are attracted by
the magnet. Unlike paramagnets, ferromagnets retain their magnetic
properties even at high temperatures. For example, iron can be heated up to
before it loses its magnetic properties. Nickel loses its magnetic
properties when heated to over .

This phenomenon was observed by Marie Curie, and the temperature at

which a ferromagnetic material loses its magnetic properties is known as
“Curie temperature”.
 689

Like paramagnetism, ferromagnetism is associated with the interaction

between the spins of unpaired electons. The distinct feature of a
ferromagnetic material is that all the spins of the unpaired electrons are
aligned in the same direction. Why this happens cannot be explained by
Classical Physics; only Quantum mechanics offers a satisfactory explanation.
But, anyways, this is what happens. For instance, in an iron atom, there are five
unpaired electrons, the spin of which are all aligned. In addition, the field of
the aligned spins is so strong that influences its neighbouring atoms to align
their spins in the same direction with the original atom. It seems that there is
“a collective behaviour” of the atoms. Furthermore, it is found that this
collective behaviour is restricted to relatively small regions of the material,
called “domains”. Of course, even though macroscopically, each domain is a
rather small region, still it is large enough to contain a huge number of atoms.

Typically, while the size of a domain volume may range between

to , still each domain may contain, on the average, atoms or
more.

A typical domain pattern is shown in Fig. 20-2.

Fig. 20-2: Typical pattern of domains in Ferromagnetic materials.

As we see, in each domain, the spins are practically all aligned up to the
same direction, but the alignments from domain to domain vary, and since
the domains are randomly oriented, the overall magnetization of the
ferromagnet is practically zero.
 690

However, when the ferromagnet is placed in an externally magnetic field ,

then, all the domains, collectively, are rotated and line up with the applied
field . The ferromagnet now, macroscopically, exhibits magnetic properties.

If the applied field is now removed, we notice that the ferromagnet does
not lose its macroscopic magnetic properties. This phenomenon, in which the
state of a ferromagnetic material depends also from the “magnetic past
history” of the material, is known as “magnetic hysteresis”. This phenomenon
will be studied further in section 20-6.

20-3) Magnetic Field Produced by a Magnetized Object

a) In our so far analysis, we have assumed that static magnetic fields are
generated by steady current distributions, in free space (vacuum). The sources
of the fields are, usually, current carrying conductors. The current flowing in
the conductors is due to the motion of free or conduction electrons. For
reasons to be explained shortly, let us call the current density associated
with the motion of free electrons, ( for surface current densities). The
starting point in our analysis was the Biot-Savart law, and then, using
superposition (integration over the current sources), we could compute, (at
least theoretically), the magnetic field at any point in space.

b) Let us now assume that a piece of material is immersed within a

magnetic field in free space, generated by some current density . As we
have already explained, the atomic magnetic dipole moments (due to the spin
or the orbiting electrons), tend to line up to the applied field . In other
words, the tiny magnetic dipoles, collectively, behave as sources of
macroscopically observed magnetic fields, which however, are not due to the
motion of free electrons, on the contrary, are due to the “magnetic
polarization” of the material.

The total magnetic field, now, will be the superposition of two fields; the
initially existing magnetic field in free space, which causes the magnetic
polarization and the magnetic field resulting from the magnetic polarization.

The situation is analogous to the one described in section 10-2, (Dielectrics

in the presence of an electric field).
 691

c) Let us consider a small volume , within the material, containing tiny

magnetic dipole moments . We may define the “average
magnetic dipole moment per unit volume” as

Note that the unit of is the same as the unit of , (section 16-4).

The physical meaning of the limit is that must be considered very small,
but at the same time, large enough as to contain a great number of tiny dipole
moments. The vector is called “the magnetization vector”.

As we shall show shortly, a magnetization vector varying as a function of

the spatial coordinates, is equivalent to a macroscopic current density, which
we shall call , (magnetization current). Recall that a spatially varying
polarization vector , (the electric analogue of ), gives rise to “bound charge
density”, as shown in equation (10-3-5).

d) To start our analysis, we assume that within a volume , bounded by a

closed surface , there exists a magnetization , which is considered to be a
known function of the spatial coordinates, i.e. .

Fig. 20-3: The magnetic field of a magnetized object.

 692

We want to compute the vector magnetic potential , at the field

point , due to the magnetization .

According to formula (15-4-3), the vector magnetic potential due to

the magnetization is:

where .

In formula (*) we are still using the permeability of free space , since the
effects of the magnetic polarization have been taken into account in the
magnetization vector .

Since ,

Now, using the vector identity

and formula (**), we may express formula (*) as follows:

If we now apply the vector identity (for a proof see Problem 20-3-1),

(where is the unit out normal vector at each point of the surface ), with
, equation (****) becomes:
 693

Comparing formula (20-3-2) with equations (15-2-1) and (15-2-3), we may

attribute a simple, physical interpretation to the integrals in (20-3-2):

1) The term represents a current density , whose

contribution to the vector potential is given by the volume integral in
(20-3-2) over the source volume , while

2) The term represents a surface current density ,

whose contribution to the vector potential is given by the surface
integral in (20-3-2) over the surface , bounding the volume .

The current densities and , caused by the “magnetic polarization” of

the material when immersed in an externally applied field, are called
“magnetization currents” or “bound currents”, (the terminology is similar to
the “bound charges” or “polarization charges” in dielectrics). Note that the
magnetization currents are not associated to the motion of free or
conduction electrons, i.e. to the free currents, exactly like the polarization
charges are not associated with free charges.

Recapitulation: A magnetic field is produced in free space by moving free

electrons, (for example, by current carrying conductors). A piece of material is
immersed in this magnetic field. Due to the torques exerted on the orbiting
electrons and the spin of the electrons, there will be an alignment of the tiny
magnetic dipole moments in the direction of the applied field .
Macroscopically, the material acquires a magnetization , (total magnetic
dipole moment per unit volume). Without magnetic field applied, due to the
random thermal motion, we expect (as defined in (20-3-1)), to be
approximately equal to zero. However, when the magnetic field is applied, all
the tiny dipole moments are oriented in a direction, more or less parallel to the
field , and in this case, usually, . This means that there is another
magnetic field, produced by the magnetization , in accordance to the
equation (20-3-2).
 694

The total magnetic field in space, now, will be the superposition of two
fields, the initial, externally applied field plus the magnetic field due to the
vector , i.e. to the magnetization of the material, (an identical approach was
followed when we studied the placement of a dielectric material in an
externally applied electric field ).

Once the magnetization is known, (or somehow evaluated), then the

field of the magnetized material is identical to the field that would be
produced by a magnetization current density and a magnetization
surface current density .

Example 20-3-1: Find the vector magnetic potential of a uniformly

magnetized sphere of radius .

Solution

Fig. 20-4: Vector magnetic potential of a uniformly magnetized sphere.

Since is constant, inside the sphere, , ,

and there is no contribution to from the volume integral in (20-3-2).
However, we do have a contribution from the surface integral in (20-3-2).

since is a constant vector. Also, on the spherical surface ,

, and formula (*) yields:
 695

We have dealt with, exactly the same, surface integral in formula (**), in
Example 15-2-5, where we found:

By virtue of (***), equation (**) yields:

Example 20-3-2: Find the magnetic field of the uniformly magnetized

sphere of the preceding example, (An alternative solution in Ex.21-2-1).

Solution

Since and formula (****) in the preceding example

becomes:

since .

Having determined the vector magnetic potential , the magnetic

field . Using the expression of in spherical coordinates, we
finally find, (for detailed calculations, see Problem 20-3-2):
 696

Based on formula (**) we can make the following two remarks:

1) The magnetic field inside the sphere is uniform, and

2) Comparing the second formula in (**) with formula (15-4-5), the

magnetic field of a dipole, we conclude that outside ( ), the uniformly
magnetized sphere behaves as a magnetic dipole with magnetic dipole
moment .

PROBLEMS

20-3-1) Show the vector identity: ,

where is a closed surface, bounding the volume .

Hint: Starting with the well known Gauss-Ostrogradsky Theorem

set , where is a constant vector, and then use the vector

identity: .

20-3-2) In Example 20-3-2, start with the expression of the vector magnetic
potential , in equations (*) to derive the magnetic field in equation (**).

20-3-3) An infinitely long cylindrical rod, of radius and its axis coincident
with the axis, carries a magnetization . Find: a) and b)
The magnetic field everywhere in space.

(Ans: inside the cylindrical rod and zero outside,

, ) inside and zero outside.
 697

Hint: .

20-3-4) The region between two concentric spheres of radii and

(assume ), carries a uniform magnetization . Show that the
magnetic field inside the smallest sphere is zero, i.e. for .

Hint: Assume that the sphere carries a magnetization and

the sphere carries a magnetization . The magnetic field
inside the sphere is the superposition of the two fields, one produced by
and the other produced by , etc.

20-3-5) An infinitely long cylindrical shell with inner radius and outer
radius , carries a magnetization . Find: a) and b) the
magnetic field everywhere in space.

(Ans: a) inside the shell and zero outside, on the

cylindrical surface , , and on the cylindrical surface ,

, b) .

Hint: On , on , .

20-4) The Magnetic Field Intensity

a) If is the magnetic field in free space, produced by a current density ,

then, according to Ampere’s circuital law, (section 16-3), we must have:

We point out that in equation (*), is the magnetic field in free space, (no
matter present), while is the free (or conduction) current density.

b) Let us now assume that an object is immersed in an already existing

magnetic field. As we explained in the preceding section, the object gets
magnetized. This results in a magnetization current . The total
current density within the object is
 698

where is the current density due to the magnetization of the material

and is the current density due to everything else, (for example motion of free
charges inside conductors). Equation (*) has to be modified accordingly, in
order to include all the currents, magnetization plus everything else. We thus
arrive at the following expression of Ampere’s circuital law, in the presence of
a material:

If we introduce a new vector defined by

equation (20-4-2) becomes:

where is the current density due to conduction electrons.

The vector is called “the magnetic field intensity” or just the “magnetic
field ”. The unit of is , as it is the unit of .

In free space, , (since in free space there is no matter to be

“magnetically polarized”), formula (20-4-3) reduces to , and we thus
recover the definition of the field in vacuum, (see section 16-4, eq. (16-4-1)).

In terms of the field , Ampere’s circuital law, takes the following form:
 699

An important remark: It is worth of noting, once more, that the current

density and of course the total current in (20-4-5) is associated with all types of
currents except magnetization currents. This is important, since we do not
have any control on the magnetization currents, which are induced when a
material is immersed in a magnetic field. The effects of the magnetization are
incorporated in the field , which is controlled by the “free currents”, i.e. by
the transport of conduction electrons in conductors.

The situation is analogous to the vector in Electrostatics, which

determines the density of “free charges”, which, by definition, includes
everything except the bound charges, induced when a dielectric is placed in an
electric field, ( ). In that sense, we may consider in
Magnetostatics to be the magnetic counterpart of the vector in
Electrostatics.

For a wide class of linear and isotropic materials, the induced

magnetization is proportional to the field , i.e.

where is a dimensionless number, called the “magnetic susceptibility”

of the material. Equation (20-4-3) takes now the form

and if we define “the relative permeability of the material” ,

equation (20-4-7) becomes:

The constant is called “the permeability of the material” and is measured

in , just like , since is dimensionless.

The magnitude of is usually much less than , ( ). For

diamagnetic materials , while for paramagnetic materials . It
follows that for diamagnetic materials and , while for
paramagnetic materials and .
 700

Equation (20-4-8) shows that for diamagnetic and paramagnetic materials

the field is directly proportional to , the constant of proportionality being
the permeability of the material. In ferromagnetic materials, this is no longer
true. Ferromagnetic materials are non linear and anisotropic. It is customary,
though, to write , and treat as a function of .

In particular, for anisotropic materials, we may still write , but now,

is a “ matrix”, known as “the permeability tensor”, defined as

In terms of the permeability tensor, equation (20-4-8) is now written as

In an anisotropic material, it is possible, the component, for example, to

depend on the , and the and the components, etc.

Example 20-4-1: Within a homogeneous, linear and isotropic material with

, a magnetic field is applied. Find and
.

Solution
 701

Example 20-4-2: A long solenoid of radius and turns

per meter, carries a current . The region is filled
with a magnetic material with , while the region
is filled with another magnetic material with . Find the magnetic flux
through a cross section of the solenoid.

Solution

Fig. 20-5: Cross section of the solenoid with a plane perpendicular to its
axis ( axis).

The field inside the solenoid, produced by the current , is given by

formula (17-1-2), i.e. .

In the region , , while in region

, .

The total magnetic flux through the cross section of the solenoid is:
 702

and substituting numerical values we find, .

Example 20-4-3: In Fig. 20-6, the whole space is divided into four regions by
means of the planes and . Surface currents and
flow on the planes, as shown in Fig. 20-6. Find the magnetic field in all
regions. Given: .

Solution

Fig. 20-6: Calculation of the magnetic field in four regions.

The free surface currents and produce magnetic field and

respectively, which, according to formula (14-4-4) will be:
 703

The total magnetic field in each region is obtained by superposition of

the individual fields in each region. The result is shown in the following table:

Total field

Region 4
Region 3
Region 2
Region 1

Having found in each region, the magnetic field is obtained easily:

Region 1: .

Region 2: .

Region 3: .

Region 4: .

PROBLEMS

20-4-1) In Example 20-4-3, find the magnetic susceptibility and the

magnetization vector in each region.

(Ans: , and ).

20-4-2) In Fig. 20-6 it is given: ,

, and . Find the magnetic
field and the magnetization vector in each region.

20-4-3) a) Is associated with macroscopic transport of electric charge? b)

Show that .

(Ans: a) No, b) Since , ).

 704

20-4-4) Within a homogeneous, linear and isotropic material with

, a magnetic field is applied. Find: and
.

20-4-5) Calculate the magnetization of a material in which and

, (Ans: ).

20-4-6) Calculate the magnetization and the field within a material for
which , containing , each having a magnetic
dipole moment of .

20-4-7) A magnetic field is applied in a material with .

Find: and .

(Ans: ).

20-4-8) A coaxial cable whose cross section is shown in Fig. 20-7, carries a
current in the positive direction. The current is uniformly
distributed over the cross section of the inner conductor. The region
is filled with a material of , while the region
is filled with a material of . Find and
everywhere inside the coaxial cable, ( ).

Fig. 20-7: Cross section of a current carrying coaxial cable.

20-4-9) For experimental Physicists, which one of the fields and is more
useful? Justify your answer. Repeat for the electric vectors and .
 705

(Ans: Field , since the sources of this field are free electric currents
through a wire or a coil, for example, and we do have control over these
currents. Similarly, since is controlled by voltage, is more useful).

20-5) Boundary Conditions Between two Magnetic Media

a) The vectors and at the boundary between two different magnetic

materials satisfy certain “boundary conditions”, which result as a direct
consequence of the “solenoidal” property of the field , ( ),
and Ampere’s circuital law of the field , ( ).

Let us, for definiteness, consider two magnetic materials with

permeabilities and , as shown in Fig. 20-8.

Fig. 20-8: Boundary conditions between two different magnetic media.

The boundary conditions at the interface between the two magnetic media,
as we shall show, are the following:

1) For the field :

 706

where is the normal component of in medium , and is the

normal component of in medium . This shows that the normal
component of the field is continuous across the interface.

2) For the field :

If is the surface current density on the interface, then

where is the unit vector normal to the interface, pointing from medium
towards the medium .

Assuming that the (free) surface current density on the interface is zero
( ), which is a realistic assumption in most of the cases, then (20-5-2)
reduces to the following:

where the subscript designates the tangential component of the field .

This shows that, provided , the tangential component of the field is
continuous across the interface.

To show equation (20-5-1) we consider the infinitesimal cylinder in Figure

20-8. Due to the solenoidal property of the field , we have:

or, assuming that , (so that we actually have the magnetic field at the
point , just above and right below the interface), the third term vanishes, and
thus we obtain

and this competes the proof.

 707

To prove (20-5-2) we apply Ampere’s circuital law on the rectangle shown in

Fig. 20-8. With reference to the coordinate system shown in the figure, we
have:

and assuming that , equation (*) yields:

Since , equation (**) becomes,

and using the vector identity ,

(for a proof see Problem 20-5-1), equation (***) yields:

which implies that , or in general, since coincides with

the unit normal at the point , pointing from medium towards the
medium , , and this completes the proof.

If , we obtain , or equivalently, .

b) Refraction of the magnetic field lines at the boundary between two

different magnetic media.

Equations (20-5-1) and (20-5-3) may be combined to show how the field
lines are refracted at the boundary between two magnetic media. Recall that,
by definition, the field lines are always tangent to the field.

Let us consider Fig. 20-9, where and are the angles of and ,
respectively, with the normal to the surface. Assuming that there is no surface
current on the boundary surface, ( ), equations (20-5-1) and (20-5-3)
yield:
 708

Fig. 20-9: Refraction of the magnetic field lines.

From (20-5-4) it follows:

For a proof, see Problem 20-5-2.

These equations allow us to find the fields and on one side of the
boundary, if we know them on the other side.

Two special cases are worth of mentioning.

Case 1: , i.e. the field enters in a region of much greater

permeability. Then, from (20-5-5) it follows, , so
practically, , even for small angles of incidence .
 709

For example, if we assume that , (air) and ,

then, from (20-5-5) it follows that , and
. This leads to the following, useful conclusion.

A magnetic field exists in free space, air for example, with .A

ferromagnetic material, say iron, with is placed inside the magnetic
field. Then, as shown, the field lines inside the iron will be refracted at
approximately from the normal to the surface, as in Fig. 20-10.

Fig. 20-10: Refraction of field lines entering a ferromagnetic substance.

Well inside the iron, the magnetic field is practically zero. This phenomenon
is known as “magnetic shielding”.

Case 2: , i.e. the field enters in a region of much smaller

permeability. Equation (20-5-5) implies that , and
hence . The refracted lines are almost perpendicular to the boundary
surface.

Example 20-5-1: In Fig. 20-9, . If the angle of incidence is

find the refraction angle .

Solution

Application of formula (20-5-5) yields:

 710

Example 20-5-2: In Example 20-5-1, assume that and

. Find and .

Solution

From formula (20-5-6) we have:

In region 1: .

In region 2:
i.e. .

Example 20-5-3: Region 1 where has permeability , while

region 2 where has permeability . A uniform magnetic field
exists in region 1. Find the fields and in
region 2. Assume that on the boundary the surface current density is
.

Solution

Fig. 20-11: Refraction across an interface with surface current density .

In region 1: .

In region 2: let . The field will be:

 711

The continuity of the normal component of the field (equation (20-5-1)),

implies:

To find the tangential component of we apply equation (20-5-2):

The field , while the field

.

PROBLEMS

20-5-1) If are any three vectors, show the identity:

20-5-2) Starting with equation (20-5-4) show eqs. (20-5-5) and (20-5-6).

20-5-3) Two homogeneous, linear and isotropic materials have an interface

at . For , and . In the
region , . Assume that on the interface , . Find ,
and .

(Ans: )
 712

20-5-4) In problem 20-5-3, find the angle of incidence and the angle of
refraction , (see Fig. 20-9), and show that equation (20-5-5) is satisfied. Then
find using (20-5-6) and compare with the value of as obtained
from the expression of in Problem 20-5-3.

(Ans: ).

Hint: ).

20-5-5) Two infinite slabs are located in free space as shown in Fig. 20-12. If
in region 1, find the angle between and in
all regions. Assume on all interfaces.

Fig. 20-12: Refraction of field lines by two infinite slabs.

(Ans: ).

20-5-6) In problem 20-5-5, find using two different methods:

a) Find first the vectors and and then find their magnitudes, and b)
use formula (20-5-6). The results must be identical.

20-5-7) Assuming that , show that the boundary conditions for the
magnetization vector are:
 713

20-6) Hysteresis Loop in Ferromagnetic Materials

As we have pointed out, in the preceding paragraphs, ferromagnetic

materials are non linear materials. The simple relation and the
resulting equation , which applies well for diamagnetic and
paramagnetic materials, does not hold true for ferromagnetic materials. Of
course, we may, sometimes write , even for ferromagnetic materials,
but in this case we understand that is not a constant, on the contrary does
depend on , i.e. .

To study the dependence of from , in a ferromagnetic material, we shall

perform a typical experiment.

Fig. 20-13: Hysteresis loop.

At the center of a solenoid we place a piece of a ferromagnetic material, in

the form of a small cylinder. As we know, the magnetic field inside the
 714

solenoid is approximately uniform, i.e. , where is the number

of turns per unit length and is the current of the solenoid, (see eq. (17-1-2)).
We can, therefore, control the magnetic field , both in magnitude and
direction by controlling the current . The result of the experiment is shown in
the second figure in Fig. 20-13. Initially, the field when , (i.e. when
). Increasing , by increasing the current , the field increases as well,
not linearly, but more or less as shown in the figure, (curve ). Reaching at
the point , we note that further increase of , and therefore of , does not
result in an increase of . The material has reached a state of “magnetic
saturation”.

If we now start reducing , by reducing the current , we note that the

original curve is not retraced, but instead, the relation between and
follows the curve . At the point , , since , but . The
material has acquired a permanent magnetization.

Reversing the direction of , by reversing the direction of the current , the

relation follows the curve . At the point , the field has a negative
value , but the corresponding field is zero. Further decrease of , results
in the curve . At the point we reach again a state of saturation.

We continue our experiment, by increasing now from . We

note that we do not trace the original curve , but, instead we obtain
the new curve . In particular, for the points and , the remarks
made for the points and respectively, hold true.

The loop shown in Fig. 20-13 is known as the “hysteresis loop” of the
ferromagnetic material. This loop shows that the magnetization of a
ferromagnetic material does not depend only on the current value of , but
also from the “past history” of the material.

For example, the field corresponding to , (see Fig. 20-13), could be

either , or or even , depending on how we have reached at the field
, (from , or from , etc), i.e. does not depend solely on
but also on how we have reached at .

The hysteresis loop furnishes a simple method to make a permanent

magnet. We place a virgin (unmagnetized) sample of ferromagnetic material
 715

inside a solenoid, and we start increasing the field by increasing the current
in the solenoid. When magnetic saturation is reached, we start decreasing ,
by decreasing , until (and therefore ) vanishes, (point on the hysteresis
loop in Fig. 20-13). If we now remove the material from the solenoid, the
magnetization remains, i.e. the material has become a permanent magnet.

In Fig. 20-13, is called “residual flux density”, reached when ,

while is called “the coercive force”, reached when .

SUPPLEMENTARY PROBLEMS

20-1) Within a homogeneous, linear and isotropic material with

there exists a uniform magnetic field . Find: and .

(Ans: ).

20-2) In a certain magnetic material and .

Find and .

20-3) The plane is the interface between the two regions, region 1
( ) with and region 2 ( ) with . Given that at the
origin , determine at the origin. Assume that on the
interface, , where .

(Ans: ).

20-4) Consider the four regions: Region 1, , with , region 2,

, with , region 3, , with and region 4,
, with . On the interface , , on ,
, while on , . Find and in all
regions, (see Example 20-4-3).

20-5) A long solenoid has a radius and . The

current of the solenoid is . The region inside the solenoid
has , while the region has . Determine so that
a total flux of is present.

(Ans: ).
 716

20-6) In the region , and , while

everywhere else and . Find and , everywhere in space.

20-7) Region 1, ( ) has , while region 2, ( ) has .

A uniform magnetic field of magnitude exists in region 2, in the
direction . Find the fields and in region 1.

(Ans: ,

).

20-8) The plane divides the space into two regions.

Region 1, which contains the origin, has , while region 2 has .
A field exists in region 1. Find and in region 2.

20-9) In the preceding problem, find the angles of incidence and refraction,
( and in Fig. 20-9), and verify that equations (20-5-5) and (20-5-6) are
satisfied.

20-10) In spherical coordinates, region 1, has ,

region 2, has , and region 3, has
. If , find and in all regions.

20-11) In a homogeneous, linear and isotropic magnetic material, show that

.
 717

CHAPTER 21: PERMANENT MAGNETS - MAGNETIC CIRCUITS

21-1) Permanent Magnets

a) We are all familiar with permanent magnets. We know, for example, that
a piece of iron is strongly attracted towards a magnet, and this, of course,
means that a magnetic field exists in the neighborhood of the magnet. We also
know that a magnet can deflect a compass, something that leads to the same
conclusion, i.e. a magnetic field is produced by the magnet.

However, one might think that, since we do not see any free currents on
the permanent magnets, the field should be zero everywhere, ( ). But,
if this were true, then, inside the magnet we would have , and
outside the magnet we should have , i.e. the magnetic field
outside the magnet would be zero, something which, of course, simply is not
true. Magnetic field exists both inside and outside the magnet.

To explain this “paradox” we think as follows.

Since, indeed, we do not have any free currents on permanent magnets, it

is true that

This equation holds true, both inside and outside the permanent magnet.

Also, since , (see equation (20-4-3)), the solenoidal

property of the field implies:

In summary, the field in a permanent magnet is determined from the

equations
 718

These two equations, in conjunction with the boundary conditions which

must be satisfied by the field on the boundary between the magnet and the
surrounding free space, determine everywhere, (inside and outside the
magnet). In particular, in the space outside, the field is not zero, and
therefore the field outside, is not zero as well.

b) As an alternative approach, the field of a permanent magnet could be

evaluated from the vector magnetic potential , by means of the equation
, provided that the magnetization of the magnet is known. We
have followed this method of approach in Examples 20-3-1 and 20-3-2.

c) Another important remark, regarding permanent magnets, is the

following: Within the magnet, the field has direction opposite to the
direction of the field . As a result, while the field lines of are closed, (an
inevitable conclusion of the solenoidal property, ( )), the field lines of
are open, with their terminal points on the surface of the magnet. The
surface of the magnet where the lines emerge from is the “North Pole”
while the surface where the lines dive in is the “South Pole”.

Indeed, let us consider Fig. 21-1.

Fig. 21-1: Lines of the fields and .

 719

Since , (no free currents present), it follows from Stokes Theorem

that , where is an arbitrary closed curve in space. Also, we
know that the lines of the field are closed. This means that the integral
is either positive (if the integration is in the positive sense, i.e., if
is in the direction of ), or negative (if the integration is in the negative sense,
i.e., if is in the direction of ). Now, if were just proportional to
everywhere, then would, likewise be either positive or negative, but

never zero. But this cannot be true since we know that .

Of course, outside the magnet, is proportional to , since . But

in order to be true, it follows that inside the magnet and
must point in opposite directions. This is shown clearly in Fig. 21-1. The lines
of emerge from the North Pole and dive in the South Pole of the magnet.

21-2) Scalar Magnetic Potential

a) In regions where , the field satisfies the equation .

In permanent magnets there are no free currents. Therefore, the field in a
permanent magnet satisfies . This, in turn, implies that can be
derived as the gradient of a scalar function , i.e.

since then , (is satisfied identically). The scalar

function is called “the scalar magnetic potential”. Since is
measured in , it follows from (21-2-1) that is measured in .

The “magnetic voltage ” between two points and in space, is

defined as

The magnetic voltage is measured in , just like the scalar magnetic

potential. It follows, from equation (21-2-1), that is not uniquely determined,
 720

on the contrary, it is determined up to an arbitrary constant , (just like the

electric potential ). This allows us to chose an arbitrary point in
space, and set . The point is the “reference point” for the magnetic
potentials. Then the potential of any other point can be considered as the
magnetic voltage between this point and the reference point, since

We may visualize a static magnetic field in terms of its “equipotential

surfaces” defined by the equation

To each value of the constant, there corresponds one equipotential surface

of the field. The lines of the field are perpendicular to the equipotential
surfaces. This follows from the defining equation .

b) In order a scalar function to represent the scalar magnetic

potential of a magnetic field, should satisfy the following conditions:

1) must be well defined, finite and continuous at all points in space,

2) The general equation must be satisfied, i.e.

Provided that is constant, or, more generally that , (see

equation (21-1-3)), equation (21-2-5) reduces to the following:

It follows that the scalar magnetic potential , in regions where is

constant or , satisfies Laplace’s equation. All methods and
techniques, developed in Chapter 12 for the solution of Laplace’s equation in
one and two dimensions, can be used to determine .

3) On the boundary between two magnetic media, with permeabilities

and , the following boundary conditions should be satisfied:
 721

If and are the scalar magnetic potentials in the regions and

respectively, then equation (21-2-7) implies,

( represents the directional derivative of in a direction perpendicular

to the boundary, while represents the directional derivative of along the
tangent to the boundary). Since everywhere over the boundary
surface, the second equation in (21-2-8) is automatically satisfied.

Example 21-2-1: Find the magnetic field produced by a uniformly

magnetized sphere.

(The same problem was solved in Example 20-3-2, by an entirely different

method).

Solution

Fig. 21-2: Magnetic field of a uniformly magnetized sphere.

 722

Let be the scalar magnetic potential in region 1, (inside the

sphere), and be the scalar magnetic potential in region 2, (outside the
sphere). We further assume that the permeability of the sphere is , while the
permeability of region 2 is (air). Due to the azimuthal symmetry involved,
does not depend on . Note that inside the sphere, .

The simplest solution of Laplace’s equation , in spherical

coordinates, was found in Example 12-5-1 to be:

where are constants to be determined from the appropriate

boundary conditions.

Since must remain finite at the center of the sphere, ( ), we must

chose . Also, since must remain finite as , we must chose
. Equations (*) and (**) are therefore simplified to the following:

On the boundary , must be continuous, i.e.

Also, on the boundary , , (equation (21-2-7)). Region 2 is

air, permeability , and the normal component of is

Inside the sphere, due to the ferromagnetic material, we cannot just write
. We must use the more general expression ,
from which , i.e.
 723

The condition now implies that , and using

equation (****) we finally find:

The scalar magnetic potentials and , (from equation (***)), are:

The magnetic field inside the sphere, is

The field inside the sphere is uniform. We have obtained exactly the
same result, but with a different method, in Example 20-3-2.

The magnetic field outside the sphere is

This expression of is identical, of course, to the expression found in

Example 20-3-2. For the derivation of from the scalar magnetic potential
, see Problem 21-2-4.

Example 21-2-2: The region is occupied by a ferromagnetic

material in which . Find the magnetic field everywhere.
 724

Solution

Fig. 21-3: Slab permanent magnet.

There are three regions, region 1, ( ), region 2, ( ) and region

3, ( ). If we designate by the scalar magnetic potentials in
regions 1, 2, 3 respectively, then

Due to the symmetry involved, and depend only on , so

, and . Solving these
differential equations we find:

Since and should remain bounded everywhere, we must take

and . Equations (**) are simplified to the following:

Imposing the boundary conditions

 725

(where is the normal component of on the

boundary , etc), we find, (let the reader verify the calculations),

and finally,

The potentials are expressed in terms of an arbitrary constant . However,

this is not a problem, since, as we know, the potential is determined up to an
arbitrary constant.

The potentials in regions 1 and 3, outside the slab, are constant, and
therefore the fields and in these regions is zero. Inside the slab,
however,

In summary:

PROBLEMS

21-2-1) Show that .

Hint: Express and in terms of the unit vectors and .

21-2-2) Show that in any static magnetic field .

21-2-3) A spherical ball of radius , positioned in free space, carries a

magnetization . Find the fields and everywhere in space.
 726

(Ans:

Hint: The scalar potential in region 2, (outside), satisfies Laplace’s

equation , and since, , Laplace’s equation becomes
. In region 1, (inside), we must have . Also, due
to the spherical symmetry, .

21-2-4) In Example 21-2-1, starting with the expression of derive

the magnetic field in region 2.

21-3) Magnetic Circuits - Magnetomotive Force (mmf)

a) In order to explain what is meant by the term “magnetic circuit”, let us

consider, once more, a long solenoid of radius and length , having turns
and carrying a current , as shown in Fig. 17-1 (b). The magnetic field inside
the solenoid, being approximately uniform, is given by equation (17-1-2), i.e.

The magnetic flux through a cross section of the solenoid is given by

equation (15-3-1), i.e.

The quantity is the cross sectional area of the solenoid. If the cross
section of the solenoid is not circular, but of any other arbitrary shape, then
formula (21-3-2) becomes:

where is the cross sectional area of the solenoid.

 727

Formula (21-3-3) is analogous to the formula , expressing the

current flowing through a conductor of length , cross sectional area and
conductivity , (see equation (13-8-2). Comparing this equation with (21-3-3)
we may say that the magnetic flux is analogous to the current , is a term
analogous to the driving voltage , or emf, and for this reason is called
“magnetomotive force ” or simply mmf. The denominator in (21-3-3) is
analogous to the resistance , and is termed “the reluctance ” of the
magnetic circuit, i.e.

If the interior of the solenoid is occupied with a material of permeability ,

then the reluctance will be

In terms of the magnetomotive force and the reluctance , eq. (21-3-3) is

written equivalently as

This equation resembles the well known Ohm’s Law ( ).

The reluctance, in general, depends on the permeability and the

geometrical characteristics of the magnetic circuit.

b) Most often a magnetic circuit consists of a ferromagnetic core with an air

gap, as shown in Fig. 21-4. A magnetic flux is produced, by means of a
magnetomotive force . We assume that all the magnetic flux will be
confined to the ferromagnetic core. This is a reasonable assumption, since the
permeability of the core is usually a few hundred times of . Also, assuming
that the air gap is very small, the flux through the core will be the same as the
flux through the air gap. The magnetic flux remains constant throughout the
magnetic circuit. Again, this is not entirely correct, but it is a useful
approximation, used widely when working with magnetic circuits.
 728

Fig. 21-4: Typical magnetic circuit with air gap.

If we call the magnetic field inside the core and the magnetic field in
the air gap, then application of Ampere’s circuital law over a closed curve
coinciding with the axis of the torus yields:

where is the part of the closed path inside the core and is the part of
the closed curve in the air gap. By virtue of equation (17-2-2), the field inside
the core is approximately equal to , (constant magnitude in the
direction). Also, since the air gap is very small, the field is uniform.
Equation (21-3-7) now becomes:

The magnetic flux , which is constant throughout the magnetic circuit, is:

1) Inside the core:

2) Inside the air gap: Working similarly we find,

 729

By virtue of equations (21-3-9) and (21-3-10), equation (21-3-8) becomes:

This equation leads to the “equivalent magnetic circuit” as shown in Figure

21-4. Reluctances in series are replaced by one total, equal to the sum of the
individual reluctances (just like resistances). Similarly, reluctances connected in
parallel are combined as if they were resistances, (see Example 21-3-2).

The same equation also resembles Ohm’s law ( ), if we make

the correspondence

c) In general, a magnetic circuit may contain more than one mmf, a core
and an air cap. Drawing “the equivalent magnetic circuit”, we may analyze the
circuit using laws similar to the “Kirchhoff’s laws” with which all Electrical
Engineering students are familiar. Let us for example, consider the following
magnetic circuit.

Fig. 21-5: Magnetic circuit with two mmf.

In the “equivalent magnetic circuit”, the arrows next to the magnetomotive

forces show the direction of the magnetic fields produced by the
corresponding currents.
 730

Kirchhoff’s laws for the equivalent magnetic circuit are the following:

For a proof of equations in (21-3-12), see Problem 21-3-5.

Two important remarks:

1) A typical problem in magnetic circuits is to determine the solenoid

current, i.e. the mmf required to produce a given value of the magnetic field
in the air gap.

2) As we have already mentioned, all the formulas derived in this Chapter

are approximate, since we have made a number of simplifying assumptions,
like, for example, have neglected the leakage of the flux out of the core, have
assumed that there are no fringing effects on the air gap, etc. Still, the results
obtained, though approximate, are in pretty good agreement with
experimental measurements and this justifies the validity of our analysis.

Example 21-3-1: Find the total reluctance between the ends of two iron
blocks connected in series, as in Fig. 21-6, assuming that the field is uniform
throughout the blocks and perpendicular to the ends. Assume that the
permeability of the first block is and the permeability of the
second block is .

Solution

Fig. 21-6: Reluctances in series.

The reluctance of the first block is, (formula (21-3-5)):

 731

Similarly, the reluctance of the second block is:

Since the two blocks are in series, the total reluctance will be

Example 21-3-2: Find the total reluctance between the ends of two iron
blocks connected in parallel, as in Fig. 21-7. Assume that field is uniform in
each block and perpendicular to the ends.

Solution

Fig. 21-7: Reluctances in parallel.

The total reluctance is found from the equation

Example 21-3-3: An iron ring has a mean length and a cross

sectional area . Find the mmf required to produce a magnetic field
. Assume that .
 732

Solution

Fig. 21-8: Toroidal magnetic circuit.

From equation (21-3-6) the required mmf is,

The reluctance is , while , and substituting in (*) yields,

The coil could be 180 turns with current 1 A, or 1800 turns with 0.1 A.

Example 21-3-4: In the ring of Fig. 21-8, in the preceding Example, an air
gap is introduced. Find the mmf required to produce the same
magnetic field in the air gap.

Solution

The length of the core is and the length of the air gap is . The
corresponding reluctances are:

and since these two reluctances are in series, the total reluctance shall be

The required mmf to produce , in the air gap, is

 733

Example 21-3-5: Magnetic circuit of a permanent magnet.

Fig. 21-9: Evaluation of a permanent magnet.

The magnetic circuit in Fig. 21-9 is composed of the permanent magnet 1,

the two legs 2 and 3, made of mild steel and the air gap of length and cross
sectional area . One problem of practical interest is to determine the
dimensions of the magnet, i.e. its length and its cross sectional area , so
that the magnetic field in the air gap to have a given value .

In order to solve the problem, we have to make some simplifying

assumptions. First, we assume that the magnetic flux remains the same
throughout the magnetic circuit. Second, we assume that the permeability of
the mild steel is so large that the field inside the legs 2 and 3.
Finally, we assume that the magnetization curve of the permanent
magnet is known, at least in the second quadrant, as shown in Fig. 21-10.

Since remains constant throughout the magnetic circuit, the following

equation holds true:

Also, since there are no free currents present, Ampere’s circuital law
around the closed path , yields (recall that in legs 2 and 3, ):
 734

From (*) and (**) we find easily that

Eliminating between (*) and (***) results in the following equation,

Since are known quantities, equation (****) shows that the

fields and , within the magnet are linearly related. The “operating point”
is thus obtained as the intersection of the magnetization curve with the
straight line , as shown in Fig. 21-10.

Fig. 21-10: Graphical determination of the operating point ( ).

Since the magnet is a non linear medium, the solution is obtained

graphically. Graphical solutions are typical when working with non linear
elements.

Now, having determined , and with and given, we may determine

from equation (***) the length of the magnet. Note that the cross sectional
area is not determined. We may, for simplicity, assume that . The cross
 735

section of the magnet may be determined, provided that some additional

information is given, (see Problem 21-3-3).

Example 21-3-6: The silicon-steel transformer core shown in Fig. 21-11 has
a 200- turn coil and a constant cross section . The mean lengths are
and . What coil current will produce a magnetic
field of in leg 3?

Solution

Fig. 21-11: A silicon-steel transformer.

Let us call the flux densities in regions 1, 2 and 3 respectively, and

the respective fluxes. Obviously,

Since the cross sectional area is constant, equation (*) implies that

Also, from Ampere’s circuital law, we have:

Since the silicon-steel is a non-linear material, we have to make extensive

use of the magnetization curve of the material. The field
and from the curve of the silicon-steel, we find . Then,
from equation (***) we find , and again from the
curve, . Then, from equation (*) we find ,
and from the curve, .

Now, from Ampere’s circuital law, we have:

 736

Remark: Working with magnetic circuits containing ferromagnetic materials

(non-linear), we must use the magnetization curve ( ) of the
ferromagnetic material. These curves can be found in books dealing with
magnetic properties of materials.

PROBLEMS

21-3-1) In the ring of Fig. 21-8, an air gap is introduced. Find the
mmf required to produce a magnetic field in the air gap. Assume
that the mean length and the cross sectional area is .
Assume .

(Ans: ).

21-3-2) In Example 21-3-5, if is the volume of the magnet and

is the volume of the air gap, show that .

21-3-3) The magnetization curve of Fig. 21-10, in Example 21-3-5, has the
analytic expression , where and are assumed to be
known constants. Determine the dimensions of a magnet, of the minimum
volume possible, producing magnetic field in the air gap.

(Ans: ).

Hint: Use the result found in Ex. 21-3-2. To find the minimum volume , set
, etc.

21-3-4) A toroid has an iron core with cross sectional area and
inner radius of . A current coil provides an mmf of . If an air gap
is introduced, find and in the air gap. What is the field prior to
introducing the air gap? Assume that for the iron, .

21-3-5) Starting with first principles, justify the equations in formula (21-3-
12).
 737

SUPPLEMENTARY PROBLEMS

21-1) The curve for a ferromagnetic material has the analytic

expression , (initial curve OKLS in Fig. 20-13). a) Find
the saturation value of the flux density , b) This ferromagnetic material is
used as the core of a toroid with mean length and cross sectional
area . If an air gap of length is introduced, find the field
and the flux , provided that the excitation is .

(Ans: ).

Hint: The resulting equation is a third degree equation for , the field in
the core. Use trial and error to solve this equation.

21-2) A ring made of silicon-steel has a cross sectional area and

a mean length . Find the mmf required to produce a field
. From the curve for the silicon-steel, at
.

21-3) In problem 21-2 find the mmf required by first evaluating the
reluctance of the ring, (Ans: ).

21-4) Repeat problem 21-2, if an air gap of length is introduced.

21-5) Spherical shell in a uniform field .

Fig. 21-12: A spherical shell in a uniform magnetic field .

 738

A spherical shell of inner radius and outer radius is placed in a uniform

field . The regions inside and outside the shell are occupied by air
(permeability ), while the spherical shell itself is occupied by a material of
constant permeability . Find the magnetic field in the inner cavity, (in
region 1).

(Ans:

Note that the field in region 1 is uniform. Also, note that while the electric
field in a cavity inside a conductor is zero, (electrostatic shielding), as we
have shown in section 9-5, the situation is rather different for static magnetic
fields. In the later case, the magnetic field inside the cavity is not zero.

Hint: Let the magnetic potentials in regions 1, 2 and 3

respectively. Each magnetic potential satisfies Laplace’s equation . Due
to the azimuthal symmetry, . Solve Laplace’s equation in each
region and apply the appropriate boundary conditions, etc.

21-6) In Problem 21-5, find the magnetic field in regions 2 and 3.

21-7) In Problem 21-5, assume . If is the thickness of

the shell, plot versus . Note that when , , while
when , .

21-8) A sphere of radius and permeability is placed in a uniform

magnetic field , existing in free space ( ). Find the field
everywhere.

(Ans: Inside the sphere the field is uniform ).

Hint: See Example 21-2-1.

21-9) An infinitely long cylinder of radius , with its axis coincident with the
axis, carries a magnetization . Solve for the scalar magnetic
potential .
 739

(Ans: , ).

Hint: . Solve Laplace’s equation in each region and apply the

appropriate boundary conditions. For the solution of Laplace’s equation in
cylindrical coordinates, when there is no dependence on , see Ex. 12-6-2.

21-10) In Problem 21-9, find the magnetic fields and everywhere. Note
that the field inside the cylinder is uniform.
 740

CHAPTER 22: SELF INDUCTANCE - MUTUAL INDUCTANCE - ENERGY

STORED IN MAGNETIC FIELDS

22-1) Self - Inductance

a) Suppose we have a closed loop carrying a current , as in Fig. 22-1.

Fig. 22-1: Self-inductance of a closed, current carrying loop.

We assume that the current carrying conductor has a small, but non zero
radius (otherwise, the field would become infinite in the vicinity of the
conductor).

The magnetic flux through any open surface , whose perimeter is , is

where is the local unit vector, normal to the surface element .

The vector at the position of is found from the Biot-Savart law, i.e.
 741

By virtue of (22-1-2) equation (22-1-1) becomes:

It follows that the magnetic flux is proportional to the current , (which

produces the flux), and that the coefficient of proportionality depends solely
on the medium (air in our case with permeability ) and the geometry of the
configuration.

The constant of proportionality is called “the self-inductance ” of the loop,

and from equation (22-1-3),

The inductance is measured in , and this is actually the definition of

the . So, the self inductance is measured in . In terms of the
self inductance, the flux through a closed loop is

If the loop exists within a linear and homogeneous medium with constant
permeability , the self inductance is given by the same formula (22-1-4)
where is replaced by .

In practice, the evaluation of from (22-1-4) is rather difficult, if not

impossible, for loops of arbitrary shape. In the next sections we shall develop
another method to find the self inductance, which method is related to the
energy stored in the magnetic field. Recall that we have done something
similar for the computation of the capacitance of various types of capacitors.

b) A useful formula for , in terms of the vector magnetic potential ,

results as follows:
 742

(by virtue of Stokes theorem), and therefore,

c) Self inductance of solenoids and toroids.

Let us consider a solenoid of length , radius , having turns and carrying

a current , as in Fig. 17-1 (b). The cross sectional area of the solenoid is
. Assuming that the turns of the solenoid are packed closely together,
the same amount of flux links each one of the turns, where is the
magnetic flux through the cross section of the solenoid. In order to apply
formula (22-1-5) to find the inductance of the solenoid, we have to use
in the numerator, instead of just , since the “effective area” of the circuit is
now . The quantity is called “the flux linkage ”, i.e.

For a circular loop, (solenoid with one turn), and .

The inductance of a solenoid is found by applying (22-1-5):

The “per unit length inductance” is

Reasoning similarly, we find the inductance of a toroid of turns and inner

radius ,
 743

Comment: In many practical solenoids or toroids, the turns are not packed
closely together. An evaluation of the “effective area”, in such cases, is difficult
or even impossible. The situation is remedied by considering various scaling
factors resulting from experiments and experience.

Example 22-1-1: Find the self-inductance of a solenoid having length

, radius and turns.

Solution

Application of formula (22-1-8) yields:

Example 22-1-2: A toroidal coil has inner radius , cross sectional

area and turns. Find its self-inductance.

Solution

Application of formula (22-1-10) yields:

PROBLEMS

22-1-1) Find the self-inductance of a solenoid having length ,

radius and turns, (Ans: ).

22-1-2) A toroidal coil has inner radius , cross sectional area

and turns. Find its self-inductance.

22-2) Mutual Inductance – Neumann’s Formula

a) Let us now assume that in the neighborhood of a current carrying closed

filament we place another closed loop , as shown in Fig. 22-2. Let also
and be any two surfaces bounded by and respectively.
 744

Some of the field lines, produced by the current , will cross the surface ,
i.e. there will be a magnetic flux through due to the current in loop .

Fig. 22-2: Mutual inductance between two loops.

If we call the magnetic flux through the surface , due to the current
flowing in loop , then

However, from the Biot-Savart law,

From (*) and (**) it follows:

 745

Formula (22-2-1) shows that the flux is proportional to the current ,

producing the flux, and the coefficient of proportionality depends solely on the
medium and the geometry of the configuration. The coefficient of
proportionality is called the mutual inductance between the two loops,
i.e.

Mutual inductance is measured in , just like the self inductance.

The magnetic flux through , due to a current flowing in loop , is

thus expressed in terms of the mutual inductance as

b) The Neumann’s formula for the mutual inductance.

Formula (22-2-2) for the mutual inductance can be written in a different,

more symmetric form, which reveals the deeper physical meaning of the
mutual inductance. Indeed, if is the vector magnetic potential due to the
current ,

In view of (22-2-4) equation (*) becomes:

 746

Formula (22-2-4) is known as the Neumann’s formula. It is a symmetric

formula. It follows that if a current flows in loop , then the mutual
inductance will be identical with . For this reason, we can
drop the subscripts and just talk about the mutual inductance between the
two loops, i.e.

It is clear that depends on the medium and the relative displacement

between the two loops. If the medium is not air ( ), but another medium
with constant permeability , then (22-2-5) still holds true, provided that is
replaced by .

From our so far analysis it follows that

From a physical point of view, the magnetic flux through the loop due
to a current flowing in loop , is identical to the flux through the loop
when the same current flows in loop .

c) Two current carrying conductors.

Fig. 22-3: Flux related to two current carrying closed loops.

 747

Let us consider two current carrying, closed, filamentary conductors, as in

Fig. 22-3. If is the flux through due to the current , and the flux
through due to the current , then

where is the magnetic field due to and is the magnetic field due
to . According to our previous analysis,

The total flux through the surface will be:

While the self-inductance in (22-2-9) is always a positive quantity,

could be either positive or negative. The mutual inductance is said to be
positive if the flux links loop 1 in the same direction as the flux , i.e. if
has the same sign as , otherwise will be negative. If
is negative, then will be negative as well. Obviously, the sign of the
mutual inductance can be changed from positive to negative, or vice versa,
simply by changing the direction of flow of either or .

Example 22-2-1: In Fig. 22-3 it is given: , ,

and . What is the total flux through ?

Solution

Application of formula (22-2-10) yields:

Example 22-2-2: Two circular loops of radii and lie on the planes
and , respectively, with their centers on the axis. Assuming that ,
find and and verify that .
 748

Solution

Let us assume that a current flows in the circular loop of radius . The
magnetic field at the point , due to , is given in equation (14-4-5):

Fig. 22-4: Mutual inductance between to circular loops.

Since , we may assume that the field over the surface of the small
loop is approximately uniform, and equal to the value of the field at the
center of the loop, i.e. equal to . The flux through the small loop,
due to the current in loop 1, will be:

Let us now assume that the same current flows in the small loop. We may
treat this small loop as a magnetic dipole with moment , (for the
 749

current direction shown). Part of the magnetic flux produced by will cross
the loop 1, and therefore,

Taking into account the expression for the vector potential of a magnetic
dipole (equation (15-4-3)) and the fact that , equation (***)
yields:

and since and ,

From (**) and (****) it follows that , as expected.

Example 22-2-3: Find the mutual inductance between an infinite straight

filament and the rectangular loop of Fig. 22-5.

Fig. 22-5: Mutual inductance between a filament and a rectangular loop.

Solution

The field of an infinite filamentary conductor, carrying a current is:

 750

The flux linking the rectangle is

Example 22-2-4: A circular loop of radius lies entirely inside a long

solenoid having turns, length and radius . The circular loop is
perpendicular to the axis of the solenoid. Find the mutual inductance between
the solenoid and the circular loop.

Solution

Fig. 22-6: Mutual inductance between a solenoid and a circular loop.

 751

There are two ways to work the problem. The first is to drive a current in
the small loop (radius ), find the flux linking the solenoid, and then determine
. This is the hard way. Mathematics would be extremely difficult (if not
impossible), to be carried out, (let the reader think how could we evaluate the
flux linking the solenoid, due to a current in the small loop). Fortunately we can
go the other way around. Drive a current in the solenoid and find the flux
linking the small loop. This is very easy to be done.

The magnetic field inside the solenoid, carrying current is uniform and is
given by the expression

The flux through the surface of the small loop will be

Having now determined , we may find easily the flux linking the solenoid
if we drive a current in the small loop. Indeed,

PROBLEMS

22-2-1) In Example 22-2-4 replace the small loop by a small solenoid, lying
entirely inside the big solenoid. Assume that the small solenoid has turns
and radius . Find the mutual inductance between the two solenoids.

(Ans: ).

22-2-2) A solenoid of turns, length and radius exists in a

region with . A second, smaller solenoid of turns, length
and radius is centered coaxially inside the first solenoid. Find the self-
inductance of each solenoid and the mutual inductance between them.
 752

22-2-3) Work Example 22-2-4 assuming that the small circle is tilted by
with respect to the plane.

(Ans: ).

22-2-4) Circuit 1 consists of two long filamentary conductors, lying on the

plane, at and . Circuit 2 is a rectangular loop as shown in
Figure 22-7. Find the mutual inductance between the two circuits.

Fig. 22-7: Evaluation of the mutual inductance.

22-2-5) In Figure 22-7, find the mutual inductance between circuit 1 and
circuit 3, (Ans: ).

22-3) Energy Stored in Magnetic Fields – Magnetic Energy Density -

Inductance in terms of the Magnetic Energy

a) In Chapter 6 we have shown that every electric field contains energy,

which is actually the work expended to build the charge distribution, which, in
turn, generates the electric field. We found that the electric energy density ,
i.e. the electric energy per unit volume ( ) is given by the expression

where is the permittivity of the medium.

 753

The total electric energy stored within a volume is found by integrating

the energy density over this volume, i.e.

b) Similar considerations apply for magnetic fields. Magnetic fields, as we

know by now, are generated by current distributions. The work expended to
build a certain current distribution, reappears as energy stored in the
magnetic field thus generated. So, every magnetic field contains magnetic
energy. The magnetic energy per unit volume is called “magnetic energy
density” ( ).

If and are the field vectors in a region of permeability , then the

magnetic energy density is given by the expression

The total magnetic energy stored within a volume will be

In free space, (air, approximately), .

c) An equivalent expression for the magnetic energy.

Let us assume that we have a current density distribution within a

volume , bounded by a closed surface . This current distribution generates a
magnetic field everywhere in space. The magnetic energy stored in the
volume , according to (22-3-2), will be:

or, since , ( being the vector magnetic potential),

 754

Fig. 22-8: Magnetic energy stored in space.

Making use of the vector identity

it follows that

since , and equation (22-3-3) becomes:

Using the divergence theorem, the first volume integral in (22-3-5) can be
converted to a surface integral over the closed surface , and we thus obtain:
 755

Equation (22-3-6) gives the magnetic energy stored within the volume .
However, the magnetic field exists everywhere, inside and outside the volume
. The total magnetic energy, stored in all space, shall be equal to

All space contains the volume , where , and the rest of the space
where . To compute the magnetic energy in all space from (22-3-6) it
suffices to consider in (22-3-6), that the integration surface becomes bigger
and bigger. The integral remains unaffected in this process, since
outside of . However, the first term in (22-3-6), i.e. the surface integral
tends to zero, as , since the magnitude of drops off as ,
while the surface of the sphere increases as ; thus the integrand varies as
, approaches zero as approaches infinity, and the contribution from the
surface integral vanishes. Equation (22-3-6) now becomes:

Conclusion: The current distribution within the volume , generates a

magnetic field everywhere in space, both inside and outside the volume .
The energy of the magnetic field stored in all space, can be computed either by
formula (22-3-8), (integration over the region of the current sources), or by
formula (22-3-7), (integration over the whole space). Both must yield identical
results, even though the integration takes place over different regions.

Recall that we have faced a similar situation when calculated the energy
stored in electric fields, (see Chapter 6).

In practice, if we want to find the magnetic energy stored within some

volume , we find first the magnetic field , then find the energy density
 756

, and finally we find the energy by integrating over the volume

, i.e.

d) Inductance in terms of the magnetic energy.

The theory we have developed thus far, furnishes a convenient method for
the evaluation of the inductance of a circuit. As we have shown, the magnetic
energy stored in the magnetic field generated by a current distribution is
given by the formula

Let us now assume that we have a circuit, composed by a closed

filamentary conductor of small cross section, carrying a current . Then the
volume integral becomes a line integral over , and as we know, may be
replaced by . Formula (*) for becomes:

or, since , (formula (22-1-6)), , from which,

This formula for the inductance is of importance, and is used frequently

for the evaluation of the inductance . Notice the similarity with the formula
for the capacitance of a capacitor (formula (11-2-3)).

We close this section by posing a reasonable question: We know that

magnetic forces do no work. If this is the case, then where the energy stored
in magnetic fields come from? The answer to this question will be postponed
 757

until Chapter 24, where we shall show that the agent responsible for the
magnetic energy is an electric field. Not a static electric field, but a time
varying electric field generated in the process of building up the magnetic
field. More, in Chapters 24 and 25.

Example 22-3-1: A long coaxial cable of length carries a current on the

surface of the inner conductor of radius . The current returns back along the
outer cylindrical surface of radius . Find the inductance of the cable.

Solution

Fig. 22-9: Inductance of a coaxial cable.

Due to the cylindrical symmetry, , and application of Ampere’s

circuital law yields:

The energy density is .

The magnetic energy stored in the cable is, according to (22-3-9):

 758

The inductance of the cable is, (formula (22-3-11)),

The “inductance per unit length” of the cable is

Example 22-3-2: In the coaxial cable of Fig. 22-9, assume that the current
flows in the positive direction, uniformly distributed over the inner conductor
of radius , and returns back over the cylindrical surface . Find the
inductance of the cable.

Solution

Again, due to the cylindrical symmetry, . Since the current is

uniformly distributed over the cross section of the inner conductor, the current
density is . Application of Ampere’s circuital law yields, (for
detailed calculations see Problem 22-3-1),

The energy density in the region is , while the

energy density in the region is .

 759

Working as in Example 22-3-1, we find the total magnetic energy , and

from this we find the inductance of the cable, (for detailed calculations see
Problem 22-3-2),

The term is called “internal inductance” of the cable, since it is due

to the magnetic energy stored in the interior conductor, while the term
is called “external inductance” of the cable.

PROBLEMS

22-3-1) In Example 22-3-2, apply Ampere’s circuital law to derive the

expressions of in formula (*).

22-3-2) In Example 22-3-2, start with equation (*) and derive equation (**).

22-3-3) Find the energy stored in a solenoid with turns, cross sectional
area and length , carrying a current . Then, using formula (22-3-11), find its
inductance and compare your answer with formula (22-1-8).

(Ans: ).

22-3-4) Using formula (22-3-11) find the inductance of a toroidal coil, and
compare your answer with formula (22-1-10).

22-3-5) In Example 22-3-2, . Find provided that the internal

inductance is of the external inductance.

(Ans: ).

22-3-6) In Example 22-3-2, assume , and .

a) Find the inductance per unit length, and b) If find the energy
stored per unit meter.

22-3-7) The cross section of a long coaxial cable with a plane perpendicular
to its axis is shown in Fig. 22-10. A current flows through the inner conductor
and returns back through the outer conductor. Assuming that the current is
 760

uniformly distributed, find a) The magnetic field everywhere and b) The

inductance per unit length.

Fig. 22-10: Inductance of the coaxial cable.

(Ans:

22-3-8) In Problem 22-3-7, it is given:

and length of cable . How much work does it take to establish
a current ?

Hint: , where .

22-4) Lifting Force of a Magnet

Using energy considerations we may calculate the lifting force of

electromagnets or permanent magnets. The lifting force of an electromagnet
can be very great. To illustrate the main idea, let us consider a U-shaped
electromagnet, as in Fig. 22-11. This electromagnet consists of a U-shaped iron
yoke with turns per unit length, tightly wrapped around the yoke. The
magnetic field intensity in the yoke is therefore , and from the
magnetization curve the field can be obtained. Let us call this value .
 761

If an iron bar approaches the poles of the magnet, it is strongly attracted

towards the magnet. We can determine the lifting force applying the principle
of energy balance.

Fig. 22-11: Lifting force of an electromagnet.

If the air gaps ( ) are small, we may assume that the field , in the air gaps,
is uniform. The energy of the magnetic field stored in both air gaps is

Under the action of the magnetic forces , the bar is lifted by a small
distance , towards the magnet. The decrease in magnetic energy shall be

and energy balance requires that this decrease must be equal to the work
done by the magnetic forces, i.e.
 762

The lifting force of a permanent magnet is found (approximately) from the

same formula.

Example 22-4-1: In Fig. 22-11 find the lifting force, assuming

and .

Solution

and since , the lifting force of the magnet is

.

PROBLEMS

22-4-1) In Fig. 22-11, assuming that and the magnetic flux

through the circuit is , what is the maximum bar weight which can be
lifted? (Ans: ).

22-4-2) In Fig. 22-11 the iron block weights , each pole has a face area
and an effective air gap . What mmf should the coil
supply for the electromagnet to lift the iron block? (Neglect the reluctance of
the electromagnet and the car), (Ans: ).

22-4-3) Repeat Problem 22-4-2, assuming that the iron block has an
effective path length of , an effective area of and an effective
permeability of , while the U shaped magnet has an effective length of
, an effective area of and an effective permeability of .

(Ans: ).

SUPPLEMENTARY PROBLEMS

22-1) A solenoid of 1500 turns has a length of and a radius of

. If the solenoid carries a current , find the magnetic energy
density and the total magnetic energy stored in the solenoid. Then, using
formula find the inductance of the solenoid.

(Ans: ).
 763

22-2) Two large, parallel, plane conducting sheets, separated by a distance

, carry a sheet current density of , in opposite directions. Find
the magnetic energy density between the sheets and the magnetic energy
stored between the sheets over an area of .

22-3) Determine the energy density stored in a medium with by

the field . What is the magnetic energy
stored in a volume of ?

(Ans: ).

22-4) Find the inductance of a toroidal coil, having a square cross section
and turns, packed closely together. The cross section of the toroid is shown
in Fig. 22-12, and its height is .

Fig. 22-12: Inductance of a toroidal coil with rectangular cross section.

(Ans: ).

22-5) a) The cross section of a long coaxial cable by a plane perpendicular to

its axis is shown in Fig. 22-13. A current flows outwards in the inner
conductor , uniformly distributed over its cross section, and returns
through the outer cylindrical surface . The space is occupied
by air, permeability , while the space is occupied by a medium of
permeability . Find the self-inductance per meter of the coaxial cable.

b) Assuming that the length of the cable is , ,

and , what is the work required to establish a current of ?
 764

Fig. 22-13: Inductance of a coaxial cable.

(Ans: , ).

22-6) Find the inductance per unit length of an infinite transmission line, as
shown in Fig. 22-14.

Fig. 22-14: Inductance of a long transmission line.

Assume permeability of the wires and permeability of surrounding space

. The current is uniformly distributed over the cross sections of the wires.

(Ans: ).
 765

22-7) In Problem 22-6, assume that the length of the transmission line is
. Find the inductance , and the work required to establish a current of
.

22-8) A wire of circular cross section of radius is bent to form a ring of

radius . Assuming that the permeability of the wire is , and that , find
the internal inductance of the ring.

(Ans: ).

22-9) Find the mutual inductance between the circle

and an infinite filament coincident with the axis.

Fig. 22-15: Mutual inductance between a circle and a long filament.

(Ans: ).

22-10) a) A charged particle of charge moves in free space with constant

velocity . If and are the electric and the magnetic fields
respectively, at a point whose position vector with respect to is , use
formula (14-2-4) to show that , b) Regarding the charged
particle as a conducting sphere of radius , use formula (22-3-7) to show that
the total magnetic energy associated with the moving particle is
.

Hint: The electric field of a conducting sphere of radius , carrying a charge

, is outside the sphere and is zero inside.

22-11) In Problem 22-10, the charged particle is an electron moving with

speed . Assuming that the total magnetic energy of the electron is equal to its
 766

kinetic energy, show that the radius of the electron is .

Substituting numerical values for the constants we find the radius of the
electron, .

22-12) In Fig. 22-11, assuming that and the magnetic flux

through the circuit is , what is the maximum bar weight which can be
lifted?

22-13) Find the self-inductance of a toroidal coil of rectangular cross

section, inner radius , outer radius , height , with
turns, (Ans: ).

22-14) Using formula (**) for the magnetic field in Example 20-3-2, find
the total magnetic energy associated with a uniformly magnetized sphere of
radius .

22-15) Conducting planes at and carry surface currents .

a) Find the magnetic energy density everywhere in space, b) Find the magnetic
energy stored between the planes per unit area, c) Find the energy stored in
the volume of a sphere of radius centered at the origin.

(Ans: a) , b) ( ,

c) ).

22-16) Show that the interaction energy between two magnetic dipoles
and , separated by a distance , is

Fig. 22-16: Interaction energy between two magnetic dipoles.

 767

22-17) Two circular loops of the same radius , lie on the planes and
, with their centers on the axis. Assuming that , show that the
mutual inductance between the loops is .

22-18) If currents and of the same direction, flow in the circular loops
of Problem 22-17, use the result found in Problem 22-16 to show that the two
loops are attracted with a force .
 768

CHAPTER 23: MAXWELL’S EQUATIONS FOR STATIC ELECTRIC AND

MAGNETIC FIELDS

In this chapter we summarize the main results and equations, obtained in

the preceding chapters.

23-1) Maxwell’s Equations for Static Fields

Thus far we have developed the fundamental equations for static electric
and magnetic fields. These equations, which relate the electric field and the
magnetic field with their sources (charge density) and (current density)
respectively, are the following:

a) Point Form:

b) Integral Form:

These equations are called Maxwell’s equations for static electric and
magnetic fields. Maxwell’s equations are supplemented by the following
constitutive equations:
 769

An important comment: The two “divergence equations”, (i.e.

and ), are valid for static and time varying fields, but the two “curl
equations” hold true only for static fields.

Recall that static fields, in general, are produced by charge and current
densities which do not depend on time, (see Section 13-13).

23-2) Potential Formulation of Electrostatics and Magnetostatics

a) Equation implies that the electric field can be derived as the

negative gradient of a scalar function of the spatial coordinates, the electric
potential , i.e.

The potential satisfies Poisson’s equation

or, Laplace’s equation in charge free regions, ( ),

The usual approach to find the electric field , produced by a given charge
distribution, is to find first , by solving either Poisson’s or Laplace’s equation
(subject to the appropriate boundary conditions), and then find from
equation (23-2-1).

The “solution” of Poisson’s equation , is expressed in integral

form, as follows, (see equation (5-5-4)):
 770

b) Equation implies that the magnetic field can be derived as the

curl of a vector function of the spatial coordinates, the vector magnetic
potential , i.e.

The vector magnetic potential satisfies Poisson’s equation (see equation

(16-5-5)),

or, Laplace’s equation in current free regions, ( ),

The “solution” of Poisson’s equation , is expressed in integral

form, as follows (see equation (15-2-1)):

To find the magnetic field produced by a current distribution, we usually

find first the associated vector potential , using (23-2-8), and then find
. We prefer this approach, since most of the times, the integral in
(23-2-8), expressing in terms of the sources , is evaluated easier than the
integral expressing in terms of , (direct evaluation of from the Biot-Savart
law).

23-3) Boundary Conditions

a) At the boundary between two dielectrics, the normal and the tangential
components of the electric fields satisfy the following conditions:

If on the boundary, which is usually the case, then the second

equation in (23-3-1) becomes, .
 771

b) Similarly, at the boundary between two magnetic media, the normal and
the tangential component of the magnetic fields satisfy the conditions:

where is a unit vector, normal to the boundary, and pointing from

medium 1 towards medium 2. If , which is usually the case, then the
second equation in (23-3-2) becomes, , i.e. .

23-4) The Continuity Equation

The continuity equation is a very general equation, which expresses in

mathematical terms the conservation of charge.

In static cases, (no dependence on time), the point form of the continuity
equation becomes , while the integral form of the same equation
becomes . This means that, in static cases, the total current
entering a closed surface is zero. This statement is in fact Kirchhoff’s Current
Law, as applied to circuit analysis. Kirchhoff’s Voltage Law, (the algebraic sum
of voltages around any closed path is zero), results from the equation
.

Comment: While Kirchhoff’s Laws are exact for static cases only, still, they
can be applied with negligible error in time varying cases, provided that the
time variation is not rapid. As we say, Kirchhoff’s Laws apply well for slowly
varying fields. Alternating currents circuits (AC Analysis) is an important topic
for Electrical Engineers, with many practical applications. Kirchhoff’s Laws are
used in the analysis of AC circuits, even though the household currents
alternate at a rate of 50 times per second.
 772

In the rest of this book we shall derive Maxwell’s equations for time
varying fields. These equations are the most general equations of the
Electromagnetic fields. Time-varying Electromagnetic fields are generated
whenever the charge density and the current density do depend on time.

As we might expect, the general Maxwell’s equations, reduce to equations

(23-1-1) or (23-1-2) when there is no time variation of the quantities involved,
i.e. under static or steady conditions.
 773

PART C: TIME - VARYING

FIELDS AND MAXWELL’S
EQUATIONS
 774

CHAPTER 24: FARADAY’S LAW OF INDUCTION - DISPLACEMENT

CURRENT

24-1) Introduction

In 1819, Hans Christian Oersted, a Danish Physicist, demonstrated very

clearly that a current carrying filament generates a magnetic field in the
surrounding space. This shows that electric currents produce magnetic fields.
Faraday, a prominent British Physicist, thought that, if currents produce
magnetic fields, then, magnetic fields should, somehow, produce electric
currents. Performing a series of brilliant experiments, Faraday was finally able
to demonstrate, in 1831, that time - changing magnetic fields can generate
electric currents.

The main idea in Faraday’s experiment is illustrated in Fig. 24-1.

Fig. 24-1: Faraday’s experiment.

Loop is connected to a battery , by means of a switch . Loop , close

to Loop , is connected to a sensitive galvanometer. The result of the
experiment showed that there is no current in loop 2, when a steady current
flows in loop1. However, every time the switch opens or closes, the sensitive
galvanometer shows a current in loop 2. When the switch closes the
galvanometer shows a momentary deflection; when the switch remains closed
the galvanometer shows no current, while when the switch opens, the
galvanometer shows again a momentary deflection but in the opposite
direction. We deduce that a changing magnetic field, resulting from the
opening or the closing of the switch, produces a current in loop 2.
 775

A slight modification of this experiment showed that whenever a closed

circuit moved inside a time changing magnetic field, a current flowed in the
circuit. This is a situation where we do have a current not due to batteries
connected to the circuit.

Faraday continued with his experiments, and finally he was able to

demonstrate the following important and very general result: Whenever the
magnetic flux through a circuit changes, in any possible way, an
electromotive force (emf - an electrical voltage), is induced in the circuit. This
is known as “the Faraday’s law of induction”. The flux through a loop can
change either by keeping the loop fixed and changing the magnetic field, (as
for example in Fig. 24-1), or by moving the loop inside a constant magnetic
field, or perhaps by a combination of both (changing magnetic field and
moving, simultaneously, the loop). What matters is actually the relative
motion between the loop and the magnetic field.

The main effect of the induction is the production of an emf (electrical

voltage). If the loop is conducting and closed, then the emf will produce a
current, (induced current).

The electromotive force (emf) was defined in Section 13-11. If we call the
“force per unit charge” then, from formulas (13-11-4) and (13-11-6), we have:

Note that, in formula (24-1-1), the electric field responsible for circulating
the electrons, and thus producing a current, is not a static electric field, since
for a static electric field . The electric field in this case, is a time
varying field.

24-2) Faraday’s Law - Lenz’s Law

a) Summarizing the results of Faraday’s experiments and the vast amount of

experimental data obtained since then, we may state Faraday’s law of
induction as follows:
 776

Note that the emf , contrary to what happens in Electrostatics, does

depend on the closed path . Recall that, in Electrostatics, the line integral
, identically, for any closed path. In time - varying cases, this is no
longer true. Also, in Electrostatics, the fact that , implies that the
voltage between two points A and B, defined as is
independent from the path connecting the points A and B. In time varying
cases, this is no longer true. The voltage between two points A and B depends
on the path connecting these two points.

The closed path in (24-2-1) does not have to be, necessarily, a closed
conducting path. It could be an open conducting path, for example a closed
filamentary conductor including a capacitor, or it could be a geometrical closed
path in space. In any case, formula (24-2-1) shows that the line integral
represents a number (a voltage around the path), equal to the time
rate of the flux through any surface having as a boundary, in accordance
to (24-2-1). If it happens, the closed path , to be a conducting path, then the
induced emf produces a current (an induction current), according to
, where is the ohmic resistance of the loop.

In applying equation (24-2-1), in practice, we face the following problem:

Which direction around the closed path are we supposed to integrate?
But, of course, at the same time, we are faced with one more question; what is
the direction of the unit normal vector ? At each there are two unit
vectors normal to . If is one of these two vectors, then the other one will
be . To be consistent with the application of (24-2-1), we have to apply
the “right hand rule”: If the fingers of our right hand define the direction of
integration around the closed path , then the thumb indicates the
direction of . This is illustrated in Fig. 24-2.

Since , if comes out positive, (assuming that is a conducting

loop), then will be positive as well, meaning that the induction current flows
 777

in the positive direction, but if comes out negative, then flows in the
negative direction.

Fig. 24-2: “Right hand rule” relating direction of integration with the
direction of .

Note that the emf which produces the induction current , is not
concentrated in a region in space, (like a battery in a circuit), but it is
distributed all over the closed loop .

b) Lenz’s Law

Lenz’s law is an alternative way to deduce the direction of the induced

currents. It was proposed by Heinrich Lenz, in 1834, and states that “the
induced current in a closed conducting loop is in such a direction as to oppose
to the cause that produced it”. For example, if the induced current is produced
by an increase of the magnetic flux, then direction of the induced current
would be such as to produce a flux opposing the original increase of the flux,
(which produced the induced current). Lenz’s law accounts for the negative
sign in equation (24-2-1), (see Example 24-4).

c) Let us now consider a closed path which bounds an open surface , as

in Fig. 24-3. Then, from equation (24-2-1) we have:
 778

The term may be converted to an equivalent surface integral, by

means of Stokes Theorem, and equation (*) becomes:

or, since the loop is stationary,

This equation must be true for any open surface , having boundary the
closed path , and therefore, the integrand must be zero identically, i.e.

Fig. 24-3: Derivation of Maxwell’s equation (24-2-2).

Equation (24-2-2) is one of Maxwell’s equations, for time - varying fields, in

point form. The integral form of this equation is of course, equation (24-2-1).
 779

For static fields, there is no time dependence, , and (24-2-1) and

(24-2-2) reduce to the Electrostatic equations, and .

Example 24-2-1: Let us consider a fixed circular loop of radius , lying on

the plane, as in Fig. 24-4. Assuming that a magnetic field
exists in free space, find: a) The induced emf around the loop, and b) The
electric field over the loop.

Solution

Fig. 24-4: Induced emf around a circular loop in a time varying field .

a) The magnetic flux through the circular loop is:

Application of formula (24-2-1) yields:

b) Due to the cylindrical symmetry, , and from (24-2-1) we have:

 780

Another method to find : Since , (due to the cylindrical

symmetry), Maxwell’s equation yields, (let the reader verify it):

where is an arbitrary function of and . However, since must

remain bounded at the center, ( ), we must take , and
therefore,

At , we obtain the electric field over the circular loop, i.e.

which is identical to the expression of in equation (***).

Example 24-2-2: The ideal Transformer.

An ideal transformer consists of two coils wound on the same soft iron core,
as in Fig. 24-5. An alternating voltage source is connected to
the “primary winding”, consisting of turns, and a pure resistor (a load) is
connected to the “secondary winding” consisting of turns.
 781

The current provided by the voltage source, produces an alternating

magnetic flux , all of which, we assume, links the secondary winding.

Fig. 24-5: The ideal transformer.

The total flux linking the turns of the primary coil is (see Eq. (22-1-7)),
, and similarly, the total flux linking the turns of the secondary coil
is , where is the flux through a cross section of the core.

With the switch open, there is no current on the secondary circuit, but,
still, there is an induced voltage nevertheless, determined by Faraday’s law:

The primary voltage, again, by Faraday’s law, is :

From (*) and (**) we find:

If , i.e. , we speak of a “step up transformer” while if ,

i.e. , we speak of a “step down transformer”.

When the switch is closed, then, there will be, of course, a current in the
secondary coil. Assuming that there are no power losses in the core, (or, at
 782

least that the power losses are very small as compared to the transmitted
power), power balance requires

Equations (24-2-3) and (24-2-4) are the fundamental equations of an ideal

transformer.

In a more realistic model, things are not so simple. For example, the current
in the secondary coil, produces itself a flux, added to the flux of the primary
coil, etc. But, for simple calculations, formulas (24-2-3) and (24-2-4) are
adequate.

Example 24-2-3: The AC generator.

Fig. 24-6: The ideal AC generator.

Let us consider a rectangle of area , rotating about the axis, with angular
speed . We further assume that a uniform magnetic field exists in
space. If at time , the rectangle lies on the plane, the flux through the
surface , at , will be . At any subsequent time
instant , the flux through shall be . The induced emf, at
the terminals and , as obtained from Faraday’s law is
 783

The rotating rectangle in uniform magnetic field is the simplest type of an

AC generator. The surface could, of course, be any other plane figure. If,
instead of having just one rectangle, we had turns, packed closely together,
then the induced voltage in (*) would be multiplied by .

Example 24-2-4: Lenz’s law.

In Fig. 24-7, find the direction of the induced current in the loop, assuming:
a) The flux through the loop increases and b) The flux through the loop
decreases.

Fig. 24-7: Lenz’s law and direction of induced currents.

a) The induced current is due to the increase of the magnetic flux through
the surface bounded by the loop. According to Lenz’s law, the induced current
should be in such a direction as to produce a magnetic field opposing to the
increase of the magnetic flux (which produced the induced current). This
induced magnetic field should, therefore, point downwards, and the direction
of the current which produces such a field is that shown in Fig. 24-7 (a).

b) Reasoning similarly, the direction of the induced current is that shown in

Fig. 24-7 (b).

Comment: It follows that, according to Lenz’s law, the loop “strives” to

keep a constant flux through it. It reacts to any change of the flux by sending
an induced current in such a direction as to counteract the cause that
produced it.
 784

Example 24-2-5: The Betatron.

The betaron is a machine used to accelerate electrons. The principle of

operation of the betaron is illustrated below.

Fig. 24-8: Principle of operation of the Betatron.

Let us suppose that a magnetic field, pointing in the direction, increases

linearly with time, i.e. that . According to Faraday’s law, this time
varying magnetic field produces an emf around the loop , and therefore an
electric field , (due to the cylindrical symmetry), which can be
found as follows:

The energy gained by an orbiting electron, a distance from the origin,

during one complete revolution is

To get an order of magnitude of , let us assume that

and . Then, according to (**), the energy
gained in one revolution is found to be .
 785

Assuming that the period of the orbit is , then, in

there are revolutions, and the energy gained by the electron
in is .

PROBLEMS

24-2-1) A wire of resistance is bend to form a circle of radius

. The circle is placed, perpendicularly, to the magnetic field
. What is the induction current ?

(Ans: ).

24-2-2) Repeat Problem 24-2-1, provided that the plane of the circle forms
with the magnetic field.

24-2-3) A magnetic field exists in free space. Solve for

the electric field a distance from the origin, (Ans: ).

24-2-4) In Problem 24-2-3, what is the emf around the loop

?

24-2-5) The induced voltage at the terminals, in Fig. 24-9, is constant and
equal to . Determine at , if at . Given:
, , .

Fig. 24-9: Computation of the induced voltage.

(Ans: ).
 786

24-2-6) The magnetic field increases exponentially with time.

Find the electric field induced and the emf around the circle at
.

(Ans: ).

24-3) Conductors Moving in a Uniform Magnetic Field (Motional emf)

a) A very general result, deduced from Faraday’s law, is that a potential

difference is induced across a conductor moving in a magnetic field. To
illustrate this important fact, let us consider Fig. 24-10.

Fig. 24-10: Motional emf.

A magnetic field exists in free space, vertically to the page, and directed
from the reader towards the page. A straight conductor of length is moving
with velocity , perpendicular to the field . The charged particles, inside the
conductor, electrons and positive ions, experience a Lorentz force
. However, since the ions are more than 1800 times heavier than the
electrons, we assume that the positive ions remain stationary and only the
electrons move under the action of the Lorentz force. If is the
(negative) charge of the electrons, then the Lorentz force acting on them is
, and is directed from towards . Each electron is pushed to
the terminal , which now becomes negatively charged. On the terminal we
have a deficiency of electrons, and this terminal becomes positively charged. A
 787

static electric field is now established, with field lines emerging from and
diving at . Each electron is acted upon by two forces, the Lorentz force,
(which remains constant during the process), and the electrostatic force
, pointing from towards , which gradually increases. The
movement of the electrons ceases, when the electrostatic force becomes
equal, in magnitude, with the Lorentz force, i.e. when

We thus see that a potential difference, given by (24-3-1), is induced when

a conductor is moving in a magnetic field.

If forms an angle with the magnetic field , then equation (24-3-1)

becomes,

b) We now assume that the conductor slides freely with velocity along the
conducting bars OM and NT, as shown in Fig. 24-11.

Fig. 24-11: Calculation of the motional emf.

 788

In this case, electrons are moving from , and this

constitutes a current in the counterclockwise direction, as shown in the
Figure, (since the charge of the electron is negative). When the electrons start
moving from D to N, etc, the electrostatic field weakens, and the Lorentz force
takes over, keep sending electrons to the end D, which then move towards N,
etc. The cycle is repeated over and over. Still, each electron is acted upon by an
electrostatic force and a magnetic force (the Lorentz force). The emf which
produces the current in the loop is, (see Section 13-11),

Note that only the magnetic force is responsible for the emf, since the line
integral of the electrostatic field , over any closed path, is zero, (recall that
any electrostatic field is conservative, meaning that ).

Assuming that and , then , and eq. (24-3-3)

yields:

or, since ,

since is the magnetic flux through the surface OCDN.

Since this kind of emf is related to the motion of the conductor, it is

referred to as “motional emf”.

Comment: In deriving the formula for the motional emf, we have assumed
that the magnetic field is uniform. In the most general case though, we may
have a circuit moving in a time - varying magnetic field . We have two kinds
of changes occurring simultaneously (motion of circuit and changing magnetic
field). Then the induced emf is given by the formula:
 789

It can be shown that equation (24-3-5) is exactly equivalent to the Faraday’s

law

Example 24-3-1: In Fig. 24-11, the length of the conductor is , its

velocity is and the magnetic field is . Find
the induced emf and the direction of the induced current .

Solution

From Lenz’s law, the induced current should flow in the counterclockwise
direction.

Example 24-3-2: A conductor CD of length slides with velocity along the

conducting bars KL and MN. The system exists inside a uniform field ,
perpendicular to the plane KLMN, ( plane). Assume that the resistance of
the conductor is , while the resistance of the frame KLMN is negligible. a)
Find the motional emf, b) Find the force (magnitude and direction) required
to keep the conductor moving and c) Verify that the mechanical power
required to move the conductor is equal to the electrical power dissipated in
the resistor .

Solution

Fig. 24-12: Calculation of motional emf.

 790

The motional emf induced is

The induced current , controlled by the resistance of the conductor, is

Application of Lenz’s law shows that the direction of the current is the one
shown in Fig. 24-12. The current carrying conductor CD experiences a Laplace
force ,( ), whose direction is as shown in the Figure. To
maintain the motion of the conductor, a mechanical force must be
applied. The mechanical force is in the same direction with the velocity .

The mechanical power required to keep the conductor moving is

The electrical power dissipated on the resistance is

From (***) and (****) it follows that .

Comment: Given that magnetic forces do no work, it is obvious that the

energy dissipated in the resistor is supplied by the agent who is pulling the
conductor, (mechanical force). The mechanical power is converted into heat
in the resistance .

Example 24-3-3: A solenoid of cross sectional area and

turns lies within a uniform magnetic field , with its axis
parallel to the field lines. A core of soft iron with is inserted in the
solenoid. Assuming that the time of insertion is , find the
induced emf at the ends of the solenoid. If the instantaneous current is
what is the resistance of the coil?
 791

Solution

Before the insertion, the total flux linking the turns of the solenoid is

After the insertion, the total flux is

The change of the flux, which takes place in , is

The magnitude of the induced emf, according to Faraday’s law, is:

The resistance is .

PROBLEMS

24-3-1) In Fig. 24-13 the bar is moving at a speed in a

magnetic field , perpendicular to the plane of the figure and pointing
upwards. Find the induced voltage as a function of time , provided that
at . Work this problem with two methods, a) Use the motional emf
formula (24-3-1) and b) Use the Faraday law. Which method is easier?

(Ans: ).

24-3-2) Repeat Problem 24-3-1, assuming now that .

24-3-3) Repeat Problem 24-3-1, assuming that the magnetic field is

, where , (use formula (24-3-5)).

(Ans: ).
 792

Fig. 24-13: Computation of motional emf.

24-3-4) The rails in Fig. 24-14 have a resistance of . A conducting

bar moves at a constant speed of in a uniform magnetic field
. Find the magnitude of the induced current as a function of time,
as well as its direction. Assume that the sliding bar starts from the left side at
, (Ans: ).

Fig. 24-14: Computation of the induced current .

24-3-5) Repeat Problem 24-3-4, assuming that , where

, (Ans: ).

24-3-6) a) A solenoid of cross sectional area has turns and lies

in a uniform field , with its axis parallel to the field lines. The
solenoid rotates so that in a time interval of its axis becomes
 793

perpendicular to the field lines. Find the induced emf, b) Repeat if in its new
position the axis of the solenoid forms an angle of with the field lines.

24-4) Displacement Current

In the preceding section he have used Faraday’s law to derive Maxwell’s

equation , which shows clearly that a time varying magnetic
field generates an electric field. It is natural, now, to ask the reverse question:
Does a time varying electric field generates a magnetic field? And if yes, how
are these two fields related?

To investigate this problem, let us consider Ampere’s circuital law for static
fields, (the second equation in (23-1-1)), and the continuity eq. (23-4-1)):

The first equation has been derived for time constant fields, whereas the
second equation is true for static and time varying fields (it expresses the
conservation of electric charge).

Even though it is not so obvious, there exists an inconsistency between the

two equations in (24-4-1). The inconsistency results from the well known fact
that the divergence of the curl is identically equal to zero.

So, taking the divergence of the first equation in (24-4-1) yields

This equation is consistent with the continuity equation in static cases,

when there is no time variation involved, but cannot possibly be true for time
varying situations. Ampere’s circuital law has to be amended, so as
to hold for time varying fields as well.

The problem was solved by Maxwell, by his ingenious introduction of, what
we call today, “displacement current”. In order to rescue the continuity
equation, Maxwell assumed that a modification should be made to the
Ampere’s circuital law. In fact, he assumed that we may write Ampere’s
circuital law in the form
 794

where must be determined so as to make this equation consistent with

the continuity equation. Taking the divergence of both sides of this equation
yields:

and substituting by , (from the continuity equation),

results in the following:

The simplest possible solution for is

With this choice of , Ampere’s circuital law takes the form:

Remarks: a) Since , (the displacement vector), is measured in , the

term is measured in , i.e. in , i.e. in ,
as expected. In other words, the physical significance of the term is
current density, and is therefore called “displacement current density”.

b) Note that the most general solution of equation (***) for is

where is an arbitrary vector field, since

 795

But, Maxwell accepted the simplest solution, . His choice has,

since then, been confirmed by a vast amount of experimental data. The
predictions of the theory are in complete agreement with all the experimental
results.

The integral form of (24-4-3) is obtained, as usually, with the aid of Stokes
theorem. If we consider an open surface in space, having a closed loop as
its boundary, then from (24-4-3) we obtain:

The term represents the conduction current through the

surface , while the term represents the displacement

current through . In the light of these two definitions, equation (24-4-4) may
be written as

To elaborate a little more on (24-4-5), and on the physical significance of

, let us consider Fig. 24-15. A time varying magnetic field produces a time
changing magnetic flux through , which in turn, produces an emf around the
loop. Assuming that the boundary of the surface is a filamentary conductor
which connects the two plates of a capacitor, the emf around the loop will
produce an induced current .
 796

Let be the closed loop shown in Fig. 24-15. There are many open
surfaces in space having as a boundary. For example, one such surface is
the surface , punctured by the filamentary conductor.

Fig. 24-15: The physical significance of the displacement current.

Application of Ampere’s circuital law over the closed path which is the
boundary of the surface yields:

where is the conduction current enclosed by .

Let us now consider the open surface . This surface has the same
boundary with , but, on the contrary to what happens with , there is
no conduction current through , since the filamentary conductor does not
pierce . In this case , and (24-4-4) reads,
 797

The displacement current through the capacitor is .

Assuming that is the voltage between the plates of the capacitor, is the
distance between the plates, is the area of the plates and is the
permittivity of the material between the plates, then, as we know,

and, in this case,

since the capacitance of the capacitor is , (formula (11-2-1)).

This formula, expressing the capacitor current in terms of the applied

voltage, is a well known formula in circuit analysis. From formulas (24-4-6) and
(24-4-7) it follows that the conduction current is equal to the displacement
current through the capacitor. In obtaining (24-4-8) we have made the
(reasonable) assumption that exists, actually, only between the capacitor
plates and is zero outside, i.e. we have assumed that the small leaking fields
outside of the capacitor are so small that we may neglect them.

In summary: The integral form of Maxwell’s equation (24-4-4) holds true for
an arbitrary closed path , which is the boundary of an open surface . The
integral has a unique and well defined value which is equal to the
total current crossing the surface. For some surfaces the current is entirely
conduction current, but for some other surfaces, for example, for surfaces
passing through the plates of a capacitor, the circulation of around is
equal to the total displacement current through the surface.

Comment: In static cases or in time varying cases where the rate of change
of is not large, (i.e. for slowly time varying fields), , and this
implies that . In such cases the current through a surface is conduction
current. However, at higher frequencies, (for example optical frequencies), the
displacement current predominates over the conduction current.
 798

In free space, (vacuum), the conduction current is zero, and the magnetic
field , as determine by (24-4-7), is due exclusively to displacement currents,
i.e. to the time variation of the electric field .

Example 24-4-1: Compare the amplitudes of the conduction current density

and the displacement current density , for the electric field
in sea water, , , .

Solution

From equation (13-4-6), (Ohm’s law in point form), we have:

The displacement current density is

From (*) and (**) it follows:

Example 24-4-2: Find the displacement current density associated with the
magnetic field , in free space.

Solution

Since , in vacuum, equation (24-4-3) becomes:

Example 24-4-3: A voltage source is applied to a spherical

capacitor, as shown in Fig. 24-16. The capacitor is filled with a perfect dielectric
 799

with permittivity and permeability . Find the total displacement current

of the capacitor, and thus verify formula (24-3-8).

Solution

Fig. 24-16: Displacement current in a spherical capacitor.

The spherical capacitor was studied in Section 11-2. We assume that the
formulas found there, valid for static cases, still hold true in our case, where
the voltage depends on time, (see comment at the end of the Example).

Therefore, the electric field inside the capacitor shall be, (eq. (11-3-2),

The displacement current density is

The total displacement current through the capacitor, flowing from

towards , shall be, (note that in this case ),
 800

or, since the capacitance , (see formula (11-3-4)),

Formula (**) is a well known formula for the capacitor current in terms of
the applied voltage.

Comment: Faraday’s law and the displacement current, strictly speaking,

pertain to time varying fields. However, provided that the rate of change of
the fields is not large, then . Maxwell’s equation
reduces to the equation , describing static fields, where, however
now, the quantities involved ( and ) do depend on time. This state, where
Electrostatics and Magnetostatics rules and formulas can be applied even for
time varying fields, not changing very rapidly, is known as “quasistatic”.

We may, for example, use the formula obtained for the electric field in a
spherical capacitor, assuming static conditions, (formula (11-3-2), to the case
where , (as we did in Example 24-4-3). Or, we may use Biot-
Savart law to find the magnetic field produced by a filamentary conductor
carrying a time varying current , etc. The results thus obtained, even
though approximate, are extremely good approximations with negligible
errors. An electrical engineer, working with household frequencies, ( in
Europe, or in U.S.A), may safely apply all formulas obtained in
Electrostatics and Magnetostatics, to analyze and design the circuits and
networks he is working on.

However, when the time variation of the fields is extremely rapid, then,
the quasistatic approximation breaks down, it is no longer valid. This is the
regime of electromagnetic radiation. In this regime, we may no longer assume
that and , and Maxwell’s equations in their complete
form must be used to analyze the problem at hand.
 801

PROBLEMS

24-4-1) Compare the amplitudes of the conduction current density and

the displacement current density , for the electric field
in Germanium, , ,
, (Ans: ).

24-4-2) Consider the cylindrical capacitor in Fig. 11-8. For an applied voltage
, find the total displacement current of the capacitor,
and thus verify formula (24-4-8).

Hint: See Example 24-4-3.

24-4-3) The displacement current density in air ( and ) is

. Find the vectors and , and
determine the angular frequency .

(Ans: ,
).

24-4-4) A magnetic field in a region in free space is given, in spherical

coordinates, as . Use Maxwell’s equations
to find the electric field .

Hint: Start with , express the curl in spherical coordinates,

etc.

SUPPLEMENTARY PROBLEMS

24-1) A filamentary conductor coincident with the axis, carries a current

. Assuming quasistatic conditions, find the magnetic field .

(Ans: ).

24-2) Find the electric field produced by the magnetic field .

24-3) If is the inductance of a solenoid, carrying a time changing current

, show that the voltage induced in the solenoid is given by the formula
 802

, (this is also a well known formula, in circuit analysis, for the

voltage of an inductor).

24-4) If , in free space, ( is the speed of light),

find: a) The magnetic flux through a square of side , lying on the plane
and centered at the origin, and b) The induced emf around the square in the
counterclockwise direction.

24-5) Faraday’s disk: A thin conducting disk, ( ), rotates with

angular speed about a shaft, in a uniform magnetic field , perpendicular to
the plane of the disk, as shown in Fig. 24-17. Find the emf induced between the
shaft and the circumference of the disk.

(Ans: ).

Fig. 24-17: Faraday’s disk.

Hint: Consider the Lorentz force on an electron, at some point on the

disk. This force points in the radial direction, inwards or outwards, depending
on and . Then, , etc.

24-6) A long filamentary conductor coincident with the axis carries an

alternating current and returns back through a coaxial
cylindrical surface of radius . Find the electric field everywhere.

(Ans: ).
 803

Hint: From Ampere’s circuital law, find , and

then, from Maxwell’s equation , find .

24-7) In Fig. 24-18, , (directed out of the page), and the position
of the sliding bar is given by , (in ). Find the voltage as a
function of time . Use two methods to determine : a) The formula for the
motional emf and b) Faraday’s law . The results must, of course,
be identical.

Fig. 24-18: Computation of the induced emf in a sliding bar.

(Ans: ).

24-8) A rod long revolves at a constant angular speed of in

a uniform magnetic field . The axis of rotation passes through one
end of the rod and is parallel to the magnetic field. Find the induced emf at the
ends of the rod, (Ans: ).

24-9) Two coils have a mutual inductance . Given that the

current in the first coil is , find the emf induced in the
second coil, (Ans: ).

24-10) The speed of an airplane is . Find the induced emf at the

ends of the wing if the vertical component of the magnetic field of the Earth is
and the distance between the ends of the wings wing is .

24-11) A coil consisting of turns rotates in a uniform magnetic field

. The axis of rotation is perpendicular to the axis of the coil and to
 804

the direction of the magnetic field. The period of rotation of the coil is
and its cross sectional area is . Find the amplitude of the induced emf.

(Ans: ).

24-12) A circular wire of radius is placed in a uniform magnetic field

. The plane of the circle is perpendicular to the magnetic field. What
is the induced emf if the field is switched off during .

24-13) A coil consisting of turns of wire is placed in a uniform field

. The resistance of the coil is and its cross sectional area is
. The angle between the axis of the coil and the magnetic field is .
What electric charge will pass through the coil if the magnetic field is switched
off? (Ans: ).

Hint: Show that, in general, .

24-14) A circular wire is placed on top of a solenoid with its axis being
vertical. When a DC voltage source is connected to the solenoid the circular
wire jumps (momentarily) in the air. Explain.

Hint: From Lenz’s law, the current in the circular wire and the solenoid
current flow in opposite directions, and as we know, opposite currents repel,
(see Example 18-1-1).

24-15) In Fig. 24-19, as the magnet approaches the stationary metal ring, an
emf is induced, which in turn, gives rise to an induced current. What is the
nature of the force responsible for circulating the electrons around? Can this
force be magnetic? (Ans: The force cannot be magnetic since the ring is
stationary; it must therefore be electric. Recall that ).

Fig. 24-19: Permanent magnet approaching a stationary ring.

 805

CHAPTER 25: MAXWELL’S EQUATIONS (TIME VARYING FIELDS)

25-1) Point and Integral Form of Maxwell’s Equations

In the light of Faraday’s law of induction and the displacement current

density, Maxwell’s equation for static fields (Chapter 23), are modified as
follows for time varying fields:

a) Point form:

b) Integral form:

For static fields, there is no time variation of the quantities involved, and
therefore all the partial derivatives with respect to the time vanish. In this
case, we recover Maxwell’s equations for static fields, (equations (23-1-1) and
(23-1-2)).

Maxwell’s equations are supplemented by the (already known equations),

 806

For linear and homogeneous materials, , and

equations (25-1-3) take the more familiar form,

Maxwell’s equations are consistent with the continuity equation

.

Maxwell’s equations relate the electric and magnetic fields with their
sources, charge density and current density . Being Partial Differential
Equations (P.D.E), Maxwell’s equations admit an infinite number of solutions.
To hope for a unique solution, when working in a concrete problem, we have
to consider the appropriate Boundary and/or Initial conditions, which, when
applied, lead to a unique solution. Boundary conditions for time varying fields
will be studied in Section 25-4.

Usually, we prefer to work with Maxwell’s equations in point form. Once a

coordinate system is chosen, (Cartesian, or Cylindrical, or Spherical, or any
other orthogonal coordinate system), we express the curl and the divergence
in this particular system, and split the vector equations into scalar equations,
one for its direction, (the choice of the coordinate system is dictated by the
geometry of the problem at hand). Note that each one the curl equations give
three scalar equations. These equations are interrelated and the mathematical
manipulation is not easy, as a matter of fact it is quite difficult, (see Example
25-3-1 and Problems 25-3-1 and 25-3-2).

25-2) Maxwell’s Equations in Free Space (Vacuum)

In free space, and therefore , (conduction current density),

, (no free charge in vacuum), and . In this case,
Maxwell’s equations (25-1-1) become:
 807

Note how crucial the displacement current density is.

Equations (25-2-1) show clearly that a time changing magnetic field produces a
time changing electric field (the first curl equation), and then, the time
changing electric field produces a time changing magnetic field (the second
curl equation), which in turn produces an electric field, etc, etc.

In general, we may say that every time varying magnetic field produces an
electric field with closed lines and every time varying electric field produces a
magnetic field with closed lines, (Recall that the lines of static electric fields
are open).

Fig. 25-1: Electric field with closed lines, (not electrostatic field).

By a suitable transformation, it can be shown that Maxwell’s equations in

free space lead to a wave equation, i.e. the electric field and the magnetic
field satisfy a wave type of equation. We shall show this important fact in
the next section. Maxwell’s equations predict the existence of
“electromagnetic waves” in free space, traveling at the speed of light. Since
light was known to be a wave phenomenon, it was immediately realized that
light is an electromagnetic wave, of very high frequency though, , or
more). Soon after Maxwell’s prediction, in a number of experiments conducted
by Hertz , in 1988, it was demonstrated very clearly that electromagnetic
waves do exist and travel at the speed of light. One can transmit
 808

electromagnetic waves from one place and receive them at another place. The
first “wireless communication” was achieved.

25-3) The Wave Equation for and

Taking the curl of both sides of the second equation in (25-2-1), we obtain:

(The operators and commute since the operator operates on

spatial coordinates and operates on time). Making use of the vector
identity

(since form the fourth equation in (25-2-1)), equation (*)

becomes:

Since and , the quantity

and equation (***) becomes . Similarly, taking the curl of

both sides of the first equation in (25-2-1) we obtain an identical equation for
. We have thus shown that and satisfy the equations:
 809

The type of equation satisfied by and is known as “a wave equation”.

Comments: a) The wave equation in a medium with permittivity and

permeability , is the same as in (25-3-2), where now the speed of light in the
medium is .

b) The speed of light (and of all Electromagnetic waves) in vacuum, is

. Since and are universal constants,
the speed of light, in vacuum, should also be a universal constant, not related
to any particular frame of reference. This seems to be in a direct conflict with
classical Physics, in the regime of which, the speed of a particle or a wave must
necessarily refer to a particular frame of reference, (assumed to be at rest). It
seemed that there was a contradiction between classical physics and Maxwell’s
theory of Electromagnetism. These, seemingly opposing views, were reconciled
by Einstein, in 1905, in his paper entitled “On the Electrodynamics of Moving
Bodies”, where he introduced his Special Theory of Relativity. Einstein
accepted Galileo’s principle of relativity, (all inertial systems are appropriate
for the description of the laws of nature), and that the speed of light in vacuum
is a universal constant, as deduced from Maxwell’s equations. Special relativity
and all its subsequent deductions about time dilation, length contraction,
equivalence of mass and energy, etc, are based on these two axioms.

Example 25-3-1: An electric field exists in

free space. If , use the wave equation to determine .
Then use Maxwell’s equations to find .

Solution

The electric field must satisfy the corresponding wave equation

. We have:

Similarly,
 810

By virtue of (*) and (**), the wave equation becomes:

and in order this equation to be satisfied identically, we must choose

, i.e. .

To find the magnetic field we start with the equation ,

which when expanded in a Cartesian coordinates yields:

and integrating with respect to the time, we find:

Note that and satisfy the two divergence equations ( and

).

Example 25-3-2: Assuming that and

, expand Maxwell’s equations in free space in Cartesian
coordinates, (each one of the field components is supposed to be a function of
position and time ).
 811

Solution

Similarly, starting with we find:

The equation becomes:

and similarly, becomes:

Equations (*), (**), (***) and (****) are Maxwell’s equations in expanded
form, in Cartesian coordinates.
 812

Expressing the curl and the divergence in cylindrical and in spherical

coordinates, we may expand Maxwell’s equations in these coordinate systems,
(see Problems 25-3-1and 25-3-2).

PROBLEMS

25-3-1) Expand Maxwell’s equations in free space, in cylindrical coordinates,

(assume and ).

(Ans:

25-3-2) Expand Maxwell’s equations in free space, in spherical coordinates,

(assume and ).

25-3-3) Starting with , derive the continuity equation

.
 813

25-3-4) An electric field exists in free space. If

, use the wave equation to determine . Then use
Maxwell’s equations to find .

25-3-5) A time varying electromagnetic field exists in a linear and

homogeneous medium of conductivity and permeability . Assuming that
the time variation is slow, show that the current density satisfies the
differential equation .

Hint: Slow time variation means that . Then, starting with

and taking the curl of both sides yields:

Note that , since from the continuity equation,

and equation (*) becomes, , i.e. , etc.

25-4) Boundary Conditions for Time - Varying Fields

It is well known that the electromagnetic fields and will, in

general, be discontinuous at the boundary between two different materials. In
section 23-3 we have stated the boundary conditions for static fields. In this
section we shall develop the boundary conditions for time varying fields.
Boundary conditions are necessary when two or more materials are present.
Maxwell’s equations must be solved separately in each region and then, the
solutions at the boundaries must be related by the appropriate boundary
conditions. This leads to unique solutions in each region.
 814

The boundary conditions for time varying fields are actually the same as
their static counterparts. Also their derivations are very similar to the
corresponding derivations for static fields stated in section 10-6, (for and ),
and section 20-5, (for and ).

Therefore, we just state the boundary conditions for time varying fields and
leave their derivations in Problem 25-4-1.

Let the surface be the boundary between two different media 1 and 2, as
in Fig. 25-2. If and are the electromagnetic fields in
regions 1 and 2 respectively, across the boundary, then the following boundary
conditions hold true, (note that is a normal to the boundary unit vector,
pointing from region 2 towards region 1):

Fig. 25-2: Boundary conditions for time varying fields.

a) Normal components of .

If , (surface charge density), then (25-4-1) is simplified to the

following:
 815

b) Tangential components of .

c) Normal component of .

d) Tangential component of .

If on the boundary, then (25-4-5) is simplified to the following:

A case of particular interest arises when medium 2, in Fig. 25-2, is a perfect

conductor. As we have shown in Electrostatics, the electric field inside a
perfect conductor is zero. The same is true for perfect conductors even in the
time varying case. The time varying electric field and magnetic field are
zero inside a perfect conductor. Thus the boundary conditions, for time
varying fields, when medium 2 in Fig. 25-2 is a perfect conductor are the
following:

For a justification of formulas (25-4-7), see Problem 25-4-3.

Example 25-4-1: The unit vector is directed

from medium 2 with , towards medium 1 with
. If at a
point in region 2 adjacent to the boundary, find and . Assume that
on the boundary.
 816

Solution

Let and
the fields and , in region 1. Since , application of (25-4-1) yields:

As a vector,

The normal component of is

The tangential component of the field is determined from (25-4-3), i.e.

The tangential component of is:

 817

In Summary:

The field is,

while the field .

Note that the frequency of the wave remains unaffected in the transition
from medium 2 to medium 1.

PROBLEMS

25-4-1) Starting with Maxwell’s equations in integral form, deduce the

boundary conditions (25-4-1) up to (25-4-6).

Hint: See the corresponding derivations for static fields in sections 10-6 and
20-5.

25-4-2) Region 2 ( ) has , while region 1

( ) is free space. The electric field in region 2 is ,
( ). a) Find , b) Find the magnetic field , c) Find and in region 1.
Assume that on the boundary , and .

25-4-3) Justify the boundary conditions (25-4-7), when region 2 is a perfect

conductor.

Hint: A perfect conductor, by definition, has an unlimited supply of free

electrons, (which means that the conductivity ), while, at the same time,
the current density remains finite. This can be true only if we assume
inside the conductor, (since ). Then from Faraday’s law, we deduce that
the magnetic field inside the perfect conductor.

25-4-4) Region 2 ( ) is a perfect conductor, while region1 ( ) is free

space. Assuming that at some time instant , on the boundary ,
and , find and at the
same time instant just above the boundary, (i.e. at all points ( )).
 818

Hint: Apply the boundary conditions in formula (25-4-7).

25-5) Scalar Potential and Vector Potential for Time - Varying Fields,
The Lorentz Condition

a) In Electrostatics, the fundamental law implies that can be

derived as the gradient of a scalar potential , i.e. we may write ,
(since, then, the equation is satisfied identically). For
similar reasons, in Magnetostratics, implies that can be derived as
the curl of a vector potential , i.e. we may write , (since, then,
is satisfied identically).

b) In time varying fields the situation is a bit different. We know that time
varying electromagnetic fields are governed by Maxwell’s equations, repeated
hereunder for convenience:

Since holds true even for time changing fields, we may write

and then, equation (1) may be written as

This equation implies that

(since, then, ), and therefore

 819

Equations (*) and (**) show that even in the time varying case, the
electromagnetic fields and can be expressed in terms of a scalar
potential and a vector potential , in accordance to equations (*) and (**).

Equations (*) and (**) were derived as a consequence of Maxwell’s

equations (1) and (4). What about the other two equations?

Substituting (*) and (**) into equation (2), in Maxwell’s equations, we

obtain:

Similarly, substituting (*) and (**) into equation (3), in Maxwell’s equations,
we obtain:
 820

Equations (25-5-1) and (25-5-2) are differential equations relating the

scalar potential and vector potential with their sources and , which in
general, depend on both, position and time.

Note that in static cases, equations (25-5-2) and (25-5-1) reduce to

Poisson’s equation and respectively.

c) The Lorentz condition

We note that the quantity appears in equation (25-5-1) and

in equation (25-5-2). Recall that the curl of is already specified in equation
(*), i.e. . However, as we know, in order to completely determine
we must also specify its divergence , (see Example 25-5-2). In our case, we
choose in such a way as to obtain the simplest possible expression of
equations (25-5-1) and (25-5-2). An obvious choice towards this direction is to
take

Equation (25-5-3) is known as the “Lorentz condition”. With this choice for
the divergence of , i.e. by taking , the two equations become:

One great advantage in equations (25-5-4) is that that the differential

equations for and are uncoupled. The potential is determined from
only and similarly the vector potential is determined from only. Once and
are found, (by solving the differential equations in (25-5-4), then and are
determined from equations (**) and (*) respectively.

Example 25-5-1: As we have shown, in time varying fields, the electric field
and the magnetic field , are given by the expressions:
 821

Show that if is an arbitrary scalar field, (depending on position and

time), then and are not affected if is replaced by and is
replaced by .

Solution

since .

The invariance of (*) under the transformation

is known as “gauge invariance” and the equations in (**) are known as

“gauge transformations”.

Example 25-5-2: a) Consider the vector field . The curl of

is known and given by the expression . Is this information
adequate to specify uniquely? b) If, in addition to the curl, the divergence of
is given by the expression , can we specify ?

Solution
 822

From the first equation in (*) we have, , where

is an arbitrary function of and . Then,

and finally, , where is an arbitrary function of . The

conclusion is that a vector field cannot be specified by its curl alone.

b) The additional information, , however, enables us to

determine . Indeed,

Conclusion: The vector field is specified by its curl and its divergence, (not
by the curl alone or by the divergence alone). The constant may be
determined if the value of the field at one point is given. For example, if
, then , and .

PROBLEMS

25-5-1) Determine the vector field given that:

(Ans: ).
 823

25-6) The Retarded Potentials

a) Before we can determine the fields and from eqs. ( )

and , respectively, we must first find the potentials and .
These potentials are, in principle, obtained as the solutions of eqs. (25-5-4).
Assuming the charge density and the current density to be known functions
of position and time, equations in (25-5-4) are partial differential equations for
and , respectively. At this point, we shall not give the rigorous derivations
of the solutions; instead, we shall merely present the solutions and state some
general important remarks and conclusions.

b) Let us first formulate our problem, more precisely.

Within a volume , bounded by a closed surface , there exists a charge

density and a current density , (source coordinates are denoted
by , and field coordinates are denoted by , see Section 1-13).

At a point in space, we want to determine the scalar potential

and the vector potential .

Fig. 25-3: Retarded potentials and at the point .

 824

The key point in our further analysis is that electromagnetic waves travel at
a finite speed in space, (see comments in Section 25-3). In free space, the
speed of the wave is actually the speed of light, .
In another homogeneous and isotropic medium with permittivity and
permeability , the speed of the wave shall be .

In classical Physics, the speed of light was assumed to be infinite, ( ).

This implies that any electromagnetic disturbance, (signal), generated at some
point in space, is transmitted to a distant point instantaneously. However, in
view of the fact that the speed of a wave is finite, we have to take into
account the “time retardation” between the time a wave is generated and the
time the wave reaches at another point in space.

For example, in Fig. 25-3, the scalar potential and the magnetic
potential at the point , at the time instant , is due not to the value
of the charge density and the current density at the same time instant , but to
their values at a previous time instant (earlier), since the wave
needs a time of to propagate from to , and this implies the following
two expressions for and :

and integrating over the volume we obtain:

The expressions for and in formula (25-6-1) are known as “retarded

potentials” since they include the retarded time .
 825

Rigorous analysis shows that the integral expression for in (25-6-1) is a

solution of the first equation in (25-5-4) and the integral expression for in
(25-6-1) is a solution of the second equation in (25-5-4).

Retarded potentials are very useful, and used frequently, in problems

related to the electromagnetic radiation from known sources and/or .

Example 25-6-1: Consider a small current element carrying a current

in the direction, as in Fig. 25-4. Determine the vector
magnetic potential at the point .

Solution

For a filamentary conductor, is replaced by , and the

formula for in (25-6-1) becomes:

Fig. 25-4: Vector potential due to an alternating current element.

 826

In deriving equation (*) we have assumed that the point is far away from
the current element, so we may make the reasonable approximation ,
and since is constant it comes out of the integral.

If we call , formula (*) becomes:

or, since ,

PROBLEMS

25-6-1) In Example 25-6-1 find the magnetic field at the point .

Hint: Start with , and use the expression for in spherical

coordinates. The computations are tedious but straightforward.

25-7) The Poynting Vector and Transmission of Electromagnetic Energy

a) This section is devoted to a general treatment of energy and power in

Electromagnetic fields. We know that, in general, a charge moving in an
electromagnetic field with velocity , experiences a force, the Lorentz force,
given by the formula

The charged particles, under the action of , may gain energy. This energy
increase is obtained at the expense of the field energy, (the energy of the field
decreases). The reverse process is also possible. If the energy of a particle
decreases, the energy of the field increases.
 827

Let us assume that a charged particle is moving with speed , inside an

Electromagnetic field . Then the particle experiences a force , in
accordance to equation (*), and the power delivered to the particle is

since is perpendicular to the vector , and therefore their dot

product vanishes, i.e. , (recall that magnetic forces do no
work).

b) Let us now consider a continuous charge distribution moving with

velocity , as shown in Fig. 25-5.

Fig. 25-5: Energy balance in Electromagnetic fields.

The power delivered to the charge , inside the volume ,(

), shall be:

since , (conduction current density). The total power delivered by

the fields to the volume , is therefore,
 828

Using Maxwell’s equation , we may express in terms

of field quantities exclusively, and (25-7-3) becomes:

Using the vector identity

equation (25-7-4) is written equivalently, as

Assuming that and do not depend on time, then, for linear materials,
and similarly, , and equation (25-7-5)
becomes:

or, converting the volume integral in the left side of the equation to a
surface integral over the closed surface , bounding the volume , we obtain:
 829

This equation, which actually expresses the conservation of energy, is to be

interpreted as follows:

1) From equation (25-7-3),

represents the total Joule heat loss, within the volume , (see
Section 13-7).

2) The term , represents the “energy per unit volume

stored in the magnetic field”, i.e. represents “magnetic energy density”. For
linear media,

3) The term , represents the “energy per unit volume

stored in the electric field”, i.e. represents “electric energy density”. For linear
media,

4) Since the right side of equation (25-7-6) represents power, ( ),

the left side should also represent power, and this means that the unit of
must be . Indeed, check by direct calculation that is
measured in . We may thus define the “Poynting vector ” by means of
the formula:

With being the out normal vector over the closed surface , the integral
of the normal component of , over the closed surface, ( ),
gives the total electromagnetic power leaving the volume , through the
closed surface . This power results from the time rate of decrease of the total
energy stored within the volume.

In general, therefore, we may interpret the Poynting vector , as

a vector giving the direction (and the magnitude as well), of the
 830

instantaneous electromagnetic power density (power per unit area) at any

point in space. Of course, only the normal component of contributes to the
power transmitted through a surface.

The Poynting vector was originally introduced by the British Physicist John.
H. Poynting, in 1883.

Example 25-7-1: A cylindrical resistor of radius and length carries a

current , uniformly distributed over its cross sectional area. Find the electric
field inside the resistor, the magnetic field on its outer surface, and then
integrate the pointing vector over the surface of the resistor to find the total
power dissipated.

Solution

Fig. 25-6: Power loss in a resistor.

The electric field inside the resistor is , while application of

Ampere’s circuital law yields the magnetic field on the surface :

The pointing vector on the surface of the resistor is,

This equation shows that the Poynting vector , at the surface , points
radially into the resistor; we have a power flow into the resistor. The total
power flowing into the resistor is found by integrating over the surface, i.e.
 831

which is a well known result for the power dissipated on a resistor.

Note that there is no contribution to the power from the top and the
bottom surface, (why?).

Example 25-7-2: a) Find the Poynting vector associated with the electric
field in free space, b) Find the time average
of the Poynting vector over one period, c) Find the average power
crossing a surface of placed perpendicularly to the axis.

Solution

To find the Poynting vector, we must know and . Since is given, can
be obtained from Maxwell’s equation , i.e.

The Poynting vector is:

The average of the Poynting vector over one period , ( ), is:

 832

(since, , for detailed calculations see Pr. 25-7-1).

Also, since , as shown in Example 25-3-1, equation (***) becomes:

The average power flowing through the area of , in the negative

direction, is:

PROBLEMS

25-7-1) Show that ,( ).

25-7-2) a) Find the Poynting vector associated with the magnetic field
in free space, b) Find the average of the Poynting
vector over one period, c) Find the average power crossing the surface
specified by: .

Hint: See Example 25-7-2.

25-7-3) A parallel plates capacitor consists of two circular, conducting

plates, each of radius . The charge on the plates varies as .
Find: a) The electric field between the plates, b) The magnetic field between
the plates, c) The Poynting vector , in the region between the plates.

Fig. 25-7: Poynting vector between the plates of a capacitor.

 833

(Ans: , , ).

Hint: From the boundary conditions for perfect conductors, (eq. (25-4-7)),
, i.e. , etc. To find in the region between the
two plates, start with Maxwell’s equation , (since ), or, in
integral form, , where is a circle between the
plates, with its plane perpendicular to the axis.

25-8) Energy Associated With a Set of Currents

Let us consider the following problem.

Suppose we have closed, current carrying filamentary conductors, as

shown in Fig. 25-8.

Fig. 25-8: Current carrying filamentary conductors.

Question: How much work is required to establish the set of currents

?

The work required to establish the set of currents must be equal to the
energy stored in the magnetic field, in space, produced by the currents, (we
assume that ohmic heating is negligible). The energy stored in the magnetic
field is given by equation (22-3-8), which reads,
 834

The volume , in equation (*), includes all the current elements . For
filamentary conductors, is replaced by , whereas the volume is
composed of the sum of all the “volumes” occupied by the filamentary
conductors, i.e.

However, the integral gives the total magnetic flux through the
surface bounded by , (formula (15-3-1). If we call the total magnetic flux
linking the surface , bounded by the filamentary conductor , then
, and formula (25-8-1) becomes:

Note that the total magnetic flux linking, for example, the filamentary
conductor , is due to all currents . Indeed, if we call the self-
inductance of , the mutual inductance between and , …, the
mutual inductance between and , then

Similar expressions hold for . In view of these expressions,

equation (25-8-2) becomes:

This equation gives the work required to establish the set currents
.
 835

(Here we have used for the mutual inductance of the loops

and , instead of the symbol introduced in Section 22-2, and for the
self-inductance of the loop ).

According to formula (25-8-2), the energy stored in a solenoid, (an

inductor), carrying current , is

Recall that the energy stored in a capacitor is given by a similar formula,

which is

Formula (25-8-4) implies that the inductor current must be a continuous

function of time, i.e. at any time instant , . To justify this
assertion, let us assume that . Then, from (25-8-4), we must have
. If this were true, then, at zero time ( )
an infinite amount of power should be supplied. But infinite amount of power
lacks physical meaning, and therefore, we are obliged to accept that the
inductor current cannot be a discontinuous function of time.

Reasoning similarly, we may show that the capacitor voltage must be a

continuous function of time, i.e. , (see Problem 25-8-1).

The continuity of the inductor current and the capacitor voltage, usually
serve as initial conditions when analyzing transients in circuits containing
inductors and capacitors, (see Example 25-8-2).

Note: As we have shown, the capacitor current is given by the formula

, and the inductor voltage is given by the formula .
From these formulas it follows that the capacitor current and the inductor
voltage could be discontinuous functions of time.

Example 25-8-1: Apply formula (25-8-3) for the case of two circuits carrying
currents and . Then show that .
 836

Solution

Since , equation (*) implies:

Since , the trinomial in is positive for all , provided that its

discriminant is less than or equal to zero, i.e. when

Example 25-8-2: In the R-L circuit of Fig. 25-9, find the current .

Fig. 25-9: A series R-L circuit.

 837

Solution

The switch closes at , so for the voltage source applies to the

circuit. Application of Kirchhoff’s Voltage Law yields:

and expressing the voltages in terms of the current , we have:

Equation (**) is a first order differential equation for the sought for current
. Solving this equation we find

where is the arbitrary constant of integration. To determine , we need

one initial condition, which is obtained as follows; the inductor current must be
a continuous function of time, and in particular, . The ,
since for the voltage source is not connected, while is obtained
from (***) with , i.e.

and substituting in (***) yields:

The inductor voltage is

Note that the inductor voltage is discontinuous at , since ,

whereas , as obtained from (*****), for .
 838

Example 25-8-3: In the R-L circuit of Figure 25-9, find an expression for the
power supplied to the inductor.

Solution

If we call the power supplied to the inductor L, then:

PROBLEMS

25-8-1) Starting with equation (25-8-5) justify why the capacitor voltage
must be a continuous function of time.

25-8-2) In Example 25-8-2 find an expression for the energy stored in the
inductor as a function of time.

25-8-3) In Example 25-8-2, plot and versus . Which of

these functions are continuous?

25-8-4) In Fig. 25-9, the inductor is replaced with an uncharged capacitor .

Find the current and the capacitor voltage .

(Ans: ).

SUPPLEMENTARY PROBLEMS

25-1) In free space . Find the associated

magnetic field and the Poynting vector . Given: ,
.

(Ans: ).

25-2) In Pr. 25-1 find the average value of Poynting vector (over one
period ), and then find the average power through the triangle ,
provided that and .
 839

Hint: The unit vector, normal to the plane ABC, is .

25-3) In free space and the associated

magnetic field . Given that ,
find and .

(Ans: ).

25-4) The plane divides the whole space in two regions.

Region 2 contains the origin and has , whereas region
1 has . If ,
, find and .

25-5) Starting with Maxwell’s equations in a material with and present,

show that and satisfy the following wave equations ( and are assumed
to be constants),

Note that in free space, where and these equations are

reduced to the wave equations obtained in Section 25-3.

25-6) A voltage is applied between the two plates of a

capacitor formed by two circular parallel plates of radius . Find the electric
field , the magnetic field and the displacement current density ,
between the plates.

Hint: See Example 25-7-3.

25-7) A filamentary conductor extends from to , in free space

and caries a current in the positive direction. Find the vector
magnetic potential at the point .

(Ans: ).
 840

Hint: See Example 25-6-1.

25-8) In the RLC circuit of Fig. 25-10, the switch closes at . Assuming
that the capacitor is initially uncharged and that the initial inductor current is
zero, find the capacitor charge for , provided that .

(Ans: , where

).

Hint: Application of KVL, for , yields the following differential equation

for the capacitor charge :

The initial conditions are: , which leads to .

25-9) Repeat Pr. 25-8, provided now that .

25-10) Repeat Pr. 25-8, provided that .

Fig. 25-10: Series RLC circuit.

25-11) In the LC circuit of Fig. 25-11, the switch closes at . The initial
voltage of the capacitor is while the initial inductor current is zero. For
, (i.e. after the closing of the switch), find the voltage , the capacitor
current and the inductor current .
 841

Fig. 25-11: LC circuit, (Thomson’s oscillating circuit).

(Ans: ,

where ).

Hint: For , and differentiating with respect to

yields:

or, since and ,

where .

The initial conditions are: and . The last condition

follows from the fact that , or since , , or
, i.e. .

25-12) In Problem 25-11, if and are the energies stored in the

capacitor and the inductor, respectively, show that
and give a physical justification of the result.

25-13) Work Pr. 25-11, assuming that the initial capacitor voltage is zero
and the initial inductor current is , (flowing downwards).
 842

(Ans: ).

25-14) In free space the retarded potentials are given by

and ,( is the speed of light in vacuum). Show

that and satisfy Lorentz’s condition ( ). Then find

and and verify that these fields satisfy Maxwell’s equations in free space.
 843

APPENDIX: AXIOMATIC DERIVATION OF MAXWELL’S EQUATIONS

a) The Four Fundamental Postulates

In this book, Maxwell’s equations were derived following, more or less, the
so called, historical approach. By this we mean that all the fundamental
principles, ideas and laws, as obtained from the relevant experiments, were
introduced gradually, following the chronological order these laws were
discovered. The student acquires the necessary knowledge and the
mathematical background needed, step by step. For example, the starting
point in our analysis was Coulomb’s law in Electrostatics and Biot-Savart law in
Magnetostatics. Based on these two fundamental laws, we derived, gradually,
Maxwell’s equations for static fields. Then, in the light of Faradays law of
induction and the concept of the displacement current (introduced by Maxwell
himself), we modified the static form of Maxwell’s equations as to apply for
time – varying field as well. The final form of these equations was presented in
Chapter 25.

However, there is another way of deriving Maxwell’s equations, known as

the axiomatic approach. Here, we summarize the totality of our knowledge
and experience about electromagnetic phenomena, in four postulates. As we
shall show, from these four postulates and the principle of conservation of
energy we can derive Maxwell’s equations in their most general form. Then,
following the reverse course, we may obtain the laws of Electrostatics, the laws
of Magnetostatics, the field of steady currents, etc.

The four postulates are the following:

Postulate 1: Every charge distribution generates an electric field in space,

characterized by a vector , (called electric field), which is a well defined
function of position and time. The energy expended for the formation of the
charge distribution, is found redistributed in the space occupied by the electric
field. If we consider a small volume at the center of which the electric field
is , then the differential electric energy stored within is given by the
formula:
 844

where is the “dielectric constant of the medium”, a constant which

characterizes the ability of the medium to allow the installation of electric
fields within it.

The electric energy stored within a finite volume , is therefore equal to

Postulate 2: Every current distribution generates a magnetic field in space,

characterized by a vector , (called magnetic field), which is a well defined
function of position and time. The energy expended for the formation of the
current distribution, is found redistributed in the space occupied by the
magnetic field. If we consider a small volume at the center of which the
magnetic field is , then the differential magnetic energy stored within
is given by the formula:

where is the “permeability of the medium”, a constant which

characterizes the ability of the medium to allow the installation of a magnetic
fields within it.

The magnetic energy stored within a finite volume , is therefore equal

to

Postulate 3 (Relaxation of the electric field): The electric energy stored in

any electric field is converted continuously into heat, known as “Joule Heat or
Joule loss”, in honor of James Prescott Joule (1818-1889). Heat loss is an
irreversible phenomenon since heat, continuously escapes from the field and
does not return back. If is the electric field at the center of a small volume
, then the amount of heat generated within , during a time interval
, is given by the formula:
 845

where is the “relaxation constant of the medium”, (notice that the units
of must be ). Formula 5 is simplified to the following,

where is called the conductivity of the medium. Notice that in

free space (vacuum) and for perfect dielectrics, and therefore .
This means that an electric field , once installed within a perfect dielectric, (or
in vacuum), stays there forever, since there is no conversion of electric energy
to heat. On the other hand, for perfect conductors, and hence ,
and this, in turn, implies that, for static configurations, the electric field
inside a perfect conductor must be zero, (see Section 13-6, Decay of free
charges).

Postulate 4: In every Electromagnetic Field there exists a continuous flow of

energy, from the positions of the field sources, (charges and/or currents),
towards the loads, which may be situated at a finite or at an infinite distance
from the sources. The “power density or power flux ”, at some position of
the field, is defined as a vector quantity whose magnitude is equal to the
power per unit area, placed normally to the direction of the energy flow,
whereas its direction coincides with the direction of transmission of the EM
energy. The power density is given by the formula:

(Notice that is, in fact, the Poynting vector, introduced in Section 25-7).

b) Derivation of Maxwell’s Equations

Let us now consider a volume , bounded by a closed surface , as shown

in Fig. 1. As usually, we call and the electric energy and the magnetic
energy, respectively, stored within the volume at the time instant . The
total EM energy stored within , at the time instant , is therefore,
 846

Any change in the total stored energy between the time instants and
, i.e. during the time interval , is caused by:

1) The conversion to heat part of the electric energy stored in , and

2) The flow of EM energy stored in to the surrounding space, through the

closed surface , which bounds the volume .

Fig. 1: Energy stored in the volume .

If we therefore call:

the increase in the electric energy stored within the volume , during
the time interval ,

the increase in the magnetic energy stored within the volume ,

during the time interval ,

the heat generated within the volume , during the time interval ,

the amount of EM energy transmitted, within the time interval ,

through the closed surface , in the surrounding space,

then, energy balance requires that the heat generated within , during the
time interval , and the energy transmitted outwards, during the same time
interval , should only come from a corresponding decrease of the electric
and the magnetic energy stored within , during , i.e.
 847

The heat generated within the volume , during the time interval , is
(since ),

The EM energy transmitted through , during the time interval , is

The increase in the electric energy stored within , during , is

or, since ,

and similarly,

In view of equations (*), (**), (***) and (****), equation (9) becomes, (let
the reader check it):
 848

or, transforming the surface integral to a volume integral with the aid of
Gauss’s theorem,

Since this equation should be true for any, arbitrary volume , it follows
that the integrand must be zero identically, i.e.

Now, using the vector identity

and substituting in (11) we find:

Since this equation must be true for any, arbitrary fields and , it follows
that the following equations must be satisfied, identically:

If we define
 849

the equations in formula (13) become,

which are, in fact, the two “curl” Maxwell’s equations.

To derivation of the two “divergence” equations, appearing in Maxwell’s

equations, follow from equation (15). Indeed, if we take the divergence of both
sides of the second equation, we obtain:

or, since the divergence of the curl is identically equal to zero,

This means that the does not vary with time. Before the installation
of a magnetic field in a region in space, , and therefore, .
After the installation of a magnetic field in the region, the , since it
cannot change with time, must be zero as well.

Now, taking the divergence of both sides of the first equation in (15), results
in the following:

and comparing with the continuity equation , (which

expresses the conservation of the electric charge), it follows that .
 850

c) Electrostatics, Magnetostatics and the Field of Steady Currents as

Special Cases of Maxwell’s Equations

Having thus obtained the general form of Maxwell’s equations, using the
“axiomatic approach”, we may now go backwards and derive the general laws
of Electrostatics, Magnetostatics and the Field of Steady Currents, as special
cases.

1) Static state, by definition, is a state where there is no time variation of

the quantities involved, and, also, for the maintenance of this static state there
is no need for a continuous energy supply, (see Section 13-1).

No dependence on time means that the partial derivatives with respect to

the time vanish, ( ). In this case, the equations in (15) become:

Further simplifications of these equations, (for static fields), follow from the
fact that at no position, in a static field, there is conversion of electric energy
into heat, (since this would require a continuous supply of energy to keep the
electric field energy constant, but, for static states, by definition, there is no
energy supply). We are thus forced to accept that, in static fields, ,
(from eq. (6)). In turn, this means that, for static electric fields, either , or
. The first condition, implies that static electric fields ( ) can be
developed only within perfect dielectrics or in vacuum, ( ). The second
condition implies that inside good conductors ( ), the electric field must
be zero ( ). Of course, we have already derived these properties, but from
a different point of view.

Since , it follows that , for static fields, and, finally, we obtain:

 851

The fact that these two equations are independent, shows that static
electric and static magnetic fields may exist simultaneously, but
independently one from the other, in the same region in space.

2) Steady state, by definition, is a state where there is no time variation of

the quantities involved, but, for the maintenance of this steady state a
continuous energy supply is required, (see Section 13-1).

In this case, Maxwell’s equations in (15) become:

The first equation in (19) is, in fact, Ampere’s circuital law, which, in Chapter
16, was derived from the Biot-Savart law.

Contrary to what happens in static fields, in steady fields there are regions
where , (energy is continuously supplied to compensate for the heat
loss). In these regions, , and this shows that in steady fields, we
may have conductors ( ) in the interior of which an electric field exists
( ). This electric field sets the electric charges in motion, and this results in
a current flowing within the conductors, (see also Section 13-13).

In steady fields, the electric field is irrotational (as it is in static fields),

because , but the magnetic field is not irrotational, because
.

d) Quasi-static Approximation

A very useful approximation of Maxwell’s equations, known as “Quasi-static

approximation”, is obtained when the conduction current ( ) is much greater
than the displacement current ( ). In this case, Maxwell’s equations in
(15) become:
 852

Note that in equation (20), the fields and are functions of the
spatial coordinates and the time . Still, in the Quasi-static approximation
regime, we may use , (which strictly speaking holds true for time
independent fields), to determine, for instance, the time varying field
produced by a conductor carrying an alternating current. The quasi-static
approximation breaks down as we move to very high frequencies, where now
the displacement current predominates over the conduction current, (see
Comment in Example 24-4-3).

e) Derivation of Coulomb’s Law from Maxwell’s Equations

Coulomb’s law was the first experimental law studied in this textbook. The
development of Electrostatics was, in fact, based on Coulomb’s law. We
conclude this book by going back, and presenting a derivation of Coulomb’s
law, but, this time, starting with the Maxwell’s equations.

1) Let us consider a static charge distribution in a region characterized by a

constant permittivity . A point charge is placed at some point of the region.
Electrical forces act on the point charge and it is our aim to find a general
expression for the force exerted on .

Assuming that the point charge , under the action of the electric force ,
is displaced by a small , then, the work done by will be . Energy
balance requires that this differential work should be at the expense of the
energy stored in the electric field. If is positive, then the energy stored
in the field must be reduced by an equal amount. If is negative, then
the energy stored in the field must be increased by an equal amount.

If we call the incremental variation of the energy stored in the electric

field, while the point charge is displaced by , then we must have:

or, since, in general, ,

 853

and since must be true for every , (arbitrary), we conclude that, in

general,

2) Let us now consider a fixed point charge , positioned at the origin, as

in Fig. 2.

Fig. 2: Derivation of Coulomb’s law.

A second charge is placed at a point , determined by its radius vector .

Starting with the fundamental equation of the static electric field , it
follows that the electric field may be expressed as the (negative) gradient of a
scalar function , i.e. , as it was shown in Chapter 5. Also, the
potential must satisfy Laplace’s equation, i.e. ,( ), as it was also
shown in Chapter 5. This equation was solved for a point charge in Section 5-4,
the result being

and the electric field at the point , due to the charge , is

Finally, we shall use the expression for the energy stored in the field
generated by two point charges and , separated by a distance , as in
 854

Figure 2. This energy is given by formula (6-2-1), which in our case, (two point
charges), becomes:

Note that in our so far analysis, we have not made any reference to
Coulomb’s law. Everything was, in fact, derived from the fundamental
equations and .

Now, using formulas (21), and (24), we obtain an expression for the force
exerted on by the charge :

and this completes the derivation of the Coulomb’s law, starting with
Maxwell’s equations.