Physics II by Halliday Resnick

Download as pdf or txt
Download as pdf or txt
You are on page 1of 677

SELECTED PHYSICAL CONSTANTS

I See Appendix A for a more complete list I

Speed of light c 3.00 X 108 meters/sec 1.86 X 10 5


miles/sec

Mass-energy relation 931 Mev /amu = 8.99 X 10 16 joules/kg

Gravitational constant G 6.67 X 10- 11 nt-m 2 /kg 2

Universal gas constant R 8.31 joules/mole °K = 1.99 cal/mole °K


= 0.0823 li-atm/mole °K

Triple point of water Ttr 273.16 °K

Permeability constant µo 1.26 X 10- 5 henry /meter

Permittivity constant EO 8.85 X 10- 12 farad/meter

Avogadro's constant No 6.02 X 10 23 molecules/mole

Boltzmann's constant k 1.38 X 10- 23 joule/molecule °K

Planck's constant h 6.63 X 10- 34 joule-sec

Elementary charge e 1.60 X 10- 19 coul

Electron rest mass me 9.11 x 10- 31 kg

Electron charge-to-mass ratio e/me 1.76 X 10 11 coul/kg

Proton rest mass mp 1.67 x 10- 27 kg

Electron magnetic moment µe 9.27 X 10- 24 joule/tesla


SELECTED PHYSICAL PROPERTIES

Density of air (STP) 1.29 kg/meter 3

Density of water (20° C) 1.00 X 10 3 kg/meter 3

Density of mercury (20° C) 13.5 X 10 3 kg/meter 3

Speed of sound in dry air (STP) 331 meters/sec = 1090 ft/sec

Acceleration of gravity (standard) 9.81 meters/sec 2 = 32.2 ft/sec 2

Standard atmosphere 1.01 X 10 5 nt/meter 2 = 14.7 lb/in. 2


= 760 mm-Hg
Mean radius of the earth 6.37 X 10 6 meters = 3960 miles

}Tean earth-sun distance 1.49 X 10 8 km = 92.9 X 10 6 miles

.l\Iean earth-moon distance 3.80 X 10 5 km = 2.39 X 10 5 miles

Mass of earth 5.98 X 10 24 kg

Heat of fusion of water (0° C, 1 atm) 79.7 cal/gm

Heat of vaporization of ,rnter (100° C, 1 atm) 539 cal/gm

Melting point of ice 0.00° C = 273.15° K

Ratio of specific heats (,) for air (20° C) 1. 40

Wavelength of the sodium yellow doublet 5892 A

Index of refraction of water (@, ,58!)2 A) 1. 33

Index of refraction of crown glass (@ 5892 A) 1.52


Phy sics PART II
Physics
PART II
DAVID HALLIDAY ROBERT RESNICK
Professor of Physics Professor of Physics
University of Pittsburgh Rensselaer Polytechnic Institute

John Wiley & Sons, Inc. New York• London• Sydney


BOOKS BY HALLIDAY (D.)

Introductory Nuclear Physics. Second Edition


BOOKS BY HALLIDAY (D.) AND RESNICK (R.)
Physics for Students of Science and Engineering
Part II, Second Edition
Physics for Students of Science and Engineering
Parts I First Edition and II Second Edition Combined
Physics, Part II
Physics, Parts I and II Combined
BOOKS BY RESNICK
Introduction to Special Relativity
BOOKS BY RESNICK (R.) AND HALLIDAY (D.)
Physics for Students of Science and Engineering
Part I
Physics, Part I

10

Copyright © 1960, 1962 by John Wiley & Sons, Inc.

All rights reserved. This book or any part


thereof must not be reproduced in any form
without the written permission of the publisher.

Library of Congress Catalog Card Number: 62-15336

Printed in the United States of America


Physics, Part II, is a revised printing of the second edition of Physics for
Students of Science and Engineering. In this (1966) new printing of Part II,
the following revisions have been made.
1. A supplementary set of 142 problems has been provided. These have
been carefully assembled so as to permit, when combined with the numerous
problems in the original set, a wider choice of area of interest and level of
difficulty than heretofore available.
2. A new Topical Supplement, on the differential form of Maxwell's
equations and the electromagnetic wave equation, has been written for
reference of students whose instructors choose to discuss this more
advanced material.
3. The original appendices have been expanded and modernized and
some new ones have been added in order to increase the usefulness and
accuracy of such reference material. Front and back end-papers have been
changed accordingly.
4. Pagination, numbered references, symbols and units, have been
changed, where necessary, for consistency with the new Part I and the
above revisions of Part II.
Preface to the Second Edition

This revision of Part II of Physics for Students of Science and Engineering,


which we hope will make our book more useful to students and instructors,
is based on classroom experience at many institutions during the last two
years. The basic outline of the book and its underlying philosophy remain
unchanged. We have, however, modified the treatment of many topics,
among them Ampere's, Faraday's, Gauss's and Lenz's laws and Huygens'
principle, to make them clearer and more explicit than before. The experi-
mental basis of the laws of physics has received increased stress, and new
experimental material, such as the radiation pressure measurements of
Nichols and Hull, optical masers, relativistic particle accelerators, and fiber
optics illustrate the basic concepts. More than half the figures have been
redrawn for greater clarity, and new figures and photographs have been
inserted in key places. The questions and problems have been re-evalu-
ated, and more than 200 are new.
In a few places we have supplied new sections either to extend or to
improve the understanding of the basic material. For example, there are
sections in reduced type on the electric vectors (E, D, and P) and on the
magnetic vectors (B, H, and M). RC circuits are discussed, the Compton
effect has been included in the treatment of the wave-particle duality, and
a discussion of electromagnetic fields as seen by observers in relative motion
has been added.
Other significant changes are (1) references to the literature available to
students have been supplemented and brought up to date; (2) the sign
vii
viii PREFACE

conventions for geometrical optics have been put on a more physical basis;
(3) attention has been called to examples or problems that contain impor-
tant new material by using italicized titles; and (4) we have further sys-
temized and simplified the notation in several places, particularly in wave
optics.
We continue to be indebted to Dr. Benjamin Chi for assistance in pre-
paring the illustrations and for his detailed criticism of the optics chapters.
One of us (DH) wishes to thank Professor G. Wataghin for courtesies
extended at the University of Turin while this revision was in progress.
Finally, we extend special thanks to those students and instructors who
have been kind enough to write to us about their experiences with the first
edition.
David Halliday
University of Pittsburgh
January 1, 1962 Pittsburgh, Pennsylvania, 15213
Robert Resnick
Rensselaer Polytechnic Institute
Troy, New York, 12181
Preface to First Edition

The time lag between developments in basic science and their application
to engineering practice has shrunk enormously in the past few decades.
The base of engineering, once largely empirical, is now largely scientific.
Today the need is to stress principles rather than specific procedures, to
select areas of contemporary interest rather than of past interest, and to
condition the student to the atmosphere of change he will encounter during
his career. These developments require a revision of the traditional course
in general physics for engineers and scientists.
The most frequent criticisms made in varying degrees of textbooks used
in such a course are these: (a) the content is encyclopedic in that topics are
not treated with sufficient depth, the discussions are largely descriptive
rather than explanatory and analytical, and too many topics are surveyed;
(b) the content is not sufficiently "modern," and applications are drawn
mostly from past engineering practice rather than from contemporary
physics; (c) the organization of the material is too compartmentalized to
reveal the essential unity of physics and its principles; (d) the approach is
highly deductive and does not stress sufficiently the connection between
theory and experiment. Of course, it is unlikely that a textbook will ever
be written that is not criticized on one ground or another.
In writing this textbook we have been cognizant of these criticisms and
have given much thought to ways of meeting them. We have considered
the possibility of reorganizing the subject matter. The adoption of an
atomic approach from the beginning or a structure built around energy in
ix
x PREFACE

its various aspects suggest themselves. We have concluded that our goals
can best be achieved by modifying the selection and treatment of topics
within the traditional organization. To shuffle freely the cards of subject
matter content or to abandon entirely a sequence which represents the
growth of physical thought invites both a failure to appreciate the New-
tonian and Maxwellian syntheses of classical physics and a superficial
understanding of modern physics. A solid underpinning of classical
physics is essential to build the superstructure of contemporary physics in
our opinion.
To illustrate how we hope to achieve our goals within this framework, we
present here the principal features of our book.

1. Many topics are treated in greater depth than has been customary
heretofore, and much contemporary material has been woven into the body
of the text. For example, gravitation, kinetic theory, electromagnetic
waves, and physical optics, among others, are treated in greater depth.
Contemporary topics, such as atomic standards, collision cross section,
intermolecular forces, mass-energy conversion, isotope separation, the Hall
effect, the free-electron model of conductivity, nuclear stability, nuclear
resonance, and neutron diffraction, are discussed where they are pertinent.
To permit this greater depth and inclusion of contemporary material, we
have omitted entirely or treated only indirectly much traditional material,
such as simple machines, surface tension, viscosity, calorimetry, change of
state, humidity, pumps, practical engines, musical scales, architectural
acoustics, electrochemistry, thermoelectricity, motors, alternating-current
circuits, electronics, lens aberrations, color, photometry, and others.
2. We have tried to reveal the unity of physics in many ways. Through-
out the book we stress the general nature of key ideas common to all areas
of physics. For example, the conservation laws of energy, linear momen-
tum, angular momentum, and charge are used repeatedly. Wave concepts
and properties of vibrating systems, such as resonance, are used in mechan-
ics, sound, electromagnetism, optics, atomic physics, and nuclear physics.
The field concept is applied to gravitation, fluid flow, electromagnetism,
and nuclear physics.
The interrelation of the various disciplines of physics is emphasized by
the use of physical and mathematical analogies and by similarity of method.
For example, the correspondences between the mass-spring system and the
LC circuit or between the acoustic tube and the electromagnetic cavity
are emphasized, and the interweaving of microscopic and macroscopic
approaches is noted in heat phenomena and electrical and magnetic
phenomena. We have tried to make a smooth transition between particle
mechanics and kinetic theory, stressing that, in their classical aspects, both
belong to the Newtonian synthesis. We have also sought a smooth transi-
tion between electromagnetism and wave optics, pointing frequently to the
Maxwellian synthesis.
PREFACE xi

We discuss the limitations of classical ideas and the domain of their


validity, and we emphasize the generalizing nature of contemporary ideas
applicable in a broader domain. Throughout we aim to show the relation
of theory to experiment and to develop an awareness of the nature and uses
of theory.
3. Our approach to quantum physics is not the traditional descriptive
one. Rather we seek to develop the contemporary concepts fairly rigor-
ously, at a length and depth appropriate to an introductory course. In the
early chapters we pave the way by pointing to the limitations of classical
theory, by stressing the aspects of classical physics that bear on contempo-
rary physics, and by choosing illustrative examples that have a modern
flavor. Thus we stress fields rather than circuits, particles rather than
extended bodies, and wave optics rather than geometrical optics. Among
the illustrative examples are molecular potential energy curves, binding
energy of a deuteron, nuclear collisions, the nuclear model of the atom, the
Thomson atom model, molecular dipoles, drift speed of electrons, stability
of betatron orbits, nuclear magnetic resonance, the red shift, and others too
numerous to mention.
The point of view is that of developing the fundamental ideas of quantum
physics. The customary descriptive chapter on nuclear physics is, for
example, not present. Instead, the wave-particle duality, the uncertainty
principle, the complementarity principle, and the correspondence principle
are stressed.
4. The mathematical level of our book assumes a concurrent course in
calculus. The derivative is introduced in Chapter 3 and the integral in
Chapter 7. The related physical concepts of slope and area under a curve
are developed steadily. Calculus is used freely in the latter half of the
book. Simple differential equations are not avoided, although no formal
procedures are needed or given for solving them. Vector notation and
vector algebra, including scalar and vector products, are used throughout.
Displacement is taken as the prototype vector, and the idea of invariance
of vector relations is developed.
5. The number of problems is unusually large, but few are "plug-in"
problems. Many involve extensions of the text material, contemporary
applications, or derivations. The questions at the end of each chapter are
intended to be thought-provoking; they may serve as the basis for class
discussion, for essay papers, or for self-study. Only rarely can the ques-
tions be answered by direct quotation from the text.
6. The book contains an unusually large number of worked-out examples,
with the "plug-in" variety used only to emphasize a numerical magnitude.
Algebraic, rather than numerical, solutions are stressed. Examples some-
times extend the text treatment or discuss the fine points, but usually they
are applications of the principles, often of contemporary physics.
xii PREFACE

7. The textbook has been designed to fit physics courses of various


lengths. In small print there is a great deal of supplementary material of
an advanced, historical, or philosophical character, to be omitted or
included to varying degrees depending on interest and course length. In
addition, many chapters may be regarded as optional. Each teacher will
make his own choice. At our institutions Chapter 14 (statics of rigid
bodies) and Chapters 41 and 42 (geometrical optics) are omitted. Other
possibilities suggested, depending on emphasis or depth desired, or the
nature of succeeding courses, are Chapter 12 (rotational dynamics), Chap-
ters 17 and 18 (fluids), Chapter 24 (kinetic theory-II), Chapter 32 (emf
and circuits), Chapter 46 (polarization), and Chapters 47 and 48 (quantum
physics).
8. We have adopted the mks system of units throughout, although the
British engineering system is also used in mechanics. Having observed the
gradual exclusion, year by year, of the cgs system from advanced textbooks,
we have seen fit to limit ourselves to the bare definition of the basic cgs
quantities. An extensive list of conversion factors appears in Appendix G.
We wish to thank the engineering and science students at both Rensselaer
Polytechnic Institute and the University of Pittsburgh who have borne with
us through two successive preliminary editions. Constructive criticisms
from our colleagues at each institution and from some eight reviewers have
resulted in many changes. Benjamin Chi of R.P.I. has been of major
service in all aspects of the preparation of the manuscript. Finally, we
express our deep appreciation to our wives, not only for aid in typing and
proofreading but for the patience and encouragement without which this
book might never have been written.
David Halliday
University of Pittsburgh
January 1960 Pittsburgh, Pennsylvania, 15213
Robert Resnick
Rensselaer Polytechnic Institute
Troy, New York, 12181
Contents

26 CHARGE AND MATTER 647


26-1 Electromagnetism -A Preview 647
26-2 Electric Charge 648
26-3 Conductors and Insulators 649
26-4 Coulomb's Law 650
26-5 Charge Is Quantized 654
26-6 Charge and Matter 655
26-7 Charge Is Conserved 660

27 THE ELECTRIC FIELD 663


27-1 The Electric Field 663
27-2 The Electric Field Strength E 665
27-3 Lines of Force 665
27-4 Calculation of E 670
27-5 A Point Charge in an Electric Field 675
27-6 A Dipole in an Electric .Field 678

28 GAUSS'S LAW 686


28-1 Flux of the Electric Field 686
28-2 Gauss's Law 690
28-3 Gauss's Law and Coulomb's Law 690
28-4 An Insulated Conductor 692
28-5 Experimental Proof of Gauss's and Coulomb's Laws 693
28-6 Gauss's Law-Some Applications 695
28-7 The Nuclear Model of the Atom 700
xiii
xiv CONTENTS

29 ELECTRIC POTENTIAL 708


29-1 Electric Potential 708
29-2 Potential and Field Strength 711
29-3 Potential Due to a Point Charge 715
29-4 A Group of Point Charges 716
29-5 Poteptial Due to a Dipole 718
29-6 Electric Potential Energy 722
29-7 Calculation of E from V 724
29-8 An Insulated Conductor 730
29-9 The Electrostatic Generator 733

30 CAPACITORS AND DIELECTRICS 741


30-1 Capacitance 741
30-2 Calculating Capacitance 7 45
30-3 Paralle-Plate Capacitor with Dielectric 748
30-4 Dielectrics-An Atomic View 750
30-5 Dielectrics and Gauss's Law 753
30-6 Three Electric Vectors 756
30-7 Energy Storage in an Electric Field 760

31 CURRENT AND RESISTANCE 770


31-1 Current and Current Density 770
31-2 Resistance, Resistivity, and Conductivity 774
31-3 Ohm's Law 778
31-4 Resistivity-An Atomic View 780
31-5 Energy Transfers in an Electric Circuit 783

32 ELECTROMOTIVE FORCE AND CIRCUITS 789


32-1 Electromotive Force 789
32-2 Calculating the Current 792
32-3 Other Single-Loop Circuits 793
32-4 Potential Differences 795
32-5 Multiloop Circuits 798
32-6 Measuring Currents and Potential Differences 800
32-7 The Potentiometer 801
32-8 RC Circuits 802

33 THE MAGNETIC FIELD 814


33-1 The Magnetic Field 814
33-2 The Definition of B 816
33-3 Magnetic Force on a Current 819
33-4 Torque on a Current Loop 823
33-5 The Hall Effect 827
33-6 Circulating Charges 829
33-7 The Cyclotron 831
33-8 Thomson's Experiment 835
CONTENTS xv
34 AMPERE'S LAW 844

34-1 Ampere's Law 844


34-2 B Near a Long Wire 848
34-3 Magnetic Lines of Induction 851
34-4 Two Parallel Conductors 853
34-5 B for a Solenoid 856
34-6 The Biot-Savart Law 859

35 FARADAY'S LAW 870

35-1 Faraday's Experiments 870


35-2 Faraday's Law of Induction 872
35-3 Lenz's Law 873
35-4 Induction-A Quantitative Study 875
35-5 Time-Varying Magnetic Fields 880
35-6 The Betatron 885
35-7 Induction and Relative Motion 888

36 INDUCTANCE 899

36-1 Inductance 899


36-2 Calculation of Inductance 901
36-3 An LR Circuit 902
36-4 Energy and the Magnetic Field 907
36-5 Energy Density and the Magnetic Field 910

37 MAGNETIC PROPERTIES OF MATTER 915

37-1 Poles and Dipoles 915


37-2 Gauss's Law for Magnetism 920
37-3 Paramagnetism 920
37-4 Diamagnetism 923
37-5 Ferromagnetism 926
37-6 Nuclear Magnetism 931
37-7 Three Magnetic Vectors 934

38 ELECTROMAGNETIC OSCILLATIONS 943

38-1 LC Oscillations 943


38-2 Analogy to Simple Harmonic Motion 947
38-3 Electromagnetic Oscillations-Quantitative 948
38-4 Forced Oscillations and Resonance 952
38-5 Lumped and Distributed Elements 954
38-6 Electromagnetic Cavity Oscillator 955
38-7 Induced Magnetic Fields 959
38-8 Displacement Current 962
38-9 Maxwell's Equations 963
38-10 Maxwell's Equations and Cavity Oscillations 963
xvi CONTENTS

39 ELECTROMAGNETIC WAVES 971


39-1 Transmission Lines 971
39-2 Coaxial Cable-Fields and Currents 974
39-3 Waveguide 977
39-4 Radiation 979
39-5 Traveling Waves and Maxwell's Equations 982
39-6 The Poynting Vector 986

40 NATURE AND PROPAGATION OF LIGHT 993


40-1 Light and the Electromagnetic Spectrum 993
40-2 Energy and Momentum 995
40-3 The Speed of Light 998
40-4 Moving Sources and Observers 1003
40-5 Doppler Effect 1006

41 REFLECTION AND REFRACTION-


PLANE WAVES AND PLANE SURFACES 1013
41-1 Reflection and Refraction 1013
41-2 Huygens' Principle 1018
41-3 Huygens' Principle and the Law of Reflection 1020
41-4 Huygens' Principle and the Law of Refraction 1021
41-5 Total Internal Reflection 1024
41-6 Fermat's Principle 1028

42 REFLECTION AND REFRACTION-


SPHERICAL WAVES AND SPHERICAL SURFACES 1035
42-1 Geometrical Optics and Wave Optics 1035
42-2 Spherical Waves-Plane Mirror 1038
42-3 Spherical Waves-Spherical Mirror 1041
42-4 Spherical Refracting Surface 1049
42-5 Thin Lenses 1055

43 INTERFERENCE 1068
43-1 Young's Experiment 1068
43-2 Coherence 1074
43-3 Intensity in Young's Experiment 1077
43-4 Adding Wave Disturbances 1080
43-5 Interference from Thin Films 1083
43-6 Phase Changes on Reflection 1089
43-7 Michelson's Interferometer 1090
43-8 Michelson's Interferometer and Light Propagation 1092

44 DIFFRACTION 1099
44-1 Introduction 1099
44-2 Single Slit 1102
44-3 Single Slit-Qualitative 1105
CONTENTS xvii
44-4 Single Slit-Quantitative 1108
44-5 Diffraction at a Circular Aperture 1112
44-6 Double Slit 1115

45 GRATINGS AND SPECTRA 1123

45-1 Introduction 1123


45-2 Multiple Slits 1123
45-3 Diffraction Gratings 1128
45-4 Resolving Power of a Grating 1132
45-5 X-ray Diffraction 1134
45-6 Bragg's Law 1140

46 POLARIZATION 1147

46-1 Polarization 1147


46-2 Polarizing Sheets 1150
46-3 Polarization by Reflection 1153
46-4 Double Refraction 1155
46-5 Circular Polarization 1163
46-6 Angular Momentum of Light 1167
46-7 Scattering of Light 1167
46-8 Double Scattering 1169

47 LIGHT AND QUANTUM PHYSICS 1173

4 7-1 Sources of Light 1173


47-2 Cavity Radiators 1174
47-3 Planck's Radiation Formula 1177
47-4 Photoelectric Effect 1179
47-5 Einstein's Photon Theory 1182
47-6 The Compton Effect 1184
47-7 Line Spectra 1188
47-8 The Hydrogen Atom 1191
47-9 The Correspondence Principle 1196

48 WAVES AND PARTICLES 1200

48-1 Matter Waves 1200


48-2 Atomic Structure and Standing Waves 1203
48-3 Wave Mechanics 1204
48-4 Meaning of '1i 1208
48-5 The Uncertainty Principle 1210

SUPPLEMENTARY TOPIC V 1

The Differential Form of Maxwell's Equations 'and the Electromag-


netic "\Vave Equation

SUPPLEMENTARY PROBLEMS 9
xviii CONTENTS

APPENDICES 29

A Definition of Standards and Fundamental and Derived Physical


Constants 29
B Miscellaneous Terrestrial Data 32
C The Solar System 34
D Periodic Table of the Elements 36
E The Particles of Physics 37
F Symbols, Dimensions, and Units for Physical Quantities 39
G Conversion Factors 41
H' Mathematical Symbols and the Greek Alphabet 49
Mathematical Formulas 52
J Values of Trigonometric Functions 53
K Nobel Prize Winners in Physics 55
L The Gaussian System of Units 58
ANSWERS TO PROBLEMS 61
ANSWERS TO SUPPLEMENTARY PROBLEMS 68
INDEX 73
Physics PART II
Charge and Matter
CHAPTER 26

26-1 Electromagnetism-A Preview


The science of electricity has its roots in the observation, known to Thales
of Miletus in 600 B.c., that a rubbed piece of amber will attract bits of straw.
The study of magnetism goes back to the observation that naturally occur-
ring "stones" (that is, magnetite) will attract iron. These two sciences de-
veloped quite separately until 1820, when Hans Christian Oersted (1777-
1851) observed a connection between them, namely, that an electric current
in a wire can affect a magnetic compass needle (Section 33-1).
The new science of electromagnetism was developed further by many
workers, of whom one of the most important was Michael Faraday (1791-
1867). It fell to James Clerk Maxwell (1831-1879) to put the laws of electro-
magnetism in the form in which we know them today. These laws, often
called Maxwell's equations, are displayed in Table 38-3, which the student
may want to examine at this time. These laws play the same role in electro-
magnetism that Newton's laws of motion and of gravitation do in mechanics.
Although Maxwell's synthesis of electromagnetism rests heavily on the
work of his predecessors, his own contribution is central and vital. Maxwell
deduced that light is electromagnetic in nature and that its speed can be
found by making purely electric and magnetic measurements. Thus the
science of optics was intimately connected with those of electricity and of
magnetism. The scope of Maxwell's equations is remarkable, including as it
does the fundamental principles of all large-scale electromagnetic and optical
devices such as motors, cyclotrons, electronic computers, radio, television,
microwave radar, microscopes, and telescopes.
647
648 CHARGE AND MATTER Chap. 26
The development of classical electromagnetism did not end with Maxwell.
The English physicist Oliver Heaviside (1850-1925) and especially the Dutch
physicist H. A. Lorentz (1853-1928) contributed substantially to the clari-
fication of Maxwell's theory. Heinrich Hertz (1857-1894) * took a great
step forward when, more than twenty years after Maxwell set up his theory,
he produced in the laboratory electromagnetic "Maxwellian waves" of a kind
that we would now call short radio waves. It remained for Marconi and
others to exploit this practical application of the electromagnetic waves of
Maxwell and Hertz.
Present interest in electromagnetism takes two forms. At the level of
engineering applications Maxwell's equations are used constantly and
universally in the solution of a wide variety of practical problems. At the
level of the foundations of the theory there is a continuing effort to extend its
scope in such a way that electromagnetism is revealed as a special case of a
more general theory. Such a theory would also include (say) the theories of
gravitation and of quantum physics. This grand synthesis has not yet been
achieved.

26-2 Electric Charge


The rest of this chapter deals with electric charge and its relationship to
matter. We can show that there are two kinds of charge by rubbing a glass
rod with silk and hanging it from a long silk thread as in Fig. 26-1. If a
second glass rod is rubbed with silk and held near the rubbed end of the first
rod, the rods will repel each other. On the other hand, a hard-rubber rod
rubbed with fur will attract the glass rod. Two hard-rubber rods rubbed with
fur will repel each other. We explain these facts by saying that rubbing a
rod gives it an electric charge and that the charges on the two rods exert
forces on each other. Clearly the charges on the glass and on the hard rubber
must be different in nature.
Benjamin Franklin (1706-1790), who, among his other achievements, was
the first American physicist, named the kind of electricity that appears on
the glass positive and the kind that appears on the hard rubber negative;
these names have remained to this day. We can sum up these experiments
by saying that like charges repel and unlike charges attract.
Electric effects are not limited to glass rubbed with silk or to hard rubber
rubbed with fur. Any substance rubbed with any other under suitable
conditions will become charged to some extent; by comparing the unknown
charge with a charged glass rod or a charged hard-rubber rod, it can be labeled
as either positive or negative.
The modern view of bulk matter is that, in its normal or neutral state, it
contains equal amounts of positive and negative electricity. If two bodies
like glass and silk are rubbed together, a small amount of charge is trans-
ferred from one to the other, upsetting the electric neutrality of each. In this
case the glass would become positive, the silk negative.

* "Heinrich Hertz," by P. and E. Morrison, Scientific American, December 1957.


Sec. 26-3 CONDUCTORS AND INSULATORS 649

Silk
thread

Fig. 26-1 Two positively charged


glass rods repel each other.

26-3 Conductors and Insulators


A metal rod held in the hand and rubbed with fur will not seem to develop
a charge. It is possible to charge such a rod, however, if it is furnished with
a glass or hard-rubber handle and if the metal is not touched with the hands
while rubbing it. The explanation is that metals, the human body, and the
earth are conductors of electricity and that glass, hard rubber, plastics, etc.,
are insulators (also called dielectrics).
In conductors electric charges are free to move through the material,
whereas in insulators they are not. Although there are no perfect insulators,
the insulating ability of fused quartz is about 1025 times as great as that of
copper, so that for many practical purposes some materials behave as if they
were perfect insulators.
In metals a fairly subtle experiment called the Hall effect (see Section 33-5)
shows that only negative charge is free to move. Positive charge is as im-
mobile as it is in glass or in any other dielectric. The actual charge carriers in
metals are the free electrons. When isolated atoms are combined to form a
metallic solid, the outer electrons of the atom do not remain attached to
individual atoms but become free to move throughout the volume of the
solid. For some conductors, such as electrolytes, both positive and negative
charges can move.
650 CHARGE AND MATTER Chap. 26
A class of materials called semiconductors is intermediate between conductors and
insulators in its ability to conduct electricity. Among the elements, silicon and
germanium are well-known examples. In semiconductors the electrical conductivitv
can often be greatly increased by adtling very small amounts of other elements; trac;s
of arsenic or boron are often added to silicon for this purpose. Semiconductors have
many practical applications, among which is their use in the construction of tran-
sistors. The mode of action of semiconductors cannot be described adequately with-
out some understanding of the basic principles of quantum physics.

26-4 Coulomb's Law


Charles Augustin de Coulomb (1736-1806) in 1785 first measured electrical
attractions and repulsions quantitatively and deduced the law that governs
them. His apparatus, shown in Fig. 26-2, resembles the hanging rod of
Fig. 26-1, except that the charges in Fig. 26-2 are confined to small spheres
a and b.
If a and b are charged, the electric force on a will tend to twist the sus-
pension fiber. Coulomb canceled out this twisting effect by turning the
suspension head through the angle e needed to keep the two charges at the
particular distance apart in which he was interested. The angle e is then a
relative measure of the electric force acting on charge a. The device of
Fig. 26-2 is called a torsion balance; a similar arrangement was used later by
Cavendish to measure gravitational attractions (Section 16-3).
Coulomb's first experimental results can be represented by

1
Fo::-·
r2

F is the magnitude of the force that acts on each of the two charges a and b;
r is their distance apart. These forces, as Newton's third law requires, act
along the line joining the charges but point in opposite directions. Note that
the magnitude of the force on each charge is the same, even though the
charges may be different.
Coulomb also studied how the electrical force varied with the relative size
of the charges on the spheres of his torsion balance. For example, if we touch
a charged conducting sphere to an exactly similar but uncharged conducting
sphere, the original charge must divide equally between the spheres. By
such techniques Coulomb extended the inverse square relationship to

F o:: q1q2, (26-1)


r2

where q1 and q2 are relative measures of the charges on spheres a and b.


Equation 26-1, which is called Coulomb's law, holds only for charged objects
whose sizes are much smaller than the distance between them. vVe often say
that it holds only for point charges.
Coulomb's law resembles the inverse square law of gravitation which was
already more than 100 years old at the time of Coulomb's experiments; q
plays the role of m in that law. In gravity, however, the forces are always
COULOMB'S LAW 651

t.,.;-iiiin--Fiber

Fig. 26-2 Coulomb's torsion balance


from his 1785 memoir to the French
Academy of Sciences.

attractive; this corresponds to the fact that there are two kinds of electricity
but (apparently) only one kind of mass.
Our belief in Coulomb's law does not rest quantitativel y on Coulomb's
experiments. Torsion balance measurement s are difficult to make to an
accuracy of better than a few per cent. Such measurement s could not, for
example, convince us that the exponent in Eq. 26-1 is exactly 2 and not, say,
2.01. In Section 28-5 we show that Coulomb's law can also be deduced from
an indirect experiment which shows that the exponent in Eq. 26-1 lies
652 CHARGE AND MATTER Chap. 26
between the limits 2 ± 2 X 10- 9 • This has recently been extended by
Williams, Faller and Hill to 2 ± 3 X 10- 16 •
Although we have established the physical concept of electric charge, we
have not yet defined a unit in which it may be measured. It is possible to do
so operationally by putting equal charges q on the spheres of a torsion balance
and by measuring the magnitude F of the force that acts on each when the
charges are a measured distance r apart. One could then define q to have a
unit value if a unit force acts on each charge when the charges are separated
by a unit distance and one can give a name to the unit of charge so defined.*
For practical reasons having to do with the accuracy of measurements, the
unit of charge in the mks system is not defined using a torsion balance but is
derived from the unit of electric current. If the ends of a long wire are
connected to the terminals of a battery, it is common knowledge that an
electric current i is set up in the wire. We visualize this current as a flow of
charge. The mks unit of current is the ampere (abbr. amp). In Section 34-4
we describe the operational procedures in terms of which the ampere is
defined.
The mks unit of charge is the coulomb (abbr. coul). A coulomb is defined as
the amount of charge that flows through a given cross section of a wire in 1 second
if there is a steady current of 1 ampere in the wire. In symbols

q = it,

where q is in coulombs if i is in amperes and t is in seconds. Thus, if a wire


is connected to an insulated metal sphere, a charge of 10-6 coul can be put
on the sphere if a current of 1.0 amp exists in the wire for 10-6 sec.
~ Example I. A copper penny has a mass of 3.1 gm. Being electrically neutral,
it contains equal amounts of positive and negative electricity. What is the magnitude
q of these charges? A copper atom has a positive nuclear charge of 4.6 X 10- 18
coul and a negative electronic charge of equal magnitude.
The number N of copper atoms in a penny is found from the ratio

N m
No= M'

where No is Avogadro's number, m the mass of the coin, and M the atomic weight
of copper. This yields, solving for N,

N = (6.0 X 10 23 atoms/mole)(3.l gm) = 2 _9 X 1022 atoms.


64 gm/mole
The charge q is

q = (4.6 X 10-18 coul/atom)(2.9 X 1022 atoms) = 1.3 X 105 coul.

In a 100-watt, 110-volt light bulb the current is 0.91 amp. The student should verify
that it would take 40 hr for a charge of this amount to pass through this bulb. ....

* This scheme is the basis for the definition of the unit of charge called the statcoulomb.
However, in this book we do not use this unit or the systems of units of which it is a part;
see Appendix L, however.
Sec. 26-4 COULOMB'S LAW 653

Equation 26-1 can be written as an equality by inserting a constant of pro-


portionality. Instead of writing this simply as, say, k, it is usually written
in a more complex way as 1/41rE0 or

F = _1_2_1q2_ (26-3)
41rEo r 2

Certain equations that are derived from Eq. 26-3, but are used more often
than it is, will be simpler in form if we do this.
In the mks system we can measure q1 , q2 , r, and Fin Eq. 26-3 in ways that
do not depend on Coulomb's law. Numbers with units can be assigned to
them. There is no choice about the so-called permittivity constant Eo; it must
have that value which makes the right-hand side of Eq. 26-3 equal to the
left-hand side. This (measured) value turns out to be *

Eo = 8.85418 X 10-12 coul 2/nt-m 2.


In this book the value 8.9 X 10-12 coul 2/nt-m 2 will be accurate enough for
all problems. For direct application of Coulomb's law or in any problem in
which the quantity 1/4n0 occurs we may use, with sufficient accuracy for
this book,
1/41rE0 = 9.0 X 109 nt-m 2 /coul 2 .

IJJli,, Example 2. Let the total positive and the total negative charges in a copper
penny be separated to a distance such that their force of attraction is 1.0 lb ( = 4.5
nt). How far apart must they be?
We have (Eq. 26-3)
F = _l_ q1q2_
41rEo r 2

Putting q1q2 = q2 (see Example 1) and solving for r yields

l/4no . ~9.0 X 109 nt-m 2/coul2


r = q ~ - - = 1. 3 X 10" cou1
F 4.5 nt

= 5.8 X 109 meters = 3.6 X 106 miles.

This suggests that it is not possible to upset the electrical neutrality of gross objects
by any very large amount. What would be the force between the two charges if they
were placed 1.0 meter apart? ~

If more than two charges are present, Eq. 26-3 holds for every pair of
charges. Let the charges be q1 , q2 , and q3 , etc.; we calculate the force exerted
on any one (say q1 ) by all the others from the vector equation

F1 = F12 + F13 + F14 + · · ·, (26-4)

where F 12 , for example, is the force exerted on q1 by q2 •


* For practical reasons this value is not actually measured by direct application of
Eq. 26-3 but in an equivalent although more circuitous way that is described in Section
30-2.
654 CHARGE AND MATTER Chap. 26
IJli,,- Example 3.
Figure 26-3 shows three charges qi, q2, and q3 • What force acts on
q1? Assume that q1 = -1.0 X 10-6 coul, q2 = +3.0 X 10-6 coul, q3 = -2.0 X
10- 6 coul, r12 = 15 cm, r13 = 10 cm, and e = 30°.

Fig. 26-3 Example 3. Showing the forces ex-


erted on q1 by q2 and q3.

From Eq. 26-3, ignoring the signs of the charges, since we are interested only in the
magnitudes of the forces,

(9.0 X 109 nt-m 2 /coul2)(1.0 X 10- 6 coul)(3.0 X 10-6 coul)


(1.5 X 10-1 meter) 2
= 1.2 nt

F 13 -_ (9.0 X 10 nt-m /coul2)(1.0 X 10- coul)(2.0 X 10- coul)


9 2 6 6
and
(1.0 X 10- meter)
1 2

= 1.8 nt.
The directions of F 12 and F 13 are as shown in the ~gure.
The components of the resultant force F 1 acting i:m q1 (see Eq. 26-4) are

Fix = F12x + F13x = F12 + F13 sin 8

= 1.2 nt + (1.8 nt)(sin 30°) = 2.1 nt


and F1y = F12y + F13y = 0- F13 cos 8

= -(1.8nt)(cos30°) = -l.6nt.
The student should find the magnitude of F 1 and the angle it makes with the x-axis .
...
26-5 Charge Is Quantized
In Franklin's day electric charge was thought of as a continuous fluid, an
idea that was useful for many purposes. The atomic theory of matter, how-
ever, has shown.tJiat !!_uids themselves, such as water and air, are not con-
tinuous but are made up of atoms. Experiment shows that the "electric
fluid" is not continuous either but that it is made up of integral multiples
of a certain minimum electric charge. This fundamental charge, to which
Sec. 26-6 CHARGE AND MATTER 655

we give the symbol e, has the magnitude 1.60210 X 10-19 coul. Any physi-
cally existing charge q, no matter what its origin, can be written as ne where
n is a positive or a negative integer.
When a physical property such as charge exists in discrete "packets"
rather than in continuous amounts, the property is said to be quantized.
Quantization is basic to modern physics. The existence of atoms and of
particles like the electron and the proton indicates that mass is quantized
also. Later the student will learn that several other properties prove to be
quantized when suitably examined on the atomic scale; among them are
energy and angular momentum.
The quantum of charge e is so small that the "graininess" of electricity does
not show up in large-scale experiments, just as we do not realize that the air
we breathe is made up of atoms. In an ordinary 110-volt, 100-watt light
bulb, for example, 6 X 10 18 elementary charges enter and leave the filament
every second.
There exists today no theory that predicts the quantization of charge (or
the quantization of mass, that is, the existence of fundamental particles such
as protons, electrons, pions, etc.). Even assuming quantization, the classi-
cal theory of electromagnetism and Newtonian mechanics are incomplete
in that they do not correctly describe the behavior of charge and matter on
the atomic scale. The classical theory of electromagnetism, for example,
describes correctly what happens when a bar magnet is thrust through a
closed copper loop; it fails, however, if we wish to explain the magnetism of
the bar in terms of the atoms that make it up. The more detailed theories of
quantum physics are needed for this and similar problems.

26-6 Charge and Matter


Matter as we ordinarily experience it can be regarded as composed of three
kinds of elementary particles, the proton, the neutron, and the electron.
Table 26-1 shows their masses and charges. Note that the masses of the
neutron and the proton are approximately equal but that the electron is
lighter by a factor of about 1840.
Atoms are made up of a dense, positively charged nucleus, surrounded by a

Table 26-1

PROPERTIES OF THE PROTON, THE NEUTRON, AND THE ELECTRON

Particle Symbol Charge Mass

Proton p +e 1.67252 x 10-27 kg


Neutron n 0 1.67482 x 10-27 kg
Electron e- -e 9.1091 x 10-31 kg
656 CHARGE AND MATTER Chap. 26

Fig. 26-4 An atom, suggesting the electron cloud and, above, an enlarged view of the
nucleus.

cloud of electrons; see Fig. 26-4. The nucleus varies in radius from about
1 X 10- 15 meter for hydrogen to about 7 X 10- 15 meter for the heaviest
atoms. The outer diameter of the electron cloud, that is, the diameter of the
the atom, lies in the range 1-3 X 10- 10 meter, about 105 times larger than
the nuclear diameter.
lllll- Example 4. The distance r between the electron and the proton in the hydrogen
atom is about 5.3 X 10- 11 meter. What are the magnitudes of (a) the electrical force
and (b) the gravitational force between these two particles?
Sec. 26-6 CHARGE AND MATTER 657
From Coulomb's law,

F = _1_ q1q2
e 41rEo r 2

(9.0 X 109 nt-m 2 /coul2)(1.6 X 10- 19 coul) 2


(5.3 X 10- 11 meter)2
= 8.1 X 10-s nt.
The gravitational force is given b? Eq. 16-1, or

F = Gm1~
g r2

(6.7 X 10-11 nt-m 2 /kg 2)(9.1 X 10-31 kg)(l.7 X 10- 27 kg)


(5.3 X 10- 11 meter)2
= 3.7 X 10-47 nt.
Thus the electrical force is about 1039 times stronger than the gravitational force. <11111

The significance of Coulomb's law goes far beyond the description of the
forces acting between charged balls or rods. This law, when incorporated
into the structure of quantum physics, correctly describes (a) the electric
forces that bind the electrons of an atom to its nucleus, (b) the forces that
bind atoms together to form molecules, and (c) the forces that bind atoms 01
molecules together to form solids or liquids. Thus most of the forces of our
daily experience that are not gravitational in nature are electrical. A force
transmitted by a steel cable is basically an electrical force because, if we pass
an imaginary plane through the cable at right angles to it, it is only the
attractive electrical interatomic forces acting between atoms on opposite
sides of the plane that keep the cable from parting. vVe ourselves are an
assembly of nuclei and electrons bound together in a stable configuration by
Coulomb forces.
In the atomic nucleus we encounter a new force which is neither gravita-
tional nor electrical in nature. This strong attractive force, which binds
together the protons and neutrons that make up the nucleus, is called simply
the nuclear force. If this force were not present, the nucleus would fly apart at
once because of the strong Coulomb repulsion force that acts between its
protons. The nature of the nuclear force is only partially understood today
and forms the central problem of present-day researches in nuclear physics.
~ Example 5. What repulsive Coulomb force exists between two protons in a
nucleus of iron? Assume a separation of 4.0 X 10- 15 meter.
From Coulomb's law,

F = __1__ q1q2
4m,o r 2
(9.0 X 109 nt-m 2/coul2)(1.6 X 10- 19 coul) 2
(4.0 X 10- 15 meter)2
= 14 nt.
658 CHARGE AND MATTER Chap. 26
This enormous repulsive force must be more than compensated for by the strong
attractive nuclear forces. This example, combined with Example 4, shows that
nuclear binding forces are much stronger than atomic binding forces. Atomic binding
forces are, in turn, much stronger than gravitational forces for the same particles
separated by the same distance. <1111

The repulsive Coulomb forces acting between the protons in a nucleus make the
nucleus less stable than it otherwise would be. The spontaneous emission of alpha
particles from heavy nuclei and the phenomenon of nuclear fission are evidences of
this instability.
The fact that heavy nuclei contain significantly more neutrons than protons is still
another effect of the Coulomb forces. Consider Fig. 26-5 in which a particular atomic
species is represented by a circle, the coordinates being Z, the number of protons in
the nucleus (that is, the atomic number), and N, the number of neutrons in the nucleus
(that is, the neutron number). Stable nuclei are represented by filled circles and radio-
active nuclei, that is, nuclei that disintegrate spontaneously, emitting electrons or
a-particles, by open circles. Note that all elements (iron, for example, for which
Z = 26; see arrow) exist in a number of different forms, called isotopes.
Figure 26-5 shows that light nuclei, for which the Coulomb forces are relatively
unimportant,* lie on or close to the line labeled "N = Z" and thus have about equal
numbers of neutrons and protons. The heavier nuclei have a pronounced neutron
excess, U 238 having 92 protons and 238 - 92 or 146 neutrons. t In the absence of
Coulomb forces we would assume, extending the N = Z rule, that the most stable
nucleus with 238 particles would have 119 protons and 119 neutrons. However, such
a nucleus, if assembled, would fly apart at once because of Coulomb repulsion. Rela-
tive stability is found only if 27 of the protons are replaced by neutrons, thus diluting
the total Coulomb repulsion. Even in U 238 Coulomb repulsion is still very important
because (a) this nucleus is radioactive and emits a-particles, and (b) it may break
up into two large fragments (:fission) if it absorbs a neutron; both processes result in
separation of the nuclear charge and are Coulomb repulsion effects. Figure 26-5
shows that all nuclei with Z > 83 have no stable isotopes.
We have pointed out that matter, as we ordinarily experience it, is made up of
electrons, neutrons, and protons. Nature exhibits much more variety than this,
however. No fewer than 28 distinct elementary particles are now known, most of
them having been discovered since 1940, either in the penetrating cosmic rays that
come to us from beyond our atmosphere or in the reaction products of giant cyclotron-
like devices.
Appendix E, which lists some properties of these particles, shows that, like the more
familiar particles of Table 26-1, their charges are quantized, the quantum of charge
again being e. An understanding of the nature of these particles an<l of their relation-
ships to each other is perhaps the most significant research goal of modern physics.

* Coulomb forces are important in relation to the strong nuclear attractive forces
only for large nuclei, because Coulomb repulsion occurs between every pair of protons in
the nucleus but the attractive nuclear force does not. J n U 238 , for example, every proton
exerts a force of repulsion on each of the other Hl protons. However, each proton (and
neutron) exerts a nuclear attraction on only a small number of other neutrons and protons
that happen to be near it. As we proceed to larger nuclei, the amount of energy associated
with the repulsive Coulomb forces increases much faster than that associated with the
attractive nuclear forces.
t The superscript in this notation is the mass number A ( = N + Z). This is the total
number of particles in the nucleus.
CHARGE AND MATTER 659

150~~-----,,c~~....,----~--,.c--~---,,s-~~r--~--r;:~~--r-::~-::.-----r-:--~!'7"~--,
~
~

10

Fig. 26-5 The filled circles represent stable nuclei, the open ones radioactive nuclei.
This figure does not show the more recently discovered radioactive nuclei, which appear
throughout the chart, extending up to Z = 103.
660 CHARGE AND MATTER Chap. 26
26-7 Charge Is Conserved
When a glass rod is rubbed with silk, a positive charge appears on the rod.
Measurement shows that a negative charge of equal magnitude appears on
the silk. This suggests that rubbing does not create charge but merely
transfers it from one object to another, disturbing slightly the electrical
neutrality of each. This hypothesis of the conservation of charge has stood up
under close experimental scrutiny both for large-scale events and at the
atomic and nuclear level; no exceptions have ever been found.
An interesting example of charge conservation comes about when an
electron (charge = -e) and a positron (charge = +e) are brought close to
each other. The two particles may simply disappear, converting all their
rest mass into energy according to the well-known E = mc 2 relationship;
this annihilation process was described in Section 8-9. The energy appears
in the form of two oppositely directed gamma rays, which are similar in
character to X-rays. The net charge is zero both before and after the
event so that charge is conserved. Rest mass is not conserved, being
turned completely into energy.
Another example of charge conservation is found in radioactive decay, of
which the following process is typical:
u23s - Thz34 + He4. (26-5)
The radioactive "parent" nucleus, U 238 , contains 92 protons (that is, its
atomic number Z = 92). It disintegrates spontaneously by emitting an a-
particle (He4 ; Z = 2) transmuting itself into the nucleus Th 234 , with Z = 90.
Thus the amount of charge present before disintegration ( +92e) is the same
as that present after the disintegration.
A final example of charge conservation is found in nuclear reactions, of
which the bombardment of Ca44 with cyclotron-accelerated protons is
typical. In a particular collision a neutron may emerge from the nucleus,
leaving Sc 44 as a "residual" nucleus:
Ca44 +p ----; Sc44 + n.
The sum of the atomic numbers before the reaction (20 + 1) is exactly equal
to the sum of the atomic numbers after the reaction (21 + 0). Again charge
is conserved.

QUESTIONS

1. You are given two metal spheres mounted on portable insulating supports. Find a
way to give them equal and opposite charges. You may use a glass rod rubbed with silk
but may not touch it to the spheres. Do the spheres have to be of equal size for your
method to work?
2. A charged rod attracts bits of dry cork dust which, after touching the rod, often
jump violently away from it. Explain.
3. If a charged glass rod is held near one end of an insulated uncharged metal rod as in
Fig. 26-6, electrons are drawn to one end, as shown. Why does the flow of electrons cease?
l'here is an almost inexhaustible supply of them in the metal rod.
Chap. 26 PROBLEMS 661

4. In Fig. 26-6 does any net electri-


cal force act on the metal rod? Ex-
plain.
5. An insulated rod is said to carry
an electric charge. How could you Metal
verify this and determine the sign of
the charge?
6. Why do electrostatic experiments
Insulating
not work well on humid days?
support
7. A person standing on an insu-
lated stool touches a charged, insu-
lated conductor. Is the conductor
discharged completely?
8. (a) A positively charged glass
Fig. 26-6
rod attracts a suspended object. Can
we conclude that the object is neg-
atively charged? (b) A positively charged glass rod repels a suspended object. Can we
conclude that the object is positively charged?
\!. Is the Coulomb force that one charge exerts on another changed if other charges
are brought nearby?
10. The quantum of charge is 1.60 X 10-19 coul. Is there a corresponding single quan-
tum of mass?
11. Verify the fact that the decay schemes for the elementary particles in Appendix E
are consistent with charge conservation.
12. What does it mean to say that a physical quantity is (a) quantized or (b) conserved?
Give some examples.
13. A nucleus U 238 splits into two identical parts. Are the nuclei so produced likely
to be stable or radioactive?

PROBLEMS

1. Protons in the cosmic rays strike the earth's upper atmosphere at a rate, averaged
over the earth's surface, of 0.15 protons/cm 2-sec. What total current does the earth receive
from beyond its atmosphere in the form of incident cosmic ray protons? The earth's
radius is 6.4 X 106 !Jleters.
~ .. A point cha;ge of +3.0 X 10-5 coul is 12 cm distant from a second point charge of
-1.5 X 10-6 coul. Calculate the magnitude and direction of the force on each charge.
--8:-· Two similar balls of mass m are hung from silk threads of
length land carry similar charges q as in Fig. 26-7. Assume that
(J is so small that tan (J can be replaced by its approximate equal,
sine. To this approximation show that

x = (___<[}__ ) Ya
21rEomg

where x is the separation between the balls. If l = 120 cm, m =


10 gm, and x = 5.0 cm, what is q?
4. Assume that each ball in Problem 3 is losing charge at the rate
q q of 1.0 X 10-9 coul/sec. At what instantaneous relative speed
( = dx/dt) do the balls approach each other initially?
.5. Three small balls, each of mass 10 gm, are suspended sep-
Fig. 26-7 arately from a common point by silk threads, each 1.0 meter
662 CHARGE AND MATTER Chap. 26
long. The balls are identically charged and hang at the corners of an equilateral triangle
0.1 meter long on a side. What is the charge on each ball?
6. In Fig. 26-8 what is the resultant force on the charge 1n the lower left corner of the
square? Assume that q = 1.0 X 10- 7 coul and a -== 5.0 cm.
+q a -q

a a

+2q a -2q
Fig. 26-8

7. A charge Q is placed at each of two opposite corners of a square. A charge q is placed


at each of the other two corners. (a) If the resultant electrical force on Q is zero, how are
Q and q related? (b) Could q be chosen to make the resultant force on every charge zero?
,-..-&,.. How far apart must two protons be if the electrical repulsive force acting on either
one is equal to its weight?
'··9. (a) What equal positive charges would have to be placed on the earth and on the moon
to neutralize their gravitational attraction? (b) Do you need to know the lunar distance to
solve this problem? -{s).::'.How many tons of hydrogen would be needed to provide the
positive charge calculated ina?
10. A certain charge Q is to be divided into two parts, q and Q - q. What is the relation-
ship of Q to q if the two parts, placed a given distance apart, are to have a maximum
Coulomb repulsion?
11. Each of two small spheres is charged positively, the combined charge being 5.0 X
10-5 coul. If each sphere is repelled from the other by a force of 1.0 newton when the
spheres are 2.0 meters apart, how is the total charge distributed between the spheres?
12. Two equal positive point charges are separated by a distance 2a. A point test
charge is located in a plane which is normal to the line joining these charges and midway
between them. (a) Calculate the radius r of the circle of symmetry in this plane for which
the force on the test charge has a maximum value. (b) What is the direction of this force,
assuming a positive test charge.
13. A cube of edge a carries a point charge q at each corner. (a) Show that the magni-
tude of the resultant force on any one of the charges is
0.26lq 2
F=--·
Eoa2

(b) What is the direction of F relative to the cube edges?


14. Estimate roughly the number of coulombs of positive charge in a glass of water.
15. (a) How many electrons would have to be removed from a penny to leave it with a
charge of + 10-7 coul? (b) What fraction of the electrons in the penny does this correspond
to?
16. The radius of a copper nucleus is about 5.1 X 10-13 cm. Calculate the density of
the material that makes up the nucleus. Does your answer seem reasonable? (The atomic
weight of copper is 64 gm/mole; ignore the mass of the electrons in comparison to that of
-tp.e nucleus.)
'17. In the radioactive decay of U 238 (see Eq. 26-5) the center of the emerging a-particle
is, at a certain instant, 9 X 10- 15 meter from the center of the residual nucleus Th 234 •
At this instant (a) what is the electrostatic force on the a-particle and (b) what is its
acceleration?
The Electric Field
CHAPTER 27

27-1 The Electric Field


'\Yith every point in space near the earth we can associate a gravitational
iield strength vector g (see Eq. 16-12). This is the gravitational acceleration
that a test body, placed at that point and released, would experience. If m
is the mass of the body and F the gravitational force acting on it, g is given by

g = F/m. (27-1)

This is an example of a vector field. For points near the surface of the earth
the field is often taken as uniform; that is, g is the same for all points.
The flow of water in a river provides another example of a vector field,
i:·alled a flow field (see Section 18-7). Every point in the water has associated
with it a vector quantity, the velocity v with which the water flows past the
point. If g and v do not change with time, the corresponding fields are
described as stationary. In the case of the river note that even though the
water is moving the vector v at any point does not change with time for
,teady-flow conditions.
The space surrounding a charged rod is affected by the presence of the rod,
and we speak of an electric field in this space. In the same way we speak of
a magnetic field in the space around a bar magnet. In the classical theory
of electromagnetism the electric and magnetic fields are central concepts.
Before Faraday's time, the force acting between charged particles was
thought of as a direct and instantaneous interaction between the two
particles. This action-at-a-distance view was also held for magnetic and for
graYitational forces. Today we prefer to think in terms of electric fields as
iollmrn:
1. Charge q1 in Fig. 27-1 sets up an electric field in the space around itself.
This field is suggested by the shading in the figure; later we shall show how to
represent electric fields more concretely.
663
664 THE ELECTRIC FIELD Chap. 27

Fig. 27-1 Charge q1 sets up a field


that exerts a force F on charge q2 .

2. The field acts on charge q2 ; this shows up in the force F that q2 ex-
periences.
The field plays an intermediary role in the forces between charges. There
are two separate problems: (a) calculating the fields that are set up by given
distributions of charge and (b) calculating the forces that given fields will
exert on charges placed in them. We think in terms of

charge ;;:::::: field

and not, as in the action-at-a-distance point of view, in terms of

charge ;;:::::: charge.

In Fig. 27-1 we can also imagine that q2 sets up a field and that this field acts
on q1 , producing a force -Fon it. The situation is completely symmetrical,
each charge being immersed in a field associated with the other charge.
If the only problem in electromagnetism was that of the forces between
stationary charges, the field and the action-at-a-distance points of view
would be perfectly equivalent. Suppose, however, that q1 in Fig. 27-1
suddenly accelerates to the right. How quickly does the charge q2 learn
that q1 has moved and that the force which it (q2 ) experiences must increase?
Electromagnetic theory predicts that q2 learns about q1 's motion by a field
disturbance that emanates from q1 , traveling with the speed of light. The
action-at-a-distance point of view requires that information about q1 's
acceleration be communicated instantaneously to q2 ; this is not in accord
with experiment. Accelerating electrons in the antenna of a radio transmitter
influence electrons in a distant receiving antenna only after a time l/ c where l
is the separation of the antennas and c is the speed of light.
Sec. 27-3 LINES OF FORCE 665

27-2 The Electric Field Strength E


To define the electric field operationally, we place a small test body carry-
ing a test charge q0 (assumed positive for convenience) at the point in space
that is to be examined, and we measure the electrical force F (if any) that
acts on this body. The electric field strength Eat the point is defined as*
E = F/q 0 . (27-2)
Here E is a vector because F is one, q0 being a scalar. The direction of E
is the direction of F, that is, it is the direction in which a resting positive
charge placed at the point would tend to move.
The definition of gravitational field strength g is much like that of electric field
strength, except that the mass of the test body rather than its charge is the property
of interest. Although the units of g are usually written as meters/sec 2, they could
also be written as nt/kg (Eq. 27-1); those for E are nt/coul (Eq. 27-2). Thus both
g and E are expressed as a force divided by a property (mass or charge) of the test
body .
..,._ Example 1. What is the magnitude of the electric field strength E such that an
electron, placed in the field, would experience an electrical force equal to its weight?
From Eq. 27-2, replacing qo bye and F by mg, where mis the electron mass, we have

(9.1 X 10- 31 kg)(9.8 meters/sec 2)


1.6 X 10- 19 coul
= 5.6 X 10- 11 nt/coul.
This is a very weak electric field. Which way will E have to point if the electric force
is to cancel the gravitational force? _.,..
In applying Eq. 27-2 we must use a test charge as small as possible. A large test
charge might disturb the primary charges that are responsible for the field, thus
changing the very quantity that we are trying to measure. Equation 27-2 should,
strictly, be replaced by
. F
E = 1I l l - · (27-3)
Qo->O qo

This equation instructs us to use a smaller and smaller test charge q0 , evaluating the
ratio F /qo at every step. The electric field Eis then the limit of this ratio as the size
of the test charge approaches zero.

27-3 Lines of Force


The concept of the electric field as a vector was not appreciated by Michael
Faraday, who always thought in terms of lines of force. The lines of force
still form a convenient way of visualizing electric-field patterns. We shall
use them for this purpose but we shall not employ them quantitatively.
* This definition of E, though conceptually sound and quite appropriate to our present
purpose, is rarely carried out in practice because of experimental difficulties. E is normally
found by calculation from more readily measurable quantities such as the electric po-
tential; see Section 29-7.
666 THE ELECTRIC FIELD Chap. 27
The relationship between the (imaginary) lines of force and the electric
field strength vector is this :
1. The tangent to a line of force at any point gives the direction of Eat that
point.
2. The lines of force are drawn so that the number of lines per unit cross-
sectional area is proportional to the magnitude of E. Where the lines are close
together Eis large and where they are far apart Eis small.
It is not obvim,s that it is possible to draw a continuous set of lines to meet
these requirements. Indeed, it turns out that if Coulomb's law were not
true it would not be possible to do so; see Problem 4.
Figure 27-2 shows the lines of force for a uniform sheet of positive charge.
We assume that the sheet is infinitely large, which, for a sheet of finite dimen-
sions, is equivalent to considering only those points whose distance from the
sheet is small compared to the distance to the nearest edge of the sheet. A
positive test charge, released in front of such a sheet, would move a,vay from
the sheet along a perpendicular line. Thus the electric field strength vector
at any point near the sheet must be at right angles to the sheet. The lines of
force are uniformly spaced, which means that E has the same magnitude for
all points near the sheet.
Figure 27--3 sho-ws the lines of force for a negatively charged sphere. From
symmetry, the lines must lie along radii. They point inward because a free
positive charge would be accelerated in this direction. The electric field Eis
not constant but decreases with increasing distance from the charge. This is
evident in the lines of force, which are farther apart at greater distances.
From symmetry, Eis the same for all points that lie a given distance from the
center of the charge.
~ Example 2. In Fig. 27-3 how does E vary with the distance r from the center
of the charged sphere?
Suppose that N lines originate on the sphere. Draw an imaginary concentric
sphere of radius r; the number of lines per unit cross-sectional area at every point on
the sphere is N /47rr 2 • Since Eis proportional to this, we can write that
E ex 1/r2 •
We derive an exact relationship in Section 27-4. How should E vary with distance
from an infinitely long uniform cylinder of charge? .,..

-----------tj-i-------;--

------1-1--------'l>-- Fig. 27-2 Lines of force for a section of an


infinitely large sheet of positive charge.

----------11~.______
; __
LINES OF FORCE 667

Fig. 27-3 Lines of force for a negatively


charged sphere.

Figures 27-4 and 27-5 show the lines of force for two equal like charges and
for two equal unlike charges, respectively. Michael Faraday, as we have
said, used lines of force a great deal in his thinking. They were more real
for him than they are for most scientists and engineers today. It is possible

Fig. 27-4 Lines of force for two equal positive charges.


668 THE ELECTRIC FIELD Chap. 27

Fig. 27-5 Lines of force for equal but oppoE'ite charges.

to sympathize with Faraday's point of view. Can we not almost "see" the
charges being pushed apart in Fig. 27-4 and pulled together in Fig. 27-5 by
the lines of force? The student should compare Fig. 27-5 with Fig. 18-15,
which represents a flow field. Figure 27-6 shows a representation of lines of
force around charged conductors, using grass seeds suspended in an insulating
liquid.
Lines of force give a vivid picture of the way E varies through a given region of
space. However, the equations of electromagnetism (see Table 38-3) are written in
terms of the electric field strength E and other field vectors and not in terms of the
lines of force. The electric field E varies in a perfectly continuous way as any path in
the field is traversed; see Fig. 27-7a. This kind of physical continuity, however, is
different from the continuity of the lines of force, which has no real physical meaning.
To illustrate this point we assert that the lines of force need not be continuous but can
be drawn as in Fig. 27-7b. This figure agrees with the two properties of lines of force
laid down at the beginning of this section. This point of view has been especially
emphasized by Joseph Slepian. *
* See American Journal of Physics, 19, 87 (1951).
Fig. 27-6 Photographs of the patterns of electric lines of force around (a) a charged
plate (compare Fig. 27-2), and (b) two rods with equal and opposite charges (compare
fig. 27-5). The patterns were made by suspending grass seed in an insulating liquid.
, Courtesy Educational Services Incorporated, vVatertown, Mass.)
669
670 THE ELECTRIC FIELD Chap. 27

(a)

(b)

Fig. 27-7 (a) E varies continuously as we move along any path AB in the field set up by
q. (b) Lines of force need not be continuous.

27-4 Calculation of E
Let a test charge q0 be placed a distance r from a point charge q. The
magnitude of the force acting on q0 is given by Coulomb's law, or

The electric field strength at the site of the test charge is given by Eq. 27-2, or
F 1 q
E=-=~-· (27-4)
qo 41reo r2
Sec. 27-4 CALCULATION OF E 671

The direction of Eis on a radial line from q, pointing outward if q is positive


and inward if q is negative.
To find E for a group of point charges: (a) Calculate En due to each charge
at the given point as if it were the only charge present. (b) Add these separately
calculated fields vectorially to find the resultant field E at the point. In
equation form,
E = E1 + E2 + Ea + · · · = ~En n = 1, 2, 3, .... (27-5)
The sum is a vector sum, taken over all the charges.
If the charge distribution is a continuous one, the field it sets up at any
point P can be computed by dividing the charge into infinitesimal elements
dq. The field dE due to each element at the point in question is then calcu-
lated, treating the elements as point charges. The magnitude of dE (see
Eq. 27-4) is given by
1 dq
dE=--, (27-6)
41re 0 r 2

where r is the distance from the charge element dq to the point P. The re-
sultant field at P is then found by adding (that is, integrating) the field
contributions due to all the charge elements, or,

(27-7)

The integration, like the sum in Eq. 27-5, is a vector operation; in Example 5
we will see lrw such an integral is handled in a simple case .
.... Exampl4 3. An electric dipole. Figure 27-8 shows a positive and a negative
charge of eqdal magnitude q placed a distance 2a apart, a configuration called an elec-

Fig. 27-8 Example 3.

-q

tric dipole. The pattern of lines of force is that of Fig. 27-5, which also shows an
electric dipole. What is the field E due to these charges at point P, a distance r along
the perpendicular bisector of the line joining the charges? Assume r » a.
Equation 27-5 gives the vector equation
E = E1 + E2,
where, from Eq. 27-4, *
1 q
E1 = E2 = - - - - - ·
41reo a 2 r2 +
* Note that the r's in Eq. 27-4 and in this equation have different meanings
672 THE ELECTRIC FIELD Chap. 27
The vector sum of E1 and E2 points vertically downward and has the magnitude
E = 2E 1 cos e.
From the figure we see that
a
cose = ---;:===
,v'a2 + r2
Substituting the expressions for E 1 and for cos e into that for E yields

E = ___'!:_ q a 1 2aq
47l'Eo (a2 + r2) Va2 + r2 47r€o (a 2 + r )%2

If r » a, we can neglect a in the denominator; this equation then reduces to

(27-8a)

The essential properties of the charge distribution in Fig. 27-8, the magnitude of
the charge q and the separation 2a between the charges, enter Eq. 27-8a only as a
product. This means that, if we measure E at various distances from the electric
dipole (assuming r » a), we can never deduce q and 2a separately but only the
product 2aq; if q were doubled and a simultaneously cut in half, the electric field at
large distances from the dipole would not change.
The product 2aq is called the electric dipole moment p. Thus we can rewrite this
equation for E, for distant points along the perpendicular bisector, as
1 p
E = - - 3· (27-8b)
47l'Eo r
The result for distant points along the dipole axis (see Problem 10) and the general
result for any distant point (see Problem 23) also contain the quantities 2a and q only
as the product 2aq ( = p). The variation of E with r in the general result for distant
points is also as 1/r3 , as in Eq. 27-8b.
The dipole of Fig. 27-8 is two equal and opposite charges placed close to each other
so that their separate fields at distant points almost, but not quite, cancel. On this
point of view it is easy to understand that E(r) for a dipole varies as 1/r3 (E1. 27-8b),
whereas for a point charge E(r) drops off more slowly, namely as 1/r2 (Eq. 27-4).
Example 4. Figure 27-9 shows a charge q1 ( = +1.0 X 10-6 coul) 10 cm from
a charge q2 ( = +2.0 X 10-6 coul). At what point on the line joining the two
charges is the electric field strength zero?

Fig. 27-9 Example 4.

The point must lie between the charges because only here do the forces exerted by
q1 and q2 on a test charge oppose each other. If E 1 is the electric field strength due
to q1 and fa that due to q2, we must have
E1 = E2
1 q1 1 q2
or (see Eq. 27-4)
47!'1:o x2 47!'1:o (l - x) 2
Sec. 27-4 CALCULATION OF E 673
where x is the distance from q1 and l equals 10 cm. Solving for x yields
l 10 cm
x=-----
1 + V2 = 4.1 cm.
1+~
The student should supply the missing steps. On what basis was the second root of
the resulti{')uadratic equation discarded?
Example-_.~J \ Figure 27-10 shows a ring of charge q and of radius a. Calculate E
for points on the axis of the ring a distance x from its center.
Consider a differential element of the ring of length ds, located at the top of the
ring in Fig. 27-10. It contains an element of charge given by
ds
dq = q~
21ra
where 21ra is the circumference of the ring. This element sets up a differential electric
field dE at point P.
The resultant field E at Pis found by integrating the effects of all the elements that
make up the ring. From symmetry this resultant field must lie along the ring axis.
Thus only the component of dE parallel to this axis contributes to the final result.
The component perpendicular to the axis is canceled out by an equal but opposite
component established by the charge element on the opposite side of the ring.
Thus the general vector integral (Eq. 27-7)

E = fdE
becomes a scalar integral E = f dE cos e.

(J

Fig. 27-10 Example 5


674 THE ELECTRIC FIELD Chap. 27
The quantity dE follows from Eq. 27-6, or

dE _ 1 dq _ 1 (q ds) 1
- 41r1:o ~ - 41rfo 21r-;: a 2 x2 • +
x
From Fig. 27-10 we have cos (J = .
va2 + x2
Noting that, for a given point P, x has the same value for all charge elements and is
not a variable, we obtain

E = f dE cos 8 = f~- q ds
47rfo (21ra)(a 2 + x 2) v a2 + :c2
x

= ~- qx fds
41rfo (21ra)(a 2 + x2)% 3
·
The integral is simply the circumference of the ring ( = 21ra), so that

E = _1_ qx
41rfo (a 2 x 2 )% +
Does this expression for E reduce to an expected result for x O? For x » a we
can neglect a in the denominator of this equation, yielding
1 q
E~---·
- 41rfo x 2

This is an expected result (compare Eq. 27-4) because at great enough distances the
ring behaves like a point charge q.
Example 6. Line of charge. Figure 27-11 shows a section of an infinite line
of charge whose linear charge density (that is, the charge per unit length, measured

r
y

x
X ---------C:±··;..f-':::;-.:.+E-:-.-:-::£:+-.-=E=-:+::Z:-=:+E::-:-:-[t;:+:,:;:,:,.,+3--,-:+£:::-:;E+:Ej:·+E·-:·:,:-·+·-:J:·-:+C:·-·-·E+·::£·-:+·z,.:·:+E.~ E I
0 dx

Fig. 27-11 Example 6. A section of an infinite line of charge.

in coul/meter) has the constant value;\, Calculate the field Ea distance y from the
line.
The magnitude of the field contribution dE due to charge element dq ( = A dx) is
given, using Eq. 27-6, by
dE = _1_ dq = _1_ ;\ dx .
41r1:o r 2 41r1:o y 2 x2 +
Sec. 27-5 A POINT CHARGE IN AN ELECTRIC FIELD 675

The vector dE, as Fig. 27-11 shows, has the components

dEx = -dE sin 8 and dEu = dE cos 8.

The minus sign in front of dEx indicates that dEx points in the negative x direction.
The x and y components of the resultant vector Eat point Pare given by

Ex= f dEx = -i=-oox=+oo


sin8dE and Eu = f dEu = f x=+oo

X=-OO
cos 8 dE.

Ex must be zero because every charge element on the right has a corresponding
element on the left such that their field contributions in the x direction cancel.
Thus E points entirely in the y direction. Because the contributions to Eu from the
right- and left-hand halves of the rod are equal, we can write

.[=
x=+oo
E = Eu = 2 0
cos edE.

Note that we have changed the lower limit of integration and have introduced a factor
of two.
Substituting the expression for dE into this equation gives

E = ~-
A f x=oo
cos8
dx
·
27rE:o x=O Y2 + X2

From Fig. 27-11 we see that the quantities 8 and x are not independent. We must
eliminate one of them, say x. The relation between x and 8 is (see figure)
x = y tan e.
Differentiating, we obtain dx = y sec 2 ed8.
Substituting these two expressions leads finally to
A 10=1r/2
E = ~- cos 8 d8.
271"«:oY o=o .
The student should check this step carefully, noting that the limits must now be on 8
and not on x. For example as x --, + oo, 8 --, 7r/2, as Fig. 27-11 shows. This
equation integrates readily to

E = 2~oY I sin e [2 = 27r~oY.

The student may wonder about the usefulness of solving a problem involving an
infinite rod of charge when any actual rod must have a finite length (see Problem 15).
However, for points close enough to finite rods and not near their ends, the equation
that we have just derived yields results that are so close to the correct values that the
difference can be ignored in many practical situations. It is usually unnecessary to
solve exactly every geometry encountered in practical problems. Indeed, if idealiza-
tions or approximations are not made, the vast majority of significant problems of
all kinds in physics and engineering cannot be solved at all. <1111

27-5 A Point Charge in an Electric Field


An electric field will exert a force on a charged particle given by (Eq. 27-2)
F = Eq.
676 THE ELECTRIC FIELD Chap. 27
This force will produce an acceleration
a= F/m,
where m is the mass of the particle. We will consider two examples of the
acceleration of a charged particle in a uniform electric field. Such a field can
be produced by connecting the terminals of a battery to two parallel metal
plates which are otherwise insulated from each other. If the spacing between
the plates is small compared with the dimensions of the plates, the field
between them will be fairly uniform except near the edges. Note that in
calculating the motion of a particle in a field set up by external charges the
field due to the particle itself (that is, its self-field) is ignored. For example,
the earth's gravitational field can have no effect on the earth itself but only
on a second object, say a stone, placed in that field.
~ Example 7. A particle of mass m and charge q is placed at rest in a uniform
electric field (Fig. 27-12) and released. Describe its motion.

P1
+ + + + + + + + + + + + + + +

Vl !E Fig. 27-12 A charge is released from


rest in a uniform electric field set up
between two oppositely charged metal
plates P1 and P2.
V2
- - - - - - - - - - - - -
p2
I,
The motion resembles that of a body falling in the earth's gravitational field. The
(constant) acceleration is given by
F qE
a= - = --·
m m
The equations for uniformly accelerated motion (Table 3-1) then apply. With
Vo =O, they are
qEt
v =at=-,
m
i 2 qEt2
y = z;at = -2m'
2qEy
and v2 = 2ay. = -m

The kinetic energy attained after moving a distance y is found from

K = !mv2 = !m c~Y) = qEy.


Sec. 27-5 A POINT CHARGE IN AN ELECTRIC FIELD 677

This result also follows directly-from the work-energy theorem because a constant
force qE acts over a distance y.
Example 8. Deflecting an electron beam. Figure 27-13 shows an electron of mass
m and charge e projected with speed v0 at right angles to a uniform field E. Describe
its motion.
p

Fig. 27-13 Example 8. An elec-


tron is projected into a uniform elec-
tric field.

The motion is like that of a projectile fired horizontally in the earth's gravitational
field. The considerations of Section 4-3 apply, the horizontal (x) and vertical (y)
motions being given by
X = Vot

and

eE
Eliminating t yields y = --x2 (27-9)
2mvo 2
for the equation of the trajectory.
When the electron emerges from the plates in Fig. 27-13, it travels (neglecting
gravity) in a straight line tangent to the parabola at the exit point. We can let it fall
on a fluorescent screen S placed some distance beyond the plates. Together with
other electrons following the same path, it will then make itself visible as a small
luminous spot; this is the principle of the electrostatic cathode-ray oscilloscope.
Example 9. The electric field between the plates of a cathode-ray oscilloscope is
1.2 X 10 4 nt/coul. What deflection will an electron experience if it enters at right
angles to the field with a kinetic energy of 2000 ev ( = 3.2 X 10- 15 joule), a typical
value? The deflecting assembly is 1.5 cm long.
Recalling that Ko = 1mv 0 2 , we can rewrite Eq. 27-9 as
eEx 2
y=--·
4Ko
If x1 is the horizontal position of the far edge of the plate, y 1 will be the corresponding
deflection (see Fig. 27-13), or
eEx1 2
YI=~-
4Ko
(1.6 X 10-19 coul)(l.2 X 104 nt/coul)(l.5 X 10- 2 meter) 2
(4)(3.2 X 10- 15 joule)
= 3.4 X 10- 4 meter = 0.34 mm.
The deflection measured, not at the deflecting plates but at the fluorescent screen, is
much larger.
678 THE ELECTRIC FIELD Chap. 27
Example 10. A positive point test charge q0 is placed halfway between two equal
positive charges q. What force acts on it at or near this point P'?
From symmetry the force at the point is zero so that the particle is in equilibrium;
the nature of the equilibrium remains to be found. Figure 27-14 (compare Fig. 27-4)
z

Fig. 27-14 Example 10. The electric field at four points


near a point P which is centered between two equal posi-
tive charges.

shows the E vectors for four points near P. If the test charge is moved along the z axis,
a restoring force is brought into play; however, the equilibrium is unstable for motion
in the x-y plane. Thus we have the three-dimensional equivalent of saddle point
equilibrium; see Fig. 14-8. What is the nature of the equilibrium for a negative test
charge? ....

27-6 A Dipole in an Electric Field


An electric dipole moment can be regarded as a vector p whose magni-
tude p, for a dipole like that described in Example 3, is the product 2aq of the
magnitude of either charge q and the distance 2a between the charges. The
direction of p for such a dipole is from the negative to the positive charge.
The vector nature of the electric dipole moment permits us to cast many
expressions involving electric dipoles into concise form, as we shall see.
Figure 27-15a shows an electric dipole formed by placing two charges +q
and - q a fixed distance 2a apart. The arrangement is placed in a uniform
external electric field E, its dipole moment p making an angle ewith this field.
Two equal and opposite forces F and - F act as shown, where
F = qE.
The net force is clearly zero, but there is a net torque about an axis through
0 (see Eq. 12-2) given by
T = 2F(a sine) = 2aF sine.
Combining these two equations and recalling that p (2a) (q), we obtain
T = 2aqE sin e= pE sin e. (27-10)
Thus an electric dipole placed in an external electric field E experiences a
torque tending to align it with the field. Equation 27-10 can be written in
vector form as
T = p xE, (27-11)
the appropriate vectors being shown in Fig. 27-15b.
Sec. 27-6 A DIPOLE IN AN ELECTRIC FIELD 679

Fig. 27-15 (a-) An electric dipole in -F<f---ar:


a uniform external field. (b) An ----------------
oblique view, illustrating T = p X E.
(a)

(b)

Work (positive or negative) must be done by an external agent to change


the orientation of an electric dipole in an external field. This work is stored
aE potential energy U in the system consisting of the dipole and the arrange-
ment used to set up the external field. If 8 in Fig. 27-15a has the initial
value 80, the work required to turn the dipole axis to an angle 8 is given (see
Table 12-2) from

w= J i e
dW =
eo
T d8 = u,

where r is the torque exerted by the agent that does the work. Combining
this equation with Eq. 27-10 yields

U = l 80
e
pE sin 8 d8 = pE iOo
e
sin 8 d8

Since we are interested only in changes in potential energy, we choose the


reference orientation 80 to have any convenient value, in this case 90°. This
gives
U = -pE cos 8 (27-12)
or, in vector form,
u = -p·E. (27-13)
680 THE ELECTRIC FIELD Chap. 27
IJIJI> Example 11. An electric dipole consists of two opposite charges of magnitude
q = 1.0 X 10-6 coul separated by d = 2.0 cm. The dipole is placed in an external
field of 1.0 X I 10 5 nt/ coul.
(a) What maximum torque does the field exert on the dipole? The maximum
torque is found by putting e = go 0 in Eq. 27-10 or
T = pE sine = qdE sin e
= (1.0 X 10-6 coul)(0.020 meter)(l.O X 10 5 nt/coul)(sin go 0 )
= 2.0 X 10-3 nt-m.
(b) How much work must an external agent do to turn the dipole end for end,
starting from a position of alignment (8 = O)? The work is the difference in potential
energy U between the positions e = 180° and e = 0. From Eq. 27-12,
W = Ul8oo - Uo 0 = (-pE cos 180°) - (-pE cos 0)
= 2pE = 2qdE
= (2)(1.0 X 10- 6 coul)(0.020 meter)(l.O X 105 nt/coul)
= 4.0 X 10- 3 joule.

QUESTIONS

1. Name as many scalar fields and vector fields as you can.


2. (a) In the gravitational attraction between the earth and a stone, can we say that
the earth lies in the gravitational field of the stone? (b) How is the gravitational field
due to the stone related to that due to the earth?
3. A positively charged ball hangs from a long silk thread. We wish to measure E at a
point in the same horizontal plane as that of the hanging charge. To do so, we put a
poRitive test charge q0 at the point and measure F /q 0 . Will F /q 0 be less than, equal to, or
greater than Eat the point in question?
4. Taking into account the quantization of electric charge (the single electron pro-
viding the basic charge unit), how can we justify the procedure suggested by Eq. 27-3?
5. In Fig. 27-5 the force on the lower charge points up and is finite. The crowding of
the lines of force, however, suggests that E is infinitely great at the site of this (point)
charge. A charge immersed in an infinitely great field should have an infinitely great force
acting on it. What is the solution to this dilemma?
6. Electric lines of force never cross. Why?
7. In Fig. 27-4 why do the lines of force around the edge of the figure appear, when
extended backwards, to radiate uniformly from the center of the figure?
8. Figure 27-2 shows that E has the same value for all points in front of an infinite
uniformly charged sheet. Is this- reasonable? One might think that the field should be
stronger near the sheet because the charges are so much closer.
9. Two point charges of unknown magnitude and sign are a distance d apart. The
electric field strength is zero at one point between them, on the line joining them. What
can you conclude about the charges?
10. Compare the way E varies with r for (a) a point charge (Eq. 27-4), (b) a dipole (Eq.
27-8a), and (c) a quadrupole (Problem 18).
11. If a point charge q of mass mis released from rest in a nonuniform field, will it follow
a line of force?
12. An electric dipole is placed in a nonuniform electric field. Is there a net force on it?
13. An electric dipole is placed at rest in a uniform external electric field, as in Fig.
27-15a, and released. Discuss its motion.
Chap. 27 PROBLEMS 681

14. An electric dipole has its dipole moment p aligned with a uniform external electric
field E. (a) Is the equilibrium stable or unstable? (b) Discuss the nature of the equilib-
rium if p and E point in opposite directions.

PROBLEMS

1. Sketch qualitatively the lines of force associated with a thin, circular, uniformly
charged disk of radius R. (Hint: Consider as limiting cases points very close to the surface
and points very far from it.) Show the lines only in a plane containing the axis of the disk.
-~-~. (a) Sketch qualitatively the lines of force associated with these long parallel lines
;. ,of charge, in a perpendicular plane. Assume that the intersections of the lines of
charge with such a plane form an equilateral triangle and that each line of charge has
the same linear charge density "II. (coul/meter). (b) Discuss the nature of the equilib-
rium of a test charge placed on the central axis of the charge assembly.
3. In Fig. 27-4 consider any two lines of force leaving the upper charge. If the angle
between their tangents for points near the charge is e, it becomes () /2 at great distances.
Verify this statement and explain it. (Hint: Consider how the lines must behave both
close to either charge and far from the charges.)
4. Assume that the exponent in Coulomb's law is not "two" but n. Show that for
n >"' 2 it is impossible to construct lines that will have the properties listed for lines of force
in Section 27-3. For simplicity, treat an isolated point charge.
~ - What is the magnitude of a point charge chosen so that the electric field 50 cm
away has the magnitude 2.0 nt/coul?
6. Two equal and opposite charges of magnitude 2.0 X 10-7 coul are 15 cm apart. (a)
What are the magnitude and direction of Eat a point midway between the charges? (b)
What force (magnitude and direction) would act on an electron placed there?
~ - Two point charges are a distanced apart (Fig. 27-16). Plot E(x), assuming x = 0 at
the left-hand charge. Consider both positive and negative values of x. Plot E as positive
if E points to the right and negative if E points to the left. Assume q1 = + 1.0 X 10- 6
coul, q2 = +3.0 X 10-s coul, and d = 10 cm.

+q

a a


I d !• p
x-
i.ql x _q_2_ _ _ _ ~J +Q a
-Q
Fig. 27-16 . Fig. 27-17

8. Three charges are arranged m an equilateral triangle as in Fig. 27-17. What is


the direction of the force on +q?
'-4 Two point charges of magnitude +2.0 X 10-7 coul and +8.5 X 10-s coul are 12 cm
apart. (a) What electric field does each produce at the site of the other? (b) What force
acts on each? ,;,
10. Axial field due to an electric dipole. In Fig. 27-8, consider a point a distance r from
the center of the dipole along its axis. (a) Show that, at large values of r, the electric field is
1 p
E =---,
21reo r 3
682 THE ELECTRIC FIELD Chap. 27
which is twice the value given for the conditions of Example 3. (b) What is the direction
of E?
11. In Fig. 27-8 assume that both charges are positive. (a) Show that Eat point P in
that figure, assuming r » a, is given by
1 2q
E = ---·
41reo r 2

(b) What is the direction of E? (c) Is it reasonable that E should vary as r- 2 here and ac
r- 3 for the dipole of Fig. 27-8?
12. (a) In Fig. 27-18 locate the point (or points) at which the electric fieldstrength iE
zero. (b) Sketch qualitatively the lines of force. Take a = 50 cm.

+q a -2q

IE a
~1

-5q +2q • -q a
+2q
Fig. 27-18 Fig. 27-19

13. What is E in magnitude and direction at the center of the square of Fig. 27-19?
Assume that q = 1.0 X 10-8 coul and a = 5.0 cm.
T4. Two point charges of unknown magnitude and sign are placed a distance d apart.
(a) If it is possible to have E = 0 at any point not between the charges but on the line
joining them, what are the necessary conditions and where is the point located? (b) Is it
possible, for any arrangement of two point charges, to find two points (neither at infinity)
at which E = O; if so, under what conditions?
15. A thin nonconducting rod of finite length l carries a total charge q, spread uniformly
along it. Show that Eat point Pon the perpendicular bisector in Fig. 27-20 is given by

E = _q~ 1
2noy v z2 + 4y2
Show that as l --> oo this result approaches that of Example 6.
p

Fig. 27-20
Chap. 27 PROBLEMS 683
16. A thin nonconducting rod is bent to form the arc of a circle of radius a and subtends
an angle Bo at the center of the circle. A total charge q is spread uniformly along its length.
Find the electric field strength at the center of the circle in terms of a, q, and Bo.
17. A nonconducting hemispherical cup of inner radius a has a total charge q spread
uniformly over its inner surface. Find the electric field at the center of curvature.
18. Electric quadrupole. Figure 27-21 shows a typical electric quadrupole. It consists
of two dipoles whose effects at external points do not quite cancel. Show that the value of
E on the axis of the quadrupole for points distant r from its center (assumer » a) is given
by

where Q ( = 2qa 2 ) is called the quadrupole moment of the charge distribution.

~----P

+q

l+P
}-2q
~-p
+q

Fig. 27-21

19. An electron is constrained to move along the axis of the ring of charge in Example 5.
Show that the electron can perform oscillations whose frequency is given by

This formula holds only for small oscillations, that is, for x « a in Fig. 27-10. (Hint:
Show that the motion is simple harmonic and use Eq. 15-11.)
20. For the ring of charge in Example 5, show that the maximum value of E occurs at
x = a/v2.
21. Consider the ring of charge of Example 5. Suppose that the charge q is not distri-
buted uniformly over the ring but that charge _q 1 is distributed uniformly over half the
circumference and charge q2 is distributed uniformly over the other half. Let qi + q2 = q.
(a) Find the component of the electric field at any point on the axis directed along the axis
and compare with the uniform case of Example 5. (b) Find the component of the electric
field at any point on the axis perpendicular to the axis and compare with the uniform case
of Example 5.
'22. A thin circular disk of radius a is charged uniformly so as to have a charge per unit
area of u. Find the electric field on the axis of the disk at a distance r from the disk.
684 THE ELECTRIC FIELD Chap. 27
23. Field due to an electric dipole. Show that the components of E due to a dipole are
given, at distant points, by
1 3pxy
Ex= - -
4Ho (x 2 +
y 2)%

1 p(2y 2 - x 2)
E =-- ,
Y 4Ho (x 2 + 2
y )%

where x and y are coordinates of a point in Fig. 27-22. Show that this general result
includes the special results of Eq. 27-8b and of Problem 10.

y p

I
+q IY
I
a +----1o"-----.:..:x_ ___._I _ _ _ _ x
a

-q

Fig. 27-22

24. What is the magnitude and direction of an electric field that will balance the weight
of (a) an electron and (b) an alpha particle?
25. A particle having a charge of -2.0 X 10-9 coul is acted on by a downward electric
force of 3.0 X 10-5 newton in a uniform electric field. (a) What is the strength of the
electric field'? (b) What is the magnitude and direction of the electric force exerted on a
proton placed in this field? (c) What is the gravitational force on the proton? (d) What is
the ratio of the electric to the gravitational forces in this case?
26. (a) What is the acceleration of an electron in a uniform electric field of 106 nt/coul?
(b) How long would it take for the electron, starting from rest, to attain one-tenth the
speed of light? (c) What considerations limit the applicability of Newtonian mechanics
Jo such problems?
"'·27 An electron moving with a speed of 5.0 X 108 cm/sec is shot parallel to an electric
field of strength 1.0 X 103 nt/coul arranged so as to retard its motion. (a) How far will
the electron travel in the field before coming (momentarily) to rest, and (b) how much
time will elapse? (c) If the electric field ends abruptly after 0.8 cm, what fraction of its
initial kinetic energy will the electron lose in traversing it?
28. An electron is projected as in Fig. 27-23 at a speed of 6.0 X 106 meters/sec and at
an angle e of 45°; E = 2.0 X 103 nt/coul (directed upward), d = 2.0 cm, and l = 10.0 cm
(a) Will the electron strike either of the plates? (b) If it strikes a plate, where does it do so?
29. Dipole in a nonuniform field. Derive an expression for dE /dz at a point midway
between two equal positive charges, where z is the distance from one of the charges, meas-
ured along the line joining them. Would there be a force on a small dipole placed at this
point, its axis being aligned with the z axis? Recall that E = 0 at this point.

T
d
_J_

Fig. 27-23
Chap. 27 PROBLEMS 685
30. Oil drop experiment. R. A. Millikan set up ar1 apparatus (Fig. 27-24) in which a
tiny, charged oil drop, placed in an electric field E, could be "balanced" by adjusting E
until the electric force on the drop was equal and opposite to its weight. If the radius of
the drop is 1.64 X 10-4 cm and E at balance is 1.92 X 105 nt/coul, (a) what charge is on
the drop? (b) Why did Millikan not try to balance electrons in his apparatus instead of
oil drops? The density of the oil is 0.851 gm/cm3 . (Millikan first measured the electronic
charge in this way. He measured the drop radius by observing the limiting speed that
the drops attained when they fell in air with the electric field turned off. He charged
the oil drops by irradiating them with bursts of X-rays.)

Microscope

Fig. 27-24 Millikan's oil drop apparatus. Charged oil drops from the atomizer fall through
the hole in the central plate.

31. In a particular early run (1911), Millikan observed that the following measured
charges, among others, appeared at different times on a single drop:
6.563 X 10-19 coul 13.13 X 10-19 coul 19.71 X 10-19 coul
8.204 X 10-19 coul 16.48 X 10-19 coul 22.89 X 10-19 coul
11.50 X 10-19 coul 18.08 X 10-19 coul 26.13 X 10-19 coul
What value for the elementary charge e can be deduced from these data?
32. An electric field E with an average magnitude of about 150 nt/coul points downward
in the earth's atmosphere. We wish to "float" a sulfur sphere weighing 1.0 lb in this field
by charging it. (a) What charge (sign and magnitude) must be used? (b) Why is the
experiment not practical? Give a qualitative reason supported by a very rough numerical
calculation to prove your point.
Gauss's Law
CHAPTER 28

28-1 Flux of the Electric Field


Flux (symbol <I>) is a property of any vector field; it refers to a hypothetical
surface in the field, which may be closed or open. For a flow field the flux
(<I>v) is measured by the number of streamlines that cut through such a sur-
face. For an electric field the flux (<I> E) is measured by the number of lines
of force that cut through such a surface.
For closed surfaces we shall see that <l>E is positive if the lines of force point
outward everywhere and negative if they point inward. Figure 28-1 shows
two equal and opposite charges and their lines of force. Curves 8 1 , 8 2 , 8 3 , and
84 are the intersections with the plane of the figure of four hypothetical closed
surfaces. From the statement just given, <I>E is positive for surface 8 1 and
negative for 8 2 • The flux of the electric field is important because Gauss's
law, one of the four basic equations of electromagnetism (see Table 38-3), is
expressed in terms of it. Although the concept of flux may seem a little
abstract at first, the student will soon see its value in solving problems.
To define <l>E precisely, consider Fig. 28-2, which shows an arbitrary closed
surface immersed in an electric field. Let the surface be divided into ele-
mentary squares D.8, each of which is small enough so that it may be con-
sidered to be plane. Such an element of area can be represented as a vector
D.S, whose magnitude is the area D.8; the direction of D.S is taken as the
outward-drawn normal to the surface.
At every square in Fig. 28-2 we can also construct an electric field vector
E. Since the squares have been taken to be arbitrarily small, E may be taken
as constant for all points in a given square.
686
Sec. 28-1 FLUX OF THE ELECTRIC FIELD 687

Fig. 28-1 Two equal and opposite charges. The dashed lines represent hypothetical
closed surfaces.

The vectors E and ~S that characterize each square make an angle ewith
each other. Figure 28-2b shows an enlarged view of the three squares on the
surface of Fig. 28-2a marked x, y, and z. Nate that at x, e > go 0 ; at y, e =
go 0 ; and at z, e < go 0.
A semiquantitative definition of flux is

(28-1)

which instructs us to add up the scalar quantity E · ~S for all elements of area
into which the surface has been divided. For points such as x in Fig. 28-2 the
contribution to the flux is negative; at y it is zero and at z it is positive.
Thus if E is everywhere outward, e < go 0, E · ~S will be positive, and <FE for
the entire surface will be positive; see Fig. 28-1, surface S 1 . If E is every-
where inward, e > go 0 , E· ~Swill be negative, and <FE for the surface will be
negative; see Fig. 28-1, surface S2 . From Eq. 28-1 we see that the appro-
priate mks unit for <FE is the newton-meter 2 / coul.
The exact definition of electric flux is found in the diff""rential limit of Eq.
68 8 G AU ss 's LA W Chap. 28
e
e by an in te gr al over the surfac
rfac
e sum over the su
ds. Heplacing th
28el--1
yi
(2 8- 2)
divided
tl1e su rf ac e in question is to be is
at
hi s su rf ac e in te gral indicates th d th at th e sc al ar qu an ti ty E ·d S
T dS an surface.
elements of ar ea ke n for th e entire
into infinitesimal an d th e su m ta
for each element
to be evaluated

A.S

(x) (y ) (z )

(b )
t,
ot dc field. (b) Tl m ,e el em en
ic al su rf ac e ;m meraed ;n an de
-2 (a) A hy po
th et
e, shown enlarged
.
figar
of ea on this surfac
. 28
Sec. 28-1 FLUX OF THE ELECTRIC FIELD 689

Fig. 28-3 Example 1. A cylindrical surface immersed in a uniform field E parallel to


its axis.

The circle on the integral sign indicates that the surface of integration is a
closed surface.*
.... Example I. Figure 28-3 shows a hypothetical cylinder of radius R immersed in
a uniform electric field E, the cylinder axis being parallel to the field. What is <I> E
for this closed surface?
The flux cf,E can be written as the sum of three terms, an integral over (a) the left
cylinder cap, (b) the cylindrical surface, and (c) the right cap. Thus

=f (a)
E·dS +f (b)
E·dS +f (c)
E·dS.

For the left cap, the angle 8 for all points is 180°, E has a constant value, and the
vectors dS are all parallel. Thus

f (a)
E·dS = f E cos 180° dS
= -E dS f = -ES,

where S ( = 1rR 2) is the cap area. Similarly, for the right cap,

f (c)
E·dS = +ES,

the angle 8 for all points being zero here. Finally, for the cylinder wall,

f (b)
E·dS = 0,

because 8 = 90°, hence E·dS = 0 for all points on the cylindrical surface. Thus

<l>E = -ES+ 0 +ES= 0.


* Similarly, a circle on a line integral sign indicates a closed path. It will be clear from
the context and from the differential element (dS in this case) whether we are dealing
with a surface integral or a line integral.
690 GAUSS'S LAW Chap. 28
28-2 Gauss's Law
Gauss's law, which applies to any closed hypothetical surface (called a
Gaussian surface), gives a connection between cf>E for the surface and the net
charge q enclosed by the surface. It is
(28-3)

or, using Eq. 28-2, i:ofE·dS = q. (28-4)

The fact that cf>E proves to be zero in Example 1 is predicted by Gauss's law
because no charge is enclosed by the Gaussian surface in Fig. 28-3 (q = 0).
Note that q in Eq. 28-3 (or in Eq. 28-4) is the net charge, taking its alge-
braic sign into account. If a surface encloses equal and opposite charges,
the flux cf>E is zero. Charge outside the surface makes no contribution to the
value of q, nor does the exact location of the inside charges affect this value.
Gauss's law can be used to evaluate E if the charge distribution is so
symmetric that by proper choice of a Gaussian surface we can easily evaluate
the integral in Eq. 28-4. Conversely, if Eis known for all points on a given
closed surface, Gauss's law can be used to compute the charge inside. If E
has an outward component for every point on a closed surface, cf>E, as Eq.
28-2 shows, \vill be positive and, from Eq. 28-4, there must be a net positive
charge within the surface (see Fig. 28-1, surface 8 1 ). If E has an inward
component for every point on a closed surface, there must be a net negative
charge within the surface (see Fig. 28-1, surface 8 2 ). Surface 8 3 in Fig. 28-1
encloses no charge, so that Gauss's law predicts that cf>E = 0. This is con-
sistent with the fact that lines of E pass directly through surface 8 3 , the con-
tribution to the integral on one side canceling that on the other. What
would be the value of cf>E for surface 8 4 in Fig. 28-1, which encloses both
charges?

28-3 Gauss's Law and Coulomb's Law


Coulomb's law can be deduced from Gauss's law and symmetry considera-
tions. To do so, let us apply Gauss's law to an isolated point charge q as in
Fig. 28-4. Although Gauss's law holds for any surface whatever, informa-
tion can most readily be extracted for a spherical surface of radius r centered

Fig. 28-4 A spherical Gaussian surface of radius


r surrounding a point charge q.
GAUSS'S LAW AND COULOMB'S LAW 691

on the charge. The advantage of this surface is that, from symmetry, E must
be normal to it and must have the same (as yet unknown) magnitude for all
points on the surface.
In Fig. 28-4 both E and dS at any point on the Gaussian surface are
directed radially outward. The angle between them is zero and the quantity
E·dS becomes simply E dS. Gauss's law (Eq. 28-4) thus reduces to

eofE·dS = EofEdS = q.

Because Eis constant for all points on the sphere, it can be factored from
inside the integral sign, leaving

where the integral is simply the area of the sphere.* This equation gives
EoE (41rr 2 ) = q
1 q
or E = 41rEo r2. (28-5)

Equation 28-5 gives the magnitude of the electric field strength E at any
point a distance r from an isolated point charge q. The direction of E is
already known from symmetry.
Let us put a second point charge q0 at the point at which E is calculated.
The magnitude of the force that acts on it (see Eq. 27-2) is
F = Eq 0 .
Combining with Eq. 28-5 gives
1 qqo
F=--,
41re 0 r2
which is precisely Coulomb's law. Thus we have deduced Coulomb's law
from Gauss's law and considerations of symmetry.
Gauss's law is one of the fundamental equations of electromagnetic theory
and is displayed in Table 38-3 as one of Maxwell's equations. Coulomb's law
is not listed in that table because, as we have just proved, it can be deduced
from Gauss's law and from simple assumptions about the symmetry of E due
to a point charge.
It is interesting to note that writing the proportionality constant in Coulomb's law
as l/41rEo (see Eq. 26-3) permits a particularly simple form for Gauss's law (Eq. 28-3).
If we had written the Coulomb law constant simply ask, Gauss's law would have to
be written as (1/41rk)<'h = q. We prefer to leave the factor 41r in Coulomb's law so
that it will not appear in Gauss's law or in other much used relations that will be de-
rived later.
* The usefulness of Gauss's law depends on our ability to find a surface over which,
from symmetry, both E and e (see Fig. 28-2) have constant values. Then E cos e can be
factored out of the integral and E can be found simply, as in this example.
692 GAUSS'S LAW Chap. 28
28-4 An Insulated Conductor
Gauss's law can be used to make an important prediction, namely: An
excess charge, placed on an insulated conductor, resides entirely on its outer sur-
face. This hypothesis.was shown to be true by experiment (see Section 28-5)
before either Gauss's law or Coulomb's law were advanced. Indeed, the
experimental proof of the hypothesis is the experimental foundation upon
which both laws rest: We have already pointed out that Coulomb's torsion
balance experiments, although direct and convincing, are not capable of great
accuracy. In showing that the italicized hypothesis is predicted by Gauss's
law, we are simply reversing the historical situation.
Figure 28-5 is a cross section of an insulated conductor of arbitrary shape
carrying an excess charge q. The dashed lines show a Gaussian surface that
lies a small distance below the actual surface of the conductor. Although the
Gaussian surface can be as close to the actual surface as we wish, it is im-
portant to keep in mind that the Gaussian surface is inside the con-
ductor.
·when an excess charge is placed at random on an insulated conductor, it
will set up electric fields inside the conductor. These fields act on the charge
carriers of the conductor and cause them to move, that is, they set up internal
currents. These currents redistribute the excess charge in such a way that
the internal electric fields are automatically reduced in magnitude. Even-
tually the electric fields inside the conductor become zero everywhere, the
currents automatically stop, and electrostatic conditions prevail. This
redistribution of charge normally takes place in a time that is negligible for
most purposes. What can be said about the distribution of the excess
charge when such electrostatic conditions have been achieved"?
If, at electrostatic equilibrium, Eis zero everywhere inside the conductor,
it must be zero for every point on the Gaussian surface. This means that the
flux <l>E for this surface must be zero. Gauss's law then predicts (see Eq.
28-3) that there must be no net charge inside the Gaussian surface. If the
excess charge q is not inside this surface, it can only be outside it, that is, it
must be on the actual surface of the conditctor.

Silk thread

Fig. 28-5 An insulated conductor.


Sec. 28-5 EXPERIMENTAL PROOF OF GAUSS'S AND COULOMB'S LAWS 693

(a) (b) (c)

Fig. 28-6 The entire charge on the ball is transferred to the outside of the can. This state-
ment and the discussion of the first paragraph of Section 28-5 are strictly correct only if the
can is provided with a conducting lid which can be closed after the ball is inserted.

28-5 Experimental Proof of Gauss's and Coulomb's Laws


Let us turn to the experiments that prove that the hypothesis of Section
28-4. is true. For a simple test, charge a metal ball and lower it with a silk
thread deep into a metal can as in Fig. 28-6. Touch the ball to the inside of
the can; when the ball is removed from the can, all its charge will have
vanished. When the metal ball touches the can, the ball and can together
form an "insulated conductor" to which the hypothesis of Section 28-4
applies. That the charge moves entirely to the outside surface of the can
can be shown by touching a small insulated metal object to the can; only
on the outside of the can will it be possible to pick up a charge.
Benjamin Franklin seems to have been the first to notice that there can be no charge
inside an insulated metal can. In 1755 he wrote to a friend:
I electrified a silver pint cann, on an electric stand, and then lowered into it a
cork-ball, of about an inch diameter, hanging by a silk string, till the cork touched
the bottom of the cann. The cork was not attracted to the inside of the cann as it
would have been to the outside, and though it touched the bottom, yet when drawn
out, it was not found to be electrified by that touch, as it would have been by touch-
ing the outside. The fact is singular. You require the reason; I do not know
it . . . .
About ten years later Franklin recommended this "singular fact" to the attention
of his friend Joseph Priestley (1733-1804). In 1767 (about twenty years before
Coulomb's experiments) Priestley checked Franklin's observation and, with remark-
able insight, realized that the inverse square law of force followed from it. Thus the
694 GAUSS'S LAW Chap. 28
indirect approach is not only more accurate than the direct approach of Section 26-4
but was carried out earlier.
Priestley, reasoning by analogy with gravitation, said that the fact that no electric
force acted on Franklin's cork ball when it was surrounded by a deep metal can is
similar to the fact (see Section 16-6) that no gravitational force acts on a mass inside
a spherical shell of matter; if gravitation obeys an inverse square law, perhaps the
electrical force does also. In Priestley's words:
May we not infer from this [that is, Franklin's experiment] that the attraction of
electricity is subject to the same laws with that of gravitation and is therefore ac-
cording to the squares of the distances; since it is easily demonstrated that were the
earth in the form of a shell, a body in the inside of it would not be attracted to one
side more than another?
Michael Faraday also carried out experiments designed to show that excess charge
resides on the outside surface of a conductor. In particular, he built a large metal-
covered box which he mounted on insulating supports and charged with a powerful
electrostatic generator. In Faraday's words:
I went into the cube and lived in it, and using lighted candles, electrometers, and
all other tests of electrical states, I could not find the least influence upon them . . .
though all the time the outside of the cube was very powerfully charged, and large
sparks and brushes were darting off from every part of its outer surface.
Henry Cavendish (1731-1810) carried out an improved version of the
experiment of Fig. 28-6. With the instruments available to him, Cavendish
proved experimentally that the exponent in the force law lay, with high
probability, between 2.02 and 1.98. Cavendish, however, did not publish his
results so that almost nobody knew about them at the time. Maxwell re-
peated Cavendish's experiment with more accuracy and set these limits as
2.00005 and 1.99995. In 1936 Plimpton and Lawton repeated the experi-
ment again; they set the probability limits as 2.000000002 and 1.999999998.
In 1971 Williams, Faller, and Hill set the limits as 2.0000000000000003
and 1. 9999999999999997.
Figure 28-7 is an idealized sketch of the apparatus of Plimpton and Law-
ton. It consists in principle of two concentric metal shells, A and B, the
former being 5 ft in diameter. The inner shell contains a sensitive elec-
trometer E connected so that it will indicate whether any cha.rge moves be-
tween shells A and B.
By throwing switch S to the left, a substantial charge can be placed on the
sphere assembly. If any of this charge moves to shell B, it will have to pass
through the electrometer and will cause a deflection, which can be observed
optically using telescope T, mirror M, and windows W.
However, when the switch S is thrown alternately from left to right, thus
connecting the shell assembly either to the battery or to the ground, no effect
is observed on the galvanometer. This is the strongest experimental evidence
to date that the hypothesis of Section 28-4 is correct. Knowing the sensi-
tivity of their electrometer, Plimpton and Lawton calculated that the ex-
ponent in Coulomb's law lies, with high probability, between the limits
already stated.
Sec. 28-6 GAUSS'S LAW-SOME APPLICATIONS 695

,&
<JFfififiidi~----------
M~
·', c
T I .
Eye I
I
I

Fig. 28-7 The apparatus of Plimpton and Lawton.

28-6 Gauss's Law-Some Applications


Gauss's law can be used to calculate E if the symmetry of the charge dis-
tribution is high. One example of this, the calculation of E for a point charge,
has already been discussed (Eq. 28-5). Here we present other examples.
1111- Example 2. Spherically symmetric charge distribution. Figure 28-8 shows a
spherical distribution of charge of radius R. The charge density p (that is, the charge
per unit volume, measured in coul/meter 3) at any point depends only on the distance
of the point from the center and not on the direction, a condition called spherical
symmetry. Find an expression for E for points (a) outside and (b) inside the charge dis-
tribution. Note that the object in Fig. 28-8 cannot be a conductor or, as we have
seen, the excess charge will reside on its surface.
Applying Gauss's law to a spherical Gaussian surface of radius r in Fig. 28-8a (see
Section 28-3) leads exactly to Eq. 28-5, or

E = -~i.2
41rEo r
where q is the total charge. Thus for points outside a spherically symmetric dis-
696 GAUSS'S LAW Cl)ap. 28

(a) (b)

Fig. 28-8 Example 2. A spherically symmetric charge distribution, showing two Gauss-
ian surfaces. The density of charge, as the shading suggests, varies with distance from
the center but not with direction.

tribution of charge, the electric field has the value that it would have if the charge
were concentrated at its center. This reminds us that a sphere of mass m behaves
gravitationally, for outside points, as if the mass were concentrated at its center. At
the root of this similarity lies the fact that both Coulomb's law and the law of gravita-
tion are inverse square laws. The gravitational case was proved in detail in Section
16-6; the proof using Gauss's law in the electrostatic case is certainly much simpler.
Figure 28-8b shows a spherical Gaussian surface of radius r drawn inside the charge
distribution. Gauss's law (Eq. 28-4) gives

Eof E·dS = EoE(47rr 2 ) = q'


1 q'
or E=--,
471"Eo r 2

in which q' is that part of q contained within the sphere of radius r. The part of q
that lies outside this sphere makes no contribution to Eat radius r. This corresponds,
in the gravitational case (Section 16-6), to the fact that a spherical shell of matter
exerts no gravitational force on a body inside it.
An interesting special case of a spherically symmetric charge distribution is a uni-
form sphere of charge. For such a sphere, which would be suggested by uniform
shading in Fig. 28-8, the charge density p would have a constant value for all points
within a sphere of radius R and would be zero for all points outside this sphere. For
points inside such a uniform sphere of charge we can put

or q' = q (iY
where !71"R 3 is the volume of the spherical charge distribution. The expression for E
then becomes
1 qr
E=------· (28-6)
471"EO R 3
Sec. 28-6 GAUSS'S LAW-SOME APPLICATIONS 697
This equation becomes zero, as it should, for r = 0. Note that Eqs. 28-5 and 28-6
give the same result, as they must, for points on the surface of the charge distribution
(that is, if r = R). Note that Eq. 28-6 does not apply to the charge distribution of
Fig. 28-8b because the charge density, suggested by the shading, is not constant in
that case.
Example 3. The Thomson atom model. At one time the positive charge in the
atom was thought to be distributed uniformly throughout a sphere with a radius of
about 1.0 X 10- 10 meter, that is, throughout the entire atom. Calculate the electric
field strength at the surface of a gold atom (Z = 79) on this (erroneous) assump-
tion. Neglect the effect of the electrons.
The positive charge of the atom is Ze or (79)(1.6 X 10- 19 coul). Equation 28-5
yields, for E at the surface,
1 q
E=---
4rrEo r 2

(9.0 X 109 nt-m 2/coul2)(79)(1.6 X 10- 19 coul)


(1.0 X 10- 10 meter) 2

= 1.1 X 10 13 nt/ coul.

Figure 28-9 is a plot of E as a function of distance from the center of the atom, using
Eqs. 28-5 and 28-6. We see that E has its maximum value on the surface and de-
creases linearly to zero at the center (see Eq. 28-6). Outside the sphere E decreases
as the inverse square of the distance (see Eq. 28-5).
Example 4. The Rutherford, or nuclear, atom. We shall see in Section 28-7 that
the positive charge of the atom is not spread uniformly throughout the atom (see
Example 3) but is concentrated in a small region (the nucleus) at the center of the
atom. For gold the radius of the nucleus is about 6.9 X 10- 15 meter. What is the
electric field strength at the nuclear surface? Again neglect effects associated with
the atomic electrons.
The problem is the same as that of Example 3, except that the radius is much
smaller. This will make the electric field strength at the surface larger, in proportion
to the ratio of the squares of the radii. Thus

_ 13 . (1.0 X 10- 10 meter)2


E - (1.1 X 10 nt/coul) (6 9 _ )
. X 10 16 meter 2
= 2.3 X 1021 nt/ coul.

This is an enormous electric field, much stronger than could be produced and main-
tained in the laboratory. It is about 108 times as large as the field calculated in
Example 3.

1.2
::::,
0
~ 0.8
c: Fig. 28-9 Example 3. The electric
,....
c,:,
field due to the positive charge in a
8 0.4 gold atom, according to the (erro-
Pt:!.
neous) Thomson model.
0 .....
0 1.0 2.0 3.0
10
r, 10- meter
698 GAUSS'S LAW Chap. 28

Fig. 28-10 Example 5. An infinite rod of charge, showing a cylindrical Gaussian surface.

Example 5. Line of charge. Figure 28-10 shows a section ~ an infinite rod of


charge, the linear charge density A (that is, the charge per unit length, measured in
coul/meter) being constant for all points on the line. Find an expression for E at a
distance r from the line.
From symmetry, E due to a uniform linear charge can only be radially directed.
As a Gaussian surface we choose a circular cylinder of radius rand length h, closed at
each end by plane caps normal to the axis. E is constant over the cylindrical surface
and the flux of E through this surface is E(21rrh) where 21rrh is the area of the surface.
There is no flux through the circular caps because E here lies in the surface at every
point.
The charge enclosed by the Gaussian surface of Fig. 28-10 is Ah. Gauss's law
(Eq. 28-4),
l:ofE·dS = q,

then becomes l:oE(21rrh) = Ah,

whence E = _A__ (28-7)


21rl:or
The direction of Eis radially outward for a line of positive charge.
Note how much simpler the solution using Gauss's law is than that using integration
methods, as in Example 6, Chapter 27. Note too that the solution using Gauss's law
is possible only if we choose our Gaussian surface to take full advantage of the radial
symmetry of the electric field set up by a long line of charge. We are free to choose
any surface, such as a cube or a sphere, for a Gaussian surface. Even though
Gauss's law holds for all such surfaces, they are not all useful for the problem at
hand; only the cylindrical surface of Fig. 28-10 is appropriate in this case.
Gauss's law has the property that it provides a useful technique for calculation only
in problems that have a certain degree of symmetry, but in these problems the solu-
tions are strikingly simple.
Example 6. A sheet of charge. Figure 28-11 shows a portion of a thin, nonconduct-
ing, infinite sheet of charge, the surface charge density <I (that is, the charge per unit
area, measured in coul/meter 2 ) being constant. \Vhat is Eat a distance r in front of
the plane?
A convenient Gaussian surface is a "pill box" of cross-sectional area A and height
2r, arranged to pierce the plane as shown. From symmetry, E points at right angles
GAUSS'S LAW-SOME APPLICATIONS 699

Fig. 28-11 Example 6. An infinite


sheet of charge pierced by a cylindri-
cal Gaussian surface. The cross
section of the cylinder need not be
circular, as shown, but can have an
arbitrary shape.

to the end caps and away from the plane. Since E does not pierce the cylindrical
surface, there is no contribution to the flux from this source. Thus Gauss's law,

eofE·dS = q

becomes Eo(EA + EA) = uA


where uA is the enclosed charge. This gives

E = _!:___, (28-8)
21:0

Note that Eis the same for all points on each side of the plane; compare Fig. 27-2.
Although an infinite sheet of charge cannot exist physically, this derivation is still
useful in that Eq. 28-8 yields substantially correct results for real (not infinite) charge
sheets if we consider only points not near the edges whose distance from the sheet is
small compared to the dimensions of the sheet.
Example 7. A charged conductor. Figure 28-12 shows a conductor carrying on its
surface a charge whose surface charge density at any point is u; in general u will vary
from point to point. What is E for points a short distance above the surface?
--......_____

Fig. 28-12 Example 7. A charged insulated conductor,


showing a Gaussian surface. The cross section of the sur-
face need not be circular, as shown, but can have an arbi-
trary shape.

A
700 GAUSS'S LAW Chap. 28
The direction of E for points close to the surface is at right angles to the surface,
pointing away from the surface if the charge is positive. If E were not normal to the
surface, it would have a component lying in the surface. Such a component would
act on the charge carriers in the conductor and set up surface currents. Since there
are no such currents under the assumed electrostatic conditions, E must be normal
to the surface.
The magnitude of E can be found from Gauss's law using a small flat "pill box" of
cross section A as a Gaussian surface. Since E equals zero everywhere inside the
conductor (see Section 28-4), the only contribution to cI>E is through the plane cap
of area A that lies outside the conductor. Gauss's law

eofE·dS = q

becomes eo(EA) = a-A


where a-A is the net charge within the Gaussian surface. This yields
. (}"
E = -· (28-9)
Eo

Comparison with Eq. 28-8 shows that the electric field is twice as great near a
conductor carrying a charge whose surface charge density is u as that near a noncon-
ducting sheet with the same surface charge density. The student should compare
the Gaussian surfaces in Figs. 28-11 and 28-12 carefully. In Fig. 28-11 lines of force
leave the surface through each end cap, an electric field existing on both sides of the
sheet. In Fig. 28-12 the lines of force leave only through the outside end cap, the
inner end cap being inside the conductor where no electric field exists. If we assume
the same surface charge density and cross-sectional area A for the two Gaussian
surfaces, the enclosed charge ( = uA) will be the same. Since, from Gauss's law,
the flux cI>E must then be the same in each case, it follows that E ( = cI>E/ A) must
be twice as large in Fig. 28-12 as in Fig. 28-11. It is helpful to note that in Fig.
28-11 half the flux emerges from one side of the surface and half from the other,
whereas in Fig. 28-12 all the flux emerges from the outside surface. ~

28-7 The Nuclear Model of the Atom


Ernest Rutherford (1871-1937) was first led, in 1911, to assume that the
atomic nucleus existed when he tried to interpret some experiments carried
out at the University of Manchester by his collaborators H. Geiger and E.
Marsden.* The results of Examples 3 and 4 played an important part in
Rutherford's analysis of these experiments.
These workers, at Rutherford's suggestion, allowed a beam of a-particles t
to strike and be deflected by a thin film of a heavy element such as gold.
They counted the number of particles deflected through various angles </>.
Figure 28-13 shows the experimental setup schematically. Figure 28-14
shows the paths taken by typical a-particles as they scatter from a gold
atom; the angles <f> through which the a-particles are deflected range from
Oto 180° as the character of the collision varies from "grazing" to "head-on."

* See "The Birth of the Nuclear Atom," E. N. da C. Andrade, Scientific American,


November 1956. See also Example 5, Chapter 10.
ta-Particles are helium nuclei that are emitted spontaneously by some radioactive
materials such as radium. They move with speeds of the order of one-thirtieth that of
light when so emitted.
Sec. 28-7 THE NUCLEAR MODEL OF THE ATOM 701

Fig. 28-13 Experimental arrangement for studying the scattering of a-particles. Parti-
cles from radioactive source Sare allowed to fall on a thin metal "target" T; a-particles
scattered by the target through an (adjustable) angle 1> are counted by detector D.

The electrons in the gold atom, being so light, have almost no effect on the
motion of an oncoming a-particle; the electrons are themselves strongly de-
flected, just as a swarm of insects would be by a stone hurled through them.
Any deflection of the a-particle must be caused by the repulsive action of the
positive charge of the gold atom, which is known to possess most of the mass
of the atom.

Incident
a-particles

Fig. 28-14 The deflection of the incident a-particles depends on the nature of the nuclear
collision. (From Andrade, Scientific American, November 1956.)
702 GAUSS'S LAW Chap. 28
At the time of these experiments most physicists believed in the so-called
"plum pudding" model of the atom that had been suggested by J. J. Thom-
son (1856-1940). In this view (see Example 3) the positive charge of the
atom was thought to be spread out through the whole atom, that is, through.
a spherical volume of radius about 10- 10 meter. The electrons were thought
to vibrate about fixed centers inside this sphere.
Rutherford showed that this model of the atom was not consistent with
the a-scattering experiments and proposed instead the nuclear model of the
atom that we now accept. Here the positive charge is confined to a very
much smaller sphere whose radius is about 10-14 meter (the nucleus). The
electrons move around this nucleus and occupy a roughly spherical volume
of radius about 10-10 meter. This brilliant deduction by Rutherford laid
the foundation for modern atomic and nuclear physics.
The feature of the a-scattering experiments that attracted Rutherford's
attention at once was that a few a-particles are deflected through very large
angles, up to 180°. To scientists accustomed to thinking in terms of the
"plum pudding" model, this was a very surprising result. In Rutherford's
words: "It was quite the most incredible event that ever happened to me in
my life. It was almost as incredible as if you had fired a 15-inch shell at a
piece of tissue paper and it came back and hit you."
The a-particle must pass through a region in which the electric field
strength is very high indeed in order to be deflected so strongly.* Example
3 shows that, in Thomson's model, the maximum electric field strength is
1.1 X 10 13 nt/coul. Compare this with the value calculated in Example 4
for a point on the surface of a gold nucleus (2.3 X 1021 nt/coul). Thus the
deflecting force acting on an a-particle can be up to 108 times as great if the
positive charge of the atom is compressed into a small enough region (the
nucleus) at the center of the atoms. Rutherford made his hypothesis about
the existence of nuclei only after a much more detailed mathematical analysis
than that given here.

QUESTIONS

1. A point charge is placed at the center of a spherical Gaussian surface. Is <l>E changed
(a) if the surface is replaced by a cube of the same volume, (b) if the sphere is replaced by a
cube of one-tenth the volume, (c) if the charge is moved off-center in the original sphere,
still remaining inside, (d) if the charge is moved just outside the original sphere, (e) if a
second charge is placed near, and outside, the original sphere, and (f) if a second charge is
placed inside the Gaussian surface?
2. By analogy with <l>E, how would you define the flux <I>g of a gravitational field? What
is the flux of the earth's gravitational field through the boundaries of a room, assumed to
contain no matter?
3. In Gauss's law,

is Ethe electric field intensity attributable to the charge q?


* The chance that a big deflection can result from the combined effects of many small
deflections can be shown to be very small.
Chap. 28 PROBLEMS 703
4. Show that Eq. 18-3 illustrates what might be called Gauss's law for incompressible
fluids, or
<l>v = f v·dS = 0.

5. A surface encloses an electric dipole. What can you say about 'PE for this surface?
6. Suppose that a Gaussian surface encloses_ no net charge. Does Gauss's law require
that E equal zero for all points on the surface? Is the converse of this statement true, that
is, if E equals zero everywhere on the surface, does Gauss's law require that there be no net
charge inside?
7. Would Gauss's law hold if the exponent in Coulomb's law were not exactly two?
8. Does Gauss's law, as applied in Section 28-4, require that all the conduction electrons
in an insulated conductor reside on the surface?
9. In Section 28-4 we assumed that E equals zero everywhere inside a conductor.
However, there are certainly very large electric fields inside the conductor, at points close
to the electrons or to the nuclei. Does this invalidate the proof of Section 28-4?
10. It is sometimes said that excess charge resides entirely on the outer surface of a
conductor because like charges repel and try to get as far away as possible from one an-
other. Comment on this plausibility argument.
11. Is Gauss's law useful in calculating the field due to three equal charges located at
the corners of an equilateral triangle? Explain.
12. The use of line, surface, and volume densities of charge to calculate the charge
contained in an element of a charged object implies a continuous distribution of charge,
whereas, in fact, charge on the microscopic scale is discontinuous. How, then, is this pro-
cedure justified? ... _...
13. Is E necessarily zero inside a charged rubber balloon if the balloon is (a) spl:iJffcal or
(b) sausage-shapeg.? For each shape assume the charge to be distributed uniformly over
the surface. '
14. A spherical rubber balloon carries a charge that is uniformly distributed over its
surface. How does E vary for points (a) inside the balloon, (b) at the surface of the balloon,
and (c) outside the balloon, as the balloon is blown up?
15. As you penetrate a uniform sphere of charge, E should decrease because less charge
lies inside a sphere drawn through the observation point. On the other hand, E should
increase because you are closer to the center of this charge. Which effect predominates
and why?
16. Given a spherically symmetric charge distribution (not of uniform density of charge),
is E necessarily a maximum at the surface? Comment on various possibilities.
17. An atom is normally electrically neutral. Why then should an a-particle be de-
flected by the atom under any circumstances?
18. If an a-particle, fired at a gold nucleus, is deflected through 135 °, can you conclude
(a) that any force has acted on the a-particle or (b) that any net work has been done on it?

PROBLEMS

_l. Calculate 'PE through a hemisphere of radius R. The field of E is uniform and is
parallel to the axis of the hemisphere.
2. In Example 1 compute 'PE for the cylinder if it is turned so that its axis is perpendicu-
lar to the electric field. Do not use Gauss's law.
3. A plane surface of area A is inclined so that its normal makes an angle fJ with a uni-
fo.~m field of E. Calculate 'PE for this surface.
4. A point charge of 1.0 X 10-6 coul is at the center of a cubical Gaussian surface 0.50
meter on edge. What is 'PE for the surface?
704 GAUSS'S LAW Chap. 28
5. Charge on an originally uncharged insulated conductor is separated by holding a
positively charged rod nearby, as in Fig. 28-15. What can you learn from Gauss's law
about the flux for the five Gaussian surfaces shown? The induced negative charge on the
conductor is equal to the positive
charge on the rod.

/
--
->-&-"Gauss's law for gravitation"
....... .......
18
,,,,,...----- ___ 8
y _ __.,. . /
- 1-<1>, = - 1- -'g·dS = - m,
I / ..,.--- 84 ---- ..... 4,rG 4,rG 1'
/ 1 1/ 1,...-.81. . , -----
Conductor
8 '
1....---\ '
1
I I ;++ / :
r\ ~ ...+ r -:~
l __; I where m is the enclosed mass and
I
1
\ 1
\ ,_
', __ _________
, ........ ...,-,
__.

____ __
\. ...._ _
.,.,,/
__.. I I
I G is the universal gravitation
constant (Section 16-3). Derive
......
......
---- Newton's law of gravitation from
this.
Fig. 28-15 _.;r. Figure 28-16 shows a point
· charge of 1.0 X 10-7 coul at the
center of a spherical cavity of radius 3.0 cm in a piece of metal. Use Gauss's law to find
the electric field at point a, halfway from the center to the surface, and at point b.
/R An uncharged spherical thin metailic
/ shell has a point charge q at its center. Give
expressions for the electric field (a) inside the
shell, and (b) outside the shell, using Gauss's
law. (c) Has the shell any effect on the field
due to q? (d) Has the presence of q any effect
on the shell? (e) If a second point charge is
held outside the shell, does this outside charge
experience a force? (f) Does the inside
charge experience a force? (g) Is there a con-
tradiction with Newton's third law here?
,,,.,.,9: Two large nonconducting sheets of pos-
./ itive charge face each other as in Fig. 28-17.
What is Eat points (a) to the left of the sheets,
(b) between them, and (c) to the right of the
sheets? Assume the same surface charge
density u for each sheet. Assume a « l and Fig. 28-16
do not consider points near the edges.
(Hint: E at any point is the vector sum of the separate electric field strengths set up
by each sheet.) ~ a 4
+
+
!- a ~ ..
.

l l
+
+
+
+
+
+

J
+
+
+
+
+

Fig. 28-17 Fig. 28-18


Chap. 28 PROBLEMS 705

10. Two large metal plates face each other as in Fig. 28-18 and carry charges with
surface charge density +<1 and -<1, respectively, on their inner surfaces. What is Eat
points (a) to the left of the sheets, (b) between them, and (c) to the right of the sheets.
Assume a « l and do not consider points near the edges.
11. Two large metal plates of area 1.0 meter2 face each other. They are 5.0 cm apart
and carry equal and opposite charges on their inner surfaces. If E between the plates is
55 nt/coul, what is the charge on the plates? Neglect edge effects. See Problem 10.
12. A thin-walled metal sphere has a radius of 25 cm and carries a charge of 2.0 X 10-7
coul. Find E for a point (a) inside the sphere, (b) just outside the sphere, and (c) 3.0 meters
from the center of the sphere.
13. A 100-ev electron is·fired directly toward a large metal plate that has a surface
charge density of -2.0 X 10-6 coul/meter2 • From what distance must the electron be
fired if it is to just fail to strike the plate?
·-··44. Charge is distributed uniformly throughout an infinitely long cylinder of radius R.
Show that E at a distance r from the cylinder
axis (r < R) is given by

pr
E=-
2eo

where pis the density of charge (coul/meter3).


What result do you expect for r > R?
~-Hr.'Figure 28-19 shows a spherical non-
. conducting shell of charge of uniform density
p (coul/meter3). Plot E for distances r from
the center of the shell ranging from zero to
30 cm. Assume that p = 1.0 X 10-6 coul/
meter 3, a = 10 cm, and b = 20 cm.
16. Figure 28-20 shows a section through
a long, thin-walled metal tube of radius R,
carrying a charge per unit length X on its
surface. Derive expressions for E for var-
ious distances r from the tube axis, consider- Fig. 28-19
ing both r > R and r < R. Plot your re-
sults for the range r = 0 to r = 5 cm, assuming that X = 2.0 X 10-s coul/meter and
R = 3.0 cm.
_......----1-7-:··Figure 28-21 shows a section through two long concentric cylinders of radii a and b.
~ The cylinders carry equal and opposite charges per unit length X. Hsing Gauss's law

Fig. 28-20 Fig. 28-21


706 GAUSS'S LAW Chap. 28
prove (a) that E = 0 for r > band for r < a and (b) that between the cylinders Eis given
by
1 X
E=--·
21rEo r

18. In Problem 17 a positron revolves in a circular path of radius r, between and con-
centric with the cylinders. What must be its kinetic energy K? Assume a = 2.0 cm,
b = 3.Q ..crn, and X = 3.0 X 10-8 coul/meter.
~ A long conducting cylinder carrying a total charge +q is surrounded by a conducting
cylindrical shell of total charge -2q. as shown in cross section in Fig. 28-22. Use Gauss's

Fig. 28-22

law to find (a) the electric field strength at points outside the conducting shell, (b) the
distribution of the charge on the conducting shell, and (c) the electric field strength in the
region between the cylinders. List the assumptions made in arriving at your answers.
20. A thin metallic spherical shell of radius a carries a charge qa. Concentric with it is
another thin metallic spherical shell of radius b (b > a) carrying a charge qb. Use Gauss's
law to find the electric field strength at radial points r where {a) r < a; (b) a < r < b; (c)
r > b. (d) How is the charge on each shell distributed between the inner and outer
surfaces of that shell?
[2l~'A small sphere whose mass mis 1.0 X 10-3 gm carries a charge q of 2.0 X 10-8 coul.
It hangs from a silk thread which makes an angle of 30° with a large, uniformly charged
sheet as in Fig. 28-23. Calculate the surface charge density u for the sheet.

+
.u
+

+
+
+ ()
+
+
+
+
+

+ m,q
+
+

Fig. 28-23
Chap. 28 PROBLEMS 707

22. Equation 28-9 (E = u / EO) gives the electric field at points near a charged conducting
surface. Show that this equation leads to a familiar result when applied to a conducting
sphere of radius r, carrying a charge q.
23. An a-particle, approaching the surface of a nucleus of gold, is a distance equal to
one nuclear radius (6.9 X 10-u meter) away from that surface. What is the electro-
static force on the a-particle and its acceleration at that point? The mass of the a-
particle, which may be treated here as a point, is 6.7 X 10-27 kg.
24. A gold foil used in a Rutherford scattering experiment is 3 X 10-0 cm thick.
(a) What fraction of its surface area is "blocked out" by gold nuclei, assuming a nuclear
radius of 6.9 X 10-15 meter? Assume that no nucleus is screened by any other. (b)
What fraction of the volume of the foil is occupied by the nuclei? (c) What fills all the
rest of the space in the foil?
25. The electric field components in Fig. 28-24 are Ex = bx72 , Ey = Ez = 0, in which
b = 800 nt/coul-m%. Calculate (a) the flux 'PE through the cube and (b) the charge within
the cube. Assume that a = 10 cm.

Fig. 28-24
Electric Potential
CHAPTER 29

29-1 Electric Potential


The electric field around a charged rod can be described not only by a
(vector) electric field strength E but also by a scalar quantity, the electric
potential V. These quantities are intimately related, and often it is only a
matter of convenience which is used in a given problem.
To find the electric potential difference between two points A and B in an
electric field, we move a test charge q0 from A to B, always keeping it in
equilibrium, and we measure the work W AB that must be done by the agent
moving the charge. The electric potential difference * is defined from

WAB
VB - VA=--· (29-1)
qo
The work WAB may be (a) positive, (b) negative, or (c) zero. In these cases
the electric potential at B will be (a) higher, (b) lower, or (c) the same as the
electric potential at A.
The mks unit of potential difference that follows from Eq. 29-1 is the
joule/coul. This combination occurs so often that a special unit, the volt,
is used to represent it; that is,

1 volt = 1 joule/ coul.


* This definition of potential difference, though conceptually sound and suitable for
our present purpose, is rarely carried out in practice because of technical difficulties.
Equivalent and more technically feasible methods are usually adopted.
708
Sec. 29-1 ELECTRIC POTENTIAL 709

Usually point A is chosen to be at a large (strictly an infinite) distance


from all charges, and the electric potential VA at this infinite distance is
arbitrarily taken as zero. This allows us to define the electric potential at a
point. Putting VA = 0 in Eq. 29-1 and dropping the subscripts leads to
w
V=-, (29-2)
qo
where W is the work that an external agent must do to move the test charge
q0 from infinity to the point in question. The student should keep in mind
that potential differences are of fundamental concern and that Eq. 29-2 de-
pends on the arbitrary assignment of the value zero to the potential VA at
the reference position (infinity); this reference potential could equally well
have been chosen as any other value, say -100 volts. Similarly any other
agreed-upon point could be chosen as a reference position. In many circuit
problems the earth is taken as a reference of potential and assigned the value
zero.
Bearing in mind the assumptions made about the reference position, we
see from Eq. 29-2 that V near an isolated positive charge is positive because
positive work must ,be done by an outside agent to push a (positive) test
charge in from infinity. Similarly, the potential near an isolated negative
charge is negative because an outside agent must exert a restraining force on
(that is, must do negative work on) a (positive) test charge as it comes in
from infinity. Electric potential as defined in Eq. 29-2 is a scalar because W
and q0 in that equation are scalars.
Both W AB and VB - VA in Eq. 29-1 are independent of the path followed
in moving the test charge from point A to point B. If this were not so, point
B would not have a unique electric potential (with respect to point A as a
defined reference position) and the concept of potential would have limited
usefulness.
We can easily prove that potential differences are path-independent for the
special case shown in Fig. 29-1. This figure illustrates the case in which the
two points A and B are in a field set up by a spherical charge q; the two points
are further chosen, for simplicity, to lie along a radial line. Although our

Fig. 29-1 A test charge qo is


moved from A to B in the field of
charge q along either of two paths.
The open arrows show E at three
points on path II.

q
710 ELECTRIC POTENTIAL Chap. 29
path-independence proof applies only to this special case, it illustrates the
general principles involved.
Point A in Fig. 29-1 may be taken as a defined reference point, and we
imagine a positive test charge q0 moved by an external agent from A to B.
We consider two paths, path I being a radial line between A and Band path
II being a completely arbitrary path between these two points. The open
arrows on path II show the electric force per unit charge that would act at
various points on a test charge q0 .
Path II may be approximated by a broken path made up of alternating
elements of arc and of radius. Since these elements can be arbitrarily small,
the broken path can be made arbitrarily close to the actual path. On path
II the external agent does work only along the radial segments because along
the arcs the force F and the displacement di are at right angles, F · di being
zero in such cases. The sum of the work done on the radial segments that
make up path II is the same as the work done on path I because each path
has the same array of radial segments. Since path II is arbitrary, we have
proved that the work done is the same for all paths connecting A and B.
Although this proof holds only for the special case of Fig. 29-1, the potential
difference is path-independent for any two points in any electrostatic field.
We discussed path independence in Section 8-2 for the general class of conserv-
ative forces; electrostatic forces, like gravitational forces, are conservative.
The locus of points, all of which have the same electric potential, is called
an equipotential surface. A family of equipotential surfaces, each surface
corresponding to a different value of the potential, can be used to give a
general description of the electric field in a certain region of space. We have
seen earlier (Section 27-3) that electric lines of force can also be used for this
purpose; in later sections (see, for example, Fig. 29-15) we explore the inti-
mate connection between these two ways of describing the electric field.
No work is required to move a test charge between any two points on an
equipotential surface. This follows from Eq. 29-1,
WAB
VB - VA=--,
qo
because W AB must be zero if VA = VB· This is true, because of the path
independence of potential difference, even if the path connecting A and B
does not lie entirely in the equipotential surface.
Figure 29-2 shows an arbitrary family of equipotential surfaces. The
work to move a charge along paths I and II is zero because all these paths
begin and end on the same equipotential surface. The work to move a
charge along paths I' and II' is not zero but is the same for each path because
the initial and the final potentials are identical; paths I' and II' connect the
same pair of equipotential surfaces.
From symmetry, the equipotential surfaces for a spherical charge are a
family of concentric spheres. For a uniform field they are a family of planes
at right angles to the field. In all cases (including these two examples) the
equipotential surfaces are at right angles to the lines of force and thus to E
Sec. 29-2 POTENTIAL AND FIELD STRENGTH 711

Fig. 29-2 Portions of four equipotential surfaces. The heavy lines show four paths along
which a test charge is moved.

(see Fig. 29-15). If E were not at right angles to the equipotential surface, it
would have a component lying in that surface. Then work would have to
be done in moving a test charge about on the surface. Work cannot be done
if the surface is an equipotential, so E must be at right angles to the surface.
There is a strong analogy between electrostatic forces and gravitational forces,
based on the fact that their fundamental laws are inverse square laws (see Eqs.
26-3 and 16-1) :
and F = Gm1m2.
g r2

Thus we can define the gravitational potential Vg (compare Eq. 29-2) from

w
Vg = - ,
m

where Wis the work required to move a test body of mass m from infinity to the point
in question. Gravitational equipotential surfaces can also be constructed; they prove
to be everywhere at right angles to the gravitational field strength vector g. For a
uniform gravitational field, such as that near the surface of the earth, these surfaces
are horizontal planes. This correlates with the facts that (a) no net work is required
to move a stone of mass m between two points with the same elevation and (b) the
same net work is required to move a stone along any path starting on a given hori-
zontal surface and ending on another.

29-2 Potential and Field Strength


Let A and B in Fig. 29-3 be two points in a uniform electric field E, set up
by an arrangement of charges not shown, and let A be a distanced in the field
direction from B. Assume that a positive test charge q0 is moved, by an
external agent and without acceleration, from A to B along the straight line
connecting them.
The electric force on the charge is q0 E and points down. To move the
charge in the way we have described we must counteract this force by
712 ELECTRIC POTENTIAL Chap. 29
applying an external force F of the same magnitude but directed upward.
The work W done by the agent that supplies this force is
WAB = Fd = q0 Ed. (29-3)
Substituting this into Eq. 29-1 yields

WAB
VB - vA = - - = Ed. (29-4)
qo
This equation shows the connection between potential difference and field
strength for a simple special case. Note from this equation that another mks
unit for Eis the volt/meter. The student may wish to prove that a volt/
meter is identical with a nt/coul; this latter unit is the one first presented for
E in Section 27-2.
In Fig. 29-3 B has a higher potential than A. This is reasonable because
an external agent would have to do positive work to push a positive test
charge from A to B. Figure 29-3 could be used as it stands to illustrate the
act of lifting a stone from A to Bin the uniform gravitational field near the
earth's surface.
What is the connection between V and E in the more general case in which
the field is not uniform and in which the test body is moved along a path that
is not straight, as in Fig. 29-4? The electric field exerts a force q0 E on the test
charge, as shown. To keep the test charge from accelerating, an external
agent must apply a force F chosen to be exactly equal to -q 0 E for all posi-
tions of the test body.
If the external agent causes the test body to move through a displacement
dl along the path from A to B, the element of work done by the external agent

1
Fig. 29-3 A test charge qo is moved
from A to B in a uniform electric
d field E by an external agent that
exerts a force Fon it.

'V I
-'------+-·A E
Sec. 29-2 POTENTIAL AND FIELD STRENGTH 713

Fig. 29-4 A test charge qo is


moved from A to B in a nonuni-
form electric field by an external
agent that exerts a force F on it.

is F · di. To find the total work W AB done by the external agent in moving
the test charge from A to B, we add up (that is, integrate) the work contri-
butions for all the infinitesimal segments into which the path is divided.
This leads to

Such an integral is called a line integral. Note that we have substituted -q 0 E


for its equal, F.
Substituting this expression for W AB into Eq. 29-1 leads to

WAB JB
VB - VA= - - = - E·dl. (29-5)
qo A

If point A is taken to be infinitely distant and the potential VA at infinity is


taken to be zero, this equation gives the potential Vat point B, or, dropping
the subscript B,
V = - i B

00
E·dl. (29-6)

These two equations allow us to calculate the potential difference between


any two points (or the potential at any point) if Eis known at various points
in the field .
...,. Example 1. In Fig. 29-3 calculate VB - VA using Eq. 29-5. Compare the
result with that obtained by direct analysis of this special case (Eq. 29-4).
In moving the test charge the element of path dl always points in the direction of
714 ELECTRIC POTENTIAL Chap. 29
motion; this is upward in Fig. 29-3. The electric field E in this figure points down
so that the angle e between E and dl is 180°.
Equation 29-5 then becomes

vB - vA = - L B B
E·dl = - LE cos 180° dl =LE dl.
B

Eis the constant for all parts of the path in this problem and can thus be taken out-
side the integral sign, giving
B
vB - vA = EL dl = Ed,

which agrees with Eq. 29-4, as it must.


Example 2. In Fig. 29-5 let a test charge qo be moved without acceleration from
A to B over the path shown. Compute the potential difference between A and B.

B ---+---+--~-+------1 --..._ C

Fig. 29-5 Example 2. A test


charge qo is moved along path ACE
in a uniform electric field by an ex-
ternal agent.

For the path AC we have e = 135° and, from Eq. 29-5,

Ve - VA= - IACE·dl = - ICA E cos 135° dl = VE 2 ICA dl.


The integral is the length of the line AC which is y2d. Thus

Ve - VA= ~(V2d) = Ed.

Points Band C have the same potential because no work is done in moving a charge
between them, E and di being at right angles for all points on the line CB. In other
words, B and C lie on the same equipotential surface at right angles to the lines of
force. Thus
v B - vA = v c - vA = Ed.
This is the same value derived for a direct path connecting A and B, a result to be
expected because the potential difference between two points is path independent. ....
Sec. 29-3 POTENTIAL DUE TO A POINT CHARGE 715
29-3 Potential Due to a Point Charge
Figure 29-6 shows two points A and B near an isolated point charge q.
For simplicity we assume that A, B, and q lie on a straight line. Let us com-
pute the potential difference between points A and B, assuming that a test
charge q0 is moved without acceleration along a radial line from A to B.
In Fig. 29-6 E points to the right and di, which is always in the direction of
motion, points to the left. Therefore, in Eq. 29-5,

E·dl = E cos 180° dl = -E dl.

However, as we move a distance dl to the left, we are moving in the direction


of decreasing r because r is measured from q as an origin. Thus

dl = -dr.

Combining yields E·dl = E dr.

Fig. 29-6 A test charge qo is moved by an external agent from A to Bin the field set up
by a point charge q.
716 ELECTRIC POTENTIAL Chap. 29
Substituting this into Eq. 29-5 gives

-JB E·dl = - frBEdr.


A rA
Combining with Eq. 27-4,
1 q
E=--
47rEo r2

leads to (29-7)

Choosing reference position A. to be at infinity (that is, letting rA ~ co),


choosing VA = 0 at this position, and dropping the subscript B leads to

1 q
V=--· (29-8)
47l"Eo r

This equation shows clearly that equipotential surfaces for an isolated point
charge are spheres concentric with the point charge (see Fig. 2t)-15:r). A
study of the derivation will show that this relation also holds for points ex-
ternal to spherically symmetric charge distributions.
~ Example 3. What must the magnitude of an isolated positive point charge be
for the electric potential at 10 cm from the charge to be + 100 volts?
Solving Eq. 29-8 for q yields

q = V47rEor = (100 volts)(47r)(8.9 X 10- 12 coul2/nt-m 2)(0.10 meter)

= 1.1 X 10- 9 coul.

This charge is comparable to charges that can be produced by friction.

Example 4. What is the electric potential at the surface of a gold nucleus'? The
radius is 6.6 X 10- 15 meter and the atomic number Z = 79.
The nucleus, assumed spherically symmetrical, behaves electrically for external
points as if it were a point charge. Thus we can use Eq. 29-8, or, recalling that the
proton charge is 1.6 X 10- 19 coul,

V = -~-<]_ = (9.0 X 109 nt-m 2/coul 2)(?9)(1.6 X 10- 19 coul)


47rEo r 6.6 X 10- 10 meter

= 1.7 X 107 volts.

29-4 A Group of Point Charges


The potential at any point due to a group of point charges is found by (a)
calculating the potential V n due to each charge, as if the other charges were
not present, and (b) adding the quantities so obtained, or (see Eq. 29-8)

(29-9)
A GROUP OF POINT CHARGES 717

where qn is the value of the nth charge and r n is the distance of this charge
from the point in question. The sum used to calculate Vis an algebraic sum
and not a vector sum like the one used to calculate E for a group of point
charges (see Eq. 27-5). Herein lies an important computational advantage
of potential over electric field strength.
If the charge distribution is continuous, rather than being a collection of
points, the sum in Eq. 29-9 must be replaced by an integral, or

V = f dV = -41re1 fdq-,r
0

,vhere dq
is a differential element of the charge distribution, r is its distance
from the point at which V is to be calculated, and dV is the potential it
establishes at that point.
~ Example 5. What is the potential at the center of the square of Fig. 29-7?
Assume that q1 = + 1.0 X 10-s coul, qz = -2.0 X 10-s coul, q3 = +3.0 X 10-s
coul, q4 = +2.0 X 10-s coul, and a = 1.0 meter.

a p a
Fig. 29-7 Example 5

The distance r of each charge from Pis a/ y2 or 0.71 meter. From Eq. 29-9

v = L vn = _1_ q1 + q2 + qa + q4
n 47rEO r
(9.0 X 109 nt-m 2/coul2)(1.0 - 2.0 + 3.0 + 2.0) X 10-s coul
0.71 meter
= 500 volts.
Is the potential constant within the square? Does any point inside have a negative
potential? Can you sketch roughly the intersection of the plane of Fig. 29-7 with
the equipottntial surface corresponding to zero volts?
,;.,C-"''"". ·--~-;, _,. ·,

<\, Example 6:) A charged disk.


Find the electric potential for points on the axis of a
unifurrrrtrcharged circular disk whose surface charge density is <7 (see Fig. 29-8).
Consider a charge element dq consisting of a flat circular strip of radius y and width
dy. We have
dq = <7(21ry)(dy),

where (21ry)(dy) is the area of the strip: All parts of this charge element are the same
distance r' ( = V y2 + r 2) from axial point P so that their contribution dV to the
718 ELECTRIC POTENTIAL Chap. 29

Fig. 29-8 Example 6. A point


P on the axis of a uniformly
charged circular disk of radius a.

electric potential at P is given by Eq. 29-8, or

dV = _I_ dq = _I <J'21ry dy
41re 0 r' 47re0 V y2 + r2
The potential Vis found by integrating over all the strips into which the disk can be
divided (Eq. 29-10) or
V = Jav = _<!__
2eo Jo
ra (y + r
2 2) -%y dy

= _<!__ ( ~ - r).
2eo

This general result is valid for all values of r. In the special case of r » a the quan-
tity V a2 + r2 can be approximated as

~ = r(1+ ;:y 2
= r(1 + i ~: + ···) ~ r + ~:,
in which the quantity in parentheses in the second member of this equation has been
expanded by the binomial theorem (see Appendix I). This equation means that V
becomes

where q ( = <J'1ra 2) is the total charge on the disk. This limiting result is expected
because the disk behaves like a point charge for r » a. ...._

29-5 Potential Due to a Dipole


Two equal charges, q, of opposite sign, separated by a distance 2a, consti-
tute an electric dipole; see Example 3, Chapter 27. The electric dipole
moment p has the magnitude 2aq and points from the negative charge to the
positive charge. Here we derive an expression for the electric potential V
at any point of space due to a dipole, provided only that the point is not too
close to the dipole.
A point Pis specified by giving the quantities rand() in Fig. 29-9. From
symmetry, it is clear that the potential will not change as point P rotates
about the z axis, rand() being fixed. Thus we need only find V(r,()) for any
plane containing this axis; the plane of Fig. 29-9 is such a plane. Applying
Sec. 29-5 POTENTIAL DUE TO A DIPOLE 719

Eq. 29-9 gives


V=LVn=V1+V2
n =
1
41rEo
(q q) =
r1 - r2
q r 2 - r1
41rEo r1r2 '

which is an exact relationship.


·we now limit consideration to points such that r» 2a. These approxi-
mate relations then follow from Fig. 29-9:
and
and the potential reduces to
q 2a cos e 1 p cos e
v = ------ (29-11)
41rEo r2 41rEo r2
in which p ( = 2aq) is the dipole moment. Note that V vanishes everywhere
in the equatorial plane (e = 90°). This reflects the fact that it takes no work
to bring a test charge in from infinity along the perpendicular bisector of the
dipole. For a given radius, V has its greatest positive value fore = 0 and its
greatest negative value for e = 180°. Note that the potential does not de-
pend separately on q and 2a but only on their product p.
It is convenient to call any assembly of charges, for which V at distant points is
given by Eq. 29-11, an electric dipole. Two point charges separated by a small dis-
tance behave this way, as we have just proved. However, other charge configurations
also obey Eq. 29-11. Suppose that by measurement at points outside an imaginary
box (Fig. 29-10) we find a pattern of lines of force that can be described quantitatively
by Eq. 29-11. We then declare that the object inside the box is an electric dipole,
that its axis is the line zz', and that its dipole moment p points vertically upward.
Many molecules have electric dipole moments. That for H20 in its vapor state is
6.1 X 10-30 coul-m. Figure 29-11 is a representation of this molecule, showing the
three nuclei and the surrounding electron cloud. The dipole moment pis represented

Fig. 29-9 A point Pin the field of an electric dipole.


720 ELECTRIC POTENTIAL Chap. 29

Fig. 29-10 If an object inside the


spherical box sets up the electric
field shown (described quantita-
tively by Eq. 29-11), it is an electric
dipole.

by the arrow on the axis of symmetry of the molecule. In this molecule the effective
center of positive charge does not coincide with the effective center of negative charge.
It is precisely because of this separation that the dipole moment exists.
Atoms, and many molecules, do not have permanent dipole moments. However,
dipole moments may be induced by placing any atom or molecule in an external
electric field. The action of the field (Fig. 29-12) is to separate the centers of positive
and of negative charge. We say that the atom becomes polarized and acquires an
induced electric dipole moment. Induced dipole moments disappear when the electric
field is removed.
Electric dipoles are important in situations other than atomic and molecular ones.
Radio and radar antennas are often in the form of a metal wire or rod in which elec-
trons surge back and forth periodically. At a certain time one end of the wire or rod
will be negative and the other end positive. Half a cycle later the polarity of the ends
is exactly reversed. This is an oscillating electric dipole. It is so named because its
dipole moment changes in a periodic way with time.

Fig. 29-11 A schematic repre-


sentation of a water molecule,
showing the three nuclei, the elec-
tron cloud, and the orientation of
the dipole moment.
Sec. 29-5 POTENTIAL DUE TO A DIPOLE 721

Fig. 29-12 (a) An atom, show-


ing the nucleus and the electron
cloud. The center of negative
charge coincides with the center
of positive charge, that 1s, with
the nucleus. (b) If an external
field E is applied, the electron
cloud is distorted so that the cen-
ter of negative charge, marked
by the dot, and the center of posi-
tive charge no longer coincide.
An electric dipole appears .

.... Example 7. An electric quadrupole. An electric quadrupole, of which Fig. 27-21


is an example, consists of two electric dipoles so arranged that they almost, but not
quite, cancel each other in their electric effects at distant points.* Calculate V(r)
for points on the axis of this quadrupole.
Applying Eq. 29-9 to Fig. 27-21 yields

1 ( -q - -
V=LVn=- 2q+ q- )
n 47!'eo r - a r r +a
q 2a 2
41reo (r - a)(r)(r + a)
Assuming that r » a allows us to put a = 0 in the sum and difference terms in the
denominator, yielding
1 Q
V=--,
47!'eo r 3

where Q ( = 2qa 2) is the electric quadrupole moment of the charge assembly of Fig.
27-21. Note that V varies (a) as 1/r for a point charge (see Eq. 29-8), (b) as 1/r2 for
a dipole (see Eq. 29-11), and (c) as 1/r3 for a quadrupole.
Note too that (a) a dipole is two equal and opposite charges that do not coincide in
space so that their electric effects at distant points do not quite cancel, and (b) a

* See Problem 18, Chapter 27.


722 ELECTRIC POTENTIAL Chap. 29
quadrupole is two equal and opposite dipoles that do not coincide in space so that
their electric effects at distant points again do not quite cancel. This pattern can be
extended to define higher orders of charge distribution such as octupoles.
The potential at points at distances from an arbitrary charge distribution (continu-
ous or discrete) that are large compared with the size of the distribution can always be
written as the sum of separate potential distributions due to (a) a single charge-
sometimes, in this context, called a monopole-(b) a dipole, (cJ a quadrupole, etc.
This process is called an expansion in multipoles and is a very useful technique in
problem solving. <11111

29-6 Electric Potential Energy *


If we raise a stone from the earth's surface, the work that we do against
the earth's gravitational attraction is stored as potential energy in the system
earth + stone. If we release the stone, the stored potential energy changes
steadily into kinetic energy as the stone drops. After the stone comes to rest
on the earth, this kinetic energy, equal in magnitude just before the time of
contact to the originally stored potential energy, is transformed into heat
energy in the system earth + stone.
A similar situation exists in electrostatics. Consider two charges q1 and q2
a distance r apart, as in Fig. 29-13. If we increase the separation between

·f----------..~
ql

l--r J
q2

Fig. 29-13

them, an external agent must do work that will be positive if the charges are
opposite in sign and negative otherwise. The energy represented by this
work can be thought of as stored in the system q1 + q2 as electric potential
energy. This energy, like all varieties of potential energy, can be trans-
formed into other forms. If q1 and q2 , for example, are charges of opposite
sign and we release them, they will accelerate toward each other, transform-
ing the stored potential energy into kinetic energy of the accelerating masses.
The analogy to the earth + stone system is exact, save for the fact that
electric forces may be either attractive or repulsive whereas gravitational
forces are always attractive.
VVe define the electric potential energy of a system of point charges as the
work required to assemble this system of charges by bringing them in from an
infinite distance. ,v
e assume that the charges are all at rest when they are
infinitely separated, that is, they have no initial kinetic energy. \
In Fig. 29-13 let us imagine q2 removed to infinity and at rest. The

* In mechanics the concept of potential energy (of compressed springs, falling masses,
etc.) is more commonly used than the concept of potential. In electrostatics the reverse is
true, electric potential being perhaps a more common concept than electric potential
energy. In what follows the student must be careful not to confuse these quite different
quantities, potential and potential energy.
Sec. 29-6 ELECTRIC POTENTIAL ENERGY 723

electric potential at the original site of q2 , caused by q1, is given by Eq. 29-8,
or

If q2 is moved in from infinity to the original distance r, the work required is,
from the definition of electric potential, that is, from Eq. 29-2,

(29-12)

Combining these two equations and recalling that this work W is precisely
the electric potential energy U of the system q1 +
q2 yields

U (= W) = _l_q1q2. (29-13)
4'1l"eo r12

The subscript of r emphasizes that the distance involved is that between the
point charges q1 and qz.
For systems containing more than two charges the procedure is to compute
the potential energy for every pair of charges separately and to add the
results algebraically. This procedure rests on a physical picture in which (a)
charge q1 is brought into position, (b) q2 is brought from infinity to its posi-
tion near q1 , (c) q3 is brought from infinity to its position near q1 and q2 , etc.
The potential energy of continuous charge distributions (an ellipsoid of charge,
for example) can be found by dividing the distribution into infinitesimal elements dq,
treating each such element as a point charge, and using the procedures of the preced-
ing paragraph, with the summation process replaced by an integration. We have
not considered such problems in this text.

~ Example 8. Two protons in a nucleus of U 238 are 6.0 X 10- 15 meter apart.
What is their mutual electric potential energy?
From Eq. 29-13

U = _1_ q1q2 = (9.0 X 109 nt-m 2/coul2)(1.6 X 10-19 coul) 2


4'1!"Eo r 6.0 X 10- 15 meter

= 3.8 X 10- 14 joule = 2.4 X 10 5 ev.


Example 9. Three charges are arranged as in Fig. 29-14. What is their mutual
potential energy? Assume that q = 1.0 X 10-7 coul and a = 10 cm.

Fig. 29-14 Example 9.


Three charges are fixed
rigidly, as shown, by ex-
ternal forces.
724 ELECTRIC POTENTIAL Chap. 29
The total energy of the configuration is the sum of the energies of each pair of
particles. From Eq. 29-13,

U = U12 +U +U
13 23

= _1_ [< +q)(-4q) + (+q)( +2q) + (-4q)( +2q)J


41l'Eo a a a
10 q2
----
41l'Eo a

(9.0 + 10 9 nt-m 2/coul2)(10)(1.0 X 10-7 coul) 2


-9.0 X 10-3 joule.
0.10 meter
The fact that the total energy is negative means that negative work would have to be
done to assemble this structure, starting with the three charges separated and at rest
at infinity. Expressed otherwise, 9.0 X 10-3 joule of work must be done to dismantle
this structure, removing the charges to an infinite separation from one another.
When, as is common practice, infinity is taken as the zero of electric potential, a
positive potential energy (as in Example 8) corresponds to repulsive electric forces and
a negative potential energy (as in this example) to attractive electric forces. If the
protons in Example 8 were not held in place by attractive (nonelectrical) nuclear
forces, they would move away from each other. If the three particles in this example
were released from their fixed positions, in which they are held by external forces,
they would move toward each other. ..._

29-7 Calculation of E from V


We have stated that V and E are equivalent descriptions and have deter-
mined (Eq. 29-6) how to calculate V from E. Let us now consider how to
calculate E if Vis known throughout a certain region.
This problem has already been solved graphically. If Eis known at every
point in space, the lines of force can be drawn; then a family of equipotentials
can be sketched in by drawing surfaces at right angles. These equipotentials
describe the behavior of V. Conversely, if Vis given as a function of posi-
tion, a set of equipotential surfaces can be drawn. The lines of force can
then be found by drawing lines at right angles, thus describing the behavior
of E. It is the mathematical equivalent of this second graphical process that
we seek here. Figure 29-15 shows some examples of lines of force and of the
corresponding equipotential surfaces.
Figure 29-16 shows the intersection with the plane of the figure of a family
of equipotential surfaces. The figure shows that E at a typical point P is at
right angles to the equipotential surface through P, as it must be.
Let us move a test charge q0 from P along the path marked ~I to the
equipotential surface marked V + ~ V. The work that must be done by the
agent exerting the force F (see Eq. 29-1) is q0 ~ V.
From another point of view we can calculate the work from *

* We assume that the equipotentials are so close together that F is constant for all
parts of the path M. In the limit of a differential path (di) there will be no difficulty.
Sec. 29-7 CALCULATION OF E FROM V 725

where Fis the force that must be exerted on the charge to overcome exactly
the electrical force q0 E. Since F and q0 E have opposite signs and are equal in
magnitude,

These two expressions for the work must be equal, which gives

~v
or E cos O = - · (29-14)
~l

Now 1y cos () is the component of E in the direction -1 in Fig. 29-16; the


quantity -E cos (), which we call Ez, would then be the component of E in
the +I direction. In the differential limit Eq. 29-14 can then be written as

dV
Ez= - - · (29-15)
dl

In words, this equation says: If we travel through an electric field along a


straight line and measure Vas we go, the rate of change of V with distance
that we observe, when changed in sign, is the component of E in that direc-
tion. The minus sign implies that E points in the direction of decreasing V,
as in Fig. 29-16. It is particularly clear from Eq. 29-15 that appropriate
units for E are the volts/meter.

There will be one direction I for which the quantity dV /dl is a maximum. From
Eq. 29-15, E1 will also be a maximum for this direction and will in fact be E itself.
Thus
(29-16)

The maximum value of dV /dl at a given point is called the potential gradient at that
point. The direction I for which dV /dl has its maximum value is always at right
angles to the equipotential surface, corresponding to() = 0 in Fig. 29-16.
If we take the direction I to be, in turn, the directions of the x, y, and z axes, we
can find the three components of Eat any point, from Eq. 29-15.

av av Ez = - - ·
av (29-17)
Ex= - ax; Ey = - ay; az
Thus if Vis known for all points of space, that is, if the function V(x, y, z) is known,
the components of E, and thus E itself, can be found by taking derivatives.*

* The symbol av /ax is a partial derivative. It implies that in taking this derivative of the
function V(x, y, z) the quantity xis to be viewed as a variable and y and z are to be regarded
as constants. Similar considerations hold for av/ ay and av /az.
726 ELECTRIC POTENTIAL Chap. 29

/
/
/
/

/
/
/
/

(a)

(b)
Sec. 29-7 CALCULATION OF E FROM V 727

(c)

Fig. 29-15 Equipotential surfaces (dashed lines) and lines of force (solid lines) for (a) a
point charge, (b) a uniform electric field, and (c) an electric dipole. In all figures there is a
constant difference of potential ~ V between adjacent equipotential surfaces. Thus from
Eq. 29-14, written for the case of() = 180° as ~l = -~ V /E, the surfaces will be relatively
close together where E is relatively large and relatively far apart where E is small. Simi-
larly (see Section 27-3) the lines of force are relatively close together where Eis large and
far apart where Eis small. See discussion and figures of Section 18-7 for other examples.
728 ELECTRIC POTENTIAL Chap. 29

Fig. 29-16 A test charge q0 is


moved from one equipotential
surface to another along an arbi-
trarily selected direction marked 1.
V+ 2~V

V+~V

V-2~V

... Example 10. Calculate E(r) for a point charge q, using Eq. 29-16 and assuming
that V(r) is given as (see Eq. 29-8)
1 q
V=--·
41r1:o r

From symmetry, E must be directed radially outward for a (positive) point charge.
Consider a point Pin the field a distance r from the charge. It is clear that dV / dl at P
has its greatest value if the direction l is identified with that of r. Thus, from Eq.
29-16,

- - _q _!!:__
- 41r1:o dr r
(!) -- __1 !£.
41r1: r0 2

This result agrees exactly with Eq. 27-4, as it must.

Example II. Efor a dipole. Figure 29-17 shows a (distant) point Pin the field
of a dipole located at the origin of an xy-axis system. Vis given by Eq. 29-11, or

V = _1_ p cos~.
41r1:o r 2

Calculate E as a function of position.


From symmetry, E, for points in the plane of Fig. 29-17, lies in this plane. Thus
it can be expressed in terms of its components Ex and Ey. Let us first express the
potential function in rectangular coordinates rather than polar coordinates, making
use of
y
and cos 8 = (x2 + y2)Y2

The result is
Sec. 29-7 CALCULATION OF E FROM V 729
We find Eu from Eq. 29-17, recalling that xis to be treated as a constant in this
calculation:

Ey --
av p (x2 + y2)% _ yJ(x2 + y2)"Y2(2y)
ay 41rt=o (x2+ y2)3

p x2 - 2y2
- 41rt=o (x2 + y2)%

Note that putting x = 0 describes points along the dipole axis (that is, the y axis),
and the expression for Ey reduces to
E - 2p __!___
Y - 41rt=o y 3

This result agrees exactly with that found in Chapter 27 (see Problem 10), for, from
symmetry, Ex equals zero on the dipole axis.
Putting y = 0 in the expression for Ey describes points in the median plane of the
dipole and yields
p 1
Ey=----,
41reo x3

which agrees exactly with the result found in Chapter 27 (see Example 3), for, again
from symmetry, Ex equals zero in the median plane. The minus sign in this equation
indicatef\ that E points in the negative y direction (see Fig. 29-10).
The component Ex is also found from Eq. 29-17, recalling that y is to be taken as
a constant during this calculation:

av - _JJ_Jj__ ( - !) (x 2 + y2) -%(2x)


ax 41rt=o

3p xy
41reo (x2 +
y2)~1

As expected, Ex vanishes both on the dipole axis (x = O) and in the median plane
(y = 0); see Fig. 29-10. "41

(}
Fig. 29-17 Showing a point P (} r
in the field of an electric dipole p. y

x
!~--------'----~x
730 ELECTRIC POTENTIAL Chap. 29
29-8 An Insulated Conductor
We proved in Section 28-4, using Gauss's law, that after a steady state is
reached an excess charge q placed on an insulated conductor will move to its
outer surface. We now assert that this charge q will distribute itself on this
surface so that all points of the conductor, including those on the surface and
those inside, have the same potential.
Consider any two points A and Bin or on the conductor. If they were not
at the same potential, the charge carriers in the conductor near the point of
lower potential would tend to move toward the point of higher potential. We
have assumed, however, that a steady-state situation, in which such currents
do not exist, has been reached; thus all points, both on the surface and inside

VI
~
0
>
i:t:)
0
r-i

fig. 29-18 (a) The potential and


1 2 3 4 (b) the electric field strength, for
r, meter points near a conducting spherical
shell of 1.0 meter radius carrying
(a)
+
a charge of 1.0 X 10-6 coul.
12

--
~
....
Q)
Q)

E
VI
8
0
>
i:t:)

0 4
r-i

1 2 3 4
r, meter
(b)
Sec. 29-8 AN INSULATED CONDUCTOR 731

it, must have the same potential. Since the surface of the conductor is an
equipotential surface, E for points on the surface must be at right angles to
the surface.
We saw in Section 28-4 that a charge placed on an insulated conductor will
spread over the surface until E equals zero for all points inside. We now have
an alternative way of saying the same thing; the charge will move until all
points of the conductor (surface points and interior points) are brought to the
same potential, for if V is constant in the conductor then E is zero every-
where in the conductor (E 1 = -dV /dl).
Figure 29-18a is a plot of potential against radial distance for an isolated
spherical conducting shell of LO-meter radius carrying a positive charge of
1.0 X 10-6 coul. For points outside the shell V(r) can be calculated from
Eq. 29-8 because the charge q behaves, for such points, as if it were concen-
trated at the center of the sphere. Equation 29-8 is correct right up to the
surface of the shell. Now let us push the test charge through the surface,
assuming that there is a small hole, and into the interior. No extra work is
needed because no electrical forces act on the test charge once it is inside the
shell. Thus the potential everywhere inside is the same as that on the sur-
face, as Fig. 29-18a shows.
Figure 29-18b shows the electric field strength for this same sphere. Note
that E equals zero inside. The lower of these curves can be derived from the
upper by differentiation with respect tor, using Eq. 29-16; the upper can be
derived from the lower by integration with respect to r, using Eq. 29-6.
Figure 2~F18 holds without change if the conductor is a solid sphere rather
than a spherical shell as we have assumed. It is constructive to compare
Fig. 29-18b (conducting shell or sphere) with Fig. 28-9, which holds for a
nonconducting sphere. The student should try to understand the difference
between these two figures, bearing in mind that in the first the charge lies on
the surface whereas in the second it was assumed to have been spread uni-
formly throughout the volume of the sphere.
Finally we note that, as a general rule, the charge density tends to be high
on isolated conducting surfaces whose radii of curvature are small, and con-
versely. For example, the charge density tends to be relatively high on
sharp points and relatively low on plane regions on a conducting surface.
The electric field strength Eat points immediately above a charged surface
is proportional to the charge density u so that E may also reach very high
values near sharp points. Glow discharges from sharp points during thunder-
storms are a familiar example. The lightning rod acts in this way to
neutralize charged clouds and thus prevent lightning strokes.
We can examine the qualitative relationship between u and the curvature of the
surface in a particular case by considering two spheres of different radii connected
by a very long fine wire; see Fig. 29-19. Suppose that this entire assembly is raised
732 ELECTRIC POTENTIAL Chap. 29

Wire

Fig. 29-19 Two spheres connected by a long fine wire.

to some arbitrary potential V. The (equal) potentials of the two spheres are, from
Eq. 29-8, *

which yields (29-18)

where q1 is the charge on the sphere of radius R1 and q2 is the charge on the sphere of
radius R2.
The surface charge densities for each sphere are given by

and

Dividing gives
CT1 q1 Ri
CT2 = ~ Yil
Combining with Eq. 29-18 yields

which is consistent with our qualitative statement above. Note that the larger sphere
has the larger total charge but the smaller charge density.

The fact that er, and thus E, can become very large near sharp points is
important in the design of high-voltage equipment. Corona discharge can
result from such points if the conducting object is raised to high potential and
surrounded by air. Normally air is thought of as a nonconductor. However,
it contains a small number of ions produced, for example, by the cosmic rays.
A positively charged conductor will attract negative ions from the surround-
ing air and thus will slowly neutralize itself.

* Equation 29-8 holds only for an isolated point charge or spherically symmetric charge
distribution. The spheres must be assumed to be so far apart that the charge on either
one has a negligible effect on the distribution of charge on the other.
Sec. 29-9 THE ELECTROSTATIC GENERATOR 733

If the charged conductor has sharp points, the value of E in the air near the
points can be very high. If the value is high enough, the ions, as they are
drawn toward the conductor, will recC'ive such large accelerations that, by
collision with air molecules, they will produce vast additional numbers of
ions. The air is thus made much more conducting, and the discharge of the
conductor by this corona discharge may be very rapid indeed. The air
surrounding sharp conducting points may even glow visibly because of light
emitted from the air molecules during these collisions.

29-9 The Electrostatic Generator


The electrostatic generator was conceived by Lord Kelvin in 1890 and put
into useful practice in essentially its modern form by R. J. Van de Graaff in
1931. It is a device for producing electric potential differences of the order of
several millions of volts. Its chief application in physics is the use of this
potential difference to accelerate charged particles to high energies. Beams
of energetic particles made in this way can be used in many different "atom-
smashing" experiments. The technique is to let a charged particle "fall"
through a potential difference V, gaining kinetic energy as it does so.
Let a particle of (positive) charge q move in a vacuum under the influence
of an electric field from one position A to another position B whose electric
potential is lower by V. The electric potential energy of the system is re-
duced by qV because this is the work that an external agent would have to do
to restore the system to its original condition. This decrease in potential
energy appears as kinetic energy of the particle, or

K = qV. (29-19)

K is in joules if q is in coulombs and V in volts. If the particle is an elec-


tron or a proton, q will be the quantum of charge e.
If we adopt the quantum of charge e as a unit in place of the coulomb, we
arrive at another unit for energy, the electron volt, which is used extensively
in atomic and nuclear physics. By substituting into Eq. 29-19,

1 electron volt = (1 quantum of charge)(l volt)

(1.60 X 10-19 coul)(l.00 volt)

1.60 X 10- 19 joule.

The electron volt can be used interchangeably with any other energy unit.
Thus a 10-gm object moving at 1000 cm/sec can be said to have a kinetic
energy of 3.1 X 10 18 ev. Most physicists would prefer to express this result
as 0.50 joule, the electron volt being inconveniently small. In atomic and
nuclear problems, however, the electron volt (ev) and its multiples the Mev
( = 10 6 ev), the Gev ( = 109 ev) and the Tev ( = 10 12 ev) are the usual units
of choice.
734 ELECTRIC POTENTIAL Chap. 29
~ Example 12. The electrostatic generator. Figure 29-20, which illustrates the
basic operating principle of the electrostatic generator, shows a small sphere of radius
r placed inside a large spherical shell of radius R. The two spheres carry charges q
and Q, respectively. Calculate their potential difference.
The potential of the large sphere is caused in part by its own charge and in part
because it lies in the field set up by the charge q on the small sphere. From Eq. 29-8,

V R = __
l
47!'Eo
('lR + i)
R
·
The potential of the small sphere is caused in part by its own charge and in part
because it is inside the large sphere; see Fig. 29-18a. From Eq. 29-8,

The potential difference is


Vr - V R = _SL_ (~ -
471'Eo r R
!) ·
Thus, assuming q is positive, the inner sphere will always be higher in potential than
the outer sphere. If the spheres are connected by a fine wire, the charge q will flow
entirely to the outer sphere, regardless of the charge Q that may already be present.

Silk thread

Fig. 29-20 Example 12. A small


charged sphere of radius r is suspended
inside a charged spherical shell whose
outer surface has radius R.

Insulator
Sec. 29-9 THE ELECTROSTATIC GENERATO~ 735

Removable tank cover~


Equipotential shield,
Electronic circuits 68" 1.0.

Built-in 2 kw High voltage terminal,


power supply 38" diam

400 lb/in.2 steel


~--pressure tank

Insulating column Field control rods


18' long -----

Insulating belt

Manhole

Movable platform

Lead shielding

Dry ice trap - - -


~ Analyzing magnet

Adjustable magnet base

Feet

Fig. 29-21 An electrostatic generator at MIT capable of producing 9-Mev protons. The
proton beam is accelerated vertically downward, being deflected into a horizontal plane
by the analyzing magnet shown at the bottom. (Courtesy of J. G. Trump.)
736 ELECTRIC POTENTIAL Chap. 29
From another I?oint of view, we note that since the spheres when electrically con-
nected. form a ~mgle conductor at electrostatic equilibrium there can be only a single
potential. This means that Vr - V R = 0, which can occur only if q = O.
In actual ele?trosta~ic genera~ors charge is_ carried into the shell on rapidly moving
belts made of msulatmg material. Charge 1s "sprayed" onto the belts outside the
shell by corona discharge from a series of sharp metallic points connected to a source
of moderately high potential difference. Charge is removed from the belts inside the
shell by a similar series of points connected to the shell. Electrostatic generators can
be built commercially to accelerate protons to energies up to 10 Mev, using a single
acceleration. Figure 29-21 shows a schematic diagram of an electrostatic generator
at MIT that can produce 9-Mev protons.
Generators can be built in which the accelerated particles are subject to successive
accelerations. An electrostatic generator at the University of Pittsburgh subjects
particles to three successive accelerations.

QUESTIONS

1. Are we free to call the potential of the earth + 100 volts instead of zero? What effect
would such an assumption have on measured values of (a) potentials and (b) potential
differences?
2. What would happen to a person on an insulated stand if his potential was increased
by 10,000 volts?
3. Do electrons tend to go to regions of high potential or of low potential?
4. Suppose that the earth has a net charge that is not zero. Is it still possible to adopt
the earth as a standard reference point of potential and to assign the potential V = 0 to it?
5. Does the potential of a positively charged insulated conductor have to be positive?
Give an example to prove your point.
6. Can two different equipotential surfaces intersect?
7. If E equals zero at a given point, must V equal zero for that point? Give some ex-
amples to prove your answer.
8. If you know E at a given point, can you calculate V at that point? If not, what
further information do you need?
9. If V equals a constant throughout a given region of space, what can you say about
E in that region?
10. In Section 16-6 we saw that the gravitational field strength is zero inside a spherical
shell of matter. The electrical field strength is zero not only inside an isolated charged
spherical conductor but inside an isolated conductor of any shape. Is the gravitational
field strength inside, say, a cubical shell of matter zero? If not, in what respect is the
analogy not complete?
11. How can you insure that the electric potential in a given region of space will have
a constant value?
12. An isolated conducting spherical shell carries a negative charge. What will happen
if a positively charged metal object is placed in contact with the shell interior? Assume
that the positive charge is (a) less than, (b) equal to, and (c) greater than the negative
charge in magnitude.
13. An uncharged metal sphere suspended by a silk thread is placed in a uniform ex-
ternal electric field E. What is the magnitude of the electric field for points inside the
sphere? Is your answer changed if the sphere carries a charge?
14. A charge is placed on an insulated conductor in the form of a perfect cube. What
will be the relative charge density at various points on the cube (surfaces, edges, corners);
what will happen to the charge if the cube is in air?
Chap. 29 PROBLEMS 737

PROBLEMS

1. An infinite charged sheet has a surface charge density u of 1.0 X 10-7 coul/meter2 .
How far apart are the equipotential surfaces whose potentials differ by 5.0 volts?
2. A charge q is distributed uniformly throughout a nonconducting spherical volume of
radius R. (a) Show that the potential a distance a from the center, where a < R, is
given by
q(3R 2 - a 2)
V= .
81rEoR 3
(b) Is it reasonable that, according to this expression, V is not zero at the center of the
sphere?
/---3:-A charge of 10-s coul can be produced by simple rubbing. To what potential would
such a charge raise an insulated conducting sphere of 10-cm radius?
4. Consider a point charge with q = 1.5 X 10-s coul. (a) What is the radius of an
equipotential surface having a potential of 30 volts? (b) Are surfaces whose potentials
differ by a constant amount (say 1.0 volt) evenly spaced in radius?
5. In Fig. 29-22, locate the points (a) where V = 0 and (b) where E = 0. Consider
only points on the axis and choose d = 1.0 meter.

fig. 29-22

6. In Fig. 29-22 (see Problem 5) sketch qualitatively (a) the lines of force and (b) the
intersections of the equipotential surfaces with the plane of the figure. (Hint: Consider
the behavior close to each point charge and at considerable distances from the pair of
charges.)
~-In Fig. 29-~derive an expression for VA - VB· Does your result reduce to the
expected answer when d = O? When q = O?

B ••---11~-------•A
q
(a)

----1-E- d ---------;.1


+q •
-3q
q
(b)

fig. 29-23 fig. 29-24

,.$. A point charge has q = + 1.0 X 10-6 coul. Consider point A which is 2.0 meters
distant and point B which is 1.0 meter distant in a direction diametrically opposite, as in
Fig. 29-24a. (a) What is the potential difference VA - VB? (b) Repeat if points A and B
are located as in Fig. 29-24b.
9. Calculate the dipole moment of a water molecule under the assumption that all ten
electrons in the molecule circulate- symmetrically about the oxygen atom, that the OH
distance is 0.96 X 10-s cm, and that the angle between the two OH bonds is 104 °. Com-
pare with the value quoted on p. 719; see Fig. 29-11.
738 ELECTRIC POTENTIAL Chap. 29
10. For the charge configuration of Fig. 29-25, show that V(r) for points on the vertical
axis, assuming r » a, is given by
V = _l _
41Teo
(g + 2qa)
r r2
.

Is this an expected result? (Hint: The charge configuration can be viewed as the sum of
an isolated charge and a dipole.)

-,--____.p

+q

+q

-q

Fig. 29-25

11. In a typical lightning flash the potential difference between discharge points is
about 109 volts and the quantity of charge transferred is about 30 coul. How much ice
would it melt at 0°C if all the energy released could be used for this purpose?
12. Calculate (a) the electric potential established by the nucleus of a hydrogen atom
at the mean distance of the circulating electron (r = 5.3 X 10-11 meter), (b) the electric
potential energy of the atom when the electron is at this radius, and (c) the kinetic energy
of the electron, assuming it to be moving in a circular orbit of this radius centered on the
nucleus. (d) How much energy is required to ionize the hydrogen atom? Express all
energies in electron volts.
13. What is the electric potential energy of the charge configuration of Fig. 29-7? Use
the numerical values of Example 5.
14. (a) A spherical drop of water carrying a charge of 3 X 10-11 coul has a potential of
500 volts at its surface. What is the radius of the drop? (b) If two such drops, of the
same charge and radius, combine to form a single spherical drop, what is the potential at
the surface of the new drop so formed?
15. If the earth had a net charge equivalent to 1 electron/meter2 of surface area, (a)
what would the earth's potential be? (b) What would the electric field due to the earth
be just outside its surface?
16. Figure 29-26 shows an idealized representation of a U 238 nucleus (Z = 92) on the
verge of fission. Calculate (a) the repulsive force acting on each fragment and (b) the
mutual electric potential energy of the two fragments. Assume that the fragments are

Fig. 29-26
Chap. 29 PROBLEMS 739
equal in size and charge, spherical, and just touching. The radius of the initially spherical
U 238 nucleus is 8.0 X 10-15 meter. Assume that the material out of which nuclei are
made has a constant density.
17. In the Millikan oil drop experiment (see Fig. 27-24) an electric field of 1.92 X 105
nt/coul is maintained at balance across two plates separated by 1.50 cm. Find the
potential difference between the plates.
/--Hl':'(a) Show that the electric potential at a point on the axis of a ring of charge of
radius a, computed directly from Eq. 29-10, is given by

V=-1- q
4no V x2 + a2
(b) From this result derive an expression for E at axial points; compare with the direct
calculation of E in Example 5, Chapter 27.
~ In Example 6 the potential at an axial point for a charged disk was shown to be

V = __!!__ ( V a 2 + r2 - r).
2eo
From this result show that E for axial points is given by

E---
u ( 1- r ) ·
- 2eo Va2 + r2
Does this expression for E reduce to an expected result for (a) r » a and (b) for r = O?
20. (a) Starting from Eq. 29-11, find the magnitude Er of the radial component of the
electric field due to a dipole. (b) For what values of e is Er zero?
21. Can a conducting sphere 10 cm in radius hold a charge of 4 X 10-6 coul in air
without breakdown? The dielectric strength (minimum field required to produce break-
down) of air at 1 atm is 3 X 106 volts/meter.
22. A Geiger counter has a metal cylinder 2.0 cm in diameter along whose axis is
stretched a wire 0.005 in. in diameter. If 850 volts are applied between them, what is the
electric field strength at the surface of (a) the wire and (b) the cylinder?
23. Two metal spheres are 3.0 cm in radius and carry charges of + 1.0 X 10-s coul and
-3.0 X 10-s coul, respectively, assumed to be uniformly distributed. If their centers
are 2.0 meters apart, calculate (a) the potential of the point halfway between their centers
and (b) the potential of each sphere.
-----24-:"' In Fig. 29-19 let R 1 = 1.0 cm and R2 = 2.0 cm. Before the spheres are connected
by the fine wire, a charge of 2.0 X 10-7 coul is placed on the smaller sphere, the larger
sphere being uncharged. Calculate (a) the charge, (b) the charge density, and (c) the
potential for each sphere after they are connected.
25. The metal object in Fig. 29-27 is a figure of revolution about the horizontal axis.
If it is charged negatively, sketch roughly a few equipotentials and lines of force. Use
physical reasoning rather than mathematical analysis.

Axis

Fig. 29-27
740 ELECTRIC POTENTIAL Chap. 29
26. Devise an arrangement of three point charges, separated by finite distances, that
has zero electric potential energy.
27. Derive an expression for the work required to put the four charges together as
indicated in Fig. 29-28.

+q -q
a ...

a a

-q a +q

Fig. 29-28

28. A gold nucleus contains a positive charge equal to that of 79 protons. An a-particle
(Z = 2) has a kinetic energy K at points far from this charge and is traveling directly
toward the charge. The particle just touches the surface of the charge (assumed spherical)
and is reversed in direction. Calculate K, assuming a nuclear radius of 5.0 X 10-15
meter. The actual a-particle energy used in the experiments of Rutherford and his col-
laborators was 5.0 Mev. What do you conclude?
29. What is the potential gradient, in volts/meter, at a distance of 10-12 meter from
the center of the gold nucleus? What is the gradient at the nuclear surface?
30. For the spheres of Fig. 29-19, what is the ratio of electric field strengths at the sur-
face?
31. (a) Through what potential difference must an electron fall, according to Newtonian
mechanics, to acquire a speed v equal to the speed c of light? (b) Newtonian mechanics
fails as v ---> c. Therefore, using the correct relativistic expression for the kinetic energy

K = m 0 c2 [
Vl -
1
(v/c) 2
- 1 J
in place of the Newtonian expression K = tm v
0 2, determine the actual electron speed ac-
quired in falling through the potential difference computed in (a). Express this speed
as an appropriate fraction of the speed of light.
32. Two thin insulated concentric conducting spheres of radii R1 and R2 carry
charges q1 and q2. Derive expressions for E(r) and V(r), where r is the distance from
the center of the spheres. Plot E(r) and V(r) from r = 0 tor = 4.0 meters for R1 = 0.50
meter, R2 = 1.0 meter, q 1 = +2.0 X 10-6 coul, and q2 = +1.0 X 10-6 coul. Compare
with Fig. 29-18.
33. Let the potential difference between the shell of an electrostatic generator and the
point at which charges are sprayed onto the moving belt be 3.0 X 106 volts. If the belt
transfers charge to the shell at the rate of 3.0 X 10- 3 coul/sec, what power must be pro-
vided to drive the belt, considering only electrical forces?
34. (a) How much charge is required to raise an isolated metallic sphere of LO-meter
radius to a potential of 1.0 X 10 6 volts? Repeat for a sphere of 1.0-cm radius. (b) Why
use a large sphere in an electrostatic generator since the same potential can be achieved
for a smaller charge with a small sphere?
35. An alpha particle is accelerated through a potential difference of one million volts
in an electrostatic generator. (a) What kinetic energy does it acquire? (b) What kinetic
energy would a proton acquire under these same circumstances? (c) Which particle would
acquire the greater speed, starting from rest?
Capacitors and Dielectrics
CHAPTER 30

30-1 Capacitance
In Section 29-3 we showed that the potential of a charged conducting
sphere, assumed to be completely isolated with no other bodies (conducting
or nonconducting) nearby, is given by
1 q
v +' = ---, (30-1)
41reo R
in which q is the charge on the sphere and R is the sphere radius. The sub-
script on Vindicates that we assume the charge to be positive. We represent
this potential in Fig. 30-1 by the line marked V +'· The line marked V 00
in that figure represents the potential of an infinitely distant reference posi-
tion; it has been assigned the value zero, following the usual convention.

V'
+

Potential difference -- - - ~ - - - - ] Potential difference


(= V') when
conductors
are far apart
I
V°"(=0)__,?1'
---- -
(=V)when
conductors
are nearby

-----
V_'

Fig. 30-1 The potential difference between two conductors that carry constant, equal,
and opposite charges is reduced as the conductors are brought closer together.
741
742 CAPACITORS AND DIELECTRICS Chap. 30
Let us now imagine a second sphere of radius R, carrying a negative charge
-q and located a large distance (>>R) from the first sphere so that each may
still be considered to be electrically isolated. The potential of the second
sphere is given by
1 q
V_' = - ----, (30-2)
4'71"Eo R
and this quantity also is represented in Fig. 30- 1.
The potential difference V' between the two spheres is
1 2q
V' = V +' - V _, = - - - ·
4noR
This shows that V', the potential difference, and q, the magnitude of the
charge on either sphere, are proportional to each other. We may rewrite the
equation as
q = (2'71"E 0R) V' = C'V', (30-3)
in which the proportionality constant in parentheses is called the capacitance
of the two spheres and assigned the symbol C'.
Let us move the two spheres close together. The presence of each will now
spoil the spherical symmetry of the lines of force emanating from the other.
Lines from a given sphere, which, for large sphere separations, radiated uni-
formly in all directions to infinity, now terminate, in part, on the other
sphere. Under these conditions Eqs. 30-1 and 30-2 no longer apply, since
they were derived (see Section 29-3) on the assumption that spherical
symmetry existed, permitting the useful application of Gauss's law.
A positive charge brought near to an isolated object serves to raise the
potential of that object and a negative charge serves to lower it, as the
student can see by considering the work required to move a positive test
charge from infinity to points near such charges. Thus the potential of the
positively charged sphere will be lowered by the presence nearby of the nega-
tively charged sphere, from V +' to some lower value V + Similarly the
potential of the negative sphere will be raised from V _, to a higher value V -·
These new potentials are shown in Fig. 30-1, the potential changes for each
sphere being suggested by the vertical arrows.
From Fig. 30-1, although the charges on the spheres have not changed, it is
clear that the potential difference between the spheres has been considerably
reduced. Put another way, the capacitance of the system of two spheres (see
Eq. 30-3), defined from
q
c =-, (30-4)
v
has been made considerably larger than its initial value C' by bringing the
spheres closer together.
It is also possible to use Eq. 30-4 to define the capacitance of a single iso-
lated conductor such as a sphere. In such cases one may imagine that the
Sec. 30-1 CAPACITANCE 743
second "plate," carrying an equal and opposite charge, is a conducting sphere
of very large-essentially infinite-radius centered about the conductor.
The potential of this infinitely distant sphere, according to the usual conven-
tion for potential measurements, is zero. The capacitance of an isolated
sphere of radius R is given from Eqs. 30-4 and 30-1 as
q
C = - = 41r1: 0R.
v
The mks unit of capacitance that follows from Eq. 30-4 is the coul/volt.
A special unit, the farad, is used to represent it. It is named in honor of
Michael Faraday who, among other contributions, developed the concept of
capacitance. Thus
1 farad = 1 coul/volt.
The submultiples of the farad, the microfarad (1 µf = 10-5 farad) and the
micromicrofarad (1 µµf = 10- 12 farac!), are more convenient units in practice.
An analogy can be made between a capacitor carrying a charge q and a rigid con-
tainer of volume 'U containing µ moles of an ideal gas.
The gas pressure pis directly proportional toµ for a fixed temperature, according
to the ideal gas law ~Eq. 23-2)
µ = (:r) p.
For the capacitor (Eq. 30-4)
q = (C)V.
Comparison shows that the capacitance C of the capacitor, assuming a fixed tem-
perature, is analogous to the volume 'U of the container.
Note that any amount of charge can be put on the capacitor, and any mass of gas
can be put in the container, up to certain limits. These correspond to electrical
breakdown ("arcing over") for the capacitor and to rupture of the walls for the con-
tainer.
Figure 30-2 shows a more general case of two nearby conductors, which are
now permitted to be of any shape, carrying equal and opposite charges. Such
an arrangement is called a capacitor, the conductors being called plates. The
equal and opposite charges might be established by connecting the plates
momentarily to opposite poles of a battery. The capacitance C of any
capacitor is defined from Eq. 30-4 in which we remind the student that Vis
the potential difference between the plates and q is the magnitude of the charge
on either plate; q must not be taken as the net charge of the capacitor, which is
zero. The capacitance of a capacitor depends on the geometry of each plate,
their spatial relationship to each other, and the medium in which the plates
are immersed. For the present, we take this medium to be a vacuum.
Capacitors are very useful devices, of great interest to physicists and
engineers. For example:
1. In this book we stress the importance of fields to the understanding of
natural phenomena. A capacitor can be used to establish desired electric
field configurations for various purposes. In Section 27-5 we described the
deflection of an electron beam in a uniform field set up by a capacitor, al-
744 CAPACITORS AND DIELECTRICS Chap. 30

Fig. 30-2 Two insulated conductors carrying equal and opposite charges form a capacitor.

though we did not use this term in that section. In later sections we discuss
the behavior of dielectric materials when placed in an electric field (provided
conveniently by a capacitor) and we shall see how the laws of electromagne-
tism can be generalized to take the presence of dielectric bodies more readily
into account.
2. A second important concept stressed in this book is energy. By ana-
lyzing a charged capacitor we show that electric energy may be considered to
be stored in the electric field between the plates and indeed in any electric
field, however generated. Because capacitors can confine strong electric
fields to small volumes, they can serve as useful devices for storing energy.
In many electron synchrotrons, which are cyclotron-like devices for accele-
rating electrons, energy accumulated and stored in a large bank of capacitors
over a relatively long period of time is made available intermittently to
accelerate the electrons by discharging the capacitor in a much shorter time.
Many researches and devices in plasma physics also make use of bursts of
energy stored in this way.
3. The electronic age could not exist without capacitors. They are used, in
conjunction with other devices, to reduce voltage fluctuations in electronic
povver supplies, to transmit pulsed signals, to generate or detect electro-
magnetic oscillations at radio frequencies, and to provide time delays. In
most of these applications the potential difference between the plates will not
be constant, as we assume in this chapter, but will vary with time, often in a
sinusoidal or a pulsed fashion. In later chapters we consider some aspects
of the capacitor used as a circuit element.
Sec. 30-2 CALCULATING CAPACITANCE 745

30-2 Calculating Capacitance


Figure 30-3 shows a parallel-plate capacitor formed of two parallel con-
ducting plates of area A separated by a distanced. If we connect each plate
to the terminal of a battery, a charge +q will appear on one plate and a
charge -q on the other. If dis small compared with the plate dimensions,
the electric field strength E between the plates will be uniform, which means
that the lines of force will be parallel and evenly spaced. The laws of electro-
magnetism (see Problem 20, Chapter 35) require that there be some "fring-
ing" of the lines at the edges of the plates, but for small enough d it can be
neglected for our present purpose.
We can calculate the capacitance of this device using Gauss's law. Figure
30-3 shows (dashed lines) a Gaussian surface of height h closed by plane caps
of area A that are the shape and size of the capacitor plates. The flux of Eis
zero for the part of the Gaussian surface that lies inside the top capacitor
plate because the electric field inside a conductor carrying a static charge is
zero. The flux of E through the wall of the Gaussian surface is zero because,
to the extent that the fringing of the lines of force can be neglected, E lies in
the wall.
This leaves only the face of the Gaussian surface that lies between the
plates. Here Eis constant and the flux <f>E is simply EA. Gauss's law gives
(30-5)
The work required to carry a test charge q0 from one plate to the other can
be expressed either as q0 V (see Eq. 29-1) or as the product of a force q0 E
times a distanced or q0 Ed. These expressions must be equal, or
V = Ed. (30-6)
More formally, Eq. 30-6 is a special case of the general relation (Eq. 29-5;
see also Example 1, Chapter 29)
V = - f E·dl,

where Vis the difference in potential between the plates. The integral may

Gaussian
surface

Fig. 30-3 A parallel-plate capacitor with plates of area A. The dashed line represents a
Gaussian surface whose height is hand whose top and bottom caps are the same shape and
size as the capacitor plates.
746 CAP A CITO RS AND DIELECTRICS Chap. 30
be taken over any path that starts on one plate and ends on the other because
each platf is an equipotential surface and the electrostatic force is path
independeht. Although the simplest path between the plates is a perpen-
dicular straight line, Eq. 30-6 follows no matter what path of integration
we choose.
If we substitute Eqs. 30-5 and 30-6 into the relation C = q/V, we obtain
q eoEA eoA
C = - = - = -· (30-7)
V Ed d
Equation 30-7 holds only for capacitors of the parallel-plate type; different
formulas hold for capacitors of different geometry.
In Section 26-4 we stated that e0 , which we first met in connection with
Coulomb's law, was not measured in terms of that law because of experi-
mental difficulties. Equation 30-7 suggests that e0 might be measured by
building a capacitor of accurately known plate area and plate spacing and
determining its capacitance experimentally by measuring q and V in the
relation C = q/V. Thus Eq. 30-7 can be solved for e0 and a numerical value
found in terms of the measured quantities A, d, and C; e0 has been measured
accurately in this way .
.... Example 1. The parallel plates of an air-filled capacitor are everywhere 1.0 mm
apart. What must the plate area be if the capacitance is to be 1.0 farad?
From Eq. 30-7

A _ ~(_}_ _ (1.08 9X 10-


3 meter)(l.O farad) _
11 108 2
- eo - . X 10-12 12;
cou nt-m 2 - . X meter .

This is the area of a square sheet more than 6 miles on edge; the farad is indeed a large
unit.
Example 2. A cylindrical capacitor. A cylindrical capacitor consists of two
coaxial cylinders (Fig. 30-4) of radius a and band length Z. What is the capacitance
of this device? Assume that the capacitor is very long (that is, that l » b) so that
fringing of the lines of force at the ends can be ignored for the purpose of calculating
the capacitance.

Fig, 30-4 Example 2. A cross section


of a cylindrical capacitor. The dashed
circle is a cross section of a cylindrical
Gaussian surface of radius rand length l.
Sec. 30-2 CALCULATING CAPACITANCE 747
As a Gaussian surface construct a coaxial cylinder of radius r and length l, closed
by plane caps. Gauss's law
EofE·dS = q

gives EoE(21rr)(l) = q,
the flux being entirely through the cylindrical surface and not through the end caps.
Solving for E yields
q
E=--·
21rEorl

The potential difference between the plates is given by Eq. 29-5 [note that E and
dl ( = dr) point in opposite directions] or

V = -fbE·dl =fbE dr =fb_q_ dr = _q_ ln ~-


a a a 21rEol r 21rEol a
Finally, the capacitance is given by

C = !l = 2nol
V ln (b/a)
Like the relation for the parallel-plate capacitor (Eq. 30-7), this relation also depends
only on geometrical factors.
Example 3. Capacitors in parallel. Figure 30-5 shows three capacitors connected
in parallel. What single capacitance C is equivalent to this combination? "Equiva-
lent" means that if the parallel combination and the single capacitance were each
in a box with wires a and b connected to terminals, it would not be possible to dis-
tinguish the two by electrical measurements external to the box.

Fig, 30-5 Example 3. Three


a 1
v
1ql I q2 q3

capacitors in parallel.
b 1 IC1 I
C2 Cs

The potential difference across each capacitor in Fig. 30-5 will be the same. This
follows because all of the upper plates are connected together and to terminal a,
whereas all of the lower plates are connected together and to terminal b. Applying
the relation q = CV to each capacitor yields
and
The total charge q on the combination is
q = q1 + qz + q3
= (C1 + C2 + C3)V.
The equivalent capacitance C is

This result can easily be extended to any number of parallel-connected capacitors.


J
748 CAPACITORS AND 'DIELECTRICS Chap. 30
Example 4. Capacitors in series. Figure 30-6 shows three capacitors connected in
series. What single capacitance C is "equivalent" (see Example 3) to this combina-
tion?
For capacitors connected as shown, the magnitude q of the charge on each plate
must be the same. This is true because the net charge on the part of the circuit en-
closed by the dashed line in Fig. 30-6 must be zero; that is, the charge present on
these plates initially is zero aild connecting a battery between a and b will only
produce a charge separation, the net charge on these plates still being zero. Assuming
that neither C1 nor C2 "sparks over," there is no way for charge to enter or leave the
region enclosed by the dashed line.

V1 V2 V3

i~----1~
IC1 C2 I
Fig. 30-6 Example 4. Three capacitors in
-q ! +q -q ! +q -q +q senes.
L ____ _J

~-----v-----.a
a b

Applying the relation q = CV to each capacitor yields


and V3 = q/C3.
The potential difference for the series combination is

The equivalent capacitance is


q 1
C=v=1 1 1
C1 + C2 + C3
1 I 1 1
or C= C1 + C2 + C3.
The equivalent series capacitance is always less than the smallest capacitance in the
chain. -<1111

30-3 Parallel-Plate Capacitor with Dielectric


Equation 30-7 holds only for a parallel-plate capacitor with its plates in a
vacuum. Michael Faraday, in 1837, first investigated the effect of filling the
space between the plates with a dielectric, say mica or oil. In Faraday's
words:
The question may be stated thus: suppose A an electrified plate of metal sus-
pended in air, and B and C two exactly similar plates, placed parallel to and on
each side of A at equal distances and insulated; A will then induce equally toward
Band C [that is, equal charges will appear on these plates]. If in this position of
the plates some other dielectric than air, as shell-lac, be introduced between A and
Sec. 30-3 PARALLEL-PLATE CAPACITOR WITH DIELECTRIC 749

+++++ +++++
;w inmm•"-•nwm1 m
B

(a)

++ ++

(b)

Fig. 30-7 (a) Battery B supplies the same potential difference to each capacitor; the one
on the right has the higher charge. (b) Both capacitors carry the same charge; the one on
the right has the lower potential difference, as indicated by the meter readings.

C, will the induction between them remain the same? Will the relation of C and B
to A be unaltered, notwithstanding the difference of the dielectrics interposed
between them?

Faraday answered this question by constructing two identical capacitors,


in one of which he placed a dielectric, the other containing air at normal
pressure. When both capacitors were charged to the same potential dijJerence,
Faraday found by experiment that the charge on the one containing the dielectric
was greater than that on the other; see Fig. 30-7a.
Faraday measured the relative charges on the plates of the two capacitors by
touching a metal ball (fitted with an insulating handle) to the plates, thus sampling
the charge quantitatively. He then put this ball in a Coulomb torsion balance and
measured the force of Coulomb repulsion on a second (standard) charged ball mounted
on the balance arm.

Since q is larger, for the same V, if a dielectric is present, it follows from the
relation C = q/V that the capacitance of a capacitor increases if a dielectric is
placed between the plates. The ratio of the capacitance with the dielectric * to
that without is called the dielectric constant K of the material; see Table 30-1.
Instead of maintaining the two capacitors at the same potential difference,
we can place the same charge on them, as in Fig. 30-7!:i. Experiment then
shows that the potential difference Va between the plates of the right-hand
capacitor is smaller than that for the left-hand capacitor by the factor 1/ K, or

Va= Vo/K.
We are led once again to conclude, from the relation C = q/V, that the effect
of the dielectric is to increase the capacitance by a factor K.

* Assumed to fill completely the space between the plates.


750 CAP A CITO RS AND DIELECTRICS Chap. 30

Table 30-1
PROPERTIES OF SOME DIELECTRICS

Dielectric
Dielectric Strength *
Material Constant (kv /mm)

Vacuum 1.00000 00

Air 1.00054 0.8


Water 78
Paper 3.5 14
Ruby mica 5.4 160
Amber 2.7 90
Porcelain 6.5 4
Fused quartz 3.8 8
Pyrex glass 4.5 13
Bakelite 4.8 12
Polyethylene 2.3 50
Polystyrene 2.6 25
Teflon 2.1 60
Neoprene 6.9 12
Pyranol oil 4.5 12
Titanium dioxide 100 6

* This isthe maximum potential gradient that may exist in the dielectric without
the occurrence of electrical breakdown. Dielectrics are often placed between con -
ducting plates to permit a higher potential difference to be applied between them
than would be possible with air as the dielectric.

For a parallel-plate capacitor we can write, as an experimental result,


Kt:oA
C=-· (30-8)
d
Equation 30-7 is a special case of this relation found by putting K I, cor-
responding to a vacuum between the plates. Experiment shows that the
capacitance of all types of capacitor is increased by the factor K if the space
between the plates is filled with a dielectric. Thus the capacitance of any
capacitor can be written as
C = KEoL,
where L depends on the geometry and has the dimensions of a length. For a
parallel-plate capacitor (see Eq. 30-7) L is A/d; for a cylindrical capacitor
(see Example 2) it is 21rl/ln (b/a).

3Q-4 Dielectrics-An Atomic View


We now seek to understand, in atomic terms, what happens when a di-
electric is placed in an electric field. There are two possibilities. The mole-
cules of some dielectrics, like water, have permanent electric dipole moments.
Sec. 30-4 DIELECTRICS-AN ATOMIC VIEW 751

In such materials (called polar) the electric dipole moments p tend to align
themselves with an external electric field, as in Fig. 30-8; see also Section
27-6. Because the molecules are in constant thermal agitation, the degree of
alignment will not be complete but will increase as the applied electric field is
increased or as the temperature is decreased.
Whether or not the molecules have permanent electric dipole moments,
they acquire them by induction when placed in an electric field. In Section
29-5 we saw that the external electric field tends to separate the negative and
the positive charge in the atom or molecule. This induced electric dipole
moment is present only when the electric field is present. It is proportional
to the electric field (for normal field strengths) and is created already lined up
with the electric field as Fig. 29-12 suggests.
Let us use a parallel-plate capacitor, carrying a fixed charge q and not
connected to a battery (see Fig. 30-7b ), to provide a uniform external electric
field E 0 into which we place a dielectric slab. The over-all effect of alignment
and induction is to separate the center of positive charge of the entire slab
slightly from the center of negative charge. The slab, as a whole, although
remaining electrically neutral, becomes polarized, as Fig. 30-9b suggests.
The net effect is a pile-up of positive charge on the right face of the slab and of
negative charge on the left face; within the slab no excess charge appears in
any given volume element. Since the slab as a whole remains neutral, the
positive induced surface charge must be equal in magnitude to the negative
induced surface charge. Note that in this process electrons in the dielectric

-
-------\.:~----:;=;:::-----r-,-;-.>n----~

@ ~
~@ e---:
@
G,3
G{J)
(j)
(a) (b)

Fig. 30-8 (a) :VIolecules with a permanent electric dipole moment, showing their random
orientation in the absence of an external electric field. (b) An electric field is applied,
producing partial alignment of the dipoles. Thermal agitation prevents complete align-
ment.
752 CAPACITORS AND DIELECTRICS Chap. 30

E 0 =0 i-------N
E0 I

(a) (c)

Fig. 30-9 (a) A dielectric slab, showing the random distribution of plus and minus
charges. (b) An external field Eo, established by putting the slab between the plates of a
parallel-plate capacitor (not shown), separates the center of plus charge in the slab slightly
from the center of minus charge, resulting in the appearance of surface charges. No net
charge exists in any volume element located in the interior of the slab. (c) The surface
charges set up a field E' which opposes the external field Eo associated with the charges on
the capacitor plates. The resultant field E ( = Eo + E') in the dielectric is thus less than
Eo.

are displaced from their equilibrium positions by distances that are con-
siderably less than an atomic diameter. There is no transfer of charge over
macroscopic distances such as occurs when a current is set up in a conductor.
Figure 30-9c shows that the induced surface charges will always appear
in such a way that the electric field set up by them (E') opposes the external
electric field E 0 . The resultant field in the dielectric E is the vector sum of
E 0 and E'. It points in the same direction as E 0 but is smaller. If a d1'electric
is placed in an electric field, induced surface charges appear which tend to
weaken the original field within the dielectric.
This weakening of the electric field reveals itself in Fig. 30-7b as a reduc-
tion in potential difference between the plates of a charged isolated capacitor
when a dielectric is introduced between the plates. The relation V = Ed
for a parallel-plate capacitor (see Eq. 30-6) holds whether or not dielectric is
present and shows that the reduction in V described in Fig. 30-7b is directly
connected to the reduction in E described in Fig. 30-9. More specifically, if a
dielectric slab is introduced into a charged parallel-plate capacitor, then
Eo Vo
- K (30-9)
E Va
where the symbols on the left refer to Fig. 30-9 and the symbols VO and Va
refer to Fig. 30-7b. *
Induced surface charge is the explanation of the most elementary fact of
static electricity, namely, that a charged rod will attract uncharged bits of
* Equation 30-9 does not hold if the battery remains connected while the dielectric
slab is introduced. In this case V (hence E) could not change. Instead, the charge q on
the capacitor plates would increase by a factor K, as Fig. 30-7 a suggests.
Sec. 30-5 DIELECTRICS AND GAUSS'S LAW 753

Fig. 30-10 A charged rod attracts an uncharged piece of paper because unbalanced
forces act on the induced surface charges.

paper, etc. Figure 30-10 shows a bit of paper in the field of a charged rod.
Surface charges appear on the paper as shown. The negatively charged end
of the paper will be pulled toward the rod and the positively charged end will
be repelled. These two forces do not have the same magnitude because the
negative end, being closer to. the rod, is in a stronger field and experiences a
stronger force. The net effect is an attraction. A dielectric body in a uni-
form electric field will not experience a net force.

30-5 Dielectrics and Gauss's Law


So far our use of Gauss's law has been confined to situations in which no di-
electric was present. Now let us apply this law to a parallel-plate capacitor
filled with a dielectric of dielectric constant K.
Figure 30-11 shows the capacitor both with and without the dielectric. It
is assumed that the charge q on the plates is the same in each case. Gaussian
surfaces have been drawn after the fashion of Fig. 30-3.

-r-q
(a)

(b)

Fig. 30-11 A parallel-plate capacitor (a) without and (b) with a dielectric. The charge q
on the plates is assumed to be the same in each case.
754 CAP A CITO RS AND DIELECTRICS Chap. 30
If no dielectric is present (Fig. 30-lla), Gauss's law (see Eq. 30-5) gives

t:ofE·dS = t:oEoA = q
q
or Eo=-· (30-10)
t:oA
If the dielectric is present (Fig. 30-llb), Gauss's law gives

t:ofE·dS = t:0 EA = q - q'


q q'
or E=---, (30-11)
t:oA t:oA
in which -q', the induced surface charge, must be distinguished from q, the so-
called free charge on the plates. These two charges, both of which lie within
the Gaussian surface, are opposite in sign; q - q' is the net charge within the
Gaussian surface.
Equation 30-9 shows that in Fig. 30-11

Eo
E=-·
K

Combining this with Eq. 39-10, we have

Eo q
E=-=-· (30-12)
K Kt:oA
Inserting this in Eq. 30-11 yields

q q'
(30-13a)

or (30-13b)

This shows correctly that the induced surface charge q' is always less in
magnitude than the free charge q and is equal to zero if no dielectric is present,
that is, if K = 1.
Now we write Gauss's law for the case of Fig. 30-llb in the form

(30-14)

q - q' again being the net charge within the Gaussian surface. Substituting
from Eq. 30-13b for q' leads, after some rearrangement, to

(30-15)
Sec. 30-5 DIELECTRICS AND GAUSS'S LAW 755

This important relation, although derived for a parallel-plate capacitor, is


true generally and is the form in which Gauss's law is usually written when
dielectrics are present. Note the following:
1. The flux integral now contains a factor K.
2. The charge q contained within the Gaussian surface is taken to be the
free charge only. Induced surface charge is deliberately ignored on the right
side of this equation, having been taken into account by the introduction of K
on the left side. Equations 30-14 and 30-15 are completely equivalent for-
mulations.
..... Example 5. Figure 30-12 shows a dielectric slab of thickness b and dielectric
constant K placed between the plates of a parallel-plate capacitor of plate area A and
separation d. A potential difference VO is applied with no dielectric present. The
battery is then disconnected and the dielectric slab inserted. Assume that A = 100
cm2, d = 1.0 cm, b = 0.50 cm, K = 7.0, and Vo = 100 volts and (a) calculate the
capaciiance Co before the slab is inserted.

Fig, 30-12 Example 5. A parallel-plate capacitor containing a dielectric slab.

From Eq. 30-7, Co is found:

C0 _ :_()_A _ (8.9 X 10- coul2/nt-m )(10- meter ) _ 8 9 f


12 2 2 2

- d - 10-2 meter - · µµ ·

(b) Calculate the free charge q.


From Eq. 30-4,
q = Co Vo = (8.9 X 10- 12 farad)(lOO volts) = 8.9 X 10-10 coul.
Because of the technique used to charge the capacitor, the free charge remains un-
changed as the slab is introduced. If the charging battery had not been disconnected,
this would not be the case.
(c) Calculate the electric field strength in the gap.
Applying Gauss's law in the form given in Eq. 30-15 to the Gaussian surface of
Fig. 30-12 (upper plate) yields

eof KE·dS = e E A = q,
0 0

or E 0 = _Cf___ = 8 ·9 X io-io coul 1.0 X 10 4 volts/meter.


eoA (8.9 X 10- 12 coul2/nt-m 2 )(10- 2 meter 2 )
Note that we put K = 1 here because the surface over which we evaluate flux integral
does not pass through any dielectric. Note too that E 0 remains unchanged when the
756 CAP A CITO RS AND DIELECTRICS Chap. 30
slab is introduced; this derivation takes no specific account of the presence of the
dielectric.
(d) Calculate the electric field strength in the dielectric.
Applying Eq. 30-15 to the Gaussian surface of Fig. 30-12 (lower plate) yields

fof KE·dS = foKEA = q.


Note that K appears here because the surface cuts through the dielectric and that only
the free charge q appears on the right. Thus we have

E = __51__ = Eo = 1.0 X 104 volts/meter = 0.14 X 104 volts/meter.


KfoA K 7.0
(e) Calculate the potential difference between the plates.
Applying Eq. 29-5 to a straight perpendicular path from the lower plate (L) to
the upper one ( U) yields

V = -
u
L
E·dl = -
u
L
E cos 180° dl =
u
L
E dl = E 0(d - b) + Eb.

Numerically
V = (1.0 X 104 volts/meter)(5 X 10- 3 meter)
+ (0.14 X 104 volts/meter)(5 X 10- 3 meter) = 57 volts.
This contrasts with the original applied potential difference of 100 volts; compare Fig.
30-7b.
(f) Calculate the capacitance with the slab in place.
From Eq. 30-4,
C = !!___ = 8.9 X 10- 10 coul = 16 µµf.
V 57 volts
When the dielectric slab is introduced, the potential difference drops from 100 to 57
volts and the capacitance rises from 8.9 to 16 µµf, a factor of 1.8. If the dielectric
slab had filled the capacitor, the capacitance would have risen by a factor of K ( = 7.0)
to 62 µµf. _,..

3o-6 Three Electric Vectors


For all situations that we encounter in this book our discussion of the behavior of
dielectrics in an electric field is adequate. However, the problems that we treat are
simple ones, such as that of a rectangular slab placed at right angles to a uniform ex-
ternal electric field. For more difficult problems, such as that of finding E at the
center of a dielectric ellipsoid placed in a (possibly nonuniform) external electric
field, it greatly simplifies the labor and leads to deeper insight if we introduce a new
formalism. We do so largely so that students who take a second course in electro-
magnetism will have some familiarity with the concepts.
Let us rewrite Eq. 30-13a, which applies to a parallel-plate capacitor containing a
dielectric, as
!!___
A
= fo (-L)
KfoA
+ q' .
A
(30-16)

The quantity in parentheses (see Eq. 30-12) is simply the electric field strength E in
the dielectric. The last term in Eq. 30-16 is the induced surface charge per unit area.
We call it the electric polarization P, or
q'
p = -· (30-17)
A
Sec. 30-6 THREE ELECTRIC VECTORS 757

(P= 0)

-q'

+q'

-q
(a)

'-----r-----1
D
(b)

Fig. 30-13 (a) Showing D, EOE, and P in the dielectric (upper right) and in the gap
(upper left) for a parallel-plate capacitor. (b) Showing samples of the lines associated with
D (free charge), EOE (all charges), and P (polarization charge).

The name is suitable because the induced surface charge q' appears when the dielectric
is polarized.
The electric polarization P can be defined in an equivalent way by multiplying the
numerator and denominator in Eq. 30-17 by d, the thickness of the dielectric slab in
Fig. 30-11,
q'd
P=-· (30-18)
Ad

The numerator is the product q'd of the magnitude of the (equal and opposite)
polarization charges by their separation. It is thus the induced electric dipole mo-
ment of the dielectric slab. Since the denominator Ad is the volume of the slab, we
see that the electric polarization can also be defined as the induced electric dipole
moment per unit volume in the dielectric. This definition suggests that since the
electric dipole moment is a vector the electric polarization is also a vector, its magni-
tude being P. The direction of Pis from the negative induced charge to the positive
induced charge, as for any dipole. In Fig. 30-13, which shows a capacitor with a
dielectric slab filling half the space between the plates, P points down.
We can now rewrite Eq. 30-16 as

q
A= EoE + P. (30-19)
758 CAPACITORS AND DIELECTRICS Chap. 30
The quantity on the right occurs so often in electrostatic problems that we give it the
special name electric displacement D, or
D =EoE + P
in which D = !L.
A
The name has historical significance only.
Since E and Pare vectors, D must also be one, so that in the more general case we
have
D = EoE + P. (30-21)
In Fig. 30-13 all three vectors point down and each has a constant magnitude for
every point in the dielectric (and also at every point in the air gap) so that the vector
nature of Eq. 30-21 is not very important in this case. In more complicated prob-
lems, however, E, P, and D may vary in magnitude and direction from point to point.
From their definitions we see the following:
1. D (see Eq. 30-20b) is connected with the free charge only. We can represent the
vector field of D by lines of D, just as we represent the field of E by lines of force.
Figure 30-13b shows that the lines of D begin and end on the free charges.
2. P (see Eq. 30-17) is connected with the polarization charge only. It is also pos-
sible to represent this vector field by lines. Figure 30-13b shows that the lines of
P begin and end on the polarization (.;!larges.
3. Eis connected with all charges that are actually present, whether free or polari-
zation. The lines of E reflect the presence of both kinds of charge, as Fig. 30-13b
shows. Note (Eqs. 30-17 and 30-20b) that the units for P and D (coul/meter 2 ) differ
from those of E (nt/coul).
The electric field vector E, which is what determines the force that acts on a suita-
bly placed test charge, remains of fundamental interest. D and P are auxiliary vec-
tors useful as aids in the solution of problems more complex than that of Fig. 30-13.
The vectors D and P can both be expressed in terms of E alone. A convenient
starting point is the identity

Comparison with Eqs. 30-12 and 30-20b shows that this, extended to vector form,
can be written as
D = KEoE. (30-22)
We can also write the polarization (see Eqs. 30-17 and 30--13b) as

P=~=!(1-D-
Since q/ A = D, we can rewrite this, using Eq. 30-22 and casting the result into vector
form, as
p = Eo(K - l)E. (30-23)
This shows clearly that in a vacuum (K = 1) the polarization vector P is zero.*

* Certain waxes, when polarized in their molten state, retain a permanent polarization
after solidifying, even though the external polarizing field is removed. Electrets, manu-
factured in this way, are the electrostatic analog of permanent magnets in that they
possess a gross permanent electric dipole moment. Materials from which electrets can be
constructed are called ferroelectric. Electrets do not obey Eq. 30~23 because they have a
nonvanishing value of P even though E = 0.
Sec. 30-6 THREE ELECTRIC VECTORS 759
Equations 30-22 and 30-23 show that for isotropic materials, to which a single
dielectric constant K can be assigned, D and P both point in the direction of E at any
given point.
The definition of D given by Eq. 30-22 allows us to write Eq. 30-15, that is, Gauss's
law in the presence of a dielectric, simply as

fD·dS = q, (30-24)

where, as before, q represents the free charge only, the induced surface charges being
excluded .
.... Example 6. In Figure 30-13, using data from Example 5, calculate E, D, and
P: (a) in the dielectric and (b) in the air gap.
(a) The electric field in the dielectric is calculated in Example 5 to be 1.43 X 10 3
volts/meter. From Eq. 30-22,

D = K(;oE

= (7.0)(8.9 X 10-12 coul2/nt-m 2 )(1.43 X 10 3 volts/meter)


= 8.9 X 10-s coul/meter 2

and, from Eq. 30-23,

P = (;o(K - l)E

= (8.9 X 10-12 coul2 /nt-m 2)(7 .0 - 1)(1.43 X 103 volts/meter)


= 7.5 X 10-s coul/meter2 •

(b) The electric field E 0 in the air gap is calculated in Example 5 to be 1.00 X 10 4
volts/meter. From Eq. 30-22,

Do= KEoEo

= (1)(8.9 X 10- 12 coul2/nt-m 2)(1.00 X 104 volts/meter)

= 8.9 X 10-s coul/meter 2

and, from Eq. 30-23, recalling that K = 1 in the air gap,


Po = eo(K - l)Eo = 0.

Note that P vanishes outside the dielectric, D has the same value in the dielectric
and in the gap, and E has different values in the dielectric and in the gap. The student
should verify that Eq. 30-21 (D = e0E + P) is correct both in the gap and in the di-
electric.
It can be shown from Maxwell's equations that no matter how complex the prob-
lem the component of D normal to the surface of the dielectric has the same value on
each side of the surface. In this problem D itself is normal to the surface, there being
no component but the normal one. It can also be shown that the component of E
tangential to the dielectric surface has the same value on each side of the surface.
This boundary condition, like the one for D, is trivial in this problem, both tangential
components being zero. In more complex problems these boundary conditions on D
and E are very important. Table 30-2 summarizes the properties of the electric
vectors E, D, and P. <1111
760 CAP A CITO RS AND DIELECTRICS Chap. 30

Table 30-2

THREE ELECTRIC VECTORS

Associated Boundary
Name Symbol
with Condition

Electric field strength E All charges Tangential component


continuous

Electric displacement D Free charges only Normal component


continuous

Polarization p Polarization Vanishes in a


(electric dipole moment charges only vacuum
per unit volume)

Defining equation for E F = qE Eq. 27-2

General relation among the three vectors D = E:oE +P Eq. 30-21

Gauss's law when dielectric media are present fD·dS = q Eq. 30--24
(q = free charge only)

Empirical relations for certain dielectric D = KEoE Eq. 30-22


materials* P = (K - l)E:oE Eq. 30-23

* Generally true, with K independent of E, except for certain materials called


ferroelectrics; see footnote on page 758.

30-7 Energy Storage in an Electric Field


In Section 29-6 we saw that all charge configurations have a certain
electric potential energy U, equal to the work W (which may be positive or
negative) that must be done to assemble them from their individual com-
ponents, originally assumed to be infinitely far apart and at rest. This po-
tential energy reminds us of the potential energy stored in a compressed
spring or the gravitational potential energy stored in, say, the earth-moon
system.
For a simple example, work must be done to separate two equal and
opposite charges. This energy is stored in the system and can be recovered if
the charges are allowed to come together again. Similarly, a charged
capacitor has stored in it an electrical potential energy U equal to the work W
required to charge it. This energy can be recovered if the capacitor is
allowed to discharge. We can visualize the work of charging by imagining
Sec. 30-7 ENERGY STORAGE IN AN ELECTRIC FIELD 761

that an external agent pulls electrons from the positive plate and pushes
them onto the negative plate, thus bringing about the charge separation;
normally the work of charging is done by a battery, at the expense of its store
of chemical energy.
Suppose that at a time ta charge q'(t) has been transferred from one plate
to the other. The potential difference V(t) between the plates at that
moment will be q'(t)/C. If an extra increment of charge dq' is transferred,
the small amount of additional work needed will be

dW = V dq = ( ~) dq'.

If this process is continued until a total charge q has been transferred, the
total work will be found from

W = I dW = i q q'
o C dq' =
1 q2
2C . (30-25)

From the relation q = CV we can also write this as

w (= U) = tcv 2 . (30-26)

It is reasonable to suppose that the energy stored in a capacitor resides in the


electric field. As q or V in Eqs. 30-25 and 30-26 increase, for example, so
does the electric field E; when q and V are zero, so is E.
In a parallel-plate capacitor, neglecting fringing, the electric field has the
same value for all points between the plates. Thus the energy density u,
which is the stored energy per unit volume, should also be uniform; u (see
Eq. 30-26) is given by
u
u=-=--,
Ad Ad

where Ad is the volume between the plates. Substituting the relation C


= KE: 0 A/d (Eq. 30-8) leads to

However, V /dis the electric field strength E, so that

(30-27)

Although this equation was derived for the special case of a parallel-plate
capacitor, it is true in general. If an electric field E exists at any point in
space, we can think of that point as the site of stored energy in amount, per unit
volume, of !KE: 0 E 2 .
762 CAPACITORS AND DIELECTRICS Chap. 30
._ Example 7. A capacitor C1 is charged to a potential difference Vo. This charging
battery is then removed and the capacitor is connected as in Fig. 30-14 to an un-
charged capacitor C2•
(a) What is the final potential difference V across the combination?

.----oz___~ Fig. 30-14 Example 7. C1 is charged and then


connected to C2 by closing switch S.

Cz

The original charge q0 is now shared by the two capacitors. Thus


qo = q1 + q2.
Applying the relation q = CV to each of these terms yields
C1Vo = C1V + C2V
C1
or V = Vo C1 + C2
This suggests a way to measure an unknown capacitance (G 2 , say) in terms of a known
one.
(b) What is the stored energy before and after the switch in Fig. 30-14 is thrown?
The initial stored energy is

The final stored energy is


U -_ 1
2
c1V 2 + 21 c2V 2 = 1
2( 1
C + C2) ( Ci+
V oC 1C )
2

Thus U is less than U0 ! The "missing" energy appears as heat in the connecting
wires as the charges move through them.
Example 8. A parallel-plate capacitor has plates with area A and separation d.
A battery charges the plates to a potential difference Vo. The battery is then dis-
connected, and a dielectric slab of thickness d is introduced. Calculate the stored
energy both before and after the slab is introduced and account for any difference.
The energy U o before introducing the slab is
Uo = !CoVo2•
After the slab is in place, we have

C = KCo and V= Vo/K

and thus U = i
2CV
2
= i
2K o
C (Vo)2
----; = ;1 U O·

The energy after the slab is introduced is less by a factor 1/K. The "missing" energy
would be apparent to the person who inserted the slab. He would feel a "tug" on the
slab and would have to restrain it if he wished to insert the slab without acceleration.
This means that he would have to do negative work on it, or, alternatively, that the
condenser + slab system would do positive work on him. This positive work is

w = uo - u = tc ov 02 ( 1- ~ )-
Sec. 30-7 ENERGY STORAGE IN AN ELECTRIC FIELD 763

Figure 30-15 shows how the forces that do this work arise, in terms of attraction be-
tween the free charge on the plates and the induced surface charges that appear on
the slab when it is introduced between the plates.

Fig. 30-15 When we introduce a dielectric slab into a charged capacitor, as shown, forces
arise which tend to pull the slab into the capacitor.

Example 9. A conducting sphere of radius R, in a vacuum, carries a charge q.


(a) Compute the total electrostatic energy stored in the surrounding space.
At any radius r from the center of the sphere (assuming r > R) E is given by
1 q
E = - - 2·
47T'E:o r

The energy density at any radius r is found from Eq. 30-27, assuming K = 1, or

The energy dU that lies in a spherical shell between the radii r and r + dr is
q2 dr
dU = (47rr 2 )(dr)u = -8 - 2 ,
7T'E:o r
where (47rr2)(dr) is the volume of the spherical shell. The total energy U is found by
integration, or
q2 J,oo dr
f
U = d U = 8no R r 2 = 87rE:oR.
q2

Note that this relation follows at once from Eq. 30-25 (U = q2/2C), where C (seep.
651) is the capacitance ( = 47T'E=oR) of an isolated sphere of radius R.
(b) What is the radius Ro of a spherical surface such that half the stored energy lies
within it?
In the equation just given we put

!.[! -
q2
-~
J,Ro dr-
2 ' - 871'€0 R r2

q2 q2 ( 1 1)
or
167T'E:oR = 87T'E:o R - R~ '
which yields, after some rearrangement,

Ro= 2R.
764 CAP A CITO RS AND DIELECTRICS Chap. 30

QUESTIONS

1. A capacitor is connected across a battery. (a) Why does each plate receive a charge
of exactly the same magnitude? (b) Is this true even if the plates are of different sizes?
2. Can there be a potential difference between two adjacent conductors that carry the
same positive charge?
3. The relation rr ex 1/R suggests that the charge placed on an
Foil isolated conductor concentrates on points and avoids flat surfaces,
---------'--=
where R = oo. How do we reconcile this with Fig. 30-3 in which
the charge is definitely on the flat surface of either plate?
4. A sheet of aluminum foil of negligible thickness is placed be-
tween the plates of a capacitor as in Fig. 30-16. What effect has it
Fig. 30-16 on the capacitance if (a) the foil is electrically insulated and (b) the
foil is connected to the upper plate.
5. Discuss similarities and differences when (a) a dielectric slab and (b) a conducting
slab are inserted between the plates of a parallel-plate capacitor. Assume the slab thick-
nesses to be one-half the plate separation.
6. An oil-filled parallel-plate capacitor has been designed to have a capacitance C and
to operate safely at or below a certain maximum potential difference V m without arcing
over. However, the designer did not do a good job and the capacitor occasionally arcs
over. What can be done to redesign the capacitor, keeping C and Vm unchanged and
using the same dielectric?
7. Would you expect the dielectric constant, for substances containing permanent
molecular electric dipoles, to vary with temperature?
8. What is your estimate of the amount by which the center of positive charge and the
center of negative charge are displaced in a situation like that of Fig. 30-9b? An intuitive
guess is all that is called for.
9. For a given potential difference does a capacitor store more or less charge with a di-
electric than it does without a dielectric (vacuum)? Explain in terms of the microscopic
picture of the Rituation.
10. An isolated conducting sphere is given a positive charge. Does its mass increase,
decrease, or remain the same?
11. A dielectric, slab is inserted in one end of a charged parallel-plate capacitor (the
plates being hori7ontal and the charging battery having been disconnected) and then re-
leased. Describe what happens. Neglect friction.
12. A capacitor is charged by using a battery, which is then disconnected. A dielectric
slab is then slipped between the plates. De-
scribe qualitatively what happens to the charge,
the capacitance, the potential difference, the

I
electric field strength, and the stored energy.
13. While a capacitor remains connected to
a battery, a dielectric slab is slipped between
the plates. Describe qualitatively what hap-
pens to the charge, the capacitance, the
potential difference, the electric field strength,
I
and the stored energy. Is work required to Fig. 30-17
insert the slab?
14. Two identical capacitors are connected as shown in Fig. 30-17. A dielectric slab is
slipped between the plates of one capacitor, the battery remaining connected. Describe
qualitatively what happens to the charge, the capacitance, the potential difference, the
electric field strength, and the stored energy for each capacitor.
15. Show that the dielectric constant of a conductor can be taken to be infinitely great.
Chap. 30 PROBLEMS 765

PROBLEMS
/
l. A potential difference of 300 volts is applied to a 2.0-µf capacitor and an 8.0-µf capaci-
tor connected in series. (a) What are the charge and the potential difference for each
capacitor? (b) The charged capacitors are reconnected with their positive plates together
and their negative plates together, no external voltage being applied. What are the charge
and the potential difference for each? (c) The charged capacitors in (a) are reconnected
with plates of opposite sign together. What are the charge and the potential difference for
each?
2. Calculate the capacitance of the earth, viewed as a spherical conductor of radius 6400
km.
){ A 100-µµf capacitor is charged to a potential difference of 50 volts, the charging
b~ttery then being disconnected. The capacitor is then connected, as in Fig. 30-14, to a
second capacitor. If the measured potential difference drops to 35 volts, what is the
capacitance of this second capacitor?
4. If we solve Eq. 30-7 for eo, we see that its mks units are farads/meter. Show that
these units are equivalent to those obtained earlier for eo, namely coul2 /nt-m 2.
5. Figure 30-18 shows two capacitors in series, the rigid center section of length b being "'
movable vertically. Show that the equivalent capacitance of the series combination is
independent of the position of the center section and is given by
eoA
C=--·
a-b

f i l_j
a

L-,....---
Fig. 30-18 Fig. 30-19

6. In Fig. 30-19 a variable air capacitor of the type used in tuning radios is shown.
Alternate plates are connected together, one group being fixed in position, the other group
being capable of rotation. Consider a pile of n plates of alternate polarity, each having an
area A and separated from adjacent plates by a distanced. Show that this capacitor has a
maximum capacitance of
(n - l)eoA
C= .
d
, 7.'A spherical capacitor consists of two concentric spherical shells of radii a and b,;,;
with b > a. Show that its capacitance is
ab
C = 41reo--·
b- a
8. A capacitor has square plates, each of side a, making an angle of () with each other
as shown in Fig. 30-20. Show that for small () the capacitance is given by

C = eoa 2 (l _a())
d 2d
766 CAP A CITO RS AND DIELECTRICS Chap. 30
(Hint: The capacitor may be divided into differential strips which are effectively in
parallel.)

Fig. 30-20

9. Suppose that the two spherical shells of a spherical capacitor have their radii ap-
proximately equal. Under these conditions the device approximates a parallel plate
capacitor with b - a = d. Show that the formula in Problem 7 does indeed reduce to
Eq. 30-7 in this case.
10. A parallel-plate capacitor has circular plates of 8.0-cm radius and 1.0-mm separation.
Wha,tcharge will appear on the plates if a potential difference of 100 volts is applied'?
/n. In Fig. 30-21 find the equivalent capacitance of the combination. Assume that
Ci = 10 .uf, C2 = 5 µf, Cs = 4 µ,f, and V = 100 volts.
12. In Fig. 30-21 suppose that capacitor Cs breaks down electrically, becoming equiva-
lent to a conducting path. What changes in (a) the charge and (b) the potential dif-
ference occur for capacitor C1 ?

I
1
v
C1
I
C2

I
j C31
Fig. 30-21

13. Find the effective capacitance between points x and y in Fig. 30-22. Assume that
C2 = 10 µ,f and that the other capacitors are all 4.0 µ,f. (Hint: Apply a potential difference
V between x and y and write down all the relationships that involve the charges and poten-
tial differences for the separate capacitors.)

Fig. 30-22
Chap. 30 PROBLEMS 767
14. If you have available several 2.0-µf capacitors, each capable of withstanding 200
volts without breakdown, how would you assemble a combination having an equivalent
capaci~nce of (a) 0.40 µfor of (b) 1.2 µf, each capable of withstanding 1000 volts?
VIn Fig. 30-23 find the equivalent capacitance of the combination. Assume that
Ci = 10 µf, C2 = 5 µf, C3 = 4 µf, and V = 100 volts.

I
1
v
l
l
C3
C2
I
Fig. 30-23

16. In Fig. 30-24 the battery B supplies 12 volts. (a) Find the charge on each capacitor
when switch 8 1 is closed and (b) when switch 8 2 is also closed. Take C1 = 1 µf, C 2 = 2 µf,
C3 = 3 µf, and C4 = 4 µf.

....______ ,1-B-----0

Fig. 30-24

<. 17. A parallel-plate capacitor is filled with two dielectrics as in Fig. 30-25. Show that
the capacitance is given by

Check this formula for all the limiting cases that you can think of. (Hint: Can you justify
regarding this arrangement as two capacitors in parallel?)

Fig. 30-25 Fig. 30-26

18. A parallel-plate capacitor is filled with two dielectrics as in Fig. 30-26. Show that
the capacitance is given by
768 CAPACITORS AND DIELECTRICS Chap. 30
Check this formula for all the limiting cases that you can think of. (Hint: Can you justify
regarging this arrangement as two capacitors in series?)
ft~ A dielectric slab of thickness b is inserted between the plates of a parallel-plate
capacitor of plate separation d. Show that the capacitance is given by
KEOA
C=----~
Kd - b(K - 1)

(Hint: Derive the formula following the pattern of Example 5.) Does this formula predict
the correct numerical result of Example 5? Does the formula seem reasonable for the
special cases of b = 0, K = 1, and b = d?
20. A slab of copper of thickness b is thrust into a parallel-plate capacitor as shown in
Fig. 30-27; it is exactly halfway between the plates. What is the capacitance before and
after the slab is introduced?

Fig. 30-27

21. For making a capacitor you have available two plates of copper, a sheet of mica
(thickness = 0.10 mm, K = 6), a sheet of glass (thickness = 2.0 mm, K = 7), and a slab of
paraffin (thickness = 1.0 cm, K = 2). To obtain the largest capacitance, which sheet
(or sheets) should you place between the copper plates?
22. In Example 5, suppose that the 100-volt battery remains connected during the time
that the dielectric slab is being introduced. Calculate (a) the charge on the capacitor
plates, (b) the electric field in the gap, (c) the electric field in the slab, and (d) the capaci-
tance. For all of these quantities give the numerical values before and after the slab is
introduced. Contrast your results with those of Example 5 by constructing a tabular
listing.
23. A parallel-plate capacitor has a capacitance of 100 µ,µ,f, a plate area of 100 cm 2, and a
mica dielectric. At 50 volts potential difference, calculate (a) E in the mica, (b) the free
charge on the plates, and (c) the induced surface charge.
24. Two parallel plates of area 100 cm 2 are each given equal but opposite charges of
8.9 X 10-7 coul. Within the dielectric material filling the space between the plates the
electric field strength is 1.4 X 106 volts/meter. (a) Find the dielectric constant of the
material. (b} Determine the magitude of the charge induced on each dielectric surface.
25. Hard rubber has a dielectric constant of 2.8 and a dielectric strength of 18 X 106
volts/meter. If it is used as the dielectric material in a parallel-plate capacitor, what
minimum area may the plates of the capacitor have in order that the capacitance be
7.0 X 10-2 µ,f and that the capacitor be able to withstand a potential difference of 4000
volts?
26. A cylindrical capacitor has radii a and b as in Fig. 30-4. Show that half the stored
electric potential energy lies within a cylinder whose radius is

r = Vab.
27. An isolated metal sphere whose diameter is 10 cm has a potential of 8000 volts.
What is the energy density at the surface of the sphere?
28. A parallel-plate capacitor has plates of area A and separation d and is charged to a
potential difference V. The charging battery is then disconnected and the plates are pulled
Chap. 30 PROBLEMS 769
apart until their separation is 2d. Derive expressions in terms of A, d, and V for (a) the
new potential difference, (b) the initial and the final stored energy, and (c) the work re-
quired to separate the plates.
29. Show that the plates of a parallel-plate capacitor attract each other with a force
given by

Prove this by calculating the work necessary to increase the plate separation from x to
x + dx.
30. In the capacitor of Example 5 the dielectric slab fills half the space between the
plates. (a) What per cent of the energy is stored in the air gaps? (b) What per cent is
stored in the slab?
31. A parallel-plate air capacitor has a capacitance of 100 µµf. (a) What is the stored
energy if the applied potential difference is 50 volts? (b) Can you calculate the energy
density for points between the plates?
32. For the capacitors of Problem 1, compute the energy stored for the three different
connections of parts (a), (b), and (c). Compare your answers and explain any differences.
33. A charge q is placed on the surface of an originally uncharged soap bubble of radius
Ro. Because of the mutual repulsion of the charged surface, the radius is increased to a
somewhat larger value R. Show that
q = [3s27r 2eopRoR(R2 + RoR + Rl)]Y2
in which p is the pressure of the atmosphere. Find q for p = 1.00 atm, Ro = 2.00 cm,
and R = 2.10 cm. (Hint: The work done by the bubble in pushing back the atmosphere
must equal the decrease in the stored electric field energy that accompanies the expansion,
from the conservation of energy principle.)
34. Two capacitors (2.0 µf and 4.0 µf) are connected in parallel across a 300-volt poten-
tial difference. Calculate the total stored energy in the system.
35. A parallel-connected bank of 2000 5.0-µf capacitors is used to store electric energy.
What does it cost to charge this bank to 50,000 volts, assuming a rate of 2¢/kw-hr?
36. In Fig. 30-21 find (a) the charge, (b) the potential difference, and (c) the stored energy
for each capacitor. Assume the numerical values of Problem 11.
37. In Fig. 30-23 find (a) the charge, (b) the potential difference, and (c) the stored energy
for each capacitor. Assume the numerical values of Problem 15.
Current and Resistance
CHAPTER 31

31-1 Current and Current Density


The free electrons in an isolated metallic conductor, such as a length of
copper wire, are in random motion like the molecules of a gas confined to a
container. They have no net directed motion along the wire. If a hypo-.
thetical plane is passed through the wire, the rate at which electrons pass
through it from right to left is the same as the rate at which they pass through
from left to right; the net rate is zero.
If the ends of the wire are connected to a battery, an electric field will be
set up at every point within the wire. If the potential difference maintained
by the battery is 10 volts and if the wire (assumed uniform) is 5 meters long,
the strength of this field at every point will be 2 volts/meter. This field E
will act on the electrons and will give them a resultant motion in the direc-
tion of - E. We say that an electric current i is established; if a net charge q
passes through any cross section of the conductor in time t the current,
assumed constant, is
i = q/t. (31-1)
The appropriate mks units are amperes for i, coulombs for q, and seconds for
t. The student will recall (Section 26-4) that Eq. 31-1 is the defining equa-
tion for the coulomb and that we have not yet given an operational definition
of the ampere; we do so in Section 34-4.
If the rate of flow of charge with time is not constant, the current varies
with time and is given by the differential limit of Eq. 31-1, or
i = dq/dt. (31-2)
In the rest of this chapter we consider only constant currents.
770
Sec. 31-1 CURRENT AND CURRENT DENSITY 771

The current i is the same for all cross sections of a conductor, even though
the cross-sectional area may be different at different points. In the same way
the rate at which water (assumed incompressible) flmvs past any cross section
of a pipe is the same even if the cross sectiorr varies. The water flows faster
where the pipe is smaller and slower where it is larger, so that the volume
rate, measured perhaps in gal/min, remains unchanged. This constancy of
the electric current follows because charge must be conserved; it does not
pile up steadily or drain away steadily from any point in the conductor under
the assumed steady-state conditions. In the language of Section 18-3 there
are no "sources" or "sinks" of charge.
The existence of an electric field inside a conductor does not contradict
Section 28-4, in which we asserted that E equals zero inside a conductor. In
that section, which dealt with a state in which all net motion of charge had
stopped (electrostatics), we assumed that the conductor was insulated and
that no potential difference was deliberately maintained between any two
points on it, as by a battery. In this chapter, which deals with charges in
motion, we relax this restriction.
The electric field that acts on the electrons in a conductor does not pro-
duce a net acceleration because the electrons keep colliding with the atoms
(strictly, ions) that make up the conductor. This array of ions, coupled
together by strong spring-like forces of electric origin, is called the lattice
(see Fig. 21-5). The over-all effect of these collisions is to transfer kinetic
energy from the accelerating electrons into vibrational energy of the lattice.
The electrons acquire a constant average drift speed vd in the direction - E.
The analogy is to a marble rolling down a long flight of stairs and not to a
marble falling freely from the same height. In the first case the acceleration
caused by the (gravitational) field is effectively canceled by the decelerating
effects of collisions with the stair treads so that, under proper conditions, the
marble rolls down the stairs with zero average acceleration, that is, at con-
stant average speed.
Although in metals the charge carriers are electrons, in electrolytes or in
gaseous conductors they may also be positive or negative ions or both. A
convention for labeling the directions of currents is needed because charges
of opposite sign move in opposite directions in a given field. A positive
charge moving in one direction is equivalent in nearly all external effects to a
negative charge moving in the opposite direction. Hence, for simplicity and
algebraic consistency, we assume that all charge carriers are positive and we
draw the current arrows in the direction that such charges would move. If the
charge carriers are negative, they simply move opposite to the direction of
the current arrow (see Fig. 31-1). When we encounter a case (as in the Hall
effect; see Section 33-5) in which the sign of the charge carriers makes a
difference in the external effects, we will disregard the convention and take
the actual situation into account.
Current i is a characteristic of a particular conductor. It is a macroscopic
quantity, like the mass of an object, the volume of an object, or the length
of a rod. A related microscopic quantity is the current density j. It is a
772 CURRENT AND RESISTANCE Chap. 31

<~---E

Fig. 31-1 Electrons drift in a direction opposite to the electric field in a conductor.

vector and is characteristic of a point inside a conductor rather than of the


conductor as a whole. If the current is distributed uniformly across a con-
ductor of cross-sectional area A, the magnitude of the current density for
all points on that cross section is

j = i/A. (31-3)

The vector j at any point is oriented in the direction that a positive charge
carrier would move at that point. An electron at that point would move in
the direction - j.
The general relationship between j and i is that, for a particular surface in
a conductor, i is the flux of the vector j over that surface, or

i = fj·dS, (31-4)

where dS is an element of surface area and the integral is taken over the
surface in question. Equation 31-3 (written as i = jA) is a special case of
this relationship in which the surface of integration is a cross section of the
conductor and in which j is constant over this surface and at right angles
to it. However, Eq. 31-4 may be applied to any surface through which we
wish to know the current. Equation 31-4 shows clearly that i is a scalar be-
cause the integrand j ·dS is a scalar.

The arrow often associated with the current in a wire does not indicate that current
is a vector but merely shows the sense of charge flow. Positive charge carriers either
move in a certain direction along the wire or in the opposite direction, these two possi-
bilities being represented by+ or - in algebraic equations. Note that (a) the current
in a wire remains unchanged if the wire is bent, tied into a knot, or otherwise distorted,
and (b) the arrows representing the sense of currents do not in any way obey the laws
of vector addition.

The drift speed va of charge carriers in a conductor can be computed from


the current density j. Figure 31-1 shows the conduction electrons in a wire
moving to the right at an assumed constant drift speed vd. The number of
conduction electrons in the wire is nAl where n is the number of conduction
electrons per unit volume and Al is the volume of the wire. A charge of
magnitude
q = (nAl)e
Sec. 31-1 CURRENT AND CURRENT DENSITY 773
passes out of the wire, through its right end, in a time t given by
l

The current i is given by


q nAle
i = - = - - = nAeva.
t l/va
Solving for va and recalling thatj = i/A (Eq. 31-3) yields
. .
J
i
Va= - - = - · (31-5)
nAe ne
~ Example 1. An aluminum wire whose diameter is 0.10 in. is welded end to end
to a copper wire with a diameter of 0.064 in. The composite wire carries a steady
current of 10 amp. What is the current density in each wire?
The current is distributed uniformly over the cross section of each conductor, ex-
cept near the junction, which means that the current density may be taken as con-
stant for all points within each wire. The cross-sectional area of the aluminum wire is
0.0079 in. 2 Thus, from Eq. 31-3,

. - i - 10 amp - 1300 ;· 2
JAl - A - 0.0079 in.2 - amp m.

The cross-sectional area of the copper wire is 0.0032 in. 2 Thus


. i 10 amp 3 100 amp /'m. 2
Jcu = A = 0.0032 in.2 =

The fact that the wires are of different materials does not enter into consideration
here.
Example 2. What is va for the copper wire in Example 1?
We can write the current density for the copper wire as 480 amp/cm 2. To compute
n we start from the fact that there is one free electron per atom in copper. The num-
ber of atoms per unit volume is dN 0 /M where dis the density, No is Avogadro's
number, and M is the atomic weight. The number of free electrons per unit volume
is then
dN 0 (9.0 gm/cm 3)(6.0 X 10 23 atoms/mole)(l electron/atom)
n =--=
M 64 gm/mole
= 8.4 X 1022 electrons/cm 3.
Finally, Va is, from Eq. 31-5,
j 480 amp/cm 2
Va= - = 22
ne (8.4 X 10 electrons/cm 3)(1.6 X 10-19 coul/electron)
= 3.6 X 10- 2 cm/sec.
It takes 28 sec for the electrons in this wire to drift 1.0 cm. Would you have guessed
that Va was so low? The drift speed of electrons must not be confused with the speed
at which changes in the electric field configuration travel along wires, a speed which
approaches that of light. When pressure is applied to one end of a long water-filled
tube, a pressure wave travels rapidly along the tube. The speed at which water moves
through the tube is much lower, however. <11111
774 CURRENT AND RESISTANCE Chap. 31
31-2 Resistance, Resistivity, and Conductivity
If the same potential difference is applied between the ends of a rod of
copper and of a rod of wood, very different currents result. The characteristic
of the conductor that enters here is its resistance. We define the resistance of
a conductor (often called a resistor; symbol ...1\/M-) between two points by
applying a potential difference V between those points, measuring the
current i, and dividing:
R = V /i. (31-6)

If Vis in volts and i in amperes, the resistance R will be in ohms.


The flow of charge through a conductor is often compared with the flow of
water through a pipe, which occurs because there is a difference in pressure
between the ends of the pipe, established perhaps by a pump. This pressure
difference can be compared with the potential difference established between
the ends of a resistor by a battery. The flow of water (ft 3 /sec, say) is com-
pared with the current (coul/sec or amp). The rate of flow of water for a
given pressure difference is determined by the nature of the pipe. Is it long
or short? Is it narrow or wide? Is it empty or filled, perhaps with gravel?
These characteristics of the pipe are analogous to the resistance of a con-
ductor.

Primary standards of resistance, kept at the Nationa~ Bureau of Standards, are


spools of wire whose resistances have been accurately measured. Because resistance
varies with temperature, these standards, when used, are placed in an oil bath at a
controlled temperature. They are made of a special alloy, called manganin, for which
the change of resistance with temperature is very small. They are carefully annealed
to eliminate strains, which also affect the resistance. These primary standard resis-
tors are used chiefly to calibrate secondary standards for other laboratories.
Operationally, the primary resistance standards are not measured by using Eq.
31-6 but are measured in an indirect way which involves magnetic fields. Equation
31-6 is, in fact, used to measure V, by setting up an accurately known current i (using
a current balance; see Section 34-4) in an accurately known resistance R. This opera-
tional procedure for potential difference is the one normally used in place of the con-
ceptual definition introduced in Section 29-1, in which one measures the work per
unit charge required to move a test charge between two points.

Related to resistance is the resistivity p, which is a characteristic of a ma-


terial rather than of a particular specimen of a material; it is defined, for iso-
tropic materials,* from
E
p = --:-· (31-7)
J

The resistivity of copper is 1.7 X 10-s ohm-m; that of fused quartz is about
10 16 ohm-m. Few physical properties are measurable over such a range of
values; Table 31-1 lists some values for common metals.

* These are materials whose properties (electrical in this case) do not vary with direction
in the material.
Sec. 31-2 RESISTANCE, RESISTIVITY, AND CONDUCTIVITY 775

Table 31-1

PROPERTIES OF METALS AS CONDUCTORS

Temperature
Resistivity
Coefficient of Density, Melting
(at 20°C),
Resistivity,* gm/cm 3 Point, °C
ohm-m
a, per C 0

Aluminum 2.8 x 10-8 3.9 x 10- 3 2.7 659


Copper 1.7 x 10-8 3.9 x 10-3 8.9 1080
Carbon (amorphous) 3.5 x 10- 5 -5 x 10- 4 1.9 3500
Iron 1.0 x 10- 7 5.0 x 10- 3 7.8 1530
Manganin 4.4 x 10-7 1 X 10- 5 8.4 910
Nickel 6.8 x 10- 8 6 x 10- 3 8.9 1450
Silver 1.6 x 10- 8 3.8 x 10-3 10.5 960
Steel 1.8 x 10- 7 3 x 10-3 7.7 1510
Wolfram (tungsten) 5.6 x 10- 8 4.5 x 10-3 19 3400

* This quantity, defined from


1 dp
a= --- (31-8)
P dT

is the fractional change in resistivity (dp/ p) per unit change in temperature. It


varies with temperature, the values here referring to 20°C. For copper (a = 3.9 X
10- 3/C 0 ) the resistivity increases by 0.39 per cent for a temperature increase of 1 C 0
near 20°C. Note that a for carbon is negative, which means that the resistivity de-
creases with increasing temperature.

Consider a cylindrical conductor, of cross-sectional area A and length Z,


carrying a steady current i. Let us apply a potential difference V between
its ends. If the cylinder cross sections at each end are equipotential surfaces,
the electric field strength and the current density will be constant for all
points in the cylinder and will have the values

v
E=- and J
l A

The resistivity p may then be written as

E V/l
p = - = -·
j i/A
But V /i is the resistance R which leads to

l
R = p-· (31-9)
A
776 CURRENT AND RESISTANCE Chap. 31
V, i, and R are macroscopic quantities, applying to a particular body or extended
region. The corresponding microscopic quantities are E, j, and p; they have values at
every point in a body. The macroscopic quantities are related to each other by Eq.
31-6 (V = iR) and the microscopic quantities by Eq. 31-7, which can be written in
vector form as E = jp. •
The macroscopic quantities can be found by integrating over the microscopic
quantities, using relations already given, namely

(31-4)

and Vab = - f E·dl.


a
b
(29-5)

The integral in Eq. 31-4 is a surface integral, carried out over any cross section of the
conductor. The integral in Eq. 29-5 is a line integral carried out along an arbitrary
line drawn along the conductor, connecting any two equipotential surfaces, identified
by a and b. For a long wire connected to a battery equipotential surface a might be
chosen as a cross section of the wire near the positive battery terminal and b might
be a cross section near the negative terminal.
The resistance of a conductor between a and b can be expressed in microscopic
terms by dividing the two equations, or

R = V_ab = - ibE·dl
'/,

If the conductor is a long cylinder of cross section A and length l and if points a and b
are its ends, the foregoing equation for R (see Eq. 31-7) reduces to

which is Eq. 31-9.


The macroscopic quantities V, i, and Rare of primary interest when we are making
electrical measurements on real conducting objects. They are the quantities that one
reads on meters. The microscopic quantities E, j, and pare of primary importance
when we are concerned with the fundamental behavior of matter (rather than of
specimens of matter), as we usually are in the research area of solid state physics.
Thus Section 31-4 deals appropriately with an atomic view of the resistivity of a metal
and not of the resistance of a metallic specimen. The microscopic quantities are also
important when we are interested in the interior behavior of irregularly shaped con-
ducting objects.
..... Example 3. A rectangular carbon block has dimensions 1.0 cm X 1.0 cm X 50
cm. (a) ·what is the resistance measured between the two square ends? (b) Between
two opposing rectangular faces? The resistivity of carbon at 20°C is 3.5 X 10-5
ohm-m.
(a) The area of a square end is 1.0 cm 2 or 1.0 X 10-4 meter 2 • Equation 31-9 gives
for the resistance between the square ends:

R = P _!_ = (3.5 X 10-5 ohm-m)(0.50 meter) = O.lS ohm.


A 1.0 X 10- 4 meter 2
(b) For the resistance between opposing rectangular faces (area = 5.0 X 10-3
meter 2), we have
R = P _!_ = (3.5 X 10- 5 ohm-m)(10- 2 meter) = 7 _0 X 10 _ 5 ohm.
A 5.0 X 10- 3 meter 2
Sec. 31-2 RESISTANCE, RESISTIVITY, AND CONDUCTIVITY 777
Thus a given conductor can have a number of resistances, depending on how the
potential difference is applied to it. The ratio of resistances for these two cases is
2600. We assume in each that the potential difference is applied to the block in such
a way that the surfaces between which the resistance is desired are equipotential.
Otherwise Eq. 31-9 would not be valid. ~

Figure 31-2 shows (solid curve) how the resistivity of copper varies with tempera-
ture. Sometimes, for practical use, such data are expressed in equation form. If we
are interested in only a limited range of temperatures extending, say, from Oto 500°0,
we can fit a straight line to the curve of Fig. 31-2, making it pass through two arbi-
trarily selected points; see the dashed line. We choose the point labeled To, p 0 in
the figure as a reference point, To being 0°0 in this case and p0 being 1.56 X 10-s
ohm-m. The resistivity p at any temperature T can be found from the empirical
equation of the dashed straight line in Fig. 31-2, which is
p = po[l + a(T - To)]. (31-10)
This relation shows correctly that p - po as T - T 0•
If we solve Eq. 31-10 for a, we obtain
_ 1 p - Po
a=- .
Po T - To

Comparison with Eq. 31-8 shows that a is a mean temperature coefficient of resistivity
for a selected pair of temperatures rather than the temperature coefficient of re-
sistivity at a particular temperature, which is the definition of a. For most practical
purposes Eq. 31-10 gives results that are within the acceptable range of accuracy.

T, °C
-200 0 200 400 600 800
8

E
I
E
.c
0
00 4
I
0
.....
c:i:

200 400 600 800 1000 1200


T, °K

Fig. 31-2 The resistivity of copper as a function of temperature. The dashed line is an
ap,proximation chosen to fit the curve at the two circled points. The point marked To, p0
is chosen as a reference point.
778 CURRENT AND RESISTANCE Chap. 31

0.16

(/)

E
-§ 0.08 Fig. 31-3 The resistance of mer-
cury disappears below about 4 °K.
~

2 4 6
T, °K

The curve of Fig. 31-2 does not go to zero at the absolute zero of temperature, even
though it appears to do so, the residual resistivity at this temperature being 0.02 X
10-s ohm-m. For many substances the resistance does become zero at some low
temperature. Figure 31-3 shows the resistance of a specimen of mercury for temper-
atures below 6°K. In the space of about 0.05 K 0 the resistance drops abruptly to an
immeasurably low value. This phenomenon, called superconductivity,* was dis-
covered by Kamerlingh Onnes in the Netherlands in 1911. The resistance of ma-
terials in the superconducting state seems to be truly zero; currents, once established
in closed superconducting circuits, persist for weeks without diminution, even though
there is no battery in the circuit. If the temperature is raised slightly above the
superconducting point, such currents drop rapidly to zero.

31-3 Ohm's Law


Let us apply a variable potential difference V between the ends of a 100-
foot coil of #18 copper wire. For each applied potential difference, let us
measure the current i and plot it against Vas in Fig. 31-4. The straight line
that results means that the resistance of this conductor is the same no matter
what applied voltage is used to measure it. This important result, which holds
for metallic conductors, is known as Ohm's law. We assume that the tem-
perature of the conductor is essentially constant throughout the measure-
ments.
Many conductors do not obey Ohm's law. Figure 31-5, for example, shows
a V--i plot for a type 2A3 vacuum tube. The plot is not straight and the
resistance depends on the voltage used to measure it. Also, the current for
this device is almost vanishingly small if the polarity of the applied potential
difference is reversed. For metallic conductors the current reverses direc-
tion when the potential difference is reversed, but its magnitude does not
change.
Figure 31-6 shows a typical V--i plot for another nonohmic device, a
thermistor. This is a semiconductor (see p. 650) with a large and negative
temperature coefficient of resistivity a (see Table 31-1) that varies greatly
with temperature. We note that two different currents through the ther-
mistor can correspond to the same potential difference between its ends.
* See "Superconductivity" by B. T. Matthias, Scientific American, p. 92, November
1957.
Sec. 31-3 OHM'S LAW 779
0.8

0.6

a. .
Fig. 31-4 The current in a par- ~ 0.4 .•
ticular copper conductor as a func- .....-
tion of potential difference. This
conductor obeys Ohm's law.
0.2

o-
0 0.2 0.4 0.6
V, volts

80

a.
E
ro
c,:,
I Fig. 31-5 The current in a type
0
...... 2A3 vacuum tube as a function of
..... potential difference. This conduc-
20
tor does not obey Ohm's law.

100 200 300 400


V, volts

4.0

Fig. 31-6 A plot of current as a 3.0


function of potential difference in
a Western Electric 1-B thermistor. c.
The curve shows how the voltage E
<ti

across the thermistor varies as the ""I 2.0


current through it is increased. 0
.....
The shape of the curve can be ac- .....
counted for in terms of the large
negative temperature coefficient
1.0
of resistivity of the material of
which the device is made.

10 20 30
V, volts
780 CURRENT AND RESISTANCE Chap. 31
Thermistors are often used to measure the rate of energy flow in microwave
beams by allowing the microwave beam to fall on the thermistor and heat it.
The rel&,tively small temperature rise so produced results in a relatively large
change in resistance, which serves as a measure of the microwave power.
Modern electronics, and therefore much of the character of our present tech-
nological civilization, depends in a fundamental way on the fact that many
conductors, such as vacuum tubes, crystal rectifiers, thermistors, and
transistors, do not obey Ohm's law.
We stress that the relationship V = iR is not a statement of Ohm's law. A
conductor obeys Ohm's law only if its V - i curve is linear, that is, if R is
independent of V and i. The relationship R = V / i remains as the general
definition of the resistance of a conductor whether or not the conductor
obeys Ohm's law.
The microscopic equivalent of the relationship V = iR is Eq. 31-7, or E
= jp. A conducting material is said to obey Ohm's law if a plot of E versus
j is linear, that is, if the resistivity pis independent of E and j. Ohm's law
is a specific property of certain materials and is not a general law of electro-
magnetism, for example, like Gauss's law.
A close analogy exists between the fl.ow of charge because of a potential difference
and the fl.ow of heat because of a temperature difference. Consider a thin electrically
conducting slab of thickness Lix and area A. Let a potential difference Li V be main-
tained between opposing faces. The current i is given by Eqs. 31-6 (i = V/R) and
31-9 (R = pl/A), or
. LiV LiVA
?,=-=--·
R pLix
In the limiting case of a slab of thickness dx this becomes
i = !A dV
p dx
dq dV
or - = - a A -- , (31-11)
dt dx
where a ( = 1/p) is the conductivity of the material. Since positive charge flows in
the direction of decreasing V, we introduce a minus sign into Eq. 31-11, that is,
dq/dt is positive when dV /dx is negative.
The analogous heat fl.ow equation (see Section 22-4) is

dQ = -kA dT, (31-12)


dt dx
which shows that k, the thermal conductivity, corresponds to a and dT /dx, the tem-
perature gradient, corresponds to dV / dx, the potential gradient. There is more than
a formal mathematical analogy between Eqs. 31-11 and 31-12. Both heat energy
and charge are carried by the free electrons in a metal; empirically, a good electrical
conductor (silver, say) is also a good heat conductor and conversely.

31-4 Resistivity-an Atomic View


One can understand why metals obey Ohm's law on the basis of simple
classical ideas. If these ideas are modified when necessary by the require-
ments of quantum physics, it is possible to go further and to calculate theo-
retical values of the resistivity p for various metals. These calculations are
Sec. 31-4 RESISTIVITY-AN ATOMIC VIEW 781

not simple, but when they have been carried out the agreement with the
experimental value of p has usually been good.
In a metal the valence electrons are not attached to individual atoms but
are free to move about within the lattice and are called conduction electrons.
In copper there is one such electron per atom, the other 28 remaining bound
to the copper nuclei to form ionic cores.
The speed distribution of conduction electrons can be described correctly
only in terms of quantum physics. For our purposes, however, it suffices to
consider only a suitably defined average speed ii; for copper ii = 1.6 X 108
cm/sec. In the absence of an electric field, the directions in which the elec-
trons move are completely random, like those of the molecules of a gas con-
fined to a container.
The electrons collide constantly with the ionic cores of the conductor,
that is, they interact ·.vith the lattice, often suffering sudden changes in speed
and direction. These collisions remind us of the collisions of gas molecules
confined to a container. As in the case of molecular collisions, we can des-
cribe electron-lattice collisions by a mean free path [, where [ is the average
distance that an electron travels between collisions.*
In an ideal metallic crystal at 0°K electron-lattice collisions would not occur,
according to the predictions of quantum physics, that is, l .- oo as T .- 0°K for
ideal crystals. Collisions take place in actual crystals because (a) the ionic cores at
any temperature Tare vibrating about their equilibrium positions in a random way,
(b) impurities, that is, foreign atoms, may be present, and (c) the crystal may contain
lattice imperfections, such as rows of missing atoms and displaced atoms. On this
view it is not surprising that the resistivity of a metal can be increased by (a) raising
its temperature, (b) adding small amounts of impurities, and (c) straining it severely,
as by drawing it through a die, to increase the number of lattice imperfections.
When an electric field is applied to a metal, the electrons modify their
random motion in such a way that they drift slowly, in the opposite direction
to that of the field, with an average drift speed vd. This drift speed is much
less than the effective average speed ii mentioned above (see Example 2).
Figure 31-7 suggests the relationship between these two speeds. The solid
lines suggest a possible random path followed by an electron in the absence of
an applied field; the electron proceeds from x toy, making six collisions on the
way. The dashed curves show how this same event might have occurred if
an electric field E had been applied. Note that the electron drifts steadily to
the right, ending at y' rather than at y. In preparing Fig. 31-7, it has been
assumed that the drift speed vd is 0.02ii; actually, it is more like 10-10ii, so
that the "drift" exhibited in the figure is greatly exaggerated.
The drift speed vd can be calculated in terms of the applied electric field E
and of ii and [. When a field is applied to an electron in the metal it will
experience a force eE which will impart to it an acceleration a given by
Newton's second law, eE
a=-·
m
* It can be shown that collisions between electrons occur only rarely and have little
effect on the resistivity.
782 CURRENT AND RESISTANCE Chap. 31

Fig. 31-7 The solid lines show an electron moving from x to y, making six collisions.
The dashed curves show what the electron path might have been in the presence of an
electric field E. Note the steady drift in the direction of - E.

Consider an electron that has just collided with an ion core. The col-
lision, in general, will momentarily destroy the tendency to drift and the
electron will have a truly random direction after the collision. At its next
collision the electron's velocity component in the - E direction will have
changed, on the average, by a(Z/v) where l/v is the mean time between
collisions. We call this the drift speed vd, or

Vd = a(J)v
= eE!.
mv
(31-13)

The electron's motion through the conductor is analogous to the constant rate of
fall of a stone in water. The gravitational force Fg on the stone is opposed by a
viscous resisting force that is proportional to the velocity, or
Fg =mg= "bv,
where bis a viscous coefficient (see Section 15-8). Thus the constant terminal speed
of the stone is

We can rewrite Eq. 31-13 as


Va = (~fj) FE,
where FE ( = eE) is the electrical force. Comparison of these equations shows that
the equivalent "viscous coefficient" for the motion of an electron in a particular con-
ductor is [/ mv. If l is short, the conductor exhibits a greater "viscous effect" on
the electron motion, and the drift speed Va is proportionally lower.
We may express vd in terms of the current density (Eq. 31-5) and combine
with Eq. 31-13 to obtain
J eEl
Vd = - = --·
ne mv
Sec. 31-5 ENERGY TRANSFERS IN AN ELECTRIC CIRCUIT 783

Combining this with Eq. 31-7 (p = E/j) leads finally to


mv (31-14)
p = --·
ne 2 l
Equation 31-14 can be taken as a statement that metals obey Ohm's law
if we can show that v and [ do not depend on the applied electric field E. In
this case p will not depend on E, which (see Section 31-3) is the criterion that
a material obey Ohm's law. The quantities v and l depend on the speed dis-
tribution of the conduction electrons. We have seen that this distribution is
affected only slightly by the application of even a relatively large electric
field, since vis of the order of 108 cm/sec and Vd (see Example 2) only of the
order of 10-2 cm/sec, a ratio of 10 10 . We may be sure that whatever the
values of v and l are (for copper at 20°C, say) in the absence of a field they
remain essentially unchanged when the field is applied. Thus the right side
of Eq. 31-14 is independent of E and the material obeys Ohm's law. The
numerical calculation of p from Eq. 31-14 is hampered by the difficulty of
calculating l, although the calculation has been carried out in a number of
cases .
..... Example 4. What are (a) the mean time T between collisions and (b) the mean
free path for free electrons in copper?
(a) From Eq. 31-14 (see also Example 2), we have
[ m (9.1 X 10- 31 kg)
T =- = ~ = ~~~~-:--~~c--~~~~~~-----,----~~~~--~~
ii ne 2 p (8.4 X 10 28 /meter 3 )(1.6 X 10- 19 coul) 2 (1.7 X 10-s ohm-m)
= 2.5 X 10- 14 sec.
(b) The mean free path is
[ = TV = (2.5 X 10- 14 sec)(l.6 X 10 8 cm/sec) = 4.0 X 10- 5 cm.
This is about 200 ionic diameters.

31-5 Energy Transfers in an Electric Circuit


Figure 31-8 shows a circuit consisting of a battery B connected to a
"black box." A steady current i exists in the connecting wires and a steady
potential difference Vab exists between the terminals a and b. The box might
contain a resistor, a motor, or a storage battery, among other things.

i-

Fig. 31-8 A battery B sets up a current in a circuit


containing a "black box."
B +

-i
784 CURRENT AND RESISTANCE Chap. 31
Terminal a, connected to the positive battery terminal, is at a higher
potential than terminal b. If a charge dq moves through the box from a to b,
this charge will decrease its electric potential energy by dq Vab (see Section
29-6). The conservation-of-energy principle tells us that this energy is
transferred in the box from electric potential energy to some other form.
What that other form will be depends on what is in the box. In a time dt the
energy d U transferred inside the box is then

dU = dq Vab =i dt Vab·

We find the rate of energy transfer P by dividing by the time, or

dU
P = - = iVab· (31-15)
dt

If the device in the box is a motor, the energy appears largely as mechanical
work done by the motor; if the device is a storage battery that is being
charged, the energy appears largely as stored chemical energy in this second
battery.
If the device is a resistor, we assert that the energy appears as heat in the
resistor. To see this, consider a stone of mass m that falls through a height h.
It decreases its gravitational potential energy by mgh. If the stone falls in a
vacuum or-for practical purposes-in air, this energy is transformed into
kinetic energy of the stone. If the stone falls in water, however, its speed
eventually becomes constant, which means that the kinetic energy no longer
increases. The potential energy that is then steadily being made available
as the stone falls at constant speed appears as thermal energy in the stone
and the surrounding water. It is the viscous, friction-like drag of the
water on the surface of the stone that stops the stone from accelerating, and
it is at this surface that thermal energy appears.
The course of the electrons through the resistor is much like that of the
stone through water. The electrons travel with a constant drift speed va and
thus do not gain kinetic energy. The electric potential energy that they lose
is transferred to the resistor as heat. On a microscopic scale this can be
understood in that collisions between the electrons and the lattice (see Fig.
21-5) increase the amplitude of the thermal vibrations of the lattice; on a
macroscopic scale this corresponds to a temperature increase. This effect,
which is thermodynamically irreversible, is called Joule heating.
For a resistor we can combine Eqs. 31-15 and 31-6 (R = V/i) and obtain
either
(31-16)

vz
or P=-· (31-17)
R
Sec. 31-5 ENERGY TRANSFERS IN AN ELECTRIC CIRCUIT 785
Note that Eq. 31-15 applies to electrical energy transfer of all kinds; Eqs.
31-16 and 31-17 apply only to the transfer of electrical energy to heat energy
in a resistor. Equations 31-16 and 31-17 are known as Joule's law. This
law is a particular way of writing the conservation-of-energy principle for the
special case in which electrical energy is transferred into heat energy.
The unit of power that follows from Eq. 31-15 is the volt-amp. It can be
written as

1 volt-amp =
:,.._ 1
1 v~t-a ~ p ( 1 joule ) ( 1 c\ul )
1 v~t X 1 c~l 1 a~p X 1 sec
= 1 joule/sec.

The first conversion factor in parenthesis comes from the definition of the
volt (Eq. 29-1); the second comes from the definition of the coulomb. The
joule/sec is such a common unit that it is given a special name of its own, the
watt; see Section 7-7. Power is not an exclusively electrical concept, of
course, and we can express in watts the power ( = F·v) expended by an agent
that exerts a force F while it moves with a velocity v.
.... Example 5. You are given a 20-ft length of heating wire made of the special
alloy Nichrome; it has a resistance of 24 ohms. Can you obtain more heat by winding
one coil or by cutting the wire in two and winding two separate coils? In each case
the coils are to be connected individually across a 110-volt line.
The power P for the single coil is given by Eq. 31-17 :

V2 (110 volts) 2
P = ~R = 2 h = 500 watts.
4 oms

The power for a coil of half the length is given by

P' = (110 volts)2 = 1000 watts.


12 ohms

There are two "half-coils," so that the total power obtained by cutting the wire in
half is 2000 watts, or four times that for the single coil. This would seem to suggest
that we could buy a 500-watt heating coil, cut it in half, and rewind it to obtain
2000 watts. Why is this not a practical idea? <Ill!
786 CURRENT AND RESISTANCE Chap. 31

QUESTIONS

1. Name other physical quantities that, like current, are scalars having a sense repre-
sented by an arrow in a diagram.
2. What conclusions can you draw by applying Eq. 31-4 to a closed surface through
which a number of wires pass in random directions, carrying steady currents of different
sizes?
3. A potential difference V is applied to a circular cylinder of carbon by clamping it
between circular copper electrodes, as in Fig. 31-9. Discuss the difficulty of calculating
the resistance of the carbon cylinder, using the relation R = pL/ A.

Fig. 31-9

4. How would you measure the resistance of a pretzel-shaped conductor? Give specific
details to clarify the concept.
5. Discuss the difficulties of testing whether the filament of a light bulb obeys Ohm's
law.
6. Does the relation V = iR apply to nonohmic resistors?
7. The temperature coefficient of resistance of a thermistor is negative and varies
greatly with temperature. Account qualitatively for the shape of the curve of i versus 11
for the thermistor of Fig. 31-6.
8. A potential difference 11 is applied to a copper wire of diameter d and length l. What
is the effect on the electron drift speed of (a) doubling V, (b) doubling l, and (c) doubling d'?
9. If the drift speeds of the electrons in a conductor under ordinary circumstances are
so slow (see Example 2), why do the lights in a room turn on so quickly after the switch is
closed?
10. Can you think of a way to measure the drift speed for electrons by timing their
travel along a conductor?
11. Why are the dashed ·white lines in Fig. 31-7 curved slightly?
12. A current i enters the top of a copper sphere of radius Rand leaves at a diametrically
opposite point. Are all parts of the sphere equally effective in dissipating Joule heat?
13. What special characteristics must (a) heating wire and (b) fuse wire have? '
Chap. 31 PROBLEMS 787

14. Equation 31-16 (P = i 2R) seems to suggest that the rate of Joule heating in a
resistor is reduced if the resistance is made less; Eq. 31-17 (P = V 2 /R) seems to suggest
just the opposite. How do you reconcile this apparent paradox?
15. Is the filament resistance lower or higher in a 500-watt light bulb than in a 100-watt
bulb? Both bulbs are designed to operate on 110 volts.
16. Five wires of the same length and diameter are connected in turn between two points
maintained at constant potential difference. Will heat be developed at the fastest rate in
the wire of (a) the smallest or (b) the largest resistance?

PROBLEMS

1. A current of 5 amp exists in a 10-ohm resistance for 4 min. (a) How many coulombs
and (b) how many electrons pass through any cross section of the resistor in this time?
_,2:" A current is established in a gas discharge tube when a sufficiently high potential
difference is applied across the two electrodes in the tube. The gas ionizes; electrons move
toward the positive terminal and positive ions toward the negative terminal. What are the
magnitude and sense of the current in a hydrogen discharge tube in which 3.1 X 1018
electrons and 1.1 X 10 18 protons move past a cross-sectional area of the tubP each second?
3. A copper wire and an iron wire of the same length have the same potential difference
applied to them. (a) What must be the ratio of their radii if the current is to be the same?
(b) Can the current density be made the same by suitable choices of the radii?
4. A current i enters one corner of a square sheet of copper and leaves at the opposite
corner. Sketch arrows for various points within the square to represent the relative values
of j.,. Intuitive guesses rather than detailed mathematical analysis are called for.
/5. The belt of an electrostatic generator is 50 cm wide and travels at 30 meters/sec.
The belt carries charge into the sphere at a rate corresponding to 10-4 amp. Compute
the surface charge density on the belt.
6. A square aluminum rod is 1.0 meter long and 5.0 mm on edge. (a) What is the resist-
ance between its ends? (b) What must be the diameter of a circular LO-meter copper rod
if its resistance is to be the same?
7. A wire with a resistance of 6.0 ohms is drawn out so that its new length is three
times its original length. Find the resistance of the longer wire, assuming that the resis-
tivity and density of the material are not changed during the drawing process.
8. A copper wire and an iron wire of equal length l and diameter d are joined and a
potential difference V is applied between the ends of the composite wire. Calculate (a)
the electric field strength in each wire, (b) the current density in each wire, and (c) the
potential difference across each wire. Assume that l = 10 meters, d = 2.0 mm, and V =
100 volts.
9. A rod of a certain metal is 1.00 meter long and 0.550 cm in diameter. The resistance
between its ends (at 20°C) is 2.87 X 10-3 ohm. A round disk is formed of this same ma-
terial, 2.00 cm in diameter and 1.00 mm thick. (a) What is the resistance between the
opposing round faces? (b) What is the material?
10. Steel trolley-car rail has a cross-sectional area of 7.1 in. 2 What is the resistance of 10
miles of single track? The resistivity of the steel is 6.0 X 10-7 ohm-m.
11. (a) At what temperature would the resistance of a copper conductor be double its
resistance at 0°C? (b) Does this same temperature hold for all copper conductors, regard-
less of shape or size?
12. It is desired to make a long cylindrical conductor whose temperature coefficient of
resistance at 20°C will be close to zero. (a) If such a conductor is made by assembling
alternate disks of iron and carbon, what is the ratio of the thickness of a carbon disk to
that of an iron disk? Assume that the temperature remains essentially the same in each
disk. (b) What is the ratio of the rate of Joule heating in a carbon disk to that in an iron
disk?
788 CURRENT AND RESISTANCE Chap. 31
13. When a metal rod is heated, not only its resistance but also its length and its cross-
sectional area change. The relation R = pl/ A suggests that all three factors should be
taken into account in measuring p at various temperatures. If the temperature changes by
1.0 C 0 , what per cent changes in R, l, and A occur for a copper conductor. What con-
clusion do you draw? The coefficient of linear expansion is 1.7 X 10-5/C 0 •
14. The copper windings of a motor have a resistance of 50 ohms at 20°C, when the
motor is idle. After running for several hours the resistance rises to 58 ohms, What is
the te;nperature of the windings?
~ (a) Using data from Fig. 31-5, plot the resistance of the vacuum tube as a function of
applied potential difference. (b) Repeat for the thermistor of Fig. 31-6.
16. A small but measurable current of 10-10 amp exists in a copper wire whose diameter
is 0.10 in. Calculate the electron drift speed.
17. Heat is developed in a resistor at a rate of 100 watts when the current is 3.0 amp.
What is the resistance in ohms?
18. A potential difference of 1.0 volt is applied to a 100-ft length of #18 copper wire
(diameter = 0.040 in.). Calculate (a) the current, (b) the current density, (c) the electric
field strength, and (d) the rate of Joule heating.
19. The National Board of Fire Underwriters has fixed safe current-carrying capacities
for various sizes and types of wire. For #10 rubber-coated copper wire (wire diameter =
0.10 in.) the maximum safe current is 25 amp. At this current, find (a) the current density,
(b) the electric field strength, (c) the potential difference for 1000 ft of wire, and (d) the rate
of Joule heating for 1000 ft of wire.
20. A 500-watt immersion heater is placed in a pot containing 2.0 liters of water at 20°C.
(a) How long will it take to bring the water to boiling temperature, assuming that 80% of
the available energy is absorbed by the water? (b) How much longer will it take to boil
half the water away?
21. A nichrome heater dissipates 500 watts when the applied potential difference is 110
volts and the wire temperature is 800°C. How much power would it dissipate if the wire
temperature were held to 200°C by immersion in a bath of cooling oil? The applied poten-
tial difference remains the same; a for nichrome is about 4 X 10-4;c 0 •
22. A beam of 16-Mev deuterons from a cyclotron falls on a copper block. The beam is
equivalent to a current of 15 X 10-6 amp. (a) At what rate do deuterons strike the block?
(b) At what rate is heat produced in the block?
23. A "500-watt" heating unit is designed to operate from a 115-volt line. (a) By what
percentage will its heat output drop if the line voltage drops to 110 volts? Assume no
change in resistance. (b) Taking the variation of resistance with temperature into account,
would the actual heat output drop be larger or smaller than that calculated in (a)?
24. Show that p, the power per unit volume transformed into Joule heat in a resistor,
can be written as
p = j2p or
Electromotive Force
and Circuits
CHAPTER 32

32-1 Electromotive Force


There exist in nature certain devices such as batteries and electric gene-
rators which are able to maintain a potential difference between two points
to which they are attached. Such devices are called seats of electromotive
force (abbr. emf). In this chapter we do not discuss their internal construc-
tion or detailed mode of action but confine ourselves to describing their gross
electrical characteristics and to exploring their usefulness in electric circuits.
Figure 32-la shows a seat of emf B, represented by a battery, connected to
a resistor R. The seat of emf maintains its upper terminal positive and its
lower terminal negative, as shown by the + and - signs. In the circuit ex-
ternal to B positive charge carriers would be driven in the direction shown by
the arrows marked i. In other words, a clockwise current would be set up.
An emf is represented by an arrow which is placed next to the seat and
points in the direction in which the seat, acting alone, would cause a positive
charge carrier to move in the external circuit. A small circle is drawn on the
tail of an emf arrow so that it will not be confused with a current arrow.
A seat of emf must be able to do work on charge carriers that enter it. In
the circuit of Fig. 32-la, for example, the seat acts to move positive charges
from a point of low potential (the negative terminal) through the seat to a
point of high potential (the positive terminal). This reminds us of a pump,
which can cause water to move from a place of low gravitational potential to
a place of high potential.
In Fig. 32-la a charge dq passes through any cross section of the circuit in
time dt. In particular, this charge enters the seat of emf 8 at its low-potential
789
790 ELECTROMOTIVE FORCE AND CIRCUITS Chap. 32

o]: B R
li
-+------ i
Fig. 32-1 (a) A simple electric
(a) circuit and (b) its gravitational
analog.
''"'i@ ,':.,

a
"I i
h

'\\'' ~11, l ,
'' ,,;

(b)

end and leaves at its high-potential end. The seat must do an amount of
work dW on the (positive) charge carriers to force them to go to the point of
higher potential. The emf 8 of the seat is defined from
8 = dW/dq. (32-1)
The unit of emf is the joule/coul (see Eq. 29-1) which is the volt. We
might be inclined to say that a battery has an emf of 1 volt if it maintains a
difference of potential of 1 volt between its terminals. This is true only
under certain conditions, which we describe in Section 32-4.
If a seat of emf does work on a charge carrier, energy must be transferred
within the seat. In a battery, for example, chemical energy is transferred
into electrical energy. Thus we can describe a seat of emf as a device in
which chemical, mechanical, or some other form of energy is changed (re-
versibly) into electrical energy. The chemical energy provided by the
battery in Fig. 32-la is stored in the electric and the magnetic * fields that
surround the circuit. This stored energy does not increase because it is being
drained away, by transfer to Joule heat in the resistor, at the same rate at
which it is supplied. The electric and magnetic fields play an intermediary
role in the energy transfer process, acting as a storage reservoir.
Figure 32-lb shows a gravitational analog of Fig. 32-la. In the top figure
the seat of emf B does work on the charge carriers. This energy, stored
temporarily as electromagnetic field energy, appears eventually as Joule heat
in resistor R. In the lower figure the man, in lifting the bowling balls from
* A current in a wire is surrounded by a magnetic field, and this field, like the electric
field, can also be viewed as a site of stored energy (see Section 36-4).
Sec. 32-1 ELECTROMOTIVE FORCE 791

fig. 32-2 Two batteries, a resistor,


and a motor, connected in a single-
loop circuit. It is given that GB > GA,

the floor to the shelf, does work on them. This energy is stored temporarily
as gravitational field energy. The balls roll slowly and uniformly along the
shelf, dropping from the right end into a cylinder full of viscous oil. They
sink to the bottom at constant speed, are removed by a trapdoor mechanism
not shown, and roll back along the floor to the left. The energy put into the
system by the man appears eventually as heat in the viscous fluid. The
energy supplied by the man comes from his own internal (chemical) energy.
The circulation of charges in Fig. 32-la will stop eventually if battery Bis not
charged; the circulation of bowling balls in Fig. 32-lb will stop eventually if
the man does not replenish his store of internal energy by eating.
Figure 32-2 shows a circuit containing two (ideal) batteries, A and B, a
resistor R, and an (ideal) electric motor employed in lifting a weight. The
batteries are connected so that they tend to send charges around the circuit
in opposite directions; the actual direction of the current is determined by B,
which supplies the larger potential difference. The energy transfers in this
~oott~ ·

Stored
electric
and
magnetic
field
energy
792 ELECTROMOTIVE FORCE AND CIRCUITS Chap. 32
The chemical energy in B is being steadily depleted, the energy appearing in
the three forms shown on the right. Battery A is being "charged" while
battery B is being discharged. Again, the electric and magnetic fields that
surround the circuit act as an intermediary.
It is part of the definition of an emf that the energy transfer process be reversible,
at least in principle. The student will recall that a reversible process is one that
passes through equilibrium states; its course can be reversed by making an infinites-
imal change in the environment of the system; see Section 25-2. A battery, for
example, can either be on charge or discharge; a generator can be driven mechanically,
producing electric energy, or it can be operated backward as a motor. The (rever-
sible) energy transfers here are
electrical ;_:::: chemical
and electrical ;_:::: mechanical.
Joule heating is an electric energy transfer that is not reversible. We can easily heat
a conductor by supplying electric energy to it, but it is not possible to set up a
current in a closed copper loop by heating the loop uniformly. Because of this lack
of reversibility, we do not associate an emf with Joule heating.

32-2 Calculating the Current


In a time dt an amount of energy given by i 2 R dt will appear in the resistor
of Fig. 32- la as Joule heat. During this same time a charge dq ( = i dt) will
have moved through the seat of emf, and the seat will have done work on this
charge (see Eq. 32-1) given by
dW = 8dq = 8i dt.
From the conservation of energy principle, the work done by the seat must
equal the Joule heat, or
8i dt = i2R dt.
Solving for i, we obtain
i = 8/R.
vVe can also derive Eq. 32-2 by considering that if electric potential is to
have any meaning a given point can have only one value of potential at any
given time. If ,ve start at any point in the circuit of Fig. 32-la and, in imagi-
nation, go around the circuit in either direction, adding up algebraically the
changes in potential that we encounter, we must arrive at the same potential
when we return to our starting point. In other words, the algebraic sum of the
changes in potential encountered in a complete traversal of the circuit must be
zero.
In Fig. 32-la let us start at point a, whose potential is Va,* and traverse
the circuit clockwise. In going through the resistor, there is a change in
potential of -iR. The minus sign shows that the top of the resistor is higher
* The actual value of Va depends on assumptions made in the definition of potential
(as described in Section 29-1). The numerical value of Va is not important because, as in
most electric circuit situations, we are concerned here with di.fferences of potential. Point
a in Fig. 32-la (or any other single point in that figure) could be connected to ground
(symbol~) and assigned the potential Va = O, following a common practice.
Sec. 32-3 OTHER SINGLE-LOOP CIRCUITS 793
in potential than the bottom, which must be true, because positive charge
carriers move of their own accord from high to low potential. As we tra-
verse the battery from bottom to top, there is an increase of potential +e
because the battery does (positive) work on the charge carriers, that is, it
moves them from a point of low potential to one of high potential. Adding
the algebraic sum of the changes in potential to the initial potential Va must
yield the identical value Va, or
Va - iR +e= Va.
We write this as
-iR +e= 0,
which is independent of the value of Va and which asserts explicitly tha~ the
algebraic sum of the potential changes for a complete circuit traversal is zero.
This relation leads directly to Eq. 32-2.
These two ways to find the current in single-loop circuits, based on the
conservation of energy and on the concept of potential, are completely
equivalent because potential differences are defined in terms of work and
energy (Section 29-1). The statement that the sum of the changes in poten-
tial encountered in making a complete loop is zero is called Kirchhoff's second
rule; for brevity we call it the loop theorem. It must always be borne in mind
that this theorem is simply a particular way of stating the law of conserva-
tion of energy for electric circuits.
To prepare for the study of more complex circuits, let us examine the rules
for finding potential differences; these rules follow from the previous dis-
cussion. They are not meant to be memorized but rather to be so thoroughly
understood that it becomes trivial to re-derive them on each application.
1. If a resistor is traversed in the direction of the current, the change in
potential is -iR; in the opposite direction it is +iR.
2. If a seat of emf is traversed in the direction of the emf, the change in
potential is +e;in the opposite direction it is -e.

32-3 Other Single-Loop Circuits


Figure 32-3a shows a circuit which emphasizes that all seats of emf have
an intrinsic internal resistance r. This resistance cannot be removed-al-
though we would usually like to do so-because it is an inherent part of the
device. The figure shows the internal resistance rand the emf separately, al-
though, actually, they occupy the same region of space.
If we apply the loop theorem, starting at band going around clockwise, we
obtain

or +e - ir - iR = 0.
The student should compare these equations with Fig. 32-3 b, which shows
the changes in potential graphically. In writing these equations, note that
we traversed rand R in the direction of the current and e in the direction of the
794 ELECTROMOTIVE FORCE AND CIRCUITS Chap. 32

ii le :] it R

(a)

o------+- (:;
b r a R b
•I I ,N\ti
I I
• NNV'rI
I

I I
I I I
I I I I
I I I
.....
"' (:;
I I
0
>
I I
I
ro ir I
:p Va
c::
.....
(I)

0
c.. (,
f,- ir(=Vab)

vb

The seat of emf


The external _j
a
resistor li
b
(b)

Fig. 32-3 A single-loop circuit. The rectangular block is a seat of emf with internal re-
sistance r. (b) The same circuit is drawn for convenience as a straight line. Directly
below are shown the changes in potential that one encounters in traversing the circuit
clockwise from point b.

emf. The same equation follows if we start at any other point in the circuit
or if we traverse the circuit in a counterclockwise direction. Solving for i
gives e
i =--- (32-3)
R +r
~ Example I. Resistors in series. Resistors in series are connected so that there
is only one conducting path through them, as in Fig. 32-4. What is the equivalent
resistance R of this series combination? The equivalent resistance is the single re-
sistance R which, substituted for the series combination between the terminals ab,
will leave the current i unchanged.
Applying the loop theorem (going clockwise from a) yields
-iR1 - iR2 - iR3 +8=0
or 'I,=
. e •
R1 + R2 + R3
Sec. 32-4 POTENTIAL DIFFERENCES 795
i~
a R1

t
Fig. 32-4 Example 1. Three resistors are connected
in series between terminals a and b. lr.
b R3
---i

For the equivalent resistance R


. e
i =R
or R = R1 + R2 + R3. (32-4)
The extension to more than three resistors is clear. .....

32-4 Potential Differences


We often want to compute the potential difference between two points in a
circuit. In Fig. 32-3a for example, what is the relationship between the po-
tential difference Vab ( = Va - Vb) between points b and a and the fixed
circuit parameters e, r, and R? To find this relationship, let us start from
point band traverse the circuit to point a, passing through resistor R against
the current. If Vb and Va are the potentials at band a, respectively, we have

because we experience an increase in potential in traversing a resistor against


the current arrow. We rewrite this relation as

which tells us that Vab, the potential difference between points a and b, has
the magnitude iR and that point a is more positive than point b. Combining
this last equation with Eq. 32-3 yields
R
Vab = e · (32-5)
R +r
To sum up: To find the potential difference between any two points in a cir-
cuit start at one point and traverse the circuit to the other, following any
path, and add up algebraically the potential changes encountered. This
algebraic sum will be the potential difference. This procedure is similar to
that for finding the current in a closed loop, except that here the potential
differences are added up over part of a loop and not over the whole loop.
The potential difference between any two points can have only one value;
thus we must obtain the same answer for all paths that connect these points.
796 ELECTROMOTIVE FORCE AND CIRCUITS Chap. 32
If we consider two points on the side of a hill, the measured difference in
gravitational potential (that is, in altitude) between them is the same no
matter what path is followed in going from one to the other. In Fig. 32-3a
let us calculate Vab, using a path passing through the seat of emf. We have
Vb+ 8 - ir = Va
or (see also Fig. 32-3b)
Vab = Va - Vb= +0 - ir.
Again, combining with Eq. 32-3 leads to Eq. 32-5.
The terminal potential difference of the battery Vab, as Eq. 32-5 shows, is
less than 8 unless the battery has no internal resistance (r = 0) or if it is on
open circuit (R = oo); then Vab is equal to 8. Thus the emf of a device is
equal to its terminal potential difference when on open circuit .
.,_. Example 2. In Fig. 32-5a let 8 1 and 8 2 be 2.0 volts and 4.0 volts, respectively;
let the resistances r 1, r 2, and R be 1.0 ohm, 2.0 ohms, and 5.0 ohms, respectively.
What is the current?

(a)

---H b :,__

1---- Seat of emf 2----External resistor


c

------''>+<c--- Seat of emf 1 ~


M--
a c
o~'- . . . . - - - r - - - - - - + - - - - - - - - - + - - - - - - , - - - . - _ . _ _

-1
~
0
> -2
m
~ -3 =
iR Vbc
~ -------,,.--""'-- - - - _ ___±
-4

-5
(b)

Fig. 32-5 (a) A single-loop circuit. (b) The same circuit is shown schematically as a
straight line, the potential differences encountered in traversing the circuit clockwise from
point a being displayed directly below. In the lower figure the potential of point a was
assumed to be zero for convenience.
Sec. 32-4 POTENTIAL DIFFERENCES 797
Emfs 81 and 82 oppose each other, but because 82 is larger it controls the direction
of the current. Thus i will be counterclockwise. The loop theorem, going clockwise
from a, yields
-82 + ir2 + iR + ir1 + 81 = 0.
The student should check that the same result is obtained by going around counter-
clockwise. He ilhould also compare this equation carefully with Fig. 32-5b, which
shows the potential changes graphically.
Solving for i yields
. 82 - 81 , 4.0 volts - 2.0 volts
i= =
R + r 1 + r 2 5.0 ohms+ 1.0 ohm+ 2.0 ohms
= 0.25 amp.
It is not necessary to know in advance what the actual direction of the current is.
To show this, let us assume that the current in Fig. 32-5a is clockwise, an assumption
that we know is incorrect. The loop theorem then yields (going clockwise from a)
- 82 - ir2 - iR - ir1 + 81 = 0
. Si - 82
or 1, = .
+ +
R r1 r2
Substituting numerical values (see above) yields -0.25 amp for the current. The
minus sign tells us that the current is in the opposite direction from the one we have
assumed.
In more complex circuit problems involving many loops and branches it is often
impossible to know in advance the correct directions for the currents in all parts of
the circuit. We can assume directions for the currents at random. Those currents
for which positive numerical values are obtained will have the correct directions;
those for which negative values are obtained will be exactly opposite to the assumed
directions. In all cases the numerical values will be correct.
Example 3. What is the potential difference (a) between points b and a in Fig.
32-5a? (b) Between points a and c?
(a) For points a and b we start at band traverse the circuit to a, obtaining
Vab ( = Va - Vb) = -ir2 + 82 = -(0.25 amp)(2.0 ohms)+ 4.0 volts
= +3.5 volts.
Thus a is more positive than band the potential difference (3.5 volts) is less than the
emf (4.0 volts); see Fig. 32-5b.
(b) For points c and a, we start at c and traverse the circuit to a, obtaining
Vac ( = Va - Ve) = +8 1 + ir1 = +2.0 volts + (0.25 amp)(l.O ohm)
= +2.25 volts.
This tells us that a is at a higher potential than c. The terminal potential difference
of Si (2.25 volts) is larger than the emf (2.0 volts); see Fig. 32-5b. Charge is being
forced through 8 1 in a direction opposite to the one in which it would send charge if
it were acting by itself; if 8 1 is a storage battery, it is being charged at the expense of
82.
Let us test the first result by proceeding from b to a along a different path, namely,
through R, r1, and 81. We have
Vab = iR + ir1 + 81 = (0.25 amp)(5.0 ohms)
+ (0.25 amp)(l.O ohm) + 2.0 volts = +3.5 volts,
which is the same as the earlier result. ....
798 ELECTROMOTIVE FORCE AND CIRCUITS Chap. 32
32-5 Multiloop Circuits
Figure 32-6 shows a circuit containing two loops. For simplicity, we have
neglected the internal resistances of the batteries. There are two junctions,
b and d, and three branches connecting these junctions. The branches are
the left branch bad, the right branch bed, and the central branch bd. If the
emfs and the resistances are given, what are the currents in the various
branches?
We label the currents in the branches as i 1 , i 2 , and i 3 , as shown. Current
i 1 has the same value for any cross section of the left branch from b to d.
Similarly, i 2 has the same value everywhere in the right branch and i 3 in the
central branch. The directions of the currents have been chosen arbitrarily.
The careful reader will note that i 3 must point in a direction opposite to the
one we have shown. We have deliberately drawn it in wrong to show how
the formal mathematical procedures will always indicate this to us.
The three currents i 1 , i 2 , and i 3 carry charge either toward junction d or
away from it. Charge does not accumulate at junction d, nor does it drain
away from this junction because the circuit is in a steady-state condition.
Thus charge must be removed from the junction by the currents at the same
rate that it is brought into it. If we arbitrarily call a current approaching
the junction positive and the one leaving the junction negative, then

This equation suggests a general principle for the solution of multiloop


circuits: At any junction the algebraic sum of the currents must be zero. This
junction theorem is also known as Kirchhoff's first rule. Nate that it is simply
a statement of the conservation of charge. Thus our basic tools for solving
circuits are (a) the conservation of energy (seep. 793) and (b) the conserva-
tion of charge.
For the circuit of Fig. 32-6, the junction theorem yields only one relation-
ship among the three unknowns. Applying the theorem at junction b leads
to exactly the same equation, as the student should verify. To solve for
the three unknowns, we need two more independent equations; they can be
found from the loop theorem.
In single-loop circuits there is only one conducting loop around which to
apply the loop theorem, and the current is the same in all parts of this loop.

01
-----.0 ~
82
a b c

Fig. 32-6 A multiloop circuit.


R1 !i1 R3 ii3 R2 ti2

d
Sec. 32-5 MULTILOOP CIRCUITS 799

In multiloop circuits there is more than one loop, and the current in general
will not be the same in all parts of any given loop.
If we traverse the left loop of Fig. 32-6 in a counterclockwise direction,
the loop theorem gives
(32-6)
The right loop gives
(32-7)

These two equations, together with the relation derived earlier with the
junction theorem, are the three simultaneous equations needed to solve for
the unknowns i 1 , i 2 , and i 3 . Doing so yields

81 (R2 + R3) - 82R3 (32-8a)


R1R2 + R2R3 + R1R3
. 81R3 - 82(R1 + R3)
i2 = ' (32-8b)
R1R2 + R2R3 + R1R3

and (32-8c)

The student should supply the missing steps. Equation 32-8c shows that
no matter what numerical values are given to the emfs and to the resistances
the current i 3 will always have a negative value. This means that it will
always point up in Fig. 32-6 rather than down, as we deliberately assumed.
The currents i 1 and i 2 may be in either direction, depending on the particular
numerical values given.
The student should verify that Eqs. 32-8 reduce to sensible conclusions
in special cases. For R 3 = oo, for example, we find

. 81 - 82
i1 = i2 = and
R1 + R2
What do these equations reduce to for R 2 = oo?
The loop theorem can be applied to a large loop consisting of the entire
circuit abcda of Fig. 32-6. This fact might suggest that there are more equa-
tions than we need, for there are only three unknowns and we already have
three equations written in terms of them. However, the loop theorem yields
for this loop

which is nothing more than the sum of Eqs. 32-6 and 32-7. Thus this large
loop does not yield another independent equation. It will never be found in
solving multiloop circuits that there are more independent equations than
variables.
800 ELECTROMOTIVE FORCE AND CIRCUITS Chap. 32
..... Example 4. Resistors in parallel. Figure 32-7 shows three resistors connected
across the same seat of emf. Resistances across which the identical potential dif-
ference is applied are said to be in parallel. What is the equivalent resistance R of
this parallel combination? The equivalent resistance is that single resistance which,
substituted for the parallel combination between terminals ab, would leave the
current i unchanged.

Fig. 32-7 Example 4. Three re-


sistors are connected in parallel
betweeen terminals a and b.

~ b
i

The currents in the three branches are


.
i1 = - ,
v and
.
1,3
v
= -,
R1 R3
where Vis the potential difference that appears between points a and b. The total
current i is found by applying the junction theorem to an extended junction con-
sisting of the topmost horizontal wire in Fig. 32-7 .

'!,
. = i1. + i2. + 1,3. = v(1
-
R1
+ -R21 + -R31) .
If the equivalent resistance is used instead of the parallel combination, we have

. v
i=-·
R
Combining these two equations gives
1 1 1 1
(32-9)
R= R1 + R2 + R3.
This formula can easily be extended to more than three resistances. Note that the
equivalent resistance of a parallel combination is less than any of the resistances that
make it up. "illl

32-6 Measuring Currents and Potential Differences


A meter to measure currents is called an ammeter (or a milliammeter or micro-
arnmeter, depending on the size of the current to be measured). To determine the
current in a wire, it is necessary to break or cut the wire and to insert the ammeter,
so that the current to be measured passes through the meter (see Fig. 32-8).*
It is essential that the resistance RA of the ammeter be small compared to other
resistances in the circuit. Otherwise the act of inserting the meter will in itself change
* The meter must be connected so that the direction of current through it (assuming
positive charge carriers) is into the meter terminal marked +. Otherwise the meter will
deflect in a direction opposite to that intended.
Sec. 32-7 THE POTENTIOMETER 801

Fig. 32-8 An ammeter (A) is connected to read


the current in a circuit, and a voltmeter (V) is con-
nected to read the potential difference across re-
sistor R1.

the current to be measured. An ideal ammeter would have zero resistance. In


the circuit of Fig. 32-8 the required condition, assuming that the voltmeter is not
connected, is
RA << r + R1 + R2.
A meter to measure potential differences is called a voltmeter (or a millivoltmeter
or microvoltmeter). To find the potential difference between two points in a circuit,
it is necessary to connect one of the voltmeter terminals to each of the circuit points,
without breaking the circuit (see Fig. 32-8). *
It is essential that the resistance of the voltmeter Rv be large compared to any
circuit resistance across which the voltmeter is connected. Otherwise the meter will
itself constitute an important circuit element and will alter the circuit current and
the potential difference to be measured. An ideal voltmeter would have an infinite
resistance. In Fig. 32-8 the required condition is
Rv » R1
In measuring potential difference in electronic circuits, where the effective circuit
resistance may be of the order of 106 ohms or higher, it becomes necessary to use a
vacuum-tube voltmeter, which is an electron-tube device designed specifically to have an
extremely high effective resistance between its input terminals.

32-7 The Pot&ntiomehr


Figure 32-9 shows the rudiments of a potentiometer, which is a device for measuring
an unknown emf 8.,. The currents and emfs are marked as shown. Applying the
loop theorem to loop abed yields
-8., - ir + (io - i)JR = 0,
where io - i, by application of the junction theorem at a, is the current in resistor R
and f is the fraction of the uniform resistor R corresponding to the position of the sliding
contact at d. Solving for i yields
. iofR - 8.,
i = .
JR +r
This relation shows that if the sliding contact at d is moved to a position defined by
!:,,, where
iofxR = 8., (32-10)

the current i in the branch abed becomes zero. To balance the potentiometer in this way,
the sliding contact must be adjusted manually until the sensitive meter G reads zero.

* The voltmeter terminal marked + must be connected to the point of higher potential.
Otherwise the meter will deflect in a direction opposite to that intended.
802 ELECTROMOTIVE FORCE AND CIRCUITS Chap. 32
~
ex
b

it Fig. 32-9 Rudiments of a poten-


tiometer. The long resistor in the
a e center of the figure has resistance
i0- i R between points a and e

iol

The emf can be obtained from Eq. 32-10 if the current io is known becausef., can
be read from the calibrated markings along resistor R. However, it is standard
practice to replace 8., by a known standard emf Bs, and once again to adiust the
sliding contact to the zero-current condition. This yields-, assuming the current i 0
remains unchanged,
iof.R = e•.
Combining the last two equations yields
f,,
8., = Ba J.' (32-11)

which allows us to compare emfs with precision. Note that the internal resistance r
of the emf plays no role. In practice, potentiometers are conveniently packaged
units, containing a standard cell which, after calibration at the National Bureau of
Standards or elsewhere, serves as a convenient known standard seat of emf e•.
Switching arrangements for replacing the unknown emf by the standard and arrange-
ments for ascertaining that the current io remains constant are also incorporated
and the markings on resistor R, which is usually an array of several adjustable
resistors, are arranged to read 8., directly.

32-8 RC Circuits
The preceding sections dealt with circuits in which the circuit elements
were resistors and in which the currents did not vary with time. Here we
introduce the capacitor as a circuit element, which will lead us to the con-
cept of time-varying currents. In Fig. 32-10 let switch S be thrown to posi-
tion a. What current is set up in the single-loop circuit so formed? Let us
apply conservation of energy principles.
In time dt a charge dq ( = i dt) moves through any cross section of the
circuit. The work done by the seat of emf ( = e dq; see Eq. 32-1) must
equal the energy that appears as Joule heat in the resistor during time

Fig. 32-10 An RC circuit.


Sec. 32-8 RC CIRCUITS 803

dt ( = i 2 R dt) plus the increase in the amount of energy U that is stored in


the capacitor [ = dU = d(q 2 /2C); see Eq. 30-25]. In equation form

e dq = i 2 R dt + d(;;)
or e dq = i2R dt + !!_ dq.
c
Dividing by dt yields
dq ·2 q dq
e-=iR+--·
dt c dt
But dq/dt is simply i, so that this equation becomes

e= iR + i. (32-12)
c
This equation also follows from the loop theorem, as it must, since the
loop theorem was derived from the conservation of energy principle. Start-
ing from point x and traversing the circuit clockwise, we experience an in-
crease in potential in going through the seat of emf and decreases in potential
in traversing the resistor and the capacitor, or

e- iR - !!._ = O
c '
which is identical with Eq. 32-12.
We cannot immediately solve Eq. 32-12 because it contains two variables,
q and i, which, however, are related by
. dq
i = -· (32-13)
dt
Substituting for i into Eq. 32-12 gives
dq q
S=R-+-· (32-14)
dt c
Our task now is to find the function q(t) that satisfies this differential equa-
tion. Although this particular equation is not difficult to solve, we choose
to avoid mathematical complexity by simply presenting the solution, which is
q = CS(l - e-t!Rc). (32-15)
We can easily test whether this function q(t) is really a solution of Eq. 32-14 by
substituting it into that equation and seeing whether an identity results. Differ-
entiating Eq. 32-15 with respect to time yields

dq (- ") _ f, -t/RC
dt - i - Re · (32-16)

Substituting q (Eq. 32-15) and dq/dt (Eq. 32-16) into Eq. 32-14 yields an identity,
as the student should verify. Thus Eq. 32-15 is a solution of Eq. 32-14.
804 ELECTROMOTIVE FORCE AND CIRCUITS Chap. 32

<I)
.c 8
E
..Q
:::i
0
u
e
-~
E
C< 4

2 4 6 8 10
t, milliseconds
(a)

<I)
4
~
QJ
c.
E
-~
E
..... 2

4 6 8 10
t, milliseconds
(b)

Fig. 32-11 If, in Fig. 32-10, we assume that R = 2000 ohms, C = 1.0 µ,f, and e = 10
volts, then (a) shows the variation of q with t during the charging process and (b) the varia-
tions of i with t. The time constant is RC = 2.0 X 10-3 sec.

Figure 32-11 shows some plots of Eqs. 32-15 and 32-16 for a particular
case. Study of these plots and of the corresponding equations shows that
(a) at t = 0, q = 0 and i = S/R, and (b) as t ----,. ex:;, q ----,. Cs and i ----,. 0:
that is, the current is initially S/R and finally zero; the charge on the capaci-
tor plates is initially zero and finally CS.
Sec. 32-8 RC CIRCUITS 805
The quantity RC in Eqs. 32-15 and 32-16 has the dimensions of time
(since the exponent must be dimensionless) and is called the capacitative
time constant of the circuit. It is the time at which the charge on the capaci-
tor has increased to within a factor of (1 - e- 1 ) ( = 63%) of its equilibrium
value. To show this, we putt = RC in Eq. 32-15 to obtain
q = GS(l - e- 1) = 0.63GS.
Since GS is the equilibrium charge on the capacitor, corresponding tot ~ oo,
the foregoing statement follows .
.,_ Example 6. After how many time constants will the energy stored in the capaci-
tor in Fig. 32-10 reach one-half its equilibrium value?
The energy is given by Eq. 30-25, or
1 2
U = 2Cq'
the equilibrium energy U«) being (1/2C)(CS) 2 • From Eq. 32-15, we can write the
energy as
U = __!__ (CS)2(1 - e -t/Rc)2
2C
or U = U«i(l - e-t/Rc)2.
Putting U = !-U«) yields
! = (1 _ e-t/RC)2
and solving this relation for t yields finally
t = 1.22 RC = 1.22 time constants.
Figure 32-11 shows that if a resistance is included in the circuit the rate of increase
of the charge of a capacitor toward its final equilibrium value is delayed in a way
measured by the time constant RC. With no resistor present (RC = O), the charge
would rise immediately to its equilibrium value. Although we have shown that this
time delay follows from an application of the loop theorem to RC circuits, it is im-
portant that the student develop a physical understanding of the causes of the delay.
When switch S in Fig. 32-10 is closed on a, the resistor experiences instantaneously
an applied potential difference of S, and an initial current of S/R is set up. Initially,
the capacitor experiences no potential difference because its initial charge is zero,
the potential difference always being given by q/C. The flow of charge through the
resistor starts to charge the capacitor, which has several effects. First, the existence
of a capacitor charge means that there must now be a potential difference ( = q/C)
across the capacitor; this, in turn, means that the potential difference across the
resistor must decrease by this amount, since the sum of the two potential differences
must always equal S. This decrease in the potential difference across R means that
the charging current is reduced. Thus the charge of the capacitor builds up and the
charging current decreases until the capacitor is fully charged. At this point the
full emf Sis applied to the capacitor, there being no potential drop (i = 0) across the
resistor. This is precisely the reverse of the initial situation. The student should
review the derivations of Eqs. 32-15 and 32-16 and should study Fig. 32-11 with
the qualitative arguments of this paragraph in mind.
Assume now that the switch S in Fig. 32-10 has been in position a for a
time t such that t >> RC. The capacitor is then fully charged for all practical
purposes. The switch Sis then thrown to position b. How do the charge of
the capacitor and the current vary with time?
806 ELECTROMOTIVE FORCE AND CIRCUITS Chap. 32
With the switch S closed on b, there is no emf in the circuit and Eq. 32-12
for the circuit, with 8 = 0, becomes simply
iR + '!_ = 0. (32-17)
c
Putting i = dq/dt allows us to write, as the differential equation of the circuit
(compare Eq. 32-14),
R dq +i = 0. (32-18a)
dt c
The solution is
(32-18b)

as the student may readily verify by substitution, q0 being the initial charge
on the capacitor. The capacitative time constant RC appears in this ex-
pression for capacitor discharge as well as in that for the charging process
(Eq. 32-15). We see that at a time such that t = RC the capacitor charge
is reduced to q0 e- 1 , which is 37% of the initial charge q0 .
The current during discharge follows from differentiating Eq. 32-18b, or

i = dq = - qo e-t/RC (32-19)
dt RC

The negative sign shows that the current is in the direction opposite to that
shown in Fig. 32-10. This is as it should be, since the capacitor is dis-
charging rather than charging. Since q0 = CB, we can write Eq. 32-19 as

'l = _ ! e-t/RC
R '
in which 8/R appears as the initial current, corresponding to t = 0. This
is reasonable because the initial potential difference for the fully charged
capacitor is 8.
The behavior of the RC circuit of Fig. 32-10 during charge and discharge can be
studied with a cathode-ray oscilloscope. This familiar laboratory device can display
on its fluorescent screen plots of the variation of potential with time. Figure 32-12
shows the circuit of Fig. 32-10 with connections made to display (a) the potential
difference V c across the capacitor and (b) the potential difference V R across the
resistor as functions of time. V c and V Rare given by

and VR = (R)i,
the former being proportional to the charge and the latter to the current.
Figure 32-13 shows oscillograph plots of V c and V R that result when, in effect,
switch Sin Fig. 32-10 is thrown regularly back and forth between positions a and b,
being left in each position for a time equal to several time constants. Intervals
during which the charge is building up are labeled ch and those during which it is
decaying are labeled dis.
RC CIRCUITS 807

VR (= iR)

0
1 GI,
1_
~I:
1---c-r> Vc(=H

Fig. 32-12 The circuit of Fig. 32-10 with connections made to display the potential
variations across the resistor and the capacitor on a cathode ray oscilloscope.

a II II
(a)
rJ
• m ~

Ill II II
(b)
II II
I

(c)

Fig. 32-13 In Fig. 32-10 switch Sis thrown periodically, by electronic means, between
positions a and b. The variations with time of the potential differences across (a) the ca-
pacitor, and (b) the resistor are shown, as displayed on a cathode ray oscilloscope. (c) The
appearance of the screen when the oscilloscope is connected to display the sum of V R and
V c. (Courtesy E. K. Hege, Rensselaer Polytechnic Institute.)
808 ELECTROMOTIVE FORCE AND CIRCUITS Chap. 32
The "charge" intervals in plot a (see Eq. 32-15) are represented by

V = (~) q = 8(1 - e-t/RC)

and the "discharge" intervals (see Eq. 32-lSb) by

V = (~) q = f;e-t!RC.

Note that the current, as indicated by plot b, is in opposite directions during the
charge and discharge intervals, in agreement with Eqs. 32-16 and 32-19.
In plot c in Fig. 32-13 the oscillograph has been connected to show the algebraic
sum of plots a and b. According to the loop theorem this sum should equal 8 during
the charge intervals and should be zero during the discharge intervals, when the
battery is no longer in the circuit, that is,

VR + V c = 8 during charge (see Eq. 32-12)


V R + V c = 0 during discharge (see Eq. 32-17).
Plot c is in exact agreement with this expectation.

QUESTIONS

1. Does the direction of the emf provided by a battery depend on the direction of cur-
rent flow through the battery?
2. Figure 32-lb shows a gravitational analog of a single-loop electric circuit. Is the
source of "gravitational emf" in this figure reversible as far as energy exchanges are con-
cerned?
3. Discuss in detail the statement that the energy method and the loop theorem method
for solving circuits are perfectly equivalent.
4. It is possible to generate a 10,000-volt potential difference by rubbing a pocket comb
with wool. Why is this large voltage not dangerous when the much lower voltage pro-
vided by an ordinary electric outlet is very dangerous?
5. Devise a method for measuring the emf and the internal resistance of a battery.
6. A 25-watt, 110-volt bulb glows at normal brightness when connected across a bank
of batteries. A 500-watt, 110-volt bulb glows only dimly when connected across the same
bank. Explain.
7. Under what circumstances can the terminal potential difference of a battery exceed
its emf?
8. What is the difference between an emf and a potential difference?
9. In the flow of incompressible fluids, what are the analogies to (a) the loop theorem
and (b) the junction theorem?
10. Compare and contrast the formulas for the effective values of (a) capacitors and (b)
resistors, in series and in parallel.
11. Does the time required for the charge on a capacitor in an RC circuit to build up to a
given fraction of its equilibrium value depend on the value of the applied emf?
12. Devise a method whereby an RC circuit can be used to measure very high re-
sistances.
Chap. 32 PROBLEMS 809

PROBLEMS

1. A 5.0-amp current is set up in an external circuit by a 6.0-volt storage battery for


6.0 min. By how much is the chemical energy of the battery reduced?
/2: The current in a simple series circuit is 5 amp. When an additional resistance of 2
ohms is inserted, the current drops to 4 amp. What was the resistance of the original
circuit?
3. In Example 2 an ammeter whose resistance is 0.05 ohm is inserted in the circuit.
What per cent change in the current results because of the presence of the meter?
4. In Fig. 32-3a put 0 = 2.0 volts and r = 100 ohms. Plot (a) the current, and (b) the
potential difference across R, as functions of R over the range O to 500 ohms. Make both
plots on the same graph. (c) Make a third plot by multiplying together, for each value of
R, the two curves plotted. What is the physical significance of this plot?
.. · 5-: (a) In the circuit of Fig. 32-3a show that the power delivered to Ras Joule heat is a
maximum when R is equal to the internal resistance r of the battery. (b) Show that this
maximum power is P = 02 /4r.
6. Heat is to be generated in a 0.10-ohm resistor at the rate of 10 watts by connecting it
to a battery whose emf is 1.5 volts. (a) What is the internal resistance of the battery?
(b) What potential difference exists across the resistor?
7. (a) In Fig. 32-14 what value must R have if the current in the circuit is to be 0.001
amp? Take 0 1 = 2.0 volts, 0 2 = 3.0
volts, and r 1 = r2 = 3.0 ohms. (b) What
is the rate of Joule heating in R?
01 t
l
1e 2

8. A wire of resistance 5.0 ohms is con-


nected to a battery whose emf 0 is 2.0
volts and whose internal resistance is 1.0
ohm. In 2.0 min (a) how much energy
is transferred from chemical to electric R
form? (b) How much energy appears in
the wire as Joule heat? (c) Account for
the difference between (a) and (b). Fig. 32-14
9. In Fig. 32-6 calculate the potential
difierence between points c and d by as many paths as possible. Assume that 01 = 4.0
volts, 02 = 1.0 volt, R 1 = R 2 = 10 ohms, and Ra = 5 ohms.
10. In Fig. 32-5 calculate the potential difference between a and c by considering a path
that contains R and 82•
.. Jl. (a) In Fig. 32-15 what is the equivalent resistance of the network shown? (b) What
are the currents in each resistor? Put R1 ~ 100 ohms, R2 = Ra = 50 ohms, R4 = 75
ohms, and 0 = 6.0 volts.

Fig. 32-15 Fig. 32-16

12. In Fig. 32-16 imagine an ammeter inserted in the branch containing Ra. (a) What
will it read, assuming 0 = 5.0 volts, R 1 = 2.0 ohms, R 2 = 4.0 ohms, and Ra = 6.0 ohms?
(b) The ammeter and the source of emf are now physically interchanged. Show that the
810 ELECTROMOTIVE FORCE AND CIRCUITS Chap. 32
ammeter reading remains unchanged. This reciprocity relationship holds for any circuit
that contains only one source of emf.
13. Two batteries of emf e and internal resistance r are connected in parallel across a
resistor Ras in Fig. 32-20b. (a) For what value of R is the power delivered to the resistor
a maximum? (b) What is the maximum power?
14. By using only two resistance coils-singly, in series, or in parallel-a student is able
to obtain resistances of 3, 4, 12, and 16 ohms. What are the separate resistances of the
coils?
_/15: The Wheatstone bridge. In Fig. 32-17 Rs is to be adjusted in value until points
a and b are brought to exactly the same potential. (One tests for this condition by mo-
mentarily connecting a sensitive meter between a and b; if these points are at the same
potential, the meter will not deflect.) Show that when this adjustment is made the fol-
lowing relation holds:

Unknown resistors (Rx) can be measured in terms of standards (Rs) using this device,
which is called a Wheatstone bridge.

Fig. 32-17

16. If points a and bin Fig. 32-17 are connected by a wire of resistance r, show that the
current in the wire is
e(R, - Rx)
i = (R + 2r)(R, + Rx) + 2R,Rx 1

where e is the emf of the battery. Assume that R 1 and R2 are equal (R1 = R2 = R) and
that Ro equals zero. Is this formula consistent with the result of Problem 15?
,,...ft- Two resistors, R1 and R2, may be connected either in series or parallel across a
(resistanceless) battery with emf e. We desire the Joule heating for the parallel combina-
tion to be five times that for the series combination. If R1 equals 100 ohms, what is R2?
18. Four 100-watt heating coils are to be connected in all possible series-parallel combi-
nations and plugged into a 100-volt line. What different rates of heat dissipation are
possible?
19. What is the equivalent resistance between the terminal points x and y of the circuits
shown in (a) Fig. 32-18a, (b) Fig. 32-18b, and (c) Fig. 32-18c? Assume that the re-
sistance of each resistor is 10 ohms.
Chap. 32 PROBLEMS 811

x
(a) (b) (c)

Fig. 32-18

_20. In Fig. 32-19 find the current in each resistor and the potential difference between
a and b. Put lh = 6.0 volts, 82 = 5.0 volts, 83 = 4.0 volts, R1 = 100 ohms, and R 2 = 50
ohms.

(a)

(b)

Fig. 32-19 Fig. 32-20

21. You are given two batteries of emf 8 and internal resistance r. They may be con-
nected either in series or in parallel and are used to establish a current in a resistor R, as in
Fig. 32-20. Derive expressions for the current in R for both methods of connection.
Which connection yields the larger current, considering various values of R / r?
22. (a) In Fig. 32-21 what power appears as Joule heat in R 1? In R 2? In R 3 ? (b)
What power is supplied by t; 1? By 82? (c) Discuss the energy balance in this circuit. As-
sume that 81 = 3.0 volts, 82 = 1.0 volt, R 1 = 5.0 ohms, R2 = 2.0 ohms, and R 3 = 4.0
ohms.

Fig. 32-21
812 ELECTROMOTIVE FORCE AND CIRCUITS Chap. 32
23. For manual control of the current in a circuit, a student uses a parallel combination
of variable resistors of the sliding contact type, as in Fig. 32-22, with R 1 = 20R 2 . (a) What
procedure is used to adjust the current to the desired value? (b) Why is the parallel com-
bination better than a single-variable resistor?

;24'. Resistance measurement. A voltmeter (resistance Rv) and an ammeter (resistance


RA) are connected to measure a resistance R, as in Fig. 32-23a. The resistance is given by
R = V /i, where Vis the voltmeter reading and i is the current in the resistor R. Some of
the current registered by the ammeter (i') goes through the voltmeter so that the ratio of
the meter readings ( = V /i') gives only an apparent resistance reading R'. Show that
R and R' are related by
1 1 1
- =---·
R R' Rv
Note that if Rv » R, then R ::::'. R'.

R R

(a) (b)

Fig. 32-23

.25. Resistance measurement. If meters are used to measure resistance, they may also
be connected as they are in Fig. 32-23b. Again the ratio of the meter readings gives only
an apparent resistance R'. Show that R' is related to R by

R = R' - RA,
in which RA is the ammeter resistance. Note that if RA « R, then R ""R'.
26. In Fig. 32-8 assume that 8 = 5.0 volts, r = 2.0 ohms, R1 = 5.0 ohms, and R2 =
4.0 ohms. If RA = 0.10 ohm, what per cent error is made in reading the current? Assume
that the voltmeter is not present.
27. In Fig. 32-8 assume that 8 = 5.0 volts, r = 20 ohms, R1 = 50 ohms, and R2 =
40 ohms. If Rv = 1000 ohms, what per cent error is made in reading the potential dif-
ferences across R 1? Ignore the presence of the ammeter.
Chap. 32 PROBLEMS 813

,2.s-:·(~) Find the three currents in Fig. 32-24. (b) Find Vab· Assume that R1 = 1.0
ohm, R 2 = 2.0 ohms, e1 = 2.0 volts, and 02 = es = 4.0 volts.
R1 a
R1

63

G1

Gz
R1

Fig. 32-24
.,,.,r_,,.,,,..""'

,29: How many time constants must elapse before a capacitor in an RC circuit is charged
to within 1.0 per cent of its equilibrium charge?
30. rn the circuit of Fig. 32-25 let i1, i2, and is be the currents through resistors R1, R2,
and Rs, respectively, and let Vi, V2, Vs, and V c be the corresponding potential differences
across the resistors and across the capacitor C. (a) Plot qualitatively as a function of time
after switch Sis closed the currents and voltages listed above. (b) After being closed for
a large number of time constants, the switch S is now opened. Plot qualitatively as a
function of time after the switch is opened the currents and voltages listed above.

Rs

Fig. 32-25

31. Show that the units of RC are indeed time units, that is, that 1 ohm X 1 farad = 1
sec.
32. Prove that when switch S in Fig. 32-10 is thrown from a to b all the energy stored
in the capacitor is transformed into Joule heat in the resistor. Assume that the capacitor
is fully.charged before the switch is thrown.
~ A 3.0 X 106-ohm resistor and a 1.0-µf capacitor are connected in a single-loop
circuit with a seat of emf with e = 4.0 volts. At 1.0 sec after the connection is made, what
are the rates at which (a) the charge of the capacitor is increasing, (b) energy is being stored
in the capacitor, (c) Joule heat is appearing in the resistor, and (d) energy is being delivered
by the seat of emf?
The Magnetic Field
CHAPTER 33

33-1 The Magnetic Field


The science of magnetism grew from the observation that certain "stones"
(magnetite) would attract bits of iron. The word magnetism comes from the
district of Magnesia in Asia Minor, which is one of the places at which the
stones were found. Figure 33-1 shows a modern permanent magnet, the
lineal descendant of these natural magnets. Another "natural magnet" is
the earth itself, whose orienting action on a magnetic compass needle has
been known since ancient times.
In 1820 Oersted first discovered that a current in a wire can also produce
magnetic ~fleets, namely, that it can change the orientation of a compass
needle. We pointed out in Section 26-1 how this important discovery linked
the then separate sciences of magnetism and electricity. The magnetic ef-

Fig. 33-1 A permanent magnet. Lines


of magnetic induction leave the north pole
face, marked N, and enter the south pole
face on the other side of the air gap.

814
Sec. 33-1 THE MAGNETIC FIELD 815

Fig. 33-2 A research-type electromagnet showing


iron frame F, pole faces P, and coils C. The pole
faces are 12 in. in diameter. (Courtesy Varian
Associates.)

feet of a current in a wire can be intensified by forming the wire into a coil of
many turns and by providing an iron core. Figure 33-2 shows how this is
done in a large electromagnet of a type commonly used for research involving
magnetism.
We define the space around a magnet or a current-carrying conductor as
the site of a magnetic field, just as we defined the space near a charged rod
as the site of an electric field. The basic magnetic field vector B, which we
define in the following section, is called the magnetic induction: * it can be
represented by lines of induction, just as the electric field was represented by
lines of force. As for the electric field (see Section 27-3), the magnetic field
vector is related to its lines of induction in this way:

1. The tangent to a line of induction at any point gives the direction of B


at that point.
2. The lines of induction are drawn so that the number of lines per unit
cross-sectional area is proportional to the magnitude of the magnetic field
vector B. Where the lines are close together Bis large and where they are
far apart B is small.
As for the electric field, the field vector B is of fundamental importance,
the lines of induction simply giving a graphic representation of the way B
varies throughout a certain region of space.
The flux cf>s for a magnetic field can be defined in exact analogy with the
flux cf>E for the electric field, namely

cI>s = J
B·dS, (33-1)

in which the integral is taken over the surface (closed or open) for which cI>B
is defined.

* Magnetic field strength would be a more suitable name for B, but it has been usurped
for historical reasons by another vector connected with the magnetic field (see Section
37-7).
816 THE MAGNETIC FIELD Chap. 33
33-2 The Definition of B
This chapter is not concerned with the causes of the magnetic field; we
seek to determine (a) whether a magnetic field exists at a given point and
(b) the action of this field on charges moving through it. As for the electric
field, a particle of charge q0 serves as a test body. We assume that there
is no electric field present, which means that, neglecting gravity, no force
will act on the test body if it is placed at rest at the point in question.
Let us fire a positive test charge with arbitrary velocity v through a point
P. If a sideways deflecting force F acts on it, we assert that a magnetic field is
present at P and we define the magnetic induction B of this field in terms of
F and other measured quantities.
If we vary the direction of v through point P, keeping the magnitude of v
unchanged, we find, in general, that although F will always remain at right
angles to v its magnitude F will change. For a particular orientation of v
(and also for the opposite orientation -v) the force F becomes zero. We
define this direction as the direction of B, the specification of the sense of B
(that is, the way it points along this line) being left to the more complete
definition of B that we give below.
Having found the direction of B, we are now able to orient v so that the
test charge moves at right angles to B. We will find that the force Fis now
a maximum, and we define the magnitude of B from the measured magnitude
of this maximum force F .1., or
F ..L
B = --· (33-2)
qov
Let us regard this definition of B (in which we have specified its magnitude
and direction, but not its sense) as preliminary to the complete vector defini--
tion that we now give: If a positive test charge q0 is fired with velocity v through
a point P and if a (sideways) force F acts on the moving charge, a magnetic
induction B is present at point P, where B is the vector that satisfies the relation
F = q0 v x B, (33-3a)
v, q0 , and F being measured quantities. The magnitude of the magnetic
deflecting force F, according to the rules for vector products, is given by *
F = q0vB sin 8, (33-3b)
in which () is the angle between v and B.
Figure 33-3 shows the relations among the vectors. We see that F, being
at right angles to the plane formed by v and B, will always be at right angles
to v (and also to B) and thus will always be a sideways force. Equation 33-3a
is consistent with the observed facts that (a) the magnetic force vanishes as
v ~ 0, (b) the magnetic force vanishes if v is either parallel or antiparallel
to the direction of B (in these cases() = 0 or 180° and v xB = 0), and (c) if
v is at right angles to B (8 = 90°), the deflecting force has its maximum
value, given by Eq. 33-2, that is, F .1. = q0 vB.
* The student may wish to review Section 2-4, which deals with vector products.
Sec. 33-2 THE DEFINITION OF B 817
z

x
Fig. 33-3 Illustrating F = qov X B (Eq.
33-3a).

This definition of B is similar in spirit, although more complex, than the definition
of the electric field strength E, which we can cast into this form: If a positive test
charge q0 is placed at point P and if an (electric) force Facts on the stationary charge, an
electric field E is present at P, where E is the vector satisfying the relation
F = qoE,
q0 and F being measured quantities. In defining E, the only characteristic direction
to appear is that of the electric force FE which acts on the positive test body; the
direction of Eis taken to be that of FE. In defining B, two characteristic directions
appear, those of v and of the magnetic force FB; they prove always to be at right
angles.
In Fig. 33-4 a positive and a negative electron are created at point Pin a
bubble chamber. A magnetic field is perpendicular to the chamber, pointing
out of the plane of the figure (symbol O). * The relation F = q0 v x B
(Eq. 33-3a) shows that the deflecting forces acting on the two particles are
as indicated in the figure. These deflecting forces would make the tracks
deflect as shown.
The unit of B that follows from Eq. 33-3 is the (nt/coul)/(meter/sec).
This is given the special name weber/meter 2 , or tesla. Recalling that a
couljsec is an ampere,
1 nt 1 nt
1 weber/meter2 = - - - - - - - = - - -
coul (meter/sec) amp-m
An earlier unit for B, still in common use, is the gauss; the relationship is
1 weber/meter2 = 104 gauss.
The weber is used to measure cI>B, the flux of B; see Eq. 33-1.
The fact that the magnetic force is always at right angles to the direction
of motion Illeans that (for steady magnetic fields) the work done by this
* The syr.1bol © indicates a vector into the page, the X being thought of as the tail of
an arrow; the symbol O indicates a vector out of the page, the dot being thought of as the
tip of an arrow.
818 THE MAGNETIC FIELD Chap. 33

Fig. 33-4 A bubble chamber is a device for rendering visible, by means of small bubbles,
the tracks of charged particles that pass through the chamber. The figure is a photograph
taken with such a chamber immersed in a field of magnetic induction B and exposed to
radiations from a large cyclotron-like accelerator. The curved Vat point Pis formed by
a positive and a negative electron, which deflect in opposite directions in the magnetic
field. The spirals Sare the tracks of three low-energy electrons. (Courtesy E. 0. Lawrence
Radiation Laboratory, University of California.)

force on the particle is zero. For an element of the path of the particle of
length di, this work dW is FB·dl; dW is zero because FB and di are always
at right angles. Thus a static magnetic field cannot change the kinetic
energy of a moving charge; it can only deflect it sideways.
If a charged particle moves through a region in which both an electric field
and a magnetic field are present, the resultant force is found by combining
Eqs. 27-2 and 33-3a, or
(33-4)
This 1s sometimes called the Lorentz relation in tribute to H. A. Lorentz
Sec. 33-3 MAGNETIC FORCE ON A CURRENT 819

who did so much to develop and clarify the concepts of the electric and mag-
netic fields.
~ Example 1. A uniform field of magnetic induction B points horizontally from
south to north; its magnitude is 1.5 webers/meter 2 • If a 5.0-Mev proton moves
vertically downward through this field, what force will act on it?
The kinetic energy of the proton is
K = (5.0 X 106 ev)(l.6 X 10-19 joule/ev) = 8.0 X 10- 13 joule.
Its speed can be found from the relation K = imv 2 , or

_ /2K _ /(2)(8.0 X 10-13 joule) _ 3 1 101 t I


v - '\J m - '\j 1. 7 X 10 _ 27 kg - . X me ers sec.

Equation 33-3b gives


F = qvB sin() = (1.6 X 10-19 coul)(3.1 X 107 meters/sec)(l.5 webers/meter2)(sin 90°)
= 7 .4 X 10-12 nt.
The student can show that this force is about 4 X 10 14 times greater than the weight
of the proton.
The relation F = qv x B shows that the direction of the deflecting force is to the
east. If the particle had been negatively charged, the deflection would have been to
the west. This is predicted automatically by Eq. 33-3a if we substitute -e for qo. ~

33-3 Magnetic Force on a Current


A current is an assembly of moving charges. Because a magnetic field
exerts a sideways force on a moving charge, we expect that it will also exert a
sideways force on a wire carrying a current. Figure 33-5 shows a length l
of wire carrying a current i and placed in a field of magnetic induction B.
For simplicity we have oriented the wire so that the current density vector
j is at right angles to B.
The current i in a metal wire is carried by the free (or conduction) elec-
trons, n being the number of such electrons per unit volume of the wire.
The magnitude of the average force on one such electron is given by Eq.
33-3b, or, since o = 90°,
F' = q0vB sin () = evaB
where Vd is the drift speed. From the relation va = j/ne (Eq. 31-5),
.) B
F' = e ( _!__ B = J_ ·
ne n

Fig. 33-5 A wire carrying a


current i is placed at right an-
gles to a field of magnetic in-
duction B.
820 THE MAGNETIC FIELD Chap. 33
The length l of the wire contains nAl free electrons, Al being the volume of
the section of wire of cross section A that we are considering. The total force
on the free electrons in the wire, and thus on the wire itself, is
'B
F = (nAl)F' = nAlJ_·
n
Since jA is the current i in the wire, we have

F = ilB. (33-5)

The negative charges, which move to the right in the wire of Fig. 33-5,
are equivalent to positive charges moving to the left, that is, in the direction
of the current arrow. For such a positive charge the velocity v would point
to the left and the force on the wire, given by Eq. 33-3a (F = q0 v x B)
points up, out of the page. This same conclusion follows if we consider the
actual negative charge carriers for which v points to the right but q0 has a
negative sign. Thus by measuring the sideways magnetic force on a wire
carrying a current and placed in a magnetic field we cannot tell whether the
current carriers are negative charges moving in a given direction or positive
charges moving in the opposite direction.
Equation 33-5 holds only if the wire is at right angles to B. We can ex-
press the more general situation in vector form as

F = ii xB, (33-6a)

where I is a (displacement) vector that points along the (straight) wire in


the direction of the current. Equation 33-6a is equivalent to the relation
F = q0 v x B (Eq. 33-3a); either can be taken as a defining equation for B.
The student should note that the vector I in Fig. 33-5 points to the left
and that the magnetic force F ( = ii x B) points up, out of the page. This
agrees with the conclusion obtained by analyzing the forces that act on the
individual charge carriers.
If we consider a differential element of a conductor of length dl, the force
dF acting on it can be found, by analogy with Eq. 33-6a, from

dF = i di xB. (33-6b)

By integrating this formula in an appropriate way we can find the force F


on a nonlinear conductor.
~ Example 2. A wire bent as shown in Fig. 33-6 carries a current i and is placed
in a uniform field of magnetic induction B that emerges from the plane of the figure.
Calculate the force acting on the wire. The magnetic field is represented by lines of
induction, shown emerging from the page. The dots show that the sense of Bis up,
out of the page.
The force on each straight section, from Eq. 33-6a, has the magnitude

Fi= Fa= ilB


and points down as shown by the arrows in the figure.
Sec. 33-3 MAGNETIC FORCE ON A CURRENT 821
• • • • • • • 0 0
• 0 0 0

• 0 0 0
• • 0 0
• • 0 0

dF sin(} dF /B
• 0
• 0
I
\ • 0
• •
I

• • • 0
. ·M . -

(}
\

\\
d(J

R
. 0 0

0
___ j __ \ ~ - - - -
>I O
I~ _j
• • • • • 0
• • • • • • •
F1 F3
• 0 0 0
• • • 0
• 0
• 0 0

Fig. 33-6 Example 2.

A segment of wire of length dl on the arc has a force dF on it whose magnitude is


dF = iB dl = iB(R d8)
and whose direction is radially toward 0, the center of the arc. Only the downward
component of this force is effective, the horizontal component being canceled by an
oppositely directed component associated with the corresponding arc segment on the
other side of O. Thus the total force on the semicircle of wire about O points down
and is
F2 = frdF sin (J = L,r(iBR d8) sine = iBR fo'rsin edO = 2iBR.

The resultant force on the whole wire is

F = F 1 + F2 + F& = 2ilB + 2iBR = 2iB(l + R).


Notice that this force is the same as that acting on a straight wire of length 2l + 2R .
....
Figure 33-7 shows the arrangement used by Thomas, Driscoll, and Hipple at the
National Bureau of Standards to measure the magnetic induction provided by a
laboratory magnet such as that of Fig. 33-2. The rectangle is a coil of nine turns
whose width a and length bare about 10 cm and 70 cm, respectively. The lower end
of the coil is placed in the field of magnetic induction B and the upper end is hung
from the arm of a balance; B enters the plane of the figure at right angles.
An accurately known current i of about 0.10 amp is set up in the coil in the direction
shown, and weights are placed in the right-hand balance pan until the system is
balanced. The magnetic force F ( = il x B; see Eq. 33-6a) on the bottom leg of the
coil points upward, as shown in the figure. Equation 33-5 also shows that the force
on each wire at the bottom of the coil is iaB. Since there are nine wires, the total force
on the bottom leg of the coil is 9iaB. The forces on the vertical sides of the coil
( = ii x B) are sideways; because they are equal and opposite, they cancel and pro-
duce no effect.
After balancing the system, the experimenters reversed the direction of the current,
which changed the sign of all the magnetic forces acting on the coil. In particular, F
then pointed downward, which caused the rest point of the balance to move. A mass
a

,------------ -/B -,
Ix x x xilx x x I
x x x x
I F I
xxxxxxxxl
1 xxxx
lxxxxxxxxxxxxl
Ix x x x x x xix x x x x I
I I
Jxxxxxxxxxxxxl
L---------------~

Fig. 33-7 Apparatus used to measure B. The zero point of the balance is observed by
means of a light beam reflected from the mirror attached to the balanre beam.

822
Sec. 33-4 TORQUE ON A CURRENT LOOP 823
m of about 8.78 gm had to be added to the left balance pan to restore the original rest
point. The change in force when the current is reversed is 2F, and this must equal
the weight added to the left balance pan, or
mg= 2(9iaB) = 18iaB.
This gives

B = mg = (8.78 X 10-a kg)(9.80 meters/sec2) = 0.48 weber/meter2 = 4800 gauss.


18ai (18)(0.10 meter)(0.10 amp)
The Bureau of Standards' workers made this measurement with much more care than
these approximate numbers suggest. In one series of measurements, for example,
they found a magnetic induction of 4697.55 gauss.

33-4 Torque on a Current Loop


Figure 33-8 shows a rectangular loop of wire of length a and width b
placed in a uniform field of induction B, with sides 1 and 3 always normal
to the field direction. The normal nn' to the plane of the loop makes an
angle e with the direction of B.
Assume the current to be as shown in the figure. Wires must be provided to lead
the current into the loop and out of it. If these wires are twisted tightly together,
there will be no net magnetic force on the twisted pair because the currents in the
two wires are in opposite directions. Thus the lead wires may be ignored. Also,
some way of supporting the loop must be provided. Let us imagine it to be suspended
from a long string attached to the loop at its center of mass. In this way the loop
will be free to turn, through a small angle at least, about any axis through the center
of mass.

The net force on the loop is the resultant of the forces on the four sides of
the loop. On side 2 the vector I points in the direction of the current and has
the magnitude b. The angle between I and B for side 2 (see Fig. 33-Sb) is

F1 F1
x x x x x x x x x x x x
x x x x x x x x x x x x
1 n'
x x x x x x
~ x x

xxxx xxxxxxxxx n

:::Lxx
x x x x x x 8x xxx:::
xxxx
3
B
Fa B
a
(a) (b)

Fig. 33-8 A rectangular coil carrying a current i is placed in a uniform external magnetic
field.
824 THE MAGNETIC FIELD Chap. 33
90° - 8. Thus the magnitude of the force on this side is
F 2 = ibB sin (90° - 8) = 1:bB cos 8.
From the relation F = ii x B (Eq. 33-6a), we find the direction of F2 to be
out of the plane of Fig. 33-Sb. The student can show that the force F 4 on
side 4 has the same magnitude as F 2 but points in the opposite direction.
Thus F 2 and F 4 , taken together, have no effect on the motion of the loop.
The net force they provide is zero, and, since they have the same line of
action, the net torque due to these forces is also zero.
The common magnitude of F 1 and F 3 is iaB. These forces, too, are op-
positely directed so that they do not tend to move the coil bodily. As Fig.
33-Sb shows, however, they do not have the same line of action if the coil is
in the position shown; there is a net torque, which tends to rotate the coil
clockwise about the line xx'. The coil can be supported on a rigid axis that
lies along xx', with no loss of its freedom of motion. This torque can be
represented in Fig. 33-Sb by a vector pointing into the figure at point x' or
in Fig. 33-Sa by a vector pointing along the xx' axis from right to left.
The magnitude of the torque T' is found by calculating the torque caused
by F 1 about axis xx' and doubling it, for F 3 exerts the same torque about this
axis that F 1 does. Thus
r' = 2(iaB) (~) (sin 8) = iabB sin 0. ~
This torque acts on every turn of the coil. If there are N turns, the torque
on the entire coil is
r = Nr' = NiabB sin() = NiAB sin O, (33-7)
in which A, the area of the coil, is substituted for ab.
This equation can be shown to hold for all plane loops of area A, whether
they are rectangular or not. A torque on a current loop is the basic operating
principle of the electric motor and of most electric meters used for measuring
current or potential difference.
9ll- Example 3. A galvanometer. Figure 33-9 shows the rudiments of a galvanometer,
which is a device used to measure currents. The coil is 2.0 cm high and 1.0 cm wide;
it has 250 turns and is mounted so that it can rotate about a vertical axis in a uniform
radial magnetic field with B = 2000 gauss. A spring Sp provides a countertorque
that cancels out the magnetic torque, resulting in a steady angular deflection <J, cor-
responding to a given current i in the coil. If a current of 1.0 X 10-4 amp produces
an angular deflection of 30°, what is the torsional constant K of the spring (see Eq.
15-21)'?
Equating the magnetic torque to the torque caused by the spring (see Eq. 33-7)
yields
r = NiAB sin() = K<f,
or
NiAB sin8
K=----
<f,
(250)(1.0 X 10-4 amp)(2.0 X 10-4 meter2)(0.20 weber/meter 2)(sin 90°)
30°
= 3.3 X 10-s nt-m/deg.
Sec. 33-4 TORQUE ON A CURRENT LOOP 825
Scale

Fig. 33-9 Example 3. The


elements of a galvanometer,
showing the coil, the helical
spring Sp, and pointer P.

Note that the normal to the plane of the coil (that is, the pointer P) is always at
right angles to the (radial) magnetic field so that() = 90°. .....

A current loop orienting itself in an external magnetic field reminds us of


the action of a compass needle in such a field. One face of the loop behaves
like the north pole of the needle; * the other face behaves like the south pole.
Compass needles, bar magnets, and current loops can all be regarded as
magnetic dipoles. We show this here for the current loop, reasoning entirely
by analogy with electric dipoles.
A structure is called an electric dipole if (a) when placed in an external
electric field it experiences a torque given by Eq. 27-11,

T = p xE, (33-8)

where pis the electric dipole moment, and (b) it sets up a field of its own at
distant points, described qualitatively by the lines of force of Fig. 29-10
and quantitatively by Eq. 29-11. These two requirements are not in-
dependent; if one is fulfilled, the other follows automatically.
The magnitude of the torque described by Eq. 33-8 is

r = pE sin 0, (33-9)

where e is the angle between p and E. Let us compare this with Eq. 33-7,
the expression for the torque on a current loop:

r = (NiA)B sin 0. (33-7)


* The north pole of a compass needle is the end that points toward the geographic
north.
826 THE MAGNETIC FIELD Chap. 33
In each case the appropriate field (E or B) appears, as does a term sin 0.
Comparison suggests that NiA in Eq. 33-7 can be taken as the magnetic
dipole momentµ, corresponding top in Eq. 33-9, or
µ = NiA. (33-10)
Equation 33-7 suggests that we write the torque on a current loop in vector
form, in analogy with Eq. 33-8, or
T = µ. xB. (33-11)
The magnetic dipole moment of the loop µ. must be taken to lie along the axis
of the loop; its direction is given by the following rule: Let the fingers of the
right hand curl around the loop in the direction of the current; the extended
right thumb will then point in the direction ofµ.. The student should check
carefully that, if µ. is defined by this rule and Eq. 33-10, Eq. 33-11 cor-
rectly describes in every detail the torque acting on a current loop in an
external field (see Fig. 33-8).
Since a torque acts on a current loop, or other magnetic dipole, when it is
placed in an external magnetic field, it follows that work (positive or nega-
tive) must be done by an external agent to change the orientation of such a
dipole. Thus a magnetic dipole has potential energy associated with its
orientation in an external magnetic field. This energy may be taken to be
zero for any arbitrary position of the dipole. By analogy with the assump-
tion made for electric dipoles in Section 27-6, we assume that the magnetic
energy U is zero when µ. and B are at right angles, that is, when O = 90°.
This choice of a zero-energy configuration for U is arbitrary because we are
interested only in the changes in energy that occur when the dipole is rotated.
The magnetic potential energy in any position Ois defined as the work that
an external agent must do to turn the dipole from its zero-energy position
(0 = 90°) to the given position 0. Thus

U =
()
f r do =
()
f NiAB sin Odo = µB
i() sin Odo = -µB cos o,
.Jgoo .Jgoo goo
in which Eq. 33-7 is used to substitute for r. In vector symbolism this rela-
tion can be written as
U = -µ.·B, (33-12)
which is in perfect correspondence with Eq. 27-13, the expression for the
energy of an electric dipole in an external electric field,
u= -p·E.
.._ Example 4. A circular coil of N turns has an effective radius a and carries a
current i. How much work is required to turn it in an external magnetic field B from
a position in which 8 equals zero to one in which 8 equals 180°? Assume that N = 100,
a = 5.0 cm, i = 0.10 amp, and B = 1.5 webers/meter2 •
The work required is the difference in energy between the two positions, or, from
Eq. 33-12,
W = Uo= 1so 0 - Uo=o = (-µB cos 180°) - (-µB cos 0) = 2µB.
Sec. 33-5 THE HALL EFFECT 827
Butµ = NiA, so that
W = 2NiAB = 2N i('1ra 2)B

= (2)(100)(0.10 amp)(1r)(5 X 10-2 meter) 2(1.5 webers/meter2) = 0.24 joule. ....

33-5 The Hall Effect


In 1879 E. H. Hall devised an experiment that gives the sign of the charge
carriers in a conductor; seep. 771. Figure 33-10 shows a flat strip of copper,
carrying a current i in the direction shown. As usual, the direction of the
current arrow, labeled i, is the direction in which the charge carriers would
move if they were positive. The current arrow can represent either positive
charges moving down (as in Fig. 33-lOa) or negative charges moving up
(as in Fig. 33-lOb). The Hall effect can be used to decide between these
two possibilities.
A field of magnetic induction B is set up at right angles to the strip by
placing the strip between the polefaces of an electromagnet. This field
exerts a deflecting force F on the strip (given by ii x B), which points to the
right in the figure. Since the sideways force on the strip is due to the side-
ways forces on the charge carriers (given by qv xB), it follows that these
carriers, whether they are positive or negative, will tend to drift toward the
right in Fig. 33-10 as they drift along the strip, producing a transverse Hall
potential d(fferenr;e V xy between points such as x and y. The sign of the
charge carriers is determined by the sign of this Hall potential difference.
If the carriers are positive, y will be at a higher potential than x; if they are
negative, y will be at a lower potential than x. Experiment shows that in
metals the charge carriers are negative.

x x x x

x x

:,1 y x
(B :,1 y x
(B

x x x x

x x x x

(a) (b)

Fig. 33-10 A current i is set up in a copper strip placed in a field of magnetic induction
B, assuming (a) positive carriers and (b) negative carriers.
828 THE MAGNETIC FIELD Chap. 33
To analyze the Hall effect quantitatively, let us use the free-electron model of a
metal, the same model used in Section 31-4 to explain resistivity. The charge carriers
Mn be assumed to move along the conductor with a certain constant drift speed l'a·
The magnetic deflecting force that causes the moving charge carriers to drift toward
the right edge of the strip is given by qva xB (see Eq. 33-3a).
The charge carriers do not build up without limit on the right edge of the strip
because the displacement of charge gives rise to a transverse Hall electric .field Ey,
which acts, inside the conductor, to oppose the sideways drift of the carriers. This
Hall electric field is another manifestation of the Hall potential difference and is
related to it by
EH= Vxy/d.
Eventually an equilibrium is reached in which the sideways magnetic deflecting
force on the charge carriers is just canceled by the oppositely directed electric force
qEy caused by the Hall electric field, or
qEH + qva x B = 0,
which can be written
EH= -va xB. (33-13)
This equation shows explicitly that if EH and B are measured va can be determined
both in magnitude and direction; given the direction of va, the sign of the charge
carriers follows at once, as Fig. 33-10 shows.
The number of charge carriers per unit volume (n) can also be found from Hall effect
measurements. If we write Eq. 33-13 in terms of magnitudes, for the case in which
va and B are at right angles, we obtain EH = vaB. Combining this with Eq. 31-5
(vd = j/ne) leads to
j jB
EH= -B or n=--· (33-14)
ne eEH
The agreement between experiment and Eq. 33-14 is rather good for monovalent
metals, as Table 33-1 shows.

Table 33-1

NUMBER OF CONDUCTION ELECTRONS PER UNIT VOLUME

Based on Hall Calculated, Assuming


Metal Effect Data, One Electron/Atom,
1022;cm3 1022;cm3

Li 3.7 4.8
Na 2.5 2.6
K 1.5 1.3
Cs 0.80 0.85
Cu 11 8.4
Ag 7.4 6.0
Au 8.7 5.9

For nonmonovalent metals, for iron and similar magnetic materials, and for so-
called semiconductors such as germanium, the simple interpretation of the Hall effect
in terms of the free-electron model is not valid. A theoretical interpretation of the
Sec. 33-6 CIRCULATING CHARGES 829
Hall effect based on modern quantum physics gives a reasonable agreement with
experiment in all cm-es.

.... Example 5. A copper strip 2.0 cm wide and 1.0 mm thick is placed in a magnetic
field with B = 1.5 webers/meter 2 , as in Fig. 33-10. If a current of 200 amp is set
up in the strip, what Hall potential difference appears across the strip?
From Eq. 33-14,
jB
EH=-,
ne
Vxy . i i
but EH= - - and J = - = -,
d A dh

where h is the thickness of the strip. Combining these equations gives

V xy = iB = (200 amp)(l.5 webers/meter 2)


neh (8.4 X 1028 /meter 3)(1.6 X 10- 19 coul)(l.O X 10-3 meter)

= 2.2 X 10-5 volt = 22µv.


These potential differences are not large. See p. 773 for the calculation of n. .....

33-6 Circulating Charges


Figure 33-11 shows a negatively charged particle introduced with velocity
v into a uniform field of magnetic induction B. We assume that vis at right
angles to B and thus lies entirely in the plane of the figure. The relation
F = qv x B (Eq. 33-3a) shows that the particle will experience a sideways
deflecting force of magnitude qvB. This force will lie in the plane of the
figure, which means that the particle cannot leave this plane.
This reminds us of a stone held by a rope and whirled in a horizontal
circle on a smooth surface. Here, too, a force of constant magnitude, the
tension in the rope, acts in a plane and at right angles to the velocity. The
charged particle, like the stone, also moves with constant speed in a circular

x x x x x x x x x

x x x x x x x x

x x x x x x

x x x x x x x x

x x x
~ x x
x r x x

Fig. 33-11 A charge -q circu- x x x x ~x x x x x x


culates at right angles to a uniform x x x x x x x x x x x x
magnetic field.
x x x x x x

x x x x x x
)""B
x x x x x x

x x x x x x

x x x x x x x x x x x x
830 THE MAGNETIC FIELD Chap. 33
path. From ~ewton's second law we have
mv 2 mv
qvB = - or r = -,
r qB
which gives the radius of the path. The three spirals in Fig. 33-4 show rela-
tively low-energy electrons in a bubble chamber. The paths are not circles
because the electrons lose energy by collisions in the chamber as they move.
The angular velocity w is given by v/r or, from Eq. 33-15,
v qB
w = - = -,
r m

The frequency "measured, say, in rev/sec, is given by


w qB
"= - = --· (33-16)
21r 21rm

Note that" does not depend on the speed of the particle. Fast particles move
in large circles (Eq. 33-1.5) and slow ones in small circles, but all require the
same time T (the period) to complete one revolution in the field.
The frequency " is a characteristic frequency for the charged particle in
the field and may be compared to the characteristic frequency of a swinging
pendulum in the earth's gravitational field or to the characteristic frequency
of an oscillating mass-spring sy::;tem. It is sometimes called the cyclotron
frequency of the particle in the field because particles circulate at this fre-
quency in the cyclotron.
~ Example 6. A 10-ev electron is circulating in a plane at right angles to a uniform
field of magnetic induction of 1.0 X 10- 4 weber/meter 2 ( = 1.0 gauss).
(a) What is its orbit radius?

~2:.
The velocity of an electron whose kinetic energy is K can be found from

v=

The student should verify that this yields 1.9 X 10 6 meters/sec for v. Then, from
Eq. 33-15,
r= ~ = (9.1 X 10- 31 kg)(l.9 X 106 meters/sec) = O.ll meter = 11 cm.
qB (1.6 X 10-19 coul)(l.O X 10-4 weber/meter 2)
(b) What is the cyclotron frequency? From Eq. 33-16,
" = qB = (1.6 X 10- 19 coul)(l.O X 10-4 weber/meter 2 ) = 2 .8 X 106 rev/sec.
21rm (21r)(9.1 X 10-3 1 kg)
(c) What is the period of revolution T?

T = v1 = 2.8 X 101 rev/sec = 3·6 X 10 -7 sec.


6

Thus an electron requires 0.36 µsec to make 1 revolution in a LO-gauss field.


(d) What is the direction of circulation as viewed by an observer sighting along
the field?
In Fig. 33-11 the magnetic force qv x B must point radially inward, since it provides
the centripetal force. Since B points into the plane of the paper, v would have to
Sec. 33-7 THE CYCLOTRON 831

point to the left at the position shown in the figure if the charge q were positive.
However, the charge is an electron, with q = - e, which means that v must point to
the right. Thus the charge circulates clockwise as viewed by an observer sighting in
the direction of B. _,..

33-7 The Cyclotron


The cyclotron, first put into operation by Ernest Lawrence (1902-1958) in
1932, accelerates charged particles, such as protons or deuterons, * to high
energies so that they can be used in atom-smashing experiments. Figure
33-12 shows the University of Pittsburgh cyclotron.

Fig. 33-12 The University of Pittsburgh cyclotron. Note vacuum chamber V, magnet
frame F, magnetic pole faces P, magnet coils C, and the deuteron beam emerging into the
air of the laboratory. The rule is 6 ft long. (Courtesy A. J. Allen.)

In an ion source at the center of the cyclotron molecules of deuterium


are bombarded with electrons whose energy is high enough (say 100 ev) so
that many positive ions are formed during the collisions. Many of these ions
are free deuterons, which enter the cyclotron proper through a small hole
in the wall of the ion source and are available to be accelerated.
* Deuterons are the nuclei of heavy hydrogen.
832 THE MAGNETIC FIELD Chap. 33
The cyclotron uses a modest potential difference for accelerating (say 105
volts), but it requires the ion to pass through this potential difference a
number of times. To reach 10 Mev with 105 volts accelerating potential
requires 100 passages. A magnetic field is used to bend the ions around so
that they may pass again and again through the same accelerating potential.
Figure 33-13 is a top view of the part of the cyclotron that is inside the
vacuum tank marked V in Fig. 33-12. The two D-shaped objects, called
dees, are made of copper sheet and form part of an electric oscillator which
establishes an accelerating potential difference across the gap between the
dees. The direction of this potential difference is made to change sign some
millions of times per second.
The dees are immersed in a magnetic field (B 1.6 webers/meter2 ) whose r-..1

direction is out of the plane of Fig. 33-13. The field is set up by a large
electromagnet, marked Fin Fig. 33-12. Finally, the space in which the ions
move is evacuated to a pressure of about 10- 5 mm-Hg. If this were not
done, the ions would continually collide with air molecules.
Suppose that a deuteron, emerging from the ion source, finds the dee that
it is facing to be negative; it will accelerate toward this dee and will enter it.
Once inside, it is screened from electrical forces by the metal walls of the
<lees. The magnetic field is not screened by the <lees so that the ion bends in
a circular path whose radius, which depends on the velocity, is given by
Eq. 33-15, or mv
r = --·
qB
After a time t 0 the ion emerges from the dee on the other side of the ion
source. Let us assume that the accelerating potential has now changed sign.
Thus the ion again faces a negative dee, is further accelerated, and again
describes a semicircle, of somewhat larger radius (see Eq. 33-15), in the dee.

Beam

Def le ctor
plate \
t: "Dee"

- ,.
r---
I
/ / i\
..., ~

"' ~
.8
,.. :: '"'''....
·c3
<J)

v ····· ·····
.....
0
•·

: L_
:

~ :; ""Dee"
= ········

Fig. 33-13 The elements of a cyclotron showing the ion source Sand the dees. The de-
flector plate, held at a suitable negative potential, deflects the particles out of the dee
system.
Sec. 33-7 THE CYCLOTRON 833

The time of passage through this dee, however, is still to. This follows because
the period of revolution T of an ion circulating in a magnetic field does not
depend on the speed of the ion; see Eq. 33-16. This process goes on until
the ion reaches the outer edge of one dee where it is pulled out of the system
by a negatively charged deflector plate.
The key to the operation of the cyclotron is that the characteristic fre-
quency v at which the ion circulates in the field must be equal to the fixed
frequency v 0 of the electric oscillator, or
V = VO,

This resonance condition says that if the energy of the circulating ion is to
increase energy must be fed to it at a frequency v 0 that is equal to the
natural frequency v at which the ion circulates in the field. In the same
way we feed energy to a swing by pushing it at a frequency equal to the
natural frequency of oscillation of the swing.
From Eq. 33-16 (v = qB/21rm), we can rewrite the resonance equation as

qB
- = VO, (33-17)
21rm
Once we have selected an ion to be accelerated, q/m is fixed; usually the
oscillator is designed to work at a single frequency vo. We then "tune" the
cyclotron by varying B until Eq. 33-17 is satisfied and an accelerated beam
appears.
The energy of the particles produced in the cyclotron depends on the radius
R of the dees. From Eq. 33-15 (r = mv/qB), the velocity of a particle cir-
culating at this radius is given by
qBR
v = --·
m
The kinetic energy is then
· q2B2R2
K = !mv 2 (33-18)
2m
~ Example 7. The University of Pittsburgh cyclotron has an oscillator frequency
of 12 X 106 cycles/sec and a dee radius of 21 in. (a) What value of magnetic induc-
tion B is needed to accelerate deuterons?
From Eq. 33-17, v 0 = qB/21rm, so that
B = 21rvom = (21r)(12 X 10 6/sec)(3.3 X 10- 27 kg) = 1. 6 webers/meter2.
q 1.6 X 10- 19 coul
Note that the deuteron has the same charge as the proton but (very closely) twice the
mass.
(b) What deuteron energy results?
From Eq. 33-18,
K = q2B 2R 2 = (1.6 X 10-19 coul)2(1.6 webers/meter2)2(21 X 0.0254 meter)2
2m (2)(3.3 X 10-27 kg)

-- (2 .8 X 10 -12 JOU
. 1e) ( 1. X 1 ev ) -
10 _ 19 joule - 17 Mev.
T

6
The cyclotron fails to operate at high energies because one of its assumptions, that
the frequency of rotation of an ion circulating in a magnetic field is independent of its
834 THE MAGNETIC FIELD Chap. 33
speed, is true only for speeds much less than that of light. As the particle speed
increases, we must use the relativistic mass min Eq. 33-16. The relativistic mass
increases with velocity (Eq. 8-20) so that at high enough speeds 11 decreases with
velocity. Thus the ions get out of step with the electric oscillator, and eventually
the energy of the circulating ion stops increasing.
Another difficulty associated with the acceleration of particles to high energies is
that the size of the magnet that would be required to guide such particles in a circular
path is very large. For a 30-Gev proton, for example, in a field of 15,000 gauss the
radius of curvature is 65 meters. A magnet of the cyclotron type of this size (about
430 ft in diameter) would be prohibitively expensive. Incidentally, a 30-Gev proton
has a speed equal to 0.99998 that of light.
Both the relativistic and the economic limitations have been removed by tech-
niques that can be understood in terms of Eq. 33-17 in which mis now taken to be
the relativistic mass, given by Eq. 8-20, or

mo
m= '
Vl - (v/c) 2

v being the speed of the particle and c being that of light.


As the particle speed increases, the relativistic mass m also increases. To maintain
the equality in Eq. 33-17, and thus insure resonance, one may decrease the oscilla-
tor frequency 11 0 as the particle (assumed to be a proton) accelerates in such a way
that the product 11 0m remains constant. Accelerators that use this technique are
called synchrocyclotrons.
To remove the magnet cost limitation one can vary both B and 110 in a cyclic
fashion in such a way that not only is Eq. 33-17 satisfied at all times but the particle
orbit radius remains constant during the acceleration process. This permits the
use of an annular (or ring-shaped) magnet, rather than the conventional cyclotron
type, at great saving in cost. With the two variables Band 110 at our disposal, it is
possible to preserve two equalities during the acceleration process, one being Eq.
33-17 and the other being the relation
v = woRo = (21r11o)Ro
in which Ro is the desired (fixed) orbit radius. Accelerators that use this tech-
niql\e are called synchrotrons. Table 33-2 shows some characteristics of the accel-

Table 33-2
THE CERN PROTON SYNCHROTRON

Orbit diameter 560 ft


Vacuum chamber cross section 5.5 in. X 2.7 in.
Maximum magnetic field 1.4 webers/meter2
Frequency range per cycle 7 me/sec
Pulse rate 20/min
Maximum proton energy 28Gev
Energy gain per cycle 54 kev
Distance traveled by a proton 5 X 104 miles
Protons per pulse 1011
Cost $28 x 106
Sec. 33-8 THOMSON'S EXPERIMENT 835

erator built at Geneva, Switzerland, by the European Council for Nuclear Research
(CERN), and embodying these principles.

33-8 Thomson's Experiment


In 1897 J. J. Thomson, working at the Cavendish Laboratory in Cam-
bridge, measured the ratio of the charge e of the electron to its mass m by
observing its deflection in both electric and magnetic fields. The dis-
covery of the electron is usually said to date from this historic experiment,
although H. A. Lorentz and P. Zeeman (1865-1943), during the previous
year, had measured this same quantity for electrons bound in atoms, using
a method entirely different from Thomson's.
Figure 33-14 shows a modernized version of Thomson's apparatus.
electrons are emitted from hot filament F and accelerated by an applied
potential difference V. They then enter a region in which they move at right
angles to an electric field E and a field of magnetic induction B; E and B
are themselves at right angles to each other. The beam is made visible as a
spot of light when it strikes fluorescent screen S. The entire region in which
the electrons move is highly evacuated so that collisions with air molecules
will not occur.
The resultant force on a charged particle moving through an electric and a
magnetic field is given by Eq. 33-4, or
F = q0 E + q0v x B.
Study of Fig. 33-14 shows that the electric field deflects the particle upward
and the magnetic field deflects it downward. If these deflecting forces are to
cancel (that is, if F = O), this equation. for this problem, reduces to
eE = evB
or E = vB. (33-19)

Fig. 33-14 Electrons from the heated filament Fare accelerated by a potential difference
V and pass through a hole in the screen C. After passing through a region in which per-
pendicular electric and magnetic fields are present, they strike the fluorescent screen S.
836 THE MAGNETIC FIELD Chap. 33

Thus for a given electron speed v the condition for zero deflection can be
satisfied by adjusting E or B.
When Thomson's experiment is repeated today, usually as a student
experiment, the procedure* is: (a) note the position of the undeflected beam
spot, with E and B both equal to zero; (b) apply a fixed electric field E,
measuring on the fluorescent screen the deflection so caused; and (c) apply
a magnetic field and adjust its value until the beam deflection is restored to
zero.
In Section 27-5 we saw that the deflection y of an electron in a purely
electric field (step b), measured at the far edge of the deflecting plates, 1s
given by Eq. 27-9, or, with small changes in notation,
eEl2
y=--,
2mv 2
where v is the electron speed and l is the length of the deflecting plates; y is
not measurable directly, but it may be calculated from the measured dis-
placement of the spot on the screen if the geometry of the apparatus is
known. Thus y, E, and l are known; the ratio e/m and the velocity v are
unknown. We cannot calculate e/m until we have found the velocity, which
is the purpose of step c above.
If (step c) the electric force is set equal and opposite to the magnetic force,
the net force is zero and we can write (Eq. 33-19)
E
v = -·
B
Substituting this equation into the equation for y and solving for the ratio
e/m leads to
e 2yE
(33-20)
m = B 2 z2 '

in which all the quantities on the right can be measured. Thomson's value
for e/m was 1.7 X 10 11 couljkg, in excellent agreement with the modern
value of 1.75890 X 10 11 coul/kg.

QUESTIONS

1. Of the three vectors in the equation F = qv x B, which pairs are always at right
angles? Which may have any angle between them?
2. Why do we not simply define the magnetic induction B to point in the direction of the
magnetic force that acts on the moving charge?
3. Imagine that you are sitting in a room with your back to one wall and that an electron
beam, traveling horizontally from the back wall toward the front wall, is deflected to your
right. What is the direction of the field of magnetic induction that exists in the room?

* Thomson, no doubt for practical reasons, did not use this null method. He applied
the fields E and B separately and adjusted them to obtain the same beam deflection.
Equation 33-20 still holds. See Thomson's original paper: Philosophical Magazine,
44, 293 (1897).
Chap. 33 PROBLEMS 837
4. If an electron is not deflected in passing through a certain region of space, can we be
sure that there is no magnetic field in that region?
5. If a moving electron is deflected sideways in passing through a certain region of space,
can we be sure that a magnetic field exists in that region? .
6. A beam of protons is deflected sideways. Could this deflection be caused (a) by an
electric field? (b) By a magnetic field? (c) If either could be responsible, how would you
be able to tell which was present?
7. A conductor, even though it is carrying a current, has zero net charge. Why, then,
does a magnetic field exert a force on it?
8. Equation 33-11 (T = I' x B) shows that there is no torque on a current loop in an
external magnetic field if the angle between the axis of the loop and the field is (a) 0° or
(b) 180°. Discuss the nature of the equilibrium (that is, is it stable, neutral, or unstable?)
for these two positions.
9. In Example 4 we showed that the work required to turn a current loop end-for-end
in an external magnetic field is 2µB. Does this hold no matter what the original orienta-
tion of the loop was?
10. Imagine that the room in which you are seated is filled with a uniform magnetic
field with B pointing vertically upward. A circular loop of wire has its plane horizontal
and is mounted so that it is free to rotate about a horizontal axis. For what direction
of current in the loop, as viewed from above, will the loop be in stable equilibrium with
respect to forces and torques of magnetic origin?
11. A rectangular current loop is in an arbitrary orientation in an external magnetic
field. Is any work required to rotate the loop about an axis perpendicular to its plane?
12. (a) In measuring Hall potential differences, why must we be careful that points x
and yin Fig. 33-10 are exactly opposite to each other? (b) If one of the contacts is mov-
able, what procedure might we follow in adjusting it to make sure that the two points are
properly located?
13. A uniform magnetic field fills a certain cubical region of space. Can an electron be
fired into this cube from the outside in such a way that it will travel in a closed circular
path inside the cube?
14. Imagine the room in which you are seated to be filled with a uniform magnetic field
with B pointing vertically downward. At the center of the room two electrons are
suddenly projected horizontally with the same speed but in opposite directions. (a) Discuss
their motions. (b) Discuss their motions if one particle is an electron and one a positron.
15. In Fig. 33-4, why are the low-energy electron tracks spirals? That is, why does the
radius of curvature change in the constant magnetic field in which the chamber is im-
mersed?
16. What are the primary functions of (a) the electric field and (b) the magnetic field
in the cyclotron?
17. For Thomson's e/rn experiment to work properly (Section 33-8), is it essential that
the electrons have a. fairly constant speed?
PROBLEMS

_/l: The electrons in the beam of a television tube have an energy of 12 kev. The tube is
oriented so that the electrons move horizontally from south to north. The vertical com-
ponent of the earth's magnetic field points down and has B = 5.5 X 10-0 weber/meter2.
(a) In what direction will the beam deflect? (b) What is the acceleration of a given elec-
tron? (c) How far will the beam deflect in moving 20 cm through the television tube?
~fn a nuclear experiment a 1.0-Mev proton moves in a uniform magnetic field in a
circular path. What energy must (a) an alpha particle and (b) a deuteron have if they are
to circulate in the same orbit?
3. A wire 1.0 meter long carries a current of 10 amp and makes an angle of 30° with
a uniform magnetic field with B = 1.5 webers/meter2. Calculate the magnitude and di-
rection of the force on the wire.
838 THE MAGNETIC FIELD Chap. 33
/-4'<A wire of 60 cm length and mass 10 gm is suspended by a pair of flexible leads in a
magnetic field of induction 0.40 weber/meter2• What are the magnitude and direction
of the current required to remove the tension in the supporting leads? See Fig. 33-15.

x x x x x x

x x x x x x

x x x x x x

x x x x x x
~60cm-

Fig. 33-15

5. Express magnetic induction B and magnetic flux cp in terms of the fundamental


dimen,,sions M, L, T, and Q (mass, length, time, and charge).
,6: A metal wire of mass m slides without friction on two rails spaced a distance d apart,
as in Fig. 33-16. The track lies in a verti-
B cal uniform field of magnetic induction B.
;;;;·····~·.,.::;:::jl:::;~~l==:;;:~az:~$BI~-, A constant current i flows from generator G
along one rail, across the wire, and back
d down the other rail. Find the velocity
(speed and direction) of the wire as a
function of time, assuming it to be at rest
at t = 0.
7. A U-shaped wire of mass m and length
l is immersed with its two ends in mercury
Fig. 33-16 (Fig. 33-17). The wire is in a homogeneous
field of magnetic induction B. If a charge,
that is, a current pulse q = f dt,
i is sent through the wire, the wire will jump up. Cal-
culate, from the height h that the wire reaches, the size of the charge or current pulse,
assuming that the time of the current pulse is very small in comparison with the time of
flight. Make use of the fact that impulse of force equals f F dt, which equals mv. (Hint:
Try to relate f i dt to f F dt.) Evaluate q for B = 0.1 weber/meter 2, m = 10 gm, l = 20
cm, and h = 3 meters.
x x x XB x x

x x x x x x

x x x x x x

Hg

Fig. 33-17
Chap. 33 PROBLEMS 839

8. Figure 33-18 shows a wire ring of radius a at right angles to the general direction of
a radially symmetric diverging magnetic field. The magnetic induction at the ring is
everywhere of the same magnitude B, and its direction at the ring is everywhere at an
angle 8 ,vith a normal to the plane of the ring. The twisted lead wires have no effect on
the problem. Find the magnitude and direction of the force the field exerts on the ring if
the ring carries a current i as shown in the figure.

Fig. 33-18

9. A copper rod weighing 0.30 lb rests on two rails 1.0 ft apart and carries a current of
50 amp from one rail to the other. The coefficient of starting friction is 0.60. What is the
smallest magnetic field that would cause the bar to slide and what is its direction?
10. Figure 33-19 shows a wire of arbitrary shape carrying a current i between points a
and b. The wire lies in a plane at right angles to a uniform field of magnetic induction B.
Prove that the force on the wire is the same as that on a straight wire carrying a current i
directly from a to b. (Hint: Replace the wire by a series of "steps" parallel and perpen-
dicular to the straight line joining a and b.)

x x x x x x
b 10 cm
x x /,x x
/
x
/
x
rB
X/ x x
/
/
x }/ x x x x
a B
x x x x x x

x x x x x x 5cm

Fig. 33-19 Fig. 33-20

~ · · Figure 33-20 shows a rectangular 1 wenty-turn loop of wire, 10 cm by 5.0 cm. It


carries a current of 0.10 amp and is hinged at one side. What torque (direction and magni-
tude) acts on the loop if it is mounted with its plane at an angle of 30° to the direction of
a uniform field of magnetic induction 0.50 weber/meter 2?
12. Prove that the relation r = N AiB sin 8 holds for closed loops of arbitrary shape.
(Hint: Replace the loop of arbitrary shape by an assembly of adjacent long, thin-approx-
imately rectangular-loops which are equivalent to it as far as the distribution of current
is concerned.)
,.J.3<A length L of wire carries a current i. Show that if the wire is formed into a circular
coil the maximum torque in a given magnetic field is developed when the coil has one turn
only and the maximum torque has the value

T = -1- £2'iB.
4'11'
840 THE MAGNETIC FIELD Chap. 33
14. Figure 33-21 shows a wooden cylinder with a mass m of 0.25 kg, a radius R, and a
length l of 0.1 meter with N equal to ten turns of wire wrapped around it longitudinally,
so that the plane of the wire loop contains the axis of the cylinder. What is the least cur-
rent through the loop that will prevent the cylinder from rolling down an inclined plane
whose surface is inclined at an angle () to the horizontal, in the presence of a vertical field
of magnetic induction 0.5 weber /meter 2, if the plane of the windings is parallel to the
inclined plane?

Fig. 33-21 Fig. 33-22

15. A current i, indicated by the crosses in Fig. 33-22, is established in a strip of copper
of height h and width w. A uniform field of magnetic induction Bis applied at right angles
to the strip. (a) Calculate the drift speed VJ for the electrons. (b) What are the magni-
tude and direction of the magnetic force F acting on the electrons? (c) What would the
magnitude and direction of a homogeneous electric field E have to be in order to counter-
balance the effect of the magnetic field? (d) What is the voltage V necessary between two
sides of the conductor in order to create this field E? Between which sides of the conductor
would this voltage have to be applied? (e) If no electric field is applied from the outside,
the electrons will be pushed somewhat to one side and therefore will give rise to a uniform
electric field Eu across the conductor until the forces of this electrostatic field EH balance
the magnetic forces encountered in part (b). What will be the magnitude and direction of
the field EH? Assume that n, the number of conduction electrons per unit volume, is
1.1 X 10 29 /meter 3 and that h = 0.02 meter, w = 0.1 cm, i = 50 amp, and B = 2 webers/
meter 2•
16. (a) Show that the ratio of the Hall electric field EH to the electric field E responsible
for the current is
EH B
-=-·
E nep

(b) What is the angle between EH and E? (c) Evaluate this ratio for the conditions of
Example 5.
17. A proton, a deuteron, and an a-particle, accelerated through the same potential
difference, enter a region of uniform magnetic field, moving at right angles to B. (a)
Compare their kinetic energies. (b) If the radius of the proton's circular path is 10 cm,
what are the radii of the deuteron and the a-particle paths?
18. A proton, a deuteron, and an a-particle with the same kinetic energies enter a re-
gion of uniform magnetic field, moving at right angles to B. Compare the radii of their
circul~J:,paths.
_.,..,,Hf'.'. An a-particle travels in a circular path of radius 0.45 meter in a magnetic field with
.B = 1.2 webers/meter 2 . Calculate (a) its speed, (b) its period of revolution, (c) its kinetic
energy, and (d) the potential difference through which it would have to be accelerated to
achieve this energy.
Chap. 33 PROBLEMS 841

20. An electron is accelerated through 15,000 volts and is then allowed to circulate at
right angles to a uniform magnetic field with B = 250 gauss. What is its path radius?
21. A beam of electrons whose kinetic energy is K emerges from a window at the end
of an accelerator tube. There is a metal plate a distance d from this window and at
right angles to the direction of the emerging beam. Show that we can stop the beam
from hitting the plate, if we apply a magnetic field B such that

B";>
: -. (2mK)Hi
~
ed

in which m and e are the electron mass and charge. How should B be oriented?
22. Show that the radius of curvature of a charged particle moving at right angles to a
magnetic field is proportional to its momentum.
23. What uniform magnetic field can be set up in space to permit a proton of speed
1.0 X 107 meters/sec to move in a circle the size of the earth's equator?
24. A deuteron in a large cyclotron is moving in a magnetic field with B = 1.5 webers/
meter 2 and an orbit radius of 2.0 meters. Because of a grazing collision with a target,
the deuteron breaks up, with a negligible loss of kinetic energy, into a proton and a neu-
tron. Discuss the subsequent motions of each. Assume that the deuteron energy is
shared equally by the proton and neutron at breakup.
25. A 2-kev positron is projected into a uniform field of induction B of 0.10 weber /meter 2
with its velocity vector making an angle of 89° with B. Convince yourself that the path
will be a helix, its axis being the direction of .B. Find the period, the pitch p, and the radius
r of the helix; see Fig. 33-23.

L /
/
/,.,..,,.---- --
I
v

I
I
I

__
/
//
_,,.,,...,,..,,.

---
Fig. 33-23

26. (a) In a magnetic field with B = 0.50 weber/meter2, for what path radius will an
electron circulate at 0.1 the speed of light? (b) What will its kinetic energy be?
842 THE MAGNETIC FIELD Chap. 33
27. Time-of-flight spectrometer. S. A. Goudsmit has devised a method for measuring
accurately the masses of heavy ions by timing their period of circulation in a known
magnetic field. A singly charged ion of iodine makes 7 rev in a field of 4.5 X 10-2 weber/
meter 2 in about 1.29 X 10-3 sec. What (approximately) is its mass in kilograms? Actu-
ally, the mass measurements are carried out to much greater accuracy than these approxi-
mate data suggest.
28. Mass spectrometer. Figure 33-24 shows an arrangement used by Dempster to mea-
sure the masses of ions. An ion of mass Mand charge +q is produced essentially at rest in
source S, a chamber in which a gas discharge is taking place. The ion is accelerated by
potential difference V and allowed to enter a field of magnetic induction B. In the field it
moves in a semicircle, striking a photographic plate at distance x from the entry slit and
being recorded. Show that the mass M is given by
B2q
M = - x2 •
8V

Fig. 33-24

29. Zeeman effect. In Bohr's theory of the hydrogen atom the electron can be thought
of as moving in a circular orbit of radius r about the proton. Suppose that such an atom
is placed in a magnetic field, with the plane of the orbit at right angles to B. (a) If the
electron is circulating clockwise, as viewed by an observer sighting along B, will the angular
frequency increase or decrease? (b) What if the electron is circulating counterclockwise?
Assume that the orbit radius does not change. [Hint: The centripetal force is now parti-
ally electric (FE) and partially magnetic (FB) in origin.]
30. In Problem 29 show that the change in frequency of rotation caused by the mag-
netic field is given approximately by
Be
~II=±--,
47r7n
Such frequency shifts were actually observed by Zeeman in 1896. (Hint: Calculate the
frequency of rotation without the magnetic field and also with it. Subtract, bearing in
mind that because the effect of the magnetic field is very small some--but not all-terms
containing B can be set equal to zero with little error.)
Chap. 33 PROBLEMS 843
31. The University of Pittsburgh cyclotron is normally adjusted to accelerate deuterons.
(a) What energy of protons could it produce, using the same oscillator frequency as that
used for deuterons? (b) What magnetic induction would be required? (c) What energy
of protons could be produced if the magnetic induction was left at the value used for deu-
terons? (d) What oscillator frequency would then be required? (e) Answer the same
questions for a-particles.
32. Estimate the total path length traversed by a deuteron in the University of Pitts-
burgh cyclotron during the acceleration process. Assume an accelerating potential be-
tween the <lees of 80,000 volts.
33. In a synchrocyclotron producing 400-Mev protons, what must the ratio be of the
oscillator frequency at the beginning of an accelerating cycle to that at the end? Such
a proton has a speed of 0.70c, where c is the speed of light.
34. A 10-kev electron moving horizontally enters a region of space in which there is a
downward-directed electric field of magnitude 100 volts/cm. (a) What are the magnitude
and direction of the (smallest) field of magnetic induction that will allow the electron to
continue to move horizontally? Ignore gravitational forces, which are rather small.
(b) Is it possible for a proton to pass through this combination of fields undeflected? If
so, under what circumstances?
35. An electric field of 1500 volts/meter and a magnetic field of 0.40 weber/meter2 act
on a moving electron to produce no force. (a) Calculate the minimum electron speed v.
(b) Draw the vectors E, B, and v.
Ampere's Law
CHAPTER 34

34-1 Ampere's Law


One class of problems involving magnetic fields, dealt with in Chapter 33,
concerns the forces exerted by a magnetic field on a moving charge or on a
current-carrying conductor and the torque exerted on a magnetic dipole.
A second class concerns the production of a magnetic field by a current-
carrying conductor or by moving charges. This chapter deals with prob-
lems of this second class.
The discovery that currents produce magnetic effects was made by Oersted
in 1820. Figure 34-1, which shows a wire surrounded by a number of small
magnets, illustrates a modification of his experiment. If there is no current
in the wire, all the magnets are aligned with the earth's magnetic field.
When a strong current is present, the magnets point so as to suggest that
the magnetic lines of induction form closed circles around the wire. This
view is strengthened by the experiment of Fig. 34-2, which shows iron filings
on a horizontal glass plate, through the center of which a current-carrying
conductor passes.
Today we write the quantitative relationship between current i and the
magnetic field B as
(34-1)

which is known as Ampere's law. Ampere, being an advocate of the action-


at-a-distance point of view, did not formulate his results in terms of fields;
this was first done by Maxwell. Ampere's law, including an important ex-
844
Sec. 34-1 AMPERE'S LAW 845

Fig. 34-1 An array of compass needles near a wire


/ ~
!
carrying a strong current. The black ends of the
compass needles are their north poles. The dot
shows the current emerging from the page. As
usual, the direction of a current is taken as the di-
rection of flow of positive charge.
1 ~

/
' ~

Fig. 34-2 Iron filings around a wire carrying a strong current. (Courtesy Physical
Science Study Committee.)
846 AMPERE'S LAW Chap. 34
tension of it made later by Maxwell, is one of the basic equations of electro-
magnetism (see Table 38-3).
We can gain an appreciation of the way Ampere's law developed histori-
cally by considering a hypothetical experiment which has, in fact, much in
common with experiments that were actually carried out. The experiment
consists of measuring B at various distances r from a long straight wire of
circular cross section and carrying a current i. This can be done by making
quantitative the qualitative observation of Fig. 34-1.
Let us put a small compass needle a distance r from the wire. Such a
needle, a small magnetic dipole, tends to line up with an external magnetic
field, with its north pole pointing in the direction of B. Figure 34-1 makes
it clear that B at the site of the dipole is tangent to a circle of radius r cen-
tered on the wire.
If the current in the wire of Fig. 34-1 is reversed in direction, all the com-
pass needles would reverse end-for-end. This experimental result leads to
the "right-hand rule" for finding the direction of B near a wire carrying a
current i: Grasp the wire with the right hand, the thumb pointing in the direction
of the current. The fingers will curl around the wire in the direction of B.
Let us now turn the dipole through an angle Ofrom its equilibrium position.
To do this, we must exert an external torque just large enough to overcome
the restoring torque T that will act on the dipole. T, 0, and B are related by
Eq. 33-11 (T = µ xB), which can be written in terms of magnitudes as
T = µ,B sin O (34-2)
and in which µ, is the magnitude of the magnetic moment of the dipole, 0
being the angle between the vectors µ and B. Even though we may not
know the value of µ, for the compass needle, we may take it to be a constant,
independent of the position or orientation of the needle. Thus by measur-
ing T and O in Eq. 34-2 we can obtain a relative measure of B for various
distances rand for various currents i in the wire. The experimental results
can be described by the proportionality
i
B o: -· (34-3)
r
We can convert this proportionality into an equality by inserting a propor-
tionality constant. As for Coulomb's law, and for similar reasons (see Sec-
tion 26-4), we do not write this constant simply ask but in a more complex
form, namely µ, 0 /21r, in which µ, 0 is called the permeability constant.* Equa-
tion 34-3 then becomes
µ,oi
B=--, (34-4)
21rr

which we choose to write in the form


(34-5)
* µo has no connection with the dipole moment µ that appears in Eq. 34-2.
Sec. 34-1 AMPERE'S LAW 847

Fig. 34-3 A circular path of integration surrounding a wire. The central dot suggests a
current i in the wire emerging from the page. Note that the angle between B and dI is
zero so that B·dl = B dl.

The left side of Eq. 34-5 can easily be shown to be f B·dl for a path
consisting of a circle of radius r centered on the wire. For all points on this
circle B has the same (constant) magnitude Band di, which is always tan-
gent to the path of integration, points in the same direction as B, as Fig.
34-3 shows. Thus

f B·dl = f B dl Bfdl
= = (B)(21rr),

f dl being simply the circumference of the circle. In this special case,


therefore, we can write the experimentally observed connection between the
field and the current as
(34-1)

which is Ampere's law. A host of other experiments suggests that Eq. 34-1
is true in general * for any magnetic field configuration, for any assembly
of currents, and for any path of integration.
In applying Ampere's law in the general case, we construct a closed linear
path in the magnetic field as shown in Fig. 34-4. This path is divided into
elements of length di, and for each element the quantity B·dl is evaluated.
Recall that B · di has the magnitude B dl cos () and can be interpreted as the
product of dl and the component of B ( = B cos (J) parallel to dl.
The inte-

* We must modify Eq. 34-1 if a time-varying electric field is present within the path
of integration. In this chapter we assume that if electric fields are present, they are
constant in magnitude and direction.
848 AMPERE'S LAW Chap. 34

Fig. 34-4 A path of integration in a magnetic field.

gral is the sum of the quantities B · di for all path elements in the complete
loop. The term i on the right of Eq. 34-1 is the net current (strictly,
f j·dS) that passes through the area bounded by the closed path.
The permeability constant in Ampere's law has an assigned value of
µ0 = 411' X 10- 7 weber/amp-meter
Both this and the permittivity constant ( e0 ) occur in electromagnetic
formulas when the mks system of units is used.
The student may wonder why eo in Coulomb's law is a measured quantity, whereas
µo in Ampere's law is an assigned quantity. The answer is that the ampere, which is
the mks unit for the current i in Ampere's law, is defined by a laboratory technique
(the current bal,ance) that involves forces exerted by magnetic fields and in which this
same constant µ. 0 appears. In effect, as we show in detail in Section 34-4, the size of
the current that we agree to define as one ampere is adjusted so that µo may have
exactly the value assigned to it above. In Coulomb's law, on the other hand, the
quantities F, q, and rare measured in ways in which the constant eo plays no role.
This constant must then take on the particular value that makes the left side of
Coulomb's law equal to the right side; no arbitrary assignment is possible.

34-2 B Near a long Wire


We have seen that the lines of magnetic induction for a long straight wire
carrying a current i are concentric circles centered on the wire and that B
at a distance r from the wire is given by Eq. 34--4:

µoi
B=-· (34-4)
211'r
Sec. 34-2 B NEAR A LONG WIRE 849

We may regard this as an experimental result consistent with, and readily


derivable from, Ampere's law.
It is interesting to compare Eq. 34-4 with the expression for the electric field near
a long line of charge, or
1 ;\
E=--· (28-7)
2'71"eo r
In each case there are multiplying constants, namely µ.o/211" and 1/211"1:o, and factors
describing the device responsible for the field, namely i and A. Finally, each field
varies as 1/r.
Equation 28-7 may be derived from Gauss's law by relating the electric field at a
Gaussian surface to the net charge within this surface. The (surface) integral in
Gauss's law is evaluated for a closed cylindrical surface to which the lines of E are
everywhere perpendicular.
Equation 34-4 may be derived from Ampere's law by relating the magnetic field
at a path of integration to the net current that pierces this path. The (line) integral
in Ampere's law is evaluated for a closed circular path to which the lines of B are
everywhere tangent .
..,._ Example 1 Derive an expression for B at a distance r from the center of a long
cylindrical wire of radius R, where r < R. The wire carries a current i 0 , distributed
uniformly over the cross section of the wire.

Fig. 34-5 Example 1. A circular path of


integration inside a wire. A current i 0, dis-
tributed uniformly over the cross section of
the wire, emerges from the page.

Figure 34-5 shows a circular path of integration inside the wire. Symmetry
suggests that B is tangent to the path as shown. Ampere's law,

f B·dl = µ.oi,
11"T2
gives (B)(2'11"r) = µ.oio 11"R 2 •

since only the fraction of the current that passes through the path of integration is
included in the factor ion the right. Solving for B and dropping the subscript on the
current yields

At the surface of the wire (r = R) this equation reduces to the same expression as that
found by putting r = R in Eq. 34-4 (B = µoi/21rR).
850 AMPERE'S LAW Chap. 34
Example 2. Figure 34-6 shows a flat strip of copper of width a and negligible
thickness carrying a current i. Find the magnetic field at a distance R from the
center of the strip, at right angles to the strip.

Fig. 34-6 Example 2. A flat strip of


width a carries a current i.

x=O

Let us subdivide the strip into long infinitesimal filaments of width dx, each of
which may be treated as a wire carrying a current di given by i(dx/a). The field
contribution dB at point P in Fig. 34-6 is given, for the element shown, by the dif-
ferential form of Eq. 34-4, or

dB _ µo di _ µo i(dx/a)
- 271" ---;: - 27!" R sec((

Note that the vector dB is at right angles to the line marked r.


Only the horizontal component of dB, namely dB cos(), is effective, the vertical
component being canceled by the contribution associated with a symmetrically
located filament on the other side of the origin. Thus B at point P is given by the
(scalar) integral
B fdB cos () µoi(dx/a) ()
= = f cos
27rR sec()

= 2~R
µoi f sec dx
2 ().
Sec. ~4-3 MAGNETIC LINES OF INDUCTION 851
The variables x and 8 are not independent, being related by
x=Rtan8
or dx = R sec 2 () d8.
Bearing in mind that the limits on 8 are ±tan- 1 (a/2R) and eliminating dx from this
expression for B, we find
B = µ 0i f
R sec 2 8 d8
21raR sec 2 8
µ i
= ~ 0
I +tan-la/2R
d8 =
µoi a
-tan- 1 - ·
21ra -tan-'a/2R 1ra 2R
At points far from the strip, a/2R is a small angle, for which tan- 1 a ""a. Thus we
have, as an approximate result,

B "";~ (2~) = ; ; ~-

This result is expected because at distant points the strip cannot be distinguished
from a cylindrical wire (see Eq. 34-4). ....

34-3 Magnetic Lines of Induction


Figure 34-7 shows the lines of magnetic induction representing the field of
B near a long straight wire. Note the increase in the spacing of the lines
with increasing distance from the wire. This represents the 1/r decrease in
B predicted by Eq. 34-4.
Figure 34-8 shows the resultant lines of magnetic induction associated
with a current in a wire that is oriented at right angles to a uniform external
field Be. At any point the resultant magnetic induction B will be the vector
sum of Be and Bi, where Bi is the magnetic induction set up by the current
in the wire. The fields Be and Bi tend to cancel above the wire and to re-
enforce each other below the wire. At point Pin Fig. 34-8 Be and Bi cancel
exactly. Very near the wire the field is represented by circular lines and is
essentially Bi.
Michael Faraday, who originated the concept of lines of induction, en-
dowed them with more reality than they are currently given. He imagined

Fig. 34-7 Lines of B near a long cylindrical


wire. A current i, suggested by the central
dot, emerges from the page.
852 AMPERE' s LA w Chap. 34

),
--------------- J

Fig. 34-8 Lines of B near a long current-carrying wire immersed in a uniform external
field Be that points to the right. The current i is emerging from the page.

that, like stretched rubber bands, they represent the site of mechanical
forces. On this picture can we not visualize that the wire in Fig. 34-8 will be
pushed up? Today we use lines of induction largely for purposes of visualiza-
tion. For quantitative calculations we use the field vectors, describing the
force on the wire in Fig. 34-8, for example, from the relation F = ii x B.
In applying this relation to Fig. 34-8, we recall that B is always the
external field in which the wire is immersed; that is, it is Be and thus points
to the right. Since I points out of the page, the magnetic force on the wire
( = il x Be) does indeed point up. It is necessary to use only the external
field in such calculations because the field set up by the current in the wire
cannot exert a force on the wire, just as the gravitational field of the earth
cannot exert a force on the earth itself but only on another body. In Fig.
34-7, for example, there is no magnetic force on the wire because no external
magnetic field is present.
Faraday's idea of lines of induction was instrumental in overthrowing the older
action-at-a-distance theory of magnetic (and electric) attraction. Like many new
ideas, it was not immediately accepted. In 1851, for example, Faraday wrote:

I cannot refrain from again expressing my conviction of the truthfulness of the


representation, which the idea of lines of force affords in regard to magnetic action.
All the points that are experimentally established in regard to that action-i.e., all
that is not hypothetical-appear to be well and truly represented by it.
On the other hand, four years later another well-known British scientist, Sir
George Airy, wrote:

I declare that I can hardly imagine anyone who practically and numerically
knows this agreement [with the action-at-a-distance theory] to hesitate an instant
in the choice between this simple and precise action, on the one hand, and anything
so vague as lines of force, on the other hand.
Sec. 34-4 TWO PARALLEL CONDUCTORS 853

34-4 Two Parallel Conductors


Figure 34-9 shows two long parallel wires separated by a distance d and
carrying currents ia and ib. It is an experimental fact, noted by Ampere
only one week after word of Oersted's experiments reached Paris, that two
such conductors attract each other.
Some of Ampere's colleagues thought that in view of Oersted's experiment this
attraction between two conductors was an obvious result and did not need to be
proved. They reasoned that if wire a and wire b each exert forces on a compass
needle they should exert forces on each other. This conclusion is wrong. When he
heard it, Arago, a contemporary of Ampere, drew two iron keys from his pocket and
replied, "Each of these keys attracts a magnet. Do you believe that they therefore
also attract each other?"
Wire a in Fig. 34-9 will produce a field of induction Ba at all nearby
points. The magnitude of Ba, due to the current ia, at the site of the second
wire is, from Eq. 34-4,

The right-hand rule shows that the direction of Ba at wire b is down, as


shown in the figure.
Wire b, which carries a current ib, finds itself immersed in an external field
of magnetic induction Ba, A length l of this wire will experience a sideways
magnetic force ( = ii x B) whose magnitude is

(34-6)

The vector rule of signs tells us that Fb lies in the plane of the wires and points
to the left in Fig. 34-9.
We could have started with wire b, computed the field of induction which
it produces at the site of wire a, and then computed the force on wire a. The

Fig. 34-9 Two parallel wires that carry par-


allel currents attract each other.

ib
854 AMPERE'S LAW Chap. 34
force on wire a would, for parallel currents, point to the right. The forces
that the two wires exert on each other are equal and opposite, as they must
be according to Newton's law of action and reaction. For antiparallel cur-
rents the two wires repel each other.
This discussion reminds us of our discussion of the electric field between
two point charges in Section 27-1. There we saw that the charges act on
each other through the intermediary of the electric field. The conductors in
Fig. 34-9 act on each other through the intermediary of the magnetic field.
We think in terms of
current ~ field

and not, as in the action-at-a-distance point of view, in terms of

current ~ current.

The attraction between long parallel wires is used to define the ampere.
Suppose that the wires are 1 meter apart (d = 1.0 meter) and that the two
currents are equal (ia = ib = i). If this common current is adjusted until,
by measurement, the force of attraction per unit length between the wires
is 2 X 10-7 nt/meter, the current is defined to be 1 ampere. From Eq. 34-6,

F (47r X 10-7 weber/amp-m)(l amp) 2


(27r) ( 1 meter)
= 2 X 10-7 nt/meter
as expected.*
At the National Bureau of Standards primary measurements of current are made
with a current balance. This consists of a carefully wound coil placed between two
other coils, as in Fig. 34-10. The outer pair of coils is fastened to the table, and the
inner one is hung from the arm of a balance. The coils are so connected that the
current to be measured exists, as a common current, in all three of them.
The coils exert forces on one another-just as the parallel wires of Fig. 34-9 do-
which can be measured by loading weights on the balance pan. The current is
defined in terms of this measured force and the dimensions of the coils. The current
balance is perfectly equivalent to the long parallel wires of Fig. 34-9 but is a much
more practical arrangement. Current balance measurements are used primarily to
standardize other, more convenient, secondary methods of measuring currents.
.... Example 3. A long horizontal rigidly supported wire carries a current ia of 100
amp. Directly above it and parallel to it is a fine wire that carries a current ib of 20
amp and weighs 0.0050 lb/ft ( = 0.073 nt/meter). How far above the lower wire
should this second wire be strung if we hope to support it by magnetic repulsion?
To provide a repulsion, the two currents must point in opposite directions. For
equilibrium, the magnetic force per unit length must equal the weight per unit length
and must be oppositely directed.

* Note that µ 0 appears in this relation used to define the ampere. As stated on page
848, µo is assigned the (arbitrary) value of 41r X 10-7 weber /amp-m, and the size of the
current that we define as 1 ampere is adjusted to give the required force of attraction per
unit length.
Sec. 34-4 TWO PARALLEL CONDUCTORS 855

Fig. 34-10 A current balance.

Solving Eq. 34-6 for d yields


d = µoiaib = (411" X 10-7 weber/amp-m)(lOO amp)(20 amp)
2?r(F /l) (2'11")(0.073 nt/meter)
= 5.5 X 10-3 meter = 5.5 mm.
We assume that the wire diameters are much smaller than their separation. This
assumption is necessary because in deriving Eq. 34-6 we tacitly assumed that the
magnetic induction produced by one wire is uniform for all points within the second
wire.
Is the equilibrium of the suspended wire stable or unstable against vertical dis-
placements? This can be tested by displacing the wire vertically and examining how
the forces on the wire change.
Suppose that the fine wire is suspended below the rigidly supported wire. How may
it be made to "float"? Is the equilibrium against vertical displacements stable or
unstable?
Example 4. Two parallel wires a distanced apart carry equal currents i in oppo-
site directions. Find the magnetic induction for points between the wires and at a
distance x from one wire.
Study of Fig. 34-11 shows that Ba due to the current ia and Bb due to the current
ib point in the same direction at P. Each is given by Eq. 34-4 (B = µ 0i/2?rr) so that

B = Ba + Bb = µoi
-
(1 + -1- )·
-
271" x d- x
856 AMPERE' s LAw Chap. 34

Fig. 34-11 Example 4.

This relationship does not hold for points inside the wires because Eq. 34-4 is not
valid there. <1111

34-5 B for a Solenoid


A solenoid is a loi;ig wire wound in a close-packed helix and carrying a cur-
rent i. We assume that the helix is very long compared with its diameter.
What is the nature of the field of B that is set up?
For points close to a single turn of the solenoid, the observer is not aware
that the wire is bent in an arc. The wire behaves magnetically almost like
a long straight wire, and the lines of B due to this single turn are almost
concentric circles.
The solenoid field is the vector sum of the fields set up by all the turns that
make up the solenoid. Figure 34-12, which shows a "solenoid" with widely
spaced turns, suggests that the fields tend to cancel between the wires. It
suggests also that, at points inside the solenoid and reasonably far from the
wires, B is parallel to the solenoid axis. In the limiting case of adjacent
square tightly packed wires, the solenoid becomes essentially a cylindrical
current sheet and the requirements of symmetry then make the statement
just given necessarily true. We assume that it is true in what follows.

•P

Fig. 34-12 A loosely wound solenoid.


Sec. 34-5 B FOR A SOLENOID 857

Fig. 34-13 A solenoid of finite length. The right end, from which lines of B emerge,
behaves like the north pole of a compass needle. The left end behaves like the south pole.

For points such as P in Fig. 34-12 the field set up by the upper part of
the solenoid turns (marked O) points to the left and tends to cancel the
field set up by the lower part of the solenoid turns (marked ®), which points
to the right. As the solenoid becomes more and more ideal, that is, as it ap-
proaches the configuration of an infinitely long cylindrical current sheet, the
field of induction at outside points approaches zero. Taking the external
field to be zero is not a bad assumption for a practical solenoid if its length is
much greater than its diameter and if we consider only external points near
the central region of the solenoid, that is, away from the ends. Figure 34-13
shows the lines of induction for a real solenoid, which is far from ideal in
that the length is not much greater than the diameter. Even here the spac-
ing of the lines of induction in the central plane shows that the external field
is much weaker than the internal field.
Let us apply Ampere's law,
f B·dl = µ.oi,

to the rectangular path abed in the ideal solenoid of Fig. 34-14. We write
the integral f B·dl as the sum of four integrals, one for each path segment:
f B·dl = ib +ic
B·dl B·dl + f B·dl + iaB·dl.
d

Fig. 34-14 A section of an ideal


solenoid, made of adjacent
square turns, equivalent to an
.................................
infinitely long cylindrical current a --- ---- b
sheet.
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
858 AMPERE'S LAW Chap. 34
The first integral on the right is Bh, where Bis the magnitude of the magnetic
induction inside the solenoid and his the arbitrary length of the path from a
to b. Note that path ab, though parallel to the solenoid axis, need not coin-
cide with it.
The second and fourth integrals are zero because for every element of
these paths Bis at right angles to the path. This makes B·dl zero and thus
the integrals are zero. The third integral, which includes the part of the
rectangle that lies outside the solenoid, is zero because we have taken B as
zero for all external points for an ideal solenoid.
Thus f B·dl for the entire rectangular path has the value Bh. The net
current i that passes through the area bounded by the path of integration is
not the same as the current i 0 in the solenoid because the path of integration
encloses more than one turn. Let n be the number of turns per unit length;
then
i = i 0 (nh).
Ampere's law then becomes

or (34-7)
Although Eq. 34-7 was derived for an infinitely long ideal solenoid, it
holds quite well for actual solenoids for internal points near the center of
the solenoid. It shows that B does not depend on the diameter or the length
of the solenoid and that B is constant over the solenoid cross section. A
solenoid is a practical way to set up a known uniform magnetic field for
experimentation, just as a parallel-plate capacitor is a practical way to set
up a known uniform electric field.
.... Example 5. A solenoid is 1.0 meter long and 3.0 cm in mean diameter. It
has five layers of windings of 850 turns each and carries a current of 5.0 amp.
(a) What is B at its center? From Eq. 34-7,

B = µ 0i 0n = (47r X 10-7 weber/amp-m)(5.0 amp)(5 X 850 turns/meter)


= 2.7 X 10-2 weber/meter 2 •
We can use Eq. 34-7 even if the solenoid has more than one layer of windings because
the diameter of the windings does not enter.
(b) What is the magnetic flux PB for a cross section of the solenoid at its center?
To the extent that B is constant, we can calculate the flux from

<J:,B = f B·dS = BA,


where A is the effective cross-sectional area. Let us assume that A represents the
area of a circular disk whose diameter is the mean diameter of the windings (3.0 cm).
The effective area can then be shown to be 7 .1 X 10-4 meter 2, and
<l>B = BA = (2.7 X 10-2 weber/meter 2)(7.1 X 10-4 meter 2)
= 1.9 X 10- 5 weber.
Sec. 34-6 THE BIOT-SA Y ART LAW 859
Example 6. A toroid. Figure 34-15 shows a toroid, which may be described as a
solenoid of finite length bent into the shape of a doughnut. Calculate B at interior
points.

Fig. 34-15 Example 6. A toroid.

From symmetry the lines of B form concentric circles inside the toroid, as shown
in the figure. Let us apply Ampere's law to the circular path of integration of radius r:

f B·dl = µoi

or (B)(27rr) = µoioN,
where io is the current in the toroid windings and N is the total number of turns.
This gives
B = µo ioN_
27f r
In contrast to the solenoid, Bis not constant over the cross section of a toroid. Show
from Ampere's law that B equals zero for points outside an ideal toroid. .._

34-6 The Biot-Savart Law


Ampere's law can be used to calculate magnetic fields only if the sym-
metry of the current distribution is high enough to permit the easy evalua-
tion of the line integral f B·dl. This requirement limits the usefulness of
the law in practical problems. The law does not fail; it simply becomes
difficult to apply in a useful way.
Similarly, in electrostatics, Gauss's law can be used to calculate electric
fields only if the symmetry of the charge distribution is high enough to per-
mit the easy evaluation of the surface integral f E·dS. We can, for exam-
ple, use Gauss's law to find the electric field due to a long uniformly charged
rod but we cannot apply it usefully to an electric dipole, for the symmetry is
not high enough in this case.
To compute E at a given point for an arbitrary charge distribution, we
divided the distribution into charge elements dq and (see Section 27-4) we
used Coulomb's law to calculate the field contribution dE due to each ele-
860 AMPERE'S LAW Chap. 34

•b __ x'
------
•C

r dB
p

Fig. 34-16 The current element di establishes a magnetic field contribution dB at point P.

ment at the point in question. We found the field Eat that point by adding,
that is, by integrating, the field contributions dE for the entire distribution.
We now describe a similar procedure for computing Bat any point due to
an arbitrary current distribution. We divide the current distribution into
current elements and, using the law of Biot and Savart (which we describe
below), we calculate the field contribution dB due to each current element
at the point in question. We find the field B at that point by integrating
the field contributions for the entire distribution.
Figure 34-16 shows an arbitrary current distribution consisting of a cur-
rent i in a curved wire. The figure also shows a typical current element; it
is a length di of the conductor carrying a current i. Its direction is that of
the tangent to the conductor (dashed line). A current element cannot exist
as an isolated entity because a way must be provided to lead the current into
the element at one end and out of it at the other. Nevertheless, we can
think of an actual circuit as made up of a large number of current elements
placed end to end.
Let P be the point at which we want to know the magnetic induction dB
associated with the current element. According to the Biot-Savart law, dB
is given in magnitude by
µ 0 i dl sine
dB=- , (34-8)
41!' r2
where r is a displacement vector from the element to P and e is the angle
between this vector and di. The direction of dB is that of the vector dI x r.
In Fig. 34-16, for example, dB at point P for the current element shown is-
directed into the page at right angles to the plane of the figure. Note that
Eq. 34-8, being an inverse square law that describes the magnetic induction
due to a current element, may be viewed as the magnetic equivalent of
Coulomb's law, which is an inverse square law that describes the electric
field due to a charge element.
The law of Biot and Savart may be written in vector form as
µ 0 i dI x r
dB=-~~. (34-9)
411' r3
Sec. 34-6 THE BIOT-SAVART LAW 861

This formulation reduces to that of Eq. 34-8 when expressed in terms of


magnitudes; it also gives complete information about the direction of dB,
namely that it is the same as the direction of the vector di x r.
The resultant field at P is found by integrating Eq. 34-9, or

(34-10)

where the integral is a vector integral.


..,_ Example 7. A long straight wire. We illustrate the law of Biot and Savart by
applying it to find B due to a current i in a long straight wire. We discussed this
problem at length in connection with Ampere's law in Section 34-1.

11

l_ : ~() '

dx

r x

Fig. 34-17 Example 7.


l~R·~ p
dB

11
11
11
11

Figure 34-17, a side view of the wire, shows a typical current element dx. The
magnitude of the contribution dB of this element to the magnetic field at P is found
from Eq. 34-8, or
dB = µoi dx sin () .
41r r2

The directions of the contributions dB at point P for all elements are the same,
namely, into the plane of the figure at right angles to the page. Thus the vector
integral of Eq. 34-10 reduces to a scalar integral, or

B = fdB = !!'.f!i
41r
fx=+"'
x=-oo
sin~ dx.
r

Now, x, (), and rare not independent, being related (see Fig. 34-17) by

r = Vx2 + R2
and sin () [ = sin (1r - 0)] = V R ,
x2 + R2
862 AMPERE'S LAW Chap. 34
so that the expression for B becomes
µoi f+oo R dx
B = 4; -oo (x2 + R2)%
µoi I X 1x=+oo
= 41rR (x 2 +R 2) 72 x=-oo

= µo £.
21r R
This is the result that we arrived at earlier for this problem (see Eq. 34-4). The law
of Biot and Savart will always yield results that are consistent with Ampere's law
and with experiment.
This problem reminds us of its electrostatic equivalent. We derived an expression
for E due to a long charged rod, using Gauss's law (Section 28-6); we also solved this
problem by integration methods, using Coulomb's law (Section 27-4).
Example 8. A circular current loop. Figure 34-18 shows a circular loop of radius
R carrying a current i. Calculate B for points on the axis.

dl

dB..L ___ dB

Fig. 34-18 Example 8. A ring of radius R carrying a current i.

The vector di for a current element at the top of the loop points perpendicularly
out of the page. The angle Obetween di and r is 90°, and the plane formed by dl and r
is normal to the page. The vector dB for this element is at right angles to this plane
and thus lies in the plane of the figure and at right angles to r, as the figure shows.
Let us resolve dB into two components, one, dB II, along the axis of the loop and
another, dB ..l, at right angles to the axis. Only dB I contributes to the total induction
B at point P. This follows because the components dB11 for all current elements lie
on the axis and add directly; however, the components dB..l point in different direc-
tions perpendicular to the axis, and their resultant for the complete loop is zero, from
symmetry. Thus

where the integral is a simple scalar integration over the current elements.
For the current element shown in Fig. 34-18 we have, from the Biot-Savart law
(Eq. 34-8),
dB = l!.9'!. dl sin 90°.
41r r2
Sec. 34-6 THE BIOT-SAV ART LAW 863
We also have dB11 = dB cos a.

dB _ µoi cos a dl
Combining gives II - 41rr2

Figure 34-18 shows that r and a are not independent of each other. Let us express
each in terms of a new variable x, the distance from the center of the loop to the point
P. The relationships are
r = 2 yR +
x2 .

R R
= - =
and COSCI'.
r yR2 + x2 ·
Substituting these values into the expression for dB11 gives

µoiR
dB II = 41r(R2 + x2) % dl.

Note that i, R, and x have the same values for all current elements. Integrating this
equation, noting thatf dl is simply the circumference of the loop ( = 21rR), yields

(34-11)

...
If we put x » R in Example 8 so that points close to the loop are not
considered, Eq. 34-11 reduces to
µoi·R2
B=--·
2x3

Recalling that 1rR 2 is the area A of the loop and considering loops with N
turns, we can write this equation as

µ 0 (NiA) µo µ
B =- -
21r x 3 - 21r x3 '

whereµ is the magnetic dipole moment of the current loop. This reminds us
of the result derived in Problem 10, Chapter 27 [E = (1/2n 0 )(p/x 3 )], which
is the formula for the electric field strength on the axis of an electric dipole.
Thus we have shown in two ways that a current loop can be regarded as a
magnetic dipole: It experiences a torque given by T = µ. x B when placed in
an external magnetic field (Eq. 33-11); it generates its own magnetic field
given, for points on the axis, by the equation just developed.
864 AMPERE' s LAw Chap. 34
Table 34-1 is a summary of the properties of electric and magnetic dipoles.

Table 34-1

SOME DIPOLE EQUATIONS

Property Dipole Type Equation

Torque in an external field electric T = pXE

magnetic T = t,LXB

Energy in an external field electric U = -p·E

magnetic U = -1,L·B

Field at distant points along axis electric E = _l_p_


21reo x3

magnetic B = µ,o .!!__


21r x3

Field at distant points along electric E = _1_}!___


perpendicular bisector 41reo x3

magnetic B = µ,o .!!__


41r x 3

~ Example 9. In the Bohr model of the hydrogen atom the electron circulates
around the nucleus in a path of radius 5.1 X 10- 11 meter at a frequency v of 6.8 X
10 15 rev/sec. (a) What value of Bis set up at the center of the orbit?
The current is the rate at which charge passes any point on the orbit and is given by
i = ev = (1.6 X 10-19 coul)(6.8 X 10 15/sec) = 1.1 X 10-3 amp.
Bat the center of the orbit is given by Eq. 34-11 with x = 0, or

B = µ,oiR2 = '!.oi
2(R2 + x2) u 2R

(41r X 10- 7 weber/amp-m)(l.1 X 10- 3 amp)


(2)(5.1 X 10- 11 meter)
= 14 webers/meter 2 •
(b) What is the equivalent magnetic dipole moment? From Eq. 33-10,
µ, = NiA = (1)(1.1 X 10- 3 amp)(1r)(5.1 X 10-11 meter) 2
= 9.0 X 10-24 amp-m 2 •
Chap. 34 QUESTIONS 865

QUESTIONS

1. Can the path of integration around which we apply Ampere's law pass through a
conductor?
2. Suppose we set up a path of integration around a cable that contains twelve wires
with different currents (some in opposite directions) in each wire. How do we calculate i
in Ampere's law in such a case?
3. Apply Ampere's law qualitatively to the three paths shown in Fig. 34-19.

-, ,-,
I I
,_,
I
I
I
I
III •b
I
I I
I ,-,
,_,
II I

Fig. 34-19

4. Is B constant in magnitude for points that lie on a given line of induction?


5. Discuss and compare Gauss's law and Ampere's law.
6. A ~urrent is set up in a long copper pipe. Is there a magnetic field (a) inside and (b)
outside the pipe?
7. Equation 34-4 (B = µ 0i/27rr) suggests that a strong magnetic field is set up at points
near a long wire carrying a current. Since there is a current i and a magnetic field B, why
is there not a force on the wire in accord with the equation F = ii X B?
8. In electronics, wires that carry equal but opposite currents are often twisted together
to reduce their magnetic effect at distant points. Why is this effective?
9. A beam of 20-Mev protons emerges from a cyclotron. Is a magnetic field associated
with these particles?
10. Test the "floating" wire of Example 3 for equilibrium under horizontal displace-
ments. Consider that the wire floats above the rigidly supported wire and also below it.
Summarize the equilibrium situation for both wire positions and for both vertical and
horizontal displacements.
11. Explain qualitatively the forces of interaction between parallel wires carrying
parallel or antiparallel currents in terms of Faraday's lines of induction representation.
12. Comment on this statement: "The magnetic induction outside a long solenoid
cannot be zero, if only for the reason that the helical nature of the windings produces a
field for external points like that of a straight wire along the solenoid axis."
13. A current is sent through a vertical spring from whose lower end a weight is hanging;
what will happen? ·
14. Does Eq. 34-7 (B = µ 0in) hold for a solenoid of square cross section?
15. What is the direction of the magnetic fields at points a, b, and c in Fig. 34-16 set up
by the particular current element shown?
16. In a circular loop of wire carrying a current i, is B uniform for all points within the
loop?
17. Discuss analogies and differences between Coulomb's law and the Biot-Savart law.
18. Equation 34-9 gives the law of Biot and Savart in vector form. Write its electro-
static equivalent [that is, Eq. 27-6, or dE = dq/(47reor2 )] in vector form.
19. How might you measure the magnetic dipole moment of a compass needle?
20. What is the basis for saying that a current loop is a magnetic dipole?
866 AMPERE'S LAW Chap. 34

PROBLEMS

1. A #10 bare copper wire (0.10 in. in diameter) can carry a current of 50 amp without
overheating. For this current, what is B at the surface of the wire?
2. A surveyor is using a compass 20 ft below a power line in which there is a steady
current of 100 amp. Will this interfere seriously with the compass reading? The hori-
zontal component of the earth's magnetic field at the site is 0.2 gauss.
3. A long straight wire carries a current of 50 amp. An electron, traveling at 10 7 meters/
sec, is 5.0 cm from the wire. What force acts on the electron if the electron velocity is
directed (a) toward the wire, (b) parallel to the wire, and (c) at right angles to the direc-
tions defined by (a) and (b)?
4. A conductor consists of an infinite number of adjacent wires, each infinitely long and
carrying a current i. Show that the lines of B will be as represented in Fig. 34-20 and that
B for all points in front of the infinite current sheet will be given by

B = }µ 0ni,

where n is the number of conductors per unit length. Derive both by direct application
of Ampere's law and by considering the problem as a limiting case of Example 2.

Fig. 34-20 Fig. 34-21

'-&. Figure 34-21 shows a hollow cylindrical conductor of radii a and b which carries a
current i uniformly spread over its cross section. (a) Show that the magnetic field B for
points inside the body of the conductor (that is, a < r < b) is given by

Check this formula for the limiting case of a = 0. (b) Make a rough plot of the general
behavior of B(r) from r = 0 tor ----> oo.
Chap. 34 PROBLEMS 867
6. A long coaxial cable consists of two concentric conductors with the dimensions shown
in Fig. 34-22. There are equal and opposite currents i in the conductors. (a) Find the
magnetic induction Bat r within the inner conductor (r < a). (b) Find B between the two
conductors (a < r < b). (c) Find B within the outer conductor (b < r < c). (d) Find B
outside the cable (r > c).

Fig. 34-22 Fig. 34-23

· 7. A long copper wire carries a current of 10 amp. Calculate the magnetic flux per
meter of wire for a plane surface S inside the wire, as in Fig. 34-23.
8. Two long, parallel #10 copper wires (diameter = 0.10 in.) carry currents of 10 amp
in opposite directions. (a) If their centers are 2.0 cm apart, calculate the flux per meter
that exists in the space between the axes of the wires. (b) What fraction of the flux in
(a) lies inside the wires? (c) Repeat the calculation of (a) for parallel currents.
9. A long wire carrying a current of 100 amp is placed in a uniform external magnetic
field of 50 gauss. The wire is at right angles to this external field. Locate the points at
wpich the resultant magnetic field is zero.
l,Q. Two long wires a distanced apart carry equal antiparallel currents i, as in Fig. 34-24.
Show that B at point P, which is equidistant from the wires, is given by
2µoid
B = -----
1r(4R2 + d2)

r
d R p
a a

L. Fig. 34-24 Fig. 34-25

11. Four long #10 copper wires are parallel to each other, their cross section forming a
square 20 cm on edge. A 20-amp current is set up in each wire in the direction shown in
Fig. 34-25. What are the magnitude and direction of Bat the center of the square?
12. In Problem 11 what is the force per meter acting on the lower left wire, in magnitude
and direction?
868 AMPERE'S LAW Chap. 34
,,,·13. Figure 34-26 shows a long wire carrying a cur-
rent of 30 amp. The rectangular loop carries a current
of 20 amp. Calculate the resultant force acting on
the loop. Assume that a = 1.0 cm, b = 8.0 cm, and
l = 30 cm.
14. Derive the solenoid equation (Eq. 34-7) starting
from the expression for the field on the axis of a circular
loop (Example 8). (Hint: Subdivide the solenoid into
0..
E 0.. a series of current loops of infinitesimal thickness and
ro E
0 ro integrate.)
(Y) 0
N 15. A square loop of wire of edge a carries a current i.
Show that the value of B at the center is given by

1 B=
2viµoi
1ra
.

16.. A square loop of wire of edge a carries a current i.


(a) Sliow that B for a point on tlie axis of the loop and
a distance x from its center is given by
4µoia 2
B = ----~----~.
1r(4x
2 2 2 +
2 a ) ( 4x + 2a ) Y2
(b) Does this reduce to tlie result of Problem 15 for
x = O? (c) Does the square loop behave like a dipole
for points such that x » a? If so, what is its dipole
moment?
17. A straight wire segment of length l carries a
Fig. 34-26 current i. (a) Show that the field of induction B to
be associated with this segment, at a distance R from
the segment along a perpendicular bisector (see Fig. 34-27), is given in magnitude by
µoi l
B =------~
21rR cz + 4R 2) v.
2

!I>) DoeR thiR expreRRion reduce to an expected result as l ~ oo?


18. The wire shown in Fig. 34-28 carries a current i. What
is the magnetic induction at the center C of the semicircle
arising from (a) each straight segment of length l, (b) the
semicircular segment of radius R, and (c) the entire wire?
19. (a) A wire in the form of a regular polygon of n sides
is just enclosed by a circle of radius a. If the current in this

Fig. 34-27 Fig. 34-28

wire is i, show that the magnetic induction at the center of the circle is given by

µoni
B = - tan (1r/n).
21ra

(b) Show that as n ~ oo this result approaches that of a circular loop.


Chap. 34 PROBLEMS 869
20. (a) Show that B at the center of a rectangle of
length l and width d, carrying a current i, is given by
I

B=-
2µoi (l 2 + d 2) v, .
I
7r ld
!
(b) What does B reduce to for l » d? Is this a result
tliat you expect? Ip
21. Helmholtz coils. Two 300-turn coils are ar-
-T-
ranged a distance apart equal to their radius, as in
Fig. 34-29. For R = 5.0 cm and i = 50 amp, plot B
as a function of distance x along the common axis
over the range x = - 5 cm to x = +5 cm, taking
I

I
R
l
1
x = 0 at point P. (Such coils provide an especially
I
uniform field of B near point P.)
22. In Problem 21 let the separation of the coils be a I x
variable z. Show that if z equals R, then not only the R
first derivative (dB /dx) but also the second (d 2B /dx 2)
of B is zero at point P. This accounts for the uni-
Fig. 34-29
formity of B near point P for this particular coil sep-
aration.
23. A long "hairpin" is formed by bending a piece of wire as shown in Fig. 34-19. If a
10-amp current is set up, what are the direction and magnitude of B at point a? At
point b? Take R = 0.50 cm.
24. Calculate B at point P in Fig. 34-30. Assume that i = 10 amp and a = 8.0 cm.
25. A plastic disk of radius R has a charge q uniformly dis-
a/4 tributed over its surface. If the disk is rotated at an angular
frequency w about its axis, show that (a) the induction at the

l
a/4 center of the disk is
B = µowq
21rR
a
and (b) the magnetic dipole moment of the disk is
,...__________.J µ =
wqR 2
-4-·
e : - - a------:.
i-.e-l I

(Hint: The rotating disk is equivalent to an array of current


Fig. 34-30 loops; see Example 8.)
26. You are given a length l of wire in which a current
may be established. The wire may be formed into a circle or a square. Which yields th
larger value for Bat the central point? See Problem 15.
27. A circular copper loop of radius 10 cm carries a current of 15 amp. At its center ls
placed a second loop of radius 1.0 cm, having 50 turns and a current of 1.0 amp. (a) What
magnetic induction B does the large loop set up at its center? (b) What torque acts on thf
small loop? Assume that the planes of the two
loops are at right angles and that the induction B N
provided by the large loop is essentially uniform
throughout the volume occupied by the small loop.
28. Show that it is impossible for a uniform
magnetic field B to drop abruptly to zero as one
moves at right angles to it, as suggested by the
B a
horizontal arrow in Fig. 34-31 (see point a). In
actual magnets fringing of the lines of force always
occurs, which means that B approaches zero in a
continuous and gradual way. (Hint: Apply Am- s
pen~'s law to the rectangular path shown by the
dashed lines.) Fig. 34-31
Faraday's Law
CHAPTER 35

35-1 Faraday's Experiments


For some physical laws it is hard to find experiments that lead in a direct
and convincing way to the formulation of the law. Gauss's law, for example,
emerged only slowly as the common factor with whose aid all electrostatic
experiments could be interpreted and correlated. In Chapter 28 we found
it best to state Gauss's law first and then to show that the underlying experi-
ments were consistent with it.
Faraday's law of electromagnetic induction, which is one of the basic equa-
tions of electromagnetism (see Table 38-3), is different in that there are a
number of simple experiments from which the law can be-and was-de-
duced directly. Such experiments were carried out by Michael Faraday in
England in 1831 and by Joseph Henry in the United States at about the
same time.
Figure 35-1 shows the terminals of a coil connected to a galvanometer.
Normally we would not expect this instrument to deflect because there
seems to be no electromotive force in this circuit; but if we push a bar mag-
net toward the coil, with its north pole facing the coil, a remarkable thing
happens. While the magnet is moving, the galvanometer deflects, showing
that a current has been set up in the coil. If the magnet is held stationary
with respect to the coil, the galvanometer does not deflect. If the magnet
is moved away from the coil, the galvanometer again deflects, but in the
opposite direction, which means that the current in the coil is in the opposite
direction. If we use the south pole end of a magnet instead of the north
pole end, the experiment works as described but the deflections are reversed.
Further experimentation shows that what matters is the relative motion of the
870
Sec. 35-1 FARADAY'S EXPERIMENTS 871

Fig. 35-1 Galvanometer G deflects while the


magnet is moving with respect to the coil.

magnet and the coil. It makes no difference whether the magnet is moved
toward the coil or the coil toward the magnet.
The current that appears in this experiment is called an induced current
and is said to be set up by an induced electromotive force. Faraday was able
to deduce from experiments like this the law that gives their magnitude and
direction. Such emfs are very important in practice. The chances are good
that the lights in the room in which you are reading this book are operated
from an induced emf produced in a commercial electric generator.
In another experiment the apparatus of Fig. 35-2 is used. The coils are
placed close together but at rest with respect to each other. When the
switch S is closed, thus setting up a steady current in the right-hand coil, the
galvanometer deflects momentarily; when the switch is opened, thus in-
terrupting this current, the galvanometer again deflects momentarily, but in
the opposite direction. No gross objects are moving in this experiment.
In Faraday's words:
When the contact was made, there was a sudden and very slight effect at the
galvanometer, and there was also a similar slight effect when the contact with the
battery was broken. But whilst the voltaic current was continuing to pass through
the one helix, no galvanometrical appearances nor any effect like induction upon
the other helix could be perceived, although the active power of the battery was
proved to be very great . . . .
Experiment shows that there will be an induced emf in the left coil of
Fig. 35-2 whenever the current in the right coil is changing. It is the rate
at which the current is changing and not the size of the current that is significant.

Fig. 35-2 Galvanometer G deflects momen-


tarily when switch S is closed or opened.
872 FARADAY'S LAW Chap. 35
35-2 Faraday's Law of Induction
Faraday had the insight to perceive that the change in the flux cf>e of
magnetic induction for the left coil in the preceding experiments is the im-
portant common factor. This flux may be set up by a bar magnet or a cur-
rent loop. Faraday's law of induction says that the induced emf 8 in a cir-
cuit is equal to the negative rate at which the flux through the circuit is
changing. If the rate of change of flux is in webers/sec, the emf 8 will be in
volts. In equation form
c,= - - · (35-1)
dt

This equation is called Faraday's law of induction. The minus sign is an


indication of the direction of the induced emf, a matter we discuss further in
Section 35-3.
If Eq. 35-1 is applied to a coil of N turns, an emf appears in every turn
and these emfs are to be added. If the coil is so tightly wound that each
turn can be said to occupy the same region of space, the flux through each
turn will then be the same. The flux through each turn is also the same for
(ideal) toroids and solenoids (see Section 34-5). The induced emf in all such
devices is given by
dcf>B
e= -N- = (35-2)
dt

where Nif,B measures the so-called flux linkages in the device.


Figures 35-1 and 35-2 suggest that there are at least two ways in which we
can make the flux through a circuit change and thus induce an emf in that
circuit. The coil that is connected to the galvanometer cannot tell in which
of these experiments it is participating; it is aware only that the flux passing
through its cross-sectional area is changing. The flux through a circuit can
also be changed by changing its shape, that is, by squeezing or stretching it .

.... Example 1. A long solenoid has 200 turns/ cm and carries a current of 1.5 amp;
its diameter is 3.0 cm. At its center we place a 100-turn, close-packed coil of diameter
2.0 cm. This coil is arranged so that B at the center of the solenoid is parallel to its
axis. The current in the solenoid is reduced to zero and then raised to 1.5 amp in
the other direction at a steady rate over a period of 0.050 sec. What induced emf
appears in the coil while the current is being changed?
The induction Bat the center of the solenoid is given by Eq. 34-7, or

B = µ 0ni = (41r X 10-7 weber/amp-m)(200 X 10 2 turns/meter)(l.5 amp)


= 3.8 X 10-2 weber/meter 2 •
The area of the coil (not of the solenoid) is 3.1 X 10-4 meter 2 • The initial flux if,B
through each turn of the coil is given by
cf>B = BA = (3.8 X 10-2 weber/meter 2)(3.1 X 10-4 meter 2) = 1.2 X 10-5 weber.

The flux goes from an initial value of 1.2 X 10-5 weber to a final value of -1.2 X
10-5 weber. The change in flux Llcf>B for each turn of the coil during the 0.050-sec
Sec. 35-3 LENZ'S LAW 873
period is thus twice the initial value. The induced emf is given by

6 = _ N t:,.il!B = _ (100)(2 X 1.2 X 10- weber) = _ 4 _8 X 10 _ 2 volt


5
-48 mv.
6.t 0.050 sec
The minus sign deals with the direction of the emf, as we explain below.

35-3 Lenz's Law


So far we have not specified the directions of the induced emfs. Although
these directions can be found from a formal analysis of Faraday's law, we
prefer to find them from the conservation-of-energy principle which, in this
context, takes the form of Lenz's law: The induced current will appear in such
a direction that it opposes the change that produced it.* The minus sign in
Faraday's law suggests this opposition. In mechanics the energy principle

Fig. 35-3 If the magnet is moved toward the loop, the induced current points as shown,
setting up a magnetic field that opposes the motion of the magnet.

often allows us to draw conclusions about mechanical systems without an-


alyzing them in detail. We use the same approach here.
Lenz's law refers to induced currents, which means that it applies only to
closed circuits. If the circuit is open, we can usually think in terms of what
would happen if it were closed and in this way find the direction of the in-
duced emf.
Consider the first of Faraday's experiments described in Section 35-1.
Figure 35-3 shows the north pole of a magnet and a cross section of a nearby
conducting loop. As we push the magnet toward the loop, an induced cur-
rent is set up in the loop. What is its direction?
* Heinrich Friedrich Lenz (1804-1865) deduced this law in 1834. Faraday also dis-
covered how to determine the directions of induced emfs, but he did not express his re-
sults as succinctly as Lenz.
874 FARADAY'S LAW Chap. 35
A current loop sets up a magnetic field at distant points like that of a
magnetic dipole, one face of the loop being a north pole, the opposite face
being a south pole. The north pole, as for bar magnets, is that face from
which the lines of B emerge. If, as Lenz's law predicts, the loop in Fig. 35-3
is to oppose the motion of the magnet toward it, the face of the loop toward
the magnet must become a north pole. The two north poles-one of the
current loop and one of the magnet-will repel each other. The right-hand
rule shows that for the magnetic field set up by the loop to emerge from the
right face of the loop the induced current must be as shown. The current
will be counterclockwise as we sight along the magnet toward the loop.
When we push the magnet toward the loop (or the loop toward the mag-
net), an induced current appears. In terms of Lenz's law this pushing is
the "change" that produces the induced current, and, according to this law,
the induced current will oppose the "push." If we pull the magnet away
from the coil, the induced current will oppose the "pull" by creating a south
pole on the right-hand face of the loop of Fig. 35-3. To make the right-hand
face a south pole, the current must be opposite to that shown in Fig. 35-3.
Whether we pull or push the magnet, its motion will always be automatically
opposed.
The agent that causes the magnet to move, either toward the coil or
away from it, will always experience a resisting force and will thus be re-
quired to do work. From the conservation-of-energy principle this work
done on the system must be exactly equal to the Joule heat produced in the
coil, since these are the only two energy transfers that occur in the system.
If we move the magnet more rapidly, we will have to do work at a faster
rate and the rate of the Joule heating will increase correspondingly. If we
cut the loop and then perform the experiment, there will be no induced cur-
rent, no Joule heating, no force on the magnet, and no work required to
move it. There will still be an emf in the loop, but, like a battery connected
to an open circuit, it will not set up a current.
If the current in Fig. 35-3 were in the opposite direction to that shown, the face
of the loop toward the magnet would be a south pole, which would pull the bar
magnet toward the loop. We would only need to push the magnet slightly to start
the process and then the action would be self-perpetuating. The magnet would
accelerate toward the loop, increasing its kinetic energy all the time. At the same
time Joule heat would appear in the loop at a rate that would increase with time.
This would indeed be a something-for-nothing situation! Needless to say, it does
not occur.

Let us apply Lenz's law to Fig. 35-3 in a different way. Figure 35-4
shows the lines of B for the bar magnet.* On this point of view the "change"
is the increase in <PB through the loop caused by bringing the magnet nearer.
The induced current opposes this change by setting up a field that tends to
oppose the increase in flux caused by the moving magnet. Thus the field

* There are two fields of Bin this problem-one connected with the current loop and one
with the bar magnet. The student must always be certain which one is meant.
Sec. 35-4 INDUCTION-A QUANTITATIVE STUDY 875

Fig. 35-4 In moving the magnet to-


ward the loop, we increase il>B through
the loop.

due to the induced current must point from left to right through the plane
of the coil, in agreement with our earlier conclusion.
It is not significant here that the induced field opposes the magnet field
but rather that it opposes the change, which in this case is the increase in
<l>B through the loop. If we withdraw the magnet, we reduce <l>B through the
loop. The induced field will now oppose this decrease in <l>B (that is, the
change) by re-enforcing the magnet field. In each case the induced field
opposes the change that gives rise to it.

35-4 Induction-A Quantitative Study


The example of Fig. 35-4, although easy to understand qualitatively, does
not lend itself to quantitative calculations. Consider then Fig. 35-5, which
shows a rectangular loop of wire of width Z, one end of which is in a uniform
field B pointing at right angles to the plane of the loop. This field of B may
be produced in the gap of a large electromagnet like that of Fig. 33-2. The
dashed lines show the assumed limits of the magnetic field. The experiment
consists in pulling the loop to the right at a constant speed v.

fx x x x x x- x x x x YB
Ix x x Ix x x x ; 2 x x x I
1x x x x x x x x x i x x J

Ix x x ; x x x x x x x I
J F1 I
~ x x x x x x x x xi
~ x x x x x x x x x xi
I I
Ix x x x x x x x x i x x

:x x x x x x x F3 x x x I
[.:'.___ __'! _ ~ ~- _'!_ ~ _x_ .2..__X- ~

Fig. 35-5 A rectangular loop is pulled out of a magnetic field with velocity v.
876 FARADAY'S LAW Chap. 35
Note that the situation described by Fig. 35-5 does not differ in any essen-
tial particular from that of Fig. 35-4. In each case a conducting loop and
a magnet are in relative motion; in each case the flux of the field of the mag-
net through the loop is being caused to change with time.
The flux cf>s enclosed by the loop in Fig. 35-5 is

cf>s = Blx,

where lx is the area of that part of the loop in which Bis not zero. The emf
8 is found from Faraday's law, or

dcf>s d dx
8 = - - = - - (Blx) -Bl-= Blv (35-3)
dt dt dt '

where we have set -dx/dt equal to the speed v at which the loop is pulled
out of the magnetic field. Nate that the only dimension of the loop that
enters into Eq. 35-3 is the length l of the left end conductor. As we shall
see later, the induced emf in Fig. 35-5 may be regarded as localized here.
An induced emf such as this, produced by pulling a conductor through a
magnetic field, is sometimes called a motional emf.
The emf Blv sets up a current in the loop given by

. 8 Elv
i = - = -, (35-4)
R [(

where R is the loop resistance. From Lenz's law, this current (and thus 8)
must be clockwise in Fig. 35-5; it opposes the "change" (the decrease in
cf>s) by setting up a field that is parallel to the external field within the loop.
The current in the loop will cause forces Fi, F 2 , and F 3 to act on the three
conductors, in accord with Eq. 33-6a, or

F=ilxB. (35-5)

Because F 2 and F 3 are equal and opposite, they cancel each other; F 1 , which
is the force that opposes our effort to move the loop, is given in magnitude
from Eqs. 35-5 and 35-4 as

The agent that pulls the loop must do work at the steady rate of
B2l2v2
P = F1v = - - · (35-6)
R
From the principle of the conservation of energy, Joule heat must appear in
the resistor at this same rate. We introduced the conservation-of-energy
principle into our derivation when we wrote down the expression for the
current (Eq. 35-4); the student will recall that the relation i = 8/R for
Sec. 35-4 INDUCTION-A QUANTITATIVE STUDY 877

single-loop circuits is a direct consequence of this principle. Thus we should


be able to write down the expression for the rate of Joule heating in the loop
with the expectation that we will obtain a result identical with Eq. 35-6.
Recalling Eq. 35-4, we put
2
PJ=iR= -
. (Blv) R =B-l-v,
2 2 2 2

R R
which is indeed the expected result. This example provides a quantitative
illustration of the conversion of mechanical energy (the work done by an ex-
ternal agent) into electrical energy (the induced emf) and finally into thermal
energy (the Joule heating).
Figure 35-6 shows a side view of the coil in the field. In Fig. 35-6a the
coil is stationary; in Fig. 35-6b we are moving it to the right; in Fig. 35-6c
we are moving it to the left. The lines of induction in these figures repre-
sent the resultant fiel,d produced by the vector addition of the field B 0 due to
the magnet and the field Bi due to the induced current, if any, in the coil.

(a)

----------* (b)

-+---- (c)

Fig. 35-6 Side view of a rectangular loop in a magnetic field showing the loop (a) at rest,
(b) being pulled out, and (c) being pushed in.
878 FARADAY'S LAW Chap. 35
These lines suggest convincingly that the agent moving the coil always ex-
periences an opposing force.
~ Example 2. Figure 35-7 shows a rectangular loop of resistance R, width l, and
length a being pulled at constant speed v through a region of thickness d in which
a uniform field of induction B is set up by a magnet.

,---------1
I x x x x x I

rr:-:ix x~B
r
l
:x
I
xx x xi
I
Fig. 35-7 Example 2. A rectangular loop is
caused to move with a velocity v through a
L Ix xx x
.;>V
xi magnetic field. The position of the loop is
measured by x, the distance between the
I x x Ix x x : effective left edge of the field B and the right
~x_J I end of the loop.
I x
L
x x
_________
x x I
_J

----d---~

(a) Plot the flux <l>B through the loop as a function of the coil position x. Assume
that l = 4 cm, a = 10 cm, d = 15 cm, R = 16 ohms, B = 2.0 webers/meter2, and
v = 1.0 meter/sec.
The flux <I>B is zero when the loop is not in the field; it is Bla when the loop is
entirely in the field; it is Blx when the loop is entering the field and Bl[a - (x - d)J
when the loop is leaving the field. These conclusions, which the student should
verify, are shown graphically in Fig. 35-8a.
(b) Plot the induced emf f,,
The induced emf f, is given by f, = -d<I>B/dt, which can be written as
c, = _ d<J> B = _ d<J> B dx = _ d<J> B V
dt dxdt dx'
where d<I>B/dx is the slope of the curve of Fig. 35-8a. c.(x) is plotted in Fig. 35-8b.
Lenz's law, from the same type of reasoning as that used for Fig. 35-5, shows that
when the coil is entering the field the emf f, acts counterclockwise as seen from above.
Note that there is no emf when the coil is entirely in the magnetic field because the
flux <I>B through the coil is not changing with time, as Fig. 35-8a shows.
(c) Plot the rate P of Joule heating in the loop.
This is given by P = f, 2/R. It may be calculated by squaring the ordinate of the
curve of Fig. 35-8b and dividing by R. The result is plotted in Fig. 35-8c.
If the fringing of the magnetic field, which cannot be avoided in practice (see
Problem 34-28), is taken into account, the sharp bends and corners in Fig. 35-8 will
be replaced by smooth curves. What changes would occur in the curves of Fig. 35-8
if the coil were open circuited?
Example 3. A copper rod of length L rotates at angular frequency w in a uniform
field of magnetic induction B as shown in Fig. 35-9. Find the emf f, developed be-
tween the two ends of the rod.
If a wire of length di is moved at velocity vat right angles to a field B, a motional
emf de. will be developed (see Eq. 35-3) given by
de. = Bv dl.
Sec. 35-4 INDUCTION-A QUANTITATIVE STUDY 879

Fig. 35-8 Example 2.

The rod of Fig. 35-9 may be divided into elements of length dl, the linear speed v of
each element being wl. Each element is perpendicular to B and is also moving in a
direction at right angles to B so that

S = f de = LL
Bv dl = £
L
B(wl) dl = !BwL2.

For a second approach, consider that at any instant the flux enclosed by the sector
aOb in Fig. 35-9 is given by
<l>B = BA = B(!L20),
880 FARADAY'S LAW Chap. 35

__,,.... --x
_...----
x x x --
x- ......._
/
/x x x x x x x
/
/ x x x x x x x x
/
/ x x x x x x x
I
/x x x x x x x x x x
I
x x x x x x x x x x x
I
Ix x x x x x x x x x x x x
I
Ix x x x x x x x x x x
I
Ix x x x x x x XO
Ix x x x x x x x x x x x x x x
I
\
x x I
\x x x x x x x x x x x x Bx x x x x/
\ I
\ x x x x x x x x x x x x x x x
I
\
\x x x x x x x x x x x x x x x I
\ I
\ x x x x x x x x x x x x x /
\ /
x x x x x x x x x x x
""' "' /
/

- ----- -
x x x x x x x /
"x x/
....... ......._ /
/
~ x x x x-

Fig. 35-9 Example 3.

where !L20 can be shown to be the area of the sector. Differentiating gives

dcf,B = lBL2 d() = lB L2


dt 2 dt 2 W.

From Faraday's law, this is precisely the magnitude of 8, which agrees with the result
just derived. "41

35-5 Time-Varying Magnetic Fields

So far we have considered emfs induced by the relative motion of magnets


and coils. In this section we assume that there is no physical motion of
gross objects but that the magnetic field may vary with time. If a conduct-
ing loop is placed in such a time-varying field, the flux through the loop will
change and an induced emf will appear in the loop. This emf will set the
charge carriers in motion, that is, it will induce a current.
From a microscopic point of view we can say, equally well, that the chang-
ing flux of B sets up an induced electric field E at various points around the
loop. These induced electric fields are just as real as electric fields set up
by static charges and will exert a force F on a test charge q0 given by F =
qoE. Thus we can restate Faraday's law of induction in a loose but m-
formative way as: A changing magnetic field produces an electric field.
Sec. 35-5 TIME-VARYING MAGNETIC FIELDS 881

To fix these ideas, consider Fig. 35-10, which shows a uniform field of in-
duction B at right angles to the plane of the page. We assume that B is
increasing in magnitude at the same constant rate dB/dt at every point.
This could be done by causing the current in the windings of the electro-
magnet that establishes the field to increase with time in the proper way.
The circle of arbitrary radius r shown in Fig. 35-10 encloses, at any in-
stant, a flux ipB· Because this flux is changing with time, an induced emf
given by 8 = -dipB/dt will appear around the loop. The electric fields E
induced at various points of the loop must, from symmetry, be tangent to
the loop. Thus the electric lines of force that are set up by the changing
magnetic field are in this case concentric circles.
If we consider a test charge q0 moving around the circle of Fig. 35-10, the
work W done on it per revolution is, in terms of the definition of an emf,
simply 8q0 . From another point of view, it is (q0 E)(21rr), where q0 E is the
force that acts on the charge and 27rf is the distance over which the force
acts. Setting the two expressions for W equal and canceling q0 yields

8 = E21rr. (35-7)
In a more general case than that of Fig. 35-10 we must write

8 = f E·dl. (35-8)

If this integral is evaluated for the conditions of Fig. 35-10, we obtain Eq.
35-7 at once. If Eq. 35-8 is combined with Eq. 35-1 (8 = -dipB/dt),
Faraday's law of induction can be written as

f E·dl = - dipB'
dt
(35-9)

which is the form in which this law is expressed in Table 38-3 .

.,,,..,1(-x-x---....
/ .......

/ ........

/ x x x '\,
I \
x x x x x \
I \
Fig. 35-10 The induced electric fields
{xxR~rx \
at four points produced by an in-
creasing magnetic field. I xE x x x x x x x,
\
\x xx x x rI
\ I
{ x x x /
'
.... ~
........
_____
x x
E
x ,.,-x
'/
/
,X.._B
882 FARADAY'S LAW Chap. 35
..... Example 4. Let B in Fig. 35-10 be increasing at the rate dB/ dt. Let R be
the radius of the cylindrical region in which the magnetic field is assumed to exist.
What is the magnitude of the electric field Eat any radius r? Assume that dB/dt =
0.10 weber/m 2-sec and R = 10 cm.
(a) For r < R, the flux <I>s through the loop is
<I>s = B(1rr 2 ).
Substituting into Faraday's law (Eq. 35-9),
,{: d<I>s
yE·dl = - -dt
d<I>s dB
yields (E)(21rr) = - -dt = - 2
(1rr) dt.
1 dB
Solving for E yields E= - -r-·
2 dt
The minus sign is retained to suggest that the induced electric field E acts to oppose
the change of the magnetic field. Note that E(r) depends on dB/dt and not on B.
Substituting numerical values, assuming r = 5 cm, yields, for the magnitude of E,

E = dB
!r -d = (!)(5 X 10-2 meter) (0.10 2weber) = 2.5 X 10-3 volt/meter.
t m -sec
(b) For r >R the flux through the loop is

<Ps = f B·dS = B(1rR 2).

This equation is true because B·dS is zero for those points of the loop that lie outside
the effective boundary of the magnetic field.
From Faraday's law (Eq. 35-9),
d<I>s dB
(E)(21rr) = - dt = - (1rR 2 ) dt.

1 R 2 dB
Solving for E yields E=-2rdt.
These two expressions for E(r) yield the same result, as they must, for r = R. Figure
35-11 is a plot of the magnitude of E(r) for the numerical values given. ~

6
....
Q)
+-'
Q)
E
:;::.- 4
0
>
,:,::,
I 2
0
,-t

~
0 ',<,

0 10 20 30 40
r, cm
Fig. 35-11 Example 4. If the fringing of the field in Fig. 35-10 were to be taken into
account, the result would be a rounding of the sharp cusp at r = R ( = 10 cm).
Sec. 35-5 TIME-VARYING MAGNETIC FIELDS 883

Fig. 35-12 Showing the circular lines of


E from an increasing magnetic field. The
four loops are imaginary paths around
which an emf can be calculated.

In applying Lenz's law to Fig. 35-10, imagine that a circular conducting


loop is placed concentrically in the field. Since <I>B through this loop is in-
creasing, the induced current in the loop will tend to oppose this "change"
by setting up a magnetic field of its own that points up within the loop.
Thus the induced current i must be counterclockwise, which means that the
lines of the induced electric field E, which is responsible for the current,
must also be counterclockwise. If the magnetic field in Fig. 35-10 were
decreasing with time, the induced current and the lines of force of the in-
duced electric field E would be clockwise, again opposing the change in <l>B.
Figure 35-12 shows four of many possible loops to which Faraday's law
may be applied. For loops 1 and 2, the induced emf 8 is the same because
these loops lie entirely within the changing magnetic field and thus have the
same value of d<I>B/dt. Note that even though the emf 8 ( = f E·dl) is the
same for these two loops, the distribution of electric fields E around the per-
imeter of each loop, as indicated by the electric lines of force, is different.
For loop 3 the emf is less because <I>B and d<I>B/dt for this loop are less, and
for loop 4 the induced emf is zero.
The induced electric fields that are set up by the induction process are not associ-
ated with charges but with a changing magnetic flux. Although both kinds of electric
fields exert forces on charges, there is a difference between them. The simplest
manifestation of this difference is that lines of E associated with a changing magnetic
flux can form closed loops (see Fig. 35-12); lines of E associated with charges cannot
but can always be drawn to start on a positive charge and end on a negative charge.
Equation 29-5, which defined the potential difference between two points a and b, is
wb
Vb - Va = _a = - fbE·dl.
qo a
We have insisted that if potential is to have any useful meaning this integral (and
W ab) must have the same value for every path connecting a with b. This proved to be
true for every case examined in earlier chapters.
An interesting special case comes up if a and bare the same point. The path con-
necting them is now a closed loop; Va must be identical with Vb and this equation
reduces to
fE·dl = 0. (35-10)
884 FARADAY'S LAW Chap. 35

However, when changing magnetic flux is present, f E·dl is precisely not zero but
is, according to Faraday's law (see Eq. 35-9), -dif>B/dt. Electric fields associated
with stationary charges are conservative, but those associated with changing magnetic
fields are nonconservative; see Section 8-2. Electric potential, which can be defined
only for a conservative force, has no meaning for electric fields produced by induction.

35-6 The Betatron


The betatron is a device used to accelerate electrons to high speeds by
allowing them to be acted upon by induced electric fields that are set up by a
changing magnetic flux. It provides an excellent illustration of the "reality"
of such induced fields. The energetic electrons can be used for fundamental
research in physics or to produce penetrating X-rays which are useful in
cancer therapy and in industry.
Figure 35-13 shows the 100-Mev betatron at the General Electric Com-
pany. At this energy the electron speed is 0.999986c, where c is the speed
of light, so that relativistic mechanics must certainly be used in the analysis
of its operation. Figure 35-14 shows a vertical cross section through the
central part of the betatron to which the man in Fig. 35-13 is pointing.
The magnetic field in the betatron has several functions: (a) it guides the
electrons in a circular path; (b) it accelerates the electrons in this path;

Fig. 35-13 A 100-Mev betatron. M shows the magnet and D the region in which the
"doughnut" is located. (Courtesy General Electric Company.)
Sec. 35-6 THE BETATRON 885

Fig. 35-14 Cross section of a betatron, showing magnet M, coils C, and "doughnut" D.
Electrons emerge from the page at the left and enter it at the right.

(c) it keeps the radius of the orbit in which the electrons are moving a con-
stant; (d) it introduces the electrons into the orbit initially and removes
them from the orbit after they have reached full energy; and finally (e) it
provides a restoring force that resists any tendency for the electrons to leave
their orbit, either vertically or radially. It is remarkable that it is possible
to do all these things by proper shaping and control of the magnetic field.
The object marked D in Fig. 35-14 is an evacuated glass "doughnut"
inside which the electrons travel. Their orbit is a circle at right angles to the
plane of the figure. The electrons emerge from the plane at the left ( ·) and
enter it at the right (x). In the General Electric machine the radius of the
electron path is 33 in. The coils C and the 130-ton steel magnet shown in
Fig. 35-13 provide the magnetic flux that passes through the plane of this
orbit.
The current in coils C is made to alter periodically, 60 times/sec, to produce
a changing flux through the orbit, shown in Fig. 35-15. Here «l>B is taken as
positive when B is pointing up, as in Fig. 35-14. If the electrons are to
circulate in the direction shown, they must do so during the positive half-
cycle, marked ac in Fig. 35-15. The student should verify this (see Section
33-6). The electrons are accelerated by electric fields set up by the changing
flux. The direction of these induced fields depends on the sign of dif>B/dt
and must be chosen to accelerate, and not to decelerate, the electrons. Thus
only half the positive half-cycle in Fig. 35-15 can be used for acceleration;
it will prove to be ab.
The average value of dif>B/dt during the quarter-cycle ab is the slope of the
dashed line, or
1.8 weber
- - - - - = 430 volts.
4.2 X 10-3 sec
886 FARADAY'S LAW Chap. 35
Acceleration
Deceleration

Fig. 35-15 The flux through the


orbit of a betatron, during one
cycle. Rotation of the electrons
I in the desired direction (counter-
20
clockwise as viewed from above in
Fig. 35-14) is possible only during
half-cycle ac.

Counter-
clockwise Clockwise
rotation rotation

From Faraday's law (Eq. 35-1), this is also the emf in volts. The electron
will thus increase its energy by 430 ev every time it makes a trip around the
orbit in the changing flux. If the electron gains only 430 ev of energy per
revolution, it must make about 230,000 rev to gain its full 100 Mev. For an
orbit radius of 33 in., this corresponds to a length of path of some 750 miles.

The betatron provides a good example of the fact that electric potential has no
meaning for electric fields produced by induction. If a potential exists, it must be true
that, as Eq. 35-10 shows, f E·dl = 0 for any closed path. In the betatron, however,
this integral, evaluated around the orbit, is precisely not zero but is, in our example,
430 volts. It must not be thought, of course, that the betatron violates the conserva-
tion of energy principle. The gain in kinetic energy of the circulating electron (430
ev/rev) must be supplied by an identifiable energy source. It comes, in fact, from
the generator that energizes the magnet coils, thus providing the changing magnetic
field. The energy is transmitted to the electron through the intermediary of this
changing field.

~ Example 5. In the betatron of Fig. 35-13, which is the "accelerating" quarter-


cycle?
Let us assume that it is ab in Fig. 35-15, during which <I>n through the orbit is
increasing. If a conducting loop were placed to coincide with the orbit, an induced
current would appear in the loop to oppose the tendency of <I>n to increase. This
means that a magnetic field would be set up that would oppose the field of the large
magnet. Thus E would point outward at the right side of "doughnut" D in Fig.
35-14 and inward on the left side. The force ( - eE) acting on the electron is in the
opposite direction to E because of the negative charge of the electron. Thus the
tangential force acting on the electron is in the same direction as that at which it
circulates in its orbit; this means that the speed of the electron will increase, as
desired. The student should go through this same analysis carefully, assuming
(incorrectly, as it will turn out) that the accelerating half-cycle is be in Fig. 35-15
rather than ab.
Sec. 35-7 INDUCTION AND RELATIVE MOTION 887

35-7 Induction and Relative Motion


Faraday's law, in the form 8 = -difJB/dt, gives correctly the induced emf no
matter whether the change in ifJB is produced by moving a coil, moving a magnet,
changing the strength of a magnetic field, changing the shape of a conducting loop,
or in other ways. However, observers who are in relative motion with respect to
each other, even though they would all agree on the nun1erical value of the emf,
would give different microscopic descriptions of the induction process. In elec-
tromagnetic systems, as well as in m.:ichanical systems, it is important that the
state of motion of the observer with respect to his environment be made perfectly
clear.
y

x
1
v
x x x x x x

x x x x x x

y'

1---- - - - - --1
jx xx xx x 1

I x x x x x
I Fe_,
I x x + x x x

-v I
x x x x x

(b)

Fig. 35-16 A closed conducting loop in relative motion with respect to a magnetic
field. (a) An observer S, fixed with respect to the magnet that produces the field B,
sees the loop moving to the right. ( b) An observer S', fixed with respect to the loop,
sees the magnet moving to the left.

Figure 35-16 shows a closed loop which is caused to move at velocity v with
respect to a magnet that provides a uniform field Bin the region shown. We con-
sider first an observer, identified as S, who is at rest with respect to the magnet used
to establish the field B; see Fig. 35- l 6a. The induced emf in this case is called a
motional emf because the conducting loop is moving with respect to this observer.
Consider a positive charge carrier at the center of the left end of the conducting
loop. To observer S, this charge, constrained to move to the right along with the
888 FARADAY'S LAW Chap. 35
loop, is a charge q moving, for an instant at least when the loop has just been set
in motion, with a velocity v in the magnetic field B and as such it experiences a
sideways magnetic deflecting force given by Eq. 33-3a (F = qv X B). This
force causes the carriers to move upward along the conductor, so that they acquire
a drift velocity Vd also, as shown in Fig. 35-16a.
The resultant equilibrium speed of the carrier when the velocity v of the loop
has become constant is now V, which we find by adding v and va vectorially in
:Fig. 35-16a. The magnetic deflecting force Fm is, as always, at right angles to
the resultant velocity V of the carrier and is given by

Fm= qV X B (35-13)

Now Fm acting alone would tend to push the carriers through the left wall of the
conductor. Because this does not happen the conductor wall must exert a normal
force Non the carriers(see Fig. 35-16a)of magnitude such that va lies parallel to
the axis of the wire; in other words, N exactly cancels the horizontal component of
Fm, leaving only the component Fm cos e that lies along the direction of the con-
ductor. This component of force on the carrier is also cancelled out, this time
by the average force Fi associated with the internal collisions that the carrier
experiences as it drifts with (constant) speed Vd through the wire.
The kinetic energy of the charge carrier as it drifts through the wire remains
constant. This is consistent with the fact that the resultant force acting on the
charge carrier ( = Fm+ Fi+ N) is zero. The work done by Fm is zero because
magnetic forces, acting at right angles to the velocity of a moving charge, can do
no work on that charge. Thus the (negative) work done on the carrier by the
average internal collision force Fi must be exactly cancelled by the (positive)
work done on the carrier by the force N. In the last analysis N is exerted by the
agent who pulls the loop through the magnetic field, and the mechanical energy
expended by this agent appears as heat energy in the loop, as we have seen in
Section 35-4.
Let us then calculate the work dW done on the carrier in time dt by the force N;
it is
dW = N(v dt) (35-14)

in which v dt is the distance that the loop (and the carrier) have moved to the right
in Fig. 35-16a in time dt. We can write for N (see Fig. 35-16a and Eq. 35-13).

N = Fm sine = (qVB)(vd/V) = qBvd (35-15)

Substituting Eq. 35-15 into Eq. 35-14 yields

dW = (qBvd)(v dt)
= (qBv) (vd dt) = qBv dl (35-16)

in which dl( = vd dt) is the distance the carrier drifts along the conductor in
time dt.
The work done on the carrier as it makes a complete circuit of the loop is found
by integrating Eq. 35-16 around the loop and is

W = ¢dW = qBvl (35-17)


Sec. 35-7 INDUCTION AND RELATIVE MOTION 889
This follows because work contributions for the top and the bottom of the loops
are opposite in sign and cancel and no work is done in those portions of the loop
that lie outside the magnetic field.
An agent that does work on charge carriers, thus establishing a current in a
closed conducting loop, can be viewed as an emf. We can write, making use of
Eq. 35-17,

8 = W = qBvl = Blv (35-18)


q q '
which is, of course, the same result that we derived from Faraday's law of induction;
see Eq. 35-3. Thus a motional emf is intimately connected with the sideways
deflection of a charged particle moving through a magnetic field.
We now consider how the situation of Fig. 35-16 would appear to an observer S'
who is at rest with respect to the loop. To this observer, the magnet is moving to
the left in Fig. 35-16b with velocity -v, and the charge q is at rest as far as its
left-to-right motion is concerned. However S', like S, observes that the charge
drifts clockwise around the loop and he measures the same emf 8 that S measures.
S' accounts for this, at the microscopic level, by postulating that an electric field
E is induced in the loop by the action of the moving magnet. This induced field
E, which has the same origin as the induced fields that we discussed in Section
35-5, exerts a force on the charge carrier given by qE.
The induced field E, which exists in the end of the loop only, is associated with
an emf 8 and generates a current in the closed loop. Note that, in any closed loop
in which there is a current, an internal electric field must exist at every point at
which charges are moving. These electric fields however are set up by the emf,
as in the case of a closed loop connected to a battery, and are not induced by the
motion of the magnet. It is only this induced field E that we associate with the
emf, through the relation (Eq. 35-8)

8=E·dl
which reduced to
8 = El (35-19)

in this case. This is so because no induced electric field appears in the upper and
lower bars, because of the nature of their motions, and none appears in the part
of the loop outside the magnetic field.
The emfs given by Eqs. 35-19 and 35-18 must be identical because the relative
motion of the loop and the magnet ii; identical in the two cases shown in Fig. 35-16.
Equating these relations yields
El= Blv,

E = vB. (35-20a)

In Fig. 35-16b the vector E points upward along the axis of the left end of the
conducting loop because this is the direction in which positive charges are observed
to drift. The directions of v and B are clearly shown in this figure. We see,
then, that Eq. 35-20a is consistent with the more general vector relation

E = v X B. (35-20b)
890 FARADAY'S LAW Chap. 35
We have not proved Eq. 35-20b except for the special case of Fig. 35-16; neverthe-
less it proves to be true in general, that is, no matter what the angle between v
and B.
An experiment carried out in 1926 by the German physicist Wilhelm Wien (1864-
1928) gives concrete support to our descriptions of the effect of the motion of an
observer on the nature of the electric and magnetic fields that he observes. The
radiations emitted from atoms vary slightly in wavelength if the atom is immersed
in either a magnetic or an electric field. The atom can thus be used as a probe
to examine the nature of such fields. Wien fired a beam of atoms with velocity
v through a magnetic field B. The radiations emitted were identical in the distribu-
tion of their wavelengths with those that would have been emitted by a resting
atom immersed in (a) the magnetic fa 11 B and (b) an induced electric field E given
by Eq. 35-20b.
We interpret Eq. 35-20b in the following way: Observer S fixed with respect to
the magnet is aware only of a magnetic field. The force to him arises from the
motion of the charges through B. Observer S' fixed on the charge carrier is aware
of an electric field E also and attributes the force on the charge (at rest with respect
to him initially) to the electric field. S says the force is of purely magnetic origin
and S' says the force is of purely electric origin. From the point of view of S,
the induced emf is given by §(v X B) · di. From the point of view of S', the
same induced emf is given by §E · di, where E is the ( induced) electric vector
that he observes at points along the circuit.
For a third observer S" who judges that both the magnet and the loop are
moving, the forc0 tending to move charges around the loop is neither purely
electric nor purely magnetic, but a bit of each. In summary, in the equation

F/q=E+vXB,

different observers form different assessments of E, B, and v but, when these are
combined, all observers form the same assessment of F /q and all obtain the same
value for the induced emf in the loop (this depends only on the relative motion).
That is, the total force is the same for all observers, but each observer forms a
different estimate of the separate electric and magnetic forces contributing to the
same total force .
.... Example 7. In Fig. 35-16 assume that B = 2.0 webers/meter2, l = 10 cm,
and v = 1.0 meter/sec. Calculate (a) the induced electric field observed by S',
and (b) the emf induced in the loop.
(a) The electric field, which is apparent only to observer S', is associated with
the moving magnet and is given in magnitude (see Eq. 35-20a) by

E = vB
= (1.0 meter/sec)(2.0 webers/meter 2 )
= 2.0 volt/meter.

(b) Observer S would calculate the induced (motional) emf from

8 = Blv
= (2.0 webers/meters 2) (1.0 X 10- 1 meter) (1.0 meter/sec)
= 0.20 volt.
Chap. 35 QUESTIONS 891

Observer S' would not regard the emf as motional and would use the relationship

8 = El

= (2.0 volt/meter)(l.O X 10- 1 meter)

= 0.20 volt.
As must be the case, both observers agree as to the numerical vailue of the emf. <II(

QUESTIONS

1. The north pole of a magnet is moved away from a metallic ring, as in Fig. 35-17. In
the part of the ring farthest from the reader, which way does the current point?

Fig. 35-17

2. Eddy currents. A sheet of copper is placed in a magnetic field as shown in Fig. 35-18.
If we attempt to pull it out of the field or push it further in, an automatic resisting force
appears. Explain its origin. (Hint: Currents, called eddy currents, are induced in the
sheet in such a way as to oppose the motion.)

---
/
/
----- x x x ------ "'-- '--.

/
/
x
/
x x x x x
"x ~ B
I \
I x x x \
x x x x
I \
I x x x x x
I
I x x x x x
I
I
\ x x x x x
\
\ x x x x x I
\ I
\ x x x x x x /
"x /
/
""" " '-- '----- x x x/
---- ------- -----

Fig. 35-18
892 FARADAY'S LAW Chap. 35
3. Electromagnetic shielding. Consider a conducting sheet lying in a plane perpendicular
to a magnetic field B, as shown in Fig. 35-19. (a) If B suddenly changes, the full change
in Bis not immediately detected in region P. Explain. (b) If the resistivity of the sheet is
zero, the change is not ever detected at P.
Explain. (c) If B changes periodically at
high frequency and the conductor is made
of a material of low resistivity, the re-
gion near P is almost completely shielded
from the changes in flux. Explain. (d)
Is such a conductor useful as a shield
from static magnetic fields? Explain.

XX X•X
B

P Conducting xx:\: xx x
==*'"-~,#\¥%:%\\\\\'"'"* sheet
x :1:::
I

x
x,x
x x

Fig. 35-19 Fig. 35-20

4. Magnetic damping. A strip of copper is mounted as a pendulum about O in Fig.


35-20. It is free to swing through a magnetic field normal to the page. If the strip has
slots cut in it as shown, it can swing freely through the field. If a strip without slots is
substituted, the vibratory motion is strongly damped. Explain. (Hint: Use Lenz's law;
consider the paths that the charge carriers in the strip must follow if they are to oppose the
motion.)
5. Two conducting loops face each other a distanced apart (Fig. 35-21). An observer
sights along their common axis. If a clockwise current i is suddenly established in the
larger loop, (a) what is the direction of the induced current in the smaller loop? (b)
What is the direction of the force (if any) that acts on the smaller loop?

-0-
I
·--·-

~d_j R

Fig. 35-21 Fig. 35-22

6. What is the direction, if any, of the conventional current through resistor R in Fig.
35-22 (a) immediately after switch Sis closed, (b) some time after switch S was closed, and
(c) immediately after switch S is opened. (d) When switch S is held closed, which end
of the coil acts as a north pole?
Chap. 35 QUESTIONS 893

7. A current-carrying solenoid is moved toward a conducting loop as in Fig. 35-23.


What is the direction of circulation of current in the loop as we sight toward it as shown?

----"v>

Fig. 35-23

8. If the resistance R in the left-hand circuit of Fig. 35-24 is increased, what is the
direction of the induced current in the right-hand circuit?

R A

Fig. 35-24 Fig. 35-25

9. In Fig. 35-25 the movable wire is moved to the right, causing an induced current as
shown. What is the direction of B in region A?
10. A loop, shown in Fig. 35-26, is removed from the magnet by pulling it vertically
upward. (a) What is the direction of the induced current? (b) Is a force required to remove
the loop? (c) Does the total amount of Joule heat produced in removing the loop depend
on the time taken to remove it?

Fig. 35-26

11. A magnet is dropped down a long vertical copper tube. Show that, even neglecting
air resistance, the magnet will reach a constant terminal velocity.
12. A magnet is dropped from the ceiling along the axis of a copper loop lying flat on the
floor. If the falling magnet is photographed with a time sequence camera, what differences,
if any, will be noted if (a) the loop is at room temperature and (b) the loop is packed in dry
ice?
894 FARADAY'S LAW Chap. 35
13. A copper ring and a wooden ring of the same dimensions are placed so that there is
the same changing magnetic flux through each. How do the induced electric fields in each
ring compare?
14. In Fig. 35-12 how can the induced emfs around paths 1 and 2 be identical? The
induced electric fields are much weaker near path 1 than near path 2, as the spacing of the
lines of force shows. See also Fig. 35-11.
15. In a certain betatron the electrons rotate counterclockwise as seen from above. In
what direction must the magnetic field point and how must it change with time while the
electron is being accelerated?
16. Why can a betatron be used for acceleration only during one-quarter of a cycle?
17. To make the electrons in a betatron orbit spiral outward, would it be necessary to
increase or to decrease the central flux? Assume that B at the orbit remains essentially
unchanged.
18. A cyclotron is a so-called resonance device. Does a betatron depend on resonance?

PROBLEMS

1. A hundred turns of insulated copper wire are wrapped around an iron cylinder of
cross-sectional area 0.001 meter 2 and are connected to a resistor. The total resistance in
the circuit is 10 ohms. If the longitudinal magnetic induction in the iron changes from
1 weber /meter 2 in one direction to 1 weber /meter 2 in the opposite direction, how much
charge flows through the circuit?
· 2. In Fig. 35-27 a closed copper coil with 100 turns and a total resistance of 5.0 ohms
is placed outside a solenoid like that of Example 1. If the current in the solenoid is changed
as in that example, what current appears in the coil?

Fig. 35-27

3. A circular loop of wire 10 cm in diameter is placed with its normal making an angle
of 30° v·ith the direction of a uniform 5000-gauss magnetic field. The loop is "wobbled"
so that its normal rotates about the field direction at the constant rate of 100 rev /min;
the angle between the normal and the field direction ( = 30°) remains unchanged during
this process. What emf appears in the loop?
4. A uniform field of induction B is normal to the plane of a circular ring 10-cm in
diameter made of #10 copper wire (diameter = 0.10 in.). At what rate must B change
with time if an induced current of 10 amp is to appear in the ring?
5. A uniform field of induction B is changing in magnitude at a constant rate dB /dt.
You are given a mass m of copper which is to be drawn into a wire ot radius r and formed
into a circular loop of radius R. Show that the induced current in the loop does not depend
Chap. 35 PROBLEMS 895
on the size of the wire or of the loop and, assuming B perpendicular to the loop, is given by

m dB
i = ~---,
47rp0 dt

where p is the resistivity and f, the density of copper.


6. You are given 50 cm of #18 copper wire (diameter = 0.040 in.). It is formed into a
circular loop and placed at right angles to a uniform magnetic field that is increasing with
time at the constant rate of 100 gauss/sec. At what rate is Joule heat generated in the
loop?
7. A small bar magnet is pulled rapidly through a conducting loop, along its axis.
Sketch qualitatively (a) the induced current and (b) the rate of Joule heating as a function
of the position of the center of the magnet. Assume that the north pole of the magnet
enters the loop first and that the magnet moves at constant speed. Plot the induced
<turrent as positive if it is clockwise as viewed along the path of the magnet.
's",Alternating current generator. A rectangular loop of N turns and of length a and width
bis rotated at a frequency Pin a uniform field of induction B, as in Fig. 35-28. (a) Show
that an induced emf given by
G = 21CPNbaB sin 21CPt = GO sin 21CPt

appears in the loop. This is the principle of the commercial alternating-current generator.
(b) Design a loop that will produce an emf with GO = 150 volts when rotated at 60 rev/sec
in a field of magnetic induction of 5000-gauss.

fig. 35-28

9. A stiff wire bent into a semicircle of radius R is rotated with a frequency Pin a uni-
form field of induction B, as shown in Fig. 35-29. What are the amplitude and fre-
quency of the induced voltage and of the induced current when the internal resistance
of the meter Mis RM and the remainder of the circuit has negligible resistance?

x x x x x x

x x )( x x x
R

t
x > x ~x x

x x x x x x x

x x x x x x x

x x x x x x x

'--------i M i--------'

fig. 35-29
896 FARADAY'S LAW Chap. 35
10. A circular copper disk 10 cm in diameter rotates at 1800 rev /min about an axis
through its center and at right angles to the disk. A uniform field of induction B of 10,000
gauss is perpendicular to the disk. What potential difference develops between the axis
of the disk and its rim?
11. Figure 35-30 shows a copper rod moving with velocity v parallel to a long straight
wire carrying a current i. Calculate the induced emf in the rod, assuming that v = 5.0
meters/sec, i = 100 amp, a = 1.0 cm, and b = 20 cm.

Fig. 35-30 Fig. 35-31

12. A metal wire of mass rn slides without friction on two rails spaced a distance d
apart, as in Fig. 35-31. The track lies in a vertical uniform field of induction B. (a) A
constant current i flows from generator G along one rail, across the wire, and back down the
other rail. Find the velocity (speed and direction) of the wire as a function of time, as-
suming it to be at rest at t = 0. (b) The generator is replaced by a battery with constant
ernf 8. The velocity of the wire now approaches a constant final value. What is this ter-
minal speed? (c) What is the current in part (b) when the terminal speed has been reached?
13. In Fig. 35-32 the magnetic flux through the loop perpendicular to the plane of the
coil and directed into the paper is varying according to the relation
<f>s = 6t2 + 7t + 1,
where <f>R is in milliwebers (1 milliweber = 10-3 weber) and tis in seconds. (a) What is
the magnitude of the emf induced in the loop when t = 2 sec? (b) What is the direction
of the current through R?
x x x x x

x x

x x x x Ax x x x
D
x x x x x x x x
50cm
x x x x x x
c
R x x Bx x x x

Fig. 35-32 Fig, 35-33

14. In Fig. 35-33 a conducting rod AB makes contact with the metal rails AD and BC
which are 50 cm apart in a uniform magnetic field of induction 1.0 weber /meter 2 perpen-
dicular to the plane of the paper as shown. The total resistance of the circuit ABCD is 0.4
ohm (assumed constant). (a) What is the magnitude and direction of the emf induced in
the rod when it is moved to the left with a velocity of 8 meters/sec? (b) What force is re-
quired to keep the rod in motion? (c) Compare the rate at which mechanical work is done
by the force F with the rate of development of heat in the circuit.
15. A square wire of length l, mass rn, and resistance R slides without friction down
parallel conducting rails of negligible resistance, as in Fig. 35-34. The rails are connected
Chap. 35 PROBLEMS 897
to each other at the bottom by a resistanceless rail parallel to the wire, so that the wire
and rails form a closed rectangular conducting loop. The plane of the rails makes an
angle e with the horizontal, and a uniform vertical field of magnetic induction B exists
throughout the region. (a) Show that the wire acquires a steady-state velocity of magni-
tude
mgR sine
v = .
B 2t2 cos 2 e

(b) Prove that this result is consistent with the conservation-of-energy principle. (c)
What change, if any, would there be if B were directed down instead of up?

Fig. 35-34 Fig. 35-35

16. Prove that if the flux of magnetic induction through the coil of N turns of Fig.
35-35 changes in any way from <1>1 to <1>2, then the charge q that flows through the circuit
of total resistance R is given by
N(<1>2 - <1>1)
q=
R
~igure 35-36 shows a uniform field of induction B confined to a cylindrical volume
of radius R. Bis decreasing in magnitude at a constant rate of 100 gauss/sec. What is
the instantaneous acceleration (direction and magnitude) experienced by an electron
placed at a, at b, and at c? Assume r = 5.0 cm. (The necessary fringing of the field
beyond R will not change your answer as long as there is axial symmetry about a perpen-
dicular axis through b.)

// x
........ -------....
x x ....... ,
' "-
//
/x
/
/

x
x
x
x
t\ x
x R x
;>, x\
II ~
r \I
IX xx bx x xi
I r (B I
\x x x I x x x/
\ l I
\\ x x xa x x //
' '-
' x
.................. ______x x_....,,.,.,. /
/ /

Fig. 35-36

18. A uniform magnetic field of induction B fills a cylindrical volume of radius R. A


metal rod of length l is placed as shown in Fig. 35-37. If Bis changing at the rate dB/dt,
898 FARADAY'S LAW Chap. 35
show that the emf that is produced by the changing magnetic field and that acts between
the ends of the rod is given by

G = ~ ~~ R2 - GY.
/
/
/
/
----- --
x x x
......
'
' '-
I
I
/
x x x x x ""\
\
Ix x x x x x ,::"'x \
I
I B \
I
I x x x x x x I
I i I
\ R
\X x x x x x I
I
\
x x x x I
""'- /

' ....... .......x__


/
/

---- x,,... /

Fig. 35-37

19. Some measurements of the maximum magnetic induction as a function of radius


for the General Electric Company betatron are as follows:

r, cn1 B, gauss r, cn1 B, gauss


0 4000 81.2 4090
10.2 9500 83.7 4000
68.2 9500 88.9 3810
73.2 5280 91.4 3720
75.2 4510 93.5 3600
77.3 4280 95.5 3400

Show by graphical analysis that the relation B= 2BR is satisfied at the orbit radius,

• R = 84 cm. (Hint: Note that B


_
= -12 LR
B(r)(21rr) dr and evaluate the integral
o 1rR
graphically.)
20. Prove that the electric field E in a charged parallel-plate capacitor cannot drop
abruptly to zero as one moves at right angles to it, as suggested by the arrow in Fig. 35-38
(see point a). In actual capacitors fringing of the lines of force always occurs, which means
that E approaches zero in a continuous and gradual way. See Problem 34-28. (Hint:
Apply Faraday's law to the rectangular path shown by the dashed lines.)
+q
.......
r -,
I I

E --,>.
I 1¥'(a
I I
I I
L _J

- q ..A

Fig. 35-38
Inductance
CHAPTER 36

36-1 Inductance
If two coils are near each other, a current i in one coil will set up a flux <I>B
through the second coil. If this flux is changed by changing the current, an
induced emf will appear in the second coil according to Faraday's law.
However, two coils are not needed to show an inductive effect. An induced
emf appears in a coil if the current in that same coil is changed. This is called
self-induction and the electromotive force produced is called a self-induced
emf. It obeys Faraday's law of induction just as other induced emfs do.
Consider first a "close-packed" coil, a toroid, or the central section of a
long solenoid. In all three cases the flux <I>B set up in each turn by a current
i is essentially the same for every turn. Faraday's law for such coils (Eq.
35-2) d(N<I>B)
8=----
dt
shows that the number of flux linkages N<I>B (N being the number of turns)
is the important characteristic quantity for induction. For a given coil,
provided no magnetic materials such as iron are nearby, this quantity is
proportional to the current i, or
N<I>B = Li, (36-2)
in which L, the proportionality constant, is called the inductance of the device.
From Faraday's law (see Eq. 36-1) the induced emf can be written as
d(N<I>s)
8=---- (36-3a)
dt
899
900 INDUCTANCE Chap. 36

8
Written in the form L=--, (36-3b)
di/dt

this relation may be taken as the defining equation for inductance for coils
of all shapes and sizes, whether or not they are close-packed and whether or
not iron or other magnetic material is nearby. It is analogous to the defining
relation for capacitance, namely
q
c =-·
v
If no iron or similar materials are nearby, L depends only on the geometry
of the device. In an inductor (symbol / OOOO\.) the presence of a magnetic
field is the significant feature, corresponding to the presence of an electric
field in a capacitor.
The unit of inductance, from Eq. 36-:3b, is the volt-sec/amp. A special
name, the henry, has been given to this combination of units, or

1 henry = 1 volt-sec/amp.
The unit of inductance is named after Joseph Henry (1797-1878), an Ameri-
can physicist and a contemporary of Faraday. Henry independently dis-
covered the law of induction at about the same time Faraday did. The units
millihenry (1 mh = 10-3 henry) and microhenry (1 µh = 10-6 henry) are
also commonly used.
The direction of a self-induced emf can be found from Lenz's law. Suppose
that a steady current i, produced by a battery, exists in a coil. Let us sud-
denly reduce the (battery) emf in the circuit to zero. The current i will
start to decrease at once; this decrease in current, in the language of Lenz's
law, is the "change" which the self-induction must oppose. To oppose the
falling current, the induced emf must point in the same direction as the
current, as in Fig. 36-la. When the current in a coil is increased, Lenz's
law shows that the self-induced emf points in the opposite direction to that of
the current, as in Fig. 36-lb. In each case the self-induced emf acts to op-

Fig. 36-1 In (a) the current i is


decreasing and in (b) it is increas-
ing. The self-induced emf GL
opposes the change in each case.
Sec. 36-2 CALCULATION OF INDUCTANCE 901

pose the change in the current. The minus sign in Eq. 36-3 shows that 8
and di/dt are opposite in sign, since Lis always a positive quantity.

36-2 Calculation of Inductance


It has proved possible to make a direct calculation of capacitance in terms
of geometrical factors for a few special cases, such as the parallel-plate ca-
pacitor. In the same way, it is possible to calculate the self-inductance L
for a few special cases.
For a close-packed coil with no iron nearby, we have, from Eq. 36-2,

(36-4)

Let us apply this equation to calculate L for a section of length l near the
center of a long solenoid. The number of flux linkages in the length l of the
solenoid is
NiJ>s = (nl) (BA),
where n is the number of turns per unit length, B is the magnetic induction
inside the solenoid, and A is the cross-sectional area. From Eq. :34-7, B is
given by

Combining these equations gives


NiJ>s = µ 0 n 2 liA.
Finally, the inductance, from Eq. 36-4, is
NiJ>B 2
L = -.- = µ 0n lA. (36-5)
i

The inductance of a length l of a solenoid is proportional to its volume (lA)


and to the square of the number of turns per unit length. K ote that it de-
pends on geometrical factors only. The proportionality to n 2 is expected.
If the number of turns per unit length is doubled, not only is the total num-
ber of turns N doubled but also the flux through each turn q,B is also doubled,
an over-all factor of four for the flux linkages NiJ>B, hence also a factor of
four for the inductance (Eq. 36-4).
~ Example 1. Derive an expression for the inductance of a toroid of rectangular
cross section as shown in Fig. 36-2. Evaluate for N = 103 , a = 5.0 cm, b = 10 cm,
and h = 1.0 cm.
The lines of B for the toroid are concentric circles. Applying Ampere's law,

fB·dl = µoi,

to a circular path of radius r yields


(B)(2-n) = µoioN,
where N is the number of turns and io is the current in the toroid windings; recall that
902 INDUCTANCE Chap. 36

Fig. 36-2 Example 1. A cross


section of a toroid, showing the
current in the windings and the
magnetic field.

i in Ampere's law is the total current that passes through the path of integration.
Solving for B yields
B = µ,oioN.
21rr
The flux cf> s for the cross section of the toroid is

cf>s f
= B·dS = .C(B)(h dr) = ib µ,;~~ h dr

µ,oioNhfb -
=~~-
dr= -µ,oioNh b
~ - n1 - ,
21r a r 21r a
where h dr is the area of the elementary strip shown in the figure.
The inductance follows from Eq. 36-4, or

L = Ncf>s = µ,oN 2h In~--


io 271' a
Substituting numerical values yields
L = (47r X 10-7 weber/amp-m)(l0 3)2(1.0 X 10-2 meter) In 10 X 10-2 meter
21r 5 X 10- 2 meter
= 1.4 X 10-3 weber/amp = 1.4 mh. .....

36-3 An LR Circuit
In Section 32-8 we saw that if an emf 8 is suddenly introduced, perhaps
by using a battery, into a single loop circuit containing a resistor R and a
capacitor C the charge does not build up immediately to its final equilibrium
value ( = C8) but approaches it in an exponential fashion described by
Eq. 32-15, or
(36-6)
The delay in the nse of the charge is described by the capacitative time
constant Tc, defined from
Tc= RC. (36-7)
If in this same circuit the battery emf 8 is suddenly removed, the charge
does not immediately fall to zero but approaches zero m an exponential
fashion, described by Eq. 32-18b, or
(36-8)
The same time constant Tc describes the fall of the charge as well as its rise.
AN LR CIRCUIT 903

10
a S c R

Fig. 36-3 An LR eiccuit.

(.,

An analogous delay in the rise or fall of the current occurs if an emf f, is


suddenly introduced into, or removed from, a single loop circuit containing
a resistor R and an inductor L. When the switch S in Fig. 36-3 is closed on
a, for example, the current in the resistor starts to rise. If the inductor were
not present, the current would rise rapidly to a steady value 0/R. Because
of the inductor, however, a self-induced emf f,L appears in the circuit and,
from Lenz's law, this emf opposes the rise of current, which means that it
opposes the battery emf f, in polarity. Thus the resistor responds to the
difference between two emfs, a constant one f, due to the battery and a var-
iable one f,L ( = -L di/dt) due to self-induction. As long as this second emf
is present, the current in the resistor will be less than 0/R.
As time goes on, the rate at which the current increases becomes less rapid
and the self-induced emf GL, which is pr'oportional to di/dt, becomes smaller.
Thus a ti:ne delay is introduced, and the current in the circuit approaches
the value c,/R asymptotically.
When the switch S in Fig. :36-:3 is thrown to a, the circuit reduces to that
of Fig. 36-4. Let us apply the loop theorem, starting at x in this figure and
going clockwise around the loop. For the direction of current shown, x will
be higher in potential than y, which means that we encounter a drop in
potential of -iR as we traverse the resistor. Point y is higher in potential
than point z because, for an increasing current, the induced emf will oppose
the rise of the current by pointing as shown. Thus as we traverse the in-
ductor from y to z we encounter a drop in potential of - L(di/ dt). We en-
counter a rise in potential of + f, in traversing the battery from z to x. The
loop theorem thus gives
di
-iR - L - + r, = 0
dt
di
or L- + iR = c,. (36-9)
dt

-- i

Fig. 36-4 The circuit of Fig. 36-3 just after switch


8 is closed on a.
L
904 INDUCTANCE Chap. 36
Equation 36-9 is a differential equation involving the variable i and its first
derivative di/dt. We seek the function i(t) such that when it and its first
derivative are substituted in Eq. 36-9 the equation is satisfied.
Although there are formal rules for solving various classes of differential
equations (and Eq. :36-9 can, in fact, be easily solved by direct integration,
after rearrangement) we often find it simpler to guess at the solution, guided
by physical reasoning and by previous experience. Any proposed solution
can be tested by substituting it in the differential equation and seeing
whether this equation reduces to an identity.
The solution to Eq. 36-9 is, we assert,
e
i · -_ - (1 - e-RtfL) . (36-10)
R
To test this solution by substitution, we find the derivative di/dt, which is

di G -Rt/L
- =-e (36-11)
dt L
Substituting i and di/dt into Eq. 36-9 leads to an identity, as the student
can easily verify. Thus Eq. 36-10 is a solution of Eq. 36-9. Figure 36-5
shows how the potential difference V R across the resistor ( = iR; see Eq.
36-10) and V L across the inductor ( = L di/dt; see Eq. 36-11) vary with
time for particular values of e, L, and R. "The student should compare this
figure carefully with the corresponding figure for an RC circuit (Fig. 32-11).
We can rewrite Eq. 36-10 as

(36-12)

in which r L, the inductive time constant, is given by


TL= L/R. (36-13)
The student should note the correspondence between Eqs. 36-12 and 36-6.
To show that the quantity TL ( = L/R) has the dimensions of time, we put

1 henry = 1 h~ry ( 1 v\Jt-sec ) (1 o~-a~p) = 1 sec.


ohm o~ 1 he\ry-a~p 1 v~t
The first quantity in parenthesis is a conversion factor based on the defining
equation for inductance [L = -e/(di/dt); Eq. :36-3b]. The second conver-
sion factor is based on the relation V = iR.
The physical significance of the time constant follows from Eq. 36-12.
If we putt = TL = L/R in this equation, it reduces to
e e e
i = - (1 - e- 1 ) = (1 - 0.37) - = 0.63 -·
R R R
Thus the time constant TL is that time at which the current in the circuit
Sec. 36-3 AN LR CIRCUIT 905

4
en
~
a,
c.
E
.!!!
.E
...;- 2

2 4 6 8 10
t, milliseconds
(a)

-
0
>
en

2 4 6 8 10
t, milliseconds
(b)
Fig. 36-5 If in Fig. 36-3 we assume that R = 2000 ohms, L = 4 henrys and e = 10
volts, then (a) shows the variation of i with t during the current buildup after switch Sis
closed on a, and (b) the variation of V L with t. The time constant is LIR = 2.0 X 10-3
sec.

will reach a value within 1/e (about 37%) of its final equilibrium value (see
Fig. 36-5).
If the switch S in Fig. 36-3, having been left in position a long enough
for the equilibrium current 8/R to be established, is thrown to b, the effect
is to remove the battery from the circuit. The differential equation that
906 INDUCTANCE Chap. 36
governs the subsequent decay of the current in the circuit can be found by
putting 8 = 0 in Eq. 36-9, or
di
L-+ iR = 0. (36-14)
dt
The student can show by the test of substitution that the solution of this
differential equation is
8
i = _ e-t/TL (36-15)
R
Just as for the RC circuit, the behavior of the circuit of Fig. 36-3 can be investi-
gated experimentally, using a cathode-ray oscilloscope. If switch S in this figure is
thrown periodically between a and b, the applied emf alternates between the values
8 and zero. If the terminals of an oscilloscope are connected across b and c in Fig.
36-3, the oscilloscope will display the waveform of this applied emf on its screen, as
in Fig. 36-6c. There is a practical problem that R and L for an actual coil are
physically inseparable; we are assuming an "ideal" resistor and an "ideal" inductor.
If the terminals of the oscilloscope are connected across the resistor, the waveform
displayed (Fig. 36-6a) will be that of the current in the circuit, since the potential
drop across R, which determines the oscilloscope deflection, is given by V R = iR.
During the intervals marked inc in Fig. 36-6. the current is increasing and the wave-
form (see Eq. 36-12) is given bv
V R ( = iR) = 8(1 - e-t/TL).

k-inc~dec~inc~dec-4-inc~

Fig. 36-6 Oscilloscope photograph showing the variation with time of (a) the potential
drop V R across the resistor, (b) the potential drop V L across the inductor, and (c) the
applied emf 8. During the intervals marked inc the current is increasing; during those
marked dee it is decreasing. Compare with Fig. 32-13. (Courtesy E. K. Hege.)
Sec. 36-4 ENERGY AND THE MAGNETIC FIELD 907
During the intervals marked dee, the current is decreasing and V R (see Eq. 36-15) is
given by

Note that both the growth and the decay of the current are delayed.
If the oscilloscope terminals are connected across the inductor, the screen will show
a plot of the potential difference across it as a function of time (Fig. 36-6b). While
the current is increasing, the equation of the trace (see Eq. 36-11) should be

When the current is decreasing, V Lis given in terms of the time derivative of Eq.
36-15 and is
VL
di
dt(= L-) =
-se-tfTL
.

Note that V L is opposite in sign when the current is increasing (di/dt positive)
and when it is decreasing (di/dt negative), as is true also for the induced emf SL[=
-L(di/dt) = -VLJ,
Examination of Fig. 36-6 shows that at any instant the sum of curves a and b
always yields curve c. This is an expected consequence of the loop theorem, as Eq.
36-9 shows.

l)IJIJ.. Example 2. A solenoid has an inductance of 50 henrys and a resistance of 30

ohms. If it is connected to a 100-volt battery, how long will it take for the current
to reach one-half its final equilibrium value?
The equilibrium value of the current is reached as t ---+ c,:i; from Eq. 36-12 it is
8/R. If the current has half this value at a particular time t 0, this equation becomes

I_§_ = _§_ (I _ e-tofTL)


2R R .
Solving for to yields
L
to= 7Lln 2 = 0.69e,'

Putting TL = L/R and using the values given, this reduces to

50 henrys)
t 0 = 0.69r L = 0.69 ( 30 ohms = 1.2 sec.

36-4 Energy and the Magnetic Field


In Section 30-7 we saw that the electric field could be viewed as the site of
stored energy, the energy per unit volume being given, in a vacuum, by

uE = 1
zEo E2 ,

where Eis the electric field strength at the point in question. Although this
formula was derived for a parallel-plate capacitor, it holds for all kinds of
electric field configurations.
Energy can also be stored in a magnetic field. For example, two parallel
wires carrying currents in the same direction attract each other, and to pull
908 INDUCTANCE Chap: 36
them further apart work must be done. It is useful to think that this ex-
pended energy is stored in the magnetic field between and around the wires.
The energy can be recovered from the field if the wires are allowed to move
back to their original separation. In the electrostatic case the same argu-
ment was applied to the pulling apart of two unlike charges and in the gravi-
tational case to the pulling apart of two masses.
To derive a quantitative expression for the storage of energy in the mag-
netic field, consider Fig. 36-4, which shows a source of emf 8 connected to a
resistor R and an inductor L.
di
8 = iR + L-, (36-9)
dt
is the differential equation that describes the growth of current in this circuit.
We stress that this equation follows immediately from the loop theorem and
that the loop theorem in turn is an expression of the principle of conservation
of energy for single-loop circuits. If we multiply each side of Eq. 36-9 by i,
we obtain
di
Si = i 2 R + Li - , (36-16)
dt
which has the following physical interpretation in terms of work and energy:
1. If a charge dq passes through the seat of emf 8 in Fig. 36-4 in time dt,
the seat does work on it in amount 8 dq. The rate of doing work is (8 dq)/dt,
or 8i. Thus the left term in Eq. 36-16 is the rate at which the seat of emf
delivers energy to the circuit.
2. The second term in Eq. 36-16 is the rate at which energy appears as
Joule heat in the resistor.
3. Energy that does not appear as Joule heat must, by our hypothesis, be
stored in the magnetic field. Since Eq. 36-16 represents a statement of the
conservation of energy for LR circuits, the last term must represent the rate
dUB/dt at which energy is stored in the magnetic field, or

dUB di
- - = Li-· (36-17)
dt dt
We can write this as dUB = Li di.
Integrating yields
UB = ius ii
dUB = Li di = !Li2, (36-18)

which represents the total stored magnetic energy in an inductance L carry-


ing a current i.
This relation can be compared with the expression for the energy asso-
ciated with a capacitor C carrying a charge q, namely
1 q2
UE = --·
2C
ENERGY AND THE MAGNETIC FIELD 909

Here the energy is stored in an electric field. In each case the expression for
the stored energy was derived by setting it equal to the work that must be
done to set up the field.
.... Example 3. A coil has an inductance of 5.0 henrys and a resistance of 20 ohms.
If a 100-volt emf is applied, what energy is stored in the magnetic field after the
current has built up to its maximum value 8/ R?
The maximum current is given by
. 8 100 volts
i = R= 20 ohms = 5 ·0 amp.

The stored energy is given by Eq. 36-18:


UB = !Li 2 = !(5.0 henrys)(5.0 amp) 2 = 63 joules.
Note that the time constant for this coil ( = L/R) is 0.25 sec. After how many time
constants will half of this equilibrium energy be stored in the field?
Example 4. A 3.0-henry inductor is placed in series with a 10-ohm resistor, an
emf of 3.0 volts being suddenly applied to the combination. At 0.30 sec (which is
one inductive time constant) after the contact is made, (a) what is the rate at which
energy is being delivered by the battery?
The current is given by Eq. 36-12, or

i = %(1 _ e-tlrL),

which at t = 0.30 sec ( = TL) has the value

. (3.0 volts) 1)
i =
10 ohms (1 - e- = 0.189 amp.

The rate P 8 at which energy is delivered by the battery is


Pe= 8i
= (3.0 volts)(0.189 amp)
= 0.567 watt.
(b) At what rate does energy appear as Joule heat in the resistor? This is given by
PJ = i 2R
= (0.189 amp) 2(10 ohms)
= 0.357 watt.
(c},At what rate PB is energy being stored in the magnetic field? This is given by
the last term in Eq. 36-16, which requires that we know di/dt. Differentiating Eq.
36-12 yields

At t = TL we have
di I
( 33.0h
volts ) _ 1 _
e - 0.37 amp sec.
dt .0 enrys
910 INDUCTANCE Chap. 36
From Eq. 36-17, the desired rate is
p _dUB_L_di
B - dt - 1, dt
= (3.0 henrys)(0.189 amp)(0.37 amp/sec)
= 0.210 watt.
Note that as required by the principle of conservation of energy (see Eq. 36-16)
Pi= PJ + PB,
or 0.567 watt = 0.357 watt + 0.210 watt
= 0.567 watt.

36-5 Energy Density and the Magnetic Field


We now derive an expression for the density of energy u in a magnetic
field. Consider a length l near the center of a very long solenoid; Al is the
volume associated with this length. The stored energy must lie entirely
within this volume because the magnetic field outside such a solenoid is
essentially zero. Moreover, the stored energy must be uniformly distributed
throughout the volume of the solenoid because the magnetic field is uniform
everywhere inside.
UB
Thus we can write UB = -
Al
or, smce
IL·2
2 i
UB = --·
Al
To express this in terms of the magnetic field, we can substitute for Lin this
equation, using the relation L = µ 0n 2 lA (Eq. 36-5). Also we can solve Eq.
34-7 (B = µ 0 in) for i and substitute in this equation. Doing so yields
finally
1 B2
UB = --· (36-19)
2 µo

This equation gives the energy density stored at any point (in a vacuum or in
a nonmagnetic substance) where the magnetic induction is B. The equation
is true for all magnetic field configurations, even though it was derived by
considering a special case, the solenoid. Equation 36-19 is to be compared
with Eq. 30-27,
E2 ,
uE = 2I t:o,, (K = 1) (36-20)
which gives the energy density (in a vacuum) at any point in an electric
field. K ote that both UB and UE are proportional to the square of the appro-
priate field quantity, B or E.
The solenoid plays a role with relationship to magnetic fields similar to the
role the parallel-plate capacitor plays with respect to electric fields. In each
case we have a simple device that can be used for setting up a uniform field
ENERGY DENSITY AND THE MAGNETIC FIELD 911

throughout a well-defined region of space and for deducing, in a simple way,


some properties of these fields.
~ Example 5. A long coaxial cable (Fig. 36-7) consists of two concentric cylinders
with radii a and b. Its central conductor carries a steady current i, the outer con-
ductor providing the return path. Calculate the energy stored in the magnetic
field between the conductors for a length l of such a cable.

Fig. 36-7 Example 5. Cross section of


a coaxial cable, showing steady currents
in the central and outer conductors.

In the space between the two conductors Ampere's law

f B·dl = µoi

leads to (B)(21rr) = µei

or B = µoi.
21rr
Ampere's law shows further that the magnetic field is zero for points outside the
outer conductor (why?). Magnetic fields exist inBide each of the conductors but
we are concerned here only with energy stored between the conductors.

The energy density for points between the conductors, from Eq. 36-19, is

u = _1_ B2 = _!_ ( µoi ) 2


2µo 2µo 21rr
Consider a volume element dV consisting of a cylindrical shell whose radii are r and
r +
dr and whose length is Z. The energy dU contained in it is

dU = u dV = µoi 2 (21rrl)(dr) = µoi 2l dr_


81r 2r 2 41r r
The total stored magnetic energy is found by integration, or

u = fdu = µoi2Z
41r
Ia
b dr
r
= µoi2Z ln ~'
41r a

which is the desired expression.


912 INDUCTANCE Chap. 36
Example 6. Compare the energy required to set up, in a cube 10 cm on edge,
(a) a uniform electric field of 10 5 volts/meter and (b) a uniform magnetic field of 1
weber /meter 2 ( = 10 4 gauss). Both these fields would be judged reasonably large but
they are readily available in the laboratory.
(a) In the electric case we have, where VO is the volume of the cube,

UE = UE Vo= i-1:0E 2Vo


= (0.5)(8.9 X 10- 12 coul2/nt-m 2)(10 5 volts/meter)2(0.l meter) 3
= 4.5 X 10-5 joule.
(b) In the magnetic case, from Eq. 36-19, we have
B2 (1 weber/meter 2)2(0.1 meter) 3
UB = UB Vo = - Vo = - - - - - - - - - - - - - -
2µ0 (2)(41r X 10-7 weber/amp-m)
= 400 joules.
In terms of fields normally available in the laboratory, much larger amounts of energy
can be stored in a magnetic field than in an electric one, the ratio being about 107 in
this example. Conversely, much more energy is required to set up a magnetic field
of reasonable laboratory magnitude than is required to set up an electric field of
similarly reasonable magnitude. ,._

QUESTIONS

1. Two coils are connected in series. Does their equivalent inductance depend on their
geometrical relationship to each other?
2. Is the inductance per unit length for a solenoid near its center (a) the same as, (b) less
than, or (c) greater than the inductance per unit length near its ends?
3. Two solenoids, A and B, have the same diameter and length and contain only one
layer of windings, with adjacent turns touching, insulation thickness being negligible.
Solenoid A contains many turns of fine wire and solenoid B contains fewer turns of heavier
wire. (a) Which solenoid has the larger inductance? (b) Which solenoid has the larger
inductive time constant?
4. If the flux passing through each turn of a coil is the same, the inductance of the coil
may be computed from L = N<PB/i (Eq. 36-4). How might one compute L for a coil for
which this assumption is not valid?
5. If a current in a source of emf is in the direction of the emf, the energy of the source
derreases; if a current is in a direction opposite to the emf (as in charging a battery), the
energy of the source increases. Do these statements apply to the inductor in Fig. 36-la
and 36-lb?
6. Show that the dimensions of the two expressions for L, N<P3/i (Eq. 36-4) and 8/
(di/dt) (Eq. 36-3b), are the same.
7. You are given N turns of wire connected in series. How should the turns be arranged
to obtain the maximum self-inductance?
Chap. 36 PROBLEMS 913
8. Does the time required for the current in a particular LR circuit to build up to any
given fraction of its equilibrium value depend on the value of the applied emf?
9. A steady current is set up in a coil with a very large inductive time constant. When
the current is interrupted with a switch, a heavy arc tends to appear at the switch blades.
Explain. (Note: Interrupting currents in highly inductive circuits can be dangerous.)
10. In an LR circuit like that of Fig. 36-4 can the self-induced emf ever be larger than
the battery emf?
11. In an LR circuit like that of Fig. 36-4 is the current in the resistance always the same
as the current in the inductance?
12. In the circuit of Fig. 36-3 the self-induced emf is a maximum at the instant the
switch is closed on a. How can this be since there is no current in the inductance at this
instant?
13. Give some arguments to show that energy can be stored in a magnetic field.
14. The switch in Fig. 36-3 is thrown from a to b. What happens to the energy stored
in the inductor?
15. In a toroid is the energy density larger near the inner radius or near the outer radius?

PROBLEMS

1. A IO-henry inductor carries a steady current of 2.0 amp. How can a 100-volt self-
induced emf be made to appear in the inductor?
,,%.'.Two inductances L 1 and L 2 are connected in series. and are separated by a large dis-
tance. (a) Show that the equivalent inductance Lis L1 + L2. (b) Why must their sep-
aration be large?
3. Show that if two inductors with equal inductance L are connected in parallel the
equivalent inductance of the combination 1s !L. The inductors are separated by a large
distance .
..4:/Two long parallel wires whose centers are a distance d apart carry equal currents in op-
posite directions. Show that, neglecting the flux within the wires themselves, the induc-
tance of a length l of such a pair of wires is given by
µold-a
L = -ln--,
7r a
where a is the wire radius. See Example 4, Chapter 34.
_,,,&:· A long thin solenoid can be bent into a ring to form a toroid. Show that if the
solenoid is long and thin enough the equation for the inductance of a toroid (see Example
1) reduces to that for a solenoid (Eq. 36-5).
6. A solenoid is wound with a single layer of #10 copper wire (diameter, 0.10 in.). It is
4.0 cm in diameter and 2.0 meters long. What is the inductance per unit length for the
solenoid near its center? Assume that adjacent wires touch and that insulation thickness
is negligible.
/7: The inductance of a close-packed coil of 400 turns is 8 mh. What is the magnetic
flux through the coil when the current is 5 X 10-3 amp?
8. A wooden toroidal core with'a square cross section has an inner radius of 10 cm and
an outer radius of 12 cm. It is wound with one layer of n8 wire (diameter, 0.040 in.;
resistance, 160 ft/ohm). What are (a) the inductance and (b) the inductive time
constant? Ignore the thickness of the insulation.
_.--9: The current in an LR circuit builds up to one--third of its steady-state value in 5.0
sec. What is the inductive time constant?
10. How many "time constants" must we wait for the current in an LR circuit to build
up to within 0.1 per cent of its equilibrium value?
11. The switch S in Fig. 36-3 is thrown from b to a. After one inductive time constant
show that (a) the total energy transformed to Joule heat in the resistor is 0.168S2rL/R
914 INDUCTANCE Chap. 36
and that (b) the energy stored in the magnetic field is 0.200S 2rL/R. (c) Show that the
equilibrium energy stored in the magnetic field is 0.500c 2rL/R.
12. Show that the inductive time constant TL can also be defined as tpe time that would
be required for the current in an LR circuit to reach its equilibrium value if it continued to
increase at its initial rate.
)-3: A 50-volt potential difference is suddenly applied to a coil with L = 50 mh and
R = 180 ohms. At what rate is the current increasing after 0.001 sec?
14. A coil with an inductance of 2.0 henrys and a resistance of 10 ohms is suddenly
connected to a resistanceless battery with S = 100 volts. At 0.1 sec after the connection
is made, what are the rates at which (a) energy is being stored in the magnetic field, (b)
Joule heat is appearing, and (c) energy is being delivered by the battery?
15. A coil with an inductance of 2.0 henrys and a resistance of 10 ohms is suddenly
connected to a resistanceless battery with 8 = 100 volts. (a) What is the equilibrium
current? (b) How much energy is stored in the magnetic field when this current exists in
the coil?
16. Prove that when switch Sin Fig. 36-3 is thrown from a to ball the energy stored in
the inductor appears as Joule heat in the resistor.
17. A circular loop of wire 5.0 cm in radius carries a current of 100 amp. What is the
energy density at the center of the loop?
18. What is the magnetic energy density at the center of a circulating electron in the
hydrogen atom (see Example 9, Chapter 34)?
~ A long wire carries a current i, uniformly distributed over a cross-section of the
wire. Show that the magnetic energy per unit length stored within the wire equals
µ 0i 2 /l61r. Note that it does not depend on the wire diameter.
20. Show that the self-inductance for a length l of a long wire associated with the flux
inside the wire only is µol/81r. Assume uniform distribution of the current, as in
Problem 19.
21. The coaxial cable of Example 5 has a = 1.0 mm, b = 4.0 mm, and c = 5.0 mm
(c is the radius of the outer surface of the outer conductor). It carries a current of 10 amp
in the inner conductor and an equal but oppositely directed return current in the outer
conductor. Calculate and compare the stored magnetic energy per meter of cable length
(a) within the central conductor, (b) in the space between the conductors, and (c) within
the outer conductor.
22. A length of #10 copper wire carries a current of 10 amp. Calculate (a) the magnetic
energy density and (b) the electric energy density at the surface of the wire. The wire
diameter is 0.10 in. and its resistance per unit length is 1.0 ohm/1000 ft.
-~·' What must be the strength of a uniform electric field if it is to have the same energy
density as that possessed by a 5000-gauss magnetic field?
Magnetic Properties
of Matter
CHAPTER 37

37-1 Poles and Dipoles


In electricity the isolated charge q is the simplest structure that can exist.
If two such charges of opposite sign are placed near each other, they form an
electric dipole, characterized by an electric dipole moment p. In magnetism
isolated magnetic "poles," which would correspond to isolated electric
charges, apparently do not exist. The simplest magnetic structure is the
magnetic dipole, characterized by a magnetic dipole moment µ. Table 34-1
summarizes some characteristics of electric and magnetic dipoles.
A current loop, a bar magnet, and a solenoid of finite length arc examples
of magnetic dipoles. Their magnetic dipole moments can be measured by
placing the dipole in an external magnetic field B, measuring the torque T
that acts on it, and computingµ from Eq. 33-11, or
T = µ XB. (37-1)
Alternatively, we can measure B due to the dipole at a point along its axis a
(large) distance r from its center and compute µ from the expression in
Table 34-1, or
B = µo !!:___, (37-2)
271" r3
Figure 37-1, which shows iron filings sprinkled on a sheet of paper under
which there is a bar magnet, suggests that this dipole might be viewed as two
"poles" separated by a distance d. However, all attempts to isolate these
poles fail. If the magnet is broken, as in Fig. 37-2, the fragments prove to be
dipoles and not isolated poles. If we break up a magnet into the electrons
915
916 MAGNETIC PROPERTIES OF MATTER Chap. 37

Fig. 37-1 A bar magnet is a mag-


netic dipole. The iron filings sug-
gest the lines of B in Fig. 37-4a.
(Courtesy Physical Science Study
Committee.)

and nuclei that make up its atoms, it will be found that even these ele-
mentary particles are magnetic dipoles. Figure 37-3 contrasts the electric
and the magnetic characteristic of the free electron.
All electrons have a characteristic "spin" angular momentum about a
certain axis, which has the value of
Ls = 0.52723 X 10-34 joule-sec.
This is suggested by the vector Ls in Fig. 37-3b. Such a spinning charge can
be viewed classically as being made up of infinitesimal current loops. Each
such loop is a tiny magnetic dipole, its moment being given by (Eq. 33-10)
µ = NiA, (37-3)

Fig. 37-2 If a bar magnet is


broken, each fragment becomes
a small dipole.
Sec. 37-1 POLES AND DIPOLES 917

where i is the equivalent current in each infinitesimal loop and A is the loop
area. The number of turns, N, is unity for each loop. The magnetic dipole
moment of the spinning charge can be found by integrating over the mo-
ments of the infinitesimal current loops that make it up; see Problem 2.
Although this model of the spinning electron is too mechanistic and is not
in accord with modern quantum physics, it remains true that the magnetic
dipole moments of elementary particles are closely connected with their
intrinsic angular momenta, or spins. Those particles and nuclei whose spin
angular momentum is zero (the a-particle, the pion, the 0 16 nucleus, etc.)
have no magnetic dipole moment. The "intrinsic" or "spin" magnetic mo-

(a) (b)

Fig. 37-3 (a) The lines of E and (b) the lines of B for an electron. The magnetic dipole
moment of the electron, µ.z, is directed opposite to the spin angular momentum vector, L 8 •

ment of the electron must be distinguished from any additional magnetic


moment it may have because of its orbital motion in an atom; see Example 2.
liJi. Example 1. Devise a method for measuring µ for a bar magnet.
(a) Place the magnet in a uniform external magnetic field B, with µ making an
angle e with B. The magnitude of the torque acting on the magnet (see Eq. 37-1)
is given by
T = µB sine.
Clearlyµ can be learned if T, B, and e are measured.
(b) A second technique is to suspend the magnet from its center of mass and to
allow it to oscillate about its stable equilibrium position in the external field B. For
small oscillations, sine can be replaced bye and the equation just given becomes
T = -(µB)e = -Ke,
where K is a constant. The minus sign has been inserted to show that T is a restoring
torque. Since T is proportional toe, the condition for simple angular harmonic motion
is met. The frequency v is given by the reciprocal of Eq. 15-24, or

v = 2~~i 2~/P·
=
With this equationµ can be found from the measured quantities v, B, and I.
918 MAGNETIC PROPERTIES OF MATTER Chap. 37
Example 2. An electron in an atom circulating in an assumed circular orbit of
radius r behaves like a tiny current loop and has an orbital magnetic dipole moment *
usually represented by µz. Derive a connection between µ 1 and the orbital angular
momentum Lz.
Newton's second law (F = ma) yields, if we substitute Coulomb's law for F,

1 e2 mv 2
- - - =ma= - -
41T'E:o r 2 r

or v= (37-4)
The angular velocity w is given by

The current for the orbit is the rate at which charge passes any given point, or

The orbital dipole moment µ 1 is given from Eq. 37-3 if we put N = 1 and A = 1T'r 2 , or

(37-5)

The orbital angular momentum Lz is

Lz = (mv)r.

Combining with Eq. 37-4 leads to

Finally, eliminating r between this equation and Eq. 37-5 yields

/J,z = Lz ( 2;)·

which shows that the orbital magnetic moment of an electron is proportional to its
orbital angular momentum.
For r = 5.1 X 10- 11 meter, which corresponds to hydrogen in its normal state,
we have from Eq. 37-5
e2 r-;;:-
µz = 4 ~;;m
= (1.6 X 10- 19 coul) 2 / 5.1 X 10- 11 meter
4 \J (11")(8.9 X 10- 12 coul2/nt-m 2)(9.1 X 10- 31 kg)
= 9.1 X 10-24 amp-m 2 • _,.

* This must not be confused with the magnetic dipole moment µs of the electron spin,
which is also present.
I

/
I
\
"
-"°"°

(a) (b) (c)

Fig. 37-4 Lines of B (a) for a bar magnet and (b) for a short solenoid. (c) Lines of E for an electric dipole. At large enough distances all
three fields vary like those for a dipole. The four dashed curves are interse<>tions with the plane of the figure of closed Gaussian surfaces.
Note that <f>B equals zero for (a) or (b). <f>E equals zero for surfaces like II in (c), which do not contain any charge, but <f>E is not zero for sur-
faces like I.
920 MAGNETIC PROPERTIES OF MATTER Chap. 37
37-2 Gauss's Law for Magnetism
Gauss's law for magnetism, which is one of the basic equations of electro-
magnetism (see Table 38-3), is a formal way of stating a conclusion that
seems to be forced on one by the facts of magnetism, namely, that isolated
magnetic poles do not exist. This equation asserts that the flux 'PB through
any closed Gaussian surface must be zero, or

'PB = f B·dS = 0, (37-6)

where the integral is to be taken over the entire closed surface. We con-
trast this with Gauss's law for electricity, which is

The fact that a zero appears at the right of Eq. 37-6, but not at the right of
Eq. 37-7, means that in magnetism there is no counterpart to the free charge
q in electricity.
Figure 37-4a shows a Gaussian surface that encloses one end of a bar
magnet. Note that the lines of B enter the surface for the most part inside
the magnet and leave it for the most part outside the magnet. There is
thus an inward (or negative) flux inside the magnet and an outward (or
positive) flux outside it. The total flux for the whole surface is zero. In
Section 37-7 the interested student will find a fuller discussion of why the
lines of B for a bar magnet are drawn in the way shown.
Figure 37-4b shows a similar surface for a solenoid of finite length which,
like a bar magnet, is also a magnetic dipole. Here, too, 'PB equals zero.
Figures 37-4a and b show clearly that there are no "sources" of B; that is,
there are no points from which lines of B emanate. Also, there are no "sinks"
of B; that is, there are no points toward which B converges. In other words,
there are no free magnetic poles.
Figure 37-4c shows a Gaussian surface (I) surrounding the positive end of
an electric dipole. Here there is a net flux of the lines of E. There is a
"source" of E; it is the charge q. If q is negative, we have a "sink" of E
because the lines of E end on negative charges. For surfaces like surface II
in Fig. 37-4c for which the charge inside is zero, the flux of E over the surface
is also zero.

37-3 Paramagnetism
Magnetism as we know it in our daily experience is an important but
special branch of the subject called ferromagnetism; we discuss this in Sec-
tion 37-5. Here we discuss a weaker form of magnetism called paramag-
netism.
For most atoms and ions, the magnetic effects of the electrons, including
both their spins and orbital motions, exactly cancel so that the atom or ion
Sec. 37-3 PARAMAGNETISM 921

is not magnetic. This is true for the rare gases such as neon and for ions *
such as cu+, which makes up ordinary copper. These materials do not ex-
hibit paramagnetism. For other atoms or ions the magnetic effects of the
electrons do not cancel, so that the atom as a whole has a magnetic dipole
moment µ. Examples are found among the so-called transition elements,
such as Mn++, the rare earths, such as Gd+++, and the actinide elements,
such as u++++.
If a sample of N atoms, each of which has a magnetic dipole moment µ,
is placed in a magnetic field, the elementary atomic dipoles tend to line up
with the field. For perfect alignment, the sample as a whole would have a
magnetic dipole moment of Nµ. However, the aligning process is seriously
interfered with by the collisions that take place between the atoms if the
sample is a gas and by temperature vibrations if the sample is a solid. The
importance of this thermal agitation effect may be measured by comparing
two energies: one ( = JkT) is the mean kinetic energy of translation of a
gas atom at temperature T; the other ( = 2µ,B) is the difference in energy
between an atom lined up with the magnetic field and one pointing in the
opposite direction. As Example 3 shows, the effect of the collisions at
ordinary temperatures and fields is very great. The sample acquires a mag-
netic moment when placed in an external magnetic field, but this moment
is usually much smaller than the maximum possible moment N µ.
~ Example 3. A paramagnetic gas, whose atoms (see Example 2) have a magnetic
dipole moment of about 10-23 amp-m 2 , is placed in an external magnetic field of
magnitude 1 weber/meter 2 • At room temperature (T = 300°K) calculate and com-
pare Ur, the mean kinetic energy of translation ( = }kT), and UB, the magnetic
energy ( = 2µ,B) :
Ur = fkT = (t )(1.38 X 10-23 joule/°K) (300°K) = 6 X 10-21 joule,
UB = 2µ,B = (2)(10-23 amp-m 2 )(1 weber/meter 2 ) = 2 X 10-23 joule.
Because Ur equals 300 UB, we see that energy exchanges in collisions can interfere
seriously with the alignment of the dipoles with the external field. ....
If a specimen of a paramagnetic substance is placed in a nonuniform magnetic
field, such as that near the pole of a strong magnet, it will be attracted toward the
region of higher field, that is, toward the pole. We can understand this by drawing
an analogy with the corresponding electric case of Fig. 37-5, which shows a dielectric
specimen (a sphere) in a nonuniform electric field. The net electric force points to
the right in the figure and is
Fe = q(E + fiE) - q(E - fiE) = q(2fiE),
which can be written as

Fe = (q fix) 2fiE = p (2fiE) r-..; p (~~) .


fix fix dx max
Here p ( = q fix) is the induced electric dipole moment of the sphere. In the differen-
tial limit of a very small sphere (2fiE/ fix) approaches (dE/dx)max, the gradient of
the electric field at the center of the sphere.
* Cu+ indicates a copper atom from which one electron has been removed; AI+++ indi-
cates an aluminum atom from which three electrons have been removed, etc.
922 MAGNETIC PROPERTIES OF MATTER Chap. 37
In the corresponding magnetic case we have, by analogy,

F dB)
- µ ( -- (37-8)
m - dx rnax
Thus, by measuring the magnetic force Fm that acts on a small paramagnetic speci-
men when it is placed in a nonuniform magnetic field whose field gradient (dB/dx)max
is known, we can learn its magnetic dipole moment µ. The magnetization M is
defined as the magnetic moment per unit volume, or

M=~,
v
where V is the volume of the specimen. It is a vector because µ., the dipole moment
of the specimen, is a vector.
In 1895 Pierre Curie (1859-1906) discovered experimentally that the magnetization
M of a paramagnetic specimen is directly proportional to B, the effective value of
magnetic induction in which the specimen is placed, and inversely proportional to
the temperature, or
B
M = C-, (37-9)
T
in which C is a constant. This equation is known as Curie's law. The law is physi-
cally reasonable in that increasing B tends to align the elementary dipoles in the speci-
men, that is, to increase M, whereas increasing T tends to interfere with this align-
ment, that is, to decrease M. Curie's law is well verified experimentally, provided
that the ratio B/T does not become too large.

E = E 0 - AE E = E 0 + AE
Fig. 37-5 A dielectric sphere in a nonuniform electric field.
DIAMAGNETISM 923

0.75

x
• 1.30° K
.. 2.00°K
"'
E
x 3.00°K
~ 0.50
• 4.21° K
~ -Theory

0.25

10 20 30 40
B/T, 10 3 gauss/degree

Fig. 37-6 The ratio M /M max for a paramagnetic salt (chromium potassium alum) in
various magnetic fields and at various temperatures. The curve through the experimental
points is a theoretical curve calculated ±rom modern quantum physics. (From measure-
ments by W. E. Henry.)

M cannot increase without limit, as Curie's law implies, but must approach a
value M max ( = µN /V) corresponding to the complete alignment of the N dipoles
contained in the volume V of the specimen. Figure 37-6 shows this saturation
effect for a sample of CrK(S04)2 · 12H20. The chromium ions are responsible for all
the paramagnetism of this salt, all the other elements being paramagnetically inert.
To achieve 99.5% saturation, it is necessary to use applied magnetic fields as high
as 50,000 gauss and temperatures as low as l.3°K. Note that for more readily
achievable conditions, such as B = 10,000 gauss and T = 10°K, the abscissa in
Fig. 37-6 is only 1.0 so that Curie's law would appear to be well obeyed for this
and for all lower values of B/T. The curve that passes through the experimental
points in this figure is calculated from a theory based on modern quantum physics;
it is in excellent agreement with experiment.

37-4 Diamagnetism
In 1846 Michael Faraday discovered that a specimen of bismuth brought near to
the pole of a strong magnet is repelled. He called such substances diamagnetic.
Diamagnetism, present in all substances, is such a feeble effect that its presence is
masked in substances made of atoms that have a net magnetic dipole moment,
that is, in paramagnetic or ferromagnetic substances.
Figures 37-7a and b show an electron circulating in a diamagnetic atom at angular
frequency w0 in an assumed circular orbit of radius r. Each electron is moving under
the action of a centripetal force FE of electrostatic origin where, from Newton's
second law,
FE = ma = mwo 2r. (37-10)

Each rotating electron has an orbital magnetic moment, but for the atom as a whole
the orbits are randomly oriented so that there is no net magnetic effect. In Fig.
37-7a, for example, the magnetic dipole moment l,Lz points into the page; in Fig.
924 MAGNETIC PROPERTIES OF MATTER Chap. 37
-e -e

FE FE

w~
B=O
~
(a) (b)

x x x FB x x x x x x x x x x
V2
x x x x x x x x x x
FB
x x x x x x x x x x
FE FE
x x x x x x x x x x x x x x

x x x x x x x x x x x x x x

x x x x x·-
~x
1 = w0 - Aw)
x x x x x
Ax 2 (= w0 + Aw)
x x x x x x x~x x x x x x x

B
(c) (d)

Fig. 37-7 (a) An electron circulating in an atom. (b) An electron circulating in the op-
posite direction. (c) A magnetic field is introduced, decreasing the linear speed of the
electron in (a), that is, v1 < vo. (d) The magnetic field increases the linear speed of the
electron in (b ), that is, v2 > vo.

37-7b it points out and the net effect for the two orbits sh01vn is cancellation. This
cancellation is also shown at the left in Fig. 37-8.
If an external field B is applied as in Fig. 37-7c and d, an additional force, given
by -e(v x B), acts on the electron. This magnetic force acts always at right angles
to the direction of motion; its magnitude is
FB = evB = e(wr)B. (37-11)
The student should show that in Fig. 37-7c FB and FE point in opposite directions
and that in Fig. 37-7d they point in the same direction. Note that since the centrip-
etal force changes when the magnetic field is applied (the radius can be shown
to remain constant), the angular velocity must also change; thus w in Eq. 37-11
differs from wo in Eq. 37-10.
Applying Newton's second law to Figs. 37-7c and d, and allowing for both direc-
tions of circulation, yields for the resultant forces on the electrons
FE± FB = ma = mw 2r.
Sec. 37-4 DIAMAGNETISM 925
Substituting Eqs. 37-10 and 37-11 into this equation yields

or w2 =F ( ~ ) w - wo 2 = 0. (37-12)

This quadratic equation can be solved for w, the new angular velocity. Rather
than doing this, we take advantage of the fact (presented without proof; see Problem
7) that w differs only slightly from wo, even in the strongest external magnetic fields.
Thus
w = wo 6.w + (37-13)
where 6.w <<wo. Substituting this equation into Eq. 37-12 yields
[wo2 + 2wo 6.w + (6.w)2] ± [{Jwo + {36.w] - wo 2 = 0,
where /3 is a convenient abbreviation for eB/m. The two terms w02 cancel each other;
the terms (6.w) 2 and {JAw are small compared to the remaining terms and may be set
equal to zero with only small error. This leads, as an excellent approximation, to
,,..__., 1 _ eB
6.w = =F 2/J - =F 2m (37-14)

eB
or, from Eq. 37-13, W = WO =f= --·
2m
Thus the effect of applying a magnetic field is to increase or decrease (depending on
the direction of circulation) the angular velocity. This, in turn, increases or decreases
the orbital magnetic moment of the circulating electron (see Example 2).
In Fig. 37-7c the angular velocity is reduced (because the centripetal force is
reduced) so that the magnitude of the magnetic moment is reduced. In Fig. 37-7d,
however, the angular velocity is increased so that the magnitude of µ 1 is increased.
These effects are shown on the right in Fig. 37-8, where it will be noted that the two
magnetic moments no longer cancel.
We see that if a magnetic field B is applied to a diamagnetic substance (zero net
magnetic moment in absence of applied field), a magnetic moment will be induced
whose direction (out of the plane of Fig. 37-7) is opposite to B; see also Fig. 37-8.
This is precisely the reverse of paramagnetism, in which the (permanent) magnetic
dipoles tend to point in the same direction as the applied field.
We can now understand why a diamagnetic specimen is repelled when brought
near to the pole of a strong magnet. If the pole is a north pole, there exists a non-
uniform magnetic field of induction with B pointing away from the pole. If a sphere

Fig. 37-8 The magnetic moments of the two b d µl + t:.µ


oppositely circulating electrons in an atom cancel
when there is no external magnetic field, as in (a),
but do not cancel when a field is applied, as in (b).
Note that the resultant moment in (b) points in
the opposite direction to B. Compare carefully
with Fig. 37-7.

B=O
(a) (b)
926 MAGNETIC PROPERTIES OF MATTER Chap. 37
made of a diamagnetic material (bismuth, say) is brought near to this pole, the mag-
netization M that is induced in it points toward the pole, that is, opposite to B.
Thus the side of the sphere closest to the magnet behaves like a north pole and is
repelled by the nearby north pole of the magnet. For a paramagnetic sphere, the
vector M points along the direction of B and the side of the sphere nearest to the
magnet is a south pole, which is attracted to the north pole of the magnet.
~ Example 4. Calculate the change in magnetic moment for a circulating electron,
as described in Example 2, if a magnetic field of induction B of 2.0 webers/meter2
acts at right angles to the plane of the orbit.
We obtainµ from Eq. 37-3, or

µ = NiA = (l)(ev)(7rr 2 ) = (1) (~:) (7rr 2) = !er 2w.


The change in µ is Aµ = !er 2 Aw
eB) e2Br 2
or, from Eq. 37-14, Aµ = ± !er 2 ( 2m = ± 4m ·
Substituting numbers yields
A _ (1.6 X 10- 19 coul) 2(2.0 webers/meter 2)(5.l X 10- 11 meter) 2
µ - ± (4)(9.1 x 10- 31 kg)
= ±3.7 X 10-29 amp-m 2 •
In Example 2 the moment µ 1 was 9.1 X 10-24 amp-m 2, so that the change induced by
even a strong external magnetic field is rather small, the ratio Aµ/ µz being about
4 x 10-6 . ..

37-5 Ferromagnetism
For five elements (Fe, Co, Ni, Gd, and Dy) and for a variety of alloys of
these and other elements a special effect occurs which permits a specimen
to achieve a high degree of magnetic alignment in spite of the randomizing
tendency of the thermal motions of the atoms. In such materials, described
as ferromagnetic, a special form of interaction called exchange coupling occurs
between adjacent atoms, coupling their magnetic moments together in rigid
parallelism.* Modern quantum physics successfully predicts that this will
occur only for the five elements listed. If the temperature is raised above a
certain critical value, called the Curie temperature, the exchange coupling
suddenly disappears and the materials become simply paramagnetic. For
iron the Curie temperature is 1043°K. Ferromagnetism is evidently a
property not only of the individual atom or ion but also of the interaction
of each atom or ion with its neighbors in the crystal lattice (see Fig. 21-5)
of the solid.
Figure 37-9 shows a magnetization curve for a specimen of iron. To obtain
such a curve, we form the specimen, assumed initially unmagnetized, into a
ring and wind a toroidal coil around it as in Fig. 37-10, to form a so-called
Rowland ring. When a current i is set up in the coil, if the iron core is not
present, a field of induction is set up within the toroid given by (Eq. 34-4)
(37-15)
* Exchange coupling, a purely quantum effect, cannot be "explained" in terms of
classical physics.
Sec. 37-5 FERRO MAGNETISM 927

1.0

0.8
)(

"'
E
:i 0.6
Fig. 37-9 A magnetization curve
for iron.
ES.
~ 0.4

0.2

2 4 6 8 10 12 14
B 0 , 10- 4 weber/meter2

where_ n is the number of turns per unit length for the toroid. Although
this formula was derived for a long solenoid, it can be applied to a toroid
if d « r in Fig. 37-10. Because of the iron core, the actual value of Bin the
toroidal space will exceed B 0 , by a large factor in many cases, since the
elementary atomic dipoles in the core line up with the applied field B 0 ,
thereby setting up their own field of induction. Thus we can write

(37-16)
where BM is the magnetic induction due to the specimen; it is proportional
to the magnetization M of the specimen. Often BM» B 0 .
The field B 0 is proportional to the current in the toroid and can be cal-
culated readily, using Eq. 37-15; B can be measured in a way that is de-
scribed below. An experimental value for BM can be derived from Eq.

Fig. 37-10 A Rowland ring,


~,,,,i,~-B showing a secondary coil S.

d
928 MAGNETIC PROPERTIES OF MATTER Chap. 37

Fig. 37-11 (a) B for a Rowland


ring as the current in the wind-
ings is increased from zero during
an interval t.t. (b) The corre-
sponding induced current in the
secondary coil. Both curves are
idealized; in practice, the sharp
corners would be rounded off.

37-16. It has a maximum value BM.max corresponding to complete align-


ment of the atomic dipoles in the iron. Thus we can plot, as in Fig. 37-9,
the fractional degree of alignment ( = BM/BM, max) as a function of Bo. For
this specimen a value of 96.5% saturation is achieved at B 0 = 0.13 weber/
meter 2 ( = 1300 gauss; this point is about 16 ft to the right of the origin in
the figure); increasing B 0 to 1.0 weber/meter2 ( = 10,000 gauss; about 120
ft to the right in Fig. 37-9) increases the fractional saturation only to
97.7%.
To measure Bin the system of Fig. 37-10, let the current in the toroid windings be
increased from zero to i. The flux through the secondary coil Swill change by BA,
where A is the area of the toroid. While the flux is changing, an induced emf will
appear in coil S, according to Faraday's law. For simplicity, we assume that the
current in the toroid is so adjusted that B increases linearly with time for an interval
D..t, as shown in Fig. 37-lla. The emf in coil S during this interval, from Faraday's
law,* will then be
b..<I>s NBA
8 = N~- = ---,
D..t D..t
where N is the number of turns in coil S. This emf will set up a current is in coil S
given by
. 8 NBA
is= R = R D..t

B = (i., At)R = qR ,
or
NA NA
in which R is the resistance of coil Sand is At is the charge q that passes through this
coil during time At. If a so-called ballistic galvanometer is connected to S, its deflection
will be a measure of the charge q. Thus it is possible to find B for any value of the
current i in the toroid windings. In practice, it is not necessary that the curve B(t)
in Fig. 37-lla be linear during the interval At.
* We ignore the minus sign because we are concerned only with the magnitude of 8.
Sec. 37-5 FERRO MAGNETISM 929

b
Fig. 37-12 A magnetization curve (ab) for a speci-
men of iron and an associated hysteresis loop
(ebcde).

Magnetization curves for ferromagnetic materials do not retrace themselves as we


increase and then decrease the toroid current. Figure 37-12 shows the following
operations with a Rowland ring: (1) starting with the iron unmagnetized (point a),
increase the toroid current until Bo ( = µoni) has the value corresponding to point b;
(2) reduce the current in the toroid winding back to zero (point c); (3) reverse the
toroid current and increase it in magnitude until point d is reached; (4) reduce the
current to zero again (point e); (5) reverse the current once more until point b is
reached again. The lack of retraceability shown in Fig. 37-12 is called hysteresis.
Note that at points c and e the iron core is magnetized, even though there is no cur-
rent in the toroid windings; this is the familiar phenomenon of permanent magnetism.
The magnetization curve for paramagnetism (Fig. 37-6) is explained in terms of the
mutually opposing tendencies of alignment with the external field and of randomiza-
tion because of the temperature motions. In ferromagnetism, however, we have
assumed that adjacent atomic dipoles are locked in rigid parallelism. Why, then,
does the magnetic moment of the specimen not reach its saturation value for very
low-even zero-values of B 0? The modern interpretation is to assume the existence
within the specimen of domains, that is, of local regions within which there is essen-
tially perfect alignment. The domains themselves, however, as Fig. 37-13 suggests,
are not parallel at moderately low values of Bo.

Fig. 37-13 The separate mag-


netic domains in an unmagnet-
ized polycrystalline ferromag-
netic sample are oriented to pro-
duce little external effect. Each
domain, however, is made up
ot completely aligned atomic
dipoles, as suggested by the
arrows. The heavy boundaries
define the crystals that make up
the solid; the light boundaries
define the domains within the
crystals.
930 MAGNETIC PROPERTIES OF MATTER Chap. 37

Fig. 37-14 Domain patterns for a


single crystal of iron containing
3.8% silicon. The white lines
show the boundaries between the
domains. These boundaries are
regular rather than irregular, as in
Fig. 37-13, because the specimen
is a single crystal. In Fig. 37-13
the specimen is made up of many
crystallites or grains. (Courtesy
H. J. Williams, Bell Telephone
Laboratories.)

Figure 37-14 shows some domain photographs, taken by sprinkling a colloidal


suspension of finely powdered iron oxide on a properly etched single crystal of iron.
The domain boundaries, which are thin regions in which the alignment of the ele-
mentary dipoles changes from a certain direction in one domain to an entirely di:ff erent
direction in the other, are the sites of intense but highly localized and nonuniform
magnetic fields. The suspended colloidal particles are attracted to these regions.
Although the atomic dipoles in the individual domains are completely aligned, the
specimen as a whole may have a very small resultant magnetic moment. This is
the state of affairs in an unmagnetized iron nail.
As we magnetize a piece of iron by placing it in an external magnetic field, two
effects take place. One is a growth in size of the domains that are favorably oriented
at the expense of those that are not, as in Fig. 37-15. Second, the direction of orien-
tation of the dipoles within a domain may swing around as a unit, becoming closer to
the field direction. Hysteresis comes about because the domain boundaries do not
move completely back to their original positions when the external field Bo is removed.
Two other types of magnetism, closely related to ferromagnetism, are antiferro-
magnetism and ferrimagnetism (note spelling). In antiferromagnetic substances, of

t---1
0.01 mm
(a) (b) (c)

Fig. 37-15 (a) A boundary between two domains, with the magnetization in each domain
as shown by the white arrows. (b) If an external magnetic field pointing from left to right
is imposed on the specimen, the upper domain will grow at the expense of the lower. The
domain boundary will move down as the elementary dipoles reverse themselves. (c) The
process continues. The boundary has moved across a region in which there is a crystal
imperfection. (Courtesy H. J. Williams, Bell Telephone Laboratories.)
Sec. 37-6 NUCLEAR MAGNETISM 931
·••·• ·•·•·•· ···•· ······
••i

•••••

(a)
..•• ,. :: · : 11 •11·····

Fig. 37-16 Showing how elementary mag-


netic dipoles are oriented by the interatomic (b)
.... : 'J .,.
exchange coupling in (a) ferromagnetism, i ·· ·· ······ ••••••

(b) antiferromagnetism, and (c) ferrimag- ······ ······


,,. .
·······
·····

netism.
······ ······ "''

(c) . ······
······
.................. ······ ······
................. .... ······
1i1ii:!i!il
iii!!!
····· ······ ·····
•••••• ••••• •••••
............
.... ······
' ······ ..., .... ······
······ ·•···
•·•·•·•·• ·•·•·•·•·•·•••·•
.·•·.·.·..·.... ·..······· ······
····· ·•·······•·····•·•·•··· ···········

which Mn0 2 is an example, the exchange coupling to which we referred on page 834
serves to lock adjacent ions into rigid antiparallelism (see Fig. 37-16b). Such ma-
terials exhibit very little gross external magnetism. However, if they are heated
above a certain temperature, called the Neel temperature, the exchange coupling
ceases to act and the material becomes paramagnetic. In ferrimagnetic substances,
of which iron ferrite is an example, two different kinds of magnetic ions are present.
In iron ferrite the two ions are Fe++ and Fe+++. The exchange coupling locks the
ions into a pattern like that of Fig. 37-16c, in which the external effects are inter-
mediate between ferromagnetism and antiferromagnetism. Here, too, the exchange
coupling disappears if the material is heated above a certain characteristic tempera-
ture.

37-6 Nuclear Magnetism


Many nuclei have magnetic dipoles, and the possibility arises that a specimen of
matter may exhibit gross external magnetic effects associated with its nuclei. How-
ever, nuclear magnetic moments are several orders of magnitude smaller than those
associated with the electronic motions in an atom or ion. The magnetic moment of
an electron associated with its spin, for example, exceeds that of the proton (the
nucleus of hydrogen) by a factor of 660.
Gross external effects for nuclear magnetism are smaller than the corresponding
(electronic) paramagnetic effects by the square of ratios of this order of magnitude,
because (a), all else being equal, the external magnetism is reduced by such a ratio,
but (b) the very fact that the magnetic dipole moment of the nucleus is smaller
means that (see Example 3) the thermal vibrations are proportionally (to a good
approximation) more effective in reducing the degree of alignment of the elementary
dipoles in an external magnetic field; thus all else is not equal and the ratio enters
twice.
Techniques such as the Rowland ring (Fig. 37-10) are far too insensitive to detect
nuclear magnetism. We describe here a nuclear resonance technique by means of
which nuclear magnetism can readily reveal itself. This method is also vastly useful
for studying paramagnetism, ferromagnetism, antiferromagnetism, and ferrimagne-
tism, in all of which cases the magnetic effects are associated not with the nuclei but
with the atomic electrons. The nuclear-resonance technique was developed in 1946
by E. M. Purcell and his co-workers at Harvard. Simultaneously and independently,
F. Bloch and his co-workers at Stanford discovered a very similar method. For these
achievements the two physicists received a Nobel prize.
932 MAGKETIC PROPERTfES OF MATTER Chap. 37

I I

Fig. 37-17 (a) A spinning proton


precessing in an external magnetic
field and (b) a spinning top precess-
ing in an external gravitational
field. Lp and L 1 are the two angular
B momentum vectors.

(a) (b)

We focus our attention on the problem of measuring the magnitude µ of the mag-
netic moment of the proton. In principle, this can be done by placing a specimen
containing protons in an external field of magnetic induction B and by measuring
the energy ( = 2µB) required to turn the protons end for end. A rigorously correct
description of the procedures cannot be given without using quantum physics. The
description given, although based entirely on classical physics, nevertheless leads to
the correct conclusions.
Figure 37-17a shows a spinning proton with its axis making an angle() with a uni-
form external magnetic field B. Figure 37-17b shows a spinning top with its axis
making an angle() with a uniform external gravitational field g. In each case there is
a torque that tends to align the axis of the spinning object with the field. For the
proton (Eq. 33-11) it is given by
Tp = µB sine. (37-17a)
For the top it is given by Tt = mgr sin(), (37-17b)
where r locates the center of mass of the top and m is its mass.
In Section 13-2, we saw that the spinning top processes about a vertical axis with
an angular frequency given by
mgr
Wt=-,
Lt
in which Lt is the spin angular momentum of the top.
The proton, which has a quantized spin angular momentum Lp, will also precess
about the direction of the (magnetic) field because of the action of the (magnetic)
torque. The student should derive the expression for the frequency of precession,
being guided by the derivation of Section 13-2, but using the magnetic torque
(Eq. 37-17a) instead of the gravitational torque (Eq. 37-17b). The relation is
µB
w = --· (37-18b)
P Lp

... Example 5. What is the precession frequency of a proton in a magnetic field of


0.5 weber/meter 2 ?
The quantitiesµ and LP in Eq. 37-18b are 1.4 X 10-26 amp-m 2 and 0.53 X 10- 34
joule-sec. This equation then yields

Vp = ~E = _!!_1!__ = (1.4 X 10- 26 amp-m 2)(0.5? weber/meter 2) = 2 _1 X 107 c s.


21r 21rLP (21r)(0.53 X 10- 34 Joule-sec) p
This frequency ( = 21 me/sec) is in the radio-frequency range.
Sec. 37-6 NUCLEAR MAGNETISM 933

I\
I ta
I \
Fig. 37-18 In the nuclear magnetic resonance method a small oscil- I \
lating magnetic field Bose is placed at right angles to a steady field B. I \
I B \
I \

It is possible to change the energy of any system in periodic motion if we allow an


external influence to act on it at the same frequency as that of its motion. This is the
familiar resonar>ce condition. As an "external influence" for the precessing proton,
we use a small alternating magnetic field Bose arranged to be at right angles to the
steady field B. This oscillating field combines vectorially with the steady field so
that the resultant field rocks back and forth between the limits shown by the dashed
lines in Fig. 37-18. Typical values for B and for the amplitude of Bose are 5000
gauss and 1 gauss, respectively, so that the rocking angle a in the figure is quite
small. If the angular frequency w0 of the oscillating field is chosen equal to the
angular precession frequency wp of the proton, it turns out that the precessing proton
can absorb energy. An increase in energy means an increase in () in Fig. 37-17 a.
The resonance condition
µB
WO= - - (37-19)
Lp

can be used to measureµ. We place the spinning proton in a known field B, apply a
"perturbing field" at right angles to it, and vary the angular frequency w0 of this
perturbing field until resonance occurs. It is possible to tell when Eq. 37-19 is
satisfied because, at resonance, many spinning protons will tend to turn end for end
in the field, absorbing energy which can be detected by appropriate electronic tech-
niques.
Figure 37-19 is a schematic diagram of an experimental arrangement. The protons,
present as hydrogen nuclei in a small vial V of water, are immersed in a strong steady
magnetic field caused by the electromagnet whose pole faces N and S are shown. A
rapidly alternating current in the small coil C provides the (horizontal) weak, per-
turbing magnetic field Bose· This current is provided by a radio-frequency oscillator

Resonance
Oscillator detector

Frequency Oscilloscope
control knob

Fig. 37-19 An arrangement to observe nuclear resonance. The oscillating field ls


horizontal within the coil.
934 MAGNETIC PROPF,HTIES OF MATTER Chap. 37

Fig. 37-20 An oscilloscope photograph of


a proton resonance peak showing energy
absorbed from the oscillator versus oscil-
lator frequency. (From Bloembergen et
al., Phys. Rev., 73, 679.)

whose angular frequency w0 can be varied; an electronic "resonance detector," also


connected to the oscillator, serves to indicate when energy is being drained from the
oscillator and used to "flip the protons." In principle, the oscillator angular fre-
quency w 0 is varied until the resonance detector shows that Eq. 37-19 is satisfied
(see Fig. 37-20). The magnetic moment µ can then be determined by measuring
B and wo. Surprisingly enough, magnetic moments can be measured in this and simi-
lar ways to a much greater accuracy than we can measure µ for a bar magnet. For
the proton we have
µP = 1.41044 X 10- 26 amp-m 2 •

37-7 Three Magnetic Vectors


In Chapter 30 we saw that if a dielectric is placed in an electric field polarization
charges will appear on its surface. These surface charges, which find their origin in
the elementary electric dipoles (permanent or induced) that make up the dielectric,
set up a field of their own that modifies the original field. For the simple case dis-
cussed in Chapter 30-a dielectric slab in a parallel-plate capacitor-this complica-
tion can readily be handled in terms of the electric field strength vector E and some
knowledge of the electric properties of the slab material, such as the dielectric con-
stant. For more complex problems we asserted that it was useful to introduce two
other (subsidiary) electric vectors, the electric polarization P and the electric displace-
ment D. Table 30-2 shows some of the characteristics of these three vectors.
In magnetism we find a similar situation. If magnetic materials are placed in a
field of induction, the elementary magnetic dipoles (permanent or induced) will act
to set up a field of induction of their own that will modify the original field. For the
simple case discussed in this chapter-a Rowland ring with a ferromagnetic core-
this complication can readily be handled in terms of the magnetic induction vector B
and some knowledge of the magnetic properties of the ring material, such as that
provided by Fig. 37-9. For more complex problems we find it useful to introduce two
other (subsidiary) magnetic vectors, the magnetization M and the magnetic .field
strength H. We do so largely so that the student who takes a second course in elec-
tromagnetism will have some familiarity with them.
Consider a Rowland ring carrying a current i 0 in its windings and designed so
that its core, assumed to be iron, can be removed. The magnetic induction B,
measured by the methods of Section 37-5, will be much greater when the core is in
place than when it is not, assuming that the current in the windings remains un-
changed.
Sec. 37-7 THREE MAGNETIC VECTORS 935
Physically we can understand the large value of B in the iron core in terms of the
alignment of the elementary dipoles in the iron. A hypothetical slice out of the iron
core, as in Fig. 37-21b, has a magnetic moment dµ equal to the vector sum of all of
the elementary dipoles contained in it. We define our first subsidiary vector, the
magnetization M, as the magnetic moment per unit volume of the core material.
For the slice of Fig. 37-21b we have
dµ = M(A dl),
where (A dl) is the volume of the slice, A being the cross section of the core.
When we discussed Ampere's law in Chapter 34, we assumed that no magnetic
materials were present. If we apply this law, namely

f B·dl = µoi, (37-20)

to the circular path of integration shown in Fig. 37-21a, we obtain


(B)(21rro) = µo(Noio), (37-21)
in which r 0 is the mean radius of the core, N O is the number of turns, and io is the
current in each turn. We see at once that Ampere's law, in the form of Eq. 37-20,

Fig. 37-21 (a) A Rowland ring with an


iron core. (b) A slice of the core, showing
its magnetic moment dµ caused by the
(a)
alignment of the elementary magnetic di-
poles in the iron.

\
\

dl

1-+-~~-~ I
I dµ,
(b)
936 MAGNETIC PROPERTIES OF MATTER Chap. 37
is not valid when magnetic materials are present. Equation 37-21 predicts that, since
the right side is the same whether or not the core is in place, the induction B should
also be the same, a prediction not in accord with experiment.
We can increase B in the absence of the iron core to the value that it has ,vhen the
core is in place if we increase the current in the windings by an amount i 111 .0 • The
magnetization of the iron core is thus equivalent in its ejfect on B to such a hypothetical
current increase. We choose to modify Ampere's law by arbitrarily inserting a
magnetizing current term i 111 on the right, obtaining

f B·dl = µo(i + i111). (37-22)

If we give i111 a suitable value when the iron core is in place, it is clear that Ampere's
law, in this new form, can remain valid. It remains to relate this (largely hypo-
thetical) magnetizing current to something more physical, the magnetization M. *
Applying Eq. 37-22 to the iron ring of Fig. 37-2la yields

(B)(21rro) = µo(Noio) + µo(Noi111,o). (37-23)

We can relate i111,o to the magnetization M if we recall (Eq. 33-10) that the magnetic
moment of a magnetic dipole in the form of a current loop is given by

µ = NiA,

where N is the number of turns in the loop, i is the loop current, and A is the loop
area. Let us use this equation to find what increase i111.o in current in the windings
around the slice of Fig. 37-2lb would produce a magnetic moment equivalent to that
actually produced by the alignment of elementary dipoles in the slice. We have

M(A dl) = ( No 2 !~J (i111.o)(A),

the quantity in the first parentheses on the right being the number of turns associ-
ated with the slice of thickness dl. This reduces to

(37-24)

Substituting this into Eq. 37-23 yields

(B)(21rro) = µo(Noio) + µo(M)(21rro). (37-25)

We now choose to generalize from the special case of the Rowland ring by writing
Eq. 37-25 as

or

* It is possible to give reality to the magnetizing current by viewing it as a real current


that flows around the magnet at its surface, being the resultant macroscopic effect of all
the microscopic current loops that constitute the atomic electron orbits. This Amperian
current viewpoint however does not take the magnetization due to electron spin readily
into account. Since we do not attempt to measure magnetizing currents experimentally,
other than through their (postulated) magnetic effects, we prefer to view the magnetizing
current as a convenient formalism.
Sec. 37-7 THREE MAGNETIC VECTORS 937
The quantity (B - µ 0M) / µ 0 occurs so often in magnetic situations that we give it a
special name, the magnetic .field strength H, or

H = B ·- µoM
µo

which we write as B = µoH + µoM. (37-26)

Ampere's law can now be written in the simple form

fH·dI = i, (37-27)

which holds in the presence of magnetic materials and in whi1;n i is the true current only,
that is, it does not include the magnetizing current. This reminds us that the
electric displacement vector D permitted us to write Gauss's law for the case in which
dielectric materials are present, in a form involving free charges only, that is, not
polarization charges; see Table 30-2.
We state without proof (see Problems 12 and 13) that at a boundary between two
media (1) the component of H tangential to the surface has the same value on each
side of the surface * and (2) the component of B perpendicular to the surface has the
same value on each side of the surface. These boundary conditions are of great value
in solving complex problems.
To find Hin our Rowland ring, let us apply Ampere's law in the generalized form
of Eq. 37-27. We have
(H)(21rro) = N oio,

where io is the (true) current in the windings. This gives

H = 2
No).
( 1rro io =
.
mo, (37-28)

in which n is the number of turns per unit length. Since we have not introduced
any information describing the core into Eq. 37-27, the value of H computed from
Eq. 37-28 is independent of the core material.
B can be measured experimentally by the method of Section 37-5 and M can then
be calculated from Eq. 37-26. The student should note in passing (see Eq. 37-15)
that the abscissa Bo in Fig. 37-9 is proportional to H ( = µ 0H), the ordinate being
proportional to B. Curves such as this and that of Fig. 37-12 are called B-H curves.
Let us assume that we have made measurements of H, B, and M for a wide variety
of magnetic materials, using either the technique described or an equivalent one.
For paramagnetic and diamagnetic materials we would find, as an experimental result,
that B is directly proportional to H, or

(37-29)
in which Km, the permeability of the magnetic medium, is a constant for a given tem-
perature and density of the material. Eliminating B between Eqs. 37-29 and 37-26
allows us to write
M = (Km - l)H, (37-30)

which is another expression of the linear or proportional character of paramagnetic


and diamagnetic materials.

* Assuming; that there are no true current~ at the surface, as there are in the Rowland
ring of Fig. 37-21a, for example.
938 MAGNETIC PROPERTIES OF MATTER Chap. 37

(c) ~r: r~
B .uoH

M =O

(a) (b)

Fig. 37-22 (a) The lines of H and (b) the lines of B for a permanent magnet.
Note that the lines of H change direction at the boundary. The closed dashed
curves are paths of integration around which Ampere's law may be applied. The
relation B = µoH + µoM is shown to be satisfied for (c) a particular outside point p
and (d) a particular inside point q.

For a vacuum, in which there are no magnetic dipoles present to be aligned, the
magnetization M must be zero. Putting M = 0 in Eq. 37-26 leads to

B = µoH (a vacuum). (37-31)

Comparison with Eq. 37-29 shows that a vacuum must be described by Km = 1.


Equation 37-30 verifies that the magnetization vanishes if we put Km equal to unity.
For paramagnetic materials Km is slightly greater than unity. For diamagnetic
materials it is slightly less than unity; Eq. 37-30 shows that this requires M and H
to be oppositely directed, a fact discussed at length in Section 37-4.
In ferromagnetic materials the relationship between B and H is far from linear,
as Figs. 37-9 and 37-12 show. Experimentally, Km proves to be a function not only
of the value of H but also, because of hysteresis, of the magnetic and thermal history
of the specimen.*
An interesting special case of ferromagnetism is the permanent magnet, for which
H, M, and B all have nonvanishing values inside the magnet even though there is no
true current. Figure 37-22 shows typical lines of B and H associated with such a
magnet. The lines of B may be drawn as continuous loops, the boundary condition
(2), mentioned above, being satisfied where the lines enter and leave the magnet.
Equation 37-22 shows that the vector B is associated with the total current, both
true and magnetizing. In Fig. 37-22a f B·dl around any loop such as that shown
1:)y the dashed curve is not zero and must be associated with a hypothetical magne-
tizing current iM imagined to circulate around the magnet at its surface; actual or
true currents (i) do not exist in this problem. Figure 37-22b shows that H reverses
direction at the boundary. Since H (see Eq. 37-27) is associated with true currents
only, we must have
lines.
f H·dl = 0 around any loop such as that shown by the dashed
* In the dielectric case there are waxy materials, called f erroelectrics, for which the rela-
tionship between D and E is nonlinear, which exhibit hysteresis, and from which quasi-
permanent electric dipoles (electrets) can be constructed. However, most commonly useful
dielectric materials are linear whereas most commonly useful magnetic materials are non-
linear.
THREE MAGNETIC VECTORS 939

The reversal of H at the boundary makes this possible. Note that M and H point
in opposite directions within the magnet. Table 37-1 summarizes the properties of
the three vectors B, H, and M.

Table 37-1
THREE MAGNETIC VECTORS

Name Symbol Associated with Boundary Condition

Magnetic induction B All currents Normal component


continuous

Magnetic field strength H True currents only Tangential component


continuous t
----
Magnetization M Magnetization Vanishes in a vacuum
(magnetic dipole moment currents only
per unit volume)

Defining equations for B F = qv xB Eq. 33-3a


or = ii xB Eq. 33-6a

General relation among the three vectors B = µoH + µoM Eq. 37-26

Ampere's law when magnetic materials are


present
f H·dl = i Eq. 37-27

(i = true current only)

Empirical relations for certain magnetic B = KmµoH Eq. 37-29


materials * M = (Km - l)H Eq. 37-30

* For paramagnetic and diamagnetic materials only, if Km is to be independent of H.


t Assuming no true currents exist at the boundary .

.. Example 6. In the Rowland ring the (true) current io in the windings is 2.0 amp
and the number of turns per unit length (n) in the toroid is 10 turns/cm. B, measured
by the technique of Section 37-5, is 1.0 weber/meter 2 • Calculate (a) H, (b) M, and
(c) the magnetizing current i 111 ,0 both when the core is in place and when it is removed.
(d) For these particular operating conditions, what is Km?
(a) His independent of the core material and may be found from Eq. 37-28:

H = ni
= (10 turns/cm)(2.0 amp)

= 2.0 X 10 3 amp/meter.
940 MAGNETIC PROPERTIES OF MATTER Chap. 37
(b) M is zero when the core is removed. With the core in place, we may solve
Eq. 37-26 for M, obtaining for the magnitude of M,

M = B - µoH
µo
(1.0 weber/meter 2) - (41r X 10-7 weber/amp-m)(2.0 X 10 3 amp/meter)
(41r X 10-7 weber/amp-m)
= 7.9 X 10 5 amp/meter.
(c) The effective magnetizing current follows from Eq. 37-24:

7.9 X 10 5 amp/meter
1.0 X 103 turns/meter
= 790 amp.

An additional current of this amount in the windings would produce the same value
of B in the absence of a core as that obtained, by alignment of the elementary di-
poles, with the core in place.
(d) The permeability can be found from Eq. 37-29, or

B
Km= µoH
1.0 weber/meter 2
(41r X 10-7 weber/amp-m)(2.0 X 103 amp/meter)
= 397.
We emphasize that this value of Km holds only for the special conditions of this
experiment. ...ii

QUESTIONS

1. Two iron bars are identical in Rppearance. One is a magnet and one is not. How can
you tell them apart? You are not permitted to suspend either bar as a compass needle or
to use any other apparatus.
2. How could you reverse the magnetism of a compass needle?
3. Two iron bars always attract, no matter the combination in which their ends are
brought near each other. Can you conclude that one of the bars must be unmagnetized?
4. If we sprinkle iron filings on a particular bar magnet, they cling both to the ends and
to the middle. Sketch roughly the lines of B, both outside and inside the magnet.
5. The earth is a huge magnetic dipole. (a) Is the magnetic pole in the Northern
Hemisphere a north or a south magnetic pole? (b) ln the Northern Hemisphere do the
magnetic lines of force associated with the earth's magnetic field point toward the earth's
surface or away from it?
6. Cosmic rays are charged particles that strike our atmosphere from some external
source. We find that more low-energy cosmic rays reach the earth at the north and south
magnetic poles than at the (magnetic) equator. Why is this so?
7. How might the magnetic dipole moment of the earth be measured?
8. Give three reasons for believing that the flux <f>B of the earth's magnetic field is
greater through the boundaries of Alaska than through those of Texas.
Chap. 37 PROBLEMS 941
9. The neutron, which has no charge, has a magnetic dipole moment. Is this possible
on the basis of classical electromagnetism, or does this evidence alone indicate that classical
electromagnetism has broken down?
10. Is the magnetirntion at saturation for a paramagnetic substance very much different
from that for a saturated ferromagnetic substance of about the same size?
11. Explain why a magnet attracts an unmagnetized iron object such as a nail.
12. Does any net force or torque act on (a) an unmagnetized iron bar or (b) a permanent
bar magnet when placed in a uniform magnetic field?
13. A nail is placed at rest on a smooth table top near a strong magnet. It is released
and attracted to the magnet. What is the source of the kinetic energy it has just before
it strikes the magnet?
14. The magnetization induced in a given diamagnetic sphere by a given external
magnetic field does not vary with temperature, in sharp contrast to the situation in para-
magnetism. Is this understandable in terms of the description that we have given of the
origin of diamagnetism?
15. Compare the magnetirntion curves for a paramagnetic substance (Fig. 37-6) and
for a ferromagnetic substance (Fig. 37-9). What would a similar curve for a diamagnetic
substance look like? Do you think that it would show saturation effects in strong applied
fields (say 10 weber /meter2)?
16. Distinguish between the precession frequency and the cyclotron frequency of a
proton in a magnetic field.
17. In our discussion of nuclear magnetism we said that energy absorption occurs
because the dipoles are turned end for end. However, a given dipole might initially be
lined up either with the field or against it. In the first case there would be an absorption
of energy, but in the second case there would be a release of energy, each amount being 2µB.
Why do we observe a net absorption? These two events would seem to cancel.
18 . .Discuss similarities and differences in Tables 30-2 and 37-1.

PROBLEMS

__-Y.'The earth has a magnetic dipole moment of 6.4 X 1021 amp-m 2 • (a) What current
would have to be set up in a single turn of wire going around the earth at its magnetic
equator if we wished to set up such a dipole? (b) Could such an arrangement be used to
cancel out the earth's magnetism at points in space well above the earth's surface? (r)
On the earth's surface?
,:l: · · Assume that the electron is a small sphere of radius R, its charge and mass being
spread uniformly throughout its volume. Such an electron has a "spin" angular mo-
mentum L of 0.53 X 10-34 joule-sec and a magnetic moment µ of 9.3 X 10-24 amp-m 2 •
Show that e/m = 2µ/L. Is this prediction in agreement with experiment? (Hint: The
spherical electron must be divided into infinitesimal current loops and an expression for
the magnetic moment found by integration. This model of the electron is too mechan-
istic to be in the spirit of quantum physics.)
3. Calculate (a) the electric field strength and (b) the magnetic induction at a point
1.0 A (one angstrom unit) away from a proton, measured along its axis of spin. The mag-
netic moment of the proton is 1.4 X 10-25 amp-m2 .
4. A Rowland ring is formed of ferromagnetic material. It is circular in cross section,
with an inner radius of 5.0 cm and an outer radius of 6.0 cm and is wound with 400 turns
of wire. (a) What current must be set up in the windings to attain Bo = 2 X 10-4 weber/
meter2 in Fig. 37-9? (b) A secondary coil wound around the toroid has 50 turns and has
a resistance of 8.0 ohms. If, for this value of Bo, we have BM = 800B 0 , how much charge
moves through the secondary coil when the current in the toroid windings is turned on?
5. The dipole moment associated with an atom of iron in an iron bar is 1.8 X 10-23
amp-m 2 . Assume that all the atoms in the bar, which is 5 cm long and has a cross-sec-
942 MAGNETIC PROPERTIES OF MA'l'TER Chap. 37
tional area of 1 cm 2 , have their dipole moments aligned. (a) What is the dipole moment
of the bar? (b) What torque must be exerted to hold this magnet at right angles to an
external field of 15,000 gauss? The density of iron is 7.9 gm/cm 3 •
6. Can you give an explanation of diamagnetism based tin Faraday's law of induction?
In Figs. 37-7a and b, for example, what inductive effects can be expected as the magnetic
field is built up from zero to the final value B?
7. Prove that ~w «wo in Eq. 37-13.
8. Show that, classically, a spinning positive charge will have a spin magnetic moment
that points in the same direction as its spin angular momentum.
-- 9: Assume that the nuclei (protons) in 1 gm of water could all be aligned. What mag-
netic induction B would be produced 5.0 cm from the sample, along its alignment axis?
10. It is possible to measure e/m for the electron by measuring (a) the cyclotron fre-
quency 11e of electrons in a given magnetic field and (b) the precession frequency 11p of
protons in the same field. Show that the relation is

e lie µ.,
m 11pLa

Since µ. 8 and L. for the proton are accurately known, this experiment gives us our most
precise value of e/m today.
11. Dipole-dipole interactiori. The exchange coupling mentioned in Section 37-5 as
being responsible for ferromagnetism is not the mutual magnetic interaction energy be-
tween two elementary magnetic dipoles. To show this (a) compute B a distance a ( = 1.0
A) away from a dipole of moment µ. ( = 1.8 X 10-23 amp-m 2); (b) compute the energy
( = 2µ.B) required to turn a second similar dipole end for end in this field. What do you
conclude about the strength of this dipole-dipole interaction? Compare with the results
of Example 3. (Note: for the same distance, the field in the median plane of a dipole is
only half as large as on the axis; see Eq. 37-2.)
'"'12', .Boundary condition for B. Prove that at the boundary between two media the
normal component of B has the same value on each side of the surface. (Hint: Construct
a closed Gaussian surface shaped like a flat pillbox with one face in each medium and apply
Gauss's law for magnetism.)
13, Boundary condition for H. Prove that at the boundary between two media the
tangential component of H has the same value on each side of the surface, assuming that
there is no current at the surface. (Hint: Construct a closed rectangular loop, the two
opposite longer sides being parallel to the surface, with one side in each medium. Use
Ampere's law in the form that applies when magnetic materials are present.)
14. The magnetic energy density can be shown to be given in its most general form as

µ.B = !B·H.
Does this reduce to a familiar result for a vacuum?
Electromagnetic Oscillations
CHAPTER 38

38-1 LC Oscillations
The LC system of Fig. 38-1 resembles a mass-spring system (see Fig. 8-4)
in that, among other things, each system has a characteristic frequency of
oscillation. To see this, we assume that initially the capacitor C in Fig.
38-la carries a charge qm and the current i in the inductor is zero. At this
instant the energy stored in the capacitor is given by Eq. 30-25, or
1 qm2
UE = --· (38-1)
2 C
The energy stored in the inductor, given by
UB -_ 21£;2
•, (38-2)
is zero because the current is zero. The capacitor now starts to discharge
through the inductor, positive charge carriers moving counterclockwise, as
shown in Fig. 38-lb. This means that a current i, given by dq/dt and point-
ing down in the inductor, is established.
As q decreases, the energy stored in the electric field in the capacitor also
decreases. This energy is transferred to the magnetic field that appears
around the inductor because of the current i that is building up there. Thus
the electric field decreases, the magnetic field builds up, and energy is trans-
ferred from the former to the latter.
At a time corresponding to Fig. 38-lc, all the charge on the capacitor will
have disappeared. The electric field in the capacitor will be zero, the energy
stored there having been transferred entirely to the magnetic field of the
943
944 ELECTRO MAG NET IC OSCILLATIONS Chap. 38

(b) (d)

I
•• ••
UB
\

Orr,.)
_I a- I
.._______...... (f)

I_
UB UE
Fig. 38-1 Showing eight stages in a cycle of oscillation of an LC circuit. The bar graphs
below each figure show the stored magnetic and electric potential energy. The vertical
arrows on the induewr axis show the current. The student should compare this figure in
detail with Fig. 8-4, to which it exactly corresponds.

inductor. According to Eq. 38-2, there must then be a current-and indeed


one of maximum value-in the inductor. Note that even though q equals
zero the current (which is dq/dt) is not zero at this time.
The large current in the inductor in Fig. 38-lc continues to transport
positive charge from the top plate of the capacitor to the bottom plate, as
shown in Fig. 38-ld; energy now flows from the inductor back to the capaci-
tor as the electric field builds up again. Eventually, the energy will have
been transferred completely back to the capacitor, as in Fig. 38-le. The
situation of Fig. 38-le is like the initial situation, except that the capacitor
is charged in the opposite direction.
Sec. 38-1 LC OSCILLATIONS 945
The capacitor will start to discharge again, the current now being clock-
wise, as in Fig. 38-lf. Reasoning as before, we see that the circuit eventually
returns to its initial situation and that the process continues at a definite
frequency v (measured, say, in cycles/sec) to which corresponds a definite
angular frequency w ( = 27rv and measured, say, in radians/sec). Once
started, such LC oscillations (in the ideal case described, in which the cir-
cuit conb;ris no resistance) continue indefinitely, energy being shuttled
back and forth between the electric field in the capacitor and the magnetic
field in the inductor. Any configuration in Fig. 38-1 can be set up as an
initial condition. The oscillations will then continue from that point, pro-
ceeding clockwise around the figure. The student should compare these
oscillations carefully with those of the mass-spring system described in
Fig. 8-4.
To measure the charge q as a function of time, we can measure the variable
potential difference V c(t) that exists across capacitor C. The relation

shows that V c is proportional to q. To measure the current we can insert a


small resistance R in the circuit and measure the potential difference across
it. This is proportional to i through the relation
vR = (R)i.
We assume here that Risso small that its effect on the behavior of the cir-
cuit is negligible. Both q and i, or more correctly V c and V R, which are pro-
portional to them, can be displayed on a cathode-ray oscilloscope. This
instrument can plot automatically on its screen graphs proportional to q(t)
and i(t), as in Fig. 38-2 .

-
.-.
Co)
<::,,
(a)
II Time
'""
~

I
ace g ace gab
~I I I I I I I I I
(b) -~-~ ~ )

~ ~~Time

Fig. 38-2 A drawing of an oscilloscope screen showing potential differences proportional


to (a) the charge and (b) the current, in the circuit of Fig. 38-1, as a function of time. The
letters indicate corresponding phases of oscillation in that figure. Note that because
i = dq/dt the lower curve is proportional to the derivative of the upper.
946 ELECTROMAGNETIC OSCILLATIONS Chap. 38
~ Example I. A l.0-µf capacitor is charged to 50 volts. The charging battery is
then disconnected and a 10-mh coil is connected across the capacitor, so that LC
oscillations occur. What is the maximum current in the coil? Assume that the
circuit contains no resistance.
The maximum stored energy in the capacitor must equal the maximum stored
energy in the inductor, from the conservation-of-energy principle. This leads, from
Eqs. 38-1 and 38-2, to
!2qm2 - l.L.
c-2tm,
2

where im is the maximum current and qm is the maximum charge. Note that the
maximum current and the maximum charge do not occur at the same time but one-
fourth of a cycle apart; see Figs. 38-1 and 38-2. Solving for im and substituting
CVo for qm gives

. /c /1.0 x 10- 5 farad


tm = Vo '\JI, = (50 volts) '\J 10 X 10 _ 3 henry = 0.50 amp.

In an actual LC circuit the oscillations will not continue indefinitely be-


cause there is always some resistance present that will drain away energy
by Joule heating. The oscillations, once started, will die away, as in Fig.
38-3. This figure should be compared to Fig. 15-19, which shows the decay
of the mechanical oscillations of a mass-spring system caused by frictional
damping.
It is possible to have sustained electromagnetic oscillations if arrange-
ments are made to supply, automatically and periodically (once a cycle,
say), enough energy from an outside source to compensate for that lost to
Joule heat. We are reminded of a clock escapement, which is a device for
feeding energy from a spring or a falling weight into an oscillating pendulum,
thus compensating for frictional losses that would otherwise cause the oscil-
lations to die away. Oscillators whose frequency v may be varied between
certain limits are commercially available as packaged units over a wide range
of frequencies, extending from low audio-frequencies (lower than 10 cycles/
sec) to microwave frequencies (higher than 10 10 cycles/sec).

Fig. 38-3 A photograph of an oscil-


loscope trace showing how the oscil-
lations in an LRC circuit die away
because energy is lost to Joule heat
in the resistor. The figure is a plot
of the potential difference across the
resistor as a function of time.
Sec. 38-2 ANALOGY TO SIMPLE HARMONIC MOTION 947

38-2 Analogy to Simple Harmonic Motion


Figure 8-4 shows that in a mass-spring system performing simple har-
monic motion, as in an oscillating LC circuit, two kinds of energy occur.
One is potential energy of the compressed or extended spring; the other is
kinetic energy of the moving mass. These are given by the familiar formulas
in the first column of Table 38-1. The table suggests that a capacitor is in

Table 38-1

SOME ENERGY FORMULAS

Mechanical Electromagnetic

1 q2
sprmg capacitor UE = --
2C
mass inductor
v = dx/dt i = dq/dt

some formal way like a spring and an inductor is like a mass and that certain
electromagnetic quantities "correspond" to certain mechanical ones, namely,
q corresponds to x,
i corresponds to v,
C corresponds to 1/k,
L corresponds to m.
Comparison of Fig. 38-1, which shows the oscillations of the LC circuit,
with Fig. 8-4, which shows the oscillations in a mass-spring system, indicates
how close the correspondence is. Note how v and i correspond in the two
figures; also x and q. Note, too, how in each case the energy alternates
between two forms, magnetic and electric for the LC system, and kinetic
and potential for the mass-spring system.
In Section 15-3 we saw that the natural angular frequency of oscillation
of an undamped mass-spring system is

"' - 2,,-, - Jf.


The method of correspondences suggests that to find the natural frequency
for the LC circuit k should be replaced by 1/C and m by L, obtaining

/-1
w = 21rv = ~LC. (38-3)

This formula is indeed correct, as we show in the next section.


948 ELECTROMAGNETIC OSCILLATIONS Chap. 38
38-3 Electromagnetic Oscillations-Quantitative
We now derive an expression for the frequency of oscillation of an LC
circuit, our derivation being based on the conservation-of-energy principle.
The total energy U present at any instant in an oscillating LC circuit is
given by
1 2 1 q2
U = UB + UE = 2Li + --,
2C
which expresses the fact that at any arbitrary time the energy is stored
partly in the magnetic field in the inductor and partly in the electric field
in the capacitor. If we assume the circuit resistance to be zero, there is no
energy transfer to Joule heat and U re,nains constant with time, even
though i and q vary. In more formal language, dU /dt must be zero. This
leads to
(38-4)

Now, q and i are not independent variables, being related by


. dq
1,=-·
dt
di d2 q
Differentiating yields
dt dt 2
Substituting these two expressions into Eq. 38-4 leads to
d2 q 1
L-+-q = 0. (38-5)
dt 2 C
This is the differential equation that describes the oscillations of a (re-
sistanceless) LC circuit. To solve it, the student should note that Eq. 38-5
is mathematically of exactly the same form as Eq. 15-6,
d2 x
m -2 + kx = 0, (15-3)
dt
which is the differential equation for the mass-spring oscillations. Funda-
mentally, it is by comparing these two equations that the correspondences
on p. 947 arise.
The solution of Eq. 15-6 proved to be
x = A cos (wt + ct,), (15-5)
where A ( = Xm) is the amplitude of the motion and cf, is an arbitrary phase
angle. Since q corresponds to x, we can write the solution of Eq. 38-5 as
q = qm COS (wt + cf,), (38-6)
where w is the still unknown angular frequency of the electromagnetic
oscillations.
ELECTROMAGNETIC OSCILLATIONS-QUANTITATIVE 949

We can test whether Eq. 38-6 is indeed a solution of Eq. 38-5 by substi-
tuting it and its second derivative in that equation. To find the second
derivative, we write
dq .
~ = i = -wqm sin (wt </>) +
(38-7)
dt
dzq
and -w 2 qm cos (wt + </>).
dt 2
Substituting q and d 2 q/dt 2 into Eq. 38-5 yields
1
--Lw 2 qm COS (wt + </>) + - qm COS (wt + </>) = 0.
c
Canceling qm cos (wt + <I>) and rearranging leads to

Thus, if w is given the constant value 1/yLC, Eq. 38-6 is indeed a solution
of Eq. 38-5. This expression for w agrees with Eq. 38-3, which was arrived
at by the method of correspondences.
The phase angle cp in Eq. 38--6 is determined by the conditions that
prevail at t = 0. If the initial condition is as represented by Fig. 38-la,
then we put </> = 0 in order that Eq. 38-6 may predict q = qm at t = 0.
What initial condition in Fig. 38-1 is implied if we select </> = 90°?
..... Example 2. (a) In an oscillating LC circuit, what value of charge, expressed in
terms of the maximum charge, is present on the capacitor when the energy is shared
equally between the electric and the magnetic field? (b) How much time is required
for this condition to arise, assuming the capacitor to be fully charged initially?
Assume that L = 10 mh and C = 1.0 ,uf.
(a) The stored energy and the maximum stored energy in the capacitor are,
respectively,
and

Substituting UE = !UE,m yields


q2 1 qm2
or
2C 2 2C
(b) To find the time, we write, assuming ct, = 0 in Eq. 38-6,

q = qm COS wt = V1 2 qm,
1 7r 7r
which leads to wt = cos- 1 --= = - or t = -·
y2 4 4w
The angular frequency w is found from Eq. 38-3, or

w = ~2 = ~~10 X 10-3 henry;(l.O X 10-6 farad) = 1.0 X 104 radians/sec.


950 ELECTROMAGNETIC OSCILLATIONS Chap. 38
The time t is then
7r 11'
t = 4~ = (4)(1.0 X 10 4 radians/sec) = 7 ·9 X io-5 sec.
What is the frequency v in cycles/sec?
The stored electric energy in the LC circuit, using Eq. 38-6, is
1 q2 qm 2 2
UE = (38-8)
2 C = 2C cos (wt+ <f,),
and the magnetic energy, using Eq. 38-7, is
Us = !Li2 = !Lw 2qm 2 sin 2 (wt + <f,).
Substituting the expression for w (Eq. 38-3) into this last equation yields
2
Us qm . 2 (
= - sm wt+ <f,). (38-9)
2C

Fig. 38-4 The stored magnetic and electric energy in the circuit of Fig. 38-1. Note that
their sum is a constant.

Figure 38-4 shows plots of UE(t) and Us(t) for the case of q, = 0. Note
that (a) the maximum values of UE and Us are the same ( = qm 2 /2C); (b)
at any instant the sum of UE and Us is a constant ( = qm 2 /2C); and (c) when
UE has its maximum value, Us is zero and conversely. This analysis sup-
ports the qualitative analysis of Section 38-1. The student should compare
this discussion with that given in Section 15-4 for the energy transfers in
a mass-spring system .
.... Example 3. The LCR circuit. (a) Derive an expression for the quantity q(t)
for a single-loop circuit containing a resistance R as well as an inductance L and a
capacitance C. (b) After how long a time will the charge oscillations decay to half-
amplitude if L = 10 mh, C = 1.0 µf, and R = 0.1 ohm'?
(a) If U is the total stored field energy, we have, as before,

U is no longer constant but rather


[; = ! Li2 + ~ f.
dU = -i2R
dt ,
Sec. 38-3 ELECTROMAGNETIC OSCILLATIONS-QUANTITATIVE 951

the minus sign signifying that the stored energy U decreases with time, being con-
verted to Joule heat at the rate i 2R. Combining these two equations leads to

Li di+ Sf__ ~q_ = - i 2R.


dt c dt

Substituting dq/dt for i and d 2q/dt 2 for di/dt leads finally to


d 2q dq 1
L dt? + R dt + Cq = O,

which is the differential equation that describes the damped LC oscillations. If we


put R = 0, this equation reduces, as expected, to Eq. 38-5.
The student should compare this differential equation for damped LC oscillations
with Eq. 15-37, or
(15-28)

which describes damped mass-spring oscillations. Once again the equations are
mathematically identical, the resistance R corresponding to the mechanical damping
constant b.
The solution of the LCR circuit follows at once, by correspondence, from the
solution of Eq. 15-37. It is (see Eqs. 15-38 and 15-39) for R reasonably small, and
for an initial condition in which the capacitor has a maximum charge
q = qme-Rtf 2L cos w't, (38-10)

where (38-11)

Note that Eq. 38-10, which can be described as a cosine function with an expo-
nentially decreasing amplitude, is the equation of the decay curve of Fig. 38-3. Note,
too (Eq. 38-11), that the presence of resistance reduces the oscillation frequency.
These two equations reduce to familiar results as R ----, 0.
(b) The oscillation amplitude will have decreased to half when the amplitude
factor e-Rtf 2L in Eq. 38-10 has the value one-half, or

which leads readily to

t = ~!:_ ln 2 = (2)(10 X 10-3 henry)(0.69) = 0.14 sec.


R 0.10 ohm

The angular frequency, from Eq. 38-11, is

, / 1 ( 0.10 ohm )2
w = '\J (10 X 10- henry)(l.O X 10- farad) - 2 X 10 X 10- henry
3 5 3

= V 108 radians/ sec 2 - 25 radians/ sec 2 = 1.0 X 10 4 radians/ sec.

Note that the second term is rather small, so that in this case, as in many practical
cases, the resistance has a negligible effect on the frequency. The student should
show that 0.14 sec, the time at which the oscillations decrease to half-amplitude,
corresponds to about 220 cycles of oscillation. The damping is much less severe
than that illustrated in Fig. 38-3. ..._
952 ELECTROMAGNETIC OSCILLATIONS Chap. 38
38-4 Forced Oscillations and Resonance
Figure 38-5 shows an LCR circuit containing a sinusoidally varying emf S(t) given
by
8 = Sm cos w11t,
in which w" can be varied at will. The emf S(t) might be provided by a variable-
frequency oscillator. What amplitudes of electromagnetic oscillations are set up in
this circuit for various angular frequencies w" of the "driving force"?
The problem corresponds to that of forced oscillations in the damped mass-spring
system of Section 15-10. The differential equation describing that motion is

d 2x
m dt2 + b dx k
dt + X =
F
m
II
cos W t, (15-31)

where w" is the angular frequency of the external periodic driving force applied to the
system and Fm is its amplitude.
For the circuit of Fig. 38-5, the differential equation that follows from the cor-
respondences of p. 947 and the additional reasonable correspondence of 8 to F, is

L d2q
dt2 + R dqdt + (;1 q = C'
um COS W
II
t. (38-12)

This equation can also be derived by applying the energy conservation principle to
the circuit of Fig. 38-5.

Fig. 38-5 An LCR circuit containing a sinusoidal emf


of angular frequency w".

f,(t)

We can use the correspondences of p. 947 to write down the solution to Eq. 38-12.
Starting from Eq.15-41 and making the appropriate substitutions, we obtain

q= 8; sin (w 11
t - </>), (38-13)

where

and </>, the phase angle between the "driving force" and the "response," is given by
Rw 11
"*'=cos
"" -1
0 .
As often as not, we are interested in the current i(t) in the circuit, rather than the
charge; the current corresponds to the velocity v(t) of the moving mass in Section
15-10. We can find i(t) by differentiating Eq. 38-13, or

i
. = -dq = w"Sm
-- ( ,, t - 'I'
COS W
"") = 1-m
. COS W
( ,, t - 'I'
"") .
dt G
Sec. 38-4 FORCED OSCILLATIONS AND RESONANCE 953
The amplitude im of the current oscillations is given, from these equations, by

(38-14)

Inspection of Eq. 38-14 shows that the current (not the charge; see Question 5) will
have its maximum amplitude when
w"L = _!__, (38-15a)
w"C

which can be written as w" = ~2~· (38-15b)

Comparison with Eq. 38-3 shows that the maximum amplitude of the current oscilla-
tions occurs when the frequency w" of the applied emf is exactly equal to the natural
(undamped) frequency w of the system.
At resonance (w" = w) the amplitude of the current oscillations is determined
entirely by the resistance; this follows by combining Eqs. 38-14 and 38-15a, or
. Sm
im = R (at resonance).

Figure 38-6 shows im as a function of w" for an oscillating LCR circuit containing
three different values of resistance. Note that the smaller the resistance, the sharper
the resonance curve. The sharpness of such a curve is measured by its half-width,
which is the difference between two frequencies, each of which corresponds to a cur-
rent amplitude of one-half the maximum current amplitude. The half-width of the
curve for R = 10 ohms in Fig. 38-6 is shown in that figure by the arrow marked Liw.

1.0
L = lOOµh
0.9 C = lOOµµf
E = lOmv
0.8

0.7 R = IO ohms

~ 0.6
~
.,.,
I 0.5
0
.....
-
!:: 0.4

0.3

0.2

0.1

0.90 0.95 1.00 1.05 1.10


w ", 10 7/sec

Fig. 38-6 The current amplitude as a function of frequency in the circuit of Fig. 38-5.
The arrow marked Liw on the curve for R = 10 ohms is the half-width of that curve.
954 ELECTROMAGNETIC OSCILLATIONS Chap. 38
38-5 Lumped and Distributed Elements
In the oscillating mass-spring system the two kinds of energy involved
appear in separate parts of the system, the potential energy being stored in
the spring and the kinetic energy in the moving mass. An acoustic cavity
resonator, such as an organ pipe, is a mechanical oscillating system in which
the two forms of energy are not separated in space. Kinetic energy, asso-
ciated with moving air in the cavity, and potential energy, associated with
compressions or rarefactions of the air, can both be present throughout the
volume of the cavity. Such a cavity is an example of an oscillating system
with distributed rather than lumped (as in the mass-spring system) elements.
A similar distinction exists in electromagnetic systems. The LC circuit
of Fig. 38-1 is an example of lumped elements, in that the two kinds of
energy are stored in rather different places; the circuit is described by giving
the (lumped) system parameters Land C. Actually, in modern engineering
practice and in physics research electromagnetic systems with distributed
elements play a fundamental role.
Figure 38-7, a series of "snapshots" taken one-eighth of a cycle apart,
shows the pressure and velocity variations in the fundamental mode of a
particular acoustic resonator. There is a pressure node at the center and a
pressure antinode at each end. There is a velocity * node at each end and
a velocity antinode at the center. When the pressure variation is the great-
est, the velocity is zero (Figs. 38-7a and e). When the pressure is uniform,
the velocities have their maximum values (Figs. 38-7c, and g).
The energy in the acoustic resonator alternates between kinetic energy
associated with the moving gas and potential energy associated with the
compression and rarefaction of the gas. In Figs. 38-7c and g the energy is
all kinetic and in Figs. 38-7a and e it is all potential. In intermediate phases
it is part of each.
The kinetic energy of a small mass t::.m of the gas, which is moving parallel to the
cylinder axis with a speed Vg, is !t::.mvg 2, The kinetic energy density, that is, the
kinetic energy per unit volume, is

2 iv Vg
1 t::.m 2 1 2
UK = = zpoVg ,

in which t::. V is the volume of the gas element and po is the mean density of the gri,s.
The potential energy per unit volume in the gas, that is, the potential energy dens~ty,
associated with the compressions and rarefactions of the gas may be given by

Here B is the bulk modulus of elasticity of the gas and t::.p/Po, which is positive for a
compression and negative for a rarefaction, is the fractional change in gas density.

* The velocity of interest here is the directed velocity Vg of small volume elements of
the gas which are, however, large enough to contain a great number of molecules. The
thermal velocities of the molecules have no directional preference and are ignored.
Sec. 38-6 ELECTROMAGNETIC CAVITY OSCILLATOR 955

(c)

<+& . ···4.·.·.··,--i:,······i
~ =+~iti ~
~-
:;-r;-.. ~·:_:~. .:. ·~.. F.~
..~-:·_:~·-: . .·_~}.: .:
(b) ·.:::. :::: : UK I Up . ~> ::::···:':t>:: (d)

! \
(a) (e)

I
(h)
r:;..
~
:·..:~: .<J- .
,:,:... .......~·.· .-. .
<J,,- . . . ~ - -
:-:-:-·.

·-- .. --.:;:/ (f)


:":"}/~::::.~: :._?-:·. (g) <l- . :·~· ·: . ~-

Fig. 38-7 Showing eight stages in a cycle of oscillation of a cylindrical acoustic resonant
cavity. The bar graphs below each figure show the kinetic and potential energy. The
arrows represent the directed velocities of small volume elements of the gas. Compare
with Fig. 38-1.

The angular frequency of oscillation for the cavity of Fig. 38-7, in the fundamental
(or lowest frequency) mode shown in that figure, is found from
27rV 7rV
W1 = 271'V = -~ = T'
where vis the speed of sound in the gas and l is the length of the cavity. From Eq.
20-1 we may write v as yB 0 /p 0 • Note that in the above we have put 'A = 2Z, cor-
responding to the fundamental mode. What are the frequencies w2, wa, etc., of the
higher frequency modes?

38-6 Electromagnetic Cavity Oscillator


Consider now a second closed cylinder, of radius a and length Z, whose
walls are made of copper or some other good conductor. A system of oscil-
lating electric and magnetic fields can be set up in such a cavity even if, as
in the common case that we consider, the cavity contains no material me-
dium. Such an electromagnetic cavity resonator is a distributed electromag-
netic oscillator, in contrast to an LC circuit, which is a lumped system. As
956 ELECTROMAGNETIC OSCILLATIONS Chap. 38
for the acoustic resonator, many modes of oscillation with discrete frequen-
cies are possible; we describe only the fundamental mode, which has the
lowest frequency. The cavity oscillations can be set up by suitably con-
necting the cavity, through a small hole in its side wall, to a source of elec-
tromagnetic radiation such as a magnetron. If the cavity dimensions are
of the order of a few centimeters, the resonant frequencies will be of the
order of 10 10 cycles/sec. This is far higher than the acoustic frequencies in
cavities of the same size, reflecting the fact that the speed of electromagnetic
disturbances in free space ( = 3 X 108 meters/sec) is much greater than the
speed of sound in air (about 350 meters/sec).
Figure 38-8, which is a series of "snapshots" taken one-eighth of a cycle
apart, shows, by the horizontal lines, how the electric field E varies with
time in the cavity. The electric lines of force originate on positive charges
at one end of the cylinder and terminate on negative charges at the other
end. As E changes with time, eventually reversing itself, currents flow from
one end of the cylinder to the other on the inner cylinder wall. At any point

(c)

.................
. . ... ... . ..
. . .- . .... . -.
x x x x
xxxxxxxx
xxxxxxxxx-xxxxxx -

(b) 11------a;---x---il x (d)


lr-------o>~-------j
x x x xxxxxx

UBg 111J1i1 UE

t \
(a)
~ ~(e)
~

L!, E
I
x
Ix
x x x x x"T/"
.. . . ... . . ..
x x
(h) (f)
(g)
xxxxxxxxxxxxxxx
xxxxxxxx
x x x x

. ... ... . .
................•

Fig. 38-8 Showing eight stages in a cycle of oscillation of a cylindrical electromagnetic


resonant cavity. The bar graphs below each figure show the stored electric and magnetic
energy. The dots and crosses represent circular lines of B; the horizontal lines represent E.
Compare with Fig. 38-7.
Sec. 38-6 ELECTROMAGNETIC CAVITY OSCILLATOR 957

in the cavity, energy is stored in the electric field in an amount per unit
volume given by Eq. 30-27, or
1 E2 (38-16)
UE = zEo .

Figure 38-8 also shows, by the dots and crosses, how the magnetic field B
varies with time. The magnetic lines form circles about the cylinder axis.
Note that the magnetic field has a maximum value when the electric field is
zero, and conversely. At any point in the cavity energy is stored in the
magnetic field in an amount per unit volume given by Eq. 36-19, or

1 2
UB = -B. (38-17)
2µo
Thus, as in the LC circuit, energy is shuttled back and forth between the
electric and the magnetic fields. However, these fields no longer occupy
completely separate regions of space.
We state without proof that the angular frequency of oscillation for the electro-
magnetic cavity of Fig. 38-8 is, in the fundamental mode shown in that figure,
1.19c
W1 = --,
a
in which a is the cavity radius and c is the speed of electromagnetic radiations in
free space. We will see in Section 39-5 that c may be written in terms of electro-
magnetic quantities as 1/ ~ - As the field patterns of Fig. 38-8 suggest, the
resonant frequency of oscillation of the cavity, for the mode of oscillation shown,
depends only on the radius of the cavity and not on its length.
Table 38-2 summarizes some characteristics of the four oscillating systems that
we have discussed so far. For lumped systems it gives expressions for the two kinds

Table 38-2
FouR OscrLLATING SYSTEMS

Mechanical Systems Electromagnetic Systems

Lumped Mass + spring LC circuit


systems UK= 1mv2 Us= !Li 2
Up= !kx 2 UE =!(1/C)q2
w =~(1/C)
w =~; L

Distributed Acoustic cavity Electromagnetic cavity


systems UK= !Pov/ Us = !Cl/ µo)B 2
Up = !B(b.p/ p 0) 2 UE =
!EoE2
3.14v
W1 = - - ·
l ' v=l
Po
W1 =-a-;
1.19c
c = ~E:olµo
958 ELECTROMAGNETIC OSCILLATIONS Chap. 38
of energy involved and for the (single) oscillation frequency. For the distributed
systems it gives expressions for the two kinds of energy density involved and for the
oscillation frequency in the fundamental mode. The student should study carefully
all the correspondences, similarities, and differences that occur in the table.
The expression given above for the fundamental resonant angular frequency w 1 of
the cavity is derived by applying the basic equations of electromagnetism (Table
38-3) to the space inside the cavity and by invoking this boundary condition: E
must be zero inside the cavity wall and can have no tangential component anywhere on
the cavity wall. If this were not true, an infinite current would be set up in the as-
sumed resistanceless wall by the tangential component of E. This boundary con-
dition is similar in spirit to the requirement that a clamped string must have zero
amplitude of oscillation at the points at which it is clamped or that a velocity node
must exist at the end walls of an acoustic cavity resonator. Granted this boundary
condition, on E, it can be shown that, assuming a perfectly conducting wall, (a) no
time-varying magnetic field can exist inside the cavity wall and (b) no time-varying
currents can exist inside the wall. A tangential magnetic field can exist on the surface,
however; surface charges can exist, and surface currents can flow.

IJJl,- Example 4. In the cavity of Fig. 38-8, what is the relationship between the
"average" value of E throughout the cavity, measured at the instant corresponding
to Fig. 38-8a, to the "average" value of B, measured at the instant corresponding to
Fig. 38-8c?
At the first instant the energy is all electric and at the second it is all magnetic.
The total energy U, found by integrating the energy density over the volume of the
cavity, must be the same at these two instants, or

U = f UE,mdV = f UB,mdV,
where dV is a volume element in the cavity and UE,m and UB,m are the maximum values
of UE and of UB at the site of this volume element; these maximum values occur one-
fourth of a cycle apart, as Fig. 38-8 shows. Substituting Eqs. 38-16 and 38-17 leads
to

I
foEm2 dV =
2
I
Bm2 dV
2µo

or

f
The quantity Em 2 dV can be written as Em 2V, where V is the cavity volume and
Em 2 is the average value of Em 2 throughout the cavity. Treating Bm in the same way
leads to

or, taking square roots,

We can represent V Bm 2 by Brms, a so-called "root-mean-square" value of Bm.


In computing Brms, note that the averaging is done throughout the volume of the
cavity, at the instant corresponding to Fig. 38-8c. It is not a time average for a
particular point in the cavity. Doing the same for E yields

Brms = ~ Erms = V(41r x 10-7 weber/amp-m)(8.9 x 10-12 coul2/nt-m 2) Erms

= (3.3 X 10- 9 sec/meter)Erms·


Sec. 38-7 INDUCED MAGNETIC FIELDS 959
If Erms equals 104 volts/meter, a reasonable value, then

Brms = (3.3 X 10-9 sec/meter)(10 4 volts/meter)

= 3.3 X 10-5 weber/meter 2 = 0.33 gauss.

What is the total stored energy in the cavity under these conditions, assuming the
cavity to be 10 cm long and 3.0 cm in diameter?
The student will recall that in Example 6, Chapter 36, we showed the energy
density for a magnetic field of "ordinary" laboratory magnitude (say, 1 weber/meter2 )
to be enormously greater than that for an electric field of "ordinary" magnitude
(say, 10 5 volts/meter). This fact is consistent with the present example. <1111

38-7 Induced Magnetic Fields


To understand the electromagnetic cavity oscillations in terms of electro-
magnetic theory, we must complete our description of the basic equations of
electromagnetism by introducing a new concept, namely, that a changing
electric field produces a magnetic field. This concept, which is the symmetrical
counterpart of Faraday's law of induction, is of fundamental significance.
We will develop this concept by a symmetry argument and will let the
agreement with experiment of our final conclusions speak for itself. This
comparison with experiment, which is worked out largely in Chapters 39
and 40, forms one of the chief experimental bases of electromagnetic theory.
A central achievement was the demonstration that the experimentally
measured speed c of visible light in free space could be related to purely
electromagnetic quantities by
1
(38-18)

This demonstration not only revealed optics as a branch of electromagnetism


but led directly to the concept of the electromagnetic spectrum, which in
turn resulted in the discovery of radio waves.
Figure 38-9a shows a uniform electric field E filling a cylindrical region of
space. It might be produced by a circular parallel-plate capacitor, as sug-
gested in Fig. 38-9b. We assume that E is increasing at a steady rate
dE / dt, which means that charge must be supplied to the capacitor plates at a
steady rate; to supply this charge requires a steady current i into the pos-
itive plate and an equal steady current i out of the negative plate.
If a sufficiently delicate experiment could be performed, it would be found
that a magnetic field is set up by this changing electric field. Figure 38-9a
shows B for four arbitrary points. Figure 38-9 suggests a beautiful example
of the symmetry of nature. A changing magnetic field induces an electric
field (Faraday's law); now we see that a changing electric field induces a
magnetic field.
To describe this new effect quantitatively, we are guided by analogy with
Faraday's law of induction,
d<Ps
--, (38-19)
dt
960 ELECTROMAGNETIC OSCILLATIONS Chap. 38

E
.,,,..,,,,. .,..-- ------ -.... .............. + { -
/x x x x x,
/ // ',,.,/E + -

/ 'Ac- + -
/ X x x x,
I B ' + -
I \
/ xx xx xx\ + -
I r , -
+
Ix
I
x x x x< x x x\
J + -
I
IX X X X X X X X XI
I i + -- i
I I
+ -
\ R /
\xx xx x I + -
\ I
\
\ I
I
+ -
,x x x x x/
', B / + -
', ;/
+ -
_________
'~.... ......... x x x ,,..,,. .,,."'>""
-
+
(a) (b)

Fig. 38-9 (a) Showing the induced magnetic fields Bat four points, produced by a chang-
ing electric field E. The electric field is increasing in magnitude. Compare Fig. 35-10.
(b) Such a changing electric field may be produced by charging a parallel-plate capacitor as
shown.

which asserts that an electric field (left term) is produced by a changing


magnetic field (right term). For the symmetrical counterpart we might·
write *
(38-20)

Equation 38-20 asserts that a magnetic field (left term) can be produced by
a changing electric field (right term). The student should compare carefully
Fig. 35-10, which illustrates the production of an electric field by a changing
magnetic field, with Fig. 38-9a. In each case the appropriate flux <PB or <PE
is increasing. However, experiment shows that the lines of E in Fig. 35-10
are counterclockwise, whereas those of B in Fig. 38-9a are clockwise. This
difference requires that the minus sign of Eq. 38-19 be omitted from Eq.
38-20.
In Section 34-1 we saw that a magnetic field can also be set up by a cur-
rent in a wire. We described this quantitatively by Ampere's law:

in which i is the conduction current passing through the loop around which
the line integral is taken. Thus there are at least two ways of setting up

* Our system of units requires that we insert the constants eo and µo in Eq. 38-20. In
some unit systems they would not appear.
Sec. 38-7 INDUCED MAGNE'rIC FIELDS 961

a magnetic field: (a) by a changing electric field and (b) by a current. In


general, both possibilities must be allowed for, or *

(38-21)

Maxwell is responsible for this important generalization of Ampere's law.


In Chapter 34 we assumed that no changing electric fields were present so
that the term dcI>E/dt in Eq. 38-21 is zero. In the discussion just given we
assumed that there were no conduction currents in the space containing the
electric field. Thus the term i in Eq. 38-21 is zero. We see now that each
of these situations is a special case.
~ Example 5. A parallel-plate capacitor with circular plates is being charged as
in Fig. 38-9. (a) Derive an expression for the induced magnetic field at various radii
r. Consider both r < Rand r > R.
From Eq. 38-20,
f B·dl = dcpE
µ01=0 - - ,
dt
we can write, for r < R,
d dE
(B)(21rr) = µ01=0 dt [(E)(1rr 2 )] = µ01=01rr 2 dt.
dE
Solving for B yields B =
1
wo1:ordt (r < R).
For r > R, Eq. 38-20 yields
(B)(21rr) = d [(E)(1rR 2)]
µ01:0 dt = µ01:01rR 2 (dE)
-dt ,

or B = µ01:oR~ dE
2r dt
(r > R).
(b) Find Bat r = R for dE/dt = 10 12 volts/m-sec and for R = 5.0 cm. At r = R
the two equations for B reduce to the same expression, or

= (!)(41r X 10-7 weber/amp-m)(8.9 X 10- 12 coul2/nt-m 2)


(5.0 X 10- 2 meter)(l0 12 volts/m-sec)
= 2.8 X 10- 7 weber/meter 2 = 0.0028 gauss.
This shows that the induced magnetic fields in this example are so small that they
can scarcely be measured with simple apparatus, in sharp contrast to induced electric
fields (Faraday's law), which can be demonstrated easily. This experimental dif-
'- Actually, there is a third way of setting up a magnetic field, by the use of magnetized
bodies. In Section 37-7 we saw that this could be accounted for by inserting a magnetizing
current term iM on the right side of Ampere's law. This law would then read, in its full
generality,

f B·dl = µ.oEo -dt.


dif>E
+ µ.oi + µ.oiM.
.

In all that follows we assume that no magnetic materials are present so that iM = 0.
962 ELECTROMAGNETIC OSCILLATIONS Chap. 38
ference is in part due to the fact that induced emfs can easily be multiplied by using
a coil of many turns. No technique of comparable efficiency exists for magnetic
fields. In experiments involving oscillations at very high frequencies dE/dt above
can be very large, resulting in significantly larger values of the induced magnetic
field. ~

38-8 Displacement Current


Equation 38-21 shows that the term 1:0 d<PE/dt has the dimensions of a
current. Even though no motion of charge is involved, there are advantages
in giving this term the name displacement* current. Thus we can say that a
magnetic field can be set up either by a conduction current i or by a displace-
ment current id ( = 1: 0 d<PE/dt), and Eq. 38-21 can be rewritten as t

f B·dl = µ 0(id + i). (38-22)

The concept of displacement current permits us to retain the notion that


current is continuous, a principle established for steady conduction currents
in Section 31-1. In Fig. 38-9b, for example, a current i enters the positive
plate and leaves the negative plate. The conduction, current is not continuous
across the capacitor gap because no charge is transported across this gap.
However, the displacement current id in the gap will prove to be exactly i,
thus retaining the concept of the continuity of current.
To calculate the displacement current, recall (see Eq. 30-5) that <PE in the
gap is given by

q
<PE = -·
Eo

But (see Eqs. 38-21 and 38-22)

Therefore the displacement current in the gap is identical ,vith the conduc-
tion current in the lead wires.

* The word "displacement" was introduced for historical reasons that need not concern
us here.
t We may write this more generally, taking the presence of magnetic materials into
account, as

See the footnote on p. 961.


Sec. 38-10 MAXWELL'S EQUATIONS AND CAVITY OSCILLATIONS 963
~ Example 6. What is the displacement current for the capacitor of Example 5?
From the definition of displacement current,
. d~E d dE
id = Eo -dt = Eo dt [(E)('1rR 2)] = 1:01rR 2 dt
= (8.9 X 10- 12 coul2/nt-m 2)(1r)(5.0 X 10-2 meter)2(10 12 volts/m-sec)
= 0.070 amp.
Even though this displacement current is reasonably large, it produces only a small
magnetic field (see Example 5) because it is spread out over a large area. <1111

38-9 Maxwell's Equations


Equation 38-21 completes our presentation of the basic equations of elec-
tromagnetism, called Maxwell's equations. They are summarized in Table
38-3. All equations of physics that serve, as these do, to correlate experi-
ments in a vast area and to predict new results have a certain beauty about
them and can be appreciated, by those who understand them, on an aesthetic
level. This is true for Newton's laws of motion, for the laws of thermody-
namics, for the theory of relativity, and for the theories of quantum physics.
As for Maxwell's equations, the German physicist Ludwig Boltzmann (quot-
ing a line from Goethe) wrote "Was it a god who wrote these lines. . . ." In
more recent times J. R. Pierce,* in a book chapter entitled "Maxwell's Won-
derful Equations" says: "To anyone who is motivated by anything beyond
the most narrowly practical, it is worth while to understand Maxwell's equa-
tions simply for the good of his soul." The scope of these equations is re-
markable, including as it does the fundamental operating principles of all
large-scale electromagnetic devices such as motors, cyclotrons, electronic
computers, television, and microwave radar.

38-10 Maxwell's Equations and Cavity Oscillations


In this section we show how the oscillations of an electromagnetic cavity
can be understood in terms of Maxwell's equations. A completely formal
treatment, which is beyond our scope here, would start from these equations
and would end with mathematical expressions for the variation of B and E
with time and with position in the cavity for all modes of oscillation of the
cavity. We confine ourselves to the fundamental mode only, illustrated in
Fig. 38-8, for which we postulated the variations of B and E given in that
figure; we will show that these postulated fields are completely consistent
with Maxwell's equations.
Figure 38-10 presents two views of the cavity of Fig. 38-8d, in which both
electric and magnetic fields are present. Study of Fig. 38-8 reveals that B
is decreasing in magnitude and E is increasing at this phase of the cycle of
oscillation. Let us apply Faraday's law,

* Electrons, Waves and lv! essages, Hanover House, 1956. This book is recommended as
collateral reading in electromagnetism.
Table 38-3
°''°.s:,.
THE BASIC EQUATIONS OF ELECTROMAGNETISM (MAXWELL'S EQUATIONS)*

Text
Name Equation Describes Crucial Experiment
Reference
-
lctj
t"
lctj
Cl
Gauss's law for 1:0 fE·dS = q Charge and the electric field 1. Like charges repel and unlike charges attract, as Chapter 28 >-3
~
electricity the inverse square of their separation. 0
l'. A charge on an insulated conductor moves to its ~
0
outer surface. 'Z
~
H

Gauss's law for f B·dS = 0 The magnetic field 2. It is impossible to create an isolated magnetic Section 37-2 Cl
0
magnetism pole. Ul
Cl
H
t"
t"
Ampere's law (as ex- fB·dl The magnetic effect of a 3. The speed of light can be calculated from purely Section 39-5
~
tended by Maxwell) changing electric field or electromagnetic measurements. H
0
= µo ( Eo dcJ>E
dt + i ') of a current 3'. A current in a wire sets up a magnetic field near Chapter 34 ~
al
' the wire.

Faraday's law of in- f E·dl = - dPs The electrical effect of a 4. A bar magnet, thrust through a closed loop of Chapter 35
duction dt changing magnetic field wire, will set up a current in the loop.
Q
:::;"
~
* Written on the assumption that no dielectric or magnetic material is present. "Cl
c,.,
00
Sec. 38-10 MAXWELL'S EQUATIONS AND CAVITY OSCILLATIONS 965

r--- h
-
·-
• I •
· ~-· •
r
. . . . . . . . ..•t • ... ..
E e • - •J•
. . . .I. .
--=---t>- - ~. --

x x x x x x x

x x x x x x x x x x x
X- x x x x x x x x x x x x x x x

(a) (b)

Fig. 38-10 Two cross sections of an electromagnetic resonant cavity at a phase of oscilla-
tion corresponding to Fig. 38-Sd. (a) The dashed circle is a path suitable for applying
Ampere's law. (b) The dashed rectangle is a path suitable for applying Faraday's law.

to the rectangle of dimensions h and a - r. There is a definite flux <l'>B


through the rectangular path in question, and this flux is decreasing with
time because B is decreasing. The line integral above is

f E·dl = hE(r),

in which E(r) is the value of E at a radius r from the center of the cavity.
Note that E equals zero for the upper leg of the integration path (which
lies in the cavity wall) and that E and di are at right angles on the two side
legs. Combining these equations yields

1 d<l'>B
E(r) = - - - · (38-23)
h dt

Equation 38-23 shows that E(r) depends on the rate at which cf>B through
the path shown is changing with time and that it has its maximum value
when d<I>B/dt is a maximum. This occurs when Bis zero, that is, when Bis
changing its direction; the student will recall that a sine or cosi,1e is changing
most rapidly, that is, it has the steepest tangent, at the instant it crosses the
axis between positive and negative values. Thus the electric field pattern
in the cavity will have its maximum value when the magnetic field is zero
everywhere, consistent with Figs. 38-8a and e and with the concept of the
interchange of energy between electric and magnetic forms. The student
should demonstrate, by applying Lenz's law, that the electric field in Fig.
38-lOb points to the right as shown, if the magnetic field is decreasing.
Figure 38-lOa shows an end view of the cavity; the electric lines of force
are entering the page at right angles to the page and the magnetic lines form
clockwise circles. Let us apply Ampere's law in the form
'
(38-21)
966 ELECTROMAGNETIC OSCILLATIONS Chap. 38
to the circular path of radius r shown in the figure. ~ o charge is transported
through the ring so that the conduction current i in Eq. 38-21 is zero. The
line integral on the left is (B) (21rr) so that the equation reduces to

(38-24)

Equation 38-24 shows that the magnetic field B(r) is proportional to the
rate at which the electric flux cf>E through the ring is changing with time.
The field B(r) has its maximum value when dcf>E/dt is at its maximum; this
occurs when E = 0, that is, when E is reversing its direction. Thus we see
that B has its maximum value when E is zero for all points in the cavity.
This is consistent with Figs. 38-8c and g and with the concept of the inter-
change of energy between electric and magnetic forms. A comparison with
Fig. :"38-9a, which, like Fig. 38- lOa, corresponds to an increasing electric
field, shows that the lines of B are indeed clockwise, as viewed along the di-
rection of the electric field.
Comparison of Eqs. 38-23 and 38-24 suggests the complete interdepend-
ence of B and E in the cavity. As the magnetic field changes with time, it
induces the electric field in a way described by Faraday's law. The electric
field, which also changes with time, induces the magnetic field in a way
described by Maxwell's extension of Ampere's law. The oscillations, once
established, sustain each other and would continue indefinitely were it not
for losses due to Joule heating in the cavity walls or leakage of energy from
openings that might be present in the walls. In Chapter 39 we show that
this interplay of B and E occurs not only in standing electromagnetic waves
in cavities but also in traveling electromagnetic waves, such as radio waves
or visible light.

Let us now analyze the currents-both conduction and displacement-that occur


in the cavity and examine their connections to the electric and the magnetic fields.
Figure 38-11 shows two views of the cavity, taken at an instant corresponding to that
of Fig. 38-10. For simplicity the fields E and Bare not shown; the arrows represent
currents.
Since Eis increasing at this instant, the positive charge on the left end cap must be
increasing. Thus there must be conduction currents in the walls pointing from right
to left in Fig. 38-llb. These currents are also shown by the dots (representing the
tips of arrows) near the cavity walls in Fig. 38-lla.
Bearing in mind that fo dcf>E/dt is a displacement current, we can write Eq. 38-24 as

µo dcf>E µo .
B(r) = -- fo --- = -- ia.
21rr dt 21rr

This equation stresses that B in the cavity is associated with a displacement current;
compare Eq. 34-4. Applying the right-hand rule in Fig. 38-lOa shows that the dis-
placement current ia must be directed into the plane of the figure if it is to be associ-
ated with the clockwise lines of B that are present.
The displacement current is represented in Fig. 38-llb by the arrows that point to
the right and in Fig. 38-lla by the crosses that represent arrows entering the page.
Study of Fig. 38-11 shows that the current is continuous, being directed up the
Chap. 38 QUESTIONS 967

--+ --+ --+ --+ --+ --+

--+ --+ --+ --+ --+ --+


--+ --+ --+ ---+ --+ --+

(a) (b)

Fig. 38-11 The cavity of Fig. 38-10 showing (a) the conduction current coming up the
walls and displacement current going down the cavity volume and (b) the displacement
current (black arrowheads) in the volume of the cavity and the conduction currents (white
arrowheads) at the walls. The arrows in each case represent current densities. Note
the continuity of current (conduction + displacement), that is, it is possible to form
closed current loops. (For a truly resistanceless cavity the conduction current would
lie entirely on the surface, requiring a modification of our definition of current density;
we leave this small complication for the student to consider.)

walls as a conduction current and then back down through the volume of the cavity as
a displacement current. If we apply Ampere's law as extended by Maxwell,

f B·dl = µo(ia + i), (38-25)

to the circular path of radius r 1 in Fig. 38-lla, we see that B at that path is due
entirely to the displacement current, the conduction current i within the path being
zero.
For the path of radius r 2, the net current enclosed is zero because the conduction
current in the walls is exactly equal and opposite to the displacement current in the
cavity volume. Since i equals ia in magnitude, but is oppositely directed, it follows
from Eq. 38-25 that B must be zero for all points outside the cavity, in agreement
with observation.

QUESTIONS

1. Why doesn't the LC circuit of Fig. 38-1 simply stop oscillating when the capacitor
has been completely discharged?
2. How might you start an LC circuit into oscillation with its initial condition· being
represented by Fig. 38-lc? Devise a switching scheme to bring this about.
3. In an oscillating LC circuit, assumed resistanceless, what determines (a) the fre-
quency and (b) the amplitude of the oscillations?
4. Tabulate as many mechanical or electric systems as you can think of that possess a
natural frequency, along with the formula for that frequency if given in the text.
5. Resonance in LC circuits, judged by Eq. 38-14 and Fig. 38-6, occurs when the fre-
quency w" of the "driving force" is exactly equal to the (undamped) natural frequency
of the LC circuit. In Section 15-10 we saw that resonance for mass-spring systems,
judged by Eq. 15-41 and Fig. 15-19, occurs when w" is close to, but not exactly equal to,
the (undamped) natural frequency of the mass-spring system. Is there a failure of the
principle of correspondence here?
968 ELECTROMAGNETIC OSCILLATIONS Chap. 38
6. Suppose that, in a low frequency oscillating LCR circuit, one could display on the
screen of a cathode-ray oscilloscope a potential difference proportional to the temperature
of the (specially designed) resistor as it changed with time in response to the hes.ting effect
of the oscillating current. What would the connection be between the frequency of this
temperature-time waveform and that of the oscillations of current, as measured by the
potential difference across the resistor?
7. Discuss the periodic flow of energy, if any, from point to point in an acoustic resonant
cavity.
8. Can a given circuit element (a capacitor, say) behave like a "lumped" element under
some circumstances and like a "distributed" element under others?
9. List as many (a) lumped and (b) distributed mechanical oscillating systems as you
can.
10. An air-filled acoustic resonant cavity and an electromagnetic resonant cavity of the
same size have resonant frequencies that are in the ratio of 106 or so. Which has the higher
frequency and why?
11. What constructional difficulties would you encounter if you tried to build an LC
circuit of the type shown in Fig. 38-1 to oscillate (a) at 0.01 cps, or (b) at 1010 cycles/sec?
12. Could an electromagnetic cavity of the type shown in Fig. 38-8 be designed to
operate at 60 cycles/sec? If so, give some details of its construction.
13. The "sharpness of tuning" of a copper electromagnetic cavity can be considerably
ipcreased by immersing it in liquid air. Explain. Be guided by Fig. 38-6, which shows
"tuning curves" for an LC circuit.
14. Electromagnetic cavities are often silver plated on the inside. Explain.
15. Why is Faraday's law of induction more familiar than its symmetrical counterpart,
Eq. 38-20?
16. Why is the quantity EO d4!E/dt referred to as a (displacement) current1
17. In Fig. 38-lc a displacement current is needed to maintain continuity of current in
the capacitor. How can one exist, since there is no charge on the capacitor?
18. Discuss the symmetries that appear between (a) the first two and (b) the second two
of Maxwell's equations.
19. At what parts of the cycle will (a) the conduction current and (b) the displacement
current in the cavity of Fig. 38-8 be zero?
20. Discuss the time variation during one complete cycle of the charges that appear
at various points on the inner walls of the oscillating electromagnetic cavity of Fig. 38-8.

PROBLEMS

l. You are given a 10-mh inductor and two capacitors, of 5.0- and 2.0-µ.f capacitance.
What resonant frequencies can be obtained by connecting these elements in various ways?
2. Given a 1.0-mh inductor, how would you make it oscillate at 1.0 X 106 cycles/sec?
3. An inductor is connected across a capacitor whose capacitance can be varied by
turning a knob. We wish to make the frequency of the LC oscillations vary linearly with
the angle of rotation of the knob, going from 2 X 105 cycles/sec to 4 X 105 cycles/sec as
the knob turns through 180°. If L = 1.0 mh, plot Casa function of angle for the 180°
rotation.
4. A 10-henry coil has a resistance of 180 ohms. What size of capacitor must be put in
series with it if the combination is to "resonate" when connected to a 60-cycle/sec outlet?
5. Derive the differential equation for an LC circuit (Eq. 38-5) using the loop theorem.
6. Derive Eq. 38-12, the differential equation for forced oscillations in an LCR circuit,
from the principle of conservation of energy.
7. Show that the q-oscillations in a damped LC circuit (see Eq. 38-10) reach half
their initial amplitude in a time given approximately by 0.69 (L/R).
8. A circuit has L = 10 mh and C = 1.0 µ.f. How much resistance must be inserted in
the circuit to reduce the (undamped) resonant frequency by 0.01 %?
Chap. 38 PROBLEMS 969
9. Suppose that in an oscillating LCR circuit the amplitude of the charge oscillations
drops to one-half its initial value after n cycles. Show that the fractional reduction in
the frequency of resonance, caused by the presence of the resistor, is given to a close
approximation by
w - w' 0.0061
--w-=~,

which is independent of L, C, or R. Apply to the decay curve of Fig. 38-3.


10. Show that, for low damping, the current in a damped LC circuit is given approxi-
mately by

-1 R
in which </>=tan --·
2Lw'
Start from Eq. 38-10.
11. "Q" for a circuit. In the damped LC circuit of Example 3 show that the fraction
of the energy lost per cycle of oscillation, l}U /U, is given to a close approximation by
21rR/wL. The quantity wL/R is often called the "Q" of the circuit (for "quality"). A
"high-Q" circuit has low resistance, low fractional energy loss per cycle ( = 21r /Q) and
(see Fig. 38-6) a sharp resonance or "tuning" curve.
12. Show that the amplitude of the charge oscillations in an oscillating LCR circuit h1
given by
8m
~;====================
qm =
~( w"2L -iY + (w"R)2

For what value of w" will qm be a maximum?


13. An LCR circuit has L = 1.0 henry, C = 20 µ.f, and R = 20 ohms. For what
frequency w" of an applied emf will it resonate with maximum response? At what fre-
quencies will the response be one-half its maximum value?
14. Show that the fractional half-width of the resonance curves of Fig. 38-6 is given,
to a close approximation, by

in which w is the resonant frequency and /}w is the width of the resonance peak at i = Mm.
Note (see Problem 11) that this expression may be written as y'3/Q which shows clearly
that a "high-Q" circuit has a sharp resonance peak, that is, a small /}w/w.
15. A resistor-inductor-capacitor combination R1, Li, C1 (connected in series) exhibits
resonance at the same frequency as a second combination R2, L2, C2. If the two com-
binations are now connected in series, at what frequency would the whole circuit resonate?
16. In Example 5 show that the displacement current density ja is given, for r < R, by

dE
Jd = EO
dt'
17. Prove that the displacement current in a parallel-p~ate capacitor can be written as

dV
ia = C--·
dt

18. You are given a 1.0-µ.f capacitor. How would you establish an (instantaneous) dis-
placement current of 1.0 amp in the space between its plates?
19. In Example 5 how does the displacement current through a concentric circular loop
of radius r vary with r? Consider both r < Rand r > R.
970 ELECTROMAGNETIC OSCILLATIONS Chap. 38
20. In microscopic terms the principle of continuity of current may be expressed as

f (j + ja) ·dS = 0,

in which j is the conduction current density and ja is the displacement current density.
The integral is to be taken over any closed surface; the equation essentially says that
whatever current flows into the enclosed volume must also flow out. (a) Apply this equa-
tion to the surface shown by the dashed lines in Fig. 38-12 shortly after switch Sis closed.
(b) Apply it to various surfaces that may be drawn in the cavity of Fig. 38-11, including
some that cut the cavity walls.

Fig. 38-12

21. A cylindrical electromagnetic cavity 5.0 cm in diameter and 7.0 cm long is oscillating
in the mode shown in Fig. 38-8. (a) Assume that, for points on the axis of the cavity,
Em = 10 4 volts/meter. For such axial points what is the maximum rate (dE /dt)m at which
E changes? (b) Assume that the average value of (dE /dt)m, for all points over a cross
section of a cavity, is about one-half the value found above for axial points. On this
assumption, what is the maximum value of Bat the cylindrical surface of the cavity?
--:22:· Collect and tabulate expressions for the following four quantities, considering both
r < R and r > R. Copy down the derivations side by side and study them as interesting
applications of Maxwell's equations to problems having cylindrical symmetry.
(a) B(r) for a current i in a long wire of radius R (see Section 34-2).
(b) E(r) for a long uniform cylinder of charge of radius R (see Section 28-6; also Problem
14, Chapter 28).
(c) B(r) for a parallel-plate capacitor, with circular plates of radius R, in which E is
changing at a constant rate (see Section 38-7).
(d) E(r) for a cylindrical region of radius R in which a uniform magnetic field B 1s
changing at a constant rate (see Section 35-5).
Electromagnetic Waves *
CHAPTER 39

39-1 Transmission Line


In Chapter 38 we studied electromagnetic energy confined, as a standing
wave, to a restricted region of space, the interior of an electromagnetic
resonant cavity. Such energy can also be transferred from place to place as
a traveling wave. An arrangement of conductors for facilitating such trans-
£ers is called a transmission line. Figure 39-1 shows one type of line, a
coaxial cable, its input end being connected to a switch S. For the time
being we assume that the cable is infinitely long and that the cable elements
have zero resistance.
When Sis closed on b, the central and the outer conductors are at the same
potential. If the switch is then thrown to a, a potential difference V sud-
denly appears between these elements. This potential difference does not
appear instantaneously all along the line but is propagated with a finite

s
a b

~ 1____c...____

Fig. 39-1 An electromagnetic pulse can be sent along the coaxial cable by throwing
switch S from b to a.
971

* This subject is treated in greater depth in Supplementary Topic V at the end of


the book.
972 ELECTROMAGNETI C WAVES Chap. 39

speed c that will turn out to be exactly that of light, assuming a resistanceless
line. Figure 39-2a shows that the potential difference between the con-
ductors at a distance l along the line suddenly rises, at a time given by t =
l/c, from zero to a value determined by the battery emf. We can also con-
sider the variation of V with position x along the line at a given time t 1
after closing the switch. Figure 39-2b shows such an instantaneous "snap-
shot." It, too, suggests a traveling "wavefront" moving along the line at
speed c. At t = t 1 the signal has not yet reached points where x > ct 1 .

Fig. 39-2 (a) The variation with time of the potential difference between the conductors
of a coaxial cable at a distance l from the input end. (b) An instantaneous "snapshot" of
the pulse in the cable at a certain time t1. (c) The waveform if switch S in Fig. 39-1 is
periodically thrown between a and b. (d) The waveform if switch S is replaced by an
electromagnetic oscillator with a sinusoidal output.
Sec. 39-1 TRANSMISSION LINE 973

If switch Sis periodically thrown from b to a and back again, a wave dis-
turbance like that of Fig. 39-2c is propagated. This suggests that if the
battery and switch arrangement is replaced by an electromagnetic oscillator
with a sinusoidal output of frequency v a wave like that of Fig. 39-2d will
be propagated.
A traveling wave in a resistanceless transmission line will exhibit a wave-
length A given by
c
/\ = -·
v

If the oscillator frequency is 60 cycles/sec, the common commercial power


frequency, the wavelength is 5 X 106 meters, which is about 3000 miles. At
this low frequency traveling waves are not apparent in any line of normal
length. By the time the polarity of the oscillator has changed appreciably,
the energy fed into the line at the oscillator end has been delivered to the
load.
Frequencies in the radio or the microwave range are much higher and the
wavelengths correspondingly smaller. Commercial television frequencies as
established by the Federal Communications Commission range from 54 X
10 6 to 980 X 10 6 cycles/sec. In terms of wavelength this is a range of 5.6
to 0.31 meter. At these wavelengths the patterns of potential difference in
the transmission lines used to send television signals across the country can
be described aptly as traveling waves. Microwaves, used in radar systems
and for communication purposes, have even smaller wavelengths, in the
range of about 20 cm to about 0.5 mm.
These considerations suggest another way of viewing the difference be-
tween lumped and distributed circuit elements. A system is "distributed"
if the wavelength is about the same size as, or less than, the dimensions of the
system. If the wavelength is much larger than the dimensions of the
system, we are dealing with lumped components. A transmission line 50
meters long would be a lumped system for electromagnetic radiation at 60
cycles/sec (A = 5 X 10 6 meters) but a distributed system at 108 cycles/sec
(A = 3 meters). In a lumped system the circuit analysis is normally carried
out in terms of lumped system parameters such as L, C, and R; in a distrib-
uted system the analysis is often carried out in terms of the fields that are
set up and the charges and currents that are related to them.

~ Example 1. A potential difference given by

Vo= V m sin wt

is applied between the terminals of a long resistanceless transmission line; the fre-
quency P ( = w/21r) is 3 X 109 cycles/sec. Write an equation for V(t) at a point P
which is 1.5 wavelengths down the line from the oscillator.
The general equation for a wave traveling in the x direction (see Eq. 19-10) is

V = V m sin (wt - kx),


974 ELECTROMAGNETIC WA YES Chap. 39
where k ( = 271"/A) is the wave number. At x = 0 this gives correctly the time vari-
ation of the input terminal potential difference. At x = 1.5;\ we have

Vp V m sin [ wt - (~,_~) (1.5;\) J= V m sin (wt - 37r)

- V m sin wt.
Thus VP is always equal in magnitude to Vo but is opposite in sign. What is the
wavelength in this example? ~

..... x xx xx . .... x xx x x

••• xxx • •• xxx --c-C>


x x

x ...----~ ~ - - - - ~ (a) (b)

- <1---
i --I>

<1---
l <1---

--I>
~ t >Displacement

<1---
current

(c)

Fig. 39-3 (a) The electric and magnetic fields in a coaxial cable, showing a wave traveling
to the right at speed c. (b) A cross-sectional view at a plane through xx in (a); the wave is
emerging from the page. (c) Conduction current (open arrows) and displacement currents
(filled arrows) associated with the wave in (a); the arrows in each case represent current
density vectors.

39-2 Coaxial Cable-Fields and Currents


Figures 39-3a and b are "snapshots" of the electric and magnetic field
configurations in a coaxial cable. The electric field is radial and the magnetic
field forms concentric lines about the central conductor. The entire pattern
moves along the line, assumed resistanceless, at speed c.
The field patterns in this figure obey the boundary condition required for a
line that is assumed to be resistanceless, namely, that E for all points on
either conducting surface has no tangential component (see p. 957). The
field patterns can be deduced mathematically from Maxwell's equations by
Sec. 39-2 COAXIAL CABLE-FIELDS AND CURRENTS 975

imposing this requirement. The configuration shown is the simplest of


many different wave patterns that can travel along the line. The coaxial
cable, unlike the electromagnetic cavity of Fig. 38-8, is not a resonant de-
vice. The angular frequency w of waves that travel along it can be varied
continuously, as is the case for all traveling waves, such as transverse waves
in a long stretched cord.
Figure 39-3c shows the currents in the cable at the instant corresponding
to Figs. 39-3a and b. The arrows parallel to the cable axis represent conduc-
tion currents in the central and the outer conductors. The vertical arrows
with filled heads represent displacement currents that exist in the space be-
tween the conductors. Note that the conduction current and the displace-
ment current arrows form closed loops, preserving the concept of the con-
tinuity of current.

~ Example 2. Verify that the displacement current represented in Fig. 39-3c is


consistent with the pattern of Band E shown in Fig. 39-3a.
Consider a small surface element ~S shown edge-on in Fig. 39-3; it is shown as
viewed from above in Fig. 39-4. This hypothetical element is stationary with respect
to the cable while the field configuration moves through it at speed c. Figure 39-4a
shows the electric lines of force in the vicinity of this element. It is clear from sym-
metry that, at the instant shown, the net flux <l>E through this area is zero. However,
even though <t>E is zero in magnitude, it is, at this instant, changing at its most rapid

{E'\
x x x • • •
x •
x x x • • •
x ...-----.-.0is
x x • •
x •
x
x
x
x

x


·~


x •
x x • •
x •
x x x • • •
Fig. 39-4 The area element Sin Fig. 39-3 (a)
enlarged and viewed from above, showing
the adjacent (a) electric and (b) magnetic
fields.

(b)
976 ELECTROMAGNETIC WAVES Chap. 39
rate, since Eat the element Mis at the very moment of reversing its direction as the
wave moves through. Thus the displacement current, which is given by
. d<!JE
ia = t:o--•
dt
also has its maximum value.
Figure 39-4b shows the magnetic field in the vicinity of the element of area. Let
us apply the generalized form of Ampere's law,

f B·dl = µo(i + ia),


to the element. The conduction current i is zero since no charge is transported
through !::i.S. The displacement current id is not zero, having, in fact, its maximum
value. Thus, since the right side of this equation ( = µoid) does not vanish, the left
side must not vanish. Study of Fig. 39-4b shows thatfB·dl around the boundary
of this square has, indeed, a nonzero value. Thus the field and displacement current
configurations of Figs. 39-3a-c are consistent. We have not discussed the direc-
tion of ia; that is, does it point into the plane of Fig. 39-4, as Fig. 39-3c asserts, or
out of it? We leave this as a question for the student. He may be guided by con-
sidering the direction of the displacement current in Figs. 38-9a and b.

Example 3. Show that the conduction currents in Fig. 39-3c are appropriately
related to the magnetic field pattern.

/
/ -------- Conduction
/
/ ~i­ , current
I \
I \
I I Fig. 39-5 Example 3. The coaxial cable of
I \
I I Fig. 39-3b, showing the conduction currents
I I in the central and outer conductors. The
I wave is emerging from the page.
I
\
\

' "'
----~ ---------

Figure 39-5 shows a cross section of the cable at a plane through xx in Fig. 39-3a.
Let us apply Ampere's law,

f B·dl = µo(i + ia),


to the ring of radius r. The displacement current through this ring is zero, this
current being at right angles to the central conductor. The conduction current i
in the central conductor, shown by the X's in the figure, does pass through this
ring so that the equation becomes
(B)(21rr) = µoi,
or B = M':_.
21rr

Note that Bis related to the current in the central conductor by the usual right-hand
rule and that the expression for Bis that found earlier (Eq. 34-4) for a long straight
wire carrying a steady current.
Sec. 39-3 WAVEGUIDE 977

·. ·:.:itil .'.:;:4iAl ;::-::.;::/il .J::::i4YJJlti)f-~-r(~~!,~\{\\.+\t{fll\i:.::i:;-;--{> v

"Loudspeaker

Fig. 39-6 An acoustic transmission line, showing a sound wave traveling to the right.
The small arrows with filled heads show the directed drift velocities for small volume ele-
ments of the gar5. Compare Fig. 39-3a.

If we apply Ampere's law to the large ring of radius r', we must put the net conduc-
tion current equal to zero because the current in the outer conductor, shown by the
dots, is equal and opposite to that in the inner conductor. This means that B must
be zero for points outside the cable, in agreement with experiment. -11111

It is interesting to compare the electromagnetic oscillations in a typical traveling


wave, such as that of Fig. 39-3, to those in a cavity resonator, such as that of Fig.
38-8. The latter oscillations are an electromagnetic standing wave. In a traveling
wave E and B are in phase, which means that at a given position along the trans-
mission line they reach their maxima at the same time. However, Fig. 38-8 shows
that at a given position in a cavity resonator E and Breach their maxima one-fourth
of a cycle apart; they are 90° out of phase.
A complete analogy exists in mechanical systems. In the acoustic resonator of
Fig. 38-7 the time variations of pressure and velocity for the standing acoustic wave
are also 90° out of phase, in exact correspondence to the electromagnetic cavity
oscillations of Fig. 38-8. The acoustic analogy to a transmission line (Fig. 39-6)
would be an infinitely long gas-filled tube, one end being connected to an acoustic
oscillator such as a loud-speaker. The entire configuration of Fig. 39-6 moves to
the right with speed v. The pressure variations, suggested by the dots, and the
instantaneous velocities, suggested by the arrows, are in phase, just as are E and B
in the coaxial cable of Fig. 39-3.

39-3 Waveguide
It is possible to send electromagnetic waves through a hollow metal pipe. We
assume that the inner walls of such a pipe, or waveguide as it is called, are resistance-
less and that the cross section is rectangular.
Figure 39-7 shows a typical electric and magnetic field pattern. We imagine that
a microwave oscillator is connected to the left end and sends electromagnetic energy
down the guide. Figure 39-7a shows a side view of the guide and Fig. 39-7b, a top
view; Fig. 39-7c shows the cross section. As for the coaxial cable, the field patterns
are such that E has no tangential component for any point on the inner surface of the
guide. The fields E and B are in phase, again like the coaxial cable.
As for all traveling waves, the angular frequency w of electromagnetic waves
traveling down a guide can be varied continuously. In a waveguide of given dimen-
sions, however, there exists, for every mode of transmission, that is, for every pattern
of E and B, a so-called cutoff frequency wo. A given guide will not transmit waves in a
given mode if their frequency is below the cutoff value for that mode in that guide.
The field patterns of Fig. 39-7 show the dominant mode for a rectangular guide; this
is the mode with the lowest cutoff frequency. Given the frequency w of electromag-
netic waves to be transmitted, it is common practice to select a guide whose dimen-
sions are such that w is larger than the cutoff frequency w0 for the dominant mode
978 ELECTROMAGNETIC WAVES Chap. 39

HIH
E\
~HtH
r a 1_1_
(a)
HtmH Hlt*tH 1Itl*tH7 (c)
I I
Xg
I I
I
I
I
ph
~-t
I I I I I
(b)
~~jf l iji 1 tijijt 1 .. 11M1+
1
I
1Jij~ --i>
Vph
a
t

Fig. 39-7 A waveguide, showing (a) a side view of the lines of E, (b) a top view of the
lines of B, and (c) a cross-sectional view of the lines of E. In (c) the wave is emerging from
the page. For simplicity the lines of Bare not shown in (a) and (c), nor are the lines of E
shown in (b).

but smaller than the cutoff frequencies of all other modes. Under these conditions
the dominant mode of propagation is the only one possible.
In a (resistanceless) coaxial cable the wave patterns travel at speed c. In the
acoustic transmission line of Fig. 39-6 (assumed "resistanceless") the waves also
travel at a speed v, which is the same as the propagation speed in an infinite medium.
In a waveguide, however, the speed is not c. In waveguides we must distinguish.
between (a) the phase speed Vph, which is the speed at which the wave patterns of
Fig. 39-7 travel, and (b) the group speed Vgr, which is the speed at which electromag-
netic energy or information-carrying "signals" travel along the guide. These speeds,
which are identical for electromagnetic waves in a coaxial cable and for acoustic
waves in a tube, are different for waves in a waveguide.
The phase speed is not directly measurable. The wave pattern is a repetitive struc-
ture, and there is no way to distinguish one wave maximum from another. The waves
can be observed to enter one end of the guide and to leave at the other, but there is no
way to identify a particular wave maximum so that its passage down the guide can
be timed. We can put a "signal" on the wave by increasing the power level of the
oscillator for a short time. This power pulse could be timed as it passes through
the guide, but there is no guarantee that it travels at the same speed as the wave
pattern and, indeed, it does not. The speed of such signals or markers is the speed
at which energy is propagated, that is, the group speed.
From Maxwell's equations it can be shown that the phase speed and the group
speed for the mode of Fig. 39-7 are
c
(39-1)

and (39-2)

in which a is the width of the guide and A the free-space wavelength. Note that as
a ---+ CXJ, which corresponds to free-space conditions, Vph = Vgr = c.
The phase speed Vph is greater than the velocity of light, the group speed Vgr being
correspondingly less. In relativity theory we learn that no speed at which signals or
energy travel can be faster than that of light. However, signals or energy cannot be
transmitted down a guide at speeds exceeding c; they travel with speed Vgr which is
always less than c so there is no conflict with the theory of relativity.
Sec. 39-4 RADIATION 979
The wavelength X in Eqs. 39-1 and 39-2 is the wavelength that would be measured
for the oscillations in free space, that is,
c
A=-, (39-3)
v

where c is the speed in free space and v is the frequency. For waves of a given fre-
quency, the wavelength exhibited in a guide (Ag) must differ from the free-space
wavelength A because the speed Vph has changed. The so-called guide wavelength Ag
is given by
A = Vph = Vph = A Vph.
g v c/'t. c
From Eq. 39-1 this yields
(39-4)

Thus the guide wavelength, which is the wavelength exhibited by the field patterns
in Fig. 39-7, is larger than the free-space wavelength .

.... Example 4. What must be the width a of a rectangular guide such that the
energy of electromagnetic radiation whose free-space wavelength is 3.0 cm travels
down the guide (a) at 95% of the speed of light? (b) At 50% of the speed of light?
From Eq. 39-2 we have

Vgr c
= 0.95c = ~l - (:aY·
Solving for a yields a = 4.8 cm; repeating for Vgr = 0.50c yields a = 1.7 cm.
This formula illustrates the cutoff phenomenon described above. If X = 2a, then
Vgr = 0 and energy cannot travel down the guide. For the radiation considered in
this example A = 3.0 cm, so that the guide must have a width a of at least ix 3.0 cm =
1.5 cm if it is to transmit this wave. The guide whose width we calculated in (a)
above can transmit radiations whose free-space wavelength is 2 x 4.8 cm = 9.6 cm
or less. <1111

39-4 Radiation
The acoustic transmission line of Fig. 39-6 cannot be infinitely long. Its
far end may be sealed by a solid cap or left open, or it may have a flange,
a horn, or some similar device mounted on it. If the far end is not sealed,
energy will escape into the medium beyond. This is called acoustic radiation.
In general, some energy will also be reflected back down the transmission
line. If acoustic radiation is desirable, the designer's .task is to fashion a
termination (that is, an "acoustic antenna") for the transmission line such
that the smallest possible fraction of the incident energy will be reflected
back down the line. Such a termination might take the form of a flared
horn. Acoustic radiation, of course, requires a medium such as air in order
to be propagated.
An electromagnetic transmission line such as a coaxial cable or a wave-
guide can also be terminated in many ways, and energy can escape from the
end of the line into the space beyond. In contrast to sound waves, a physical
980 ELECTROMAGNETIC WA YES Chap. 39

Fig. 39-8 An electric dipole antenna on the end


of a coaxial cable.

medium is not required. Thus electromagnetic energy can be radiated from


the end of the transmission line, to form a traveling electromagnetic wave
in free space.
Figure 39-8 shows an effective termination for a coaxial cable; it consists
of two wires arranged as shown and is called an electric dipole antenna. The
potential difference between the two conductors alternates sinusoidally as
the wave reaches them, the effect being that of an electric dipole whose di-
pole moment p varies with time.
Figure 39-9 shows such a dipole, represented by two equal and opposite
charges. Its dipole moment, represented by the arrows marked p in the

E
+Q

l j r
V+= Q V+++
i=0
v_ f-
p
p
t= 0 t=tT
(a) (b)

V+i f-
j '(
v_f
+
- j !p
v_

V+++
p

p
t=!T
(c)

t=iT
(d)

Fig. 39-9 Showing how radiation escapes from an oscillating electric dipole. The veloci-
ties of the charges, the electric dipole moment, the equivalent current, and the lines of E
are shown at four stages of oscillation, one-eighth of a period apart. The lines of B are not
shown. (Adapted from Attwood, Electric and Magnetic Fields, John Wiley and Sons,
third edition, 1949.)
Sec. 39-4 RADIATION 981

figure, oscillates sinusoidally as the charges oscillate; in Fig. 39--9c it is, on


the verge of reversing its direction. The four views, one-eighth of a cycle
apart, show how the electric lines of force break away from the dipole and
form closed loops that travel through free space with speed c. The oscil-
lating charges constitute a current, which is represented by the arrows
marked i in Fig. 39-9. These oscillating currents generate a field of B
which, for simplicity, is not represented in the figure. The lines of B as
well as those of E also form closed loops that move away from the dipole
with speed c. These traveling electric and magnetic fields, which, as we
shall see, are strongly interdependent, constitute electromagnetic radiation.

Fig. 39-10 An instantaneous view of an electromagnetic wave as seen by an observer at


P in Fig. 39-9d.

Figure 39-10 shows a part of the wavefront as it would appear to an ob-


server at point Pin Fig. 39-9d. The wave is moving directly out of the page.
One-half a period later the observer at P will see a field pattern like that
of Fig. 39-10, except that the directions of both the electric and the magnetic
fields will be reversed. The speed c of the wave in free space is given by
c = v")\, which can be written as
w
c = -, (39-5)
k

where w, the angular frequency, and k, the wave number, are related to the
frequency v and the wavelength")<.. by

and k = 21r/A.
982 ELECTROMAGNETIC WAVES Chap. 39
39-5 Traveling Waves and Maxwell's Equations
In earlier sections we have postulated the existence of certain magnetic
and electric field distributions, in resonant cavities, coaxial cables, and wave-
guides, and we have shown that these postulated distributions are consist-
ent with Maxwell's equations, as are the distributions of conduction and
displacement currents associated with the fields. The student who pursues
his studies of electromagnetism will learn how to derive mathematical ex-
pressions for E and B by subjecting Maxwell's equations to the boundary
conditions appropriate to the problem at hand. In this section we continue
our program by showing that the postulated patterns of E and B for a
traveling electromagnetic wave are completely consistent with Maxwell',s
equations. In doing so, we will be able to show that the speed of such waves
in free space is that of visible light and thus that visible light is itself an
electromagnetic wave.
If the observer at P in Fig. 39-9d is a considerable distance from the
source, the wavefronts described by the electric and magnetic fields that
reach him (see Fig. 39-10) will be planes and the wave that moves past him
will be a plane wave (see Section 19-2). Figure 39-11 shows a "snapshot"
of a plane wave traveling in the x direction. The lines of E are parallel to
the z axis and those of B are parallel to the y axis. The values of B and E
for this wave depend only on x and t (not on y or z). We postulate that
they are given in magnitude by
B = Bm sin (kx - wt) (39-6)
and E = Em sin (kx - wt). (39-7)
Figure 39-12 shows two sections through the three-dimensional diagram of
Fig. 39-11. In Fig. 39-12a the plane of the page is the xz plane and in

Fig. 39-11 A plane electromagnetic wave traveling to the right at speed c. Lines of B
are parallel to they axis; those of E are parallel to the z axis. The shaded rectangles on
the right refer to Fig. 39-12.
Sec. 39-5 TRAVELING WAVES AND MAXWELL'S EQUATIONS 983

a
z E
E,//~+~

. .. .
) • x x
x x' • • h x"' x
x x
x x .. • x
x
x
x

(a)

{E
c......
y B (B+dB

~ ~.. .
x x
x x x x

... .
x x ~ ~ x
h ~
.. . x
x
x

x
x
x

I
x
x

x
x

x
__J l-ax
(b)

Fig. 39-12 The wave of Fig. 39-11 viewed (a) in the xz plane and (b) in the xy plane.

Fig. 39-12b it is the xy plane. Note that, as for the traveling waves in a
coaxial cable (Fig. 39-3a) and in a waveguide (Fig. 39-7), E and B are in
phase, that is, at any point through which the wave is moving they reach
their maximum values at the same time.
The shaded rectangle of dimensions dx and h in Fig. 39-12a is fixed in
space. As the wave passes over it, the magnetic flux <flB through the rectangle
will change, which will give rise to induced electric fields around the rectangle,
according to Faraday's law of induction. These induced electric fields are,
in fact, simply the electric component of the traveling wave.
Let us apply Lenz's law to this induction process. The flux <flB for the
shaded rectangle of Fig. 39-12a is decreasing with time because the wave is
moving through the rectangle to the right and a region of weaker magnetic
fields is moving into the rectangle. The induced field will act to oppose this
change, which means that if we imagine that the boundary of the rectangle
is a conducting loop a counterclockwise induced current would appear in it.
This current would produce a field of B that, within the rectangle, would
point out of the page, thus opposing the decrease in <flB. There is, of course,
no conducting loop, but the net induced electric field E does indeed act
counterclockwise around the rectangle because E + dE, the magnitude of
Eat the right edge of the rectangle, is greater than E, the magnitude of E
at the left edge. Thus the electric field configuration is entirely consistent
with the concept that it is induced by the changing magnetic field.
984 ELECTROMAGNETIC WAVES Chap. 39
For a more detailed analysis let us apply Faraday's law of induction, or

£ E·dl = - dips, (39-8)


j dt
going counterclockwise around the shaded rectangle of Fig. 39-12a. There
is no contribution to the integral from the top or bottom of the rectangle
because E and di are at right angles here. The integral then becomes

f E·dl = [(E + dE)(h)] - [(E)(h)] = dE h.

The flux ips for the rectangle is


ips = (B)(dx h),
where Bis the magnitude of Bat the rectangular strip and dx his the area
of the strip. Differentiating gives
dips dB
- = hdx-·
dt dt
From Eq. 39-8 we have then
dB
dEh = -hdx-,
dt
dE dB
or (39-9)
dx dt
Actually, both Band E are functions of x and t; see Eqs. 39-6 and 39-7.
In evaluating dE/dx, it is assumed that tis constant because Fig. 39-12a is
an "instantaneous snapshot." Also, in evaluating dB/dt it is assumed that
x is constant since what is required is the time rate of change of B at a
particular place, the strip in Fig. 39-12a. The derivatives under these cir-
cumstances are called partial derivatives, and a somewhat different notation
is used for them; see the footnote on p. 725. In this notation Eq. 39-9
becomes
aE aB
(39-10)
ax at
The minus sign in this equation is appropriate and necessary, for, although
Eis increasing with x at the site of the shaded i'ectangle in Fig. 39-12a, B
is decreasing with t. Since E(x,t) and B(x,t) are known (see Eqs. 39-6 and
39-7), Eq. 39-10 reduces to
kEm cos (kx - wt) = wBm cos (kx - wt),
w Em
or (see Eq. 39-5) -=-=c. (39-lla)
k Bm
Thus the speed of the wave c is the ratio of the amplitudes of the electric
and the magnetic components of the wave. From Eqs. 39-6 and 39-7 we
Sec. 39-5 TRAVELING WAVES AND MAXWELL'S EQUATIONS 985

see that the ratio of amplitudes is the same as the ratio of instantaneous
values, or
E = cB. (39-llb)
This important result will be useful in later sections.
We now turn our attention to Fig. 39-12b, in which the flux <I>E for the
shaded rectangle is decreasing with time as the wave moves through it.
According to Maxwell's third equation (with i = 0, because there are no
conduction currents in a traveling electromagnetic wave),

f B·dl =
d<l>E
µoEo - ,
dt
(39-12)

this changing flnx will induce a magnetic field at points around the periphery
of the rectangle. This induced magnetic field is simply the magnetic com-
ponent of the electromagnetic wave. Thus, as in the cavity resonator of
Section 38-10, the electric and the magnetic components of the wave are
intimately connected with each other, each depending on the time rate of
change of the other.
Comparison of the shaded rectangles in Fig. 39-12 shows that for each
the appropriate flux, <I>s or <I>E, is decreasing with time. However, if we pro-
ceed counterclockwise around the upper and lower shaded rectangles, we
see that f E·dl is positive, whereas f B·dl is negative. This is as it should
be. If the student will compare Figs. 35-10 and 38-9a, he will be reminded
that although the fluxes <I>s and <l>E in those figures are changing with time
in the same way (both are increasing) the lines of the induced fields, E and
B, respectively, circulate in opposite directions.
The integral in Eq. 39-12, evaluated by proceeding counterclockwise
around the shaded rectangle of Fig. 39-126, is

f B·dl = [-(B + dB)(h)] + [(B)(h)] = -h dB,

where B is the magnitude of B at the left edge of the strip and B + dB is


its magnitude at the right edge.
The flux <I>E through the rectangle of Fig. 39-12b is
<I>E = (E)(h dx).
d<I>E dE
Differentiating gives - = hdx-·
dt dt
Equation 39-12 can thus be written

-h dB = µ 0 E0 ( h dx dd~)
or, substituting partial derivatives,
aB aE
- - = µoEo-· (39-13)
ax at
986 ELECTROMAGNETIC WAVES Chap. 39
Again, the minus sign in this equation is appropriate and necessary, for,
although B is increasing with x at the site of the shaded rectangle in Fig.
39-12b, E is decreasing with t.
Combining this equation with Eqs. 39-6 and 39-7 yields
-kBm cos (kx - wt) = -µoEowEm cos (kx - wt),

k 1
or (see Eq. 39-5) (39-14)

Eliminating Em/Bm between Eqs. 39-lla and 39-14 yields


1
(39-15)

Substituting numerical values yields


1
c= ~-;=============================================
V (47r 7 12
X 10- weber/amp-m)(8.9 X 10- 2 coul /nt-m
2 )

= 3.0 X 108 meters/sec, (39-16)


which is the speed of light in free space! This emergence of the speed of
light from purely electromagnetic considerations is the crowning achieve-
ment of Maxwell's electromagnetic theory. Maxwell made this prediction
before radio waves were known and before it was realized that light was elec-
tromagnetic in nature. His prediction led to the concept of the electro-
magnetic spectrum, which we discuss in Chapter 40, and to the discovery
of radio waves by Heinrich Hertz in 1890. It made it possible to discuss
optics as a branch of electromagnetism and to derive its fundamental laws
from Maxwell's equations.
A conclusion as fundamental as Eq. 39-15 must be subject to rigorous
experimental verification. Of the three quantities in that equation, one, µ 0 ,
has an assigned value, namely, 471" X 10-7 weber/amp-m. The speed of
light c is one of the most precisely measured physical constants, having the
presently accepted value of 2.997924 X 108 meters/sec. The remaining
quantity, Eo, can be measured by making measurements on an accurately
constructed parallel-plate capacitor, as described in Section 30-2. The best
measured value, by Rosa and Dorsey of the National Bureau of Standards
(U.S.A.) in 1906, is 8.84025 X 10-12 coul 2 /nt-m 2 . To this accuracy Eq.
:39-15 is completely verified. Our confidence in electromagnetic theory,
bolstered by numerous successful predictions and agreements with experi-
ment, is now such that we reverse the foregoing procedure and calculate
our presently accepted value of Eo (see Appendix A) from the measured
speed of light, using Eq. 39-15.

39-6 The Poynting Vector


One of the important characteristics of an electromagnetic wave is that
it can transport energy from point to point. As we show below, the rate of
energy fl.ow per unit area in a plane electromagnetic wave can be described
Sec. 39-6 THE POYNTING VECTOR 987

by a vector S, called the Poynting vector after John Henry Poynting (1852-
1914), who first pointed out its properties. We define S from
1
S = -ExB. (39-17)
µo
In the mks system S is expressed in watts/meter 2 ; the direction of S gives
the direction in which the energy moves. The vectors E and B ref er to
their instantaneous values at the point in question. If Eq. 39-17 is applied
to the traveling plane electromagnetic wave of Fig. 39-11, it is clear that
E x B, hence S, point in the direction of propagation. Note, too, that S
points parallel to the axis for all points in the coaxial cable of Fig. 39-3.
We get meaningful results if we extend the Poynting vector concept to
other electromagnetic situations involving either traveling or standing elec-
tromagnetic waves, as we will see in Examples 5 and 6. If we extend it to
circuit situations involving steady or almost steady currents and lumped
circuit elements, we are led to some interesting conclusions, which we ex-
plore in Problems 11, 17, and 18 .
.... Example 5. Analyze energy flow in the cavity of Fig. 38-8, using the Poynting
vector.
Study of Fig. 39-13 shows that when the energy is all electric (Figs. 39-13a and e)
it is concentrated along the axis, because this is the region in which E has its maximum

. . . .
111:.: ·:.:.:.:.:.:

fr-.-• .------.- • • .,---~--


~ I; x xx,
llixxxxxxxxxxxx_~x
, x, x : ~ rc========
,'
I• • ( c)
I~~··
~-----2';
x-
xx x

(b) (d)

I \

(a) (e)

\ I

~
xx xxxx xxxx xx
x xx x x / x
xx ;xl ~
(h) (f)
. . ... ... . ..
................
(g)

Fig. 39-13 Example 5. Energy surges back and forth periodically between the central
region of the cavity and the region near the walls, as indicated by the Poynting vector S.
988 ELECTROMAGNETIC WA YES Chap. 39
value. When the energy is all magnetic (Figs. 39-13c and g), it is concentrated near
the walls. Thus the energy surges back and forth periodically between the central
region and the region near the walls. The figure shows, by the open arrows, the
direction of Sat various points in the cavity and at various times in the cycle. Note
that S equals zero for Figs. 39-13a, c, e, and g, which is appropriate because at these
instants of time the field configurations are momentarily stationary and energy
is not flowing. A pendulum bob at the end of its swing and at the bottom of its
trajectory forms a mechanical analogy. The student should verify from Eq. 39-17
that these arrows point in the correct directions. ....

Figure 39-14 can be used to derive the Poynting relation for the special case of a
traveling plane electromagnetic wave. It shows a cross section of a traveling plane
wave, along with a thin "box" of thickness dx and area A. The box, a mathematical
construction, is fixed with respect to the axes while the wave moves through it.

.. •

• •

x

'x I
r
x
x

X
x

• •
• •
• •
I x

x x
x

x
x
x
7
.
• • x
x
x • •

x x
x
x
-J k---dx

Fig. 39-14 A plane wave is traveling to the right at speed c; compare Fig. 39-12b. The
dashed rectangle in this figure represents a three-dimensional box, of area A and thickness
dx, that extends at right angles to the plane of the figure.

At any instant the energy stored in the box, from Eqs. 30-27 and 36-19, is

dU = dUE + dUB = (uE + UB)(A dx)

= ( !eoE2 + ~Lo B 2 ) A dx, (39-18)

where A dx is the volume of the box and E and B are the instantaneous values of the
field vectors in the box.
Using Eq. 39-llb (E = cB) to eliminate one of the E's in the first term in Eq. 39-18
and one of the B's in the second term leads to

dU = [ !eoE(cB) + 2~ B 0
(~) J A dx
(µoeoc 2 + l)(EBA dx)
2µoc

From Eq. 39-15, however, µ 0e0c2 = 1, so that

dU = EBA dx
µoc
Sec. 39-6 THE POYNTING VECTOR 989

This energy dU will pass through the right face of the box in a time dt equal to
dx/c. Thus the energy per unit area per unit time, which is S, is given by

S = dU = EBA dx = J_ EB
dt A µoc (dx/c) A µo ·
This is exactly the prediction of the more general relation Eq. 39-17 for a traveling
plane wave.
This relation refers to values of S, E, and Bat any instant of time. We are usually
more interested in the average value of S, taken over one or more cycles of the wave.
An observer making intensity measurements on a wave moving past him would
measure this average value S. We can easily show (see Example 6) that Sis related
to the maximum values of E and B by

.... Example 6. An observer is at a great distance r from a point light source whose
. power output is P 0 • Calculate the magnitudes of the electric and the magnetic fields.
Assume that the source is monochromatic, that it radiates uniformly in all directions,
and that at distant points it behaves like the traveling plane wave of Fig. 39-11.
The power that passes through a sphere of radius r is (S) (47rr2 ), where S is the average
value of the Poynting vector at the surface of the sphere. This power must equal Po,
or

From the definition of S (Eq. 39-17), we have

Using the relation E = cB (Eq. 39-llb) to eliminate B leads to

S = _I E2.
µoc
The average value of E 2 over one cycle is !Em 2 , since E varies sinusoidally (see Eq.
39-7). This leads to

or

For Po = 103 watts and r = 1.0 meter this yields

Em = _1_ /(10 3 watts)(47r X 10-7 weber/amp-m)(3 X 108 meters/sec)


(1.0m) \j 271"

= 240 volts/meter.
The relationship Em = cBm (Eq. 39-lla) leads to

_ Em _ 240 volts/meter _ _
Bm -
c
- 3 X 108 t I
me ers sec
- 8 X 10 7 weber/meter.2
990 ELECTROMAGNETIC WA YES Chap. 39
Note that Em is appreciable as judged by ordinary laboratory standards but that Bm
(= 0.008 gauss) is quite small. .....

QUESTIONS

1. In the coaxial cable of Fig. 39-1, what are the directions of the conduction current
(a) in the central conductor and (b) in the outer conductor, shortly after the switch is
thrown to position a? Consider points that have been reached by the wavefront of Fig.
39-2a and b and those that have not.
2. Compare a coaxial cable and a waveguide, used as a transmission line. Point out
both similarities and differences.
3. What is the relation between the wavelength in the cable and that in free space for a
coaxial cable?
4. Can traveling waves with a continuous range of wavelengths be sent down (a) a
coaxial cable and (b) a waveguide? Can standing waves with a continuous range of wave-
lengths be set up in a resonant cavity? Develop mechanical or acoustical analogies to
support your answers.
5. If a certain wavelength is larger than the cutoff wavelength for a guide in its dominant
mode, can energy be sent down it in any other mode?
6. Explain why the term fO dif>Eldt is needed in Ampere's equation to understand the
propagation of electromagnetic ,vaves.
7. In the equation c = 1 / ~ (Eq. 39-15), how can c always have the same value if
/.lo is arbitrarily assigned and fO is measured?
8. Is it conceivable that electromagnetic theory might some day be able to predict the
value of c (3 X 108 meters/sec), not in terms of µ 0 and fO, but directly and numerically
without recourse to any measurements?
9. What is the direction of the displacement current in Fig. 39--4? Give an argument
to support your answer.
10. In a coaxial cable is the energy transported in the conductors, through the agency
of the currents, or in the space between them, through the agency of the fields?

PROBLEMS

1. Using Gauss's law, sketch the instantaneous charges that appear on the conductors
of the coaxial cable of Fig. 39-3 and show that this pattern of charges is appropriately
related to the conduction currents shown in Fig. 39-3c.
2. For a rectangular guide of width 3.0 cm, plot the phase speed, the group speed, and
the guide wavelength as a function of the free-space wavelength. Assume the dominant
mode.
3. For a rectangular guirl.e of width 3.0 cm, what must the free-space wavelength of
radiation be if it is to require 1.0 µsec ( = 10-6 sec) for energy to traverse a 100-meter
length of guide? What is the phase speed under these circumstances?
4. Under what conditions will the guide wavelength in the guide of Fig. 39-7 be double
the free-space wavelength?
··. 5. How does the displacement current vary with space and time in a traveling plane
electromagnetic wave?
6. Prove that for any point in an electromagnetic wave such as that of Fig. 39-11 the
density of energy stored in the electric field equals that stored in the magnetic field.
7. A resonant cavity is constructed by closing each end of the coaxial cable of Fig. 39-3
with a metal cap. The cavity contains three half-waves. Describe the patterns of E and B
Chap. 39 PROBLEMS 991

that occur, assuming the same mode of oscillation as that shown in Fig. 39-3. (Hint:
Remember that E can have no tangential component at a conducting surface and that
Band E must be 90° out of phase.)
8. If a coaxial cable has resistance, energy must flow from the fields into the conducting
surfaces to provide the Joule heating. How must the electric lines of force of Fig. 39-3a
be modified in this case? (Hint: The Poynting vector near the surface must have a com-
ponent pointing toward the surface.)
9. What guide wavelength does 10-cm radiation (free-space wavelength) exhibit in a
rectangular guide whose width is 6.0 cm? Assume the dominant mode. What is the cutoff
wavelength for this guide?
10. Sketch five more figures to complete the sequence of Fig. 39-9, showing radiation
from an oscillating dipole. Include an indication of the lines of B in your drawings.
11. Figure 39-15 shows a long resistanceless transmission line, delivering power from a
battery to a resistive load. A steady current i exists as shown. (a) Sketch qualitatively
the electric and magnetic fields around the line, and (b) show that, according to the
Poynting vector point of view, energy travels from the battery to the resistor through the
space around the line and not through the line itself. (Hint: Each conductor in the line
is an equipotential surface, since the line has been assumed to have no resistance.)

f,1+
),

R~
--E
i
Fig. 39-15

12. Analyze the flow of energy in the waveguide of Fig. 39-7, using the Poynting
vector.
13. A cube of edge a has its edges parallel to the x-, y-, and z-axes of a rectangular
coordinate system. A uniform electric field E is parallel to the y-axis and a uniform
magnetic field B is parallel to the x-axis. Calculate (a) the rate at which, according to
the Poynting vector point of view, energy may be said to pass through each face of the
cube and (b) the net rate at which the energy stored in the cube may be said to change.
~4.: A #10 copper wire (diameter, 0.10 in.; resistance per 1000 ft, 1.00 ohm) carries a
current of 25 amp. Calculate E, B, and S for a point on the surface of the wire.
15. Sunlight strikes the earth, outside its atmosphere, with an intensity of 2.0 cal/cm 2-
min. Calculate Em and Bm for sunlight, assuming it to be a wave like that of Fig. 39-10.
16. A plane radio wave has Em "' 10-4 volt/meter. Calculate (a) Bm and (b) the in-
tensity of the wave, as measured by S.
· '·17- Figure 39-16 shows a cylindrical resistor of length l, radius a, and resistivity p,
carrying a current i. (a) Show that the Poynting vector Sat the surface of the resistor is
everywhere directed normal to the surface, as shown. (b) Show that the rate Pat which
energy flows into the resistor through its cylindrical surface, calculated by integrating the
Poynting vector over this surface, is equal to the rate at which Joule heat is produced;
that is,
-J S·dA = i 2R,

where dA is an element of area of the cylindrical surface. This shows that, according to
the Poynting vector point of view, the energy that appears in a resistor as Joule heat does
not enter it through the connecting wires but through the space around the wires and the
resistor. Note that S and dA are oppositely directed. (Hint: E is parallel to the axis
of the cylinder, in the direction of the current; B forms concentric circles around the
cylinder, in a direction given by the right-hand rule.)
992 ELECTROMAGNETI C WAVES Chap. 39

s s d

Fig. 39-16 Fig. 39-17


,,
18. Figure 39-17 shows a parallel-plate capacitor being charged. (a) Show that the
Poynting vector S points everywhere radially into the cylindrical volume. (b) Show that
the rate Pat which energy flows into this volume, calculated by integrating the Poynting
vector over the cylindrical boundary of this volume, is equal to the rate at which the
stored electrostatic energy increases; that is, that

-f S·dA =ad~ (hoE2),

where ad is the volume of the capacitor and !~ 2 is the energy density for all points within

that volume. This analysis shows that, according to the Poynting vector point of view,
the energy stored in a capacitor does not enter it through the wires but through the space
around the wires and the plates. (Hint: To find S we must first find B, which is the
magnetic field set up by the displacement current during the charging process; see Fig.
38-9. Ignore fringing of the lines of E.)
Nature and Propagation of Light
CHAPTER 40

40-1 Light and the Electromagnetic Spectrum


Light was shown by Maxwell to be a component of the electromagnetic
spectrum of Fig. 40-1. All these waves are electromagnetic in nature and
have the same speed c in free space. They differ in wavelength (and thus
in frequency) only, which means that the sources that give rise to them and
the instruments used to make measurements with them are rather different.*
The electromagnetic spectrum has no definite upper or lower limit. The
labeled regions in Fig. 40-1 represent frequency intervals within which a
common body of experimental technique, such as common sources and com-
mon detectors, exists. All such regions overlap. For example, we can pro-
duce radiation of wavelength 10-3 meter either by microwave techniques
(microwave oscillators) or by infrared techniques (incandescent sources).

Frequency, cycles/sec
102 104 106 108 1010 1012 1014 1016 1018 1020 1022

Power Microwaves Visible X-rays


------
Radio Infrared
-- -----
Ultraviolet Gamma rays

106 104 102 1 10-2 10-4 10-6 10-8 10-10 10-12 10-14
Wavelength, meters

Fig. 40-1 The electromagnetic spectrum. Note that the wavelength and frequency
scales are logarithmic.
* For a report of electromagnetic waves with wavelengths as long as 1.9 X 107 miles
the student should consult an article by James Heirtzler in the Scientific American for
March 1962.
993
994 NATURE AND PROPAGATION OF LIGHT Chap. 40

violet blue green yellow orange red


100

80

->,

>
:;:::
'vi 60
c:
Q)
(J)

Q)
>
:;::: 40
"'
al
0:::

20

0
400 450 500 550 600 650 700
Wavelength, mµ

Fig. 40-2 The relative eye sensitivity of an assumed standard observer at different wave-
lengths for normal levels of illumination. The shaded areas represent the (continuously
graded) color sensations for normal vision.

"Light" is defined here as radiation that can affect the eye. Figure 40-2,
which shows the relative eye sensitivity of an assumed standard observer to
radiations of various wavelengths, shows that the center of the visible region
is about 5.55 X 10-7 meter. Light of this wavelength produces the sensa-
tion of yellow-green.*
In optics we often use the micron (abbr. µ) the millimicron (abbr. mµ),
and the Angstrom (abbr. A) as units of wavelength. They are defined from
1 µ = 10-5 meter
1 mµ = 10-9 meter
1 A = 10-10 meter.
Thus the center of the visible region can be expressed as 0.555 µ, 555 mµ, or
5550A.
The limits of the visible spectrum are not well defined because the eye
sensitivity curve approaches the axis asymptotically at both long and short
wavelengths. If the limits are taken, arbitrarily, as the wavelengths at
which the eye sensitivity has dropped to 1% of its maximum value, these
limits are about 4300 A and 6900 A, less than a factor of two in wavelength.
The eye can detect radiation beyond these limits if it is intense enough. In
many experiments in physics one can use photographic plates or light-sensi-
tive electronic detectors in place of the human eye.
* See "Experiments in Color Vision" by Edwin H. Land, Scientific American, May
1959, and especially "Color and Perception: the Work of Edwin Land in the Light of
Current Concepts" by M. H. Wilson and R. W. Brocklebank, Contemporary Physics,
December 1961, for a fascinating discussion of the problems of perception and the distinc-
tion between color as a characteristic of light and color as a perceived property of objects.
Sec. 40-2 ENERGY AND MOMENTUM 995

40-2 Energy and Momentum


Energy is carried by electromagnetic waves from the sun to the earth or
from an open fire to a hand placed nearby. The transport of energy by such
a wave in free space was described in Section 39-6 by the Poynting vector S,
or
1
S = -ExB, (40-1)
µo

where E and Bare the instantaneous values of the electric and magnetic field
vectors.
Less familiar is the fact that electromagnetic waves may also transport
linear momentum. In other words, it is possible to exert a pressure (a
radiation pressure *) on an object by shining a light on it. Such forces must
be small in relation to forces of our daily experience because we do not
ordinarily notice them. The first measurement of radiation pressure was
made in 1901-1903 by Nichols and Hull in this country and by Lebedev in
Russia, about thirty years after the existence of such effects had been pre-
dicted theoretically by Maxwell.
Let a parallel beam of light fall on an object for a time t, the incident light
being entirely absorbed by the object. If energy U is absorbed during this
time, the momentum p delivered to the object is given, according to Max-
well's prediction, by
p
u
= - (total absorption), (40-2a)
c
where c is the speed of light. The direction of p is the direction of the in-
cident beam. If the light energy U is entirely reflected, the momentum de-
livered will be twice that given above, or
2U
p=- (total reflection). (40-2b)
c
In the same way, twice as much momentum is delivered to an object when
a perfectly elastic tennis baJl is bounced from it as when it is struck by a
perfectly inelastic ball of the same mass and speed. If the light energy U
is partly reflected and partly absorbed, the delivered momentum will lie
between U /c and 2U /c.
~ Example 1. A parallel beam of light with an energy flux S of 10 watts/cm 2 falls
for 1 hr on a perfectly reflecting plane mirror of 1.0-cm2 area. (a) What momentum
is delivered to the mirror in this time and (b) what force acts on the mirror?
(a) The energy that is reflected from the mirror is
C = (10 watts/cm 2)(1.0 cm 2)(3600 sec) = 3.6 X 104 joules.
The momentum delivered after 1 hr's illumination is
_ 2U _ (2)(3.6 X 104 joules) _ 2 4
P - -- -
_ k I
- . X 10 4 g-m sec.
c 3 X 10 8 meters/sec
* See "Radiation Pressure," G. E. Henry, Scientific American, p. 99, June 1957.
996 NATURE AND PROPAGATION OF LIGHT Chap. 40
(b) From Newton's second law, the average force on the mirror is equal to the
average rate at which momentum is delivered to the mirror, or

F = -p = 2.4 X 10- kg-m/sec = 6 .,~ X 10 _ 8 nt.


4

t 3600 sec
This is a small force.

::-,.Jichols and Hull, in 190:3, measured radiation pressures and verified


Eq. 40-2, using a torsion balance technique. They allowed light to fall on
mirror M in Fig. 40-3; the radiation pressure caused the balance arm to
turn through a measured angle 8, twisting the torsion fiber F. Assuming a
suitable calibration for their torsion fiber, the experimenters could arrive at
a numerical value for this pressure. Nichols and Hull measured the in-
tensity of their light beam by allowing it to fall on a blackened metal disk
of known absorptivity and by measuring the temperature rise of this disk.
In a particular run these experimenters measured a radiation pressure of
7.01 X 10- 5 nt/meter 2 ; for their light beam, the value predicted, using Eq.
40-2, was 7.05 X 10-5 nt/meter 2 , in excellent agreement. Assuming a
mirror area of 1 cm 2 , this represents a force on the mirror of only 7 X 10-10
nt, about 100 times smaller than the force calculated in Example 1.
The success of the experiment of Nichols and Hull was the result in large
part of the care they took to eliminate spurious deflecting effects caused by
changes in the speed distribution of the molecules in the gas surrounding the
mirror. These changes were brought about by the small rise in the tempera-
ture of the mirror as it absorbed light energy from the incident beam. This
"radiometer effect" is responsible for the spinning action of the familiar toy
radiometers when placed in a beam of sunlight. In a perfect vacuum such
effects would not occur, but in the best vacuums available in 1903 radiometer
effects were present and had to be taken specifically into account in the de-
sign of the experiment.
To demonstrate the transport of momentum from :VIaxwell's equations in a par-
ticular case, let a plane electromagnetic wave traveling in the z direction fall on a

Torsion
suspension

Fig. 40-3

Mirrors

Incident
light
beam ef/
M
M'
Sec. 40-2 ENERGY AND MOMENTUM 997
y

Fig. 40-4 An incident plane light


wave falls on an electron in a thin
resistive sheet. Instantaneous Incident wave
values of E, B, the electron veloc-
ity v, and the radiation force F z are
shown.

large thin sheet of a material of high resistivity as in Fig. 40-4. A small part of the
incident energy will be absorbed within the sheet, but most of it will be transmitted
if the sheet is thin enough.*
The incident wave E and B vary with time at the sheet as
E = Em sin wt (4v-3)
and B = Bm sin wt (40-4)
where E is parallel to the ±y axis and B is parallel to the ±x axis.
In Section 31-4 we saw that the effect of a (constant) electric force ( = -eE) on a
conduction electron in a metal was to make it move with a (constant) drift speed vd.
The electron behaves as if it is immersed in a viscous fluid, the electric force acting on
it being counterbalanced by a "viscous" force, which may be taken as proportional
to the electron speed. Thus for a constant field E, after equilibrium is established,
eE = bvd, (40-5)
where bis a resistive damping coefficient. The electron equilibrium speed, dropping
the subscript d, is thus
eE
v = -· (40-6)
b
If the applied electric field varies with time and if the variation is slow enough,
the electron speed can continually readjust itself to the changing value of E so that its
speed continues to be given essentially by its equilibrium value (Eq. 40-6) at all
times. These readjustments are more rapidly made the more viscous the medium,
just as a stone falling in air reaches a constant equilibrium rate of descent only rela-
tively slowly but one falling in a viscous oil does so quite rapidly. We assume that the
sheet in Fig. 40-4 is so viscous, that is, that its resistivity is so high, that Eq. 40-6
remains valid even for the rapid oscillations of E in the incident light beam.
As the electron vibrates parallel to the y axis, it experiences a second force due to
the magnetic component of the wave. This force F z ( = -ev x B) points in the z
direction, being at right angles to the plane formed by v and B, that is, the xy plane.
The instantaneous magnitude of F 2 is given by
e2EB
F-• = evB = - b- · (40-7)

* Some of the incident energy will also be reflected, but the reflected wave is of such low
intensity that it can be ignored in the derivation that follows; see Optics by B. Rossi,
Addison-Wesley Publishing Company, p. 411, 1957, from which this derivation is adapted.
998 NATURE AND PROPAGATION OF LIGHT Chap .. 40
Fz always points in the positive z direction because v and B reverse their directions
simultaneously; this force is, in fact, the mechanism by which the radiation pressure
acts on the sheet of Fig. 40-4.
From Newton's second law, Fz is the rate dpe/dt at which the incident wave de-
livers momentum to each electron in the sheet, or
dpe = e2EB. (40-8)
dt b
Momentum is delivered at this rate to every electron in the sheet and thus to the
sheet itself. It remains to relate the momentum transfer to the sheet to the absorption
of energy within the sheet.
The electric field component of the incident wave does work on each oscillating
electron at an instantaneous rate (see Eq. 40-6) given by

dUe = FEv = (eE) (eE) = e2E2.


& b b
Note that the magnetic force F always being at right angles to the velocity v, does
2,

no work on the oscillating electron. Equation 39- llb shows that for a plane wave in
free space B and E are related by
E = Be.
Substituting above for one of the E's leads to

dUe e2 EBc
- - = ---· (40-9)
dt b

This equation represents the rate, per electron, at which energy is absorbed from the
incident wave.
Comparing Eqs. 40-8 and 40-9 shows that

dpe = ~ dU_!_.
dt c dt

dt = 2 it?:_!/_!_ dt
Integrating yields i t dpe
O dt C dt
O '
Ue
or Pe=--, (40-10)
c
where Pe is the momentum delivered to a single electron in any given time t and Ue
is the energy absorbed by that electron in the same time interval. Multiplying each
side by the number of free electrons in the sheet leads to Eq. 40-2a.
Although we derived this relation (Eq. 40-10) for a particular kind of absorber,
no characteristics of the absorber-for example, the resistive damping coefficient
b-remain in the final expression. This is as it should be because Eq. 40-10 is a
general property of radiation absorbed by any material.

40-3 The Speed of Light *


Light travels so fast that there is nothing in our daily experience to suggest
that its speed is not infinite. It calls for considerable insight even to ask
"How fast does light travel?" Galileo asked himself this question and tried
to answer it experimentally. His chief work, Two New Sciences, published

* See "The Speed of Light," J. H. Rush, Scientific American, p. 67, August 1955.
Sec. 40-3 THE SPEED OF LIGHT 999

in the Nether lands in 16:38, is written in the form of a conversation among


three fictitious persons called Salviati, Sagredo, and Simplicio. Here is part
of what they say about the speed of light.
Simplicio: Everyday experience shows that the propagation of light is instan-
taneous; for when we see a piece of artillery fired, at a great distance, the flash
reaches our eyes without lapse of time; but the sound reaches the ear only after a
noticeable interval.
Sagredo: Well, Simplicio, the only thing I am able to infer from this familiar bit
of experience is that sound, in reaching our ear, travels more slowly than light; it
does not inform me whether the coming of the light is instantaneous or whether,
although extremely rapid, it still occupies time . . . .
Sagredo, who evidently is Galileo himself, then describes a possible method
for measuring the speed of light. He and an assistant stand facing each other
some distance apart, at night. Each carries a lantern which can be covered
or uncovered at will. Galileo started the experiment by uncovering his lan-
tern. When the light reached the assistant he uncovered his own lantern,
whose light was then seen by Galileo. Galileo tried to measure the time
between the instant at which he uncovered his own lantern and the instant
at which the light from his assistant's lantern reached him. For a 1-mile
separation we now know that the round trip travel time would be only 11 X
10-5 sec. This is much less than human reaction times, so the method fails.
To measure a large velocity directly, we must either measure a small time interval
or use a long base line. This situation suggests that astronomy, which deals with great
distances, might be able to provide an experimental value for the speed of light;
this proved to be true. Although it would be desirable to time the light from the sun
as it travels to the earth, there is no way of knowing when the light that reaches us at
any instant left the sun; we must use subtler astronomical methods.
Note, however, that microwave pulses are quite regularly reflected from the moon;
this gives a 7.68 X 10 8-meter base line (there and back) for timing purposes. The
speed of light (and of microwaves) is so well known now from other experiments that
these measurements are used to measure the lunar distance accurately. Microwave
signals have also been reflected from Venus.
In 1675 Ole Roemer, a Danish astronomer working in Paris, made some observa-
tions of the moons of Jupiter (see Problem 9) from which a speed of light of 2 X 108
meters/sec may be deduced. About fifty years later James Bradley, an English
astronomer, made some astronomical observations of an entirely different kind from
which a value of 3.0 X 10 8 meters/sec may be deduced.
In 1849 Hippolyte Louis Fizeau (1819-1896), a French physicist, first
measured the speed of light by a nonastronomical method, obtaining a value
of 3.13 X 108 meters/sec. Figure 40--5 shows Fizeau's apparatus. Let us
first ignore the toothed wheel. Light from source Sis made to converge by
lens L 1 , is reflected from mirror M 1 , and forms in space at F an image of the
source. Mirror 211 1 is a so-called "half-silvered mirror"; its reflecting coating
is so thin that only half the light that falls on it is reflected, the other half
being transmitted.
Light from the image at F enters lens L 2 and emerges as a parallel beam;
after passing through lens L 3 it is reflected back along its original direction by
mirror 1112 . In Fizeau's experiment the distance l between Af 2 and F was
1000 NATURE AND PROPAGATION OF LIGHT Chap. 40

Toothed
wheel

M;no:f<:f= =========Lt2:::>t~;::Lt4==:
-1.- - - - - z - ~

s
Fig. 40-5 Fizeau's apparatus for measuring the speed of light.

8630 meters or .5.36 miles. When the light strikes mirror Af 1 again, some will
be transmitted, entering the eye of the observer through lens L 4 .
The observer will see an image of the source formed by light that has
traveled a distance 2Z between the wheel and mirror M 2 and back again. To
time the light beam a marker of some sort must be put on it. This is done by
"chopping" it with a rapidly rotating toothed wheel. Suppose that during
the round-trip travel time of 2Z/c the wheel has turned just enough so that,
when the light from a given "burst" returns to the wheel, point Fis covered
by a tooth. The light will hit the face of the tooth that is toward M 2 and
will not reach the observer's eye.
If the speed of the wheel is exactly right, the observer will not see any of
the bursts because each will be screened by a tooth. The observer measures
c by increasing the angular speed w of the wheel from zero until the image of
source S disappears. Let e be the angular distance from the center of a gap
to the center of a tooth. The time -needed for the wheel to rotate a distance
e is the round-trip travel time 2Z/c. In equation form,
e 2wl
or c = --· (40-11)
w c e
This "chopped beam" technique, suitably modified, is used today to measure
the speeds of neutrons and other particles .
..... Example 2. The wheel used by Fizeau had 720 teeth. What is the smallest angu-
lar speed at which the image of the source will vanish?
The angle eis 1/1440 rev; solving Eq. 40-11 for w gives
ce
w = - =
(3.00 X 10 8 meters/sec)(l/1440 rev)
12.1 rev/sec.
...,.
""'Ill
2l (2)(8630 meters)
The French physicist Foucault (1819-1868) greatly improved Fizeau's
method by substituting a rotating mirror for the toothed wheel. The Amer-
ican physicist Albert A. Michelson (1852-1931) conducted an extensive
series of measurements of c, extending over a fifty-year period, using this
technique.
Sec. 40-3 THE SPEED OF LIGHT 1001

Table 40-1

THE SPEED OF ELECTROMAGNETIC RADIATION IN FREE SPACE


(Some selected measurements)

Uncer-
Speed,
Date Experimenter Country Method tainty,
km/sec
km/sec

1600(?) Galileo Italy Lanterns and "If not instantane-


shutters ous, it is extraor-
dinarily rapid"
1675 Roemer France Astronomical 200,000
1729 Bradley England Astronomical 304,000
1849 Fizeau France Toothed wheel 313,300
1862 Foucault France Rotating mirror 298,000 500
1876 Cornu France Toothed wheel 299,990 200
1880 Michelson U.S.A. Rotating mirror 299,910 50
1883 Newcomb England Rotating mirror 299,860 30
1883 Michelson U.S.A. Rotating mirror 299,853 60
1906 Rosa and Dorsey U.S.A. Electromagnetic 299, 781 10
theory
1923 Mercier France Standing waves on 299,782 15
wires
1926 Michelson U.S.A. Rotating mirror 299,796 4
1928 Karolus and Mit- Germany Kerr cell 299,778 10
telstaedt
1932 Michelson, Pease, U.S.A. Rotating mirror 299,774 11
and Pearson
1940 Huettel Germany Kerr cell 299,768 10
1941 Anderson U.S.A. Kerr cell 299,776 14
1950 Bergstrand Sweden Geodimeter 299,792.7 0.25
1950 Essen England Microwave cavity 299,792.5 3
1950 Houston Scotland Vibrating crystal 299,775 9
1950 Bol and Hansen U.S.A. Microwave cavity 299,789.3 0.4

1951 Aslakson U.S.A. Shoran radar 299,794.2 1.9


1952 Rank, Ruth, and U.S.A. Molecular spectra 299,776 7
Ven der Sluis
1952 Froome England Microwave inter- 299,792.6 0.7
ferometer
1954 Florman U.S.A. Radio interferom- 299,795.l 3.1
eter
1954 Rank, Shearer, U.S.A. Molecular spectra 299,789.8 3.0
and Wiggins
1956 Edge Sweden I Geodimeter I 299,792.9 0.2
1002 NATURE AND PROPAGATION OF LIGHT Chap. 40
We must view the speed of light within the larger framework of the speed
of electromagnetic radiation in general. It is a significant experimental
confirmation of Maxwell's theory of electromagnetism that the speed in
free space of waves in all parts of the electromagnetic spectrum has the
same value c. Table 40-1 shows some selected measurements that have
been made of the speed of electromagnetic radiation since Galileo's day. It
stands as a monument to man's persistence and ingenuity. Note in the last
column how the uncertainty in the measurement has improved through the
years. Note also the international character of the effort and the variety
of methods.
The task of arriving at a single "best" value of c from the many listed in
the table is difficult, for it involves a careful study of each reported measure-
ment and a selection from among them, based on the reported uncertainty
and the selector's judgment of the probable presence or absence of hidden
error. In the final averaging measurements with small uncertainties will be
given more weight than those with large ones. By careful an3'.lysis of such
measurements the "best" value, as of 1964, is
c = 2.997925 X 10 8 meters/sec.
The uncertainty of measurement is less than 0.000003 X 10 8 meters/sec
or 0.0001 %.
Since about 1940, nearly all precise measurements of c have been made in the
microwave or the short radio wave region of the electromagnetic spectrum. We
describe here the "microwave cavity method" used by Essen in England and by Bol
and Hansen in the U.S.A. It employs standing electromagnetic waves confined to a
cavity rather than traveling waves in free space.
It is possible to convert a section of waveguide such as that of Fig. 39-7 into a
resonant cavity by closing it with two metal caps; see Fig. 40-6. The pattern of
oscillations in the cavity is closely related to that in the guide and exhibits the same
"guide wavelength" Ag, The guide wavelength is related to the cavity length l by
A _ 2l
g - n n = 1, 2, 3, · · ·, (40-12)

which is the same relationship used for acoustic waves in closed pipes; n ( = 3 for Fig.
40-6) gives the number of half-waves contained in the cavity.

I I

~I . ~
l~-1
I
I EI
I
;

~
I
j~
(a)
Fig. 40-6 A resonant cavity con-
structed from a section of waveguide;
compare Fig. 39-7. For simplicity
the lines of E are not shown in (a) and
those of B are not shown in (b).
I f mn f ! t nn6 ! I fIDf f 1 (b)
I I
I I
Sec. 40-4 MOVING SOURCES AND OBSERVERS 1003
The procedure is to measure Ag for such a cavity, which has been tuned to resonance,
and then, using Eq. 39-4,
Ag= A ,
~l - (tJ2 (40-13)

calculate the free-space wavelength A. From the measured resonant frequency, the
speed c can be found from c = Av.
.... Example 3. Essen of the National Physical Laboratory in England made a
resonant cavity measurement of the speed of electromagnetic waves. His cavity was
made of a circular waveguide rather than a rectangular one; it can be shown that, for
the oscillation pattern used by him, the geometrical factor 2a in Eq. 40-13 must be
replaced by l.64062R, where R is the guide radius. The cavity radius was 3.25876
cm; the cavity length was 15.64574 cm and it proved to resonate at 9.498300 X 109
cycles/sec. At resonance it was determined that there were eight half-waves in the
cavity. What value of c results?
From Eq. 40-12, computing only an approximate result,
, _ 2l _ (2)(15.6 cm) _ 3
Ag - n - 8 - .90 cm.

Substituting into Eq. 40-13, suitably modified for a circular waveguide, yields

3.90 cm= / .

'\J 1 - ((1.64)(~.26 cm)Y


Solving this equation for A yields A = 3.15 cm. Finally, we have
c = Av = (3.15 cm)(9.50 X 109 cycles/sec) = 2.99 X 108 meters/sec.

For practical reasons Essen analyzed his data in a more roundabout way than that
given. His final result, based on many measurements under different conditions and
carried to much greater accuracy than that illustrated in the above example, was
2.997925 X 108 meters/sec with an uncertainty of 0.000030 X 108 meters/sec. <111111

4o-4 Moving Sources and Observers


When we say that the speed of sound in dry air at 0°C is 331.7 meters/sec,
we imply a reference frame fixed with respect to the air mass. When we say
that the speed of light in free space is 2.997925 X 108 meters/sec, what
reference frame is implied? It cannot be the medium through which the
light wave travels because, in contrast to sound, no medium is required.
The concept of a wave requiring no medium was abhorrent to the physi-
cists of the nineteenth century, influenced as they then were by a false
analogy between light waves and sound waves or other purely mechanical
disturbances. These physicists postulated the existence of an ether, which
was a tenuous substance that filled all space and served as a medium of
transmission for light. The ether was required to have a vanishingly small
density to account for the fact that it could not be observed by any known
means in an evacuated space.
The ether concept, although it proved useful for many years, did not
survive the test of experiment. In particular, careful attempts to measure
1004 NATURE AND PROPAGATION OF LIGHT Chap. 40

Fig. 40-7 Observers S and S', who


are in relative motion, each observe
a light pulse P. The pulse is emitted
from a source, not shown, that is at
rest in the S' frame of reference.

the speed of the earth through the ether always gave the result of zero.*
Physicists were not willing to believe that the earth was permanently at
rest in the ether and that all other bodies in the universe were in motion
through it. Other hypotheses about the nature of the propagation of light
also proved unsatisfactory for one reason or another.
Einstein in 1905 resolved the difficulty of understanding the propagation
of light by making a bold postulate: If a number of observers are moving
(at uniform velocity) with respect to each other and to a source of light and
if each observer measures the speed of the light emerging from the source,
they will all obtain the same value. This is the fundamental assumption of
Einstein's theory of relativity. It does away with the need for an ether by
asserting that the speed of light is the same in all reference frames; none is
singled out as fundamental. The theory of relativity, derived from this
postulate, has been subject to many experimental tests, from which agree-
ment with the predictions of theory has always emerged. These agreements,
extending over half a century, lend strong support to Einstein's basic
postulate about light propagation.
Figure 40-7 focuses specifically on the fundamental problem of light
propagation. A source of light, at rest in reference frame S', emits a light
pulse P whose speed v' is measured by an observer at rest in this same frame.
From the point of view of an observer in reference frame S, frame S' and its
associated observer are moving in the positive x direction at speed u. Ques-
tion: What speed v would observer S measure for the light pulse P? Ein-
stein's hypothesis asserts that each observer would measure the same speed
c, or that
v = v' = c.

This hypothesis contradicts the classical law of addition of velocities (see


Section 4-6), which asserts that

v = v' + u. (40-14)

This law, which is so familiar that it seems (incorrectly) to be intuitively


true, is in fact based on observations of gross moving objects in the world
about us. Even the fastest of these-an earth- satellite, say-is moving at

* See Section 43-7, which describes the crucial experiment of Michelson and Morley.
Sec. 40-4 MOVING SOURCES AND OBSERVERS 1005

a speed that is quite small compared to that of light. The body of experi-
mental evidence that underlies Eq. 40-14 thus represents a severely re-
stricted area of experience, namely, experiences in which v' << c and u « c.
If we assume that Eq. 40-14 holds for all particles regardless of speed, we
are making a gross extrapolation. Einstein's theory of relativity predicts
that this extrapolation is indeeJ not valid and that Eq. 40-14 is a limiting
case of a more general relationship that holds for light pulses and for ma-
terial particles, whatever their speed, or

v =
v' +u . (40-15)
1 + v'u/c2
Equation 40-15 is quite indistinguishable from Eq. 40-14 at low speeds,
that is, when v' « c and u « c; see Example 4.
If we apply Eq. 40-15 to the case in which the moving object is a light
pulse, and if we put v' = c, we obtain

c+u
v = = c.
1 + cu/c2
This is consistent, as it must be, with the fundamental assumption on which
the derivation of Eq. 40-15 is based; it shows that both observers measure
the same speed c for light. Equation 40-14 predicts (incorrectly) that the
speed measured by S will be c + u. Figure 40-8 shows that the (correct)
Eq. 40-15 and the (approximate) Eq. 40-14 cannot be distinguished from
each other at speeds that are small compared to the speed of light.

0.8

0.6 .,.g, ~

--
'-'
:::,
~
,.:

'lt

~
~
-::5

0.4 ~
& J;>.....
t::
r.J
0.2

0
0 0.2 0.4 0.6 0.8 1.0
u/c

Fig. 40-8 The speed of a particle P, as seen by observer S in Fig. 40-7, for the special
case of v' = u. All speeds are expressed as a ratio to c, the speed of light. The vertical
arrow corresponds to 5 X 107 miles/hr.
1006 NATURE AND PROPAGATION OF LIGHT Chap. 40
IJJII- Example 4. Suppose that v' = u = 25,000 miles/hr. What per cent error is
made in using Eq. 40-14 rather than Eq. 40-15 to calculate v? The speed of light is
6.7 X 108 miles/hr.
Equation 40-14 gives

v = v' +u = 25,000 miles/hr + 25,000 miles/hr = 50,000 miles/hr.


Equation 40-15 gives
v'
v=----
u +
l +
v'u/c 2
25,000 miles/hr +
25,000 miles/hr
(25,000 miles/hr) 2
1 + (6.7 X 10 8 miles/hr)2

50,000 miles/hr
1.0000000014

Even at 25,000 miles/hr the error in Eq. 40-14 is immeasurably small.

Example 5. Two electrons are ejected in opposite directions from radioactive


atoms in a sample of radioactive material. Let each electron have a speed, as
measured by a laboratory observer, of 0.6c (this corresponds to a kinetic energy of 130
kev). What is the speed of one electron as seen from the other?
Equation 40-14 gives
v = v' +u = 0.6c + 0.6c = l .2c.
Equation 40-15 gives
v = v' + u = 0.6c + 0.6c = O.SSc.
1 + v'u/c 2 1 + (0.6c)2/c 2

This example shows that for speeds that are comparable to c, Eqs. 40-14 and 40-15
yield rather different results. A wealth of indirect experimental evidence points to
the latter result as being correct. ....

40-5 Doppler Effect


We have seen that the same speed is measured for light no matter what the
relative speeds of the light source and the observer are. The measured fre-
quency and wavelength will change, but always in such a way that their
product, which is the velocity of light, remains constant. Such frequency
shifts are called Doppler shifts, after Johann Doppler (1803-1853), who first
predicted them.
In Section 20-7 we showed that if a source of sound is moving away from
an observer at a speed u, the frequency heard by the observer (see Eq.
20-10, which has been rearranged with u substituted for V 8 ) is

v' =
1
1
v-~-
+ u/v
l l. sound wave
2. observer fixed in medium
3. source receding from observer
(40-16)

In this equation vis the frequency heard when the source is at rest and v is
the speed of sound.
Sec. 40-5 DOPPLER EFFECT 1007

If the source is at rest in the transmitting medium but the observer is


moving away from the source at speed u, the observed frequency (see Eq.
20-9b, in which u has been substituted for v0 ) is

1. sound wave
[ 2. source fixed in medium (40-17)
3. observer receding from source

Even if the relative separation speeds u of the source and the observer are
the same, the frequencies predicted by Eqs. 40-16 and 40-17 are different.
This is not surprising, because a sound source moving through a medium in
which the observer is at rest is physically different from an observer moving
through that medium with the source at rest, as comparison of Figs. 20-10
and 20-11 shows.
We might be tempted to apply Eqs. 40-16 and 40-17 to light, substituting
c, the speed of light, for v, the speed of sound. For light, as contrasted with
sound, however, it has proved impossible to identify a medium of transmis-
sion relative to which the source and the observer are moving. This means
that "source receding from observer" and "observer receding from source"
are physically identical situations and must exhibit exactly the same Doppler
frequency. As applied to light, either Eq. 40-16 or Eq. 40-17 or both must
be incorrect. The Doppler frequency predicted by the theory of relativity
is, in fact,
1. light wave
1 - u/c
v . 2. source and observer (40-18)
Vl - (u/c) 2 separating

In all three of the foregoing equations we obtain the appropriate relations


for the source and the observer approaching each other if we replace u by - u.
Equations 40-16, 40-17, and 40-18 are not so different as they seem if the
ratio u/c is small enough. This was made clear in Example 3, Chapter 20,
for the first two of these equations. Let us expand Eqs. 40-16, 40-17 and
40-18 by the binomial theorem, as in the example referred to. The equations
then become, substituting c for v,

v' v[1-~+(:Y+···l (40-16a)

vI v(1-:} (40-17a)

and v' = v [ 1 - : + ~ (: Y
+ ··· l (40-18a)

The ratio u/c for all available monochromatic light sources, even those of
atomic dimensions, is small. This means that successive terms in these
1008 NATURE AND PROPAGATION OF LIGHT Chap. 40

equations become small rapidly and, depending on the accuracy required,


only a limited number of terms need be retained.
Under nearly all circumstances the differences among these three equatiom
are not important. Nevertheless, it is of extreme interest to carry out at
least one experiment precisely enough to serve as a test of Eq. 40-18a and
thus, in part, of the theory of relativity.

H. E. Ives and G. R. Stilwell carried out such a precision experiment in 1938.


They sent a beam of hydrogen atoms, generated in a gas discharge, down a tube at
speed u, as in Fig. 40-9a. They could observe light emitted by these atoms in a
direction opposite to u (atom 1, for example) using a mirror, and also in a direction
parallel to u (atom 2, for example). With a precision spectrograph, they could
photograph a particular characteristic spectrum line in this light, obtaining, on a
frequency scale, the lines marked v 1' and v/ in Fig. 40-9b. It is also possible to
photograph, on the same photographic plate, a line corresponding to light emitted
from resting atoms; such a line appears as v in Fig. 40-9b. A fundamental meas-
ured quantity in this experiment is fiv/v, defined from

(40-19)

(see Fig. 40-9b). It measures the extent to which the frequency of the light from
resting atoms fails to lie halfway between the frequencies v1 ' and v2 '. Table 40-2
shows that the measured results agree with the formula predicted by the theory of
relativity (Eq. 40-lSa) and not with the classical formula borrowed from the theory of
sound propagation in a material medium (Eq. 40-16a).
Ives and Stilwell did not present their experimental results as evidence for the
support of Einstein's theory of relativity but rather gave them an alternative the-
oretical explanation. Modern observers, looking not only at their excellent experi-
ment but at the whole range of experimental evidence, now give the Ives-Stilwell
experiment the interpretation we have described for it above.

100 v 30 kv

.-------11111 1111111111111111

Mirror 1
~2u

~
Acceleration
arc region region (a)

(b)

Fig. 40-9 The Ives-Stilwell experimPnt.


Sec. 40-5 DOPPLER EFFECT 1009

Table 40-2

THE IVES-STILWELL EXPERIMENT*

Speed of moving atoms ( = u), 106 meters/sec


~v, 10-5
v
0.865 1.01 1.15 1.33

I
I
Theoretical value according to I
classical theory (Eq. 40-l 6a) 1.67 2.26 2.90 3.94
I
Theoretical value according to the
theory of relativity (Eq. 40-18a) 0.835 1.13 1.45 1.97

Experimental value 0.762 1.1 1.42 1.9

* See Eq. 40-19; the table shows only part of the data taken by Ives and Stilwell.

The Doppler effect for light finds many applications in astronomy, where
it is used to determine the speeds at which luminous heavenly bodies are
moving toward us or receding from us. Such Doppler shifts measure only
the radial or line-of-sight components of the relative velocity. All galaxies *
for which such measurements have been made (Fig. 40-10) appear to be
receding from us, the recession velocity being greater for the more distant
galaxies; these observations are the basis of the expanding-universe concept.

(a) (b)

Fig. 40-10 (a) The central spot is a nebula in the constellation Corona Borealis; it is
130,000,000 light years distant.· (b) The central streak shows the distribution in wave-
length of the light emitted from this nebula. The two vertical dark bands show the
presence of calcium. The horizontal arrow shows that these calcium lines occur at longer
wavelengths than those for terrestrial light sources containing calcium, the length of the
arrow representing the wavelength shift. Measurement of this shift shows that the galaxy
is receding from us at 13,400 miles/sec. The lines above and below the central streak
represent light from a terrestrial source, used to establish a wavelength scale. (Courtesy
Mount Wilson and Mount Palomar Observatories.)
* See "The Red-Shift," Allen R. Sandage, Scientific American, p. 171, September 1956.
1010 NATURE AND PROPAGATION OF LIGHT Chap. 40
~ Example 6. Certain characteristic wavelengths in the light from a galaxy in
the constellation Virgo are observed to be increased in wavelength, as compared with
terrestrial sources, by about 0.4%. What is the radial speed of this galaxy with
respect to the earth? Is it approaching or receding?
If A is the wavelength for a terrestrial source, then

X' = 1.004X.

Since we must have X'v' Xv = c, we can write this as

v' = 0.996v.

This frequency shift is so small that, in calculating the source velocity, it makes no
practical difference whether we use Eq. 40-16, 40-17, or 40-18. Using Eq. 40-17
we obtain

v' = 0.996v = v ( 1 - ~) ·

Solving yields u/c = 0.004, or u = (0.004)(3 X 10 8 meters/sec) = 1.2 X 106 meters/


sec or 2.7 X 10 6 miles/hr. The galaxy is receding; had u turned out to be negative,
the galaxy would have been moving toward us. <1111

QUESTIONS

1. How might an eye-sensitivity curve like that of Fig. 40-2 be measured?


2. Why are danger signals in red, when the eye is most sensitive to yellow-green?
3. Comment on this definition of the limits of the spectrum of visible light given by a
physiologist: "The limits of the visible spectrum occur when the eye is no better adapted
than any other organ of the body to serve as a detector."
4. How can an object absorb light energy without absorbing momentum?
5. A searchlight sends out a parallel beam of light. Does the searchlight experience any
force associated with the emission of light?
6. Name two historic experiments, in addition to the radiation pressure measurements
of Nichols and Hull, in which a torsion balance was used. Both are described in this
book, one in Part 1 and one in Part 2.
7. Show that for complete absorption of a parallel beam of light the radiation pressure
on the absorbing object is given by p = S/c, where Sis the magnitude of the Poynting
vector and c is the speed of light in free space.
8. How could Galileo test experimentally that reaction times were an overwhelming
source of error in his attempt to measure the speed of light, described on p. 999?
9. It has been suggested that the velocity of light may change slightly in value as time
goes on. Can you find any evidence for this in Table 40-1?
10. A friend asserts that Einstein's postulate (that the speed of light is not affected by
the uniform motion of the source or the observer) must be discarded because it violates
"common sense." How would you answer him?
11. In a vacuum, does the speed of light depend on (a) the wavelength, (b) the frequency,
(c) the intensity, (d) the speed of the source, or (e) the speed of the observer?
12. Can a galaxy be so distant that its recession speed equals c? If so, how can we see
the galaxy? That is, will its light ever reach us?
Chap. 40 PROBLEMS l 011

PROBLEMS

1. (a) At what wavelengths does the eye sensitivity have half its maximum value?
(b) What are the frequency and the period of the light for which the eye is most sensitive?
2. It has been proposed that a spaceship might propel itself in the solar system by radi-
ation pressure, using a large sail made of aluminum foil. How large must the sail be if the
radiation force is to be equal in magnitude to the sun's gravitational attraction? Assume
that the mass of the ship + sail is 100 slugs, that the sail is perfectly reflecting, and that
the sail is oriented at right angles to the sun's rays. The sun's mass is 1.97 X 10 30 kg.
'--3. Radiation from the sun striking the earth has an intensity of 1400 watts/meter2 •
Assuming that the earth behaves like a flat disk at right angles to the sun's rays and that
all the incident energy is absorbed, calculate the force on the earth due to radiation
pressure. Compare it with the force due to the sun's gravitational attraction.
4. Prove, for a plane wave at normal incidence on a plane surface, that the radiation
pressure on the surface is equal to the energy density in the beam outside the surface.
This relation holds no matter what fraction of the incident energy is reflected.
5. Prove, for a stream of bullets striking a plane surface at right angles, that the "pres-
sure" is twice the (kinetic) energy density in the stream above the surface; assume that the
bullets are completely "absorbed"
by the surface. Contrast this with
the behavior of light (Problem 4).
6. A small spaceship whose mass,
with occupant, is 100 slugs is drift-
ing in outer space, where no gravi-
tational field exists. If it shines a
searchlight, which radiates 10 4 watts,
into space, what speed would the
ship attain in one day because of
the reaction force associated with
the momentum carried away by the
light beam?
7. What is the radiation pressure
l.Ometer away from a 500-watt light
bulb? Assume that the surface on
which the pressure is exerted faces
the bulb and is perfectly absorbing
and that the bulb radiates uniformly
in all directions.
8. The uncertainty of the distance
to the moon, as measured by the
reflection of radar waves from it, is
about 0.5 mile. Assuming that this
uncertainty is associated only with
the measurement of the elapsed
time, what uncertainty in this time
is implied?
·· 9. Roemer's method for measur-
ing the speed of light consisted in Fig. 40-11
observing the apparent times of re-
volution of one of the moons of Jupiter. The true period of revolution is 42.5 hr. (a)
Taking into account the finite speed of light, how would you expect the apparent time of
revolution to alter as the earth moves in its orbit from point x to point y in Fig. 40-11?
(b) What observations would be needed to compute the speed of light? Neglect the mo-
tion of .Jupiter in its orbit. Figure 40-11 is not drawn to scale.
1012 NATURE AND PROPAGATION OF LIGHT Chap. 40
10. Suppose that light is timed over a 1-mile base line and its speed is measured to the
accuracy quoted on p. 1002. How large an error in the length of the base line could be
tolerated, assuming other sources of error to be negligible?
11. For what value of u/c does Eq. 40-17 differ from Eq. 40-18 by 1 %?
=-~~ The "red shift" of radiation from a distant nebula consists of the light (H.y), known
to have a wavelength of 4340 X 10-s cm when observed in the laboratory, appearing to
have a wavelength of 6562 X 10-8 cm. (a) What is the speed of the nebula in the line of sight
relative to the earth? (b) Is it approaching or receding?
· t3. The difference in wavelength between an incident microwave beam and one re-
flected from an approaching or receding car is used to determine automobile speeds on the
highway. (a) Show that if vis the speed of the car and 11 the frequency of the incident
beam, the change of frequency is approximately 2v11/c, where c is the speed of the electro-
magnetic radiation. (b) For microwaves of frequency 2450 megacycles/sec, what is the
change of frequency per mile/hr of speed?
14. Show that, for slow speeds, the Doppler shift can be written in the approximate
form
d>.. u
-=-,
>.. c
where .:l>.. is the change in wavelength.
15. The period of rotation of the sun at its equator is 24.7 days; its radius is 7.0 X 108
meters. What Doppler wavelength shifts are expected for characteristic wavelengths in
the vicinity of 5500 A emitted from the edge of the sun's disk?
16. An earth satellite, transmitting on a frequency of 40 X 106 cycles/sec (exactly),
passes directly over a radio receiving station at an altitude of 250 miles and at a speed of
18,000 miles/hr. Plot the change in frequency, attributable to the Doppler effect, as a
function of time, counting t = 0 as the instant the satellite is over the station. (Hint:
The speed u in the Doppler formula is not the actual velocity of the satellite but its com-
ponent in the direction of the station. Use the nonrelativistic formula (Eq. 40--16a) and
neglect the curvature of the earth and of the satellite orbit.)
17. A rocketship is receding from the earth at a speed of 0.2c. A light in the rocketshlp
appears blue to passengers on the ship. What color would it appear to be to an observer
on the earth? See Fig. 40-2.
18. In the experiment of Ives and Stilwell the speed u of the hydrogen atoms in a par-
ticular run was 8.61 X 105 meters/sec. Calculate L'.111 1, L'.1112, and L'.111/11, on the assump-
tions that (a) Eq. 40--18a is correct and (b) that Eq. 40--16a is correct; compare your
results with those given in Table 40--2 for this speed. Retain the first three terms only
in Eqs. 40--18a and 40--16a.
Reflection and Refraction-
Plane Waves
and Plane Surfaces
CHAPTER 41

41-1 Reflection and Refraction


In Fig. 41-la a light beam falling on a water surface is both reflected from
the surface and bent (that is, refracted) as it enters the water. The incident
beam is represented in Fig. 41-lb by a single line, the incident ray, parallel to
the direction of propagation. The incident beam is assumed in Fig. 41-lb to
be a plane wave, the wavefronts being normal to the incident ray. The re-
flected and refracted beams are also represented by rays. The angles of
incidence (0 1 ), of reflection (0 1 '), and of refraction (0 2 ) are measured between
the normal to the surface and the appropriate ray, as shown in the figure.

Fig. 41-1 (a) A photograph showing reflection and refraction at an air-water interface.
(b) A representation using rays.
1013
1014 REFLECTION AND REFRACTION Chap. 41
The laws governing reflection and refraction can easily be found from ex-
periment:
1. The reflected and the refracted rays lie in the plane formed by the in-
cident ray and the normal to the surface at the point of incidence, that is,
the plane of Fig. 41- lb.
2. For reflection:
(41-1)
3. For refraction:
sin 01
-.-- = n21, (41-2)
sm 02
where n 21 is a constant called the index of refraction of medium 2 with respect
to medium 1. Table 41-1 shows the indices of refraction for some common

Table 41-1

SOME INDICES OF REFRACTION *


(For 'A = 5890 A)

Index
Medium of
Refraction

Water 1.33
Ethyl alcohol 1.36
Carbon bisulfide 1.63
Air (1 atm and 20°C) 1.0003
Methylene iodide 1.74
Fused quartz 1.46
Glass, crown 1.52
Glass, dense flint 1.66
Sodium chloride 1.53
Polyethylene 1.50-1.54
Fluorite 1.43

* Measured with respect to a vacuum. The index with respect to air (except, of
course, the index of air itself-see item 4) will be negligibly different in most cases.

substances with respect to a vacuum for a wavelength of 5890 A (sodium


light).
The index of refraction of one medium with respect to another generally
varies with wavelength, as Fig. 41-2 shows. Because of this fact refraction,
unlike reflection, can be used to analyze a beam of light into its component
wavelengths. Figure 41-3, taken from Newton's Opticks, shows how
Newton, using glass prism ABC, formed a spectrum of sunlight entering his
window through a small hole at F.
Sec. 41-1 REFLECTION AND REFRACTION 1015

1.480

1.475.
c:
21.470
u
Fig. 41-2 The index of refrac- e
tion of fused quartz at 18°C, ~ 1.465
with respect to a vacuum. 0
x
Q)
-g 1.460

1.455

1.450
3000 4000 5000 6000 7000 8000
Wavelength, A (in air at STP)

The law of reflection was known to Euclid. That of refraction was dis-
covered experimentally by Willebrod Snell (1591-1626) and deduced from
the early corpuscular theory of light by Rene Descartes (1596-1650). The
law of refraetion is known as Snell's law or (in France) as Descartes' law.
The laws of reflection and refraction can be derived from Maxwell's
equations, which means that these laws should hold for all regions of the
electromagnetic spectrum. Figure 41-4a shows an experimental setup for in-
vestigating the reflection of microwaves from a large metal sheet. Figure
41-4b shows the reading of the detector as a function of the angular position
of the mirror. The existence of a reflected beam at the proper angle confirms
the law of reflection for microwaves. There is ample experimental evidence
that Eqs. 41-1 and 41-2 correctly describe the behavior of reflected and
refracted beams in all parts of the electromagnetic spectrum.
It is common knowledge that a polished steel surface will form a well-defined reflected
beam if an incident beam falls on it, but a sheet of paper wi.11 reflect light
more or less in all directions (diffuse reflection). It is largely by diffuse reflection that
we see nonluminous objects around us. The difference between diffuse and specular
(that is, mirror-like) reflection is a matter of surface roughness; a reflected beam will

Fig. 41-3 Sunlight from a hole Fin a screen is refracted by prism ABC, forming a spec-
trum on screen MN; from Newton's Opticks (1704).
1016 REFLECTION AND REFRACTION Chap. 41

a~!~~~a ~M:tal
( </> Ym1rror
~ - - - - --
M.IC Incident ray I
I
oscillator "'-.,ip ara bo11c
. l
I
reflector I
Reflected ray t
Microwave~
(a) detector LJ
~100
"C
ro
....
Q)

....
+"'
0
(.)
60
Q)
~ 40
"C
Q)
>
:;:::; 20
ro
Q)
a::
10 20 30 40 50 60 70 80 90
ct,, degrees
(b)

Fig. 41-4 (a) An apparatus to test the law of reflection for microwaves. (b) A reflected
beam (for "}-. ~ 10 cm) appears for the expected orientation of the mirror.

be formed only if the average depth of the surface irregularities of the reflector is
substantially less than the wavelength of the incident light. This criterion of surface
roughness has different implications in different regions of the electromagnetic spec-
trum. The bottom of a cast-iron skillet for example is a good reflector for microwaves
of wavelength 0.5 cm but it is not a good reflector for visible light (that is, one cannot
shave by it).
A second requirement for the existence of a reflected beam is that the transverse
dimensions of the reflector must be substantially larger than the wavelength of the
incident beam. If a beam of visible light falls on a polished metal disk the size of a
dime, a reflected beam will be formed. However, if the same disk is placed in a beam
of short radio waves with, say, 'A. = 1.0 meter, radiation will be scattered in all
directions from it, and no well-defined unidirectional beam will appear. We inves-
tigate this phenomenon of diffraction in Chapter 44. The requirements that surfaces
be "smooth" and "large" also apply to the formation of refracted beams. If these
two requirements are not met the description of reflection and refraction in terms of
rays, whose behavior is governed by Eqs. 41-1 and 41-2, is not valid .
.... Example 1. Figure 41-5 shows an incident ray i striking a plane mirror MM'
at angle of incidence 0. Trace this ray.
The reflected ray makes an angle () with the normal at b and falls as an incident ray
on mirror M' M". Its angle of incidence ()' on this mirror is 7r /2 - fJ. A second
Sec. 41-1 REFLECTION AND REFRACTION 1017

reflected ray r' makes an angle 0' with the normal erected at b'. Rays i and r' are
antiparallel for any value of 0. To see this, note that

¢ = 1r - 20' = 1r - 2 (~ - 0) = 20.
Two lines are parallel if their opposite interior angles for an intersecting line (¢ and 20)
are equal.
Repeat the problem if the angle between the mirrors is 120° rather than 90°.

Fig. 41-5 Example 1.

Example 2. An incident ray in air falls on the plane surface of a block of quartz
and makes an angle of 30° with the normal. This beam contains two wavelengths,
4000 and 5000 A. The indices of refraction for quartz with respect to air (nqa) at
these wavelengths are 1.4702 and 1.4624, respectively. What is the angle between
the two refracted beams?
From Eq. 41-2 we have, for the 4000-A beam,
sin 01 = nqa sin 02,
or sin 30° = (1.4702) sin 02,
which leads to 02 = 19.88°.
For the 5000-A beam we have
sin 30° = (1.4624) sin 02',
or 02' = 19.99°.
The angle !:i.O between the beams is 0.11 °, the shorter wavelength component being
bent through the larger angle, that is, having the smaller angle of refraction.
1018 REFLECTION AND REFRACTION Chap. 41

Fig. 41-6 Example 3.

Example 3. An incident ray falls on one face of a glass prism in air as in Fig. 41-6.
The angle eis so chosen that the emerging ray also makes an angle e with the normal
to the other face. Derive an expression for the index of refraction of the prism
material with respect to air.
Note that L abc = a, the two angles having their sides mutually perpendicular.
Therefore
a = !ct>, (41-3)
The deviation angle i/; is the sum of the two opposite interior angles in triangle
aed, or
iJ; = 2ce - a).
Substituting !ct> for a and solving for e yields
e = !Cf+ ct>). (41-4)
At point a, e is the angle of incidence and a the angle of refraction. The law of
refraction (see Eq. 41-2) is
sin e = nga sin a,
in which nga is the index of refraction of the glass with respect to air.
From Eqs. 41-3 and 41-4 this yields
. f+ct>
--
Slil - =
. ct>
nga Sln 2
2

or
sin !(iJ;+ </>)
nga = sin (cp/2)
which is the desired relation. This equation holds only fore so chosen that the light
ray passes symmetrically through the prism. For this condition the deviation angle
i/; is a minimum; if e is either increased or decreased, a larger deviation will be pro-
duced. .,..

41-2 Huygens' Principle


A theory of light would not be accepted if it were not able to predict the
well-established laws of reflection and refraction. These laws can be derived
from Maxwell's equations, but mathematical complexity prevents us from
Sec. 41-2 HUYGENS' PRINCIPLE 1019

doing so. Fortunately, these and several other laws of optics can be derived
on the basis of a simpler but less comprehensive theory of light, put forward
by the Dutch physicist Christian Huygens in 1678. This theory simply
assumes that light is a wave rather than, say, a stream of particles. It says
nothing about the nature of the wave and, in particular-since Maxwell's
theory of electromagnetism appeared only after the lapse of a century-
gives no hint of the electromagnetic character of light. Huygens did not
know whether light was a transverse wave or a longitudinal one; he did
not know the wavelengths of visible light; he had little knowledge of the
speed of light. K evertheless, his theory was a useful guide to experiment for
many years and remains useful today for pedagogic and certain other prac-
tical purposes. We must not expect it to yield the same wealth of detailed
information that Maxwell's more complete electromagnetic theory does.
Huygens' theory is based on a geometrical construction, called Huygens'
principle, that allows us to tell where a given wavefront will be at any time
in the future if we know its present position; it is: All points on a wavefront
can be considered as point sources for the production of spherical secondary
wavelets. After a time t the new p0sition of the wavefront will be the surface of
tangency to these secondary wavelets.
We illustrate this by a trivial example: Given a wavefront (ab in Fig. 41-7)
in a plane wave in free space, where will the wavefront be a time t later?
Following Huygens' principle, we let several points on this plane (see dots)
serve as centers for secondary spherical wavelets. In a time t the radius of
these spherical waves is ct, where c is the speed of light in free space. The
plane of tangency to these spheres at time t is represented by de. As we

Fig. 41-7 The propagation of


a plane wave in free space is
described by the Huygens con-
struction. Note that the ray
(horizontal arrow) representing
the wave is perpendicular to the
wavefronts.
1020 REFLECTION AND REFRACTION Chap. 41
expect, it is parallel to plane ab and a perpendicular distance ct from it.
Thus plane wavefronts are propagated as planes and with speed c. Note
that the Huygens method involves a three-dimensional construction and that
Fig. 41-7 is the intersection of this construction with the plane of the page.
We might expect that, contrary to observation, a wave should be radiated back-
ward as well as forward from the dots in Fig. 41-7. This result is avoided by assuming
that the intensity of the spherical wavelets is not uniform in all directions but varies
continuously from a maximum in the forward direction to a minimum of zero in the
back direction. This is suggested by the shading of the circular arcs in Fig. 41-7.
Huygens' method can be applied quantitatively to all wave phenomena; see Problem 1.
The method was put on a firm mathematical footing by Augustin Fresnel (1788-1827).

41-3 Huygens' Principle and the Law of Reflection


Figure 41-Sa shows three wavefronts in a plane wave falling on mirror
MM'. For convenience they are chosen to be one wavelength apart. Note
that e1 , the angle between the wavefronts and the mirror, is the same as the
angle between the incident ray and the normal to the mirror. In other

(a)

(b)

Fig. 41-8 The reflection of a plane


wave as described by the Huygens
construction.

(c)

(d)
Sec. 41-4 HUYGENS' PRINCIPLE AND THE LAW OF REFRACTION 1021

words, f)i is the angle of incidence. The three wavefronts are related to each
other by the Huygens construction, as in Fig. 41-7.
In Fig. 41-8b a Huygens wavelet centered on point a will expand to
include point l after a time ).../c. Light from point pin this same wavefront
cannot move beyond the mirror but must expand upward as a spherical
Huygens wavelet. Setting a compass to radius A and swinging an arc about
p provides a semicirrle to which the reflected wavefront must be tangent.
Since point l must lie on the new wavefront, this tangent must pass through l.
Note that the angle Bi' between the wavefront and the mirror is the same as
the angle between the reflected ray and the normal to the mirror. In other
words, 81 ' is the angle of reflection.
Consider right triangles alp and a'lp. They have side lp in common and
side al ( = )...) is equal to side a'p. The two right triangles are thus congruent
and we may conclude that

as required by the law of reflection. If the student recalls that the Huygens
construction is three-dimensional and that the arcs shown represent segments
of spherical surfaces, he will be able to convince himself that the reflected
ray lies in the plane formed by the incident ray and the normal to the mirror,
that is, the plane of Fig. 41-8. This is also a requirement of the law of re-
flection; see p. 1014.

41-4 Huygens' Principle and the Law of Refraction


Figure 41-9 shows four stages in the refraction of three successive wave-
fronts in a plane wave falling on an interface between air (medium 1) and
glass (medium 2). For convenience, we assume that the incident wavefronts
are separated by )... 1 , the wavelength as measured in medium 1. Let the
speed of light in air be v1 and that in glass be v2 . We assume that
(41-5)
This assumption about the speeds is vital to the derivation that follows.
It was not possible to test it experimentally because of technical difficulties
until 1850, when the assumption was shown by Foucault to be correct.
The wavefronts in Fig. 41-9a are related to each other by the Huygens
construction of Fig. 41-7. As in Fig. 41-8, 81 is the angle of incidence.
In Fig. 41-9b consider the time ( = )...i/v 1 ) during which a Huygens wavelet
from point e moves to include point c. Light from point h, traveling through
glass at a reduced speed (recall the assumption of Eq. 41-5) will move a
shorter distance
(41-6)

during this time. The refracted wavefront must be tangent to an arc of this
radius centered on h. Since c lies on the new wavefront, the tangent must
pass through this point, as shown. K ote that 82 , the angle between the
refracted wavefront and the air-glass interface, is the same as the angle
1022 REFLECTION AND REFRACTION Chap. 41

(a)

(b)

Fig. 41-9 The refraction of a plane


wave as described by the Huygens
construction; the reflected wave is
omitted for simplicity. Note the
change in wavelength on refraction.
(c)

between the refracted ray and the normal to this interface. In other words,
82 is the angle of refraction. Note, too, that the wavelength in glass ()\ 2 ) 1s
less than the wavelength in air (;>.. 1 ).
For the right triangles hce and hce' we may write
A1
sin 81 = - (for hce)
he

. A2
and sm 82 = - (for hce').
he
Dividing and using Eq. 41-6 yields

a constant. (41-7)
Sec. 41-4 HUYGENS' PRINCIPLE AND THE LA w OF REFRACTION 1023

The law of refraction, as stated in Eq. 41-2, is

sin 81
-.-- = n21, (41-2)
sm 82
so that n 21 is now revealed as the ratio of the speeds of light in the two
media, or
(41-8)

We may rewrite Eq. 41-7 as

(41-9)

fn which c is the speed of light in free space. The quantities (c/v 1 ) and
(c/v 2 ) (see Eq. 41-8) are the indices of refraction of medium 1 and of medium
2, respectively, with respect to a vacuum. Introducing the symbols n 1 and
n 2 for these quantities allows us to write the law of refraction as

(41-10)

If we assume that the medium above the glass in Fig. 41-9 is a vacuum
rather than air, the speed v1 becomes c and the wavelength, called ;\ 1 in
Fig. 41-9, assumes a value;\ that is characteristic of the wave in free space.
Equation 41-6 may thus be written

V2 ;\
;\2 = ; \ - = - · (41-11)
c n2

This shows specifically that the wavelength of light in a material medium is


less than the wavelength of the same wave in a vacuum. Figure 41-9 shows
clearly the difference in wavelength in the two media.
The application of Huygens' principle to refraction requires that if a light ray is
bent toward the normal in passing from air to an optically dense medium then the
speed of light in that optically dense medium (glass, say) must be less than that in air;
see Eq. 41-5. This requirement holds for all wave theories of light. For the early
particle theory of light put forward by Newton, refraction can be explained only if
the speed of light in the medium in which light is bent toward the normal (the
optically dense medium) is greater than that in air. The dense medium was thought
to exert attractive forces on the light "corpuscles" as they neared the surface, speeding
them up and changing their direction to cause them to make a smaller angle with the
normal. Figure 41-10 shows a figure from a 1637 work of Rene Descartes, in which he
makes an analogy between the refraction of light and the motion of a tennis ball on
entering a medium in which it moves more slowly.
An experimental comparison of the speed of light in water and in air is decisive,
therefore, between the wave and corpuscular theories of light. Such a measurement
was first carried out by Foucault in 1850; he showed conclusively that light travels
more slowly in water than in air, thus ruling out the corpuscular theory of Newton.
1024 REFLECTION AND REFRACTION-II Chap. 41

Fig. 41-10 According to the early


(incorrect) corpuscular theory of
light, AB! is the path of a ray on
entering a medium in which its speed
is less. The (correct) wave theory
predicts that for the ray shown the
speed in the lower medium (below
BE) must be greater; from Des-
cartes' La Dioptrique (1637).

41-5 Total Internal Reflection


Let light rays in an optically dense medium (glass, say) fall on a surface
on the other side of which is a less optically dense medium (air, say); see
Fig. 41-11. As the angle of incidence () is increased, a situation is reached
(see ray e) at which the refracted ray points along the surface, the angle of
refraction being 90°. For angles of incidence larger than this critical angle 8c
no refracted ray exists, giving rise to a phenomenon called total internal
reflection.
The critical angle is found by putting 82 = 90° in the law of refraction
(see Eq. 41-10):

or (41-12)

Fig. 41-11 Showing the total internal reflection of light from a source S; the critical angle
is Oc,
Sec. 41-5 TOTAL INTERNAL REFLECTION
1025

For glass and air sin ec = (1.00/1.50) = 0.667, which yields ec = 41.8°.
Total internal reflection does not occur when light originates in the medium
of lower index of refraction.
Light can be "piped" from one point to another with little loss by allowing it to
enter one end of a rod of transparent plastic. The light will undergo total internal

Fig. 41-12 A bundle of tapered fibers (belcrw) is placed over the letter S. Above, with
the aid of a mirror, we see that the image, reduced in size, is transmitted to the top of the
bundle by total internal reflection in the individual fibers. (Courtesy Dr. N. S. Kapany,
Optics Technology, Inc.)

reflection at the boundary of the rod and will follow its contour, emerging at its far
end. Images may be transferred from one location to another, using a bundle of fine
glass fibers, each fiber transmitting a small fraction of the image.* Such bundles can
be made in which a (flexible) seven-foot length delivers half the energy entering it at
the far end. In a fiber in such a bundle a typical ray may undergo 48,000 reflections.
Most of the energy loss is due to absorption within the glass, reflection being almost
* See "Fiber Optics," by N. S. Kapany, Scientific American, November 1960.
1026 REFLECTION AND REFRACTION Chap. 41

Fig. 41-13 A photomicrograph of light emerging from the end of a bundle of fibers. The
fiber diameters approach the wavelength of light so that each fiber acts like an "optical
waveguide." We have here convincing visual evidence that light is an electromagnetic
wave. (Courtesy of Dr. N. S. Kapany, Optics Technology, Inc.)

truly total. Measurable amounts of light can be transmitted through single fibers
150 feet long.
Fiber optics techniques make possible many useful optical devices for transmitting
and transforming luminous images. Figure 41-12 shows a short fiber bundle con-
structed so that the fibers taper in diameter along its length. The wide end is shown
placed over the letter Sin the printed word "OPTICS." We see, with the aid of a
mirror placed above the bundle, that a letter S, reduced in size, has been transmitted
by tofal internal reflection through the bundle to its narrow end.
Figure 41-13 is an enlarged view of a cross section of such a fiber bundle in .which
the diameters of the individual fibers are made so small that they are of the order of
magnitude of the wavelength cf light. This condition violates the spirit of our
assumption of p. 1016, namely that the transverse dimensions of reflecting and
refracting surfaces would be large compared to the wavelength of light. Conse-
quently, a description of the reflect.ion and refraction of light in terms of rays, as in
Figs. 41-lb and 41-11, is not possible. Figure 41-13 is readily interpreted, however,
on the basis of the electromagnetic wave theory of light and provides convincing
pictorial supporting evidence for that theory. The fibers behave like waveguides *
* A dielectric rod can serve as a waveguide in the same way that a hollow metal pipe
can, and it has similar properties.
Sec. 41-5 TOTAL INTERNAL REFLECTION 1027

(see Section 39-3), and the patterns of darkness and light represent the distribution
of the E and B vectors for various modes of oscillation of the electromagnetic waves
traveling down the "guides."

~ Example 4. Figure 41-14a shows a triangular prism of glass, a ray incident


normal to one face being totally reflected. If 81 is 45°, what can you conclude about
the index of refraction n of the glass?

Fig. 41-14 Example 4.

(a) (b)

The angle 81 must be equal to or greater than the critical angle Oc where Oc is given
by Eq. 41-12:
. n2 1
Sln Oc = - = - ,
n1 n

in which, for all practical purposes, the index of refraction of air ( = n2) is set equal
to unity. Suppose that the index of refraction of the glass is such that total internal
reflection just occurs, that is, that Oc = 45°. This would mean

1
n = . 450 = 1.41.
sm

Thus the index of refraction of the glass must be equal to or larger than 1.41. If it
were less, total internal reflection would not occur.

Example 5. What happens if the prism in Example 4 (assume that n = 1.50) is


immersed in water (n = 1.33)? See Fig. 41-14b.
The new critical angle, given by Eq. 41-12, is

. n2 1.33
sm Oc = n 1 = l. 50 = 0.887,

which corresponds to Oc = 62.5°. The actual angle of incidence ( = 45°) is less than
this so that we do not have total internal reflection.
There is a reflected ray, with an angle of reflection of 45 °, as Fig. 41-14b shows.
There is also a refracted ray, with an angle of refraction given by

(l.50)(sin 45°) = (1.33) sin 82,

which yields 02 = 52.9°.

Maxwell's equations permit us to calculate how the incident energy is divided


between the reflected and the refracted beams. Figure 41-15 shows the theoretical
1028 REFLECTION AND REFRACTION Chap. 41

~ 60f--~-+~~l""'?.?'4""""-:.:_::_4-~--+~~-+-~~t--~---+--J----i
:,,;;
~
~ 40t--~-+~~L....<LL----,-....it11,rLl.i_J_~--+~~-+-~~t--~-f------\----i
w

Reflected wave
0 0° 10° 20° 30° 40° 50° 60° 70° 80° 90°
Angle of incidence, ()

~
~ 40t--~-t-~~-t-~------i-~~4--it~~~~""-'-,--""""""'~
w

Reflected wave

10° 20° 30° 40° 50° 60° 70° 80° 90°


Angle of incidence, ()

Fig. 41-15 (a) The per cent of the energy reflected and refracted when an incident wave
in air falls on glass (n = 1.50). (b) The same for the incident wave in glass, showing total
internal reflection.

prediction for (a) a light beam in air falling on a glass-air interface and (b) a light beam
in glass falling on such an interface. Figure 41-15a shows that for angles of incidence
up to about 50°, less than 10% of the light energy is reflected. At grazing incidence,
however (that is, angles of incidence near 90°), the surface becomes an excellent
reflector. We are all familiar with the high reflecting power of a wet road for light
from automobile headlights that strikes near grazing incidence.
Figure 41-15b shows clearly that at a certain critical angle (41.8° in this case; see
Eq. 41-12) all the light is reflected. For angles of incidence appreciably below this
value, about 4% of the energy is reflected.

41-6 Fermat's Principle


In 1650 Pierre Fermat discovered a remarkable principle which we often express
today in these terms: A light ray travelirig from one point to another will follow a path
such that, compared with nearby paths, the time required is either a minimum or a
maxim um or will remain urichanged (that is, it will be stationary).
The laws of reflection and refraction can readily be derived from this principle.
Sec. 41-6 FERMAT'S PRINCIPLE 1029

Fig. 41-16 A ray from A passes through B after reflection at P.

Figure 41-16 shows two fixed points A and Band a ray APB connecting them.*
The total length l of this ray is

l = y~2+ x2 + V b2 + (d - x) 2'
where x locates the point P at which the ray touches the mirror.
According to Fermat's principle, P will have a position such that the time of travel
of the light must be a minimum (or a maximum or must remain unchanged). Ex-
pressed in another way, the total length l of the ray must be a minimum (or a max-
imum or must remain unchanged). In either case, the methods of the calculus re-
quire that dl/dx be zero. Taking this derivative yields

!! = (!)(a2 + x2)-1f2(2x) + ![b2 + (d - x)2J-112(2)(d - x)(-1) = 0,

which can be rewritten as


x d-x
va2 + x2 Vb + (d
2 - xi
Comparison with Fig. 41-16 shows that this can be written as
sin (Ji = sin 01',
or
which is the law of reflection.
To prove the law of refraction from Fermat's principle, consider Fig. 41-17, which
shows two points A and Bin two different media and a ray APB connecting them.
The time t is given by

Using the relation n = c/v this can be written as

t =
n1Z1 + n2l2 =
l

c c
The quantity l ( = n 1Z1 + n2l 2 ) is called the optical path length of the ray. Equation
41-11 (written as A = n"'An) shows that the optical path length is equal to the length
* We assume that ray APB lies in the plane of the figure; see Problem 20.
1030 REFLECTION AND REFRACTION Chap. 41

Fig. 41-17 A ray from A passes through B after refraction at P.

that this same number of waves would have if the medium were a vacuum. The
optical path length must not be confused with the geometrical path length, which is
Z1 + Z2.
Fermat's principle requires that l be a minimum (or a maximum or must remain
unchanged) which, in turn, requires that x be so chosen that dl/dx = 0. The optical
pa th length is
l = n1li + n2Z2 = n1Va 2 x 2 + n2Vb 2 +(d - x)~ +--
Differentiating yields
dl
~
dx
= n1(!)(a 2 +x 2) - 1 i 2 (2x) + n2(!)[b 2 + (d - x) 2J- 1i 2 (2)(d - x)(-1) = 0
I

which can be rewritten as


x d - x
n1 ~,=== = n2 ·
Va2 2 +x Vb + (d -
2 x) 2
Comparison with Fig. 41-17 shows that this, in turn, can be written as
n1 sin 81 = n2 sin 82,
which is the law of refraction.
In each of the examples of this section the time required, or, what is equivalent,
the optical path length, proves to be a minimum. Problem 22 describes a case in
which it may be a maximum, a minimum, or stationary.

QUESTIONS

1. Discuss the propagation of spherical and of cylindrical waves, using Huygens'


principle.
2. Does Huygens' principle apply to sound waves in air?
3. If Huygens' principle predicts the laws of reflection and refraction, why is it necessary
or desirable to view light as an electromagnetic wave, with all its attendant complexity?
Chap. 41 PROBLEMS 1031

4. Would you expect sound waves to obey the laws of reflection and of refraction obeyed
by light waves?
5. A street light, viewed by reflection across a body of water in which there are ripples,
appears very elongated. Explain.
6. The light beam in Fig. 41-la is broadened on entering the water. Explain.
7. By what per cent does the speed of blue light in fused quartz differ from that of red
light?
8. Can (a) reflection phenomena or (b) refraction phenomena be used to determine the
wavelength of light?
9. How did Foucault's measurement of the speed of light in water decide between the
wave and the particle theories of light?
10. Why does a diamond "sparkle" more than a glass imitatiah cut to the same shape?
11. Is it plausible that the wavelength of light should change in passing from air into
glass but that its frequency should not? Explain.
12. How can one determine the indices of refraction of the media in Table 41-1 relative
to water, given the data in that table?
13. You are given a cube of glass. How can you find the speed of light (from a sodium
light source) in this cube?
14. Describe and explain what a fish sees as he looks in various directions above his
"horizon."
15. Design a periscope, taking advantage of total internal reflection. What are the
advantages compared with silvered mirrors?
16. What characteristics must a material have in order to serve as an efficient "light
pipe"?
17. Can the optical path length between two points ever be less than the geometrical
path length between those points?

PROBLEMS

1. One end of a stick is dragged through water at a speed v which is greater than the
speed u of water waves. Applying Huygens' construction to the water waves, show that
a conical wavefront is set up and that its half-angle a is given by

sin a = u/v.

This is familiar as the bow wave of a ship or the shock wave caused by an object moving
through air with a speed exceeding that of sound, as in Fig. 20---12 .
.-%.'·Prove that if a mirror is rotated through an angle a, the reflected beam is rotated
through an angle 2a. Is this result reasonable for a = 45°?
3. What is the speed in fused quartz of light of wavelength 5500 A; see Fig. 41-2?
4. The wavelength of yellow sodium light in air is 5890 A. (a) What is its frequency?
(b) What is its wavelength in glass whose index of refraction is 1.52? (c) From the results
of (a) and (b) find its speed in this glass.
5. The speed of yellow sodium light in a certain liquid is measured to be l.92 X 10 8
meters/sec. What is the index of refraction of this liquid, with respect to air, for sodium
light?
6. Suppose that the speed of light in air has been measured with an uncertainty of, say,
1 km/sec. In calculating the speed in vacuum, suppose that it is not certain whether n
for air is 1.00029 or 1.00030. (a) How much extra uncertainty is introduced into the
calculated value for c? (b) Estimate how accurately n should be known for this purpose.
~- In Fig. 41-6 show by graphical ray tracing, using a protractor, that if e for the
incident ray is either increased or decreased, the deviation angle f is increased. The
symmetrical situation shown in this figure is called the position of minimum deviation.
1032 REFLECTION AND REFRACTION Chap. 41
""8: A ray of light is incident normally on the face ab of a glass prism (n = 1.52), as
shown in Fig. 41-18. (a) Assuming that the prism is immersed in air, find the largest value
for the angle cp so that the ray is
totally reflected at face ac. (b)
Find cp if the prism is immersed
in water.
9. Show that for a thin prism
(¢ small) and light not far from
normal incidence (01 small) the de-
viation angle is independent of the
angle of incidence and is equal to
Fig. 41-18 (n - 1) cp (see Fig. 41-6).
10. A glass prism with an apex
angle of 60° has n = 1.60. (a) What is the smallest angle of incidence for which a ray can
enter one face of the prism and emerge from the other? (b) What angle of incidence would
be required for the ray to pass through the prism symmetrically, as in Fig. 41-6?
11. A 60° prism is made of fused quartz. A ray of light falls on one face, making an
angle of 45° with the normal. Trace the ray through the prism graphically with some
care, showing the paths traversed by rays representing (a) blue light, (b) yellow-green light,
and (c) red light. See Figs. 40-2 and 41-2.
--1?. Ptolemy, who lived at Alexandria toward the end of the first century A.D., gave
the following measured values for the angle of incidence 0 1 and the angle of refraction 02
for a light beam passing from air to water:

01 02 01 02
100 7°45 1 50° 35°0'
20° 15°30 1 60° 40°30 1
30° 22°30 1 70° 45°30 1
40° 29°0 1 80° 50°0 1
Are these data consistent with Snell's law; if so, what index of refraction results? These
data are interesting as the oldest recorded physical measurements.
----13, Prove that a ray of light
incident on the surface of a sheet of
plate glass of thickness t emerges
from the opposite face parallel to
its initial direction but displaced
sideways, as in Fig. 41-19. Show
that, for small angles of incidence
0, this displacement is given by

n - 1
x = te~-
n

where n is the index of refraction


and O is measured in radians.
14. A plane wave of white light
traveling in fused quartz strikes a
plane surface of the quartz, mak-
ing an angle of incidence 0. Is it
possible for the internally reflected Fig. 41-19
beam to appear (a) bluish or (b)
reddish? Roughly what value of O must be used? (Hint: White light will appear bluish
if wavelengths corresponding to red are removed from the spectrum.)
15. A point source of light is placed a distance h below the surface of a large deep lake.
(a) Show that the fraction! of the light energy that escapes directly from the water surface
Chap. 41 PROBLEMS 1033
1s independent of hand is given by
1
f = 2n2
where n is the index of refraction of water. (Note: Absorption within the water and reflec-
tion at the surface-except where it is total-have been neglected.) (b) Evaluate this
fraction for n = 1.33.
16. Figure 41-20 shows a constant-deviation prism. Although made of one piece of
glass, it is equivalent to two 30°-60°-90° prisms and one 45°-45°-90° prism. White light
is incident in the direction i. 1:1 1 is changed by rotating the prism so that, in turn, light of
any desired wavelength may be made to follow the path shown, emerging at r. Show that,
if sin 1:1 1 = !n, then 1:1 2 = 1:1 1 and beams i and r are at right angles.

Fig. 41-20

17. A drop of liquid may be placed on a semicircular slab of glass as in Fig. 41-21.
Show how to determine the index of refraction of the liquid by observing total internal
reflection. The index of refraction of the glass is unknown and must also be determined.
Is the range of indices of refraction that can be measured in this way restricted in any
sense?

Fig. 41-21
1034 REFLECTION AND REFRACTION Chap. 41
18. A light ray falls on a square glass slab as in Fig. 41-22. What must the index of
refraction of the glass be if total internal reflection occurs at the vertical face?

Fig. 41-22

19. A glass cube has a small spot at its center. What parts of the cube face must be
covered to prevent the spot from being seen, no matter what the direction of viewing?
What fraction of the cube surface must be so covered? Assume a cube edge of 1.0 cm
and an index of refraction of 1.50. (Neglect the subsequent behavior of an internally
reflected ray.)
20. Using Fermat's principle, prove that the reflected ray, the incident ray, and the
normal lie in one plane.
21. Prove that the optical path lengths for reflection and refraction in Figs. 41-16
and 41-17 are minima when compared with other nearby paths connecting the same two
points.
Reflection and Refraction-
Spherical Waves
and Spherical Surfaces
CHAPTER 42

42-1 Geometrical Optics and Wave Optics


In Chapter 41 we described the reflection and refraction of plane waves
at plane surfaces. In this chapter we consider the more general case of
spherical waves falling on spherical reflecting and refracting surfaces. All
of the results of Chapter 41 will emerge as special cases of the results of this
chapter, since a plane can be viewed as a spherical surface with an infinite
radius of curvature.
Both in Chapter 41 and in this chapter we make extensive use of rays.
Although a ray is a convenient construction, it proves impossible to isolate
one physically. Figure 42-la shows schematically a plane wave of wave-
length "A falling on a slit of width a = 5"A. We find that the light flares out
into the geometrical shadow of the slit, a phenomenon called diffraction.
Figures 42-lb (a = 3"A) and 42-lc (a = "A) show that diffraction becomes
more pronounced as a/"A ------, 0 and that attempts to isolate a single ray from
the incident plane wave are futile.
Figure 42-2 shows water waves in a shallow ripple tank, produced by
tapping the water surface periodically and automatically with the edge of a
flat stick. We see that the plane wave so generated flares out by diffraction
when it encounters a gap in a barrier placed across it. Diffraction is char-
acteristic of waves of all types. We can hear around corners, for example,
because of the diffraction of sound waves.
The diffraction of waves at a slit (or at an obstacle such as a wire) is ex-
pected from Huygens' principle. Consider the portion of the wavefront
that arrives at the position of the slit in Fig. 42-1. Every point on it can be
1035
1036 REFLECTION AND REFRACTION-II Chap. 42

Fig. 42-1 An attempt to isolate a ray by reducing the slit width a fails because of diffrac-
tion, which becomes more pronounced as a/}.. approaches zero.

viewed as the site of an expanding spherical Huygens' wavelet. The "bend-


ing" of light into the region of the geometrical shadow is associated with the
blocking off of Huygens' wavelets from those parts of the incident wavefront
that lie behind the slit edges.
Figure 42-3 was made by allowing parallel light to fall on a slit placed
50 cm in front of a photographic plate. In Fig. 42-3a the slit width was
about 6 X 10-3 mm. The central band of light is much wider than this,
showing that light has "flared out" into the geometric shadow of the slit.
In addition, many secondary maxima, omitted from Fig. 42-1 for simplicity,
appear. Figure 42-3b shows what happens when the slit width is reduced
by a factor of two. The central maximum becomes wider, in agreement with
Fig. 42-1. Figure 42-3c shows the effect of reducing the slit width by an
additional factor of 7, to 4 X 10-4 mm. The central maximum is now much
wider and the secondary maxima, whose intensities relative to the central
Sec. 42-1 GEOMETRICAL OPTICS AND WA VE OPTICS 1037

Fig. 42-2 Diffraction of water waves at a slit in a ripple tank. Note that the slit width
is about the same size as the wavelength. (Courtesy of Educational Services Incorporated.)

maximum have been deliberately overemphasized by long exposure, are very


evident.
Diffraction can be ignored if the ratio a/'A. is large enough, a being a measure
of the smallest sideways dimension of the slit or obstacle. If a » 'A, light
appears to travel in straight lines which can be represented by rays that obey
the laws of reflection (Eq. 41-1) and refraction (Eq. 41-2). In Chapter 41
this condition, called geometrical optics, prevailed, the lateral dimensions of
all mirrors, prisms, etc., being much greater than the wavelength. We as-
sume in this chapter also that the conditions for geometrical optics are
satisfied.

(a)

(b)

(c)

Fig. 42-3 (a) The intensity of light diffracted from a slit of width a ,..__, 6 X 10-3 mm
and falling on a screen 50 cm beyond. (b) The slit width is reduced by a factor of two.
(c) The slit width is further reduced by an additional factor of seven. Note that secondary
maxima, made prominent in this case by deliberate overexposure, appear on either side
of the central maximum. These secondary maxima have been omitted from Fig. 42-1
for simplicity.
1038 REFLECTION AND REFRACTION-II Chap. 42
If the requirement for geometrical optics is not met, we cannot describe
the behavior of light by rays but must take its wave nature specifically into
account. This subject is called wave optics; it includes geometrical optics as
an important limiting case. We will treat wave optics in succeeding chapters.

42-2 Spherical Waves-Plane Mirror


Figure 42-4 shows a point source of light 0, the oliject, placed a distance o
in front of a plane mirror. The light falls on the mirror as a spherical wave
represented in the figure by rays emanating from 0. * At the point at which
each ray strikes the mirror we construct a reflected ray. If the reflected rays
are extended backward, they intersect in a point I which is the same distance

Fig. 42-4 A point object O


forms a virtual image I in a
plane mirror. The rays appear
to emanate from I, but actually
light energy does not pass
through this point.

behind the mirror that the object O is in front of it; I is called the image of 0.
Images may be real or virtual. In a real image light energy actually passes
through the image point; in a virtual image the light behaves as though it
diverges from the image point, although, in fact, it does not pass through
this point; see Fig. 42--4. Images in plane mirrors are always virtual. We
know from daily experience how "real" such a virtual image appears to be
and how definite is its location in the space behind the mirror, even though
this space may, in fact, be occupied by a brick wall.
Figure 42-5 shows two rays from Fig. 42_:4. One strikes the mirror at v,
along a perpendicular line. The other strikes it at an arbitrary point a,
making an angle of incidence e with the normal at that point. Elementary
* In our discussion of reflection from mirrors in Chapter 41 (see Fig. 41-8) we assumed
an incident plane wave; the incident rays are parallel to each other in that case.
Sec. 42-2 SPHERICAL WA VES~PLANE MIRROR 1039

Fig. 42-5 Two rays from Fig.


42-4; ray Oa makes an arbi-
trary angle IJ with the normal.

geometry shows that the angles aOv and alv are also equal to fJ. Thus the
right triangles aOva and alva are congruent and
0 = -i, (42-1)
in which the minus sign is arbitrarily introduced to show that I and Oare on
opposite sides of the mirror. Equation 42-1 does not involve e, which means
that all rays striking the mirror pass through I when extended backward, as
we have seen above. Beyond assuming that the mirror is truly plane and
that the conditions for geometrical optics hold, we have made no approxi-
mations in deriving Eq. 42-1. A point object produces a point im?,ge in a
plane mirror, with o = -i, no matter how large the angle () in Fig. 42-5.
Because of the finite diameter of the pupil of the eye, only rays that lie
fairly close together can enter the eye after reflection at a mirror. For the
eye position shown in Fig. 42-6 only a small patch of the mirror near point a
is effective in forming the image; the rest of the mirror may be covered up
or removed. If the eye is moved to another location, a different patch of the
mirror will be effective; the location of the virtual image I will remain un-
changed, however, as long as the object remains fixed.
If the object is an extended source such as the head of a person, a virtual
image is also formed. From Eq. 42-1, every point of the source has an image
point that lies an equal distance directly behind the plane of the mirror.
Thus the image reproduces the object point by point.
Images in plane mirrors differ from objects in that left and right are inter-
changed. The image of a printed page is different from the page itself.
Similarly, if a top is made to spin clockwise, the image, viewed in a vertical
mirror, will seem to spin counterclockwise. Figure 42-7 shows an image of
a left hand, constructed by using point-by-point application of Eq. 42-1; the
image has the symmetry of a right hand.*
* See "The Overthrow of Parity" by Philip Morrison, Scientific American, April, 1957,
for a discussion of the distinction in nature between right and left.
1040 REFLECTION AND REFRACTION-II Chap. 42

Fig. 42-6 A pencil of rays from O enters the eye after reflection at the mirror. Only a
small portion of the mirror near a is effective. The small arcs represent portions of
spherical wavefronts.

Fig. 42-7 A plane mirror reverses right and left.


Sec. 42-3 SPHERICAL WA YES-SPHERICAL MIRROR 1041
... Example 1. How tall must a vertical mirror be if a person 6 ft high is to be able
to see his entire length? Assume that his eyes are 4 in. below the top of his head.
Figure 42-8 shows the paths followed by light rays leaving the top of the man's
head and the tips of his toes. These rays, chosen so that they will enter the eye e
after reflection, strike the vertical mirror at points a and b, respectively. The mirror
need occupy only the region between these two points. Calculation shows that b is
2 ft, 10 in. and a is 5 ft, 10 in. above the floor. The length of the mirror is thus 3 ft, or
half the height of the person. Note that this height is independent of the distance
between the person and the mirror. Mirrors that extend below point b show reflec-
tions of the floor between the person and the mirror.

4
~
Fig. 42-8 Example 1. A person can view his +:
full-length image in a mirror that is only half 3 "So
his height. '<ii
I
2

1
0
~

Example 2. Two plane mirrors are placed at right angles, and a point object
O is located on the perpendicular bisector, as shown in Fig. 42-9a. Locate the images.
Images 11 and 12 are formed in mirrors ab and cd, respectively. There is also a third
image; it may be considered to be the image of 1 1 in mirror cd or the image of 1 2 in
mirror ab. The three images and the object O lie on a circle whose center is on the
line of intersection of the mirrors and whose plane is at right angles to that line.
In viewing 1 3 the light entering the observer's eye is reflected twice after leaving the
source. Figure 42-9b shows a typical bundle of rays. In viewing l 1 or l 2, the light
is reflected only once, as in Fig. 42-6. ..ill

42-3 Spherical Waves-Spherical Mirror


In Fig. 42-10 a spherical light wave from a point object O falls on a con-
cave spherical mirror whose radius of curvature is r. * A line through O and
the center of curvature C makes a convenient reference axis.
A ray from O that makes an arbitrary angle a with this axis intersects the
axis at I after reflection from the mirror at a. A ray that leaves O along the
axis will be reflected back along itself at v and will also pass through I. Thus,
for these two rays at least, I is the image of O; it is a real image because light
energy actually passes through I. Let us find the location of I.
A useful theore.:n is that the exterior angle of a triangle is equal to the sum
of the two opposite interior angles. Applying this to triangles OaCO and

* A spherical shell, viewed from inside, is everywhere concave; viewed from outside it is
everywhere convex. In this chapter concave and convex will always be judged from the
point of view of an observer sighting along the direction of the incident light.
1042 REFLECTION AND REFRACTION~n Chap. 42

Fig. 42-9 Example 2. (a) Ob-


ject O has three virtual images.
(b) A typical bundle of rays
used to view I 3 .

Fig. 42-10 Two rays from O


converge after reflection in a
spherical concave mirror, form-
ing a real image at /.
Sec. 42-3 SPHERICAL WA YES-SPHERICAL MIRROR 1043

Oa!O in Fig. 42-10 yields


!J=a+8
and 'Y = a+ 28.
Eliminating 8 between these equations leads to
a+ 'Y = 2/). (42-2)
In radian measure we can write angles a, /J, and 'Y as
av av
a~-=-
-vO o
av av
!J=-=- (42-3)
vC r
av av
'Y "--' - = -·
vl i
Note that only the equation for /J is exact, for the reason that the center of
curvature of arc av is at C and not at O or I. However, the equations for
a and for 'Y are approximately correct if these angles are sufficiently small.
In all that follows we assume that the rays diverging from the object make only
a small angle a with the axis of the mirror. Such rays, which lie close to the
mirror axis, are called paraxial rays. We did not find it necessary to make
such an assumption for plane mirrors. Substituting these equations into
Eq. 42-2 and canceling av yields
1 1 2
-+---: = -, (42-4)
o i r
in which o is the object distance and i is the image distance. Both these dis-
tances are measured from the vertex of the mirror, which is the point v at
which the axis intercepts the mirror.
Significantl~, Eq. 42-4 does not contain a (or /J, 'Y, or 8), so that it holds
for all rays that strike the mirror provided that they are sufficiently paraxial.
In an actual case the rays can be made as paraxial as one likes by putting a
circular diaphragm in front of the mirror, centered about the vertex v; this
will impose a certain maximum value of a.
As a in Fig. 42-10 is permitted to become larger, it will become less true
that a point object will form a point image; the image will become extended
and fuzzy. ~ o sharp criterion for deciding whether a given ray is paraxial
can be laid down. If the maximum permitted value of a is reduced, the rays
will become more paraxial and the image will become shatper. Unfortu-
nately, the image will also become fainter because less total light energy
will be reflected from the mirror. A compromise must often be made be-
tween image brightness and image quality.
As for plane mirrors, the image (real or virtual) in a spherical mirror can
be seen only if the eye is located so that light rays from the object can enter
1044 REFLECTION AND REFRACTION-II Chap. 42

Fig. 42-11 The eye must be


properly located to see image I.

it after reflection. In Fig. 42-11 a bundle of light rays is shown entering


the eye in position x; only the small patch of mirror near a is effective for
this eye position. If the observer moves his eye to position y, the image
will vanish for him because the mirror does not exist near point a'.
Although Eq. 42-4 was derived for the special case in which the object is
located beyond the center of curvature, it is generally true, no matter where
the object is located. It is also true for convex mirrors, as in Fig. 42-12.
In applying Eq. 42-4, we must be careful to follow a consistent convention
of signs for o, i, and r. As the basis for the sign conventions to be used in
this book, we start from this statement:
In Fig. 42-10, in which light diverges from a real object, falls on
a concave mirror, and converges after reflection to form a real image, the
quantities o, i, and r in Eq. 42-4 are given positive numerical values.
Figure 42-10 was used to derive Eq. 42-4 and the student should associate
them in his mind as an aid in getting the signs correct.
Let us fix our minds on the side of the mirror from which incident light
comes. Because mirrors are opaque, the light, after reflection, must remain
on this side, and if an image is formed here it will be a real image. There-

Fig. 42-12 Two rays from O


diverge after reflection in a

-----
spherical convex mirror, form-
ing a virtual image at I, the
point from which they appear
to originate. Compare Fig.
42-10.
SPHERICAL WAYES-SPHERICAL MIRROR 1045

fore, we call the side of the mirror from which the light comes the R-side (for
real image). We call the back of the mirror the V-side (for virtual image),
because images formed on this side of the mirror must be virtual, no light
energy being present on this side.*
In the indented statement above we associated real images with positive
image distances. This suggests our first sign convention:
1. The image distance i is positive if the image (real) lies on the R-side
of the mirror, as in Fig. 42-10; i is negative if the image (virtual) lies on the
V-side, as in Fig. 42-12.
If the mirror in Fig. 42-10, which is concave as viewed from the direction
of the incident light, is made convex, the rays will diverge after reflection
and will form a virtual image, as Fig. 42-12 shows. Thus the indented
statement above suggests our second sign convention:
2. The radius of curvature r is positive if the center of curvature of the
mirror lies on the R-side, as in Fig. 42-10; r is negative if the center of curva-
ture lies on the V-side, as in Fig. 42-12.
The student should not commit these sign conventions to memory but
should deduce them in each case from the basic statement on p. 1044, using
Fig. 42-10 as a mnemonic aid.
For all cases in this book the object distance o is to be taken as positive. In
systems of two or more mirrors (or combinations of mirrors and refracting surfaces)
it is possible to arrange that converging light falls on the mirror. In such cases the
object is called virtual and the object distance o is negative; we limit our discussion
here to real objects .
..,.. Example 3. A convex mirror has a radius of curvature of 20 cm. If a point
source is placed 14 cm away from the mirror, as in Fig. 42-12, where is the image?
A rough graphical construction, applying the law of reflection at a in the figure,
shows that the image will be on the V-side of the mirror and thus will be virtual.
We may verify this quantitatively and analytically from Eq. 42-4, noting that r is
negative here because the center of curvature of the mirror is on its V-side. We have
1 1 2
-+-:=-
O i r
1 1 2
or
+14 cm +i = -20 cm'
which yields i = - 5.8 cm, in agreement with the graphical prediction. The nega-
tive sign for i reminds us that the image is on the V-side of the mirror and thus is
virtual. ....
When parallel light falls on a mirror (Fig. 42-13), the image point (real or
virtual) is called the focal point F of the mirror. The focal length f is the
distance between F and the vertex. If we put o ---,. oc:i in Eq. 42-4, thus
insuring parallel incident light, we have
i. = 1
2r =
f.
* This nomenclature may seem unnecessarily cumbersome at this point. We adopt it
because of the later extension of these ideas to refracting surfaces and lenses.
Fig. 42-13 (a) The focal point for a concave spherical mirror, showing both the rays
and the wavefronts. F and C lie on the R-side, the focal point is real, and the focal
length f of the mirror is positive (as is r). (b) The same, except that the incident light
makes an angle a with the mirror axis; the rays are focused at a point in the focal plane
F'F". (c) Same as (a) except that the mirror is convex; F and C lie on the V-side of the
mirror. The focal point is virtual and the focal length! is negative (as is r).
1046
Sec. 42-3 SPHEHICAL WA VES-SPHEHICAL MIRROR 1047

Equation 42-4 can then be rewritten

-1 + -1 = -,1 (42-5)
O i j
where f, liker, is taken as positive for mirrors whose centers of curvature are
on the R-sidc (that is, for concave, or converging mirrors; see Fig. 42-l:3a)
and negative for those whose centers of curvature arc on the V-sidc (that is,
for convex, or diverging mirrors; see Fig. 42-13c). Figure 42-l:3b shows an
incident plane ,vave that makes a small angle a with the mirror axis. The
rays are focused at a point in the focal plane of the mirror. This is a plane
at right angles to the mirror axis at the focal point.

Fig. 42-14 The image of an extended object in (a) a concave mirror and (b) a convex
mirror is located graphically. Any two of the three special rays shown are sufficient.

We now consider objects that are not points. Figure 42-14 shows a candle
in front of (a) a concave mirror and (b) a convex mirror. We choose to draw
the mirror axis through the foot of the candle and, of course, through the
center of curvature. The image of any off-axis point, such as the tip of the
candle, can be found graphically, using the following facts:

l. A ray that strikes the mirror after passing (either directly or upon being
extended) through the center of curvature C returns along itself (ray x in
Fig. 42-1:3). Such rays strike the mirror at right angles.
2. A ray that strikes the mirror parallel to its axis passes (or will pass when
extended) through the focal point (ray y).
1048 REFLECTION AND REFRACTIO:',l" ~u Chap. 42
3. A ray that strikes the mirror after passing (either directly or upon being
extended) through the focal point emerges parallel to the axis (ray z).

Figure 42-1.5 shows a ray (dve) that originates on the tip of the object
candle of Fig. 42-14a, is reflected from the mirror at point v, and passes
through the tip of the image candle. The law of reflection demands that
this ray make equal angles e with the mirror axis as shown. For the two
similar right triangles in the figure we can write

ce vc
bd vb

The quantity on the left (apart from a question of sign) is the lateral mag-
nification m of the mirror. Since we want to represent an inverted image by
a negative magnification, we arbitrarily define m for this case as - (ce/bd).
Since vc = i and vb = o, we have at once

m= (42-6)
0

This equation gives the magnification for spherical and plane mirrors under
all circumstances. For a plane mirror, o = - i and the predicted magnifica-
+
tion is 1 which, in agreement with experience, indicates an erect image the
same size as the object.
Images in spherical mirrors suffer from several "defects" that arise because the
assumption of paraxial rays is never completely justified. In general, a point source
will not produce a point image; see Problem 5. Apart from this, distortion arises
because the magnification varies somewhat with distance from the mirror axis, Eq.
42-6 being strictly correct only for paraxial rays. Superimposed on these defects are
diffraction effects which come about because the basic assumption of geometrical optics,
that light travels in straight lines, must always be considered an approximation.

Fig. 42-15 A particular ray for the arrangement of Fig. 42-14, used to show that the
lateral magnification mis given by -i/o.
Sec. 42-4 SPHERICAL REFRACTING SURFACE 1049
i/f

I \
3~-2.ox .
\ Concave mirror,

2
1
real invert d image

~ Ix ~
Convex mirror,
virtual erect image +---1
x-x
+0.25
I + 0·3 - ;+o.5o
--x~
I x +LO
I ) '
o/f
-3 -2 -1 0 1 2 3
Plane mirror

-1~y2.a
\ I

-2-x
!.+!.=!. +3.0 \ Concave mirror,
O i f virtual erect image
-3 '----X
+4.0

Fig. 42-16 A graphical representation of the mirror formula (Eq. 42-·5). The numbers
on the curves give the lateral magnifications (Eq. 42-6).

Figure 42-16 summarizes the predictions of the mirror equation (Eq. 42-5)
and of the lateral magnification equation (Eq. 42-6). For convex mirrors
the image is always virtual. For concave mirrors the image is virtual if the
object is inside the focal point (o/f < 1) and real otherwise (o/f > 1). As
o ---t OC:J for both types of mirror, the image approaches the focal point
(i/f ---t 1). A plane mirror (m = +1) is represented in the figure by a
point at the origin. It is appropriate that this point form the central point
of the lower-left branch of Fig. 42-16 because both convex and concave
mirrors approach plane mirrors as r is increased in magnitude, that is, as
f ---tOC) (or as o/f - 0 in Fig: 42-16, assuming a fixed object distance).

42-4 Spherical Refracting Surface


Figure 42-17 shows a point source O near a convex spherical refracting
surface of radius of curvature r. The surface separates two media whose
indices of refraction differ, that of the medium in which the incident light falls
on the surface being n 1 and that on the other side of the surface being n 2 .
From Owe draw a line through the center of curvature C of the refracting
surface, thus establishing a convenient axis which intercepts the surface at
vertex v. From O we draw a ray that makes a small but arbitrary angle a
with the axis and strikes the refracting surface at a, being refracted according
to
1050 REFLECTIO~ A~D REFRACTION-II Chap. 42

Fig. 42-17 Two rays from O converge after refraction at a spherical surface, forming a
real image at I.

The refracted ray intersects the axis at I. A ray from O that travels along
the axis will not be bent on entering the surface and will also pass through I.
Thus, for these two rays at least, I is the image of 0.
As in the derivation of the mirror equation, we use the theorem that the
exterior angle of a triangle is equal to the sum of the two opposite interior
angles. Applying this to triangles COaC and !Cal yields

81 = a+ /3
and (42-8)

As a is made small, angles /3, 'Y, 81 , and 82 in Fig. 42-17 also become small.
lVe at once assume that a, hence all these angles, are arbitrarily small, this
paraxial ray assumption was also made for spherical mirrors. Replacing
the sines of the angles by the angles themselves-since the angles are re-
quired to be small-permits us to write the law of refraction as

(42-9)

Combining Eqs. 42-8 and 42-9 leads to

Eliminating 81 between this equation and Eq. 42-7 leads, after rearrange-
ment, to
(42-10)
Sec. 42-4 SPHEHICAL REFRACTING SURFACE 1051

In radian measure the angles a, (3, and 'Y in Fig. 42-17 are
av
0

av
(3 = - (42-11)
r
av
'Y r-..; - : - •

Only the second of these equations is exact. The other two are approximate
because I and O arc not the centers of circles of which av is an arc. However,
for paraxial rays (a small enough) the inaccuracies in Eq. 42-11 can be made
as small as desired.
Substituting Eqs. 42-11 into Eq. 42-10 leads readily to
n1 n2 n2 - n1
-+-:-=--- (42-12)
O i r
This equation holds whenever light is refracted from point objects at spheri-
cal surfaces, assuming only that the rays are paraxial. As with the mirror
formula, care must be taken to use Eq. 42-12 with consistent signs for o, i,
and r. Once again we establish our sign conventions by physical reasoning
from a particular case, that of Fig. 42-17:
In Fig. 42-17, in which light diverges from a real object, falls on a
convex refracting surface, and converges after refraction to form a real
image, the quantities o, i, and r in Eq. 42-12 have positive numerical
values.
Figure 42-17 was used to derive Eq. 42-12, and the student should asso-
ciate them in his mind as an aid in getting the signs correct. This basic
statement is quite similar to that which we made for mirrors on p. 1044.
We fix our attention on the side of the refracting surface from which the
incident light falls on the surface. In contrast to mirrors, the light energy
passes through a refracting surface to the other side, and if a real image is
formed it must appear on the far side, which we call the R-side. The side
from which the incident light comes is called the V-side because virtual
images must appear here. Figure 42-18 suggests this important distinction
between reflection and refraction.
In the indented statement above we associated real images with positive-
image distances. Thus we are led to the sign convention:

1. The image distance i is positive if the image (real) is on the R-side of


the refracting surface, as in Fig. 42-17; i is negative if the image (virtual)
lies on the V-side, as in Fig. 42-19.
The refracting surface in Fig. 42-17 is convex. If it is made concave
(still assuming that n 2 > n 1 ), the rays will diverge after refraction and form
1052 REFLECTION AND REFRACTION-II Chap. 42
R-side V-side
Incident light

Reflected light
Fig. 42-18 Real images are formed on the same
(concave or side as the incident light for mirrors but on the
convex mirror)
opposite side for refracting surfaces and thin
V-side I R-side lenses. This is so because the incident light is
Incident light n1 I n2 reflected back by mirrors but is transmitted
I through by refracting surfaces.
I
I Refracted light
I
(concave or convex
refracting surface
or thin lens)

a virtual image, as Fig. 42-19 shows. Thus we are led to our second sign
convention:
2. The radius of curvature r is positive if the center of curvature of the
refracting surface lies on the R-side, as in Fig. 42-17; r is negative if the
center of curvature is on the V-side, as in Fig. 42-19.
The sign conventions for refracting surfaces are the same as for mirrors
(p. 1045), the fundamental difference between the two situations being
absorbed in the definitions of R-side and V-side in Fig. 42-18. This differ-
ence is easily remembered on physical grounds.
For all cases in this book the object distance o is to be taken as positive. In
systems of two or more refracting surfaces (or combinations of refracting surfaces
and mirrors) it is possible to arrange that converging light falls on the refracting
surface. In such cases the object is called virtual and the object distance o is negative;
we limit our discussions here to real objects.

Fig. 42-19 Two rays from O diverge after refraction at a spherical surface, forming a
virtual image at I.
Sec. 42-4 SPHERICAL REFRACTING SURFACE 1053
~ Example 4. Locate the image for the geometry shown in Fig. 42-17, assuming
the radius of curvature to be 10 cm, n2 to be 2.0, and n 1 to be 1.0. Let the object be
20 cm to the left of v.
From Eq. 42-12,

we have
1.0 + 2.0 = 2.0 - 1.0_
+20 cm i +10 cm
Note that r is positive because the center of curvature of the surface lies on the R-side.
This relation yields i = +40 cm in agreement with the graphical construction. The
light energy actually passes through I so that the image is real, as indicated by the
positive sign for i.
Example 5. An object is immersed in a medium with n 1 = 2.0, being 15 cm from
the spherical surface whose radius of curvature is -10 cm, as in Fig. 42-20; r is
negative because C lies on the V-side. Locate the image.

Fig. 42-20 Two rays from O


appear to originate from I
(virtual image) after refraction
at a spherical surface.

Figure 42-20 shows a ray traced through the surface by applying the law of refrac-
tion at point a. A second ray from O along the axis emerges undeflected at v. The
image I is found by extending these two rays backward; it is virtual.
From Eq. 42-12,

we have 2.0 + _!_:2 = 1.0 - 2.0'


+15 cm i -10 cm
which yields i = -30 cm, in agreement with Fig. 42-20 and with the sign conven-
tions. Note that n 1 always refers to the medium on the side of the surface from
which the light comes.
Example 6. What is the relationship between i and o if the refracting surface is
plane?
A plane surface has an infinite radius of curvature. Puttmg r - rI.J in Eq. 42-12
leads to
1054 REFLECTION AND REFRACTION-II Chap. 42

~.
Fig. 42-21 Refraction at a
plane surface at near-normal
incidence, showing a pencil of
rays and the corresponding
wavefronts entering the pupil.
(a) Source in air and (b) source
in water.
Sec. 42-5 THIN LENSES 1055
Figure 42-21 illustrates the situation graphically (a) for an object in air as seen from
below water and (b) for an object in water with air above. This shows that a diver,
looking upward at, say, an overhanging tree branch, will think it higher than it is by
the factor 1.33/1.00. Similarly, an observer in air will think that objects on the
bottom of a water tank are closer to the surface than they actually are, in the ratio
1.00/1.33. These considerations, being based on Eq. 42-12, hold only for paraxial
rays, which means that the incident rays can make only a small angle with the normal;
this angle has been exaggerated in the figure for clarity .. Note again that n1 is always
identified with the medium that lies on the side of the surface containing the incident
light. .....

42-5 Thin Lenses


In most refraction situations there is more than one refracting surface.
This is true even for a spectacle lens, the light passing from air into glass
and then from glass into air. In microscopes, telescopes, cameras, etc.,
there are often many more than two surfaces.
Figure 42-22a shows a thick glass "lens" of length l whose surfaces are
ground to radii r' and r". A point object O' is placed near the left surface as
shown. A ray leaving 0' along the axis is not deflected on entering or leav-
ing the lens because it falls on each surface along a normal.
A second ray leaving 0', at an arbitrary angle a with the axis, strikes the
surface at point a', is refracted, and strikes the second surface at point a".
The ray is again refracted and crosses the axis at I", which, being the inter-
section of two rays from 0", is the image of point O', formed after refraction
at two surfaces.
Figure 42-22b shows the first surface, which forms a virtual image of O' at
I'. To locate I', we use Eq. 42-12,
n1 n2 n2 - n1
-+-:-=-~
o i r
Putting n 1 = 1.0 and n 2 = n and bearing in mind that the image distance
is negative (that is, i = -i' in Fig. 42-22b), we obtain
1 n n - 1
(42-13)
o' i' r'
In this equation i' will be a positive number because we have arbitrarily
introduced the minus sign appropriate to a virtual image.
Figure 42-22c shows the second surface. Unless an observer at point a"
were aware of the existence of the first surface, he would think that the light
striking that point originated at point I' in Fig. 42-22b and that the region
to the left of the surface was filled with glass. Thus the (virtual) image I'
formed by the first surface serves as a real object 0" for the second surface.
The distance of this object from the second surface is
0 11 = i' + l. (42-14)
In applying Eq. 42-12 to the second surface, we insert n 1 = n and n 2 =
1.0 because the object behaves as if it were imbedded in glass. If we use
1056 REFLECTION AND REFRACTION-II Chap. 42

Fig. 42-22 (a) Two rays from 0' intersect at I" (real image) after refraction at two spheri-
cal surfaces. (b) The first surface and (c) the second surface shown separately. The
quantities a and n have been exaggerated for clarity.

Eq. 42-14, Eq. 42-12 becomes


n 1 1- n
~-+-=~-·
i' + l i" r"
(42-15)

Let us now assume that the thickness l of the "lens" in Fig. 42-22 is so
small that it can be neglected in comparison with other linear quantities in
Sec. 42-5 THIN LENSES 1057

this figure (such as o', i', o", i", r', and r"). In all that follows we make this
thin-lens approximation. Putting l = 0 in Eq. 42-15 leads to
n 1 n - 1
-+-= - - - · (42-16)
i' i" r"
Adding Eqs. 42-13 and 42-16 leads to

~ +~
o' i"
= (n - 1) (~ -
r'
2-) ·
r"
Finally, calling the original object distance simply o and the final image
distance simply i leads to

~ +~
o i
= (n - 1) (~ -
r'
2-) ·
r"
(42-17)

This equation holds only for paraxial rays and only if the lens is so thin that
it essentially makes no difference from which surface of the lens the quanti-
ties o and i are measured. In Eq. 42-17 r' refers to the first surface struck
by the light as it traverses the lens and r" to the second surface.
The sign conventions for Eq. 42-17 are the same as those for mirrors and
for single refracting surfaces. Because the lens is assumed to be thin, we
ref er to the R-side and the V-side of the lens itself (see Fig. 42-18) rather
than those of its separate surfaces. The sign conventions then are the fol-
lowing:
1. The image distance i is positive if the image (real) lies on the R-side
of the lens, as in Fig. 42-23a; i is negative if the image (virtual) lies on the
V-side of the lens, as in Fig. 42-23b.
2. The radii of curvature r' and r" are positive if their respective centers

Fig. 42-23 Illustrations used to establish sign conventions for thin lenses.
1058 REFLECTION AND REFRACTION~II Chap. 42
of curvature lie on the R-side of the lens. They are negative if their centers
lie on the V-side. In Fig. 42-23a, r' is positive and r" is negative; in Fig.
42-23b, r' is negative and r" is positive.

Figure 42-24a and c shows parallel light from a distant object falling on
a thin lens. The image location is called the second focal point F 2 of the lens.
The distance from F 2 to the lens is called the focal length f. The first focal
point for a thin lens (F 1 in figure) is the object position for which the image

Fig. 42-24 (a) Parallel light passes through the second focal poiut F2 of a converging
lens. (b) The incident light makes an angle a with the lens axis, the rays being focused
in the focal plane F'F". (c) Parallel light, passing through a diverging lens, seems to
originate at the second focal point F 2• C' and C" are centers of curvature for the lens
surfaces; F 1 is the first focal point.
Sec. 42-5 THIN LENSES 1059

is at infinity. For thin lenses the first and second focal points are on opposite
sides of the lens and are equidistant from it.
The focal length can be computed from Eq. 42-17 by inserting o ------} oo
and i = f. This yields

! = (n - 1) (~ - --1:_) · (42-18)
f r' r"
This relation is called the lens maker's equation because it allows us to com-
pute the focal length of a lens in terms of the radii of curvature and the index
of refraction of the material. Combining Eqs. 42-17 and 42-18 allows us to
write the thin-lens equation as
1 1 1
- +- = -· (42-19)
O i f
Figure 42-24b shows parallel incident rays that make a small angle a with
the lens axis; they are brought to a focus in the focal plane F' F", as shown.
This is a plane normal to the lens axis at the focal point.
In Fig. 42-24a we note that all rays in the figure contain the same number
of waveler{gths; in other words, they have the same optical path lengths; see
Section 41-6. The optical path lengths are the same because the wavefronts
are surfaces over which the wave disturbance has the same constant value
and because all the rays shown pass through the same number of wavefronts.
IJI,, Example 7. The lenses of Fig. 42-24 have radii of curvature of magnitude 40 cm
and are made of glass with n = 1.65. Compute their focal lengths.
Since C' lies on the R-side of the lens in Fig. 42-24a, r' is positive ( = +40 cm).
Since C" lies on the V-side, r" is negative ( = -40 cm). Substituting in Eq. 42-18
yields

-1 =
f
(n - 1) (1- - -
r'
1)
r"
= ( 1.65 - 1) (+401cm - ~- 1
-40 cm
) ,

or f = +31 cm.
A positive focal length indicates that in agreement with Fig. 42-24a the focal point
F 2 is on the R-side of the lens and parallel incident light converges after refraction
to form a real image.
In Fig. 42-24c C' lies on the V-side of the lens so that r' is negative ( = -40 cm).
Since r" is positive ( = +40 cm), Eq. 42-17 yields
f = -31 cm.
A negative focal length indicates that in agreement with Fig. 42-24c the focal point
F 2 is on the V-side of the lens and incident light diverges after refraction to form a
virtual image. ...i
The location of the image of an extended object such as a candle (Fig. 42-
25) can be found graphically by using the following three facts:
1. A ray parallel to the axis and falling on the lens passes, either directly
or when extended, through the second focal point (ray x in Fig. 42-25).
2. A ray falling on a lens after passing, either directly or when extended,
1060 REFLECTION AND REFRAC'l'JON~II Chap. 42
through the first focal point will emerge from the lens parallel to the axis
(ray y).
3. A ray falling on the lens at its center will pass through undeflected.
There is no deflection because the lens, near its center, behaves like a thin
piece of glass with parallel sides. The direction of the light rays is not
changed and the sideways displacement can be neglected because the lens
thickness has been assumed to be negligible (ray z; see also Problem 13,
Chapter 41).
Figure 42-26, which represents part of Fig. 42-25a, shows a ray passing
from the tip of the object through the center of curvature to the tip of the
image. For the similar triangles abe and dee we may write
de de
ab ae
The right side of this equation is i/ o and the left side is -m, where mis the

Fig. 42-25 Showing the graphical location of images for three thin lenses.
Sec. 42-5 THIN LENSES 1061

Fig. 42-26 Two rays for the situation of Fig. 42-25a.

lateral magnification. The minus sign is required because we wish m to be


negative for an inverted image. This yields
i
m = - -,
0

which holds for all types of thin lenses and for all object distances .
.... Example 8. A converging thin lens has a focal length of +24 cm. An object is
placed 9.0 cm from the lens as in Fig. 42-25b; describe the image.
From Eq. 42-19,
1 1 1
;+i=r
1 1 1
we have -+-9-.0-cm-+ i = +24 cm'
which yields i = -14.4 cm, in agreement with the figure. The minus sign means that
the image is on the V-side of the lens and is thus virtual.
The lateral magnification is given by
i -14.4cm
m = - - = - - - - - = +1.6,
o +9.0 cm
again in agreement with the figure. The plus signifies an erect image.
Images formed by lenses suffer from defects similar to those discussed for
mirrors on p. 1048. There are effects connected with the failure of a point
object to form a point image, with the variation of magnification with dis-
tance from the lens axis, and with diffraction. For lenses, but not for mir-
rors, there are also chromatic aberrations associated with the fact that the
refracting properties of the lens vary with wavelength because the index of
refraction of the lens material does. If a point object on the lens axis emits
white light, the image, neglecting other lens defects, will be a series of colored
points spread out along the axis. We have all seen the colored images pro-
duced by inexpensive lenses. A great deal of ingenious optical engineering
goes into the design of lenses (more commonly lens systems) in which the
various lens defects are minimized. The lens surfaces are normally not
spherical.
1062 REFLECTION AND REFRACTION-II Chap. 42
i/f

1 _\
3--2.ox
\ Converging lens,
real inverted image
2 -1.0 x'-..._I
Diverging lens, -o;o-
virtual erect image - - 1
x-'
+0.25 +OJj"""--x~
I I +0.50
x
+1.0
o/f

\f
-3 -2 -1 0 2 3

-1 -y1.o
1 1 1
-2 +30, Converging lens,
-+-
O i
=-f virtual erect image
3 +4.0x

Fig. 42-27 A graphical representation of the thin-lens formula (Eq. 42-19). The num-
bers on the curves give the lateral magnifications (Eq. 42-20). Compare Fig. 42-16.

Figure 42-27 is a graphical representation of the object-image relation-


ships for thin lenses; see Eqs. 42-19 and 42-20. The fact that it is identical
in form with Fig. 42-16, which holds for mirrors, is not surprising because
the basic equations for mirrors and thin lenses (Eqs. 42-19 and 42-5; Eqs.
42-20 and 42-6) are identical. A converging lens of a given focal length is
quantitatively similar to a concave mirror of the same focal length, the dif-
ference being that the lens forms real images on the opposite side from the
(real) object and the mirror forms them on the same side.

QUESTIONS

1. If a mirror reverses right and left, why doesn't it reverse up and down?
2. Is it possible to photograph a virtual image?
3. What approximations were made in deriving the mirror equation (Eq. 42-4):
1 1 2
-+-:-=-?
o i r
4. Can you think of a simple test or observation to prove that the law of reflection is
the same for all wavelengths, under conditions in which geometrical optics prevails?
5. Under what conditions will a spherical mirror, which may be concave or convex, form
(a) a real image, (b) an inverted image, and (c) an image smaller than the object?
6. An unsymmetrical thin lens forms an image of a point object on its axis. Is the image
location changed if the lens is reversed?
7. Why has a lens two focal points and a mirror only one?
Chap. 42 PROBLEMS 1063

8. Under what conditions will a thin lens, which may be converging or diverging, form
(a) a real image, (b) an inverted image, and (c) an image smaller than the object?
9. Does the apparent depth of an object below water depend on the angle of view of the
observer in air? Explain and illustrate with ray diagrams.
10. A skin diver wants to use an air-filled plastic bag as a converging lens for underwater
use. Sketch a suitable cross section for the bag.
11. What approximations were made in deriving the thin lens equation (Eq. 42-19):
1 1 1
-+-=-?
O i f
12. Under what conditions will a thin lens have a lateral magnification (a) of -1 and (b)
of +1?
13. How does the focal length of a glass lens for blue light compare with that for red
light, assuming the lens is (a) diverging and (b) converging?
14. Does the focal length of a lens depend on the medium in which the lens is immersed?
Is it possible for a given lens to act as a converging lens in one medium and a diverging
lens in another medium?
15. Are the following statements true for a glass lens in air? (a) A lens that is thicker
at the center than at the edges is a converging lens for parallel light. (b) A lens that is
thicker at the edges than at the center is a diverging lens for parallel light. Explain and
illustrate, using wavefronts.
16. Under what conditions would the lateral magnification, (m = -i/o) for lenses and
mirrors become infinite? Is there any practical significance to such a condition?
17. Light rays are reversible. Discuss the situation in terms of objects and images if
all rays in Figs. 42-10, 42-14, 42-17, 42-19, 42-23, and 42-25 are reversed in direction.
18. What significance can be given to the origin of coordinates in the graphical repre-
sentation of the thin-lens formula (Fig. 42-27)?
19. In connection with Fig. 42-24a, we pointed out that all rays originating on the same
wavefront in the incident wave have the same optical path length to the image point.
Discuss this in connection with Fermat's principle (Section 41-6).

PROBLEMS

1. Solve Example 2 if the angle between the mirrors is (a) 45°, (b) 60°, (c) 120°, the object
always being placed on the bisector of the mirrors.
2. Two plane mirrors make an angle of 90° with each other. What is the largest
number of images of an object placed between them that can be seen by a properly placed
eye? The object need not lie on the mirror bisector.
3. A small object is 10 cm in front of a plane mirror. If you stand behind the object,
30 cm from the mirror, and look at its image, for what distance must you focus your eyes?
4. A small object O is placed one-third of the way between two parallel plane mirrors
as in Fig. 42-28. Trace appropriate bundles of rays for viewing the four images that lie
closest to the object.

Fig. 42-28
1064 REFLECTION AND REFRAC'I'ION--II Chap. 42
5. Redraw Fig. 42-29 on a large sheet of paper and trace carefully the reflected rays,
using the law of reflection. Is a point focus formed? Discuss.

Fig. 42-29

6. Fill in this table, each column of which refers to a spherical mirror. Check your
results by graphical analysis. Distances are in centimeters; if a number has no plus or
minus sign in front of it, it may have either sign.

a b c d e f g h

Type concave convex


------- -------

f 20 +20 20
----

r -40 40

i -10 4
---

0 +10 +10 +30 +60 +24


----------

m +1 -0.5 +0.10 0.50


------

Real image? no
---

Erect image? no

7. A short linear object of length l lies on the axis of a spherical mirror, a distance o
from the mirror. (a) Show that its image will have a length l' where

l' = l (_f-)
0 -f
2
-

(b) Show that the longitudinal magnification m'( = l' /l) is equal to m 2 where mis the lateral
magnification discussed in Section 42-3. (c) Is there any condition such that, neglecting
all mirror defects, the image of a small cube would also be a cube?
8. As an example of the effect of relaxing the paraxial ray assumption find, analytically,
Chap. 42 PROBLEMS 1065
the apparent depth of a swimming pool having water (n = 1.33) of actual depth 8.0 ft (a)
when viewed at normal incidence and (b) when viewed at an angle of 30° from the normal.
9. A layer of water (n = 1.33) 2.0 cm thick floats on carbon tetrachloride (n = 1.46)
4.0 cm thick. How far below the water surface, viewed at normal incidence, does the
bottom of the tank seem to be?
~O. Fill out the following table, each column of which refers to a spherical surface
separating two media with different indices of refraction. Distances are measured m
centimeters.
a b c d e f g h

n1 1.0 1.0 1.0 1.0 1.5 1.5 1.5 1.5

n2 1.5 1.5 1.5 1.0 1.0 1.0

0 +10 +10 +20 +10 +10 +100

i -13 +600 -20 -6 -7.5 +600

r +30 +30 -20 -30 +30 -30

Real image?

Draw a figure for each situation and construct the appropriate rays graphically. Assume
a point object.
···11-... fill in this table, each column of which refers to a thin lens, to the extent possible.
Check .your results by graphical analysis. Distances are in centimeters; if a number
(except in row n) has no plus sign or minus sign in front of it, it may have either sign.

a b c d e f g h i

Type converging
------------ --- ------

f 10 +10 10 10
--------- ------------

r' +30 -30 -30


------ --- --- ---------
r" -30 +30 -60
------ ---------------

i
------ --- ------------ ---

0 +20 +5 +5 +5 +10 +10 +10 +10 +10


--- --- ------------ ---
n 1.5 1.5 1.5
--- ------------ --- ---

m >1 <1 0.5 0.5


---------------------

Real image? yes


------------ --- ------

Erect image? yes

Draw a figure for each situation and construct the appropriate rays graphically. Assume
a finite object.
1066 REFLECTION AND REFRACTION-II Chap. 42
12. Define and locate the first and second focal points (see p.1058) for a single spherical
refracting surface such as that of Fig. 42-17.
13. A luminous object and a screen are a fixed distance D apart. (a) Show that a con-
verging lens of focal length f will
form a real image on the screen for
two positions that are separated by

d = VD(D - 4f).
(b) Show that the ratio of the two
image sizes for these two positions
is

14. Using the lens maker's equa-


tion (Eq. 42-18), decide which of
the thin lenses in Fig. 42-30 is con-
Fig. 42-30 verging and which diverging for
parallel incident light.
15. Two thin lenses of focal length Ji andf2 are in contact. Show that they are equiva-
lent to a single thin lens with a focal length given by

f = hh .
h+h
16. Show that the distance between an object and its real image formed by a thin
converging lens is always greater than four times the focal length of the lens.
17. A double-convex lens is to be made of glass with an index of refraction of 1.50.
One surface is to have twice the radius of curvature of the other and the focal length is
to be 6.0 cm. What are the radii?
18. The formula
1 1 1
-+-=-
O i f
is called the Gaussian form of the thin lens formula. Another form of this formuia, the
Newtonian form, is obtained by considering the distance x from the object to the first focal
point and the distance x' from the second focal point to the image. Show that
xx' = f 2•
19. A parallel incident beam falls on a solid glass sphere at normal incidence. Locate
the image in terms of the index of refraction n and the sphere radius r.
20. An erect object is placed a distance in front of a converging lens equal to twice the
focal length Ji of the lens. On the other side of the lens is a converging mirror of focal
length h separated from the lens by a distance 2(!1 + f2). (a) Find the location, nature
and relative size of the final image. (b) Draw the appropriate ray diagram. See Fig.
42-31.

Fig. 42-31
Chap. 42 PROBLEMS 1067

21. An object is placed at a center of curvature of a double-concave lens, both of


whose radii of curvature are the same. (a) According to the convention used in this book,
what are the signs of the two radii of curvature? (b) Find the location of the image in
terms of the radius of curvature r and the index of refraction, n, of the glass. (c) Describe
the nature of the image. (d) Verify your result with a ray diagram.
22. Two lenses with their focal points are shown in Fig. 42~32. With the given
object, find the image formed by lens 1 and, using this as an object for lens 2, find the
final image. Discuss the nature of the two images. Yerify your conclusions by copying
the fig;ure to scale and drawing a ray diagram.

+ '
Lens 1

It
Ii
Object

Fig. 42-32

23. An object is 20 cm to the left of a lens with a focal length of + 10 cm. A second lens
of focal length +12.5 cm is 30 cm to the right of the first lens. (a) Using the image formed
by the first lens as the object for the second, find the location and relative size of the final
image. (b) Verify your conclusions by drawing the lens system to scale and constructing
a ray diagram. (c) Describe the final image.
Interference
CHAPTER 43

43-1 Young's Experiment


In Section 19-7 we saw that if two waves of the same frequency travel in
approximately the same direction and have a phase difference that remains
constant with time, they may combine so that their energy is not distributed
uniformly in space but is a maximum at certain points and a minimum
(possibly even zero) at others. The demonstration of such interference ef-
fects for light by Thomas Young in 1801 first established the wave theory of
light on a firm experimental basis. Young was able to deduce the wave-
length of light from his experiments, the first measurement of this important
quantity.
Young allowed sunlight to fall on a pinhole 8 0 punched in a screen A in
Fig. 43-1. The emerging light spreads out by diffraction (see Section 42-1)
and falls on pinholes 8 1 and 8 2 punched into screen B. Again diffraction
occurs and two overlapping spherical waves expand into the space to the
right of screen B.
The condition for geometric optics, namely that a » )I. where a is the di-
ameter of the pinholes, is definitely not met in this experiment. The pin-
holes do not cast geometrical shadows but act as sources of expanding
Huygens' wavelets. We are dealing here (and in succeeding chapters) with
wave optics rather than with geometrical optics.
Figure 43-2, taken from an 1803 paper of Young, shows the region between
screens B and C. The blackening represents the minima of the wave dis-
turbance; the white space between represents the maxima. If you hold
the page with your eye close to the left edge and look at a grazing angle
along the figure, you will see that along lines marked by x's there is cancella-
1068
Sec. 43-1 YOUNG'S EXPERIMENT 1069

Fig. 43-1 Showing how Thomas Young produced an interference pattern by allowing
diffracted waves from pinholes S1 and S2 to overlap on screen C.

x x x x

Fig. 43-2 Thomas Young's original drawing showing interference effects in overlapping
waves. Place the eye near the left edge and sight at a grazing angle along the figure.
(From Thomas Young, Phil. Transactions, 1803.)
1070 INTERFERENCE Chap. 43
tion of the wave; between them there is reinforcement. If a screen is placed
anywhere across the superimposed waves, we expect to find alternate bright
and dark spots on it. Figure 43-3 shows a photograph of such interjerence

Fig. 43-3 Interference fringes for monochromatic light, made with an


arrangement like that of Fig. 43~ 1, using long narro-w slits rather than
pinholes.

fringes; in keeping with modern technique, long narrow slits rather than pin-
holes were used in preparing this figure.
Interference is not limited to light waves but is a characteristic of all wave
phenomena. Figure 43-4, for example, shows the interference pattern of

Fig. 43-4 The interference of water waves in a ripple tank. There is destructive inter-
ference along the lines marked "Line of nodes" and constructive interference between
these lines. (Courtesy Physical Science Study Committee.)
Sec. 43-1 YOUNG'S EXPERIMENT 1071

Fig. 43-5 Rays from S 1 and S 2 combine at P. The light falling on screen B has been
taken as parallel. Actually, D » d, the figure being distorted for clarity.

water waves in a shallow ripple tank. The waves are generated by two vibra-
tors that tap the water surface in synchronism, producing two expanding
spherical waves.
Let us analyze Young's experiment quantitatively, assuming that the
incident light consists of a single wavelength only. In Fig. 43-5 P is an
arbitrary point on the screen, a distance r 1 and r 2 from the narrow slits 8 1
and 8 2 , respectively. Let us draw a line from 8 2 to b in such a way that the
lines P8 2 and Pb are equal. If d, the slit spacing, is much smaller than the
distance D between the two screens (the ratio d/D in the figure has been
exaggerated for clarity), 8 2 b is then almost perpendicular to both r 1 and r 2 .
This means that angle 8 18 2 b is almost equal to angle PaO, both angles being
marked e in the figure. This is equivalent to saying that the lines r 1 and r 2
may be taken as parallel.
We often put a lens in front of the two slits, as in Fig. 43-6, the screen C
being in the focal plane of the lens. Under these conditions light focused at
P must have struck the lens parallel to the line Px, drawn from P through
the center of the (thin) lens. Under these conditions rays r 1 and r2 are
strictly parallel even though the requirement D » d is not met. The lens L
may in practice be the lens and cornea of the eye, screen C being the retina.
The two rays arriving at P in Figs. 43-5 or 43-6 from 8 1 and 8 2 are in
phase at the source slits, both being derived from the same wavefront in the
incident plane wave. Because the rays have different optical path lengths,
1072 INTERFERENCE Chap. 43

Fig. 43-6 A lens is normally used to produce interference fringes; compare with Fig. 43-5.
The figure is again distorted for clarity in that f » d in practice.

they arrive at P with a phase difference. The number of wavelengths con-


tained in Sib, which is the path difference, determines the nature of the inter-
ference at P.
To have a maximum at P, Sib ( = d sine) must contain an integral number
of wavelengths, or
Sib= m")\ m = 0, 1, 2, ... ,
which can be written as
d sin() = m")\ m = 0, 1, 2, ... (maxima). (43-1)
Note that each maximum above O in Figs. 43-5 and 43-6 has a symmetri-
cally located maximum below 0. There is a central maximum described by
m = 0.
For a minimum at P, Sib ( = d sine) must contain a half-integral number
of wavelengths, or
d sin () = (m + t) A m = 0, 1, 2, ... (minima). (43-2)
If a lens is used as in Fig. 43-6, it may seem that a phase difference should
develop between the rays beyond the plane represented by S2 b, the path
lengths between this plane and P being clearly different. In Section 42-5,
however, we saw that for such parallel rays focused by a lens the optical path
lengths are identical. Two rays with the same optical path lengths contain
the same number of wavelengths, so that no phase difference will result be-
cause of the light passing through the lens.
Sec. 43-1 YOUNG'S EXPERIMENT 1073

... Example 1. The double-slit arrangement in Fig. 43-5 is illuminated with light
from a mercury vapor lamp so filtered that only the strong green line (X = 5460 A) is
effective. The slits are 0.10 mm apart, and the screen on which the interference
pattern appears is 20 cm away. What is the angular position of the first minimum?
Of the tenth maximum?
At the first minimum we put m = 0 in Eq. 43-2, or
sin() = (m + f)X = (!)(546 X 10- 9 meter) = 0 _0027 _
d 0.10 X 10-3 meter
This value for sin () is so small that we can take it to be the value of 0, expressed in
radians; expressed in degrees it is O.16 °.
At the tenth maximum (not counting the central maximum) we must put m = 10
in Eq. 43-1. Doing so and calculating as before leads to an angular position of 3.8°.
For these conditions we see that the angular spread of the first dozen or so fringes is
small.
Example 2. In Example 1 what is the linear distance on screen C between ad-
jacent maxima?
If() is small enough, we can use the approximation

sin() ,...._, tan() ,...._, e.


y
From Fig. 43-5 we see that tanO = -·
D
Substituting this into Eq. 43-1 for sin() leads to
XD
y=m- m = 0, 1, 2, ... (maxima).
d
The positions of any two adjacent maxima are given by

and Ym+I = (m + 1) XD
d.
Their separation b..y is found by subtracting:
XD
D..y = Ym+l - Ym = d
= (546 X 10-9 meter)(20 X 10- 2 meter) = 1.og mm.
0.10 X 10-3 meter
As long as() in Figs. 43-5 and 43-6 is small, the separation of the interference fringes is
independent of m; that is, the fringes are evenly spaced. Note that if the incident
light contains more than one wavelength the separate interference patterns, which
will have different fringe spacings, will be superimposed. ~

Equation 43-1 can be used to determine the wavelength of light; to quote Thomas
Young:
From a comparison of various experiments, it appears that the breadth of the
undulations [that is, the wavelength] constituting the extreme red light must be
supposed to be, in air, about one 36 thousandth of an inch, and those of the extreme
violet about one 60 thousandth; the mean of the whole spectrum, with respect to
the intensity of light, being one 45 thousandth.
1074 INTERFERENCE Chap. 43
Young's value for the average effective wavelength present in sunlight (1/45,000 in.)
can be written as 5700 A, which agrees rather well with the wavelength at which the
eye sensitivity is a maximum, 5550 A (see Fig. 40-2). It must not be supposed that
Young's work was received without criticism. One of his contemporaries, evidently
a firm believer in the corpuscular theory of light, wrote:
We wish to raise our feeble voice against innovations that can have no other
effect than to check the progress of science, and renew all those wild phantoms of
the imagination which Bacon and Newton put to flight from her temple. This
paper contains nothing that deserves the name of either experiment or discovery.
Needless to say, posterity has decided in favor of Young.

43-2 Coherence
Analysis of the derivation of Eqs. 43-1 and 43-2 shows that a fundamental
requirement for the existence of well-defined interference fringes on screen C
in Fig. 43-1 is that the light waves that travel from 8 1 and 8 2 to any point P
on this screen must have a sharply defined phase difference cp that remains
constant with time. If this condition is satisfied, a stable, well-defined
fringe pattern will appear. At certain points P, cp will be given, independ-
ent of time, by n1r where n = 1, 3, 5, ... so that the resultant intensity
will be strictly zero and will remain so throughout the time of observation.
At other points cp will be given by n1r where n = 0, 2, 4 ... and the result-
ant intensity will be a maximum. Under these conditions the two beams
emerging from slit 8 1 and 8 2 are said to be completely coherent.
Let the source in Fig. 43-1 be removed and let slits 8 1 and 8 2 be replaced
by two completely independent light sources, such as two fine incandescent
wires placed side by side in a glass envelope. No interference fringes will
appear on screen C but only a relatively uniform illumination. We can in-
terpret this if we make the reasonable assumption that for completely inde-
pendent light sources the phase difference between the two beams arriving
at P will vary with time in a random way. At a certain instant conditions
may be right for cancellation and a short time later (perhaps 10-s sec) they
may be right for re-enforcement. This same random phase behavior holds
for all points on screen C with the result that this screen is uniformly illu-
minated. The intensity at any point is equal to the sum of the intensities
that each source 8 1 and 8 2 produces separately at that point. Under these
conditions the two beams emerging from 8 1 and 8 2 are said to be completely
incoherent.
Note that for completely coherent light beams one (1) combines the
amplitudes vectorially, taking the (constant) phase difference properly into
account, and then (2) squares this resultant amplitude to obtain a quantity
proportional to the resultant intensity. For completely incoherent light
beams, on the other hand, one (1) squares the individual amplitudes to
obtain quantities proportional to the individual intensities and then (2) adds
the individual intensities to obtain the resultant intensity. This procedure
is in agreement with the experimental fact that for completely independent
light sources the resultant intensity at every point is always greater than
the intensity produced at that point by either light source acting alone.
Sec. 43-2 COHERENCE 1075

It remains to investigate under what experimental conditions coherent or


incoherent beams may be produced and to give an explanation for coherence
in terms of the mode of production of the radiation. Consider first a parallel
beam of microwave radiation emerging from an antenna connected by a co-
axial cable to an oscillator based on an electromagnetic resonant cavity.
The cavity oscillations (see Section 38-6) are completely periodic with time
and produce, at the antenna, a completely periodic variation of E and B
with time. The radiated wave at large enough distances from the antenna
is well represented by Fig. 39-11. Note that (1) the wave has essentially
infinite extent in time, including both future times (t > 0, say) and past
times (t < 0); see Fig. 43-7a. At any point, as the wave passes by, the
wave disturbance (i.e., E or B) varies with time in a perfectly periodic way.
(2) The wavefronts at points far removed from the antenna are parallel
planes of essentially infinite extent at right angles to the propagation direc-
tion. At any instant of time the wave disturbance varies with distance
along the propagation direction in a perfectly periodic way.
Two beams generated from a single traveling wave like that of Fig. 39-11
will be completely coherent. One way to generate two such beams is to put
an opaque screen containing two slits in the path of the beam. The waves
emerging from the slits will always have a constant phase difference at any
point in the region in which they overlap and interference fringes will be
produced. Coherent radio beams can also be readily established, as can
coherent elastic waves in solids, liquids and gases. The two prongs of the
vibrating tapper in Fig. 43-4, for example, generate two coherent waves in
the water of the ripple tank.
The technique of producing two beams from a single beam (and tlius from a single
source) tests specifically whether the wavefronts in the single parallel beam are truly
planes, that is, whether all points in a plane at right angles to the direction of propa-
gation have the same phase at any given instant. By dividing the beam in another
way it is possible to test whether the beam is truly periodic over a large number of
cycles of oscillation. This can be done, as we show in detail in Section 43-6, by in-
serting in the beam at 45 ° to it a thin sheet of material possessing the property that
two beams are produced, one (which will be at right angles to the incident beam) by
reflection and a second (which will be in the direction of the incident beam) by
transmission. In the visible region such a sheet, called a half-silvered mirror, may be
formed from a glass plate by depositing on it an appropriately thin film of silver.
By appropriate use of mirrors (see Section 43-7) these two sub-beams can be re-
combined into a single beam traveling in a chosen direction. If the beams travel
different distances before they are recombined, we are comparing in the combined
beam a sample of the original beam with another sample an arbitrarily large number
of cycles away. If the original beam is truly periodic in space and time, the two sub-
beams will be completely coherent and interference fringes will be produced when
they are recombined.
If we turn from microwave sources to common sources of visible light,
such as incandescent wires or an electric discharge passing through a gas,
we become aware of a fundamental difference. In both of these sources the
fundamental light emission processes occur in individual atoms and these
atoms do not act together in a cooperative (i.e., coherent) way. The act of
1076 INTERFERENCE Chap. 43
light emission by a single atom takes, in a typical case, about 10-s sec and
the emitted light is properly described as a wavetrain (Fig. 43-7b) rather
than as a wave (Fig. 43-7a). For emission times such as these the wave-
trains are a few meters long.
Interference effects from ordinary light sources may be produced by put-
ting a very narrow slit (So in Fig. 43-1) directly in front of the source. This
insures that the wavetrains that strike slits S 1 and S 2 in screen B in this
figure originate from the same small region of the source. The diffracted
beams emerging from S 1 and S 2 thus represent the same population of wave-
trains and are coherent with respect to each other. If the phase of the light
emitted from S 0 changes, this change is transmitted simultaneously to S 1
and S2 . Thus, at any point on screen C, a constant phase difference is
maintained between the beams from these two slits and a stationary inter-
ference pattern occurs.
If the width of slit So in Fig. 43-1 is gradually increased it will be observed experi-
mentally that the maxima of the interference fringes become reduced in intensity
and that the intensity in the fringe minima is no longer strictly zero. In other words,
the fringes become less distinct. If So is opened extremely wide, the lowering of the
maximum intensity and the raising of the minimum intensity will be such that the
fringes disappear, leaving only a uniform illumination. Under these conditions we
say that the beams from 81 and 82 pass continuously from a condition of complete
coherence to one of complete incoherence. When not at either of these two limits
the beams are said to be partially coherent.
Partial coherence can also be demonstrated in two beams that are _12!'Q~_l!_ced, as
described on p. 1090 by inserting a "half-silvered mirror" in a beam at an angle of
45° to the direction of propagation. The two beams so produced, by reflection and
transmission, traverse paths of different lengths before they are recombined. If the
path difference is small compared to the average length of a wavetrain, the interference
fringes will be sharply defined and will go essentially to zero at their minima. If the
path difference is deliberately made longer, the fringes will become less distinct, and
finally, when the path difference is larger than the average length of a wavetrain, the
fringes will disappear altogether. Thus it is possible once again to progress smoothly
in an experimental arrangement from complete coherence, through partial coherence,
to complete incoherence.
The lack of coherence of the light from ordinary sources such as glowing
wires is due to the fact that the emitting atoms do not act cooperatively

(\ (\ (\ (\ (\ (\ (\ (\ (\ (\

(b)

Fig. 43-7 (a) A section of an infinite wave and (b) a wavetrain.


Sec. 4-3-3 INTENSITY IN YOUNG'S EXPERIMENT 1077

Fig. 43-8 A screen with two small holes is placed against the end of an optical maser.
(a) The light passing through the holes forms an interference pattern on a strip of photo-
graphic film placed at (b). The fact that the pattern is formed shows that the light
emitted from the maser is coherent across the beam cross section. At (b) is shown the
image as formed on a photographic strip exposed in this manner. At (c) is an intensity
plot of the film made by measuring the degree to which the film has been exposed. (Cour-
tesy of D. F. Nelson and R. J. Collins, Bell Telephone Laboratories.)

(i.e., coherently). Since 1960 it has proved possible to construct sources of


visible light in which the atoms do act cooperatively and in which the
emitted light is highly coherent. Such devices are called optical masers or
lasers; * their light output is extremely monochromatic, intense, and highly
collimated. The coherence of the emitted light can be demonstrated by
placing a screen with two holes in it in the beam emerging from the laser.
An interference pattern results, as Fig. 43-8 shows. These methods permit,
for the first time, a degree of control of visible light approaching that possi-
ble for radio and for microwaves. The practical applications of lasers, which
include the amplification of weak light signals, the use of light beams as
highly efficient carriers of information from point to point (see problem 13),
and the production of high temperatures by intense local heating, remam
to be fully exploited.

43-3 Intensity in Young's Experiment


Let us assume that the electric field components of the two waves in Fig.
43-5 vary with time at point Pas
E1 = E 0 sin wt (43-3)
and E2 = Eo sin (wt + </J) (43-4)
where w ( = 21r11) is the angular frequency of the waves and <pis the phase
difference between them. Note that <p depends on the location of point P,
* Laser, like maser, is a coined word, meaning "light amplification through stimulated
emission of radiation." The interested student is referred to an article called "Optical
Masers" by Arthur L. Schawlow, Scientific American, June 1961.
1078 INTERFERENCE Chap. 43
which, in turn, for a fixed geometrical arrangement, is described by the angle
8 in Figs. 43-5 and 43-6. We assume that the slits are so narrow that the
diffracted light from each slit illuminates the central portion of the screen
uniformly. This means that near the center of the screen E 0 is independent
of the position of P, that is, of the value of 8.
The resultant wave disturbance * at Pis found from

E = E1 + E2 (43-5)

and (see Eq. 19-17) proves to be

E = Eo sin (wt + (3), (43-6a)

where (43-6b)

and Eo = 2Eo cos (3 = Em cos (3. (43-6c)

Em, the maximum possible amplitude for Eo, is equal to twice the amplitude
of the combining waves ( = 2E0 ), corresponding to complete reinforcement.
The student should verify Eq. 43-6 carefully. The amplitude Eo of the
resultant wave disturbance, which determines the intensity of the inter-
ference fringes, will turn out to depend strongly on the value of 8, that is, on
the location of point P in Figs. 43-5 and 43-6.
In Section 19-6 we showed that the intensity of a wave I, measured per-
haps in watts/meter 2 , is proportional to the square of its amplitude. For
the resultant wave then, ignoring the proportionality constant,

Io ex E/. (43-7)

This relationship seems reasonable if we recall (Eq. 30-27) that the energy
density in an electric field is proportional to the square of the electric field
strength. This is true for rapidly varying electric fields, such as those in a
light wave, as well as for static fields.
The ratio of the intensities of two light waves is the ratio of the squares of
the amplitudes of their electric fields. If I O is the intensity of the resultant
wave at P and I O is the intensity that a single wave acting alone would
produce, then
Io=
I0
(EEo0) 2
(43-8)

Combining with Eq. 43-6c leads to

Io = 4I 0 cos 2 (3 = Im cos 2 (3. (43-9)

* The electric field E in the light wave rather than the magnetic field B is normally
identified with the "wave disturbance" because the effects of Bon the human eye and on
various light detectors are exceedingly small. Radiation pressure (Section 40-2) is one such
effect. Note too that, although Eq. 43-5 should be a vector equation, in most cases of
interest the E vectors in the two interfering waves are closely parallel so that an algebraic
equation suffices.
Sec. 43-3 INTENSITY IN YOUNG'S EXPERIMENT 1079

Note that the intensity of the resultant wave at any point P varies from
zero [for a point at which <j, ( = 2/3) = 1r, say] to Im, which is four times the
intensity 10 of each individual wave [for a point at which <j, ( = 2/3) = 0, say].
Let us compute I e as a function of the angle 8 in Figs. 43-5 or 43-6.
The phase difference <j, in Eq. 43-4 is associated with the path difference
Sib in Fig. 43-5 or 43-6. If Sib is tA, <j, will be 1r; if S 1 b is A, <j, will be 21r,
etc. This suggests that

phase difference path difference


21r A

21r
<j, = - (d sin 8),
A
or, finally, from Eq. 43-6b,
1rd
~ = !cf> = - sin 8. (43-10)
A

This expression for /3 can be substituted into Eq. 43-9 for 18, yielding the
latter quantity as a function of 8. For convenience we collect here the ex-
pressions for the amplitude and the intensity in double-slit interference.

[Eq. 43-6c] Ee= Em cos~ interference (43-lla)


[Eq. 43-9] le= Im cos2 ~ from narrow (43-llb)
slits (that is,
1 1rd
[Eq. 43-10] ~ ( = 2 cf,) = - sin O
X a« X) (43-llc)

To find the positions of the intensity maxima, we put

m = 0, 1, 2, ...
in Eq. 43-llb. From Eq. 43-llc this reduces to

d sin O = mA m = 0, 1, 2, ... (maxima),

which is the equation derived in Section 43-1 (Eq. 43-1). To find the in-
tensity minima we write

1rd sin O
- - - = (m + !)1r m = 0, 1, 2, ... (minima),
A
which reduces to the previously derived Eq. 43-2.
Figure 43-9 shows the intensity pattern for double-slit interference. The
horizontal solid line is 10 ; this describes the (uniform) intensity pattern on
the screen if one of the slits is covered up. If the two sources were incoherent
the intensity would be uniform over the screen and would be 210 ; see the
horizontal dashed line in Fig. 43-9. For coherent sources we expect the
1080 INTERFERENCE Chap. 43

~
"vi

I ~twio;r
!\--A-
- --4Io(= Im)

~
" " \ ~ \ / \ ~ \ / ~ L ~ Io(one source)
~-A~2To(t;,~o~~;~~;-
(jJ 57!" 471" 37!" 271" 7r O 7r 271" 371" 471" 57!"
E. l_ 2 ],._ 3 O l_ E._ l_
sin O (see Eq. 43- 11 c) 2 d d 2 aA d
A 1 A
2 d
1 A
2 ,I cf
A 3 A
2 d 2d 2 d

m(Eq.43-1) 2 1 0 1 2 maxima
m (Eq. 43-2) 2 1 0 0 1 2 minima

Fig. 43-9 The intensity pattern for double-slit interference. The heavy arrow in the
central peak represents the half-width of the peak. This figure is constructed on the
assumption that the two interfering waves each illuminate the central portion of the
screen uniformly, that is, / 0 is independent of position as shown.

energy to be merely redistributed over the screen, since energy is neither


created nor destroyed by the interference process. Thus the average inten-
sity in the interference pattern should be 210 , as for incoherent sources.
This follows at once if, in Eq. 43-llb, we substitute one-half for the cosine-
squared term and if we recall that Im= 41 0 . We have seen several times
that the average value of the square of a sine or a cosine term over one or more
half-cycles is one-half.

43-4 Adding Wave Disturbances


In Section 43-3 we combined two time-varying wave disturbances, namely

E 1 = E 0 sin wt (43-3)

and E2 = E 0 sin (wt + ¢), (43-4)

which have the same angular frequency wand amplitude E 0 but which have
a phase difference tj, between them. In this case the result (Eqs. 43-lla and
c) is easily obtained algebraically.
In later chapters we will want to add larger numbers of wave disturbances,
often an infinite number, with infinitesimal individual amplitudes. Since
analytic methods become more difficult in such cases we describe a graphical
method, illustrating it by rederiving Eq. 43-lla.
A sinusoidal wave disturbance such as that represented by Eq. 43-3 can
be represented graphically, using a rotating vector. In Fig. 43-lOa a vector
of magnitude E 0 is allowed to rotate about the origin in a counterclockwise
direction with an angular frequency w. Following electrical engineering
practice we call such a rotating vector a phasor. The alternating wave dis-
turbance E 1 (Eq. 43-3) is represented by the projection of this phasor on the
vertical axis.
A second wave disturbance E 2 , which has the same amplitude E 0 but a
Sec. 43-4 ADDING WAVE DISTURBANCES 1081

phase difference <fJ with respect to E 1 ,


E2 = E 0 sin (wt+ </J), (43-4)
can be represented graphically (Fig. 43-lOb) as the projection on the vertical
axis of a second phasor of magnitude E 0 which makes an angle <fJ with the
first phasor. As this figure shows, the sum E of E 1 and E 2 is the sum of the
projections of the two phasors on the vertical axis. This is revealed more
clearly if the phasors are redrawn, as in Fig. 43-lOc, placing the foot of one
arrow at the head of the other, maintaining the proper phase difference, and
letting the whole assembly rotate counterclockwise about the origin.
In Fig. 43-lOc E can also be regarded as the projection on the vertical axis
of a phasor of length Eo, which is the vector sum of the two phasors of magni-
tude E 0 . Note that the (algebraic) sum of the projections of the two phasors
is equal to the projection of the (vector) sum of the two phasors.
In most problems in optics we are concerned only with the amplitude E 0
of the resultant wave disturbance and not with its time variation. This is
because the eye and other common measuring instruments respond to the
resultant intensity of the light (that is, to the square of the amplitude) and
cannot detect the rapid time variations that characterize visible light. For
sodium light, for example (A = 5890 A), the frequency v ( = w/21r) is 5.1 X
10 14 cycles/sec. Often, then, we need not consider the rotation of the
phasors but can confine our attention to finding the magnitude of the re-
sultant phasor.

(a) (b)

Fig. 43-10 (a) A wave disturbance E1 is rep-


resented by a rotating vector or phasor.
(b) Two wave disturbances E 1 and E 2 , with
a phase difference </> between them, are so
represented. These two phasors can represent
the two wave disturbances in the double-slit
problem; see Eqs. 43-3 and 43-4. (c) Another
way of drawing (b).

(c)
1082 INTERFERENCE Chap. 43

Fig. 43-11 (a) A construction to find the amplitude


Ee of two wave disturbances of amplitude Eo and
(a)
phase difference </>. (b) The maximum possible ampli-
tude for these wave disturbances occurs for </> = 0
Em I
and has the value Em = 2Eo.

Eo
[)
Eo ~
(b)

Figure 43-1 la shows the phasors for double-slit interference at time t = 0;


compare Fig. 43-lOc. We see that
Ee = 2E0 cos f3 = Em cos {3,
in which, from the theorem that the exterior angle of a triangle (</>) is equal
to the sum of its two opposite interior angles (f3 + {3),
f3 = t<J>.
This is exactly the result arrived at earlier algebraically; compare Eqs.
43-lla and c.
In a more general case we might want to find the resultant of a number
(>2) of sinusoidally varying wave disturbances. The general procedure is
the following:

1. Construct a series of phasors representing the functions to be added.


Draw them end to end, maintaining the proper phase relationships between
adjacent phasors.
2. Construct the vector sum of this array. Its length gives the amplitude
of the resultant. The angle between it and the first phasor is the phase of
the resultant with respect to this first phasor. The projection of this phasor
on the vertical axis gives the time variation of the resultant wave disturbance .
.... Example 3. Find graphically the resultant E(t) of the following wave dis-
turbances:
E1 = 10 sin wt
E2 = 10 sin (wt + 15°)
Es = 10 sin (wt+ 30°)
E4 = 10 sin (wt+ 45°).
Figure 43-12 in which Eo equals 10, shows the assembly of four phasors that repre-
sents these functions. Their vector sum, by graphical measurement, has an ampli-
tude ER of 38 and a phase </>o with respect to E 1 of 23°. In other words
E(t) = E1 + E2 +Es+ E 4 = 38 sin (wt+ 23°).
Check this result by trigonometric calculation.
Sec. 43-5 INTERFERENCE FROM THIN FILMS 1083

.E
Fig. 43-12 Example 3. Four wave dis-
turbances are added graphically, using
the method of phasors.

43-5 Interference from Thin Films


The colors of soap bubbles, oil slicks, and other thin films are the result of
interference. Figure 43-13 shows interference effects in a thin vertical film
of soapy water illuminated by monochromatic light.
Figure 43-14 shows a film of uniform thickness d and index of refraction n,
the eye being focused on spot a. The film is illuminated by a broad source of
monochromatic light S. There exists on this source a point P such that two
rays, identified by the single and double arrows, respectively, can leave P and
enter the eye as shown, after passing through point a. These two rays follow

Fig. 43-13 A soapy water film on a wire loop,


viewed by reflected light. The black segment
at the top is not a tear. It arises because the
film, by drainage, is so thin here that there is
destructive interference between the light re-
flected from its front surface and that reflected
from its back surface. Recall that these two
waves differ in phase by 180°
1084 INTERFERENCE Chap. 43

Fig. 43-14 Interference by reflection from a thin film, assuming an extended source S.

different paths in going from P to the eye, one being reflected from the upper
surface of the film, the other from the lower surface. Whether point a ap-
pears bright or dark depends on the nature of the interference between the
two waves that diverge from a. These waves are coherent because they both
originate from the same point Pon the light source.
If the eye looks at another part of the film, say a', the light that enters the
eye must originate from a different point P' of the extended source, as sug-
gested by the dashed lines in Fig. 43-14.
For near-normal incidence (0 ,...._, 0 in Fig. 43-14) the geometrical path dif-
ference for the two rays from P will be close to 2d. We might expect the re-
sultant wave reflected from the film near a to be an interference maximum if
the distance 2d is an integral number of wavelengths. This statement must
be modified for two reasons.
First, the wavelength must refer to the wavelength of the light in the film
An and not to its wavelength in air A; that is, we are concerned with optical
path lengths rather than geometrical path lengths. The wavelengths A and
An (see Eq. 41-11) are related by
An = A/n. (43-12)
To bring out the second point, let us assume th~t the film is so thin that 2d
is very much less than a wavelength. The phase difference between the two
Sec. 43-5 INTERFERENCE FROM THIN FILMS 1085

waves would be close to zero on our assumption, and we would expect such a
film to appear bright on reflection. However, it appears dark. This is clear
from Fig. 43-13, in which the action of gravity produces a wedge-shaped
film, extremely thin at its top edge. As drainage continues, the dark area
increases in size. To explain this and many similar phenomena, we assume
that one or the other of the two rays of Fig. 43-14 suffers an abrupt phase
change of 1r ( = 180°) associated either with reflection at the air-film interface
or transmission through it. As it turns out, the ray reflected from the upper
surface suffers this phase change. The other ray is not changed abruptly in
phase, either on transmission through the upper surface or on reflection at
the lower surface.
In Section 19-9 we discussed phase changes on reflection for transverse
waves in strings. To extend these ideas, consider the composite string of
Fig. 43-15, which consists of two parts with different masses per unit length,
stretched to a given tension. If a pulse moves to the right in Fig. 43-15a,
approaching the junction, there will be a reflected and a transmitted pulse,

-I>
Initial I\ <}- ---!>
Final
Fig. 43-15 Phase changes on reflection at a
(a)
junction between two stretched composite
strings. (a) Incident pulse in heavy string ---{>

and (b) incident pulse in light string. Initial I\ -I>


Final I\
v
<1--
(b)

the reflected pulse being in phase with the incident pulse. In Fig. 43-15b
the situation is reversed, the incident pulse now being in the less massive
string. In this case the reflected pulse will differ in phase from the incident
pulse by 1r ( = 180°). In each case the transmitted pulse will be in phase with
the incident pulse.
Figure 43-15a suggests a light wave in glass, say, approaching a surface
beyond which there is a less optically dense medium (one of lower index of
refraction) such as air. Figure 43-15b suggests a light wave in air approach-
ing glass. To sum up the optical situation, when reflection occurs from an
interface beyond which the medium has a lower index of refraction, the re-
flected wave undergoes no phase change; when the medium beyond the inter-
face has a higher index, there is a phase change of 1r. * The transmitted wave
does not experience a change of phase in either case.
* These statements, which can be proved rigorously from Maxwell's equations (see also
Section 43-6), must be modified for light falling on a less dense medium at an angle such
that total internal reflection occurs. They must also be modified for reflection from
metallic surfaces.
1086 INTERFERENCE Chap. 43
We are now able to take into account both factors that determine the na-
ture of the interference, namely, differences in optical path length and phase
changes on reflection. For the two rays of Fig. 43-14 to combine to give a
maximum intensity, assuming normal incidence, we must have

2d = (m + t)An m = 0, 1, 2, ....

The term tAn is introduced because of the phase change on reflection, a


phase change of 180° being equivalent to half a wavelength. Substituting
A/n for An yields finally
2dn = (m + t)A m = 0, 1, 2, ... (maxima). (43-13)

The condition for a minimum intensity is

2dn = mA m = 0, 1, 2, ... (minima). (43-14)

These equations hold when the index of refraction of the film is either greater
or less than the indices of the media on each side of the film. Only in these
cases will there be a relative phase change of 180° for reflections at the two
surfaces. A water film in air and an air film in the space between two glass
plates provide examples of cases to which Eqs. 43-13 and 43-14 apply.
Example 5 provides a case in which they do not apply.
If the film thickness is not uniform, as in Fig. 43-13, where the film is
wedge-shaped, constructive interference will occur in certain parts of the
film and destructive interference will occur in others. Lines of maximum
and of minimum intensity will appear-these are the interference fringes.
They are called fringes of constant thickness, each fringe being the locus of
points for which the film thickness dis a constant. If the film is illuminated
with white light rather than monochromatic light, the light reflected from
various parts of the film will be modified by the various constructive or de-
structive interferences that occur. This accounts for the brilliant colors of
soap bubbles and oil slicks.

Only if the film is "thin," which implies that d is no more than a few wavelengths
of light, will fringes of the type described, that is, fringes that appear localized on the
film and associated with a variable film thickness, be possible. For very thick films
(say d ,.._., 1 cm), the path difference between the two rays of Fig. 43-14 will be many
wavelengths and the phase difference at a given point on the film will change rapidly
as we move even a small distance away from a. For "thin" films, however, the phase
difference at a also holds for reasonably nearby points; there is a characteristic "patch
brightness" for any point on the film, as Fig. 43-13 shows. Interference fringes can
be produced for thick films; they are not localized on the film but are at infinity. See
Section 43-7 .

._ Example 4. A water film (n = 1.33) in air is 3200 A thick. If it is illuminated


with white light at normal incidence, what color will it appear to be in reflected light?
By solving Eq. 43-13 for A,

A= ~ = (2)(3200 A)(l.33) = 8500 A


(maxima).
m+! m+! m+i
Sec. 43-5 INTERFERENCE FROM THIN FILMS 1087
From Eq. 43-14 the minima are given by

A = 8500 A (minima).
m
Maxima and minima occur for the following wavelengths:

m 0 (max) 1 (min) 1 (max) 2 (min) 2 (max)

A,A 17000 8500 5700 4250 3400

Only the maximum corresponding tom = l lies in the visible region (see Fig. 40-2);
light of this wavelength appears yellow-green. If white light is used to illuminate the
film, the yellow-green component will be enhanced when viewed by reflection.
Example 5. N onrefiecting glass. Lenses are often coated with thin films of trans-
parent substances like MgF 2 (n = 1.38) in order to reduce the reflection from the
glass surface, using interference. How thick a coating is needed to produce a mini-
mum reflection at the center of the visible spectrum (5500 A)?
We assume that the light strikes the lens at near-normal incidence (8 is exaggerated
for clarity in Fig. 43-16), and we seek destructive interference between rays rand r 1 •
Equation 43-14 does not apply because in this case a phase change of 180° is asso-
ciated with each ray, for at both the upper and lower surfaces of the MgF 2 film the
reflection is from a medium of greater index of refraction.
There is no net change in phase produced by the two reflections, which means that
the optical path difference for destructive interference is (m + i)A (compare Eq.
43-13), leading to
2dn = (m + })A m = 0, 1, 2, ... (minima).
Solving ford and putting m = 0 yields
d = (m +!)A=~= 5500 A = lOOO A.
2n 4n (4)(1.38)

Fig. 43-16 Example 5. Unwanted reflections from glass can be reduced by coating the
glass with a thin transparent film.
1088 INTERFERENCE Chap. 43
Example 6. Newton's rings. Figure 43-17 shows a lens of radius of curvature R
resting on an accurately plane glass plate and illuminated from above by light of
wavelength A. Figure 43-18 shows that circular interference fringes (Newton's rings)
appear, associated with the variable thickness air film between the lens and the plate.
Find the radii of the circular interference maxima.
Here it is the ray from the bottom of the (air) film rather than from the top that
undergoes a phase change of 180°, for it is the one reflected from a medium of higher
refractive index. The condition for a maximum remains unchanged, however (Eq.
43-13), and is
2d = (m+ !)>- m = 0, 1, 2, ... , (43-15)

the index of refraction of the air film being assumed to be unity. From Fig. 43-17
we can write
d = R - VR2- r2 = R - R [1- (i )2] . Yz

If r/R << 1, the square bracket can be expanded by the binomial theorem, keeping
only two terms, or
I ( r
d= R - R [1- 2 R )2 + .. .J"-' 2R.
r2
Combining with Eq. 43-15 yields

r = V(m + !)>.R m = 0, 1, 2, ... ,

Fig. 43-17 Example 6. Appa-


ratus for observing Newton's
rings.
Sec. 43-6 PHASE CHANGES ON REFLECTION 1089

Fig. 43-18 Example 6. Newton's rings. (Courtesy of Bausch and Lomb Optical Co.)

which gives the radii of the bright rings. If white light is used, each spectrum com-
ponent will produce its own set of circular fringes, the sets all overlapping. ....

43-6 Phase Changes on Reflection


G. G. Stokes (1819-1903) used the principle of optical reversibility to investigate
the reflection of light at an interface between two media. The principle states that if
there is no absorption of light a light ray that is reflected or refracted will retrace its
original path if its direction is reversed. This reminds us that any mechanical system
can run backward as well as forward, provided there is no absorption of energy be-
cause of friction, etc.
Figure 43-19a shows a wave of amplitude E reflected and refracted at a surface
separating media 1 and 2, where n2 > n1. The amplitude of the reflected wave is
r12E, in which r12 is an amplitude reflection coefficient. The amplitude of the refracted
wave is t12E, where t12is an amplitude transmission coefficient.
We consider only the possibility of phase changes of O or 180°. If r 12 = +0.5, for
example, we have a reduction in amplitude on reflection by one-half and no change
in phase. For r12 = -0.5 we have a phase change of 180° because
E sin (wt+ 180°) = -E sin wt.
Figure 43-19b suggests that if we reverse these two rays they should combine to
produce the original ray reversed in direction. Ray r 12E, identified by the single
arrows in the figure, is reflected and refracted, producing the rays of amplitudes
1090 INTERFERENCE Chap. 43

Fig. 43-19 (a) A ray is reflected and refracted at an air-glass interface. (b) The optically
reversed situation; the two rays in the lower left must cancel.

r12 2E and r12t12E. Ray t12E, identified by the triple arrows, is also reflected and re-
fracted, producing the rays of amplitudes t12t21E and t12r21E as shown. Note that r12
describes a ray in medium 1 reflected from medium 2, and r 21 describes a ray in
medium 2 reflected from medium 1. Similarly, t12 describes a ray that passes from
medium 1 to medium 2; b describes a ray that passes from medium 2 to medium 1.
The two rays in the upper left of Fig. 43-19b must be equivalent to the incident ray
of Fig. 43-19a, reversed; the two rays in the lower left of Fig. 43-19b must cancel.
This second requirement leads to
rd12E + t12r21E = 0,
or
This result tells us that if we compare a wave reflected from medium one with one
reflected from medium 2, they behave differently in that one or the other undergoes a
phase change of 180°. We must rely on experiment to show that, as we pointed out
earlier, the ray reflected from the more optically dense medium is the one that ex-
periences the phase change of 180°.

43-7 Michelson's Interferometer


An interferometer is a device that can be used to measure lengths or changes
in length with great accuracy by means of interference fringes. We describe
the form originally built by Michelson in 1881.
Consider light that leaves point P on extended source S (Fig. 43-20) and
falls on half-silvered mirror M. This mirror has a silver coating just thick
enough to transmit half the incident light and to reflect half; in the figure
we have assumed that this mirror, for convenience, possesses negligible thick-
ness. At 1l1. the light divides into two waves. One proceeds by transmission
toward mirror 1l11 ; the other proceeds by reflection toward 1l12. The waves
Sec. 43-7 MICHELSON'S INTERFEROMETER 1091

are reflected at each of these mirrors and are sent back along their directions
of incidence, each wave eventually entering the eye E. Since the waves are
coherent, being derived from the same point on the source, they will inter-
fere.
If the mirrors M 1 and M 2 are exactly perpendicular to each other, the effect
is that of light from an extended source S falling on a uniformly thick slab of
air, between glass, whose thickness is equal to d 2 - d 1 . Interference fringes
appear, caused by small changes in the angle of incidence of the light from
different points on the extended source as it strikes the equivalent air film.
For thick films a path difference of one wavelength can be brought about by a
very small change in the angle of incidence.
If 1l1.2 is moved backward or forward, the effect is to change the thickness
of the equivalent air film. Suppose that the center of the (circular) fringe
pattern appears bright and that M 2 is moved just enough to cause the first
bright circular fringe to move to the center of the pattern. The path of the
light beam striking M 2 has been changed by one wavelength. This means
(because the light passes twice through the equivalent air film) that the
mirror must have moved one-half a wavelength.
The interferometer is used to measure changes in length by counting the
number of interference fringes that pass the field of view as mirror M 2 is
moved. Length measurements made in this way can be accurate if large
numbers of fringes are counted.

Fig. 43-20 Michelson's interferometer, showing the path of a particular ray originating
at point P of an extended source S.
1092 INTERFERENCE Chap. 43
Michelson measured the length of the standard meter, kept in Paris, in
terms of the wavelength of certain monochromatic red light emitted from a
light source containing cadmium. He showed that the standard meter was
equivalent to 1,553,163.5 wavelengths of the red cadmium light.
Physicists have long speculated on the advantages of discarding the
standard meter bar as the basic standard of length and of defining the meter
in terms of the wavelength of some carefully chosen monochromatic radia-
tion. This would make the primary length standard readily available in
laboratories all over the world. It would improve the accuracy of length
measurements, since one would no longer need to compare an unknown ob-
ject with a standard object (the meter bar), using interferometer techniques,
but could measure the unknown object directly and in an absolute sense, using
these techniques. There is the additional advantage that if the standard
meter bar were destroyed it could never be replaced, whereas light sources
and interferometers will (presumably) always be available.
In 1961 such an atomic standard of length was adopted by international
agreement. Quoting from an article * describing the event:
The wavelength of the orange-red light of krypton-86 has replaced the platinum
iridium bar as the world standard of length. Formerly the wavelength of this light
was defined as a function of the length of the meter bar. Now the meter is defined
as a multiple (1,650,763.73) of the wavelength of the light.
The light from krypton-86 was used in preference to that from cadmium or
other sources because it produces sharper interference fringes in the inter-
ferometer over the long optical paths sometimes used in length measurement.

43-8 Michelson's Interferometer and Light Propagation


In Section 40-4 we presented Einstein's hypothesis, now well verified, that in free
space light is propagated with the same speed c no matter what the relative velocity
of the source and the observer may be. We pointed out that this hypothesis contra-
dicted the views of nineteenth-century physicists regarding wave propagation. It
was difficult for these physicists, trained as they were in the classical physics of the
time, to believe that a wave could be propagated without a medium. If such a
medium could be established, the speed c of light would naturally be construed as the
speed with respect to that medium, just as the speed of sound always refers to a medium
such as air.
Although no medium for light propagation was obvious, the physicists postulated
one, called the ether t and hypothesized that its properties were such that it was un-
detectable by ordinary means such as weighing.
In 1881 (24 years before Einstein's hypothesis) A. A. Michelson set himself the task
of forcing the ether, assuming that it existed, to submit to direct physical verification.
In particular, Michelson, later joined by E.W. Morley, tried to measure the speed u
with which the earth moves through the ether. Michelson's interferometer was their
instrument of choice for this now-famous Michelson-Morley experiment.
The earth together with the interferometer moving with velocity u through the
ether is equivalent to the interferometer at rest with the ether streaming through it
with velocity -u, as shown in Fig. 43-21. Consider a wave moving along the path
MM 1M and one moving along MM 2M. The first corresponds classically to a man
* Scientific American, p.75, December 1960.
t More fully, the luminiferous (or light-carrying) ether.
Sec. 43-8 MICHELSON'S INTERFEROMETER AND LIGHT PROPAGATION 1093

Fig. 43-21 The "ether" is streaming with velocity - u through Michelson's interferom-
eter. The wave speeds shown are on the basis of the (incorrect) ether hypothesis.

rowing a boat a distanced downstream and the same distance upstream; the second
corresponds to rowing a boat a distanced across a stream and back.
On the ether hypothesis the speed of light on the path MM1 is c + u; on the return
path M 1M it is c - u. The time required for the complete trip is
d 2c 2d 1
l1 = - - +c--d-u =
c+u
d2- - -2
c - u c 1 - (u/c) 2

The speed of light, on the ether hypothesis, for path MM 2 is 2 - Vc


u 2 , as Fig. 43-21
suggests. This same speed holds for the return path M2M, so that the time required
for this complete path is
2d 2d 1
t?= =- .
- vc 2 - u 2 c v1 - (u/c) 2
The difference of time for the two paths is
l:lt = t1 - t2

Assuming u/ c << 1, we can expand the quantities in the square brackets by using the
binomial theorem, retaining only the first two terms. This leads to

l:lt = 2cd { [ 1 + (~Y+ ... J- [ + ~ (~Y + ... J}


1

= 2d
c
{!2 ('.!!:)c 2
} = du2.
c3 (43-16)
1094 INTERFERENCE Chap. 43
Now let the entire interferometer be rotated through 90°. This will interchange
the roles of the two light paths, MM 1M now being the "cross-stream" path and
MM2M the "down- and upstream" path. The time difference between the two waves
entering the eye is also reversed; this changes the phase difference between the com-
bining waves and alters the positions of the interference maxima. The experiment
consists of looking for a shift of the interjerence fringes as the apparatus is rotated.
The change in time difference is 2i:it, which corresponds to a fringe shift of 2i:it/T
where T ( = A/c) is the period of vibration of the light. The expected maximum shift
in the number of fringes on a 90° rotation (see Eq. 43-16) is

i:iN = !~ =
T
2i:itc = 2d
"'A X
(!±)c 2
• (43-17)

In the Michelson-Morley interferometer let d = 11 meters (obtained by multiple


reflection in the interferometer) and )\ = 5.9 X 10-7 meter. If u is assumed to be
roughly the orbital speed of the earth, then u/c""' 10-4 , The expected maximum
fringe shift when the interferometer is rotated through 90° is then

i:iN = 2d
A
(!±)
c
2
= (2)(11 meters) (I0- 4) 2 = 0.4.
5.9 X 10- 7 meter
Even though a shift of only about 0.4 of a fringe was expected, Michelson and Morley
were confident that they could observe a shift of 0.01 fringe. They found from their
experiment, however, that there was no observable f ririge shift!
The analogy between a light wave in the supposed ether and a boat moving in
water, which seemed so evident in 1881, is simply incorrect. The derivation based on
this analogy is incorrect for light waves. When the analysis is carried through on
Einstein's hypothesis, the observed negative result is clearly predicted, the speed of
light being c for all paths. The motion of the earth around the sun and the rotation
of the interferometer have, in Einstein's view, no effect whatever on the speed of the
light waves in the interferometer.
It should be made clear that although Einstein's hypothesis is completely con-
sistent with the negative result of the Michelson-Morley experiment this experiment
standing alone cannot serve as a proof for Einstein's hypothesis. Einstein said that
no number of experiments, however large, could prove him right but that a single
experiment could prove him wrong. Our present-day belief in Einstein's hypothesis
rests on consistent agreement in a large number of experiments designed to test it.
The "single experiment" that might prove Einstein wrong has never been found.
Chap. 43 QUESTIONS 1095

QUESTIONS

1. Is Young's experiment an interference experiment or a diffraction experiment,


or both?
2. Do interference effects occur for sound waves? Recall that sound is a longitudinal
wave and that light is a transverse wave.
3. In Young's double-slit interference experiment, using a monochromatic laboratory
light source, why is screen A in Fig. 43-1 necessary? What would happen if one gradually
enlarged the hole in this screen?
4. Describe the pattern of light intensity on screen C in Fig. 43-5 if one slit is covered
with a red filter and the other with a blue filter, the incident light being white.
5. Is coherence important in reflection and refraction?
6. Define carefully, and distinguish between, the angles eand <J, that appear in Eq. 43-10.
7. If one slit in Fig. 43-5 is covered, what change occurs in the intensity of light in the
center of the screen?
8. What changes occur in the pattern of interference fringes if the apparatus of Fig. 43-5
is placed under water?
9. What are the requirements for a maximum intensity when viewing a thin film by
transmitted light?
10. In a Newton's rings experiment, is the central spot, as seen by reflection, dark or
light? Explain.
11. Why must the film of Fig. 43-14 be "thin" for us to see an interference pattern of the
type described?
12. Why do coated lenses (see Example 5) look purple by reflected light?
13. A person wets his eyeglasses to clean them. As the water evaporates he notices that
for a short time the glasses become markedly more nonreflecting. Explain.
14. A lens is coated to reduce reflection, as in Example 5. What happens to the energy
that had previously been reflected? Is it absorbed by the coating?
15. Very small changes in the angle of incidence do not change the interference con-
ditions much for "thin" films but they do change them for ''thick" films. Why?
16. The directional characteristics of a certain radar antenna as a receiver of radiation
are known. What can be said about its directional characteristics as a transmitter?
17. A person in a dark room, looking through a small window, can see a second person
standing outside in bright sunlight. The second person cannot see the first person. Is this
a failure of the principle of optical reversibility? Assume no absorption of light.
18. Why is it necessary to rotate the interferometer in the Michelson-Morley experi-
ment?
19. How is the negative result of the Michelson-Morley experiment interpreted ac-
cording to Einstein's theory of relativity?
20. If interference between light waves of different frequencies is possible, one should
observe light beats, just as one obtains sound beats from two sources of sound with slightly
different frequencies. Discuss how one might experimentally look for this possibility.
21. In Young's double-slit experiment suppose that screen A in Fig. 43-1 contained
two very narrow parallel slits instead of one. (a) Show that if the spacing between these
slits is properly chosen the interference fringes can be made to disappear. (b) Under thyse
conditions, would you call the beams emerging from slits 8 1 and 8 2 in screen B coherent?
They do not produce interference fringes. (c) Discuss what would happen to the inter-
ference fringes in the case of a single slit in screen A if the slit width were gradually in-
creased.
1096 INTERFERENCE Chap.43

PROBLEMS

1. Design a double-slit arrangement that will produce interference fringes 1 ° apart on a


distant screen. Assume sodium light (;\ = 5890 A).
~ , Two point sources in Fig. 43-22 emit coherent waves. Show that curves, such as
that given, over which the phase difference for rays r 1 and r2 is a constant are hyperbolas.
Extend the analysis to three dimensions. (Hint: A constant phase difference implies a
constant difference in length between r 1 and r2.)

Fig. 43-22

,,,'3. A double-slit arrangement produces interference fringes for sodium light (X = 5890 A)
that are 0.20° apart. What is the angular fringe separation if the entire arrangement is
immersed in water?
4. A double-slit arrangement produces interference fringes for sodium light (X = 5890 A)
that are 0.20° apart. For what wavelength would the angular separation be 10% greater?
&, A thin flake of mica (n = 1.58) is used to cover one slit of a double-slit arrangement.
The central point on the screen is occupied by what used to be the seventh bright fringe.
If X = 5500 A, what is the thickness of the mica?
6. Sodium light (X = 5890 A) falls on a double slit of separation d = 2.0 mm. D in
Fig. 43-5 is 4 cm. What per cent error is made in locating the tenth bright fringe if it is
not assumed that D » d?
,,7'., Sodium light (X = 5890 A) falls on a double slit of separation d = 0.20 mm. A thin
lens (f = + 1.0 meter) is placed near the slit as in Fig. 43-6. What is the linear fringe
separation on a screen placed in the focal plane of the lens?
Chap. 43 PROBLEMS 1097
··. 8. S 1 and S2 in Fig. 43-23 are effective point sources of radiation, excited by the same
oscillator. They are coherent and in phase with each other. Placed 4.0 meters apart,
they emit equal amounts of power in the form of LO-meter wavelength electromagnetic
waves. (a) Find the positiorn, of the first (that is, the nearest), the second, and the third
maxima of the received signal, as the detector is moved out along Ox. (b) Is the intensity
at the nearest minimum equal to zero? Justify your answer.

r 82
d
l ____ =t------X

Fig. 43-23

9. One of the slits of a double-slit system is wider than the other, so that the amplitude
of the light reaching the central part of the screen from one slit, acting alone, is twice that
from the other slit, acting alone. Derive an expression for Io in terms of e, corresponding
to Eqs. 43-llb and c.
10. Show that the half-width D.0 of the double-slit interference fringes (see arrow in Fig.
43-9) is given by
'A
D.0 =-
2d

if e is small enough so
that sine ,...., e.
11. Find the sum of the following quantities (a) by the vector method and (b) analyti-
cally:
Y1 = 10 sin wt

Y2 = 8 sin (wt + 30°).


12. Add the following quantities graphically, using the vector method:

Y1 = 10 sin wt

Y2 = 15 sin (wt + 30°)


y3 = 5 sin (wt - 45°).

13. The frequency width of a standard television channel is 4 X 106 cycles/sec. If it


were possible to transmit television signals in the visible portion of the electromagnetic
spectrum, possibly by using optical maser or laser techniques, how many such standard
channels would be available for allocation in the visible 8pectrum? Assume that this
srectrum extends from 4000 to 7000 A.
·14. In a Newton's rings experiment the radius of curvature R of the lens is 5.0 meters
andits diameter is 2.0 cm. (a) How many rings are produced? (b) How many rings would
be seen if the arrangement were immersed in water (n = 1.33)? Assume that 'A = 5890 A.
15. The diameter of the tenth bright ring in a Newton's rings apparatus changes from
1.40 to 1.27 cm as a liquid is introduced between the lens and the plate. Find the index
of refraction of the liquid.
1098 INTERFERENCE Chap. 43
16. A broad source of light (A = 6800 A) illuminates normally two glass plates 12 cm
long that touch at one end and are separated by a wire 0.048 mm in diameter at the other
(Fig. 43-24). How many bright fringes appear over the 12-cm distance?

Fig. 43-24

17. A thin film 4 X 10-5 cm thick is illuminated by white light normal to its surface.
Its index of refraction is 1.5. What wavelengths within the visible spectrum will be
intensified in the reflected beam?
18. White light reflected at perpendicular incidence from a soap film has, in the visible
spectrum, an interference maximum at 6000 A and a minimum at 4500 A, with no mini-
mum in between. If n = l.33 for the film, what is the film thickness, assumed uniform?
19. A plane wave of monochromatic light falls normally on a uniformly thin film of oil
which covers a glass plate. The wavelength of the source can be varied continuously.
Complete destructive interference of the reflected light is observed for wavelengths of
5000 and 7000 A and for no other wavelengths in between. If the index of refraction of
the oil is l.30 and that of the glass is 1.50, find the thickness of the oil film.
20. In Example 5 assume that there is zero reflection for light of wavelength 5500 A
at normal incidence. Calculate the factor by which the reflection is diminished by the
coating at 4500 and at 6500 A.
21. A plane monochromatic light wave in air falls at normal incidence on a thin film of
oil which covers a glass plate. The wavelength of the source may be varied continuously.
Complete destructive interference in the reflected beam is observed for wavelengths of
5000 and 7000 A and for no other wavelength in between. The index of refraction of glass
is 1.50. Show that the index of refraction of the oil must be less than 1.50.
22. White light reflected at perpendicular incidence from a soap bubble has, in the visi-
ble spectrum, a single interference maximum (at A = 6000 A) and a single minimum at the
violet end of the spectrum. If n = 1.33 for the film, calculate its thickness.
23. If mirror M 2 in Michelson's interferometer is moved through 0.233 mm, 792 fringes
are counted. What is the wavelength of the light?
24. A thin film with n = 1.40 for light of wavelength 5890 A is placed in one arm of a
Michelson interferometer. If a shift of 7.0 fringes occurs, what is the film thickness?
25. (a) What is the wavelength of the orange-red line of krypton-86 in angstrom units?
Take 1 A = 10- 10 meter exactly. (b) Does this question really make sense, in view of the
fact that the meter is defined in terms of this wavelength? Explain.
26. A Michelson interferometer is used with a sodium discharge tube as a light source.
The yellow sodium light consists of two wavelengths, 5890 and 5896 A. It is observed
that the interference pattern disappears and reappears periodically as one moves mirror
M2 in Fig. 43-20. (a) Explain this effect. (b) Calculate the change in path difference
between two successive reappearances of the interference pattern.
Diffraction
CHAPTER 44

44-1 Introduction
Diffraction, which is illustrated in Fig. 42-3, is the bending of light around
an obstacle such as the edge of a slit. 'Ne can see the diffraction of light by
looking through a crack between two fingers at a distant light source such
as a tubular neon sign or by looking at a street light through a cloth um-
brella. Usually diffraction effects are small and must be looked for carefully.
Also, most sources of light have an extended area so that a diffraction pat-
tern produced by one point of the source will overlap that produced by
another. Finally, common sources of light are not monochromatic. The
patterns for the various wavelengths overlap and again the effect is less
apparent.
Diffraction was discovered by Francesco Maria Grimaldi (1618-1663),
and the phenomenon was known both to Huygens (1629-1695) and to New-
ton (1642-1727). Newton did not see in it any justification for a wave
theory for light. Huygens, although he believed in a wave theory, did not
believe in diffraction! He imagined his secondary wavelets to be effective
only at the point of tangency to their common envelope, thus denying the
possibility of diffraction. In his words:
And thus we see the reasons why light . . . proceeds only in straight lines in
such a way that it does not illuminate any object except when the path from the
source to the object is open along such a line.
Fresnel (1788-1827) correctly applied Huygens' principle (which is called
the Huygens-Fresnel principle in Europe) to explain diffraction. In these
early days the light waves were believed to be mechanical waves in an all-
1099
1100 DIFFRACTION Chap. 44

Fig. 44-1 Light is diffracted at the


aperture in screen B and illuminates
screen C. The intensity at Pis found
by dividing the wavefront at B into
elementary radiators dS and combin-
ing their effects at P.

pervading ether. \Ve have seen (Section 39-5) how Maxwell (1831-1879)
showed that light waves were not mechanical in nature but electromagnetic.
Einstein (1879-1955) rounded out our modern view of light waves by elim-
inating the need to postulate an ether (see Section 40-4).
Figure 44-1 shows a general diffraction situation. Surface A is a wavefront
that falls on B, which is an opaque screen containing an aperture of arbitrary
shape; C is a diffusing screen that receives the light that passes through this
aperture. This pattern of light intensity on C can be calculated by sub-
dividing the wavefront into elementary areas dS, each of which becomes a
source of an expanding Huygens' wavelet. The light intensity at an arbi-
trary point Pis found by superimposing the wave disturbances (that is, the
E vectors) caused by the wavelets reaching P from all these elementary
radiators.
The wave disturbances reaching P differ in amplitude and in phase be-
cause (a) the elementary radiators are at varying distances from P, (b) the
light leaves the radiators at various angles to the normal to the wavefront
(&ee p. 1020), and (c) soni.e radiators are blocked by screen B; others are not.
Diffraction calculations-simple in principle-may become difficult in prac-
tice. The calculation must be repeated for every point on screen Cat which
we wish to know the light intensity. We followed exactly this program in
calculating the double-slit intensity pattern in Section 43-3. The calcula-
tion there was simple because we assumed only two elementary radiators,
the two narrow slits.
Figure 44-2a shows the general case of Fresnel diffraction, in which the
light source and/ or the screen on which the diffraction pattern is displayed
are a finite distance from the diffracting aperture; the wavefronts that fall
on the diffracting aperture in this case and that leave it to illuminate any
point P of the diffusing screen are not planes; the corresponding rays are
not parallel.
Sec. 44-1 INTRODUCTION 1101

A simplification results if source S and screen C are moved to a large


distance from the diffraction aperture, as in Fig. 44-26. This limiting case
is called Fraunhofer diffraction. The wavefronts arriving at the diffracting
aperture from the distant source S are planes, and the rays associated with
these wavefronts are parallel to each other. Fraunhofer conditions can
be established in the laboratory by using two converging lenses, as in

Fig. 44-2 (a) Fresnel diffraction. (b) Source S and screen Care moved to a large dis-
tance, resulting in Fraunhofer diffraction. (c) Fraunhofer diffraction conditions produced
by lenses, leaving source S and screen C in their original positions.
1102 DIFFRACTION Chap. 44
Fig. 44-2c. The first of these converts the diverging wave from the
source into a plane wave. The second lens causes plane waves leaving
the diffracting aperture to converge to point P. All rays that illuminate
P will leave the diffracting aperture parallel to the dashed line Px drawn
from P through the center of this second ( thin) lens. We assumed
Fraunhofer conditions for Young's double-slit experiment in Section 43-1
(see Fig. 43-5).
Although Fraunhofer diffraction is a limiting case of the more general
Fresnel diffraction, it is an important limiting case and is easier to handle
mathematically. This book deals only with Fraunhofer diffraction.
44-2 Single Slit
Figure 44-3 shows a plane wave falling at normal incidence on a long
narrow slit of width a. Let us focus our attention on the central point P 0 of
screen C. The rays extending from the slit to P 0 all have the same optical
path lengths, as we saw in Section 42-5. Since they are in phase at the
plane of the slit, they will still be in phase at P 0 , and the central point of
the diffraction pattern that appears on screen Chas a maximum intensity.

Fig. 44-3 Conditions at the central maximum of the diffraction pattern. The slit extends
a distance above and below the figure, this distance being much greater than the slit
width a.
Sec. 44-2 SINGLE SLIT 1103

Fig. 44-4 Conditions at the first minimum of the diffraction pattern.

We now consider another point on the screen. Light rays which reach P 1
in Fig. 44-4 leave the slit at an angle () as shown. Ray r 1 originates at the
top of the slit and ray r 2 at its center. If() is chosen so that the distance bb'
in the figure is one-half a wavelength, r 1 and r 2 will be out of phase and will
produce no effect at P 1 . * In fact, every ray from the upper half of the slit
will be canceled by a ray from the lower half, originating at a point a/2 below
the first ray. The point P 1 , the first minimum of the diffraction pattern,
will have zero intensity (compare Fig. 42-3).
The condition shown in Fig. 44-4 is
a A
- sin() = -,
2 2
or a sin() = A. (44-1)
As we stated earlier (see Fig. 42-1), the central maximum becomes wider as
the slit is made narrower. If the slit width is as small as one wavelength
(a = }..), the first minimum occurs at() = 90°, which implies that the central
maximum fills the entire forward hemisphere. We assumed a condition ap-
* Whatever phase relation exists between r1 and r2 at the plane represented by the
sloping dashed line in Fig. 44-4 that passes through b' also exists at P1, not being affected
by the lens (see Section 42-5).
1104 DIFFRACTION Chap. 44

Fig. 44-5 Conditions at the second minimum of the diffraction pattern.

proaching this in our discussion of Young's double-slit interference experi-


ment in Section 43- 1.
In Fig. 44-5 the slit is divided into four equal zones, with a ray leaving the
top of each zone. Let e be chosen so that the distance bb' is one-half a wave-
length. Rays r 1 and r 2 will then cancel at P 2 . Rays r 3 and r 4 will also be
half a wavelength out of phase and will also cancel. Consider four other
rays, emerging from the slit a given distance below the four rays above. The
two rays below r 1 and r 2 will cancel uniquely, as will the two rays below r 3
and r4 . We can proceed across the entire slit and conclude again that no
light reaches P 2 ; we have located a second point of zero intensity.
The condition described (see Fig. 44-5) requires that
a >-.
- sine=-,
4 2
or a sine = 2>-..
By extension, the general formula for the minima in the diffraction pattern
on screen C is
a sine = m>-. m = 1, 2, 3, ... (minima). (44--2)
Sec. 44-3 SINGLE SLIT--QUALITATIVE 1105

There is a maximum approximately halfway between each adjacent pair of


mm1ma. The student should consider the simplification that has resulted
in this analysis by examining Fraunhofer (Fig. 44-2c) rather than Fresnel
conditions (Fig. 44-2a).
..... Example 1. A slit of width a is illuminated by white light. For what value of
a will the first minimum for red light (A = 6500 A) fall at() = 30°?
At the first minimum we put m = 1 in Eq. 44-2. Doing so and solving for a yields
= m>-. = (1)(6500 A) = 13 OOO A
a sin () sin 30° ' ·
Note that the slit width must be twice the wavelength in this case.
Example 2. In Example 1 what is the wavelength>-.' of the light whose first dif-
fraction maximum (not counting the central maximum) falls at fJ = 30°, thus co-
inciding with the first minimum for red light?
This maximum is about halfway between the first and second minima. It can be
found without too much error by putting m = 1.5 in Eq. 44-2, or
a sin f) ""' 1.5>-.'.
From Example 1, however, a sin fJ = >-..

Dividing gives >-.' = ~ = 55 00 A = 4300 A.


1.5 1.5
Light of this color is violet. The second maximum for light of wavelength 4300 A will
always coincide with the first minimum for light of wavelength 6500 A, no matter
what the slit width. If the slit is relatively narrow, the angle fJ at which this overlap
occurs will be relatively large. ..,.

44-3 Single Slit-Qualitative


Figure 44-6 shows a slit of width a divided into N parallel strips of width
Lix. Each strip acts as a radiator of Huygens' wavelets and produces a
characteristic wave disturbance at point P, whose position on the screen,
for a particular arrangement of apparatus, can be described by the angle fJ.
If the strips are narrow enough-which we assume-all points on a given
strip have essentially the same optical path length to P, and therefore all
the light from the strip will have the same phase when it arrives at P. The
amplitudes LiE 0 of the wave disturbances at P from the various strips may
be taken as equal if fJ in Fig. 44-6 is not too large.
We limit our considerations to points that lie in, or infinitely close to, the plane of
Fig. 44-6. It can be shown that this procedure is valid for a slit whose length is
much greater than its width a. We made this same assumption tacitly both earlier
in this chapter and in Chapter 43; see Figs. 43-5 and 44-3, for example.
The wave disturbances from adjacent strips have a constant phase dif-
ference Li¢ between them at P given by
phase difference path difference
21r >-.

or 2 (44-3)
Li¢ = ( ; ) (Lix sin fJ),
1106 DIFFRACTION Chap. 44

Fig. 44-6 A slit of width a is divided into N strips of width ~x. The insert shows condi-
tions at the second strip more clearly. In the differential limit the slit is divided into an
infinite number of strips (that is, N ----+ co) of differential width dx. For clarity in this
and the following figure, we take N = 18.

where Lix sin() is, as the figure insert sho,vs, the path difference for rays origi-
nating at the top edges of adjacent strips. Thus, at P, N vectors with the
same amplitude LiE0 , the same frequency, and the same phase difference Li</)
between adjacent members combine to produce a resultant disturbance. We
ask, for various values of Li</) [that is, for various points Pon the screen, cor-
responding to various values of e (see Eq. 44-3)], what is the amplitude Ee
of the resultant wave disturbance? We find the answer by representing the
individual wave disturbances LiE0 by phasors and calculating the resultant
phasor amplitude, as described in Section 43-4.
At the center of the diffraction pattern () equals zero, and the phase shift
between adjacent strips (see Eq. 44-3) is also zero. As Fig. 44-7a shows,
the phasor arrows in this case are laid end to end and the amplitude of the
resultant has its maximum value Em. This corresponds to the center of the
central maximum.
Sec. 44-3 SINGLE SLIT-QUALITATIVE 1107

As we move to a value of e other than zero, f}.rp assumes a definite nonzero


value (again see Eq. 44-3), and the array of arrows is now as shown in Fig.
44-7b. The resultant amplitude E 0 is less than before. Note that the length
of the "arc" of small arrows is the same for both figures and indeed for all
figures of this series. Ase increases further, a situation is reached (Fig. 44-7c)
in which the chain of arrows curls around through 360°, the tip of the last
arrow touching the foot of the first arrow. This corresponds to E 0 = 0, that
is, to the first minimum. For this condition the ray from the top of the slit
(1 in Fig. 44-7c) is 180° out of phase with the ray from the center of the slit
(!N in Fig. 44-7c). These phase relations are consistent with Fig. 44-4,
which also represents the first minimum.
Ase increases further, the phase shift continues to increase, and the chain
of arrows coils around through an angular distance greater than 360°, as in
Fig. 44-7d, which corresponds to the first maximum beyond the central
maximum. This maximum is much smaller than the central maximum.
In making this comparison, recall that the arrows marked E 0 in Fig. 44-7
correspond to the amplitudes of the wave disturbance and not to the in-

(a)

Fig. 44-7 Conditions at (a) the


central maximum, (b) a direction
slightly removed from the central
maximum, (c) the first minimum,
and (d) the first maximum beyond
the central maximum for single-
slit diffraction. This figure corre-
sponds to N = 18 in Fig. 44---6. Eo =O

@ I

(d)
1108 DIFFRACTION Chap. 44
tensity. The amplitudes must be squared to obtain the corresponding rela-
tive intensities (see Eq. 43-7).

44-4 Single Slit-Quantitative


The "arc" of small arrows in Fig. 44-8 shows the phasors representing, in
amplitude and phase, the wave disturbances that reach an arbitrary point P
on the screen of Fig. 44-6, corresponding to a particular angle e. The
resultant amplitude at P is Ee. If we divide the slit of Fig. 44-6 into infi-
nitesimal strips of width dx, the arc of arrows in Fig. 44-8 approaches the arc
of a circle, its radius R being indicated in that figure. The length of the arc
is Em, the amplitude at the center of the diffraction pattern, for at the center
of the pattern the wave disturbances are all in phase and this "arc" becomes
a straight line as in Fig. 44-7 a.
The angle <p in the lower part of Fig. 44-8 is revealed as the difference in
phase between the infinitesimal vectors at the left and right ends of the arc
Em. This means that <p is the phase difference between rays from the top and
the bottom of the slit of Fig. 44-6. From geometry we see that <fJ is also the
angle between the two radii marked R in Fig. 44-8. From this figure we can
write

Ee = 2R sin!.
2

In radian measure </J, from the figure, is

Em
<p = -·
R

\a
a\
\
<P
\
\ R
\
\
\ Fig. 44-8 A construction used to
R \ calculate the intensity in single-slit
\
\ diffraction. The situation corre-
\ sponds to that of Fig. 44-7.
\
\
\
\
\
\
Sec. 44-4 SINGLE SLIT-QUANTITATIVE 1109

Em. c/J
Combining yields Eo = -sin-,
cp/2 2

sma
or Eo = Em--, (44-4)
a

cp
in which a=-· (44-5)
2

From Fig. 44-6, recalling that cp is the phase difference between rays from
the top and the bottom of the slit and that the path difference for these rays
is a sin (J, we have

phase difference path difference


2~ A

or cp = ( 2; ) (a sin 0).
Combining with Eq. 44-5 yields
cp ~a .
a= - = -sm8. (44-6)
2 A

Equation 44-4, taken together with the definition of Eq. 44-6, gives the
amplitude of the wave disturbance for a single-slit diffraction pattern at any
angle 0. The intensity Io for the pattern is proportional to the square of the
amplitude, or
Io = Im ( si: a Y· (
44-7)

For convenience we display together, and renumber, the formulas for the
amplitude and the intensity in single-slit diffraction.

sma
[Eq. 44-4] Eo = Em-- (44-Sa)
a
single-
[Eq. 44-7] Io= Im (sin
-- ay slit (44-Sb)
a
diffraction
~a
[Eq. 44-6] a (= !c/J) = -sin() (44-Sc)
A J
Figure 44-9 shows plots of Io for several values of the ratio a/A. Note that
the pattern becomes narrower as a/A is increased; compare this figure with
Figs. 42-1 and 42-3.
Minima occur in Eq. 44-Sb when

a= m~ m = 1, 2, 3, .... (44-9)
1110 DIFFRACTION Chap. 44

1.0

0.8
->,
.vi a=X
0.6

0.4
-c:
Q)

c:
Q)
>
:;:::;
<ti
ai
0:::
0.2 (a)

15 10 5 0 5 10 15
(J, degrees

15 10 5 0 10 15
(J, degrees

a= lOX

(c)

15 10 5 0 5 10 15
(J, degrees

Fig. 44-9 The relative intensity in single-slit diffraction for three values of the ratio a/'A.
The arrow in (b) shows the half-width !1(J of the central maximum.

Combining with Eq. 44-Sc leads to


a sine = m'J\ m = 1, 2, 3, ... (minima),
which is the result derived in the preceding section (Eq. 44-2). In that
section, however, we derived only this result, obtaining no quantitative in-
formation about the intensity of the diffraction pattern at places in which
it was not zero. Here (Eqs. 44-8) we have complete intensity information.
Sec. 44-4 SINGLE SLIT-QUANTITATIVE 1111

~ Example 3. Intensities of the secondary diffraction maxima. Calculate, approx-


imately, the relative intensities of the secondary maxima in the single-slit Fraunhofer
diffraction pattern.
The secondary maxima lie approximately halfway between the minima and are
found (compare Eq. 44-9) from
a"-' (m + !}ir m = 1, 2, 3, ....
Substituting into Eq. 44-8b yields

I = I [sin (m + !)11'] 2 ,
o m (m + !)11'
Io 1
which reduces to
Im - (m+!)211'2
This yields, form = 1, 2, 3, ... , Io/Im = 0.045, 0.016, 0.0083, etc. The successive
maxima decrease rapidly in intensity.
Example 4. Width of the central diffraction maximum. Derive the half-width
fl()of the central maximum in a single-slit Fraunhofer diffraction (see Fig. 44-9b).
The half-width is the angle between the two points in the pattern where the intensity
is one-half that at the center of the pattern.
Point x in Fig. 44-9b is so chosen that Io = !Im, or, from Eq. 44-8b,

! = (sin ax) 2 •
2 O'.x

This equation cannot be solved analytically for ax, It can be solved graphically,
as accurately as one wishes, by plotting the quantity (sin ax/ax) 2 as ordinate versus
ax as abscissa and noting the value of ax at which the curve intersects the line "one-
half" on the ordinate scale (see Problem 5). However, if only an approximate answer
is desired, it is often quicker to use trial-and-error methods.
We know that a equals 11' at the first minimum; we guess that ax is perhaps 11'/2
( = 90° = 1.57 radians). Trying this in Eq. 44-8b yields
!y_ = [sin (1r/2)] 2 = 0 4
Im 1r/2 ..

This intensity ratio is less than 0.5, so that ax must be less than 90°. After a few more
trials we find easily enough that
ax = 1.40 radians = 80°

does yield a ratio close to the correct value of 0.5.


We now use Eq. 44-Sc to find the corresponding angle fJ:

O'.x = 11"a
~
· ()
Sln x = 1.40 ,

or, noting that a/'A. = 5 for Fig. 44-9b,

. 8x = l. 40t.. = l. 40 = 0 .0892 .
sin
11"a 511'

The half-width tl8 of the central maximum (see Fig. 44-9b) is given by
tl8 = 2Bx = 2 sin- 1 0.0892 = 2 X 5.1 ° = 10.2°,
which is in agreement with the figure.
1112 DIFFRACTION Chap. 44
44-5 Diffraction at a Circular Aperture
Diffraction will occur when a wavefront is partially blocked off by an
opaque object such as a metal disk or an opaque screen containing an aper-
ture. Here we consider diffraction at a circular aperture of diameter d, the
aperture constituting the boundary of a circular lens.
Our previous treatment of lenses was based on geometrical optics, diffrac-
tion being specifically assumed not to occur. A rigorous analysis would be
based from the beginning on wave optics, since geometrical optics is always
an approximation, although often a good one. Diffraction phenomena would
emerge in a natural way from such a wave-optical analysis.
Figure 44-10 shows the image of a distant point source of light (a star)
formed on a photographic film placed in the focal plane of a converging lens.
It is not a point, as the (approximate) geometrical optics treatment suggests,
but a circular disk surrounded by several progressively fainter secondary
rings. Comparison with Fig. 42-3c leaves little doubt that we are dealing
with a diffraction phenomenon in which, however, the aperture is a circle
rather than a long narrow slit. The ratio d/'A, where dis the diameter of the
lens (or of a circular aperture placed in front of the lens), determines the scale
of the diffraction pattern, just as the ratio a/>.. does for a slit.
Analysis shows that the first minimum for the diffraction pattern of a
circular aperture of diameter d, assuming Fraunhofer conditions, is given by
A
sin(} = 1.22 -· (44-10)
d
This is to be compared with Eq. 44-1, or
A
sin(} = -,
a
which locates the first minimum for a long narrow slit of width a. The factor
1.22 emerges from the mathematical analysis when we integrate over the
elementary radiators into which the circular aperture may be divided.

Fig. 44-10 The image of a star formed by a


converging lens is a diffraction pattern. Note the
central maximum, sometimes called the Airy disk
(after Sir George Airy, who first solved the problem
of diffraction at a circular aperture in 1835), and
the circular secondary maximum. Other second-
ary maxima occur at larger radii but are too faint
to be seen.
Rayleigh's
criterion

~,I\,~ o ,,=,,~r:==--r,/ I \,--r=::+'-../s-» ~ I {J .-=-.~ .c::;::,, / 1 ' c;::--- p2"j= >,..-:::--, ()

(a) ~Jo~ (b) ~ fo ~ (c) h-lo~


Fig. 44-11 The images of two distant point objects are formed by a converging lens whose diameter ( = 10 cm) is 200,000 times the effective
wavelength ( = 5000 A). Sketches of the images as they appear in the focal plane of the lens are shown with the corresponding intensity plots below
them. (a) The angular separation of the objects (see vertical ticks) is so small that the images are not resolved. (b) The objects are farther apart
and the images meet Rayleigh's criterion for resolution. (c) The objects are still farther apart and the images are well resolved.
1114 DIFFRACTION Chap.44
In actual lenses the image of a distant point object will be somewhat larger than
that shown in Fig. 44-10 and may not have radial symmetry. This is caused by the
various lens "defects" mentioned on p. 1061. However, even if all of these defects
could be eliminated by suitable shaping of the lens surfaces or by introducing cor-
recting lenses, the diffraction pattern of Fig. 44-10 would remain. It is an inherent
property of the lens aperture and of the wavelength of light used.
The fact that lens images are diffraction patterns is important when we
wish to distinguish two distant point objects whose angular separation is
small. Figure 44-11 shows the visual appearances and the corresponding
intensity patterns for two distant point objects with small angular separa-
tions and approximately equal central intensities. In (a) the objects are
not resolved; that is, they cannot be distinguished from a single point
object. In (b) they are barely resolved and in (c) they are fully resolved.
In Fig. 44-llb the angular separation of the two point sources is such that
the maximum of the diffraction pattern of one source falls on the first mini-
mum of the diffraction pattern of the other. This is called Ray"leigh's crite-
rion. This criterion, though useful, is arbitrary; other criteria for deciding
when two objects are resolved are sometimes used. From Eq. 44-10, two
objects that are barely resolvable by Rayleigh's criterion must have an angu-
lar separation OR of
1 l.22X
()R = sin- - - ·
d
Since the angles involved are rather small, we can replace sin OR by OR, or
X
()R = 1.22-· (44-11)
d
If the angular separation 8 between the objects is greater than 8R, we can
resolve the two objects; if it is less, we cannot .
.,.. Example 5. A converging lens 3.0 cm in diameter has a focal length f of 20 cm.
(a) What angular separation must two distant point objects have to satisfy Rayleigh's
criterion? Assume that X = 5500 A.
From Eq. 44-11,

(JR = 1.22 ~ = (1.22)(5.5 X 10-1 meter) = 2.2 X 10-6 radian.


d 3.0 X 10-2 meter
(b) How far apart are the centers of the diffraction patterns in the focal plane of
the lens? The linear separation is
x = f8 = (20 cm)(2.2 X 10-5 radian) = 44,000 A.
This is 8.0 wavelengths of the light employed.
When one wishes to use a lens to resolve objects of small angular separa-
tion, it is desirable to make the central disk of the diffraction pattern as
small as possible. This can be done (see Eq. 44-11) by increasing the lens
diameter or by using a shorter wavelength. One reason for constructing
large telescopes is to produce sharper images so that celestial objects can be
examined in finer detail. The images are also brighter, not only because the
Sec. 44-6 DOUBLE SLIT 1115

energy is concentrated into a smaller diffraction disk but because the larger
lens collects more light. Thus fainter objects, for example, more distant
stars, can be seen.
To reduce diffraction effects in microscopes we often use ultraviolet light,
which, because of its shorter wavelength, permits finer detail to be examined
than would be possible for the same microscope operated with visible light.
We shall see in Chapter 48 that beams of electrons behave like waves under
some circumstances. In the el,ectron microscope such beams may have an
effective wavelength of 0.04 A, of the order of 105 times shorter than visible
light (A,...._, 5000 A). This permits the detailed examination of tiny objects
like viruses. If a virus were examined with an optical microscope, its struc-
ture would be hopelessly concealed by diffraction.

44-6 Double Slit


In Young's double-slit experiment (Section 43-1) we assume that the
slits are arbitrarily narrow (that is, a<<>.), which means that the central
part of the diffusing screen was uniformly illuminated by the diffracted waves
from each slit. When such waves interfere, they produce fringes of uniform
intensity, as in Fig. 43-9. This idealized situation cannot occur with actual
slits because the condition a<< A cannot usually be met. Waves from the
two actual slits combining at different points of the screen will have intensi-
ties that are not uniform but are governed by the diffraction pattern of a
single slit. The effect of relaxing the assumption that a<< X i11 Young's ex-
periment is to leave the fringes relatively unchanged in location but to alter
their intensities.
The interference pattern for infinitesimally narrow slits is given by Eq.
43-llb and c or, with a small change in nomenclature,

le.int = Im,int cos2 {3, (44-12)

1rd
where {3 = - sin O (44-13)
X
in which d is the distance between the center-lines of the slits.
The intensity for the diffracted wave from either slit is given by Eqs. 44-8b
and c, or, with a small change in nomenclature,

le,dif = lm,dif ( ~
sin a)
2
' (44-14)

,ra
where a= -sin 0. (44-15)
X

The combined effect is found by regarding Im.int in Eq. 44-12 as a variable


amplitude, given in fact by I O,dif of Eq. 44-14. This assumption, for the com-
bined pattern, leads to
le = Im (cos f3) 2 (sin
2
-a- , a) (44-16)
1116 DIFFRACTION Chap. 44
in which we have dropped all subscripts referring separately to interference
and diffraction.
Let us express this result in words. At any point on the screen the avail-
able light intensity from each slit, considered separately, is given by the
diffraction pattern of that slit (Eq. 44-14). The diffraction patterns for the
two slits, again considered separately, coincide because parallel rays in

- -~ .-.,.- - - .,,- - . r .,,- -~-1-·~ -.,. -,., -1.0


--re-~ __ a=X
.,,_IT-
-~- --
-" -
- >,

-
"in
c
Q)-

c
Q)
(a)
>
~
Q)

15
I
10
I
5
rr 0
I
5 10
I
15
8, degrees

/
/

Z'
' \

-
'in
c
Q)

c a= 5X (b)
Q)
I >
:;::::; \
ro
Q)
/
c:::
"
.......
15 10 5 0 5 10 15
8, degrees

I
/
'
-
\
I >, \
I

-
"in
c
Q)

I .S \
I \ (c)
I \ a= lOX
I \

15 10 5 0 5 10 15
8, degrees

Fig. 44-12 Interference fringes for a double slit with slit separation d = 50X. Three
different slit widths, described by a/X = 1, 5, and 10 are shown.
Sec. 44-6 DOUBLE SLIT 1117
r r r r

~ .?;-
'iii
c
2c
(!)
>
:;:;
ro
<ii
0:::

(a)
I I I \ I I I
15 10 5 0 5 10 15
8, degrees

(b)

8, degrees

~
VI
c
.....
(I)

.!::
(!)
.2:
.....
ro
<ii
0:::
(c)

15 10 5 0 5 10 15
8, degrees

Fig. 44-13 (a) The "interfere nce factor" and (b) the "diffracti on factor"
in Eq. 44-16
and (c) their product; compare Fig. 44-12b.

Fraunho fer diffracti on are focused at the same spot (see Fig. 44-5). Because
the two diffract ed waves are coheren t, they will interfer e.
The effect of interfer ence is to redistri bute the availabl e energy over the
screen, produci ng a set of fringes. In Section 43-1, where we assume d a << >..,
the availabl e energy was virtuall y the same at all points on the screen so
that the interfer ence fringes had virtuall y the same intensit ies (see Fig. 43-9).
If we relax the assump tion a« 'J\, the availabl e energy is not uniform over
the screen but is given by the diffract ion pattern of a slit of width a. In this
1118 DIFFRACTION Chap. 44

ti(a)
Fig. 44-14 (a) Interference fringes for a double-slit system
in which the slit width is not negligible in comparison to the
wavelength. The fringes are modulated in intensity by the
diffraction pattern of a single slit. (b) If one of the slits is
covered up, the interference fringes disappear and we see
the single slit diffraction pattern. (Courtesy G. H. Carra-
gan, Rensselaer Polytechnic Institute.)

(b)

case the interference fringes will have intensities that are determined by the
intensity of the diffraction pattern at the location of a particular fringe.
Equation 44-16 is the mathematical expression of this argument.
Figure 44-12 is a plot of Eq. 44-16 for d = 50t- and for three values of
a/t-. It shows clearly that for narrow slits (a = t-) the fringes are nearly
uniform in intensity. As the slits are widened, the intensities of the fringes
are markedly modulated by the "diffraction factor" in Eq. 44-16, that is,
by the factor (sin a/a) 2 .
Equation 44-16 shows that the fringe envelopes of Fig. 44-12 are precisely
the single-slit diffraction patterns of Fig. 44-9. This is especially clear in
Fig. 44-13, which shows, for the curve of Fig. 44-12b, (a) the "interference
factor" in Eq. 44-16 (that is, the factor cos2 /3), (b) the "diffraction factor"
(sin a/a.)2, and (c) their product.
If we put a = 0 in Eq. 44-16, then (see Eq. 44-15) a = 0 and sin a/a"'
a-->0
a/a = 1. Thus this equation reduces, as it must, to the intensity equation
for a pair of vanishingly narrow slits (Eq. 44-12). If we put d = 0 in Eq.
44-16, the two slits coalesce into a single slit of width a, as Fig. 44-15 shows;
d = 0 implies {3 = 0 (see Eq. 44-13) and cos2 {3 = 1. Thus Eq. 44-16 re-
duces, as it must, to the diffraction equation for a single slit (Eq. 44-14).
Figure 44-14 shows some actual double-slit interference photographs.
The uniformly spaced interference fringes and their intensity modulation
by the diffraction pattern of a single slit is clear. If one slit is covered up,
as in Fig. 44-14b, the interference fringes disappear and we see the diffrac-
tion pattern of a single slit .
._ Example 6. Starting from the curve of Fig. 44-12b, what is the effect of (a)
increasing the slit width, (b) increasing the slit separation, and (c) increasing the
wavelength?
(a) If we increase the slit width a, the envelope of the fringe pattern changes so
that its central peak is sharper (compare Fig. 44-12c). The fringe spacing, which
depends on d/t., does not change.
(b) If we increased, the fringes become closer together, the envelope of the pattern
remaining unchanged.
(c) If we increase A, the envelope becomes broader and the fringes move further
apart. Increasing )\ is equivalent to decreasing both of the ratios a/t- and d/t-. The
general relationship of the envelope to the fringes, which depends only on d/ a, does
not change with wavelength.
Sec. 44-6 DOUBLE SLIT 1119

Example 7. In double-slit Fraunhofer diffraction what is the fringe spacing on a


screen 50 cm away from the slits if they are illuminated with blue light C\ = 4800 A),
if d = 0.10 mm, and if the slit width a = 0.02 mm? What is the linear distance from
the central maximum to the first minimum of the fringe envelope?
The intensity pattern is given by Eq. 44-16, the fringe spacing being determined
by the interference factor cos 2 fJ. From Example 2, Chapter 43, we have
XD
t::.y=-,
d

where D is the distance of the screen from the slits. Substituting yields

A _ (480 X 10-9 meter)(50 X 10-2 meter) _ 2 4 _3 t _


~Y - - • X 10 me er - 2 .4 mm.
0.10 X 10- 3 meter

The distance to the first minimum of the envelope is determined by the diffraction
factor (sin a/a) 2 in Eq. 44-16. The first minimum in this factor occurs for a = 1r.
From Eq. 44-15, •
sin 8 = aX = ~ = 480 X 10-9 meter = 0 _024 _
1ra a 0.02 X 10- 3 meter

This is so small that we can assume that 8 "' sin 8 "' tan 8, or

y = D tan 8"' D sin 8 = (50 cm)(0.024) = 1.2 cm.

There are about ten fringes in the central peak of the fringe envelope.
Example 8. What requirements must be met for the central maximum of the
envelope of the double-slit Fraunhofer pattern to contain exactly eleven fringes?
The required condition will be met if the sixth minimum of the interference factor
(cos2 fJ) in Eq. 44-16 coincides with the first minimum of the diffraction factor
(sin a/a) 2•
The sixth minimum of the interference factor occurs when

in Eq. 44-12.
The first minimum in the diffraction term occurs for

a = 7r.

Dividing (see Eqs. 44-13 and 44-15) yields

fJ d 11
~ = - = -·
a a 2

This condition depends only on the slit geometry and not on the wavelength. For
long waves the pattern will be broader than for short waves, but there will always be
eleven fringes in the central peak of the envelope. <11111

The double-slit problem as illustrated in Fig. 44-12 combines interference


and diffraction in an intimate way. At root both are superposition effects
and depend on adding wave disturbances at a given point, taking phase dif-
ferences properly into account. If the waves to be combined originate from
a finite (and usually small) number of elementary coherent radiators, as in
Young's double-slit experiment, we call the effect interference. If the waves
to be combined originate by subdividing a wave into infinitesimal coherent
1120 DIFFRACTION Chap. 44
radiators, as in our treatment of a single slit (Fig. 44-6), we call the effect
diffraction. This distinction between interference and diffraction is con-
venient and useful. However, it should not cause us to lose sight of the fact
that both are superposition effects and that often both are present simul-
taneously, as in Young's experiment.

QUESTIONS

1. Why is the diffraction of sound waves more evident in daily experience than that of
light waves?
2. Why do radio waves diffract around buildings, although light waves do not?
3. A loud-speaker horn has a rectangular aperture 4 ft high and 1 ft wide. Will the
pattern of sound intensity be broader in the horizontal plane or in the vertical?
4. A radar antenna is designed to give accurate measurements of the height of an air-
craft but only reasonably good measurements of its direction in a horizontal plane. Must
the height-to-width ratio of the radar reflector be less than, equal to, or greater than unity?
5. A person holds a single narrow vertical slit in front of the pupil of his eye and looks
at a distant light source in t,he form of a long heated filament. Is the diffraction pattern
that he sees a Fresnel or a Fraunhofer pattern?
6. In a single-slit Fraunhofer diffraction, what is the effect of increasing (a) the wave-
length and (b) the slit width?
7. Sunlight falls on a single slit of width 104 A. Describe qualitatively what the result-
ing diffraction pattern looks like.
8. In Fig. 44-5 rays r1 and r3 are in phase; so are r2 and r4. Why isn't there a maximum
intensity at P2 rather than a minimum?
9. Describe what happens to a Fraunhofer single-slit diffraction pattern if the whole
apparatus is immersed in water.
10. Distinguish clearly between(), a, and</> in Eq. 44-8c.
11. Do diffraction effects occur for virtual images as well as for real images? Explain.
12. Do diffraction effects occur for images formed by (a) plane mirrors and (b) spherical
mirrors? Explain.
13. If we were to redo our analysis of the properties of lenses in Section 42-5 by the
methods of geometrical optics but without restricting our considerations to paraxial rays
and to "thin" lenses, would diffraction phenomena, such as that of Fig. 44-10, emerge
from the analysis? Discuss.
14. Distinguish carefully between interference and diffraction in Young's double-slit
experiment.
15. In what way are interference and diffraction similar? In what way are they dif-
ferent?
16. In double-slit interference patterns such as that of Fig. 44-14a we said that the
interference fringes were modulated in intensity by the diffraction pattern of a single slit.
Could we reverse this statement and say that the diffraction pattern of a single slit is
intensity-modulated by the interference fringes? Discuss.
Chap. 44 PROBLEMS 1121

PROBLEMS

1. In a single-slit diffraction pattern the distance between the first minimum on the
right and the first minimum on the left is 5.2 mm. The screen on which the pattern is
displayed is 80 cm from the slit and the wavelength is 5460 A. Calculate the slit width.
"2. A plane wave (A = 5900 A) falls on a slit with a = 0.40 mm. A converging lens
(f = +70 cm) is placed behind the slit and focuses the light on a screen. What is the linear
distance on the screen from the center of the pattern to (a) the first minimum and (b) the
second minimum?
3. A single slit is illuminated by light whose wavelengths are Aa and Ab, so chosen that the
first diffraction minimum of Aa coincides with the second minimum of Ab, (a) What
relationship exists between the two wavelengths? (b) Do any other minima in the two
patterns coincide?
4. (a) Show that the values of a at which intensity maxima for single-slit diffraction
occur can be found exactly by differentiating Eq. 44-8b with respect to a and equating to
zero, obtaining the condition
tan a = a.
(b) Find the values of a satisfying this relation by plotting graphically the curve y = tan a
and the straight line y = a and finding their intersections. (c) Find the (nonintegral)
values of m corresponding to successive maxima in the single-slit pattern. Note that the
secondary maxima do not lie exactly halfway between minima.
5. In Example 4 solve the transcendental equation

! = (sin ax) 2
2 ax
graphically for ax, to an accuracy of three significant figures.
6. (a) In Fig. 44-7d, why is E9, which represents the first maximum beyond the central
maximum, not vertical? (b) Calculate the angle it makes with the vertical, assuming the
~it to be divided into infinitesimal strips of width dx .
.,Z, What is the half-width of a diffracted beam for a slit whose width is (a) 1, (b) 5, and
(c) io wavelengths?
8. (a) A circular diaphragm 0.60 meter in diameter oscillates at a frequency of 25,000
cycles/sec in an underwater source of sound for submarine detection. Far from the source
the sound intensity is distributed as a Fraunhofer diffraction pattern for a circular hole
whose diameter equals that of the diaphragm. Take the speed of sound in water to be
1450 meters/sec and find the angle between the normal to the diaphragm and the direc-
tion of the first minimum. (b) Repeat for a source having an (audible) frequency of 1000
cycles/sec.
9. The two headlights of an approaching automobile are 4 ft apart. At what maximum
distance will the eye resolve them? Assume a pupil diameter of 5.0 mm and A = 5500 A.
Assume also that diffraction effects and not retinal structure limit the resolution.

10. The wall of a large room is covered with acoustic tile in which small holes are drilled
5.0 mm from center to center. How far can a person be from such a tile and still distin-
guish the individual holes, assuming ideal conditions? Assume the diameter of the oupil
to be 4.0 mm and A to be 5500 A.
11. (a) How small is the angular separation of two stars if their images are barely
resolved by the Thaw refracting telescope at the Allegheny Observatory in Pittsburgh?
The lens diameter is 30 in. and its focal length is 46 ft. Assume A = 5000 A. (b) Find the
distance between these barely resolved stars if each of them is 10 light years distant from
the earth. (c) For the image of a single star in this telescope, find the diameter of the
first dark ring in the diffraction pattern, as measured on a photographic plate placed at the
1122 DIFFRACTION Chap. 44
focal plane. Assume that the star image structure is associated entirely with diffraction
at the lens aperture and not with (small) lens ·'errors."
12. Find the separation of two points on the moon's surface that can just be resolved
by the 200-in. telescope at Mount Palomar, assuming that this distance is determined by
diffraction effects. The distance from the earth to the moon is 240,000 miles.
13. Construct qualitative vector diagrams like those of Fig. 44-7 for the double-slit
interference pattern. For simplicity, consider d = 2a (see Fig. 44-15). Can you interpret
the main features of the intensity pattern this way?

Fig. 44-15

14. Suppose that, as in Example 8, the envelope of the central peak contains eleven
fringes. How many fringes lie between the first and second minima of the envelope?
15. Ford = 2a in Fig. 44-15, how many interference fringes lie in the central diffraction
envelope?
16. If we put d = a in Fig. 44-15, the two slits coalesce into a single slit of width 2a.
Show that Eq. 44-16 reduces to the diffraction pattern for such ~slit.
17. (a) Design a double-slit system in which the fourth fringe, not counting the central
maximum, is missing. (b) What other fringes, if any, are also missing?
Gratings and Spectra
CHAPTER 45

45-1 Introduction
In connection with Young's experiment (Sections 43-1 and 43-3) we
discussed the interference of two coherent waves formed by diffraction at
two elementary radiators (pinholes or slits). In our first treatment we as-
sumed that the slit width was much less than the wavelength, so that light
diffracted from each slit illuminated the observation screen essentially uni-
formly. Later, in Section 44-6, we took the slit width into account and
showed that the intensity pattern of the interference fringes is modulated
by a "diffraction factor" (sin a/a) 2 (see Eq. 44-16).
Here we extend our treatment to cases in which the number N of radia-
tors or diffracting centers is larger-and usually much larger-than two.
We consider two situations:
1. An array of N parallel equidistant slits, called a diffraction grating.
2. A three-dimensional array of periodically arranged radiators-the
atoms in a crystalline solid such as NaCL In this case the average spacing
between the elementary radiators is so small that interference effects must
be sought at wavelengths much smaller than those of visible light. We
speak of X-ray diffraction.
In each case we distinguish carefully between the diffracting properties
of a single radiator (slit or atom) and the interference of the waves dif-
fracted, coherently, from the assembly of radiators.

45-2 Multiple Slits


A logical extension of Young's double-slit interference experiment is to
increase the number of slits from two to a larger number N. An arrangement
like that of Fig. 45-1, usually involving many more slits, is called a diffrac-
1123
1124 GRATINGS AND SPECTRA Chap. 45

Fig. 45-1 An idealized diffraction grating containing five slits. The slit width a is shown
for convenience to be considerably smaller than A, although this condition is not realized
in practice. The figure is distorted in that f is much greater than d in practice.

tion grating. As for a double slit, the intensity pattern that results when
monochromatic light of wavelength>.. falls on a grating consists of a series of
interference fringes. The angular separations of these fringes are determined
by the ratio A/ d, where d is the spacing between the centers of adjacent slits.
The relative intensities of these fringes are determined by the diffraction pat-
tern of a single grating slit, which depends on the ratio A/ a, where a is the
slit width.
Figure 45-2, which compares the intensity patterns for N = 2 and N = 5,
shows clearly that the "interference" fringes are modulated in intensity by
a "diffraction" envelope, as in Fig. 44-14. Figure 45-3 presents a theoretical
calculation of the intensity patterns for a few fringes near the centers of the
patterns of Fig. 45-2. These two figures show that increasing N (a) does
not change the spacing between the (principal) interference fringe maxima,

Fig. 45-2 Intensity patterns for "grat-


ings" with (a) N = 2 and (b) N = 5 for
the same value of d and \. Note how
the intensities of the fringes are modu-
(a) N = 2 lated by a diffraction envelope as in Fig.
44-14; thus the assumption a«\ is not
realized in these actual "gratings." For
N = 5 three very faint secondary max-
ima, not visible in this photograph, ap-
(b) N = 5 pear between each pair of adjacent pri-
mary maxima.
Sec. 45-2 MULTIPLE SLITS 1125

.i::'
"vi
c 100 -
a.> 80 N=2
c 60
> 40
a.>
-
20
.;:::;
"' 0
a, I I I', I I l.1 I I I I I I I Ll...L_LJ...L.[...Li...J',W

er -40 -30 -20 -10 0 10 20 30 40


8, degrees
(a)

::!~l ~.il~ ... A


er -40 -30 -20
Ilk-±+
-10
I'
~

e, degrees
(b)
'.
0
N=5
d h _h n
I
d
I

10
I
±:I +d
40
,J

Fig. 45-3 Calculated intensity patterns for (a) a two-slit and (b) a five-slit grating for
the same value of d and JI.. This figure shows the sharpening of the principal maxima
and the appearance of faint serondary maxima for N > 2. The letters on the five-slit
pattern refer to Fig. 45-5. The figure assumes slits with a « 'A so that the principal
maxima are of uniform intensity.

provided d and A remain unchanged, (b) sharpens the (principal) maxima,


and (c) introduces small secondary maxima between the principal maxima.
Three such secondaries are present (but not readily visible) between each
pair of adjacent principal maxima in Fig. 45-2b.
A principal maximum in Fig. 45-1 will occur when the path difference
between rays from adjacent slits ( = d sin 0) is given by
d sin O = m")\ m = 0, 1, 2, ... (principal maxima), (45-1)
where mis called the order number. This equation is identical with Eq. 43-1,
which locates the intensity maxima for a double slit. The locations of the
(principal) maxima are thus determined only by the ratio V d and are inde-
pendent of N. As for the double slit, the ratio a/A determines the relative
intensities of the principal maxima but does not alter their locations appre-
ciably.
The sha.rpening of the principal maxima as N is increased can be under-
stood by a graphical argument, using phasors. Figures 45-4a and b show
conditions at any of the principal maxima for a two-slit and a nine-slit grat-
ing. The small arrows represent the amplitudes of the wave disturbances
arriving at the screen at the position of each principal maximum. For
simplicity we consider the central principal maximum only, for which m = 0,
and thus O = 0, in Eq. 45-1.
Consider the angle !i0 0 corresponding to the position of zero intensity that
lies on either side of the central principal maximum. Figures 45-4c and d
show the phasors at this point. The phase difference between waves from
adjacent slits, which is zero at the central principal maximum, must increase
by an amount ti¢ chosen so that the array of phasors just closes on itself,
1126 GRATINGS AND SPECTRA Chap. 45
yielding zero resultant intensity. For N = 2, b..cp = 21r/2 ( = 180°); for
N = 9, b..cp = 21r/9 ( = 40°). In the general case it is given by
21r
b..cp = - ·
N
This increase in phase difference for adjacent waves corresponds to an
increase in the path difference b..Z given by
phase difference path difference
21r t,.

or b..l = (;71") b..cp = (2t,.1r) ( ; ) = ~.


From Fig. 45-1, however, the path difference b..l at the first minimum is also
given by d sin b..00 , so that we can write
t,.
d sin Ll8 0 = -,
N
t,.
or sin b..80 = - ·
Nd
Since N >> 1 for actual gratings, sin b..00 will ordinarily be quite small (that
is, the lines will be sharp), and we may replace it by b..80 to good approxima-
tion, or
(central principal maximum). (45-2)

This equation shows specifically that if we increase N for a given >.. and d,

~
E2m
(a)

(d)

Fig. 45-4 Drawings (a) and (b) show conditions at the central principal maximum for a
two-slit and a nine-slit grating, respectively. Drawings (c) and (d) show conditions at the
minimum of zero intensity that lies on either side of this central principal maximum. In
going from (a) to (c) the phase shift between waves from adjacent slits changes by 180°
(!'>.</> = 21r/2); in going from (b) to (d) it changes by 40° (!'>.</> = 21r/9).
Sec. 45-2 MULTIPLE SLITS 1127
E 8(= Em) \
(a) \

~
(360°)
(i)
(240°)

d
I
I

''
,[µ&
(e)
I / ~q:_ (280°)
E9
(j)
I E9 (230°)
I I

--a2_0.'._
I
(b)
(330°) (f) I
(270°)
"'" I

/~
l E8 (k)
I (216°)
I
I I E9 \
I \
\
(c) --~ (g) \
1.90· (260°)

'-
(300°)
\
\
-~ E9 '-
(l)
(200°)

"' ' '-


'-
170° (m)

/
/ --~ (190°)

(d)
(n)
(288°) (180°)
/ ~180°
Eo

Fig. 45-5 The figures taken in sequence from (a) to (n) and then from (n) to (a) show con-
ditions as the intensity pattern of a five-slit grating is traversed from the central principal
maximum to an adjacent principal maximum. Phase differences between waves from
adjacent slits are shown directly or, when going from (n) to (a), in parentheses. Principal
maxima occur at (a), secondary maxima at, or near, (h) and (n), and points of zero intensity
at (d) and (k). Compare Fig. 45-3b.

~e0 will decrease, which means that the central principal maximum becomes
sharper.
We state without proof,* and for later use, that for principal maxima other than
the central one (that is, for m ,t: 0) the angular distance between the position em
of the principal maximum of order m and the minimum that lies on either side is
given by
")..
~em=----
Nd cos &m
(any principal maximum). (45-3)

For the central principal maximum we have m = 0, em = 0, and ~&m = ~&0, so that
Eq. 45-3 reduces, as it must, to Eq. 45-2.
The origin of the secondary maxima that appear for N > 2 can also be
understood using the phasor method. Figure 45-Sa shows conditi9ns at
the central principal maximum for a five-slit grating. The vectors are in
* See Problem 15.
1128 GRATINGS AND SPECTRA Chap. 45
phase. As we depart from the central maximum, e in Fig. 45-1 increases
from zero and the phase difference between adjacent vectors increases from
21r .
zero to !::.¢ = - d sm e. Successive figures show how the resultant wave
A
amplitude Ee varies with !::.¢. The student should verify by graphical con-
struction that a given figure represents conditiens for both !::.¢ and 21r - !::.¢.
Thus we start at !::.¢ = 0, proceed to !::.¢ = 180°, and then trace backward
through the sequence, following the phase differences shown in parentheses,
until we reach !::.¢ = 360°. This sequence corresponds to traversing the
intensity pattern from the central principal maximum to an adjacent one.
Figure 45-5, which should be compared with Fig. 45-3b, shows that for
N = 5 there are three secondary maxima, corresponding to!::.¢ = 110°, 180°,
and 250°. The student should make a similar analysis for N = 3 and should
show that only one secondary maximum occurs. In actual gratings, which
commonly contain 10,000 to 50,000 "slits," the secondary maxima lie so
close to the principal maxima or are so reduced in intensity that they cannot
be distinguished from them experimentally.

45-3 Diffraction Gratings


The grating spacing d for a typical grating that contains 12,000 "slits"
distributed over a 1-in. width is 2.54 cm/12,000, or 21,000 A. Gratings are
often used to measure wavelengths and to study the structure and intensity
of spectrum lines. Few devices have contributed more to our knowledge of
modern physics.
Gratings are made by ruling equally spaced parallel grooves on a glass or a
metal * plate, using a diamond cutting point whose motion is automatically
controlled by an elaborate ruling engine. Once such a master grating has
been prepared, replicas can be formed by pouring a collodion solution on the
grating, allowing it to harden, and stripping it off. The stripped collodion,
fastened to a flat piece of glass or other backing, forms a good grating.
Figure 45-6 shows a cross section of a common type of grating ruled on glass. In
the rudimentary grating of Fig. 45-1 open slits were separated by opaque strips; the
amplitude of the wave disturbance varied in a periodic way as the grating was crossed,
dropping to zero on the opaque strips. The grating of Fig. 45-6 is transparent every-
where, so that there is little periodic change in amplitude as the grating is crossed.
The effect of the rulings is to change the optical thickness of the grating in a periodic
way, rays traversing the grating between the rulings (bin Fig. 45-6) containing more
wavelengths than rays traversing the grating in the center of the rulings (a in Fig.
45-6). This results in a periodic change of phase as one crosses the grating at right
angles to the rulings. Reflection gratings also depend for their operation on a periodic
change in phase of the reflected wave as one crosses the grating, the change in ampli-
tude under these conditions being negligible. The principal maxima for phase
gratings, assuming that the incident light falls on the grating at right angles, can be
given by the same formula derived earlier for idealized amplitude or slit gratings,
* Gratings ruled on metal are called reflection gratings because the interference effects
are viewed in reflected rather than in transmitted light. Many reseaich gratings are
the reflection type; commonly they are ruled on the surface of a concave mirror, which
eliminates the need for lenses.
Sec. 45-3 DIFFRACTION GRATINGS 1129

Fig. 45-6 An enlarged cross section of


a diffraction grating ruled on glass.
Such gratings, in which the phase of
the emerging wave changes as one
crosses the grating, are called phase
gratings.

namely
d sin 8 = m>.. m=0,1,2 ... ,
where d is the distance between the rulings and the integer m is called the order of
the particular principal maximum. Essentially all gratings used in the visible
spectrum, whether of the transmission type, as in Fig. 45-6, or the reflection type, are
phase gratings.
Figure 45-7 shows a simple grating spectroscope, used for viewing the
spectrum of a light source, assumed to emit a number of discrete wavelengths,
or spectrum lines. The light from source Sis focused by lens L 1 on a slit S 1
placed in the focal plane of lens L 2 . The parallel light emerging from col-
limator C falls on grating G. Parallel rays associated with a particular in-
terference maximum occurring at angle 8 fall on lens L 3 , being brought to a
focus in plane F-F'. The image formed in this plane is examined, using a
magnifying lens arrangement E, called an eyepiece. A symmetrical inter-

Fig. 45-7 A simple type of grating spectroscope used to analyze the wavelengths of the
light emitted by source S.
1130 GRATINGS AND SPECTRA Chap. 45
ference pattern is formed on the other side of the central position, as shown
by the dotted lines. The entire spectrum can be viewed by rotating tele-
scope T through various angles. Instruments used for scientific research or
in industry are more complex than the simple arrangement of Fig. 45-7.
They invariably employ photographic or photoelectric recording and are
called spectrographs. Figure 47-12 shows a small portion of the spectrum of
iron, produced by examining the light produced in an arc struck between iron
electrodes, using a research type spectrograph with photographic recording.
Each line in the figure represents a different wavelength that is emitted from
the source.
Grating instruments can be used to make absolute measurements of wave-
length, since the grating spacing din Eq. 45-1 can be measured accurately
with a traveling microscope. Several spectra are normally produced in such
instruments, corresponding tom = ± 1, ±2, etc., in Eq. 45-1 (see Fig. 45-8).
This may cause some confusion if the spectra overlap. Further, this multi-
plicity of spectra reduces the recorded intensity of any given spectrum line
because the available energy is divided among a number of spectra.
This disadvantage of the grating instrument can be overcome by shaping the
profile of the grating grooves so that a large fraction of the light is thrown into a
particular order on a particular side (for a given wavelength). This technique, called
blazing, so alters the diffracting properties of the individual grooves (by controlling
their profiles) that the light of wavelength X diffracted by a single groove has a sharp
peak of maximum intensity at a selected angle() ( ~ 0).
Light can also be analyzed into its component wavelengths if the grating
in Fig. 45-7 is replaced by a prism. In a prism spectrograph each wavelength

Fig. 45-8 Example 1. The spectrum of white light as viewed in a grating instrument
like that of Fig. 45-7. The different orders, identified by the order number m, are shown
separated vertically for clarity. As actually viewed, they would not be so displaced.
The central line in each order corresponds to }.. = 5500 A.
Sec. 45-3 DIFFRACTION GRATINGS 1131

in the incident beam is deflected through a definite angle 8, determined by


the index of refraction of the prism material for that wavelength. Curves
such as Fig. 41-2, which gives the index of refraction of fused quartz as a
function of wavelength, show that the shorter the wavelength, the larger
the angle of deflection 8. Such curves vary from substance to substance and
must be found by measurement. Prism instruments are not adequate for
ac~urate absolute measurements of wavelength because the index of refrac-
tion of the prism material at the wavelength in question is usually not known
precisely enough. Both prism and grating instruments make accurate com-
parisons of wavelength, using a suitable comparison spectrum such as that
shown in Fig. 47-12, in which careful absolute determinations have been
made of the wavelengths of the spectrum lines. The prism instrument has
an advantage over a grating instrument (unblazed) in that its light energy
is concentrated into a single spectrum so that brighter lines may be produced.
.... Example I. A grating with 8000 rulings/in. is illuminated with white light at
perpendicular incidence. Describe the diffraction pattern. Assume that the wave-
length of the light extends from 4000 to 7000 A.
The grating spacing d is 2.54 cm/8000, or 31,700 A. The central or zero-order
maximum corresponds tom = 0 in Eq. 45-1. All wavelengths present in the incident
light are superimposed at 8 = 0, as Fig. 45-8 shows.
The first-order diffraction pattern corresponds tom = 1 in Eq. 45-1. The 4000-A
line occurs at an angle given by
_ . _ 1 m"\ _ . _ 1 (1)(4000 A) _. _1 6 _ 7 3o
8 - sm d - sm 31 ,700 A - sm 0.12 - . .

In the same way the angle for the 7000-A line is found to be 12.8°, and the entire
pattern of Fig. 45-8 can be calculated. Note that the first-order spectrum (m = 1) is
isolated but that the second-, third-, and fourth-order spectra overlap.
Example 2. A diffraction grating has 104 rulings uniformly spaced over 1 in.
It is illuminated at normal incidence by yellow light from a sodium vapor lamp.
This light contains two closely spaced lines (the well-known sodium doublet) of
wavelengths 5890.0 and 5895.9 A. (a) At what angle will the first-order maximum
occur for the first of these wavelengths?
The grating spacing d is 10-4 in., or 25,400 A. The first-order maximum corre-
sponds tom = 1 in Eq. 45-1. We thus have
. -1 ':!!'.~ - . -1 (1)(5890 A) = . -1 0 232 = 13 30
8 -- sm d - sm 25 400 A sm · · ·
'
(b) What is the angular separation between the first-order maxima for these lines?
The straightforward way to find this separation is to repeat this calculation for
A = 5895.9 A and to subtract the two angles. A difficulty, which can best be ap-
preciated by carrying out the calculation, is that we must carry a large number of
significant figures to obtain a meaningful value for the difference between the angles.
To calculate the difference in angular positions directly, let us write down Eq. 45-1,
solved for sin 8, and differentiate it, treating 8 and A as variables:
. 8 m"\
sm =a
1132 GRATINGS AND SPECTRA Chap. 45
If the wavelengths are close enough together, as in this case, d").. can be replaced by
LlA, the actual wavelength difference; d{) then becomes .6.(), the quantity we seek. This

.6.() = m .6.").. = (l)( 5 ,9 A) = 2.4 X 10-4 radian = 0.014°.


d cos{) (25,400 A)(cos 13.3°)
Note that although the wavelengths involve five significant figures our calculation,
done this way, involves only two or three, with consequent reduction in numerical
manipulation. <1111

The quantity de/d").., called the dispersion D of a grating, is a measure of


the angular separation produced between two incident monochromatic
waves whose wavelengths differ by a small wavelength interval. From this
example we see that
de m
D=-=--- (45-4)
d").. d cos{)

45-4 Resolving Power of a Grating


To distinguish light waves whose wavelengths are close. together, the
principal maxima of these wavelengths formed by the grating should be as
narrow as possible. Expressed otherwise, the grating should have a high
resolving power R, defined from
")..
R=-· (45-5)
.6.")..

Here ").. is the mean wavelength of two spectrum lines that can barely be
recognized as separate and .6.").. is the wavelength difference between them.
The smaller .6.").. is, the closer the lines can be and still be resolved; hence
the greater the resolving power R of the grating. It is to achieve a high
resolving power that gratings with many rulings are constructed.
The resolving power of a grating is usually determined by the same con-
sideration (that is, the Rayleigh criterion) that we used in Section 44-5 to
determine the resolving power of a lens. If two principal maxima are to be
barely resolved, they must, according to this criterion, have an angular
separation .6.() such that the maximum of one line coincides with the first
minimum of the other; see Fig. 44-11. If we apply this criterion, we can
show that
R = Nm, (45-6)
where N is the total number of rulings in the grating and m is the order.
As expected, the resolving power is zero for the central principal maximum
(m = 0), all wavelengths being undeflected in this order.
Let us derive Eq. 45-6. The angular separation between two principal maxima
whose wavelengths differ by .6."' is found from Eq. 45-4, which we recast as

.6,() = m .6.").. . (45-4)


d cos{)
The Rayleigh criterion (Section 44-5) requires that this be equal to the angular
separation between a principal maximum and its adjacent minimum. This is given
Sec. 45-4 RESOLVING POWER OF A GRATING 1133
from Eq. 45-3, dropping the subscript m in cos Om, as
;\
l::!..Om = Nd cose (45-3)

Equating Eqs. 45-4 and 45-3 leads to


R (=;\/I::!..;\)= Nm,
which is the desired relation .
.... Example 3. In Example 2 how many rulings must a grating have if it is barely
to resolve the sodium doublet in the third order?
From Eq. 45-5 the required resolving power is
;\ 5890 A
R = I::!.;\ = (5895.9 - 5890.0)A = lOOO.
From Eq. 45-6 the number of rulings needed is

N = : = l~OO = 330.
This is a modest requirement.

The resolving power of a grating must not be confused with its dispersion.
Table 45-1 shows the characteristics of three gratings, each illuminated with
light of;\ = 5890 A, the diffracted light being viewed in the first order (m =
1 in Eq. 45-1).
Table 45-1

SOME CHARACTERISTICS OF THREE GRATINGS


(>- = 5890 A, m = 1)

d, D
Grating N e R
A 10-3 degrees/ A

A 10,000 25.,400 13.3° 10,000 2.32


B 20,000 25,400 13.3° 20,000 2.32
c 10,000 13,700 25.5° 10,000 4.64

The student should verify that the values of D and R given in the table
can be calculated from Eqs. 45-4 and 45-6, respectively.
For the conditions of use noted in Table 45-1, gratings A and B have the
same dispersion and A and C have the same resolving power. Figure 45-9
shows the intensity patterns that would be produced by these gratings for
two incident waves of wavelengths ;\ 1 and >. 2 , in the vicinity of;\ = 5890 A.
Grating B, which has high resolving power, has narrow intensity maxima
and is inherently capable of distinguishing lines that are much closer together
in wavelength than those of Fig. 45-9. Grating C, which has high dispersion,
produces twice the angular separation between rays >. 1 and >. 2 that grating B
does.
1134 GRATINGS AND SPECTRA Chap. 45

(a)

~
~
~L
---~l
!\ Grating
C X2
I\ (c)
0 0, degrees
------- fl(} - - - - -
25.5°

Fig. 45-9 The intensity patterns for light of wavelengths X1 and X2 near 5890 A, incident
on the gratings of Table 45-1. Grating B has the highest resolving power and grating C
the highest dispersion.

lili1- Example 4. The grating of Example 1 has 8000 lines illuminated by light from
a mercury vapor discharge. (a) What is the expected dispersion, in the third order,
in the vicinity of the intense green line (X = 5460 A)? Noting that d = 31,700 A,
we have, from Eq. 45-1,

. -1 mX - . -1 (3)(5460 A) - . -1 0 517 - 31 lo
0 -- sm d - sm 31 700 A - sm · - · ·
'
From Eq. 45-4 we have

D = d c:s O = (31,700 A:(cos 31.1 o) = 1.1 X 10-4 radian/ A = 6.3 X 10-3 deg/ A.

(b) What is the expected resolving power of this grating in the fifth order? Equation
45-6 gives
R = Nm = (8000)(5) = 40,000.
Thus near X = 5460 A a wavelength difference ~X given by Eq. 45-5, or

~X = ~ = 5460 A = 0 14 A
R 40,000 . '
can be distinguished.

45-5 X-ray Diffraction


Figure 45-10 shows how X-rays are produced when electrons from a heated
filament F are accelerated by a potential difference V and strike a metal tar-
get T. X-rays are electromagnetic radiation with wavelengths of the order
of 1 A. This value is to be compared to 5500 A for the center of the visible
spectrum. For such small wavelengths a standard optical diffraction grat-
ing, as normally employed, cannot be used. For X = 1 A and d = 30,000 A,
Sec. 45-5 X-RAY DIFFRACTION 1135

Fig. 45-10 X-rays are generated when electrons from heated filament F, accelerated by
potential difference V, are brought to rest on striking metallic target T. Wis a "window"
-transparent to X-rays-in the evacuated metal container C.

for example, Eq. 45-1 shows that the first-order maximum occurs at
mA (l)(lA)
() = sin-1 - = sin- 1 4 = sin- 1 0.33 X 10-4 = 0.002°.
d 3Xl0 A
This is too close to the central maximum to be practical. A grating with
d"' A is desirable, but, since X-ray wavelengths are about equal to atomic
diameters, such gratings cannot be constructed mechanically.
In 1912 it occurred to the German physicist Max von Laue that a crystal-
line solid, consisting as it does of a regular array of atoms, might form a
natural three-dimensional "diffraction grating" for X-rays. Figure 45-11
shows that if a collimated beam of X-rays, continuously distributed in wave-
length, is allowed to fall on a crystal, such as sodium chloride, intense beams
corresponding to constructive interference from the many diffracting centers

Diffracted beams

X-ray beam
(continuous in
wavelength)

Fig. 45-11 A nonmonochromatic beam of X-rays falls on a crystal X, which may be


NaCl. Strong diffracted beams appear in certain directions, forming a so-called Laue
pattern on a photographic film S.
1136 GRATINGS AND SPECTRA Chap. 45

Fig. 45-12 Laue X-ray diffraction pattern from sodium chloride. A crystal of ordinary
table salt was used in making this plate. (Courtesy of W. Arrington and J. L. Katz,
X-ray Laboratory, Rensselaer Polytechnic Institute.)

of which the crystal is made up appear in certain sharply defined directions.


If these beams fall on a photographic film, they form an assembly of "Laue
spots." Figure 45-12, which is an actual example of these spots, shows that
the hypothesis of Laue is indeed correct. The atomic arrangements in the
crystal can be deduced from a careful study of the positions and intensities
of the Laue spots * in much the same way that we might deduce the structure
of an optical grating (that is, the detailed profile of its slits) by a study of the
positions and intensities of the lines in the interference pattern.
Figure 45-13 shows how sodium and chlorine atoms (strictly, Na+ and
c1- ions) are stacked to form a crystal of sodium chloride. This pattern,
which has cubic symmetry, is one of the many atomic arrangements exhibited
* Other experimental arrangements have supplanted the Laue technique to a consider-
able extent today; the principle remains unchanged, however.
Sec. 45-5 X-RAY DIFFRACTION 1137

by solids. The model represents the unit cell for sodium chloride. This is
the smallest unit from which the crystal may be built up by repetition in
three dimensions. The student should verify that no smaller assembly of
atoms possesses this property. For sodium chloride the length of the cube
edge of the unit cell is 5.62737 A.
Each unit cell in sodium chloride has four sodium ions and four chlorine
ions associated with it. In Fig. 45-13 the sodium ion in the center belongs
entirely to the cell shown. Each of the other twelve sodium ions shown is
shared with three adjacent unit cells so that each contributes one-fourth of
an ion to the cell under consideration. The total number of sodium ions is
then 1 + -!(12) = 4. By similar reasoning the student can show that al-
though there are fourteen chlorine ions in Fig. 45-13 only four are associated
with the unit cell shown.
The unit cell is the fundamental repetitive diffracting unit in the crystal,
corresponding to the slit (and its adjacent opaque strip) in the optical dif-

Fig. 45-13 A model showing how Na+ and c1- ions are stacked to form a unit cell of
NaCl. The small spheres represent sodium ions, the large ones chlorine. The edge a0
of the (cubical) unit cell is 5.62737 A.
1138 GRATINGS AND SPECTRA Chap. 45

• •
• •
• • • • •
• • • • •
• • • • •
(a) (b)

Fig. 45-14 (a) A section through a crystal of sodium chloride, showing the sodium and
chlorine ions. (b) The corresponding unit cells in this section, each cell being represented
by a small black square.

fraction grating of Fig. 45-1. Figure 45-14a shows a particular plane in a


sodium chloride crystal. If each unit cell intersected by this plane is repre-
sented by a small cube, Fig. 45-14b results. The student may imagine each
of these figures extended indefinitely in three dimensions.
Let us treat each small cube in Fig. 45-14b as an elementary diffracting
center, corresponding to a slit in an optical grating. The directions (but
not the intensities) of all the diffracted X-ray beams that can emerge from a
sodium chloride crystal (for a given X-ray wavelength and a given orienta-
tion of the incident beam) are determined by the geometry of this three-
dimensional lattice of diffracting centers. In exactly the same way the di-
rections (but not the intensities) of all the diffracted beams that can emerge
from a particular optical grating (for a given wavelength and orientation of
the incident beam) are determined only by the geometry of the grating, that
is, by the grating spacing d. Representing the unit cell by what is essentially
a point, as in Fig. 45-14b, corresponds to representing the slits in a diffraction
grating by lines, as we did in discussing Young's experiment in Section 43-1.
The intensities of the lines from an optical diffraction grating are deter-
mined by the diffracting characteristics of a single slit, as Fig. 44-14 shows.
In the idealized case of Fig. 45-1 these characteristics depend on the slit
width a. In practical optical gratings these characteristics depend on the
detailed shape of the profile of the grating rulings.
In exactly the same way the intensities of the diffracted beams emerging
from a crystal depend on the diffracting characteristics of the unit cell.*
Fundamentally the X-rays are diffracted by electrons, diffraction by nuclei
being negligible in most cases. Thus the diffracting characteristics of a

* For some directions in which a beam might be expected to emerge, from interference
considerations, no beam will be found because the diffracting characteristics of the unit
cell are such that no energy is diffracted in that direction. Similarly, in optical gratings
some lines, permitted by interference considerations, may not appear if their predicted
positions coincide with a null in the single-slit diffraction pattern (see Fig. 44-12).
(a)

(b)

Fig. 45-15 (a) A photograph of an oscilloscope screen arranged to display projected


electron density contours for phthalocyanine (Ca2H1sNs). Such plots, constructed elec-
trqnically from X-ray diffraction data by an analog computer, provide a vivid picture of
the structure of molecules. (Courtesy of Ray Pepinsky.) (b) A structural representation
.of the molecule phthalocyanine. The student should make a detailed comparison with
(a), locating the various atoms identified in (b). Note that hydrogen atoms, which
contain only a single electron, a.re not prominent in (a).
1139
1140 GRATINGS AND SPECTRA Chap. 45
unit cell depend on how the electrons are distributed throughout the volume
of the cell. By studying the directions of diffracted X-ray beams, we can
learn the basic symmetry of the crystal. By also studying the intensities we
can learn how electrons are distributed in the unit cell. Figure 45-15 shows
the average density of electrons projected onto a particular plane through a
unit cell of a crystal of phthalocyanine. This remarkable figure suggests
something of the full power of X-ray methods for studying the structure of
solids.

45-6 Bragg's Law


Bragg's law predicts the conditions under which diffracted X-ray beams
from a crystal are possible. In deriving it, we ignore the structure of the
unit cell, which is related only to the intensities of these beams. The dashed
sloping lines in Fig. 45-16a represent the intersection with the plane of the
figure of an arbitrary set of planes passing through the elementary diffracting
centers. The perpendicular distance between adjacent planes is d. Many
other such families of planes, with different interplanar spacings, can be
defined.
Figure 45-16b shows a plane wave that lies in the plane of the figure falling
on one member of the family of planes defined in Fig. 45-16a, the incident
rays making an angle 8 with the plane.* Consider a family of diffracted rays
lying in the plane of Fig. 45-16b and making an angle~ with the plane con-
taining the elementary diffracting centers. The diffracted rays will combine
to produce maximum intensity if the path difference between adjacent rays
is an integral number of wavelengths or
ae - bd = h(cos~ - cos&)= l>-. l = 0, 1, 2, .... (45-7)
For l = 0 this leads to ~ = e,
and the plane of atoms acts like a mirror for the incident wave, no matter
what the value of 0.
For other values of l, ~ does not equal 11, but the diffracted beam can always be
regarded as being "reflected" from a different set of planes than that shown in Fig.
45-16a with a different interplanar spacing d. Since we wish to describe each dif-
fracted beam as a "reflection" from a particular set of planes and since we are dealing
in the present argument only with the particular set of planes shown in Fig. 45-16a
we ignore all values of l other than l = 0 in Eq. 45-7. It can also be shown that a
plane of diffracting centers acts like a mirror (that is, ~ = 0) whether or not the
incident wave lies in the plane of Fig. 45-16b.

Figure 45-16c shows an incident wave striking the family of planes, a single
member of which was considered in Fig. 45-16b. For a single plane, mirror-
like "reflection" occurs for any value of e, as we have seen. To have a con-
structive interference in the beam diffracted from the entire family of planes

* In X-ray diffraction it is customary to specify the direction of a wave by giving the


angle between the ray and the plane (the glancing angle) rather than the angle between the
ray and the normal.
Sec. 45-6 BRAGG'S LAW 1141

Incident wave
x
\
Fig. 45-16 (a) A section through the
NaCl unit cell lattice of Fig. 45-14b.
The dashed sloping lines represent an

v\
I
arbitrary family of planes, with inter- h I

planar spacing d. (b) An incident


wave falls, at grazing angle O, on one
of the planes, xx', shown in (a). (c)
An incident wave falls on the entire h \
family of planes shown in (a). A
strong diffracted wave is formed. V\ \x,
(b)

Incident wave

\
~
\

\

• Diffracted
wave

(c)
1142 GRATINGS AND SPECTRA Chap. 45
in the direction 0, the rays from the separate planes must reinforce each other.
This means that the path difference for rays from adjacent planes (abc in Fig.
45-16c) must be an integral number of wavelengths or
2d sin() = m">.. m = 1, 2, 3, .... (45-8)
This relation is called Bragg's law after W. L. Bragg who first derived it.
The quantity din this equation (the interplanar spacing) is the perpendicular
distance between the planes. For the planes of Fig. 45-16a analysis shows
that d is related to the unit cell dimension a0 by

(45-9)

If an incident monochromatic X-ray beam falls at an arbitrary angle Oona


particular set of atomic planes, a diffracted beam will not result because
Eq. 45-8 will not, in general, be satisfied. If the incident X-rays are con-
tinuous in wavelength, diffracted beams will result when wavelengths given
by
2d sin()
A=--- m = 1, 2, 3,
m
are present in the incident beam (see Eq. 45-8) .
.... Example 5. At what angles must an X-ray beam with A = 1.10 A fall on the
family of planes represented in Fig. 45-16c if a diffracted beam is to exist? Assume
the material to be sodium chloride.
The interplanar spacing d for these planes is given by Eq. 45--9 or

d = __!!_o~ = 5.63 A= 252A


y'5 2.24 . .
Equation 45--8 gives
. mA (m)(l.10 A)
sm () = 2d = (2)(2.52 A) = 0.218m.

Diffracted beams are possible at() = 12.6° (m = 1), () = 25.9° (m = 2), () = 40.9°
(m = 3), and () = 60.7° (m = 4). Higher-order beams cannot exist because they
require sin() to exceed unity. Actually, the odd-order beams (m = 1, 3) prove to
have zero intensity because the unit cell in cubic crystals such as NaCl has diffracting
properties such that the intensity of the light scattered in these orders is zero (see
Problem 26). <11111

X-ray diffraction is a powerful tool for studying the arrangements of atoms


in crystals. To do so quantitatively requires that the wavelength of the
X-rays be known. In one of several approaches to this problem the unit
cell dimension for NaCl * is determined by a method that does not involve
X-rays. X-ray diffraction measurements on NaCl can then be used to de-
termine the wavelength of the X-ray beam which in turn can be used to
determine the structures of solids other than Na Cl.

* In practice, calcite (CaC0 3) proves to be more useful as a standard crystal for a num-
ber of technical reasons.
Chap. 45 QUESTIONS 1143

If p is the measured density of NaCl, we have, for the unit cell of Fig.
45-13, * recalling that each unit cell contains four NaCl "molecules,"
m 4mNaCl
p =- =
V a0 3
Here mNaCI, the mass of a NaCl molecule, is given by
M
-,
No
where M is the molecular weight of NaCl and N 0 is Avogadro's number.
Combining these two equations and solving for a0 yields

-(4M)li
ao - --
NoP
which permits us to calculate a 0 • Once a 0 is known, the wavelengths of
monochromatic X-ray beams can be found, using Bragg's law (Eq. 45-8).

QUESTIONS

1. Discuss this statement: "A diffraction grating can just as well be called an inter-
ference grating."
2. For the simple spectroscope of Fig. 45-7, show (a) that 8 increases with Xfor a grating
and (b) that 8 decreases with X for a prism.
3. You are given a photograph of a spectrum on which the angular positions and the
wavelengths of the spectrum lines are marked. (a) How can you tell whether the spectrum
was taken with a prism or a grating instrument? (b) What information could you gather
about either the prism or the grating from studying such a spectrum?
4. Assume that the limits of the visible spectrum are 4300 and 6800 A. Is it possible
to design a grating, assuming that the incident light falls normally on it, such that the
first-order spectrum barely overlaps the second-order spectrum?
5. (a) Why does a diffraction grating have closely spaced rulings? (b) Why does it
have a large number of rulings?
6. The relation R = Nm suggests that the resolving power of a given grating can be
made as large as desired by choosing an arbitrarily high order of diffraction. Discuss.
7. Show that at a given wavelength and a given angle of diffraction the resolving power
of a grating depends only on its width W ( = Nd).
8. According to Eq. 45-3 the principal maxima become wider (that is, b..8m increases) the
higher the order m (that is, the larger 8m becomes). According to Eq. 45-6 the resolving
power becomes greater the higher the order m. Explain this apparent paradox.
9. Is the pattern of Fig. 45-12 more properly described as a diffraction pattern or as an
interference pattern?
10. For a given family of planes in a crystal, can the wavelength of incident X-rays be
(a) too large or (b) too small to form a diffracted beam?
11. If a parallel beam of X-rays of wavelength)'.. is allowed to fall on a randomly oriented
crystal of any material, generally no intense diffracted beams will occur. Such beams
* This relation cannot be written down unless it is known that the structure of NaCl is
cubic. 'fhis can be determined, however, by inspection of the symmetry of the spots in
Fig. 45-12; the wavelength of the X-rays need not be known.
1144 GRATINGS AND SPECTRA Chap. 45
appear if (a) the X-ray beam consists of a continuous distribution of wavelengths rather
than a single wavelength or (b) the specimen is not a single crystal but a finely divided
powder. Explain.
12. Why cannot a simple cube of edge a 0 /2 in Fig. 45-13 be used as a unit cell for sodium
chloride?
13. How would you measure (a) the dispersion D and (b) the resolving power R for
either a prism or a grating spectrograph.

PROBLEMS
f
/
/ Given a grating with 4000 lines/cm, how many orders of entire visible spectrum
(4000-7000 A) can be produced?
2. Derive this expression for the intensity pattern for a three-slit "grating":

le = ·/J-Im(l + 4 cos¢+ 4 cos 2 ¢),


21rd sin{)
where ¢ = ----
A

Assume that a« X and be guided by the derivation of the corresponding double-slit


formula (Eq. 43-9).
3. (a) Using the result of Problem 2, show that the half-width of the fringes for a three-
slit diffraction pattern, assuming{) small enough so that sin{)"' O, is

X
6.£) =--·
-3.2d

(b) Compare this with the expression derived for the two-slit pattern in Problem 10, Chap-
ter 43. (c) Do these results support the conclusion that for a fixed slit spacing the inter-
ference maxima become sharper as the number of slits is increased?
4. Using the result of Problem 2, show that a three-slit "grating" has only one secondary
maximum. Find its location and its relative intensity.
5. A grating designed for use in the infrared region of the electromagnetic spectrum is
"blazed" to concentrate all its intensity in the first order (m = 1) for X = 80,000 A. If
visible light (4000 A < X < 7000 A) were allowed to fall on this grating, what visual ap-
pearance would the diffracted beams present?
6. The central intensity maximum formed by a grating, along with its subsidiary
secondary maxima, can be viewed as the diffraction pattern of a single "slit" whose width
is that of the entire grating. Treating the grating as a single wide slit, assuming that
m = 0, and using the methods of Section 44-4, show that Eq. 45-2 can be derived.
7. A grating has 8000 rulings/in. For what wavelengths in the visible spectrum can
fifth-order diffraction be observed?
8. A diffraction grating has 5000 rulings/in., and a strong diffracted beam is noted at
{) = 30°. (a) What are the possible wavelengths of the incident light? (b) How could you
identify them in an actual case?
9. A diffraction grating 2.0 cm wide has 6000 rulings. At what angles will maximum-in-
tensity beams occur if the incident radiation has a wavelength of 5890 A?
10. Assume that the limits of the visible spectrum are arbitrarily chosen as 4300 and
6800 A. Design a grating that will spread the first-order spectrum through an angular
range of 20 °.
11. A grating has 8000 rulings/in. and is illuminated at normal incidence by white
light. A spectrum is formed on a screen 30 cm from the grating. If a 1.0-cm square hole
is cut in the screen, its inner edge being 5.0 cm from the central maximum, what range of
wavelengths passes through the hole?
Chap. 45 PROBLEMS 1145

)2:" Assume that light is incident on a grating at an angle if; as shown in Fig. 45-17.
show that the condition for a diffraction maximum is
d(sin if; + sin e) = m">,. m = 0, 1, 2, ....

Only the special case if; = 0 has been treated in this chapter (compare Eq. 45-1).

Fig. 45-17

13. Show that in a grating with alternately transparent and opaque strips of equal
width all the even orders (except m = 0) are absent.
14. A transmission grating with d = 1.50 X 10-4 cm is illuminated at various angles
of incidence by light of wavelength 6000 A. Plot as a function of angle of incidence
(0 to 90°) the angular deviation of the first-order diffracted beam from the incident direc-
tion.
15. Derive Eq. 45-3, that is, the expression for f:j.Om, the angular distance between a
principal maximum of order m and either adjacent minimum.
16. A three-slit grating has separation d between adjacent slits. If the middle slit is
covered up, will the half-width of the intensity maxima become broader or narrower?
See Problem 3 and also Problem 10, Chapter 43 .
. 17. A grating has 40,000 rulings spread over 3.0 in. (a) What is its expected dispersion
D for sodium light(">,. = 5890 A) in the first three orders? (b) What is its resolving power
in these orders?
18. In a particular grating the sodium doublet (see Example 2) is viewed in third order
at 80° to the normal and is barely resolved. Find (a) the grating spacing and (b) the total
width of the rulings.
19. A source containing a mixture of hydrogen and deuterium atoms emits a red doublet
at ">,. = 6563 A whose separation is 1.8 A. Find the minimum number of lines needed in a
diffraction grating which can resolve these lines in the first order.
20. Show that the dispersion of a grating can be written as
tan e
D=--·
h

21. A grating has 6000 rulings/cm and is 6.0 cm wide. (a) What is the smallest wave-
length interval that can be resolved in the third order at").. = 5000 A? (b) For this wave-
length and this grating, can the resolution be improved? How?
22. Light containing a mixture of two wavelengths, 5000 A and 6000 A, is incident
normally on a diffraction grating. It is desired (1) that the first and second principal
maxima for each wavelength appear ate ~ 30°, (2) that the dispersion be as high as pos-
sible, and (3) that the third order for 6000 A be a missing order. (a) What is the separation
1146 GRATINGS AND SPECTRA Chap. 45
between adjacent slits? (b) What is the smallest possible individual slit width? (c) Name
all orders for 6000 A that actually appear on the screen with the values chosen in (a) and
(b).
23. Light of wavelength 6000 A is incident normally on a diffraction grating. Two
adjacent principal maxima occur at sin () = 0.2 and sin () = 0.3, respectively. The fourth
order is a missing order. (a) What is the separation between adjacent slits? (b) What is the
smallest possible individual slit width? (c) Name all orders actually appearing on the
screen with the values chosen in (a) and (b).
24. An optical grating with a spacing d = 15,000 A is used to analyze soft X-rays of
wavelength >,. = 5.0 A: The angle of incidence () is 90° - 'Y, where 'Y is a small angle.
The first order maximum is found at an angle() = 90° - 2(3. Find the value of (3.
25. Monochromatic X-rays (A = 1.25A) fall on a crystal of sodium chloride, making
an angle of 45° with the reference line shown in Fig. 45-18. The planes shown are
those of Fig. 45-16a, for which d = 2.52A. Through what angles must the crystal be
turned to give a diffracted beam associated with the planes shown? Assume that the
crystal is turned about an axis that is perpendicular to the plane of the page. Ignore
the possibility (see Problem 26) that some of these beams may be of zero intensity.

Incident
beam
direction

- - - - - _i_
-----t----- T
----,-----
I d

1
-----,------
1

Fig. 45-18

26. Missing orders in X-ray diffraction. In Example 5 the m = 1 beam, permitted by


interference considerations, has zero intensity because of the diffracting properties of the
unit cell for this geometry of beams and crystal. Prove this. (Hint: Show that the "re-
flection" from an atomic plane through the top of a layer of unit cells is canceled by a
"reflection" from a plane through the middle of this layer of cells. All odd-order beams
prove to have zero intensity.)
27. Assume that the incident X-ray beam in Fig. 45-18 is not monochromatic but con-
tains wavelengths in a band from 0.95 to 1.30 A. Will diffracted beams, associated with
the planes shown, occur? Assume a0 = 2. 75 A.
28. In comparing the wavelengths of two monochromatic X-ray lines, it is noted that
line A gives a first-order reflection maximum at a glancing angle of 30° to the smooth face
of a crystal. Line B, known to have a wavelength of 0.97 angstroms, gives a third-order
reflection maximum at an angle of 60° from the same face of the same crystal. Find the
wavelength of line A.
Polarization
CHAPTER 46

46-1 Polarization
Light, like all electromagnetic radiation, is predicted by electromagnetic
theory to be a transverse wave, the directions of the vibrating electric and
magnetic vectors being at right angles to the direction of propagation instead
of parallel to it as in a longitudinal wave. The transverse waves of Figs. 46-1
and 39-11 have the additional characteristic that they are plane-polarized.
This means that the vibrations of the E vector are parallel to each other for
all points in the wave. At any such point the vibrating E vector and the di-
rection of propagation form a plane, called the plane of vibration; in a plane-
polarized wave all such planes are parallel.

Fig. 46-1 An instantaneous "snapshot" of a plane-polarized wave showing the vectors


E and B along a particular ray. The wave is moving to the right with speed c. The
plane containing the vibrating E vector and the direction of propagation is a pl,ane of
vibration.
1147
1148 POLARIZATION Chap. 46
The transverse nature of light waves cannot be deduced from the inter-
ference or diffraction experiments so far described because longitudinal waves
such as sound waves also show these effects. An experimental basis for be-
lieving that light waves are transverse was provided by Thomas Young in
1817. Two of his contemporaries, Dominique-Franc;ois Arago (1786-1853)
and Augustin Jean Fresnel (1788-1827), were able, by allowing a light beam
to fall on a crystal of calcite, to produce two separate beams (see Section
46-4). Astonishingly, these beams, although coherent, produced no inter-
ference fringes but only a uniform illumination. Young deduced from this
that light must be a transverse wave and that the planes of vibration in the
two beams must be at right angles to each other. Wave disturbances that
act at right angles to each other cannot show interference effects; the student
is asked to prove this in Problem 9. Young's words to Arago were these:
I have been reflecting on the possibility of giving an imperfect explanation of
the affection of light which constitutes polarization without departing from the
genuine doctrine of undulations. It is a principle in this theory that all undu-
lations are simply propagated through homogeneous mediums in concentric
spherical surfaces like the undulations of sound, consisting simply in the direct
and retrograde motions of the particles in the direction of the radius with their
concomitant condensation and rarefactions [that is, longitudinal waves]. And
yet, it is possible to explain in this theory a transverse vibration, propagated
also in the direction of the radius, and with equal velocity, the motions of the
particles being in a certain constant direction with respect to that radius; and
this is a polarization.
Note how Young presents the possibility of a transverse vibration as a novel
idea, light having been generally-but incorrectly-assumed to be a longi-
tudinal vibration.
In a plane-polarized transverse wave it is necessary to specify two direc-
tions, that of the wave disturbance (E, say) and that of propagation. In a
longitudinal wave these directions are identical. In plane-polarized trans-
verse waves, but not in longitudinal waves, we may thus expect a lack of
symmetry about the direction of propagation. Electromagnetic waves in
the radio and microwave range exhibit this lack of symmetry readily. Such
a wave, generated by the surging of charge up and down in the dipole that
forms the transmitting antenna of Fig. 46-2, has (at large distances from the
dipole and at right angles to it) an electric field vector parallel to the dipole
axis. When this plane-polarized wave falls on a second dipole connected to a
microwave detector, the alternating electric component of the wave will
cause electrons to surge back and forth in the receiving antenna, producing
a reading on the detector. If we turn the receiving antenna through 90°
about the direction of propagation, the detector reading drops to zero. In
this orientation the electric field vector is not able to cause charge to mov8
along the dipole axis because it points at right angles to this axis. We can
reproduce the experiment of Fig. 46-2 by turning the receiving antenna of a
television set (assumed an electric dipole type) through 90° about an axis
that points toward the transmitting station.
Common sources of visible light differ from radio and microwave sources
Sec. 46-1 POLARIZATION 1149

Microwave ---[>----- Microwave


transmitter c receiver

Antenna) t Anten~n-a---~

Fig. 46-2 The vectors E in the transmitted wave are parallel to the axis of the receiving
antenna so that the wave will be detected. If the receiving antenna is rotated through
90° about the direction of propagation, no signal will be detected.

in that the elementary radiators, that is, the atoms and molecules, act inde-
pendently. The light propagated in a given direction consists of independent
wavetrains whose planes of vibration are randomly oriented about the direc-
tion of propagation, as in Fig. 46-3b. Such light, though still transverse,
is unpolarized. The random orientation of the planes of vibration produces
symmetry about the propagation direction, which, on casual study, conceals
the true transverse nature of the waves. To study this transverse nature, a
way must be found to unsort the different planes of vibration.

Fig. 46-3 (a) A plane-polarized transverse wave moving toward the reader, showing only
the electric vector. (b) An unpolarized transverse wave viewed as a random superposition
of many plane-polarized wavetrains. (c) A second, completely equivalent, description of
an unpolarized transverse wave; here the unpolarized wave is viewed as two plane-polar-
ized waves with a random phase difference. The orientation of the x and y axes about the
propagation direction is completely arbitrary.
1152 POLARIZATION Chap. 46

Fig. 46-7 Two square sheets of Polaroid are laid over a book. In (a) the axes of polar-
ization of the two sheets are parallel and light passes through both sheets. In (b) one
sheet has been rotated 90° and no light passes through. The book is opened to an illus-
tration of the Luxembourg Palace in Paris. Malus discovered the phenomenon of polar-
ization by reflection while looking at sunlight reflected off the palace windows through a
calcite crystal.

than those so far described (see Section 46-3). Equation 46-1 describes
precisely the lack of symmetry about the propagation direction that must be
exhibited by plane-polarized transverse waves. Longitudinal waves could
not possibly show such effects. Interestingly enough the human eye, under
certain conditions, can detect polarized light.*
._ Example 1. Two polarizing sheets have their polarizing directions parallel so
that the intensity Im of the transmitted light is a maximum. Through what angle
must either sheet be turned if the intensity is to drop by one-half?
From Eq. 46-1, since I = !Im, we have
!Im = Im cos 2 fJ

or fJ = cos- 1 ± 0 = ±45°, ±135°.

The same effect is obtained no matter which sheet is rotated or m which direc-
tion. .....
Historically polarization studies were made to investigate the nature of
light. Today we reverse the procedure and deduce something about the
nature of an object from the polarization state of the light emitted by or
scattered from that object. It has been possible to deduce, from studies of
* The so-called Haidinger's brushes; the interested student is referred to Concepts of
Classical Optics, John Strong, W. H. Freeman & Co., 1958.
Sec. 46-3 POLARIZATION BY REFLECTION 1153

the polarization of light reflected from them, that the grains of cosmic dust
present in our galaxy have been oriented in the weak galactic magnetic field
( ""'2 X 10-4 gauss) so that their long dimension is parallel to this field.
Polarization studies have shown that Saturn's rings consist of ice crystals.
The size and shape of virus particles can be determined by the polarization of
ultraviolet light scattered from them. Much useful information about the
structure of atoms and nuclei is gained from polarization studies of their
emitted radiations in all parts of the electromagnetic spectrum. Thus we
have a useful research technique for structures ranging in size from a galaxy
(,...._,10+ 20 meters) to a nucleus (,...__,10- 14 meter). Polarized light also has
many practical applications in industry and in engineering science.

46-3 Polarization by Reflection


Malus discovered in 1809 that light can be partially or completely polar-
ized by reflection. Anyone who has watched the sun's reflection in water,
while wearing a pair of sunglasses made of polarizing sheet, has probably
noticed the effect. It is necessary only to tilt the head from side to side,
thus rotating the polarizing sheets, to observe that the intensity of the re-
flected sunlight passes through a minimum.
Figure 46-8 shows an unpolarized beam falling on a glass surface. The
E vector for each wavetrain in the beam can be resolved into two components,
one perpendicular to the plane of incidence-which is the plane of Fig. 46-8-
and one lying in this plane. The first component, represented by the dots,
is called the er-component, from the German senkrecht, meaning perpendi-
cular. The second component, represented by the arrows, is called the
1r-component (for parallel). On the average, for completely unpolarized in-
cident light, these two components are of equal amplitude.

Fig. 46-8 For a particular angle of incidence Op, the reflected light is completely polar-
ized, as shown. The transmitted light is partially polarized.
1154 POLARIZATION Chap. 46

Fig. 46-9 Polarization of light by a stack of glass plates. Unpolarized light is incident
on a stack of glass plates at Brewster's angle 8p. (Polarization in the plane of the page is
shown by the short lines and polarization normal to the page by the dots.) All light re-
flected out of the original ray is polarized normal to the page. After passing through
several reflecting interfaces, the light in the original ray no longer contains any appreciable
component polarized normal to the page.

Experimentally, for glass or other dielectric materials, there is a particular


angle of incidence, called the polarizing angle Op, at which the reflection co-
efficient for the 1r-component is zero. This means that the beam reflected
from the glass, although of low intensity, is plane-polarized, with its plane
of vibration at right angles to the plane of incidence. This polarization of
the reflected beam can easily be verified by analyzing it with a polarizing
sheet.
The 1r-component at the polarizing angle is entirely refracted; the O"-com-
ponent is only partially refracted. Thus the transmitted beam, which is of
high intensity, is only partially polarized. By using a stack of glass plates
rather than a single plate, reflections from successive surfaces occur and the
intensity of the emerging reflected Cu-component) beam can be increased
(see Fig. 46-9). By the same token, the u-components are progressively
removed from the transmitted beam, making it more completely 1r-polarized.
At the polarizing angle it is found experimentally that the reflected and
the refracted beams are at right angles, or (Fig. 46-8)
Op + Or = 90°.
From Snell's law,
Sec. 46-4 DOUBLE REFRACTION 1155

Combining these equations leads to

or (46-2)

where the incident ray is in medium one and the refracted ray in medium
two. This can be written as
tan Op = n, (46-3)

where n ( = n 2 /n 1 ) is the index of refraction of medium two with respect to


medium one. Equation 46-3 is known as Brewster's l,aw after Sir David
Brewster (1781-1868), who deduced it empirically in 1812. It is possible to
prove this law rigorously from Maxwell's equations.

.... Example 2. We wish to use a plate of glass (n = 1.50) as a polarizer. What is the
polarizing angle? What is the angle of refraction?
From Eq. 46-3,
Op = tan- 1 1.50 = 56.3°.
The angle of refraction follows from Snell's law:
(1) sin Op = n sin Or

. = sin 56.3° = 0 555


or sm 0r 1.50 . Or = 33.7°.

~ Double Refraction
In earlier chapters we assumed that the speed of light, and thus the index
of refraction, is independent of the direction of propagation in the medium
and of the state of polarization of the light. Liquids, amorphous solids such
as glass, and crystalline solids having cubic symmetry normally show this
behavior and are said to be optically isotropic. Many other crystalline solids
are optically anisotropic (that is, not isotropic).*
Solids may be anisotropic in many properties. Mica cleaves readily in one
plane only; a cube of crystalline graphite does not have the same electric
resistance between all pairs of opposite faces; a cube of crystalline nickel
magnetizes more readily in certain directions than in others, etc. If a solid
is a mixture of a large number of tiny crystallites, it may appear to be iso-
tropic because of the random orientations of the crystallites. Powdered

* Many transparent amorphous solids such as glasses and plastics become optically
anisotropic when they are mechanically stressed. This fact is useful in engineering design
studies in that strains in gears, bridge structures, etc., can be studied quantitatively by
building plastic models, stressing them appropriately, and examining the optical anisotropy
that results, using polarization techniques. The interested student should consult "Photo-
elasticity," a chapter by H. T. Jessop in Vol. 6 of the Encyclopedia of Physics, edited by
H. Flugge (1958), Springer Verlag, Berlin.
1156 POLARIZATION Chap. 46

Fig. 46-10 Two images, one polarized 90° relative to the other, are formed by a calcite
crystal. The book on which the crystal is lying is Huygens' Opera Reliqua, wherein the
phenomenon of birefringence is discussed.

mica, for example, compacted to a solid mass with a binder, does not exhibit
the cleavage properties that characterize the crystallites making it up .
.Figure 46-10, in which a polished crystal of calcite (CaC0 3 ) is laid over
some printed letters, shows the optical anisotropy of this material; the image
appears double. Figure 46-11 shows a beam of unpolarized light falling on a
calcite crystal at right angles to one of its faces. The single beam splits into
two at the crystal surface. The "double-bending" of a beam transmitted
through calcite, exhibited in Figs. 46-10 and 46-11, is called double refraction.
If the two emerging beams in Fig. 46-11 are analyzed with a polarizing
sheet they are found to be plane-polarized with their planes of vibration at
right angles to each other, a fact discovered by Huygens in 1678. Huygens
used a second calcite crystal to investigate the polarization states of the
beams labeled o and e in the figure.
If experiments are carried out at various angles of incidence, one of the
beams in Fig. 46-11 (represented by the ordinary ray, or o-ray) will be
found to obey Snell's law of refraction at the crystal surface, just like a ray
passing from one isotropic medium into another. The second beam (repre-
sented by the extraordinary ray, or e-ray) will not. In Fig. 46-11, for ex-
ample, the angle of incidence for the incident light is zero but the angle of
refraction of the e-ray, contrary to the prediction of Snell's law, is not. In
general, the e-ray does not even lie in the plane of incidence.
This difference between the waves represented by the o- and e-rays with
respect to Snell's law can be explained in these terms:
Sec. 46-4 DOUBLE REFRACTION 1157

Fig. 46-11 A beam of unpolarized light falling on a calcite crystal is split into two beams
which are polarized at right angles to each other.

1. The a-wave travels in the crystal with the same speed v0 in all directions.
In other words, the crystal has, for this wave, a single index of refraction n 0 ,
just like an isotropic solid.
2. Thee-wave travels in the crystal with a speed that varies with direction
from v0 to a larger value (for calcite) Ve. In other words, the index of refrac-
tion, defined as c/v, varies with direction from n 0 to a smaller value (for cal-
cite) ne.
The quantities n 0 and ne are called the principal indices of refraction for the
crystal. Problem 7 suggests how to measure them. Table 46-1 shows these
indices for six doubly refracting crystals. For three of them the e-wave is
slower; for the other three it is faster. Some doubly refracting crystals (mica,
topaz, etc.) are more complex optically than calcite and require three prin-
cipal indices of refraction for a complete description of their optical proper-
ties. Crystals whose basic crystal structure is cubic (see Fig. 45-13) are
optically isotropic, requiring only one index of refraction.
The behavior for the speeds of the two waves traveling in calcite is sum-
marized by Fig. 46-12, which shows two wave surfaces spreading out from
an imaginary point light source S imbedded in the crystal. The a-wave sur-
1158 POLARIZATION Chap. 46

Table 46-1
PRINCIPAL INDICES OF REFRACTION OF
SEVERAL DO UBL y REFRACTING CRYSTALS
(For sodium light, ;\ = 5890 A)

Crystal Formula no ne ne - no

Ice H20 1.309 1.313 +0.004


Quartz Si02 1.544 1.553 +0.009
Wurz~te ZnS 2.356 2.378 +0.022
Calcite CaCOa 1.658 1.486 -0.172
Dolomite CaO · MgO · 2C02 1.681 1.500 -0.181
Siderite FeO·C02 1.875 1.635 -0.240

face is a sphere, as we would expect if the medium were isotropic. The e-


wave surface is an ellipsoid of revolution about a characteristic direction in
the crystal called the optic axis. The two wave surfaces represent light hav-
ing two different polarization states. If we consider for the present only rays
lying in the plane of Fig. 46-12, then (a) the plan€ of polarization for the o-
rays is perpendicular to the figure, as suggested by the dots, and (b) that for
thee-rays coincides with the plane of the figure, as suggested by the short lines.
We describe the polarization states more fully at the end of this section.

Fig. 46-12 Huygens' wave surfaces generated by a point source S imbedded in calcite.
The polarization states for three o-rays and three e-rays are shown by the dots and lines,
respectively. Note that in general (ray Sb) the bars representing the polarization direc-
tion are not perpendicular 't'o the e-rays.
Sec. 46-4 DOUBLE REFRACTION 1159

Fig. 46-13 A calcite crystal; a is 78° 13';


f3 is 102° 21'.

Figure 46-13, which shows a typical calcite crystal that may be obtained
by cleavage from a naturally occurring crystal, shows how to locate the
optic axis. The edges of calcite crystals may have any lengths but the
angles at which the edges intersect always have one or another of two values,
78° 13' or 102° 21'. The optic axis is found by erecting a line at either of
the two corners where three obtuse angles meet (the "blunt" corners), mak-
ing equal angles with the crystal edges. Any line in the crystal parallel to this
line is also an optic axis.
We can use Huygens' principle to study the propagation of light waves in
doubly refracting crystals. Figure 46-14a shows the special case in which
unpolarized light falls at normal incidence on a calcite slab cut from a crystal
in such a way that the optic axis is normal to the surface. Consider a wave-
fron t that, at time t = 0, coincides with the crystal surface. Following
Huygens, we may let any point on this surface serve as a radiating center
for a double set of Huygens' wavelets, such as those in Fig. 46-12. The
plane of tangency to these wavelets represents the new position of this wave-
front at a later time t. The incident beam in Fig. 46-l 4a is propagated
through the crystal without deviation at speed v0 • The beam emerging
from the slab will have the same polarization character as the incident beam.
The calcite slab, in these special circumstances only, behaves like an iso-
tropic material, and no distinction can be made between the o- and the
e-waves.
Figure 46-14b shows two views of another special case, namely, unpolarized
incident light falling at right angles on a slab cut so that the optic axis is
parallel to its surface. In this case also the incident beam is propagated
without deviation. However, we can now identify o- and e-waves that travel
through the crystal with different speeds, v0 and Ve, respectively. These
waves are polarized at right angles to each other.
Some doubly refracting crystals have the interesting property, called dichroism,
in which one of the polarization components is strongly absorbed within the crystal,
1160 POLARIZATION Chap. 46
the other being transmitted with little loss. Dichroism, illustrated in Fig. 46-15,
is the basic operating principle of the commercial Polaroid sheet. The many small
crystallites, imbedded in a plastic sheet with their optic axes parallel, have a polar-
izing action equivalent to that of a single large crystal slab.

Figure 46-14c shows unpolarized light falling at normal incidence on a


calcite slab cut so that its optic axis makes an arbitrary angle with the
crystal surface. Two spatially separated beams are produced, as in Fig.
46-11. They travel through the crystal at different speeds, that for the o-
wave being v0 and that for the e-wave being intermediate between v0 and Ve.

Fig. 46-14 Unpolarized light falls at normal incidence on slabs cut from a calcite crystal.
The Huygens' wavelets are appropriate sections of the figure of revolution about the optic
axis represented by Fig. 46-15. (a) No double refraction or speed difference occurs.
(I>) No double refraction occurs but there is a speed difference. (c) Both double refraction
and a speed difference occur. (d) Same as (c) but showing the polarization states and the
emerging rays.
Sec. 46-4 DOUBLE REFRACTION 1161

Fig. 46-15 Showing the absorption of


one polarization component inside a
dichroic crystal of the type used in
Polaroid sheets.

Note that ray xa represents the shortest optical path for the transfer of light
energy from point x to thee-wavefront. Energy transferred along any other
ray, in particular along ray xb, would have a longer transit time, a conse-
quence of the fact that the speed of e-waves varies with direction.* Figure
46-14d represents the same case as Fig. 46- l 4c. It shows the rays emerging
from the slab, as in Fig. 46-11, and makes clear that the emerging beams are
polarized at right angles to each other, that is, they are cross-polarized.
We now seek to understand, in terms of the atomic structure of optically anisotropic
crystals, how cross-polarized light waves with different speeds can exist. Light is
propagated through a crystal by the action of the vibrating E vectors of the wave on
the electrons in the crystal. These electrons, which experience electrostatic restoring
forces if they are moved from their equilibrium positions, are set into forced periodic
oscillation about these positions and pass along the transverse wave disturbance that
constitutes the light wave. The strength of the restoring forces may be measured
by a force constant k, as for the simple harmonic oscillator discussed in Chapter
15 (see Eq.
In optically isotropic materials the force constant k is the same for all directions
of displacement of the electrons from their equilibrium positions. In doubly re-
fracting crystals, however, k varies with direction. For electron displacements that
lie in a plane at right angles to the optic axis k has the constant value k 0 , no matter
how the displacement is oriented in this plane. For displacements parallel to the

* The student who has not previously read Section 41-6 on Fermat's principle may care
to do so now.
1162 POLARIZATION Chap. 46

Fig. 46-16 Two views of a one-dimensional me-


chanical model for double refraction.

(a) (b)

optic axis, k has the larger value (for calcite) ke.* Note carefully that the speed
of a wave in a crystal is determined by the direction in which the E vectors vibrate
and not by the direction of propagation. It is the transverse E-vector vibrations
that call the restoring forces into play and thus determine the wave speed. Note too
that the stronger the restoring force, that is, the larger k, the faster the wave. For
waves traveling along a stretched cord, for example, the restoring force for the trans-
verse displacements is determined by the tension F in the cord. Equation 19-12
shows that an increase in F means an increase in the wave speed v.
Figure 46--16, a long weighted "tire chain" supported at its upper end, provides a
one-dimensional mechanical analogy for double refraction. It applies specifically
too- and e-waves traveling at right angles to the optic axis, as in Fig. 46--14b. If the
supporting block is oscillated, as in Fig. 46--16a, a transverse wave travels along the
chain with a certain speed. If the block is oscillated lengthwise, as in Fig. 46--16b,
another transverse wave is also propagated. The restoring force for the second wave
is greater than for the first, the chain being more rigid in the plane of Fig. 46--16b
than in the plane of Fig. 46--16a. Thus the second wave travels along the chain with
a greater speed.
In the language of optics we would say that the speed of a transverse wave in the
chain depends on the orientation of the plane of vibration of the wave. If we oscillate
the top of the chain in a random way, the wave disturbance at a point along the chain
can be described as the sum of two waves, polarized at right angles and traveling
with different speeds. This corresponds exactly to the optical situation of Fig. 46--14b.
* For doubly refracting crystals with ne > n 0 (see Table 46-1) k for displacements
parallel to the optic axis is smaller than for those at right angles to it. Also, for crystals
with three principal indices of refraction, there will be three principal force constants.
Such crystals have two optic axes and are called biaxial. The crystals listed in Table 46-1
have only a single optic axis and are called uniaxial.
Sec. 46-5 CIRCULAR POLARIZATION 1163
For waves traveling parallel to the optic axis, as in Fig. 46-14a, or for waves in
optically isotropic materials, the appropriate mechanical analogy is a single weighted
hanging chain. Here there is only one speed of propagation, no matter how the upper
end is oscillated. The restoring forces are the same for all orientations of the plane
of polarization of waves traveling along such a chain.
These considerations allow us to understand more clearly the polarization states
of the light represented by the double-wave surface of Fig. 46-12. For the (spherical)
o-wave surface, the E-vector vibrations must be everywhere at right angles to the
optic axis. If this is so, the same force constant ko will always be operative, and the
o-waves will travel with the same speed in all directions. More specifically, if we
draw a ray in Fig. 46-12 from S to the o-wave surface, considered three-dimension-
ally (that is, as a sphere), the E-vector vibrations will always be at right angles to
the plane defined by this ray and the optic axis. Thus these vibrations will always
be at right angles to the optic axis.
For the (ellipsoidal) e-wave surface, the E-vector vibrations in general have a
component parallel to the optic axis. For rays such as Sa in Fig. 46-12 or for the
e-rays of Fig. 46-14b, the vibrations are completely parallel to this axis. Thus a
relatively strong force constant (in calcite) k. is operative, and the wave speed v.
will be relatively high. For e-rays, such as Sb in Fig. 46-12, the parallel component
of the E-vector vibrations is less than 100%, so that the corresponding wave speed
will be less than v.. For ray Sc, in Fig. 46-12, the parallel component is zero, and
the distinction between o- and e-rays disappears.

46-5 Circular Polarization


Let plane-polarized light of angular frequency w ( = 21rv) fall at normal in-
cidence on a slab of calcite cut so that the optic axis is parallel to the face of
the slab, as in Fig. 46-17. The two waves that emerge will be plane-polarized
at right angles to each other, and, if the incident plane of vibration is at 45°
to the optic axis, they will have equal amplitudes. Since the waves travel
through the crystal at different speeds, there will be a phase difference </> be-
tween them when they emerge from the crystal. If the crystal thickness is
chosen so that (for a given frequency of light) </> = go0 , the slab is called a
quarter-wave plate. The emerging light is said to be circu,larly polarized.
In Section 15-7 we saw that the two emerging plane-polarized waves just
described (vibrating at right angles with a go0 phase difference) can be repre-
sented as the projections on two perpendicular axes of a vector rotating with

Plane-polarized
light

Fig. 46-17 Plane-polarized light falls on a


doubly refracting slab of thickness x cut
with its optic axis parallel to the surface.
The plane of vibration of the incident light
is oriented to make an angle of 45° with
the optic axis.
1164 POLARIZATION Chap. 46
x

(a)

(b)
EE
2
3
EB z

4
EB5
EE
6
EE
7
EE
8

Fig. 46-18 (a) Two plane-polarized waves of equal amplitude and at right angles to each
other are moving in the z direction. They differ in phase by 90°; where one wave has
maximum values, the other is zero. (b) Views of the resultant amplitude of the approach-
ing wave as seen by observers located at the positions shown on the z axis. Note that
each observer will see the resultant vector rotate clockwise with time.

angular frequency w about the propagation direction. These two descrip-


tions of circularly polarized light are completely equivalent. Figure 46-18
clarifies the relationship between these two descriptions.
_.. Example 3. A quartz quarter-wave plate is to be used with sodium light (A =
5890 A). What must its thickness be?
Two waves travel through the slab at speeds corresponding to the two principal
indices of refraction given in Table 46-1 (ne = 1.553 and n 0 = 1.544). If the crystal
thickness is x, the number of wavelengths of the first wave contained in the crystal is
Ne= __;r:__ = xne,
Ae A
where Ae is the wavelength of thee-wave in the crystal and A is the wavelength in air.
For the second wave the number of wavelengths is
No= __;r:__ = xno,
Ao A
where Ao is the wavelength of the o-wave in the crystal. The difference Ne - N 0
must be one-fourth, or
1 x
4 = ~ (ne - no).
This equation yields
x = A = 5890 A = 0.016 mm.
4(ne - n 0 ) (4)(1.553 - 1.544)
Sec. 46-5 CIRCULAR POLARIZATION 1165
This plate is rather thin; most quarter-wave plates are made from mica, splitting the
sheet to the correct thickness by trial and error.
Example 4. A beam of circularly polarized light falls on a polarizing sheet. De-
scribe the emerging beam.
The circularly polarized light, as it enters the sheet, can be represented by
Ex= Em sin wt
and Ey = Em COS wt,
where x and y represent arbitrary perpendicular axes. These equations correctly
represent the fact that a circularly polarized wave is equivalent to two plane-polarized
waves with equal amplitude and a 90° phase difference.
The resultant amplitude in the incident circularly polarized wave is

Ecp = VEx 2 + Ey2 = VEm 2 (sin 2 wt+ cos 2 wt) = Em,


an expected result if the circularly polarized wave is represented as a rotating vector.
The resultant intensity in the incident circularly polarized wave is proportional to
Em 2, or
(46-4)
Let the polarizing direction of the sheet make an arbitrary angle() with the x axis
as shown in Fig. 46-19. The instantaneous value of the plane-polarized wave trans-
mitted by the sheet is
E = Ey sin() + Ex cos()
= Em cos wt sin () + Em sin wt cos ()
= Em sin (wt+ 0).
The intensity of the wave transmitted by the sheet is proportional to E 2 , or
I <X Em 2 sin 2 (wt+ 0).
The eye and other measuring instruments respond only to the average intensity J,
which is found by replacing sin 2 (wt + ()) by its average value over one or more cycles
( = !), or
I- <X 2IEm 2 •
Comparison with Eq. 46-4 shows that inserting the polarizing sheet reduces the in-
tensity by one-half. The orientation of the sheet makes no difference, since() does not

Fig. 46-19 Circularly polarized light


falls on a polarizing sheet. Ex and Ey
are instantaneous values of the two --x
components, their maximum values be-
ing Em,
1166 POLARIZATION Chap. 46
appear in this equation; this is to be expected if circularly polarized light is represented
by a rotating vector, all azimuths about the propagation direction being equivalent.
Inserting a polarizing sheet in an unpolarized beam has just the same effect, so that a
simple polarizing sheet cannot be used to distinguish between unpolarized and cir-
cularly polarized light.
Example 5. A beam of light is thought to be circularly polarized. How may this
be verified?
Insert a quarter-wave plate. If the beam is circularly polarized, the two com-
ponents will have a phase difference of 90° between them. The quarter-wave plate
will introduce a further phase difference of ±90° so that the emerging light will have
a phase difference of either zero or 180°. In either case the light will now be plane-
polarized and can be made to suffer complete extinction by rotating a polarizer in its
path.
Does the quarter-wave plate have to be oriented in any particular way to carry
out this test?
Example 6. A plane-polarized light wave of amplitude Eo falls on a calcite quarter-
wave plate with its plane of vibration at 45° to the optic axis of the plate, which is
taken as they axis; see Fig. 46-20. The emerging light will be circularly polarized.
In what direction will the rotating electric vector appear to rotate? The direction of
propagation is out of the page.
The wave component whose vibrations are parallel to the optic axis (thee-wave)
can be represented as it emerges from the plate as

Ey = (Eo cos 45°) sin wt = ~ E 0 sin wt = Em sin wt.

The wave component whose vibrations are at right angles to the optic axis (the
o-wave) can be represented as
Ex = (Eo sin 45°) sin (wt - 90°) = - ~Eo cos wt = -Em cos wt,

the 90° phase shift representing the action of the quarter-wave plate. Note that Ex

ly \y
I

I
I
(a) (b)

Fig. 46-20 Plane-polarized light falls from behind on a quarter-plate oriented so that
the light emerging from the page is circularly polarized. In this case the electric vector
Em rotates clockwise as seen by an observer facing the light source.
Sec. 46-7 SCATTERING OF LIGHT 1167
reaches its maximum value one-fourth of a cycle later than Ey does, for, in calcite,
wave Ex (the o-wave) travels more slowly than wave Ey (thee-wave).
To decide the direction of rotation, let us locate the tip of the rotating electric
vector at two instants of time, (a) t = 0 and (b) a short time t1 later chosen so that wti
is a small angle. At t = 0 the coordinates of the tip of the rotating vector (see Fig.
46-20a) are
and

At t = ti these coordinates become, approximately,

Figure 46-20b shows that the vector that represents the emerging circular polarized
light is rotating clockwise; by convention such light is called right-circularly polarized,
the observer always being considered to face the light source.
If the plane of vibration of the incident light in Fig. 46-20 is rotated through
±90°, the emerging light will be left-circularly polarized. ~ -·-.

46-6 Angular Momentum of Light


That light waves can deliver linear momentum to an absorbing screen or to a mirror
is in accord with classical electromagnetism, with quantum physics, and with experi-
ment. The facts of circular polarization suggest that light so polarized might also
have angular momentum associated with it. This is indeed the case; once again the
prediction is in accord with classical electromagnetism and with quantum physics.
Experimental proof was provided in 1936 by Beth, who showed that when circularly
polarized light is produced in a doubly refracting slab the slab experiences a reaction
torque.
The angular momentum carried by light plays a vital role in understanding the
emission of light from atoms and of -y-rays from nuclei. If light carries away angular
momentum as it leaves the atom, the angular momentum of the residual atom must
change by exactly the amount carried away; otherwise the angular momentum of the
isolated system atom plus light will not be conserved.
Classical and quantum theory both predict that if a beam of circularly polarized
light is completely absorbed by an object on which it falls, an angular momentum
given by
(46-5)
w

is transferred to the object, where U is the amount of absorbed energy and w the angu-
lar frequency of the light. The student should verify that the dimensions in Eq.
46-5 are consistent.

46-7 Scattering of Light


A light wave, falling on a transparent solid, causes the electrons in the
solid to oscillate periodically in response to the time-varying electric vector
of the incident wave. The wave that travels through the medium is the
resultant of the incident wave and of the radiations from the oscillating
electrons. The resultant wave has a maximum intensity in the direction of
the incident beam, falling off rapidly on either side. The lack of sideways
scattering, which would be essentially complete in a large "perfect" crystal,
1168 POLARIZATION Chap. 46
comes about because the oscillating charges in the medium act cooperatively
or coherently.
When light passes through a gas, we find much more sideways scattering.
The oscillating electrons in this case, being separated by relatively large
distances and not being bound together in a rigid structure, act independently
rather than cooperatively. Thus the rigid cancellation of wave disturbances
that are not in the forward direction is less likely to occur; there is more
sideways scattering.
Light scattered sideways from a gas can be wholly or partially polarized,
even though the incident light is unpolarized. Figure 46-21 shows an un-
polarized beam moving upward on the page and striking a gas atom at a.
The electrons at a will oscillate in response to the electric components of
the incident wave, their motion being equivalent to two oscillating dipoles
whose axes are represented by the arrow and the dot at a. An oscillating
dipole does not radiate along its own line of action. Thus an observer at b
would receive no radiation from the dipole represented by the arrow at a.
The radiation reaching him would come entirely from the dipole represented
by the dot at a; thus this radiation would be plane-polarized, the plane of
vibration passing through the line ab and being normal to the page.
Observers at c and d would detect partially polarized light, since the dipole
represented by the arrow at a would radiate somewhat in these directions.
Observers viewing the transmitted or the back-scattered light would not
detect any polarization effects because both dipoles at a would radiate equally
in these two directions.
A familiar example is the scattering of sunlight by the molecules of the
earth's atmosphere. If the atmosphere were not present, the sky would ap-
pear black except when we looked directly at the sun. This has been verified

Fig. 46-21 Light is polarized either partially (c and d) or completely (b) by scattering
from a gas molecule at a.
Sec. 46-8 DOUBLE SCATTERING 1169

by measurements made in rockets and satellites above the atmosphere. We


can easily check with a polarizer that the light from the cloudless sky is at
least partially polarized. This fact is used in polar exploration in the so-
called solar compass. In this device we establish direction by noting the
nature of the polarization of the scattered sunlight. As is well known, mag-
netic compasses are not useful in these regions. It has been learned * that
bees orient themselves in their flights between their hive and the pollen
sources by means of polarization of the light from the sky; bees' eyes contain
built-in polarization-sensing devices.
It still remains to be explained why the light scattered from the sky is pre-
dominantly blue and why the light received directly from the sun-par-
ticularly at sunset when the length of the atmosphere that it must traverse is
greatest-is red. The cross section of an atom or molecule for light scatter-
ing depends on the wavelength, blue light being scattered more effectively
than red light. Since the blue light is largely scattered, the transmitted light
will have the color of normal sunlight with the blues largely removed; it is
therefore more reddish in appearance.
The fact that the scattering cross section for blue light is higher than that for red
light can be made reasonable. An electron in an atom or molecule is bound there by
strong restoring forces. It has a definite natural frequency, like a small mass sus-
pended in space by an assembly of springs. The natural frequency for electrons in
atoms and molecules is usually in a region corresponding to violet or ultraviolet light.
When light is allowed to fall on such bound electrons, it sets up forced oscillations
at the frequency of the incident light beam. In mechanical resonant systems it is
possible to "drive" the system most effectively if we impress on it an external force
whose frequency is as close as possible to that of the natural resonant frequency. In
the case of light the blue is closer to the natural resonant frequency of the bound
electron than is the red light. Therefore, we would expect the blue light to be more
effective in causing the electron to oscillate, and thus it will be more effectively
scattered.

46-8 Double Scattering


When X-rays were discovered in 1898, there was much speculation whether they
were waves or particles. In 1906 they were established as transverse waves by Charles
Glover Barkla (1877-1944) by means of a polarization experiment.
When the unpolarized X-rays strike scattering block S 1 in Fig. 46-22, they set the
electrons into oscillatory motion. The considerations of the preceding section require
that the X-rays scattered toward the second block be plane-polarized as shown in the
figure. Let this wave be scattered from the second scattering block, and let us ex-
amine the radiation scattered from it by rotating a detector D in a plane at right
angles to the line joining the blocks. The electrons will oscillate parallel to each
other, and the positions of maximum and zero intensity will be as shown. A plot of
detector reading as a function of the angle cp supports the hypothesis that X-rays are
transverse waves. If the X-rays were a stream of particles or a longitudinal wave,
these effects could by no means be so readily understood. Thus Barkla's important
experiment established that X-rays are a part of the electromagnetic spectrum.

* See Scientific American, July 1955, and Bees: Their Vision, Chemical Sense, and
Language, K. von Frisch, Cornell University Press, 1950.
1170 POLARIZATION Chap. 46
Second-scattered
wave, maximum
intensity

wave, zero
intensity

Incident
unpolarized
wave

Fig. 46-22 A double-scattering experiment used by Barkla to show that X-rays are
transverse waves.

In his later studies the student will learn that beams of particles such as electrons,
protons, and pions can be viewed as waves. Scattering (including double scat-
tering) techniques are often used to investigate the polarization characteristics of
such beams.

QUESTIONS

1. Why do sunglasses made of polarizing materials have a marked advantage over those
that simply depend on absorption effects?
2. Unpolarized light falls on two polarizing sheets so oriented that no light is trans-
mitted. If a third polarizing sheet is placed between them, can light be transmitted?
3. Can polarization by reflection occur if the light is incident on the interface from the
side with the higher index of refraction (glass to air, for example)?
4. Is the optic axis of a doubly refracting crystal simply a line or a direction in space?
Has it a direction sense, like an arrow? What about the characteristic direction of a
polarizing sheet?
5. Devise a way to identify the polarizing direction of a sheet of Polaroid.
6. If ice is doubly refracting (see Table 46-1), why don't we see two images of objects
viewed through an ice cube?
7. From Table 46-1, would you expect a quarter-wave plate made from calcite to be
thicker than one made from quartz?
8. Does the e-wave in doubly refracting crystals always travel at a speed given by
c/ne?
9. In Fig. 46-14a and b describe qualitatively what happens if the incident beam falls
on the crystal with an angle of incidence that is not zero. Assume in each case that the
incident beam remains in the plane of the figure.
10. Devise a way to identify the direction of the optic axis in a quarter-wave plate.
11. If plane-polarized light falls on a quarter-wave plate with its plane of vibration
making an angle of (a) 0° or (b) 90° with the axis of the plate, describe the transmitted
light. (c) If this angle is arbitrarily chosen, the transmitted light is called elliptically
polarized; describe such light.
12. What would be the action of a half-wave plate (that is, a plate twice as thick as a
Chap. 46 PROBLEMS 1171

quarter-wave plate) on (a) plane-polarized light (assume the plane of vibration to be at 45°
to the optic axis of the plate), (b) circularly polarized light, and (c) unpolarized light?
13. You are given an object which may be (a) a disk of grey glass, (b) a polarizing sheet,
(c) a quarter-wave plate, or (d) a half-wave plate (see Question 12). How could you
identify it?
14. Can a plane-polarized light beam be represented as a sum of two circularly polarized
light beams of opposite rotation? What effect has changing the phase of one of the cir-
cular components on the resultant beam?
15. How can a right-circularly polarized light beam be transformed into a left-circularly
polarized beam?
16. Could (a) a radar beam and (b) a sound wave in air be circularly polarized?
17. A beam of light is said to be unpolarized, plane-polarized, or circularly polarized.
How could you choose among them experimentally?
18. A parallel beam of light is absorbed by an object placed in its path. Under what
circumstances will (a) linear momentum and (b) angular momentum be transferred to the
object?
19. When observing a clear sky through a polarizing sheet, one finds that the intensity
varies by a factor of two on rotating the sheet. This does not happen when one views a
cloud through the sheet. Can you devise an explanation?

PROBLEMS

1. Unpolarized light falls on two polarizing sheets placed one on top of the other. What
must be the angle between the characteristic directions of the sheets if the intensity of the
transmitted light is (a) one-third the maximum intensity of the transmitted beam or (b)
one-third the intensity of the incident beam? Assume that the polarizing sheet is ideal,
that is, that it reduces the intensity of unpolarized light by exactly 50%.
2. An unpolarized beam of light is incident on a group of four polarizing sheets which
are lined up so that the characteristic direction of each is rotated by 30° clockwise with
respect to the preceding sheet. What fraction of the incident intensity is transmitted?
3. Describe the state of polarization represented by these sets of equations:

(a) Ex = E sin (kz - wt)

E11 = E cos (kz - wt),

(b) Ex = E cos (kz - wt)

E11 = E cos ( kz - wt + ~),


(c) Ex = E sin (kz - wt)

E 11 = -E sin (kz - wt).

4. (a) At what angle of incidence will the light reflected from water be completely
polarized? (b) Does this angle depend on the wavelength of the light?
5. Calculate the range of polarizing angles for white light incident on fused quartz.
Assume that the wavelength limits are 4000 and 7000 A and use the dispersion curve of
Fig. 41-2.
6. A narrow beam of unpolarized light falls on a calcite crystal cut with its optic axis
as shown in Fig. 46-23. (a) Fort = 1.0 cm and for (Ji = 45°, calculate the perpendicular
distance between the two emerging rays x and y. (b) Which is the o-ray and which the
e-ray? (c) What are the states of polarization of the emerging rays? (d) Describe what
1172 POLARIZATION Chap. 46
happens if a polarizer is placed in the incident beam and rotated. (Hint: Inside the crystal
the E-vector vibrations for one ray are always perpendicular to the optic axis and for the
other ray they are always parallel. The two rays are described by the indices n 0 and ne;
in this plane each ray obeys Snell's law.)

Fig. 46-23

7. A prism is cut from calcite so that the optic axis is parallel to the prism edge as shown
in Fig. 46-24. Describe how such a prism might be used to measure the two principal
indices of refraction for calcite. (Hint: See hint in
Problem 6; see also Example 3, Chapter 41.)
8. How thick must a sheet of mica be if it
is to form a quarter-wave plate for yellow light
(X = 5890 A)? Mica cleaves in such a way that the
appropriate indices of refraction, for transmission
at right angles to the cleavage plane, are 1.6049
and 1.6117.
9. Prove that two plane-polarized light waves of
equal amplitude, their planes of vibration being at
right angles to each other, cannot produce inter-
ference effects. (Hint: Prove that the intensity of
the resultant light wave, averaged over one or more
Fig. 46-24
cycles of oscillation, is the same no matter what
phase difference exists between the two waves.)
10. Show that in a parallel beam of circularly polarized light the angular momentum
per unit volume Lv is given by
p
Lv = -,
we

where Pis the power per unit area (watts/cm2, say) of the beam. Start from Eq. 46-5.
11. Assume that a parallel beam of circularly polarized light whose intensity is 100 watts
is absorbed by an object. At what rate is angular momentum transferred to the object?
If the object is a flat disk of diameter 5.0 mm and mass 1.0 X 10-2 gm, after how long a
time (assuming it is free to rotate about its axis) would it attain an angular speed of 1.0
rev /sec? Assume a wavelength of 5000 A.
Light and Quantum Physics
CHAPTER 47

47-1 Sources of Light


We have studied the propagation, reflection, refraction, diffraction, polari-
zation, scattering, and interference of light. This chapter deals in part with
the production of light and with the way that such studies led, in 1900, to
the birth of modern quantum physics.
The most common light sources are heated solids and gases through which
an electric discharge is passing. The tungsten filament of an incandescent
lamp and the familiar neon sign are examples in each category. By analyzing
the light from a source with a spectrometer, we can learn how strongly it
radiates at various wavelengths. Figure 4 7-1, which is typical of spectra for
heated solids, shows the results of such measurements for a heated tungsten
ribbon at 2000°K.
The ordinate CR>, in Fig. 47-1 is called the spectral radiancy, defined so that
the quantity CR>, dX is the rate at which energy is radiated per unit area of sur-
face for wavelengths lying in the interval)\ to )\ + dX. Typical units for CR\
are watts/cm 2-µ; the corresponding units of CR\ dX are watts/cm 2 . In meas-
uring CR>,, all radiation emerging into the forward hemisphere is included.
Sometimes we wish to discuss the radiated energy without regard to its
wavelength. An appropriate quantity here is the radiancy CR, defined as the
rate per unit surface area at which energy is radiated into the forward
hemisphere, appropriate units being watts/cm 2 . It can be found by in-
tegrating the radiation present in all -wavelength intervals:

(47-1)
1174 LIGHT AND QUANTUM PHYSICS Chap. 47

~ 301--~~~-+-~~~-+--'~~~-+-~~~-1-~~~--l
I
IN
E
..g
~
;
~201--~~-++-~~~-+-~--'',--+-~~~-1-~~~--1
e5

Fig. 47-1 The spectral radiancy of tungsten at 2000°K. The dashed curve refers to a
cavity radiator at the same temperature. One micron (1 µ) = 10-6 meter = 10 4 A.

The radiancy CR can be interpreted as the area under the plot of%. against X.
In Fig. 47-1 this area-and thus CR-is 23.5 watts/cm 2 • The student should
note the formal similarity between such curves and the Maxwell speed dis-
tribution curve of Section 24-2.
For every material there exists a family of spectral radiancy curves like
that of Fig. 47-1, one curve for every temperature. If such families of
curves are compared, no obvious regularities stand out. A quantitative
understanding in terms of a basic theory presents serious difficulties. Fortu-
nately, it is possible to work with an idealized heated solid, called a cavity
radiator. Its light-emitting properties prove to be independent of any par-
ticular material and to vary in a simple way with temperature. In much the
same way it proved convenient earlier to deal with an ideal gas rather than to
analyze the properties of the infinite variety of real gases. The cavity
radiator is the ideal solid as far as its light-emitting characteristics are con-
cerned. We shall describe in the next two sections how the theoretical study
of cavity radiation in 1900 by the German physicist Max Planck (1858-
1947) laid the foundations of modern quantum physics.
47-2 Cavity Radiators
Let us construct a cavity in each of three metal blocks through the walls of
which a small hole is drilled. Let the blocks be made of any suitable materi-
als; for example, tungsten, tantalum, and molybdenum. Let each block be
Sec. 47-2 CAVITY RADIATORS 1175

raised to the same uniform temperature (say 2000°K) as determined by a


suitable thermometer. Finally, let us observe the blocks by their emitted
light in a- dark room. Measurements of CR and CRx show the following:

1. The radiation from the cavity interior is always more intense than the
radiation from the outside wall. Comparison of the two curves in Fig. 47-1
makes this clear for tungsten. For the three materials given, at 2000°K the
ratio of the radiancy for the outside surface to that for the cavity is 0.259
(tungsten), 0.212 (molybdenum), and 0.232 (tantalum).
2. At a given temperature the radiancy of the hole is identical for all three
radiators, in spite of the fact that the radiancies of the outer surfaces are
different. At 2000°K the cavity radiancy (that is, the hole radiancy) is
90.0 watts/ cm 2 •
3. In contrast to the radiancy of the outer surfaces, the cavity radiancy
CRc varies with temperature in a simple way, namely as

(47-2)

where u is a universal constant (the Stefan-Boltzmann constant) whose


measured value is 5.67 X 10-s watt/(meter 2 )(°K 4 ). The radiancy of the
outer surfaces varies with temperature in a more complicated way and is
different for different materials. It is often written as
CR = eCRc = euT4, (47-3)
where e, the emissivity, depends on the material and the temperature.
4. CRx for the cavity radiation varies with temperature in the way shown
in Fig. 47-2. These curves depend only on the temperature and are quite
independent of the material and of the shape and size of the cavity.
Figure 47-3 shows an actual cavity, consisting of a hollow thin-walled
cylinder of tungsten heated by sending an electric current through it. The
cylinder is mounted in an evacuated glass bulb, and a tiny hole is drilled

140 .
fl
I
120
\
<J)
100 \
:!:
c
:::,
\
c::,
Fig. 47-2 The spectral radiancy for cavity
radiation at three different temperatures.
-:e ....
(ti

(ti

-
80

60 \
\
~
,<

40
"-
' y::
v
-,,.•4000° K
3000°K
~

t--
~

"
~2000°K
20 I >-
['-... ~
/
I
,V II -~
1 2 3 4
Wavelength, microns
1176 LIGHT AND QUANTUM PHYSICS Chap. 47

Fig. 47-3 Photograph of an incan-


descent tungsten tube with a small
hole drilled in its wall. The radiation
emerging from the hole is cavity radi-
ation.

through the cylinder wall. It is clear from the photograph that the radiancy
of the cavity interior is greater than that of the cavity walls.
Many of the facts just given about cavity radiation can be deduced from
Fig. 47-4, which shows two cavities made of different materials, of arbitrary
shapes, and with the same wall temperature T. Radiation, described by
CRA, goes from cavity A to cavity B and radiation described by CRB moves
in the opposite direction. If these two rates of energy transfer are not equal,
one end of the composite block will start to heat up and the other end will
start to cool down, which is a violation of the second law of thermodynamics.
(Why?) Thus we must have
(47-4)
where CRc describes the total radiation for all cavities.
Not only the total radiation but also the distribution of radiant energy
with wavelength must be the same for each cavity in Fig. 47-4. This can
be shown by placing a filter between the two cavity openings, so chosen that
it permits only a selected narrow band of wavelengths to pass. Applying
the same argument, we can show that we must have
(47-5)
where CR;>,,c is a spectral radiancy characteristic of all cavities.

Fig. 47-4 Two radiant cavities


initially at the same temperature
are placed together as shown.
Sec. 47-3 PLANCK'S RADIATION FORMULA 1177

47-3 Planck's Radiation Formula


A theoretical explanation for the cavity radiation was the outstanding un-
solved problem in physics during the years before the turn of the present
century. A number of capable physicists advanced theories based on clas-
sical physics, which, however, had only limited success. Figure 47-5, for
example, shows the theory of Wien; the fit to the experimental points is
reasonably good, within the experimental error of the data, but definitely
not exact. Wien's formula is

where c1 and c2 are constants that must be determined empirically by fitting


the theoretical formula to the experimental data.

>,
u
c:
.!!!
""O
~

~
0Q,)
CL.
(/)

0 1 2 3 4 5 6
Wavelength, microns

Fig. 47-5 The circles show the experimental sp~ctral radiancy data of Coblentz for
cavity radiation. The theoretical formulas of Wien and Planck are also shown, Planck'i,
providing an excellent fit to the data.

In 1900 Max Planck pointed out that if Wien's formula were modified in
a simple way it would prove to fit the data precisely. Planck's formula, an-
nounced to the Berlin Physical Society on October 19, 1900, was
C1 1
(RA= - --- (47-6)
A5 ec2/AT -1

This formula, though interesting and important, was still empirical at that
stage and did not constitute a theory.
Planck sought such a theory in terms of a detailed model of the atomic
processes taking place at the cavity walls. He assumed that the atoms that
1178 LIGHT AND QUANTUM PHYSICS Chap. 47
make up these walls behave like tiny electromagnetic oscillators, each with
a characteristic frequency of oscillation. The oscillators emit electromag-
netic energy into the cavity and absorb electromagnetic energy from it.
Thus it should be possible to deduce the characteristics of the cavity radia-
tion from those of the oscillators with which it is in equilibrium.
Planck was led to make two radical assumptions about the atomic oscil-
lators. As eventually formulated, these assumptions are the following:

1. An oscillator cannot have any energy but only energies given by *


E = nhv, (47-7)
where v is the oscillator frequency, h is a constant (now called Planck's
constant), and n is a number (now called a quantum number) that can take
on only integral values. Equation 47-7 asserts that the oscillator energy is
quantized.
2. The oscillators do not radiate energy continuously, but only in "jumps,"
or quanta. These quanta of energy are emitted when an oscillator changes
from one to another of its quantized energy states. Thus, if n changes by one
unit, Eq. 47-7 shows that an amount of energy given by
D..E = D..nhv = hv (47-8)
is radiated. As long as an oscillator remains in one of its quantized states
(or stationary states as they are called), it neither emits nor absorbs energy.

These assumptions were radical ones and, indeed, Planck himself resisted
accepting them wholeheartedly for many years. In his words, "My futile
attempts to fit the elementary quantum of action [that is, the quantity h]
somehow into the classical theory continued for a number of years, and they
cost me a great deal of effort."
Consider the application of Planck's hypotheses to a large-scale oscillator
such as a mass-spring system or an LC circuit. It would be a stoutly de-
fended common belief that oscillations in such systems could take place with
any value of total energy and not with only certain discrete values. In the
decay of such oscillations (by friction in the mass-spring system or by re-
sistance and radiation in the LC circuit), it would seem that the mechanical
or electromagnetic energy would decrease in a perfectly continuous way and
not by "jumps.'' There is no basis in everyday experience, however, to dis-
miss Planck's assumptions as violations of "common sense," for Planck's
constant proves to have a very small value, namely

h = 6.625 X 10-34 joule-sec.


The following example makes this clear.

* Later developments show that the correct formula for a harmonic oscillator is
E = (n + !)hv.
This change makes no difference to Planck's conclusions, however.
Sec. 47-4 PHOTOELECTRIC EFFECT 1179
Ill- Example 1. A mass-spring system has a mass m = 1.0 kg and a spring constant
k = 20 nt/meter and is oscillating with an amplitude of 1.0 cm. (a) If its energy is
quantized according to Eq. 47-7, what is the quantum number n? (b) If n changes by
unity, what fractional change in energy occurs?
(a) From Eq. 15-8 the frequency is

1 /k 1 /20 nt/meter
.,, = 271" '\J m = 2 '\J
71" l.O kg = 0.71 cycles/sec.

From Eq. 8-8 the mechanical energy is

E = !kx;.ax = !(20 nt/meter)(10- 2 meter) 2 = 1.0 X 10-3 joule.


From Eq. 47-7 the quantum number is

n = E = 1.0 _x 10-3 joule = 2 _1 X 1030 .


hv (6.6 X 10-34 Joule-sec)(0.71 cycles/sec)
(b) If n changes by unity, the fractional change in energy is given by dividing Eq.
17-8 by Eq. 47-7, or
D..E = hv = 1 = _, 10 _ 30
E nhv n ·
Thus for large-scale oscillators the quantum numbers are enormous and the quantized
nature of the energy of the oscillations will not be apparent. Similarly, we are not
aware in large-scale experiments of the discrete nature of mass and the quantized
nature of charge, that is, of the existence of atoms and electrons. <llil

On the basis of his two assumptions, Planck was able to derive his radiation
law (Eq. 47-6) entirely from theory, receiving a Nobel prize for this accom-
plishment in 1918. His theoretical expressions for the hitherto empirical
constants c1 and c2 were
he
and

where k is Boltzmann's constant (see Section 23-5) and c is the speed of light.
By inserting the experimental values for c1 and c2 , Planck was able to derive
the values of both hand k. Planck described his theory to the Berlin Physi-
cal Society on December 14, 1900. Quantum physics dates from that day.
Planck's ideas soon received re-enforcement from Einstein, who, in 1905,
applied the concepts of energy quantization to a new area of physics, the
photoelectric effect.
Before discussing this effect, it is important to realize that although Planck
had quantized the energies of the oscillators in the cavity walls he still treated
the radiation within the cavity as an electromagnetic wave. Einstein's analy-
sis of the photoelectric effect first pointed out the inadequacy of the wave
picture of light in certain situations.

47-4 Photoelectric Effect


Figure 47-6 shows an apparatus used to study the photoelectric effect.
Monochromatic light, falling on metal plate A, will liberate photoelectrons,
which can be detected as a current if they are attracted to metal cup B by
1180 LIGHT AND QUANTUM PHYSICS Chap. 47

Fig. 47-6 An apparatus used to study the photoelectric effect. V cannot only be varied
continuously but can be reversed in sign by a switching arrangement not shown.

means of a potential difference V applied between A and B. Galvanometer


G serves to measure this photoelectric current.
Figure 47-7 (curve a) is a plot of the photoelectric current in an apparatus
like that of Fig. 47-6, as a function of the potential difference V. If V is
made large enough, the photoelectric current reaches a certain limiting value
at which all photoelectrons ejected from plate A are collected by cup B.
If V is reversed in sign, the photoelectric current does not immediately
drop to zero, which proves that the electrons are emitted from A with a
finite velocity. Some will reach cup B in spite of the fact that the electric
field opposes their motion. However, if this reversed potential difference is
made large enough, a value VO (the stopping potential) is reached at which
the photoelectric current does drop to zero. This potential difference V 0,
multiplied by electron charge, measures the kinetic energy K max of the fastest
ejected photoelectron. In other words,
Kmax = eV 0 . (47-9)
Here Kmax turns out to be independent of the intensity of the light as shown
by curve bin Fig. 47-7, in which the light intensity has been reduced to one-
half.
Figure 47-~ shows the stopping potential V 0 as a function of the frequency
of the incident light for sodium. Note that there is a definite cutoff frequency
Sec. 47-4 PHOTOELECTRIC EFFECT 1181

Fig. 47-7 Some data taken with the apparatus of Fig. 47-6. The applied potential
difference Vis called positive when the cup B in Fig. 47-6 is positive with respect to the
photoelectric surface A. In curve b the incident light intensity has been reduced to one-
half that of curve a.

v0 , below which no photoelectric effect occurs. These data were taken by


R. A. Millikan (1868~1953), whose painstaking work on the photoelectric
effect won him the Nobel prize in 1923. Because the photoelectric effect is
largely a surface phenomenon, it is necessary to avoid oxide films, grease, or
other surface contaminants. Millikan devised a technique to cut shavings
from the metal surface under vacuum conditions, a "machine shop in vacuo"
as he called it.
Three major features of the photoelectric effect cannot be explained m
terms of the wave theory of light:

3.0.

0
...
en

~ 2.0
ro
:;::::
A plot of Millikan's
Fig. 47-8
measurements of the stopping po- ...
c:
Cl)

0
c..
tential at various frequencies for Ill)
sodium. The cutoff frequency vo c:
·a.
c..
1.0
is 4.39 X 1014 cycles/sec.
.8
(/)

0
4.0 6.0 8.0 10.0 12.0
Frequency, 1014 cycles/sec
1182 LIGHT AND QUANTUM PHYSICS Chap. 47
1. Wave theory suggests that the kinetic energy of the photoelectrons
should increase as the light beam is made more intense. However, Fig. 47-7
shows that K max ( = eV 0) is independent of the light intensity; this has been
tested over a range of intensities of 107 .
2. According to the wave theory, the photoelectric effect should occur for
any frequency of the light, provided only that the light is intense enough.
However, Fig. 47-8 shows that there exists, for each surface, a characteristic
cutoff frequency v0 . For frequencies less than this, the photoelectric effect
disappears, no matter how intense the illumination.
3. If the energy of the photoelectrons is "soaked up" from the incident
wave by the metal plate, it is not likely that the "effective target area" for
an electron in the metal is much more than a few atomic diameters. Thus,
if the light is feeble enough, there should be a measurable time lag (see Ex-
ample 2) between the impinging of the light on the surface and the ejection
of the photoelectron. During this interval the electron should be "soaking
up" energy from the beam until it had accumulated enough energy to escape.
However, no detectable time lag has ever been measured. This disagreement is
particularly striking when the photoelectric substance is a gas; under these
circumstances the energy of the emitted photoelectron must certainly be
"soaked out of the beam" by a single atom.
.... Example 2. A metal plate is placed 5 meters from a monochromatic light source
whose power output is 10- 3 watt. Consider that a given ejected photoelectron may
collect its energy from a circular area of the plate as large as ten atomic diameters
(10- 9 meter) in radius. The energy required to remove an electron through the metal
surface is about 5.0 ev. Assuming light to be a wave, how long would it take for such a
"target" to soak up this much energy from such a light source?
The target area is 1r (10- 9 meter)2 or 3 X 10- 18 meter 2 ; the area of a 5-meter
sphere centered on the light source is 41r (5 meters)2 ,..._, 300 meters 2 • Thus, if the
light source radiates uniformly in all directions, the rate P at which energy falls on
the target is given by
3 X 10-18 meter 2 ) .
P = (10- 3 watt) ( 300 meters 2 = 10-23 Joule/sec.

Assuming that all this power is absorbed, we may calculate the time required from

t= ( 5_ ev
10- Joule/sec
23
) (1.6 X 10-
1 ev
19
joul~) ,..._, 20 hr.

However, no detectable time lag can be measured under any circumstances.

47-5 Einstein's Photon Theory


Einstein succeeded in explaining the photoelectric effect by making a
remarkable assumption, namely, that the energy in a light beam travels
through space in concentrated bundles, called photons. The energy E of a
single photon (see Eq. 47-8) is given by
E = hv. (47-10)
Recall that Planck believed that light, although emitted from its source
discontinuously, travels through space as an electromagnetic wave. Ein-
Sec. 47-5 EINSTEIN'S PHOTON THEORY 1183

stein's hypothesis suggests that light traveling through space behaves not like
a wave at all but like a particle. Millikan, whose experiments verified
Einstein's ideas in every detail, spoke of Einstein's "bold, not to say reckless,
hypothesis.''
Applying the photon concept to the photoelectric effect, Einstein wrote

hv = Eo + Kmax (47-11)

where hv is the energy of the photon. Equation 47-11 says that a photon
carries an energy hv into the surface. Part of this energy (E 0 ) is used in
causing the electron to pass through the metal surface. The excess energy
(hv - E 0 ) is given to the electron in the form of kinetic energy; if the electron
does not lose energy by internal collisions as it escapes from the metal, it
will exhibit it all as kinetic energy after it emerges. Thus K max represents
the maximum kinetic energy that the photoelectron can have outside the
surface; in nearly all cases it will have less energy than this because of internal
losses.
Consider how Einstein's photon hypothesis meets the three objections
raised against the wave-theory interpretation of the photoelectric effect. As
for objection 1 (the lack of dependence of Kmax on the intensity of illumina-
tion), there is complete agreement of the photon theory with experiment.
Doubling the light intensity merely doubles the number of photons and thus
doubles the photoelectric current; it does not change the energy ( = hv) of the
individual photons or the nature of the individual photoelectric processes
described by Eq. 47-11.
Objection 2 (the existence of a cutoff frequency) follows from Eq. 47-11.
If K max equals zero, we have

hv0 = E 0 ,

which asserts that the photon has just enough energy to eject the photo-
electrons and none extra to appear as kinetic energy. This quantity E 0 is
called the work function of the substance. If v is reduced below v0 , the indi-
vidual photons, no matter how many of them there are (that is, no matter
how intense the illumination), will not have enough energy to eject photo-
electrons.
Objection 3 (the absence of a time lag) follows from the photon theory
because the required energy is supplied in a concentrated bundle. It is not
spread uniformly over a large area, as in the wave theory.
Although the photon hypothesis certainly fits the facts of photoelectricity,
it seems to be in direct conflict with the wave theory of light which, as we
have seen in earlier chapters, has been verified in many experiments. Our
modern view of the nature of light is that it has a dual character, behaving
like a wave under some circumstances and like a particle, or photon, under
others. We discuss the wave-particle duality at length in Chapter 48. Mean-
while, let us continue our studies of the firm experimental foundation on
which the photon concept rests.
1184 LIGHT AND QUANTUM PHYSICS Chap. 47
Let us rewrite Einstein's photoelectric equation (Eq. 47-11) by substitut-
ing eV 0 for Kmax (see Eq. 47-9). This yields, after rearrangement,
h E0
V 0 = -v - ~· (47-12)
e e
Thus Einstein's theory predicts a linear relationship between VO and v, in
complete agreement with experiment; see Fig. 47-8. The slope of the experi-
mental curve in this figure should be h/e, or
h ab 2.20 volt - 0.65 volt
.~--------- = 3.9 X 10-15 volt-sec.
e be (10 X 10 14 - 6 X 10 14 ) cycles/sec
We can find h by multiplying this ratio by the electron charge e,
h = (3.9 X 10-15 volt-sec) (1.6 X 10-19 coul) = 6.2 X 10-34 joule-sec.
From a more careful analysis of this and other data, including data taken
with lithium surfaces, Millikan found the value h = 6.57 X 10-34 joule-sec,
with an accuracy of about 0.5%. This agreement with the value of h derived
from Planck's radiation formula is a striking confirmation of Einstein's pho-
ton concept.
~ Example 3. Deduce the work function for sodium from Fig. 47-8.
The intersection of the straight line in Fig. 47-8 with the horizontal axis is the cut-
off frequency vo. Substituting these values yields
Eo = hv0 = (6.63 X 10- 34 joule-sec)(4.39 X 10 14 cycles/sec)
= 2.92 X 10-19 joule = 1.82 ev.

47-6 The Compton Effect


Compelling confirmation of the concept of the photon as a concentrated
bundle of energy was provided in 1923 by A. H. Compton (1892-1962) who
earned a Nobel prize for this work in 1927. * Compton allowed a beam of
X-rays of sharply defined wavelength A to fall on a graphite block, as in Fig.
47-9, and he measured, for various angles of scattering, the intensity of the
scattered X-rays as a function of their wavelength. Figure 47-10 shows his
experimental results. We see that although the incident beam consists es-
sentially of a single wavelength A the scattered X-rays have intensity peaks
at two wavelengths; one of them is the same as the incident wavelength, the
other, A1 , being larger by an amount LlA. This so-called Compton shift LlA
varies with the angle at which the scattered X-rays are observed.
The presence of a scattered wave of wavelength A1 cannot be understood
if the incident X-rays are regarded as an electromagnetic wave like th~t of
Fig. 39-11. On this picture the incident wave of frequency v causes electrons.
in the scattering block to oscillate at that same frequency. These oscillating
electrons, like charges surging back and forth in a small radio transmitting
*Foran historical account of Compton's researches read "The Scattering of X Rays as
Particles," A.H. Compton, p. 817, Am. J. Phys., December 1961.
Sec. 47-6 THE COMPTON EFFECT 1185
Fig. 47-9 Compton's experimental X-ray
arrangement. Monochromatic X-rays source
of wavelength).. fall on a graphite scat-
terer. The distribution of intensity
with wavelength is measured for X-rays
scattered at any selected angle <p. The
scattered wavelengths are measured by
observing Bragg reflections from a
~>!+!~t-~
t+~Jt IJ ~
Scattm,
~:~:,
crystal; see Eq. 45-8. Their intensities
are measured by a detector, such as an Collimating j Detecto~
ionization chamber. slits

antenna, radiate electromagnetic waves that again have this same frequency
v. Thus, on the wave picture the scattered wave should have the same fre-
quency v and the same wavelength X as the incident wave.
Compton* was able to explain his experimental results by postulating
that the incoming X-ray beam was not a wave but an assembly of photons
of energy E ( = hv) and that these photons experienced billiard-ball-like col-
lisions with the free electrons in the scattering block. The "recoil" photons
emerging from the block constitute, on this view, the scattered radiation.
Since the incident photon transfers some of its energy to the electron with
* P. W. Debye simultaneously and independently offered the same interpretation.

Primary
<P= oo

<P = 45°
Fig. 47-10 Compton's experimental results. The solid
vertical line on the left corresponds to the wavelength )..,
that on the right to )..'. Results are shown for four different
angles of scattering <p. Note that the Compton shift .6.).. for
'P = 90° is h/moc = 0.242 A.

0.750
A,A
1186 LIGHT AND QUANTUM PHYSICS Chap. 47
y y

Photon Electron
-~ x
A v =0

Before After

Fig. 47-11 A photon of wavelength A is incident on an electron at rest. On collision, the


photon is scattered at an angle 'fl with increased wavelength A', while the electron moves
off with speed v in direction 8.

which it collides, the scattered photon must have a lower energy E'; it must
therefore have a lower frequency v'( = E' /h), which implies a larger wave-
length A'(= e/v'). This point of view accounts, at least qualitatively, for the
wavelength shift LlA. Notice how different this particle model of X-ray
scattering is from that based on the wave picture. Now let us analyze a
single photon-electron collision quantitatively.
Figure 47-11 shows a collision between a photon and an electron, the elec-
tron assumed to be initially at rest and essentially free, that is, not bound to
the atoms of the scatterer. Let us apply the law of conservation of
energy to this collision. Since the recoil electrons may have a speed v that
is comparable with that of light, we must use the relativistic expression
for the kinetic energy of the electron. From Eqs. 47-10 and 8-21 we may
write
hv = hv' 2 + (m - mo)e ,

in which the second term on the right is the relativistic expression for the
kinetic energy of the recoiling electron, m being the relativistic mass and m 0
the rest mass of that particle. Substituting e/A for v (and e/A' for v') and
using Eq. 8-20 to eliminate the relativistic mass m leads us to

he = he+ m e2 ( 1 _ 1) (47-13)
A A' 0
Vl - (v/e) 2 ·

Now let us apply the (vector) law of conservation of linear momentum to


the collision of Fig. 47-11. We first need an expression for the momentum
of a photon. In Section 40-2 we saw that if an object completely absorbs
an energy U from a parallel light beam that falls on it the light beam, accord-
ing to the wave theory of light, will simultaneously transfer to the object a
linear momentum given by U /e. On the photon picture we imagine this
momentum to be carried along by the individual photons, each photon trans-
porting linear momentum in amount p = hv / c, where hv is the photon en-
ergy. Thus, if we substitute A for e/v, we can write
E hv h
p= - (47-14)
e c A
Sec. 47-6 THE COMPTON EFFECT 1187

This conclusion, that the momentum of a photon is given by h/A, may also
be deduced from the theory of relativity.
For the electron, the relativistic expression for the linear momentum is
given by Eq. 9-13, or
m 0v
Pe= ·
Vl - (v/c) 2

We can then write for the conservation of the x component of linear mo-
mentum
h h m0v
- = - cos cp
A A'
V +
1 - (v/c)
2 cos (} (47-15)

and for they component

h m 0v .
0 = - sin cp - sm (}. (47-16)
A' Vl - (v/c) 2

Our immediate aim is to find .:1A ( = ").. ' - A), the wavelength shift of the
scattered photons, so that we may compare it with the experimental results
of Fig. 47-10. Compton's experiment did not involve observations of the
recoil electron in the scattering block. Of the five collision variables (A, ")...', v,
cp, and 0) that appear in the three equations (47-13, 47-15, and 47-16) we
may eliminate two. We chose to eliminate v and 0, which deal only with the
electron, thereby reducing the three equations to a single relation among the
variables.
Carrying out the necessary algebraic steps (see Problem 15) leads to this
simple result:
h
.:1A (= A' - A) = - (1 - cos cp). (47-17)
m 0c

Thus the Compton shift .:1A depends only on the scattering angle cp and not
on the initial wavelength A. Equation 47-17 predicts within experimental
error the experimentally observed Compton shifts of Fig. 47-10. Note from
the equation that .:1A varies from zero (for cp = 0, corresponding to a "graz-
ing" collision in Fig. 47-11, the incident photon being scarely deflected) to
2h/m0 c (for cp = 180°, corresponding to a "head-on" collision, the incident
photon being reversed in direction).
It remains to explain the presence of the peak in Fig. 47-10 for which the
wavelength does not change on scattering. This peak can be understood as
resulting from a collision between a photon and electrons bound in an ionic
core in the scattering block. During photon collisions the bound electrons
behave like the free electrons that we considered in Fig. 47-11, with the ex-
ception that their effective mass is much greater. This is because the ionic
core as a whole recoils during the collision. The effective mass M for a carbon
scatterer is approximately the mass of a carbon nucleus. Since this nucleus
1188 LIGHT AND QUANTUM PHYSICS Chap. 47
contains 6 protons and 6 neutrons, we have approximately that M = 12 X
1840m0 = 22,000m 0 . If we replace m 0 by Min Eq. 47-17, we see that the
Compton shift for collisions with tightly bound electrons is immeasurably
small.
As in the cavity radiation problem (see Eq. 47-7) and the photoelectric effect (see
Eq. 47-11), Planck's constant h is centrally involved in the Compton effect. The
quantity h is the central constant of quantum physics. In a universe in which h = 0
there would be no quantum physics and classical physics would be valid in the sub-
atomic domain. In particular, as Eq. 47-17 shows, there would be no Compton effect
(that is, ~A = 0) in such a universe.
~ Example 4. X-rays with A.= 1.00 A are scattered from a carbon block. The
scattered radiation is viewed at 90° to the incident beam. (a) What is the Compton
shift~;\? (b) What kinetic energy is imparted to the recoiling electron?
(a) Putting~ = 90° in Eq. 47-17, we have, for the Compton shift,

h
~A = - (1 - cos 'P)
moc
6.63 x 10- 34 joule-sec (1 - cos goo)
(9.11 X 10-31 kg)(3.00 X 10 8 meters/sec)
= 2.43 X 10- 12 meter = 0.0243 A.
(b) If we put K for the kinetic energy of the electron, we can write Eq. 47-13 as
he he
~ = ;\' + K.
Since A' =A+ ~A, we obtain
he he
A A+ ~A+ K,
which reduces to
hc~A
K = A(A + ~A)
(6.63 X 10-34 joule-sec)(3.00 X 10 8 meters/sec)(2.43 X 10- 12 meter)
(LOO X 10-10 meter)(l.00 + 0.024) X 10-10 meter
= 4.73 X 10- 17 joule = 295 ev.
The student may show that the initial photon energy E in this case ( = hv = he/A) is
12,400 ev so that the photon lost about 2.3% of its energy in this collision. A photon
whose energy was ten times as large ( = 124,000 ev) can be shown to lose 23% of its
energy in a similar collision. This follows from the fact that ~A does not depend on
the initial wavelength. Hence more energetic X-rays, which have smaller wave-
lengths, will experience a larger per cent increase in wavelength and thus a larger per
cent loss in energy. ...

47-7 Line Spectra


We have seen how Planck successfully explained the nature of the radiation
from heated solid objects of which the cavity radiator formed the prototype.
Such radiations form continuous spectra and are contrasted with line spectra
such as that of Fig. 47-12, which shows the radiation emitted from iron ions
Sec. 47-7 LINE SPECTRA 1189

Wavelength, A
3450 3500

Fig. 47-12 A small portion of the spectrum of iron, in the region 3400 to 3500 A.

and atoms in an electric arc struck between iron electrodes. We shall see
that Planck's quantization ideas, suitably extended, lead to an understand-
ing of line spectra also. The prototype for the study of line spectra is that of
atomic hydrogen; being the simplest atom it has the simplest spectrum.
Line spectra are common in all parts of the electromagnetic spectrum.
Figure 47-13 shows a spectrum of the ,y-rays (XI"./ 10- 12 meter) emitted
from a particular radioactive nucleus, an isotope of mercury. Figure 47-14
shows a spectrum of X-rays (X r-,..; 10-10 meter) emitted from a molybdenum
target when struck by a 35-kev electron beam. The sharp emission lines are
superimposed on a continuous background.
Figure 47-15 shows a spectrum associated with the molecule HCl. It
occurs in the infrared, with A 10-5 meter. This is an absorption spectrum
I"./

rather than an emission spectrum, as in Fig. 47-12. Experiment shows that


isolated atoms and molecules absorb radiation, as well as emit it, at discrete
wavelengths.
Figure 47-16 shows a portion of the absorption spectrum of ammonia
(NH 3 ) in the microwave region (A 10-2 meter). Finally, Fig. 47-17
I"./

shows how radiation in the radio-frequency region (A r-,..; 43 meters) is ab-


sorbed by hydrogen molecules placed in a magnetic field.

300
,J
250

c::
.E
200 r
.
Fig. 47-13 A wavelength plot for a gamma ray .!:: 150
emitted by the nucleus Hg 198 . (From data by Du- ~ :::,
Mond and co-workers.) u
0
100 i I
!
50
I \
i \
·-·-•,,,._.. . . ·:-r-·
\

0
3.10 3.05 3.00
-12
Wavelength (uncorrected), 10 m
1190 LIGHT AND QUANTUM PHYSICS Chap. 47
40

30 Fig. 47-14 X-rays from a molyb-


,<
denum target struck by 35-kev

-
~
~ electrons. Note the two sharp
.iii 20 lines rising above a broad con-

-
c Q) tinuous base. The wavelength
c of the most intense line is 7.1 X
10 10-11 meter or 0.71 A. (From
data by Ulrey.)

Wavelength, A

100
~ 80
c:
0 60 i.
:.::::;
a.
...... 40 • I

Cl)
0
..c 20 • /I I II n
<
0 1.0 lJ'" u ' '"" uVII\.
3.8 3.7 3.6 3.5 3.4 3.3
Wavelength, microns

Fig. 47-15 An absorption spectrum of the HCl molecule near A = 3.5 X 10-6 meter =
3.5 µ. (From data by E. S. Imes.)

s::
0
:g_
Fig. 47-16 An oscilloscope trace show- £
..c
ing one strong line and four weak lines (IJ

in the absorption spectrum of ammonia ~


a,
at microwave frequencies. w
s::

Frequency

b
vo-s,~
l
ooov
-uv"' r '\ ( o·o
--, ,}
fl"O Fig. 47-17 A portion of the ab-
sorption spectrum of the protons

-
"iii '
c in molecular hydrogen at A ,..__, 43
Q)

c meters. In this technique the fre-


E quency is left fixed and the sample
(l:J
QJ is placed in a magnetic field, which
Ill
is varied to scan the spectrum.
(From data by Kellogg, Rabi, and
1600 1650 1700 Zacharias.)
Magnetic field, gauss
Sec. 47-8 THE HYDROGEN ATOM 1191

47-8 The Hydrogen Atom


Figure 47-18 shows the spectrum of hydrogen. A theoretical interpreta-
tion of the hydrogen spectrum, based on quantum ideas, was first given by
Niels Bohr in 1913.
We might wish to associate the frequency of an emitted spectrum line
with the frequency of an electron revolving in an orbit inside the atom.
Classical electromagnetism predicts that charges will radiate energy when
they are accelerated. In this way electromagnetic waves are emitted from a
radio transmitting antenna in which electrons are caused to surge back and
forth. This radiation represents a loss of energy for the moving electrons
which, in a radio antenna, is compensated for by supplying energy from an
oscillator. In an isolated atom, however, no energy is supplied from external
sources. We would expect the frequency of the electron and thus that of the
emitted radiation to change continuously as the energy drains away. This
prediction of classical theory cannot be reconciled with the existence of sharp
spectrum lines. Thus classical physics cannot explain the hydrogen, or any
other, spectrum.
Bohr circumvented this difficulty by assuming that, like Planck's oscil-
lators, the hydrogen atom exists in certain stationary states in which it does
not radiate. Radiation occurs only when the atom makes a transition from
one state, with energy Ek, to a state with lower energy Ej. In equation
form
(47-18)

where hv is the quantum of energy carried away by the photon that is emitted
from the atom during the transition.
To learn the allowed frequencies predicted by Eq. 47-18, it is necessary to
know the energies of the various stationary states in which a hydrogen atom
can exist. This calculation was first carried out by Bohr on the basis of a
specific model for the hydrogen atom put forward by him. Bohr's model
was highly successful for hydrogen and had a tremendous influence on the

Fig. 47-18 The spectrum of hydrogen. It consists of a number of series of lines, three
of which are shown. Within each series the spectrum lines follow a regular pattern, ap-
proaching a so-called series limit at the short-wave end of the series.
1192 LIGHT AND QUANTUM PHYSICS Chap. 47
further development of the subject; it is now regarded as an important pre-
liminary stage in the development of a more complete theory of quantum
physics.
Let us assume that the electron in the hydrogen atom moves in a circular
orbit of radius r centered on its nucleus. We assume that the nucleus, which
is a single proton, is so massive that the center of mass of the system is es-
sentially at the position of the proton. Let us calculate the energy E of such
an atom.
Writing Newton's second law for the motion of the electron, we have (using
Coulomb's law)
F = ma,

e2 v2
or --=m-·
4'll"E0 r 2 r

This allows us to calculate the kinetic energy of the electron, which is


2
1 2 e
K = 2 mv = - - · (47-19)
811"Eor

The potential energy U of the proton-electron system is given by

e2
U = V(- e) = - - - , (47-20)
4'll"Eor

where V ( = e/411"Eor) is the potential of the proton at the radius of the elec-
tron.
The total energy E of the system is
e2
(47-21)

Since the orbit radius can apparently take on any value, so can the energy
E. The problem of quantizing E reduces to that of quantizing r.
Every property of the orbit is fixed if the radius is given. Equations 47-19,
47-20, and 47-21 show this specifically for the energies K, U, and E. From
Eq. 47-19 we can show that the linear speed v for the electron is also given in
terms of r by

v= \ ) ~ ·
te2 (47-22)

The rotationalfrequeru;y v0 follows at once from

v
~·o = - = (47-23)
211"r
Sec. 47-8 THE HYDROGEN ATOM 1193

The linear momentum p follows from Eq. 47-22:

= mv = ~ (47-24)
p
'14;;;;;
The angular momentum Lis given by

(47-25)

Thus if r is known, the orbit parameters K, U, E, v, v0 , p, and L are also


known. If any one of these quantities is quantized, all of them must be.
At this stage Bohr had no rules to guide him and so made (after some in-
direct reasoning which we do not reproduce) a bold hypothesis, namely, that
the necessary quantization of the orbit parameters shows up most simply
when applied to the angular momentum and that, specifically, L can take
on only values given by
h
L= n - n = 1, 2, 3, .... (47-26)
21r

Planck's constant appears again in a fundamental way; the integer n is a


quantum number.
Combining Eqs. 47-25 and 47-26 leads to

n = 1, 2, 3, ... ' (47-27)

which tells how r is quantized. Substituting Eq. 47-27 into Eq. 47-21 pro-
duces
me 4 1
E = - -~ - n = 1, 2, 3, ... , (47-28)
8Eo2h2 n2

which gives directly the energy values of the allowed stationary states.
Figure 47-19 shows the energies of the stationary states and their as-
sociated quantum numbers. Equation 47-27 shows that the orbit radius
increases as n 2 . The upper level in Fig. 47-19, marked n = ex:;, corresponds
to a state in which the electron is completely removed from the atom (that
is, E = 0 and r = ex:;). Figure 47-19 also shows some of the quantum jumps
that take place between the different stationary states.
Combining Eqs. 47-18 and 47-28 allows us to write a completely theoreti-
cal formula for the frequencies of the lines in the hydrogen spectrum. It is

v - ~(2-j2 - ~)
8Eo2h3 k2
(47-29)

in which j and k are integers describing, respectively, the lower and the
upper stationary states. The corresponding wavelengths can easily be
1194 LIGHT AND QUANTUM PHYSICS Chap. 47

Fig. 47-19 An energy level diagram for hydrogen showing the quantum number n for
each level and some of the transitions that appear in the spectrum. An infinite number
of levels is crowded in between the levels marked n = 6 and n = oo. Compare this figure
carefully with Fig. 47-18.
Sec. 47-8 THE HYDROGEN ATOM 1195

found from X = c/v. Table 47-1 shows some wavelengths so calculated; it


should be compared carefully with Figs. 47-18 and 47-19.

Table 47-1

THE H YD ROG EN SPECTRUM

(Some selected lines)

Quantum Number
Name
of Wavelength, A
series
j (lower state) k (upper state)

1 2 1216
1 3 1026
Lyman 1 4 970
1 5 949
1 6 940
1 00 912

2 3 6563
2 4 4861
Balmer 2 5 4341
2 6 4102
2 7 3970
2 00 3650

3 4 18751
3 5 12818
Paschen 3 6 10938
3 7 10050
3 8 9546
3 00 8220

1J11,, Example 5. Calculate the binding energy of the hydrogen atom (the energy
binding the electron to the nucleus) from Eq. 47-28.
The binding energy is numerically equal to the energy of the lowest state in Fig.
47-19. The largest negative value of E in Eq. 47-28 is found for n = 1. This yields
me 4
E= ---
8Eih2
(9.11 X 10-31 kg)(l.60 X 10- 19 coul) 4
(8)(8.85 X 10- 12 coul 2 /nt-m 2)2(6.63 X 10-34 joule-sec) 2
-2.17 X 10-18 joule = -13.6 ev,
which agrees with the experimentally observed binding energy for hydrogen.
1196 LIGHT AND QUANTUM PHYSICS Chap. 47
47-9 The Correspondence Principle
Although all theories in physics have limitations, they usually do not
break down abruptly but in a continuous way, yielding results that agree less
and less well with experiment. Thus the predictions of Newtonian mechanics
become less and less accurate as the speed is made to approach that of light.
A similar relationship must exist between quantum physics and classical
physics; it remains to find the circumstances under which the latter theory
is revealed as a special case of the former.
The radius of the lowest energy state in hydrogen (the so-called ground
state) is found by putting n = 1 in Eq. 47-27; it turns out to be 5.3 X 10-11
meter. If n = 10,000, however, the radius is (10,000) 2 times as large or 5.3
mm. This "atom" is so large that we suspect that its behavior should be
accurately described by classical physics. Let us test this by computing the
frequency of the emitted light on the basis of both classical and quantum
assumptions. These calculations should differ at small quantum numbers
but should agree at large quantum numbers. The fact that quantum physics
reduces to classical physics at large quantum numbers is called the correspond-
ence principle. This principle, credited to Niels Bohr, was very useful during
the years in which quantum physics was being developed. Bohr, in fact,
based his theory of the hydrogen atom on correspondence principle argu-
ments.
Classically, the frequency of the light emitted from an atom is equal to v0 ,
its frequency of revolution * in its orbit. This can be expressed in terms of a
quantum number n by combining Eqs. 47-23 and 47-27 to obtain

me4 2
Vo=----· (47-30)
8t:02h3 n3

Quantum physics predicts that the frequency v of the emitted light is


given by Eq. 47-29. Considering a transition between an orbit with quan-
tum number k = n and one with j = n - 1 leads to

me4
v = 8t:02h3
[ 1 1
(n - 1)2 - ;)i
J

me4 [ 2n - 1 J (47-31)
= 8Eo 2 h3 (n - 1) 2n 2 •

As n - oo the expression in the square brackets above approaches 2/n3 so


that v - v0 as n - oo. Table 47-2 illustrates this example of the cor-
respondence principle.

* Integral multiples of the frequency also exist but may be ignored without affecting the
present argument.
Chap. 47 QUESTIONS 1197

Table 47-2
THE CORRESPONDENCE PRINCIPLE AS APPLIED TO THE
HYDROGEN ATOM

Frequency of
Frequency of I Transition to
Quantum Revolution in
Next Lowest Differ-
Number, Orbit
State ence, %
n (Eq. 47-30)
(Eq. 47-31)
cycles/sec
cycles/sec

2 8.20 x 10 14 24.6 x 10 14 67
5 5.26 x 10 13 7.38 x 10 13 29
10 6.57 x 10 12 7.72 x 10 12 14
50 5.25 x 10 10 5.42 x 10 10 3
100 6.578 x 109 6.677 x 109 1.5
1,000 6.5779 x 106 6.5878 X 10 6 0.15
10,000 6.5779 x 10 3 6.5789 x 10 3 0.015

QUESTIONS

1. "Pockets" formed by the coals in a coal fire seem brighter than the coals themselves.
Is the temperature in such pockets appreciably higher than the surface temperature of an
exposed glowing coal?
2. The relation R = <rT 4 (Eq. 47-2) is exact for true cavities and holds for all tempera-
tures. Why don't we use this relation as the basis of a definition of temperature at, say,
100°C?
3. Do all incandescent solids obey the fourth-power law of cemperature, as Eq. 47-3
seems to suggest?
4. A hole in the wall of a cavity radiator is sometimes called a black body. Why?
5. It is stated that if we look into a cavity whose walls are maintained at a constant
temperature no details of the interior are visible. Does this seem reasonable?
6. How can a photon energy be given by E = hv (Eq. 47-10) when the very presence of
the frequency v in the formula implies that light is a wave?
7. In the photoelectric effect, why does the existence of a cutoff trequency speak in favor
of the photon theory and against the wave theory?
8. Why are photoelectric measurements so sensitive to the nature of the photoelectric
surface?
9. Does Einstein's theory of photoelectricity, in which light is postulated to be a pho-
ton, invalidate Young's interference experiment?
10. List and discuss carefully the assumptions made by Planck in connection with
the cavity radiation problem, by Einstein in connection with the photoelectric effect, and
by Bohr in connection with the hydrogen atom problem.
11. In Bohr's theory for the hydrogen atom orbits, what is the implication of the fact
that the potential energy is negative and is greater in magnitude than the kinetic energy?
12. Can a hydrogen atom absorb a photon whose energy exceeds its binding energy
(13.6 ev)?
1198 LIGHT AND QUANTUM PHYSICS Chap. 47
13. Discuss Example 1 in terms of the correspondence principle.
14. According to classical mechanics, an electron moving in an orbit should be able to do
so with any angular momentum whatever. According to Bohr's theory of the hydrogen
atom, however, the angular momentum is quantized according to L = nh/2r. Reconcile
these two statements, using the correspondence principle.

PROBLEMS
1. At what wavelength does a cavity radiator at 6000°K radiate most per unit wave-
length? Solve either analytically or graphically.
2. Using Fig. 47-1, verify that the emissivity of tungsten at 2000°K is 0.259.
3. A cavity radiator at 6000°K has a hole 0.10 mm in diameter drilled in its wall. At
what rate do photons in the range 5500 to 5510 A escape from this hole?
4. A cavity whose walls are held at 4000°K has a circular aperture 5.0 mm in diameter.
(a) At what rate does energy in the visible range (defined to extend from 0.40 to 0.70 µ.)
escape from this hole? (b) What fraction of the total radiation escaping from the cavity
does this represent? Solve either analytically or graphically.
5. Show that Wien's law (p. 1177) is a special case of Planck's law (Eq. 47--6) for short
wavelengths or low temperatures.
6. Solar radiation falls on the earth at a rate of 2.0 cal/cm2-min. How many photons/
cm2-min is this, assuming an average wavelength of 5500 A?
7. A spectral emission line, important in radioastronomy, has a wavelength of 21 cm.
To what photon energy does this correspond?
8. The energy required to remove an electron from sodium is 2.3 ev. Does sodium show
a photoelectric effect for orange light, with JI. = 6800 A?
9. Light of a wavelength 2000 A falls on an aluminum surface. In aluminum 4.2 ev
are required to remove an electron. What is the kinetic energy of (a) the fastest and (b)
the slowest emitted photoelectrons? (c) What is the stopping potential?. (d) What is the
cutoff wavelength for aluminum?
10. The work function for a clean lithium surface is 2.3 ev. Make a rough plot of the
stopping potential Vo versus the frequency of the incident light for such a surface.
11. In Example 2 suppose that the "target" is a single gas atom of 1.0 A radius and that
the intensity of the light source is reduced to 10-5 watt. If the binding energy of the most
loosely bound electron in the atom is 20 ev, what time lag for the photoelectric effect is
expected on the basis of the wave theory of light?
12. A 100-watt sodium vapor lamp radiates uniformly in all directions. (a) At what
distance from the lamp will the average density of photons be 10/cm3? (b) What is the
average density of photons 2.0 meters from the lamp? Assume the light to be mono-
chromatic, with JI. = 5890 A.
13. Show, by analyzing a collision between a photon and a free electron (using rela-
tivistic mechanics), that it is impossible for a photon to give all of its energy to the free
electron. In other terms, the photoelectric effect cannot occur for completely free elec-
trons; the electrons must be bound in a solid or in an atom.
14. Calculate the per cent change in photon energy for a Compton collision with ,pin
Fig. 47-11 equal to 90° for radiation in (a) the microwave range, with JI. = 3.0 cm, (b)
the visible range, with JI. = 5000 A, (c) the X-ray range, with JI. = 1.00 A, and (d) the
gamma ray range, the energy of the gamma-ray photons being 1.0 Mev. What are your
conclusions about the importance of the Compton effect in these various regions of the
electromagnetic spectrum, judged solely by the criterion of energy loss in a single Compton
encounter?
15. Carry out the necessary algebra to eliminate v and fJ from Eqs. 47-13, 47-15, and
47-16 to obtain the Compton shift relation (Eq. 47-17).
16. (a) Using Bohr's formula, calculate the three longest wavelengths in the Balmer
series. (b) Between what wavelength limits does the Balmer series lie?
Chap. 47 PROBLEMS 1199
17. In the ground state of the hydrogen atom, according to Bohr's theory, what are (a)
the quantum number, (b) the orbit radius, (c) the angular momentum, (d) the linear
momentum, (e) the angular velocity, (f) the linear speed, (g) the force on the electron, (h)
the acceleration of the electron, (i) the kinetic energy, (j) the potential energy, and (k) the
total energy?
18. How do the quantities (b) to (k) in Problem 17 vary with the quantum number?
19. How much energy is required to remove an electron from a hydrogen atom in a state
with n = 8?
20. A hydrogen atom is excited from a state with n = 1 to one with n = 4. (a) Cal-
culate the energy that must be absorbed by the atom. (b) Calculate and display on an
energy-level diagram the different photon energies that may be emitted if the atom returns
to its n = 1 state. (c) Calculate the recoil speed of the hydrogen atom, assumed initially
at rest, if it makes the transition from n = 4 ton = 1 in a single quantum jump.
21. A hydrogen atom in a state having a binding energy (this is the energy required to
remove an electron) of 0.85 ev makes a transition to a state with an excitation energy
(this is the difference in energy between the state and the ground state) of 10.2 ev. (a)
Find the energy of the emitted photon. (b) Show this transition on an energy-level dia-
gram for hydrogen, labeling the appropriate quantum numbers.
22. Show on an energy-level diagram for hydrogen the quantum numbers corresponding
to a transition in which the wavelength of the emitted photon is 1216 A.
23. A neutron, with kinetic energy of 6.0 ev, collides with a resting hydrogen atom in
its ground state. Show that this collision must be elastic (that is, kineti.c energy must
be conserved).
24. Apply Bohr's theory to singly ionized helium, that is, to a helium atom with one
electron removed. What 1·8lationship exists between this spectrum and the hydrogen
spectrum?
25. Using Bohr's theory, calculate the energy required to remove the electron from
singly ionized helium.
26. Positronium. Apply Bohr's theory to the positronium atom. This consists of a
positive and a negative electron revolving around their center of mass, which lies halfway
between them. (a) What relationship exists between this spectrum and the hydrogen
spectrum? (b) What is the radius of the ground state orbit? (Hint: It will be necessary
to analyze this problem from first principles because this "atom" has no nucleus; both
· particles revolve about a point halfway between them.)
27. Muonic atoms. Apply Bohr's theory to a muonic atom, which consists of a nucleus
of charge Ze with a negative muon (an elementary particle with a charge of -e and a mass
m that is 207 times as large as the electron mass; see Appendix F) circulating about it.
Calculate (a) the radius of the first Bohr orbit, (b) the ionization energy, and (c) the wave-
length of the most energetic photon that can be emitted. Assume that the muon is
circulating about a hydrogen nucleus (Z = 1).
28. If an electron is rotating in an orbit at frequency vo, classical electromagnetism
predicts that it will radiate energy not only at this frequency but also at 2vo, 3vo, 4vo, etc.;
see footnote on p. 1104. Show that this is also predicted by Bohr's theory of the hydrogen
atom in the limiting case of large quantum numbers.
29. In Table 47-2 show that the quantity in the last column is given by

lOO(v - vo) 150


----~--
jl n
for large quantum numbers.
30. If the angular momentum of the earth due to its motion around the sun were
quantized according to Bohr's relation L = nh/21r, what would the quantum number be?
Could such quantization be detected if it existed?
Waves and Particles
CHAPTER 48

48-1 Matter Waves


In 1924 Louis de Broglie of France reasoned that (a) nature is strikingly
symmetrical in many ways; (b) our observable universe is composed entirely
of light and matter; (c) if light has a dual, wave-particle nature, perhaps
matter has also. Since matter was then regarded as being composed of
particles, de Broglie's reasoning suggested that one should search for a wave-
like behavior for matter.
De Broglie's suggestion might not have received serious attention had he
not predicted what the expected wavelength of the so-called matter waves
would be. We recall that about 1680 Huygens put forward a wave theory of
light that did not receive general acceptance, in part because Huygens was
not able to state what the wavelength of the light was. When Thomas
Young rectified this defect in 1800, the wave theory of light started on its
way to acceptance.
De Broglie assumed that the wavelength of the predicted matter waves
was given by the same relationship that held for light namely, Eq. 47-14, or
h
X = -,
p
which connects the wavelength of a light wave with the momentum of the
associated photons. The dual nature of light shows up strikingly in this
equation and also in Eq. 47-10 (E = hv). Each equation contains within its
structure both a wave concept (v and X) and a particle concept (E and p)
De Broglie predicted that the wavelength of matter waves would also be given by
Eq. 48-1, where p would now be the momentum of the particle of matter.
1200
Sec. 48-1 MATTER WAVES 1201
llJ,, Example 1. What wavelength is predicted by Eq. 48-1 for a beam of electrons
whose kinetic energy is 100 ev?
The velocity of the electrons is found from K = imv2, or

v = /~Ji = /(2)(100 ev)(l.6 X 10-19 joule/ev)


\/ m \/ 9.1 X 10- 31 kg
= 5.9 X 106 meters/sec.
The momentum follows from
p = mv = (9.1 X 10-31 kg)(5.9 X 106 meters/sec) = 5.4 X 10-24 kg-m/sec.
The wavelength (called the de Broglie wavelength) is found from Eq. 48-1 or
X = ~ = 6.6 X 10-34 joule-sec = 1. 2 A.
p 5.4 X 10- 24 kg-m/sec
This is the same order of magnitude as the size of an atom or the spacing between
adjacent planes of atoms in a solid. ....

~F

::en~G
beam)

Fig. 48-1 The apparatus of Davisson and Germer. Electrons from filament Fare accel-
erated by a variable potential difference V. After "reflection" from crystal C they are
collected by detector D.

In 1926 Elsasser pointed out that the wave nature of matter might be
tested in the same way that the wave nature of X-rays was first tested,
namely, by allowing a beam of electrons of the appropriate energy to fall on
a crystalline solid. The atoms of the crystal serve as a three-dimensional
array of diffracting centers for the electron "wave"; we should look for strong
diffracted peaks in certain characteristic directions, just as for X-ray diffrac-
tion.
This idea was tested by C. J. Davisson and L. H. Germer in this country
and by G. P. Thomson in Scotland.* Figure 48--1 shows the apparatus of
* For an historical account of Thomson's researches see "Early Work in Electron
Diffraction," Sir George Thomson, p. 821, Am. J Phys., December, 1961.
1202 WAVES AND PARTICLES Chap. 48

..... Fig. 48-2· The collector current in


c detector D in Fig. 48-1 as a func-
~
....
::::, tion of the kinetic energy of the
u
incident electrons, showing a dif-
fraction maximum. The angle <f>
in Fig. 48-1 is adjusted to 50°. If
an appreciably smaller or larger
value is used, the diffraction max-
imum disappears.

35 40 45 50 55 60 65 70 75
Kinetic energy, ev

Davisson and Germer. Electrons from a heated filament are accelerated


by a variable potential difference V and emerge from the "electron gun" G
with kinetic energy eV. This electron beam is allowed to fall at normal
incidence on a single crystal of nickel at C. Detector Dis set at a particular
angle </> and readings of the intensity of the "reflected" beam are taken at
various values of the accelerating potential V. Figure 48-2 shows that a
strong beam occurs at </> = 50° for V = 54 volts.
All such strong "reflected" beams can be accounted for by assuming that
the electron beam has a wavelength, given by X = h/p, and that "Bragg
reflections" occur from certain families of atomic planes precisely as de-
scribed for X-rays in Section 45-5.
Figure 48-3 shows such a Bragg reflection, obeying the Bragg relationship
mX = 2d sin() m = 1, 2, 3, .... (48-2)

Fig. 48-3 The strong diffracted beam at <f> = 50° and V = 54 volts arises from wave-like
"reflection" from the family of atomic planes shown, for d = 0.91 A. The Bragg angle
(} is 65°. For simplicity, refraction of the diffracted wave as it leaves the surface is ignored.
Sec. 48-2 ATOMIC STRUCTURE AND STANDING WAY"1S 1203

For the conditions of Fig. 48-3 the effective interplanar spacing d can be
shown by X-ray analysis to be 0.91 A. Since cp equals 50°, it follows that
O equals 90° - ! X 50° or 65°. The wavelength to be calculated from Eq.
48-2, if we assume m = 1, is
").. = 2d sine = 2(0.91 A) (sin 65°) = 1.65 A.
The wavelength calculated from the de Broglie relationship A = h/p is, for
54-ev electrons (see Example 1), 1.64 A. This excellent agreement, com-
bined with much similar evidence, is a convincing argument for believing that
electrons are wave-like in some circumstances.
Not only electrons but all other particles, charged or uncharged, show
wave-like characteristics. Beams of slow neutrons from nuclear reactors
are routinely used to investigate the atomic structure of solids. Figure 48-4
shows a "neutron diffraction pattern" for finely powdered lead.
The evidence for the existence of matter waves with wavelengths given
by Eq. 48-1 is strong indeed. Nevertheless, the evidence that matter is

.....
<J)
::,
500
c 400
.E
.... 300
<J)
a.
<JI 200
c
.... 100
0
.....
::,
<J) 0
z 19° 25° 31 ° 37° 43° 49° 55° 61° 67°
20, degrees

Fig. 48-4 A diffraction pattern for powdered lead, using a monochromatic neutron beam
from a nuclear reactor. The peaks represent "Bragg reflections" from the various atomic
planes, IJ being the corresponding "Bragg angle."

composed of particles remains equally strong; see Fig. 10-10. Thus, for
matter as for light, we must face up to the existence of a dual character;
matter behaves in some circumstances like a particle and in others like a wave.

48-2 Atomic Structure and Standing Waves


The motion of electrons in beams is not bounded or limited in the beam
direction. We can make an analogy to a sound wave in a long gas-filled
tube, a wave traveling down a long string, or an electromagnetic wave in a
long waveguide. All four cases can be described by appropriate traveling
waves and, significantly, waves of any wavelength (within a certain range)
can be propagated.
Let these last three waves be bounded by imposing physical restrictions.
For the sound wave this corresponds to inserting end walls on a section of
the long gas-filled pipe, thus forming an acoustic resonant cavity (Section
38-5). For the waves in the string it corresponds to removing a finite sec-
tion of string and clamping it at each end, as a violin string (Section 20-5).
1204 WAVES AND PARTICLES Chap. 48

Fig. 48-5 Showing how an electron wave can be


adjusted in wavelength to fit an integral number of
times around the circumference of a Bohr orbit of
radius r. This concept, like the Bohr orbit concept, is
now regarded as oversimplified.

For the electromagnetic wave it corresponds to inserting end caps on a finite


length of waveguide, thus forming an electromagnetic resonant cavity
(Section 38-6).
Two important changes occur: (a) the motions are now represented by
standing rather than traveling waves and (b) only certain wavelengths (or fre-
quencies) can now exist. This quantization of the wavelength is a direct result
of bounding or limiting the wave. We expect that if electrons are limited in
their motions by being localized in an atom that (a) the electron motion can
be represented by a standing matter wave, and (b) the electron motion will
become quantized, that is, its energy can take on only certain discrete values.
De Broglie was able to derive the Bohr quantization condition for angular
momentum by applying proper boundary conditions to matter waves in the
hydrogen atom. Figure 48-5 suggests an instantaneous "snapshot" of a
standing matter wave associated with an orbit of radius r. The de Broglie
wavelength (A = h/p) has been chosen so that the orbit of radius r contains
an integral number n of the matter waves, or
21rr 21rr
-=--=n n = 1, 2, 3,
A (h/p)
This leads at once to
h
L=pr=n- n = 1, 2, 3, ... '
21r
which is the Bohr quantization condition for L.

48-3 Wave Mechanics


The idea that the stationary states in atoms correspond to standing matter
waves was taken up by Erwin Schrodinger in 1926 and used by him as the
foundation of wave mechanics, one of several equivalent formulations of
quantum physics.
An important quantity in wave mechanics is the wave function w, which
measures the "wave disturbance" of matter waves. For waves on strings
the "wave disturbance" may be measured by a transverse displacement y;
for sound waves it may be measured by a pressure variation p; for electro-
magnetic waves it may be measured by the electric field vector E.
Sec. 48-3 WAVE MECHANICS 1205

We make the physical meaning of the wave disturbance w clear in Section


48-4. Meanwhile, let us study the wave function w(x, t) for a simple, one-
dimensional problem, that of the possible motions a particle of mass m con-
fined between rigid walls of separation las in Fig. 48-6b. The wave function
can be obtained by analogy with a known mechanical problem, that of the
natural modes of vibration of a string of length l, clamped at each end as in
Fig. 48-6a.
In the vibrating string the boundary conditions require that nodes exist at
each end. This means that the wavelength A must be chosen so that
A
l = n- n = 1, 2, 3, ... ,
2
or that the wavelength A is "quantized" by the requirement that
2l
A= - n = 1, 2, 3, .... (48-3)
n
The wave disturbance for the string is represented by a standing wave
whose equation was shown in Section 19-9 to be
y = 2ym sin kx cos wt,
where w ( = 21l'v) is the angular frequency of the wave and k ( = 211' /A) is the
wave number. Since A is quantized, k must be also, or
21!' n1l'
k = - = -- n = 1, 2, 3, ....
A l
which leads to
y = [ 2ym sin n;x] cos wt n = 1, 2, 3, .... (48-4)

(a)

~\
Fig. 48-6 (a) A stretched string of length l
clamped between rigid supports. (b) A
particle of mass m and velocity v confined to
move between rigid walls a distance l apart.

19-11--1> (b)
m v
1206 WAVES AND PARTICLES Chap. 48

Fig. 48-7 Three quantized modes of vibration for the string of Fig. 48-6a. The figure
also represents three of the quantized wave functions for the particle of Fig. 48-6b. The
ordinate is a displacement amplitude in the first case and a wave-function amplitude in
the second.

Inspection of Eq. 48-4 shows that no matter what value of n is selected


nodes exist at x = 0 and at x = l, as required by the boundary conditions.
Figure 48-7 shows plots of the quantity in the square brackets in this equa-
tion (the amplitude of the standing wave) for the modes of vibration of the
string corresponding to n = 1, 2, and 3.
Consider now the particle confined between rigid walls. Since the walls
are assumed to be perfectly rigid, the particle cannot penetrate them so that
'Ir, which represents the particle motion in some way not yet clearly specified
(see Section 48-4), must vanish at x = 0 and x = l. The allowed wave-
lengths of the matter waves must be given by Eq. 48-3, or
Sec. 48-3 WA VE MECHANICS 1207

2Z
A=-·
n
Replacing Aby h/p (see Eq. 48-1) leads to
·nh
p = --, (48-5)
2Z
which shows that the linear momentum of the particle is quantized. The
momentum p ( = mv) is related to the energy E (which is entirely kinetic
and is equal to !mv2 ) by
p = -v'w. (48-6)
Combining Eqs. 48-5 and 48-6 leads to the quantization condition for E,
or
h2
E = n2 - - n = 1, 2, 3, .... (48-7)
8ml2

The particle cannot have any energy, as we would expect classically, but
only energies given by Eq. 48-7.
The matter wave is described, in strict analogy with Eq. 48-4, by

'11" =
n1rx]
['11" m sin -Z- cos wt n = 1, 2, 3, .... (48-8)

Figure 48-7 can serve equally well to show how the amplitude of the standing
matter waves for the ~tates of motion corresponding ton = 1, 2, and 3 varies
throughout the box. We see clearly in this problem how the act of localizing
or bounding a particle leads to energy quantization.
~ Example 2. Consider an electron (m = 9.1 X 10-31 kg) confined by electrical
forces to move between two rigid "walls" separated by 1.0 X 10-9 meter, which is
about five atomic diameters. Find the quantized energy values for the three lowest
stationary states.
From Eq. 48-7, for n = 1, we have
_ h2 2 _ 2 (6.6 X 10-34 joule-sec)2
E - n 8ml 2 - (l) (8)(9.1 X 10- 31 kg)(l.O X 10-9 meter) 2
= 6.0 X 10-20 joule = 0.38 ev.
The energies for the next two states (n = 2 and n = 3) are 2 2 X 0.38 ev = 1.5 ev and
3 2 X 0.38 ev = 3.4 ev.

Example 3. Consider a grain of dust (m = 1.0 µgm = 1.0 X 10-9 kg) confined to
move between two rigid walls separated by 0.1 mm ( = 10-4 meter). Its speed is only
10-5 meter/sec, so that it requires 100 sec to cross the gap. What quantum number
describes this motion?
The energy is
E ( = K) = imv 2 = i(10- 9 kg)(l0- 6 meter/sec) 2
= 5 X 10-22 joule.
1208 WAVES AND PARTICLES Chap. 48
Solving Eq. 48-7 for n yields

n -- v
- ~ E l-h -- v
8m1c., - I( 8) (10 -9 k g )( 5 X 10
-22 . 1)
JOU e.
( 10-4 meter
6 .6 X _ .
10 34 JOU 1e-sec
)

= 3 x 1014.
Even in these extreme conditions the quantized nature of the motion would never be
apparent; we cannot distinguish experimentally between n = 3 X 10 14 and n = 3
X 10 14 + 1. Classical physics, which fails completely for the problem of Example 2,
works extremely well for this problem. .,..
48-4 Meaning of '11
Max Born first suggested that the quantity '11 2 at any particular point is a
measure of the probability that the particle will be near that point. More ex-
actly, if a volume element dV is constructed at that point, the probability
that the particle will be found in the volume element at a given instant is
'112 dV. This interpretation of '11 provides a statistical connection between

=38e,
1~ -
0 ~ x
--
1z Fig. 48-8 The "probability functions"
for three states of motion of the parti-
cle of Fig. 48-6b, along with the corre-
sponding quantized energies for the
conditions of Example 2. The horizon-
tal lines show the predictions of classi-
cal mechanics, in which the probability
function is cons.tant for all positions of
the particle.
I""' ~ = 3
~=3•"
O x l

the wave and the associated particle; it tells us where the particle is likely to
be, not where it is.
For the particle confined between rigid walls the probability that the par-
ticle will lie between two planes that are distance x and x + dx from one wall
(see Fig. 48-8) is given by
n1r
'11 2 dx = 'Vm 2 sin2 - x cos 2 wt dx.
l
Since we are more interested in the particle's spatial location than in its time
behavior, we average '1' 2 over one cycle of the motion. This is equivalent
to replacing cos 2 wt by its average value for one cycle, namely one-half, or
- n1rx
'112 = 1-'¥ 2 sin2 - - n = 1, 2, 3, .... (48-9)
2 m l

Figure 48-8 shows '11 2 for the three stationary states corresponding to
n = 1, 2, and 3. Note that for n = 1 the particle is more likely to be near
Sec. 48-4 MEANING OF '1i° 1209

the center than the ends. This is in sharp contradiction to the results of
classical physics, according to which the particle has the same probability of
being located anywhere between the walls, as shown by the horizontal line
in Fig. 48-8.
The problem of a particle confined between rigid walls has little real applica-
tion in physics. We would prefer to illustrate the wave mechanics of Schro-
dinger by applying it to a more experimentally realizable situation, such
as the hydrogen atom. Only mathematical complexity prevents us from
doing this. We state without proof that when this problem is solved by
wave mechanics the motion of the electron in the ground state of the atom,
defined by putting n = 1 in Eq. 47-28, is described by the following wave
function,
'1i" = f _~
'\j 1ra3
e-r/a COS wt '
(48-10)

h2eo
where a=---·
1rme2

Putting n = 1 in Eq. 47-27 shows that a is the radius of the ground-state


orbit in Bohr's theory. This special interpretation has little meaning in wave
mechanics; a is taken here merely as a convenient unit of length when deal-
ing with atomic problems, having the value 0.529 A.
Ill- Example 4. Consider two hypothetical spherical shells centered on the nucleus
of a hydrogen atom with radii r and r + dr. What is the probability P(r) that the
electron will lie between these shells, as a function of r?
This probability is '1i"2 dV, where dV is the volume between the shells, or 41rr 2 dr.
Thus
'1i" 2 dV = ( ~ wt)\41rr dr) = P(r) dr.
1r!3 e-r/a cos 2

Averaging over the time (that is, replacing cos wt by!) yields for the probability P(r)
2

-- 4r 2
P(r) = -3 e-2r1a.
a
Figure 48-9 shows a plot of this function. Note that the most probable location for
the electron corresponds to the first Bohr radius. Thus in wave mechanics we do not
say that the electron in the n = 1 state in hydrogen goes around the nucleus in a
circular orbit of 0.529 A radius but only that the electron is more likely to be found at
this distance from the nucleus than at any other distance, either larger or smaller. <111111

Fig. 48-9 The probability function


for the ground state of the hydrogen
atom, as calculated from wave me-
chanics. The separation between
the nucleus and the electron is r; a
is the radius of the first Bohr orbit
(0.529 X 10-10 meter), used here
merely as a convenient unit of dis- 1.0 2.0 3.0 4.0
tance. r/a
1210 WAVES AND PARTICLES Chap. 48
48-5 The Uncertainty Principle
Only those quantities that can be measured have any real meaning in
physics. If we could focus a "super" microscope on an electron in an atom
and see it moving around in an orbit, we would declare that such orbits have
meaning. However, we shall show that it is fundamentally impossible to
make such an observation-even with the most ideal instruments that could
conceivably be constructed. Therefore, we declare that such orbits have no
physical meaning.
We observe the moon traveling around the earth by means of the sunlight
that it reflects in our direction. Now light transfers linear momentum to an
object from which it is reflected. In principle, this reflected light would dis-
turb the course of the moon in its orbit, although a little thought shows that
this disturbing effect is negligible.
For electrons the situation is quite different. Here, too, we can hope
to "see" the electron only if we reflect light, or another particle, from it. In
this case the recoil that the electron experiences when the light (photon)
bounces from it completely alters the electron's motion in a way that cannot
be a voided or even corrected for.
It is not surprising that the probability curve of Fig. 48-9 is the most de-
tailed information that we can hope to obtain, by measurement, about the
distribution of negative charge in the hydrogen atom. If orbits such as
those envisaged by Bohr existed, they would be broken up completely in our
attempts to verify their existence. Under these circumstances, we prefer to
say that it is the probability function, and not the orbits, that represents
physical reality.

Fig. 48-10 An incident beam of electrons is diffracted at the slit in screen A, forming a
typical diffraction pattern on screen B. If the slit is made narrower, the pattern becomes
wider.
Sec. 48-5 THE UNCERTAINTY PRINCIPLE 1211
Our inherent inability to describe the motions of electrons in a classical way finds
expression in the uncertainty principle, enunciated by Werner Heisenberg in 1927.
To formulate this principle, consider a beam of monoenergetic electrons of speed v0
moving from left to right in Fig. 48-10. Let us set ourselves the task of measuring
the position of a particular electron in the vertical (y) direction and also its velocity
component Vy in this direction. If we succeed in carrying out these measurements
with unlimited accuracy, we can then claim to have established the position and
motion of the electron (or one component of it at least) with precision. However, we
shall see that it is impossible to make these two measurements simultaneously with un-
limited accuracy.
To measure y we block the beam with an absorbing screen A in which we put a slit
of width fj.y. If an electron gets through the slit, its vertical position must be known
to this accuracy. By making the slit narrower, we can improve the accuracy of this
vertical position measurement as much as we wish.
Since the electron is a wave, it will undergo diffraction at the slit, and a photo-
graphic plate placed at B in Fig. 48-10 will reveal a typical diffraction pattern. The
existence of this diffraction pattern means that there is an uncertainty in the values of
Vy possessed by the electrons emerging from the slit. Let Vya be the value of Vy that
corresponds to an electron landing at the first minimum on the screen, marked by
point a and described by a characteristic angle ea. We take Vya as a rough measure of
the uncertainty /j.vy in v11 for electrons emerging from the slit.
The first minimum in the diffraction pattern is given by Eq. 44-2, or
. e
sm a = fj.y.
;\
If we assume that ea is small enough, we can write this equation as
;\
ea !'-' fj.y. (48-11)

To reach point a, Vya ( = /j.vy) must be such that

ea~ - /j.vy. (48-12)


Vo
Combining Eqs. 48-11 and 48-12 leads to
jj.vy ;\
-;;- = fj.y,
which we rewrite as (48-13)
Now;\, the wavelength of the electron beam, is given by h/p or h/mvo; substituting
this into Eq. 48-13 yields
/j.vy fj.y "-' -hvo .
mvo
We rewrite this as (48-14)
In Eq. 48-14 fj.py ( = m /j.vy) is the uncertainty in our knowledge of the vertical
momentum of the electrons; fj.y is the uncertainty in our knowledge of their vertical
position. The equation tells us that, since the product of these uncertainties is a
constant, we cannot measure py and y simultaneously with unlimited accuracy.
If we want to improve our measurement of y (that is, if we want to reduce jj.y), we
use a finer slit. However (see Eq. 48-11), this will produce a wider diffraction pattern.
A wider pattern means that our knowledge of the vertical momentum component of
the electron has deteriorated, or, in other words, fj.py has increased; this is exactly
what Eq. 48-14 predicts.
1212 WAVES AND PARTICLES Chap. 48
The limits on measurement imposed by Eq. 48-14 have nothing to do with the
crudity of our measuring instruments. We are permitted to postulate the existence
of the finest conceivable measuring equipment. Equation 48-14 represents a funda-
mental limitation, imposed by nature.
Equation 48-14 is a derivation, for a special case, of a general principle known as
the uncertainty principle. As applied to position and momentum measurements, it
asserts that
tipx tix h >
tipy tiy > h
tip. tiz > h. (48-15)
Thus no component of the motion of an electron, free or bound, can be described with un-
limited precision.
Planck's constant h probably appears nowhere that has more deep-seated signi-
ficance than in Eq. 48-15. If this product had been zero instead of h, the classical ideas
about particles and orbits would be correct; it would then be possible to measure both
momentum and position with unlimited precision. The fact that h appears means
that the classical ideas are wrong; the magnitude of h tells us under what circum-
stances these classical ideas must be replaced by quantum ideas. Gamow * has
speculated, in an interesting and readable fantasy, what our world would be like if the
constant h were much larger than it is, so that nonclassical ideas would be apparent to
our sense perceptions.
l)JI> Example 5. An electron has a speed of 300 meters/sec, accurate to 0.01 %. With
what fundamental accuracy can we locate the position of this electron?
The electron momentum is
p = mv = (9.1 X 10-31 kg)(300 meters/sec) = 2.7 X 10-28 kg-m/sec.
The uncertainty in momentum is given to be 0.01 % of this, or
tip = (0.0001)(2.7 X 10-28 kg-m/sec) = 2.7 X 10-32 kg-m/sec.
The minimum uncertainty in position, from Eq. 48-15, is
tix = !!__ = 6.6 X 10-34 joule-sec
tip 2.7 X 10- 32 kg-m/sec
= 2.4 cm.
If the electron momentum has really been determined by measurement to have the
accuracy stated, there is no hope whatever that its position can be known to any
better accuracy than that stated, namely about 1 in. The concept of the electron as a
tiny dot is not very valid under these circumstances.
l)JI> Example 6. A bullet has a speed of 300 meters/sec, accurate to 0.01 %. With
what fundamental accuracy can we locate its position? Its mass is 50 gm ( = 0.05 kg).
This example is the same as Example 5 in every respect save the mass of the particle
involved. The momentum is
p = mv = (0.05 kg)(300 meters/sec) = 15 kg-m/sec
and tip = (0.0001)(15 kg-m/sec) = 1.5 X 10-3 kg-m/sec.
Equation 48-15 yields
6.6 X 10-34 joule-sec _ t
= 4 .4 X 10 31 me er.
A
L>.X =
1.5 X 10-3 kg-m/s~c
* Mr. Tompkins in Wonderland, Macmillan, 1940.
Chap. 48 QUESTIONS 1213

This is so far beyond the possibility of measurement (a nucleus is only about 10-15
meter in diameter) that we can assert that for heavy objects like bullets the un-
certainty principle sets no limits whatever on our measuring procedures. Once again
the correspondence principle shows us how quantum physics reduces to classical
physics under the appropriate circumstances. .._
The uncertainty relation shows us why it is possible for both light and matter to
have a dual, wave-particle, nature. It is because these two views, so obviously op-
posite to each other, can never be brought face to face in the same experimental
situation. If we devise an experiment that forces the electron to reveal its wave
character strongly, its particle character will always be inherently fuzzy. If we
modify the experiment to bring out the particle character more strongly, the wave
character necessarily becomes fuzzy. Matter and light are like coins that can be
made to display either face at will but not both simultaneously. Niels Bohr first
pointed out in his principle of complementarity how the ideas of wave and of particle
complement rather than contradict each other.

QUESTIONS
1. How can the wavelength of an electron be given by A = h/p when the very presence
of the momentum pin this formula implies that the electron is a particle?
2. How could Davisson and Germer be sure that the "54-volt" peak of Fig. 48-2 was a
first-order diffraction peak, that is, that m = 1 in Eq. 48-2?
3. In a repetition of Thomson'r- experiment for measuring e/m for the electron (see
Section 33-8), a beam of 104-ev electrons is collimated by passage through a slit of width
0.50 mm. Why is the beam-like character of the emerging electrons not destroyed by
diffraction of the electron wave at this slit?
4. Why is the wave nature of matter not more apparent to our daily observations?
5. Apply the correspondence principle to the problem of a particle confined between
rigid walls, showing that those features which seem "strange" (that is, the quantization of
energy and the nonuniformity of the probability functions of Fig. 48--8) become undetect-
able experimentally at large quantum numbers.
6. In the n = 1 mode, for a particle confined between rigid walls, what is the probability
that the particle will be found in a small volume element at the surface of either wall?
7. A standing wave can be viewed as the superposition of two traveling waves. Can
you apply this to the problem of a particle confined between rigid walls, giving an interpre-
tation in terms of the motion of the electron?
8. What is the physical significance of the wave function 'Ir?
9. Why does the concept of Bohr orbits violate the uncertainty principle?
10. How can the predictions of wave mechanics be so exact if the only information we
have about the positions of the electrons is statistical?
11. Make up some numerical examples to show the difficulty of getting the uncertainty
principle to reveal itself during experiments with an object whose mass is about 1 gm.
12. Figure 48-8 shows that for n = 3 the probability function for a particle confined
between rigid walls is zero at two points between the walls. How can the particle ever
move across these positions? (Hint: Consider the implications of the uncertainty prin-
ciple.)
13. The uncertainty principle can be stated in terms of angular quantities (compare
Eq. 48-15) as

where t.L is the uncertainty in the angular momentum and ilcf, the uncertainty in the
angular position. For electrons in atoms the angular momentum has definite quantized
values, with no uncertainty whatever. What can we conclude about the uncertainty in the
angular position and about the validity of the orbit concept?
1214 WA YES AND PARTICLES Chap. 48

PROBLEMS

1. A bullet of mass 40 gm travels at 1000 meters/sec. (a) What wavelength can we


associate with it? (b) Why does the wave nature of the bullet not reveal itself through
diffraction effects?
2. What wavelength do we associate with a beam of neutrons whose energy is 0.025 ev?
3. An electron and a photon each have a wavelength of 2.0 A. What are their (a)
momenta and (b) energies?
4. The 50-Bev electron accelerator to be built at Stanford will provide an electron beam
of small wavelength, suitable for probing the fine details of nuclear structure by scattering
experiments. What will this wavelength be and how does it compare to the size of an
average nucleus? (Hint: At these energies it is necessary to use the extreme relativistic
relationship between momentum and energy, namely p = E/c. This is the same relation-
ship used for light (Section 40-2) and is justified whenever the kinetic energy of a particle
is very much greater than its rest energy moc2, as in this case.)
5. In the experiment of Davisson and Germer (a) at what angles would the second- and
third-order diffracted beams corresponding to the strong maximum of Fig. 48-2 occur,
provided they are present, and (b) at what angle would the first-order diffracted beam occur
if the accelerating potential were changed from 54 to 60 volts? The crystal may be rotated.
6. Make a plot of de Broglie wavelength against kinetic energy for (a) electrons and (b)
protons. Restrict the range of energy values to those in which classical mechanics applies
reasonably well. A convenient criterion is that the maximum kinetic energy on each plot
be only about, say, 5% of the rest energy moc 2 for the particular particle.
7. What is the wavelength of a hydrogen atom moving with a velocity corresponding to
the mean kinetic energy for thermal equilibrium at 20°C?
8. If an electron moves from a state represented by n = 3 in Fig. 48-8 to one repre-
sented by n = 2 in that figure, emitting electromagnetic radiation in the process, what are
(a) the energy of the emitted single photon and (b) the corresponding wavelength?
9. A particle is confined between rigid walls separated by a distance l. What is the
probability that it will be found within a distance Z/3 of one wall (a) for n = 1, (b) for
n = 2, (c) for n = 3, and (d) under the assumptions of classical physics?
10. In the ground state of the hydrogen atom show that the probability Pr that the
electron lies within a sphere of radius r is given by

Pr= 1 - e-2r/a (2r2 + 2r + 1)·


a2 a
Does this yield expected values for (a) r = 0 and (b) r = ~? (c) State clearly the difference
in meaning between this expression and that given in Section 48-4.
11. In the ground state of the hydrogen atom, what is the probability that the electron
will lie within a sphere whose radius is that of the first Bohr orbit?
12. The uncertainty in the position of an electron is given as about 0.5 A, which is the
radius of the first Bohr orbit in hydrogen. What is the uncertainty in the linear mo-
mentum of the electron?
13. Show that if the uncertainty in the location of a particle is equal to its de Broglie
wavelength the uncertainty in its velocity is equal to its velocity.
14. The highest achievable resolving power of a microscope is limited only by the
wavelength used; that is, the smallest detail that can be separated is about equal to the
wavelength. Suppose one wishes to "see" inside an atom. Assuming the atom to have a
diameter of 1.0 A, this means that we wish to resolve detail of separation about 0.1 A.
(a) If an electron microscope is used, what minimum energy of electrons is needed? (b)
If a light microscope is used, what minimum energy of photons is needed? (c) Which
microscope seems more practical for this purpose? Explain.
15. A microscope using photons is employed to locate an electron in an atom to within
a distance of 0.1 A. What is the uncertainty in the momentum of the electron located in
this way?
The Differential Form of Maxwell's
Equations and the Electromagnetic
Wave Equation
SUPPLEMENTARY TOPIC V*

V-1 Introduction
In Chapter 39 we sought to make the existence of electromagnetic waves plausi'-
ble by showing that such waves are consistent with Maxwell's equations as
expressed in Table 38-3. Here we seek to start from Maxwell's equations and
derive from them a differential equation whose solutions will describe electro-
magnetic waves. We will show directly that the speed c of such waves is given by
Eq. 39-15, or c = 1 / ~ .
We followed a similar program in Supplementary Topic III for mechanical
waves on a stretched string. Starting from Newton's laws of motion we derived
a differential equation (Eq. III-1) whose solutions (Eq. III-2) described such
waves. We showed further that the speed v of these waves is given by Eq. III-3,
orv = ~ -
In Table 38-3 we wrote Maxwell's equations as

Eo :f E · dS = q, (V-1)

:f B · dS = 0, (V-2)

:f B · di = µo(i + Eo d<PE/dt), (V-3)


and :f E. di = -d<PB/dt. (V-4)
* Supplementary Topics I to IV appear in Part I.
1
2 SUPPLEMENTARY TOPIC V

These equations are said to be written in integral form.


The field variables E and
B, which are usually the unknown quantities, appear in the integrands. Only in
a few symmetric cases (see Sections 28-6 and 34-2, for example) can we "factor
them out." In more general problems we cannot do so.
The situation is somewhat analogous to computing the density p of a body if we
know its mass m and volume r. In general these are related by the integral
equation
m = f p dr.

Only if p is a constant over all parts of the volume can we factor it out and write
p = m/r.
To carry out our program it is desirable to recast Maxwell's equations in the
form of equalities that apply at each point in space rather than as integrals that
apply to various regions of space. In other words, we wish to convert Maxwell's·
equations from the integral form of Eqs. V-1 to 4 into differential form. We will
then be able to relate E and Bat a point to the charge density and current density
at that point.

V-2 The Operator V


To transform Maxwell's equations into differential form we must deepen our
un:derstanding of vector methods and, in particular, become familiar with the
vector operator V.
In Section 29-7 we saw how to obtain the components of the (vector) electro-
static field E at any point from the (scalar) potential function V(x,y,z) by partial
differentiation. Thus,

Ex= - - ,
av --,
av and Ez =
av
ax ay az

so that the electrostatic field


E = iEx +jEy + kEz
can be written as

E= - . av- + J
(1 . av- + kav)
- · (V-5)
ax i)y az

We can write Eq. V-5 in compact vector notation as

E = -VV,

where V ("del") is a vector operator defined as

V =
.a + J.a
I -
a
- + k ~• (V-6)
ax ay i)z

This operator is useful in dealing with scalar and vector fields (see Sections 16-8
and 18-7 for examples of such fields). Given any scalar field'+' we may form a
vector field, called the gradient of'+' and written as grad'+' or V'+', simply by apply-
ing the operator V to '+'· Given a vector field U = Uxi Uyj + +
U.k we may
apply the operator V to it in two different ways. One way is to take the dot
product of V and U, yielding the scalar field called the divergence of U and written
SUPPLEMENTARY TOPIC V 3
as div U or V · U. The other way is to take the cross product of V and U, yielding
the vector field called the curl of U and written curl U or V X U. These opera-
tions may be summarized as

grad if;
. af + J. af
= Vif; = 1 -
af
- +k -
ax ay az
div u = v . u = aux + au y+ au z
ax ay az
auz auy)
curl u = v x u = i ( ay - az

+ j (aux - auz)
az ax
auy- aux)
+ I.. ( -ax - ay
Note that grad if; and curl U are vectors, whereas div U is a scalar. The student
can gain some familiarity with these operations by the following exercises: (1)
prove that curl (grad if;) = 0 and (2) prove that div (curl U) = 0.
Another frequently occuring operator is 'V 2 ("del squared"). It is simply
V · V, or, as the student can show from Eq. V-6,
a2 a2 a2
v2 = v . v = -- + ~ + ~·
ax 2 ay 2 az 2
When we apply 'V 2 to a scalar field if;, we obtain

(V-10)

For a vector field U, the operation 'V 2U is defined as

v2u = i a2ux + j a2uy + k a2uz. (V-11)


ax 2 ay 2 az 2
As an exercise the student can show that curl (curl U) - 'V 2U + grad (div U).
V-3 Maxwell's Equations in Differential Form-I

In this section we show how to cast the first two of Maxwell's equations (Eqs.
V-1, 2) into differential form. Let us apply Eq. V-1 to a differential volume ele-
ment shaped like a rectangular parallelepiped and containing a point P at (and
near) which an electric field exists (see Fig. V-la). PointP is located at x, y, z in
the reference frame of Fig. V-lb and the edges of the parallelepiped have lengths
dx, dy, and dz.
We can write the surface area vector for the rear face of the parallelepiped as
dS = - i dy dz. The minus sign enters because dS is defined to point in the
direction of the outward normal, which is defined by -i. For the front face we
have dS = +i dy dz.
4 SUPPLEMENTARY TOPIC V

4=]1dz
~
• P(x,y,z)

O>-------y
dy

(a) x (b)

Fig. V-1

If the electric field at the rear face is E, that at the front face, which is a distance
dx away from the rear face, is E +(
aE/ ax) dx, the latter term representing the
change in E associated with the change dx in x.
The flux through the entire surface of the parallelepiped is .rf E · dS and the
contribution to this flux due to these two faces alone is

(E) · (-i dy dz) + ( E + :! dx) · ( +i dy dz)

= dx dy dz ( -aE · i) = dx dy dz - a (E • i)
ax ax

= dx dy dz-·
aE,,
ax

With similar contributions from the other four faces the total electric flux becomes

...,,r E · dS = dx dy dz (a~
-
ax
+ -a~
ay
+ -a~)
az
·
From Eq. V-8 we may write this as

.rf E · dS = dx dy dz div E. (V-12)

Now the right-hand side of Eq. V-1, which gives the charge enclosed by the
surface, may be written in general as q = Jp dr and, in particular, for the differ-
ential volume element at P, as

q = p dx dy dz, (V-13)

where p is the charge per unit volume at P. Substituting Eqs. V-12 and 13 into
Eq. V-1 and canceling the common factor dx dy dz, we have finally

1:0 div E = p, (V-14)

which is Maxwell's first equation (Eq. V-1) in differential form.


Using the same technique, we can express Maxwell's second equation (Eq.
V-2) in differential form as
div B = 0. (V-15)
SUPPLEMENTARY TOPIC V 5

V-4 Maxwell's Equations in Differential Form-II


We now seek to transform Maxwell's third and fourth equations (Eqs. V-3, 4)
into differential form. We start by applying Eq. V-3 to a differential surface
element of rectangular shape at a point P in some region of a magnetic field, as
shown in Fig. V-2a. The point P is located at x, y, z in the reference frame of Fig.
z

di~ I dy
•P
0,1-------y

x
(a) (b)

Fig. V-2

V-2b and the sides of the rectangle, which is parallel to the x-y plane, have lengths
dx and dy. Going around the path, as shown by the arrows, we have
§ B · di = B · ( -j dy) (rear side)
+ B · ( +i dx) (left side)

+ ( B + ~: dx) · ( +j dy) (front side)

+ (B + ~: dy) · (-idx), (right side)

where B is the magnetic induction at P.


Collecting terms we obtain

§ B · di = dx dy (aB · j - aB ·
ax ay
i)
= dx dy [_i!__
ax
(B · j) - ~ ay (B · i) J
= dx dy (aBy _ aB.,)· (V-16)
ax ay
Now in the right-hand side of Eq. V-3 i is the current enclosed by the path and
d<PE/dt is the change in electric flux through the enclosed surface. Hence, if
J is taken to represent the current density and dS( = k dx dy) the surface area
vector, we can write
i = J · dS = J · (k dx dy) = dx dy J z (V-17)
and
d<PE aE aE
- = - · dS = - · (k dx dy)
dt at at
or
d<PE = aE. dx d . (V-18)
dt at Y
6 SUPPLEMENTARY TOPIC V

Substituting Eqs. V-16, 17, and 18 into Eq. V-3 and canceling the common factor
dx dy, we get
__!__ (aBy _ aBx) = Jz + eo aEz (V-19)
µo ax ay at
We could have proceeded exactly as above if we had started with a rectangle
parallel to the y-z plane or one parallel to the z-x plane. Each rectangle would
have given us a different component of an arbitrarily oriented differential surface
at P. Equation V-19 is obviously the z-component equation corresponding to
Eq. V-3. If we multiply it by k and add to it the two similar vector equations,
which may be obtained by cyclically permuting x, y, z and i, j, k, corresponding
to the x-component and y-component equations, we obtain

curl B = µo(J + eo aE/at), (V-20)

which is the third Maxwell equation in differential form.


Similarly, starting with Eq. V-4, we may show that
curl E = -aB/at, (V-21)

which is the fourth Maxwell equation in differential form.


We have derived four differential equations (see V-22 to V-25 below) from the
four integral equations (V-1 to V-4). It can be shown that the integral equations
can be derived from the differential equations, that is, the two sets of equations
are equivalent

V-5 The Wave Equation


We have now obtained from their integral form the four basic equations of
electromagnetism, Maxwell's equations, in differential form. Corresponding to
the integral equations, Eqs. V-1, 2, 3, and 4 respectively, we have

eo div E = p, (V-22)

div B = 0, (V-23)
curl B = µo(J + eo aE/at), (V-24)

curl E = -aB/at, (V-25)

which are four coupled partial differential equations. They apply at each point
of space in an electromagnetic field.
We will now derive the wave equation for electromagnetic waves in free space.
In free space, the charge density p and the current density J are zero, so that the
Maxwell equations there become

div E = 0,
div B = 0,
curl E = - aB/at,
and curl B = µoeo aE/at.

Let us take the curl of the equation for curl E; we obtain

aB a
curl curl E = - curl - = - - curl B.
at at
SUPPLEMENTARY TOPIC V 7
But, from above, curl B = µ 0t:o( aE/ at), so that

curl curl E = - µot:o -


a2E (V-26)
at 2
From the exercise in V-2, we know that curl curl E = -V 2E + grad div E and
from above that div E = 0. Thus,
curl curl E = -v' 2E. (V-27)
Combining Eqs. V-26 and V-27, we obtain finally

v' 2E = µot:o - ·
a2 E (V-28a)
at 2
The student, proceeding as above, should be able to show that B satisfies the same
equation, or
(V-28b)

Equations V-28 are the equations of electromagnetic wave motion. Being vector
equations, they are equivalent to six scalar equations, one for each component of
E and of B.
There are many solutions of Eqs. V-28, corresponding to different kinds of
electromagnetic waves~plane, spherical, and cylindrical waves being three
examples. Let us consider a solution in which two components of E and two of
B vanish, that is, in which
E,, = Ez = 0 and B,, = By = 0.
Equations V-28 are satsified for these assumptions. For the nonvanishing com-
ponents, EY and Bz, Eqs. V-28 reduce to (see Eq. V-11)
a2Ey a2Ey + a2Ey a2Ey
ax2 + ay2 az2 = µot:o ~ (V-29a)

and
a2Bz a2Bz a2Bz a2Bz
ax2 + -ay2 + az2 = µoEo at2 (V-29b)

If we make the additional assumption that Ev and Bz are functions of x and t


only, the simplified wave equation that results, namely

(V-29c)

is similar to Eq. III-1 for the vibrating string.


A solution to these equations, as the student may verify by substitution, is
Ey = Em sin (kx - wt) (V-30a)
and
Bz = B,,, sin (kx - wt). (V-30b)
We interpret Eqs. V-30 as an electromagnetic wave traveling in the positive
x-direction, as ill Fig. 39-11, with a speed c = w/k. The student can show that
substituting Eq. V-30a into Eq. V-29a (or Eq. V-30b into Eq. V-29b) yields

c = w/k = 1/Veoµo,
which (see Eq. 39-15) gives the speed of electromagnetic waves in free space.
Supplementary Problems

Chapter 26
I. Two free point charges +q and +4q are a distance l apart. A third charge is so
placed that the entire system is in equilibrium. Find the location, magnitude, and
sign of the third charge.
2. If the balls of Fig. 26~7 are conducting, what happens to them after one is dis-
charged? Find the new equilibrium separation.
3. Two identical conducting spheres, having charges of opposite sign, attract each
other with a force of 0.108 nt when separated by 0.5 meter. The spheres are connected
by a conducting wire, which is then removed, and thereafter repel each other with a
force of 0.036 nt. What were the initial charges on the spheres?
4. A particle of charge -q and mass m moves in a circular orbit about a fixed charge
+Q. (a) Show that the "distance cubed ex: period squared" law,
ra = Qq 1'2
1671' 3Eom
is satisfied. Note that the proportionality constant depends on the property (q/m) of
the orbiting particle. (b) What is the corresponding situation when the force is gravi-
tational rather than electrical?
5. An electron is projected with an initial speed of 3.24 X 10 5 meter /sec directly
toward a proton which is essentially at rest. If the electron is initially a very great
distance from the proton, at what distance from the proton is its speed instantaneously
equal to twice its initial value? (Hint: Use the work-energy theorem.)
6. A "dipole" is formed from a rod of length 2a and two charges, +q and -q. Two
such dipoles are oriented as shown in Fig. 26-9, their centers being separated by the

-q
r-2a~e
e +q -q
r-2a~e
e +q

J_
( _- R ------------3"-J
Fig. 26-9

9
10 SUPPLEMENTARY PROBLEMS

distance R. (a) Calculate the force exerted on the left dipole. (b) For R » a, show
that the magnitude of the force exerted on the left dipole is approximately given by
3p2
F=--
21rEoR4

where p = 2qa is the "dipole moment."

Chapter 27
1. A uniform vertical field E is established in the space bet,rnen two large parallel
plates. In this field one suspends a small conducting sphere of mass m from a string of
length l. Find the period of this pendulum when the sphere is given a charge +q if
the lower plate is charged positively; is charged negatively.
2. A charge q = 3.0 X 10- 6 coul is 30 cm from a small dipole along its perpendicular
bisector. The magnitude of the force on the charge is 5.0 X 10- 6 nt. Show on a
diagram (a) the direction of the force on the charge, (b) the direction of the force on the
dipole, and (c) determine the magnitude of the force on the dipole.
3. A thin glass rod is bent into a semicircle of radius R. A charge +Q is uniformly
distributed along the upper half and a charge - Q is uniformly distributed along the
lower half, as shown in Fig. 27-25. Find the electric field E at P, the center of the
semicircle.

+ + + + + + + + + +

Fig. 27-25 Fig. 27-26

4. A "semi-infinite" insulating rod (Fig. 27-26) carries a constant charge per unit
length of A, Show that the electric field at the point P makes an angle of 45° with the
rod. This result is independent of the distance R.
5. Find the frequency of oscillation of an electric dipole, of moment p and rotational
inertia I, for small amplitudes of oscillation about its equilibrium position in a uniform
electric field of strength E.
6. An electric dipole of moment p is placed parallel to an electric field line along the
y-axis in a nonuniform electric field (Fig. 27-27). The magnitude of the field E varies
uniformly along the y-direction as shown. (a) Show that the magnitude of the force
on the dipole is p dE/dy. (b) What is the direction of the force?

-qe e +q
R
p

Fig. 27-27 Fig. 27-28

7. One type of "electric quadrupole" is formed by four charges located at the vertices
of a square of side 2a. Point P lies a distance R from the center of the quadrupole on a
line parallel to two of sides of the square as shown in Fig. 27-28. For R » a, show that
SUPPLEMENTARY PROBLEMS 11
the electric field at P is approximately given by

3(2qa 2)
E=--·
41rEoR 4

(Hint: Treat the quadrupole as two dipoles.)

Chapter 28
I. It is found experimentally that the electric field in a large region of the earth's
atmosphere is directed vertically down. At an altitude of 300 meters the field is 60
nt / coul and at an altitude of 200 meters it is 100 nt / coul. Find the net
amount of charge contained in a cube 100 meters on edge located between 200 and
300 meters altitude. Neglect the curvature of the earth.
2, Suppose that an electric field in some region is found to have a constant direction
but to be decreasing in strength in that direction. What do you conclude about the
charge in the region?
3. Two concentric conducting spherical shells have radii R1 = 0.145 meter and
R2 = 0.207 meter. The inner sphere bears a charge -6.00 X 10- 2 coul. An electron
escapes from the inner sphere with negligible speed. Assuming that the region between
the spheres is a vacuum, compute the speed with which the electron strikes the outer
sphere.
4. (a) Two identical nonconducting spheres have radius r and are fixed with their
centers a distance R > 2r apart. If each sphere has a total charge q uniformly dis-
tributed on its surface, what is the magnitude of the electric force that either sphere
exerts on the other? (b) Suppose instead that the spheres are conductors, the same
total charge on each still being q. Will the electric force that either sphere exerts
on the other in this case be greater than, less than, or equal to the force in case (a)?
Explain.
5. The spherical region a < r < b carries a charge per unit volume of p = A/r,
where A is constant. At the center (r = 0) of the enclosed cavity is a point charge Q.
What should the value of A be so that the electric field in the
region a < r < b has constant magnitude?
6. A solid insulating sphere carries a uniform charge per unit vol-
ume of p. Let r be the vector from the center of the sphere to a
general point P within the sphere. (a) Show that the electric field
at P is given by E = pr /3eo. (b) A spherical "cavity" is removed
from the above sphere, as shown in Fig. 28-25. Using superposition
concepts show that the electric field at '111 points within the cavity Fig. 28-25
is E = pa/3e 0 (uniform field), where a is the vector connecting the
center of the sphere with the center of the cavity. Note that both these results are
independent of the radii of the sphere and the cavity.

Chapter 29
1. What is the charge density on the surface of a conducting sphere of radius 0.15
meter whose potential is 200 volts?
2. Two identical conducting spheres of radius r = 0.15 meter are separated by a
distance a = 10.0 meters. What is the charge on each sphere if the potential of one is
+1500 volts and if the other is -1500 volts?
3. Two conducting spheres, one of radius 6.0 cm and the other of radius 12.0 cm,
each have a charge of 3 X 10- 8 coul and are very far apart. If the spheres are con-
nected by a conducting wire, find (a) the direction of motion and the magnitude of the
charge transferred and (b) the final charge on and potential of each sphere.
12 SUPPLEMENTARY PROBLEMS

4. In the rectangle shown in Fig. 29-29, the sides have lengths 5.0 cm and 15.0 cm,
q1
q1 r
= -5.0 X 10- 6 coul and q2 = +2.0 X 10- 6 coul. (a) What is the electric potential

l
A at corner B? At corner A? (b) How much work is involved
in moving a third charge q3 = +3.0 X 10- 6 coul from B to A
B ~ q along a diagonal of the rectangle? (c) In this process, is
2 external work converted into electrostatic potential energy or
Fig. 29-29 vice versa? Explain.
5. Show that the potential energy of an assembly of four
charges, each of charge q, in the configuration of a triangular pyramid, ao on an edge,
is 6 X (q 2 /ao)/4?reo.
6. Three charges of +o. l coul each are placed on the corner of an equilateral tri-
angle, 1.0 meter on a side. If energy is supplied at the rate of 1.0 kw, how many days
would be required to move one of the charges onto the midpoint of the line joining the
other two?
i. Show, using the fact that an electrostatic field is a conservative field, that one
cannot create an electric field in which all the lines of force are straight parallel lines
whose density (number per unit cross-sectional area) changes in a direction at right
angles to the lines of force.
8. Two line charges are parallel to the z-axis; one, of charge per unit length +>.., is a
distance a to the right of this axis, and the other, of charge per unit length - >.., is a
distance a to the left of this axis (the lines and z-axis being in the same plane). 8ketch
some of the surfaces of equipotential.
9. A particle of (positive) charge Q is assumed to have a fixed position at P. A
second particle of mass m and (negative) charge -q moves at constant speed in a
circle of radius r1, centered at P. Derive an expression for the work W that must be
done by an external agent on the second particle in order to increase the radius of the
circle of motion, centered at P, to r 2 . Express W in terms of quantities chosen from
among m, ri, r2, q, Q, and eo only.
10. A particle of charge Q is kept in a fixed position at a point Panda second particle
of mass m, having the same charge Q, is initially held at rest a distance r1 from P. The
second particle is then released and is repelled from the first one. Determine its
velocity at the instant it is a distance r2 from P. Let Q = 3.1 X 10- 6 coul, m =
2.0 X 10- 5 kg, ri = 9.0 X 10- 4 meter, and r2 = 25 X 10- 4 meter.
11. A particle of mass m, charge q > 0, and initial kinetic energy K is projected
(from "infinity") toward a heavy nucleus of charge Q, assumed to have a fixed position

V (volts)

a h
-5 x (meters)

-12

Fig. 29-30
SUPPLEMENTARY PROBLEMS 13
in our reference frame. (a) If the "aim is perfect," how close to the center of the
nucleus is the particle when it comes instantaneously to rest? (b) With a particular
imperfect aim the particle's closest approach to the nucleus is twice the distance deter-
mined in part (a). Determine the speed of the particle at this closest distance of
approach.
12. The electric potential varies along the x-axis as shown in the graph of Fig. 29-30.
For each of the intervals shown (ignore the behavior at the end points of the intervals)
determine the x-component of the electric field and plot Ex vs. x.
13. A charge per unit length 'A is distributed uniformly along a straight-line segment
of length L. (a) Determine the electrostatic potential (chosen to be zero at infinity) at
a point P a distance y from one end of the charged segment and in line with it (see Fig.
29-31). (b) Use the result of (a) to compute the component of the electric field inten-
sity at P in the y-direction (along the line). (c) Determine the component of the elec-
tric field intensity at P in a direction perpendicular to the straight line.

0 L x

Fig. 29-31 Fig. 29-32

14. On a thin rod of length L lying along the x-axis with one end at the origin (x = 0),
as in Fig. 29-32, there is distributed a charge per unit length given by 'A = kx, where k is
a constant. (a) Taking the _electrostatic potential at infinity to be zero, find Vat the
point Pon the y-axis. (b) Determine the vertical component, Ey, of the electric field
intensity at P from the result of part (a) and also by direct calculation. (c) Why
cannot Ex, the horizontal component of the electric field at P, be found using the result
of part (a)?

Chapter 30

I. Show that the capacitance of two oppositely charged metal spheres of the same
radius, when far apart, is one-half the capacitance of one isolated sphere.
2. Two metallic spheres, radii a and b, are connected by a thin wire. Their separa-
tion is large compared with their dimensions. A charge Q is put onto this system.
(a) How much charge resides on each sphere? (b) Apply the definition of capacitance
to show that the capacitance of this system is C = 41rto (a + b).
3. N identical spherical drops of liquid are charged to the same potential V. One
large drop is formed by combining these. Show that the potential of the large drop is
N,iV.
4. Charges q1, q2, qs are placed on capacitors of capacitance C1, C2, Cs respectively,
arranged in series as shown in Fig. 30-28. Switch S is then closed. What are the final
charges qi', q2', qs' on the capacitors?
5. When switch S is thrown to the left in Fig. 30-29, the plates of the capacitor of
capacitance C1 acquire a potential difference V 0. C 2 and Cs are initially uncharged.
14 SUPPLEMENTARY PROBLEMS

The switch is now thrown to the right. What are the final charges q 1, q 2 , q 3 on the
corresponding capacitors?

Fig. 30-28 ·Fig. 30-29

6. A geiger tube is made of two long concentric metal cylinders with a gas of dielectric
constant K between them. Neglecting end effects, use Gauss' law to calculate the
capacitance of this configuration. The center rod has a radius a, the surrounding tube
a radius b, and the length l » b.
7. A parallel plate capacitor has plates of area 0.12 m 2 and a separation of 1.2 cm.
A battery charges the plates to a potential difference of 120 volts and is then discon-
nected. A dielectric slab of thickness 0.4 cm and dielectric constant 4.8 is then placed
symmetrically between the plates. In terms of to (a) find the capacitance before the slab
is inserted. (b) what is the capacitance with the slab in place? (c) what is the free charge
q before and after the slab is inserted? (d) determine the electric field strength in the
space between the plates and dielectric. (e) what is the electric field strength in the
dielectric? (J) with the slab in place what is the potential difference across the plates?
(g) how much external work is involved in the process of inserting the slab?

Chapter 31
I. A steady beam of alpha particles (q = 2e) traveling with constant kinetic energy
20 Me V carries a current 0.25 X 10- 6 ampere. (a) If the beam is directed perpendicu-
lar to a plane surface, how many alpha particles strike the surface in 3.0 sec? (b) At
any instant, how many alpha particles are there in a given 20-cm length of the beam?
(c) Through what potential difference was it
necessary to accelerate each alpha particle
from rest to bring it to an energy of 20 Me V?
2. List similarities and differences between
the flow of charge and the flow of a fluid;
between the flow of charge and the conduction
of heat.
3. Explain -why the momentum which con-
duction electrons transfer to the ions in a
./
i
metal conductor does not give rise to a result-
ant force on the conductor.
Fig. 31-10 4. A resistor is in the shape of a truncated
right circular cone (Fig. 31-10). The end
radii are a and b, the altitude isl. If the taper is small, we may assume that the current
density is uniform across any cross section. (a) Calculate the resistance of this object.
(b) Show that your answer reduces to p(l/A) for the special case of zero taper (a= b).
5. Conductors A and B, having equal lengths of 40 meters and a cross-sectional area
of 0.10 m 2 , are connected in series. A potential of 60 volts is applied across the terminal
points of the connected wires. The resistances of the wires are 40 and 20 ohms respec-
tively. Determine: (a) the resistivities of the two wires: (b) the magnitude of the elec-
SUPPLEMENTARY PROBLEMS 15
tric field in each wire; (c) the current density in each wire; (d) the potential difference
applied to each conductor.
6. A 1250-watt radiant heater is constructed to operate at 115 volts. (a) What will
be the current in the heater? (b) What is the resistance of the heating coil? (c) How
many kilocalories are created in our hour by the heater?
7. An iron wire (diameter 1 mm, length 10 cm) is placed in an evacuated chamber.
Estimate the equilibrium temperature of the wire if it carries a current of 10 amp.
Assume that all heat transfer is by radiation and that the surface of the wire radiates
according to Eq. 47-2. Take the temperature of the chamber walls to be 27° C. List
any additional assumptions that you use.

Chapter 32

1. Two batteries having the same emf e but different internal resistances ri and r 2
are connected in series to an external resistance R. Find the value of R that makes the
potential difference zero between the terminals of the first battery.
2. The section of circuit AB (see Fig. 32-26) absorbs power P = 50.0 watts and a
current i = 1.0 amp passes through it in the indicated direction. (a) What is the
potential difference between A and B? (b) If the element C does not have internal
resistance, what is its emf? (c) What is its polarity?

c
(&', r)

Fig. 32-26

3. Two light bulbs, one of resistance R and the other of resistance r( < R), are con-
nected (a) in parallel and (b) in series. Which bulb is brighter?
4. N identical batteries of emf e and internal resistance r may be connected all in
series or all in parallel. Show that each arrangement will give the same current in an
external resistor R if R = r.
5. Twelve resistors, each of resistance R ohms, form a cube (see Fig. 32-27). (a)
Find RAs, the resistance of an edge. (b) Find Rec, the equivalent resistance of a face.
(c) Find RAc, the equivalent resistance of the body diagonal.

Fig. 32-27 Fig. 32-28

6. What internal connections (resistors only) can account for the fact that there are
2 ohms resistance between any pair of the terminals on a box with N terminals? (See
Fig. 32-28.)
16 SUPPLEMENTARY PROBLEMS

7. A voltmeter and an ammeter are used to determine two unknown resistances R 1


and R2, one determination by each of the two methods shown in Fig. 32-29. The volt-
meter resistance is 3D7 ohms and the ammeter resistance is 3.62 ohms; in method (a) the
ammeter reads 0.317 amp and the voltmeter reads 28.1 volts, whereas in method (b) the
ammeter reads 0.356 amp and the voltmeter 23.7 volts. Compute R1 and R 2 •

R2

(a) (b)

Fig. 32-29 Fig, 32-30

8. An RC circuit is discharged by closing a switch at time t = 0. The initial poten-


tial difference across the capacitor is 100 volts. If the potential difference has decreased
to 1.0 volt after 10 sec, (a) what will the potential difference be 20 sec after t = O?
(h) What is the time constant of the circuit?
9. A capacitor with capacitance C = 1.0 µf and initial stored energy U o = 0.5 joule
is discharged through a resistance R = 1.0 X 10 6 ohms. (a) What is the initial charge
on the capacitor? (b) What is the current through the resistor when the discharge
starts? (c) Determine Ve, the voltage across the capacitor, and V R, the voltage across
the resistor, as a function of time. (d) Express the rate of joule heating in the resistor
as a function of time.
10. In the circuit of Fig. 32-30, 8 = 1200 volts, C = 6.50 µf, R1 = R2 = R3 =
7.30 X 10 5 ohms. With C completely uncharged, the switch S is suddenly closed
(t = 0). (a) Determine, for t = 0 and t = oo, the currents through each resistor.
(b) Draw qualitatively a graph of the potential drop V 2 across R2 from t = 0 tot = oo.
(c) What are the numerical values of V 2 at t = 0 and t = oo? (d) Give the physical
meaning of "t oo" and state a rough, but significant, numerical lower bound, in
seconds, for "t = oo" in this case.

Chapter 33

U
I. Particles 1, 2, and 3 follow the paths shown in Fig. 33-25 as they pass through the

x x: x

x x
x x x magnetic field there. What can one conclude about
each particle?
2. (a) What speed would a proton need to circle
x x x x the earth at the equator, if the earth's magnetic
1 field is everywhere horizontal there and directed
x x x x x X along longitudinal lines. Take the magnitude of the
x x x earth's magnetic field to be 0.41 X 10- weber /meter
4 2
x x x
at the equator. (b) Draw the velocity and magnetic
x x induction vectors corresponding to this situation.
x 3. An electron is accelerated through a potential
difference of 1000 volts and directed into a region
x x between two parallel plates separated by 0.02 meter
2
with a voltage difference of 100 volts between them.
Fig. 33-25 If the electron enters moving perpendicular to the
electric field between the plates, what magnetic field
is necessary perpendicular to both the electron path and the electric field so that the
electron travels in a straight line?
SUPPLEMENTARY PROBLEMS 11
4. A positive point charge Q travels in a straight line with constant speed through an
evacuated region in which there is a uniform electric field E and a uniform magnetic
field B. (a) If Eis directed vertically up and the charge travels horizontally from north
to south with speed v, determine the least value of the magnitude of B and the cor-
responding direction of B. (b) Explain why Bis not uniquely determined when E and
v alone are given. (c) Suppose the charge is a proton which enters the region after
having been accelerated through a potential difference of 3.10 X 10 5 volts. If E =
1.90 X 10 5 volt/meter, compute the value of Bin part (a). (d) If in part (c) the elec-
tric field E is turned off, determine the radius r of the circle in which the proton now
moves.
5. A certain galvanometer has a resistance of 75.3 ohms; its needle experiences a full-
scale deflection when a current 1.62 X 10- 3 amp passes through its coil. (a) Determine
the value of the auxiliary resistance required to convert the galvanometer into a volt-
meter that reads 1.000 volt at full-scale deflection. How is it to be connected? (b)
Determine the value of the auxiliary resistance required to convert the galvanometer into
an ammeter that reads 0.0500 amp at full-scale deflection. How is it to be connected?
6. In a Hall effect experiment a current of 3.0 amp lengthwise in a conductor 1.0 cm
wide, 4.0 cm long, and 10- 3 cm thick produced a transverse Hall voltage (across the
width) of 1.0 X 10- 5 volt when a magnetic field of 1.5 weber /meter 2 passed perpendicu-
larly through the thin conductor. From these data, find (a) the drift velocity of the
charge carriers and (b) the number of carriers per cubic centimeter. (c) Show on a
diagram the polarity of the Hall voltage with a given current and magnetic field direc-
tion, assuming the charge carriers are (negative) electrons.
7. (a) What is the cyclotron frequency of an electron with an energy of 100 ev in the
earth's magnetic field of 1.0 X 10- 4 webers/meters 2 ? (b) What is the radius of curva-
ture of the path of this electron if its velocity is perpendicular to the magnetic field?

Chapter 34
I. Eight wires cut the page perpendicularly at the points shown in Fig. 34-32. A
wire labeled with the integer k (k = 1, 2, . . . 8) bears the current kio. For those with
odd k, the current flows up out of the page; for
those with even k it flows down into the page.
Evaluate §B · di along the closed path shown in
the direction indicated by the single arrowhead.
2. Two long straight wires pass near one another
at right angles. If the wires are free to move,
describe what happens when currents are sent
through them.
3. Suppose, in Fig. 34-25, that the currents are all
in the same direction. What is the force per meter
(magnitude and direction) on any one wire? In the
analogous case of parallel motion of charged parti- Fig. 34-32
cles in a plasma this is known as the pinch effect.
4. A long straight conductor has a circular cross section of radius R and carries a
current i. Inside the conductor there is a cylindrical hole
of radius a whose axis is parallel to the conductor axis at
a distance b from it. Use superposition ideas, and obtain
an expression for the magnetic induction B inside the hole.
5. A straight conductor is split into identical semicircular
Fig. 34-33 turns as shown in Fig. 34-33. What is the magnetic field
at the center C of the circular loop so-formed?
6. (a) A current i flows in a straight wire of length L in the direction shown in Fig.
34-34a. Starting from the Biot-Savart law, determine the resulting magnetic induc-
18 SUPPLEMENTARY PROBLEMS

tion (Bp, BQ, BR, Bs, respectively-direction and magnitude in each case) at each of the
four points P, Q, R, S (all coplanar with the wire). (b) Using the results of part (a),
compute the magnetic induction B (magnitude and direction) resulting at the point T
?S

L
t I
I
I
I
~a-.-ec---a~
...------.---------<;>
I I

j
I
: R I
--------0 I

1
I
-<---L ~ I
I
I
L 2a

j
L
-L~
:
I
I
--------0
Q j
I

t 6p I ~---2a---~

(a) (b)

Fig. 34-34

from the current flowing as indicated in the six-sided rectilinear closed loop shown in
Fig. 34-34b. (Everything drawn is meant to lie in the same plane and all angles are
90°.)
7. Use the Biot-Savort law to calculate the magnetic induction B at C, the common
center of the semicircular arcs AD and HJ, of radii R2 and R1 respectively, forming

o,:R,
part of the circuit A DJ HA carrying current i, as shown in Fig. 34-35.

I ' I
I ', () I
I ',

A H C J D R1 C

Fig. 34-35 Fig. 34-36

8. Compute the magnetic induction B at C, the common center of the circular arcs
of the circuit carrying current i, the arcs cutting a sector of angle 8, as shown in Fig.
34-36.
9. (a) A long wire is bent into the shape shown in Fig. 34-37, without cross-contact

[}
at P. Determine the magnitude and direction
of B at the center C of the circular portion when
the current i flows as indicated. (b) The circu-
)> )>
i lar part of the wire is rotated without distortion
p about its (dashed) diameter perpendicular to
Fig. 34-37
the straight portion of the wire. The magnetic
moment associated with the circular loop is
now in the direction of the current in the straight part of the wire. Determine Bat
C in this case.

Chapter 35
1. Show that emf has the same dimensions as time rate of change of magnetic flux.
2. In Fig. 35-39, l = 2.0 meters and v = 50 cm/sec. Bis the earth's magnetic field,
directed perpendicularly out of the page and having a magnitude 6.0 X 10- 5 webers/m 2
at that place. The resistance of the circuit ADCB, assumed constant (explain how
this may be achieved approximately), is R = 1.2 X 10-5 ohms. (a) What is the emf
SUPPLEMENTARY PROBLEMS 19
induced in the circuit? (b) What is the electric field strength in the wire AB? (c)
What force does each electron in the wire experience due to the motion of the wire in
the magnetic field? (d) What is the magnitude and direction of the current in the wire?
(e) What force must an external agency exert in order to keep the wire moving with this
constant velocity? (!) Compute the rate at which the external agency is doing work.
(g) Compute the rate at which electrical energy is being converted into heat energy,

• A • • • • • •
t'x

1--. •
v
• • • •
D

l

l__ .

• • • • • •

•B • • • c.
B

Fig. 35-39 Fig. 35-40

3. Figure 35-40 shows two loops of wire having the same axis. The smaller loop is
above the larger one, a distance x which is large compared to the radius R of the larger
loop. Hence, with current i flowing as indicated in the larger loop, the consequent
magnetic induction is nearly constant throughout the plane area 1rr 2 bounded by the
smaller loop. Suppose now that xis not constant but is changing at the constant rate
dx/dt = v. (a) Determine the magnetic flux across the area bounded by the smaller
loop as a function of x. (b) Compute the emf generated in the smaller loop at the
instant when x = NR. (c) Determine the direction of the induced current flowing in
the smaller loop if v > 0.
4. A closed loop of wire consists of a pair of equal semicircles, radius 3. 70 cm, lying
in mutually perpendicular planes. The loop was formed by folding a circular loop
along a diameter until the two halves became perpendicular. A uniform magnetic
field B of magnitude 760 gauss is directed perpendicular to the fold diameter and makes
equal angles (45°) with the planes of the semicircles as shown in Fig. 35-4la. (a) The
magnetic field is reduced at a uniform rate to zero during a time interval 4.50 X 10- 3
sec. Determine the magnitude of the induced emf and the sense of the induced current
in the loop during this interval. (b) How-would the answers change if B is directed as
shown in Fig. 35-41b, perpendicular to the direction first given for it but still per-
pendicular to the "fold-diameter?"

v
B

x
x

(a) (b)

Fig. 35-41 Fig. 35-42

5. A loop of wire of area A is connected to a resistance R. The loop is exposed to a


time varying B field (see Fig. 35-42). (a) Derive an expression for the net charge
20 SUPPLEMENTARY PROBLEMS

transferred through the resistor between t = t 1 and t = t2. Show that your answer is
proportional to the difference <l>B(t2) - <l>B(ti), and is otherwise independent of the
manner in which Bis changing. (b) Suppose the change in flux, <I>s(t 2) - <l>B(t 1) is zero.
Does it then follow that no joule-heating occurred dur-
ing this time interval?
.,.-----,:~ . 6. An electromagnetic "eddy current" brake consists
of a disk of conductivity rr and thickness t rotating
about an axis through its center with a magnetic field
B applied perpendicular to the plane of the disk over a
Fig. 35-43 small area a 2 (see Fig. 35-43). If the area a 2 is at a
distance r from the axis, find an approximate expression
for the torque tending to slow down the disk at the instant its angular velocity equals w.

Chapter 36

1. Two equal solenoids (inductance L) are wired in series. (a) If they are far apart,
show that the equivalent inductance is 2L. (b) If they are "close wound," show that
the equivalent inductance is either zero or 4L depending on the direction of the windings.
(Hint: They form a single solenoid.)
2. In the circuit shown in Fig. 36-8, 8 = 10 volts, R1 = 5.0 ohms, R2 = 10 ohms,
and L = 5.0 henries. For the two separate conditions (I) switch S just closed and
(II) switch S closed for a very long time, calculate (a) the current i1 through R1,
(b) the current L2 through R2, (c) the current i through the switch, (d) the voltage
across R2, (e) the voltage across L, and (J) di2/dt.

_(,,t R1 i2 t L

L
R2
~

Fig. 36-8 Fig. 36-9

3. In Fig. 36-9, 8 = 100 volts, R1 = 10 !J, R2 = 20 !J, R3 = 30 !J and L = 2 h.


Find the value of i 1 and i 2 (a) immediately after Sis closed; (b) a long time later; (c)
immediately after Sis opened again; (d) a long time later.
4. A coaxial cable (inner radius a, outer radius b) is used as a transmission line
between a battery 8 and a resistor R. The cable acts as both a capacitor and an
inductor. If the stored electric and magnetic energies are equal, show that R must be

377
R = - In
21r
(b)
-
a
ohms.

(The quantity 377 ohms is sometimes called the impedance of free space.)
5. Each item (a) coulomb-ohm-meter/weber, (b) volt-second, (c) coulomb-ampere/
farad, (d) kilogram-volt-meter 2 / (henry-ampere) 2 , (e) (henry /farad/·~ is equal to one
of the items in the following list: meter, second, kilogram, dimensionless number,
newton, joule, volt, ohm, watt, coulomb, ampere, weber, henry, farad. Give the
equalities.
SUPPLEMENTARY PROBLEMS 21
Chapter 37
I. A total charge q is distributed uniformly on a dielectric ring of radius r. If the
ring is rotated about an axis perpendicular to its plane and through its center at an
angular speed w, find the magnitude and direction of its resulting magnetic moment.
2. (a) What is the magnetic moment due to the orbital motion of an electron in an
atom when the orbital angular momentum is one quantum unit ( = h = 1.05 X 10- 34
joule-sec). (b) The intrinsic spin magnetic moment of an electron is 0.928 X 10- 23
amp-meter 2 . What is the difference in the magnetic potential energy U between the
states in which the magnetic moment is aligned with and aligned in the opposite direc-
tion to an external magnetic field of 1.2 weber /meter 2 ? (c) What absolute temperature
would be required so that the energy difference in (b) would equal the mean thermal
energy k'l'/2?
3. An electron travels in a circular orbit about a fixed positive point charge in the
presence of a uniform magnetic field B directed normal to the plane of ~0.
its motion. The electric force has precisely N times the magnitude of
the magnetic force on the electron. (a) Determine the two possible
angular speeds of the electron's motion. (b) Evaluate these speeds
numerically if B = 4.27 X 10 3 gauss and N = 100.
4. A simple bar magnet is suspended by a string as shown in Fig. S
37-23. If a uniform magnetic field B directed parallel to the ceiling
is then established, show the resulting orientation of string and magnet.
5. An iron magnet containing iron of relative permeability 5000 N
has a flux path 1.0 meter long in the iron and an air gap 0.01 meter long Fig. 37-23
each with cross-sectional areas of 0.02 meter. 2 What current is neces-
sary in a 500 turn coil wrapped around the iron to give a flux density in the air gap of
1.8 webers/meter 2 ?

Chapter 38

1. Initially the 900 µf capacitor is charged to 100 volts, and the 100 µf capacitor is
uncharged in Fig. 38-13. (a) Describe in detail how one may charge the 100 µf capaci-

100 µfr. 10 h 900:1


tor to 300 volts using 81 and 82 appropriately.
(b) Describe in detail the mass
ical analogy of this problem.
+ spring mechan-

2. A self-consistency property of two of the Max-


well equations. Two adjacent parallelepipeds share
a common face as shown in Fig. 38-14. (a) We
Fig, 38-13 may apply Eo§E · dS = q to each of the two closed
surfaces separately. Show that, from this alone,
it follows that Eo§E · dS = q is automatically satisfied for the composite closed surface.
(b) Repeat using §B · dS = 0.

Fig. 38-14
Ir
J__ c

Fig. 38-1 S
f

3. A self-consistency property of two of the Maxwell equations. Two adjacent closed


paths abcda and efcbe share the common edge be as shown in Fig. 38-15. (a) We may
apply §E · di = -dif.>B/dt to each of these closed paths separately. Show that from
22 SUPPLEMENTARY PROBLEMS
this alone, §E · di = -d<l>B/dt is automatically satisfied for the composite closed path
. 1 dcf>E
abefcda. (b) Repeat usmg - §B · dl = i + Eo-·
µo dt
4. Using the definitions of flux, volume charge density p, and current density j, write
the four Maxwell equations in such a manner that all the fluxes, currents, and charges
appear as volume or surface integrals.
Chapter 39
I. A coaxial cable is made of a center wire of radius a surrounded by a thin metal
tube of radius b. The substance between the conductors is air. (a) Find the capaci-
tance per unit length of this coaxial cable (Hint: Imagine equal but opposite charges to
be on the wire and the tube). (b) Find the inductance per unit length of this coaxial
cable (Hint: Imagine a current 1· flowing down the center wire and back along the tube.)
2. Starting with Eqs. 39-10 and 39-13, show that E satisfies the "wave equation"
a2E a2E
OX2 = µQEO i)t2 •

Does B also satisfy the wave equation?


3. Show that (a) through (d) below satisfy Eqs. 39-10 and 39-13. In each of these,
A is a constant. (a) E = Ac(x - ct), B = A(x - ct). (b) E = Ac(x + ct) 15 , B =
-A(x + ct) 15 • (c) E = Ace<.,-ctl, B = Ae<z-ct>. (d) E = Ac In (x + ct), B =
-A In (x + ct). (e) Generalize these examples to show that E = Acf(x - ct),
B = AJ(x - ct) is a solution where f is any function of (x - ct). What is the corre-
sponding situation for functions of (x + ct)?
4. Consider the possibility of "standing waves":
E = Em (sin wt) (sin kx)
B = Bm(cos wt) (cos kx)
(a) Show that these satisfy Eqs. 39-10 and 39-13 if Em is suitably related to Bm and w
suitably related to k. What are these relationships? (b) Find the (instantaneous)
Poynting vector. (c) Show that the time average power flow across any area is zero.
(d) Describe the flow of energy in this problem.
5. An airplane flying at a distance of 10 km from a radio transmitter receives a signal
of intensity 10.0 microwatts/meter 2 . Calculate (a) the (average) electric field at the air-
plane due to this signal; (b) the (average) magnetic field at the airplane; (c) the total
power radiated by the transmitter, assuming the transmitter to radiate isotropically
and the earth to be a perfect absorber.

Fig. 39-18
SUPPLEMENTARY PROBLEMS 23
a < r < b. (b) Calculate the Poynting vector S for a < r < b. (c) By suitably
integrating the Poynting vector, show that the total power flowing across the annular
cross section a < r <bis 8 2 /R. Is this reasonable? (d) Show that the direction of S
is always from the battery to the resistor, no matter which way the battery is connected.
7. A long hollow cylinder (radius R, length l) carries a uniform charge per unit area
of u on its surface. An externally applied torque causes the cylinder to rotate at con-
stant acceleration w(t) = ed. (a) Find B within the cylinder (treat it as a solenoid).
(b) Find E at the inner surface of the cylinder. (c) Find S at the inner surface of the
cylinder. (d) Show that the flux of S entering the interior volume of the cylinder is

equal to..!!:.. (7CR 2l B 2). (Compare with Problem 18.)


dt 2µo

Chapter 40

I. A plane electromagnetic wave, with wavelength 3.0 meters, travels in free space
in the +x-direction with its electric vector E, of amplitude 300 volts/meter, directed
along the y-axis. (a) What is the frequency v of the wave? (b) What is the direction
and amplitude of the B field associated with the wave? (c) If E = Em sin (kx - wt),
what are the values of k and w for this wave? (d) What is the time averaged rate of
energy flow per unit area associated with this wave? (e) If the wave fell upon a per-
fectly absorbing sheet of area A, what momentum would be delivered to the sheet per
second and what is the radiation pressure exerted on the sheet?
2. Show that eoE X B has the dimensions of momentum/volume. (The Yector
eoE X B may be used to compute the momentum stored in the fields in the same man-
1 1 .
ner that the scalar - e0 E 2 + - B 2 may be used to compute the energy stored m the
2 2µo
fields.)
3. A particle in the solar system is under the combined influence of the sun's gravi-
tational attraction and the radiation force due to the sun's rays. Assume that the
particle is a sphere of density 1.0 gm/cm 3 and that all of the incident light is absorbed.
(a) Show that all particles with radius less than some critical radius, Ro, will be blown
out of the solar system. (b) Calculate Ro. (c) Does Ro depend on the distance from
the earth to the sun? (See the appendices for the necessary constants.)

Chapter 41
I. When an electron moves through a medium at a speed exceeding the speed of
light in that medium, it radiates electromagnetic energy (the Cerenkov effect, see Sec-
tion 20-7). What minimum speed must an electron have in a liquid of refractive index
1.54 in order to radiate?
2. Assume that the index of refraction of the earth's atmosphere varies, with dtitude
only, from the value one at the edge of the atmosphere to some larger value at the
surface of the earth. (a) Neglecting the earth's curvature, show that the apparent
angle of a star from the zenith direction is independent of how the refractive index of
the atmosphere varies with altitude and depends only on the value of n at the earth's
surface. (Hint: Compare a uniform atmosphere with one consisting of layers of
increasing refractive index.) (b) How does the earth's curvature affect the analysis?
3. A point source is 80 cm below the surface of a body of water. Find the diameter
of the largest circle at the surface through which light can emerge from the water.
4. A pole extends 2.0 meters above the bottom of a swimming pool and 0.5 meters
above the water. Sunlight is incident at 45°. What is the length of the shadow of the
pole on the bottom of the pool?
5. A given monochromatic light ray, initially m air, strikes the 90° prism at P (see
Fig. 41-24) and is refracted there and at Q to such an extent that it just grazes the
right-hand prism surface after it emerges into air at Q. (a) Determine the index of
24 SUPPLEMENTARY PROBLEMS

Fig. 41-24
v -

Fig. 41-25
- -
-

refraction, relative to air, of the prism for this wavelength in terms of the angle of
incidence 61 which gives rise to this situation. (b) Give a numerical upper bound for
the index of refraction of the prism. (c) Show, by a ray diagram, what happens if the
angle of incidence at P is slightly greater than 6 1, is slightly less than 61.
6. Two perpendicular mirrors form the sides of a vessel filled with water, as shown in
Fig. 41-25. A light ray is incident from above, normal to the water surface. (a) Show
that the emerging ray is parallel to the incident ray. Assume that there are two reflec-
tions at the mirror surfaces. (b) Repeat the analysis for the case of oblique incidence,
the ray lying in the plane of the figure. (c) Using three mirrors, state and prove the
three-dimensional analog to this problem.

Chapter 42
I. How many images of himself can an observer see in a room whose ceiling and two
adjacent walls are mirrors? Explain.
2. A thin flat plate of glass is a distance b from a convex mirror. A point source of
light Sis placed a distance a in front of the plate (see Fig. 42-33) so that its image in the
partially reflecting plate coincides with its image in the

S• _l ( mirror. If b = 7.5 cm and the focal length of the mirror is


f = -30 cm, find a and draw the ray diagram.
3. (a) A simple magnifier (a converging lens) 1s used

.L. J\ I
directly in front of the eye to examine an object. Draw a
ray diagram showing the final image on the retina of the eye.
(b) The smallest distance at which a normal eye can focus
Fig. 42-33 upon an object is 25 cm. Use this result with your ray dia-
gram to prove that the (angular) magnification of a simple
magnifier is 25 If+ 1, ,vhere f is the focal length of the magnifier lens.
4. Two thin lens, one having f = + 12.0 cm and the other f = -10.0 cm, are sepa-
rated by 7.0 cm. A small object is placed 43.5 cm from the center of the lens system on
the principal axis first on one side and next on the other side. Find the location of the
final image in each case.
5. Show that a thin converging lens of focal length f followed by a thin diverging lens
of focal length -f will bring parallel light to a focus beyond the second lens provided
that the separation of the lenses L satisfies O < L < f. (b) Does
this property change if the lenses are interchanged? (c) What
happens when L = O?
6. A concave and a convex lens are cut out of a plane parallel
block of glass as shown in Fig. 42-34. Discuss the geometry of a
beam of parallel rays incident (a) on the concave lens and (b) on
the convex lens as the distance between lenses is increased from
contact- to a large separation.
7. An object is placed 1.0 meter in front of a convergent lens, of
focal length 0.5 meter, which is 2.0 meters in front of a plane
mirror. (a) Where is the final i~age, measured from the lens, Fig. 42-34
SUPPLEMENTARY PROBLEMS 25
that would be seen by an eye looking toward the mirror through the lens? (b) Is the
final image real or virtual? (c) Is the final image erect or inverted? (d) What is the
lateral magnification?

Chapter 43
1. In a double-slit arrangement the distance between slits is 5.0 mm and the slits are
1.0 meter from the screen. Two interference patterns can be seen on the screen, one
due to light of 4800 A and the other 6000 A. What is the separation on the screen
bet,veen the third-order interference fringes of the two different patterns?
2. In Young's interference experiment in a large ripple tank (see Fig. 43-4) the
coherent vibrating sources are placed 12.0 cm apart. The distance between maxima
2.0 meters away is 18.0 cm. If the speed of ripples is 25.0 cm/sec, find the frequency
of the vibrators.
3. One slit of a double-slit arrangement is covered by a thin glass plate of refractive
index 1.4, and the other by a thin glass plate of refractive index 1. 7. The point on the
screen where the central maximum fell before the glass plates were inserted is now occu-
pied by what had been the fifth bright fringe before. Assume :\ = 4800 A and that the
plates have the same thickness t and find the value of t.
4. In Fig. 43-25, the source emits monochromatic light of wavelength:\. Sis a very
narrow slit in an otherwise opaque screen I. A plane mirror, whose surface includes
the axis of the lens shown, is located a distance h below S. Screen II is at the focal
plane of the lens. (a) Find the condition for maxima and minima brightness of fringes

Region A

*
Source

Region B
I II

Fig. 43-25

on screen II in terms of the usual angle 0, the wavelength:\, and the distance h. (b) Do
fringes appear only in region A (above the axis of the lens), only in region B (below the
axis of the lens), or in both regions A and B? Explain. (Hint: Consider the image of
S formed by the mirror.)
5. Let fi(t) = Ai sin (wt +
<J,1) and f2(t) = A2 sin (wt +
<J,2). Suppose we want to
calculate the time average of their product:

-l
1'
lT Ji
t=O
(t)h (t) dt, 1'
271"
w

(Such a problem might arise if Ji represents a current and ha voltage; fih would then
be the instantaneous power.) Show that this average is equal to one-half the dot
product of the corresponding phasors.
6. Consider the problem of determining the sum

A 1 sin (wt + <J,1) + A 2 sin (wt + <J,2) + · · · + An sin (wt + <J,n)


from the phasor diagram.
26 SUPPLEMENTARY PROBLEMS

(a) Show that the sum may always be written in the form

B sin wt+ C cos wt.


(b) Show that B 2 + C2 ~ (A 1 + A 2 + · · · + An) 2 .
(c) When does the equality sign in (b) hold?
7. A thin film of acetone (refractive index 1.25) is floated on a thick glass plate
(refractive index 1.50). Plane light waves of variable wavelength are incident normal
to the film. When one views the reflected wave it is noted that complete destructive
interference occurs at 6000 A and constructive interference at 7000 A. Calculate the
thickness of the. acetone film.
8. We wish to coat a flat piece of glass (n = 1.50) with a transparent material
(n = 1.25) so that light of wavelength 6000 A (in vacuum) incident normally is not
reflected. How can this be done?
9. Light of wavelength 6300 A is incident normally on a thin ·wedge-shaped film of
index of refraction 1.5. There are ten bright and nine dark fringes over the length of
film. By how much does the film thickness change over this length?
10. An .oil drop (n = 1.20) floats on a water (n = 1.33) surface and is observed from
above by reflected light (see Fig. 43-26). (a) Will the outer (thinnest) regions of the
drop correspond to a bright or a dark region? (b) Approximately how thick is the oil
Observer~

Water

Fig. 43-26

film where one observes the third blue region from the outside of the drop? (c) Why do
the colors gradually disappear as the oil thickness becomes larger?

Chapter 44
I. It can be shown that, except for (} = 0, a circular obstacle produces the same
diffraction pattern as a circular hole of the same diameter. Furthermore, if there are
many such obstacles located randomly, then the interference effects vanish leaving only
the diffraction associated with a single obstacle. (a) Explain why one sees a "ring"
around the moon on a foggy day. (b) Calculate the size of the water droplet in the air
if the bright ring around the moon appears to have a diameter 1.5 times that of the
moon.
2. A monochromatic beam of parallel light is incident on a "collimating" hole of
diameter x » X. Point P lies in the geometrical shadow region on a distance screen, as
shown in Fig. 44-15a. Two obstacles, shown in Fig. 44-15b, are placed in turn over the
collimating hole. A is an opaque circle with a hole in it and B is the "photographic
negative" of A. Using superposition concepts, show that the intensity at Pis identical
for each of the two diffracting objects A and B (Babinet's principle).
3. An astronaut in a satellite claims he can just barely resolve two point sources on
the earth, 100 miles below him. What is their separation, assuming ideal conditions?
Take X = 5500 A, and the pupil diameter to be 5.0 mm.
4. Under ideal conditions, estimate the linear separation of two objects on the planet
Mars which can just be resolved by an observer on earth (a) using the naked eye,
SUPPLEMENTARY PROBLEMS 27
Screen

,+.
I
-x
-t
-
A B
(a) (b)

Fig. 44-15

(b) using the 200-in. Mt. Palomar telescope. Use the following data: distance to
Mars = 50 million miles; diameter of pupil = 5.0 mm; wavelength of light = 5500 A.
5. A double-slit system (slit separation d, slit width a) is driven by two loudspeakers
as shown in Fig. 44-16. By use of a variable delay line, the phase of one of the speakers

Variable
Audio delay
source
line

Fig. 44-16

may be varied. Describe in detail what changes occur in the intensity pattern at large
distances as this phase difference is varied from zero to 21r. Take both the interference
and diffraction effects into account.
Chapter 45
I. A diffraction grating has a large number N of slits, each of width d. Let I max
denote the intensity at some principal maximum, and let h denote the intensity of the
kth adjacent secondary maxima. (a) If k « N, show from the phasor diagram that,
approximately, hi I max = ll(k + {-) 21r 2 . (Compare this with the single-slit formula.)
(b) For those secondary maxima which lie roughly midway between two adjacent
principal maxima, show that roughly hi I max = llN 2 • (c) Consider the central
principal maximum and those adjacent secondary maxima for which k « N. Show
that this part of the diffraction pattern quantitatively resembles that of one single slit
of width Nd.
2. With light from a gaseous discharge tube incident normally on a grating with a
distance 1.732 X 10- 4 cm between adjacent slit centers, a green line appears with sharp
maxima at measured transmission angles e = ± 17.6°, 37.3°, -37.1 °, 65.2°, and -65.0°.
(a) Compute the wavelength of the green line that best fits the data.
3. A narrow beam of monochromatic light strikes a grating at normal incidence and
produces sharp maxima at the follmving angles from the normal: 6°40', 13°30', 20°20',
35°40'. No other maxima appear at any angle between 0° and 35°40'. The separa-
tion between adjacent slit centers in the grating is 5.04 X 10- 4 cm. (a) Compute the
wavelength of the light used. (b) Make the most complete quantitative statement that
can be inferred from the above data concerning the width of each slit.
28 SUPPLEMENTARY PROBLEMS

4. Two spectral lines have wavelengths ).. and ).. + Li).., respectively, where Li).. « )...
Show that their angular separation MJ in a grating spectrometer is given approximately
by LiO = Li)../V (d/rn) 2 - ).. 2 , where d is the separation of adjacent slit centers and rn is
the order at which lines are observed. Notice that the angular separation is greater in
the higher orders.
5. A diffraction grating is made up of slits of width 3000 A with a 9000 A separation
between centers. The grating is illuminated by monochromatic plane waves, ).. =
6000 A, the angle of incidence being zero. (a) How many diffracted lines are there?
(b) What is the angular width of the spectral lines observed, if the grating has 1000 slits?
(c) How is the angular width of the spectral lines related to the resolving power of the
grating?
6. A diffraction grating has a resolving power R = )../Li).. = Nm. (a) Show that the
corresponding frequency range, Liv, that can just be resolved is given by Liv = c/Nm)...
(b) From Fig. 45-1, show that the "times of flight" of the two extreme rays differ by an
amount Lit = Nd sin (} /c. (c) Show that (Liv) (Lit) = 1, this relation being independent
of the various grating parameters.
7. Consider an infinite two-dimensional square lattice as in Fig. 45-14b. One inter-
planar spacing is obviously ao itself. (a) Calculate the next five smaller interplanar
spacings by sketching figures similar to Figure 45-16a. (b) Show that your answers
obey the general formula
d = ao/Vh 2 + k2
where h, k are both relatively prime integers (no common integer factor other than
unity).

Chapter 46
1. A beam of light is a mixture of plane polarized light and randomly polarized light.
When it is sent through a Polaroid sheet, it is found that the transmitted intensity can
be varied by a factor of five depending on the orientation of the Polaroid. Find the
relative intensities of these two components of the incident beam.
2. It is desired to rotate the plane of polarization of a beam of plane polarized light
by 90°. (a) How might this be done using only Polaroid sheets? (b) How many sheets
are required in order that the total intensity loss is less than 5 % ? Assume that each
Polaroid sheet is ideal.
3. A sheet of Polaroid and a quarter-wave plate are glued together in such a way
that, if the combination is placed with face A against a shiny coin, the face of the coin
can be seen when illuminated with light of appropriate wavelength. When the com-
bination is placed with face A away from the coin, the coin cannot be seen. Which
component is on face A and what is the relative orientation of the components?
4. A beam of right circularly polarized light is reflected from a mirror. (a) Is the
reflected beam right or left circularly polarized? (b) Has the direction of the associated
linear momentum of the light changed? (c) Has the direction of the associated angular
momentum of light changed? (d) Describe the reaction "felt" by the mirror.
Definition of Standards and
Fundamental and Derived
Physical Constants*
APPENDIX A

The definition of primary standards is by agreement of the International


Conference of Weights and Measures in October, 1964 in Paris. Measured and
derived values of the fundamental physical constants summarize hundreds of
physical measurements made over the years by scientists in all parts of the
world. They have been subjected to exhaustive statistical analysis and, with
their accompanying error limits (which are given as standard. deviations),
represent the best values as of 1969. For most problems in this book three
significant figures will do, and the "computational" (rounded) values may be
used. The data are those given by Taylor, Parker and Langenberg in their
recent adjustment of the fundamental constants as reported in the Reviews of
J1odern Physics, July, 1969.

* See "A Pilgrim's Progress in Search of the Fundamental Constants," by J. W. M.


Du Mond, Physics Today, October 1965.

29
30 APPENDIX A

DEFINITION OF STANDARDS

Abbre-
Standard viation Equivalent

Meter m 1,650,763.73 wavelengths in vacuo of the un-


perturbed transition 2pi 0-5d 5 in Kr 86
Kilogram kg mass of the international kilogram at Sevres,
France
Second sec 9,192,631,770 vibrations of the unperturbed
hyperfine transition 4,0--3,0 of the funda-
mental state 2S; in Cs 133
Degree Kelvin OK defined in the thermodynamic scale by assign-
ing 273.16 K to the triple point of water
Unified atomic mass unit amu rt the mass of an atom of C 12
Mole mol amount of substance containing the same num-
ber of atoms as 12 gm (exactly) of pure C 12
Standard acceleration of g 9.80665 meter/sec 2
free fall
Normal atmospheric atm 101,325 nt/meter 2
pressure
Thermochemical calorie cal 4.1840 joules
Liter li 0.001,000,028 meter 3
Inch in. 0.0254 meter (exactly)
Pound ( mass ) lb 0.453,592,37 kg
DEFINITION OF STANDARDS 31
FUNDAMENTAL AND DERIVED CONSTANTS

Best ( 1969) experimental


Name Symbol Computational value value*

Speed of light 3.00 X 10s meters/sec 2.9979250( 10)


Permeability constant µo 1.26 X 10- 6 henry/meter 4?T X 10- 7 exactly
Permitivity constant to 8.85 X 10- 12 farad/meter 8.8541853(59)
Elementary charge 1.60 X 10-19 coul 1.6021917(70)
Avogadro's number No 6.02 X 1023/mole 6.022169(40)

Electron rest mass me 9.11 x 10- 31 kg 9.109558(54)


Proton rest mass mp 1.67 x 10- 27 kg 1.672614(11)
Neutron rest mass mn 1.67 x 10- 27 kg 1.674920( 11)
Planck's constant h 6.63 X 10-34 joule sec 6.626196(50)
Electron charge/mass ratio e/me 1. 76 X 1011 coul/kg 1.7588028(54)
Quantum/charge ratio h/e 4.14 X 10- 15 joule sec/coul 4.135708( 14)
Electron Compton wavelength Ae 2.43 X 10- 12 meter 2.4263096(74)
Rydberg constant R,,, 1.10 X 10 7/meter 1.09737 312( 11)
Bohr radius ao 5.29 X 10-11 meter 5.2917715(81)
Bohr magneton µs 9.27 X 10- 24 joule/tesla 9.274096(65)
Nuclear magneton µN 5.05 X 10-27 joule/tesla 5.050951(50)
Proton magnetic moment µp 1. 41 X 10- 26 joule/tesla 1.4106203(90)
Universal gas constant R 8.31 joule/°K mole 8.31434(35)
Standard volume of ideal gas 2.24 X 10-2 meter3/mole 2.24136(30)
Boltzmann's constant k 1.38 X 10- 23 joule/°K 1.380622(59)
Stefan-Boltzmann constant (J 5.67 X 10-8 watt/meter2 °K4 5.66961(96)
Gravitational constant G 6.67 X 10- 11 nt meter 2/kg 2 6.6732(31)

* Same units and power often as the computational value. The numbers in parentheses are the standard
deviation uncertainties in the last digits of the quoted value.
Miscellaneous Terrestrial Data
APPENDIX B

Standard atmosphere 1.013 X 10 5 nt/meter 2


14.70 lb/in 2
760.0 mm-Hg
Density of dry air at STPa 1.293 kg/meter 3
2.458 X 10- 3 slug/ft 3
Speed of sound in dry air at STP 331.4 meters/sec
1089 ft/sec
742.5 miles/hr
Acceleration of gravity, g (standard value)b 9.80665 meters/sec 2
32.1740 ft/sec 2
Solar constant" 1340 watts/m 2
1.92 cal/cm 2-min
Mean total solar radiation 3.92 X 10 26 watts
Equatorial radius of earth 6.378 X 10 6 meters
3963 miles
Polar radius of earth 6.357 X 10 6 meters
3950 miles
Volume of earth 1.087 X 10 21 meter 3
3.838 x 10 22 ft 3
Radius of sphere having same volume 6.371 X 106 meters
3959 miles
2.090 X 107 ft
32
MISCELLANEOUS TERRESTRIAL DATA 33
Mean density of earth 5522 kg/meter 3
Mass of earth 5.983 x 10 24 kg
Mean orbital speed of earth 29,770 meters/sec
18.50 miles/sec
Mean angular speed of rotation of earth 7.29 X 10-5 radians/sec
Earth's magnetic field, B (at Washington, D. C.) 5. 7 X 10-5 teslad
Earth's magnetic dipole moment 8.1 X 10 22 amp-m 2

a STP = standard temperature and pressure = 0° C and 1 atm.


b This value, used for barometer corrections, legal weights, etc., was adopted by
the International Committee on Weights and Measures in 1901. It approximates
45° latitude at sea level.
• The solar constant is the solar energy falling per unit time at normal incidence
on unit area of the earth's surface.
d Tesla = weber/meter 2 •
The Solar System 0

APPENDIX C

Planet Mercury !".l Venus \;1 Earth EB, e, o Mars d" Jupier '21 Saturn 1;, Uranus ©, 1t£ Neptune tJJ Pluto E
----
Mean diameter
km 5,000 12,400 12,742 6,870 139,760 115,100 51,000 50,000 12,700?
Earth diameters 0.39 0.973 1.000 0.532 10.97 9.03 4.00 3.90 0.46
(.,)
.c,. Volume (earth
volumes) 0.06 0.92 1.00 0.15 1,318 736 64 39 0.10
Mass (earth
0.04 0.82 1.00 0.11 318.3 95.3 14.7 17.3
masses)
Density (earth
---~-
I I;!
densities) 0.69 0.89 1.00 0.70 0.24 0.13 0.23 0.29
Mean density
gm/cm 3 3.8 4.86 5.52 3.96 1.33 0.71 1.26 1. 6 ?
Surface gravity
(earth's) 0.27 0.86 1.00 0.37 2.64 1.17 0.92 1.44 ?
Velocity of
escape, km/sec 3.6 10.2 11.2 5.0 60 36 21 23 11?
Length of day
(earth days) 58.6d 30d? ld 1d37m23• 9h55m 10h38m 10.7h 15.8h ?
Planet Mercury 1;! Venus 9 Earth, EB, e, o Mars d" Jupiter '2l Saturn I? Uranus e , )!{ Neptune IV Pluto E

Period, days
87.97 224.70 365.26 686.98 4,332.59 10,759.20 30,685.93 60,187 .64 90,885
Inclination of
equator to orbit - oo? 23°27' 25°12' 3°7' 26°45' 98.0° 29° ?
Oblateness 0.00 0.00 >'l
1/296 1/192 1/15.4 1/9.5 1/14 1/45 ? :r:t'1
Atmosphere,
[f1
main con-
stituents none N2, C02, A N2, 02 N2, C02, H20 CH4, NHa CH4, NHa CH4, NHa CH4, NHa none ~ i,:j
Maximum sur- [f1
~
face tempera- [f1

ture, °K 700 700 350 320 153 138 110? 90? 80? ~
~
Distance from
Sun, 10 6 km 58 108 149 228 778 1426 2869 4495 5900
The Sun 0 329,390 earth masses, mean density 1.42, mean diameter 1,390,600 km, surface gravity 28 (earth's).
The Moon :] 0.01228 earth masses, mean density 3.36, mean diameter 3,476 km, surface gravity 0.17 (earth's), distance from earth
38 X 10 4 km.
a Adapted from Payne-Gaposchkin and Handbook of Chemistry and Physics.

w
U'I
Periodic Table of the Elements
APPE NDIX D
to have a mass of (exactly) 12 amu. For
Atomic weights are expressed in atomic mass units (amu), one atom of the isotope C being defined
12

unstable elements the mass number of the most stable or best known isotope is given in brackets.
Group-+ v VI VII VIII 0
I II III IV
Period Series 2 He
lH 4.0026
1 1 1.00797
7N 80 9F lONe
3 Li 4 Be 5B 6C 20.183
2 2 9.0122 10.811 12.01115 14.0067 15.9994 18.9984
6.939 18 A
11 Na 12 Mg 13 Al 14 Si 15 P 16 S 17 Cl
3 3 32.064 35.453 39.948
22.9898 24.312 26.9815 28.086 30.9738
23 V 24 Cr 25 Mn 26 Fe 27 Co 28 Ni
c., 19K 20 Ca 21 Sc 22 Ti 58.9332 58.71
c,,.
4 40.08 44.956 47.90 50.942 51.996 54.9380 55.847
39.102 36 Kr
4 30 Zn 31 Ga 32 Ge 33 As 34 Se 35 Br
29 Cu 79.909 83.80
5 63.54 65.37 69.72 72.59 74.9216 78.96
41 Nb 42 Mo 43 Tc 44 Ru 45 Rh 46Pd
37 Rb 38 Sr 39 Y 40 Zr 106.4
6 87.62 88.905 91.22 92.906 95.94 [99] 101.07 102.905
85.47 54 Xe
5 49 In 50 Sn 51 Sb --52T e 53 I
47 Ag 48 Cd 131.30
7 107.870 112.40 114.82 118.69 121.75 127.60 126.9044
- - 55 Cs 56 Ba 57-71 72 Hf 73 Ta 74W 75 Re 76 Os 77 Ir 78 Pt
180.948 183.85 186.2 190.2 192.2 195.09
8 132.905 137.34 e
Lanthanid 178.49
6 series•
84Po 85At 86 Rn
79Au 80 Hg 81 Tl 82 Pb 83 Bi [222]
9 196.967 200.59 204.37 207.19 208.980 (210] (210]
- -- --
87 Fr 88 Ra 89
7 10 (223] [226] Actinide
seriesb
63 Eu I 64 Gd I 65 Tb I 66
60 Nd I 61 Pm I 62 Sm I 151.96
57 La I 58 Ce I 59 Pr I 144.24 Dy I 67 Ho j 68 Er I 69 Tm I 70 Yb I 174,97
• Lanthanid · . I 138.91 157.25 158.924 162.50 164.930 167.26 168.934 173.04
71 Lu

Fm 1101 Md I 102 I 103


150.35
Bk I 98 Cf I 99 Es 1100
(145}
Pu I 95 Am I 96 Cm I 97
e series. 140.12 140.907
92 U I 93 Nf I 94
90 Th I 91 Pa I 238.04
I 89[227}Ac I232.038 Lw 110411051
b Actinide series· [242} [243) [247} [247) [251) [254) [257} [256) [254) [257]
. [231} [237
The Particles of Physics 0

APPENDIX E

I Averaged Typical
Family Particle S . Strange- No. of
Symbol Mass pm nessb Charge Antiparticle 0 lifetime, decay
name name particles
seconds products

Photon , (gamma ray) 0 1 0 0 Same 1 Infinite -


c., particle
-- -
""' Electron Electron e- 1 1
2
- -e e+ 2 Infinite -
family
-
Electron's Pe 0 1
2 - 0 lie 2 Infinite -

neutrino
---- -
Muon Muon µ- 206.77 1 - -e µ+ 2 2.212 X 10- 6 e- +Pe+ Pµ
2
family
1 -
-
Muon's I'µ O(?) 2 0 I'µ 2 Infinite -

neutrino
Mesons Pion 11"+ 273.2 0 0 +e 11"- 3 2.55 X 10-s µ+ + v,.
11"0 264.2 0 0 0 11"0 1.9 x 10- 16 'Y + 'Y

Kaon K+ !)66.6 0 +1 +e K+ 4 1.22 X 10-s 11"+ +1ro


Ko 974 0 +1 0 Ko 1.00 x 10- 10 11"+ +1r-
and 6 X 10-s
(,.)
OI)

Family I Particle Strange- Average Typical


Symbol Mass No. of
Spin Charge Antiparticle 0 lifetime decay
name name nessh particles
seconds products

Baryons Nucleon p+ (proton) 1836.12 1


2 0 +e p+ 4 Infinite
n° (neutron) 1838.65 1 0 0 no 1013 p + e- +
-
Ve
2

Lambda Ao 2182.8 1
2 -1 0 Ao 2 2.51 x 10- 10 p + 7r- ;,..
particle "d
"d
t?'J
z
Sigma par- ~+ 2327.7 2
1 -1 +e ~+
-
8. 1 x
10- 11 n + 7r+ s
ticle ~- 2340.5 2
1
-1 -e ~- 6 1.6 x
10- 10 n +1r- ~
l'."l
~o 2332 1 -1 0 ~o about 10- 20 Ao+'
2

Xi particle :g- 2580 1 -2 -e


-z- 1.3 x 10- 10
2 4 Ao + 7r-
zo 2570 1 -2 zo
2
I 0 about 10- 10 Ao + 1ro

• Adapted and modified from The World of Elementary Particles, by Kenneth W. Ford.
b This is a "quantum number" whose assignment permits an understanding of the inter-relationships of the particles.
O Antiparticles have the same mass
and spin as the particles but their charges and strangeness numbers are opposite in sign.
d The K 0 meson has two different lifetimes: all other particles have only one.
Symbols, Dimensions, and Units
for Physical Quantities
APPENDIX F

All units and dimensions are in the mksq (rationalized) system. The primary
units can be found by reading kilograms for Jl!I, meters for L, seconds for T, and
coulombs for Q. The symbols are those used in the text.
In practice, Q is defined in terms of Af, L, and T. However, the addition of Q to
the traditional iVl, L, and T enables us to avoid the use of fractional exponents in
dimensional considerations. The term 'rationalized' simply means that a factor
1/41r is separated out of Coulomb's law in order to remove the factor 41r that would
otherwise appear in many other formulas in electricity.

Quantity Symbol Dimensions Derived Units

Acceleration a LT- 2 meters/sec 2


Angular acceleration a r-2 radians/ sec 2
Angular displacement (J ~

radian
Angular frequency and speed w r-1 radians/ sec
Angular momentum L llJL2T- 1 kg-m 2 /sec .
Angular velocity (I) r-1 radians/sec
Area A,S £2 meter 2
Displacement r, d L meter

39
40 APPENDIX F

Quantity Symbol Dimensions Derived Units


Energy, total E ML2T- 2 joule
kinetic K ML2r-2 joule
potential u ML 2T- 2 joule
Force F MLT- 2 newton
Frequency v r-1 hertz = cycles/sec
Gravitational field g LT- 2 nt/kg
Gravitational potential v £2r-2 joules/kg
Length l L meter
Mass m M kilogram
Mass density p ML- 3 kg/m 3
Momentum p MLT- 1 kg-m/sec
Period T T second
Power p ML2T- 3 watt
Pressure p ML- 1T- 2 nt/m 2
Rotational inertia I ML2 kg-m 2
Time t T second
Torque T ML2r-2 nt-m
Velocity v LT- 1 meters/sec
Volume v £3 meter 3
Wavelength ">-. L meter
Work w ML2T- 2 joule
Entropy s ML2T- 2 joules/K 0
Internal energy u ML 2T- 2 joule
Heat Q ML2T- 2 joule
Temperature T - Kelvin degree
Capacitance c M-1L-2T2Q2 farad
Charge q Q coulomb
Conductivity u M-1L-sTQ2 (ohm-meter )- 1
Current i r-1Q ampere
Current density j L-2r-1Q amp/meter 2
Electric dipole moment p LQ coul-meter
Electric displacement D M-2Q coul/meter 2
Electric polarization p M-2Q coul-meter 2
Electric field E MLT- 2Q- 1 volts/meter
Electric flux <l>E MLar-2Q-1 volt-meter
Electric potential v ML2T-2Q-1 volt
Electromotive force e ML2T-2Q-1 volt
Inductance L ML2Q-2 henry
Magnetic dipole moment µ. L2r-1Q amp-meter 2
Magnetic field H LT- 1Q amp-meter
Magnetic flux <l>B ML2r-1Q-1 weber = volt-sec
Magnetic induction B MT-tQ-1 tesla = webers/meter 2
Magnetization M L-1r-1Q amp/meter
Permeability µ MLQ- 2 henrys/meter
Permittivity E M-1L-aT2Q2 farads/meter
Resistance R ML2T-1Q-2 ohm
Resistivity p ML3T-1Q-2 ohm-meter
Voltage v ML2T-2Q-1 volt
Conversion Factors 0

APPENDIX G

Conversion factors for common and not-so-common units may be read off
directly from the tables below. For example, 1 degree = 2.778 X 10- 3 revolu-
tions, so 16.7° = 16.7 X 2.778 X 10- 3 rev. The mks quantities are capitalized
in each table.
PLANE ANGLE

0 I II
RADIAN rev

1 degree = 1 60 3600 1.745 X 10- 2 2.778 X 10- 3


1 minute = 1.667 X 10- 2 1 60 2.909 X 10- 4 4.630 X 10- 5
1 second = 2.778 X 10- 4 1.667 X 10- 2 1 4.848 X 10- 6 7.716 X 10- 7
1 RADIAN= 57.30 3438 2.063 X 10 5 1 0.1592
1 revolution = 360 2.16 X 10 4 1.296 X 106 6.283 1

1 rev = 2,r radians = 360° 1° = 60' = 3600"

SOLID ANGLE

1 sphere = 41r steradians = 12.57 steradians

a Adapted in part from G. Shortley and D. Williams, Elements of Physics, Prentice-


Hall, Englewood Cliffs, New Jersey, Second Edition, 1955.

41
APPENDIX G
42
LENGTH

cm METER km in. ft mile

1 centimeter = 1 10-2 10-5 0.3937 3.281 6.214


X 10- 2 x 10- 5
1 METER= 100 1 10-3 39.37 3.281 6.214
X 10- 4
1 kilometer = 105 1000 1 3.937 3281 0.6214
X 10 4
1 inch = 2.540 2.540 2.540 1 8.333 1.578
x 10- 2 X 10- 5 X 10- 2 X 10- 5
1 foot = 30.48 0.3048 3.048 12 1 1.894
X 10- 4 X 10- 4
1 statute mile = 1.609 1609 1.609 6.336 5280 1
X 10 5 X 10 4

1 angstrom (A) = 10- 10 meter 1 light-year = 9.4600 X 1012 km 1 yard = 3 ft


1 X-unit = 10- 13 meter 1 parsec = 3.084 X 10 13 km 1 rod = 16.5 ft
1 micron = 10- 6 meter 1 fathom = 6 ft 1 mil = 10- 3 in.
1 millimicron (mµ) = 10- 9 meter
1 nautical mile = 1852 meters = 1.1508 statute miles = 6076.10 ft

AREA

METER 2 cm 2 ft 2 in. 2 circ mil

-
1 SQUARE METER = 1 10 4 10.76 1550 1.974 X 10 9
1 square centimeter = 10-4 1 1.076 X 10- 3 0.1550 1.974 X 10 5
1 square foot = 9.290 X 10- 2 929.0 1 144 1.833 X 10 8
1 square inch = 6.452 X 10- 4 6.452 6.944 X 10- 3 1 1.273 X 10 6
1 circular mill = 5.067 X 10- 10 5.067 X 10- 6 5.454 X 10- 9 7.854 X 10- 1 1

1 square mile = 27,878,400 ft 2 = 640 acres 1 acre = 43,560 ft 2


1 barn = 10- 2 s meter 2

VOLUME

METER 3 cm 3 1 ft 3 in. 3

1 CUBIC METER = 1 10 6 1000 35.31 6.102 X 10 4


1 cubic centimeter = 10-6 1 1.000 X 10- 3 3.531 X 10- 5 6.102 X 10- 2
1 liter = 1.000 X 10- 3 1000 1 3.531 X 10- 2 61.02
1 cubic foot = 2.832 X 10- 2 2.832 X 10 4 28.32 1 1728
1 cubic inch = 1.639 X 10- 5 16.39 1.639 X 10- 2 5.787 X 10- 4 1

1 U. S. fluid gallon = 4 U. S. fluid quarts = 8 U. S. pints = 128 U. S. fluid ounces = 231 in. 3
1 British imperial gallon = the volume of 10 lb of water at 62° F = 277.42 in. 3
1 liter = the volume of 1 kg of water at its maximum density = 1000.028 cm 3
CONVERSION FACTORS 43
MASS

Note: Those quantities to the right of and below the heavy lines are not mass units at all but are often
used as such. When we write, for example,

1 kg "= " 2.205 lb

this means that a kilogram is a mass that weiohs 2.205 pounds. Clearly this "equivalence" is approximate
(depending on the value of o) and is meaningful only for terrestrial measurements. Thus, care must be
employed when using the factors in the shaded portion of the table.

gm KG slug amu

1 gram = 0.001 6.852 6.024


X 10- 5 X 10 23
1 KILOGRAM 1000 6.852 6.024
x 10- 2 x 1026
1 slug = 14.59 8.789
x 1027
1 amu =

DENSITY
Note: Those quantities to the right or below the heavy line are weight densities and, as such, are dimen-
sionally different from mass densities. Care must be used. (See note for mass table.)

slug/ft 3 KG/METER 3 gm/cm 3

1 slug per ft 3 = 515.4 0.5154


1 KILOGRAM per
METER 3 =

1728

TIME

yr day hr min SEC

1 year = 1 36,5.2 8.766 X 10 3 5.259 X 10 5 3.156 X 10 7


1 day= 2.738 X 10- 3 1 24 1440 8.640 X 10 4
1 hour = 1.141 X 10- 4 4.167 X 10- 2 1 60 3600
1 minute= 1.901 X 10- 6 6.944 X 10- 4 1.667 X 10- 2 1 60
lSECOND = 3.169 X 10-s 1.157 X 10- 5 2.778 X 10- 4 1.667 X 10- 2 1
44 APPENDIX G

SPEED

ft/sec km/hr METER/SEC miles/hr cm/sec knot

1 foot per sec-


ond = 1 1.097 0.3048 0.6818 30.48 0.5925
1 kilometer per
hour= 0.9113 1 0.2778 0.6214 27.78 0.5400
1 METER per
SECOND= 3.281 3.6 1 2.237 100 1.944
1 mile per hour = 1.467 1.609 0.4470 1 44.70 0.8689
1 centimeter per
second = 3.281 X 10- 2 3.6 X 10- 2 0.01 2.237 X 10- 2 1 1.944 X 10- 2
1 knot = 1.688 1.852 0.5144 1.151 51.44 1

1 knot = 1 nautical mile/hr 1 mile/min = 88 ft/sec = 60 miles/hr

FORCE
Note: Those quantities to the right of and below the heavy lines are not force units at all but are often
used as such, especially in chemistry. For instance, if we write

1 gram-force "= " 980. 7 dynes,

we mean that a gram-mass experiences a force of 980.7 dynes in the earth's gravitational field. Thus, care
must be employed when using the factors in the shaded portion of the table .

. dyne NT lb pd!

1 dyne = 10-5 2.248 7.233


X 10- 6 X 10- 5
1 NEWTON= 0.2248 7.233
1 pound = 4.448 1 32.17

1 poundal = 0.1383

1 kgf = 9.80665 nt 1 lb = 32.17398 pd]

PRESSURE

inch of NT/
atm dyne/cm 2 cm Hg lb/in. 2 lb/ft 2
water METER 2

1 atmosphere = 1 1.013 406.8 .76 1.013 14.70 2116


X 10 6 X 10 5
1 dyne per cm 2 = 9.869 1 4.015 7.501 0.1 1.450 2.089
X 10- 7 X 10- 4 X 10- 5 x 10- 5 X 10- 3
1 inch of water at 2.458 2491 1 0.1868 249.1 3.613 5.202
40 ca= X 10- 3 X 10- 2
1 centimeter of mer- 1.316 1.333 5.353 1 1333 0.1934 27.85
cury at 0° C a = X 10- 2 X 10 4
1 NEWTON per 9.869 10 4.015 7.501 1 1.450 2.089
METER 2 = X 10- 6 X 10- 3 X 10- 4 x 10- 4 X 10- 2
1 pound per in.2 = 6.805 6.895 27.68 5.171 6.895 1 144
X 10- 2 x 104 X 10 3
1 pound per ft2 = 4.725 478.8 0.1922 3.591 47.88 6.944 1
X 10- 4 X 10- 2 X 10- 3

• Where the acceleration of gravity has the standard value 9.80665 meters/sec 2 •

1 bar = 10 6 dyne/cm 2 1 millibar = 10 3 dyne/cm 2


ENERGY, WORK, HEAT
The electron volt (ev) is the kinetic energy an electron gains from being accelerated through the potential difference of one volt in an electric field. The Mev is the kinetic
energy it gains from being accelerated through a million-volt potential difference.
The last two items in this table are not properly energy units but are included for convenience. They arise from the relativistic mass-energy equivalence formula E = mc 2
and represent the energy released if a kilogram or atomic mass unit (amu) is destroyed completely.
Again, care should be used when employing this table.

Btu erg ft-lb hp-hr JOULES cal kw-hr ev I Mev

--
1 British thermal unit = 1 I 1.055 777.9 3.929 1055 252.0 2.930 6.585 6.585
x 1015
x 1010 X 10- 4 X 10- 4 X 10 21
10-1 6.242 6.242
1 erg = I 9.481 1 7.376 3.725
x 10-14
2.389 2.778
x 10-14 0
X 10- 11 x 10-8 X 10- 8 X 10 11 X 10 5 0
1.285 1.356 1 5.051 1.356 0.3239 3.766 8.464 8.464
z
-"'1
1 foot-pound =
X 10- 3 X 10 7 X 10- 7 X 10- 7 x 1018 x 1012 t,j
i:i:i
[fl
1 horsepower-hour = 2545 2.685 1.980 1 2.685 6.414 0.7457 1.676 1.676 H
x 1013 X 10 6 X 10 6 X 10 5 x 1025 x 1019 0
z
1 JOULE= 9.481 101 0.7376 3.725 1 0.2389 2.778 6.242 6.242 >zj
X 10- 4 X 10- 7 X 10- 7 x 1018 x 1012 >
0
1 calorie = 3.968 4.186 3.087 1.559 4.186 1 1.163 2.613 2.613 8
X 10- 3 X 10 7 X 10- 6 X 10- 6 x 1019 x 1013 0
i:i:i
[fl
1 kilowatt-hour = I 3413 3.6 2.655 1.341 3.6 8.601 1 2.247 2.270
x 1013 X 10 6 X 10 6 X 10 5 x 1025 x 1019
10-6
1 electron volt = I 1.519 1.602 1.182 5.967 1.602 3.827 4.450 1
x 10-22 X 10- 12 x 10-19 x 10-26 x 10-19 X 10- 20 x 10-26
-
1 million electron volts = I 1.519 I 1.602 1.182 5.967 1.602 3.827 4.450 106

1 m-kgf = 9.807 joules 1 watt-sec 1 joule 1 nt-m 1 cm-dyne 1 erg


.I:,.
U'I
46 APPENDIX G

POWER

Btu/hr ft-lb/min ft-lb/sec hp cal/sec kw WATTS

1 British thermal 1 12.97 0.2161 3.929 7.000 2.930 0.2930


unit per hour = x 10- 4 x 10- 2 x 10- 4
1 foot-pound per 7.713 1 1.667 3.030 5.399 2.260 2.260
minute= X 10- 2 X 10- 2 X 10- 5 X 10- 3 X 10- 5 x 10- 2
1 foot-pound per 4.628 60 1 1.818 0.3239 1.356 1.356
second = X 10- 3 X 10- 3
1 horsepower = 2545 3.3 X 10 4 550 1 178.2 0.7457 745.7
1 calorie per 14.29 1.852 3.087 5.613 1 4.186 4.186
second = X 10 2 X 10- 3 X 10- 3
1 kilowatt = 3413 4.425 737.6 1.341 238.9 1 1000
x 10 4
l WATT= 3.413 44.25 0.7376 1.341 0.2389 0.001 1
x 10- 3

ELECTRIC CHARGE

abcoul amp-hr COUL faraday statcoul

1 abcoulomb (1
emu) = 1 2.778 X 10- 3 10 1.036 X 10- 4 2.998 X 10 10
1 ampere-hour = 360 1 3600 3.730 X 10- 2 1.079 X 10 13
1 COULOMB= 0.1 2.778 X 10- 4 1 1.036 X 10- 5 2.998 X 10 9
1 faraday = 9652 26.81 9.652 X 10 4 1 2.893 X 10 14
1 statcoulomb (1
esu) = 3.336 X 10- 11 9.266 X 10- 14 3.336 X 10- 10 3.456 X 10- 15 1

1 electronic charge = 1.602 X 10- 19 coulomb

ELECTRIC CURRENT

abamp AMP statamp

1 abampere (1 emu) = 1 10 2.998 X 10 10


1 AMPERE= 0.1 1 2.998 X 10 9
1 statampere (1 esu) = 3.336 X 10- 11 3.336 X 10- 10 1

ELECTRIC POTENTIAL, ELECTROMOTIVE FORCE

abv VOLTS statv

1 abvolt (1 emu) = 1 10-s 3.336 X 10- 11


1 VOLT= 10s 1 3.336 X 10- 3
1 statvolt (1 esu) = 2.998 X 10 10 299.8 1

ELECTRIC RESISTANCE

abohm OHMS statohm

1 abohm (1 emu) = 1 10-9 1.113 X 10- 21


1 OHM= 10 9 1 1.113 X 10- 12
1 statohm (1 esu) = 8.987 X 10 20 8.987 X 10 11 1
CONVERSION FACTORS 47
ELECTRIC RESISTIVITY
I
ohm-circ
abohm-cm µohm-cm ohm-cm statohm-cm OHM-M
mil/ft

1 abohm-centimeter 1 0.001 10-9 1.113 10-11 6.015


(1 emu) = x 10-21 X 10- 3
1 micro-ohm- 1000 1 10-6 1.113 10-8 6.015
centimeter = x 10-18
1 ohm-centimeter = 10 9 10 6 1 1.113 0.01 6.015
x 10-12 X 10 6
1 statohm-centimeter 8.987 8.987 8.987 1 8.987 5.406
(1 esu) = x 1020 x 1017 X 10 11 x 10 9 x 1018
1 OHM-METER = 1011 10s 100 1.113 1 6.015
x 10-10 X 10 8
1 ohm-circular mil 166.2 0.1662 1.662 1.850 1.662 1
per foot = X 10- 7 x 10-19 X 10- 9

CAPACITANCE

abf FARADS µf" statf

1 abfarad ( 1 emu) = 1 109 10.15 8.987 X 10 20


1 FARAD= 10-9 1 106 8.987 X 10 11
1 microfarad = 10-15 10-6 1 8.987 X 10 5
1 statfarad (1 esu) = 1.113 X 10- 21 1.113 X 10- 12 1.113 X 10- 6 1

a This unit is frequently abbreviated as mf.

INDUCTANCE
I
ab henry HENRYS µh mh stathenry

1 abhenry (1 emu) = 1 10-9 0.001 10-6 1.113 X 10- 21


1 HENRY= 109 1 10 6 1000 1.113 X 10- 12
1 microhenry = 1000 10-6 1 0.001 1.113 X 10- 18
1 millihenry = 106 0.001 1000 1 1.113 x 10-15
1 stathenry (1 esu) = 8.987 X 10 20 8.987 X 10 11 8.987 X 10 17 8.987 X 10 14 1

MAGNETIC FLUX

maxwell kiloline WEBER

1 maxwell (1 line or 1 emu) = 1 0.001 10-8


1 kiloline = 1000 1 10-5
1 WEBER= 10 8 105 1
I
1 esu = 299.8 webers
48 APPENDIX G

MAGNETIC INDUCTION B

WEBER/
gauss kiloline/in. 2 METER 2 milligauss
= TESLA

1 gauss (line per cm 2 ) = 1 6.452 X 10-3 10-4 1000 105


1 kiloline per in. 2 = 155.0 1 1.550 X 10- 2 1.550 X 105 1.550 X 101
1 WEBER per METER 2 =
1 TESLA= 104 64.52 1 10 7 10 9
1 milligauss = 0.001 6.452 X 10- 6 10-7 1 100
1 gamma= 10-5 6.452 X 10-s 10-9 0.01 1
I

1 esu = 2.998 X 10 6 webers/meter 2

MAGNETOMOTIVE FORCE

abamp-turn AMP-TURN gilbert


I

1 abampere-turn = 1 10 12.57
1 AMPERE-TURN = 0.1 1 1.257
1 gilbert = 7.958 X 10- 2 0.7958 1
I
1 pragilbert = 4, amp-turn 1 esu = 2.655 X 10- 11 amp-turn

MAGNETIC FIELD STRENGTH H

AMP-
amp-
abamp-turn/cm amp-turn/in. TURN/ oersted
turn/cm
METER

1 abampere-turn per
centimeter = 1 10 25.40 1000 12.57
1 ampere-turn per centi-
meter = 0.1 1 2.540 100 1.257
1 ampere-turn per inch = 3.937 X 10- 2 0.3937 1 39.37 0.4947
1 AMPERE-TURN per
METER= 0.001 0.01 2.540 X 10-2 1 1.257 X 10- 2
1 oersted = 7.958 X 10- 2 0.7958 2.021 79.58 1

1 oersted = 1 gilbert 1 esu = 2.655 X 10- 9 amp-turn/meter


1 praoersted = 41r amp-turn/met.er
Mathematical Symbols and the
Greek Alphabet
APPENDIX H

Mathematical signs and symbols


= equals
,......, equals approximately
~ is not equal to
= is identical to, is defined as
> is greater than (» is much greater than)
< is less than (« is much less than)
;;;;; is more than or equal to (or, is no less than)
;;;:; is less than or equal to (or, is no more than)
± plus or minus (e.g., V4 = ±2)
cc is proportional to (e.g., Hooke's law: F cc x, or F = -kx)
~ the sum of
x the average value of x
The Greek alphabet
Alpha A a Nu N v
Beta
Gamma
B
r
{3
'Y
Xi
Omicron
-
';:;'

0
~
0
Delta ~ 0 Pi IT 7r

Epsilon E E Rho p p
Zeta z r Sigma ~ q-
Eta H 1] Tau T T
Theta e () l ,a. Upsilon T v
Iota I Phi <I> cf,, <p
Kappa K IC Chi x x
Lambda A )\ Psi '1'" iJ;
Mu M µ Omega Q w
49
Mathematical Formulas
APPENDIX

Quadratic formula

-b + Vb 2 - 4ac
If ax 2 + bx +c= 0, then x = - - - - - - -
2a

Trigonometric functions of angle () y

sin()
x
= '!!__ cos() =-
r r

y x
tan() =- cot() = -
x y
r r
sec() = - csc () = -
x y
Fig. App. I
Pythagorean theorem
x2 + y2 = r2
Trigonometric identities
sin 2 () + cos = 1
2 () sec 2 () - tan 2 () = 1 csc 2 () - cot 2 () = 1
sin (a ± /3) = sin a cos /3 ± cos a sin /3
cos (a ± /3) = cos a cos /3 + sin a sin /3
t an (a R) = -
tJ + tan a ± tan /3
-----
- 1 + tan a tan /3
sin 2() = 2 sin () cos ()
cos 2() = cos 2 () - sin 2 () = 2 cos 2 () - 1 = 1 - 2 sin 2 ()
. ei8 _ e-i8 ei8 + e-i8
sm() = - - - - cos()=----
2i 2
e±;9 = cos () ± i sin ()

Taylor's series
x2 :r3
f(xo + x) = f(xo) + .f'(xo)x + f"(xo) -2! + f"'(xo) _:3!____ +
50
MATHEMATICAL FORMULAS 51
Series expansions (these expansions converge for -1 < x < 1, except as noted)
1
-- =
l+x
1 - x +x 2 - x3 + ··
_ ;-- x x2 x3
vl+x=l+---+-- · ·
2 8 16
1 x 3x 2 5x 3
---==1--+---+
v1 + x 2 s rn
x2 x3
ex = 1 + x + - + - + (-oo<x<oo)
2 6
. x3 xo
smx=x--+- - (-oo<x<oo)
6 120
x2 x4
cos x = 1 - - + - - (-oo<x<oo)
2 24
x3 2x 0
tanx=x+-+-+ (-71'/2 < x < 71'/2)
3 15
n n(n - 1)
(X + y)n = xn + _ xn-Iy + xn-2y2 +
1! 2!

Derivatives and indefinite integrals

In what follows, the letters u and v stand for any functions of x, and a and mare
constants. To each of the integrals should be added an arbitrary constant of
integration. A Short Table of Integrals by Peirce and Foster (Ginn and Co.) gives
a more extensive tabulation.
dx
1. - = 1 1.fdx=x
dx
d du
2. - (au)= a - 2. f au dx = a f u dx
dx dx
d du dv
3. - (u + v) = - +- 3. f (u + v) dx = f u dx +f v dx
dx dx dx
d
4. - xm = mxm-l
dx
4. f xm dx = ::\ (m -:;,!= -1)
d
5. -ln x = -
dx
1
x 5. f~ = ln JxJ
d
dx
dv
dx
du
6. -(uv) = u - + v -
dx
6. f u dv dx = uv -
dx .
f v du dx
dx
d
-/,-ex=ex
dx
7. f ex dx = ex

8. !!_ sinh x = cosh x 8. f cosh x dx = sinh x


dx
52 APPENDIX I

9. !:_ cosh x = sinh x 9. J sinh x dx = cosh x


dx
d
10. - arctan x
dx
= --
1
1
+x 2
10. f~ +
1 x2
= arctan x

d
dx
.
11. - arcsm x = _ 1
1
-v 1 - x2
11. fV dx
1 - x2
= arcsin x

d
12. - arcsec x =
dx
_1
xv x 2
1
- 1
12. fV x
dx
x2 - 1
= arcsec x

d .
13. - cos x = - sm x 13. J sin x dx = - cos x
dx

14. !:_ sin x = cos x 14. J cos x dx = sin x


dx
d
15. - tan x = sec 2 x 19. J tan x dx = ln lsec xi
dx
d
16. - cot x = - csc 2 x 20. J cot x dx = ln lsin xi
dx
d
17. - sec x = tan x sec x 21. J sec x dx = ln /sec x + tan xi
dx
d
18. - csc x = - cot x csc x 22. J csc x dx = ln Iese x - cot xi
dx

Vector products
Let i, j, k be unit vectors in the x, y, z directions. Then
i · i = j · j = k · k = 1, i · j = j · k = k · i = 0,

i X i = j X j = k X k = 0,
ix j = k, j xk = i, k xi= j.
Any vector a with components a.,, ay, a. along the x, y, z axes can be written
a = a.,i + ayj + a.k.
Let a, h, c be arbitrary vectors with magnitudes a, b, c. Then
a X (h + c) = a Xh +a X c
(sa) X h = a X (sh) = s(a X h) (s a scalar).
Let B be the smaller of the two angles between a and h. Then
a · h = h · a = axbx + ayby + a.b. = ab cos B
i j k
a xh = -h xa = ax ay a. = (aybz - byaz)i + (a.bx - b.a,,)j
b,, by b. + (a.,by - bxay)k
la X hi = ab sin B
a · (h X c) = h · (c X a) = c · (a X h)
a X (h X c) = (a· c)h - (a· h)c
Values of Trigonometric Functions
APPENDIX J

TRIGONOMETRIC FUNCTIONS

Radians Degrees Sines Cosines Tangents Cotangents

.0000 0 .0000 1.0000 .0000 co 90 1.5708


.0175 1 .0175 .9998 .OJ.75 57.29 89 1.5533
.0349 2 .0349 .9994 .0349 28.64 88 1.5359
.0524 3 .0523 .9986 .0524 19.08 87 1. 5184
.0698 4 .0698 .9976 .0699 14.30 86 1.5010
.0873 5 .0872 .9962 .0875 11.430 85 1.4835
.1047 6 .1045 .9945 .1051 9.514 84 1.4661
.1222 7 .1219 .9925 .1228 8.144 83 1.4486
.1396 8 .1392 .9903 .1405 7.115 82 1.4312
.1571 9 .1564 .9877 .1584 6.314 81 1. 4137
.1745 10 .1736 .9848 .1763 5.671 80 1.3963
.1920 11 .1908 .9816 .1944 5.145 79 1.3788
.2094 12 .2079 .9781 .2126 4.705 78 1.3614
.2269 13 .2250 .9744 .2309 4.332 77 1.3439
.2443 14 .2419 .9703 .2493 4.011 76 1.3265
.2618 15 .2588 .9659 .2679 3.732 75 1.3090
.2793 16 .2756 .9613 .2867 3.487 74 1.2915
.2967 17 .2924 .9563 .3057 3.271 73 1.2741
.3142 18 .3090 .9511 .3249 3.078 72 1.2566
.3316 19 .3256 .9455 .3443 2.904 71 1.2392

Cosines Sines Cotangents Tangents Degrees Radians

53
54 APPENDIX J

TRIGONOMETRIC FUNCTIONS (Continued)

Radians Degrees Sines Cosines Tangents Cotangents


I
.3491 20 .3420 .9397 .3640 2.748 70 1.2217
.3665 21 .3584 .9336 .3839 2.605 69 1.2043
.3840 22 .3746 .9272 .4040 2.475 68 1.1868
.4014 23 .3907 .9205 .4245 2.356 67 1.1694
.4189 24 .4067 .9135 .4452 2.246 66 1.1519

.4363 25 .4226 .9063 .4663 2.144 65 1.1345


.4538 26 .4384 .8988 .4877 2.050 64 1.1170
.4712 27 .4540 .8910 .5095 1.963 63 1.0996
.4887 28 .4695 .8829 .5317 1.881 62 1. 0821
.5061 29 .4848 .8746 .5543 1.804 61 1.0647

.5236 30 .5000 .8660 .5774 1.732 60 1.0472


.5411 31 .5150 .8572 .6009 1.664 59 1.0297
.5585 32 .5299 .8480 .6249 1.600 58 1.0123
.5760 33 .5446 .8387 .6494 1.540 57 0.9948
.5934 34 .5592 .8290 .6745 1.483 56 0.9774

.6109 35 .5736 .8192 .7002 1.428 55 0.9599


.6283 36 .5878 .8090 .7265 1.376 54 0.9425
.6458 37 .6018 .7986 .7536 1.327 53 0.9250
.6632 38 .6157 .7880 .7813 1.280 52 0.9076
.6807 39 .6293 .7771 .8098 1.235 51 0.8901

.6981 40 .6428 .7660 .8391 1.192 50 0.8727


.7156 41 .6561 .7547 .8693 1.150 49 0.8552
.7330 42 .6691 .7431 .9004 1.111 48 0.8378
.7505 43 .6820 .7314 .9325 1.072 47 0.8203
. 7679 44 .6947 . 7193 .9657 1.036 46 0.8029
.7854 45 . 7071 .7071 1.0000 1.000 45 0.7854
-
Cosines Sines Cotangents Tangents Degrees Radians
Nobel Prize Winners 1n Physics
APPENDIX K

1901 ·Wilhelm Konrad Rontgen 1845-1923 Disco.-ery of X-rays.


1902 Hendrik Antoon Lorentz 1853-1928 Influence of magnetism on the
Pieter Zeeman 1865-19-±3 phenomena of atomic radiation.
1903 Henri Becquerel 1852-1908 Discovery of natural radioacti,·ity
Pierre Curie 1850-1906 and of the radioacth·e elements
:\Iarie Curie 1867-1934 radium and polonium.
1904 John W. Strutt (Rayleigh) 1842-1919 Discovery of argon.
1905 Philipp Lenard 1862-1947 Research in cathode rays.
1906 Joseph John Thomson 1856-19-±0 Conduction of electricity through
gases.
1907 Albert •.\ . .:Siic helson 1852-1931 In.-ention of interferometer and
spectroscopic and metrological
investigations.
1908 Gabriel Lippmann 1845-1921 Photographic reproduction of
colors.
1909 Guglielmo .:Siarconi 187-1-1937 De.-eloprnent of wireless teleg-
Karl Ferdinand Braun 1850-1918 raphy.
1910 .Johannes Diderik Yan der "\Yaals 1837-1923 Equations of state of gases and
fluids.
1911 Wilhelm \rien 186-1-1928 La,n of heat radiation.
1912 ~ils Gustaf Dalen 1869-1937 Automatic coastal lighting.
1913 Heike Kamerlingh-Onnes 18,53-1926 Properties of matter at low tem-
peratures; production of liquid
helium.
1914 .:Siax .-on Laue 1879-1960 Diffraction of X-rays in crystals .
1915 "\Yilliam Henry Bragg 1862-19-±2 Study of crystal structure by
"\Yilliam La,nence Bragg 1890- means of X-rays.

1917 Charles Glo.-er Barkla 1877-lOH Disco.-ery of the characteristic


X-rays of elements.
1918 l\Iax Planck 1858-1947 Discovery of the elemental quan-
tum.
19HI Johannes Stark 1874-19.57 Discovery of the Doppler effect
in canal rays and the splitting
of spectral lines in the electric
field.
SS
56 APPENDIX K
1920 Charles Edouard Guillaume 1861-1938 Discovery of the anomalies of
nickel·steel alloys.
1921 Albert Einstein 1879-1955 Discovery of the law of the photo-
electric effect.
1922 Niels Bohr 1885-1963 Study of structure and radiations
of atoms.
1923 Robert Andrews l\lillikan 1868-1953 \York on elementary electric
charge and the photoelectric
effect.
1924 Karl l\lanne Siegbahn 1886- Discoveries in the area of X-ray
spectra.
1925 James Franck 1882-1964 Laws governing collision between
Gustav Hertz 1887 electron and atom.
1926 Jean Perrin 1870-1942 Discovery of the equilibrium of
sedimentation.
1927 Arthur H. Compton 1892-1962 Discovery of the sea ttering of
X-rays by charged particles
Charles T. R. \\·ilson 1869-1959 Im·ention of the cloud chamber,
a device t.o make visible the
paths of charged particles.
1928 Owen ·wmars Richardson 1879-1959 Discovery of the law known by
his name (the dependency of
the emission of electrons on
temperature).
1929 Louis.Victor de Broglie 1892- ,Yave nature of electrons.
1930 Chandrasekhara Raman 1888-1970 Work on the scattering of light
and discovery of the effect
known by hi& name.

1932 Werner Heisenberg 1901- Creation of quantum mechanics.


1933 Paul Adrien l\laurice Dirac 1902- Discovery of new fertile forms of
EnYin Schrodinger 1887-1961 the atomic theory.

1935 James Chadwick 1891- Discovery of the neutron.


1936 Victor Hess 1883-1964 Discovery of cosmic radiation.
Carl David Anderson 1905- Discovery of the positron.
1937 Clinton Joseph Davisson 1881-1958 Discovery of diffraction of elec-
George P. Thomson 1892- trons by crystals.
1938 Enrico Fermi 1901-1954 Artificial radioactive elements
from neutron irradiation.
1939 E. 0. Lawrence 1901-1958 Invention of the cyclotron.

1943 Otto Stern 1888-1969 \York with molecular beams and


magnetic moment of proton.
1944 Isidor Isaac Rabi 1898- Nuclear magnetic resonance.
1945 Wolfgang Pauli 1900-1958 Discovery of quantum exclusion
principle.
1946 Percy Williams Bridgman 1882-1961 High·pressure physics.
1947 Edward Appleton 1892-1965 Upper atmosphere physics and
discovery of Appleton layer.
1948 Patrick Maynard Stuart Blackett 1897- Discoveries in cosmic radiation
and nuclear physics.
1949 Hideki Yukawa 1907- Prediction of existence of meson.
1950 Cecil Frank Powell 1903- nhotographic method of studying
nuclear processes; discoveries
about mesons.
1951 John Douglas Cockcroft 1897-1967 Transmutation of atomic nuclei
Ernest Thomas Sinton ,Yalton 1903- by artificially accelerated
atomic particles.
1952 Felix Bloch 1905- Measure of magnetic fields in
Edward Mills Purcell 1912- atomic nuclei.
1953 Frits Zernike 1888-1966 Invention of phase contrast
microscopy.
1954 Max Born 1882-1970 Work in quantum mechanics and
statistical interpretation of
wave function.
NOBEL PRIZE WINNERS IN PHYSICS 57
Walther Bothe 1891-1957 Analysis of cosmic radiation using
the coincidence method.
1955 Willis E. Lamb, Jr. 1913 Fine structure of hydrogen.
Polykarp Kusch 1911- Magnetic moment of electron.
1956 John Bardeen 1908- Invention and development of
Walter H. Brattain 1902- transistor.
William B. Shockley 1910-
1957 Chen Ning Yang 1922- Non-conservation of parity and
Tsung Dao Lee 1926- work in elementary particle
theory.
1958 Pavel A. Cerenkov 1904- Discovery and interpretation of
Ilya M. Frank 1908- Cerenkov effect of radiation by
Igor Y. Tamm 1895- fast charged particles in matter.
1959 Owen Chamberlain 1920- Discovery of the antiproton.
Emilio Gino Segre 1905-
1960 Donald A. Glaser 1926- Invention of bubble chamber.
1961 Robert L. Hofstadter 1915- Electromagnetic structure of nu-
cleons from high-energy elec-
tron scattering.
Rudolf L. Mossbauer 1929 Discovery of recoilless resonance
absorption of gamma rays in
nuclei.
1962 Lev D. Landau 1908- Theory of condensed matter;
phenomena of superfluidity and
superconductivity.
1963 Eugene B. Wigner 1902- Contributions to theoretical
atomic and nuclear physics.
Maria Goeppert-Mayer 190i- Shell model theory and magic
J. H. D. Jensen 1907- numbers for the atomic nucleus.
1964 C.H. Townes 1915- Invention of the maser and theory
Nikolai Basov 1922- of coherent atomic radiation.
Aleksandr Prokhorov 1916-
1965 Richard Feynman 1918 Development of quantum electro-
Julian Schwinger 1918 dynamics
Shin-Ichiro Tomonaga 1906
1966 Allred Kastler 1902 Optical pumping studies
1967 Hana A. Bethe 1906 Energy production in staffl
The Gaussian System of Units
APPENDIX L

Much of the literature of physics is written, and continues to be written, in the


Gaussian system of units. In electromagnetism many equations have slightly
different forms depending on whether it is intended, as in this book, that mks
variables be used or that Gaussian variables be used. Equations in this book can
be cast in Gaussian form by replacing the symbols listed below under "rationalized
mks" by those listed under "Gaussian." For example, Eq. 37-26,

B = µ 0 (H + M)
becomes !!c = (
41r) (~ H
c2 41r
+ cM)
or

in Gaussian form. Symbols used in this book that are not listed below remain
unchanged. The quantity c is the speed of light.

Quantity Rationalized mks Gaussian

Permittivity constant Eo 1/4,r


Permeability constant µo 41r/c 2
Electric displacement D D/41r
Magnetic induction B B/c
Magnetic flux <(>B <(>B/ C
Magnetic field strength H cH/411'
Magnetization M cM
Magnetic dipole moment IL ctJ,

58
THE GAUSSIAN SYSTEM OF UNITS 59
In addition to casting the equations in the proper form it is of course necessary
to use a consistent set of units in those equations. Below we list some equivalent
quantities in mks and Gaussian units. This table can be used to transform units
from one system to the other.
Quantity Symbol Mks system Gaussian system

Length l 1 meter 10 2 cm
Mass m 1 kg 10 3 gm
Time t 1 sec 1 sec
Force F 1 newton 10 5 dynes
Work or Energy W,E 1 joule 10 7 ergs
Power p 1 watt 107 ergs/sec
Charge q 1 coulomb 3 X 10 9 statcoul
Current i 1 ampere 3 X 10 9 statamp
Electric field strength E 1 volt/meter i X 10- 4 statvolt/cm
Electric potential v 1 volt tio statvolt
Electric polarization p 1 coul/meter 2 3 X 10 5 statcoul/cm 2
Electric displacement D 1 coul/meter 2 121r X 10 5 statvolt/cm
Resistance R 1 ohm t X 10- 11 sec cm- 1
Capacitance c 1 farad 9 X 10 11 cm
Magnetic flux q>B 1 weber 10 8 maxwells
Magnetic induction B 1 tesla = 1 weber/ 10 4 gauss
meter 2
Magnetic field strength H 1 amp-turn/meter 41r X 10- 3 oersted
Magnetization M 1 weber/meter 2 1/41r X 10 4 gauss
Inductance I 1 henry t x 10- 11

All factors of 3 in the above table, apart from exponents, should be replaced by
(2.997925 ± 0.000003) for accurate work; this arises from the numerical value of
the speed of light. For example the mks unit of capacitance ( = 1 farad) is actu-
ally 8.98758 X 10 11 cm rather than 9 ( = 3 2 ) X 10 11 cm as listed above. This
example also shows that not only units but also the dimensions of physical quan-
tities may differ between the two systems. In the mks system (see Appendix F)
the dimensions of capacitance are M- 1L- 2 T 2 Q2 ; in the Gaussian system they are
simply L, the Gaussian standard unit of capacitance being 1 cm.
The student should consult Classical Electromagnetism, p. 611, by J. D. Jackson
(John Wiley and Sons, 1962) for a fuller treatment of units and dimensions.
Answers to Odd-Numbered
Problems

Chapter 26
1. 0.82 amp. 3. ±2.4 X 10-8 coul.
5. 2.3 X 10-s coul. 7. (a) Q = -2-y2q,
(b) no.
9. (a) 5.7 X 10 13 coul, 11. 3.8 X 10-5 coul, 1.2 X 10-5 coul.
(b) no,
(c) 660 tons.
13. (b) Along the body diagonal. 15. (a) 6.3 X 10 11 ,
(b) 7.3 X 10-13 ,
17. (a) 510 nt,
(b) 7. 7 X 1028 meters/sec 2•

Cha{er 27
.!:? -11
5. .fr X 10 coul.
9. (a) The larger charge produces a field of 13 X 104 nt/coul at the site of the small!
the smaller charge produces a field of 5.3 X 104 nt/coul at the site of the larger.
(b) 1.1 X 10-2 nt, repulsive .
. 11. E lies in the median plane and points radially away from the charge axis.
13. 1.0 X 105 nt/coul, pointing up.
17. E = q/8noa 2, pointing along the axis of symmetry and away from the hemisphe
1 qx
21. (a) E = - , 25. (a) 1.5 X 103 nt/coul,
41rfo (a2 + x2) %
(b) 2.4 X 10- 16 nt (up),
(b) E = _1_ (q1 - q2)a. (c) 1.6 X 10- 25 nt,
21r 2 fo (a 2 + x 2) % (d) 1.5 x 10 10 •
27. (a) 7.1 cm,
(b) 2.9 X 10-8 sec,
(c) 11 %.
dE Sq
29. -d = - - l3 , where l is the distance between the charges; yes.
z 11"€0

31. 1.64 X 10-19 coul.


61
62 ANSWERS TO ODD- NUMBERED PROBLEMS

Chapter 28
1. 1rR 2E (an expected result). 3. EA cos o.
7. (a) 4.0 X 106 nt/coul, 9. (a) E = ,, / eo, to the left,
(b) E = 0. (b) E = 0,
(c) E = <T/eo, to the right.
11. 4.9 X 10-10 coul. 13. 0.44 mm.

19. (a) E = _q_ (radially inward), where l is the length of the cylinders,
21rtolr
(b) -q on inner surface and -q on outer surface,

(c) E = _q_ (radially outward).


21reolr
21. ,nX 10- 9 coul/meter 2. 23. 1.9 X 102 nt,
2.9 X 1028 meters/sec 2•
25. (a) 1.1 nt-meters 2 /coul,
(b) 9.3 X 10- 12 coul.

Chapter 29
1. 0.89 mm. 3. 900 volts.
5. (a) Between the charges a distance of 25 cm from +q and outside the charges a
distance of 50 cm from +q.
(b) Outside the charges a distance of 140 cm from +q.
q d
7. VA - V s = - - - - - -
21reo a(a d) +
9. 1.9 X 10-29 coul meter; text value (0.61 X 10-29 coul meter) is lower and correct; the
assumptions made in this problem are oversimplified.
11. 99 tons. 13. -6.4 X 10-7 joule.
15. (a)-0.12 volt, 17. 2900 volts.
(b) 1.8 X 10-8 nt/coul, radially inward.
21. No. 23. (a) -180 volts,
(b) +2900 volts and -9000 volts.
27. -0.2lq 2/ eoa.
29. (a) 1.1 X 10 17 volts/meter,
(b) 4.6 X 10 21 volts/meter, assuming a nuclear radius of 5 X 10-15 meter.
31. (a) 2.6 X 105 volts, 33. 9.0 kw.

(b)
v5 c =
3 0.745 c.
35. (a) 3.2 X 10-13 joule,
(b) 1.6 X 10-13 joule,
(c) proton.

Chapter 30
1. (a) q2 = qs = 4.8 X 10-4 coul, 3. 43 µµf.
V2 = 240 volts,
Vs = 60 volts;
(b) q2 = 2.0 X 10-4 coul,
qs = 7.7 X 10-4 coul,
V2 = Vs = 96 volts;
(c) q2 = qg = zero,
V2 = Vs = zero.
11. 3.2 µf. 13. 4 µf.
15. 7.3 µf. 21. Mica.
23. Assuming K = 5.4:
(a) 104 volts/meter,
(b) +5.0 X 10-9 coul, on the positive plate,
ANSWERS TO ODD-NUMBERED PROBLEMS 63
(c) -4.1 X 10-9 coul, next to the positive plate.
25. 0.63 meter2 • 27. 0.11 joule/meter3 •
31. (a) 1.3 X 10-7 joule, 33. 7.0 X 10-5 coul.
(b) no.
35. 7.0f 37. (a) q1 = q2 = 3.3 X 10-4 coul,
qa = 4.0 X 10-4 coul;
(b) V1 = 33 volts,
V 2 = 67 volts,
Va = 100 volts;
(c) U1 = 5.4 X 10-3 joule,
U2 = 10.9 X 10-3 joule,
Ua = 2.0 X 10-2 joule.
Chapter 31
1. (a) 1200 coul, 3. (a) 2.4, iron being larger;
(b) 7.5 X 1021 electrons. (b) no.
5. 6. 7X 10-5 coul/meter2• 7. 54 ohms.
9. (a) 2.2 X 10-1 ohm, 11. (a) 260°C,
(b) nickel (p = 6.8 X 10-11 ohm -meter). (b) yes.
13. 0.39% (p), 17. 11 ohms.
0.0017% (Z),
0.0034% (A).
19. (a) 4.9 X 106 amp/meter 2, 21. 620 watts.
(b) 8.4 X 10-3 volt/meter,
(c) 2 .6 volts,
(d) 65 watts.
23. (a) 8. 7%,
(b) smaller.

Chapter 32
1. 1.1 X 104 joules. 3. -0.62%.
7. (a) 990 ohms, 9. vd - v. = 1.3 volt.
(b) 9.9 X 10-4 watt.
11. (a) 120 ohms, 13. (a) R = !r,
(b) i1 = 0.051 amp, (b) FJ'l /2r.
i2 = ia = 0.019 amp,
i4 = 0.013 amp.
17. 38 ohms. 19. (a) 10 ohms,
(b) 14 ohms,
(c) 10 ohms.
.
21. series: i
.= 2e
---·
R + 2r
parallel: i = _f,_,_ •
2R +r
The series arrangement always gives the larger current.
23. (a) Put R 1 roughly in the middle of its range; adjust current roughly with R 2 ; make
fine adjustment with R 1•
(b) Relatively large percentage changes in R 1 cause only small percentage changes in
the resistance of the parallel combination, thus permitting fine adjustment. The
ratio is 1 :21.
27. 2.7%. 29. 4.6 time constants.
33. (a) 9.6 X 10- 7 coul/ sec,
(b) 1.1 X 10-5 watt,
(c) 2.7 X 10-5 watt,
(d) ~ SI v 1 0-6 watt.
64 ANSWERS TO ODD-NUMBERED PROBLEMS

Cbapter33
1. (a) East,
(b) 6.3 X 1014 meters/sec 2,
(c) 3.0 mm.
3. 7.5 nt, perpendicular to the wire and to B.
5. See Appendix F. 7. 3.8 coul.
9. 450 gauss, perpendicular to the rod and inclined at an angle of 31 ° to the normal to
the plane of the tracks.
11. 4.3 X 10-3 nt-meter. The torque vector is parallel to the long side of the coil and
points down.
15. (a) 1.4 X 10-4 meter/sec, 17. (a) Kp = Kd = fKa,
(b) 4.5 X 10-23 nt. (down), (b) Rd= 14 cm.
(c) 2.8 X 10-4 volt/m~ter (down), (c) Ra= 14 cm.
(d) 5.7 X 10-6volt(top +,bottom-),
(e) same as (b).
19. (a) 2.6 X 107 meters/sec,
(b) 1.1 X 10-7 sec,
(c) 14 Mev,
(d) 7.0 X 106 volts.
23. 1.6 X 10-s weber /meter 2, horizontal and at right angles to the equator.
25. T = 3.6 X 10-10 sec, 27. 2.11 X 10-25 kg or 127 proton masses.
p = 0.17 mm,
r = 1.5 mm.
29. (a) Increase, 31. (a) 8.5 Mev,
(b) decrease. (b) 0.80 weber/meter2 ,
(c) 34 Mev,
(d) 24 me/sec
(e) 34 Mev, 1.6 webers/meter 2,
34 Mev, 12 me/sec.
33. 1.4. 35. 3800 meters/sec.

Chapter34
1. 7.9 X 10-3 weber/meter 2•
3. (a) 3.2 X 10-15 nt, parallel to current;
(b) 3.2 X 10-15 nt, radially outward if v is parallel to the current;
(c) zero.
7. 1.0 X 10-5 weber/meter.
9. B = 0 along a line parallel to the wire and 4.0 mm from it. If the current is horizontal
and points toward the observer and the external field points horizontally from left
to right, the line is directly above the wire.
11. 8.0 X 10-5 weber/meter 2, up. 13. 3.2 X 10-3 nt, toward the long wire.
23. (a) 1.0 X 10-3 weber/meter2 , out of figure;
(b) 8.0 X 10-4 weber/meter2, out of figure.
27. (a) 9.4 X 10-5 weber/meter2,
(b) 1.5 X 10-5 nt-meter.

Chapter 35
1. 2.0 X 10-2 coul. 3, Zero.
9. Vm = 1r2BR 2,,, 11. 3.0 X 10-4 volt.
i,,, = 1r2BR 2,,/RM
13. (a) 3.1 X 10-2 webers/sec, 17. (a) 4.4 X 107 meters/sec 2, to the right;
(b) left to right. (b) zero;
(c) 4.4 X 107 meters/sec 2, to the left.
ANSWERS TO ODD-NUMBERED PROBLEMS 65
Cllapter36
1. Let the current change at 10 amp/sec. 7. 1.0 X 10-7 weber.
9. 12 sec. 13. 27 amp/sec.
15. (a) 10 amp, 17. 0.63 joule/meter3 •
(b) 100 joules.
21. (a) 2.5 X 10-6 joule/meter, 23. 1.5 X 108 volt/meter.
(b) 14 X 10-6 joule/meter,
(c) 0.8 X 10-5 joule/meter.

Chapter 37
l. (a) 5.0 X 107 amp, 3. (a) 1.4 X 1011 volts/meter,
(b) yes, (b) 2.8 X 10-3 weber /meter2•
(c) no.
s. (a) 7.6 amp meter 2, 9. 7.5 X 10-6 weber /meter2•
(b) 11 nt-meter.
1 i. (a) 1.8 webers/meter2,
(b) 6.5 X 10-23 joule.

Chapter 38
1. 600, 710, 1100, and 1300 cycles/sec. 13. (a) 224 rad/sec.
(b) 207 or 241 rad/sec.

dE
19. id = EQ1r1° 2 -
dt
(r < R),
dE
id = Eo1rR2 -
dt
(r > R).
21. (a) 1.4 X 1014 volts/meter-sec,
(b) 9.9 X 10-6 weber/meter2 •

Chapter 39
3. 5. 7 cm, 9.0 X 108 meters/sec 9. 18 cm, 12 cm.
(=1.7c).
13. (a) a 2EB/µo
for faces parallel to xy-plane, zero for others;
(b) zero.
15. 1000 volts/meter,
3.4 X 10-6 weber/meter 2•

Chapter 40
1. (a) 5100 A and 6100 A,
(b) 5.5 X 10 14 cycles/sec and 1.8 X 10-15 sec.
3. F,ad = 6.0 X 108 nt, 7. 1.3 X 10-7 nt/meter2 •
Fgr = 3.6 X 10 22 nt.
11. 0.13. 13. (b) 7.3 (cycles/sec)/(miles/hour).
15. 3.8 X 10-2 A. 17. Yellow-orange.

Chapter 41
3. 2.05 X 108 meters/sec. 5, 1.56.
15. (b) 0.17.
19. Cover the center of each face with a circle of radius 0. 45 cm. The fraction covered
is 0.64.

Chapter 42
1. (a) 7, 3. 40 cm.
(b) 5,
(c) 2.
66 ANSWERS TO ODD-NUMBERED PROBLEMS

6. Alternate vertical columns: 7. Object at center of curvature.


(a) +, +40, -20, +2, no, yes;
(c) concave, +40, +60, -2, yes, no;
(e) convex, -20, +20, +0.5, no, yes;
(g) -20, -, -, +5, +0.80, no, yes.
9. 4.24 cm. 10. Alternate vertical columns:
(a) -18, no;
(c) +71, yes;
(e) +30, no;
(g) -26, no.
11. Alternate vertical columns:
(a)+, X, X, +20, X, -1, yes, no;
converging, +, X, X, -10, X, no, yes;
(c)
converging, +30, -15, +1.5, no, yes;
(e)
diverging, -120, -9.2, +0.92, no, yes;
(g)
converging, +3.3, X, X, +5, X, -, no.
(i)
17. 4.5 cm, 9.0 cm.

19. Assuming the light is incident from the left, a distance of - r(n - 2 )2 to the right of
the right edge of the sphere. 2( n - 1)
21. (a) R' is negative and R" is positive,
(b) i = - - r - ,
2n+l
virtual and erect.
(c)
23. (a) Coincides in location with original object and is enlarged 5.0 times,
(c) virtual and inverted.

Chapter43
1. Slit separation must be 0.034 mm. 3. 0.15°.
5. 6.4 X 10-3 mm. 7. 3.0 mm.

9. I I; [1 + 8 cos c·d :in 8)l


= 2 11. y = 17.4 sin (c,,t + 13.3°).

13. 80 million. 15. 1.21.


17. 4800 A (blue). 19. 6700 A.
23. 5880A. 25. 6057.8021 A.

Chapter44
1. 0.17 mm.
3. (a) >-a = 2>..,,,
(b) coincidences occur when mb = 2m4 •
5, 79.7°. 7. (a) 52.9°,
(b) 10.2°,
(c) 5.1 °.
9. 9100 meters. 11. (a) 0.16 sec of arc,
(b) 7.4 X 107 km,
(c) 2.2 X 10- 2 mm.
15. 3. 17. (a) Must have d = 4a,
(b) Every fourth fringe is missing.

Chapter45
1. Three complete orders.
5. The intensity would be concentrated near the twentieth order for blue and the eleventh
order for red. The orders would overlap to such an extent as to appear almost white.
ANSWERS TO ODD-NUMBERED PROBLEMS 67
7. All wavelengths shorter than 6300 A.
9. 0°, ±10°, ±21 °, ±32°, ±45°, and ±62°.
11. 5200 A to 6200 A. 17. (a) 0.0032°/A,
0.0077°/A,
0.024°/A;
(b) 40,000,
80,000,·
120,000.
19. 3600 lines.
21. (a) 4.6 X 10-2 A.
(b) No. The resolution could normally be improved by going to a higher order
diffraction, but in this case m = 3 is the highest order that can exist (assuming
that the light falls on the grating at right angles).
23. (a) 6 X 104 A,
(b) 1.5 X 104 A,
(c) 0, 1, 2, 3, 5, 6, 7, 9. The tenth order is at 8 = 90°.
25. 3°, 38 ° (clockwise); 27. Yes, n = 3 for X = 1.29 A,
15°, 31° (counterclockwise). n = .4 for X = 0.97 A.

Chapter 46
1. (a) ±55°, ±125°,
(b) ±35°, ±145°.
3. Assuming that a right-handed coordinate system is used,
(a) circular, counterclockwise as seen facing the source;
(b) elliptical, counterclockwise as seen facing the source, major axis of ellipse along
y.= x;
(c) plane, along y = -x.
5. 55°30' to 55°46'. 11. 2.7 X 10-14 kg-meter 2 /sec2,
2.1 hr.

Chapter47
l. 4.8A. 3. 2.1 X 1015 photons/sec.
7. 5.9 X 10-6 ev. 9. (a) 2.0 ev,
(b) zero,
(c) 2.0 volts,
(d) 3000 A.
11. 100 years. 17. (a) 1,
5.3 X 10-11 meter,
(b)
1.1 X 10-34 joule-sec,
(c)
2.0 X 10-24 kg-meter /sec,
(d)
4.1 X 1016 radians/sec,
(e)
(J) 2.2 X 106 meters/sec,
(g) 8.2 X 10-s nt,
(h) 9.0 X 1022 meters/sec 2
(i) + 13.6 ev,
(j) -27.2 ev,
(k) -13.6 ev.
19. +0.21 ev. 21. 2.6 ev.
25. +54 ev, assuming the electron to be in its ground state initially.
27, (a) 2.6 X 10-13 meter,
(b) 2800 ev,
(c) 4.4 A.
68 ANSWERS TO ODD-NUMBERED PROBLEMS

Chapter 48
1. (a) 1.7 X 10-35 meter.
3. (a) 3.3 X 10-24 kg-meter/sec, for each;
(b) 38 ev for the electron and 6200 ev for the photon.
5. (a) Higher orders cannot exist for this accelerating potential and for these planes.
(b) 59°; the crystal must be rotated with respect to the incident beam to satisfy
Bragg's law for this new wavelength.
7. 1.5 A. 9. (a) 0.20,
(b) 0.40,
(c) 0.33,
(d) 0.33.
11. 0.32. 15. 6.6 X 10-23 kg-meter/sec.
Answers to Odd-Numbered
Supplementary Problems

Chapter 26 3. (a) 1.0 X 10-s coul from small to


large sphere;
l
I. -{q, located - from +q. (b) small sphere: 2 X 10-s coul,
3
3000 volts,
3. ± 1.0 X 10- 5 coul, + 3.0 X 10- 5 large sphere: 4 X 10-s coul,
coul. 3000 volts.
5. 1.61 X 10- 9 meter.
9. W= +__2_[_!_ _ _!_}

J.~q, g.q·
Chapter 27 81reo ri r2
qQ
II. ( a ) - - ·
I. 21r 21r 41reoK
g--E g+-E
m m
(b) ~ ·

3. -/J-2' vertically downward.


1r ,oR 13. (a) - I n X (y + L)
-·- - ·
41reo y

(b) -
x L
.
41r,o y(y + L)
Chapter 28 (c) Zero.
I. +3.54 X 10- 6 coul.
Chapter 30
3. 1.98 X 10 7 meters/sec.
Q C1C2 + C1Ca CV
5. q1 = 1 O,
5. A = - ·
21ra 2 C1C2 + C1C3 + C2C3
q2 = q3 =
Chapter 29
I. 1.2 X 10- 8 coul/meter 2 .
69
70 ANSWERS TO ODD-NUMBERED SUPPLEMENTARY PROBLEMS

7. (a) lOcofarad,
(b) µoi ~ ' out of page at an
(b) 13.6co farad, 211'R
(c) 1200Eo coul, 1200co coul,
(d) 10 4 volts/meter, angle of 18° with page.
(e) 2.1 X 10 3 volts/meter,
(f) 88 volts, Chapter 35
(g) 19 X 10 3co joules. µoi11'r 2 R 2
3. (a)
2x 3 '
Chapter 31
3µ01rir 2v
I. (a) 2.3 X 10 12 , (b) - - - ·
2R 2N 4
(b) 5.0 X 10 3,
(c) 10 7 volts. (c) In the same sense as the current
in the larger loop.
5. (a) PA = .10 ohm-meter, PB
.05 ohm-meter, -1
(b) EA = 1.0 volt/meter, EB = .50
5. (a) q = R [<1>B(t2) - <l>B(ti)],

volts/meter, (b) No. B field could still have


(c) jA = jB = 10 amp/meter 2 , changed between t1 and t2 induc-
(d) VA = 40 volts, VB = 20 volts. ing currents.
7. T = 130° C.
Chapter 36
Chapter 32
3. (a) i1 = 3.3 amp, i2 = 3.3 amp;
I. R = r1 - r2. (b) i1 = 4.5 amp, i2 = 2.7 amp;
3. (a) r, (c) i1 = 0, i2 = 1.8 amp;
(b) R. (d) i1 = 0, i2 = 0.
5. (a) y12 R, 5. (a) meter, (b) weber, (c) watt,
(b) fR, (d) coulomb, (e) ohm.
(c) {R.
7. (a) 85.0 ohm, Chapter 37
(b) 85.0 ohm.
qwr 2 • • •
9. (a) 1.0 X 10- 3 coul., I. - 2 -, m direction of w.
(b) 1.0 X 10- 3 amp.,
(c) V C = V R = 1000e-t volts, eB
3. (a) (N ± 1) - ,
(d) e- 21 watt. rn
(b) 744 X 10 10 rad/sec; 759 X 10 10
Chapter 33
rad/sec.
I. 1 is positively charged, 2 is neutral, 5. 29 amp.
and 3 is negatively charged.
3. 2. 7 X 10- 4 weber /meter 2. Chapter 38
5. (a) 542 ohms, series, I. Let T 1 be the period of inductor and
(b) 2.52 ohms, parallel. 900 µf capacitor and T2 the period of
7. (a) 2.8 Mc/sec, inductor and 100 µf capacitor. Then:
(b) 0.34 meter. (a) close 8 2 , wait ·Ti/4; (b) quickly
close 81 then open 82; (c) wait T2/4
Chapter 34
and then open 81.
1. - lOµoio.

3.
3 Y2 µoi , toward center of square.
2 Chapter 39
411'a 271'€0
I. (a) In (b/a)
5. zero.

7. µoi
- (- - -
4 R1 R2
1 1) .
, mto page. (b)
µo In (b /a)
271'

5. (a) 6.3 X 10- 2 volt/meter,


9. (a) -µoi ( 1
2R
+ -1) , 11'
out of page.
(b) 2.1 X 10- 10 weber/metn 2 ,
(c) 1.3 X 10 4 watts.
ANSWERS TO ODD-NUMBERED SUPPLEMENTARY PROBLEMS 71
7. (a) B = µorTRw. Chapter 43
(b) E = fµOrT R 2a. I. .072 mm.
(c) S = µo rT 2 R 3wa. 3. 8 microns.
2 7. 8400 A.
(d) They are both µ01rrT 2 R 4lwa. 9. 1.89 microns.

Chapter 44
Chapter 40
I. (a) The droplets act as diffracting
I. (a) 1.0 X 10 8 cycles/sec,
obstacles producing the observed
(b) 1.0 X 10- 5 weber/meter 2,
ring.
(c) 2.1 meter- 1 , 6.3 X 10 8 rad/sec,
(b) .20 mm.
(d) 120 watts/meter 2 ,
3. 71 ft.
(e) 4.0 X 10- 7 A nt, 4.0 X 10- 7 nt/
meter 2 • Chapter 45
3. (b) About 6 X 10-7 meter (com-
parable with the wavelength of 3. (a) 5860 A.
light). (b) Since the fourth maximum is
(c) No. missing, the slit width must lie
between 1.00 and 1.69 microns.
Chapter 41 5. (a) One line to each side of the cen-
tral maximum.
I. 1.95 X 10 8 meters/sec. (b) 1.44 X 10- 4 rad.
3. 180 cm.
5. (a) v~i_+_s-in_2_0_1 1
(c) b.O = -
R
(b) 0
Chapter 46
Chapter 42 I. 67% polarized, 33% unpolarized.
I. 6. 3. Side A has the Polaroid, and the
5. (b) No. Polaroid axis must be at 45° to the
(c) Rays pass undeviated through principal axes of the quarter-wave
system. plate.
Index

Aberrations, for lenses, 1061 B, see Magnetic induction


for mirrors, 1048 Ballistic galvanometer, 928
Absorption spectrum, 1189 Barkla, C. G., 1169
Acoustic cavity, 954, 977, 1203 Bees and polarized light, 1169
Action-at-a-distance, 663, 652 Betatron, 884
Addition of velocities, 1006 Beth, 1167
Airy, Sir George, 852, 1112 B-H curve, 937
Allen, A. J., 831 Biax~al crystal, 1162
Alpha-emission, 658 Binding energy, 1195
Alpha-particle, 700 Biot-Savart law, 859
Alternating current generator, 895 Birefringence, 1156
Ammeter, 800 Birge, R. T., 1002
Ampere, 652 Bloch, F, 931
Ampere, A. M., 844, 853 Bloembergen, N., 934
Amperian current, 936 Bohr, N., 1191, 1196, 1213
Ampere's law, 844, 847, 857, 911, 936, Bohr model, 864
961, 985
Bol, 1102
Amplitude grating, 1127
Boltzmann, L., 963
Analyzer, 1150
Andrade, E. N., 700, 701 Boltzmann's constant, 1179
Angstrom, 994 Born, M., 1208
Angular momentum, of electrons, 916, Bradley, J., 999
1193, 1204 Boundary conditions, 759, 1203
of light, 1167 for Band H, 937, 942
Anisotropic, 1155 for E and D, 958
Annihilation, 660 Bragg, W. L., 1142
Antiferromagnetism, 930 Bragg's law, 1140
Arago, D. F., 761, 1148 Brewster, D., 1155
Arrington, W., 1136 Brocklebank, R. W., 994
Atomic number, 658 Bubble chamber, 818
73
74 INDEX

Calcite, 1156, 1159 Conservation of energy, in circuits, 792


Capacitance, 741, 745 798
and dielectric, 748 Conservative force, 710
and inductance, 900 Constant-deviation prism, 1033
of cylindrical capacitor, 746 Continuous spectra, 1188
of parallel-plate capacitor, 745 Convex mirror, 1041
of spherical capacitor, 765 Corona discharge, 732
Capacitative time constant, 805, 902 Correspondence principle, 1196
Capacitors, and dielectrics, 741 Coulomb, 652
as circuit elements, 802 Coulomb, C. A., 650
force on plates of, 769 Coulomb's law, 657, 1192
in parallel, 747 Critical angle, 1024
in series, 748 Curie, P., 922
Carragan, G. H., 1118 Curie temperature, 926
Cavendish, H., 650 Curie's law, 922
Cavity, acoustic, 95"4 Current, 770
electromagnetic, 963 displacement, 962
Cavity radiator, 1174 induced, 871
CERN accelerator, 834 magnetic force on a, 819
Charge, and matter, 655 measurement of, 800
conservation of, 660 sense of, 772
in circular motion, 829 Current balance, 848, 854
in electric field, 675 Current density, 770
line of, 674 Current element, 860
quantization of, 654 Current loop, B for, 861
ring of, 673 torque on, 823
two kinds of, 648 Cutoff frequency, in photoelectric effect,
Circuits, 789 1180
charge conservation in, 798 in waveguide, 977
current in, 792, 793 Cyclotron, 831
energy conservation in, 792, 798 resonance in, 833
potential differences in, 795 Cyclotron frequency, 830
RC, 802 Cylindrical capacitor, 749
Circular aperture, diffraction at, 1112
Circular polarization, 1163 D, 758, 934
Circulating charges, 829 Damped oscillations, 950
Coaxial cable, 911, 971 Davisson, C. J., 1201
currents in, 97 4 de Broglie, L., 1200, 1204
energy flow in, 987 de Broglie wavelength, 1201
fields in, 97 4 Debye, P. W., 1186
Coblentz, 1177 Dees, 832
Coherence, 107 4 Descartes, R., 1015, 1023
Complementarity, 1213 Deuteron, 831
Compton, A. H., 1184 Diamagnetism, 923, 937
Compton effect, 1184 Dichroism, 1159
Compton shift, 1184 Dielectric constant, 749, 750
Concave mirror, 1041 Dielectrics, 649, 750
Conduction electrons, 781, 828 and capacitors, 748
drift speed of, 781 and Gauss's law, 753
Conductivity, 774 Dielectric strength, 750
electrical, 780 Diffraction, 1035, 1199
thermal, 780 and Huygens' principle, 1035
Conductor, 649 and interference, 1115, 1119
insulated, 730 at a circular aperture, 1112
Conservation of charge, 660 Fraunhofer, 1101
INDEX 75
Diffraction, Fresnel, 1200 Electric field, for dipole, 671, 728
from a single slit, 1102, 1105, 1108, 1109 for line of charge, 674, 698, 849
of matter waves, 1211 for sheet of charge, 698
of neutrons, 1202 in coaxial cable, 97 4
of water waves, 1037 in waveguide, 978
X-ray, 1123, 1134 induction of, 880
Diffraction grating, 1123, 1128 of point charge, 670
Dipole, electric, see Electric dipole of ring of charge, 673
magnetic, see Magnetic dipole of spherical charge, 695
Dipole antenna, 980 Electric polarization, 756
Dipole-dipole interaction, 942 Electric potential, see Potential, electric
Dispersion, 1133 Electric quadrupole, 683, 721
Displacement, electric, 758 Electric vectors, table of, 760
Displacement current, 962 Electromagnet, 815
Distributed elements, 954 Electromagnetic cavity, 955
in acoustics, 955 energy flow in, 988
in electromagnetism, 956 Electromagnetic induction, 870
Dominant mode, 977 Electromagnetic oscillations, 943
Domains, 929, 930 energy in, 950
Doppler, J., 1006 forced, 952
Doppler effect, 1006 quantitative, 948
for light, 1007 resonance in, 952
for sound, 1006 Electromagnetic shielding, 892
Double refraction, 1155 Electromagnetic spectrum, 993
Double slit, 1104, 1115 Electromagnetic waves, 971, 982, 986,
Drift speed of electrons, 771, 781 1002
Driscoll, R. L., 821 Electromotive force, 789
Duality, of light, 1183 gravitational analog of, 790
of matter, 1200 induced, 871, 899
wave-particle, 1203 Electron, 655
Du Mond, .J., 1189 angular momentum of, 916
e/m for, 835
E, see Electric field free, 649
Eddy currents, 891 moment of, 917, 925
Einstein, A., 1004, 1092, 1094, 1100, 1179, Electron microscope, 1115
1182, 1184 Electron volt, 833
Electret, 938 Electrostatic force, 711
Electric dipole, alternating, 720, 1148, Electrostatic generator, 733, 734
1167 Elsasser, W., 1201
energy of, 679 Emissivity, 1175
field due to, 681, 684, 724 Energy, and Poynting vector, 986
in dielectrics, 750 in capacitor, 760
induced, 720, 751, 757 in circuits, 783
in nonuniform field, 684 in electric field, 760, 790
potential due to, 718 in electromagnetic cavity, 986
properties of, 864 in LC oscillations, 943, 948
torque on, 825 in magnetic field, 790, 907
water molecule as, 819 in resonant systems, 955, 956
Electric displacement, 758, 934 quantization of, 1178
Electric field, 708, 712 Energy level diagram, 1194
and potential, 724 Equipotential surface, 710, 727
definition, of, 817 Essen, 1002
energy of, 760, 912 Ether, 1003
flux of, 686 Euclid, 1015
for a conductor, 699, 771 Exchange coupling, 926, 931
76 INDEX
Expanding universe, 1009 Glancing angle, 1140
Extraordinary ray, 1156 Goudsmit, S. A., 842
Eye, sensitivity, 994 Grating, 1123, 1128
dispersion of a, 1133
Farad, 743 resolving power of, 1132
Faraday, M., 647, 743, 748, 851, 852, 870, Gravitational field, 663, 665
871, 923 Gravitational force, 711
Faraday's law, 870, 894 Grimaldi, F. M., 1099
counterpart of, 959 Group speed, in waveguide, 978
in traveling waves, 983 Guide wavelength, 979
Fermat, P., 1028
Fermat's principle, 1028, 1161 H, 815, 934, 937
Ferrimagnetism, 930 Haidinger's brushes, 1152
Ferroelectric, 938 Half-silvered mirror, 1075
Ferromagnetism, 920, 926 Half-width, of diffraction maximum, 1111
Fiber optics, 1025 of resonance curve, 953
Field, electric, see Electric field Hall, E. H., 827
magnetic, see Magnetic induction Hall effect, 649, 771, 827
Fission, 658 Hansen, 1002
Fizeau, H. L., 999 Harmonic oscillator, 1178
Flow field, 663 HCl spectrum, 1190
Flugge, H., 1155 Heaviside, 0., 648
Flux, of electric field, 686 Hege, E. K., 807, 906
of magnetic field, 815 Heisenberg, W., 1211
Flux linkages, 899 Helmholtz coils, 869
Focal length, 1058 Henry, G. E., 995
Focal plane, 1059 Henry, J., 870, 900
Focal point, 1058 Henry, W. E., 923
Forced oscillations, 952 Hertz, H., 648, 986
Foucault, .J. B. L., 1000, 1023 Hipple, J. A., 821
Fourth-power law, 1175 Hull, G. F., 995, 996
Franklin, B., 648 Huygens, C., 1019, 1099, 1156
Fraunhofer diffraction, 1101 Huygens' principle, 1018
Free charge, 754 and double refraction, 1159
Free-electron model, 828 Hydrogen atom, 1191
Fresnel, A. J., 1099, 1148 Hydrogen spectrum, 1191, 1195
Fringes of constant thickness, 1086 Hysteresis, 929

Galileo, 998 Image, 1038


Galvanometer, 824 erect, 1048
ballistic, 928 inverted, 1048
Gamma rays, 660 real, 1038
Gamow, G., 1212 virtual, 1038
Gauss, 817 Image distance, 1043
Gauss's law, 686, 690, 859 Imes, E. S., 1190
and Coulomb's law, 690 Incident ray, 1013
and dielectrics, 753, 759 Index of refraction, 1014, 1023
applications of, 695 principal, 1157
for fluids, 703 Induced magnetic fields, 959
for gravitation, 704 Induced surface charge, 751, 754, 756
for magnetism, 920 Inductance, 899
Geiger, H., 700 calculation of, 901
Generator, alternating current, 895 energy storage in, 910
Geometrical optics, 1035, 1037 of solenoid, 901
Germer, L. H., 1201 of toroid, 901
INDEX 77
Induction, and relative motion, 888 Light, polarization of, 1147
electromagnetic, 870, 875 pressure of, 995
Faraday's law of, 872, 880 propagation of, 1092
lines of, 851 scattering of, 1167
Insulated conductor, 730 sensitivity of eye to, 994
Insulator, 649 sources of, 1173
Interference, 1068 speed of, 998, 1001, 1012
and diffraction, 1115, 1119 wave theory of, 982, 1023
double-slit, 1079 Line integral, 713
fringes, 1070 Lines of force, 665
in thin films, 1083 Lines of induction, 815, 846
Interferometer, 1090 Line spectra, 1188
Internal resistance, 793 Loop theorem, 793
Interplanar spacing, 1140 Lorentz, H. A., 648, 818, 835
Isotopes, 658 Lorentz relation, 818
Isotropic, 1155 LR circuit, 902
Ives, H. E., 1008 LRC oscillations, 946, 950
Lumped elements, 954
Jessop, H. T., 1155 electromagnetic systems, 957
Joule heating, 784, 908 mechanical systems, 957
Joule's law, 785
Joyce, J., 1021, 1022 Magnesia, 814
Junction theorem, 798 Magnetic damping, 892
Magnetic dipole, 825, 915
Kapany, N. S., 1025 B for, 863
Katz, J. L., 1136 energy of, 826
Kellogg, 1190 of bar magnet, 917
Kelvin, Lord, 733 of electron, 918
Kirchhoff's first rule, 798 properties of, 864
Kirchhoff's second rule, 793 torque on, 825
Magnetic domain, 929, 930
Land, E. H., 994 ::viagnetic field, see Magnetic induction
Laser, 1027 Magnetic field strength, 815, 934, 937
Lattice, 771, 781 Magnetic force, on a current, 819
Laue spots, 1136 on a moving charge, 816
Lawrence, E. 0., 831 Magnetic induction, 663, 814, 815
Lawton, 694 definition of, 816
LC oscillations, 943 energy and, 907
analogy of to SHM, 947 energy density and, 910
Lebedev, 995 for a circular loop, 862
Lens, 1055 for a dipole, 863
equation for, 1057 for a long wire, 861
extended object for, 1059 for antiparallel wires, 867
sign conventions for, 1057 for a traveling wave, 982
Lens maker's equation, 1059 for current sheet, 866
Lenz, H. F., 873 for Helmholtz coils, 869
Lenz's law, 873, 883, 900 for parallel wires, 853
Light, 993 for polygon, 868
and quantum physics, 1173 for rectangular loop, 869
and the electromagnetic spectrum, 993 for solenoid, 856
angular momentum of, 1167 for square loop, 868
corpuscular theory of, 1023 for toroid, 859
energy of, 995 in a cavity, 955
momentum of, 995 in a coaxial cable, 97 4
moving source of, 1003 in a waveguide, 1002
78 INDEX

Magnetic induction, induced, 959 Nuclear atom, 697


time-varying, 880 Nuclear force, 657
Magnetic pole, 915, 920 Nuclear magnetism, 931
Magnetic vectors, table of, 934 Nuclear resonance, 931
Magnetism, and matter, 915 Nucleus, 655
of nuclei, 931 radius of, 656
permanent, 929
Magnetite, 814 Object, virtual, 1052
Magnetization, 922, 934 Object distance, 1043
Magnetization curve, 926, 929 Oersted, H. C., 647, 814, 844, 853
Magnetizing current, 961 Ohm, 774
Magnification, lateral, for lenses, 1061 Ohm's law, 778, 782
for mirrors, 1048 Oil drop experiment, 685
Malus, E. L., 1151, 1153 Onnes, Kamerlingh, 778
Marsden, E., 700 Optic axis, 1159
Mass-energy conversion, 660 Optical maser, 1027
Mass number, 658 Optical path, 1029, 1059, 1072
Mass spectrometer, 842 Optical reversibility, 1089
Matter, magnetism of, 9 l 5 Order number, 1125
Matter waves, 1200 Ordinary ray, 1156
Matthias, B. T., 778 Oscillations, 943, 957
Maxwell, J. C., 647, 844, 1200 damped, 950
Maxwell's equations, 647 electromagnetic, 943, 947, 963
and radiation pressure, 996 forced, 952
and waves, 982 LC, 943, 947
summary of, 963 mechanical, 947, 957
Maxwell speed, distribution, 1174 Oscilloscope, 806
Mean free path, 781
Meter, defined in wavelengths, 1092 Parallel conductors, forces between, 853
Michelson, A. A., 1000, 1090, 1092 Parallel-plate capacitor, 745
Michelson-Morley experiment, 1092 energy storage in, 761
Microfarad, 743 with dielectric, 748
Micron, 994 Paramagnetism, 920, 937
Microscope, 1049, 1115 Paraxial rays, 1043
Millikan, R. A., 685, 1181, 1183, 1184 Partial derivative, 725, 984
Mirror formula, 1047 Particles and waves, 1200
Mirrors, converging, 1047 Pepinsky, IL, 1139
diverging, 1047 Permeability, 937
Momentum of light, 995 Permeability constant, 846
Moon, radar signals from, 999 Permittivity constant, 746, 848, 986
Morley, E. W., 1092 Phase angle, 948
Morrison, P., 1039 Phase change on reflection, 1084, 1089
Morrison, P. and E., 648 Phase grating, 1128
Motional emf, 876, 888 Phase speed, 978
Multiloop circuits, 798 Phasor, 1080, 1125
Multipoles, expansion in, 722 Photoelectric effect, 1179
Photon, 1182
Neel temperature, 931 Pierce, J. R., 963
Neutron, 655 Planck, M., 1174, 1177, 1179, 1182, 1188,
Neutron diffraction, 1202 1191
Neutron number, 658 Planck's constant, 1178, 1184, 1188,
Newton, 1014, 1023, 1099 1193
Newton's rings, 1088 Planck's radiation formula, 1177
Nichols, 995, 996 Plane mirror, 1013
N onreflecting glass, 1087 spherical waves on, 1038
INDEX 79
Plane of vibration, 1147 Rabi, I. I., 1190
Plane-polarization, 1147 Radiancy, 1173
Plane waves, reflection of, 1013, 1020, Radiation, acoustic, 979
1029 electromagnetic, 980
refraction of, 1013, 1021, 1029 Radiation pressure, 995, 1078
Plimpton, 694 Radioactive decay, 658, 659
Polarization, 720, 1147 Rayleigh's criterion, 1113, 1114
by reflection, 1153 RC circuits, 802
circular, 1163 Red-shift, 1009
electric, 934 Reflected ray, 1014
plane of, 1147 Reflection, 1013
Polarizer, 1150 at plane surface, 1038
Polarizing angle, 1154 at spherical surface, 1035, 1041
Polaroid, 1150 defects in, 1048
Pole, magnetic, 915, 920 diffuse, 1015
Potential, electric, 708 Fermat's principle for, 1029
and electric field strength, 711, 724 Huygens' principle for, 1020
and induction, 883, 886 of plane wave, 1013
for a dipole, 719 of spherical wave, 1035, 1038
for a quadrupole, 721 phase change on, 1084, 1089
for charged disk, 717 polarization by, 1153
for point charges, 715, 716 sign conventions for, 1044
inside conductor, 730 specular, 1015
Potential, gravitational, 711 total internal, 1024
Potential difference, 708 Reflection grating, 1128
and induction, 883 Refracted ray, 1014
in circuits, 795 Refraction, 1013
measurement of, 800 at plane surface, 1014
path independence of, 710 at spherical surface, 1049
Potential energy, electric, 722 double, 1155
in electric circuits, 784 Fermat's principle for, 1029
of electric dipole, 679 Huygens' principle for, 1021
of magnetic dipole, 826 of plane wave, 1013
Potential gradient, 725 of spherical wave, 1035
Potentiometer, 801 sign conventions for, 1051
Poynting, J. H., 987 Relativistic mass, 834
Poynting vector, 986, 995 Relativity, 1004
Precession, 931 and cyclotron, 833
Prism, 1018, 1027 and Michelson-Morley experiment,
constant deviation, 1032 1092
Proton, 655 Resistance, 774
Ptolemy, 1032 internal, 793
Purcell, E. M., 931 measurement of, 812
Resistivity, 774, 775
Q, 968 atomic view of, 780
Quadrupole, electric, 683, 721 coefficient of, 775, 777
Quantization, in Bohr model, 1193 temperature variation of, 777
of charge, 654 Resistors, in parallel, 800
of energy, 1178 in series, 794
of momentum, 1207 Resolving power, of a grating, 1132
of wavelength, 1204, 1207 of a lens, 1112
Quantum number, 1178, 1193 Resonance, 952
Quantum physics, 1173 in cyclotron, 831
Quarter-wave plate, 1163 nuclear, 931
Quartz, index of, 1015 Right-hand rule, 846
80 INDEX

Ripple tank, coherent waves in, 1075 Thales of Miletus, 647


diffraction in, 1035 Thermistor, 779
interference in, 1070 Thin lens, 1055
Roemer, 0., 999 Thomas, H. A., 821
Rossi, B., 997 Thomson, G. P., 1201
Rowland ring, 926, 935 Thomson, J. J., 835
Rush, J. H., 998 Thomson atom model, 697
Rutherford, E., 700 Time constant, capacitative, 902
Rutherford atom, 697, 700 inductive, 904
Time-of-flight spectrometer, 842
Sandage, A. R., 1009 Toroid, 859
Scattering, double, 1169 inductance of, 901
of light, 1167 Torque, on current loop, 823
Schawlow, A. L., 1027 on electric dipole, 825
Schrodinger, E., 1204 on magnetic dipole, 826
Self-field, 676 Torsion balance, 651, 749, 996
Self-induction, 899 Total internal reflection, 1024
Shielding, electromagnetic, 892 Transmission line, 971
Sign conventions, for mirrors, 1045 acoustic, 976
for spherical surfaces, 1051 Transverse wave, 1147
for thin lenses, 1057
Simple harmonic motion, and LC oscilla- Ulrey, 1190
tions, 947 Uncertainty principle, 1210
Single slit, diffraction at, 1102, 1105, Uniaxial crystal, 1162
1108, 1109 Unit cell, 1137
Slepian, Joseph, 668
Snell, W., 1015 Vacuum tube, resistance of, 779
Solar compass, 1169 Vacuum-tube voltmeter, 801
Solenoid, 856, 919 Van de Graaff, R. J., 733
Spectra, 1123, 1130 Vector field, 663
Spectral radiancy, 1173 Velocities, addition of, 1004
Spectrograph, 1130 Venus, radar signals from, 999
Spectrum, 1015, 1130 Vertex, 1043
absorption, 1190 Virtual image, 1038
electromagnetic, 993 Volt, 708
line, 1188 Voltmeter, 801
Speed of light, 998 von Frisch, K., 1169
and electromagnetism, 959, 986 von Laue, M., 1135
Spin, 916
Standard cell, 802 Water, dipole moment of, 719
Standard meter, 1092 Watt, 785
Standing waves, 1203 Wave disturbances, addition of, 1080
Stationary state, 1178 Wave function, 1204
Stefan-Boltzmann, constant, 1175 meaning of, 1208
Stilwell, G. R., 1008 Waveguide, 977
Stokes, G. G., 1089 cutoff frequency for, 977
Stopping potential, 1180 optical, 1026
Strong, J., 1152 Wave mechanics, 1204
Superconductivity, 778 Wave number, 981
Surface charge density, 731 Wave optics, 1035
Surface integral, 688 Waves, and Maxwell's equations, 982
Synchrotron, 834 electromagnetic, 971
Waves and particles, 1200
Telescope, 1114 Weber, 817
Television frequencies, 973 Wheatstone bridge, 810
INDEX 81
Wien, W., 890 Young, T., 1068, 1073, 1148
Wien's formula, 1177 Young's experiment, 1027, 1069, 1104,
Williams, H. J., 930 1115
Wilson, M. H., 994

X-rays, as transverse waves, 1169 Zacharias, J. K., 1190


Compton effect for, 1184 Zeeman, P., 835
diffraction, 1031, 1134 Zeeman effect, 842
SELECTED NUMERICAL CONSTANTS

1r = 3.14 1r 2 = D.87 e = 2.72 e- 1 = 1/e = 0.368

ln 2 = 0.6!-J3 log e = 0.434 V2 = 1.41 V3 = 1.13

sin 30° = cos 60° = i = 0.500

cos 30° = sin 60° = V3/2 = 0.866

tan 30° = cot 60° = V3/3 = 0.577

tan 60° = cot 30° = V3 = 1.732

sin 45° = cos 45° = VZ/2 = 0.707

tan 45° = cot 45° = 1.00


SELECTED CONVERSION FACTORS
(See Appendix H for a more complete list)

180° = 1r rad
1 radian = 57.3° = 0.159 rev
1 slug = 32.2 lb (mass) = 14.6 kg
l kilogram = 2.21 lb (mass)
l pound (mass) = 0.454 kg
1 atomic mass unit = 1.66 X 10- 27 kg
l meter = 3\J.4 in. = 3.28 ft; l inch = 2.54 cm
1 mile = 5280 ft = 1.61 km
1 angstrom unit = 10- 10 meter = 0.1 mµ
1 millimicron = 10- 9 meter
1 liter = 61.0 in. 3
1 ft 3 = 28.3 Ii
l day = 86,400 sec
1 year = 3.16 X 107 sec = 365 days
l mile/hr = 1.47 ft/sec = 0.447 meter/sec
1 pound = 4.45 nt; 1 newton = 0.225 lb
1 atmosphere = 29.9 in.-Ilg = 76.0 cm-Hg = 1.01 X 10 5 nt/meter 2
1 Btu = 778 ft-lb = 252 cal = 1060 joules
1 calorie = 4. rn joules; 1 joule = 0.2:39 cal = 2. 78 X 10-7 kw-hr
1 electron volt = 1.60 X 10- 19 joule
l horsepower = 550 ft-lb/sec = 746 watts
1 weber /meter 2 = 1 tesla = 10 4 gauss

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy